├── .gitignore ├── Pictures_for_Github_only ├── Online Result.png └── mainframework.png ├── README.md ├── cldetection_utils.py ├── configs └── CLdetection2023 │ └── srpose_s2.py ├── inference_single_image.py ├── install_env.sh ├── mmpose_package └── mmpose │ ├── .gitignore │ ├── .owners.yml │ ├── .pre-commit-config.yaml │ ├── .pylintrc │ ├── .readthedocs.yml │ ├── CITATION.cff │ ├── LICENSE │ ├── MANIFEST.in │ ├── README.md │ ├── README_CN.md │ ├── configs │ ├── _base_ │ │ ├── datasets │ │ │ ├── 300w.py │ │ │ ├── aflw.py │ │ │ ├── aic.py │ │ │ ├── animalpose.py │ │ │ ├── ap10k.py │ │ │ ├── atrw.py │ │ │ ├── campus.py │ │ │ ├── cephalometric.py │ │ │ ├── cephalometric_isbi2015.py │ │ │ ├── coco.py │ │ │ ├── coco_aic.py │ │ │ ├── coco_openpose.py │ │ │ ├── coco_wholebody.py │ │ │ ├── coco_wholebody_face.py │ │ │ ├── coco_wholebody_hand.py │ │ │ ├── cofw.py │ │ │ ├── crowdpose.py │ │ │ ├── deepfashion2.py │ │ │ ├── deepfashion_full.py │ │ │ ├── deepfashion_lower.py │ │ │ ├── deepfashion_upper.py │ │ │ ├── fly.py │ │ │ ├── freihand2d.py │ │ │ ├── h36m.py │ │ │ ├── halpe.py │ │ │ ├── horse10.py │ │ │ ├── interhand2d.py │ │ │ ├── interhand3d.py │ │ │ ├── jhmdb.py │ │ │ ├── lapa.py │ │ │ ├── locust.py │ │ │ ├── macaque.py │ │ │ ├── mhp.py │ │ │ ├── mpi_inf_3dhp.py │ │ │ ├── mpii.py │ │ │ ├── mpii_trb.py │ │ │ ├── ochuman.py │ │ │ ├── onehand10k.py │ │ │ ├── panoptic_body3d.py │ │ │ ├── panoptic_hand2d.py │ │ │ ├── posetrack18.py │ │ │ ├── rhd2d.py │ │ │ ├── shelf.py │ │ │ ├── wflw.py │ │ │ └── zebra.py │ │ └── default_runtime.py │ ├── animal_2d_keypoint │ │ ├── README.md │ │ ├── rtmpose │ │ │ ├── README.md │ │ │ └── ap10k │ │ │ │ ├── rtmpose-m_8xb64-210e_ap10k-256x256.py │ │ │ │ ├── rtmpose_ap10k.md │ │ │ │ └── rtmpose_ap10k.yml │ │ └── topdown_heatmap │ │ │ ├── README.md │ │ │ ├── animalpose │ │ │ ├── hrnet_animalpose.md │ │ │ ├── hrnet_animalpose.yml │ │ │ ├── resnet_animalpose.md │ │ │ ├── resnet_animalpose.yml │ │ │ ├── td-hm_hrnet-w32_8xb64-210e_animalpose-256x256.py │ │ │ ├── td-hm_hrnet-w48_8xb64-210e_animalpose-256x256.py │ │ │ ├── td-hm_res101_8xb64-210e_animalpose-256x256.py │ │ │ ├── td-hm_res152_8xb32-210e_animalpose-256x256.py │ │ │ └── td-hm_res50_8xb64-210e_animalpose-256x256.py │ │ │ ├── ap10k │ │ │ ├── cspnext-m_udp_8xb64-210e_ap10k-256x256.py │ │ │ ├── cspnext_udp_ap10k.md │ │ │ ├── cspnext_udp_ap10k.yml │ │ │ ├── hrnet_ap10k.md │ │ │ ├── hrnet_ap10k.yml │ │ │ ├── resnet_ap10k.md │ │ │ ├── resnet_ap10k.yml │ │ │ ├── td-hm_hrnet-w32_8xb64-210e_ap10k-256x256.py │ │ │ ├── td-hm_hrnet-w48_8xb64-210e_ap10k-256x256.py │ │ │ ├── td-hm_res101_8xb64-210e_ap10k-256x256.py │ │ │ └── td-hm_res50_8xb64-210e_ap10k-256x256.py │ │ │ ├── locust │ │ │ ├── resnet_locust.md │ │ │ ├── resnet_locust.yml │ │ │ ├── td-hm_res101_8xb64-210e_locust-160x160.py │ │ │ ├── td-hm_res152_8xb32-210e_locust-160x160.py │ │ │ └── td-hm_res50_8xb64-210e_locust-160x160.py │ │ │ └── zebra │ │ │ ├── resnet_zebra.md │ │ │ ├── resnet_zebra.yml │ │ │ ├── td-hm_res101_8xb64-210e_zebra-160x160.py │ │ │ ├── td-hm_res152_8xb32-210e_zebra-160x160.py │ │ │ └── td-hm_res50_8xb64-210e_zebra-160x160.py │ ├── body_2d_keypoint │ │ ├── README.md │ │ ├── associative_embedding │ │ │ ├── README.md │ │ │ └── coco │ │ │ │ └── ae_hrnet-w32_8xb24-300e_coco-512x512.py │ │ ├── cid │ │ │ └── coco │ │ │ │ ├── cid_hrnet-w32_8xb20-140e_coco-512x512.py │ │ │ │ ├── cid_hrnet-w48_8xb20-140e_coco-512x512.py │ │ │ │ ├── hrnet_coco.md │ │ │ │ └── hrnet_coco.yml │ │ ├── dekr │ │ │ ├── README.md │ │ │ ├── coco │ │ │ │ ├── dekr_hrnet-w32_8xb10-140e_coco-512x512.py │ │ │ │ ├── dekr_hrnet-w48_8xb10-140e_coco-640x640.py │ │ │ │ ├── hrnet_coco.md │ │ │ │ └── hrnet_coco.yml │ │ │ └── crowdpose │ │ │ │ ├── dekr_hrnet-w32_8xb10-300e_crowdpose-512x512.py │ │ │ │ ├── dekr_hrnet-w48_8xb5-300e_crowdpose-640x640.py │ │ │ │ ├── hrnet_crowdpose.md │ │ │ │ └── hrnet_crowdpose.yml │ │ ├── integral_regression │ │ │ ├── README.md │ │ │ └── coco │ │ │ │ ├── ipr_res50_8xb64-210e_coco-256x256.py │ │ │ │ ├── ipr_res50_debias-8xb64-210e_coco-256x256.py │ │ │ │ ├── ipr_res50_dsnt-8xb64-210e_coco-256x256.py │ │ │ │ ├── resnet_debias_coco.md │ │ │ │ ├── resnet_debias_coco.yml │ │ │ │ ├── resnet_dsnt_coco.md │ │ │ │ ├── resnet_dsnt_coco.yml │ │ │ │ ├── resnet_ipr_coco.md │ │ │ │ └── resnet_ipr_coco.yml │ │ ├── rtmpose │ │ │ ├── README.md │ │ │ ├── body8 │ │ │ │ ├── rtmpose-l_8xb256-210e_body8-256x192.py │ │ │ │ ├── rtmpose-l_8xb256-210e_body8-384x288.py │ │ │ │ ├── rtmpose-m_8xb256-210e_body8-256x192.py │ │ │ │ ├── rtmpose-m_8xb256-210e_body8-384x288.py │ │ │ │ ├── rtmpose-s_8xb256-210e_body8-256x192.py │ │ │ │ ├── rtmpose-t_8xb256-210e_body8-256x192.py │ │ │ │ ├── rtmpose_body8.md │ │ │ │ └── rtmpose_body8.yml │ │ │ ├── coco │ │ │ │ ├── rtmpose-l_8xb256-420e_aic-coco-256x192.py │ │ │ │ ├── rtmpose-l_8xb256-420e_aic-coco-384x288.py │ │ │ │ ├── rtmpose-l_8xb256-420e_coco-256x192.py │ │ │ │ ├── rtmpose-m_8xb256-420e_aic-coco-256x192.py │ │ │ │ ├── rtmpose-m_8xb256-420e_aic-coco-384x288.py │ │ │ │ ├── rtmpose-m_8xb256-420e_coco-256x192.py │ │ │ │ ├── rtmpose-s_8xb256-420e_aic-coco-256x192.py │ │ │ │ ├── rtmpose-s_8xb256-420e_coco-256x192.py │ │ │ │ ├── rtmpose-t_8xb256-420e_aic-coco-256x192.py │ │ │ │ ├── rtmpose-t_8xb256-420e_coco-256x192.py │ │ │ │ ├── rtmpose_coco.md │ │ │ │ └── rtmpose_coco.yml │ │ │ ├── crowdpose │ │ │ │ ├── rtmpose-m_8xb64-210e_crowdpose-256x192.py │ │ │ │ ├── rtmpose_crowdpose.md │ │ │ │ └── rtmpose_crowdpose.yml │ │ │ └── mpii │ │ │ │ ├── rtmpose-m_8xb64-210e_mpii-256x256.py │ │ │ │ ├── rtmpose_mpii.md │ │ │ │ └── rtmpose_mpii.yml │ │ ├── simcc │ │ │ ├── README.md │ │ │ ├── coco │ │ │ │ ├── mobilenetv2_coco.md │ │ │ │ ├── mobilenetv2_coco.yml │ │ │ │ ├── resnet_coco.md │ │ │ │ ├── resnet_coco.yml │ │ │ │ ├── simcc_mobilenetv2_wo-deconv-8xb64-210e_coco-256x192.py │ │ │ │ ├── simcc_res50_8xb32-140e_coco-384x288.py │ │ │ │ ├── simcc_res50_8xb64-210e_coco-256x192.py │ │ │ │ ├── simcc_vipnas-mbv3_8xb64-210e_coco-256x192.py │ │ │ │ ├── vipnas_coco.md │ │ │ │ └── vipnas_coco.yml │ │ │ └── mpii │ │ │ │ └── simcc_res50_wo-deconv-8xb64-210e_mpii-256x256.py │ │ ├── topdown_heatmap │ │ │ ├── README.md │ │ │ ├── aic │ │ │ │ ├── hrnet_aic.md │ │ │ │ ├── hrnet_aic.yml │ │ │ │ ├── resnet_aic.md │ │ │ │ ├── resnet_aic.yml │ │ │ │ ├── td-hm_hrnet-w32_8xb64-210e_aic-256x192.py │ │ │ │ └── td-hm_res101_8xb64-210e_aic-256x192.py │ │ │ ├── coco │ │ │ │ ├── alexnet_coco.md │ │ │ │ ├── alexnet_coco.yml │ │ │ │ ├── cpm_coco.md │ │ │ │ ├── cpm_coco.yml │ │ │ │ ├── cspnext-l_udp_8xb256-210e_aic-coco-256x192.py │ │ │ │ ├── cspnext-l_udp_8xb256-210e_coco-256x192.py │ │ │ │ ├── cspnext-m_udp_8xb256-210e_aic-coco-256x192.py │ │ │ │ ├── cspnext-m_udp_8xb256-210e_coco-256x192.py │ │ │ │ ├── cspnext-s_udp_8xb256-210e_aic-coco-256x192.py │ │ │ │ ├── cspnext-s_udp_8xb256-210e_coco-256x192.py │ │ │ │ ├── cspnext-tiny_udp_8xb256-210e_aic-coco-256x192.py │ │ │ │ ├── cspnext-tiny_udp_8xb256-210e_coco-256x192.py │ │ │ │ ├── cspnext_udp_coco.md │ │ │ │ ├── cspnext_udp_coco.yml │ │ │ │ ├── hourglass_coco.md │ │ │ │ ├── hourglass_coco.yml │ │ │ │ ├── hrformer_coco.md │ │ │ │ ├── hrformer_coco.yml │ │ │ │ ├── hrnet_augmentation_coco.md │ │ │ │ ├── hrnet_augmentation_coco.yml │ │ │ │ ├── hrnet_coco.md │ │ │ │ ├── hrnet_coco.yml │ │ │ │ ├── hrnet_coco_aic.md │ │ │ │ ├── hrnet_dark_coco.md │ │ │ │ ├── hrnet_dark_coco.yml │ │ │ │ ├── hrnet_fp16_coco.md │ │ │ │ ├── hrnet_udp_coco.md │ │ │ │ ├── hrnet_udp_coco.yml │ │ │ │ ├── litehrnet_coco.md │ │ │ │ ├── litehrnet_coco.yml │ │ │ │ ├── mobilenetv2_coco.md │ │ │ │ ├── mobilenetv2_coco.yml │ │ │ │ ├── mspn_coco.md │ │ │ │ ├── mspn_coco.yml │ │ │ │ ├── pvt_coco.md │ │ │ │ ├── pvt_coco.yml │ │ │ │ ├── resnest_coco.md │ │ │ │ ├── resnest_coco.yml │ │ │ │ ├── resnet_coco.md │ │ │ │ ├── resnet_coco.yml │ │ │ │ ├── resnet_dark_coco.md │ │ │ │ ├── resnet_dark_coco.yml │ │ │ │ ├── resnet_fp16_coco.md │ │ │ │ ├── resnetv1d_coco.md │ │ │ │ ├── resnetv1d_coco.yml │ │ │ │ ├── resnext_coco.md │ │ │ │ ├── resnext_coco.yml │ │ │ │ ├── rsn_coco.md │ │ │ │ ├── rsn_coco.yml │ │ │ │ ├── scnet_coco.md │ │ │ │ ├── scnet_coco.yml │ │ │ │ ├── seresnet_coco.md │ │ │ │ ├── seresnet_coco.yml │ │ │ │ ├── shufflenetv1_coco.md │ │ │ │ ├── shufflenetv1_coco.yml │ │ │ │ ├── shufflenetv2_coco.md │ │ │ │ ├── shufflenetv2_coco.yml │ │ │ │ ├── swin_coco.md │ │ │ │ ├── swin_coco.yml │ │ │ │ ├── td-hm_2xmspn50_8xb32-210e_coco-256x192.py │ │ │ │ ├── td-hm_2xrsn50_8xb32-210e_coco-256x192.py │ │ │ │ ├── td-hm_3xmspn50_8xb32-210e_coco-256x192.py │ │ │ │ ├── td-hm_3xrsn50_8xb32-210e_coco-256x192.py │ │ │ │ ├── td-hm_4xmspn50_8xb32-210e_coco-256x192.py │ │ │ │ ├── td-hm_ViTPose-base-simple_8xb64-210e_coco-256x192.py │ │ │ │ ├── td-hm_ViTPose-base_8xb64-210e_coco-256x192.py │ │ │ │ ├── td-hm_ViTPose-huge-simple_8xb64-210e_coco-256x192.py │ │ │ │ ├── td-hm_ViTPose-huge_8xb64-210e_coco-256x192.py │ │ │ │ ├── td-hm_ViTPose-large-simple_8xb64-210e_coco-256x192.py │ │ │ │ ├── td-hm_ViTPose-large_8xb64-210e_coco-256x192.py │ │ │ │ ├── td-hm_ViTPose-small-simple_8xb64-210e_coco-256x192.py │ │ │ │ ├── td-hm_ViTPose-small_8xb64-210e_coco-256x192.py │ │ │ │ ├── td-hm_alexnet_8xb64-210e_coco-256x192.py │ │ │ │ ├── td-hm_cpm_8xb32-210e_coco-384x288.py │ │ │ │ ├── td-hm_cpm_8xb64-210e_coco-256x192.py │ │ │ │ ├── td-hm_hourglass52_8xb32-210e_coco-256x256.py │ │ │ │ ├── td-hm_hourglass52_8xb32-210e_coco-384x384.py │ │ │ │ ├── td-hm_hrformer-base_8xb32-210e_coco-256x192.py │ │ │ │ ├── td-hm_hrformer-base_8xb32-210e_coco-384x288.py │ │ │ │ ├── td-hm_hrformer-small_8xb32-210e_coco-256x192.py │ │ │ │ ├── td-hm_hrformer-small_8xb32-210e_coco-384x288.py │ │ │ │ ├── td-hm_hrnet-w32_8xb64-210e_coco-256x192.py │ │ │ │ ├── td-hm_hrnet-w32_8xb64-210e_coco-384x288.py │ │ │ │ ├── td-hm_hrnet-w32_8xb64-210e_coco-aic-256x192-combine.py │ │ │ │ ├── td-hm_hrnet-w32_8xb64-210e_coco-aic-256x192-merge.py │ │ │ │ ├── td-hm_hrnet-w32_coarsedropout-8xb64-210e_coco-256x192.py │ │ │ │ ├── td-hm_hrnet-w32_dark-8xb64-210e_coco-256x192.py │ │ │ │ ├── td-hm_hrnet-w32_dark-8xb64-210e_coco-384x288.py │ │ │ │ ├── td-hm_hrnet-w32_fp16-8xb64-210e_coco-256x192.py │ │ │ │ ├── td-hm_hrnet-w32_gridmask-8xb64-210e_coco-256x192.py │ │ │ │ ├── td-hm_hrnet-w32_photometric-8xb64-210e_coco-256x192.py │ │ │ │ ├── td-hm_hrnet-w32_udp-8xb64-210e_coco-256x192.py │ │ │ │ ├── td-hm_hrnet-w32_udp-8xb64-210e_coco-384x288.py │ │ │ │ ├── td-hm_hrnet-w32_udp-regress-8xb64-210e_coco-256x192.py │ │ │ │ ├── td-hm_hrnet-w48_8xb32-210e_coco-256x192.py │ │ │ │ ├── td-hm_hrnet-w48_8xb32-210e_coco-384x288.py │ │ │ │ ├── td-hm_hrnet-w48_dark-8xb32-210e_coco-256x192.py │ │ │ │ ├── td-hm_hrnet-w48_dark-8xb32-210e_coco-384x288.py │ │ │ │ ├── td-hm_hrnet-w48_udp-8xb32-210e_coco-256x192.py │ │ │ │ ├── td-hm_hrnet-w48_udp-8xb32-210e_coco-384x288.py │ │ │ │ ├── td-hm_litehrnet-18_8xb32-210e_coco-384x288.py │ │ │ │ ├── td-hm_litehrnet-18_8xb64-210e_coco-256x192.py │ │ │ │ ├── td-hm_litehrnet-30_8xb32-210e_coco-384x288.py │ │ │ │ ├── td-hm_litehrnet-30_8xb64-210e_coco-256x192.py │ │ │ │ ├── td-hm_mobilenetv2_8xb64-210e_coco-256x192.py │ │ │ │ ├── td-hm_mobilenetv2_8xb64-210e_coco-384x288.py │ │ │ │ ├── td-hm_mspn50_8xb32-210e_coco-256x192.py │ │ │ │ ├── td-hm_pvt-s_8xb64-210e_coco-256x192.py │ │ │ │ ├── td-hm_pvtv2-b2_8xb64-210e_coco-256x192.py │ │ │ │ ├── td-hm_res101_8xb32-210e_coco-384x288.py │ │ │ │ ├── td-hm_res101_8xb64-210e_coco-256x192.py │ │ │ │ ├── td-hm_res101_dark-8xb64-210e_coco-256x192.py │ │ │ │ ├── td-hm_res101_dark-8xb64-210e_coco-384x288.py │ │ │ │ ├── td-hm_res152_8xb32-210e_coco-256x192.py │ │ │ │ ├── td-hm_res152_8xb32-210e_coco-384x288.py │ │ │ │ ├── td-hm_res152_dark-8xb32-210e_coco-256x192.py │ │ │ │ ├── td-hm_res152_dark-8xb32-210e_coco-384x288.py │ │ │ │ ├── td-hm_res50_8xb64-210e_coco-256x192.py │ │ │ │ ├── td-hm_res50_8xb64-210e_coco-384x288.py │ │ │ │ ├── td-hm_res50_dark-8xb64-210e_coco-256x192.py │ │ │ │ ├── td-hm_res50_dark-8xb64-210e_coco-384x288.py │ │ │ │ ├── td-hm_res50_fp16-8xb64-210e_coco-256x192.py │ │ │ │ ├── td-hm_resnest101_8xb32-210e_coco-384x288.py │ │ │ │ ├── td-hm_resnest101_8xb64-210e_coco-256x192.py │ │ │ │ ├── td-hm_resnest200_8xb16-210e_coco-384x288.py │ │ │ │ ├── td-hm_resnest200_8xb64-210e_coco-256x192.py │ │ │ │ ├── td-hm_resnest269_8xb16-210e_coco-384x288.py │ │ │ │ ├── td-hm_resnest269_8xb32-210e_coco-256x192.py │ │ │ │ ├── td-hm_resnest50_8xb64-210e_coco-256x192.py │ │ │ │ ├── td-hm_resnest50_8xb64-210e_coco-384x288.py │ │ │ │ ├── td-hm_resnetv1d101_8xb32-210e_coco-384x288.py │ │ │ │ ├── td-hm_resnetv1d101_8xb64-210e_coco-256x192.py │ │ │ │ ├── td-hm_resnetv1d152_8xb32-210e_coco-256x192.py │ │ │ │ ├── td-hm_resnetv1d152_8xb48-210e_coco-384x288.py │ │ │ │ ├── td-hm_resnetv1d50_8xb64-210e_coco-256x192.py │ │ │ │ ├── td-hm_resnetv1d50_8xb64-210e_coco-384x288.py │ │ │ │ ├── td-hm_resnext101_8xb32-210e_coco-384x288.py │ │ │ │ ├── td-hm_resnext101_8xb64-210e_coco-256x192.py │ │ │ │ ├── td-hm_resnext152_8xb32-210e_coco-256x192.py │ │ │ │ ├── td-hm_resnext152_8xb48-210e_coco-384x288.py │ │ │ │ ├── td-hm_resnext50_8xb64-210e_coco-256x192.py │ │ │ │ ├── td-hm_resnext50_8xb64-210e_coco-384x288.py │ │ │ │ ├── td-hm_rsn18_8xb32-210e_coco-256x192.py │ │ │ │ ├── td-hm_rsn50_8xb32-210e_coco-256x192.py │ │ │ │ ├── td-hm_scnet101_8xb32-210e_coco-256x192.py │ │ │ │ ├── td-hm_scnet101_8xb48-210e_coco-384x288.py │ │ │ │ ├── td-hm_scnet50_8xb32-210e_coco-384x288.py │ │ │ │ ├── td-hm_scnet50_8xb64-210e_coco-256x192.py │ │ │ │ ├── td-hm_seresnet101_8xb32-210e_coco-384x288.py │ │ │ │ ├── td-hm_seresnet101_8xb64-210e_coco-256x192.py │ │ │ │ ├── td-hm_seresnet152_8xb32-210e_coco-256x192.py │ │ │ │ ├── td-hm_seresnet152_8xb48-210e_coco-384x288.py │ │ │ │ ├── td-hm_seresnet50_8xb64-210e_coco-256x192.py │ │ │ │ ├── td-hm_seresnet50_8xb64-210e_coco-384x288.py │ │ │ │ ├── td-hm_shufflenetv1_8xb64-210e_coco-256x192.py │ │ │ │ ├── td-hm_shufflenetv1_8xb64-210e_coco-384x288.py │ │ │ │ ├── td-hm_shufflenetv2_8xb64-210e_coco-256x192.py │ │ │ │ ├── td-hm_shufflenetv2_8xb64-210e_coco-384x288.py │ │ │ │ ├── td-hm_swin-b-p4-w7_8xb32-210e_coco-256x192.py │ │ │ │ ├── td-hm_swin-b-p4-w7_8xb32-210e_coco-384x288.py │ │ │ │ ├── td-hm_swin-l-p4-w7_8xb32-210e_coco-256x192.py │ │ │ │ ├── td-hm_swin-l-p4-w7_8xb32-210e_coco-384x288.py │ │ │ │ ├── td-hm_swin-t-p4-w7_8xb32-210e_coco-256x192.py │ │ │ │ ├── td-hm_vgg16-bn_8xb64-210e_coco-256x192.py │ │ │ │ ├── td-hm_vipnas-mbv3_8xb64-210e_coco-256x192.py │ │ │ │ ├── td-hm_vipnas-res50_8xb64-210e_coco-256x192.py │ │ │ │ ├── vgg_coco.md │ │ │ │ ├── vgg_coco.yml │ │ │ │ ├── vipnas_coco.md │ │ │ │ ├── vipnas_coco.yml │ │ │ │ ├── vitpose_coco.md │ │ │ │ └── vitpose_coco.yml │ │ │ ├── crowdpose │ │ │ │ ├── cspnext-m_udp_8xb64-210e_crowpose-256x192.py │ │ │ │ ├── cspnext_udp_crowdpose.md │ │ │ │ ├── cspnext_udp_crowdpose.yml │ │ │ │ ├── hrnet_crowdpose.md │ │ │ │ ├── hrnet_crowdpose.yml │ │ │ │ ├── resnet_crowdpose.md │ │ │ │ ├── resnet_crowdpose.yml │ │ │ │ ├── td-hm_hrnet-w32_8xb64-210e_crowdpose-256x192.py │ │ │ │ ├── td-hm_res101_8xb64-210e_crowdpose-256x192.py │ │ │ │ ├── td-hm_res101_8xb64-210e_crowdpose-320x256.py │ │ │ │ ├── td-hm_res152_8xb64-210e_crowdpose-256x192.py │ │ │ │ └── td-hm_res50_8xb64-210e_crowdpose-256x192.py │ │ │ ├── jhmdb │ │ │ │ ├── cpm_jhmdb.md │ │ │ │ ├── cpm_jhmdb.yml │ │ │ │ ├── resnet_jhmdb.md │ │ │ │ ├── resnet_jhmdb.yml │ │ │ │ ├── td-hm_cpm_8xb32-40e_jhmdb-sub1-368x368.py │ │ │ │ ├── td-hm_cpm_8xb32-40e_jhmdb-sub2-368x368.py │ │ │ │ ├── td-hm_cpm_8xb32-40e_jhmdb-sub3-368x368.py │ │ │ │ ├── td-hm_res50-2deconv_8xb64-40e_jhmdb-sub1-256x256.py │ │ │ │ ├── td-hm_res50-2deconv_8xb64-40e_jhmdb-sub2-256x256.py │ │ │ │ ├── td-hm_res50-2deconv_8xb64-40e_jhmdb-sub3-256x256.py │ │ │ │ ├── td-hm_res50_8xb64-20e_jhmdb-sub1-256x256.py │ │ │ │ ├── td-hm_res50_8xb64-20e_jhmdb-sub2-256x256.py │ │ │ │ └── td-hm_res50_8xb64-20e_jhmdb-sub3-256x256.py │ │ │ ├── mpii │ │ │ │ ├── cpm_mpii.md │ │ │ │ ├── cpm_mpii.yml │ │ │ │ ├── cspnext-m_udp_8xb64-210e_mpii-256x256.py │ │ │ │ ├── cspnext_udp_mpii.md │ │ │ │ ├── cspnext_udp_mpii.yml │ │ │ │ ├── hourglass_mpii.md │ │ │ │ ├── hourglass_mpii.yml │ │ │ │ ├── hrnet_dark_mpii.md │ │ │ │ ├── hrnet_dark_mpii.yml │ │ │ │ ├── hrnet_mpii.md │ │ │ │ ├── hrnet_mpii.yml │ │ │ │ ├── litehrnet_mpii.md │ │ │ │ ├── litehrnet_mpii.yml │ │ │ │ ├── mobilenetv2_mpii.md │ │ │ │ ├── mobilenetv2_mpii.yml │ │ │ │ ├── resnet_mpii.md │ │ │ │ ├── resnet_mpii.yml │ │ │ │ ├── resnetv1d_mpii.md │ │ │ │ ├── resnetv1d_mpii.yml │ │ │ │ ├── resnext_mpii.md │ │ │ │ ├── resnext_mpii.yml │ │ │ │ ├── scnet_mpii.md │ │ │ │ ├── scnet_mpii.yml │ │ │ │ ├── seresnet_mpii.md │ │ │ │ ├── seresnet_mpii.yml │ │ │ │ ├── shufflenetv1_mpii.md │ │ │ │ ├── shufflenetv1_mpii.yml │ │ │ │ ├── shufflenetv2_mpii.md │ │ │ │ ├── shufflenetv2_mpii.yml │ │ │ │ ├── td-hm_cpm_8xb64-210e_mpii-368x368.py │ │ │ │ ├── td-hm_hourglass52_8xb32-210e_mpii-384x384.py │ │ │ │ ├── td-hm_hourglass52_8xb64-210e_mpii-256x256.py │ │ │ │ ├── td-hm_hrnet-w32_8xb64-210e_mpii-256x256.py │ │ │ │ ├── td-hm_hrnet-w32_dark-8xb64-210e_mpii-256x256.py │ │ │ │ ├── td-hm_hrnet-w48_8xb64-210e_mpii-256x256.py │ │ │ │ ├── td-hm_hrnet-w48_dark-8xb64-210e_mpii-256x256.py │ │ │ │ ├── td-hm_litehrnet-18_8xb64-210e_mpii-256x256.py │ │ │ │ ├── td-hm_litehrnet-30_8xb64-210e_mpii-256x256.py │ │ │ │ ├── td-hm_mobilenetv2_8xb64-210e_mpii-256x256.py │ │ │ │ ├── td-hm_res101_8xb64-210e_mpii-256x256.py │ │ │ │ ├── td-hm_res152_8xb32-210e_mpii-256x256.py │ │ │ │ ├── td-hm_res50_8xb64-210e_mpii-256x256.py │ │ │ │ ├── td-hm_resnetv1d101_8xb64-210e_mpii-256x256.py │ │ │ │ ├── td-hm_resnetv1d152_8xb64-210e_mpii-256x256.py │ │ │ │ ├── td-hm_resnetv1d50_8xb64-210e_mpii-256x256.py │ │ │ │ ├── td-hm_resnext152_8xb32-210e_mpii-256x256.py │ │ │ │ ├── td-hm_scnet101_8xb64-210e_mpii-256x256.py │ │ │ │ ├── td-hm_scnet50_8xb64-210e_mpii-256x256.py │ │ │ │ ├── td-hm_seresnet101_8xb64-210e_mpii-256x256.py │ │ │ │ ├── td-hm_seresnet152_8xb32-210e_mpii-256x256.py │ │ │ │ ├── td-hm_seresnet50_8xb64-210e_mpii-256x256.py │ │ │ │ ├── td-hm_shufflenetv1_8xb64-210e_mpii-256x256.py │ │ │ │ └── td-hm_shufflenetv2_8xb64-210e_mpii-256x256.py │ │ │ └── posetrack18 │ │ │ │ ├── hrnet_posetrack18.md │ │ │ │ ├── hrnet_posetrack18.yml │ │ │ │ ├── resnet_posetrack18.md │ │ │ │ ├── resnet_posetrack18.yml │ │ │ │ ├── td-hm_hrnet-w32_8xb64-20e_posetrack18-256x192.py │ │ │ │ ├── td-hm_hrnet-w32_8xb64-20e_posetrack18-384x288.py │ │ │ │ ├── td-hm_hrnet-w48_8xb64-20e_posetrack18-256x192.py │ │ │ │ ├── td-hm_hrnet-w48_8xb64-20e_posetrack18-384x288.py │ │ │ │ └── td-hm_res50_8xb64-20e_posetrack18-256x192.py │ │ └── topdown_regression │ │ │ ├── README.md │ │ │ ├── coco │ │ │ ├── mobilenetv2_rle_coco.md │ │ │ ├── mobilenetv2_rle_coco.yml │ │ │ ├── resnet_coco.md │ │ │ ├── resnet_coco.yml │ │ │ ├── resnet_rle_coco.md │ │ │ ├── resnet_rle_coco.yml │ │ │ ├── td-reg_mobilenetv2_rle-pretrained-8xb64-210e_coco-256x192.py │ │ │ ├── td-reg_res101_8xb64-210e_coco-256x192.py │ │ │ ├── td-reg_res101_rle-8xb64-210e_coco-256x192.py │ │ │ ├── td-reg_res152_8xb64-210e_coco-256x192.py │ │ │ ├── td-reg_res152_rle-8xb64-210e_coco-256x192.py │ │ │ ├── td-reg_res152_rle-8xb64-210e_coco-384x288.py │ │ │ ├── td-reg_res50_8xb64-210e_coco-256x192.py │ │ │ ├── td-reg_res50_rle-8xb64-210e_coco-256x192.py │ │ │ └── td-reg_res50_rle-pretrained-8xb64-210e_coco-256x192.py │ │ │ └── mpii │ │ │ ├── resnet_mpii.md │ │ │ ├── resnet_mpii.yml │ │ │ ├── resnet_rle_mpii.md │ │ │ ├── resnet_rle_mpii.yml │ │ │ ├── td-reg_res101_8xb64-210e_mpii-256x256.py │ │ │ ├── td-reg_res152_8xb64-210e_mpii-256x256.py │ │ │ ├── td-reg_res50_8xb64-210e_mpii-256x256.py │ │ │ └── td-reg_res50_rle-8xb64-210e_mpii-256x256.py │ ├── body_3d_keypoint │ │ └── README.md │ ├── face_2d_keypoint │ │ ├── README.md │ │ ├── rtmpose │ │ │ ├── README.md │ │ │ ├── coco_wholebody_face │ │ │ │ ├── rtmpose-m_8xb32-60e_coco-wholebody-face-256x256.py │ │ │ │ ├── rtmpose_coco_wholebody_face.md │ │ │ │ └── rtmpose_coco_wholebody_face.yml │ │ │ ├── lapa │ │ │ │ ├── rtmpose-m_8xb64-120e_lapa-256x256.py │ │ │ │ ├── rtmpose_lapa.md │ │ │ │ └── rtmpose_lapa.yml │ │ │ └── wflw │ │ │ │ ├── rtmpose-m_8xb64-60e_wflw-256x256.py │ │ │ │ ├── rtmpose_wflw.md │ │ │ │ └── rtmpose_wflw.yml │ │ ├── topdown_heatmap │ │ │ ├── 300w │ │ │ │ ├── hrnetv2_300w.md │ │ │ │ ├── hrnetv2_300w.yml │ │ │ │ └── td-hm_hrnetv2-w18_8xb64-60e_300w-256x256.py │ │ │ ├── README.md │ │ │ ├── aflw │ │ │ │ ├── hrnetv2_aflw.md │ │ │ │ ├── hrnetv2_aflw.yml │ │ │ │ ├── hrnetv2_dark_aflw.md │ │ │ │ ├── hrnetv2_dark_aflw.yml │ │ │ │ ├── td-hm_hrnetv2-w18_8xb64-60e_aflw-256x256.py │ │ │ │ └── td-hm_hrnetv2-w18_dark-8xb64-60e_aflw-256x256.py │ │ │ ├── coco_wholebody_face │ │ │ │ ├── hourglass_coco_wholebody_face.md │ │ │ │ ├── hourglass_coco_wholebody_face.yml │ │ │ │ ├── hrnetv2_coco_wholebody_face.md │ │ │ │ ├── hrnetv2_coco_wholebody_face.yml │ │ │ │ ├── hrnetv2_dark_coco_wholebody_face.md │ │ │ │ ├── hrnetv2_dark_coco_wholebody_face.yml │ │ │ │ ├── mobilenetv2_coco_wholebody_face.md │ │ │ │ ├── mobilenetv2_coco_wholebody_face.yml │ │ │ │ ├── resnet_coco_wholebody_face.md │ │ │ │ ├── resnet_coco_wholebody_face.yml │ │ │ │ ├── scnet_coco_wholebody_face.md │ │ │ │ ├── scnet_coco_wholebody_face.yml │ │ │ │ ├── td-hm_hourglass52_8xb32-60e_coco-wholebody-face-256x256.py │ │ │ │ ├── td-hm_hrnetv2-w18_8xb32-60e_coco-wholebody-face-256x256.py │ │ │ │ ├── td-hm_hrnetv2-w18_dark-8xb32-60e_coco-wholebody-face-256x256.py │ │ │ │ ├── td-hm_mobilenetv2_8xb32-60e_coco-wholebody-face-256x256.py │ │ │ │ ├── td-hm_res50_8xb32-60e_coco-wholebody-face-256x256.py │ │ │ │ └── td-hm_scnet50_8xb32-60e_coco-wholebody-face-256x256.py │ │ │ ├── cofw │ │ │ │ ├── hrnetv2_cofw.md │ │ │ │ ├── hrnetv2_cofw.yml │ │ │ │ └── td-hm_hrnetv2-w18_8xb64-60e_cofw-256x256.py │ │ │ └── wflw │ │ │ │ ├── hrnetv2_awing_wflw.md │ │ │ │ ├── hrnetv2_awing_wflw.yml │ │ │ │ ├── hrnetv2_dark_wflw.md │ │ │ │ ├── hrnetv2_dark_wflw.yml │ │ │ │ ├── hrnetv2_wflw.md │ │ │ │ ├── hrnetv2_wflw.yml │ │ │ │ ├── td-hm_hrnetv2-w18_8xb64-60e_wflw-256x256.py │ │ │ │ ├── td-hm_hrnetv2-w18_awing-8xb64-60e_wflw-256x256.py │ │ │ │ └── td-hm_hrnetv2-w18_dark-8xb64-60e_wflw-256x256.py │ │ └── topdown_regression │ │ │ ├── README.md │ │ │ └── wflw │ │ │ ├── resnet_softwingloss_wflw.md │ │ │ ├── resnet_softwingloss_wflw.yml │ │ │ ├── resnet_wflw.md │ │ │ ├── resnet_wflw.yml │ │ │ ├── resnet_wingloss_wflw.md │ │ │ ├── resnet_wingloss_wflw.yml │ │ │ ├── td-reg_res50_8xb64-210e_wflw-256x256.py │ │ │ ├── td-reg_res50_softwingloss_8xb64-210e_wflw-256x256.py │ │ │ └── td-reg_res50_wingloss_8xb64-210e_wflw-256x256.py │ ├── fashion_2d_keypoint │ │ ├── README.md │ │ └── topdown_heatmap │ │ │ └── deepfashion2 │ │ │ ├── res50_deepfashion2.md │ │ │ ├── res50_deepfasion2.yml │ │ │ ├── td-hm_res50_1xb64-210e_deepfasion2-long-sleeved-dress-256x192.py │ │ │ ├── td-hm_res50_1xb64-210e_deepfasion2-skirt-256x192.py │ │ │ ├── td-hm_res50_1xb64-210e_deepfasion2-vest-dress-256x192.py │ │ │ ├── td-hm_res50_2xb64-210e_deepfasion2-trousers-256x192.py │ │ │ ├── td-hm_res50_3xb64-210e_deepfasion2-shorts-256x192.py │ │ │ ├── td-hm_res50_4xb64-210e_deepfasion2-short-sleeved-dress-256x192.py │ │ │ ├── td-hm_res50_4xb64-210e_deepfasion2-sling-256x192.py │ │ │ ├── td-hm_res50_4xb64-210e_deepfasion2-sling-dress-256x192.py │ │ │ ├── td-hm_res50_4xb64-210e_deepfasion2-vest-256x192.py │ │ │ ├── td-hm_res50_6xb64-210e_deepfasion2-short-sleeved-shirt-256x192.py │ │ │ ├── td-hm_res50_8xb64-210e_deepfasion2-long-sleeved-outwear-256x192.py │ │ │ ├── td-hm_res50_8xb64-210e_deepfasion2-long-sleeved-shirt-256x192.py │ │ │ └── td-hm_res50_8xb64-210e_deepfasion2-short-sleeved-outwear-256x192.py │ ├── hand_2d_keypoint │ │ ├── README.md │ │ ├── rtmpose │ │ │ ├── README.md │ │ │ ├── coco_wholebody_hand │ │ │ │ ├── rtmpose-m_8xb32-210e_coco-wholebody-hand-256x256.py │ │ │ │ ├── rtmpose_coco_wholebody_hand.md │ │ │ │ └── rtmpose_coco_wholebody_hand.yml │ │ │ └── hand5 │ │ │ │ ├── rtmpose-m_8xb256-210e_hand5-256x256.py │ │ │ │ ├── rtmpose_hand5.md │ │ │ │ └── rtmpose_hand5.yml │ │ ├── topdown_heatmap │ │ │ ├── README.md │ │ │ ├── coco_wholebody_hand │ │ │ │ ├── hourglass_coco_wholebody_hand.md │ │ │ │ ├── hourglass_coco_wholebody_hand.yml │ │ │ │ ├── hrnetv2_coco_wholebody_hand.md │ │ │ │ ├── hrnetv2_coco_wholebody_hand.yml │ │ │ │ ├── hrnetv2_dark_coco_wholebody_hand.md │ │ │ │ ├── hrnetv2_dark_coco_wholebody_hand.yml │ │ │ │ ├── litehrnet_coco_wholebody_hand.md │ │ │ │ ├── litehrnet_coco_wholebody_hand.yml │ │ │ │ ├── mobilenetv2_coco_wholebody_hand.md │ │ │ │ ├── mobilenetv2_coco_wholebody_hand.yml │ │ │ │ ├── resnet_coco_wholebody_hand.md │ │ │ │ ├── resnet_coco_wholebody_hand.yml │ │ │ │ ├── scnet_coco_wholebody_hand.md │ │ │ │ ├── scnet_coco_wholebody_hand.yml │ │ │ │ ├── td-hm_hourglass52_8xb32-210e_coco-wholebody-hand-256x256.py │ │ │ │ ├── td-hm_hrnetv2-w18_8xb32-210e_coco-wholebody-hand-256x256.py │ │ │ │ ├── td-hm_hrnetv2-w18_dark-8xb32-210e_coco-wholebody-hand-256x256.py │ │ │ │ ├── td-hm_litehrnet-w18_8xb32-210e_coco-wholebody-hand-256x256.py │ │ │ │ ├── td-hm_mobilenetv2_8xb32-210e_coco-wholebody-hand-256x256.py │ │ │ │ ├── td-hm_res50_8xb32-210e_coco-wholebody-hand-256x256.py │ │ │ │ └── td-hm_scnet50_8xb32-210e_coco-wholebody-hand-256x256.py │ │ │ ├── freihand2d │ │ │ │ ├── resnet_freihand2d.md │ │ │ │ ├── resnet_freihand2d.yml │ │ │ │ └── td-hm_res50_8xb64-100e_freihand2d-224x224.py │ │ │ ├── onehand10k │ │ │ │ ├── hrnetv2_dark_onehand10k.md │ │ │ │ ├── hrnetv2_dark_onehand10k.yml │ │ │ │ ├── hrnetv2_onehand10k.md │ │ │ │ ├── hrnetv2_onehand10k.yml │ │ │ │ ├── hrnetv2_udp_onehand10k.md │ │ │ │ ├── hrnetv2_udp_onehand10k.yml │ │ │ │ ├── mobilenetv2_onehand10k.md │ │ │ │ ├── mobilenetv2_onehand10k.yml │ │ │ │ ├── resnet_onehand10k.md │ │ │ │ ├── resnet_onehand10k.yml │ │ │ │ ├── td-hm_hrnetv2-w18_8xb64-210e_onehand10k-256x256.py │ │ │ │ ├── td-hm_hrnetv2-w18_dark-8xb64-210e_onehand10k-256x256.py │ │ │ │ ├── td-hm_hrnetv2-w18_udp-8xb64-210e_onehand10k-256x256.py │ │ │ │ ├── td-hm_mobilenetv2_8xb64-210e_onehand10k-256x256.py │ │ │ │ └── td-hm_res50_8xb32-210e_onehand10k-256x256.py │ │ │ └── rhd2d │ │ │ │ ├── hrnetv2_dark_rhd2d.md │ │ │ │ ├── hrnetv2_dark_rhd2d.yml │ │ │ │ ├── hrnetv2_rhd2d.md │ │ │ │ ├── hrnetv2_rhd2d.yml │ │ │ │ ├── hrnetv2_udp_rhd2d.md │ │ │ │ ├── hrnetv2_udp_rhd2d.yml │ │ │ │ ├── mobilenetv2_rhd2d.md │ │ │ │ ├── mobilenetv2_rhd2d.yml │ │ │ │ ├── resnet_rhd2d.md │ │ │ │ ├── resnet_rhd2d.yml │ │ │ │ ├── td-hm_hrnetv2-w18_8xb64-210e_rhd2d-256x256.py │ │ │ │ ├── td-hm_hrnetv2-w18_dark-8xb64-210e_rhd2d-256x256.py │ │ │ │ ├── td-hm_hrnetv2-w18_udp-8xb64-210e_rhd2d-256x256.py │ │ │ │ ├── td-hm_mobilenetv2_8xb64-210e_rhd2d-256x256.py │ │ │ │ └── td-hm_res50_8xb64-210e_rhd2d-256x256.py │ │ └── topdown_regression │ │ │ ├── README.md │ │ │ ├── onehand10k │ │ │ ├── resnet_onehand10k.md │ │ │ ├── resnet_onehand10k.yml │ │ │ └── td-reg_res50_8xb64-210e_onehand10k-256x256.py │ │ │ └── rhd2d │ │ │ ├── resnet_rhd2d.md │ │ │ ├── resnet_rhd2d.yml │ │ │ └── td-reg_res50_8xb64-210e_rhd2d-256x256.py │ ├── hand_3d_keypoint │ │ └── README.md │ ├── hand_gesture │ │ └── README.md │ └── wholebody_2d_keypoint │ │ ├── README.md │ │ ├── rtmpose │ │ ├── README.md │ │ └── coco-wholebody │ │ │ ├── rtmpose-l_8xb32-270e_coco-wholebody-384x288.py │ │ │ ├── rtmpose-l_8xb64-270e_coco-wholebody-256x192.py │ │ │ ├── rtmpose-m_8xb64-270e_coco-wholebody-256x192.py │ │ │ ├── rtmpose_coco-wholebody.md │ │ │ └── rtmpose_coco-wholebody.yml │ │ └── topdown_heatmap │ │ ├── README.md │ │ └── coco-wholebody │ │ ├── cspnext-l_udp_8xb64-210e_coco-wholebody-256x192.py │ │ ├── cspnext-m_udp_8xb64-210e_coco-wholebody-256x192.py │ │ ├── cspnext_udp_coco-wholebody.md │ │ ├── cspnext_udp_coco-wholebody.yml │ │ ├── hrnet_coco-wholebody.md │ │ ├── hrnet_coco-wholebody.yml │ │ ├── hrnet_dark_coco-wholebody.md │ │ ├── hrnet_dark_coco-wholebody.yml │ │ ├── resnet_coco-wholebody.md │ │ ├── resnet_coco-wholebody.yml │ │ ├── td-hm_hrnet-w32_8xb64-210e_coco-wholebody-256x192.py │ │ ├── td-hm_hrnet-w32_8xb64-210e_coco-wholebody-384x288.py │ │ ├── td-hm_hrnet-w32_dark-8xb64-210e_coco-wholebody-256x192.py │ │ ├── td-hm_hrnet-w48_8xb32-210e_coco-wholebody-256x192.py │ │ ├── td-hm_hrnet-w48_8xb32-210e_coco-wholebody-384x288.py │ │ ├── td-hm_hrnet-w48_dark-8xb32-210e_coco-wholebody-384x288.py │ │ ├── td-hm_res101_8xb32-210e_coco-wholebody-256x192.py │ │ ├── td-hm_res101_8xb32-210e_coco-wholebody-384x288.py │ │ ├── td-hm_res152_8xb32-210e_coco-wholebody-256x192.py │ │ ├── td-hm_res152_8xb32-210e_coco-wholebody-384x288.py │ │ ├── td-hm_res50_8xb64-210e_coco-wholebody-256x192.py │ │ ├── td-hm_res50_8xb64-210e_coco-wholebody-384x288.py │ │ ├── td-hm_vipnas-mbv3_8xb64-210e_coco-wholebody-256x192.py │ │ ├── td-hm_vipnas-mbv3_dark-8xb64-210e_coco-wholebody-256x192.py │ │ ├── td-hm_vipnas-res50_8xb64-210e_coco-wholebody-256x192.py │ │ ├── td-hm_vipnas-res50_dark-8xb64-210e_coco-wholebody-256x192.py │ │ ├── vipnas_coco-wholebody.md │ │ ├── vipnas_coco-wholebody.yml │ │ ├── vipnas_dark_coco-wholebody.md │ │ └── vipnas_dark_coco-wholebody.yml │ ├── demo │ ├── MMPose_Tutorial.ipynb │ ├── bottomup_demo.py │ ├── docs │ │ ├── en │ │ │ ├── 2d_animal_demo.md │ │ │ ├── 2d_face_demo.md │ │ │ ├── 2d_hand_demo.md │ │ │ ├── 2d_human_pose_demo.md │ │ │ ├── 2d_wholebody_pose_demo.md │ │ │ ├── mmdet_modelzoo.md │ │ │ └── webcam_api_demo.md │ │ └── zh_cn │ │ │ ├── 2d_animal_demo.md │ │ │ ├── 2d_face_demo.md │ │ │ ├── 2d_hand_demo.md │ │ │ ├── 2d_human_pose_demo.md │ │ │ ├── 2d_wholebody_pose_demo.md │ │ │ ├── mmdet_modelzoo.md │ │ │ └── webcam_api_demo.md │ ├── image_demo.py │ ├── inferencer_demo.py │ ├── mmdetection_cfg │ │ ├── cascade_rcnn_x101_64x4d_fpn_1class.py │ │ ├── cascade_rcnn_x101_64x4d_fpn_coco.py │ │ ├── faster_rcnn_r50_fpn_1class.py │ │ ├── faster_rcnn_r50_fpn_coco.py │ │ ├── mask_rcnn_r50_fpn_2x_coco.py │ │ ├── ssdlite_mobilenetv2-scratch_8xb24-600e_coco.py │ │ ├── ssdlite_mobilenetv2_scratch_600e_onehand.py │ │ ├── yolov3_d53_320_273e_coco.py │ │ └── yolox-s_8xb8-300e_coco-face.py │ ├── mmtracking_cfg │ │ ├── deepsort_faster-rcnn_fpn_4e_mot17-private-half.py │ │ └── tracktor_faster-rcnn_r50_fpn_4e_mot17-private.py │ ├── resources │ │ ├── demo.mp4 │ │ ├── demo_coco.gif │ │ └── sunglasses.jpg │ ├── topdown_demo_with_mmdet.py │ ├── webcam_api_demo.py │ └── webcam_cfg │ │ ├── human_animal_pose.py │ │ ├── human_pose.py │ │ └── test_camera.py │ ├── docker │ ├── Dockerfile │ └── serve │ │ ├── Dockerfile │ │ ├── config.properties │ │ └── entrypoint.sh │ ├── docs │ ├── en │ │ ├── Makefile │ │ ├── _static │ │ │ ├── css │ │ │ │ └── readthedocs.css │ │ │ └── images │ │ │ │ └── mmpose-logo.png │ │ ├── advanced_guides │ │ │ ├── codecs.md │ │ │ ├── customize_datasets.md │ │ │ ├── customize_logging.md │ │ │ ├── customize_optimizer.md │ │ │ ├── customize_transforms.md │ │ │ ├── dataflow.md │ │ │ ├── how_to_deploy.md │ │ │ ├── implement_new_models.md │ │ │ └── model_analysis.md │ │ ├── api.rst │ │ ├── collect_modelzoo.py │ │ ├── collect_projects.py │ │ ├── conf.py │ │ ├── contribution_guide.md │ │ ├── dataset_zoo │ │ │ ├── 2d_animal_keypoint.md │ │ │ ├── 2d_body_keypoint.md │ │ │ ├── 2d_face_keypoint.md │ │ │ ├── 2d_fashion_landmark.md │ │ │ ├── 2d_hand_keypoint.md │ │ │ ├── 2d_wholebody_keypoint.md │ │ │ ├── 3d_body_keypoint.md │ │ │ ├── 3d_body_mesh.md │ │ │ ├── 3d_hand_keypoint.md │ │ │ └── dataset_tools.md │ │ ├── faq.md │ │ ├── guide_to_framework.md │ │ ├── index.rst │ │ ├── installation.md │ │ ├── make.bat │ │ ├── merge_docs.sh │ │ ├── migration.md │ │ ├── notes │ │ │ ├── benchmark.md │ │ │ ├── changelog.md │ │ │ ├── ecosystem.md │ │ │ └── pytorch_2.md │ │ ├── overview.md │ │ ├── projects │ │ │ └── projects.md │ │ ├── quick_run.md │ │ ├── stats.py │ │ ├── switch_language.md │ │ ├── user_guides │ │ │ ├── configs.md │ │ │ ├── inference.md │ │ │ ├── mixed_datasets.md │ │ │ ├── prepare_datasets.md │ │ │ └── train_and_test.md │ │ ├── visualization.md │ │ └── webcam_api.rst │ ├── src │ │ └── papers │ │ │ ├── algorithms │ │ │ ├── associative_embedding.md │ │ │ ├── awingloss.md │ │ │ ├── cid.md │ │ │ ├── cpm.md │ │ │ ├── dark.md │ │ │ ├── debias_ipr.md │ │ │ ├── deeppose.md │ │ │ ├── dekr.md │ │ │ ├── dsnt.md │ │ │ ├── higherhrnet.md │ │ │ ├── hmr.md │ │ │ ├── hourglass.md │ │ │ ├── hrnet.md │ │ │ ├── hrnetv2.md │ │ │ ├── internet.md │ │ │ ├── ipr.md │ │ │ ├── litehrnet.md │ │ │ ├── mspn.md │ │ │ ├── posewarper.md │ │ │ ├── rle.md │ │ │ ├── rsn.md │ │ │ ├── rtmpose.md │ │ │ ├── scnet.md │ │ │ ├── simcc.md │ │ │ ├── simplebaseline2d.md │ │ │ ├── simplebaseline3d.md │ │ │ ├── softwingloss.md │ │ │ ├── udp.md │ │ │ ├── videopose3d.md │ │ │ ├── vipnas.md │ │ │ ├── vitpose.md │ │ │ ├── voxelpose.md │ │ │ └── wingloss.md │ │ │ ├── backbones │ │ │ ├── alexnet.md │ │ │ ├── cpm.md │ │ │ ├── higherhrnet.md │ │ │ ├── hourglass.md │ │ │ ├── hrformer.md │ │ │ ├── hrnet.md │ │ │ ├── hrnetv2.md │ │ │ ├── litehrnet.md │ │ │ ├── mobilenetv2.md │ │ │ ├── mspn.md │ │ │ ├── pvt.md │ │ │ ├── pvtv2.md │ │ │ ├── resnest.md │ │ │ ├── resnet.md │ │ │ ├── resnetv1d.md │ │ │ ├── resnext.md │ │ │ ├── rsn.md │ │ │ ├── scnet.md │ │ │ ├── seresnet.md │ │ │ ├── shufflenetv1.md │ │ │ ├── shufflenetv2.md │ │ │ ├── swin.md │ │ │ ├── vgg.md │ │ │ └── vipnas.md │ │ │ ├── datasets │ │ │ ├── 300w.md │ │ │ ├── aflw.md │ │ │ ├── aic.md │ │ │ ├── animalpose.md │ │ │ ├── ap10k.md │ │ │ ├── atrw.md │ │ │ ├── campus_and_shelf.md │ │ │ ├── coco.md │ │ │ ├── coco_wholebody.md │ │ │ ├── coco_wholebody_face.md │ │ │ ├── coco_wholebody_hand.md │ │ │ ├── cofw.md │ │ │ ├── crowdpose.md │ │ │ ├── deepfashion.md │ │ │ ├── fly.md │ │ │ ├── freihand.md │ │ │ ├── h36m.md │ │ │ ├── halpe.md │ │ │ ├── horse10.md │ │ │ ├── interhand.md │ │ │ ├── jhmdb.md │ │ │ ├── locust.md │ │ │ ├── macaque.md │ │ │ ├── mhp.md │ │ │ ├── mpi_inf_3dhp.md │ │ │ ├── mpii.md │ │ │ ├── mpii_trb.md │ │ │ ├── ochuman.md │ │ │ ├── onehand10k.md │ │ │ ├── panoptic.md │ │ │ ├── panoptic_body3d.md │ │ │ ├── posetrack18.md │ │ │ ├── rhd.md │ │ │ ├── wflw.md │ │ │ └── zebra.md │ │ │ └── techniques │ │ │ ├── albumentations.md │ │ │ ├── awingloss.md │ │ │ ├── dark.md │ │ │ ├── fp16.md │ │ │ ├── fpn.md │ │ │ ├── rle.md │ │ │ ├── smoothnet.md │ │ │ ├── softwingloss.md │ │ │ ├── udp.md │ │ │ └── wingloss.md │ └── zh_cn │ │ ├── Makefile │ │ ├── _static │ │ ├── css │ │ │ └── readthedocs.css │ │ └── images │ │ │ └── mmpose-logo.png │ │ ├── advanced_guides │ │ ├── codecs.md │ │ ├── customize_datasets.md │ │ ├── customize_logging.md │ │ ├── customize_optimizer.md │ │ ├── customize_transforms.md │ │ ├── dataflow.md │ │ ├── how_to_deploy.md │ │ ├── implement_new_models.md │ │ └── model_analysis.md │ │ ├── api.rst │ │ ├── collect_modelzoo.py │ │ ├── collect_projects.py │ │ ├── conf.py │ │ ├── contribution_guide.md │ │ ├── dataset_zoo │ │ ├── 2d_animal_keypoint.md │ │ ├── 2d_body_keypoint.md │ │ ├── 2d_face_keypoint.md │ │ ├── 2d_fashion_landmark.md │ │ ├── 2d_hand_keypoint.md │ │ ├── 2d_wholebody_keypoint.md │ │ ├── 3d_body_keypoint.md │ │ ├── 3d_body_mesh.md │ │ ├── 3d_hand_keypoint.md │ │ └── dataset_tools.md │ │ ├── faq.md │ │ ├── guide_to_framework.md │ │ ├── index.rst │ │ ├── installation.md │ │ ├── make.bat │ │ ├── merge_docs.sh │ │ ├── migration.md │ │ ├── notes │ │ ├── changelog.md │ │ ├── ecosystem.md │ │ ├── projects.md │ │ └── pytorch_2.md │ │ ├── overview.md │ │ ├── quick_run.md │ │ ├── stats.py │ │ ├── switch_language.md │ │ ├── user_guides │ │ ├── advanced_training.md │ │ ├── configs.md │ │ ├── inference.md │ │ ├── mixed_datasets.md │ │ ├── prepare_datasets.md │ │ ├── train_and_test.md │ │ ├── useful_tools.md │ │ └── visualization.md │ │ └── webcam_api.rst │ ├── mmpose │ ├── __init__.py │ ├── apis │ │ ├── __init__.py │ │ ├── inference.py │ │ ├── inferencers │ │ │ ├── __init__.py │ │ │ ├── base_mmpose_inferencer.py │ │ │ ├── mmpose_inferencer.py │ │ │ ├── pose2d_inferencer.py │ │ │ └── utils │ │ │ │ ├── __init__.py │ │ │ │ ├── default_det_models.py │ │ │ │ └── get_model_alias.py │ │ └── webcam │ │ │ ├── __init__.py │ │ │ ├── nodes │ │ │ ├── __init__.py │ │ │ ├── base_visualizer_node.py │ │ │ ├── helper_nodes │ │ │ │ ├── __init__.py │ │ │ │ ├── monitor_node.py │ │ │ │ ├── object_assigner_node.py │ │ │ │ └── recorder_node.py │ │ │ ├── model_nodes │ │ │ │ ├── __init__.py │ │ │ │ ├── detector_node.py │ │ │ │ └── pose_estimator_node.py │ │ │ ├── node.py │ │ │ ├── registry.py │ │ │ └── visualizer_nodes │ │ │ │ ├── __init__.py │ │ │ │ ├── bigeye_effect_node.py │ │ │ │ ├── notice_board_node.py │ │ │ │ ├── object_visualizer_node.py │ │ │ │ └── sunglasses_effect_node.py │ │ │ ├── utils │ │ │ ├── __init__.py │ │ │ ├── buffer.py │ │ │ ├── event.py │ │ │ ├── image_capture.py │ │ │ ├── message.py │ │ │ ├── misc.py │ │ │ └── pose.py │ │ │ └── webcam_executor.py │ ├── codecs │ │ ├── __init__.py │ │ ├── associative_embedding.py │ │ ├── base.py │ │ ├── decoupled_heatmap.py │ │ ├── integral_regression_label.py │ │ ├── megvii_heatmap.py │ │ ├── msra_heatmap.py │ │ ├── regression_label.py │ │ ├── simcc_label.py │ │ ├── spr.py │ │ ├── udp_heatmap.py │ │ └── utils │ │ │ ├── __init__.py │ │ │ ├── gaussian_heatmap.py │ │ │ ├── instance_property.py │ │ │ ├── offset_heatmap.py │ │ │ ├── post_processing.py │ │ │ └── refinement.py │ ├── datasets │ │ ├── __init__.py │ │ ├── builder.py │ │ ├── dataset_wrappers.py │ │ ├── datasets │ │ │ ├── __init__.py │ │ │ ├── animal │ │ │ │ ├── __init__.py │ │ │ │ ├── animalpose_dataset.py │ │ │ │ ├── ap10k_dataset.py │ │ │ │ ├── atrw_dataset.py │ │ │ │ ├── fly_dataset.py │ │ │ │ ├── horse10_dataset.py │ │ │ │ ├── locust_dataset.py │ │ │ │ ├── macaque_dataset.py │ │ │ │ └── zebra_dataset.py │ │ │ ├── base │ │ │ │ ├── __init__.py │ │ │ │ └── base_coco_style_dataset.py │ │ │ ├── body │ │ │ │ ├── __init__.py │ │ │ │ ├── aic_dataset.py │ │ │ │ ├── coco_dataset.py │ │ │ │ ├── crowdpose_dataset.py │ │ │ │ ├── jhmdb_dataset.py │ │ │ │ ├── mhp_dataset.py │ │ │ │ ├── mpii_dataset.py │ │ │ │ ├── mpii_trb_dataset.py │ │ │ │ ├── ochuman_dataset.py │ │ │ │ ├── posetrack18_dataset.py │ │ │ │ └── posetrack18_video_dataset.py │ │ │ ├── face │ │ │ │ ├── __init__.py │ │ │ │ ├── aflw_dataset.py │ │ │ │ ├── coco_wholebody_face_dataset.py │ │ │ │ ├── cofw_dataset.py │ │ │ │ ├── face_300w_dataset.py │ │ │ │ ├── lapa_dataset.py │ │ │ │ └── wflw_dataset.py │ │ │ ├── fashion │ │ │ │ ├── __init__.py │ │ │ │ ├── deepfashion2_dataset.py │ │ │ │ └── deepfashion_dataset.py │ │ │ ├── hand │ │ │ │ ├── __init__.py │ │ │ │ ├── coco_wholebody_hand_dataset.py │ │ │ │ ├── freihand_dataset.py │ │ │ │ ├── onehand10k_dataset.py │ │ │ │ ├── panoptic_hand2d_dataset.py │ │ │ │ └── rhd2d_dataset.py │ │ │ ├── head │ │ │ │ ├── __init__.py │ │ │ │ ├── cephalometric_dataset.py │ │ │ │ └── cephalometric_dataset_ISBI2015.py │ │ │ ├── utils.py │ │ │ └── wholebody │ │ │ │ ├── __init__.py │ │ │ │ ├── coco_wholebody_dataset.py │ │ │ │ └── halpe_dataset.py │ │ ├── samplers.py │ │ └── transforms │ │ │ ├── __init__.py │ │ │ ├── bottomup_transforms.py │ │ │ ├── common_transforms.py │ │ │ ├── converting.py │ │ │ ├── formatting.py │ │ │ ├── loading.py │ │ │ └── topdown_transforms.py │ ├── engine │ │ ├── __init__.py │ │ ├── hooks │ │ │ ├── __init__.py │ │ │ ├── ema_hook.py │ │ │ └── visualization_hook.py │ │ └── optim_wrappers │ │ │ ├── __init__.py │ │ │ └── layer_decay_optim_wrapper.py │ ├── evaluation │ │ ├── __init__.py │ │ ├── functional │ │ │ ├── __init__.py │ │ │ ├── keypoint_eval.py │ │ │ └── nms.py │ │ └── metrics │ │ │ ├── __init__.py │ │ │ ├── cephalometric_metric.py │ │ │ ├── coco_metric.py │ │ │ ├── coco_wholebody_metric.py │ │ │ ├── keypoint_2d_metrics.py │ │ │ ├── keypoint_partition_metric.py │ │ │ └── posetrack18_metric.py │ ├── models │ │ ├── __init__.py │ │ ├── backbones │ │ │ ├── __init__.py │ │ │ ├── alexnet.py │ │ │ ├── base_backbone.py │ │ │ ├── cpm.py │ │ │ ├── hourglass.py │ │ │ ├── hourglass_ae.py │ │ │ ├── hrformer.py │ │ │ ├── hrnet.py │ │ │ ├── litehrnet.py │ │ │ ├── mobilenet_v2.py │ │ │ ├── mobilenet_v3.py │ │ │ ├── monai_unet.py │ │ │ ├── mspn.py │ │ │ ├── pvt.py │ │ │ ├── regnet.py │ │ │ ├── resnest.py │ │ │ ├── resnet.py │ │ │ ├── resnext.py │ │ │ ├── rsn.py │ │ │ ├── scnet.py │ │ │ ├── seresnet.py │ │ │ ├── seresnext.py │ │ │ ├── shufflenet_v1.py │ │ │ ├── shufflenet_v2.py │ │ │ ├── swin.py │ │ │ ├── tcn.py │ │ │ ├── utils │ │ │ │ ├── __init__.py │ │ │ │ ├── channel_shuffle.py │ │ │ │ ├── ckpt_convert.py │ │ │ │ ├── inverted_residual.py │ │ │ │ ├── make_divisible.py │ │ │ │ ├── se_layer.py │ │ │ │ └── utils.py │ │ │ ├── v2v_net.py │ │ │ ├── vgg.py │ │ │ ├── vipnas_mbv3.py │ │ │ └── vipnas_resnet.py │ │ ├── builder.py │ │ ├── data_preprocessors │ │ │ ├── __init__.py │ │ │ └── data_preprocessor.py │ │ ├── heads │ │ │ ├── __init__.py │ │ │ ├── base_head.py │ │ │ ├── coord_cls_heads │ │ │ │ ├── __init__.py │ │ │ │ ├── rtmcc_head.py │ │ │ │ └── simcc_head.py │ │ │ ├── heatmap_heads │ │ │ │ ├── __init__.py │ │ │ │ ├── ae_head.py │ │ │ │ ├── cid_head.py │ │ │ │ ├── cpm_head.py │ │ │ │ ├── heatmap_head.py │ │ │ │ ├── mspn_head.py │ │ │ │ ├── srpose_head.py │ │ │ │ └── vipnas_head.py │ │ │ ├── hybrid_heads │ │ │ │ ├── __init__.py │ │ │ │ └── dekr_head.py │ │ │ └── regression_heads │ │ │ │ ├── __init__.py │ │ │ │ ├── dsnt_head.py │ │ │ │ ├── integral_regression_head.py │ │ │ │ ├── regression_head.py │ │ │ │ └── rle_head.py │ │ ├── losses │ │ │ ├── __init__.py │ │ │ ├── ae_loss.py │ │ │ ├── classification_loss.py │ │ │ ├── heatmap_loss.py │ │ │ ├── loss_wrappers.py │ │ │ └── regression_loss.py │ │ ├── necks │ │ │ ├── __init__.py │ │ │ ├── fmap_proc_neck.py │ │ │ ├── fpn.py │ │ │ ├── gap_neck.py │ │ │ └── posewarper_neck.py │ │ ├── pose_estimators │ │ │ ├── __init__.py │ │ │ ├── base.py │ │ │ ├── bottomup.py │ │ │ └── topdown.py │ │ └── utils │ │ │ ├── __init__.py │ │ │ ├── check_and_update_config.py │ │ │ ├── ckpt_convert.py │ │ │ ├── geometry.py │ │ │ ├── ops.py │ │ │ ├── realnvp.py │ │ │ ├── regularizations.py │ │ │ ├── rtmcc_block.py │ │ │ ├── transformer.py │ │ │ └── tta.py │ ├── registry.py │ ├── structures │ │ ├── __init__.py │ │ ├── bbox │ │ │ ├── __init__.py │ │ │ └── transforms.py │ │ ├── keypoint │ │ │ ├── __init__.py │ │ │ └── transforms.py │ │ ├── multilevel_pixel_data.py │ │ ├── pose_data_sample.py │ │ └── utils.py │ ├── testing │ │ ├── __init__.py │ │ └── _utils.py │ ├── utils │ │ ├── __init__.py │ │ ├── camera.py │ │ ├── collect_env.py │ │ ├── config_utils.py │ │ ├── hooks.py │ │ ├── logger.py │ │ ├── setup_env.py │ │ ├── tensor_utils.py │ │ ├── timer.py │ │ └── typing.py │ ├── version.py │ └── visualization │ │ ├── __init__.py │ │ ├── local_visualizer.py │ │ ├── opencv_backend_visualizer.py │ │ └── simcc_vis.py │ ├── model-index.yml │ ├── projects │ ├── README.md │ ├── awesome-mmpose │ │ └── README.md │ ├── example_project │ │ ├── README.md │ │ ├── configs │ │ │ └── example-head-loss_hrnet-w32_8xb64-210e_coco-256x192.py │ │ └── models │ │ │ ├── __init__.py │ │ │ ├── example_head.py │ │ │ └── example_loss.py │ ├── faq.md │ ├── mmpose4aigc │ │ ├── README.md │ │ ├── README_CN.md │ │ ├── download_models.sh │ │ ├── install_posetracker_linux.sh │ │ ├── mmpose_openpose.sh │ │ ├── mmpose_style_skeleton.sh │ │ └── openpose_visualization.py │ ├── rtmpose │ │ ├── README.md │ │ ├── README_CN.md │ │ ├── benchmark │ │ │ ├── README.md │ │ │ └── README_CN.md │ │ ├── examples │ │ │ ├── README.md │ │ │ ├── RTMPose-Deploy │ │ │ │ ├── README.md │ │ │ │ ├── README_CN.md │ │ │ │ └── Windows │ │ │ │ │ └── OnnxRumtime-CPU │ │ │ │ │ └── src │ │ │ │ │ └── RTMPoseOnnxRuntime │ │ │ │ │ ├── characterset_convert.h │ │ │ │ │ ├── main.cpp │ │ │ │ │ ├── rtmdet_onnxruntime.cpp │ │ │ │ │ ├── rtmdet_onnxruntime.h │ │ │ │ │ ├── rtmpose_onnxruntime.cpp │ │ │ │ │ ├── rtmpose_onnxruntime.h │ │ │ │ │ ├── rtmpose_tracker_onnxruntime.cpp │ │ │ │ │ ├── rtmpose_tracker_onnxruntime.h │ │ │ │ │ └── rtmpose_utils.h │ │ │ └── onnxruntime │ │ │ │ ├── README.md │ │ │ │ ├── README_CN.md │ │ │ │ ├── human-pose.jpeg │ │ │ │ ├── main.py │ │ │ │ └── requirements.txt │ │ ├── rtmdet │ │ │ ├── hand │ │ │ │ └── rtmdet_nano_320-8xb32_hand.py │ │ │ └── person │ │ │ │ ├── rtmdet_m_640-8xb32_coco-person.py │ │ │ │ └── rtmdet_nano_320-8xb32_coco-person.py │ │ └── rtmpose │ │ │ ├── animal_2d_keypoint │ │ │ └── rtmpose-m_8xb64-210e_ap10k-256x256.py │ │ │ ├── body_2d_keypoint │ │ │ ├── rtmpose-l_8xb256-420e_coco-256x192.py │ │ │ ├── rtmpose-l_8xb256-420e_coco-384x288.py │ │ │ ├── rtmpose-m_8xb256-420e_coco-256x192.py │ │ │ ├── rtmpose-m_8xb256-420e_coco-384x288.py │ │ │ ├── rtmpose-s_8xb256-420e_coco-256x192.py │ │ │ └── rtmpose-t_8xb256-420e_coco-256x192.py │ │ │ ├── face_2d_keypoint │ │ │ └── rtmpose-m_8xb64-120e_lapa-256x256.py │ │ │ ├── hand_2d_keypoint │ │ │ └── rtmpose-m_8xb32-210e_coco-wholebody-hand-256x256.py │ │ │ ├── pretrain_cspnext_udp │ │ │ ├── cspnext-l_udp_8xb256-210e_coco-256x192.py │ │ │ ├── cspnext-m_udp_8xb256-210e_coco-256x192.py │ │ │ ├── cspnext-s_udp_8xb256-210e_coco-256x192.py │ │ │ └── cspnext-tiny_udp_8xb256-210e_coco-256x192.py │ │ │ ├── pruning │ │ │ ├── README.md │ │ │ ├── README_CN.md │ │ │ ├── group_fisher_deploy_rtmpose-s_8xb256-420e_aic-coco-256x192.py │ │ │ ├── group_fisher_deploy_rtmpose-s_8xb256-420e_coco-256x192.py │ │ │ ├── group_fisher_finetune_rtmpose-s_8xb256-420e_aic-coco-256x192.py │ │ │ ├── group_fisher_finetune_rtmpose-s_8xb256-420e_coco-256x192.py │ │ │ ├── group_fisher_prune_rtmpose-s_8xb256-420e_aic-coco-256x192.py │ │ │ └── group_fisher_prune_rtmpose-s_8xb256-420e_coco-256x192.py │ │ │ └── wholebody_2d_keypoint │ │ │ ├── rtmpose-l_8xb32-270e_coco-wholebody-384x288.py │ │ │ ├── rtmpose-l_8xb64-270e_coco-wholebody-256x192.py │ │ │ └── rtmpose-m_8xb64-270e_coco-wholebody-256x192.py │ └── yolox-pose │ │ ├── README.md │ │ ├── configs │ │ ├── _base_ │ │ │ ├── datasets │ │ │ └── default_runtime.py │ │ ├── yolox-pose_l_4xb64-300e_coco.py │ │ ├── yolox-pose_m_4xb64-300e_coco.py │ │ ├── yolox-pose_s_8xb32-300e_coco.py │ │ └── yolox-pose_tiny_4xb64-300e_coco.py │ │ ├── datasets │ │ ├── __init__.py │ │ ├── bbox_keypoint_structure.py │ │ ├── coco_dataset.py │ │ └── transforms.py │ │ ├── demo │ │ ├── models │ │ ├── __init__.py │ │ ├── assigner.py │ │ ├── data_preprocessor.py │ │ ├── oks_loss.py │ │ ├── utils.py │ │ └── yolox_pose_head.py │ │ └── tools │ ├── pytest.ini │ ├── requirements.txt │ ├── requirements │ ├── albu.txt │ ├── build.txt │ ├── docs.txt │ ├── mminstall.txt │ ├── optional.txt │ ├── poseval.txt │ ├── readthedocs.txt │ ├── runtime.txt │ └── tests.txt │ ├── resources │ └── mmpose-logo.png │ ├── setup.cfg │ ├── setup.py │ ├── tests │ ├── test_apis │ │ ├── test_inference.py │ │ ├── test_inferencers │ │ │ ├── test_mmpose_inferencer.py │ │ │ └── test_pose2d_inferencer.py │ │ └── test_webcam │ │ │ ├── test_nodes │ │ │ ├── test_big_eye_effect_node.py │ │ │ ├── test_detector_node.py │ │ │ ├── test_monitor_node.py │ │ │ ├── test_notice_board_node.py │ │ │ ├── test_object_assigner_node.py │ │ │ ├── test_object_visualizer_node.py │ │ │ ├── test_pose_estimator_node.py │ │ │ ├── test_recorder_node.py │ │ │ └── test_sunglasses_effect_node.py │ │ │ ├── test_utils │ │ │ ├── test_buffer.py │ │ │ ├── test_event.py │ │ │ ├── test_image_capture.py │ │ │ ├── test_message.py │ │ │ ├── test_misc.py │ │ │ └── test_pose.py │ │ │ └── test_webcam_executor.py │ ├── test_codecs │ │ ├── test_associative_embedding.py │ │ ├── test_decoupled_heatmap.py │ │ ├── test_integral_regression_label.py │ │ ├── test_megvii_heatmap.py │ │ ├── test_msra_heatmap.py │ │ ├── test_regression_label.py │ │ ├── test_simcc_label.py │ │ ├── test_spr.py │ │ └── test_udp_heatmap.py │ ├── test_datasets │ │ ├── test_datasets │ │ │ ├── test_animal_datasets │ │ │ │ ├── test_animalpose_dataset.py │ │ │ │ ├── test_ap10k_dataset.py │ │ │ │ ├── test_atrw_dataset.py │ │ │ │ ├── test_fly_dataset.py │ │ │ │ ├── test_horse10_dataset.py │ │ │ │ ├── test_locust_dataset.py │ │ │ │ ├── test_macaque_dataset.py │ │ │ │ └── test_zebra_dataset.py │ │ │ ├── test_body_datasets │ │ │ │ ├── test_aic_dataset.py │ │ │ │ ├── test_coco_dataset.py │ │ │ │ ├── test_crowdpose_dataset.py │ │ │ │ ├── test_jhmdb_dataset.py │ │ │ │ ├── test_mhp_dataset.py │ │ │ │ ├── test_mpii_dataset.py │ │ │ │ ├── test_mpii_trb_dataset.py │ │ │ │ ├── test_ochuman_dataset.py │ │ │ │ ├── test_posetrack18_dataset.py │ │ │ │ └── test_posetrack18_video_dataset.py │ │ │ ├── test_dataset_wrappers │ │ │ │ └── test_combined_dataset.py │ │ │ ├── test_face_datasets │ │ │ │ ├── test_aflw_dataset.py │ │ │ │ ├── test_coco_wholebody_face_dataset.py │ │ │ │ ├── test_cofw_dataset.py │ │ │ │ ├── test_face_300w_dataset.py │ │ │ │ ├── test_lapa_dataset.py │ │ │ │ └── test_wflw_dataset.py │ │ │ ├── test_fashion_datasets │ │ │ │ └── test_deepfashion_dataset.py │ │ │ ├── test_hand_datasets │ │ │ │ ├── test_coco_wholebody_hand_dataset.py │ │ │ │ ├── test_freihand_dataset.py │ │ │ │ ├── test_onehand10k_dataset.py │ │ │ │ ├── test_panoptic_hand2d_dataset.py │ │ │ │ └── test_rhd2d_dataset.py │ │ │ └── test_wholebody_datasets │ │ │ │ ├── test_coco_wholebody_dataset.py │ │ │ │ └── test_halpe_dataset.py │ │ └── test_transforms │ │ │ ├── test_bottomup_transforms.py │ │ │ ├── test_common_transforms.py │ │ │ ├── test_converting.py │ │ │ ├── test_formatting.py │ │ │ ├── test_loading.py │ │ │ └── test_topdown_transforms.py │ ├── test_engine │ │ └── test_hooks │ │ │ └── test_visualization_hook.py │ ├── test_evaluation │ │ ├── test_functional │ │ │ ├── test_keypoint_eval.py │ │ │ └── test_nms.py │ │ └── test_metrics │ │ │ ├── test_coco_metric.py │ │ │ ├── test_coco_wholebody_metric.py │ │ │ ├── test_keypoint_2d_metrics.py │ │ │ ├── test_keypoint_partition_metric.py │ │ │ └── test_posetrack18_metric.py │ ├── test_models │ │ ├── test_backbones │ │ │ ├── test_alexnet.py │ │ │ ├── test_backbones_utils.py │ │ │ ├── test_cpm.py │ │ │ ├── test_hourglass.py │ │ │ ├── test_hrformer.py │ │ │ ├── test_hrnet.py │ │ │ ├── test_litehrnet.py │ │ │ ├── test_mobilenet_v2.py │ │ │ ├── test_mobilenet_v3.py │ │ │ ├── test_mspn.py │ │ │ ├── test_pvt.py │ │ │ ├── test_regnet.py │ │ │ ├── test_resnest.py │ │ │ ├── test_resnet.py │ │ │ ├── test_resnext.py │ │ │ ├── test_rsn.py │ │ │ ├── test_scnet.py │ │ │ ├── test_seresnet.py │ │ │ ├── test_seresnext.py │ │ │ ├── test_shufflenet_v1.py │ │ │ ├── test_shufflenet_v2.py │ │ │ ├── test_swin.py │ │ │ ├── test_tcn.py │ │ │ ├── test_v2v_net.py │ │ │ ├── test_vgg.py │ │ │ ├── test_vipnas_mbv3.py │ │ │ └── test_vipnas_resnet.py │ │ ├── test_heads │ │ │ ├── test_heatmap_heads │ │ │ │ ├── test_ae_head.py │ │ │ │ ├── test_cid_head.py │ │ │ │ ├── test_cpm_head.py │ │ │ │ ├── test_heatmap_head.py │ │ │ │ ├── test_mspn_head.py │ │ │ │ ├── test_rtmcc_head.py │ │ │ │ ├── test_simcc_head.py │ │ │ │ └── test_vipnas_head.py │ │ │ ├── test_hybrid_heads │ │ │ │ └── test_dekr_head.py │ │ │ └── test_regression_heads │ │ │ │ ├── test_dsnt_head.py │ │ │ │ ├── test_integral_regression_head.py │ │ │ │ ├── test_regression_head.py │ │ │ │ └── test_rle_head.py │ │ ├── test_losses │ │ │ ├── test_ae_loss.py │ │ │ ├── test_classification_losses.py │ │ │ ├── test_heatmap_losses.py │ │ │ └── test_regression_losses.py │ │ ├── test_necks │ │ │ └── test_fmap_proc_neck.py │ │ ├── test_pose_estimators │ │ │ ├── test_bottomup.py │ │ │ └── test_topdown.py │ │ └── test_utils │ │ │ └── test_check_and_update_config.py │ ├── test_structures │ │ ├── test_multilevel_pixel_data.py │ │ └── test_pose_data_sample.py │ ├── test_utils │ │ └── test_setup_env.py │ └── test_visualization │ │ └── test_pose_visualizer.py │ └── tools │ ├── analysis_tools │ ├── analyze_logs.py │ ├── get_flops.py │ └── print_config.py │ ├── dataset_converters │ ├── h36m_to_coco.py │ ├── lapa2coco.py │ ├── mat2json.py │ ├── parse_animalpose_dataset.py │ ├── parse_cofw_dataset.py │ ├── parse_deepposekit_dataset.py │ ├── parse_macaquepose_dataset.py │ ├── preprocess_h36m.py │ └── preprocess_mpi_inf_3dhp.py │ ├── dist_test.sh │ ├── dist_train.sh │ ├── misc │ ├── browse_dataset.py │ ├── keypoints2coco_without_mmdet.py │ └── publish_model.py │ ├── slurm_test.sh │ ├── slurm_train.sh │ ├── test.py │ ├── torchserve │ ├── mmpose2torchserve.py │ ├── mmpose_handler.py │ └── test_torchserver.py │ └── train.py ├── requirements.txt ├── step1_test_mmpose.py ├── step2_prepare_coco_dataset.py ├── step3_train_and_evaluation.py └── step4_test_and_visualize.py /.gitignore: -------------------------------------------------------------------------------- 1 | data 2 | MMPose-checkpoints 3 | work_dirs 4 | *.log 5 | __pycache__ 6 | *.__pycache__ 7 | *.pyc 8 | external_repos -------------------------------------------------------------------------------- /Pictures_for_Github_only/Online Result.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/5k5000/CLdetection2023/d1a01536ad892134c4dd728c87dc9ac1d87b8e11/Pictures_for_Github_only/Online Result.png -------------------------------------------------------------------------------- /Pictures_for_Github_only/mainframework.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/5k5000/CLdetection2023/d1a01536ad892134c4dd728c87dc9ac1d87b8e11/Pictures_for_Github_only/mainframework.png -------------------------------------------------------------------------------- /install_env.sh: -------------------------------------------------------------------------------- 1 | conda create -n LMD python=3.10 2 | conda activate LMD 3 | pip install -r requirements.txt 4 | pip install -U openmim 5 | cd mmpose_package/mmpose 6 | pip install -e . 7 | mim install mmengine 8 | mim install "mmcv>=2.0.0" 9 | pip install --upgrade numpy 10 | 11 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/.owners.yml: -------------------------------------------------------------------------------- 1 | assign: 2 | issues: enabled 3 | pull_requests: disabled 4 | strategy: 5 | # random 6 | daily-shift-based 7 | scedule: 8 | '*/1 * * * *' 9 | assignees: 10 | - Tau-J 11 | - LareinaM 12 | - Ben-Louis 13 | - LareinaM 14 | - Ben-Louis 15 | - Tau-J 16 | - Tau-J 17 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/.readthedocs.yml: -------------------------------------------------------------------------------- 1 | version: 2 2 | 3 | formats: 4 | - epub 5 | 6 | python: 7 | version: 3.7 8 | install: 9 | - requirements: requirements/docs.txt 10 | - requirements: requirements/readthedocs.txt 11 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/CITATION.cff: -------------------------------------------------------------------------------- 1 | cff-version: 1.2.0 2 | message: "If you use this software, please cite it as below." 3 | authors: 4 | - name: "MMPose Contributors" 5 | title: "OpenMMLab Pose Estimation Toolbox and Benchmark" 6 | date-released: 2020-08-31 7 | url: "https://github.com/open-mmlab/mmpose" 8 | license: Apache-2.0 9 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/MANIFEST.in: -------------------------------------------------------------------------------- 1 | include requirements/*.txt 2 | include mmpose/.mim/model-index.yml 3 | recursive-include mmpose/.mim/configs *.py *.yml 4 | recursive-include mmpose/.mim/tools *.py *.sh 5 | recursive-include mmpose/.mim/demo *.py 6 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/animal_2d_keypoint/README.md: -------------------------------------------------------------------------------- 1 | # 2D Animal Keypoint Detection 2 | 3 | 2D animal keypoint detection (animal pose estimation) aims to detect the key-point of different species, including rats, 4 | dogs, macaques, and cheetah. It provides detailed behavioral analysis for neuroscience, medical and ecology applications. 5 | 6 | ## Data preparation 7 | 8 | Please follow [DATA Preparation](/docs/en/dataset_zoo/2d_animal_keypoint.md) to prepare data. 9 | 10 | ## Demo 11 | 12 | Please follow [DEMO](/demo/docs/en/2d_animal_demo.md) to generate fancy demos. 13 | 14 |
15 | 16 |
17 | 18 |
19 | 20 |
21 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/animal_2d_keypoint/rtmpose/README.md: -------------------------------------------------------------------------------- 1 | # RTMPose 2 | 3 | Recent studies on 2D pose estimation have achieved excellent performance on public benchmarks, yet its application in the industrial community still suffers from heavy model parameters and high latency. 4 | In order to bridge this gap, we empirically study five aspects that affect the performance of multi-person pose estimation algorithms: paradigm, backbone network, localization algorithm, training strategy, and deployment inference, and present a high-performance real-time multi-person pose estimation framework, **RTMPose**, based on MMPose. 5 | Our RTMPose-m achieves **75.8% AP** on COCO with **90+ FPS** on an Intel i7-11700 CPU and **430+ FPS** on an NVIDIA GTX 1660 Ti GPU, and RTMPose-l achieves **67.0% AP** on COCO-WholeBody with **130+ FPS**, outperforming existing open-source libraries. 6 | To further evaluate RTMPose's capability in critical real-time applications, we also report the performance after deploying on the mobile device. 7 | 8 | ## Results and Models 9 | 10 | ### AP-10K Dataset 11 | 12 | Results on AP-10K validation set 13 | 14 | | Model | Input Size | AP | Details and Download | 15 | | :-------: | :--------: | :---: | :------------------------------------------: | 16 | | RTMPose-m | 256x256 | 0.722 | [rtmpose_cp10k.md](./ap10k/rtmpose_ap10k.md) | 17 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/animal_2d_keypoint/rtmpose/ap10k/rtmpose_ap10k.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/animal_2d_keypoint/rtmpose/ap10k/rtmpose-m_8xb64-210e_ap10k-256x256.py 3 | In Collection: RTMPose 4 | Alias: animal 5 | Metadata: 6 | Architecture: 7 | - RTMPose 8 | Training Data: AP-10K 9 | Name: rtmpose-m_8xb64-210e_ap10k-256x256 10 | Results: 11 | - Dataset: AP-10K 12 | Metrics: 13 | AP: 0.722 14 | AP@0.5: 0.939 15 | AP@0.75: 0.788 16 | AP (L): 0.728 17 | AP (M): 0.569 18 | Task: Animal 2D Keypoint 19 | Weights: https://download.openmmlab.com/mmpose/v1/projects/rtmposev1/rtmpose-m_simcc-ap10k_pt-aic-coco_210e-256x256-7a041aa1_20230206.pth 20 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/animal_2d_keypoint/topdown_heatmap/animalpose/hrnet_animalpose.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/animal_2d_keypoint/topdown_heatmap/animalpose/td-hm_hrnet-w32_8xb64-210e_animalpose-256x256.py 3 | In Collection: HRNet 4 | Metadata: 5 | Architecture: &id001 6 | - HRNet 7 | Training Data: Animal-Pose 8 | Name: td-hm_hrnet-w32_8xb64-210e_animalpose-256x256 9 | Results: 10 | - Dataset: Animal-Pose 11 | Metrics: 12 | AP: 0.740 13 | AP@0.5: 0.959 14 | AP@0.75: 0.833 15 | AR: 0.780 16 | AR@0.5: 0.965 17 | Task: Animal 2D Keypoint 18 | Weights: https://download.openmmlab.com/mmpose/animal/hrnet/hrnet_w32_animalpose_256x256-1aa7f075_20210426.pth 19 | - Config: configs/animal_2d_keypoint/topdown_heatmap/animalpose/td-hm_hrnet-w48_8xb64-210e_animalpose-256x256.py 20 | In Collection: HRNet 21 | Metadata: 22 | Architecture: *id001 23 | Training Data: Animal-Pose 24 | Name: td-hm_hrnet-w48_8xb64-210e_animalpose-256x256 25 | Results: 26 | - Dataset: Animal-Pose 27 | Metrics: 28 | AP: 0.738 29 | AP@0.5: 0.958 30 | AP@0.75: 0.831 31 | AR: 0.778 32 | AR@0.5: 0.962 33 | Task: Animal 2D Keypoint 34 | Weights: https://download.openmmlab.com/mmpose/animal/hrnet/hrnet_w48_animalpose_256x256-34644726_20210426.pth 35 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/animal_2d_keypoint/topdown_heatmap/ap10k/cspnext_udp_ap10k.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/animal_2d_keypoint/topdown_heatmap/ap10k/cspnext-m_udp_8xb64-210e_ap10k-256x256.py 3 | In Collection: UDP 4 | Metadata: 5 | Architecture: &id001 6 | - UDP 7 | - HRNet 8 | Training Data: AP-10K 9 | Name: cspnext-m_udp_8xb64-210e_ap10k-256x256 10 | Results: 11 | - Dataset: AP-10K 12 | Metrics: 13 | AP: 0.703 14 | AP@0.5: 0.944 15 | AP@0.75: 0.776 16 | AP (L): 0.71 17 | AP (M): 0.513 18 | Task: Animal 2D Keypoint 19 | Weights: https://download.openmmlab.com/mmpose/v1/projects/rtmposev1/cspnext-m_udp-ap10k_pt-in1k_210e-256x256-1f2d947a_20230123.pth 20 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/animal_2d_keypoint/topdown_heatmap/ap10k/hrnet_ap10k.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/animal_2d_keypoint/topdown_heatmap/ap10k/td-hm_hrnet-w32_8xb64-210e_ap10k-256x256.py 3 | In Collection: HRNet 4 | Metadata: 5 | Architecture: &id001 6 | - HRNet 7 | Training Data: AP-10K 8 | Name: td-hm_hrnet-w32_8xb64-210e_ap10k-256x256 9 | Results: 10 | - Dataset: AP-10K 11 | Metrics: 12 | AP: 0.722 13 | AP@0.5: 0.935 14 | AP@0.75: 0.789 15 | AP (L): 0.729 16 | AP (M): 0.557 17 | Task: Animal 2D Keypoint 18 | Weights: https://download.openmmlab.com/mmpose/animal/hrnet/hrnet_w32_ap10k_256x256-18aac840_20211029.pth 19 | - Config: configs/animal_2d_keypoint/topdown_heatmap/ap10k/td-hm_hrnet-w48_8xb64-210e_ap10k-256x256.py 20 | In Collection: HRNet 21 | Metadata: 22 | Architecture: *id001 23 | Training Data: AP-10K 24 | Name: td-hm_hrnet-w48_8xb64-210e_ap10k-256x256 25 | Results: 26 | - Dataset: AP-10K 27 | Metrics: 28 | AP: 0.728 29 | AP@0.5: 0.936 30 | AP@0.75: 0.802 31 | AP (L): 0.735 32 | AP (M): 0.577 33 | Task: Animal 2D Keypoint 34 | Weights: https://download.openmmlab.com/mmpose/animal/hrnet/hrnet_w48_ap10k_256x256-d95ab412_20211029.pth 35 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/animal_2d_keypoint/topdown_heatmap/ap10k/resnet_ap10k.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/animal_2d_keypoint/topdown_heatmap/ap10k/td-hm_res50_8xb64-210e_ap10k-256x256.py 3 | In Collection: SimpleBaseline2D 4 | Metadata: 5 | Architecture: &id001 6 | - SimpleBaseline2D 7 | - ResNet 8 | Training Data: AP-10K 9 | Name: td-hm_res50_8xb64-210e_ap10k-256x256 10 | Results: 11 | - Dataset: AP-10K 12 | Metrics: 13 | AP: 0.680 14 | AP@0.5: 0.926 15 | AP@0.75: 0.738 16 | AP (L): 0.687 17 | AP (M): 0.552 18 | Task: Animal 2D Keypoint 19 | Weights: https://download.openmmlab.com/mmpose/animal/resnet/res50_ap10k_256x256-35760eb8_20211029.pth 20 | - Config: configs/animal_2d_keypoint/topdown_heatmap/ap10k/td-hm_res101_8xb64-210e_ap10k-256x256.py 21 | In Collection: SimpleBaseline2D 22 | Metadata: 23 | Architecture: *id001 24 | Training Data: AP-10K 25 | Name: td-hm_res101_8xb64-210e_ap10k-256x256 26 | Results: 27 | - Dataset: AP-10K 28 | Metrics: 29 | AP: 0.681 30 | AP@0.5: 0.921 31 | AP@0.75: 0.751 32 | AP (L): 0.690 33 | AP (M): 0.545 34 | Task: Animal 2D Keypoint 35 | Weights: https://download.openmmlab.com/mmpose/animal/resnet/res101_ap10k_256x256-9edfafb9_20211029.pth 36 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/body_2d_keypoint/README.md: -------------------------------------------------------------------------------- 1 | # Human Body 2D Pose Estimation 2 | 3 | Multi-person human pose estimation is defined as the task of detecting the poses (or keypoints) of all people from an input image. 4 | 5 | Existing approaches can be categorized into top-down and bottom-up approaches. 6 | 7 | Top-down methods (e.g. DeepPose) divide the task into two stages: human detection and pose estimation. They perform human detection first, followed by single-person pose estimation given human bounding boxes. 8 | 9 | Bottom-up approaches (e.g. Associative Embedding) first detect all the keypoints and then group/associate them into person instances. 10 | 11 | ## Data preparation 12 | 13 | Please follow [DATA Preparation](/docs/en/dataset_zoo/2d_body_keypoint.md) to prepare data. 14 | 15 | ## Demo 16 | 17 | Please follow [Demo](/demo/docs/en/2d_human_pose_demo.md#2d-human-pose-demo) to run demos. 18 | 19 |
20 |
21 |
22 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/body_2d_keypoint/associative_embedding/README.md: -------------------------------------------------------------------------------- 1 | # Associative embedding: End-to-end learning for joint detection and grouping (AE) 2 | 3 | Associative Embedding is one of the most popular 2D bottom-up pose estimation approaches, that first detect all the keypoints and then group/associate them into person instances. 4 | 5 | In order to group all the predicted keypoints to individuals, a tag is also predicted for each detected keypoint. Tags of the same person are similar, while tags of different people are different. Thus the keypoints can be grouped according to the tags. 6 | 7 |
8 | 9 |
10 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/body_2d_keypoint/dekr/README.md: -------------------------------------------------------------------------------- 1 | # Bottom-up Human Pose Estimation via Disentangled Keypoint Regression (DEKR) 2 | 3 | 4 | 5 |
6 | DEKR (CVPR'2021) 7 | 8 | ```bibtex 9 | @inproceedings{geng2021bottom, 10 | title={Bottom-up human pose estimation via disentangled keypoint regression}, 11 | author={Geng, Zigang and Sun, Ke and Xiao, Bin and Zhang, Zhaoxiang and Wang, Jingdong}, 12 | booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition}, 13 | pages={14676--14686}, 14 | year={2021} 15 | } 16 | ``` 17 | 18 |
19 | 20 | DEKR is a popular 2D bottom-up pose estimation approach that simultaneously detects all the instances and regresses the offsets from the instance centers to joints. 21 | 22 | In order to predict the offsets more accurately, the offsets of different joints are regressed using separated branches with deformable convolutional layers. Thus convolution kernels with different shapes are adopted to extract features for the corresponding joint. 23 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/body_2d_keypoint/dekr/crowdpose/hrnet_crowdpose.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/body_2d_keypoint/dekr/crowdpose/dekr_hrnet-w32_8xb10-300e_crowdpose-512x512.py 3 | In Collection: DEKR 4 | Metadata: 5 | Architecture: &id001 6 | - DEKR 7 | - HRNet 8 | Training Data: CrowdPose 9 | Name: dekr_hrnet-w32_8xb10-300e_crowdpose-512x512 10 | Results: 11 | - Dataset: CrowdPose 12 | Metrics: 13 | AP: 0.663 14 | AP@0.5: 0.857 15 | AP@0.75: 0.714 16 | AP (E): 0.74 17 | AP (M): 0.671 18 | AP (L): 0.576 19 | Task: Body 2D Keypoint 20 | Weights: https://download.openmmlab.com/mmpose/v1/body_2d_keypoint/dekr/crowdpose/dekr_hrnet-w32_8xb10-140e_crowdpose-512x512_147bae97-20221228.pth 21 | - Config: configs/body_2d_keypoint/dekr/crowdpose/dekr_hrnet-w48_8xb5-300e_crowdpose-640x640.py 22 | In Collection: DEKR 23 | Metadata: 24 | Architecture: *id001 25 | Training Data: CrowdPose 26 | Name: dekr_hrnet-w48_8xb5-300e_crowdpose-640x640 27 | Results: 28 | - Dataset: CrowdPose 29 | Metrics: 30 | AP: 0.679 31 | AP@0.5: 0.869 32 | AP@0.75: 0.731 33 | AP (E): 0.753 34 | AP (M): 0.688 35 | AP (L): 0.593 36 | Task: Body 2D Keypoint 37 | Weights: https://download.openmmlab.com/mmpose/bottom_up/dekr/hrnet_w48_crowdpose_640x640-ef6b6040_20220930.pth 38 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/body_2d_keypoint/integral_regression/README.md: -------------------------------------------------------------------------------- 1 | # Top-down integral-regression-based pose estimation 2 | 3 | Top-down methods divide the task into two stages: object detection, followed by single-object pose estimation given object bounding boxes. At the 2nd stage, integral regression based methods use a simple integral operation relates and unifies the heatmap and joint regression differentiably, thus obtain the keypoint coordinates given the features extracted from the bounding box area, following the paradigm introduced in [Integral Human Pose Regression](https://arxiv.org/abs/1711.08229). 4 | 5 | ## Results and Models 6 | 7 | ### COCO Dataset 8 | 9 | Results on COCO val2017 with detector having human AP of 56.4 on COCO val2017 dataset 10 | 11 | | Model | Input Size | AP | AR | Details and Download | 12 | | :------------------: | :--------: | :---: | :---: | :---------------------------------------------------: | 13 | | ResNet-50+Debias-IPR | 256x256 | 0.675 | 0.765 | [resnet_debias_coco.md](./coco/resnet_debias_coco.md) | 14 | | ResNet-50+DSNT | 256x256 | 0.674 | 0.764 | [resnet_dsnt_coco.md](./coco/resnet_dsnt_coco.md) | 15 | | ResNet-50+IPR | 256x256 | 0.633 | 0.730 | [resnet_ipr_coco.md](./coco/resnet_ipr_coco.md) | 16 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/body_2d_keypoint/integral_regression/coco/resnet_debias_coco.yml: -------------------------------------------------------------------------------- 1 | Collections: 2 | - Name: DebiasIPR 3 | Paper: 4 | Title: Removing the Bias of Integral Pose Regression 5 | URL: https://openaccess.thecvf.com/content/ICCV2021/papers/Gu_Removing_the_Bias_of_Integral_Pose_Regression_ICCV_2021_paper.pdf 6 | README: https://github.com/open-mmlab/mmpose/blob/main/docs/src/papers/algorithms/debias_ipr.md 7 | Models: 8 | - Config: configs/body_2d_keypoint/integral_regression/coco/ipr_res50_debias--8xb64-210e_coco-256x256.py 9 | In Collection: DebiasIPR 10 | Metadata: 11 | Architecture: &id001 12 | - Debias 13 | - ResNet 14 | Training Data: COCO 15 | Name: ipr_res50_debias--8xb64-210e_coco-256x256 16 | Results: 17 | - Dataset: COCO 18 | Metrics: 19 | AP: 0.675 20 | AP@0.5: 0.872 21 | AP@0.75: 0.74 22 | AR: 0.765 23 | AR@0.5: 0.928 24 | Task: Body 2D Keypoint 25 | Weights: https://download.openmmlab.com/mmpose/v1/body_2d_keypoint/integral_regression/coco/ipr_res50_debias-8xb64-210e_coco-256x256-055a7699_20220913.pth 26 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/body_2d_keypoint/integral_regression/coco/resnet_dsnt_coco.yml: -------------------------------------------------------------------------------- 1 | Collections: 2 | - Name: DSNT 3 | Paper: 4 | Title: Numerical Coordinate Regression with Convolutional Neural Networks 5 | URL: https://arxiv.org/abs/1801.07372v2 6 | README: https://github.com/open-mmlab/mmpose/blob/main/docs/src/papers/algorithms/dsnt.md 7 | Models: 8 | - Config: configs/body_2d_keypoint/integral_regression/coco/ipr_res50_dsnt-8xb64-210e_coco-256x256.py 9 | In Collection: DSNT 10 | Metadata: 11 | Architecture: &id001 12 | - DSNT 13 | - ResNet 14 | Training Data: COCO 15 | Name: ipr_res50_dsnt-8xb64-210e_coco-256x256 16 | Results: 17 | - Dataset: COCO 18 | Metrics: 19 | AP: 0.674 20 | AP@0.5: 0.87 21 | AP@0.75: 0.744 22 | AR: 0.764 23 | AR@0.5: 0.928 24 | Task: Body 2D Keypoint 25 | Weights: https://download.openmmlab.com/mmpose/v1/body_2d_keypoint/integral_regression/coco/ipr_res50_dsnt-8xb64-210e_coco-256x256-441eedc0_20220913.pth 26 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/body_2d_keypoint/integral_regression/coco/resnet_ipr_coco.yml: -------------------------------------------------------------------------------- 1 | Collections: 2 | - Name: IPR 3 | Paper: 4 | Title: Integral human pose regression 5 | URL: https://arxiv.org/abs/1711.08229 6 | README: https://github.com/open-mmlab/mmpose/blob/main/docs/src/papers/algorithms/ipr.md 7 | Models: 8 | - Config: configs/body_2d_keypoint/integral_regression/coco/ipr_res50_8xb64-210e_coco-256x256.py 9 | In Collection: IPR 10 | Metadata: 11 | Architecture: &id001 12 | - IPR 13 | - ResNet 14 | Training Data: COCO 15 | Name: ipr_res50_8xb64-210e_coco-256x256 16 | Results: 17 | - Dataset: COCO 18 | Metrics: 19 | AP: 0.633 20 | AP@0.5: 0.86 21 | AP@0.75: 0.703 22 | AR: 0.73 23 | AR@0.5: 0.919 24 | Task: Body 2D Keypoint 25 | Weights: https://download.openmmlab.com/mmpose/v1/body_2d_keypoint/integral_regression/coco/ipr_res50_8xb64-210e_coco-256x256-a3898a33_20220913.pth 26 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/body_2d_keypoint/rtmpose/crowdpose/rtmpose_crowdpose.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/body_2d_keypoint/rtmpose/crowdpose/rtmpose-m_8xb64-210e_crowdpose-256x192.py 3 | In Collection: RTMPose 4 | Metadata: 5 | Architecture: 6 | - RTMPose 7 | Training Data: CrowdPose 8 | Name: rtmpose-t_8xb256-420e_coco-256x192 9 | Results: 10 | - Dataset: CrowdPose 11 | Metrics: 12 | AP: 0.706 13 | AP@0.5: 0.841 14 | AP@0.75: 0.765 15 | AP (E): 0.799 16 | AP (M): 0.719 17 | AP (L): 0.582 18 | Task: Body 2D Keypoint 19 | Weights: https://download.openmmlab.com/mmpose/v1/projects/rtmposev1/rtmpose-m_simcc-crowdpose_pt-aic-coco_210e-256x192-e6192cac_20230224.pth 20 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/body_2d_keypoint/rtmpose/mpii/rtmpose_mpii.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/body_2d_keypoint/rtmpose/mpii/rtmpose-m_8xb64-210e_mpii-256x256.py 3 | In Collection: RTMPose 4 | Metadata: 5 | Architecture: 6 | - RTMPose 7 | Training Data: MPII 8 | Name: rtmpose-m_8xb64-210e_mpii-256x256 9 | Results: 10 | - Dataset: MPII 11 | Metrics: 12 | Mean: 0.907 13 | Mean@0.1: 0.348 14 | Task: Body 2D Keypoint 15 | Weights: https://download.openmmlab.com/mmpose/v1/projects/rtmposev1/rtmpose-m_simcc-mpii_pt-aic-coco_210e-256x256-ec4dbec8_20230206.pth 16 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/body_2d_keypoint/simcc/coco/mobilenetv2_coco.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/body_2d_keypoint/simcc/coco/simcc_mobilenetv2_wo-deconv-8xb64-210e_coco-256x192.py 3 | In Collection: SimCC 4 | Metadata: 5 | Architecture: &id001 6 | - SimCC 7 | - MobilenetV2 8 | Training Data: COCO 9 | Name: simcc_mobilenetv2_wo-deconv-8xb64-210e_coco-256x192 10 | Results: 11 | - Dataset: COCO 12 | Metrics: 13 | AP: 0.62 14 | AP@0.5: 0.855 15 | AP@0.75: 0.697 16 | AR: 0.678 17 | AR@0.5: 0.902 18 | Task: Body 2D Keypoint 19 | Weights: https://download.openmmlab.com/mmpose/v1/body_2d_keypoint/simcc/coco/simcc_mobilenetv2_wo-deconv-8xb64-210e_coco-256x192-4b0703bb_20221010.pth 20 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/body_2d_keypoint/simcc/coco/vipnas_coco.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/body_2d_keypoint/simcc/coco/simcc_vipnas-mbv3_8xb64-210e_coco-256x192.py 3 | In Collection: SimCC 4 | Metadata: 5 | Architecture: &id001 6 | - SimCC 7 | - ViPNAS 8 | Training Data: COCO 9 | Name: simcc_vipnas-mbv3_8xb64-210e_coco-256x192 10 | Results: 11 | - Dataset: COCO 12 | Metrics: 13 | AP: 0.695 14 | AP@0.5: 0.883 15 | AP@0.75: 0.772 16 | AR: 0.755 17 | AR@0.5: 0.927 18 | Task: Body 2D Keypoint 19 | Weights: https://download.openmmlab.com/mmpose/v1/body_2d_keypoint/simcc/coco/simcc_vipnas-mbv3_8xb64-210e_coco-256x192-719f3489_20220922.pth 20 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/body_2d_keypoint/topdown_heatmap/aic/hrnet_aic.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/body_2d_keypoint/topdown_heatmap/aic/td-hm_hrnet-w32_8xb64-210e_aic-256x192.py 3 | In Collection: HRNet 4 | Metadata: 5 | Architecture: 6 | - HRNet 7 | Training Data: AI Challenger 8 | Name: td-hm_hrnet-w32_8xb64-210e_aic-256x192 9 | Results: 10 | - Dataset: AI Challenger 11 | Metrics: 12 | AP: 0.323 13 | AP@0.5: 0.761 14 | AP@0.75: 0.218 15 | AR: 0.366 16 | AR@0.5: 0.789 17 | Task: Body 2D Keypoint 18 | Weights: https://download.openmmlab.com/mmpose/top_down/hrnet/hrnet_w32_aic_256x192-30a4e465_20200826.pth 19 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/body_2d_keypoint/topdown_heatmap/aic/resnet_aic.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/body_2d_keypoint/topdown_heatmap/aic/td-hm_res101_8xb64-210e_aic-256x192.py 3 | In Collection: SimpleBaseline2D 4 | Metadata: 5 | Architecture: 6 | - SimpleBaseline2D 7 | - ResNet 8 | Training Data: AI Challenger 9 | Name: td-hm_res101_8xb64-210e_aic-256x192 10 | Results: 11 | - Dataset: AI Challenger 12 | Metrics: 13 | AP: 0.294 14 | AP@0.5: 0.736 15 | AP@0.75: 0.172 16 | AR: 0.337 17 | AR@0.5: 0.762 18 | Task: Body 2D Keypoint 19 | Weights: https://download.openmmlab.com/mmpose/top_down/resnet/res101_aic_256x192-79b35445_20200826.pth 20 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/body_2d_keypoint/topdown_heatmap/coco/alexnet_coco.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/body_2d_keypoint/topdown_heatmap/coco/td-hm_alexnet_8xb64-210e_coco-256x192.py 3 | In Collection: SimpleBaseline2D 4 | Metadata: 5 | Architecture: 6 | - SimpleBaseline2D 7 | - AlexNet 8 | Training Data: COCO 9 | Name: td-hm_alexnet_8xb64-210e_coco-256x192 10 | Results: 11 | - Dataset: COCO 12 | Metrics: 13 | AP: 0.448 14 | AP@0.5: 0.767 15 | AP@0.75: 0.461 16 | AR: 0.521 17 | AR@0.5: 0.829 18 | Task: Body 2D Keypoint 19 | Weights: https://download.openmmlab.com/mmpose/top_down/alexnet/alexnet_coco_256x192-a7b1fd15_20200727.pth 20 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/body_2d_keypoint/topdown_heatmap/coco/mobilenetv2_coco.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/body_2d_keypoint/topdown_heatmap/coco/td-hm_mobilenetv2_8xb64-210e_coco-256x192.py 3 | In Collection: SimpleBaseline2D 4 | Metadata: 5 | Architecture: &id001 6 | - SimpleBaseline2D 7 | - MobilenetV2 8 | Training Data: COCO 9 | Name: td-hm_mobilenetv2_8xb64-210e_coco-256x192 10 | Results: 11 | - Dataset: COCO 12 | Metrics: 13 | AP: 0.648 14 | AP@0.5: 0.874 15 | AP@0.75: 0.725 16 | AR: 0.709 17 | AR@0.5: 0.918 18 | Task: Body 2D Keypoint 19 | Weights: https://download.openmmlab.com/mmpose/v1/body_2d_keypoint/topdown_heatmap/coco/td-hm_mobilenetv2_8xb64-210e_coco-256x192-55a04c35_20221016.pth 20 | - Config: configs/body_2d_keypoint/topdown_heatmap/coco/td-hm_mobilenetv2_8xb64-210e_coco-384x288.py 21 | In Collection: SimpleBaseline2D 22 | Metadata: 23 | Architecture: *id001 24 | Training Data: COCO 25 | Name: td-hm_mobilenetv2_8xb64-210e_coco-384x288 26 | Results: 27 | - Dataset: COCO 28 | Metrics: 29 | AP: 0.677 30 | AP@0.5: 0.882 31 | AP@0.75: 0.746 32 | AR: 0.734 33 | AR@0.5: 0.920 34 | Task: Body 2D Keypoint 35 | Weights: https://download.openmmlab.com/mmpose/v1/body_2d_keypoint/topdown_heatmap/coco/td-hm_mobilenetv2_8xb64-210e_coco-384x288-d3ab1457_20221013.pth 36 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/body_2d_keypoint/topdown_heatmap/coco/pvt_coco.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/body_2d_keypoint/topdown_heatmap/coco/td-hm_pvt-s_8xb64-210e_coco-256x192.py 3 | In Collection: SimpleBaseline2D 4 | Metadata: 5 | Architecture: &id001 6 | - SimpleBaseline2D 7 | - PVT 8 | Training Data: COCO 9 | Name: td-hm_pvt-s_8xb64-210e_coco-256x192 10 | Results: 11 | - Dataset: COCO 12 | Metrics: 13 | AP: 0.714 14 | AP@0.5: 0.896 15 | AP@0.75: 0.794 16 | AR: 0.773 17 | AR@0.5: 0.936 18 | Task: Body 2D Keypoint 19 | Weights: https://download.openmmlab.com/mmpose/top_down/pvt/pvt_small_coco_256x192-4324a49d_20220501.pth 20 | - Config: configs/body_2d_keypoint/topdown_heatmap/coco/td-hm_pvtv2-b2_8xb64-210e_coco-256x192.py 21 | In Collection: SimpleBaseline2D 22 | Metadata: 23 | Architecture: *id001 24 | Training Data: COCO 25 | Name: td-hm_pvtv2-b2_8xb64-210e_coco-256x192 26 | Results: 27 | - Dataset: COCO 28 | Metrics: 29 | AP: 0.737 30 | AP@0.5: 0.905 31 | AP@0.75: 0.812 32 | AR: 0.791 33 | AR@0.5: 0.942 34 | Task: Body 2D Keypoint 35 | Weights: https://download.openmmlab.com/mmpose/top_down/pvt/pvtv2_b2_coco_256x192-b4212737_20220501.pth 36 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/body_2d_keypoint/topdown_heatmap/coco/shufflenetv1_coco.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/body_2d_keypoint/topdown_heatmap/coco/td-hm_shufflenetv1_8xb64-210e_coco-256x192.py 3 | In Collection: SimpleBaseline2D 4 | Metadata: 5 | Architecture: &id001 6 | - SimpleBaseline2D 7 | - ShufflenetV1 8 | Training Data: COCO 9 | Name: td-hm_shufflenetv1_8xb64-210e_coco-256x192 10 | Results: 11 | - Dataset: COCO 12 | Metrics: 13 | AP: 0.587 14 | AP@0.5: 0.849 15 | AP@0.75: 0.654 16 | AR: 0.654 17 | AR@0.5: 0.896 18 | Task: Body 2D Keypoint 19 | Weights: https://download.openmmlab.com/mmpose/v1/body_2d_keypoint/topdown_heatmap/coco/td-hm_shufflenetv1_8xb64-210e_coco-256x192-7a7ea4f4_20221013.pth 20 | - Config: configs/body_2d_keypoint/topdown_heatmap/coco/td-hm_shufflenetv1_8xb64-210e_coco-384x288.py 21 | In Collection: SimpleBaseline2D 22 | Metadata: 23 | Architecture: *id001 24 | Training Data: COCO 25 | Name: td-hm_shufflenetv1_8xb64-210e_coco-384x288 26 | Results: 27 | - Dataset: COCO 28 | Metrics: 29 | AP: 0.626 30 | AP@0.5: 0.862 31 | AP@0.75: 0.696 32 | AR: 0.687 33 | AR@0.5: 0.903 34 | Task: Body 2D Keypoint 35 | Weights: https://download.openmmlab.com/mmpose/v1/body_2d_keypoint/topdown_heatmap/coco/td-hm_shufflenetv1_8xb64-210e_coco-384x288-8342f8ba_20221013.pth 36 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/body_2d_keypoint/topdown_heatmap/coco/shufflenetv2_coco.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/body_2d_keypoint/topdown_heatmap/coco/td-hm_shufflenetv2_8xb64-210e_coco-256x192.py 3 | In Collection: SimpleBaseline2D 4 | Metadata: 5 | Architecture: &id001 6 | - SimpleBaseline2D 7 | - ShufflenetV2 8 | Training Data: COCO 9 | Name: td-hm_shufflenetv2_8xb64-210e_coco-256x192 10 | Results: 11 | - Dataset: COCO 12 | Metrics: 13 | AP: 0.602 14 | AP@0.5: 0.857 15 | AP@0.75: 0.672 16 | AR: 0.668 17 | AR@0.5: 0.902 18 | Task: Body 2D Keypoint 19 | Weights: https://download.openmmlab.com/mmpose/v1/body_2d_keypoint/topdown_heatmap/coco/td-hm_shufflenetv2_8xb64-210e_coco-256x192-51fb931e_20221014.pth 20 | - Config: configs/body_2d_keypoint/topdown_heatmap/coco/td-hm_shufflenetv2_8xb64-210e_coco-384x288.py 21 | In Collection: SimpleBaseline2D 22 | Metadata: 23 | Architecture: *id001 24 | Training Data: COCO 25 | Name: td-hm_shufflenetv2_8xb64-210e_coco-384x288 26 | Results: 27 | - Dataset: COCO 28 | Metrics: 29 | AP: 0.638 30 | AP@0.5: 0.866 31 | AP@0.75: 0.707 32 | AR: 0.699 33 | AR@0.5: 0.91 34 | Task: Body 2D Keypoint 35 | Weights: https://download.openmmlab.com/mmpose/v1/body_2d_keypoint/topdown_heatmap/coco/td-hm_shufflenetv2_8xb64-210e_coco-384x288-d30ab55c_20221014.pth 36 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/body_2d_keypoint/topdown_heatmap/coco/td-hm_hrnet-w32_fp16-8xb64-210e_coco-256x192.py: -------------------------------------------------------------------------------- 1 | _base_ = ['./td-hm_hrnet-w32_8xb64-210e_coco-256x192.py'] 2 | 3 | # fp16 settings 4 | optim_wrapper = dict( 5 | type='AmpOptimWrapper', 6 | loss_scale='dynamic', 7 | ) 8 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/body_2d_keypoint/topdown_heatmap/coco/td-hm_res50_fp16-8xb64-210e_coco-256x192.py: -------------------------------------------------------------------------------- 1 | _base_ = ['./td-hm_res50_8xb64-210e_coco-256x192.py'] 2 | 3 | # fp16 settings 4 | optim_wrapper = dict( 5 | type='AmpOptimWrapper', 6 | loss_scale='dynamic', 7 | ) 8 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/body_2d_keypoint/topdown_heatmap/coco/vgg_coco.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/body_2d_keypoint/topdown_heatmap/coco/td-hm_vgg16-bn_8xb64-210e_coco-256x192.py 3 | In Collection: SimpleBaseline2D 4 | Metadata: 5 | Architecture: 6 | - SimpleBaseline2D 7 | - VGG 8 | Training Data: COCO 9 | Name: td-hm_vgg16-bn_8xb64-210e_coco-256x192 10 | Results: 11 | - Dataset: COCO 12 | Metrics: 13 | AP: 0.699 14 | AP@0.5: 0.89 15 | AP@0.75: 0.769 16 | AR: 0.754 17 | AR@0.5: 0.927 18 | Task: Body 2D Keypoint 19 | Weights: https://download.openmmlab.com/mmpose/top_down/vgg/vgg16_bn_coco_256x192-7e7c58d6_20210517.pth 20 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/body_2d_keypoint/topdown_heatmap/crowdpose/cspnext_udp_crowdpose.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/body_2d_keypoint/topdown_heatmap/crowdpose/cspnext-m_udp_8xb64-210e_crowpose-256x192.py 3 | In Collection: UDP 4 | Metadata: 5 | Architecture: 6 | - UDP 7 | - CSPNeXt 8 | Training Data: CrowdPose 9 | Name: cspnext-m_udp_8xb64-210e_crowpose-256x192 10 | Results: 11 | - Dataset: CrowdPose 12 | Metrics: 13 | AP: 0.662 14 | AP (E): 0.759 15 | AP (H): 0.539 16 | AP (M): 0.675 17 | AP@0.5: 0.821 18 | AP@0.75: 0.723 19 | Task: Body 2D Keypoint 20 | Weights: https://download.openmmlab.com/mmpose/v1/projects/rtmposev1/cspnext-m_udp-crowdpose_pt-in1k_210e-256x192-f591079f_20230123.pth 21 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/body_2d_keypoint/topdown_heatmap/crowdpose/hrnet_crowdpose.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/body_2d_keypoint/topdown_heatmap/crowdpose/td-hm_hrnet-w32_8xb64-210e_crowdpose-256x192.py 3 | In Collection: HRNet 4 | Metadata: 5 | Architecture: 6 | - HRNet 7 | Training Data: CrowdPose 8 | Name: td-hm_hrnet-w32_8xb64-210e_crowdpose-256x192 9 | Results: 10 | - Dataset: CrowdPose 11 | Metrics: 12 | AP: 0.675 13 | AP (E): 0.77 14 | AP (H): 0.553 15 | AP (M): 0.687 16 | AP@0.5: 0.825 17 | AP@0.75: 0.729 18 | Task: Body 2D Keypoint 19 | Weights: https://download.openmmlab.com/mmpose/top_down/hrnet/hrnet_w32_crowdpose_256x192-960be101_20201227.pth 20 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/body_2d_keypoint/topdown_heatmap/mpii/cpm_mpii.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/body_2d_keypoint/topdown_heatmap/mpii/td-hm_cpm_8xb64-210e_mpii-368x368.py 3 | In Collection: CPM 4 | Metadata: 5 | Architecture: 6 | - CPM 7 | Training Data: MPII 8 | Name: td-hm_cpm_8xb64-210e_mpii-368x368 9 | Results: 10 | - Dataset: MPII 11 | Metrics: 12 | Mean: 0.876 13 | Mean@0.1: 0.285 14 | Task: Body 2D Keypoint 15 | Weights: https://download.openmmlab.com/mmpose/top_down/cpm/cpm_mpii_368x368-116e62b8_20200822.pth 16 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/body_2d_keypoint/topdown_heatmap/mpii/cspnext_udp_mpii.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/body_2d_keypoint/topdown_heatmap/mpii/cspnext-m_udp_8xb64-210e_mpii-256x256.py 3 | In Collection: UDP 4 | Metadata: 5 | Architecture: 6 | - UDP 7 | - CSPNeXt 8 | Training Data: MPII 9 | Name: cspnext-m_udp_8xb64-210e_mpii-256x256 10 | Results: 11 | - Dataset: MPII 12 | Metrics: 13 | Mean: 0.902 14 | Mean@0.1: 0.303 15 | Task: Body 2D Keypoint 16 | Weights: https://download.openmmlab.com/mmpose/v1/projects/rtmposev1/cspnext-m_udp-mpii_pt-in1k_210e-256x256-68d0402f_20230208.pth 17 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/body_2d_keypoint/topdown_heatmap/mpii/hourglass_mpii.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/body_2d_keypoint/topdown_heatmap/mpii/td-hm_hourglass52_8xb64-210e_mpii-256x256.py 3 | In Collection: Hourglass 4 | Metadata: 5 | Architecture: &id001 6 | - Hourglass 7 | Training Data: MPII 8 | Name: td-hm_hourglass52_8xb64-210e_mpii-256x256 9 | Results: 10 | - Dataset: MPII 11 | Metrics: 12 | Mean: 0.889 13 | Mean@0.1: 0.317 14 | Task: Body 2D Keypoint 15 | Weights: https://download.openmmlab.com/mmpose/top_down/hourglass/hourglass52_mpii_256x256-ae358435_20200812.pth 16 | - Config: configs/body_2d_keypoint/topdown_heatmap/mpii/td-hm_hourglass52_8xb32-210e_mpii-384x384.py 17 | In Collection: Hourglass 18 | Metadata: 19 | Architecture: *id001 20 | Training Data: MPII 21 | Name: td-hm_hourglass52_8xb32-210e_mpii-384x384 22 | Results: 23 | - Dataset: MPII 24 | Metrics: 25 | Mean: 0.894 26 | Mean@0.1: 0.367 27 | Task: Body 2D Keypoint 28 | Weights: https://download.openmmlab.com/mmpose/top_down/hourglass/hourglass52_mpii_384x384-04090bc3_20200812.pth 29 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/body_2d_keypoint/topdown_heatmap/mpii/hrnet_dark_mpii.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/body_2d_keypoint/topdown_heatmap/mpii/td-hm_hrnet-w32_dark-8xb64-210e_mpii-256x256.py 3 | In Collection: DarkPose 4 | Metadata: 5 | Architecture: &id001 6 | - HRNet 7 | - DarkPose 8 | Training Data: MPII 9 | Name: td-hm_hrnet-w32_dark-8xb64-210e_mpii-256x256 10 | Results: 11 | - Dataset: MPII 12 | Metrics: 13 | Mean: 0.904 14 | Mean@0.1: 0.354 15 | Task: Body 2D Keypoint 16 | Weights: https://download.openmmlab.com/mmpose/top_down/hrnet/hrnet_w32_mpii_256x256_dark-f1601c5b_20200927.pth 17 | - Config: configs/body_2d_keypoint/topdown_heatmap/mpii/td-hm_hrnet-w48_dark-8xb64-210e_mpii-256x256.py 18 | In Collection: DarkPose 19 | Metadata: 20 | Architecture: *id001 21 | Training Data: MPII 22 | Name: td-hm_hrnet-w48_dark-8xb64-210e_mpii-256x256 23 | Results: 24 | - Dataset: MPII 25 | Metrics: 26 | Mean: 0.905 27 | Mean@0.1: 0.36 28 | Task: Body 2D Keypoint 29 | Weights: https://download.openmmlab.com/mmpose/top_down/hrnet/hrnet_w48_mpii_256x256_dark-0decd39f_20200927.pth 30 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/body_2d_keypoint/topdown_heatmap/mpii/hrnet_mpii.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/body_2d_keypoint/topdown_heatmap/mpii/td-hm_hrnet-w32_8xb64-210e_mpii-256x256.py 3 | In Collection: HRNet 4 | Metadata: 5 | Architecture: &id001 6 | - HRNet 7 | Training Data: MPII 8 | Name: td-hm_hrnet-w32_8xb64-210e_mpii-256x256 9 | Results: 10 | - Dataset: MPII 11 | Metrics: 12 | Mean: 0.9 13 | Mean@0.1: 0.334 14 | Task: Body 2D Keypoint 15 | Weights: https://download.openmmlab.com/mmpose/top_down/hrnet/hrnet_w32_mpii_256x256-6c4f923f_20200812.pth 16 | - Config: configs/body_2d_keypoint/topdown_heatmap/mpii/td-hm_hrnet-w48_8xb64-210e_mpii-256x256.py 17 | In Collection: HRNet 18 | Metadata: 19 | Architecture: *id001 20 | Training Data: MPII 21 | Name: td-hm_hrnet-w48_8xb64-210e_mpii-256x256 22 | Results: 23 | - Dataset: MPII 24 | Metrics: 25 | Mean: 0.901 26 | Mean@0.1: 0.337 27 | Task: Body 2D Keypoint 28 | Weights: https://download.openmmlab.com/mmpose/top_down/hrnet/hrnet_w48_mpii_256x256-92cab7bd_20200812.pth 29 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/body_2d_keypoint/topdown_heatmap/mpii/litehrnet_mpii.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/body_2d_keypoint/topdown_heatmap/mpii/td-hm_litehrnet-18_8xb64-210e_mpii-256x256.py 3 | In Collection: LiteHRNet 4 | Metadata: 5 | Architecture: &id001 6 | - LiteHRNet 7 | Training Data: MPII 8 | Name: td-hm_litehrnet-18_8xb64-210e_mpii-256x256 9 | Results: 10 | - Dataset: MPII 11 | Metrics: 12 | Mean: 0.859 13 | Mean@0.1: 0.26 14 | Task: Body 2D Keypoint 15 | Weights: https://download.openmmlab.com/mmpose/top_down/litehrnet/litehrnet18_mpii_256x256-cabd7984_20210623.pth 16 | - Config: configs/body_2d_keypoint/topdown_heatmap/mpii/td-hm_litehrnet-30_8xb64-210e_mpii-256x256.py 17 | In Collection: LiteHRNet 18 | Metadata: 19 | Architecture: *id001 20 | Training Data: MPII 21 | Name: td-hm_litehrnet-30_8xb64-210e_mpii-256x256 22 | Results: 23 | - Dataset: MPII 24 | Metrics: 25 | Mean: 0.869 26 | Mean@0.1: 0.271 27 | Task: Body 2D Keypoint 28 | Weights: https://download.openmmlab.com/mmpose/top_down/litehrnet/litehrnet30_mpii_256x256-faae8bd8_20210622.pth 29 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/body_2d_keypoint/topdown_heatmap/mpii/mobilenetv2_mpii.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/body_2d_keypoint/topdown_heatmap/mpii/td-hm_mobilenetv2_8xb64-210e_mpii-256x256.py 3 | In Collection: SimpleBaseline2D 4 | Metadata: 5 | Architecture: 6 | - SimpleBaseline2D 7 | - MobilenetV2 8 | Training Data: MPII 9 | Name: td-hm_mobilenetv2_8xb64-210e_mpii-256x256 10 | Results: 11 | - Dataset: MPII 12 | Metrics: 13 | Mean: 0.854 14 | Mean@0.1: 0.234 15 | Task: Body 2D Keypoint 16 | Weights: https://download.openmmlab.com/mmpose/top_down/mobilenetv2/mobilenetv2_mpii_256x256-e068afa7_20200812.pth 17 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/body_2d_keypoint/topdown_heatmap/mpii/resnext_mpii.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/body_2d_keypoint/topdown_heatmap/mpii/td-hm_resnext152_8xb32-210e_mpii-256x256.py 3 | In Collection: SimpleBaseline2D 4 | Metadata: 5 | Architecture: 6 | - SimpleBaseline2D 7 | - ResNext 8 | Training Data: MPII 9 | Name: td-hm_resnext152_8xb32-210e_mpii-256x256 10 | Results: 11 | - Dataset: MPII 12 | Metrics: 13 | Mean: 0.887 14 | Mean@0.1: 0.294 15 | Task: Body 2D Keypoint 16 | Weights: https://download.openmmlab.com/mmpose/top_down/resnext/resnext152_mpii_256x256-df302719_20200927.pth 17 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/body_2d_keypoint/topdown_heatmap/mpii/scnet_mpii.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/body_2d_keypoint/topdown_heatmap/mpii/td-hm_scnet50_8xb64-210e_mpii-256x256.py 3 | In Collection: SimpleBaseline2D 4 | Metadata: 5 | Architecture: &id001 6 | - SimpleBaseline2D 7 | - SCNet 8 | Training Data: MPII 9 | Name: td-hm_scnet50_8xb64-210e_mpii-256x256 10 | Results: 11 | - Dataset: MPII 12 | Metrics: 13 | Mean: 0.888 14 | Mean@0.1: 0.29 15 | Task: Body 2D Keypoint 16 | Weights: https://download.openmmlab.com/mmpose/top_down/scnet/scnet50_mpii_256x256-a54b6af5_20200812.pth 17 | - Config: configs/body_2d_keypoint/topdown_heatmap/mpii/td-hm_scnet101_8xb64-210e_mpii-256x256.py 18 | In Collection: SimpleBaseline2D 19 | Metadata: 20 | Architecture: *id001 21 | Training Data: MPII 22 | Name: td-hm_scnet101_8xb64-210e_mpii-256x256 23 | Results: 24 | - Dataset: MPII 25 | Metrics: 26 | Mean: 0.887 27 | Mean@0.1: 0.293 28 | Task: Body 2D Keypoint 29 | Weights: https://download.openmmlab.com/mmpose/top_down/scnet/scnet101_mpii_256x256-b4c2d184_20200812.pth 30 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/body_2d_keypoint/topdown_heatmap/mpii/shufflenetv1_mpii.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/body_2d_keypoint/topdown_heatmap/mpii/td-hm_shufflenetv1_8xb64-210e_mpii-256x256.py 3 | In Collection: SimpleBaseline2D 4 | Metadata: 5 | Architecture: 6 | - SimpleBaseline2D 7 | - ShufflenetV1 8 | Training Data: MPII 9 | Name: td-hm_shufflenetv1_8xb64-210e_mpii-256x256 10 | Results: 11 | - Dataset: MPII 12 | Metrics: 13 | Mean: 0.824 14 | Mean@0.1: 0.195 15 | Task: Body 2D Keypoint 16 | Weights: https://download.openmmlab.com/mmpose/top_down/shufflenetv1/shufflenetv1_mpii_256x256-dcc1c896_20200925.pth 17 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/body_2d_keypoint/topdown_heatmap/mpii/shufflenetv2_mpii.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/body_2d_keypoint/topdown_heatmap/mpii/td-hm_shufflenetv2_8xb64-210e_mpii-256x256.py 3 | In Collection: SimpleBaseline2D 4 | Metadata: 5 | Architecture: 6 | - SimpleBaseline2D 7 | - ShufflenetV2 8 | Training Data: MPII 9 | Name: td-hm_shufflenetv2_8xb64-210e_mpii-256x256 10 | Results: 11 | - Dataset: MPII 12 | Metrics: 13 | Mean: 0.828 14 | Mean@0.1: 0.205 15 | Task: Body 2D Keypoint 16 | Weights: https://download.openmmlab.com/mmpose/top_down/shufflenetv2/shufflenetv2_mpii_256x256-4fb9df2d_20200925.pth 17 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/body_2d_keypoint/topdown_heatmap/posetrack18/resnet_posetrack18.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/body_2d_keypoint/topdown_heatmap/posetrack18/td-hm_res50_8xb64-20e_posetrack18-256x192.py 3 | In Collection: SimpleBaseline2D 4 | Metadata: 5 | Architecture: &id001 6 | - SimpleBaseline2D 7 | - ResNet 8 | Training Data: PoseTrack18 9 | Name: td-hm_res50_8xb64-20e_posetrack18-256x192 10 | Results: 11 | - Dataset: PoseTrack18 12 | Metrics: 13 | Ankl: 74.2 14 | Elb: 82.5 15 | Head: 86.5 16 | Hip: 80.1 17 | Knee: 78.8 18 | Shou: 87.7 19 | Total: 81.2 20 | Wri: 75.8 21 | Task: Body 2D Keypoint 22 | Weights: https://download.openmmlab.com/mmpose/top_down/resnet/res50_posetrack18_256x192-a62807c7_20201028.pth 23 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/body_2d_keypoint/topdown_regression/coco/mobilenetv2_rle_coco.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/body_2d_keypoint/topdown_regression/coco/td-reg_mobilenetv2_rle-pretrained-8xb64-210e_coco-256x192.py 3 | In Collection: RLE 4 | Metadata: 5 | Architecture: &id001 6 | - DeepPose 7 | - RLE 8 | - MobileNet 9 | Training Data: COCO 10 | Name: td-reg_mobilenetv2_rle-pretrained-8xb64-210e_coco-256x192 11 | Results: 12 | - Dataset: COCO 13 | Metrics: 14 | AP: 0.593 15 | AP@0.5: 0.836 16 | AP@0.75: 0.66 17 | AR: 0.644 18 | AR@0.5: 0.877 19 | Task: Body 2D Keypoint 20 | Weights: https://download.openmmlab.com/mmpose/v1/body_2d_keypoint/topdown_regression/coco/td-reg_mobilenetv2_rle-pretrained-8xb64-210e_coco-256x192-39b73bd5_20220922.pth 21 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/body_2d_keypoint/topdown_regression/mpii/resnet_rle_mpii.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/body_2d_keypoint/topdown_regression/mpii/td-reg_res50_rle-8xb64-210e_mpii-256x256.py 3 | In Collection: RLE 4 | Metadata: 5 | Architecture: 6 | - DeepPose 7 | - RLE 8 | - ResNet 9 | Training Data: MPII 10 | Name: td-reg_res50_rle-8xb64-210e_mpii-256x256 11 | Results: 12 | - Dataset: MPII 13 | Metrics: 14 | Mean: 0.861 15 | Mean@0.1: 0.277 16 | Task: Body 2D Keypoint 17 | Weights: https://download.openmmlab.com/mmpose/top_down/deeppose/deeppose_res50_mpii_256x256_rle-5f92a619_20220504.pth 18 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/body_3d_keypoint/README.md: -------------------------------------------------------------------------------- 1 | # Human Body 3D Pose Estimation 2 | 3 | 3D human body pose estimation aims at predicting the X, Y, Z coordinates of human body joints. Based on the camera number to capture the images or videos, existing works can be further divided into multi-view methods and single-view (monocular) methods. 4 | 5 | ## Data preparation 6 | 7 | Please follow [DATA Preparation](/docs/en/dataset_zoo/3d_body_keypoint.md) to prepare data. 8 | 9 | ## Demo 10 | 11 | Please follow [Demo](/demo/docs/en/3d_human_pose_demo.md) to run demos. 12 | 13 |
14 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/face_2d_keypoint/README.md: -------------------------------------------------------------------------------- 1 | # 2D Face Landmark Detection 2 | 3 | 2D face landmark detection (also referred to as face alignment) is defined as the task of detecting the face keypoints from an input image. 4 | 5 | Normally, the input images are cropped face images, where the face locates at the center; 6 | or the rough location (or the bounding box) of the hand is provided. 7 | 8 | ## Data preparation 9 | 10 | Please follow [DATA Preparation](/docs/en/dataset_zoo/2d_face_keypoint.md) to prepare data. 11 | 12 | ## Demo 13 | 14 | Please follow [Demo](/demo/docs/en/2d_face_demo.md) to run demos. 15 | 16 |
17 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/face_2d_keypoint/rtmpose/coco_wholebody_face/rtmpose_coco_wholebody_face.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/face_2d_keypoint/rtmpose/coco_wholebody_face/rtmpose-m_8xb32-60e_coco-wholebody-face-256x256.py 3 | In Collection: RTMPose 4 | Metadata: 5 | Architecture: 6 | - RTMPose 7 | Training Data: COCO-WholeBody-Face 8 | Name: rtmpose-m_8xb32-60e_coco-wholebody-face-256x256 9 | Results: 10 | - Dataset: COCO-WholeBody-Face 11 | Metrics: 12 | NME: 0.0466 13 | Task: Face 2D Keypoint 14 | Weights: https://download.openmmlab.com/mmpose/v1/projects/rtmposev1/rtmpose-m_simcc-coco-wholebody-face_pt-aic-coco_60e-256x256-62026ef2_20230228.pth 15 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/face_2d_keypoint/rtmpose/lapa/rtmpose_lapa.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/face_2d_keypoint/rtmpose/lapa/rtmpose-m_8xb64-120e_lapa-256x256.py 3 | In Collection: RTMPose 4 | Alias: face 5 | Metadata: 6 | Architecture: 7 | - RTMPose 8 | Training Data: LaPa 9 | Name: rtmpose-m_8xb64-120e_lapa-256x256 10 | Results: 11 | - Dataset: WFLW 12 | Metrics: 13 | NME: 1.29 14 | Task: Face 2D Keypoint 15 | Weights: https://download.openmmlab.com/mmpose/v1/projects/rtmposev1/rtmpose-m_simcc-lapa_pt-aic-coco_120e-256x256-762b1ae2_20230422.pth 16 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/face_2d_keypoint/rtmpose/wflw/rtmpose_wflw.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/face_2d_keypoint/rtmpose/wflw/rtmpose-m_8xb64-60e_wflw-256x256.py 3 | In Collection: RTMPose 4 | Alias: face 5 | Metadata: 6 | Architecture: 7 | - RTMPose 8 | Training Data: WFLW 9 | Name: rtmpose-m_8xb64-60e_wflw-256x256 10 | Results: 11 | - Dataset: WFLW 12 | Metrics: 13 | NME: 4.01 14 | Task: Face 2D Keypoint 15 | Weights: https://download.openmmlab.com/mmpose/v1/projects/rtmposev1/rtmpose-m_simcc-wflw_pt-aic-coco_60e-256x256-dc1dcdcf_20230228.pth 16 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/face_2d_keypoint/topdown_heatmap/300w/hrnetv2_300w.yml: -------------------------------------------------------------------------------- 1 | Collections: 2 | - Name: HRNetv2 3 | Paper: 4 | Title: Deep High-Resolution Representation Learning for Visual Recognition 5 | URL: https://ieeexplore.ieee.org/abstract/document/9052469/ 6 | README: https://github.com/open-mmlab/mmpose/blob/main/docs/src/papers/backbones/hrnetv2.md 7 | Models: 8 | - Config: configs/face_2d_keypoint/topdown_heatmap/300w/td-hm_hrnetv2-w18_8xb64-60e_300w-256x256.py 9 | In Collection: HRNetv2 10 | Metadata: 11 | Architecture: 12 | - HRNetv2 13 | Training Data: 300W 14 | Name: td-hm_hrnetv2-w18_8xb64-60e_300w-256x256 15 | Results: 16 | - Dataset: 300W 17 | Metrics: 18 | NME challenge: 5.64 19 | NME common: 2.92 20 | NME full: 3.45 21 | NME test: 4.1 22 | Task: Face 2D Keypoint 23 | Weights: https://download.openmmlab.com/mmpose/face/hrnetv2/hrnetv2_w18_300w_256x256-eea53406_20211019.pth 24 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/face_2d_keypoint/topdown_heatmap/aflw/hrnetv2_aflw.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/face_2d_keypoint/topdown_heatmap/aflw/td-hm_hrnetv2-w18_8xb64-60e_aflw-256x256.py 3 | In Collection: HRNetv2 4 | Metadata: 5 | Architecture: 6 | - HRNetv2 7 | Training Data: AFLW 8 | Name: td-hm_hrnetv2-w18_8xb64-60e_aflw-256x256 9 | Results: 10 | - Dataset: AFLW 11 | Metrics: 12 | NME frontal: 1.27 13 | NME full: 1.41 14 | Task: Face 2D Keypoint 15 | Weights: https://download.openmmlab.com/mmpose/face/hrnetv2/hrnetv2_w18_aflw_256x256-f2bbc62b_20210125.pth 16 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/face_2d_keypoint/topdown_heatmap/aflw/hrnetv2_dark_aflw.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/face_2d_keypoint/topdown_heatmap/aflw/td-hm_hrnetv2-w18_dark-8xb64-60e_aflw-256x256.py 3 | In Collection: DarkPose 4 | Metadata: 5 | Architecture: 6 | - HRNetv2 7 | - DarkPose 8 | Training Data: AFLW 9 | Name: td-hm_hrnetv2-w18_dark-8xb64-60e_aflw-256x256 10 | Results: 11 | - Dataset: AFLW 12 | Metrics: 13 | NME frontal: 1.19 14 | NME full: 1.34 15 | Task: Face 2D Keypoint 16 | Weights: https://download.openmmlab.com/mmpose/face/darkpose/hrnetv2_w18_aflw_256x256_dark-219606c0_20210125.pth 17 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/face_2d_keypoint/topdown_heatmap/coco_wholebody_face/hourglass_coco_wholebody_face.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/face_2d_keypoint/topdown_heatmap/coco_wholebody_face/td-hm_hourglass52_8xb32-60e_coco-wholebody-face-256x256.py 3 | In Collection: Hourglass 4 | Metadata: 5 | Architecture: 6 | - Hourglass 7 | Training Data: COCO-WholeBody-Face 8 | Name: td-hm_hourglass52_8xb32-60e_coco-wholebody-face-256x256 9 | Results: 10 | - Dataset: COCO-WholeBody-Face 11 | Metrics: 12 | NME: 0.0587 13 | Task: Face 2D Keypoint 14 | Weights: https://download.openmmlab.com/mmpose/face/hourglass/hourglass52_coco_wholebody_face_256x256-6994cf2e_20210909.pth 15 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/face_2d_keypoint/topdown_heatmap/coco_wholebody_face/hrnetv2_coco_wholebody_face.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/face_2d_keypoint/topdown_heatmap/coco_wholebody_face/td-hm_hrnetv2-w18_8xb32-60e_coco-wholebody-face-256x256.py 3 | In Collection: HRNetv2 4 | Metadata: 5 | Architecture: 6 | - HRNetv2 7 | Training Data: COCO-WholeBody-Face 8 | Name: td-hm_hrnetv2-w18_8xb32-60e_coco-wholebody-face-256x256 9 | Results: 10 | - Dataset: COCO-WholeBody-Face 11 | Metrics: 12 | NME: 0.0569 13 | Task: Face 2D Keypoint 14 | Weights: https://download.openmmlab.com/mmpose/face/hrnetv2/hrnetv2_w18_coco_wholebody_face_256x256-c1ca469b_20210909.pth 15 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/face_2d_keypoint/topdown_heatmap/coco_wholebody_face/hrnetv2_dark_coco_wholebody_face.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/face_2d_keypoint/topdown_heatmap/coco_wholebody_face/td-hm_hrnetv2-w18_dark-8xb32-60e_coco-wholebody-face-256x256.py 3 | In Collection: DarkPose 4 | Metadata: 5 | Architecture: 6 | - HRNetv2 7 | - DarkPose 8 | Training Data: COCO-WholeBody-Face 9 | Name: td-hm_hrnetv2-w18_dark-8xb32-60e_coco-wholebody-face-256x256 10 | Results: 11 | - Dataset: COCO-WholeBody-Face 12 | Metrics: 13 | NME: 0.0513 14 | Task: Face 2D Keypoint 15 | Weights: https://download.openmmlab.com/mmpose/face/darkpose/hrnetv2_w18_coco_wholebody_face_256x256_dark-3d9a334e_20210909.pth 16 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/face_2d_keypoint/topdown_heatmap/coco_wholebody_face/mobilenetv2_coco_wholebody_face.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/face_2d_keypoint/topdown_heatmap/coco_wholebody_face/td-hm_mobilenetv2_8xb32-60e_coco-wholebody-face-256x256.py 3 | In Collection: SimpleBaseline2D 4 | Metadata: 5 | Architecture: 6 | - SimpleBaseline2D 7 | - MobilenetV2 8 | Training Data: COCO-WholeBody-Face 9 | Name: td-hm_mobilenetv2_8xb32-60e_coco-wholebody-face-256x256 10 | Results: 11 | - Dataset: COCO-WholeBody-Face 12 | Metrics: 13 | NME: 0.0611 14 | Task: Face 2D Keypoint 15 | Weights: https://download.openmmlab.com/mmpose/face/mobilenetv2/mobilenetv2_coco_wholebody_face_256x256-4a3f096e_20210909.pth 16 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/face_2d_keypoint/topdown_heatmap/coco_wholebody_face/resnet_coco_wholebody_face.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/face_2d_keypoint/topdown_heatmap/coco_wholebody_face/td-hm_res50_8xb32-60e_coco-wholebody-face-256x256.py 3 | In Collection: SimpleBaseline2D 4 | Metadata: 5 | Architecture: 6 | - SimpleBaseline2D 7 | - ResNet 8 | Training Data: COCO-WholeBody-Face 9 | Name: td-hm_res50_8xb32-60e_coco-wholebody-face-256x256 10 | Results: 11 | - Dataset: COCO-WholeBody-Face 12 | Metrics: 13 | NME: 0.0582 14 | Task: Face 2D Keypoint 15 | Weights: https://download.openmmlab.com/mmpose/face/resnet/res50_coco_wholebody_face_256x256-5128edf5_20210909.pth 16 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/face_2d_keypoint/topdown_heatmap/coco_wholebody_face/scnet_coco_wholebody_face.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/face_2d_keypoint/topdown_heatmap/coco_wholebody_face/td-hm_scnet50_8xb32-60e_coco-wholebody-face-256x256.py 3 | In Collection: SimpleBaseline2D 4 | Metadata: 5 | Architecture: 6 | - SimpleBaseline2D 7 | - SCNet 8 | Training Data: COCO-WholeBody-Face 9 | Name: td-hm_scnet50_8xb32-60e_coco-wholebody-face-256x256 10 | Results: 11 | - Dataset: COCO-WholeBody-Face 12 | Metrics: 13 | NME: 0.0567 14 | Task: Face 2D Keypoint 15 | Weights: https://download.openmmlab.com/mmpose/face/scnet/scnet50_coco_wholebody_face_256x256-a0183f5f_20210909.pth 16 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/face_2d_keypoint/topdown_heatmap/cofw/hrnetv2_cofw.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/face_2d_keypoint/topdown_heatmap/cofw/td-hm_hrnetv2-w18_8xb64-60e_cofw-256x256.py 3 | In Collection: HRNetv2 4 | Metadata: 5 | Architecture: 6 | - HRNetv2 7 | Training Data: COFW 8 | Name: td-hm_hrnetv2-w18_8xb64-60e_cofw-256x256 9 | Results: 10 | - Dataset: COFW 11 | Metrics: 12 | NME: 3.48 13 | Task: Face 2D Keypoint 14 | Weights: https://download.openmmlab.com/mmpose/face/hrnetv2/hrnetv2_w18_cofw_256x256-49243ab8_20211019.pth 15 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/face_2d_keypoint/topdown_heatmap/wflw/hrnetv2_awing_wflw.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/face_2d_keypoint/topdown_heatmap/wflw/td-hm_hrnetv2-w18_awing-8xb64-60e_wflw-256x256.py 3 | In Collection: HRNetv2 4 | Metadata: 5 | Architecture: 6 | - HRNetv2 7 | - AdaptiveWingloss 8 | Training Data: WFLW 9 | Name: td-hm_hrnetv2-w18_awing-8xb64-60e_wflw-256x256 10 | Results: 11 | - Dataset: WFLW 12 | Metrics: 13 | NME blur: 4.59 14 | NME expression: 4.28 15 | NME illumination: 3.97 16 | NME makeup: 3.87 17 | NME occlusion: 4.78 18 | NME pose: 6.94 19 | NME test: 4.02 20 | Task: Face 2D Keypoint 21 | Weights: https://download.openmmlab.com/mmpose/face/hrnetv2/hrnetv2_w18_wflw_256x256_awing-5af5055c_20211212.pth 22 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/face_2d_keypoint/topdown_heatmap/wflw/hrnetv2_dark_wflw.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/face_2d_keypoint/topdown_heatmap/wflw/td-hm_hrnetv2-w18_dark-8xb64-60e_wflw-256x256.py 3 | In Collection: DarkPose 4 | Metadata: 5 | Architecture: 6 | - HRNetv2 7 | - DarkPose 8 | Training Data: WFLW 9 | Name: td-hm_hrnetv2-w18_dark-8xb64-60e_wflw-256x256 10 | Results: 11 | - Dataset: WFLW 12 | Metrics: 13 | NME blur: 4.56 14 | NME expression: 4.29 15 | NME illumination: 3.96 16 | NME makeup: 3.89 17 | NME occlusion: 4.78 18 | NME pose: 6.98 19 | NME test: 3.98 20 | Task: Face 2D Keypoint 21 | Weights: https://download.openmmlab.com/mmpose/face/darkpose/hrnetv2_w18_wflw_256x256_dark-3f8e0c2c_20210125.pth 22 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/face_2d_keypoint/topdown_heatmap/wflw/hrnetv2_wflw.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/face_2d_keypoint/topdown_heatmap/wflw/td-hm_hrnetv2-w18_8xb64-60e_wflw-256x256.py 3 | In Collection: HRNetv2 4 | Metadata: 5 | Architecture: 6 | - HRNetv2 7 | Training Data: WFLW 8 | Name: td-hm_hrnetv2-w18_8xb64-60e_wflw-256x256 9 | Results: 10 | - Dataset: WFLW 11 | Metrics: 12 | NME blur: 4.58 13 | NME expression: 4.33 14 | NME illumination: 3.99 15 | NME makeup: 3.94 16 | NME occlusion: 4.83 17 | NME pose: 6.97 18 | NME test: 4.06 19 | Task: Face 2D Keypoint 20 | Weights: https://download.openmmlab.com/mmpose/face/hrnetv2/hrnetv2_w18_wflw_256x256-2bf032a6_20210125.pth 21 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/face_2d_keypoint/topdown_regression/wflw/resnet_softwingloss_wflw.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/face_2d_keypoint/topdown_regression/wflw/td-reg_res50_softwingloss_8xb64-210e_wflw-256x256.py 3 | In Collection: ResNet 4 | Metadata: 5 | Architecture: 6 | - DeepPose 7 | - ResNet 8 | - SoftWingloss 9 | Training Data: WFLW 10 | Name: td-reg_res50_softwingloss_8xb64-210e_wflw-256x256 11 | Results: 12 | - Dataset: WFLW 13 | Metrics: 14 | NME: 4.44 15 | Task: Face 2D Keypoint 16 | Weights: https://download.openmmlab.com/mmpose/face/deeppose/deeppose_res50_wflw_256x256_softwingloss-4d34f22a_20211212.pth 17 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/face_2d_keypoint/topdown_regression/wflw/resnet_wflw.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/face_2d_keypoint/topdown_regression/wflw/td-reg_res50_8xb64-210e_wflw-256x256.py 3 | In Collection: ResNet 4 | Metadata: 5 | Architecture: 6 | - DeepPose 7 | - ResNet 8 | Training Data: WFLW 9 | Name: td-reg_res50_8x64e-210e_wflw-256x256 10 | Results: 11 | - Dataset: WFLW 12 | Metrics: 13 | NME: 4.88 14 | Task: Face 2D Keypoint 15 | Weights: https://download.openmmlab.com/mmpose/face/deeppose/deeppose_res50_wflw_256x256-92d0ba7f_20210303.pth 16 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/face_2d_keypoint/topdown_regression/wflw/resnet_wingloss_wflw.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/face_2d_keypoint/topdown_regression/wflw/td-reg_res50_wingloss_8xb64-210e_wflw-256x256.py 3 | In Collection: ResNet 4 | Metadata: 5 | Architecture: 6 | - DeepPose 7 | - ResNet 8 | - WingLoss 9 | Training Data: WFLW 10 | Name: td-reg_res50_wingloss_8xb64-210e_wflw-256x256 11 | Results: 12 | - Dataset: WFLW 13 | Metrics: 14 | NME: 4.67 15 | Task: Face 2D Keypoint 16 | Weights: https://download.openmmlab.com/mmpose/face/deeppose/deeppose_res50_wflw_256x256_wingloss-f82a5e53_20210303.pth 17 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/fashion_2d_keypoint/README.md: -------------------------------------------------------------------------------- 1 | # 2D Fashion Landmark Detection 2 | 3 | 2D fashion landmark detection (also referred to as fashion alignment) aims to detect the key-point located at the functional region of clothes, for example the neckline and the cuff. 4 | 5 | ## Data preparation 6 | 7 | Please follow [DATA Preparation](/docs/en/dataset_zoo/2d_fashion_landmark.md) to prepare data. 8 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/hand_2d_keypoint/README.md: -------------------------------------------------------------------------------- 1 | # 2D Hand Pose Estimation 2 | 3 | 2D hand pose estimation is defined as the task of detecting the poses (or keypoints) of the hand from an input image. 4 | 5 | Normally, the input images are cropped hand images, where the hand locates at the center; 6 | or the rough location (or the bounding box) of the hand is provided. 7 | 8 | ## Data preparation 9 | 10 | Please follow [DATA Preparation](/docs/en/dataset_zoo/2d_hand_keypoint.md) to prepare data. 11 | 12 | ## Demo 13 | 14 | Please follow [Demo](/demo/docs/en/2d_hand_demo.md) to run demos. 15 | 16 |
17 | 18 |
19 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/hand_2d_keypoint/rtmpose/coco_wholebody_hand/rtmpose_coco_wholebody_hand.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/hand_2d_keypoint/rtmpose/coco_wholebody_hand/rtmpose-m_8xb32-210e_coco-wholebody-hand-256x256.py 3 | In Collection: RTMPose 4 | Alias: hand 5 | Metadata: 6 | Architecture: 7 | - RTMPose 8 | Training Data: COCO-WholeBody-Hand 9 | Name: rtmpose-m_8xb32-210e_coco-wholebody-hand-256x256 10 | Results: 11 | - Dataset: COCO-WholeBody-Hand 12 | Metrics: 13 | AUC: 0.815 14 | EPE: 4.51 15 | PCK@0.2: 0.837 16 | Task: Hand 2D Keypoint 17 | Weights: https://download.openmmlab.com/mmpose/v1/projects/rtmposev1/rtmpose-m_simcc-coco-wholebody-hand_pt-aic-coco_210e-256x256-99477206_20230228.pth 18 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/hand_2d_keypoint/rtmpose/hand5/rtmpose_hand5.yml: -------------------------------------------------------------------------------- 1 | Collections: 2 | - Name: RTMPose 3 | Paper: 4 | Title: "RTMPose: Real-Time Multi-Person Pose Estimation based on MMPose" 5 | URL: https://arxiv.org/abs/2303.07399 6 | README: https://github.com/open-mmlab/mmpose/blob/main/projects/rtmpose/README.md 7 | Models: 8 | - Config: configs/hand_2d_keypoint/rtmpose/hand5/rtmpose-m_8xb256-210e_hand5-256x256.py 9 | In Collection: RTMPose 10 | Metadata: 11 | Architecture: &id001 12 | - RTMPose 13 | Training Data: &id002 14 | - COCO-Wholebody-Hand 15 | - OneHand10K 16 | - FreiHand2d 17 | - RHD2d 18 | - Halpe 19 | Name: rtmpose-m_8xb256-210e_hand5-256x256 20 | Results: 21 | - Dataset: Hand5 22 | Metrics: 23 | PCK@0.2: 0.964 24 | AUC: 0.839 25 | EPE: 5.06 26 | Task: Hand 2D Keypoint 27 | Weights: https://download.openmmlab.com/mmpose/v1/projects/rtmposev1/rtmpose-m_simcc-hand5_pt-aic-coco_210e-256x256-74fb594_20230320.pth 28 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/hand_2d_keypoint/topdown_heatmap/coco_wholebody_hand/hourglass_coco_wholebody_hand.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/hand_2d_keypoint/topdown_heatmap/coco_wholebody_hand/td-hm_hourglass52_8xb32-210e_coco-wholebody-hand-256x256.py 3 | In Collection: Hourglass 4 | Metadata: 5 | Architecture: 6 | - Hourglass 7 | Training Data: COCO-WholeBody-Hand 8 | Name: td-hm_hourglass52_8xb32-210e_coco-wholebody-hand-256x256 9 | Results: 10 | - Dataset: COCO-WholeBody-Hand 11 | Metrics: 12 | AUC: 0.835 13 | EPE: 4.54 14 | PCK@0.2: 0.804 15 | Task: Hand 2D Keypoint 16 | Weights: https://download.openmmlab.com/mmpose/hand/hourglass/hourglass52_coco_wholebody_hand_256x256-7b05c6db_20210909.pth 17 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/hand_2d_keypoint/topdown_heatmap/coco_wholebody_hand/hrnetv2_coco_wholebody_hand.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/hand_2d_keypoint/topdown_heatmap/coco_wholebody_hand/td-hm_hrnetv2-w18_8xb32-210e_coco-wholebody-hand-256x256.py 3 | In Collection: HRNetv2 4 | Metadata: 5 | Architecture: 6 | - HRNetv2 7 | Training Data: COCO-WholeBody-Hand 8 | Name: td-hm_hrnetv2-w18_8xb32-210e_coco-wholebody-hand-256x256 9 | Results: 10 | - Dataset: COCO-WholeBody-Hand 11 | Metrics: 12 | AUC: 0.84 13 | EPE: 4.39 14 | PCK@0.2: 0.813 15 | Task: Hand 2D Keypoint 16 | Weights: https://download.openmmlab.com/mmpose/hand/hrnetv2/hrnetv2_w18_coco_wholebody_hand_256x256-1c028db7_20210908.pth 17 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/hand_2d_keypoint/topdown_heatmap/coco_wholebody_hand/hrnetv2_dark_coco_wholebody_hand.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/hand_2d_keypoint/topdown_heatmap/coco_wholebody_hand/td-hm_hrnetv2-w18_dark-8xb32-210e_coco-wholebody-hand-256x256.py 3 | In Collection: DarkPose 4 | Metadata: 5 | Architecture: 6 | - HRNetv2 7 | - DarkPose 8 | Training Data: COCO-WholeBody-Hand 9 | Name: td-hm_hrnetv2-w18_dark-8xb32-210e_coco-wholebody-hand-256x256 10 | Results: 11 | - Dataset: COCO-WholeBody-Hand 12 | Metrics: 13 | AUC: 0.84 14 | EPE: 4.37 15 | PCK@0.2: 0.814 16 | Task: Hand 2D Keypoint 17 | Weights: https://download.openmmlab.com/mmpose/hand/dark/hrnetv2_w18_coco_wholebody_hand_256x256_dark-a9228c9c_20210908.pth 18 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/hand_2d_keypoint/topdown_heatmap/coco_wholebody_hand/litehrnet_coco_wholebody_hand.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/hand_2d_keypoint/topdown_heatmap/coco_wholebody_hand/td-hm_litehrnet-w18_8xb32-210e_coco-wholebody-hand-256x256.py 3 | In Collection: LiteHRNet 4 | Metadata: 5 | Architecture: 6 | - LiteHRNet 7 | Training Data: COCO-WholeBody-Hand 8 | Name: td-hm_litehrnet-w18_8xb32-210e_coco-wholebody-hand-256x256 9 | Results: 10 | - Dataset: COCO-WholeBody-Hand 11 | Metrics: 12 | AUC: 0.83 13 | EPE: 4.77 14 | PCK@0.2: 0.795 15 | Task: Hand 2D Keypoint 16 | Weights: https://download.openmmlab.com/mmpose/hand/litehrnet/litehrnet_w18_coco_wholebody_hand_256x256-d6945e6a_20210908.pth 17 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/hand_2d_keypoint/topdown_heatmap/coco_wholebody_hand/mobilenetv2_coco_wholebody_hand.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/hand_2d_keypoint/topdown_heatmap/coco_wholebody_hand/td-hm_mobilenetv2_8xb32-210e_coco-wholebody-hand-256x256.py 3 | In Collection: SimpleBaseline2D 4 | Metadata: 5 | Architecture: 6 | - SimpleBaseline2D 7 | - MobilenetV2 8 | Training Data: COCO-WholeBody-Hand 9 | Name: td-hm_mobilenetv2_8xb32-210e_coco-wholebody-hand-256x256 10 | Results: 11 | - Dataset: COCO-WholeBody-Hand 12 | Metrics: 13 | AUC: 0.829 14 | EPE: 4.77 15 | PCK@0.2: 0.795 16 | Task: Hand 2D Keypoint 17 | Weights: https://download.openmmlab.com/mmpose/hand/mobilenetv2/mobilenetv2_coco_wholebody_hand_256x256-06b8c877_20210909.pth 18 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/hand_2d_keypoint/topdown_heatmap/coco_wholebody_hand/resnet_coco_wholebody_hand.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/hand_2d_keypoint/topdown_heatmap/coco_wholebody_hand/td-hm_res50_8xb32-210e_coco-wholebody-hand-256x256.py 3 | In Collection: SimpleBaseline2D 4 | Metadata: 5 | Architecture: 6 | - SimpleBaseline2D 7 | - ResNet 8 | Training Data: COCO-WholeBody-Hand 9 | Name: td-hm_res50_8xb32-210e_coco-wholebody-hand-256x256 10 | Results: 11 | - Dataset: COCO-WholeBody-Hand 12 | Metrics: 13 | AUC: 0.833 14 | EPE: 4.64 15 | PCK@0.2: 0.8 16 | Task: Hand 2D Keypoint 17 | Weights: https://download.openmmlab.com/mmpose/hand/resnet/res50_coco_wholebody_hand_256x256-8dbc750c_20210908.pth 18 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/hand_2d_keypoint/topdown_heatmap/coco_wholebody_hand/scnet_coco_wholebody_hand.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/hand_2d_keypoint/topdown_heatmap/coco_wholebody_hand/td-hm_scnet50_8xb32-210e_coco-wholebody-hand-256x256.py 3 | In Collection: SimpleBaseline2D 4 | Metadata: 5 | Architecture: 6 | - SCNet 7 | Training Data: COCO-WholeBody-Hand 8 | Name: td-hm_scnet50_8xb32-210e_coco-wholebody-hand-256x256 9 | Results: 10 | - Dataset: COCO-WholeBody-Hand 11 | Metrics: 12 | AUC: 0.834 13 | EPE: 4.55 14 | PCK@0.2: 0.803 15 | Task: Hand 2D Keypoint 16 | Weights: https://download.openmmlab.com/mmpose/hand/scnet/scnet50_coco_wholebody_hand_256x256-e73414c7_20210909.pth 17 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/hand_2d_keypoint/topdown_heatmap/freihand2d/resnet_freihand2d.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/hand_2d_keypoint/topdown_heatmap/freihand2d/td-hm_res50_8xb64-100e_freihand2d-224x224.py 3 | In Collection: SimpleBaseline2D 4 | Metadata: 5 | Architecture: 6 | - SimpleBaseline2D 7 | - ResNet 8 | Training Data: FreiHand 9 | Name: td-hm_res50_8xb64-100e_freihand2d-224x224 10 | Results: 11 | - Dataset: FreiHand 12 | Metrics: 13 | AUC: 0.868 14 | EPE: 3.27 15 | PCK@0.2: 0.999 16 | Task: Hand 2D Keypoint 17 | Weights: https://download.openmmlab.com/mmpose/hand/resnet/res50_freihand_224x224-ff0799bc_20200914.pth 18 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/hand_2d_keypoint/topdown_heatmap/onehand10k/hrnetv2_dark_onehand10k.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/hand_2d_keypoint/topdown_heatmap/onehand10k/td-hm_hrnetv2-w18_dark-8xb64-210e_onehand10k-256x256.py 3 | In Collection: DarkPose 4 | Metadata: 5 | Architecture: 6 | - HRNetv2 7 | - DarkPose 8 | Training Data: OneHand10K 9 | Name: td-hm_hrnetv2-w18_dark-8xb64-210e_onehand10k-256x256 10 | Results: 11 | - Dataset: OneHand10K 12 | Metrics: 13 | AUC: 0.572 14 | EPE: 23.96 15 | PCK@0.2: 0.99 16 | Task: Hand 2D Keypoint 17 | Weights: https://download.openmmlab.com/mmpose/hand/dark/hrnetv2_w18_onehand10k_256x256_dark-a2f80c64_20210330.pth 18 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/hand_2d_keypoint/topdown_heatmap/onehand10k/hrnetv2_onehand10k.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/hand_2d_keypoint/topdown_heatmap/onehand10k/td-hm_hrnetv2-w18_8xb64-210e_onehand10k-256x256.py 3 | In Collection: HRNetv2 4 | Metadata: 5 | Architecture: 6 | - HRNetv2 7 | Training Data: OneHand10K 8 | Name: td-hm_hrnetv2-w18_8xb64-210e_onehand10k-256x256 9 | Results: 10 | - Dataset: OneHand10K 11 | Metrics: 12 | AUC: 0.567 13 | EPE: 24.26 14 | PCK@0.2: 0.99 15 | Task: Hand 2D Keypoint 16 | Weights: https://download.openmmlab.com/mmpose/hand/hrnetv2/hrnetv2_w18_onehand10k_256x256-30bc9c6b_20210330.pth 17 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/hand_2d_keypoint/topdown_heatmap/onehand10k/hrnetv2_udp_onehand10k.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/hand_2d_keypoint/topdown_heatmap/onehand10k/td-hm_hrnetv2-w18_udp-8xb64-210e_onehand10k-256x256.py 3 | In Collection: UDP 4 | Metadata: 5 | Architecture: 6 | - HRNetv2 7 | - UDP 8 | Training Data: OneHand10K 9 | Name: td-hm_hrnetv2-w18_udp-8xb64-210e_onehand10k-256x256 10 | Results: 11 | - Dataset: OneHand10K 12 | Metrics: 13 | AUC: 0.571 14 | EPE: 23.88 15 | PCK@0.2: 0.99 16 | Task: Hand 2D Keypoint 17 | Weights: https://download.openmmlab.com/mmpose/hand/udp/hrnetv2_w18_onehand10k_256x256_udp-0d1b515d_20210330.pth 18 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/hand_2d_keypoint/topdown_heatmap/onehand10k/mobilenetv2_onehand10k.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/hand_2d_keypoint/topdown_heatmap/onehand10k/td-hm_mobilenetv2_8xb64-210e_onehand10k-256x256.py 3 | In Collection: SimpleBaseline2D 4 | Metadata: 5 | Architecture: 6 | - SimpleBaseline2D 7 | - MobilenetV2 8 | Training Data: OneHand10K 9 | Name: td-hm_mobilenetv2_8xb64-210e_onehand10k-256x256 10 | Results: 11 | - Dataset: OneHand10K 12 | Metrics: 13 | AUC: 0.537 14 | EPE: 28.56 15 | PCK@0.2: 0.986 16 | Task: Hand 2D Keypoint 17 | Weights: https://download.openmmlab.com/mmpose/hand/mobilenetv2/mobilenetv2_onehand10k_256x256-f3a3d90e_20210330.pth 18 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/hand_2d_keypoint/topdown_heatmap/onehand10k/resnet_onehand10k.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/hand_2d_keypoint/topdown_heatmap/onehand10k/td-hm_res50_8xb32-210e_onehand10k-256x256.py 3 | In Collection: SimpleBaseline2D 4 | Metadata: 5 | Architecture: 6 | - SimpleBaseline2D 7 | - ResNet 8 | Training Data: OneHand10K 9 | Name: td-hm_res50_8xb32-210e_onehand10k-256x256 10 | Results: 11 | - Dataset: OneHand10K 12 | Metrics: 13 | AUC: 0.555 14 | EPE: 25.16 15 | PCK@0.2: 0.989 16 | Task: Hand 2D Keypoint 17 | Weights: https://download.openmmlab.com/mmpose/hand/resnet/res50_onehand10k_256x256-739c8639_20210330.pth 18 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/hand_2d_keypoint/topdown_heatmap/rhd2d/hrnetv2_dark_rhd2d.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/hand_2d_keypoint/topdown_heatmap/rhd2d/td-hm_hrnetv2-w18_dark-8xb64-210e_rhd2d-256x256.py 3 | In Collection: DarkPose 4 | Metadata: 5 | Architecture: 6 | - HRNetv2 7 | - DarkPose 8 | Training Data: RHD 9 | Name: td-hm_hrnetv2-w18_dark-8xb64-210e_rhd2d-256x256 10 | Results: 11 | - Dataset: RHD 12 | Metrics: 13 | AUC: 0.903 14 | EPE: 2.18 15 | PCK@0.2: 0.992 16 | Task: Hand 2D Keypoint 17 | Weights: https://download.openmmlab.com/mmpose/hand/dark/hrnetv2_w18_rhd2d_256x256_dark-4df3a347_20210330.pth 18 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/hand_2d_keypoint/topdown_heatmap/rhd2d/hrnetv2_rhd2d.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/hand_2d_keypoint/topdown_heatmap/rhd2d/td-hm_hrnetv2-w18_8xb64-210e_rhd2d-256x256.py 3 | In Collection: HRNetv2 4 | Metadata: 5 | Architecture: 6 | - HRNetv2 7 | Training Data: RHD 8 | Name: td-hm_hrnetv2-w18_8xb64-210e_rhd2d-256x256 9 | Results: 10 | - Dataset: RHD 11 | Metrics: 12 | AUC: 0.902 13 | EPE: 2.21 14 | PCK@0.2: 0.992 15 | Task: Hand 2D Keypoint 16 | Weights: https://download.openmmlab.com/mmpose/hand/hrnetv2/hrnetv2_w18_rhd2d_256x256-95b20dd8_20210330.pth 17 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/hand_2d_keypoint/topdown_heatmap/rhd2d/hrnetv2_udp_rhd2d.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/hand_2d_keypoint/topdown_heatmap/rhd2d/td-hm_hrnetv2-w18_udp-8xb64-210e_rhd2d-256x256.py 3 | In Collection: UDP 4 | Metadata: 5 | Architecture: 6 | - HRNetv2 7 | - UDP 8 | Training Data: RHD 9 | Name: td-hm_hrnetv2-w18_udp-8xb64-210e_rhd2d-256x256 10 | Results: 11 | - Dataset: RHD 12 | Metrics: 13 | AUC: 0.902 14 | EPE: 2.19 15 | PCKh@0.7: 0.992 16 | Task: Hand 2D Keypoint 17 | Weights: https://download.openmmlab.com/mmpose/hand/udp/hrnetv2_w18_rhd2d_256x256_udp-63ba6007_20210330.pth 18 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/hand_2d_keypoint/topdown_heatmap/rhd2d/mobilenetv2_rhd2d.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/hand_2d_keypoint/topdown_heatmap/rhd2d/td-hm_mobilenetv2_8xb64-210e_rhd2d-256x256.py 3 | In Collection: SimpleBaseline2D 4 | Metadata: 5 | Architecture: 6 | - SimpleBaseline2D 7 | - MobilenetV2 8 | Training Data: RHD 9 | Name: td-hm_mobilenetv2_8xb64-210e_rhd2d-256x256 10 | Results: 11 | - Dataset: RHD 12 | Metrics: 13 | AUC: 0.883 14 | EPE: 2.79 15 | PCK@0.2: 0.985 16 | Task: Hand 2D Keypoint 17 | Weights: https://download.openmmlab.com/mmpose/hand/mobilenetv2/mobilenetv2_rhd2d_256x256-85fa02db_20210330.pth 18 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/hand_2d_keypoint/topdown_heatmap/rhd2d/resnet_rhd2d.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/hand_2d_keypoint/topdown_heatmap/rhd2d/td-hm_res50_8xb64-210e_rhd2d-256x256.py 3 | In Collection: SimpleBaseline2D 4 | Metadata: 5 | Architecture: 6 | - SimpleBaseline2D 7 | - ResNet 8 | Training Data: RHD 9 | Name: td-hm_res50_8xb64-210e_rhd2d-256x256 10 | Results: 11 | - Dataset: RHD 12 | Metrics: 13 | AUC: 0.898 14 | EPE: 2.32 15 | PCK@0.2: 0.991 16 | Task: Hand 2D Keypoint 17 | Weights: https://download.openmmlab.com/mmpose/hand/resnet/res50_rhd2d_256x256-5dc7e4cc_20210330.pth 18 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/hand_2d_keypoint/topdown_regression/onehand10k/resnet_onehand10k.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/hand_2d_keypoint/topdown_regression/onehand10k/td-reg_res50_8xb64-210e_onehand10k-256x256.py 3 | In Collection: DeepPose 4 | Metadata: 5 | Architecture: 6 | - DeepPose 7 | - ResNet 8 | Training Data: OneHand10K 9 | Name: td-reg_res50_8xb64-210e_onehand10k-256x256 10 | Results: 11 | - Dataset: OneHand10K 12 | Metrics: 13 | AUC: 0.485 14 | EPE: 34.21 15 | PCK@0.2: 0.99 16 | Task: Hand 2D Keypoint 17 | Weights: https://download.openmmlab.com/mmpose/hand/deeppose/deeppose_res50_onehand10k_256x256-cbddf43a_20210330.pth 18 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/hand_2d_keypoint/topdown_regression/rhd2d/resnet_rhd2d.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/hand_2d_keypoint/topdown_regression/rhd2d/td-reg_res50_8xb64-210e_rhd2d-256x256.py 3 | In Collection: DeepPose 4 | Metadata: 5 | Architecture: 6 | - DeepPose 7 | - ResNet 8 | Training Data: RHD 9 | Name: td-reg_res50_8xb64-210e_rhd2d-256x256 10 | Results: 11 | - Dataset: RHD 12 | Metrics: 13 | AUC: 0.865 14 | EPE: 3.32 15 | PCK@0.2: 0.988 16 | Task: Hand 2D Keypoint 17 | Weights: https://download.openmmlab.com/mmpose/hand/deeppose/deeppose_res50_rhd2d_256x256-37f1c4d3_20210330.pth 18 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/hand_3d_keypoint/README.md: -------------------------------------------------------------------------------- 1 | # 3D Hand Pose Estimation 2 | 3 | 3D hand pose estimation is defined as the task of detecting the poses (or keypoints) of the hand from an input image. 4 | 5 | ## Data preparation 6 | 7 | Please follow [DATA Preparation](/docs/en/dataset_zoo/3d_hand_keypoint.md) to prepare data. 8 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/hand_gesture/README.md: -------------------------------------------------------------------------------- 1 | # Gesture Recognition 2 | 3 | Gesture recognition aims to recognize the hand gestures in the video, such as thumbs up. 4 | 5 | ## Data preparation 6 | 7 | Please follow [DATA Preparation](/docs/en/dataset_zoo/2d_hand_gesture.md) to prepare data. 8 | 9 | ## Demo 10 | 11 | Please follow [Demo](/demo/docs/en/gesture_recognition_demo.md) to run the demo. 12 | 13 | 14 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/wholebody_2d_keypoint/README.md: -------------------------------------------------------------------------------- 1 | # 2D Human Whole-Body Pose Estimation 2 | 3 | 2D human whole-body pose estimation aims to localize dense landmarks on the entire human body including face, hands, body, and feet. 4 | 5 | Existing approaches can be categorized into top-down and bottom-up approaches. 6 | 7 | Top-down methods divide the task into two stages: human detection and whole-body pose estimation. They perform human detection first, followed by single-person whole-body pose estimation given human bounding boxes. 8 | 9 | Bottom-up approaches (e.g. AE) first detect all the whole-body keypoints and then group/associate them into person instances. 10 | 11 | ## Data preparation 12 | 13 | Please follow [DATA Preparation](/docs/en/dataset_zoo/2d_wholebody_keypoint.md) to prepare data. 14 | 15 | ## Demo 16 | 17 | Please follow [Demo](/demo/docs/en/2d_wholebody_pose_demo.md) to run demos. 18 | 19 |
20 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/configs/wholebody_2d_keypoint/topdown_heatmap/coco-wholebody/cspnext_udp_coco-wholebody.yml: -------------------------------------------------------------------------------- 1 | Models: 2 | - Config: configs/wholebody_2d_keypoint/topdown_heatmap/coco-wholebody/cspnext-m_udp_8xb64-210e_coco-wholebody-256x192.py 3 | In Collection: UDP 4 | Metadata: 5 | Architecture: &id001 6 | - UDP 7 | - CSPNeXt 8 | Training Data: COCO-WholeBody 9 | Name: cspnext-m_udp_8xb64-210e_coco-wholebody-256x192 10 | Results: 11 | - Dataset: COCO-WholeBody 12 | Metrics: 13 | Body AP: 0.687 14 | Body AR: 0.735 15 | Face AP: 0.697 16 | Face AR: 0.755 17 | Foot AP: 0.680 18 | Foot AR: 0.763 19 | Hand AP: 0.46 20 | Hand AR: 0.567 21 | Whole AP: 0.567 22 | Whole AR: 0.641 23 | Task: Wholebody 2D Keypoint 24 | Weights: https://download.openmmlab.com/mmpose/v1/projects/rtmposev1/cspnext-m_udp-coco-wholebody_pt-in1k_210e-256x192-320fa258_20230123.pth 25 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/demo/resources/demo.mp4: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/5k5000/CLdetection2023/d1a01536ad892134c4dd728c87dc9ac1d87b8e11/mmpose_package/mmpose/demo/resources/demo.mp4 -------------------------------------------------------------------------------- /mmpose_package/mmpose/demo/resources/demo_coco.gif: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/5k5000/CLdetection2023/d1a01536ad892134c4dd728c87dc9ac1d87b8e11/mmpose_package/mmpose/demo/resources/demo_coco.gif -------------------------------------------------------------------------------- /mmpose_package/mmpose/demo/resources/sunglasses.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/5k5000/CLdetection2023/d1a01536ad892134c4dd728c87dc9ac1d87b8e11/mmpose_package/mmpose/demo/resources/sunglasses.jpg -------------------------------------------------------------------------------- /mmpose_package/mmpose/demo/webcam_cfg/test_camera.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | executor_cfg = dict( 3 | name='Test Webcam', 4 | camera_id=0, 5 | camera_max_fps=30, 6 | nodes=[ 7 | dict( 8 | type='MonitorNode', 9 | name='monitor', 10 | enable_key='m', 11 | enable=False, 12 | input_buffer='_frame_', 13 | output_buffer='display'), 14 | # 'RecorderNode': 15 | # This node save the output video into a file. 16 | dict( 17 | type='RecorderNode', 18 | name='recorder', 19 | out_video_file='webcam_api_output.mp4', 20 | input_buffer='display', 21 | output_buffer='_display_') 22 | ]) 23 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docker/Dockerfile: -------------------------------------------------------------------------------- 1 | ARG PYTORCH="1.8.1" 2 | ARG CUDA="10.2" 3 | ARG CUDNN="7" 4 | 5 | FROM pytorch/pytorch:${PYTORCH}-cuda${CUDA}-cudnn${CUDNN}-devel 6 | 7 | ENV TORCH_CUDA_ARCH_LIST="6.0 6.1 7.0+PTX" 8 | ENV TORCH_NVCC_FLAGS="-Xfatbin -compress-all" 9 | ENV CMAKE_PREFIX_PATH="$(dirname $(which conda))/../" 10 | 11 | # To fix GPG key error when running apt-get update 12 | RUN apt-key adv --fetch-keys https://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/3bf863cc.pub 13 | RUN apt-key adv --fetch-keys https://developer.download.nvidia.com/compute/machine-learning/repos/ubuntu1804/x86_64/7fa2af80.pub 14 | 15 | RUN apt-get update && apt-get install -y git ninja-build libglib2.0-0 libsm6 libxrender-dev libxext6 libgl1-mesa-glx\ 16 | && apt-get clean \ 17 | && rm -rf /var/lib/apt/lists/* 18 | 19 | # Install xtcocotools 20 | RUN pip install cython 21 | RUN pip install xtcocotools 22 | 23 | # Install MMEngine and MMCV 24 | RUN pip install openmim 25 | RUN mim install mmengine "mmcv>=2.0.0" 26 | 27 | # Install MMPose 28 | RUN conda clean --all 29 | RUN git clone https://github.com/open-mmlab/mmpose.git /mmpose 30 | WORKDIR /mmpose 31 | RUN git checkout main 32 | ENV FORCE_CUDA="1" 33 | RUN pip install -r requirements/build.txt 34 | RUN pip install --no-cache-dir -e . 35 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docker/serve/config.properties: -------------------------------------------------------------------------------- 1 | inference_address=http://0.0.0.0:8080 2 | management_address=http://0.0.0.0:8081 3 | metrics_address=http://0.0.0.0:8082 4 | model_store=/home/model-server/model-store 5 | load_models=all 6 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docker/serve/entrypoint.sh: -------------------------------------------------------------------------------- 1 | #!/bin/bash 2 | set -e 3 | 4 | if [[ "$1" = "serve" ]]; then 5 | shift 1 6 | torchserve --start --ts-config /home/model-server/config.properties 7 | else 8 | eval "$@" 9 | fi 10 | 11 | # prevent docker exit 12 | tail -f /dev/null 13 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/en/Makefile: -------------------------------------------------------------------------------- 1 | # Minimal makefile for Sphinx documentation 2 | # 3 | 4 | # You can set these variables from the command line, and also 5 | # from the environment for the first two. 6 | SPHINXOPTS ?= 7 | SPHINXBUILD ?= sphinx-build 8 | SOURCEDIR = . 9 | BUILDDIR = _build 10 | 11 | # Put it first so that "make" without argument is like "make help". 12 | help: 13 | @$(SPHINXBUILD) -M help "$(SOURCEDIR)" "$(BUILDDIR)" $(SPHINXOPTS) $(O) 14 | 15 | .PHONY: help Makefile 16 | 17 | # Catch-all target: route all unknown targets to Sphinx using the new 18 | # "make mode" option. $(O) is meant as a shortcut for $(SPHINXOPTS). 19 | %: Makefile 20 | @$(SPHINXBUILD) -M $@ "$(SOURCEDIR)" "$(BUILDDIR)" $(SPHINXOPTS) $(O) 21 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/en/_static/css/readthedocs.css: -------------------------------------------------------------------------------- 1 | .header-logo { 2 | background-image: url("../images/mmpose-logo.png"); 3 | background-size: 120px 50px; 4 | height: 50px; 5 | width: 120px; 6 | } 7 | 8 | table.autosummary td { 9 | width: 35% 10 | } 11 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/en/_static/images/mmpose-logo.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/5k5000/CLdetection2023/d1a01536ad892134c4dd728c87dc9ac1d87b8e11/mmpose_package/mmpose/docs/en/_static/images/mmpose-logo.png -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/en/advanced_guides/customize_logging.md: -------------------------------------------------------------------------------- 1 | # Customize Logging 2 | 3 | Coming soon. 4 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/en/advanced_guides/customize_optimizer.md: -------------------------------------------------------------------------------- 1 | # Customize Optimizer and Scheduler 2 | 3 | Coming soon. 4 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/en/advanced_guides/customize_transforms.md: -------------------------------------------------------------------------------- 1 | # Customize Data Transformation and Augmentation 2 | 3 | Coming soon. 4 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/en/advanced_guides/dataflow.md: -------------------------------------------------------------------------------- 1 | # Dataflow in MMPose 2 | 3 | Coming soon. 4 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/en/advanced_guides/how_to_deploy.md: -------------------------------------------------------------------------------- 1 | # How to Deploy MMPose Models 2 | 3 | Coming soon. 4 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/en/advanced_guides/implement_new_models.md: -------------------------------------------------------------------------------- 1 | # Implement New Models 2 | 3 | Coming soon. 4 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/en/make.bat: -------------------------------------------------------------------------------- 1 | @ECHO OFF 2 | 3 | pushd %~dp0 4 | 5 | REM Command file for Sphinx documentation 6 | 7 | if "%SPHINXBUILD%" == "" ( 8 | set SPHINXBUILD=sphinx-build 9 | ) 10 | set SOURCEDIR=. 11 | set BUILDDIR=_build 12 | 13 | if "%1" == "" goto help 14 | 15 | %SPHINXBUILD% >NUL 2>NUL 16 | if errorlevel 9009 ( 17 | echo. 18 | echo.The 'sphinx-build' command was not found. Make sure you have Sphinx 19 | echo.installed, then set the SPHINXBUILD environment variable to point 20 | echo.to the full path of the 'sphinx-build' executable. Alternatively you 21 | echo.may add the Sphinx directory to PATH. 22 | echo. 23 | echo.If you don't have Sphinx installed, grab it from 24 | echo.http://sphinx-doc.org/ 25 | exit /b 1 26 | ) 27 | 28 | %SPHINXBUILD% -M %1 %SOURCEDIR% %BUILDDIR% %SPHINXOPTS% %O% 29 | goto end 30 | 31 | :help 32 | %SPHINXBUILD% -M help %SOURCEDIR% %BUILDDIR% %SPHINXOPTS% %O% 33 | 34 | :end 35 | popd 36 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/en/notes/ecosystem.md: -------------------------------------------------------------------------------- 1 | # Ecosystem 2 | 3 | Coming soon. 4 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/en/notes/pytorch_2.md: -------------------------------------------------------------------------------- 1 | # PyTorch 2.0 Compatibility and Benchmarks 2 | 3 | Coming soon. 4 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/en/switch_language.md: -------------------------------------------------------------------------------- 1 | ## English 2 | 3 | ## 简体中文 4 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/src/papers/algorithms/ipr.md: -------------------------------------------------------------------------------- 1 | # Integral Human Pose Regression 2 | 3 | 4 | 5 |
6 | IPR (ECCV'2018) 7 | 8 | ```bibtex 9 | @inproceedings{sun2018integral, 10 | title={Integral human pose regression}, 11 | author={Sun, Xiao and Xiao, Bin and Wei, Fangyin and Liang, Shuang and Wei, Yichen}, 12 | booktitle={Proceedings of the European conference on computer vision (ECCV)}, 13 | pages={529--545}, 14 | year={2018} 15 | } 16 | ``` 17 | 18 |
19 | 20 | ## Abstract 21 | 22 | 23 | 24 | State-of-the-art human pose estimation methods are based on heat map representation. In spite of the good performance, the representation has a few issues in nature, such as not differentiable and quantization error. This work shows that a simple integral operation relates and unifies the heat map representation and joint regression, thus avoiding the above issues. It is differentiable, efficient, and compatible with any heat map based methods. Its effectiveness is convincingly validated via comprehensive ablation experiments under various settings, specifically on 3D pose estimation, for the first time. 25 | 26 | 27 | 28 |
29 | 30 |
31 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/src/papers/algorithms/simplebaseline2d.md: -------------------------------------------------------------------------------- 1 | # Simple baselines for human pose estimation and tracking 2 | 3 | 4 | 5 |
6 | SimpleBaseline2D (ECCV'2018) 7 | 8 | ```bibtex 9 | @inproceedings{xiao2018simple, 10 | title={Simple baselines for human pose estimation and tracking}, 11 | author={Xiao, Bin and Wu, Haiping and Wei, Yichen}, 12 | booktitle={Proceedings of the European conference on computer vision (ECCV)}, 13 | pages={466--481}, 14 | year={2018} 15 | } 16 | ``` 17 | 18 |
19 | 20 | ## Abstract 21 | 22 | 23 | 24 | There has been significant progress on pose estimation and increasing interests on pose tracking in recent years. At the same time, the overall algorithm and system complexity increases as well, making the algorithm analysis and comparison more difficult. This work provides simple and effective baseline methods. They are helpful for inspiring and evaluating new ideas for the field. State-of-the-art results are achieved on challenging benchmarks. 25 | 26 | 27 | 28 |
29 | 30 |
31 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/src/papers/datasets/300w.md: -------------------------------------------------------------------------------- 1 | # 300 faces in-the-wild challenge: Database and results 2 | 3 | 4 | 5 |
6 | 300W (IMAVIS'2016) 7 | 8 | ```bibtex 9 | @article{sagonas2016300, 10 | title={300 faces in-the-wild challenge: Database and results}, 11 | author={Sagonas, Christos and Antonakos, Epameinondas and Tzimiropoulos, Georgios and Zafeiriou, Stefanos and Pantic, Maja}, 12 | journal={Image and vision computing}, 13 | volume={47}, 14 | pages={3--18}, 15 | year={2016}, 16 | publisher={Elsevier} 17 | } 18 | ``` 19 | 20 |
21 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/src/papers/datasets/aflw.md: -------------------------------------------------------------------------------- 1 | # Annotated facial landmarks in the wild: A large-scale, real-world database for facial landmark localization 2 | 3 | 4 | 5 |
6 | AFLW (ICCVW'2011) 7 | 8 | ```bibtex 9 | @inproceedings{koestinger2011annotated, 10 | title={Annotated facial landmarks in the wild: A large-scale, real-world database for facial landmark localization}, 11 | author={Koestinger, Martin and Wohlhart, Paul and Roth, Peter M and Bischof, Horst}, 12 | booktitle={2011 IEEE international conference on computer vision workshops (ICCV workshops)}, 13 | pages={2144--2151}, 14 | year={2011}, 15 | organization={IEEE} 16 | } 17 | ``` 18 | 19 |
20 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/src/papers/datasets/aic.md: -------------------------------------------------------------------------------- 1 | # Ai challenger: A large-scale dataset for going deeper in image understanding 2 | 3 | 4 | 5 |
6 | AI Challenger (ArXiv'2017) 7 | 8 | ```bibtex 9 | @article{wu2017ai, 10 | title={Ai challenger: A large-scale dataset for going deeper in image understanding}, 11 | author={Wu, Jiahong and Zheng, He and Zhao, Bo and Li, Yixin and Yan, Baoming and Liang, Rui and Wang, Wenjia and Zhou, Shipei and Lin, Guosen and Fu, Yanwei and others}, 12 | journal={arXiv preprint arXiv:1711.06475}, 13 | year={2017} 14 | } 15 | ``` 16 | 17 |
18 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/src/papers/datasets/animalpose.md: -------------------------------------------------------------------------------- 1 | # Cross-Domain Adaptation for Animal Pose Estimation 2 | 3 | 4 | 5 |
6 | Animal-Pose (ICCV'2019) 7 | 8 | ```bibtex 9 | @InProceedings{Cao_2019_ICCV, 10 | author = {Cao, Jinkun and Tang, Hongyang and Fang, Hao-Shu and Shen, Xiaoyong and Lu, Cewu and Tai, Yu-Wing}, 11 | title = {Cross-Domain Adaptation for Animal Pose Estimation}, 12 | booktitle = {The IEEE International Conference on Computer Vision (ICCV)}, 13 | month = {October}, 14 | year = {2019} 15 | } 16 | ``` 17 | 18 |
19 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/src/papers/datasets/ap10k.md: -------------------------------------------------------------------------------- 1 | # AP-10K: A Benchmark for Animal Pose Estimation in the Wild 2 | 3 | 4 | 5 |
6 | AP-10K (NeurIPS'2021) 7 | 8 | ```bibtex 9 | @misc{yu2021ap10k, 10 | title={AP-10K: A Benchmark for Animal Pose Estimation in the Wild}, 11 | author={Hang Yu and Yufei Xu and Jing Zhang and Wei Zhao and Ziyu Guan and Dacheng Tao}, 12 | year={2021}, 13 | eprint={2108.12617}, 14 | archivePrefix={arXiv}, 15 | primaryClass={cs.CV} 16 | } 17 | ``` 18 | 19 |
20 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/src/papers/datasets/atrw.md: -------------------------------------------------------------------------------- 1 | # ATRW: A Benchmark for Amur Tiger Re-identification in the Wild 2 | 3 | 4 | 5 |
6 | ATRW (ACM MM'2020) 7 | 8 | ```bibtex 9 | @inproceedings{li2020atrw, 10 | title={ATRW: A Benchmark for Amur Tiger Re-identification in the Wild}, 11 | author={Li, Shuyuan and Li, Jianguo and Tang, Hanlin and Qian, Rui and Lin, Weiyao}, 12 | booktitle={Proceedings of the 28th ACM International Conference on Multimedia}, 13 | pages={2590--2598}, 14 | year={2020} 15 | } 16 | ``` 17 | 18 |
19 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/src/papers/datasets/campus_and_shelf.md: -------------------------------------------------------------------------------- 1 | # 3D Pictorial Structures for Multiple Human Pose Estimation 2 | 3 | 4 | 5 |
6 | Campus and Shelf (CVPR'2014) 7 | 8 | ```bibtex 9 | @inproceedings {belagian14multi, 10 | title = {{3D} Pictorial Structures for Multiple Human Pose Estimation}, 11 | author = {Belagiannis, Vasileios and Amin, Sikandar and Andriluka, Mykhaylo and Schiele, Bernt and Navab 12 | Nassir and Ilic, Slobodan}, 13 | booktitle = {IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR)}, 14 | year = {2014}, 15 | month = {June}, 16 | organization={IEEE} 17 | } 18 | ``` 19 | 20 |
21 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/src/papers/datasets/coco.md: -------------------------------------------------------------------------------- 1 | # Microsoft coco: Common objects in context 2 | 3 | 4 | 5 |
6 | COCO (ECCV'2014) 7 | 8 | ```bibtex 9 | @inproceedings{lin2014microsoft, 10 | title={Microsoft coco: Common objects in context}, 11 | author={Lin, Tsung-Yi and Maire, Michael and Belongie, Serge and Hays, James and Perona, Pietro and Ramanan, Deva and Doll{\'a}r, Piotr and Zitnick, C Lawrence}, 12 | booktitle={European conference on computer vision}, 13 | pages={740--755}, 14 | year={2014}, 15 | organization={Springer} 16 | } 17 | ``` 18 | 19 |
20 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/src/papers/datasets/coco_wholebody.md: -------------------------------------------------------------------------------- 1 | # Whole-Body Human Pose Estimation in the Wild 2 | 3 | 4 | 5 |
6 | COCO-WholeBody (ECCV'2020) 7 | 8 | ```bibtex 9 | @inproceedings{jin2020whole, 10 | title={Whole-Body Human Pose Estimation in the Wild}, 11 | author={Jin, Sheng and Xu, Lumin and Xu, Jin and Wang, Can and Liu, Wentao and Qian, Chen and Ouyang, Wanli and Luo, Ping}, 12 | booktitle={Proceedings of the European Conference on Computer Vision (ECCV)}, 13 | year={2020} 14 | } 15 | ``` 16 | 17 |
18 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/src/papers/datasets/coco_wholebody_face.md: -------------------------------------------------------------------------------- 1 | # Whole-Body Human Pose Estimation in the Wild 2 | 3 | 4 | 5 |
6 | COCO-WholeBody-Face (ECCV'2020) 7 | 8 | ```bibtex 9 | @inproceedings{jin2020whole, 10 | title={Whole-Body Human Pose Estimation in the Wild}, 11 | author={Jin, Sheng and Xu, Lumin and Xu, Jin and Wang, Can and Liu, Wentao and Qian, Chen and Ouyang, Wanli and Luo, Ping}, 12 | booktitle={Proceedings of the European Conference on Computer Vision (ECCV)}, 13 | year={2020} 14 | } 15 | ``` 16 | 17 |
18 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/src/papers/datasets/coco_wholebody_hand.md: -------------------------------------------------------------------------------- 1 | # Whole-Body Human Pose Estimation in the Wild 2 | 3 | 4 | 5 |
6 | COCO-WholeBody-Hand (ECCV'2020) 7 | 8 | ```bibtex 9 | @inproceedings{jin2020whole, 10 | title={Whole-Body Human Pose Estimation in the Wild}, 11 | author={Jin, Sheng and Xu, Lumin and Xu, Jin and Wang, Can and Liu, Wentao and Qian, Chen and Ouyang, Wanli and Luo, Ping}, 12 | booktitle={Proceedings of the European Conference on Computer Vision (ECCV)}, 13 | year={2020} 14 | } 15 | ``` 16 | 17 |
18 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/src/papers/datasets/cofw.md: -------------------------------------------------------------------------------- 1 | # Robust face landmark estimation under occlusion 2 | 3 | 4 | 5 |
6 | COFW (ICCV'2013) 7 | 8 | ```bibtex 9 | @inproceedings{burgos2013robust, 10 | title={Robust face landmark estimation under occlusion}, 11 | author={Burgos-Artizzu, Xavier P and Perona, Pietro and Doll{\'a}r, Piotr}, 12 | booktitle={Proceedings of the IEEE international conference on computer vision}, 13 | pages={1513--1520}, 14 | year={2013} 15 | } 16 | ``` 17 | 18 |
19 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/src/papers/datasets/crowdpose.md: -------------------------------------------------------------------------------- 1 | # CrowdPose: Efficient Crowded Scenes Pose Estimation and A New Benchmark 2 | 3 | 4 | 5 |
6 | CrowdPose (CVPR'2019) 7 | 8 | ```bibtex 9 | @article{li2018crowdpose, 10 | title={CrowdPose: Efficient Crowded Scenes Pose Estimation and A New Benchmark}, 11 | author={Li, Jiefeng and Wang, Can and Zhu, Hao and Mao, Yihuan and Fang, Hao-Shu and Lu, Cewu}, 12 | journal={arXiv preprint arXiv:1812.00324}, 13 | year={2018} 14 | } 15 | ``` 16 | 17 |
18 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/src/papers/datasets/deepfashion.md: -------------------------------------------------------------------------------- 1 | # DeepFashion: Powering Robust Clothes Recognition and Retrieval with Rich Annotations 2 | 3 | 4 | 5 |
6 | DeepFashion (CVPR'2016) 7 | 8 | ```bibtex 9 | @inproceedings{liuLQWTcvpr16DeepFashion, 10 | author = {Liu, Ziwei and Luo, Ping and Qiu, Shi and Wang, Xiaogang and Tang, Xiaoou}, 11 | title = {DeepFashion: Powering Robust Clothes Recognition and Retrieval with Rich Annotations}, 12 | booktitle = {Proceedings of IEEE Conference on Computer Vision and Pattern Recognition (CVPR)}, 13 | month = {June}, 14 | year = {2016} 15 | } 16 | ``` 17 | 18 |
19 | 20 | 21 | 22 |
23 | DeepFashion (ECCV'2016) 24 | 25 | ```bibtex 26 | @inproceedings{liuYLWTeccv16FashionLandmark, 27 | author = {Liu, Ziwei and Yan, Sijie and Luo, Ping and Wang, Xiaogang and Tang, Xiaoou}, 28 | title = {Fashion Landmark Detection in the Wild}, 29 | booktitle = {European Conference on Computer Vision (ECCV)}, 30 | month = {October}, 31 | year = {2016} 32 | } 33 | ``` 34 | 35 |
36 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/src/papers/datasets/fly.md: -------------------------------------------------------------------------------- 1 | # Fast animal pose estimation using deep neural networks 2 | 3 | 4 | 5 |
6 | Vinegar Fly (Nature Methods'2019) 7 | 8 | ```bibtex 9 | @article{pereira2019fast, 10 | title={Fast animal pose estimation using deep neural networks}, 11 | author={Pereira, Talmo D and Aldarondo, Diego E and Willmore, Lindsay and Kislin, Mikhail and Wang, Samuel S-H and Murthy, Mala and Shaevitz, Joshua W}, 12 | journal={Nature methods}, 13 | volume={16}, 14 | number={1}, 15 | pages={117--125}, 16 | year={2019}, 17 | publisher={Nature Publishing Group} 18 | } 19 | ``` 20 | 21 |
22 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/src/papers/datasets/freihand.md: -------------------------------------------------------------------------------- 1 | # Freihand: A dataset for markerless capture of hand pose and shape from single rgb images 2 | 3 | 4 | 5 |
6 | FreiHand (ICCV'2019) 7 | 8 | ```bibtex 9 | @inproceedings{zimmermann2019freihand, 10 | title={Freihand: A dataset for markerless capture of hand pose and shape from single rgb images}, 11 | author={Zimmermann, Christian and Ceylan, Duygu and Yang, Jimei and Russell, Bryan and Argus, Max and Brox, Thomas}, 12 | booktitle={Proceedings of the IEEE International Conference on Computer Vision}, 13 | pages={813--822}, 14 | year={2019} 15 | } 16 | ``` 17 | 18 |
19 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/src/papers/datasets/h36m.md: -------------------------------------------------------------------------------- 1 | # Human3.6M: Large Scale Datasets and Predictive Methods for 3D Human Sensing in Natural Environments 2 | 3 | 4 | 5 |
6 | Human3.6M (TPAMI'2014) 7 | 8 | ```bibtex 9 | @article{h36m_pami, 10 | author = {Ionescu, Catalin and Papava, Dragos and Olaru, Vlad and Sminchisescu, Cristian}, 11 | title = {Human3.6M: Large Scale Datasets and Predictive Methods for 3D Human Sensing in Natural Environments}, 12 | journal = {IEEE Transactions on Pattern Analysis and Machine Intelligence}, 13 | publisher = {IEEE Computer Society}, 14 | volume = {36}, 15 | number = {7}, 16 | pages = {1325-1339}, 17 | month = {jul}, 18 | year = {2014} 19 | } 20 | ``` 21 | 22 |
23 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/src/papers/datasets/halpe.md: -------------------------------------------------------------------------------- 1 | # PaStaNet: Toward Human Activity Knowledge Engine 2 | 3 | 4 | 5 |
6 | Halpe (CVPR'2020) 7 | 8 | ```bibtex 9 | @inproceedings{li2020pastanet, 10 | title={PaStaNet: Toward Human Activity Knowledge Engine}, 11 | author={Li, Yong-Lu and Xu, Liang and Liu, Xinpeng and Huang, Xijie and Xu, Yue and Wang, Shiyi and Fang, Hao-Shu and Ma, Ze and Chen, Mingyang and Lu, Cewu}, 12 | booktitle={CVPR}, 13 | year={2020} 14 | } 15 | ``` 16 | 17 |
18 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/src/papers/datasets/horse10.md: -------------------------------------------------------------------------------- 1 | # Pretraining boosts out-of-domain robustness for pose estimation 2 | 3 | 4 | 5 |
6 | Horse-10 (WACV'2021) 7 | 8 | ```bibtex 9 | @inproceedings{mathis2021pretraining, 10 | title={Pretraining boosts out-of-domain robustness for pose estimation}, 11 | author={Mathis, Alexander and Biasi, Thomas and Schneider, Steffen and Yuksekgonul, Mert and Rogers, Byron and Bethge, Matthias and Mathis, Mackenzie W}, 12 | booktitle={Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision}, 13 | pages={1859--1868}, 14 | year={2021} 15 | } 16 | ``` 17 | 18 |
19 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/src/papers/datasets/interhand.md: -------------------------------------------------------------------------------- 1 | # InterHand2.6M: A dataset and baseline for 3D interacting hand pose estimation from a single RGB image 2 | 3 | 4 | 5 |
6 | InterHand2.6M (ECCV'2020) 7 | 8 | ```bibtex 9 | @article{moon2020interhand2, 10 | title={InterHand2.6M: A dataset and baseline for 3D interacting hand pose estimation from a single RGB image}, 11 | author={Moon, Gyeongsik and Yu, Shoou-I and Wen, He and Shiratori, Takaaki and Lee, Kyoung Mu}, 12 | journal={arXiv preprint arXiv:2008.09309}, 13 | year={2020}, 14 | publisher={Springer} 15 | } 16 | ``` 17 | 18 |
19 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/src/papers/datasets/jhmdb.md: -------------------------------------------------------------------------------- 1 | # Towards understanding action recognition 2 | 3 | 4 | 5 |
6 | JHMDB (ICCV'2013) 7 | 8 | ```bibtex 9 | @inproceedings{Jhuang:ICCV:2013, 10 | title = {Towards understanding action recognition}, 11 | author = {H. Jhuang and J. Gall and S. Zuffi and C. Schmid and M. J. Black}, 12 | booktitle = {International Conf. on Computer Vision (ICCV)}, 13 | month = Dec, 14 | pages = {3192-3199}, 15 | year = {2013} 16 | } 17 | ``` 18 | 19 |
20 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/src/papers/datasets/locust.md: -------------------------------------------------------------------------------- 1 | # DeepPoseKit, a software toolkit for fast and robust animal pose estimation using deep learning 2 | 3 | 4 | 5 |
6 | Desert Locust (Elife'2019) 7 | 8 | ```bibtex 9 | @article{graving2019deepposekit, 10 | title={DeepPoseKit, a software toolkit for fast and robust animal pose estimation using deep learning}, 11 | author={Graving, Jacob M and Chae, Daniel and Naik, Hemal and Li, Liang and Koger, Benjamin and Costelloe, Blair R and Couzin, Iain D}, 12 | journal={Elife}, 13 | volume={8}, 14 | pages={e47994}, 15 | year={2019}, 16 | publisher={eLife Sciences Publications Limited} 17 | } 18 | ``` 19 | 20 |
21 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/src/papers/datasets/macaque.md: -------------------------------------------------------------------------------- 1 | # MacaquePose: A novel ‘in the wild’macaque monkey pose dataset for markerless motion capture 2 | 3 | 4 | 5 |
6 | MacaquePose (bioRxiv'2020) 7 | 8 | ```bibtex 9 | @article{labuguen2020macaquepose, 10 | title={MacaquePose: A novel ‘in the wild’macaque monkey pose dataset for markerless motion capture}, 11 | author={Labuguen, Rollyn and Matsumoto, Jumpei and Negrete, Salvador and Nishimaru, Hiroshi and Nishijo, Hisao and Takada, Masahiko and Go, Yasuhiro and Inoue, Ken-ichi and Shibata, Tomohiro}, 12 | journal={bioRxiv}, 13 | year={2020}, 14 | publisher={Cold Spring Harbor Laboratory} 15 | } 16 | ``` 17 | 18 |
19 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/src/papers/datasets/mhp.md: -------------------------------------------------------------------------------- 1 | # Understanding humans in crowded scenes: Deep nested adversarial learning and a new benchmark for multi-human parsing 2 | 3 | 4 | 5 |
6 | MHP (ACM MM'2018) 7 | 8 | ```bibtex 9 | @inproceedings{zhao2018understanding, 10 | title={Understanding humans in crowded scenes: Deep nested adversarial learning and a new benchmark for multi-human parsing}, 11 | author={Zhao, Jian and Li, Jianshu and Cheng, Yu and Sim, Terence and Yan, Shuicheng and Feng, Jiashi}, 12 | booktitle={Proceedings of the 26th ACM international conference on Multimedia}, 13 | pages={792--800}, 14 | year={2018} 15 | } 16 | ``` 17 | 18 |
19 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/src/papers/datasets/mpi_inf_3dhp.md: -------------------------------------------------------------------------------- 1 | # Monocular 3D Human Pose Estimation In The Wild Using Improved CNN Supervision 2 | 3 | 4 | 5 |
6 | MPI-INF-3DHP (3DV'2017) 7 | 8 | ```bibtex 9 | @inproceedings{mono-3dhp2017, 10 | author = {Mehta, Dushyant and Rhodin, Helge and Casas, Dan and Fua, Pascal and Sotnychenko, Oleksandr and Xu, Weipeng and Theobalt, Christian}, 11 | title = {Monocular 3D Human Pose Estimation In The Wild Using Improved CNN Supervision}, 12 | booktitle = {3D Vision (3DV), 2017 Fifth International Conference on}, 13 | url = {http://gvv.mpi-inf.mpg.de/3dhp_dataset}, 14 | year = {2017}, 15 | organization={IEEE}, 16 | doi={10.1109/3dv.2017.00064}, 17 | } 18 | ``` 19 | 20 |
21 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/src/papers/datasets/mpii.md: -------------------------------------------------------------------------------- 1 | # 2D Human Pose Estimation: New Benchmark and State of the Art Analysis 2 | 3 | 4 | 5 |
6 | MPII (CVPR'2014) 7 | 8 | ```bibtex 9 | @inproceedings{andriluka14cvpr, 10 | author = {Mykhaylo Andriluka and Leonid Pishchulin and Peter Gehler and Schiele, Bernt}, 11 | title = {2D Human Pose Estimation: New Benchmark and State of the Art Analysis}, 12 | booktitle = {IEEE Conference on Computer Vision and Pattern Recognition (CVPR)}, 13 | year = {2014}, 14 | month = {June} 15 | } 16 | ``` 17 | 18 |
19 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/src/papers/datasets/mpii_trb.md: -------------------------------------------------------------------------------- 1 | # TRB: A Novel Triplet Representation for Understanding 2D Human Body 2 | 3 | 4 | 5 |
6 | MPII-TRB (ICCV'2019) 7 | 8 | ```bibtex 9 | @inproceedings{duan2019trb, 10 | title={TRB: A Novel Triplet Representation for Understanding 2D Human Body}, 11 | author={Duan, Haodong and Lin, Kwan-Yee and Jin, Sheng and Liu, Wentao and Qian, Chen and Ouyang, Wanli}, 12 | booktitle={Proceedings of the IEEE International Conference on Computer Vision}, 13 | pages={9479--9488}, 14 | year={2019} 15 | } 16 | ``` 17 | 18 |
19 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/src/papers/datasets/ochuman.md: -------------------------------------------------------------------------------- 1 | # Pose2seg: Detection free human instance segmentation 2 | 3 | 4 | 5 |
6 | OCHuman (CVPR'2019) 7 | 8 | ```bibtex 9 | @inproceedings{zhang2019pose2seg, 10 | title={Pose2seg: Detection free human instance segmentation}, 11 | author={Zhang, Song-Hai and Li, Ruilong and Dong, Xin and Rosin, Paul and Cai, Zixi and Han, Xi and Yang, Dingcheng and Huang, Haozhi and Hu, Shi-Min}, 12 | booktitle={Proceedings of the IEEE conference on computer vision and pattern recognition}, 13 | pages={889--898}, 14 | year={2019} 15 | } 16 | ``` 17 | 18 |
19 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/src/papers/datasets/onehand10k.md: -------------------------------------------------------------------------------- 1 | # Mask-pose cascaded cnn for 2d hand pose estimation from single color image 2 | 3 | 4 | 5 |
6 | OneHand10K (TCSVT'2019) 7 | 8 | ```bibtex 9 | @article{wang2018mask, 10 | title={Mask-pose cascaded cnn for 2d hand pose estimation from single color image}, 11 | author={Wang, Yangang and Peng, Cong and Liu, Yebin}, 12 | journal={IEEE Transactions on Circuits and Systems for Video Technology}, 13 | volume={29}, 14 | number={11}, 15 | pages={3258--3268}, 16 | year={2018}, 17 | publisher={IEEE} 18 | } 19 | ``` 20 | 21 |
22 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/src/papers/datasets/panoptic.md: -------------------------------------------------------------------------------- 1 | # Hand keypoint detection in single images using multiview bootstrapping 2 | 3 | 4 | 5 |
6 | CMU Panoptic HandDB (CVPR'2017) 7 | 8 | ```bibtex 9 | @inproceedings{simon2017hand, 10 | title={Hand keypoint detection in single images using multiview bootstrapping}, 11 | author={Simon, Tomas and Joo, Hanbyul and Matthews, Iain and Sheikh, Yaser}, 12 | booktitle={Proceedings of the IEEE conference on Computer Vision and Pattern Recognition}, 13 | pages={1145--1153}, 14 | year={2017} 15 | } 16 | ``` 17 | 18 |
19 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/src/papers/datasets/panoptic_body3d.md: -------------------------------------------------------------------------------- 1 | # Panoptic Studio: A Massively Multiview System for Social Motion Capture 2 | 3 | 4 | 5 |
6 | CMU Panoptic (ICCV'2015) 7 | 8 | ```bibtex 9 | @Article = {joo_iccv_2015, 10 | author = {Hanbyul Joo, Hao Liu, Lei Tan, Lin Gui, Bart Nabbe, Iain Matthews, Takeo Kanade, Shohei Nobuhara, and Yaser Sheikh}, 11 | title = {Panoptic Studio: A Massively Multiview System for Social Motion Capture}, 12 | booktitle = {ICCV}, 13 | year = {2015} 14 | } 15 | ``` 16 | 17 |
18 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/src/papers/datasets/posetrack18.md: -------------------------------------------------------------------------------- 1 | # Posetrack: A benchmark for human pose estimation and tracking 2 | 3 | 4 | 5 |
6 | PoseTrack18 (CVPR'2018) 7 | 8 | ```bibtex 9 | @inproceedings{andriluka2018posetrack, 10 | title={Posetrack: A benchmark for human pose estimation and tracking}, 11 | author={Andriluka, Mykhaylo and Iqbal, Umar and Insafutdinov, Eldar and Pishchulin, Leonid and Milan, Anton and Gall, Juergen and Schiele, Bernt}, 12 | booktitle={Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition}, 13 | pages={5167--5176}, 14 | year={2018} 15 | } 16 | ``` 17 | 18 |
19 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/src/papers/datasets/rhd.md: -------------------------------------------------------------------------------- 1 | # Learning to Estimate 3D Hand Pose from Single RGB Images 2 | 3 | 4 | 5 |
6 | RHD (ICCV'2017) 7 | 8 | ```bibtex 9 | @TechReport{zb2017hand, 10 | author={Christian Zimmermann and Thomas Brox}, 11 | title={Learning to Estimate 3D Hand Pose from Single RGB Images}, 12 | institution={arXiv:1705.01389}, 13 | year={2017}, 14 | note="https://arxiv.org/abs/1705.01389", 15 | url="https://lmb.informatik.uni-freiburg.de/projects/hand3d/" 16 | } 17 | ``` 18 | 19 |
20 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/src/papers/datasets/wflw.md: -------------------------------------------------------------------------------- 1 | # Look at boundary: A boundary-aware face alignment algorithm 2 | 3 | 4 | 5 |
6 | WFLW (CVPR'2018) 7 | 8 | ```bibtex 9 | @inproceedings{wu2018look, 10 | title={Look at boundary: A boundary-aware face alignment algorithm}, 11 | author={Wu, Wayne and Qian, Chen and Yang, Shuo and Wang, Quan and Cai, Yici and Zhou, Qiang}, 12 | booktitle={Proceedings of the IEEE conference on computer vision and pattern recognition}, 13 | pages={2129--2138}, 14 | year={2018} 15 | } 16 | ``` 17 | 18 |
19 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/src/papers/datasets/zebra.md: -------------------------------------------------------------------------------- 1 | # DeepPoseKit, a software toolkit for fast and robust animal pose estimation using deep learning 2 | 3 | 4 | 5 |
6 | Grévy’s Zebra (Elife'2019) 7 | 8 | ```bibtex 9 | @article{graving2019deepposekit, 10 | title={DeepPoseKit, a software toolkit for fast and robust animal pose estimation using deep learning}, 11 | author={Graving, Jacob M and Chae, Daniel and Naik, Hemal and Li, Liang and Koger, Benjamin and Costelloe, Blair R and Couzin, Iain D}, 12 | journal={Elife}, 13 | volume={8}, 14 | pages={e47994}, 15 | year={2019}, 16 | publisher={eLife Sciences Publications Limited} 17 | } 18 | ``` 19 | 20 |
21 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/src/papers/techniques/albumentations.md: -------------------------------------------------------------------------------- 1 | # Albumentations: fast and flexible image augmentations 2 | 3 | 4 | 5 |
6 | Albumentations (Information'2020) 7 | 8 | ```bibtex 9 | @article{buslaev2020albumentations, 10 | title={Albumentations: fast and flexible image augmentations}, 11 | author={Buslaev, Alexander and Iglovikov, Vladimir I and Khvedchenya, Eugene and Parinov, Alex and Druzhinin, Mikhail and Kalinin, Alexandr A}, 12 | journal={Information}, 13 | volume={11}, 14 | number={2}, 15 | pages={125}, 16 | year={2020}, 17 | publisher={Multidisciplinary Digital Publishing Institute} 18 | } 19 | ``` 20 | 21 |
22 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/src/papers/techniques/fp16.md: -------------------------------------------------------------------------------- 1 | # Mixed Precision Training 2 | 3 | 4 | 5 |
6 | FP16 (ArXiv'2017) 7 | 8 | ```bibtex 9 | @article{micikevicius2017mixed, 10 | title={Mixed precision training}, 11 | author={Micikevicius, Paulius and Narang, Sharan and Alben, Jonah and Diamos, Gregory and Elsen, Erich and Garcia, David and Ginsburg, Boris and Houston, Michael and Kuchaiev, Oleksii and Venkatesh, Ganesh and others}, 12 | journal={arXiv preprint arXiv:1710.03740}, 13 | year={2017} 14 | } 15 | ``` 16 | 17 |
18 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/zh_cn/Makefile: -------------------------------------------------------------------------------- 1 | # Minimal makefile for Sphinx documentation 2 | # 3 | 4 | # You can set these variables from the command line, and also 5 | # from the environment for the first two. 6 | SPHINXOPTS ?= 7 | SPHINXBUILD ?= sphinx-build 8 | SOURCEDIR = . 9 | BUILDDIR = _build 10 | 11 | # Put it first so that "make" without argument is like "make help". 12 | help: 13 | @$(SPHINXBUILD) -M help "$(SOURCEDIR)" "$(BUILDDIR)" $(SPHINXOPTS) $(O) 14 | 15 | .PHONY: help Makefile 16 | 17 | # Catch-all target: route all unknown targets to Sphinx using the new 18 | # "make mode" option. $(O) is meant as a shortcut for $(SPHINXOPTS). 19 | %: Makefile 20 | @$(SPHINXBUILD) -M $@ "$(SOURCEDIR)" "$(BUILDDIR)" $(SPHINXOPTS) $(O) 21 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/zh_cn/_static/css/readthedocs.css: -------------------------------------------------------------------------------- 1 | .header-logo { 2 | background-image: url("../images/mmpose-logo.png"); 3 | background-size: 120px 50px; 4 | height: 50px; 5 | width: 120px; 6 | } 7 | 8 | table.autosummary td { 9 | width: 35% 10 | } 11 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/zh_cn/_static/images/mmpose-logo.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/5k5000/CLdetection2023/d1a01536ad892134c4dd728c87dc9ac1d87b8e11/mmpose_package/mmpose/docs/zh_cn/_static/images/mmpose-logo.png -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/zh_cn/advanced_guides/customize_datasets.md: -------------------------------------------------------------------------------- 1 | # Customize Datasets 2 | 3 | Coming soon. 4 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/zh_cn/advanced_guides/customize_logging.md: -------------------------------------------------------------------------------- 1 | # Customize Logging 2 | 3 | Coming soon. 4 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/zh_cn/advanced_guides/customize_optimizer.md: -------------------------------------------------------------------------------- 1 | # Customize Optimizer and Scheduler 2 | 3 | Coming soon. 4 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/zh_cn/advanced_guides/customize_transforms.md: -------------------------------------------------------------------------------- 1 | # Customize Data Transformation and Augmentation 2 | 3 | Coming soon. 4 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/zh_cn/advanced_guides/dataflow.md: -------------------------------------------------------------------------------- 1 | # Dataflow in MMPose 2 | 3 | Coming soon. 4 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/zh_cn/advanced_guides/how_to_deploy.md: -------------------------------------------------------------------------------- 1 | # How to Deploy MMPose Models 2 | 3 | Coming soon. 4 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/zh_cn/advanced_guides/implement_new_models.md: -------------------------------------------------------------------------------- 1 | # Implement New Models 2 | 3 | Coming soon. 4 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/zh_cn/dataset_zoo/2d_fashion_landmark.md: -------------------------------------------------------------------------------- 1 | # 2D服装关键点数据集 2 | 3 | 内容建设中…… 4 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/zh_cn/make.bat: -------------------------------------------------------------------------------- 1 | @ECHO OFF 2 | 3 | pushd %~dp0 4 | 5 | REM Command file for Sphinx documentation 6 | 7 | if "%SPHINXBUILD%" == "" ( 8 | set SPHINXBUILD=sphinx-build 9 | ) 10 | set SOURCEDIR=. 11 | set BUILDDIR=_build 12 | 13 | if "%1" == "" goto help 14 | 15 | %SPHINXBUILD% >NUL 2>NUL 16 | if errorlevel 9009 ( 17 | echo. 18 | echo.The 'sphinx-build' command was not found. Make sure you have Sphinx 19 | echo.installed, then set the SPHINXBUILD environment variable to point 20 | echo.to the full path of the 'sphinx-build' executable. Alternatively you 21 | echo.may add the Sphinx directory to PATH. 22 | echo. 23 | echo.If you don't have Sphinx installed, grab it from 24 | echo.http://sphinx-doc.org/ 25 | exit /b 1 26 | ) 27 | 28 | %SPHINXBUILD% -M %1 %SOURCEDIR% %BUILDDIR% %SPHINXOPTS% %O% 29 | goto end 30 | 31 | :help 32 | %SPHINXBUILD% -M help %SOURCEDIR% %BUILDDIR% %SPHINXOPTS% %O% 33 | 34 | :end 35 | popd 36 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/zh_cn/notes/ecosystem.md: -------------------------------------------------------------------------------- 1 | # Ecosystem 2 | 3 | Coming soon. 4 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/zh_cn/notes/pytorch_2.md: -------------------------------------------------------------------------------- 1 | # PyTorch 2.0 Compatibility and Benchmarks 2 | 3 | Coming soon. 4 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/zh_cn/switch_language.md: -------------------------------------------------------------------------------- 1 | ## 简体中文 2 | 3 | ## English 4 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/zh_cn/user_guides/train_and_test.md: -------------------------------------------------------------------------------- 1 | # 训练与测试 2 | 3 | 中文内容建设中，暂时请查阅[英文版文档](../../en/user_guides/train_and_test.md) 4 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/zh_cn/user_guides/useful_tools.md: -------------------------------------------------------------------------------- 1 | # 常用工具 2 | 3 | 中文内容建设中，暂时请查阅[英文版文档](../../en/user_guides/useful_tools.md) 4 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/docs/zh_cn/user_guides/visualization.md: -------------------------------------------------------------------------------- 1 | # 可视化 2 | 3 | 中文内容建设中，暂时请查阅[英文版文档](../../en/user_guides/visualization.md) 4 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/__init__.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | import mmcv 3 | import mmengine 4 | from mmengine.utils import digit_version 5 | 6 | from .version import __version__, short_version 7 | 8 | mmcv_minimum_version = '2.0.0rc4' 9 | mmcv_maximum_version = '2.1.0' 10 | mmcv_version = digit_version(mmcv.__version__) 11 | 12 | mmengine_minimum_version = '0.6.0' 13 | mmengine_maximum_version = '1.0.0' 14 | mmengine_version = digit_version(mmengine.__version__) 15 | 16 | assert (mmcv_version >= digit_version(mmcv_minimum_version) 17 | and mmcv_version <= digit_version(mmcv_maximum_version)), \ 18 | f'MMCV=={mmcv.__version__} is used but incompatible. ' \ 19 | f'Please install mmcv>={mmcv_minimum_version}, <={mmcv_maximum_version}.' 20 | 21 | assert (mmengine_version >= digit_version(mmengine_minimum_version) 22 | and mmengine_version <= digit_version(mmengine_maximum_version)), \ 23 | f'MMEngine=={mmengine.__version__} is used but incompatible. ' \ 24 | f'Please install mmengine>={mmengine_minimum_version}, ' \ 25 | f'<={mmengine_maximum_version}.' 26 | 27 | __all__ = ['__version__', 'short_version'] 28 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/apis/__init__.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from .inference import inference_bottomup, inference_topdown, init_model 3 | from .inferencers import MMPoseInferencer, Pose2DInferencer 4 | 5 | __all__ = [ 6 | 'init_model', 'inference_topdown', 'inference_bottomup', 7 | 'Pose2DInferencer', 'MMPoseInferencer' 8 | ] 9 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/apis/inferencers/__init__.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from .mmpose_inferencer import MMPoseInferencer 3 | from .pose2d_inferencer import Pose2DInferencer 4 | from .utils import get_model_aliases 5 | 6 | __all__ = ['Pose2DInferencer', 'MMPoseInferencer', 'get_model_aliases'] 7 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/apis/inferencers/utils/__init__.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from .default_det_models import default_det_models 3 | from .get_model_alias import get_model_aliases 4 | 5 | __all__ = ['default_det_models', 'get_model_aliases'] 6 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/apis/inferencers/utils/default_det_models.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | import os.path as osp 3 | 4 | from mmengine.config.utils import MODULE2PACKAGE 5 | from mmengine.utils import get_installed_path 6 | 7 | mmpose_path = get_installed_path(MODULE2PACKAGE['mmpose']) 8 | 9 | default_det_models = dict( 10 | human=dict(model='rtmdet-m', weights=None, cat_ids=(0, )), 11 | face=dict( 12 | model=osp.join(mmpose_path, '.mim', 13 | 'demo/mmdetection_cfg/yolox-s_8xb8-300e_coco-face.py'), 14 | weights='https://download.openmmlab.com/mmpose/mmdet_pretrained/' 15 | 'yolo-x_8xb8-300e_coco-face_13274d7c.pth', 16 | cat_ids=(0, )), 17 | hand=dict( 18 | model=osp.join( 19 | mmpose_path, '.mim', 'demo/mmdetection_cfg/' 20 | 'ssdlite_mobilenetv2_scratch_600e_onehand.py'), 21 | weights='https://download.openmmlab.com/mmpose/mmdet_pretrained/' 22 | 'ssdlite_mobilenetv2_scratch_600e_onehand-4f9f8686_20220523.pth', 23 | cat_ids=(0, )), 24 | animal=dict( 25 | model='rtmdet-m', 26 | weights=None, 27 | cat_ids=(15, 16, 17, 18, 19, 20, 21, 22, 23)), 28 | ) 29 | 30 | default_det_models['body'] = default_det_models['human'] 31 | default_det_models['wholebody'] = default_det_models['human'] 32 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/apis/webcam/__init__.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from .webcam_executor import WebcamExecutor 3 | 4 | __all__ = ['WebcamExecutor'] 5 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/apis/webcam/nodes/__init__.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from .base_visualizer_node import BaseVisualizerNode 3 | from .helper_nodes import MonitorNode, ObjectAssignerNode, RecorderNode 4 | from .model_nodes import DetectorNode, TopdownPoseEstimatorNode 5 | from .node import Node 6 | from .registry import NODES 7 | from .visualizer_nodes import (BigeyeEffectNode, NoticeBoardNode, 8 | ObjectVisualizerNode, SunglassesEffectNode) 9 | 10 | __all__ = [ 11 | 'BaseVisualizerNode', 'NODES', 'MonitorNode', 'ObjectAssignerNode', 12 | 'RecorderNode', 'DetectorNode', 'TopdownPoseEstimatorNode', 'Node', 13 | 'BigeyeEffectNode', 'NoticeBoardNode', 'ObjectVisualizerNode', 14 | 'ObjectAssignerNode', 'SunglassesEffectNode' 15 | ] 16 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/apis/webcam/nodes/helper_nodes/__init__.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from .monitor_node import MonitorNode 3 | from .object_assigner_node import ObjectAssignerNode 4 | from .recorder_node import RecorderNode 5 | 6 | __all__ = ['MonitorNode', 'ObjectAssignerNode', 'RecorderNode'] 7 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/apis/webcam/nodes/model_nodes/__init__.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from .detector_node import DetectorNode 3 | from .pose_estimator_node import TopdownPoseEstimatorNode 4 | 5 | __all__ = ['DetectorNode', 'TopdownPoseEstimatorNode'] 6 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/apis/webcam/nodes/registry.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from mmengine.registry import Registry 3 | 4 | NODES = Registry('node') 5 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/apis/webcam/nodes/visualizer_nodes/__init__.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from .bigeye_effect_node import BigeyeEffectNode 3 | from .notice_board_node import NoticeBoardNode 4 | from .object_visualizer_node import ObjectVisualizerNode 5 | from .sunglasses_effect_node import SunglassesEffectNode 6 | 7 | __all__ = [ 8 | 'ObjectVisualizerNode', 'NoticeBoardNode', 'SunglassesEffectNode', 9 | 'BigeyeEffectNode' 10 | ] 11 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/apis/webcam/utils/__init__.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from .buffer import BufferManager 3 | from .event import EventManager 4 | from .image_capture import ImageCapture 5 | from .message import FrameMessage, Message, VideoEndingMessage 6 | from .misc import (copy_and_paste, expand_and_clamp, get_cached_file_path, 7 | get_config_path, is_image_file, limit_max_fps, 8 | load_image_from_disk_or_url, screen_matting) 9 | from .pose import (get_eye_keypoint_ids, get_face_keypoint_ids, 10 | get_hand_keypoint_ids, get_mouth_keypoint_ids, 11 | get_wrist_keypoint_ids) 12 | 13 | __all__ = [ 14 | 'BufferManager', 'EventManager', 'FrameMessage', 'Message', 15 | 'limit_max_fps', 'VideoEndingMessage', 'load_image_from_disk_or_url', 16 | 'get_cached_file_path', 'screen_matting', 'get_config_path', 17 | 'expand_and_clamp', 'copy_and_paste', 'is_image_file', 'ImageCapture', 18 | 'get_eye_keypoint_ids', 'get_face_keypoint_ids', 'get_wrist_keypoint_ids', 19 | 'get_mouth_keypoint_ids', 'get_hand_keypoint_ids' 20 | ] 21 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/apis/webcam/utils/image_capture.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from typing import Union 3 | 4 | import cv2 5 | import numpy as np 6 | 7 | from .misc import load_image_from_disk_or_url 8 | 9 | 10 | class ImageCapture: 11 | """A mock-up of cv2.VideoCapture that always return a const image. 12 | 13 | Args: 14 | image (str | ndarray): The image path or image data 15 | """ 16 | 17 | def __init__(self, image: Union[str, np.ndarray]): 18 | if isinstance(image, str): 19 | self.image = load_image_from_disk_or_url(image) 20 | else: 21 | self.image = image 22 | 23 | def isOpened(self): 24 | return (self.image is not None) 25 | 26 | def read(self): 27 | return True, self.image.copy() 28 | 29 | def release(self): 30 | pass 31 | 32 | def get(self, propId): 33 | if propId == cv2.CAP_PROP_FRAME_WIDTH: 34 | return self.image.shape[1] 35 | elif propId == cv2.CAP_PROP_FRAME_HEIGHT: 36 | return self.image.shape[0] 37 | elif propId == cv2.CAP_PROP_FPS: 38 | return np.nan 39 | else: 40 | raise NotImplementedError() 41 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/codecs/__init__.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from .associative_embedding import AssociativeEmbedding 3 | from .decoupled_heatmap import DecoupledHeatmap 4 | from .integral_regression_label import IntegralRegressionLabel 5 | from .megvii_heatmap import MegviiHeatmap 6 | from .msra_heatmap import MSRAHeatmap 7 | from .regression_label import RegressionLabel 8 | from .simcc_label import SimCCLabel 9 | from .spr import SPR 10 | from .udp_heatmap import UDPHeatmap 11 | 12 | __all__ = [ 13 | 'MSRAHeatmap', 'MegviiHeatmap', 'UDPHeatmap', 'RegressionLabel', 14 | 'SimCCLabel', 'IntegralRegressionLabel', 'AssociativeEmbedding', 'SPR', 15 | 'DecoupledHeatmap' 16 | ] 17 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/codecs/utils/__init__.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from .gaussian_heatmap import (generate_gaussian_heatmaps, 3 | generate_udp_gaussian_heatmaps, 4 | generate_unbiased_gaussian_heatmaps) 5 | from .instance_property import (get_diagonal_lengths, get_instance_bbox, 6 | get_instance_root) 7 | from .offset_heatmap import (generate_displacement_heatmap, 8 | generate_offset_heatmap) 9 | from .post_processing import (batch_heatmap_nms, gaussian_blur, 10 | gaussian_blur1d, get_heatmap_maximum, 11 | get_simcc_maximum, get_simcc_normalized) 12 | from .refinement import (refine_keypoints, refine_keypoints_dark, 13 | refine_keypoints_dark_udp, refine_simcc_dark) 14 | 15 | __all__ = [ 16 | 'generate_gaussian_heatmaps', 'generate_udp_gaussian_heatmaps', 17 | 'generate_unbiased_gaussian_heatmaps', 'gaussian_blur', 18 | 'get_heatmap_maximum', 'get_simcc_maximum', 'generate_offset_heatmap', 19 | 'batch_heatmap_nms', 'refine_keypoints', 'refine_keypoints_dark', 20 | 'refine_keypoints_dark_udp', 'generate_displacement_heatmap', 21 | 'refine_simcc_dark', 'gaussian_blur1d', 'get_diagonal_lengths', 22 | 'get_instance_root', 'get_instance_bbox', 'get_simcc_normalized' 23 | ] 24 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/datasets/__init__.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from .builder import build_dataset 3 | from .dataset_wrappers import CombinedDataset 4 | from .datasets import * # noqa 5 | from .samplers import MultiSourceSampler 6 | from .transforms import * # noqa 7 | 8 | __all__ = ['build_dataset', 'CombinedDataset', 'MultiSourceSampler'] 9 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/datasets/datasets/__init__.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from .animal import * # noqa: F401, F403 3 | from .base import * # noqa: F401, F403 4 | from .body import * # noqa: F401, F403 5 | from .face import * # noqa: F401, F403 6 | from .fashion import * # noqa: F401, F403 7 | from .hand import * # noqa: F401, F403 8 | from .wholebody import * # noqa: F401, F403 9 | from .head import * 10 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/datasets/datasets/animal/__init__.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from .animalpose_dataset import AnimalPoseDataset 3 | from .ap10k_dataset import AP10KDataset 4 | from .atrw_dataset import ATRWDataset 5 | from .fly_dataset import FlyDataset 6 | from .horse10_dataset import Horse10Dataset 7 | from .locust_dataset import LocustDataset 8 | from .macaque_dataset import MacaqueDataset 9 | from .zebra_dataset import ZebraDataset 10 | 11 | __all__ = [ 12 | 'AnimalPoseDataset', 'AP10KDataset', 'Horse10Dataset', 'MacaqueDataset', 13 | 'FlyDataset', 'LocustDataset', 'ZebraDataset', 'ATRWDataset' 14 | ] 15 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/datasets/datasets/base/__init__.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from .base_coco_style_dataset import BaseCocoStyleDataset 3 | 4 | __all__ = ['BaseCocoStyleDataset'] 5 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/datasets/datasets/body/__init__.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from .aic_dataset import AicDataset 3 | from .coco_dataset import CocoDataset 4 | from .crowdpose_dataset import CrowdPoseDataset 5 | from .jhmdb_dataset import JhmdbDataset 6 | from .mhp_dataset import MhpDataset 7 | from .mpii_dataset import MpiiDataset 8 | from .mpii_trb_dataset import MpiiTrbDataset 9 | from .ochuman_dataset import OCHumanDataset 10 | from .posetrack18_dataset import PoseTrack18Dataset 11 | from .posetrack18_video_dataset import PoseTrack18VideoDataset 12 | 13 | __all__ = [ 14 | 'CocoDataset', 'MpiiDataset', 'MpiiTrbDataset', 'AicDataset', 15 | 'CrowdPoseDataset', 'OCHumanDataset', 'MhpDataset', 'PoseTrack18Dataset', 16 | 'JhmdbDataset', 'PoseTrack18VideoDataset' 17 | ] 18 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/datasets/datasets/face/__init__.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from .aflw_dataset import AFLWDataset 3 | from .coco_wholebody_face_dataset import CocoWholeBodyFaceDataset 4 | from .cofw_dataset import COFWDataset 5 | from .face_300w_dataset import Face300WDataset 6 | from .lapa_dataset import LapaDataset 7 | from .wflw_dataset import WFLWDataset 8 | 9 | __all__ = [ 10 | 'Face300WDataset', 'WFLWDataset', 'AFLWDataset', 'COFWDataset', 11 | 'CocoWholeBodyFaceDataset', 'LapaDataset' 12 | ] 13 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/datasets/datasets/fashion/__init__.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from .deepfashion2_dataset import DeepFashion2Dataset 3 | from .deepfashion_dataset import DeepFashionDataset 4 | 5 | __all__ = ['DeepFashionDataset', 'DeepFashion2Dataset'] 6 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/datasets/datasets/fashion/deepfashion2_dataset.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from mmpose.registry import DATASETS 3 | from ..base import BaseCocoStyleDataset 4 | 5 | 6 | @DATASETS.register_module(name='DeepFashion2Dataset') 7 | class DeepFashion2Dataset(BaseCocoStyleDataset): 8 | """DeepFashion2 dataset for fashion landmark detection.""" 9 | 10 | METAINFO: dict = dict(from_file='configs/_base_/datasets/deepfashion2.py') 11 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/datasets/datasets/hand/__init__.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from .coco_wholebody_hand_dataset import CocoWholeBodyHandDataset 3 | from .freihand_dataset import FreiHandDataset 4 | from .onehand10k_dataset import OneHand10KDataset 5 | from .panoptic_hand2d_dataset import PanopticHand2DDataset 6 | from .rhd2d_dataset import Rhd2DDataset 7 | 8 | __all__ = [ 9 | 'OneHand10KDataset', 'FreiHandDataset', 'PanopticHand2DDataset', 10 | 'Rhd2DDataset', 'CocoWholeBodyHandDataset' 11 | ] 12 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/datasets/datasets/head/__init__.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from .cephalometric_dataset import CephalometricDataset 3 | from .cephalometric_dataset_ISBI2015 import CephalometricDataset_ISBI2015 4 | 5 | __all__ = ['CephalometricDataset',"CephalometricDataset_ISBI2015"] 6 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/datasets/datasets/wholebody/__init__.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from .coco_wholebody_dataset import CocoWholeBodyDataset 3 | from .halpe_dataset import HalpeDataset 4 | 5 | __all__ = ['CocoWholeBodyDataset', 'HalpeDataset'] 6 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/datasets/transforms/__init__.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from .bottomup_transforms import (BottomupGetHeatmapMask, BottomupRandomAffine, 3 | BottomupResize) 4 | from .common_transforms import (Albumentation, GenerateTarget, 5 | GetBBoxCenterScale, PhotometricDistortion, 6 | RandomBBoxTransform, RandomFlip, 7 | RandomHalfBody) 8 | from .converting import KeypointConverter 9 | from .formatting import PackPoseInputs 10 | from .loading import LoadImage 11 | from .topdown_transforms import TopdownAffine 12 | 13 | __all__ = [ 14 | 'GetBBoxCenterScale', 'RandomBBoxTransform', 'RandomFlip', 15 | 'RandomHalfBody', 'TopdownAffine', 'Albumentation', 16 | 'PhotometricDistortion', 'PackPoseInputs', 'LoadImage', 17 | 'BottomupGetHeatmapMask', 'BottomupRandomAffine', 'BottomupResize', 18 | 'GenerateTarget', 'KeypointConverter' 19 | ] 20 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/engine/__init__.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from .hooks import * # noqa: F401, F403 3 | from .optim_wrappers import * # noqa: F401, F403 4 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/engine/hooks/__init__.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from .ema_hook import ExpMomentumEMA 3 | from .visualization_hook import PoseVisualizationHook 4 | 5 | __all__ = ['PoseVisualizationHook', 'ExpMomentumEMA'] 6 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/engine/optim_wrappers/__init__.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from .layer_decay_optim_wrapper import LayerDecayOptimWrapperConstructor 3 | 4 | __all__ = ['LayerDecayOptimWrapperConstructor'] 5 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/evaluation/__init__.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from .functional import * # noqa: F401,F403 3 | from .metrics import * # noqa: F401,F403 4 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/evaluation/functional/__init__.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from .keypoint_eval import (keypoint_auc, keypoint_epe, keypoint_nme, 3 | keypoint_pck_accuracy, 4 | multilabel_classification_accuracy, 5 | pose_pck_accuracy, simcc_pck_accuracy) 6 | from .nms import nms, oks_nms, soft_oks_nms 7 | 8 | __all__ = [ 9 | 'keypoint_pck_accuracy', 'keypoint_auc', 'keypoint_nme', 'keypoint_epe', 10 | 'pose_pck_accuracy', 'multilabel_classification_accuracy', 11 | 'simcc_pck_accuracy', 'nms', 'oks_nms', 'soft_oks_nms' 12 | ] 13 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/evaluation/metrics/__init__.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from .coco_metric import CocoMetric 3 | from .coco_wholebody_metric import CocoWholeBodyMetric 4 | from .keypoint_2d_metrics import (AUC, EPE, NME, JhmdbPCKAccuracy, 5 | MpiiPCKAccuracy, PCKAccuracy) 6 | from .keypoint_partition_metric import KeypointPartitionMetric 7 | from .posetrack18_metric import PoseTrack18Metric 8 | from .cephalometric_metric import CephalometricMetric 9 | 10 | __all__ = [ 11 | 'CocoMetric', 'PCKAccuracy', 'MpiiPCKAccuracy', 'JhmdbPCKAccuracy', 'AUC', 12 | 'EPE', 'NME', 'PoseTrack18Metric', 'CocoWholeBodyMetric', 13 | 'KeypointPartitionMetric', 'CephalometricMetric' 14 | ] 15 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/models/__init__.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from .backbones import * # noqa 3 | from .builder import (BACKBONES, HEADS, LOSSES, NECKS, build_backbone, 4 | build_head, build_loss, build_neck, build_pose_estimator, 5 | build_posenet) 6 | from .data_preprocessors import * # noqa 7 | from .heads import * # noqa 8 | from .losses import * # noqa 9 | from .necks import * # noqa 10 | from .pose_estimators import * # noqa 11 | 12 | __all__ = [ 13 | 'BACKBONES', 'HEADS', 'NECKS', 'LOSSES', 'build_backbone', 'build_head', 14 | 'build_loss', 'build_posenet', 'build_neck', 'build_pose_estimator' 15 | ] 16 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/models/backbones/base_backbone.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from abc import ABCMeta, abstractmethod 3 | 4 | from mmengine.model import BaseModule 5 | 6 | 7 | class BaseBackbone(BaseModule, metaclass=ABCMeta): 8 | """Base backbone. 9 | 10 | This class defines the basic functions of a backbone. Any backbone that 11 | inherits this class should at least define its own `forward` function. 12 | """ 13 | 14 | @abstractmethod 15 | def forward(self, x): 16 | """Forward function. 17 | 18 | Args: 19 | x (Tensor | tuple[Tensor]): x could be a torch.Tensor or a tuple of 20 | torch.Tensor, containing input data for forward computation. 21 | """ 22 | 23 | def train(self, mode=True): 24 | """Set module status before forward computation. 25 | 26 | Args: 27 | mode (bool): Whether it is train_mode or test_mode 28 | """ 29 | super(BaseBackbone, self).train(mode) 30 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/models/backbones/utils/__init__.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from .channel_shuffle import channel_shuffle 3 | from .inverted_residual import InvertedResidual 4 | from .make_divisible import make_divisible 5 | from .se_layer import SELayer 6 | from .utils import get_state_dict, load_checkpoint 7 | 8 | __all__ = [ 9 | 'channel_shuffle', 'make_divisible', 'InvertedResidual', 'SELayer', 10 | 'load_checkpoint', 'get_state_dict' 11 | ] 12 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/models/backbones/utils/channel_shuffle.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | import torch 3 | 4 | 5 | def channel_shuffle(x, groups): 6 | """Channel Shuffle operation. 7 | 8 | This function enables cross-group information flow for multiple groups 9 | convolution layers. 10 | 11 | Args: 12 | x (Tensor): The input tensor. 13 | groups (int): The number of groups to divide the input tensor 14 | in the channel dimension. 15 | 16 | Returns: 17 | Tensor: The output tensor after channel shuffle operation. 18 | """ 19 | 20 | batch_size, num_channels, height, width = x.size() 21 | assert (num_channels % groups == 0), ('num_channels should be ' 22 | 'divisible by groups') 23 | channels_per_group = num_channels // groups 24 | 25 | x = x.view(batch_size, groups, channels_per_group, height, width) 26 | x = torch.transpose(x, 1, 2).contiguous() 27 | x = x.view(batch_size, groups * channels_per_group, height, width) 28 | 29 | return x 30 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/models/backbones/utils/make_divisible.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | def make_divisible(value, divisor, min_value=None, min_ratio=0.9): 3 | """Make divisible function. 4 | 5 | This function rounds the channel number down to the nearest value that can 6 | be divisible by the divisor. 7 | 8 | Args: 9 | value (int): The original channel number. 10 | divisor (int): The divisor to fully divide the channel number. 11 | min_value (int, optional): The minimum value of the output channel. 12 | Default: None, means that the minimum value equal to the divisor. 13 | min_ratio (float, optional): The minimum ratio of the rounded channel 14 | number to the original channel number. Default: 0.9. 15 | Returns: 16 | int: The modified output channel number 17 | """ 18 | 19 | if min_value is None: 20 | min_value = divisor 21 | new_value = max(min_value, int(value + divisor / 2) // divisor * divisor) 22 | # Make sure that round down does not go down by more than (1-min_ratio). 23 | if new_value < min_ratio * value: 24 | new_value += divisor 25 | return new_value 26 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/models/builder.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | import warnings 3 | 4 | from mmpose.registry import MODELS 5 | 6 | BACKBONES = MODELS 7 | NECKS = MODELS 8 | HEADS = MODELS 9 | LOSSES = MODELS 10 | POSE_ESTIMATORS = MODELS 11 | 12 | 13 | def build_backbone(cfg): 14 | """Build backbone.""" 15 | return BACKBONES.build(cfg) 16 | 17 | 18 | def build_neck(cfg): 19 | """Build neck.""" 20 | return NECKS.build(cfg) 21 | 22 | 23 | def build_head(cfg): 24 | """Build head.""" 25 | return HEADS.build(cfg) 26 | 27 | 28 | def build_loss(cfg): 29 | """Build loss.""" 30 | return LOSSES.build(cfg) 31 | 32 | 33 | def build_pose_estimator(cfg): 34 | """Build pose estimator.""" 35 | return POSE_ESTIMATORS.build(cfg) 36 | 37 | 38 | def build_posenet(cfg): 39 | """Build posenet.""" 40 | warnings.warn( 41 | '``build_posenet`` will be deprecated soon, ' 42 | 'please use ``build_pose_estimator`` instead.', DeprecationWarning) 43 | return build_pose_estimator(cfg) 44 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/models/data_preprocessors/__init__.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from .data_preprocessor import PoseDataPreprocessor 3 | 4 | __all__ = ['PoseDataPreprocessor'] 5 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/models/data_preprocessors/data_preprocessor.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from mmengine.model import ImgDataPreprocessor 3 | 4 | from mmpose.registry import MODELS 5 | 6 | 7 | @MODELS.register_module() 8 | class PoseDataPreprocessor(ImgDataPreprocessor): 9 | """Image pre-processor for pose estimation tasks.""" 10 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/models/heads/__init__.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from .base_head import BaseHead 3 | from .coord_cls_heads import RTMCCHead, SimCCHead 4 | from .heatmap_heads import (AssociativeEmbeddingHead, CIDHead, CPMHead, 5 | HeatmapHead, MSPNHead, ViPNASHead) 6 | from .hybrid_heads import DEKRHead 7 | from .regression_heads import (DSNTHead, IntegralRegressionHead, 8 | RegressionHead, RLEHead) 9 | 10 | __all__ = [ 11 | 'BaseHead', 'HeatmapHead', 'CPMHead', 'MSPNHead', 'ViPNASHead', 12 | 'RegressionHead', 'IntegralRegressionHead', 'SimCCHead', 'RLEHead', 13 | 'DSNTHead', 'AssociativeEmbeddingHead', 'DEKRHead', 'CIDHead', 'RTMCCHead' 14 | ] 15 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/models/heads/coord_cls_heads/__init__.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from .rtmcc_head import RTMCCHead 3 | from .simcc_head import SimCCHead 4 | 5 | __all__ = ['SimCCHead', 'RTMCCHead'] 6 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/models/heads/heatmap_heads/__init__.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from .ae_head import AssociativeEmbeddingHead 3 | from .cid_head import CIDHead 4 | from .cpm_head import CPMHead 5 | from .heatmap_head import HeatmapHead,HeatmapHead_withSigmoid 6 | from .mspn_head import MSPNHead 7 | from .vipnas_head import ViPNASHead 8 | from .srpose_head import SRPoseHead 9 | __all__ = [ 10 | 'HeatmapHead', 'CPMHead', 'MSPNHead', 'ViPNASHead', 11 | 'AssociativeEmbeddingHead', 'CIDHead','HeatmapHead_withSigmoid','SRPoseHead' 12 | ] 13 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/models/heads/hybrid_heads/__init__.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from .dekr_head import DEKRHead 3 | 4 | __all__ = [ 5 | 'DEKRHead', 6 | ] 7 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/models/heads/regression_heads/__init__.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from .dsnt_head import DSNTHead 3 | from .integral_regression_head import IntegralRegressionHead 4 | from .regression_head import RegressionHead 5 | from .rle_head import RLEHead 6 | 7 | __all__ = [ 8 | 'RegressionHead', 9 | 'IntegralRegressionHead', 10 | 'DSNTHead', 11 | 'RLEHead', 12 | ] 13 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/models/losses/__init__.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from .ae_loss import AssociativeEmbeddingLoss 3 | from .classification_loss import BCELoss, JSDiscretLoss, KLDiscretLoss 4 | from .heatmap_loss import (AdaptiveWingLoss, KeypointMSELoss, 5 | KeypointOHKMMSELoss,FocalHeatmapLoss_5k,JointsMSELoss,KeypointL1Loss) 6 | from .loss_wrappers import CombinedLoss, MultipleLossWrapper,MultipleLossWrapper_5k 7 | from .regression_loss import (BoneLoss, L1Loss, MPJPELoss, MSELoss, RLELoss, 8 | SemiSupervisionLoss, SmoothL1Loss, 9 | SoftWeightSmoothL1Loss, SoftWingLoss, WingLoss) 10 | 11 | __all__ = [ 12 | 'KeypointMSELoss','KeypointL1Loss','KeypointOHKMMSELoss', 'SmoothL1Loss', 'WingLoss','FocalHeatmapLoss_5k', 13 | 'MPJPELoss', 'MSELoss', 'L1Loss', 'BCELoss', 'BoneLoss', 14 | 'SemiSupervisionLoss', 'SoftWingLoss', 'AdaptiveWingLoss', 'RLELoss','JointsMSELoss', 15 | 'KLDiscretLoss', 'MultipleLossWrapper', 'JSDiscretLoss', 'CombinedLoss', 16 | 'AssociativeEmbeddingLoss', 'SoftWeightSmoothL1Loss','MultipleLossWrapper_5k' 17 | ] 18 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/models/necks/__init__.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from .fmap_proc_neck import FeatureMapProcessor 3 | from .fpn import FPN 4 | from .gap_neck import GlobalAveragePooling 5 | from .posewarper_neck import PoseWarperNeck 6 | 7 | __all__ = [ 8 | 'GlobalAveragePooling', 'PoseWarperNeck', 'FPN', 'FeatureMapProcessor' 9 | ] 10 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/models/necks/gap_neck.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | import torch 3 | import torch.nn as nn 4 | 5 | from mmpose.registry import MODELS 6 | 7 | 8 | @MODELS.register_module() 9 | class GlobalAveragePooling(nn.Module): 10 | """Global Average Pooling neck. 11 | 12 | Note that we use `view` to remove extra channel after pooling. We do not 13 | use `squeeze` as it will also remove the batch dimension when the tensor 14 | has a batch dimension of size 1, which can lead to unexpected errors. 15 | """ 16 | 17 | def __init__(self): 18 | super().__init__() 19 | self.gap = nn.AdaptiveAvgPool2d((1, 1)) 20 | 21 | def init_weights(self): 22 | pass 23 | 24 | def forward(self, inputs): 25 | """Forward function.""" 26 | 27 | if isinstance(inputs, tuple): 28 | outs = tuple([self.gap(x) for x in inputs]) 29 | outs = tuple( 30 | [out.view(x.size(0), -1) for out, x in zip(outs, inputs)]) 31 | elif isinstance(inputs, list): 32 | outs = [self.gap(x) for x in inputs] 33 | outs = [out.view(x.size(0), -1) for out, x in zip(outs, inputs)] 34 | elif isinstance(inputs, torch.Tensor): 35 | outs = self.gap(inputs) 36 | outs = outs.view(inputs.size(0), -1) 37 | else: 38 | raise TypeError('neck inputs should be tuple or torch.tensor') 39 | return outs 40 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/models/pose_estimators/__init__.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from .bottomup import BottomupPoseEstimator 3 | from .topdown import TopdownPoseEstimator 4 | 5 | __all__ = ['TopdownPoseEstimator', 'BottomupPoseEstimator'] 6 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/models/utils/__init__.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from .check_and_update_config import check_and_update_config 3 | from .ckpt_convert import pvt_convert 4 | from .rtmcc_block import RTMCCBlock, rope 5 | from .transformer import PatchEmbed, nchw_to_nlc, nlc_to_nchw 6 | 7 | __all__ = [ 8 | 'PatchEmbed', 'nchw_to_nlc', 'nlc_to_nchw', 'pvt_convert', 'RTMCCBlock', 9 | 'rope', 'check_and_update_config' 10 | ] 11 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/structures/__init__.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from .bbox import (bbox_cs2xywh, bbox_cs2xyxy, bbox_xywh2cs, bbox_xywh2xyxy, 3 | bbox_xyxy2cs, bbox_xyxy2xywh, flip_bbox, 4 | get_udp_warp_matrix, get_warp_matrix) 5 | from .keypoint import flip_keypoints 6 | from .multilevel_pixel_data import MultilevelPixelData 7 | from .pose_data_sample import PoseDataSample 8 | from .utils import merge_data_samples, revert_heatmap, split_instances 9 | 10 | __all__ = [ 11 | 'PoseDataSample', 'MultilevelPixelData', 'bbox_cs2xywh', 'bbox_cs2xyxy', 12 | 'bbox_xywh2cs', 'bbox_xywh2xyxy', 'bbox_xyxy2cs', 'bbox_xyxy2xywh', 13 | 'flip_bbox', 'get_udp_warp_matrix', 'get_warp_matrix', 'flip_keypoints', 14 | 'merge_data_samples', 'revert_heatmap', 'split_instances' 15 | ] 16 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/structures/bbox/__init__.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from .transforms import (bbox_cs2xywh, bbox_cs2xyxy, bbox_xywh2cs, 3 | bbox_xywh2xyxy, bbox_xyxy2cs, bbox_xyxy2xywh, 4 | flip_bbox, get_udp_warp_matrix, get_warp_matrix) 5 | 6 | __all__ = [ 7 | 'bbox_cs2xywh', 'bbox_cs2xyxy', 'bbox_xywh2cs', 'bbox_xywh2xyxy', 8 | 'bbox_xyxy2cs', 'bbox_xyxy2xywh', 'flip_bbox', 'get_udp_warp_matrix', 9 | 'get_warp_matrix' 10 | ] 11 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/structures/keypoint/__init__.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | 3 | from .transforms import flip_keypoints 4 | 5 | __all__ = ['flip_keypoints'] 6 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/testing/__init__.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from ._utils import (get_coco_sample, get_config_file, get_packed_inputs, 3 | get_pose_estimator_cfg, get_repo_dir) 4 | 5 | __all__ = [ 6 | 'get_packed_inputs', 'get_coco_sample', 'get_config_file', 7 | 'get_pose_estimator_cfg', 'get_repo_dir' 8 | ] 9 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/utils/__init__.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from .camera import SimpleCamera, SimpleCameraTorch 3 | from .collect_env import collect_env 4 | from .config_utils import adapt_mmdet_pipeline 5 | from .logger import get_root_logger 6 | from .setup_env import register_all_modules, setup_multi_processes 7 | from .timer import StopWatch 8 | 9 | __all__ = [ 10 | 'get_root_logger', 'collect_env', 'StopWatch', 'setup_multi_processes', 11 | 'register_all_modules', 'SimpleCamera', 'SimpleCameraTorch', 12 | 'adapt_mmdet_pipeline' 13 | ] 14 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/utils/collect_env.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from mmengine.utils import get_git_hash 3 | from mmengine.utils.dl_utils import collect_env as collect_base_env 4 | 5 | import mmpose 6 | 7 | 8 | def collect_env(): 9 | env_info = collect_base_env() 10 | env_info['MMPose'] = (mmpose.__version__ + '+' + get_git_hash(digits=7)) 11 | return env_info 12 | 13 | 14 | if __name__ == '__main__': 15 | for name, val in collect_env().items(): 16 | print(f'{name}: {val}') 17 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/utils/config_utils.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from mmpose.utils.typing import ConfigDict 3 | 4 | 5 | def adapt_mmdet_pipeline(cfg: ConfigDict) -> ConfigDict: 6 | """Converts pipeline types in MMDetection's test dataloader to use the 7 | 'mmdet' namespace. 8 | 9 | Args: 10 | cfg (ConfigDict): Configuration dictionary for MMDetection. 11 | 12 | Returns: 13 | ConfigDict: Configuration dictionary with updated pipeline types. 14 | """ 15 | # use lazy import to avoid hard dependence on mmdet 16 | from mmdet.datasets import transforms 17 | 18 | if 'test_dataloader' not in cfg: 19 | return cfg 20 | 21 | pipeline = cfg.test_dataloader.dataset.pipeline 22 | for trans in pipeline: 23 | if trans['type'] in dir(transforms): 24 | trans['type'] = 'mmdet.' + trans['type'] 25 | 26 | return cfg 27 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/utils/logger.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | import logging 3 | 4 | from mmengine.logging import MMLogger 5 | 6 | 7 | def get_root_logger(log_file=None, log_level=logging.INFO): 8 | """Use `MMLogger` class in mmengine to get the root logger. 9 | 10 | The logger will be initialized if it has not been initialized. By default a 11 | StreamHandler will be added. If `log_file` is specified, a FileHandler will 12 | also be added. The name of the root logger is the top-level package name, 13 | e.g., "mmpose". 14 | 15 | Args: 16 | log_file (str | None): The log filename. If specified, a FileHandler 17 | will be added to the root logger. 18 | log_level (int): The root logger level. Note that only the process of 19 | rank 0 is affected, while other processes will set the level to 20 | "Error" and be silent most of the time. 21 | 22 | Returns: 23 | logging.Logger: The root logger. 24 | """ 25 | return MMLogger('MMLogger', __name__.split('.')[0], log_file, log_level) 26 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/utils/typing.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from typing import Dict, List, Optional, Tuple, Union 3 | 4 | from mmengine.config import ConfigDict 5 | from mmengine.structures import InstanceData, PixelData 6 | from torch import Tensor 7 | 8 | from mmpose.structures import PoseDataSample 9 | 10 | # Type hint of config data 11 | ConfigType = Union[ConfigDict, dict] 12 | OptConfigType = Optional[ConfigType] 13 | # Type hint of one or more config data 14 | MultiConfig = Union[ConfigType, List[ConfigType]] 15 | OptMultiConfig = Optional[MultiConfig] 16 | # Type hint of data samples 17 | SampleList = List[PoseDataSample] 18 | OptSampleList = Optional[SampleList] 19 | InstanceList = List[InstanceData] 20 | PixelDataList = List[PixelData] 21 | Predictions = Union[InstanceList, Tuple[InstanceList, PixelDataList]] 22 | # Type hint of model outputs 23 | ForwardResults = Union[Dict[str, Tensor], List[PoseDataSample], Tuple[Tensor], 24 | Tensor] 25 | # Type hint of features 26 | # - Tuple[Tensor]: multi-level features extracted by the network 27 | # - List[Tuple[Tensor]]: multiple feature pyramids for TTA 28 | # - List[List[Tuple[Tensor]]]: multi-scale feature pyramids 29 | Features = Union[Tuple[Tensor], List[Tuple[Tensor]], List[List[Tuple[Tensor]]]] 30 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/version.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) Open-MMLab. All rights reserved. 2 | 3 | __version__ = '1.0.0' 4 | short_version = __version__ 5 | 6 | 7 | def parse_version_info(version_str): 8 | """Parse a version string into a tuple. 9 | 10 | Args: 11 | version_str (str): The version string. 12 | Returns: 13 | tuple[int | str]: The version info, e.g., "1.3.0" is parsed into 14 | (1, 3, 0), and "2.0.0rc1" is parsed into (2, 0, 0, 'rc1'). 15 | """ 16 | version_info = [] 17 | for x in version_str.split('.'): 18 | if x.isdigit(): 19 | version_info.append(int(x)) 20 | elif x.find('rc') != -1: 21 | patch_version = x.split('rc') 22 | version_info.append(int(patch_version[0])) 23 | version_info.append(f'rc{patch_version[1]}') 24 | elif x.find('b') != -1: 25 | patch_version = x.split('b') 26 | version_info.append(int(patch_version[0])) 27 | version_info.append(f'b{patch_version[1]}') 28 | return tuple(version_info) 29 | 30 | 31 | version_info = parse_version_info(__version__) 32 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/mmpose/visualization/__init__.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from .local_visualizer import PoseLocalVisualizer 3 | 4 | __all__ = ['PoseLocalVisualizer'] 5 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/projects/example_project/configs/example-head-loss_hrnet-w32_8xb64-210e_coco-256x192.py: -------------------------------------------------------------------------------- 1 | # Directly inherit the entire recipe you want to use. 2 | _base_ = 'mmpose::body_2d_keypoint/topdown_heatmap/coco/' \ 3 | 'td-hm_hrnet-w32_8xb64-210e_coco-256x192.py' 4 | 5 | # This line is to import your own modules. 6 | custom_imports = dict(imports='models') 7 | 8 | # Modify the model to use your own head and loss. 9 | _base_['model']['head'] = dict( 10 | type='ExampleHead', 11 | in_channels=32, 12 | out_channels=17, 13 | deconv_out_channels=None, 14 | loss=dict(type='ExampleLoss', use_target_weight=True), 15 | decoder=_base_['codec']) 16 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/projects/example_project/models/__init__.py: -------------------------------------------------------------------------------- 1 | from .example_head import ExampleHead 2 | from .example_loss import ExampleLoss 3 | 4 | __all__ = ['ExampleHead', 'ExampleLoss'] 5 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/projects/mmpose4aigc/download_models.sh: -------------------------------------------------------------------------------- 1 | #!/bin/bash 2 | # Copyright (c) OpenMMLab. All rights reserved. 3 | 4 | # Create models folder 5 | mkdir models 6 | 7 | # Go to models folder 8 | cd models 9 | 10 | # Download det model 11 | wget https://download.openmmlab.com/mmpose/v1/projects/rtmpose/rtmdet_nano_8xb32-100e_coco-obj365-person-05d8511e.pth 12 | 13 | # Download pose model 14 | wget https://download.openmmlab.com/mmpose/v1/projects/rtmposev1/rtmpose-m_simcc-aic-coco_pt-aic-coco_420e-256x192-63eb25f7_20230126.pth 15 | 16 | # Go back mmpose4aigc 17 | cd .. 18 | 19 | # Success 20 | echo "Download completed." 21 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/projects/mmpose4aigc/install_posetracker_linux.sh: -------------------------------------------------------------------------------- 1 | #!/bin/bash 2 | # Copyright (c) OpenMMLab. All rights reserved. 3 | 4 | # Download pre-compiled files 5 | wget https://github.com/open-mmlab/mmdeploy/releases/download/v1.0.0/mmdeploy-1.0.0-linux-x86_64-cxx11abi.tar.gz 6 | 7 | # Unzip files 8 | tar -xzvf mmdeploy-1.0.0-linux-x86_64-cxx11abi.tar.gz 9 | 10 | # Go to the sdk folder 11 | cd mmdeploy-1.0.0-linux-x86_64-cxx11abi 12 | 13 | # Init environment 14 | source set_env.sh 15 | 16 | # If opencv 3+ is not installed on your system, execute the following command. 17 | # If it is installed, skip this command 18 | bash install_opencv.sh 19 | 20 | # Compile executable programs 21 | bash build_sdk.sh 22 | 23 | # Download models 24 | wget https://download.openmmlab.com/mmpose/v1/projects/rtmpose/rtmpose-cpu.zip 25 | 26 | # Unzip files 27 | unzip rtmpose-cpu.zip 28 | 29 | # Success 30 | echo "Installation completed." 31 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/projects/mmpose4aigc/mmpose_openpose.sh: -------------------------------------------------------------------------------- 1 | #!/bin/bash 2 | # Copyright (c) OpenMMLab. All rights reserved. 3 | 4 | INPUT_IMAGE=$1 5 | 6 | python openpose_visualization.py \ 7 | ../rtmpose/rtmdet/person/rtmdet_nano_320-8xb32_coco-person.py \ 8 | models/rtmdet_nano_8xb32-100e_coco-obj365-person-05d8511e.pth \ 9 | ../rtmpose/rtmpose/body_2d_keypoint/rtmpose-m_8xb256-420e_coco-256x192.py \ 10 | models/rtmpose-m_simcc-aic-coco_pt-aic-coco_420e-256x192-63eb25f7_20230126.pth \ 11 | --input $INPUT_IMAGE \ 12 | --device cuda:0 \ 13 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/projects/mmpose4aigc/mmpose_style_skeleton.sh: -------------------------------------------------------------------------------- 1 | #!/bin/bash 2 | # Copyright (c) OpenMMLab. All rights reserved. 3 | 4 | WORKSPACE=mmdeploy-1.0.0-linux-x86_64-cxx11abi 5 | export LD_LIBRARY_PATH=${WORKSPACE}/lib:${WORKSPACE}/thirdparty/onnxruntime/lib:$LD_LIBRARY_PATH 6 | 7 | INPUT_IMAGE=$1 8 | 9 | ${WORKSPACE}/bin/pose_tracker \ 10 | ${WORKSPACE}/rtmpose-ort/rtmdet-nano \ 11 | ${WORKSPACE}/rtmpose-ort/rtmpose-m \ 12 | $INPUT_IMAGE \ 13 | --background black \ 14 | --skeleton ${WORKSPACE}/rtmpose-ort/t2i-adapter_skeleton.txt \ 15 | --output ./skeleton_res.jpg \ 16 | --pose_kpt_thr 0.4 \ 17 | --show -1 18 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/projects/rtmpose/examples/README.md: -------------------------------------------------------------------------------- 1 | ## List of examples 2 | 3 | ### 1. RTMPose-Deploy (without MMDeploy) 4 | 5 | RTMPose-Deploy is a C++ code example that does not use MMDEPLOY for RTMPose localized deployment. 6 | 7 | - [Original Repository](https://github.com/HW140701/RTMPose-Deploy) 8 | 9 | ### 2. RTMPose inference with ONNXRuntime (Python) 10 | 11 | This example shows how to run RTMPose inference with ONNXRuntime in Python. 12 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/projects/rtmpose/examples/RTMPose-Deploy/README.md: -------------------------------------------------------------------------------- 1 | # RTMPose-Deploy 2 | 3 | [中文说明](./README_CN.md) 4 | 5 | RTMPose-Deploy is a C ++ code example that does **NOT** use MMDEPLOY for RTMPose localized deployment. 6 | 7 | At present, RTMPose-Deploy has completed the deployment of RTMDetnano and RTMPose on the Windows system. This example only contains the source code. If you want a complete project example, please refer to:[https://github.com/HW140701/RTMPose-Deploy](https://github.com/HW140701/RTMPose-Deploy) . This project provides a complete VS2019 project and release package. 8 | 9 | Subsequent will consider adding the use of C ++ Tensorrt SDK on the Windows system to deploy RTMDet-nano and RTMPose. 10 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/projects/rtmpose/examples/RTMPose-Deploy/README_CN.md: -------------------------------------------------------------------------------- 1 | # RTMPose-Deploy 2 | 3 | RTMPose-Deploy 是一个**不使用** mmdeploy 进行 RTMPose 本地化部署的 C++ 代码示例。 4 | 5 | 目前，RTMPose-Deploy 已完成在 Windows 系统上使用 OnnxRuntime CPU 对 RTMDet-nano 和 RTMPose 完成了部署。本示例只包含了源代码，如果你想要完整的项目示例可以参考：[https://github.com/HW140701/RTMPose-Deploy](https://github.com/HW140701/RTMPose-Deploy)，这个仓库提供了完整的 VS2019 工程和发布包。 6 | 7 | 后续会考虑添加在 Windows 系统上使用 C++ TensorRT SDK 对 RTMDet-nano 和 RTMPose 进行部署。 8 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/projects/rtmpose/examples/RTMPose-Deploy/Windows/OnnxRumtime-CPU/src/RTMPoseOnnxRuntime/rtmdet_onnxruntime.h: -------------------------------------------------------------------------------- 1 | #ifndef _RTM_DET_ONNX_RUNTIME_H_ 2 | #define _RTM_DET_ONNX_RUNTIME_H_ 3 | 4 | #include 5 | 6 | #include "opencv2/opencv.hpp" 7 | 8 | #include "onnxruntime_cxx_api.h" 9 | #include "cpu_provider_factory.h" 10 | #include "rtmpose_utils.h" 11 | 12 | 13 | class RTMDetOnnxruntime 14 | { 15 | public: 16 | RTMDetOnnxruntime() = delete; 17 | RTMDetOnnxruntime(const std::string& onnx_model_path); 18 | virtual~RTMDetOnnxruntime(); 19 | 20 | public: 21 | DetectBox Inference(const cv::Mat& input_mat); 22 | 23 | private: 24 | void PrintModelInfo(Ort::Session& session); 25 | 26 | private: 27 | Ort::Env m_env; 28 | Ort::Session m_session; 29 | 30 | }; 31 | 32 | #endif // !_RTM_DET_ONNX_RUNTIME_H_ 33 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/projects/rtmpose/examples/RTMPose-Deploy/Windows/OnnxRumtime-CPU/src/RTMPoseOnnxRuntime/rtmpose_onnxruntime.h: -------------------------------------------------------------------------------- 1 | #ifndef _RTM_POSE_ONNXRUNTIME_H_ 2 | #define _RTM_POSE_ONNXRUNTIME_H_ 3 | 4 | #include 5 | 6 | #include "onnxruntime_cxx_api.h" 7 | #include "cpu_provider_factory.h" 8 | #include "opencv2/opencv.hpp" 9 | 10 | #include "rtmdet_onnxruntime.h" 11 | #include "rtmpose_utils.h" 12 | 13 | class RTMPoseOnnxruntime 14 | { 15 | public: 16 | RTMPoseOnnxruntime() = delete; 17 | RTMPoseOnnxruntime(const std::string& onnx_model_path); 18 | virtual~RTMPoseOnnxruntime(); 19 | 20 | public: 21 | std::vector Inference(const cv::Mat& input_mat, const DetectBox& box); 22 | 23 | private: 24 | std::pair CropImageByDetectBox(const cv::Mat& input_image, const DetectBox& box); 25 | 26 | private: 27 | void PrintModelInfo(Ort::Session& session); 28 | 29 | private: 30 | Ort::Env m_env; 31 | Ort::Session m_session; 32 | }; 33 | 34 | #endif // !_RTM_POSE_ONNXRUNTIME_H_ 35 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/projects/rtmpose/examples/RTMPose-Deploy/Windows/OnnxRumtime-CPU/src/RTMPoseOnnxRuntime/rtmpose_tracker_onnxruntime.cpp: -------------------------------------------------------------------------------- 1 | #include "rtmpose_tracker_onnxruntime.h" 2 | 3 | RTMPoseTrackerOnnxruntime::RTMPoseTrackerOnnxruntime(const std::string& det_model_path, const std::string& pose_model_path, int dectect_interval) 4 | :m_rtm_det_ptr(nullptr), 5 | m_rtm_pose_ptr(nullptr), 6 | m_frame_num(0), 7 | m_dectect_interval(dectect_interval) 8 | { 9 | m_rtm_det_ptr = std::make_unique(det_model_path); 10 | m_rtm_pose_ptr = std::make_unique(pose_model_path); 11 | } 12 | 13 | RTMPoseTrackerOnnxruntime::~RTMPoseTrackerOnnxruntime() 14 | { 15 | } 16 | 17 | std::pair> RTMPoseTrackerOnnxruntime::Inference(const cv::Mat& input_mat) 18 | { 19 | std::pair> result; 20 | 21 | if (m_rtm_det_ptr == nullptr || m_rtm_pose_ptr == nullptr) 22 | return result; 23 | 24 | if (m_frame_num % m_dectect_interval == 0) 25 | { 26 | m_detect_box = m_rtm_det_ptr->Inference(input_mat); 27 | } 28 | 29 | std::vector pose_result = m_rtm_pose_ptr->Inference(input_mat, m_detect_box); 30 | 31 | m_frame_num += 1; 32 | 33 | return std::make_pair(m_detect_box, pose_result); 34 | } 35 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/projects/rtmpose/examples/RTMPose-Deploy/Windows/OnnxRumtime-CPU/src/RTMPoseOnnxRuntime/rtmpose_tracker_onnxruntime.h: -------------------------------------------------------------------------------- 1 | #ifndef _RTM_POSE_TRACKER_ONNXRUNTIME_H_ 2 | #define _RTM_POSE_TRACKER_ONNXRUNTIME_H_ 3 | 4 | #include "rtmdet_onnxruntime.h" 5 | #include "rtmpose_onnxruntime.h" 6 | 7 | #include 8 | #include 9 | 10 | class RTMPoseTrackerOnnxruntime 11 | { 12 | public: 13 | RTMPoseTrackerOnnxruntime() = delete; 14 | RTMPoseTrackerOnnxruntime( 15 | const std::string& det_model_path, 16 | const std::string& pose_model_path, 17 | int dectect_interval = 10 18 | ); 19 | virtual~RTMPoseTrackerOnnxruntime(); 20 | 21 | public: 22 | std::pair> Inference(const cv::Mat& input_mat); 23 | 24 | private: 25 | std::unique_ptr m_rtm_det_ptr; 26 | std::unique_ptr m_rtm_pose_ptr; 27 | unsigned int m_frame_num; 28 | DetectBox m_detect_box; 29 | int m_dectect_interval; 30 | }; 31 | 32 | #endif // !_RTM_POSE_TRACKER_ONNXRUNTIME_H_ 33 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/projects/rtmpose/examples/onnxruntime/human-pose.jpeg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/5k5000/CLdetection2023/d1a01536ad892134c4dd728c87dc9ac1d87b8e11/mmpose_package/mmpose/projects/rtmpose/examples/onnxruntime/human-pose.jpeg -------------------------------------------------------------------------------- /mmpose_package/mmpose/projects/rtmpose/examples/onnxruntime/requirements.txt: -------------------------------------------------------------------------------- 1 | loguru==0.6.0 2 | numpy==1.21.6 3 | onnxruntime==1.14.1 4 | onnxruntime-gpu==1.8.1 5 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/projects/rtmpose/rtmdet/person/rtmdet_m_640-8xb32_coco-person.py: -------------------------------------------------------------------------------- 1 | _base_ = 'mmdet::rtmdet/rtmdet_m_8xb32-300e_coco.py' 2 | 3 | checkpoint = 'https://download.openmmlab.com/mmdetection/v3.0/rtmdet/cspnext_rsb_pretrain/cspnext-m_8xb256-rsb-a1-600e_in1k-ecb3bbd9.pth' # noqa 4 | 5 | model = dict( 6 | backbone=dict( 7 | init_cfg=dict( 8 | type='Pretrained', prefix='backbone.', checkpoint=checkpoint)), 9 | bbox_head=dict(num_classes=1), 10 | test_cfg=dict( 11 | nms_pre=1000, 12 | min_bbox_size=0, 13 | score_thr=0.05, 14 | nms=dict(type='nms', iou_threshold=0.6), 15 | max_per_img=100)) 16 | 17 | train_dataloader = dict(dataset=dict(metainfo=dict(classes=('person', )))) 18 | 19 | val_dataloader = dict(dataset=dict(metainfo=dict(classes=('person', )))) 20 | test_dataloader = val_dataloader 21 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/projects/rtmpose/rtmpose/pruning/group_fisher_finetune_rtmpose-s_8xb256-420e_aic-coco-256x192.py: -------------------------------------------------------------------------------- 1 | ############################################################################# 2 | """# You have to fill these args. 3 | 4 | _base_(str): The path to your pruning config file. 5 | pruned_path (str): The path to the checkpoint of the pruned model. 6 | finetune_lr (float): The lr rate to finetune. Usually, we directly use the lr 7 | rate of the pretrain. 8 | """ 9 | 10 | _base_ = './group_fisher_prune_rtmpose-s_8xb256-420e_aic-coco-256x192.py' # noqa 11 | pruned_path = 'https://download.openmmlab.com/mmrazor/v1/pruning/group_fisher/rtmpose-s/group_fisher_prune_rtmpose-s_8xb256-420e_aic-coco-256x192.pth' # noqa 12 | finetune_lr = 4e-3 13 | ############################################################################## 14 | 15 | algorithm = _base_.model 16 | algorithm.init_cfg = dict(type='Pretrained', checkpoint=pruned_path) 17 | 18 | model = dict( 19 | _delete_=True, 20 | _scope_='mmrazor', 21 | type='GroupFisherSubModel', 22 | algorithm=algorithm, 23 | ) 24 | 25 | # restore lr 26 | optim_wrapper = dict(optimizer=dict(lr=finetune_lr)) 27 | 28 | # remove pruning related hooks 29 | custom_hooks = _base_.custom_hooks[:-2] 30 | 31 | # delete ddp 32 | model_wrapper_cfg = None 33 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/projects/rtmpose/rtmpose/pruning/group_fisher_finetune_rtmpose-s_8xb256-420e_coco-256x192.py: -------------------------------------------------------------------------------- 1 | ############################################################################# 2 | """# You have to fill these args. 3 | 4 | _base_(str): The path to your pruning config file. 5 | pruned_path (str): The path to the checkpoint of the pruned model. 6 | finetune_lr (float): The lr rate to finetune. Usually, we directly use the lr 7 | rate of the pretrain. 8 | """ 9 | 10 | _base_ = './group_fisher_prune_rtmpose-s_8xb256-420e_coco-256x192.py' 11 | pruned_path = 'https://download.openmmlab.com/mmrazor/v1/pruning/group_fisher/rtmpose-s/group_fisher_prune_rtmpose-s_8xb256-420e_coco-256x192.pth' # noqa 12 | finetune_lr = 4e-3 13 | ############################################################################## 14 | 15 | algorithm = _base_.model 16 | algorithm.init_cfg = dict(type='Pretrained', checkpoint=pruned_path) 17 | # algorithm.update(dict(architecture=dict(test_cfg=dict(flip_test=False), ))) # disable flip test # noqa 18 | 19 | model = dict( 20 | _delete_=True, 21 | _scope_='mmrazor', 22 | type='GroupFisherSubModel', 23 | algorithm=algorithm, 24 | ) 25 | 26 | # restore lr 27 | optim_wrapper = dict(optimizer=dict(lr=finetune_lr)) 28 | 29 | # remove pruning related hooks 30 | custom_hooks = _base_.custom_hooks[:-2] 31 | 32 | # delete ddp 33 | model_wrapper_cfg = None 34 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/projects/yolox-pose/configs/_base_/datasets: -------------------------------------------------------------------------------- 1 | ../../../../configs/_base_/datasets 2 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/projects/yolox-pose/configs/_base_/default_runtime.py: -------------------------------------------------------------------------------- 1 | default_scope = 'mmyolo' 2 | custom_imports = dict(imports=['models', 'datasets']) 3 | 4 | # hooks 5 | default_hooks = dict( 6 | timer=dict(type='IterTimerHook'), 7 | logger=dict(type='LoggerHook', interval=50), 8 | param_scheduler=dict(type='ParamSchedulerHook'), 9 | checkpoint=dict(type='CheckpointHook', interval=10, max_keep_ckpts=3), 10 | sampler_seed=dict(type='DistSamplerSeedHook'), 11 | visualization=dict(type='mmpose.PoseVisualizationHook', enable=False), 12 | ) 13 | 14 | # multi-processing backend 15 | env_cfg = dict( 16 | cudnn_benchmark=False, 17 | mp_cfg=dict(mp_start_method='fork', opencv_num_threads=0), 18 | dist_cfg=dict(backend='nccl'), 19 | ) 20 | 21 | # visualizer 22 | vis_backends = [dict(type='LocalVisBackend')] 23 | visualizer = dict( 24 | type='mmpose.PoseLocalVisualizer', 25 | vis_backends=vis_backends, 26 | name='visualizer') 27 | 28 | # logger 29 | log_processor = dict( 30 | type='LogProcessor', window_size=50, by_epoch=True, num_digits=6) 31 | log_level = 'INFO' 32 | load_from = None 33 | resume = False 34 | 35 | # file I/O backend 36 | file_client_args = dict(backend='disk') 37 | 38 | # training/validation/testing progress 39 | train_cfg = dict() 40 | val_cfg = dict(type='ValLoop') 41 | test_cfg = dict(type='TestLoop') 42 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/projects/yolox-pose/configs/yolox-pose_l_4xb64-300e_coco.py: -------------------------------------------------------------------------------- 1 | _base_ = ['./yolox-pose_s_8xb32-300e_coco.py'] 2 | 3 | # model settings 4 | model = dict( 5 | init_cfg=dict(checkpoint='https://download.openmmlab.com/mmyolo/v0/yolox/' 6 | 'yolox_l_fast_8xb8-300e_coco/yolox_l_fast_8xb8-300e_' 7 | 'coco_20230213_160715-c731eb1c.pth'), 8 | backbone=dict( 9 | deepen_factor=1.0, 10 | widen_factor=1.0, 11 | ), 12 | neck=dict( 13 | deepen_factor=1.0, 14 | widen_factor=1.0, 15 | ), 16 | bbox_head=dict(head_module=dict(widen_factor=1.0))) 17 | 18 | train_dataloader = dict(batch_size=64) 19 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/projects/yolox-pose/configs/yolox-pose_m_4xb64-300e_coco.py: -------------------------------------------------------------------------------- 1 | _base_ = ['./yolox-pose_s_8xb32-300e_coco.py'] 2 | 3 | # model settings 4 | model = dict( 5 | init_cfg=dict(checkpoint='https://download.openmmlab.com/mmyolo/v0/yolox/' 6 | 'yolox_m_fast_8xb32-300e-rtmdet-hyp_coco/yolox_m_fast_8xb32' 7 | '-300e-rtmdet-hyp_coco_20230210_144328-e657e182.pth'), 8 | backbone=dict( 9 | deepen_factor=0.67, 10 | widen_factor=0.75, 11 | ), 12 | neck=dict( 13 | deepen_factor=0.67, 14 | widen_factor=0.75, 15 | ), 16 | bbox_head=dict(head_module=dict(widen_factor=0.75))) 17 | 18 | train_dataloader = dict(batch_size=64) 19 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/projects/yolox-pose/datasets/__init__.py: -------------------------------------------------------------------------------- 1 | from .bbox_keypoint_structure import * # noqa 2 | from .coco_dataset import * # noqa 3 | from .transforms import * # noqa 4 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/projects/yolox-pose/datasets/coco_dataset.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from typing import Any 3 | 4 | from mmengine.dataset import force_full_init 5 | from mmyolo.registry import DATASETS 6 | 7 | from mmpose.datasets import CocoDataset as MMPoseCocoDataset 8 | 9 | 10 | @DATASETS.register_module() 11 | class CocoDataset(MMPoseCocoDataset): 12 | 13 | @force_full_init 14 | def prepare_data(self, idx) -> Any: 15 | data_info = self.get_data_info(idx) 16 | data_info['dataset'] = self 17 | return self.pipeline(data_info) 18 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/projects/yolox-pose/demo: -------------------------------------------------------------------------------- 1 | ../../demo 2 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/projects/yolox-pose/models/__init__.py: -------------------------------------------------------------------------------- 1 | from .assigner import * # noqa 2 | from .data_preprocessor import * # noqa 3 | from .oks_loss import * # noqa 4 | from .utils import * # noqa 5 | from .yolox_pose_head import * # noqa 6 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/projects/yolox-pose/models/data_preprocessor.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from typing import List, Tuple 3 | 4 | from mmdet.models import BatchSyncRandomResize 5 | from mmyolo.registry import MODELS 6 | from torch import Tensor 7 | 8 | from mmpose.structures import PoseDataSample 9 | 10 | 11 | @MODELS.register_module() 12 | class PoseBatchSyncRandomResize(BatchSyncRandomResize): 13 | """Batch random resize which synchronizes the random size across ranks. 14 | 15 | This transform is similar to `mmdet.BatchSyncRandomResize`, but it also 16 | rescales the keypoints coordinates simultaneously. 17 | """ 18 | 19 | def forward(self, inputs: Tensor, data_samples: List[PoseDataSample] 20 | ) -> Tuple[Tensor, List[PoseDataSample]]: 21 | 22 | inputs = inputs.float() 23 | h, w = inputs.shape[-2:] 24 | if self._input_size is None: 25 | self._input_size = (h, w) 26 | scale_y = self._input_size[0] / h 27 | scale_x = self._input_size[1] / w 28 | if scale_x != 1 or scale_y != 1: 29 | for data_sample in data_samples: 30 | data_sample.gt_instances.keypoints[..., 0] *= scale_x 31 | data_sample.gt_instances.keypoints[..., 1] *= scale_y 32 | 33 | return super().forward(inputs, data_samples) 34 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/projects/yolox-pose/tools: -------------------------------------------------------------------------------- 1 | ../../tools 2 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/pytest.ini: -------------------------------------------------------------------------------- 1 | [pytest] 2 | addopts = --xdoctest --xdoctest-style=auto 3 | norecursedirs = .git ignore build __pycache__ data docker docs .eggs .mim tests/legacy 4 | 5 | filterwarnings= default 6 | ignore:.*No cfgstr given in Cacher constructor or call.*:Warning 7 | ignore:.*Define the __nice__ method for.*:Warning 8 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/requirements.txt: -------------------------------------------------------------------------------- 1 | -r requirements/build.txt 2 | -r requirements/runtime.txt 3 | -r requirements/tests.txt 4 | -r requirements/optional.txt 5 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/requirements/albu.txt: -------------------------------------------------------------------------------- 1 | albumentations>=0.3.2 --no-binary qudida,albumentations 2 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/requirements/build.txt: -------------------------------------------------------------------------------- 1 | # These must be installed before building mmpose 2 | numpy 3 | torch>=1.6 4 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/requirements/docs.txt: -------------------------------------------------------------------------------- 1 | docutils==0.16.0 2 | markdown 3 | myst-parser 4 | -e git+https://github.com/gaotongxiao/pytorch_sphinx_theme.git#egg=pytorch_sphinx_theme 5 | sphinx==4.5.0 6 | sphinx_copybutton 7 | sphinx_markdown_tables 8 | urllib3<2.0.0 9 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/requirements/mminstall.txt: -------------------------------------------------------------------------------- 1 | mmcv>=2.0.0,<2.1.0 2 | mmdet>=3.0.0,<3.1.0 3 | mmengine>=0.4.0,<1.0.0 4 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/requirements/optional.txt: -------------------------------------------------------------------------------- 1 | requests 2 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/requirements/poseval.txt: -------------------------------------------------------------------------------- 1 | poseval@git+https://github.com/svenkreiss/poseval.git 2 | shapely==1.8.4 3 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/requirements/readthedocs.txt: -------------------------------------------------------------------------------- 1 | mmcv>=2.0.0rc4 2 | mmengine>=0.6.0,<1.0.0 3 | munkres 4 | regex 5 | scipy 6 | titlecase 7 | torch>1.6 8 | torchvision 9 | xtcocotools>=1.13 10 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/requirements/runtime.txt: -------------------------------------------------------------------------------- 1 | chumpy 2 | json_tricks 3 | matplotlib 4 | munkres 5 | numpy 6 | opencv-python 7 | pillow 8 | scipy 9 | torchvision 10 | xtcocotools>=1.12 11 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/requirements/tests.txt: -------------------------------------------------------------------------------- 1 | coverage 2 | flake8 3 | interrogate 4 | isort==4.3.21 5 | parameterized 6 | pytest 7 | pytest-runner 8 | xdoctest>=0.10.0 9 | yapf 10 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/resources/mmpose-logo.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/5k5000/CLdetection2023/d1a01536ad892134c4dd728c87dc9ac1d87b8e11/mmpose_package/mmpose/resources/mmpose-logo.png -------------------------------------------------------------------------------- /mmpose_package/mmpose/setup.cfg: -------------------------------------------------------------------------------- 1 | [bdist_wheel] 2 | universal=1 3 | 4 | [aliases] 5 | test=pytest 6 | 7 | [yapf] 8 | based_on_style = pep8 9 | blank_line_before_nested_class_or_def = true 10 | split_before_expression_after_opening_paren = true 11 | split_penalty_import_names=0 12 | SPLIT_PENALTY_AFTER_OPENING_BRACKET=800 13 | 14 | [isort] 15 | line_length = 79 16 | multi_line_output = 0 17 | extra_standard_library = pkg_resources,setuptools 18 | known_first_party = mmpose 19 | known_third_party = PIL,cv2,h5py,json_tricks,matplotlib,mmcv,munkres,numpy,pytest,pytorch_sphinx_theme,requests,scipy,seaborn,spacepy,titlecase,torch,torchvision,webcam_apis,xmltodict,xtcocotools 20 | no_lines_before = STDLIB,LOCALFOLDER 21 | default_section = THIRDPARTY 22 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/tests/test_apis/test_webcam/test_utils/test_event.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | import unittest 3 | from threading import Event 4 | 5 | from mmpose.apis.webcam.utils.event import EventManager 6 | 7 | 8 | class TestEventManager(unittest.TestCase): 9 | 10 | def test_event_manager(self): 11 | event_manager = EventManager() 12 | 13 | # test register_event 14 | event_manager.register_event('example_event') 15 | self.assertIn('example_event', event_manager._events) 16 | self.assertIsInstance(event_manager._events['example_event'], Event) 17 | self.assertFalse(event_manager.is_set('example_event')) 18 | 19 | # test event operations 20 | event_manager.set('q', is_keyboard=True) 21 | self.assertIn('_keyboard_q', event_manager._events) 22 | self.assertTrue(event_manager.is_set('q', is_keyboard=True)) 23 | 24 | flag = event_manager.wait('q', is_keyboard=True) 25 | self.assertTrue(flag) 26 | 27 | event_manager.wait_and_handle('q', is_keyboard=True) 28 | event_manager.clear('q', is_keyboard=True) 29 | self.assertFalse(event_manager._events['_keyboard_q']._flag) 30 | 31 | 32 | if __name__ == '__main__': 33 | unittest.main() 34 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/tests/test_apis/test_webcam/test_webcam_executor.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | import unittest 3 | 4 | from mmengine import Config 5 | 6 | from mmpose.apis.webcam import WebcamExecutor 7 | 8 | 9 | class TestWebcamExecutor(unittest.TestCase): 10 | 11 | def setUp(self) -> None: 12 | config = Config.fromfile('demo/webcam_cfg/test_camera.py').executor_cfg 13 | config.camera_id = 'tests/data/posetrack18/videos/' \ 14 | '000001_mpiinew_test/000001_mpiinew_test.mp4' 15 | self.executor = WebcamExecutor(**config) 16 | 17 | def test_init(self): 18 | 19 | self.assertEqual(len(self.executor.node_list), 2) 20 | self.assertEqual(self.executor.node_list[0].name, 'monitor') 21 | self.assertEqual(self.executor.node_list[1].name, 'recorder') 22 | 23 | 24 | if __name__ == '__main__': 25 | unittest.main() 26 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/tests/test_datasets/test_transforms/test_loading.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from unittest import TestCase 3 | 4 | import numpy as np 5 | from mmcv import imread 6 | 7 | from mmpose.datasets.transforms.loading import LoadImage 8 | 9 | 10 | class TestLoadImage(TestCase): 11 | 12 | def test_load_image(self): 13 | 14 | transform = LoadImage() 15 | results = dict(img_path='tests/data/coco/000000000785.jpg') 16 | 17 | results = transform(results) 18 | 19 | self.assertIsInstance(results['img'], np.ndarray) 20 | 21 | def test_with_input_image(self): 22 | transform = LoadImage(to_float32=True) 23 | 24 | img_path = 'tests/data/coco/000000000785.jpg' 25 | results = dict( 26 | img_path=img_path, img=imread(img_path).astype(np.uint8)) 27 | 28 | results = transform(results) 29 | 30 | self.assertIsInstance(results['img'], np.ndarray) 31 | self.assertTrue(results['img'].dtype, np.float32) 32 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/tests/test_models/test_backbones/test_alexnet.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from unittest import TestCase 3 | 4 | import torch 5 | 6 | from mmpose.models.backbones import AlexNet 7 | 8 | 9 | class TestAlexNet(TestCase): 10 | 11 | def test_alexnet_backbone(self): 12 | """Test alexnet backbone.""" 13 | model = AlexNet(-1) 14 | model.train() 15 | 16 | imgs = torch.randn(1, 3, 256, 192) 17 | feat = model(imgs) 18 | self.assertIsInstance(feat, tuple) 19 | self.assertEqual(feat[-1].shape, (1, 256, 7, 5)) 20 | 21 | model = AlexNet(1) 22 | model.train() 23 | 24 | imgs = torch.randn(1, 3, 224, 224) 25 | feat = model(imgs) 26 | self.assertIsInstance(feat, tuple) 27 | self.assertEqual(feat[-1].shape, (1, 1)) 28 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/tests/test_models/test_backbones/test_mspn.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from unittest import TestCase 3 | 4 | import torch 5 | 6 | from mmpose.models.backbones import MSPN 7 | 8 | 9 | class TestMSPN(TestCase): 10 | 11 | def test_mspn_backbone(self): 12 | with self.assertRaises(AssertionError): 13 | # MSPN's num_stages should larger than 0 14 | MSPN(num_stages=0) 15 | with self.assertRaises(AssertionError): 16 | # MSPN's num_units should larger than 1 17 | MSPN(num_units=1) 18 | with self.assertRaises(AssertionError): 19 | # len(num_blocks) should equal num_units 20 | MSPN(num_units=2, num_blocks=[2, 2, 2]) 21 | 22 | # Test MSPN's outputs 23 | model = MSPN(num_stages=2, num_units=2, num_blocks=[2, 2]) 24 | model.init_weights() 25 | model.train() 26 | 27 | imgs = torch.randn(1, 3, 511, 511) 28 | feat = model(imgs) 29 | self.assertEqual(len(feat), 2) 30 | self.assertEqual(len(feat[0]), 2) 31 | self.assertEqual(len(feat[1]), 2) 32 | self.assertEqual(feat[0][0].shape, torch.Size([1, 256, 64, 64])) 33 | self.assertEqual(feat[0][1].shape, torch.Size([1, 256, 128, 128])) 34 | self.assertEqual(feat[1][0].shape, torch.Size([1, 256, 64, 64])) 35 | self.assertEqual(feat[1][1].shape, torch.Size([1, 256, 128, 128])) 36 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/tests/test_models/test_backbones/test_v2v_net.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from unittest import TestCase 3 | 4 | import torch 5 | 6 | from mmpose.models.backbones import V2VNet 7 | 8 | 9 | class TestV2Vnet(TestCase): 10 | 11 | def test_v2v_net(self): 12 | """Test V2VNet.""" 13 | model = V2VNet(input_channels=17, output_channels=15) 14 | input = torch.randn(2, 17, 32, 32, 32) 15 | output = model(input) 16 | self.assertIsInstance(output, tuple) 17 | self.assertEqual(output[-1].shape, (2, 15, 32, 32, 32)) 18 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/tests/test_models/test_losses/test_classification_losses.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | from unittest import TestCase 3 | 4 | import torch 5 | 6 | from mmpose.models.losses.classification_loss import InfoNCELoss 7 | 8 | 9 | class TestInfoNCELoss(TestCase): 10 | 11 | def test_loss(self): 12 | 13 | # test loss w/o target_weight 14 | loss = InfoNCELoss(temperature=0.05) 15 | 16 | fake_pred = torch.arange(5 * 2).reshape(5, 2).float() 17 | self.assertTrue( 18 | torch.allclose(loss(fake_pred), torch.tensor(5.4026), atol=1e-4)) 19 | 20 | # check if the value of temperature is positive 21 | with self.assertRaises(AssertionError): 22 | loss = InfoNCELoss(temperature=0.) 23 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/tools/analysis_tools/print_config.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) OpenMMLab. All rights reserved. 2 | import argparse 3 | 4 | from mmengine import Config, DictAction 5 | 6 | 7 | def parse_args(): 8 | parser = argparse.ArgumentParser(description='Print the whole config') 9 | parser.add_argument('config', help='config file path') 10 | parser.add_argument( 11 | '--options', nargs='+', action=DictAction, help='arguments in dict') 12 | args = parser.parse_args() 13 | 14 | return args 15 | 16 | 17 | def main(): 18 | args = parse_args() 19 | 20 | cfg = Config.fromfile(args.config) 21 | if args.options is not None: 22 | cfg.merge_from_dict(args.options) 23 | print(f'Config:\n{cfg.pretty_text}') 24 | 25 | 26 | if __name__ == '__main__': 27 | main() 28 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/tools/dist_test.sh: -------------------------------------------------------------------------------- 1 | #!/usr/bin/env bash 2 | # Copyright (c) OpenMMLab. All rights reserved. 3 | 4 | CONFIG=$1 5 | CHECKPOINT=$2 6 | GPUS=$3 7 | NNODES=${NNODES:-1} 8 | NODE_RANK=${NODE_RANK:-0} 9 | PORT=${PORT:-29500} 10 | MASTER_ADDR=${MASTER_ADDR:-"127.0.0.1"} 11 | 12 | PYTHONPATH="$(dirname $0)/..":$PYTHONPATH \ 13 | python -m torch.distributed.launch \ 14 | --nnodes=$NNODES \ 15 | --node_rank=$NODE_RANK \ 16 | --master_addr=$MASTER_ADDR \ 17 | --nproc_per_node=$GPUS \ 18 | --master_port=$PORT \ 19 | $(dirname "$0")/test.py \ 20 | $CONFIG \ 21 | $CHECKPOINT \ 22 | --launcher pytorch \ 23 | ${@:4} 24 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/tools/dist_train.sh: -------------------------------------------------------------------------------- 1 | #!/usr/bin/env bash 2 | # Copyright (c) OpenMMLab. All rights reserved. 3 | 4 | CONFIG=$1 5 | GPUS=$2 6 | NNODES=${NNODES:-1} 7 | NODE_RANK=${NODE_RANK:-0} 8 | PORT=${PORT:-29500} 9 | MASTER_ADDR=${MASTER_ADDR:-"127.0.0.1"} 10 | 11 | PYTHONPATH="$(dirname $0)/..":$PYTHONPATH \ 12 | python -m torch.distributed.launch \ 13 | --nnodes=$NNODES \ 14 | --node_rank=$NODE_RANK \ 15 | --master_addr=$MASTER_ADDR \ 16 | --nproc_per_node=$GPUS \ 17 | --master_port=$PORT \ 18 | step3_train_and_evaluation.py \ 19 | $CONFIG \ 20 | --launcher pytorch ${@:3} 21 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/tools/slurm_test.sh: -------------------------------------------------------------------------------- 1 | #!/usr/bin/env bash 2 | # Copyright (c) OpenMMLab. All rights reserved. 3 | 4 | set -x 5 | 6 | PARTITION=$1 7 | JOB_NAME=$2 8 | CONFIG=$3 9 | CHECKPOINT=$4 10 | GPUS=${GPUS:-8} 11 | GPUS_PER_NODE=${GPUS_PER_NODE:-8} 12 | CPUS_PER_TASK=${CPUS_PER_TASK:-5} 13 | PY_ARGS=${@:5} 14 | SRUN_ARGS=${SRUN_ARGS:-""} 15 | 16 | PYTHONPATH="$(dirname $0)/..":$PYTHONPATH \ 17 | srun -p ${PARTITION} \ 18 | --job-name=${JOB_NAME} \ 19 | --gres=gpu:${GPUS_PER_NODE} \ 20 | --ntasks=${GPUS} \ 21 | --ntasks-per-node=${GPUS_PER_NODE} \ 22 | --cpus-per-task=${CPUS_PER_TASK} \ 23 | --kill-on-bad-exit=1 \ 24 | ${SRUN_ARGS} \ 25 | python -u tools/test.py ${CONFIG} ${CHECKPOINT} --launcher="slurm" ${PY_ARGS} 26 | -------------------------------------------------------------------------------- /mmpose_package/mmpose/tools/slurm_train.sh: -------------------------------------------------------------------------------- 1 | #!/usr/bin/env bash 2 | # Copyright (c) OpenMMLab. All rights reserved. 3 | 4 | set -x 5 | 6 | PARTITION=$1 7 | JOB_NAME=$2 8 | CONFIG=$3 9 | WORK_DIR=$4 10 | GPUS=${GPUS:-8} 11 | GPUS_PER_NODE=${GPUS_PER_NODE:-8} 12 | CPUS_PER_TASK=${CPUS_PER_TASK:-5} 13 | SRUN_ARGS=${SRUN_ARGS:-""} 14 | PY_ARGS=${@:5} 15 | 16 | PYTHONPATH="$(dirname $0)/..":$PYTHONPATH \ 17 | srun -p ${PARTITION} \ 18 | --job-name=${JOB_NAME} \ 19 | --gres=gpu:${GPUS_PER_NODE} \ 20 | --ntasks=${GPUS} \ 21 | --ntasks-per-node=${GPUS_PER_NODE} \ 22 | --cpus-per-task=${CPUS_PER_TASK} \ 23 | --kill-on-bad-exit=1 \ 24 | ${SRUN_ARGS} \ 25 | python -u tools/train.py ${CONFIG} --work-dir=${WORK_DIR} --launcher="slurm" ${PY_ARGS} 26 | -------------------------------------------------------------------------------- /requirements.txt: -------------------------------------------------------------------------------- 1 | # =================================================================== 2 | # Do not change these packages !!! 3 | # Since you are using a Docker image that already has the PyTorch library installed, there is no need to reinstall Torch. 4 | # =================================================================== 5 | arrow==1.2.3 6 | binaryornot==0.4.4 7 | build==0.10.0 8 | certifi==2022.12.7 9 | chardet==5.1.0 10 | charset-normalizer==3.1.0 11 | click==8.1.3 12 | cookiecutter==2.1.1 13 | idna==3.4 14 | imageio[tifffile]==2.27.0 15 | jinja2==3.1.2 16 | jinja2-time==0.2.0 17 | joblib==1.2.0 18 | markupsafe==2.1.2 19 | numpy==1.21.6 20 | packaging==23.1 21 | pandas==1.3.5 22 | pillow==9.5.0 23 | pip-tools==6.13.0 24 | pyproject-hooks==1.0.0 25 | python-dateutil==2.8.2 26 | python-slugify==8.0.1 27 | pytz==2023.3 28 | pyyaml==6.0 29 | requests==2.28.2 30 | scikit-learn==1.0.2 31 | scipy==1.7.3 32 | simpleitk==2.2.1 33 | six==1.16.0 34 | text-unidecode==1.3 35 | threadpoolctl==3.1.0 36 | tifffile==2021.11.2 37 | tomli==2.0.1 38 | tzdata==2023.3 39 | urllib3==1.26.15 40 | wheel==0.40.0 41 | scikit-image==0.19.3 42 | evalutils==0.3.1 43 | 44 | # =================================================================== 45 | # If you have other additional dependencies, please list them below. 46 | # =================================================================== 47 | -------------------------------------------------------------------------------- /step1_test_mmpose.py: -------------------------------------------------------------------------------- 1 | # !/usr/bin/env python 2 | # -*- coding:utf-8 -*- 3 | 4 | import mmpose 5 | print(mmpose.__version__) --------------------------------------------------------------------------------