├── CITATION.cff ├── CONTRIBUTING.md ├── LICENSE ├── MANIFEST.in ├── README.md ├── README.zh-CN.md ├── data ├── data.yaml ├── data2.yaml └── v7.yaml ├── datasets ├── data.yaml └── v8.yaml ├── doc ├── README.md ├── build_docs.py ├── build_reference.py ├── mkdocs.yml ├── mkdocs_ar.yml ├── mkdocs_de.yml ├── mkdocs_es.yml ├── mkdocs_fr.yml ├── mkdocs_hi.yml ├── mkdocs_ja.yml ├── mkdocs_ko.yml ├── mkdocs_pt.yml ├── mkdocs_ru.yml ├── mkdocs_zh.yml ├── update_translations.py └── v8.yaml ├── docker ├── Dockerfile ├── Dockerfile-arm64 ├── Dockerfile-conda ├── Dockerfile-cpu ├── Dockerfile-jetson ├── Dockerfile-python ├── Dockerfile-runner └── v5.yaml ├── example ├── README.md ├── YOLOv8-CPP-Inference │ ├── CMakeLists.txt │ ├── README.md │ ├── inference.cpp │ ├── inference.h │ └── main.cpp ├── YOLOv8-ONNXRuntime-CPP │ ├── CMakeLists.txt │ ├── README.md │ ├── inference.cpp │ ├── inference.h │ └── main.cpp ├── YOLOv8-ONNXRuntime-Rust │ ├── Cargo.toml │ ├── README.md │ └── src │ │ ├── cli.rs │ │ ├── lib.rs │ │ ├── main.rs │ │ ├── model.rs │ │ ├── ort_backend.rs │ │ └── yolo_result.rs ├── YOLOv8-ONNXRuntime │ ├── README.md │ └── main.py ├── YOLOv8-OpenCV-ONNX-Python │ ├── README.md │ └── main.py ├── YOLOv8-Region-Counter │ ├── readme.md │ └── yolov8_region_counter.py ├── YOLOv8-SAHI-Inference-Video │ ├── readme.md │ └── yolov8_sahi.py ├── YOLOv8-Segmentation-ONNXRuntime-Python │ ├── README.md │ └── main.py ├── hub.ipynb ├── tutorial.ipynb └── v4.yaml ├── main.py ├── requirements.txt ├── setup.cfg ├── setup.py └── ultralytics ├── __init__.py ├── assets ├── bus.jpg └── v9.yaml ├── cfg ├── __init__.py ├── default.yaml ├── exp1.yaml ├── models │ ├── README.md │ ├── rt-detr │ │ ├── rtdetr-l.yaml │ │ ├── rtdetr-resnet101.yaml │ │ ├── rtdetr-resnet50.yaml │ │ └── rtdetr-x.yaml │ ├── v3 │ │ ├── yolov3-spp.yaml │ │ ├── yolov3-tiny.yaml │ │ └── yolov3.yaml │ ├── v5 │ │ ├── v5.yaml │ │ ├── yolov5-p6.yaml │ │ └── yolov5.yaml │ ├── v6 │ │ └── yolov6.yaml │ ├── v8 │ │ ├── YOLOv8-AsDDet.yaml │ │ ├── YOLOv8-DynamicHead.yaml │ │ ├── YOLOv8-ImplicitHead.yaml │ │ ├── yolov8-ACmix1.yaml │ │ ├── yolov8-ACmix2.yaml │ │ ├── yolov8-ALSS.yaml │ │ ├── yolov8-ALSS1.yaml │ │ ├── yolov8-ASF-YOLO1.yaml │ │ ├── yolov8-ASF-YOLO2.yaml │ │ ├── yolov8-C2f-AdditiveBlock-CGLU.yaml │ │ ├── yolov8-C2f-DeepDBB.yaml │ │ ├── yolov8-C2f-SHSA-CGLU.yaml │ │ ├── yolov8-C2f-VSS.yaml │ │ ├── yolov8-C2fMLLABlock1.yaml │ │ ├── yolov8-C2fMLLABlock2.yaml │ │ ├── yolov8-C2f_AdditiveBlock.yaml │ │ ├── yolov8-C2f_AdditiveBlock2.yaml │ │ ├── yolov8-C2f_DUAL.yaml │ │ ├── yolov8-C2f_EFAttention.yaml │ │ ├── yolov8-C2f_GhostModule_DynamicConv.yaml │ │ ├── yolov8-C2f_MSBlock1.yaml │ │ ├── yolov8-C2f_MSBlock2.yaml │ │ ├── yolov8-C2f_MSBlock3.yaml │ │ ├── yolov8-C2f_StarNB.yaml │ │ ├── yolov8-C2f_UIB.yaml │ │ ├── yolov8-CACS_C2f.yaml │ │ ├── yolov8-CCFM-Dyhead.yaml │ │ ├── yolov8-CCFM.yaml │ │ ├── yolov8-CGA.yaml │ │ ├── yolov8-CGAFusion.yaml │ │ ├── yolov8-CLLAHead.yaml │ │ ├── yolov8-Conv_SWS.yaml │ │ ├── yolov8-DCNv3LKA1.yaml │ │ ├── yolov8-DCNv3LKA2.yaml │ │ ├── yolov8-DCNv3LKA3.yaml │ │ ├── yolov8-DenseOne.yaml │ │ ├── yolov8-DetectDeepDBB.yaml │ │ ├── yolov8-Detect_AFPN.yaml │ │ ├── yolov8-Detect_AFPN3.yaml │ │ ├── yolov8-Detect_ASFF.yaml │ │ ├── yolov8-Detect_Adown.yaml │ │ ├── yolov8-Detect_DBB.yaml │ │ ├── yolov8-Detect_DySnakeConv.yaml │ │ ├── yolov8-Detect_FASFF.yaml │ │ ├── yolov8-Detect_Partial_C.yaml │ │ ├── yolov8-Detect_SC_C.yaml │ │ ├── yolov8-Detect_dyhead3.yaml │ │ ├── yolov8-Detect_improve_s.yaml │ │ ├── yolov8-FFA_Net.yaml │ │ ├── yolov8-FRMHead.yaml │ │ ├── yolov8-FasterNet.yaml │ │ ├── yolov8-FeaturePyramidSharedConv.yaml │ │ ├── yolov8-FreqFFPN.yaml │ │ ├── yolov8-GELAN-high.yaml │ │ ├── yolov8-GELAN-low.yaml │ │ ├── yolov8-GFPN.yaml │ │ ├── yolov8-HATHead.yaml │ │ ├── yolov8-JD.yaml │ │ ├── yolov8-LADH.yaml │ │ ├── yolov8-LAE.yaml │ │ ├── yolov8-LSDECD.yaml │ │ ├── yolov8-LSKA.yaml │ │ ├── yolov8-MLCA1.yaml │ │ ├── yolov8-MLCA2.yaml │ │ ├── yolov8-MLCA3.yaml │ │ ├── yolov8-MLLA1.yaml │ │ ├── yolov8-MLLA2.yaml │ │ ├── yolov8-MLLA3.yaml │ │ ├── yolov8-MSFN.yaml │ │ ├── yolov8-MSFN1.yaml │ │ ├── yolov8-MoCAttention.yaml │ │ ├── yolov8-MoCAttention2.yaml │ │ ├── yolov8-MobileNetV3.yaml │ │ ├── yolov8-ModulatedDeformConv.yaml │ │ ├── yolov8-MousePose.yaml │ │ ├── yolov8-NAM-1.yaml │ │ ├── yolov8-NAM-2.yaml │ │ ├── yolov8-NAM-3.yaml │ │ ├── yolov8-NAM-4.yaml │ │ ├── yolov8-NAM-5.yaml │ │ ├── yolov8-PSA.yaml │ │ ├── yolov8-PSA2.yaml │ │ ├── yolov8-PSA3.yaml │ │ ├── yolov8-RCS-OSA1.yaml │ │ ├── yolov8-RCS-OSA2.yaml │ │ ├── yolov8-ReCalibrationFPN-P2345.yaml │ │ ├── yolov8-ReCalibrationFPN-P345.yaml │ │ ├── yolov8-ReCalibrationFPN-P3456.yaml │ │ ├── yolov8-RepFPN.yaml │ │ ├── yolov8-RepHead.yaml │ │ ├── yolov8-SCDown.yaml │ │ ├── yolov8-SCDown2.yaml │ │ ├── yolov8-SCSA.yaml │ │ ├── yolov8-SK.yaml │ │ ├── yolov8-SOEP.yaml │ │ ├── yolov8-SWS.yaml │ │ ├── yolov8-VanillaNet.yaml │ │ ├── yolov8-att.yaml │ │ ├── yolov8-c2f.yaml │ │ ├── yolov8-cls.yaml │ │ ├── yolov8-dysample-SCAM.yaml │ │ ├── yolov8-emo.py │ │ ├── yolov8-focus.yaml │ │ ├── yolov8-ghost-p2.yaml │ │ ├── yolov8-ghost-p6.yaml │ │ ├── yolov8-ghost.yaml │ │ ├── yolov8-leyolonano.yaml │ │ ├── yolov8-p2.yaml │ │ ├── yolov8-p6.yaml │ │ ├── yolov8-pose-p6.yaml │ │ ├── yolov8-pose.yaml │ │ ├── yolov8-rtdetr.yaml │ │ ├── yolov8-s-Detect_improve.yaml │ │ ├── yolov8-seg-p6.yaml │ │ ├── yolov8-seg.yaml │ │ ├── yolov8-starnet1.yaml │ │ ├── yolov8-starnet2.yaml │ │ ├── yolov8-starnet3.yaml │ │ ├── yolov8-starnet4.yaml │ │ ├── yolov8.yaml │ │ ├── yolov8_BasicRFB.yaml │ │ ├── yolov8_C2f-DySnakeConv.yaml │ │ ├── yolov8_C2f_Heat.yaml │ │ ├── yolov8_C2f_KAN.yaml │ │ ├── yolov8_C2f_deformable_LKA1.yaml │ │ ├── yolov8_CA.yaml │ │ ├── yolov8_CMUNeXt1.yaml │ │ ├── yolov8_CMUNeXt2.yaml │ │ ├── yolov8_CPMS.yaml │ │ ├── yolov8_EMA.yaml │ │ ├── yolov8_EMA2.yaml │ │ ├── yolov8_EMA3.yaml │ │ ├── yolov8_GFPN.yaml │ │ ├── yolov8_MCALayer.yaml │ │ ├── yolov8_MCALayer1.yaml │ │ ├── yolov8_ResBlock_CBAM_GhostConv.yaml │ │ ├── yolov8_SPPF_LSKA.yaml │ │ ├── yolov8_SPPF_LSKA2.yaml │ │ ├── yolov8_TripletAttention.yaml │ │ ├── yolov8_alpha.yaml │ │ ├── yolov8_alpha_C2f_DCN2.yaml │ │ ├── yolov8_alpha_DeepBiFPN.yaml │ │ └── yolov8jct.yaml │ └── yolov8-dysample-SCAM.yaml └── trackers │ ├── botsort.yaml │ └── bytetrack.yaml ├── data ├── __init__.py ├── annotator.py ├── augment.py ├── base.py ├── build.py ├── converter.py ├── dataset.py ├── loaders.py ├── scripts │ ├── download_weights.sh │ ├── get_coco.sh │ ├── get_coco128.sh │ └── get_imagenet.sh └── utils.py ├── engine ├── __init__.py ├── exporter.py ├── model.py ├── predictor.py ├── results.py ├── trainer.py ├── tuner.py └── validator.py ├── hub ├── __init__.py ├── auth.py ├── session.py └── utils.py ├── models ├── __init__.py ├── fastsam │ ├── __init__.py │ ├── model.py │ ├── predict.py │ ├── prompt.py │ ├── utils.py │ └── val.py ├── nas │ ├── __init__.py │ ├── model.py │ ├── predict.py │ └── val.py ├── rtdetr │ ├── __init__.py │ ├── model.py │ ├── predict.py │ ├── train.py │ └── val.py ├── sam │ ├── __init__.py │ ├── amg.py │ ├── build.py │ ├── model.py │ ├── modules │ │ ├── __init__.py │ │ ├── decoders.py │ │ ├── encoders.py │ │ ├── sam.py │ │ ├── tiny_encoder.py │ │ └── transformer.py │ └── predict.py ├── utils │ ├── __init__.py │ ├── loss.py │ └── ops.py └── yolo │ ├── __init__.py │ ├── classify │ ├── __init__.py │ ├── predict.py │ ├── train.py │ └── val.py │ ├── detect │ ├── __init__.py │ ├── predict.py │ ├── train.py │ └── val.py │ ├── model.py │ ├── pose │ ├── __init__.py │ ├── predict.py │ ├── train.py │ └── val.py │ └── segment │ ├── __init__.py │ ├── predict.py │ ├── train.py │ └── val.py ├── nn ├── DCNv4_op │ ├── DCNv4 │ │ ├── __init__.py │ │ ├── functions │ │ │ ├── __init__.py │ │ │ ├── dcnv4_func.py │ │ │ ├── flash_deform_attn_func.py │ │ │ └── table.py │ │ └── modules │ │ │ ├── __init__.py │ │ │ ├── dcnv4.py │ │ │ └── flash_deform_attn.py │ ├── MANIFEST.in │ ├── __init__.py │ ├── make.sh │ ├── scripts │ │ ├── find_best.py │ │ ├── search_bwd.sh │ │ ├── search_dcnv4.py │ │ ├── search_dcnv4_bwd.py │ │ ├── search_dcnv4_bwd_engine.py │ │ ├── search_dcnv4_engine.py │ │ ├── search_fwd.sh │ │ ├── test_dcnv4.py │ │ ├── test_dcnv4_bwd.py │ │ ├── test_flash_deform_attn.py │ │ └── test_flash_deform_attn_backward.py │ ├── setup.py │ └── src │ │ ├── cuda │ │ ├── common.h │ │ ├── dcnv4_col2im_cuda.cuh │ │ ├── dcnv4_cuda.cu │ │ ├── dcnv4_cuda.h │ │ ├── dcnv4_im2col_cuda.cuh │ │ ├── flash_deform_attn_cuda.cu │ │ ├── flash_deform_attn_cuda.h │ │ ├── flash_deform_col2im_cuda.cuh │ │ └── flash_deform_im2col_cuda.cuh │ │ ├── dcnv4.h │ │ └── vision.cpp ├── DEANet.py ├── RCSOSA.py ├── RepNCSPELAN4.py ├── __init__.py ├── autobackend.py ├── dyhead2.py ├── head │ ├── head_improve.py │ └── v8.yaml ├── mamba │ ├── AUTHORS │ ├── LICENSE │ ├── README.md │ ├── assets │ │ ├── selection.png │ │ └── ssd_algorithm.png │ ├── benchmarks │ │ └── benchmark_generation_mamba_simple.py │ ├── csrc │ │ └── selective_scan │ │ │ ├── reverse_scan.cuh │ │ │ ├── selective_scan.cpp │ │ │ ├── selective_scan.h │ │ │ ├── selective_scan_bwd_bf16_complex.cu │ │ │ ├── selective_scan_bwd_bf16_real.cu │ │ │ ├── selective_scan_bwd_fp16_complex.cu │ │ │ ├── selective_scan_bwd_fp16_real.cu │ │ │ ├── selective_scan_bwd_fp32_complex.cu │ │ │ ├── selective_scan_bwd_fp32_real.cu │ │ │ ├── selective_scan_bwd_kernel.cuh │ │ │ ├── selective_scan_common.h │ │ │ ├── selective_scan_fwd_bf16.cu │ │ │ ├── selective_scan_fwd_fp16.cu │ │ │ ├── selective_scan_fwd_fp32.cu │ │ │ ├── selective_scan_fwd_kernel.cuh │ │ │ ├── static_switch.h │ │ │ └── uninitialized_copy.cuh │ ├── evals │ │ └── lm_harness_eval.py │ ├── make.sh │ ├── mamba_ssm │ │ ├── __init__.py │ │ ├── distributed │ │ │ ├── __init__.py │ │ │ ├── distributed_utils.py │ │ │ └── tensor_parallel.py │ │ ├── models │ │ │ ├── __init__.py │ │ │ ├── config_mamba.py │ │ │ └── mixer_seq_simple.py │ │ ├── modules │ │ │ ├── __init__.py │ │ │ ├── block.py │ │ │ ├── mamba2.py │ │ │ ├── mamba2_simple.py │ │ │ ├── mamba_simple.py │ │ │ ├── mha.py │ │ │ ├── mlp.py │ │ │ └── ssd_minimal.py │ │ ├── ops │ │ │ ├── __init__.py │ │ │ ├── selective_scan_interface.py │ │ │ └── triton │ │ │ │ ├── __init__.py │ │ │ │ ├── k_activations.py │ │ │ │ ├── layer_norm.py │ │ │ │ ├── layernorm_gated.py │ │ │ │ ├── selective_state_update.py │ │ │ │ ├── ssd_bmm.py │ │ │ │ ├── ssd_chunk_scan.py │ │ │ │ ├── ssd_chunk_state.py │ │ │ │ ├── ssd_combined.py │ │ │ │ └── ssd_state_passing.py │ │ └── utils │ │ │ ├── __init__.py │ │ │ ├── generation.py │ │ │ └── hf.py │ ├── setup.py │ └── tests │ │ └── ops │ │ ├── test_selective_scan.py │ │ └── triton │ │ └── test_selective_state_update.py ├── modules │ ├── ACmix.py │ ├── AFPN.py │ ├── AFPNHead3.py │ ├── ALSS_YOLO.py │ ├── ASFFHead.py │ ├── ASFYOLO.py │ ├── AdditiveBlock_CGLU.py │ ├── AsDDet.py │ ├── C2f_GhostModule_DynamicConv.py │ ├── CACSYOLO.py │ ├── CAS_ViT.py │ ├── CLLAHead.py │ ├── CMUNeXt.py │ ├── CPMS.py │ ├── DCNv3LKA.py │ ├── DEANet.py │ ├── DSConvHead.py │ ├── DetectDeepDBB.py │ ├── Detect_Adown.py │ ├── Detect_DBB.py │ ├── Detect_LADH.py │ ├── Detect_LSDECD.py │ ├── Detect_Partial_C.py │ ├── Detect_SC_C.py │ ├── Detect_improve_s.py │ ├── DualConv.py │ ├── DynamicHead.py │ ├── FASFFHead.py │ ├── FFANet.py │ ├── FRMHead.py │ ├── FasterNet.py │ ├── FreqFusion.py │ ├── GFPN.py │ ├── GFPN │ │ └── GFPN.py │ ├── HATHead.py │ ├── HCANet.py │ ├── ImplicitHead.py │ ├── LADH.py │ ├── LANet.py │ ├── LSKAttention.py │ ├── LSMYOLO.py │ ├── LargeSeparableKernelAttention.py │ ├── LeYOLO.py │ ├── MDeformConv.py │ ├── MLCA.py │ ├── MLLA.py │ ├── MLLABlock.py │ ├── MobileNetV3.py │ ├── PSA.py │ ├── RFAConvHead.py │ ├── RepConvHead.py │ ├── RepHead.py │ ├── SCSA.py │ ├── SWS.py │ ├── VanillaNet.py │ ├── YOLOJD.py │ ├── YOLOMousePose.py │ ├── __init__.py │ ├── att.py │ ├── block.py │ ├── ca.py │ ├── conv.py │ ├── deconv.py │ ├── denseone.py │ ├── dyhead.py │ ├── dyhead2.py │ ├── dysampleSCAM.py │ ├── emo.py │ ├── head.py │ ├── head_improve.py │ ├── kan_conv.py │ ├── mamba_vss.py │ ├── mcattn.py │ ├── mobilenetv4block.py │ ├── rep_block.py │ ├── repfpn.py │ ├── starnet.py │ ├── transformer.py │ ├── utils.py │ └── vHeat.py ├── neck │ └── GFPN.py ├── ops_dcnv3 │ └── dist │ │ └── DCNv3-1.1-cp38-cp38-linux_x86_64.whl ├── rfb.py ├── tasks.py └── yolov10.py ├── solutions ├── __init__.py ├── ai_gym.py ├── heatmap.py └── object_counter.py ├── trackers ├── README.md ├── __init__.py ├── basetrack.py ├── bot_sort.py ├── byte_tracker.py ├── track.py └── utils │ ├── __init__.py │ ├── gmc.py │ ├── kalman_filter.py │ └── matching.py └── utils ├── __init__.py ├── autobatch.py ├── benchmarks.py ├── callbacks ├── __init__.py ├── base.py ├── clearml.py ├── comet.py ├── dvc.py ├── hub.py ├── mlflow.py ├── neptune.py ├── raytune.py ├── tensorboard.py └── wb.py ├── checks.py ├── dist.py ├── downloads.py ├── errors.py ├── files.py ├── instance.py ├── loss-origin.py ├── loss.py ├── metrics-origin.py ├── metrics.py ├── ops.py ├── patches.py ├── plotting.py ├── tal-origin.py ├── tal.py ├── torch_utils.py ├── triton.py └── tuner.py /CITATION.cff: -------------------------------------------------------------------------------- 1 | cff-version: 1.2.0 2 | preferred-citation: 3 | type: software 4 | message: If you use this software, please cite it as below. 5 | authors: 6 | - family-names: Jocher 7 | given-names: Glenn 8 | orcid: "https://orcid.org/0000-0001-5950-6979" 9 | - family-names: Chaurasia 10 | given-names: Ayush 11 | orcid: "https://orcid.org/0000-0002-7603-6750" 12 | - family-names: Qiu 13 | given-names: Jing 14 | orcid: "https://orcid.org/0000-0003-3783-7069" 15 | title: "YOLO by Ultralytics" 16 | version: 8.0.0 17 | # doi: 10.5281/zenodo.3908559 # TODO 18 | date-released: 2023-1-10 19 | license: AGPL-3.0 20 | url: "https://github.com/ultralytics/ultralytics" 21 | -------------------------------------------------------------------------------- /MANIFEST.in: -------------------------------------------------------------------------------- 1 | include *.md 2 | include requirements.txt 3 | include LICENSE 4 | include setup.py 5 | include ultralytics/assets/bus.jpg 6 | include ultralytics/assets/zidane.jpg 7 | include tests/*.py 8 | recursive-include ultralytics *.yaml 9 | -------------------------------------------------------------------------------- /data/data.yaml: -------------------------------------------------------------------------------- 1 | # dataset path 2 | path:/content/improved-v8/ 3 | train: ./data/images/train 4 | val: ./data/images/val 5 | test: ./data/images/test 6 | 7 | # number of classes 8 | nc: 5 9 | 10 | # class names 11 | names: ['D00','D10','D20','D40','Repair'] 12 | -------------------------------------------------------------------------------- /data/data2.yaml: -------------------------------------------------------------------------------- 1 | # dataset path 2 | path:/kaggle/working/ 3 | train: ./data/images/train 4 | val: ./data/images/val 5 | test: ./data/images/test 6 | 7 | # number of classes 8 | nc: 5 9 | 10 | # class names 11 | names: ['D00','D10','D20','D40','Repair'] 12 | -------------------------------------------------------------------------------- /data/v7.yaml: -------------------------------------------------------------------------------- 1 | v7 2 | -------------------------------------------------------------------------------- /datasets/data.yaml: -------------------------------------------------------------------------------- 1 | # dataset path 2 | path: /content/improved-v8/ 3 | train: ./data/images/train 4 | val: ./data/images/val 5 | test: ./data/images/test 6 | 7 | # number of classes 8 | nc: 5 9 | 10 | # class names 11 | names: ['D00','D10','D20','D40','Repair'] 12 | -------------------------------------------------------------------------------- /datasets/v8.yaml: -------------------------------------------------------------------------------- 1 | v8 2 | -------------------------------------------------------------------------------- /doc/v8.yaml: -------------------------------------------------------------------------------- 1 | yyyyyyyyyyyyyyyyyyyyyy 2 | -------------------------------------------------------------------------------- /docker/Dockerfile-conda: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # Builds ultralytics/ultralytics:latest-conda image on DockerHub https://hub.docker.com/r/ultralytics/ultralytics 3 | # Image is optimized for Ultralytics Anaconda (https://anaconda.org/conda-forge/ultralytics) installation and usage 4 | 5 | # Start FROM miniconda3 image https://hub.docker.com/r/continuumio/miniconda3 6 | FROM continuumio/miniconda3:latest 7 | 8 | # Downloads to user config dir 9 | ADD https://ultralytics.com/assets/Arial.ttf https://ultralytics.com/assets/Arial.Unicode.ttf /root/.config/Ultralytics/ 10 | 11 | # Install linux packages 12 | RUN apt update \ 13 | && apt install --no-install-recommends -y libgl1 14 | 15 | # Copy contents 16 | ADD https://github.com/ultralytics/assets/releases/download/v0.0.0/yolov8n.pt . 17 | 18 | # Install conda packages 19 | # mkl required to fix 'OSError: libmkl_intel_lp64.so.2: cannot open shared object file: No such file or directory' 20 | RUN conda config --set solver libmamba && \ 21 | conda install pytorch torchvision pytorch-cuda=11.8 -c pytorch -c nvidia && \ 22 | conda install -c conda-forge ultralytics mkl 23 | # conda install -c pytorch -c nvidia -c conda-forge pytorch torchvision pytorch-cuda=11.8 ultralytics mkl 24 | 25 | 26 | # Usage Examples ------------------------------------------------------------------------------------------------------- 27 | 28 | # Build and Push 29 | # t=ultralytics/ultralytics:latest-conda && sudo docker build -f docker/Dockerfile-cpu -t $t . && sudo docker push $t 30 | 31 | # Run 32 | # t=ultralytics/ultralytics:latest-conda && sudo docker run -it --ipc=host $t 33 | 34 | # Pull and Run 35 | # t=ultralytics/ultralytics:latest-conda && sudo docker pull $t && sudo docker run -it --ipc=host $t 36 | 37 | # Pull and Run with local volume mounted 38 | # t=ultralytics/ultralytics:latest-conda && sudo docker pull $t && sudo docker run -it --ipc=host -v "$(pwd)"/datasets:/usr/src/datasets $t 39 | -------------------------------------------------------------------------------- /docker/Dockerfile-runner: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # Builds GitHub actions CI runner image for deployment to DockerHub https://hub.docker.com/r/ultralytics/ultralytics 3 | # Image is CUDA-optimized for YOLOv8 single/multi-GPU training and inference tests 4 | 5 | # Start FROM Ultralytics GPU image 6 | FROM ultralytics/ultralytics:latest 7 | 8 | # Set the working directory 9 | WORKDIR /actions-runner 10 | 11 | # Download and unpack the latest runner from https://github.com/actions/runner 12 | RUN FILENAME=actions-runner-linux-x64-2.309.0.tar.gz && \ 13 | curl -o $FILENAME -L https://github.com/actions/runner/releases/download/v2.309.0/$FILENAME && \ 14 | tar xzf $FILENAME && \ 15 | rm $FILENAME 16 | 17 | # Install runner dependencies 18 | ENV RUNNER_ALLOW_RUNASROOT=1 19 | ENV DEBIAN_FRONTEND=noninteractive 20 | RUN ./bin/installdependencies.sh && \ 21 | apt-get -y install libicu-dev 22 | 23 | # Inline ENTRYPOINT command to configure and start runner with default TOKEN and NAME 24 | ENTRYPOINT sh -c './config.sh --url https://github.com/ultralytics/ultralytics \ 25 | --token ${GITHUB_RUNNER_TOKEN:-TOKEN} \ 26 | --name ${GITHUB_RUNNER_NAME:-NAME} \ 27 | --labels gpu-latest \ 28 | --replace && \ 29 | ./run.sh' 30 | 31 | 32 | # Usage Examples ------------------------------------------------------------------------------------------------------- 33 | 34 | # Build and Push 35 | # t=ultralytics/ultralytics:latest-runner && sudo docker build -f docker/Dockerfile-runner -t $t . && sudo docker push $t 36 | 37 | # Pull and Run in detached mode with access to GPUs 0 and 1 38 | # t=ultralytics/ultralytics:latest-runner && sudo docker run -d -e GITHUB_RUNNER_TOKEN=TOKEN -e GITHUB_RUNNER_NAME=NAME --ipc=host --gpus '"device=0,1"' $t 39 | -------------------------------------------------------------------------------- /docker/v5.yaml: -------------------------------------------------------------------------------- 1 | yyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyy 2 | -------------------------------------------------------------------------------- /example/YOLOv8-CPP-Inference/CMakeLists.txt: -------------------------------------------------------------------------------- 1 | cmake_minimum_required(VERSION 3.5) 2 | 3 | project(Yolov8CPPInference VERSION 0.1) 4 | 5 | set(CMAKE_INCLUDE_CURRENT_DIR ON) 6 | 7 | # CUDA 8 | set(CUDA_TOOLKIT_ROOT_DIR "/usr/local/cuda") 9 | find_package(CUDA 11 REQUIRED) 10 | 11 | set(CMAKE_CUDA_STANDARD 11) 12 | set(CMAKE_CUDA_STANDARD_REQUIRED ON) 13 | # !CUDA 14 | 15 | # OpenCV 16 | find_package(OpenCV REQUIRED) 17 | include_directories(${OpenCV_INCLUDE_DIRS}) 18 | # !OpenCV 19 | 20 | set(PROJECT_SOURCES 21 | main.cpp 22 | 23 | inference.h 24 | inference.cpp 25 | ) 26 | 27 | add_executable(Yolov8CPPInference ${PROJECT_SOURCES}) 28 | target_link_libraries(Yolov8CPPInference ${OpenCV_LIBS}) 29 | -------------------------------------------------------------------------------- /example/YOLOv8-CPP-Inference/README.md: -------------------------------------------------------------------------------- 1 | # YOLOv8/YOLOv5 Inference C++ 2 | 3 | This example demonstrates how to perform inference using YOLOv8 and YOLOv5 models in C++ with OpenCV's DNN API. 4 | 5 | ## Usage 6 | 7 | ```bash 8 | git clone ultralytics 9 | cd ultralytics 10 | pip install . 11 | cd examples/YOLOv8-CPP-Inference 12 | 13 | # Add a **yolov8\_.onnx** and/or **yolov5\_.onnx** model(s) to the ultralytics folder. 14 | # Edit the **main.cpp** to change the **projectBasePath** to match your user. 15 | 16 | # Note that by default the CMake file will try and import the CUDA library to be used with the OpenCVs dnn (cuDNN) GPU Inference. 17 | # If your OpenCV build does not use CUDA/cuDNN you can remove that import call and run the example on CPU. 18 | 19 | mkdir build 20 | cd build 21 | cmake .. 22 | make 23 | ./Yolov8CPPInference 24 | ``` 25 | 26 | ## Exporting YOLOv8 and YOLOv5 Models 27 | 28 | To export YOLOv8 models: 29 | 30 | ```commandline 31 | yolo export model=yolov8s.pt imgsz=480,640 format=onnx opset=12 32 | ``` 33 | 34 | To export YOLOv5 models: 35 | 36 | ```commandline 37 | python3 export.py --weights yolov5s.pt --img 480 640 --include onnx --opset 12 38 | ``` 39 | 40 | yolov8s.onnx: 41 | 42 | ![image](https://user-images.githubusercontent.com/40023722/217356132-a4cecf2e-2729-4acb-b80a-6559022d7707.png) 43 | 44 | yolov5s.onnx: 45 | 46 | ![image](https://user-images.githubusercontent.com/40023722/217357005-07464492-d1da-42e3-98a7-fc753f87d5e6.png) 47 | 48 | This repository utilizes OpenCV's DNN API to run ONNX exported models of YOLOv5 and YOLOv8. In theory, it should work for YOLOv6 and YOLOv7 as well, but they have not been tested. Note that the example networks are exported with rectangular (640x480) resolutions, but any exported resolution will work. You may want to use the letterbox approach for square images, depending on your use case. 49 | 50 | The **main** branch version uses Qt as a GUI wrapper. The primary focus here is the **Inference** class file, which demonstrates how to transpose YOLOv8 models to work as YOLOv5 models. 51 | -------------------------------------------------------------------------------- /example/YOLOv8-ONNXRuntime-Rust/Cargo.toml: -------------------------------------------------------------------------------- 1 | [package] 2 | name = "yolov8-rs" 3 | version = "0.1.0" 4 | edition = "2021" 5 | 6 | # See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html 7 | 8 | [dependencies] 9 | clap = { version = "4.2.4", features = ["derive"] } 10 | image = { version = "0.24.7", default-features = false, features = ["jpeg", "png", "webp-encoder"] } 11 | imageproc = { version = "0.23.0", default-features = false } 12 | ndarray = { version = "0.15.6" } 13 | ort = {version = "1.16.3", default-features = false, features = ["load-dynamic", "copy-dylibs", "half"]} 14 | rusttype = { version = "0.9", default-features = false } 15 | anyhow = { version = "1.0.75"} 16 | regex = { version = "1.5.4" } 17 | rand = { version ="0.8.5" } 18 | chrono = { version = "0.4.30" } 19 | half = { version = "2.3.1" } 20 | dirs = { version = "5.0.1" } 21 | ureq = { version = "2.9.1" } 22 | -------------------------------------------------------------------------------- /example/YOLOv8-ONNXRuntime-Rust/src/main.rs: -------------------------------------------------------------------------------- 1 | use clap::Parser; 2 | 3 | use yolov8_rs::{Args, YOLOv8}; 4 | 5 | fn main() -> Result<(), Box> { 6 | let args = Args::parse(); 7 | 8 | // 1. load image 9 | let x = image::io::Reader::open(&args.source)? 10 | .with_guessed_format()? 11 | .decode()?; 12 | 13 | // 2. model support dynamic batch inference, so input should be a Vec 14 | let xs = vec![x]; 15 | 16 | // You can test `--batch 2` with this 17 | // let xs = vec![x.clone(), x]; 18 | 19 | // 3. build yolov8 model 20 | let mut model = YOLOv8::new(args)?; 21 | model.summary(); // model info 22 | 23 | // 4. run 24 | let ys = model.run(&xs)?; 25 | println!("{:?}", ys); 26 | 27 | Ok(()) 28 | } 29 | -------------------------------------------------------------------------------- /example/YOLOv8-ONNXRuntime/README.md: -------------------------------------------------------------------------------- 1 | # YOLOv8 - ONNX Runtime 2 | 3 | This project implements YOLOv8 using ONNX Runtime. 4 | 5 | ## Installation 6 | 7 | To run this project, you need to install the required dependencies. The following instructions will guide you through the installation process. 8 | 9 | ### Installing Required Dependencies 10 | 11 | You can install the required dependencies by running the following command: 12 | 13 | ```bash 14 | pip install -r requirements.txt 15 | ``` 16 | 17 | ### Installing `onnxruntime-gpu` 18 | 19 | If you have an NVIDIA GPU and want to leverage GPU acceleration, you can install the onnxruntime-gpu package using the following command: 20 | 21 | ```bash 22 | pip install onnxruntime-gpu 23 | ``` 24 | 25 | Note: Make sure you have the appropriate GPU drivers installed on your system. 26 | 27 | ### Installing `onnxruntime` (CPU version) 28 | 29 | If you don't have an NVIDIA GPU or prefer to use the CPU version of onnxruntime, you can install the onnxruntime package using the following command: 30 | 31 | ```bash 32 | pip install onnxruntime 33 | ``` 34 | 35 | ### Usage 36 | 37 | After successfully installing the required packages, you can run the YOLOv8 implementation using the following command: 38 | 39 | ```bash 40 | python main.py --model yolov8n.onnx --img image.jpg --conf-thres 0.5 --iou-thres 0.5 41 | ``` 42 | 43 | Make sure to replace yolov8n.onnx with the path to your YOLOv8 ONNX model file, image.jpg with the path to your input image, and adjust the confidence threshold (conf-thres) and IoU threshold (iou-thres) values as needed. 44 | -------------------------------------------------------------------------------- /example/YOLOv8-OpenCV-ONNX-Python/README.md: -------------------------------------------------------------------------------- 1 | # YOLOv8 - OpenCV 2 | 3 | Implementation YOLOv8 on OpenCV using ONNX Format. 4 | 5 | Just simply clone and run 6 | 7 | ```bash 8 | pip install -r requirements.txt 9 | python main.py --model yolov8n.onnx --img image.jpg 10 | ``` 11 | 12 | If you start from scratch: 13 | 14 | ```bash 15 | pip install ultralytics 16 | yolo export model=yolov8n.pt imgsz=640 format=onnx opset=12 17 | ``` 18 | 19 | _\*Make sure to include "opset=12"_ 20 | -------------------------------------------------------------------------------- /example/v4.yaml: -------------------------------------------------------------------------------- 1 | yyyyyyy 2 | -------------------------------------------------------------------------------- /main.py: -------------------------------------------------------------------------------- 1 | from ultralytics import YOLO 2 | 3 | # 安装命令 4 | # python setup.py develop 5 | 6 | # 数据集示例百度云链接 7 | # 链接：https://pan.baidu.com/s/19FM7XnKEFC83vpiRdtNA8A?pwd=n93i 8 | # 提取码：n93i 9 | 10 | if __name__ == '__main__': 11 | # # 直接使用预训练模型创建模型. 12 | # model = YOLO('yolov8n.pt') 13 | # model.train(**{'cfg':'ultralytics/cfg/exp1.yaml', 'data':'dataset/data.yaml'}) 14 | 15 | # # 使用yaml配置文件来创建模型,并导入预训练权重. 16 | # model = YOLO('ultralytics/cfg/models/v8/yolov8.yaml') 17 | # model.load('yolov8n.pt') 18 | # model.train(**{'cfg':'ultralytics/cfg/exp1.yaml', 'data':'dataset/data.yaml'}) 19 | 20 | # # 模型验证 21 | # model = YOLO('runs/detect/yolov8n_exp/weights/best.pt') 22 | # model.val(**{'data':'dataset/data.yaml'}) 23 | 24 | # # 模型推理 25 | # model = YOLO('runs/detect/train53/weights/best.pt') 26 | # model.predict(source='datasets/test/images', **{'save':True}) 27 | 28 | model = YOLO('ultralytics/cfg/models/v8/yolov8_EMA.yaml') 29 | model.load('yolov8n.pt') 30 | # model = YOLO('yolov8n.pt') 31 | model.train(**{'cfg':'ultralytics/cfg/exp1.yaml', 'data':'datasets/data.yaml'}) 32 | 33 | # model = YOLO('ultralytics/cfg/models/v8/yolov8.yaml') 34 | # model.train(data='datasets/data.yaml') 35 | -------------------------------------------------------------------------------- /requirements.txt: -------------------------------------------------------------------------------- 1 | # Ultralytics requirements 2 | # Example: pip install -r requirements.txt 3 | 4 | # Base ---------------------------------------- 5 | matplotlib>=3.3.0 6 | numpy>=1.22.2 # pinned by Snyk to avoid a vulnerability 7 | opencv-python>=4.6.0 8 | pillow>=7.1.2 9 | pyyaml>=5.3.1 10 | requests>=2.23.0 11 | scipy>=1.4.1 12 | torch>=1.8.0 13 | torchvision>=0.9.0 14 | tqdm>=4.64.0 15 | 16 | # Logging ------------------------------------- 17 | # tensorboard>=2.13.0 18 | # dvclive>=2.12.0 19 | # clearml 20 | # comet 21 | 22 | # Plotting ------------------------------------ 23 | pandas>=1.1.4 24 | seaborn>=0.11.0 25 | 26 | # Export -------------------------------------- 27 | # coremltools>=7.0 # CoreML export 28 | # onnx>=1.12.0 # ONNX export 29 | # onnxsim>=0.4.1 # ONNX simplifier 30 | # nvidia-pyindex # TensorRT export 31 | # nvidia-tensorrt # TensorRT export 32 | # scikit-learn==0.19.2 # CoreML quantization 33 | # tensorflow>=2.4.1,<=2.13.1 # TF exports (-cpu, -aarch64, -macos) 34 | # tflite-support 35 | # tensorflowjs>=3.9.0 # TF.js export 36 | # openvino-dev>=2023.0 # OpenVINO export 37 | 38 | # Extras -------------------------------------- 39 | psutil # system utilization 40 | py-cpuinfo # display CPU info 41 | thop>=0.1.1 # FLOPs computation 42 | # ipython # interactive notebook 43 | # albumentations>=1.0.3 # training augmentations 44 | # pycocotools>=2.0.6 # COCO mAP 45 | # roboflow 46 | -------------------------------------------------------------------------------- /ultralytics/__init__.py: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | 3 | __version__ = '8.0.227' 4 | 5 | from ultralytics.models import RTDETR, SAM, YOLO 6 | from ultralytics.models.fastsam import FastSAM 7 | from ultralytics.models.nas import NAS 8 | from ultralytics.utils import SETTINGS as settings 9 | from ultralytics.utils.checks import check_yolo as checks 10 | from ultralytics.utils.downloads import download 11 | 12 | __all__ = '__version__', 'YOLO', 'NAS', 'SAM', 'FastSAM', 'RTDETR', 'checks', 'download', 'settings' 13 | -------------------------------------------------------------------------------- /ultralytics/assets/bus.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/scuzyq/improved-v8/839f704f400296d49a18746c660b4089ff41f46b/ultralytics/assets/bus.jpg -------------------------------------------------------------------------------- /ultralytics/assets/v9.yaml: -------------------------------------------------------------------------------- 1 | v9 2 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/rt-detr/rtdetr-resnet101.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # RT-DETR-ResNet101 object detection model with P3-P5 outputs. 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n-cls.yaml' will call yolov8-cls.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | l: [1.00, 1.00, 1024] 9 | 10 | backbone: 11 | # [from, repeats, module, args] 12 | - [-1, 1, ResNetLayer, [3, 64, 1, True, 1]] # 0 13 | - [-1, 1, ResNetLayer, [64, 64, 1, False, 3]] # 1 14 | - [-1, 1, ResNetLayer, [256, 128, 2, False, 4]] # 2 15 | - [-1, 1, ResNetLayer, [512, 256, 2, False, 23]] # 3 16 | - [-1, 1, ResNetLayer, [1024, 512, 2, False, 3]] # 4 17 | 18 | head: 19 | - [-1, 1, Conv, [256, 1, 1, None, 1, 1, False]] # 5 20 | - [-1, 1, AIFI, [1024, 8]] 21 | - [-1, 1, Conv, [256, 1, 1]] # 7 22 | 23 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 24 | - [3, 1, Conv, [256, 1, 1, None, 1, 1, False]] # 9 25 | - [[-2, -1], 1, Concat, [1]] 26 | - [-1, 3, RepC3, [256]] # 11 27 | - [-1, 1, Conv, [256, 1, 1]] # 12 28 | 29 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 30 | - [2, 1, Conv, [256, 1, 1, None, 1, 1, False]] # 14 31 | - [[-2, -1], 1, Concat, [1]] # cat backbone P4 32 | - [-1, 3, RepC3, [256]] # X3 (16), fpn_blocks.1 33 | 34 | - [-1, 1, Conv, [256, 3, 2]] # 17, downsample_convs.0 35 | - [[-1, 12], 1, Concat, [1]] # cat Y4 36 | - [-1, 3, RepC3, [256]] # F4 (19), pan_blocks.0 37 | 38 | - [-1, 1, Conv, [256, 3, 2]] # 20, downsample_convs.1 39 | - [[-1, 7], 1, Concat, [1]] # cat Y5 40 | - [-1, 3, RepC3, [256]] # F5 (22), pan_blocks.1 41 | 42 | - [[16, 19, 22], 1, RTDETRDecoder, [nc]] # Detect(P3, P4, P5) 43 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/rt-detr/rtdetr-resnet50.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # RT-DETR-ResNet50 object detection model with P3-P5 outputs. 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n-cls.yaml' will call yolov8-cls.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | l: [1.00, 1.00, 1024] 9 | 10 | backbone: 11 | # [from, repeats, module, args] 12 | - [-1, 1, ResNetLayer, [3, 64, 1, True, 1]] # 0 13 | - [-1, 1, ResNetLayer, [64, 64, 1, False, 3]] # 1 14 | - [-1, 1, ResNetLayer, [256, 128, 2, False, 4]] # 2 15 | - [-1, 1, ResNetLayer, [512, 256, 2, False, 6]] # 3 16 | - [-1, 1, ResNetLayer, [1024, 512, 2, False, 3]] # 4 17 | 18 | head: 19 | - [-1, 1, Conv, [256, 1, 1, None, 1, 1, False]] # 5 20 | - [-1, 1, AIFI, [1024, 8]] 21 | - [-1, 1, Conv, [256, 1, 1]] # 7 22 | 23 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 24 | - [3, 1, Conv, [256, 1, 1, None, 1, 1, False]] # 9 25 | - [[-2, -1], 1, Concat, [1]] 26 | - [-1, 3, RepC3, [256]] # 11 27 | - [-1, 1, Conv, [256, 1, 1]] # 12 28 | 29 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 30 | - [2, 1, Conv, [256, 1, 1, None, 1, 1, False]] # 14 31 | - [[-2, -1], 1, Concat, [1]] # cat backbone P4 32 | - [-1, 3, RepC3, [256]] # X3 (16), fpn_blocks.1 33 | 34 | - [-1, 1, Conv, [256, 3, 2]] # 17, downsample_convs.0 35 | - [[-1, 12], 1, Concat, [1]] # cat Y4 36 | - [-1, 3, RepC3, [256]] # F4 (19), pan_blocks.0 37 | 38 | - [-1, 1, Conv, [256, 3, 2]] # 20, downsample_convs.1 39 | - [[-1, 7], 1, Concat, [1]] # cat Y5 40 | - [-1, 3, RepC3, [256]] # F5 (22), pan_blocks.1 41 | 42 | - [[16, 19, 22], 1, RTDETRDecoder, [nc]] # Detect(P3, P4, P5) 43 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v3/yolov3-spp.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv3-SPP object detection model with P3-P5 outputs. For details see https://docs.ultralytics.com/models/yolov3 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | depth_multiple: 1.0 # model depth multiple 7 | width_multiple: 1.0 # layer channel multiple 8 | 9 | # darknet53 backbone 10 | backbone: 11 | # [from, number, module, args] 12 | [[-1, 1, Conv, [32, 3, 1]], # 0 13 | [-1, 1, Conv, [64, 3, 2]], # 1-P1/2 14 | [-1, 1, Bottleneck, [64]], 15 | [-1, 1, Conv, [128, 3, 2]], # 3-P2/4 16 | [-1, 2, Bottleneck, [128]], 17 | [-1, 1, Conv, [256, 3, 2]], # 5-P3/8 18 | [-1, 8, Bottleneck, [256]], 19 | [-1, 1, Conv, [512, 3, 2]], # 7-P4/16 20 | [-1, 8, Bottleneck, [512]], 21 | [-1, 1, Conv, [1024, 3, 2]], # 9-P5/32 22 | [-1, 4, Bottleneck, [1024]], # 10 23 | ] 24 | 25 | # YOLOv3-SPP head 26 | head: 27 | [[-1, 1, Bottleneck, [1024, False]], 28 | [-1, 1, SPP, [512, [5, 9, 13]]], 29 | [-1, 1, Conv, [1024, 3, 1]], 30 | [-1, 1, Conv, [512, 1, 1]], 31 | [-1, 1, Conv, [1024, 3, 1]], # 15 (P5/32-large) 32 | 33 | [-2, 1, Conv, [256, 1, 1]], 34 | [-1, 1, nn.Upsample, [None, 2, 'nearest']], 35 | [[-1, 8], 1, Concat, [1]], # cat backbone P4 36 | [-1, 1, Bottleneck, [512, False]], 37 | [-1, 1, Bottleneck, [512, False]], 38 | [-1, 1, Conv, [256, 1, 1]], 39 | [-1, 1, Conv, [512, 3, 1]], # 22 (P4/16-medium) 40 | 41 | [-2, 1, Conv, [128, 1, 1]], 42 | [-1, 1, nn.Upsample, [None, 2, 'nearest']], 43 | [[-1, 6], 1, Concat, [1]], # cat backbone P3 44 | [-1, 1, Bottleneck, [256, False]], 45 | [-1, 2, Bottleneck, [256, False]], # 27 (P3/8-small) 46 | 47 | [[27, 22, 15], 1, Detect, [nc]], # Detect(P3, P4, P5) 48 | ] 49 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v3/yolov3-tiny.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv3-tiny object detection model with P4-P5 outputs. For details see https://docs.ultralytics.com/models/yolov3 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | depth_multiple: 1.0 # model depth multiple 7 | width_multiple: 1.0 # layer channel multiple 8 | 9 | # YOLOv3-tiny backbone 10 | backbone: 11 | # [from, number, module, args] 12 | [[-1, 1, Conv, [16, 3, 1]], # 0 13 | [-1, 1, nn.MaxPool2d, [2, 2, 0]], # 1-P1/2 14 | [-1, 1, Conv, [32, 3, 1]], 15 | [-1, 1, nn.MaxPool2d, [2, 2, 0]], # 3-P2/4 16 | [-1, 1, Conv, [64, 3, 1]], 17 | [-1, 1, nn.MaxPool2d, [2, 2, 0]], # 5-P3/8 18 | [-1, 1, Conv, [128, 3, 1]], 19 | [-1, 1, nn.MaxPool2d, [2, 2, 0]], # 7-P4/16 20 | [-1, 1, Conv, [256, 3, 1]], 21 | [-1, 1, nn.MaxPool2d, [2, 2, 0]], # 9-P5/32 22 | [-1, 1, Conv, [512, 3, 1]], 23 | [-1, 1, nn.ZeroPad2d, [[0, 1, 0, 1]]], # 11 24 | [-1, 1, nn.MaxPool2d, [2, 1, 0]], # 12 25 | ] 26 | 27 | # YOLOv3-tiny head 28 | head: 29 | [[-1, 1, Conv, [1024, 3, 1]], 30 | [-1, 1, Conv, [256, 1, 1]], 31 | [-1, 1, Conv, [512, 3, 1]], # 15 (P5/32-large) 32 | 33 | [-2, 1, Conv, [128, 1, 1]], 34 | [-1, 1, nn.Upsample, [None, 2, 'nearest']], 35 | [[-1, 8], 1, Concat, [1]], # cat backbone P4 36 | [-1, 1, Conv, [256, 3, 1]], # 19 (P4/16-medium) 37 | 38 | [[19, 15], 1, Detect, [nc]], # Detect(P4, P5) 39 | ] 40 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v3/yolov3.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv3 object detection model with P3-P5 outputs. For details see https://docs.ultralytics.com/models/yolov3 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | depth_multiple: 1.0 # model depth multiple 7 | width_multiple: 1.0 # layer channel multiple 8 | 9 | # darknet53 backbone 10 | backbone: 11 | # [from, number, module, args] 12 | [[-1, 1, Conv, [32, 3, 1]], # 0 13 | [-1, 1, Conv, [64, 3, 2]], # 1-P1/2 14 | [-1, 1, Bottleneck, [64]], 15 | [-1, 1, Conv, [128, 3, 2]], # 3-P2/4 16 | [-1, 2, Bottleneck, [128]], 17 | [-1, 1, Conv, [256, 3, 2]], # 5-P3/8 18 | [-1, 8, Bottleneck, [256]], 19 | [-1, 1, Conv, [512, 3, 2]], # 7-P4/16 20 | [-1, 8, Bottleneck, [512]], 21 | [-1, 1, Conv, [1024, 3, 2]], # 9-P5/32 22 | [-1, 4, Bottleneck, [1024]], # 10 23 | ] 24 | 25 | # YOLOv3 head 26 | head: 27 | [[-1, 1, Bottleneck, [1024, False]], 28 | [-1, 1, Conv, [512, 1, 1]], 29 | [-1, 1, Conv, [1024, 3, 1]], 30 | [-1, 1, Conv, [512, 1, 1]], 31 | [-1, 1, Conv, [1024, 3, 1]], # 15 (P5/32-large) 32 | 33 | [-2, 1, Conv, [256, 1, 1]], 34 | [-1, 1, nn.Upsample, [None, 2, 'nearest']], 35 | [[-1, 8], 1, Concat, [1]], # cat backbone P4 36 | [-1, 1, Bottleneck, [512, False]], 37 | [-1, 1, Bottleneck, [512, False]], 38 | [-1, 1, Conv, [256, 1, 1]], 39 | [-1, 1, Conv, [512, 3, 1]], # 22 (P4/16-medium) 40 | 41 | [-2, 1, Conv, [128, 1, 1]], 42 | [-1, 1, nn.Upsample, [None, 2, 'nearest']], 43 | [[-1, 6], 1, Concat, [1]], # cat backbone P3 44 | [-1, 1, Bottleneck, [256, False]], 45 | [-1, 2, Bottleneck, [256, False]], # 27 (P3/8-small) 46 | 47 | [[27, 22, 15], 1, Detect, [nc]], # Detect(P3, P4, P5) 48 | ] 49 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v5/v5.yaml: -------------------------------------------------------------------------------- 1 | yyyyyyyyyyyyyyyyyyyyyyyyyy 2 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v5/yolov5.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv5 object detection model with P3-P5 outputs. For details see https://docs.ultralytics.com/models/yolov5 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov5n.yaml' will call yolov5.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] 9 | s: [0.33, 0.50, 1024] 10 | m: [0.67, 0.75, 1024] 11 | l: [1.00, 1.00, 1024] 12 | x: [1.33, 1.25, 1024] 13 | 14 | # YOLOv5 v6.0 backbone 15 | backbone: 16 | # [from, number, module, args] 17 | [[-1, 1, Conv, [64, 6, 2, 2]], # 0-P1/2 18 | [-1, 1, Conv, [128, 3, 2]], # 1-P2/4 19 | [-1, 3, C3, [128]], 20 | [-1, 1, Conv, [256, 3, 2]], # 3-P3/8 21 | [-1, 6, C3, [256]], 22 | [-1, 1, Conv, [512, 3, 2]], # 5-P4/16 23 | [-1, 9, C3, [512]], 24 | [-1, 1, Conv, [1024, 3, 2]], # 7-P5/32 25 | [-1, 3, C3, [1024]], 26 | [-1, 1, SPPF, [1024, 5]], # 9 27 | ] 28 | 29 | # YOLOv5 v6.0 head 30 | head: 31 | [[-1, 1, Conv, [512, 1, 1]], 32 | [-1, 1, nn.Upsample, [None, 2, 'nearest']], 33 | [[-1, 6], 1, Concat, [1]], # cat backbone P4 34 | [-1, 3, C3, [512, False]], # 13 35 | 36 | [-1, 1, Conv, [256, 1, 1]], 37 | [-1, 1, nn.Upsample, [None, 2, 'nearest']], 38 | [[-1, 4], 1, Concat, [1]], # cat backbone P3 39 | [-1, 3, C3, [256, False]], # 17 (P3/8-small) 40 | 41 | [-1, 1, Conv, [256, 3, 2]], 42 | [[-1, 14], 1, Concat, [1]], # cat head P4 43 | [-1, 3, C3, [512, False]], # 20 (P4/16-medium) 44 | 45 | [-1, 1, Conv, [512, 3, 2]], 46 | [[-1, 10], 1, Concat, [1]], # cat head P5 47 | [-1, 3, C3, [1024, False]], # 23 (P5/32-large) 48 | 49 | [[17, 20, 23], 1, Detect, [nc]], # Detect(P3, P4, P5) 50 | ] 51 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v6/yolov6.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv6 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/models/yolov6 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | activation: nn.ReLU() # (optional) model default activation function 7 | scales: # model compound scaling constants, i.e. 'model=yolov6n.yaml' will call yolov8.yaml with scale 'n' 8 | # [depth, width, max_channels] 9 | n: [0.33, 0.25, 1024] 10 | s: [0.33, 0.50, 1024] 11 | m: [0.67, 0.75, 768] 12 | l: [1.00, 1.00, 512] 13 | x: [1.00, 1.25, 512] 14 | 15 | # YOLOv6-3.0s backbone 16 | backbone: 17 | # [from, repeats, module, args] 18 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 19 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 20 | - [-1, 6, Conv, [128, 3, 1]] 21 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 22 | - [-1, 12, Conv, [256, 3, 1]] 23 | - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16 24 | - [-1, 18, Conv, [512, 3, 1]] 25 | - [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32 26 | - [-1, 6, Conv, [1024, 3, 1]] 27 | - [-1, 1, SPPF, [1024, 5]] # 9 28 | 29 | # YOLOv6-3.0s head 30 | head: 31 | - [-1, 1, Conv, [256, 1, 1]] 32 | - [-1, 1, nn.ConvTranspose2d, [256, 2, 2, 0]] 33 | - [[-1, 6], 1, Concat, [1]] # cat backbone P4 34 | - [-1, 1, Conv, [256, 3, 1]] 35 | - [-1, 9, Conv, [256, 3, 1]] # 14 36 | 37 | - [-1, 1, Conv, [128, 1, 1]] 38 | - [-1, 1, nn.ConvTranspose2d, [128, 2, 2, 0]] 39 | - [[-1, 4], 1, Concat, [1]] # cat backbone P3 40 | - [-1, 1, Conv, [128, 3, 1]] 41 | - [-1, 9, Conv, [128, 3, 1]] # 19 42 | 43 | - [-1, 1, Conv, [128, 3, 2]] 44 | - [[-1, 15], 1, Concat, [1]] # cat head P4 45 | - [-1, 1, Conv, [256, 3, 1]] 46 | - [-1, 9, Conv, [256, 3, 1]] # 23 47 | 48 | - [-1, 1, Conv, [256, 3, 2]] 49 | - [[-1, 10], 1, Concat, [1]] # cat head P5 50 | - [-1, 1, Conv, [512, 3, 1]] 51 | - [-1, 9, Conv, [512, 3, 1]] # 27 52 | 53 | - [[19, 23, 27], 1, Detect, [nc]] # Detect(P3, P4, P5) 54 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-C2f-SHSA-CGLU.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOPs 13 | 14 | # YOLOv8.0n backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 18 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 19 | - [-1, 3, C2f, [128, True]] 20 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 21 | - [-1, 6, C2f, [256, True]] 22 | - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16 23 | - [-1, 6, C2f_SHSA_CGLU, [512, True]] 24 | - [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32 25 | - [-1, 3, C2f_SHSA_CGLU, [1024, True]] 26 | - [-1, 1, SPPF, [1024, 5]] # 9 27 | 28 | # YOLOv8.0n head 29 | head: 30 | - [-1, 1, nn.Upsample, [None, 2, "nearest"]] 31 | - [[-1, 6], 1, Concat, [1]] # cat backbone P4 32 | - [-1, 3, C2f, [512]] # 12 33 | 34 | - [-1, 1, nn.Upsample, [None, 2, "nearest"]] 35 | - [[-1, 4], 1, Concat, [1]] # cat backbone P3 36 | - [-1, 3, C2f, [256]] # 15 (P3/8-small) 37 | 38 | - [-1, 1, Conv, [256, 3, 2]] 39 | - [[-1, 12], 1, Concat, [1]] # cat head P4 40 | - [-1, 3, C2f, [512]] # 18 (P4/16-medium) 41 | 42 | - [-1, 1, Conv, [512, 3, 2]] 43 | - [[-1, 9], 1, Concat, [1]] # cat head P5 44 | - [-1, 3, C2f, [1024]] # 21 (P5/32-large) 45 | 46 | - [[15, 18, 21], 1, Detect, [nc]] # Detect(P3, P4, P5) 47 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-C2f-VSS.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOPs 13 | 14 | # YOLOv8.0n backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 18 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 19 | - [-1, 3, C2f_VSS, [128, True]] 20 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 21 | - [-1, 6, C2f_VSS, [256, True]] 22 | - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16 23 | - [-1, 6, C2f_VSS, [512, True]] 24 | - [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32 25 | - [-1, 3, C2f_VSS, [1024, True]] 26 | - [-1, 1, SPPF, [1024, 5]] # 9 27 | 28 | # YOLOv8.0n head 29 | head: 30 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 31 | - [[-1, 6], 1, Concat, [1]] # cat backbone P4 32 | - [-1, 3, C2f, [512]] # 12 33 | 34 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 35 | - [[-1, 4], 1, Concat, [1]] # cat backbone P3 36 | - [-1, 3, C2f, [256]] # 15 (P3/8-small) 37 | 38 | - [-1, 1, Conv, [256, 3, 2]] 39 | - [[-1, 12], 1, Concat, [1]] # cat head P4 40 | - [-1, 3, C2f, [512]] # 18 (P4/16-medium) 41 | 42 | - [-1, 1, Conv, [512, 3, 2]] 43 | - [[-1, 9], 1, Concat, [1]] # cat head P5 44 | - [-1, 3, C2f, [1024]] # 21 (P5/32-large) 45 | 46 | - [[15, 18, 21], 1, Detect, [nc]] # Detect(P3, P4, P5) 47 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-C2f_EFAttention.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOPs 13 | 14 | # YOLOv8.0n backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 18 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 19 | - [-1, 3, C2f, [128, True]] 20 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 21 | - [-1, 6, C2f, [256, True]] 22 | - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16 23 | - [-1, 6, C2f, [512, True]] 24 | - [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32 25 | - [-1, 3, C2f, [1024, True]] 26 | - [-1, 1, SPPF, [1024, 5]] # 9 27 | 28 | # YOLOv8.0n head 29 | head: 30 | - [-1, 1, nn.Upsample, [None, 2, "nearest"]] 31 | - [[-1, 6], 1, Concat, [1]] # cat backbone P4 32 | - [-1, 3, C2f_EFAttention, [512]] # 12 33 | 34 | - [-1, 1, nn.Upsample, [None, 2, "nearest"]] 35 | - [[-1, 4], 1, Concat, [1]] # cat backbone P3 36 | - [-1, 3, C2f_EFAttention, [256]] # 15 (P3/8-small) 37 | 38 | - [-1, 1, Conv, [256, 3, 2]] 39 | - [[-1, 12], 1, Concat, [1]] # cat head P4 40 | - [-1, 3, C2f_EFAttention, [512]] # 18 (P4/16-medium) 41 | 42 | - [-1, 1, Conv, [512, 3, 2]] 43 | - [[-1, 9], 1, Concat, [1]] # cat head P5 44 | - [-1, 3, C2f_EFAttention, [1024]] # 21 (P5/32-large) 45 | 46 | - [[15, 18, 21], 1, Detect, [nc]] # Detect(P3, P4, P5) 47 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-C2f_MSBlock2.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOPs 13 | 14 | # YOLOv8.0n backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 18 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 19 | - [-1, 3, C2f, [128, True]] 20 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 21 | - [-1, 6, C2f, [256, True]] 22 | - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16 23 | - [-1, 6, C2f, [512, True]] 24 | - [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32 25 | - [-1, 3, C2f, [1024, True]] 26 | - [-1, 1, SPPF, [1024, 5]] # 9 27 | 28 | # YOLOv8.0n head 29 | head: 30 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 31 | - [[-1, 6], 1, Concat, [1]] # cat backbone P4 32 | - [-1, 3, C2f_MSBlock, [512]] # 12 33 | 34 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 35 | - [[-1, 4], 1, Concat, [1]] # cat backbone P3 36 | - [-1, 3, C2f_MSBlock, [256]] # 15 (P3/8-small) 37 | 38 | - [-1, 1, Conv, [256, 3, 2]] 39 | - [[-1, 12], 1, Concat, [1]] # cat head P4 40 | - [-1, 3, C2f_MSBlock, [512]] # 18 (P4/16-medium) 41 | 42 | - [-1, 1, Conv, [512, 3, 2]] 43 | - [[-1, 9], 1, Concat, [1]] # cat head P5 44 | - [-1, 3, C2f_MSBlock, [1024]] # 21 (P5/32-large) 45 | 46 | - [[15, 18, 21], 1, Detect, [nc]] # Detect(P3, P4, P5) 47 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-C2f_MSBlock3.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOPs 13 | 14 | # YOLOv8.0n backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 18 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 19 | - [-1, 3, C2f, [128, True]] 20 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 21 | - [-1, 6, C2f, [256, True]] 22 | - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16 23 | - [-1, 6, C2f, [512, True]] 24 | - [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32 25 | - [-1, 3, C2f, [1024, True]] 26 | - [-1, 1, SPPF, [1024, 5]] # 9 27 | 28 | # YOLOv8.0n head 29 | head: 30 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 31 | - [[-1, 6], 1, Concat, [1]] # cat backbone P4 32 | - [-1, 3, C2f, [512]] # 12 33 | 34 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 35 | - [[-1, 4], 1, Concat, [1]] # cat backbone P3 36 | - [-1, 3, C2f_MSBlock, [256]] # 15 (P3/8-small) 37 | 38 | - [-1, 1, Conv, [256, 3, 2]] 39 | - [[-1, 12], 1, Concat, [1]] # cat head P4 40 | - [-1, 3, C2f_MSBlock, [512]] # 18 (P4/16-medium) 41 | 42 | - [-1, 1, Conv, [512, 3, 2]] 43 | - [[-1, 9], 1, Concat, [1]] # cat head P5 44 | - [-1, 3, C2f_MSBlock, [1024]] # 21 (P5/32-large) 45 | 46 | - [[15, 18, 21], 1, Detect, [nc]] # Detect(P3, P4, P5) 47 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-C2f_StarNB.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOPs 13 | 14 | # YOLOv8.0n backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 18 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 19 | - [-1, 3, C2f_StarNB, [128, True]] 20 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 21 | - [-1, 6, C2f_StarNB, [256, True]] 22 | - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16 23 | - [-1, 6, C2f_StarNB, [512, True]] 24 | - [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32 25 | - [-1, 3, C2f_StarNB, [1024, True]] 26 | - [-1, 1, SPPF, [1024, 5]] # 9 27 | 28 | # YOLOv8.0n head 29 | head: 30 | - [-1, 1, nn.Upsample, [None, 2, "nearest"]] 31 | - [[-1, 6], 1, Concat, [1]] # cat backbone P4 32 | - [-1, 3, C2f, [512]] # 12 33 | 34 | - [-1, 1, nn.Upsample, [None, 2, "nearest"]] 35 | - [[-1, 4], 1, Concat, [1]] # cat backbone P3 36 | - [-1, 3, C2f, [256]] # 15 (P3/8-small) 37 | 38 | - [-1, 1, Conv, [256, 3, 2]] 39 | - [[-1, 12], 1, Concat, [1]] # cat head P4 40 | - [-1, 3, C2f, [512]] # 18 (P4/16-medium) 41 | 42 | - [-1, 1, Conv, [512, 3, 2]] 43 | - [[-1, 9], 1, Concat, [1]] # cat head P5 44 | - [-1, 3, C2f, [1024]] # 21 (P5/32-large) 45 | 46 | - [[15, 18, 21], 1, Detect, [nc]] # Detect(P3, P4, P5) 47 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-C2f_UIB.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOPs 13 | 14 | # YOLOv8.0n backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 18 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 19 | - [-1, 3, C2f, [128, True]] 20 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 21 | - [-1, 6, C2f, [256, True]] 22 | - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16 23 | - [-1, 6, C2f, [512, True]] 24 | - [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32 25 | - [-1, 3, C2f, [1024, True]] 26 | - [-1, 1, SPPF, [1024, 5]] # 9 27 | 28 | # YOLOv8.0n head 29 | head: 30 | - [-1, 1, nn.Upsample, [None, 2, "nearest"]] 31 | - [[-1, 6], 1, Concat, [1]] # cat backbone P4 32 | - [-1, 3, C2f_UIB, [512]] # 12 33 | 34 | - [-1, 1, nn.Upsample, [None, 2, "nearest"]] 35 | - [[-1, 4], 1, Concat, [1]] # cat backbone P3 36 | - [-1, 3, C2f_UIB, [256]] # 15 (P3/8-small) 37 | 38 | - [-1, 1, Conv, [256, 3, 2]] 39 | - [[-1, 12], 1, Concat, [1]] # cat head P4 40 | - [-1, 3, C2f_UIB, [512]] # 18 (P4/16-medium) 41 | 42 | - [-1, 1, Conv, [512, 3, 2]] 43 | - [[-1, 9], 1, Concat, [1]] # cat head P5 44 | - [-1, 3, C2f_UIB, [1024]] # 21 (P5/32-large) 45 | 46 | - [[15, 18, 21], 1, Detect, [nc]] # Detect(P3, P4, P5) 47 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-CACS_C2f.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOPs 13 | 14 | # YOLOv8.0n backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 18 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 19 | - [-1, 3, CACS_C2f, [128, True]] 20 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 21 | - [-1, 6, CACS_C2f, [256, True]] 22 | - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16 23 | - [-1, 6, CACS_C2f, [512, True]] 24 | - [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32 25 | - [-1, 3, CACS_C2f, [1024, True]] 26 | - [-1, 1, SPPF, [1024, 5]] # 9 27 | 28 | # YOLOv8.0n head 29 | head: 30 | - [-1, 1, nn.Upsample, [None, 2, "nearest"]] 31 | - [[-1, 6], 1, Concat, [1]] # cat backbone P4 32 | - [-1, 3, C2f, [512]] # 12 33 | 34 | - [-1, 1, nn.Upsample, [None, 2, "nearest"]] 35 | - [[-1, 4], 1, Concat, [1]] # cat backbone P3 36 | - [-1, 3, C2f, [256]] # 15 (P3/8-small) 37 | 38 | - [-1, 1, Conv, [256, 3, 2]] 39 | - [[-1, 12], 1, Concat, [1]] # cat head P4 40 | - [-1, 3, C2f, [512]] # 18 (P4/16-medium) 41 | 42 | - [-1, 1, Conv, [512, 3, 2]] 43 | - [[-1, 9], 1, Concat, [1]] # cat head P5 44 | - [-1, 3, C2f, [1024]] # 21 (P5/32-large) 45 | 46 | - [[15, 18, 21], 1, Detect, [nc]] # Detect(P3, P4, P5) 47 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-Conv_SWS.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOPs 13 | 14 | # YOLOv8.0n backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 18 | - [-1, 1, Conv_SWS, [128, 3, 2]] # 1-P2/4 19 | - [-1, 3, C2f, [128, True]] 20 | - [-1, 1, Conv_SWS, [256, 3, 2]] # 3-P3/8 21 | - [-1, 6, C2f, [256, True]] 22 | - [-1, 1, Conv_SWS, [512, 3, 2]] # 5-P4/16 23 | - [-1, 6, C2f, [512, True]] 24 | - [-1, 1, Conv_SWS, [1024, 3, 2]] # 7-P5/32 25 | - [-1, 3, C2f, [1024, True]] 26 | - [-1, 1, SPPF, [1024, 5]] # 9 27 | 28 | # YOLOv8.0n head 29 | head: 30 | - [-1, 1, nn.Upsample, [None, 2, "nearest"]] 31 | - [[-1, 6], 1, Concat, [1]] # cat backbone P4 32 | - [-1, 3, C2f, [512]] # 12 33 | 34 | - [-1, 1, nn.Upsample, [None, 2, "nearest"]] 35 | - [[-1, 4], 1, Concat, [1]] # cat backbone P3 36 | - [-1, 3, C2f, [256]] # 15 (P3/8-small) 37 | 38 | - [-1, 1, Conv_SWS, [256, 3, 2]] 39 | - [[-1, 12], 1, Concat, [1]] # cat head P4 40 | - [-1, 3, C2f, [512]] # 18 (P4/16-medium) 41 | 42 | - [-1, 1, Conv_SWS, [512, 3, 2]] 43 | - [[-1, 9], 1, Concat, [1]] # cat head P5 44 | - [-1, 3, C2f, [1024]] # 21 (P5/32-large) 45 | 46 | - [[15, 18, 21], 1, Detect, [nc]] # Detect(P3, P4, P5) 47 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-DenseOne.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOPs 13 | 14 | # YOLOv8.0n backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 18 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 19 | - [-1, 3, C2f, [128, True]] 20 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 21 | - [-1, 6, C2f, [256, True]] 22 | - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16 23 | - [-1, 6, C2f, [512, True]] 24 | - [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32 25 | - [-1, 3, DenseOne, [1024, 32]] 26 | - [-1, 1, SPPF, [1024, 5]] # 9 27 | 28 | # YOLOv8.0n head 29 | head: 30 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 31 | - [[-1, 6], 1, Concat, [1]] # cat backbone P4 32 | - [-1, 3, C2f, [512]] # 12 33 | 34 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 35 | - [[-1, 4], 1, Concat, [1]] # cat backbone P3 36 | - [-1, 3, C2f, [256]] # 15 (P3/8-small) 37 | 38 | - [-1, 1, Conv, [256, 3, 2]] 39 | - [[-1, 12], 1, Concat, [1]] # cat head P4 40 | - [-1, 3, C2f, [512]] # 18 (P4/16-medium) 41 | 42 | - [-1, 1, Conv, [512, 3, 2]] 43 | - [[-1, 9], 1, Concat, [1]] # cat head P5 44 | - [-1, 3, C2f, [1024]] # 21 (P5/32-large) 45 | 46 | - [[15, 18, 21], 1, Detect, [nc]] # Detect(P3, P4, P5) 47 | 48 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-DetectDeepDBB.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOPs 13 | 14 | # YOLOv8.0n backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 18 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 19 | - [-1, 3, C2f, [128, True]] 20 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 21 | - [-1, 6, C2f, [256, True]] 22 | - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16 23 | - [-1, 6, C2f, [512, True]] 24 | - [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32 25 | - [-1, 3, C2f, [1024, True]] 26 | - [-1, 1, SPPF, [1024, 5]] # 9 27 | 28 | # YOLOv8.0n head 29 | head: 30 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 31 | - [[-1, 6], 1, Concat, [1]] # cat backbone P4 32 | - [-1, 3, C2f, [512]] # 12 33 | 34 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 35 | - [[-1, 4], 1, Concat, [1]] # cat backbone P3 36 | - [-1, 3, C2f, [256]] # 15 (P3/8-small) 37 | 38 | - [-1, 1, Conv, [256, 3, 2]] 39 | - [[-1, 12], 1, Concat, [1]] # cat head P4 40 | - [-1, 3, C2f, [512]] # 18 (P4/16-medium) 41 | 42 | - [-1, 1, Conv, [512, 3, 2]] 43 | - [[-1, 9], 1, Concat, [1]] # cat head P5 44 | - [-1, 3, C2f, [1024]] # 21 (P5/32-large) 45 | 46 | - [[15, 18, 21], 1, DetectDeepDBB, [nc]] # Detect(P3, P4, P5) 47 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-Detect_AFPN.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOPs 13 | 14 | # YOLOv8.0n backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 18 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 19 | - [-1, 3, C2f, [128, True]] 20 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 21 | - [-1, 6, C2f, [256, True]] 22 | - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16 23 | - [-1, 6, C2f, [512, True]] 24 | - [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32 25 | - [-1, 3, C2f, [1024, True]] 26 | - [-1, 1, SPPF, [1024, 5]] # 9 27 | 28 | # YOLOv8.0n head 29 | head: 30 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 31 | - [[-1, 6], 1, Concat, [1]] # cat backbone P4 32 | - [-1, 3, C2f, [512]] # 12 33 | 34 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 35 | - [[-1, 4], 1, Concat, [1]] # cat backbone P3 36 | - [-1, 3, C2f, [256]] # 15 (P3/8-small) 37 | 38 | - [-1, 1, Conv, [256, 3, 2]] 39 | - [[-1, 12], 1, Concat, [1]] # cat head P4 40 | - [-1, 3, C2f, [512]] # 18 (P4/16-medium) 41 | 42 | - [-1, 1, Conv, [512, 3, 2]] 43 | - [[-1, 9], 1, Concat, [1]] # cat head P5 44 | - [-1, 3, C2f, [1024]] # 21 (P5/32-large) 45 | 46 | - [[15, 18, 21], 1, Detect_AFPN, [nc, 256]] # Detect(P3, P4, P5) 47 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-Detect_AFPN3.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOPs 13 | 14 | # YOLOv8.0n backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 18 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 19 | - [-1, 3, C2f, [128, True]] 20 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 21 | - [-1, 6, C2f, [256, True]] 22 | - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16 23 | - [-1, 6, C2f, [512, True]] 24 | - [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32 25 | - [-1, 3, C2f, [1024, True]] 26 | - [-1, 1, SPPF, [1024, 5]] # 9 27 | 28 | # YOLOv8.0n head 29 | head: 30 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 31 | - [[-1, 6], 1, Concat, [1]] # cat backbone P4 32 | - [-1, 3, C2f, [512]] # 12 33 | 34 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 35 | - [[-1, 4], 1, Concat, [1]] # cat backbone P3 36 | - [-1, 3, C2f, [256]] # 15 (P3/8-small) 37 | 38 | - [-1, 1, Conv, [256, 3, 2]] 39 | - [[-1, 12], 1, Concat, [1]] # cat head P4 40 | - [-1, 3, C2f, [512]] # 18 (P4/16-medium) 41 | 42 | - [-1, 1, Conv, [512, 3, 2]] 43 | - [[-1, 9], 1, Concat, [1]] # cat head P5 44 | - [-1, 3, C2f, [1024]] # 21 (P5/32-large) 45 | 46 | - [[15, 18, 21], 1, Detect_AFPN3, [nc, 256]] # Detect(P3, P4, P5) 47 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-Detect_ASFF.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOPs 13 | 14 | # YOLOv8.0n backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 18 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 19 | - [-1, 3, C2f, [128, True]] 20 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 21 | - [-1, 6, C2f, [256, True]] 22 | - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16 23 | - [-1, 6, C2f, [512, True]] 24 | - [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32 25 | - [-1, 3, C2f, [1024, True]] 26 | - [-1, 1, SPPF, [1024, 5]] # 9 27 | 28 | # YOLOv8.0n head 29 | head: 30 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 31 | - [[-1, 6], 1, Concat, [1]] # cat backbone P4 32 | - [-1, 3, C2f, [512]] # 12 33 | 34 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 35 | - [[-1, 4], 1, Concat, [1]] # cat backbone P3 36 | - [-1, 3, C2f, [256]] # 15 (P3/8-small) 37 | 38 | - [-1, 1, Conv, [256, 3, 2]] 39 | - [[-1, 12], 1, Concat, [1]] # cat head P4 40 | - [-1, 3, C2f, [512]] # 18 (P4/16-medium) 41 | 42 | - [-1, 1, Conv, [512, 3, 2]] 43 | - [[-1, 9], 1, Concat, [1]] # cat head P5 44 | - [-1, 3, C2f, [1024]] # 21 (P5/32-large) 45 | 46 | - [[15, 18, 21], 1, Detect_ASFF, [nc]] # Detect(P3, P4, P5) 47 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-Detect_DBB.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOPs 13 | 14 | # YOLOv8.0n backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 18 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 19 | - [-1, 3, C2f, [128, True]] 20 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 21 | - [-1, 6, C2f, [256, True]] 22 | - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16 23 | - [-1, 6, C2f, [512, True]] 24 | - [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32 25 | - [-1, 3, C2f, [1024, True]] 26 | - [-1, 1, SPPF, [1024, 5]] # 9 27 | 28 | # YOLOv8.0n head 29 | head: 30 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 31 | - [[-1, 6], 1, Concat, [1]] # cat backbone P4 32 | - [-1, 3, C2f, [512]] # 12 33 | 34 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 35 | - [[-1, 4], 1, Concat, [1]] # cat backbone P3 36 | - [-1, 3, C2f, [256]] # 15 (P3/8-small) 37 | 38 | - [-1, 1, Conv, [256, 3, 2]] 39 | - [[-1, 12], 1, Concat, [1]] # cat head P4 40 | - [-1, 3, C2f, [512]] # 18 (P4/16-medium) 41 | 42 | - [-1, 1, Conv, [512, 3, 2]] 43 | - [[-1, 9], 1, Concat, [1]] # cat head P5 44 | - [-1, 3, C2f, [1024]] # 21 (P5/32-large) 45 | 46 | - [[15, 18, 21], 1, Detect_DBB, [nc]] # Detect(P3, P4, P5) 47 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-Detect_FASFF.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOPs 13 | 14 | # YOLOv8.0n backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 18 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 19 | - [-1, 3, C2f, [128, True]] 20 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 21 | - [-1, 6, C2f, [256, True]] 22 | - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16 23 | - [-1, 6, C2f, [512, True]] 24 | - [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32 25 | - [-1, 3, C2f, [1024, True]] 26 | - [-1, 1, SPPF, [1024, 5]] # 9 27 | 28 | # YOLOv8.0n head 29 | head: 30 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 31 | - [[-1, 6], 1, Concat, [1]] # cat backbone P4 32 | - [-1, 3, C2f, [512]] # 12 33 | 34 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 35 | - [[-1, 4], 1, Concat, [1]] # cat backbone P3 36 | - [-1, 3, C2f, [256]] # 15 (P3/8-small) 37 | 38 | - [-1, 1, Conv, [256, 3, 2]] 39 | - [[-1, 12], 1, Concat, [1]] # cat head P4 40 | - [-1, 3, C2f, [512]] # 18 (P4/16-medium) 41 | 42 | - [-1, 1, Conv, [512, 3, 2]] 43 | - [[-1, 9], 1, Concat, [1]] # cat head P5 44 | - [-1, 3, C2f, [1024]] # 21 (P5/32-large) 45 | 46 | - [[15, 18, 21], 1, Detect_FASFF, [nc, 256]] # Detect(P3, P4, P5) 47 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-Detect_Partial_C.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOPs 13 | 14 | # YOLOv8.0n backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 18 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 19 | - [-1, 3, C2f, [128, True]] 20 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 21 | - [-1, 6, C2f, [256, True]] 22 | - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16 23 | - [-1, 6, C2f, [512, True]] 24 | - [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32 25 | - [-1, 3, C2f, [1024, True]] 26 | - [-1, 1, SPPF, [1024, 5]] # 9 27 | 28 | # YOLOv8.0n head 29 | head: 30 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 31 | - [[-1, 6], 1, Concat, [1]] # cat backbone P4 32 | - [-1, 3, C2f, [512]] # 12 33 | 34 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 35 | - [[-1, 4], 1, Concat, [1]] # cat backbone P3 36 | - [-1, 3, C2f, [256]] # 15 (P3/8-small) 37 | 38 | - [-1, 1, Conv, [256, 3, 2]] 39 | - [[-1, 12], 1, Concat, [1]] # cat head P4 40 | - [-1, 3, C2f, [512]] # 18 (P4/16-medium) 41 | 42 | - [-1, 1, Conv, [512, 3, 2]] 43 | - [[-1, 9], 1, Concat, [1]] # cat head P5 44 | - [-1, 3, C2f, [1024]] # 21 (P5/32-large) 45 | 46 | - [[15, 18, 21], 1, Detect_Partial_C, [nc]] # Detect(P3, P4, P5) 47 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-Detect_SC_C.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOPs 13 | 14 | # YOLOv8.0n backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 18 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 19 | - [-1, 3, C2f, [128, True]] 20 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 21 | - [-1, 6, C2f, [256, True]] 22 | - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16 23 | - [-1, 6, C2f, [512, True]] 24 | - [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32 25 | - [-1, 3, C2f, [1024, True]] 26 | - [-1, 1, SPPF, [1024, 5]] # 9 27 | 28 | # YOLOv8.0n head 29 | head: 30 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 31 | - [[-1, 6], 1, Concat, [1]] # cat backbone P4 32 | - [-1, 3, C2f, [512]] # 12 33 | 34 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 35 | - [[-1, 4], 1, Concat, [1]] # cat backbone P3 36 | - [-1, 3, C2f, [256]] # 15 (P3/8-small) 37 | 38 | - [-1, 1, Conv, [256, 3, 2]] 39 | - [[-1, 12], 1, Concat, [1]] # cat head P4 40 | - [-1, 3, C2f, [512]] # 18 (P4/16-medium) 41 | 42 | - [-1, 1, Conv, [512, 3, 2]] 43 | - [[-1, 9], 1, Concat, [1]] # cat head P5 44 | - [-1, 3, C2f, [1024]] # 21 (P5/32-large) 45 | 46 | - [[15, 18, 21], 1, Detect_SC_C, [nc]] # Detect(P3, P4, P5) 47 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-Detect_dyhead3.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOPs 13 | 14 | # YOLOv8.0n backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 18 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 19 | - [-1, 3, C2f, [128, True]] 20 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 21 | - [-1, 6, C2f, [256, True]] 22 | - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16 23 | - [-1, 6, C2f, [512, True]] 24 | - [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32 25 | - [-1, 3, C2f, [1024, True]] 26 | - [-1, 1, SPPF, [1024, 5]] # 9 27 | 28 | # YOLOv8.0n head 29 | head: 30 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 31 | - [[-1, 6], 1, Concat, [1]] # cat backbone P4 32 | - [-1, 3, C2f, [512]] # 12 33 | 34 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 35 | - [[-1, 4], 1, Concat, [1]] # cat backbone P3 36 | - [-1, 3, C2f, [512]] # 15 (P3/8-small) 37 | 38 | - [-1, 1, Conv, [256, 3, 2]] 39 | - [[-1, 12], 1, Concat, [1]] # cat head P4 40 | - [-1, 3, C2f, [512]] # 18 (P4/16-medium) 41 | 42 | - [-1, 1, Conv, [512, 3, 2]] 43 | - [[-1, 9], 1, Concat, [1]] # cat head P5 44 | - [-1, 3, C2f, [512]] # 21 (P5/32-large) 45 | 46 | - [[15, 18, 21], 1, Detect_dyhead3, [nc]] # Detect(P3, P4, P5) 47 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-Detect_improve_s.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOPs 13 | 14 | # YOLOv8.0n backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 18 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 19 | - [-1, 3, C2f, [128, True]] 20 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 21 | - [-1, 6, C2f, [256, True]] 22 | - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16 23 | - [-1, 6, C2f, [512, True]] 24 | - [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32 25 | - [-1, 3, C2f, [1024, True]] 26 | - [-1, 1, SPPF, [1024, 5]] # 9 27 | 28 | # YOLOv8.0n head 29 | head: 30 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 31 | - [[-1, 6], 1, Concat, [1]] # cat backbone P4 32 | - [-1, 3, C2f, [512]] # 12 33 | 34 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 35 | - [[-1, 4], 1, Concat, [1]] # cat backbone P3 36 | - [-1, 3, C2f, [256]] # 15 (P3/8-small) 37 | 38 | - [-1, 1, Conv, [256, 3, 2]] 39 | - [[-1, 12], 1, Concat, [1]] # cat head P4 40 | - [-1, 3, C2f, [512]] # 18 (P4/16-medium) 41 | 42 | - [-1, 1, Conv, [512, 3, 2]] 43 | - [[-1, 9], 1, Concat, [1]] # cat head P5 44 | - [-1, 3, C2f, [1024]] # 21 (P5/32-large) 45 | 46 | - [[15, 18, 21], 1, Detect_improve_s, [nc]] # Detect(P3, P4, P5) 47 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-FFA_Net.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOPs 13 | 14 | # YOLOv8.0n backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, FFA, []] # 0-P1/2 18 | - [-1, 1, Conv, [64, 3, 2]] # 1-P1/2 19 | - [-1, 1, Conv, [128, 3, 2]] # 2-P2/4 20 | - [-1, 3, C2f, [128, True]] 21 | - [-1, 1, Conv, [256, 3, 2]] # 4-P3/8 22 | - [-1, 6, C2f, [256, True]] 23 | - [-1, 1, Conv, [512, 3, 2]] # 6-P4/16 24 | - [-1, 6, C2f, [512, True]] 25 | - [-1, 1, Conv, [1024, 3, 2]] # 8-P5/32 26 | - [-1, 3, C2f, [1024, True]] 27 | - [-1, 1, SPPF, [1024, 5]] # 10 28 | 29 | # YOLOv8.0n head 30 | head: 31 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 32 | - [[-1, 7], 1, Concat, [1]] # cat backbone P4 33 | - [-1, 3, C2f, [512]] # 13 34 | 35 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 36 | - [[-1, 5], 1, Concat, [1]] # cat backbone P3 37 | - [-1, 3, C2f, [256]] # 16 (P3/8-small) 38 | 39 | - [-1, 1, Conv, [256, 3, 2]] 40 | - [[-1, 13], 1, Concat, [1]] # cat head P4 41 | - [-1, 3, C2f, [512]] # 19 (P4/16-medium) 42 | 43 | - [-1, 1, Conv, [512, 3, 2]] 44 | - [[-1, 10], 1, Concat, [1]] # cat head P5 45 | - [-1, 3, C2f, [1024]] # 22 (P5/32-large) 46 | 47 | - [[16, 19, 22], 1, Detect, [nc]] # Detect(P3, P4, P5) 48 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-FRMHead.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOP 13 | 14 | # YOLOv8.0 backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 18 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 19 | - [-1, 3, C2f, [128, True]] 20 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 21 | - [-1, 6, C2f, [256, True]] 22 | - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16 23 | - [-1, 6, C2f, [512, True]] 24 | - [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32 25 | - [-1, 3, C2f, [1024, True]] 26 | - [-1, 1, SPPF, [1024, 5]] # 9 27 | 28 | # YOLOv8.0n head 29 | head: 30 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 31 | - [[-1, 6], 1, Concat, [1]] # cat backbone P4 32 | - [-1, 3, C2f, [512]] # 12 33 | 34 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 35 | - [[-1, 4], 1, Concat, [1]] # cat backbone P3 36 | - [-1, 3, C2f, [256]] # 15 (P3/8-small) 37 | 38 | - [-1, 1, Conv, [256, 3, 2]] 39 | - [[-1, 12], 1, Concat, [1]] # cat head P4 40 | - [-1, 3, C2f, [512]] # 18 (P4/16-medium) 41 | 42 | - [-1, 1, Conv, [512, 3, 2]] 43 | - [[-1, 9], 1, Concat, [1]] # cat head P5 44 | - [-1, 3, C2f, [1024]] # 21 (P5/32-large) 45 | 46 | - [[15, 18, 21], 1, Detect_FRM, [nc]] # Detect(P3, P4, P5) 47 | 48 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-FasterNet.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOP 13 | 14 | # YOLOv8.0n backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, FasterNet, []] # 4 18 | - [-1, 1, SPPF, [1024, 5]] # 5 19 | 20 | # YOLOv8.0n head 21 | head: 22 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] # 6 23 | - [[-1, 3], 1, Concat, [1]] # 7 cat backbone P4 24 | - [-1, 3, C2f, [512]] # 8 25 | 26 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] # 9 27 | - [[-1, 2], 1, Concat, [1]] # 10 cat backbone P3 28 | - [-1, 3, C2f, [256]] # 11 (P3/8-small) 29 | 30 | - [-1, 1, Conv, [256, 3, 2]] # 12 31 | - [[-1, 8], 1, Concat, [1]] # 13 cat head P4 32 | - [-1, 3, C2f, [512]] # 14 (P4/16-medium) 33 | 34 | - [-1, 1, Conv, [512, 3, 2]] # 15 35 | - [[-1, 5], 1, Concat, [1]] # 16 cat head P5 36 | - [-1, 3, C2f, [1024]] # 17 (P5/32-large) 37 | 38 | - [[11, 14, 17], 1, Detect, [nc]] # Detect(P3, P4, P5) 39 | 40 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-FeaturePyramidSharedConv.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOPs 13 | 14 | # YOLOv8.0n backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 18 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 19 | - [-1, 3, C2f, [128, True]] 20 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 21 | - [-1, 6, C2f, [256, True]] 22 | - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16 23 | - [-1, 6, C2f, [512, True]] 24 | - [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32 25 | - [-1, 3, C2f, [1024, True]] 26 | - [-1, 1, FeaturePyramidSharedConv, [1024]] # 9 27 | 28 | # YOLOv8.0n head 29 | head: 30 | - [-1, 1, nn.Upsample, [None, 2, "nearest"]] 31 | - [[-1, 6], 1, Concat, [1]] # cat backbone P4 32 | - [-1, 3, C2f, [512]] # 12 33 | 34 | - [-1, 1, nn.Upsample, [None, 2, "nearest"]] 35 | - [[-1, 4], 1, Concat, [1]] # cat backbone P3 36 | - [-1, 3, C2f, [256]] # 15 (P3/8-small) 37 | 38 | - [-1, 1, Conv, [256, 3, 2]] 39 | - [[-1, 12], 1, Concat, [1]] # cat head P4 40 | - [-1, 3, C2f, [512]] # 18 (P4/16-medium) 41 | 42 | - [-1, 1, Conv, [512, 3, 2]] 43 | - [[-1, 9], 1, Concat, [1]] # cat head P5 44 | - [-1, 3, C2f, [1024]] # 21 (P5/32-large) 45 | 46 | - [[15, 18, 21], 1, Detect, [nc]] # Detect(P3, P4, P5) 47 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-FreqFFPN.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOPs 13 | 14 | # YOLOv8.0n backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 18 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 19 | - [-1, 3, C2f, [128, True]] 20 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 21 | - [-1, 6, C2f, [256, True]] 22 | - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16 23 | - [-1, 6, C2f, [512, True]] 24 | - [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32 25 | - [-1, 3, C2f, [1024, True]] 26 | - [-1, 1, SPPF, [1024, 5]] # 9 27 | 28 | # YOLOv8.0n head 29 | head: 30 | - [4, 1, Conv, [256]] # 10-P3/8 31 | - [6, 1, Conv, [256]] # 11-P4/16 32 | - [9, 1, Conv, [256]] # 12-P5/32 33 | 34 | - [[11, -1], 1, FreqFusion, []] # cat backbone P4 35 | - [-1, 3, C2f, [256]] # 14 36 | 37 | - [[10, -1], 1, FreqFusion, []] # cat backbone P3 38 | - [-1, 3, C2f, [256]] # 16 (P3/8-small) 39 | 40 | - [-1, 1, Conv, [256, 3, 2]] 41 | - [[-1, 14], 1, Concat, [1]] # cat head P4 42 | - [-1, 3, C2f, [512]] # 19 (P4/16-medium) 43 | 44 | - [-1, 1, Conv, [512, 3, 2]] 45 | - [[-1, 12], 1, Concat, [1]] # cat head P5 46 | - [-1, 3, C2f, [1024]] # 22 (P5/32-large) 47 | 48 | - [[16, 19, 22], 1, Detect, [nc]] # Detect(P3, P4, P5) 49 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-HATHead.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOP 13 | 14 | # YOLOv8.0n backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 18 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 19 | - [-1, 3, C2f, [128, True]] 20 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 21 | - [-1, 6, C2f, [256, True]] 22 | - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16 23 | - [-1, 6, C2f, [512, True]] 24 | - [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32 25 | - [-1, 3, C2f, [1024, True]] 26 | - [-1, 1, SPPF, [1024, 5]] # 9 27 | 28 | # YOLOv8.0n head 29 | head: 30 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 31 | - [[-1, 6], 1, Concat, [1]] # cat backbone P4 32 | - [-1, 3, C2f, [512]] # 12 33 | 34 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 35 | - [[-1, 4], 1, Concat, [1]] # cat backbone P3 36 | - [-1, 3, C2f, [256]] # 15 (P3/8-small) 37 | 38 | - [-1, 1, Conv, [256, 3, 2]] 39 | - [[-1, 12], 1, Concat, [1]] # cat head P4 40 | - [-1, 3, C2f, [512]] # 18 (P4/16-medium) 41 | 42 | - [-1, 1, Conv, [512, 3, 2]] 43 | - [[-1, 9], 1, Concat, [1]] # cat head P5 44 | - [-1, 3, C2f, [1024]] # 21 (P5/32-large) 45 | 46 | - [[15, 18, 21], 1, HATHead, [nc]] # Detect(P3, P4, P5) 47 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-JD.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOPs 13 | 14 | # YOLOv8.0n backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 18 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 19 | - [-1, 3, DSCFEM, [128]] 20 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 21 | - [-1, 6, DSCFEM, [256]] 22 | - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16 23 | - [-1, 6, DSCFEM, [512]] 24 | - [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32 25 | - [-1, 3, DSCFEM, [1024]] 26 | - [-1, 1, SPPM, [1024, 5]] # 9 27 | 28 | # YOLOv8.0n head 29 | head: 30 | - [-1, 1, nn.Upsample, [None, 2, "nearest"]] 31 | - [[-1, 6], 1, Concat, [1]] # cat backbone P4 32 | - [-1, 3, DSCFEM, [512]] # 12 33 | 34 | - [-1, 1, nn.Upsample, [None, 2, "nearest"]] 35 | - [[-1, 4], 1, Concat, [1]] # cat backbone P3 36 | - [-1, 3, DSCFEM, [256]] # 15 (P3/8-small) 37 | 38 | - [-1, 1, Conv, [256, 3, 2]] 39 | - [[-1, 12], 1, Concat, [1]] # cat head P4 40 | - [-1, 3, DSCFEM, [512]] # 18 (P4/16-medium) 41 | 42 | - [-1, 1, Conv, [512, 3, 2]] 43 | - [[-1, 9], 1, Concat, [1]] # cat head P5 44 | - [-1, 3, DSCFEM, [1024]] # 21 (P5/32-large) 45 | 46 | - [[15, 18, 21], 1, Detect, [nc]] # Detect(P3, P4, P5) 47 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-LADH.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOPs 13 | 14 | # YOLOv8.0n backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 18 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 19 | - [-1, 3, C2f, [128, True]] 20 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 21 | - [-1, 6, C2f, [256, True]] 22 | - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16 23 | - [-1, 6, C2f, [512, True]] 24 | - [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32 25 | - [-1, 3, C2f, [1024, True]] 26 | - [-1, 1, SPPF, [1024, 5]] # 9 27 | 28 | # YOLOv8.0n head 29 | head: 30 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 31 | - [[-1, 6], 1, Concat, [1]] # cat backbone P4 32 | - [-1, 3, C2f, [512]] # 12 33 | 34 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 35 | - [[-1, 4], 1, Concat, [1]] # cat backbone P3 36 | - [-1, 3, C2f, [256]] # 15 (P3/8-small) 37 | 38 | - [-1, 1, Conv, [256, 3, 2]] 39 | - [[-1, 12], 1, Concat, [1]] # cat head P4 40 | - [-1, 3, C2f, [512]] # 18 (P4/16-medium) 41 | 42 | - [-1, 1, Conv, [512, 3, 2]] 43 | - [[-1, 9], 1, Concat, [1]] # cat head P5 44 | - [-1, 3, C2f, [1024]] # 21 (P5/32-large) 45 | 46 | - [[15, 18, 21], 1, Detect_LADH, [nc]] # Detect(P3, P4, P5) 47 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-LAE.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOP 13 | 14 | # YOLOv8.0n backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 64 x 0.25 18 | - [-1, 1, LAE, []] # 1-P2/4 19 | - [-1, 3, C2f, [128, True]] 20 | - [-1, 1, LAE, []] # 3-P3/8 21 | - [-1, 6, C2f, [256, True]] 22 | - [-1, 1, LAE, []] # 5-P4/16 23 | - [-1, 6, C2f, [512, True]] 24 | - [-1, 1, LAE, []] # 7-P5/32 25 | - [-1, 3, C2f, [1024, True]] 26 | - [-1, 1, SPPF, [1024, 5]] # 9 27 | 28 | # YOLOv8.0n head 29 | head: 30 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 31 | - [[-1, 6], 1, Concat, [1]] # cat backbone P4 32 | - [-1, 3, C2f, [512]] # 12 33 | 34 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 35 | - [[-1, 4], 1, Concat, [1]] # cat backbone P3 36 | - [-1, 3, C2f, [256]] # 15 (P3/8-small) 37 | 38 | - [-1, 1, LAE, []] 39 | - [[-1, 12], 1, Concat, [1]] # cat head P4 40 | - [-1, 3, C2f, [512]] # 18 (P4/16-medium) 41 | 42 | - [-1, 1, LAE, []] 43 | - [[-1, 9], 1, Concat, [1]] # cat head P5 44 | - [-1, 3, C2f, [1024]] # 21 (P5/32-large) 45 | 46 | - [[15, 18, 21], 1, Detect, [nc]] # Detect(P3, P4, P5) 47 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-LSDECD.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOPs 13 | 14 | # YOLOv8.0n backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 18 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 19 | - [-1, 3, C2f, [128, True]] 20 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 21 | - [-1, 6, C2f, [256, True]] 22 | - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16 23 | - [-1, 6, C2f, [512, True]] 24 | - [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32 25 | - [-1, 3, C2f, [1024, True]] 26 | - [-1, 1, SPPF, [1024, 5]] # 9 27 | 28 | # YOLOv8.0n head 29 | head: 30 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 31 | - [[-1, 6], 1, Concat, [1]] # cat backbone P4 32 | - [-1, 3, C2f, [512]] # 12 33 | 34 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 35 | - [[-1, 4], 1, Concat, [1]] # cat backbone P3 36 | - [-1, 3, C2f, [256]] # 15 (P3/8-small) 37 | 38 | - [-1, 1, Conv, [256, 3, 2]] 39 | - [[-1, 12], 1, Concat, [1]] # cat head P4 40 | - [-1, 3, C2f, [512]] # 18 (P4/16-medium) 41 | 42 | - [-1, 1, Conv, [512, 3, 2]] 43 | - [[-1, 9], 1, Concat, [1]] # cat head P5 44 | - [-1, 3, C2f, [1024]] # 21 (P5/32-large) 45 | 46 | - [[15, 18, 21], 1, Detect_LSDECD, [nc, 256]] # Detect(P3, P4, P5) 47 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-MLCA3.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOP 13 | 14 | # YOLOv8.0n backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 18 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 19 | - [-1, 3, C2f, [128, True]] 20 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 21 | - [-1, 6, C2f, [256, True]] 22 | - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16 23 | - [-1, 6, C2f, [512, True]] 24 | - [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32 25 | - [-1, 3, C2f, [1024, True]] 26 | - [-1, 1, SPPF, [1024, 5]] # 9 27 | 28 | # YOLOv8.0n head 29 | head: 30 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 31 | - [[-1, 6], 1, Concat, [1]] # cat backbone P4 32 | - [-1, 3, C2f_MLCA, [512]] # 12 33 | 34 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 35 | - [[-1, 4], 1, Concat, [1]] # cat backbone P3 36 | - [-1, 3, C2f_MLCA, [256]] # 15 (P3/8-small) 37 | 38 | - [-1, 1, Conv, [256, 3, 2]] 39 | - [[-1, 12], 1, Concat, [1]] # cat head P4 40 | - [-1, 3, C2f_MLCA, [512]] # 18 (P4/16-medium) 41 | 42 | - [-1, 1, Conv, [512, 3, 2]] 43 | - [[-1, 9], 1, Concat, [1]] # cat head P5 44 | - [-1, 3, C2f_MLCA, [1024]] # 21 (P5/32-large) 45 | 46 | - [[15, 18, 21], 1, Detect, [nc]] # Detect(P3, P4, P5) 47 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-MSFN.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOPs 13 | 14 | # YOLOv8.0n backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 18 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 19 | - [-1, 3, C2f, [128, True]] 20 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 21 | - [-1, 6, C2f, [256, True]] 22 | - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16 23 | - [-1, 6, C2f, [512, True]] 24 | - [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32 25 | - [-1, 3, C2f, [1024, True]] 26 | - [-1, 1, SPPF, [1024, 5]] # 9 27 | - [-1, 1, MSFN, [1024]] # 10 28 | 29 | # YOLOv8.0n head 30 | head: 31 | - [-1, 1, nn.Upsample, [None, 2, "nearest"]] 32 | - [[-1, 6], 1, Concat, [1]] # cat backbone P4 33 | - [-1, 3, C2f, [512]] # 13 34 | 35 | - [-1, 1, nn.Upsample, [None, 2, "nearest"]] 36 | - [[-1, 4], 1, Concat, [1]] # cat backbone P3 37 | - [-1, 3, C2f, [256]] # 16 (P3/8-small) 38 | 39 | - [-1, 1, Conv, [256, 3, 2]] 40 | - [[-1, 13], 1, Concat, [1]] # cat head P4 41 | - [-1, 3, C2f, [512]] # 19 (P4/16-medium) 42 | 43 | - [-1, 1, Conv, [512, 3, 2]] 44 | - [[-1, 10], 1, Concat, [1]] # cat head P5 45 | - [-1, 3, C2f, [1024]] # 22 (P5/32-large) 46 | 47 | - [[16, 19, 22], 1, Detect, [nc]] # Detect(P3, P4, P5) 48 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-MoCAttention.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOPs 13 | 14 | # YOLOv8.0n backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 18 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 19 | - [-1, 3, C2f, [128, True]] 20 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 21 | - [-1, 6, C2f, [256, True]] 22 | - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16 23 | - [-1, 6, C2f, [512, True]] 24 | - [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32 25 | - [-1, 3, C2f, [1024, True]] 26 | - [-1, 1, SPPF, [1024, 5]] # 9 27 | - [-1, 1, MoCAttention, [1024]] # 10 28 | 29 | # YOLOv8.0n head 30 | head: 31 | - [-1, 1, nn.Upsample, [None, 2, "nearest"]] 32 | - [[-1, 6], 1, Concat, [1]] # cat backbone P4 33 | - [-1, 3, C2f, [512]] # 13 34 | 35 | - [-1, 1, nn.Upsample, [None, 2, "nearest"]] 36 | - [[-1, 4], 1, Concat, [1]] # cat backbone P3 37 | - [-1, 3, C2f, [256]] # 16 (P3/8-small) 38 | 39 | - [-1, 1, Conv, [256, 3, 2]] 40 | - [[-1, 13], 1, Concat, [1]] # cat head P4 41 | - [-1, 3, C2f, [512]] # 19 (P4/16-medium) 42 | 43 | - [-1, 1, Conv, [512, 3, 2]] 44 | - [[-1, 10], 1, Concat, [1]] # cat head P5 45 | - [-1, 3, C2f, [1024]] # 22 (P5/32-large) 46 | 47 | - [[16, 19, 22], 1, Detect, [nc]] # Detect(P3, P4, P5) 48 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-MobileNetV3.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOP 13 | 14 | # YOLOv8.0n backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, MobileNetV3, []] # 4 18 | - [-1, 1, SPPF, [1024, 5]] # 5 19 | 20 | # YOLOv8.0n head 21 | head: 22 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] # 6 23 | - [[-1, 3], 1, Concat, [1]] # 7 cat backbone P4 24 | - [-1, 3, C2f, [512]] # 8 25 | 26 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] # 9 27 | - [[-1, 2], 1, Concat, [1]] # 10 cat backbone P3 28 | - [-1, 3, C2f, [256]] # 11 (P3/8-small) 29 | 30 | - [-1, 1, Conv, [256, 3, 2]] # 12 31 | - [[-1, 8], 1, Concat, [1]] # 13 cat head P4 32 | - [-1, 3, C2f, [512]] # 14 (P4/16-medium) 33 | 34 | - [-1, 1, Conv, [512, 3, 2]] # 15 35 | - [[-1, 5], 1, Concat, [1]] # 16 cat head P5 36 | - [-1, 3, C2f, [1024]] # 17 (P5/32-large) 37 | 38 | - [[11, 14, 17], 1, Detect, [nc]] # Detect(P3, P4, P5) 39 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-PSA.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOPs 13 | 14 | # YOLOv8.0n backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 18 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 19 | - [-1, 3, C2f, [128, True]] 20 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 21 | - [-1, 6, C2f, [256, True]] 22 | - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16 23 | - [-1, 6, C2f, [512, True]] 24 | - [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32 25 | - [-1, 3, C2f, [1024, True]] 26 | - [-1, 1, SPPF, [1024, 5]] # 9 27 | - [-1, 1, PSA, [1024]] # 10 28 | 29 | # YOLOv8.0n head 30 | head: 31 | - [-1, 1, nn.Upsample, [None, 2, "nearest"]] 32 | - [[-1, 6], 1, Concat, [1]] # cat backbone P4 33 | - [-1, 3, C2f, [512]] # 13 34 | 35 | - [-1, 1, nn.Upsample, [None, 2, "nearest"]] 36 | - [[-1, 4], 1, Concat, [1]] # cat backbone P3 37 | - [-1, 3, C2f, [256]] # 16 (P3/8-small) 38 | 39 | - [-1, 1, Conv, [256, 3, 2]] 40 | - [[-1, 13], 1, Concat, [1]] # cat head P4 41 | - [-1, 3, C2f, [512]] # 19(P4/16-medium) 42 | 43 | - [-1, 1, Conv, [512, 3, 2]] 44 | - [[-1, 10], 1, Concat, [1]] # cat head P5 45 | - [-1, 3, C2f, [1024]] # 22 (P5/32-large) 46 | 47 | - [[16, 19, 22], 1, Detect, [nc]] # Detect(P3, P4, P5) 48 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-PSA3.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOPs 13 | 14 | # YOLOv8.0n backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 18 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 19 | - [-1, 3, C2f, [128, True]] 20 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 21 | - [-1, 6, C2f, [256, True]] 22 | - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16 23 | - [-1, 6, C2f, [512, True]] 24 | - [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32 25 | - [-1, 3, C2f, [1024, True]] 26 | - [-1, 1, SPPF, [1024, 5]] # 9 27 | 28 | # YOLOv8.0n head 29 | head: 30 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 31 | - [[-1, 6], 1, Concat, [1]] # cat backbone P4 32 | - [-1, 3, C2f, [512]] # 12 33 | 34 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 35 | - [[-1, 4], 1, Concat, [1]] # cat backbone P3 36 | - [-1, 3, PSA, [256]] # 15 (P3/8-small) 37 | 38 | - [-1, 1, Conv, [256, 3, 2]] 39 | - [[-1, 12], 1, Concat, [1]] # cat head P4 40 | - [-1, 3, PSA, [512]] # 18 (P4/16-medium) 41 | 42 | - [-1, 1, Conv, [512, 3, 2]] 43 | - [[-1, 9], 1, Concat, [1]] # cat head P5 44 | - [-1, 3, PAS, [1024]] # 21 (P5/32-large) 45 | 46 | - [[15, 18, 21], 1, Detect, [nc]] # Detect(P3, P4, P5) 47 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-RCS-OSA1.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOPs 13 | 14 | # YOLOv8.0n backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 18 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 19 | - [-1, 3, RCSOSA, [128, True]] 20 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 21 | - [-1, 6, RCSOSA, [256, True]] 22 | - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16 23 | - [-1, 6, RCSOSA, [512, True]] 24 | - [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32 25 | - [-1, 3, RCSOSA, [1024, True]] 26 | - [-1, 1, SPPF, [1024, 5]] # 9 27 | 28 | # YOLOv8.0n head 29 | head: 30 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 31 | - [[-1, 6], 1, Concat, [1]] # cat backbone P4 32 | - [-1, 3, RCSOSA, [512]] # 12 33 | 34 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 35 | - [[-1, 4], 1, Concat, [1]] # cat backbone P3 36 | - [-1, 3, RCSOSA, [256]] # 15 (P3/8-small) 37 | 38 | - [-1, 1, Conv, [256, 3, 2]] 39 | - [[-1, 12], 1, Concat, [1]] # cat head P4 40 | - [-1, 3, RCSOSA, [512]] # 18 (P4/16-medium) 41 | 42 | - [-1, 1, Conv, [512, 3, 2]] 43 | - [[-1, 9], 1, Concat, [1]] # cat head P5 44 | - [-1, 3, RCSOSA, [1024]] # 21 (P5/32-large) 45 | 46 | - [[15, 18, 21], 1, Detect, [nc]] # Detect(P3, P4, P5) 47 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-RCS-OSA2.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOPs 13 | 14 | # YOLOv8.0n backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 18 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 19 | - [-1, 3, RCSOSA, [128, True]] 20 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 21 | - [-1, 6, RCSOSA, [256, True]] 22 | - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16 23 | - [-1, 6, RCSOSA, [512, True]] 24 | - [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32 25 | - [-1, 3, RCSOSA, [1024, True]] 26 | - [-1, 1, SPPF, [1024, 5]] # 9 27 | 28 | # YOLOv8.0n head 29 | head: 30 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 31 | - [[-1, 6], 1, Concat, [1]] # cat backbone P4 32 | - [-1, 3, C2f, [512]] # 12 33 | 34 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 35 | - [[-1, 4], 1, Concat, [1]] # cat backbone P3 36 | - [-1, 3, C2f, [256]] # 15 (P3/8-small) 37 | 38 | - [-1, 1, Conv, [256, 3, 2]] 39 | - [[-1, 12], 1, Concat, [1]] # cat head P4 40 | - [-1, 3, C2f, [512]] # 18 (P4/16-medium) 41 | 42 | - [-1, 1, Conv, [512, 3, 2]] 43 | - [[-1, 9], 1, Concat, [1]] # cat head P5 44 | - [-1, 3, C2f, [1024]] # 21 (P5/32-large) 45 | 46 | - [[15, 18, 21], 1, Detect, [nc]] # Detect(P3, P4, P5) 47 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-ReCalibrationFPN-P2345.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P2-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] 9 | s: [0.33, 0.50, 1024] 10 | m: [0.67, 0.75, 768] 11 | l: [1.00, 1.00, 512] 12 | x: [1.00, 1.25, 512] 13 | 14 | # YOLOv8.0 backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 18 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 19 | - [-1, 3, C2f, [128, True]] 20 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 21 | - [-1, 6, C2f, [256, True]] 22 | - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16 23 | - [-1, 6, C2f, [512, True]] 24 | - [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32 25 | - [-1, 3, C2f, [1024, True]] 26 | - [-1, 1, SPPF, [1024, 5]] # 9 27 | 28 | # YOLOv8.0-p2 head 29 | head: 30 | - [[-1, 6], 1, SBA, []] # 10 (P4/16) 31 | - [-1, 3, C2f, [512]] # 11 32 | 33 | - [[-1, 4], 1, SBA, []] # 12 (P3/8) 34 | - [-1, 3, C2f, [256]] # 13 35 | 36 | - [[-1, 2], 1, SBA, []] # 14 (P2/4) 37 | - [-1, 3, C2f, [128]] # 15 38 | 39 | - [[-1, 13], 1, SBA, []] # 16 (P3/8) 40 | - [-1, 3, C2f, [256]] # 17 41 | 42 | - [[-1, 11], 1, SBA, []] # 18 (P4/16) 43 | - [-1, 3, C2f, [512]] # 19 44 | 45 | - [[-1, 9], 1, SBA, []] # 10 (P5/32) 46 | - [-1, 3, C2f, [1024]] # 21 47 | 48 | - [[15, 17, 19, 21], 1, Detect, [nc]] # Detect(P2, P3, P4, P5) -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-ReCalibrationFPN-P345.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOPs 13 | 14 | # YOLOv8.0n backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 18 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 19 | - [-1, 3, C2f, [128, True]] 20 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 21 | - [-1, 6, C2f, [256, True]] 22 | - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16 23 | - [-1, 6, C2f, [512, True]] 24 | - [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32 25 | - [-1, 3, C2f, [1024, True]] 26 | - [-1, 1, SPPF, [1024, 5]] # 9 27 | 28 | # YOLOv8.0n head 29 | head: 30 | - [[-1, 6], 1, SBA, []] # 10 (P4/16) 31 | - [-1, 3, C2f, [512]] # 11 32 | 33 | - [[-1, 4], 1, SBA, []] # 12 (P3/8) 34 | - [-1, 3, C2f, [256]] # 13 35 | 36 | - [[-1, 11], 1, SBA, []] # 14 (P4/16) 37 | - [-1, 3, C2f, [512]] # 15 38 | 39 | - [[-1, 9], 1, SBA, []] # 16 (P5/32) 40 | - [-1, 3, C2f, [1024]] # 17 41 | 42 | - [[13, 15, 17], 1, Detect, [nc]] # Detect(P3, P4, P5) -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-ReCalibrationFPN-P3456.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P6 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n-p6.yaml' will call yolov8-p6.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] 9 | s: [0.33, 0.50, 1024] 10 | m: [0.67, 0.75, 768] 11 | l: [1.00, 1.00, 512] 12 | x: [1.00, 1.25, 512] 13 | 14 | # YOLOv8.0x6 backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 18 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 19 | - [-1, 3, C2f, [128, True]] 20 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 21 | - [-1, 6, C2f, [256, True]] 22 | - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16 23 | - [-1, 6, C2f, [512, True]] 24 | - [-1, 1, Conv, [768, 3, 2]] # 7-P5/32 25 | - [-1, 3, C2f, [768, True]] 26 | - [-1, 1, Conv, [1024, 3, 2]] # 9-P6/64 27 | - [-1, 3, C2f, [1024, True]] 28 | - [-1, 1, SPPF, [1024, 5]] # 11 29 | 30 | # YOLOv8.0x6 head 31 | head: 32 | - [[-1, 8], 1, SBA, []] # 12 (P5/32) 33 | - [-1, 3, C2, [768, False]] # 13 34 | 35 | - [[-1, 6], 1, SBA, []] # 14 (P4/16) 36 | - [-1, 3, C2, [512, False]] # 15 37 | 38 | - [[-1, 4], 1, SBA, []] # 16 (P3/8) 39 | - [-1, 3, C2, [256, False]] # 17 40 | 41 | - [[-1, 15], 1, SBA, []] # 18 (P4/16) 42 | - [-1, 3, C2, [512, False]] # 19 43 | 44 | - [[-1, 13], 1, SBA, []] # 20 (P5/32) 45 | - [-1, 3, C2, [768, False]] # 21 46 | 47 | - [[-1, 11], 1, SBA, []] # 22 (P6/64) 48 | - [-1, 3, C2, [1024, False]] # 23 49 | 50 | - [[17, 19, 21, 23], 1, Detect, [nc]] # Detect(P3, P4, P5, P6) 51 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-RepFPN.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, GPL-3.0 license 2 | 3 | # Parameters 4 | nc: 80 # number of classes 5 | depth_multiple: 0.33 # scales module repeats 6 | width_multiple: 0.50 # scales convolution channels 7 | 8 | # YOLOv8.0s backbone 9 | backbone: 10 | # [from, repeats, module, args] 11 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 12 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 13 | - [-1, 3, C2f, [128, True]] 14 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 15 | - [-1, 6, C2f, [512, True]] 16 | - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16 17 | - [-1, 6, C2f, [512, True]] 18 | - [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32 19 | - [-1, 3, C2f, [512, True]] 20 | - [-1, 1, SPPF, [1024, 5]] # 9 21 | 22 | # YOLOv8.0s head 23 | head: 24 | - [-1, 1, SimConv, [256, 1, 1]] 25 | - [-1, 1, Transpose, [256]] 26 | - [[-1, 6], 1, Concat, [1]] 27 | - [-1, 12, RepBlock, [256]] 28 | 29 | - [-1, 1, SimConv, [128, 1, 1]] 30 | - [-1, 1, Transpose, [128]] 31 | - [[-1, 4], 1, Concat, [1]] 32 | - [-1, 12, RepBlock, [128]] #out 33 | 34 | - [-1, 1, SimConv, [128, 3, 2]] 35 | - [[-1, 14], 1, Concat, [1]] # cat head P4 36 | - [-1, 12, RepBlock, [256]] # 20 (P4/16-medium) 37 | 38 | - [-1, 1, SimConv, [256, 3, 2]] 39 | - [[-1, 10], 1, Concat, [1]] # cat head P5 40 | - [-1, 12, RepBlock, [512]] 41 | 42 | - [[17, 20, 23], 1, Detect, [nc]] # Detect(P3, P4, P5) 43 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-RepHead.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOPs 13 | 14 | # YOLOv8.0n backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 18 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 19 | - [-1, 3, C2f, [128, True]] 20 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 21 | - [-1, 6, C2f, [256, True]] 22 | - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16 23 | - [-1, 6, C2f, [512, True]] 24 | - [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32 25 | - [-1, 3, C2f, [1024, True]] 26 | - [-1, 1, SPPF, [1024, 5]] # 9 27 | 28 | # YOLOv8.0n head 29 | head: 30 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 31 | - [[-1, 6], 1, Concat, [1]] # cat backbone P4 32 | - [-1, 3, C2f, [512]] # 12 33 | 34 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 35 | - [[-1, 4], 1, Concat, [1]] # cat backbone P3 36 | - [-1, 3, C2f, [256]] # 15 (P3/8-small) 37 | 38 | - [-1, 1, Conv, [256, 3, 2]] 39 | - [[-1, 12], 1, Concat, [1]] # cat head P4 40 | - [-1, 3, C2f, [512]] # 18 (P4/16-medium) 41 | 42 | - [-1, 1, Conv, [512, 3, 2]] 43 | - [[-1, 9], 1, Concat, [1]] # cat head P5 44 | - [-1, 3, C2f, [1024]] # 21 (P5/32-large) 45 | 46 | - [[15, 18, 21], 1, RepHead, [nc]] # Detect(P3, P4, P5) 47 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-SCDown.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOP 13 | 14 | # YOLOv8.0n backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 18 | - [-1, 1, SCDown, [128, 3, 2]] # 1-P2/4 19 | - [-1, 3, C2f, [128, True]] 20 | - [-1, 1, SCDown, [256, 3, 2]] # 3-P3/8 21 | - [-1, 6, C2f, [256, True]] 22 | - [-1, 1, SCDown, [512, 3, 2]] # 5-P4/16 23 | - [-1, 6, C2f, [512, True]] 24 | - [-1, 1, SCDown, [1024, 3, 2]] # 7-P5/32 25 | - [-1, 3, C2f, [1024, True]] 26 | - [-1, 1, SPPF, [1024, 5]] # 9 27 | 28 | # YOLOv8.0n head 29 | head: 30 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 31 | - [[-1, 6], 1, Concat, [1]] # cat backbone P4 32 | - [-1, 3, C2f, [512]] # 12 33 | 34 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 35 | - [[-1, 4], 1, Concat, [1]] # cat backbone P3 36 | - [-1, 3, C2f, [256]] # 15 (P3/8-small) 37 | 38 | 39 | - [-1, 1, SCDown, [256, 3, 2]] 40 | - [[-1, 12], 1, Concat, [1]] # cat head P4 41 | - [-1, 3, C2f, [512]] # 18 (P4/16-medium) 42 | 43 | 44 | - [-1, 1, SCDown, [512, 3, 2]] 45 | - [[-1, 9], 1, Concat, [1]] # cat head P5 46 | - [-1, 3, C2f, [1024]] # 21 (P5/32-large) 47 | 48 | - [[15, 18, 21], 1, Detect, [nc]] # Detect(P3, P4, P5) 49 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-SCDown2.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOP 13 | 14 | # YOLOv8.0n backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 18 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 19 | - [-1, 3, C2f, [128, True]] 20 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 21 | - [-1, 6, C2f, [256, True]] 22 | - [-1, 1, SCDown, [512, 3, 2]] # 5-P4/16 23 | - [-1, 6, C2f, [512, True]] 24 | - [-1, 1, SCDown, [1024, 3, 2]] # 7-P5/32 25 | - [-1, 3, C2f, [1024, True]] 26 | - [-1, 1, SPPF, [1024, 5]] # 9 27 | 28 | # YOLOv8.0n head 29 | head: 30 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 31 | - [[-1, 6], 1, Concat, [1]] # cat backbone P4 32 | - [-1, 3, C2f, [512]] # 12 33 | 34 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 35 | - [[-1, 4], 1, Concat, [1]] # cat backbone P3 36 | - [-1, 3, C2f, [256]] # 15 (P3/8-small) 37 | 38 | 39 | - [-1, 1, Conv, [256, 3, 2]] 40 | - [[-1, 12], 1, Concat, [1]] # cat head P4 41 | - [-1, 3, C2f, [512]] # 18 (P4/16-medium) 42 | 43 | 44 | - [-1, 1, SCDown, [512, 3, 2]] 45 | - [[-1, 9], 1, Concat, [1]] # cat head P5 46 | - [-1, 3, C2f, [1024]] # 21 (P5/32-large) 47 | 48 | - [[15, 18, 21], 1, Detect, [nc]] # Detect(P3, P4, P5) 49 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-SCSA.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOPs 13 | 14 | # YOLOv8.0n backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 18 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 19 | - [-1, 3, C2f, [128, True]] 20 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 21 | - [-1, 6, C2f, [256, True]] 22 | - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16 23 | - [-1, 6, C2f, [512, True]] 24 | - [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32 25 | - [-1, 3, C2f, [1024, True]] 26 | - [-1, 1, SPPF, [1024, 5]] # 9 27 | - [-1, 1, SCSA, [1024]] # 10 28 | 29 | # YOLOv8.0n head 30 | head: 31 | - [-1, 1, nn.Upsample, [None, 2, "nearest"]] 32 | - [[-1, 6], 1, Concat, [1]] # cat backbone P4 33 | - [-1, 3, C2f, [512]] # 13 34 | 35 | - [-1, 1, nn.Upsample, [None, 2, "nearest"]] 36 | - [[-1, 4], 1, Concat, [1]] # cat backbone P3 37 | - [-1, 3, C2f, [256]] # 16 (P3/8-small) 38 | 39 | - [-1, 1, Conv, [256, 3, 2]] 40 | - [[-1, 13], 1, Concat, [1]] # cat head P4 41 | - [-1, 3, C2f, [512]] # 19 (P4/16-medium) 42 | 43 | - [-1, 1, Conv, [512, 3, 2]] 44 | - [[-1, 10], 1, Concat, [1]] # cat head P5 45 | - [-1, 3, C2f, [1024]] # 22 (P5/32-large) 46 | 47 | - [[16, 19, 22], 1, Detect, [nc]] # Detect(P3, P4, P5) 48 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-SK.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model. More improvement points for YOLOv8, please see https://github.com/iscyy/ultralyticsPro 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOPs 13 | 14 | # YOLOv8.0n backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 18 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 19 | - [-1, 3, C2f, [128, True]] 20 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 21 | - [-1, 6, C2f, [256, True]] 22 | - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16 23 | - [-1, 6, C2f, [512, True]] 24 | - [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32 25 | - [-1, 3, C2f, [1024, True]] 26 | - [-1, 3, SKAttention, [1024]] 27 | - [-1, 1, SPPF, [1024, 5]] # 9 28 | 29 | # YOLOv8.0n head 30 | head: 31 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 32 | - [[-1, 6], 1, Concat, [1]] # cat backbone P4 33 | - [-1, 3, C2f, [512]] # 12 34 | 35 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 36 | - [[-1, 4], 1, Concat, [1]] # cat backbone P3 37 | - [-1, 3, C2f, [256]] # 15 (P3/8-small) 38 | 39 | - [-1, 1, Conv, [256, 3, 2]] 40 | - [[-1, 13], 1, Concat, [1]] # cat head P4 41 | - [-1, 3, C2f, [512]] # 18 (P4/16-medium) 42 | 43 | - [-1, 1, Conv, [512, 3, 2]] 44 | - [[-1, 10], 1, Concat, [1]] # cat head P5 45 | - [-1, 3, C2f, [1024]] # 21 (P5/32-large) 46 | 47 | - [[16, 19, 22], 1, Detect, [nc]] # Detect(P3, P4, P5) 48 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-SWS.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOPs 13 | 14 | # YOLOv8.0n backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 18 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 19 | - [-1, 3, C2f, [128, True]] 20 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 21 | - [-1, 6, C2f, [256, True]] 22 | - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16 23 | - [-1, 6, C2f, [512, True]] 24 | - [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32 25 | - [-1, 3, C2f, [1024, True]] 26 | - [-1, 1, SPPF, [1024, 5]] # 9 27 | - [-1, 1, SimAMWithSlicing, [1024]] # 10 28 | 29 | # YOLOv8.0n head 30 | head: 31 | - [-1, 1, nn.Upsample, [None, 2, "nearest"]] 32 | - [[-1, 6], 1, Concat, [1]] # cat backbone P4 33 | - [-1, 3, C2f, [512]] # 13 34 | 35 | - [-1, 1, nn.Upsample, [None, 2, "nearest"]] 36 | - [[-1, 4], 1, Concat, [1]] # cat backbone P3 37 | - [-1, 3, C2f, [256]] # 16 (P3/8-small) 38 | 39 | - [-1, 1, Conv, [256, 3, 2]] 40 | - [[-1, 13], 1, Concat, [1]] # cat head P4 41 | - [-1, 3, C2f, [512]] # 19(P4/16-medium) 42 | 43 | - [-1, 1, Conv, [512, 3, 2]] 44 | - [[-1, 10], 1, Concat, [1]] # cat head P5 45 | - [-1, 3, C2f, [1024]] # 22 (P5/32-large) 46 | 47 | - [[16, 19, 22], 1, Detect, [nc]] # Detect(P3, P4, P5) 48 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-VanillaNet.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOPs 13 | 14 | # YOLOv8.0n backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, vanillanet_5, []] # 4 18 | - [-1, 1, SPPF, [1024, 5]] # 5 19 | 20 | # YOLOv8.0n head 21 | head: 22 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] # 6 23 | - [[-1, 3], 1, Concat, [1]] # 7 cat backbone P4 24 | - [-1, 3, C2f, [512]] # 8 25 | 26 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] # 9 27 | - [[-1, 2], 1, Concat, [1]] # 10 cat backbone P3 28 | - [-1, 3, C2f, [256]] # 11 (P3/8-small) 29 | 30 | - [-1, 1, Conv, [256, 3, 2]] # 12 31 | - [[-1, 8], 1, Concat, [1]] # 13 cat head P4 32 | - [-1, 3, C2f, [512]] # 14 (P4/16-medium) 33 | 34 | - [-1, 1, Conv, [512, 3, 2]] # 15 35 | - [[-1, 5], 1, Concat, [1]] # 16 cat head P5 36 | - [-1, 3, C2f, [1024]] # 17 (P5/32-large) 37 | 38 | - [[11, 14, 17], 1, Detect, [nc]] # Detect(P3, P4, P5) 39 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-att.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 4 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOPs 13 | 14 | # YOLOv8.0n backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 18 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 19 | - [-1, 3, C2f, [128, True]] 20 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 21 | - [-1, 6, C2f, [256, True]] 22 | - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16 23 | - [-1, 6, C2f, [512, True]] 24 | - [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32 25 | - [-1, 3, C2f_DCN, [1024, True]] 26 | - [-1, 1, SPPF, [1024, 5]] # 9 27 | - [-1, 1, ImprovedSimAM, []] #10 28 | 29 | # YOLOv8.0n head 30 | head: 31 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 32 | - [[-1, 6], 1, Concat, [1]] # cat backbone P4 33 | - [-1, 3, C2f, [512]] # 12 34 | 35 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 36 | - [[-1, 4], 1, Concat, [1]] # cat backbone P3 37 | - [-1, 3, C2f, [256]] # 15 (P3/8-small) 38 | 39 | - [-1, 1, Conv, [256, 3, 2]] 40 | - [[-1, 13], 1, Concat, [1]] # cat head P4 41 | - [-1, 3, C2f, [512]] # 18 (P4/16-medium) 42 | 43 | - [-1, 1, Conv, [512, 3, 2]] 44 | - [[-1, 10], 1, Concat, [1]] # cat head P5 45 | - [-1, 3, C2f, [1024]] # 21 (P5/32-large) 46 | 47 | - [[16, 19, 22], 1, Detect, [nc]] # Detect(P3, P4, P5) -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-c2f.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 4 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOPs 13 | 14 | # YOLOv8.0n backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 18 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 19 | - [-1, 3, C2f, [128, True]] 20 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 21 | - [-1, 6, C2f_DCN, [256, True]] 22 | - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16 23 | - [-1, 6, C2f_DCN, [512, True]] 24 | - [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32 25 | - [-1, 3, C2f_DCN, [1024, True]] 26 | - [-1, 1, SPPF, [1024, 5]] # 9 27 | 28 | # YOLOv8.0n head 29 | head: 30 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 31 | - [[-1, 6], 1, Concat, [1]] # cat backbone P4 32 | - [-1, 3, C2f, [512]] # 12 33 | 34 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 35 | - [[-1, 4], 1, Concat, [1]] # cat backbone P3 36 | - [-1, 3, C2f, [256]] # 15 (P3/8-small) 37 | 38 | - [-1, 1, Conv, [256, 3, 2]] 39 | - [[-1, 12], 1, Concat, [1]] # cat head P4 40 | - [-1, 3, C2f, [512]] # 18 (P4/16-medium) 41 | 42 | - [-1, 1, Conv, [512, 3, 2]] 43 | - [[-1, 9], 1, Concat, [1]] # cat head P5 44 | - [-1, 3, C2f, [1024]] # 21 (P5/32-large) 45 | 46 | - [[15, 18, 21], 1, Detect, [nc]] # Detect(P3, P4, P5) 47 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-cls.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8-cls image classification model. For Usage examples see https://docs.ultralytics.com/tasks/classify 3 | 4 | # Parameters 5 | nc: 1000 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n-cls.yaml' will call yolov8-cls.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] 9 | s: [0.33, 0.50, 1024] 10 | m: [0.67, 0.75, 1024] 11 | l: [1.00, 1.00, 1024] 12 | x: [1.00, 1.25, 1024] 13 | 14 | # YOLOv8.0n backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 18 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 19 | - [-1, 3, C2f, [128, True]] 20 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 21 | - [-1, 6, C2f, [256, True]] 22 | - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16 23 | - [-1, 6, C2f, [512, True]] 24 | - [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32 25 | - [-1, 3, C2f, [1024, True]] 26 | 27 | # YOLOv8.0n head 28 | head: 29 | - [-1, 1, Classify, [nc]] # Classify 30 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-emo.py: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOPs 13 | 14 | 15 | # 可替换的版本有如下 16 | # __all__ = ['EMO_1M', 'EMO_2M', 'EMO_5M', 'EMO_6M'] 17 | # YOLOv8.0n backbone 18 | backbone: 19 | # [from, repeats, module, args] 20 | - [-1, 1, EMO_1M, []] # 4 21 | - [-1, 1, SPPF, [1024, 5]] # 5 22 | 23 | # YOLOv8.0n head 24 | head: 25 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] # 6 26 | - [[-1, 3], 1, Concat, [1]] # 7 cat backbone P4 27 | - [-1, 3, C2f, [512]] # 8 28 | 29 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] # 9 30 | - [[-1, 2], 1, Concat, [1]] # 10 cat backbone P3 31 | - [-1, 3, C2f, [256]] # 11 (P3/8-small) 32 | 33 | - [-1, 1, Conv, [256, 3, 2]] # 12 34 | - [[-1, 8], 1, Concat, [1]] # 13 cat head P4 35 | - [-1, 3, C2f, [512]] # 14 (P4/16-medium) 36 | 37 | - [-1, 1, Conv, [512, 3, 2]] # 15 38 | - [[-1, 5], 1, Concat, [1]] # 16 cat head P5 39 | - [-1, 3, C2f, [1024]] # 17 (P5/32-large) 40 | 41 | - [[11, 14, 17], 1, Detect, [nc]] # Detect(P3, P4, P5) 42 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-p2.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P2-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] 9 | s: [0.33, 0.50, 1024] 10 | m: [0.67, 0.75, 768] 11 | l: [1.00, 1.00, 512] 12 | x: [1.00, 1.25, 512] 13 | 14 | # YOLOv8.0 backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 18 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 19 | - [-1, 3, C2f, [128, True]] 20 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 21 | - [-1, 6, C2f, [256, True]] 22 | - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16 23 | - [-1, 6, C2f, [512, True]] 24 | - [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32 25 | - [-1, 3, C2f, [1024, True]] 26 | - [-1, 1, SPPF, [1024, 5]] # 9 27 | 28 | # YOLOv8.0-p2 head 29 | head: 30 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 31 | - [[-1, 6], 1, Concat, [1]] # cat backbone P4 32 | - [-1, 3, C2f, [512]] # 12 33 | 34 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 35 | - [[-1, 4], 1, Concat, [1]] # cat backbone P3 36 | - [-1, 3, C2f, [256]] # 15 (P3/8-small) 37 | 38 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 39 | - [[-1, 2], 1, Concat, [1]] # cat backbone P2 40 | - [-1, 3, C2f, [128]] # 18 (P2/4-xsmall) 41 | 42 | - [-1, 1, Conv, [128, 3, 2]] 43 | - [[-1, 15], 1, Concat, [1]] # cat head P3 44 | - [-1, 3, C2f, [256]] # 21 (P3/8-small) 45 | 46 | - [-1, 1, Conv, [256, 3, 2]] 47 | - [[-1, 12], 1, Concat, [1]] # cat head P4 48 | - [-1, 3, C2f, [512]] # 24 (P4/16-medium) 49 | 50 | - [-1, 1, Conv, [512, 3, 2]] 51 | - [[-1, 9], 1, Concat, [1]] # cat head P5 52 | - [-1, 3, C2f, [1024]] # 27 (P5/32-large) 53 | 54 | - [[18, 21, 24, 27], 1, Detect, [nc]] # Detect(P2, P3, P4, P5) 55 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-p6.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P6 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n-p6.yaml' will call yolov8-p6.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] 9 | s: [0.33, 0.50, 1024] 10 | m: [0.67, 0.75, 768] 11 | l: [1.00, 1.00, 512] 12 | x: [1.00, 1.25, 512] 13 | 14 | # YOLOv8.0x6 backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 18 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 19 | - [-1, 3, C2f, [128, True]] 20 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 21 | - [-1, 6, C2f, [256, True]] 22 | - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16 23 | - [-1, 6, C2f, [512, True]] 24 | - [-1, 1, Conv, [768, 3, 2]] # 7-P5/32 25 | - [-1, 3, C2f, [768, True]] 26 | - [-1, 1, Conv, [1024, 3, 2]] # 9-P6/64 27 | - [-1, 3, C2f, [1024, True]] 28 | - [-1, 1, SPPF, [1024, 5]] # 11 29 | 30 | # YOLOv8.0x6 head 31 | head: 32 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 33 | - [[-1, 8], 1, Concat, [1]] # cat backbone P5 34 | - [-1, 3, C2, [768, False]] # 14 35 | 36 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 37 | - [[-1, 6], 1, Concat, [1]] # cat backbone P4 38 | - [-1, 3, C2, [512, False]] # 17 39 | 40 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 41 | - [[-1, 4], 1, Concat, [1]] # cat backbone P3 42 | - [-1, 3, C2, [256, False]] # 20 (P3/8-small) 43 | 44 | - [-1, 1, Conv, [256, 3, 2]] 45 | - [[-1, 17], 1, Concat, [1]] # cat head P4 46 | - [-1, 3, C2, [512, False]] # 23 (P4/16-medium) 47 | 48 | - [-1, 1, Conv, [512, 3, 2]] 49 | - [[-1, 14], 1, Concat, [1]] # cat head P5 50 | - [-1, 3, C2, [768, False]] # 26 (P5/32-large) 51 | 52 | - [-1, 1, Conv, [768, 3, 2]] 53 | - [[-1, 11], 1, Concat, [1]] # cat head P6 54 | - [-1, 3, C2, [1024, False]] # 29 (P6/64-xlarge) 55 | 56 | - [[20, 23, 26, 29], 1, Detect, [nc]] # Detect(P3, P4, P5, P6) 57 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-pose.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8-pose keypoints/pose estimation model. For Usage examples see https://docs.ultralytics.com/tasks/pose 3 | 4 | # Parameters 5 | nc: 1 # number of classes 6 | kpt_shape: [17, 3] # number of keypoints, number of dims (2 for x,y or 3 for x,y,visible) 7 | scales: # model compound scaling constants, i.e. 'model=yolov8n-pose.yaml' will call yolov8-pose.yaml with scale 'n' 8 | # [depth, width, max_channels] 9 | n: [0.33, 0.25, 1024] 10 | s: [0.33, 0.50, 1024] 11 | m: [0.67, 0.75, 768] 12 | l: [1.00, 1.00, 512] 13 | x: [1.00, 1.25, 512] 14 | 15 | # YOLOv8.0n backbone 16 | backbone: 17 | # [from, repeats, module, args] 18 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 19 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 20 | - [-1, 3, C2f, [128, True]] 21 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 22 | - [-1, 6, C2f, [256, True]] 23 | - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16 24 | - [-1, 6, C2f, [512, True]] 25 | - [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32 26 | - [-1, 3, C2f, [1024, True]] 27 | - [-1, 1, SPPF, [1024, 5]] # 9 28 | 29 | # YOLOv8.0n head 30 | head: 31 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 32 | - [[-1, 6], 1, Concat, [1]] # cat backbone P4 33 | - [-1, 3, C2f, [512]] # 12 34 | 35 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 36 | - [[-1, 4], 1, Concat, [1]] # cat backbone P3 37 | - [-1, 3, C2f, [256]] # 15 (P3/8-small) 38 | 39 | - [-1, 1, Conv, [256, 3, 2]] 40 | - [[-1, 12], 1, Concat, [1]] # cat head P4 41 | - [-1, 3, C2f, [512]] # 18 (P4/16-medium) 42 | 43 | - [-1, 1, Conv, [512, 3, 2]] 44 | - [[-1, 9], 1, Concat, [1]] # cat head P5 45 | - [-1, 3, C2f, [1024]] # 21 (P5/32-large) 46 | 47 | - [[15, 18, 21], 1, Pose, [nc, kpt_shape]] # Pose(P3, P4, P5) 48 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-rtdetr.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOPs 13 | 14 | # YOLOv8.0n backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 18 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 19 | - [-1, 3, C2f, [128, True]] 20 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 21 | - [-1, 6, C2f, [256, True]] 22 | - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16 23 | - [-1, 6, C2f, [512, True]] 24 | - [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32 25 | - [-1, 3, C2f, [1024, True]] 26 | - [-1, 1, SPPF, [1024, 5]] # 9 27 | 28 | # YOLOv8.0n head 29 | head: 30 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 31 | - [[-1, 6], 1, Concat, [1]] # cat backbone P4 32 | - [-1, 3, C2f, [512]] # 12 33 | 34 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 35 | - [[-1, 4], 1, Concat, [1]] # cat backbone P3 36 | - [-1, 3, C2f, [256]] # 15 (P3/8-small) 37 | 38 | - [-1, 1, Conv, [256, 3, 2]] 39 | - [[-1, 12], 1, Concat, [1]] # cat head P4 40 | - [-1, 3, C2f, [512]] # 18 (P4/16-medium) 41 | 42 | - [-1, 1, Conv, [512, 3, 2]] 43 | - [[-1, 9], 1, Concat, [1]] # cat head P5 44 | - [-1, 3, C2f, [1024]] # 21 (P5/32-large) 45 | 46 | - [[15, 18, 21], 1, RTDETRDecoder, [nc]] # Detect(P3, P4, P5) 47 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-s-Detect_improve.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOPs 13 | 14 | # YOLOv8.0n backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 18 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 19 | - [-1, 3, C2f, [128, True]] 20 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 21 | - [-1, 6, C2f, [256, True]] 22 | - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16 23 | - [-1, 6, C2f, [512, True]] 24 | - [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32 25 | - [-1, 3, C2f_DCN, [1024, True]] 26 | - [-1, 1, SPPF, [1024, 5]] # 9 27 | 28 | # YOLOv8.0n head 29 | head: 30 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 31 | - [[-1, 6], 1, Concat, [1]] # cat backbone P4 32 | - [-1, 3, C2f, [512]] # 12 33 | 34 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 35 | - [[-1, 4], 1, Concat, [1]] # cat backbone P3 36 | - [-1, 3, C2f, [256]] # 15 (P3/8-small) 37 | 38 | - [-1, 1, Conv, [256, 3, 2]] 39 | - [[-1, 12], 1, Concat, [1]] # cat head P4 40 | - [-1, 3, C2f, [512]] # 18 (P4/16-medium) 41 | 42 | - [-1, 1, Conv, [512, 3, 2]] 43 | - [[-1, 9], 1, Concat, [1]] # cat head P5 44 | - [-1, 3, C2f, [1024]] # 21 (P5/32-large) 45 | 46 | - [[15, 18, 21], 1, Detect_improve, [nc]] # Detect(P3, P4, P5) 47 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-seg-p6.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8-seg-p6 instance segmentation model. For Usage examples see https://docs.ultralytics.com/tasks/segment 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n-seg-p6.yaml' will call yolov8-seg-p6.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] 9 | s: [0.33, 0.50, 1024] 10 | m: [0.67, 0.75, 768] 11 | l: [1.00, 1.00, 512] 12 | x: [1.00, 1.25, 512] 13 | 14 | # YOLOv8.0x6 backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 18 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 19 | - [-1, 3, C2f, [128, True]] 20 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 21 | - [-1, 6, C2f, [256, True]] 22 | - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16 23 | - [-1, 6, C2f, [512, True]] 24 | - [-1, 1, Conv, [768, 3, 2]] # 7-P5/32 25 | - [-1, 3, C2f, [768, True]] 26 | - [-1, 1, Conv, [1024, 3, 2]] # 9-P6/64 27 | - [-1, 3, C2f, [1024, True]] 28 | - [-1, 1, SPPF, [1024, 5]] # 11 29 | 30 | # YOLOv8.0x6 head 31 | head: 32 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 33 | - [[-1, 8], 1, Concat, [1]] # cat backbone P5 34 | - [-1, 3, C2, [768, False]] # 14 35 | 36 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 37 | - [[-1, 6], 1, Concat, [1]] # cat backbone P4 38 | - [-1, 3, C2, [512, False]] # 17 39 | 40 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 41 | - [[-1, 4], 1, Concat, [1]] # cat backbone P3 42 | - [-1, 3, C2, [256, False]] # 20 (P3/8-small) 43 | 44 | - [-1, 1, Conv, [256, 3, 2]] 45 | - [[-1, 17], 1, Concat, [1]] # cat head P4 46 | - [-1, 3, C2, [512, False]] # 23 (P4/16-medium) 47 | 48 | - [-1, 1, Conv, [512, 3, 2]] 49 | - [[-1, 14], 1, Concat, [1]] # cat head P5 50 | - [-1, 3, C2, [768, False]] # 26 (P5/32-large) 51 | 52 | - [-1, 1, Conv, [768, 3, 2]] 53 | - [[-1, 11], 1, Concat, [1]] # cat head P6 54 | - [-1, 3, C2, [1024, False]] # 29 (P6/64-xlarge) 55 | 56 | - [[20, 23, 26, 29], 1, Segment, [nc, 32, 256]] # Pose(P3, P4, P5, P6) 57 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-seg.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8-seg instance segmentation model. For Usage examples see https://docs.ultralytics.com/tasks/segment 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n-seg.yaml' will call yolov8-seg.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] 9 | s: [0.33, 0.50, 1024] 10 | m: [0.67, 0.75, 768] 11 | l: [1.00, 1.00, 512] 12 | x: [1.00, 1.25, 512] 13 | 14 | # YOLOv8.0n backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 18 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 19 | - [-1, 3, C2f, [128, True]] 20 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 21 | - [-1, 6, C2f, [256, True]] 22 | - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16 23 | - [-1, 6, C2f, [512, True]] 24 | - [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32 25 | - [-1, 3, C2f, [1024, True]] 26 | - [-1, 1, SPPF, [1024, 5]] # 9 27 | 28 | # YOLOv8.0n head 29 | head: 30 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 31 | - [[-1, 6], 1, Concat, [1]] # cat backbone P4 32 | - [-1, 3, C2f, [512]] # 12 33 | 34 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 35 | - [[-1, 4], 1, Concat, [1]] # cat backbone P3 36 | - [-1, 3, C2f, [256]] # 15 (P3/8-small) 37 | 38 | - [-1, 1, Conv, [256, 3, 2]] 39 | - [[-1, 12], 1, Concat, [1]] # cat head P4 40 | - [-1, 3, C2f, [512]] # 18 (P4/16-medium) 41 | 42 | - [-1, 1, Conv, [512, 3, 2]] 43 | - [[-1, 9], 1, Concat, [1]] # cat head P5 44 | - [-1, 3, C2f, [1024]] # 21 (P5/32-large) 45 | 46 | - [[15, 18, 21], 1, Segment, [nc, 32, 256]] # Segment(P3, P4, P5) 47 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-starnet1.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOPs 13 | 14 | # 0-P1/2 15 | # 1-P2/4 16 | # 2-P3/8 17 | # 3-P4/16 18 | # 4-P5/32 19 | 20 | # YOLOv8.0n backbone 21 | backbone: 22 | # [from, repeats, module, args] 23 | - [-1, 1, starnet_s1 , []] # 4 24 | - [-1, 1, SPPF, [1024, 5]] # 5 25 | 26 | # YOLOv8.0n head 27 | head: 28 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] # 6 29 | - [[-1, 3], 1, Concat, [1]] # 7 cat backbone P4 30 | - [-1, 3, C2f, [512]] # 8 31 | 32 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] # 9 33 | - [[-1, 2], 1, Concat, [1]] # 10 cat backbone P3 34 | - [-1, 3, C2f, [256]] # 11 (P3/8-small) 35 | 36 | - [-1, 1, Conv, [256, 3, 2]] # 12 37 | - [[-1, 8], 1, Concat, [1]] # 13 cat head P4 38 | - [-1, 3, C2f, [512]] # 14 (P4/16-medium) 39 | 40 | - [-1, 1, Conv, [512, 3, 2]] # 15 41 | - [[-1, 5], 1, Concat, [1]] # 16 cat head P5 42 | - [-1, 3, C2f, [1024]] # 17 (P5/32-large) 43 | 44 | - [[11, 14, 17], 1, Detect, [nc]] # Detect(P3, P4, P5) 45 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-starnet2.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOPs 13 | 14 | # 0-P1/2 15 | # 1-P2/4 16 | # 2-P3/8 17 | # 3-P4/16 18 | # 4-P5/32 19 | 20 | # YOLOv8.0n backbone 21 | backbone: 22 | # [from, repeats, module, args] 23 | - [-1, 1, starnet_s2 , []] # 4 24 | - [-1, 1, SPPF, [1024, 5]] # 5 25 | 26 | # YOLOv8.0n head 27 | head: 28 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] # 6 29 | - [[-1, 3], 1, Concat, [1]] # 7 cat backbone P4 30 | - [-1, 3, C2f, [512]] # 8 31 | 32 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] # 9 33 | - [[-1, 2], 1, Concat, [1]] # 10 cat backbone P3 34 | - [-1, 3, C2f, [256]] # 11 (P3/8-small) 35 | 36 | - [-1, 1, Conv, [256, 3, 2]] # 12 37 | - [[-1, 8], 1, Concat, [1]] # 13 cat head P4 38 | - [-1, 3, C2f, [512]] # 14 (P4/16-medium) 39 | 40 | - [-1, 1, Conv, [512, 3, 2]] # 15 41 | - [[-1, 5], 1, Concat, [1]] # 16 cat head P5 42 | - [-1, 3, C2f, [1024]] # 17 (P5/32-large) 43 | 44 | - [[11, 14, 17], 1, Detect, [nc]] # Detect(P3, P4, P5) 45 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-starnet3.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOPs 13 | 14 | # 0-P1/2 15 | # 1-P2/4 16 | # 2-P3/8 17 | # 3-P4/16 18 | # 4-P5/32 19 | 20 | # YOLOv8.0n backbone 21 | backbone: 22 | # [from, repeats, module, args] 23 | - [-1, 1, starnet_s3 , []] # 4 24 | - [-1, 1, SPPF, [1024, 5]] # 5 25 | 26 | # YOLOv8.0n head 27 | head: 28 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] # 6 29 | - [[-1, 3], 1, Concat, [1]] # 7 cat backbone P4 30 | - [-1, 3, C2f, [512]] # 8 31 | 32 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] # 9 33 | - [[-1, 2], 1, Concat, [1]] # 10 cat backbone P3 34 | - [-1, 3, C2f, [256]] # 11 (P3/8-small) 35 | 36 | - [-1, 1, Conv, [256, 3, 2]] # 12 37 | - [[-1, 8], 1, Concat, [1]] # 13 cat head P4 38 | - [-1, 3, C2f, [512]] # 14 (P4/16-medium) 39 | 40 | - [-1, 1, Conv, [512, 3, 2]] # 15 41 | - [[-1, 5], 1, Concat, [1]] # 16 cat head P5 42 | - [-1, 3, C2f, [1024]] # 17 (P5/32-large) 43 | 44 | - [[11, 14, 17], 1, Detect, [nc]] # Detect(P3, P4, P5) 45 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8-starnet4.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOPs 13 | 14 | # 0-P1/2 15 | # 1-P2/4 16 | # 2-P3/8 17 | # 3-P4/16 18 | # 4-P5/32 19 | 20 | # YOLOv8.0n backbone 21 | backbone: 22 | # [from, repeats, module, args] 23 | - [-1, 1, starnet_s4 , []] # 4 24 | - [-1, 1, SPPF, [1024, 5]] # 5 25 | 26 | # YOLOv8.0n head 27 | head: 28 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] # 6 29 | - [[-1, 3], 1, Concat, [1]] # 7 cat backbone P4 30 | - [-1, 3, C2f, [512]] # 8 31 | 32 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] # 9 33 | - [[-1, 2], 1, Concat, [1]] # 10 cat backbone P3 34 | - [-1, 3, C2f, [256]] # 11 (P3/8-small) 35 | 36 | - [-1, 1, Conv, [256, 3, 2]] # 12 37 | - [[-1, 8], 1, Concat, [1]] # 13 cat head P4 38 | - [-1, 3, C2f, [512]] # 14 (P4/16-medium) 39 | 40 | - [-1, 1, Conv, [512, 3, 2]] # 15 41 | - [[-1, 5], 1, Concat, [1]] # 16 cat head P5 42 | - [-1, 3, C2f, [1024]] # 17 (P5/32-large) 43 | 44 | - [[11, 14, 17], 1, Detect, [nc]] # Detect(P3, P4, P5) 45 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOPs 13 | 14 | # YOLOv8.0n backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 18 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 19 | - [-1, 3, C2f, [128, True]] 20 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 21 | - [-1, 6, C2f, [256, True]] 22 | - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16 23 | - [-1, 6, C2f, [512, True]] 24 | - [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32 25 | - [-1, 3, C2f, [1024, True]] 26 | - [-1, 1, SPPF, [1024, 5]] # 9 27 | 28 | # YOLOv8.0n head 29 | head: 30 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 31 | - [[-1, 6], 1, Concat, [1]] # cat backbone P4 32 | - [-1, 3, C2f, [512]] # 12 33 | 34 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 35 | - [[-1, 4], 1, Concat, [1]] # cat backbone P3 36 | - [-1, 3, C2f, [256]] # 15 (P3/8-small) 37 | 38 | - [-1, 1, Conv, [256, 3, 2]] 39 | - [[-1, 12], 1, Concat, [1]] # cat head P4 40 | - [-1, 3, C2f, [512]] # 18 (P4/16-medium) 41 | 42 | - [-1, 1, Conv, [512, 3, 2]] 43 | - [[-1, 9], 1, Concat, [1]] # cat head P5 44 | - [-1, 3, C2f, [1024]] # 21 (P5/32-large) 45 | 46 | - [[15, 18, 21], 1, Detect, [nc]] # Detect(P3, P4, P5) 47 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8_C2f_KAN.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 80 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOPs 13 | 14 | # YOLOv8.0n backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 18 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 19 | - [-1, 3, C2f_KAN, [128, True]] 20 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 21 | - [-1, 6, C2f_KAN, [256, True]] 22 | - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16 23 | - [-1, 6, C2f_KAN, [512, True]] 24 | - [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32 25 | - [-1, 3, C2f_KAN, [1024, True]] 26 | - [-1, 1, SPPF, [1024, 5]] # 9 27 | 28 | # YOLOv8.0n head 29 | head: 30 | - [-1, 1, nn.Upsample, [None, 2, "nearest"]] 31 | - [[-1, 6], 1, Concat, [1]] # cat backbone P4 32 | - [-1, 3, C2f, [512]] # 12 33 | 34 | - [-1, 1, nn.Upsample, [None, 2, "nearest"]] 35 | - [[-1, 4], 1, Concat, [1]] # cat backbone P3 36 | - [-1, 3, C2f, [256]] # 15 (P3/8-small) 37 | 38 | - [-1, 1, Conv, [256, 3, 2]] 39 | - [[-1, 12], 1, Concat, [1]] # cat head P4 40 | - [-1, 3, C2f, [512]] # 18 (P4/16-medium) 41 | 42 | - [-1, 1, Conv, [512, 3, 2]] 43 | - [[-1, 9], 1, Concat, [1]] # cat head P5 44 | - [-1, 3, C2f, [1024]] # 21 (P5/32-large) 45 | 46 | - [[15, 18, 21], 1, Detect, [nc]] # Detect(P3, P4, P5) 47 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8_GFPN.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, GPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 4 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOPs 13 | 14 | # YOLOv8.0n backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 18 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 19 | - [-1, 3, C2f, [128, True]] 20 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 21 | - [-1, 6, C2f, [256, True]] 22 | - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16 23 | - [-1, 6, C2f, [512, True]] 24 | - [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32 25 | - [-1, 3, C2f, [1024, True]] 26 | - [-1, 1, SPPF, [1024, 5]] # 9 27 | 28 | # YOLOv8.0n head 29 | head: 30 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 31 | - [[-1, 6], 1, Concat, [1]] # cat backbone P4 32 | - [-1, 3, CSPStage, [512]] # 12 33 | 34 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 35 | - [[-1, 4], 1, Concat, [1]] # cat backbone P3 36 | - [-1, 3, CSPStage, [256]] # 15 (P3/8-small) 37 | 38 | - [-1, 1, Conv, [256, 3, 2]] 39 | - [[-1, 12], 1, Concat, [1]] # cat head P4 40 | - [-1, 3, CSPStage, [512]] # 18 (P4/16-medium) 41 | 42 | - [-1, 1, Conv, [512, 3, 2]] 43 | - [[-1, 9], 1, Concat, [1]] # cat head P5 44 | - [-1, 3, CSPStage, [1024]] # 21 (P5/32-large) 45 | 46 | - [[15, 18, 21], 1, Detect, [nc]] # Detect(P3, P4, P5) 47 | -------------------------------------------------------------------------------- /ultralytics/cfg/models/v8/yolov8_SPPF_LSKA.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, GPL-3.0 license 2 | # YOLOv8 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect 3 | 4 | # Parameters 5 | nc: 1 # number of classes 6 | scales: # model compound scaling constants, i.e. 'model=yolov8n.yaml' will call yolov8.yaml with scale 'n' 7 | # [depth, width, max_channels] 8 | n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs 9 | s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs 10 | m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs 11 | l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs 12 | x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOPs 13 | 14 | # YOLOv8.0n backbone 15 | backbone: 16 | # [from, repeats, module, args] 17 | - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 18 | - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 19 | - [-1, 3, C2f, [128, True]] 20 | - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8 21 | - [-1, 6, C2f, [256, True]] 22 | - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16 23 | - [-1, 6, C2f, [512, True]] 24 | - [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32 25 | - [-1, 3, C2f, [1024, True]] 26 | - [-1, 1, SPPF_LSKA, [1024, 5]] # 9 27 | 28 | # YOLOv8.0n head 29 | head: 30 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 31 | - [[-1, 6], 1, Concat, [1]] # cat backbone P4 32 | - [-1, 3, C2f, [512]] # 12 33 | 34 | - [-1, 1, nn.Upsample, [None, 2, 'nearest']] 35 | - [[-1, 4], 1, Concat, [1]] # cat backbone P3 36 | - [-1, 3, C2f, [256]] # 15 (P3/8-small) 37 | 38 | - [-1, 1, Conv, [256, 3, 2]] 39 | - [[-1, 12], 1, Concat, [1]] # cat head P4 40 | - [-1, 3, C2f, [512]] # 18 (P4/16-medium) 41 | 42 | - [-1, 1, Conv, [512, 3, 2]] 43 | - [[-1, 9], 1, Concat, [1]] # cat head P5 44 | - [-1, 3, C2f, [1024]] # 21 (P5/32-large) 45 | 46 | - [[15, 18, 21], 1, Detect, [nc]] # Detect(P3, P4, P5) 47 | -------------------------------------------------------------------------------- /ultralytics/cfg/trackers/botsort.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # Default YOLO tracker settings for BoT-SORT tracker https://github.com/NirAharon/BoT-SORT 3 | 4 | tracker_type: botsort # tracker type, ['botsort', 'bytetrack'] 5 | track_high_thresh: 0.5 # threshold for the first association 6 | track_low_thresh: 0.1 # threshold for the second association 7 | new_track_thresh: 0.6 # threshold for init new track if the detection does not match any tracks 8 | track_buffer: 30 # buffer to calculate the time when to remove tracks 9 | match_thresh: 0.8 # threshold for matching tracks 10 | # min_box_area: 10 # threshold for min box areas(for tracker evaluation, not used for now) 11 | # mot20: False # for tracker evaluation(not used for now) 12 | 13 | # BoT-SORT settings 14 | gmc_method: sparseOptFlow # method of global motion compensation 15 | # ReID model related thresh (not supported yet) 16 | proximity_thresh: 0.5 17 | appearance_thresh: 0.25 18 | with_reid: False 19 | -------------------------------------------------------------------------------- /ultralytics/cfg/trackers/bytetrack.yaml: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | # Default YOLO tracker settings for ByteTrack tracker https://github.com/ifzhang/ByteTrack 3 | 4 | tracker_type: bytetrack # tracker type, ['botsort', 'bytetrack'] 5 | track_high_thresh: 0.5 # threshold for the first association 6 | track_low_thresh: 0.1 # threshold for the second association 7 | new_track_thresh: 0.6 # threshold for init new track if the detection does not match any tracks 8 | track_buffer: 30 # buffer to calculate the time when to remove tracks 9 | match_thresh: 0.8 # threshold for matching tracks 10 | # min_box_area: 10 # threshold for min box areas(for tracker evaluation, not used for now) 11 | # mot20: False # for tracker evaluation(not used for now) 12 | -------------------------------------------------------------------------------- /ultralytics/data/__init__.py: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | 3 | from .base import BaseDataset 4 | from .build import build_dataloader, build_yolo_dataset, load_inference_source 5 | from .dataset import ClassificationDataset, SemanticDataset, YOLODataset 6 | 7 | __all__ = ('BaseDataset', 'ClassificationDataset', 'SemanticDataset', 'YOLODataset', 'build_yolo_dataset', 8 | 'build_dataloader', 'load_inference_source') 9 | -------------------------------------------------------------------------------- /ultralytics/data/scripts/download_weights.sh: -------------------------------------------------------------------------------- 1 | #!/bin/bash 2 | # Ultralytics YOLO 🚀, AGPL-3.0 license 3 | # Download latest models from https://github.com/ultralytics/assets/releases 4 | # Example usage: bash ultralytics/data/scripts/download_weights.sh 5 | # parent 6 | # └── weights 7 | # ├── yolov8n.pt ← downloads here 8 | # ├── yolov8s.pt 9 | # └── ... 10 | 11 | python - < None: 5 | self.data = {} 6 | 7 | def parse(self, line): 8 | def startswith(line, lst): 9 | for ele in lst: 10 | if line.startswith(ele): 11 | return True 12 | return False 13 | 14 | if not startswith(line, ['1', '2', '3', '4', '5', '6', '7', '8', '9']): 15 | return 16 | 17 | eles = line.strip().split() 18 | key = eles[0] 19 | if key not in self.data: 20 | self.data[key] = [] 21 | 22 | self.data[key].append([eles[1], float(eles[2])]) 23 | 24 | def sort(self): 25 | for k, v in self.data.items(): 26 | nv = sorted(v, key=lambda x: x[1]) 27 | self.data[k] = nv 28 | 29 | def display_best(self): 30 | for k, v in self.data.items(): 31 | print(f'{k} \t {v[0][0]} \t {v[0][1]:.4f} \t {v[1][0]} \t {v[1][1]:.4f}') 32 | 33 | def display_best_python(self, output): 34 | res = {} 35 | def parse(spec): 36 | d_stride = int(spec.split('/')[0]) 37 | thread = int(spec.split('/')[1].split('(')[0]) 38 | m = int(spec.split('(')[1].split(')')[0]) 39 | return d_stride, thread, m 40 | 41 | for k, v in self.data.items(): 42 | res[k] = parse(v[0][0]) 43 | 44 | with open(output, "w") as f: 45 | json.dump(res, f, indent=4) 46 | 47 | if __name__ == '__main__': 48 | parser = argparse.ArgumentParser() 49 | parser.add_argument('--input', type=str) 50 | parser.add_argument('--output', type=str) 51 | args = parser.parse_args() 52 | 53 | with open(args.input) as f: 54 | lines = f.readlines() 55 | 56 | lineparser = LineParser() 57 | for line in lines: 58 | lineparser.parse(line) 59 | lineparser.sort() 60 | lineparser.display_best() 61 | lineparser.display_best_python(args.output) -------------------------------------------------------------------------------- /ultralytics/nn/DCNv4_op/scripts/search_bwd.sh: -------------------------------------------------------------------------------- 1 | python search_dcnv4_bwd_engine.py > res_bwd.txt 2 | python find_best.py --input res_bwd.txt --output table_bwd.py -------------------------------------------------------------------------------- /ultralytics/nn/DCNv4_op/scripts/search_dcnv4_bwd_engine.py: -------------------------------------------------------------------------------- 1 | import os 2 | 3 | def factors(N): 4 | res = [] 5 | for i in range(1, N+1): 6 | if N % i == 0: 7 | res.append(i) 8 | return res 9 | 10 | if __name__ == '__main__': 11 | BATCH=64 12 | for N, Hin, Win in [(BATCH, 56, 56), (BATCH, 28, 28), (BATCH, 14, 14), (BATCH, 7, 7), 13 | (1, 200, 320), (1, 100, 160), (1, 50, 80), (1, 25, 40), (1, 64, 64)]: 14 | for group_channel in [16, 32, 64]: 15 | for group in [4, 5, 6, 7, 8]: 16 | for d_stride in [1, 2, 4]: 17 | for m in factors(N*Hin*Win): 18 | if m > 64: 19 | break 20 | block_thread = group * (group_channel//d_stride) * m 21 | if block_thread > 1024: 22 | break 23 | cmd = f"python search_dcnv4_bwd.py --n {N} --h {Hin} --w {Win} --g {group} --c {group_channel} --dstride {d_stride} --blockthread {block_thread} --multiplier {m}" 24 | os.system(cmd) -------------------------------------------------------------------------------- /ultralytics/nn/DCNv4_op/scripts/search_dcnv4_engine.py: -------------------------------------------------------------------------------- 1 | import os 2 | 3 | def factors(N): 4 | res = [] 5 | for i in range(1, N+1): 6 | if N % i == 0: 7 | res.append(i) 8 | return res 9 | 10 | if __name__ == '__main__': 11 | BATCH=64 12 | for group_channel in [16, 32, 64]: 13 | for group in [4, 5, 6, 7, 8]: 14 | for N, Hin, Win in [(BATCH, 56, 56), (BATCH, 28, 28), (BATCH, 14, 14), (BATCH, 7, 7), 15 | (1, 200, 320), (1, 100, 160), (1, 50, 80), (1, 25, 40), (1, 64, 64)]: 16 | for d_stride in [2, 4, 8, 16]: 17 | for m in factors(N*Hin*Win): 18 | if m > 64: 19 | break 20 | block_thread = group * (group_channel//d_stride) * m 21 | if block_thread > 1024: 22 | break 23 | cmd = f"python search_dcnv4.py --n {N} --h {Hin} --w {Win} --g {group} --c {group_channel} --dstride {d_stride} --blockthread {block_thread} --multiplier {m}" 24 | os.system(cmd) -------------------------------------------------------------------------------- /ultralytics/nn/DCNv4_op/scripts/search_fwd.sh: -------------------------------------------------------------------------------- 1 | python search_dcnv4_engine.py > res.txt 2 | python find_best.py --input res.txt --output table.py -------------------------------------------------------------------------------- /ultralytics/nn/DCNv4_op/src/cuda/dcnv4_cuda.h: -------------------------------------------------------------------------------- 1 | /*! 2 | ************************************************************************************************** 3 | * Deformable DETR 4 | * Copyright (c) 2020 SenseTime. All Rights Reserved. 5 | * Licensed under the Apache License, Version 2.0 [see LICENSE for details] 6 | ************************************************************************************************** 7 | * Modified from 8 | *https://github.com/chengdazhi/Deformable-Convolution-V2-PyTorch/tree/pytorch_1.0.0 9 | ************************************************************************************************** 10 | */ 11 | 12 | #pragma once 13 | #include 14 | 15 | at::Tensor dcnv4_cuda_forward( 16 | const at::Tensor &value, 17 | const at::Tensor &p_offset, 18 | const int kernel_h, const int kernel_w, const int stride_h, 19 | const int stride_w, const int pad_h, const int pad_w, const int dilation_h, 20 | const int dilation_w, const int group, const int group_channels, 21 | const float offset_scale, const int im2col_step, const int remove_center, 22 | const int d_stride, const int block_thread, const bool softmax); 23 | 24 | std::vector 25 | dcnv4_cuda_backward( 26 | const at::Tensor &value, 27 | const at::Tensor &p_offset, 28 | const int kernel_h, const int kernel_w, const int stride_h, 29 | const int stride_w, const int pad_h, const int pad_w, const int dilation_h, 30 | const int dilation_w, const int group, const int group_channels, 31 | const float offset_scale, const int im2col_step, const at::Tensor &grad_output, 32 | const int remove_center, const int d_stride, const int block_thread, 33 | const bool softmax); -------------------------------------------------------------------------------- /ultralytics/nn/DCNv4_op/src/cuda/flash_deform_attn_cuda.h: -------------------------------------------------------------------------------- 1 | /*! 2 | ************************************************************************************************** 3 | * Deformable DETR 4 | * Copyright (c) 2020 SenseTime. All Rights Reserved. 5 | * Licensed under the Apache License, Version 2.0 [see LICENSE for details] 6 | ************************************************************************************************** 7 | * Modified from 8 | *https://github.com/chengdazhi/Deformable-Convolution-V2-PyTorch/tree/pytorch_1.0.0 9 | ************************************************************************************************** 10 | */ 11 | 12 | #pragma once 13 | #include 14 | 15 | at::Tensor flash_deform_attn_cuda_forward( 16 | const at::Tensor &value, const at::Tensor &spatial_shapes, 17 | const at::Tensor &level_start_index, const at::Tensor &sampling_loc_attn, 18 | const int im2col_step, const int K, const int d_stride, const int block_thread); 19 | 20 | std::vector 21 | flash_deform_attn_cuda_backward( 22 | const at::Tensor &value, const at::Tensor &spatial_shapes, 23 | const at::Tensor &level_start_index, const at::Tensor &sampling_loc_attn, 24 | const at::Tensor &grad_output, const int im2col_step, const int K, 25 | const int d_stride, const int block_thread); -------------------------------------------------------------------------------- /ultralytics/nn/DCNv4_op/src/vision.cpp: -------------------------------------------------------------------------------- 1 | /*! 2 | ************************************************************************************************** 3 | * Deformable DETR 4 | * Copyright (c) 2020 SenseTime. All Rights Reserved. 5 | * Licensed under the Apache License, Version 2.0 [see LICENSE for details] 6 | ************************************************************************************************** 7 | * Modified from 8 | *https://github.com/chengdazhi/Deformable-Convolution-V2-PyTorch/tree/pytorch_1.0.0 9 | ************************************************************************************************** 10 | */ 11 | 12 | #include "dcnv4.h" 13 | 14 | PYBIND11_MODULE(TORCH_EXTENSION_NAME, m) { 15 | m.def("flash_deform_attn_forward", &flash_deform_attn_forward, 16 | "flash_deform_attn_forward"); 17 | m.def("flash_deform_attn_backward", &flash_deform_attn_backward, 18 | "flash_deform_attn_backward"); 19 | m.def("dcnv4_forward", &dcnv4_forward, "dcnv4_forward"); 20 | m.def("dcnv4_backward", &dcnv4_backward, "dcnv4_backward"); 21 | } 22 | -------------------------------------------------------------------------------- /ultralytics/nn/__init__.py: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | 3 | from .tasks import (BaseModel, ClassificationModel, DetectionModel, SegmentationModel, attempt_load_one_weight, 4 | attempt_load_weights, guess_model_scale, guess_model_task, parse_model, torch_safe_load, 5 | yaml_model_load) 6 | 7 | __all__ = ('attempt_load_one_weight', 'attempt_load_weights', 'parse_model', 'yaml_model_load', 'guess_model_task', 8 | 'guess_model_scale', 'torch_safe_load', 'DetectionModel', 'SegmentationModel', 'ClassificationModel', 9 | 'BaseModel') 10 | -------------------------------------------------------------------------------- /ultralytics/nn/head/v8.yaml: -------------------------------------------------------------------------------- 1 | v8 2 | -------------------------------------------------------------------------------- /ultralytics/nn/mamba/AUTHORS: -------------------------------------------------------------------------------- 1 | Tri Dao, tri@tridao.me 2 | Albert Gu, agu@andrew.cmu.edu 3 | -------------------------------------------------------------------------------- /ultralytics/nn/mamba/assets/selection.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/scuzyq/improved-v8/839f704f400296d49a18746c660b4089ff41f46b/ultralytics/nn/mamba/assets/selection.png -------------------------------------------------------------------------------- /ultralytics/nn/mamba/assets/ssd_algorithm.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/scuzyq/improved-v8/839f704f400296d49a18746c660b4089ff41f46b/ultralytics/nn/mamba/assets/ssd_algorithm.png -------------------------------------------------------------------------------- /ultralytics/nn/mamba/csrc/selective_scan/selective_scan_bwd_bf16_complex.cu: -------------------------------------------------------------------------------- 1 | /****************************************************************************** 2 | * Copyright (c) 2023, Tri Dao. 3 | ******************************************************************************/ 4 | 5 | // Split into multiple files to compile in paralell 6 | 7 | #include "selective_scan_bwd_kernel.cuh" 8 | 9 | template void selective_scan_bwd_cuda(SSMParamsBwd ¶ms, cudaStream_t stream); -------------------------------------------------------------------------------- /ultralytics/nn/mamba/csrc/selective_scan/selective_scan_bwd_bf16_real.cu: -------------------------------------------------------------------------------- 1 | /****************************************************************************** 2 | * Copyright (c) 2023, Tri Dao. 3 | ******************************************************************************/ 4 | 5 | // Split into multiple files to compile in paralell 6 | 7 | #include "selective_scan_bwd_kernel.cuh" 8 | 9 | template void selective_scan_bwd_cuda(SSMParamsBwd ¶ms, cudaStream_t stream); -------------------------------------------------------------------------------- /ultralytics/nn/mamba/csrc/selective_scan/selective_scan_bwd_fp16_complex.cu: -------------------------------------------------------------------------------- 1 | /****************************************************************************** 2 | * Copyright (c) 2023, Tri Dao. 3 | ******************************************************************************/ 4 | 5 | // Split into multiple files to compile in paralell 6 | 7 | #include "selective_scan_bwd_kernel.cuh" 8 | 9 | template void selective_scan_bwd_cuda(SSMParamsBwd ¶ms, cudaStream_t stream); -------------------------------------------------------------------------------- /ultralytics/nn/mamba/csrc/selective_scan/selective_scan_bwd_fp16_real.cu: -------------------------------------------------------------------------------- 1 | /****************************************************************************** 2 | * Copyright (c) 2023, Tri Dao. 3 | ******************************************************************************/ 4 | 5 | // Split into multiple files to compile in paralell 6 | 7 | #include "selective_scan_bwd_kernel.cuh" 8 | 9 | template void selective_scan_bwd_cuda(SSMParamsBwd ¶ms, cudaStream_t stream); -------------------------------------------------------------------------------- /ultralytics/nn/mamba/csrc/selective_scan/selective_scan_bwd_fp32_complex.cu: -------------------------------------------------------------------------------- 1 | /****************************************************************************** 2 | * Copyright (c) 2023, Tri Dao. 3 | ******************************************************************************/ 4 | 5 | // Split into multiple files to compile in paralell 6 | 7 | #include "selective_scan_bwd_kernel.cuh" 8 | 9 | template void selective_scan_bwd_cuda(SSMParamsBwd ¶ms, cudaStream_t stream); -------------------------------------------------------------------------------- /ultralytics/nn/mamba/csrc/selective_scan/selective_scan_bwd_fp32_real.cu: -------------------------------------------------------------------------------- 1 | /****************************************************************************** 2 | * Copyright (c) 2023, Tri Dao. 3 | ******************************************************************************/ 4 | 5 | // Split into multiple files to compile in paralell 6 | 7 | #include "selective_scan_bwd_kernel.cuh" 8 | 9 | template void selective_scan_bwd_cuda(SSMParamsBwd ¶ms, cudaStream_t stream); -------------------------------------------------------------------------------- /ultralytics/nn/mamba/csrc/selective_scan/selective_scan_fwd_bf16.cu: -------------------------------------------------------------------------------- 1 | /****************************************************************************** 2 | * Copyright (c) 2023, Tri Dao. 3 | ******************************************************************************/ 4 | 5 | // Split into multiple files to compile in paralell 6 | 7 | #include "selective_scan_fwd_kernel.cuh" 8 | 9 | template void selective_scan_fwd_cuda(SSMParamsBase ¶ms, cudaStream_t stream); 10 | template void selective_scan_fwd_cuda(SSMParamsBase ¶ms, cudaStream_t stream); -------------------------------------------------------------------------------- /ultralytics/nn/mamba/csrc/selective_scan/selective_scan_fwd_fp16.cu: -------------------------------------------------------------------------------- 1 | /****************************************************************************** 2 | * Copyright (c) 2023, Tri Dao. 3 | ******************************************************************************/ 4 | 5 | // Split into multiple files to compile in paralell 6 | 7 | #include "selective_scan_fwd_kernel.cuh" 8 | 9 | template void selective_scan_fwd_cuda(SSMParamsBase ¶ms, cudaStream_t stream); 10 | template void selective_scan_fwd_cuda(SSMParamsBase ¶ms, cudaStream_t stream); -------------------------------------------------------------------------------- /ultralytics/nn/mamba/csrc/selective_scan/selective_scan_fwd_fp32.cu: -------------------------------------------------------------------------------- 1 | /****************************************************************************** 2 | * Copyright (c) 2023, Tri Dao. 3 | ******************************************************************************/ 4 | 5 | // Split into multiple files to compile in paralell 6 | 7 | #include "selective_scan_fwd_kernel.cuh" 8 | 9 | template void selective_scan_fwd_cuda(SSMParamsBase ¶ms, cudaStream_t stream); 10 | template void selective_scan_fwd_cuda(SSMParamsBase ¶ms, cudaStream_t stream); -------------------------------------------------------------------------------- /ultralytics/nn/mamba/csrc/selective_scan/static_switch.h: -------------------------------------------------------------------------------- 1 | // Inspired by https://github.com/NVIDIA/DALI/blob/main/include/dali/core/static_switch.h 2 | // and https://github.com/pytorch/pytorch/blob/master/aten/src/ATen/Dispatch.h 3 | 4 | #pragma once 5 | 6 | /// @param COND - a boolean expression to switch by 7 | /// @param CONST_NAME - a name given for the constexpr bool variable. 8 | /// @param ... - code to execute for true and false 9 | /// 10 | /// Usage: 11 | /// ``` 12 | /// BOOL_SWITCH(flag, BoolConst, [&] { 13 | /// some_function(...); 14 | /// }); 15 | /// ``` 16 | #define BOOL_SWITCH(COND, CONST_NAME, ...) \ 17 | [&] { \ 18 | if (COND) { \ 19 | constexpr bool CONST_NAME = true; \ 20 | return __VA_ARGS__(); \ 21 | } else { \ 22 | constexpr bool CONST_NAME = false; \ 23 | return __VA_ARGS__(); \ 24 | } \ 25 | }() 26 | -------------------------------------------------------------------------------- /ultralytics/nn/mamba/evals/lm_harness_eval.py: -------------------------------------------------------------------------------- 1 | import torch 2 | 3 | import transformers 4 | from transformers import AutoTokenizer 5 | 6 | from mamba_ssm.models.mixer_seq_simple import MambaLMHeadModel 7 | 8 | from lm_eval.api.model import LM 9 | from lm_eval.models.huggingface import HFLM 10 | from lm_eval.api.registry import register_model 11 | from lm_eval.__main__ import cli_evaluate 12 | 13 | 14 | @register_model("mamba") 15 | class MambaEvalWrapper(HFLM): 16 | 17 | AUTO_MODEL_CLASS = transformers.AutoModelForCausalLM 18 | 19 | def __init__(self, pretrained="state-spaces/mamba-2.8b", max_length=2048, batch_size=None, device="cuda", 20 | dtype=torch.float16): 21 | LM.__init__(self) 22 | self._model = MambaLMHeadModel.from_pretrained(pretrained, device=device, dtype=dtype) 23 | self.tokenizer = AutoTokenizer.from_pretrained("EleutherAI/gpt-neox-20b") 24 | self.tokenizer.pad_token_id = self.tokenizer.eos_token_id 25 | self.vocab_size = self.tokenizer.vocab_size 26 | self._batch_size = int(batch_size) if batch_size is not None else 64 27 | self._max_length = max_length 28 | self._device = torch.device(device) 29 | 30 | @property 31 | def batch_size(self): 32 | return self._batch_size 33 | 34 | def _model_generate(self, context, max_length, stop, **generation_kwargs): 35 | raise NotImplementedError() 36 | 37 | 38 | if __name__ == "__main__": 39 | cli_evaluate() 40 | -------------------------------------------------------------------------------- /ultralytics/nn/mamba/make.sh: -------------------------------------------------------------------------------- 1 | python setup.py install 2 | 3 | # python setup.py clean --all -------------------------------------------------------------------------------- /ultralytics/nn/mamba/mamba_ssm/__init__.py: -------------------------------------------------------------------------------- 1 | __version__ = "2.0.3" 2 | 3 | from mamba_ssm.ops.selective_scan_interface import selective_scan_fn, mamba_inner_fn 4 | from mamba_ssm.modules.mamba_simple import Mamba 5 | from mamba_ssm.modules.mamba2 import Mamba2 6 | from mamba_ssm.models.mixer_seq_simple import MambaLMHeadModel 7 | -------------------------------------------------------------------------------- /ultralytics/nn/mamba/mamba_ssm/distributed/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/scuzyq/improved-v8/839f704f400296d49a18746c660b4089ff41f46b/ultralytics/nn/mamba/mamba_ssm/distributed/__init__.py -------------------------------------------------------------------------------- /ultralytics/nn/mamba/mamba_ssm/models/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/scuzyq/improved-v8/839f704f400296d49a18746c660b4089ff41f46b/ultralytics/nn/mamba/mamba_ssm/models/__init__.py -------------------------------------------------------------------------------- /ultralytics/nn/mamba/mamba_ssm/models/config_mamba.py: -------------------------------------------------------------------------------- 1 | from dataclasses import dataclass, field 2 | 3 | 4 | @dataclass 5 | class MambaConfig: 6 | 7 | d_model: int = 2560 8 | d_intermediate: int = 0 9 | n_layer: int = 64 10 | vocab_size: int = 50277 11 | ssm_cfg: dict = field(default_factory=dict) 12 | attn_layer_idx: list = field(default_factory=list) 13 | attn_cfg: dict = field(default_factory=dict) 14 | rms_norm: bool = True 15 | residual_in_fp32: bool = True 16 | fused_add_norm: bool = True 17 | pad_vocab_size_multiple: int = 8 18 | tie_embeddings: bool = True 19 | -------------------------------------------------------------------------------- /ultralytics/nn/mamba/mamba_ssm/modules/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/scuzyq/improved-v8/839f704f400296d49a18746c660b4089ff41f46b/ultralytics/nn/mamba/mamba_ssm/modules/__init__.py -------------------------------------------------------------------------------- /ultralytics/nn/mamba/mamba_ssm/modules/mlp.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) 2024, Tri Dao, Albert Gu. 2 | from torch import nn 3 | from torch.nn import functional as F 4 | 5 | 6 | class GatedMLP(nn.Module): 7 | def __init__( 8 | self, 9 | in_features, 10 | hidden_features=None, 11 | out_features=None, 12 | activation=F.silu, 13 | bias=False, 14 | multiple_of=128, 15 | device=None, 16 | dtype=None, 17 | ): 18 | factory_kwargs = {"device": device, "dtype": dtype} 19 | super().__init__() 20 | out_features = out_features if out_features is not None else in_features 21 | hidden_features = ( 22 | hidden_features if hidden_features is not None else int(8 * in_features / 3) 23 | ) 24 | hidden_features = (hidden_features + multiple_of - 1) // multiple_of * multiple_of 25 | self.fc1 = nn.Linear(in_features, 2 * hidden_features, bias=bias, **factory_kwargs) 26 | self.activation = activation 27 | self.fc2 = nn.Linear(hidden_features, out_features, bias=bias, **factory_kwargs) 28 | 29 | def forward(self, x): 30 | y = self.fc1(x) 31 | y, gate = y.chunk(2, dim=-1) 32 | y = y * self.activation(gate) 33 | y = self.fc2(y) 34 | return y 35 | -------------------------------------------------------------------------------- /ultralytics/nn/mamba/mamba_ssm/ops/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/scuzyq/improved-v8/839f704f400296d49a18746c660b4089ff41f46b/ultralytics/nn/mamba/mamba_ssm/ops/__init__.py -------------------------------------------------------------------------------- /ultralytics/nn/mamba/mamba_ssm/ops/triton/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/scuzyq/improved-v8/839f704f400296d49a18746c660b4089ff41f46b/ultralytics/nn/mamba/mamba_ssm/ops/triton/__init__.py -------------------------------------------------------------------------------- /ultralytics/nn/mamba/mamba_ssm/utils/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/scuzyq/improved-v8/839f704f400296d49a18746c660b4089ff41f46b/ultralytics/nn/mamba/mamba_ssm/utils/__init__.py -------------------------------------------------------------------------------- /ultralytics/nn/mamba/mamba_ssm/utils/hf.py: -------------------------------------------------------------------------------- 1 | import json 2 | 3 | import torch 4 | 5 | from transformers.utils import WEIGHTS_NAME, CONFIG_NAME 6 | from transformers.utils.hub import cached_file 7 | 8 | 9 | def load_config_hf(model_name): 10 | resolved_archive_file = cached_file(model_name, CONFIG_NAME, _raise_exceptions_for_missing_entries=False) 11 | return json.load(open(resolved_archive_file)) 12 | 13 | 14 | def load_state_dict_hf(model_name, device=None, dtype=None): 15 | # If not fp32, then we don't want to load directly to the GPU 16 | mapped_device = "cpu" if dtype not in [torch.float32, None] else device 17 | resolved_archive_file = cached_file(model_name, WEIGHTS_NAME, _raise_exceptions_for_missing_entries=False) 18 | return torch.load(resolved_archive_file, map_location=mapped_device) 19 | # Convert dtype before moving to GPU to save memory 20 | if dtype is not None: 21 | state_dict = {k: v.to(dtype=dtype) for k, v in state_dict.items()} 22 | state_dict = {k: v.to(device=device) for k, v in state_dict.items()} 23 | return state_dict 24 | -------------------------------------------------------------------------------- /ultralytics/nn/ops_dcnv3/dist/DCNv3-1.1-cp38-cp38-linux_x86_64.whl: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/scuzyq/improved-v8/839f704f400296d49a18746c660b4089ff41f46b/ultralytics/nn/ops_dcnv3/dist/DCNv3-1.1-cp38-cp38-linux_x86_64.whl -------------------------------------------------------------------------------- /ultralytics/solutions/__init__.py: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | -------------------------------------------------------------------------------- /ultralytics/trackers/__init__.py: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | 3 | from .bot_sort import BOTSORT 4 | from .byte_tracker import BYTETracker 5 | from .track import register_tracker 6 | 7 | __all__ = 'register_tracker', 'BOTSORT', 'BYTETracker' # allow simpler import 8 | -------------------------------------------------------------------------------- /ultralytics/trackers/utils/__init__.py: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | -------------------------------------------------------------------------------- /ultralytics/utils/callbacks/__init__.py: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | 3 | from .base import add_integration_callbacks, default_callbacks, get_default_callbacks 4 | 5 | __all__ = 'add_integration_callbacks', 'default_callbacks', 'get_default_callbacks' 6 | -------------------------------------------------------------------------------- /ultralytics/utils/callbacks/raytune.py: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | 3 | from ultralytics.utils import SETTINGS 4 | 5 | try: 6 | assert SETTINGS['raytune'] is True # verify integration is enabled 7 | import ray 8 | from ray import tune 9 | from ray.air import session 10 | 11 | except (ImportError, AssertionError): 12 | tune = None 13 | 14 | 15 | def on_fit_epoch_end(trainer): 16 | """Sends training metrics to Ray Tune at end of each epoch.""" 17 | if ray.tune.is_session_enabled(): 18 | metrics = trainer.metrics 19 | metrics['epoch'] = trainer.epoch 20 | session.report(metrics) 21 | 22 | 23 | callbacks = { 24 | 'on_fit_epoch_end': on_fit_epoch_end, } if tune else {} 25 | -------------------------------------------------------------------------------- /ultralytics/utils/errors.py: -------------------------------------------------------------------------------- 1 | # Ultralytics YOLO 🚀, AGPL-3.0 license 2 | 3 | from ultralytics.utils import emojis 4 | 5 | 6 | class HUBModelError(Exception): 7 | """ 8 | Custom exception class for handling errors related to model fetching in Ultralytics YOLO. 9 | 10 | This exception is raised when a requested model is not found or cannot be retrieved. 11 | The message is also processed to include emojis for better user experience. 12 | 13 | Attributes: 14 | message (str): The error message displayed when the exception is raised. 15 | 16 | Note: 17 | The message is automatically processed through the 'emojis' function from the 'ultralytics.utils' package. 18 | """ 19 | 20 | def __init__(self, message='Model not found. Please check model URL and try again.'): 21 | """Create an exception for when a model is not found.""" 22 | super().__init__(emojis(message)) 23 | --------------------------------------------------------------------------------