├── README.md
├── __pycache__
└── engine_pretrain.cpython-310.pyc
├── assets
├── method.png
└── teaser.png
├── configs
├── config
│ ├── clip_base.yml
│ ├── clip_base_eval.yml
│ ├── clip_large.yml
│ └── clip_large_eval.yml
├── default_clip_base.yml
└── default_clip_large.yml
├── dataset
├── __pycache__
│ ├── data_utils.cpython-310.pyc
│ ├── egodataset.cpython-310.pyc
│ └── ek100dataset.cpython-310.pyc
├── data_utils.py
├── egodataset.py
└── ek100dataset.py
├── engine_pretrain.py
├── environment.yml
├── evaluation
├── eval_egomcq.py
├── eval_egtea.py
├── eval_ekcls.py
└── eval_mir.py
├── exps
├── eval_egomcq.sh
├── eval_egtea.sh
├── eval_ekcls.sh
├── eval_mir.sh
├── pretrain.sh
└── pretrain_large.sh
├── main_pretrain.py
├── model
├── __pycache__
│ ├── clip.cpython-310.pyc
│ ├── loss.cpython-310.pyc
│ ├── timesformer.cpython-310.pyc
│ └── transformer.cpython-310.pyc
├── clip.py
├── loss.py
├── timesformer.py
└── transformer.py
├── output_dir
├── events.out.tfevents.1740572442.SH-IDC1-10-140-37-2.156563.0
├── events.out.tfevents.1740573180.SH-IDC1-10-140-37-41.259551.0
├── events.out.tfevents.1740573310.SH-IDC1-10-140-37-41.261269.0
└── events.out.tfevents.1740573743.SH-IDC1-10-140-37-41.119535.0
├── requirements.txt
└── util
├── __pycache__
├── config.cpython-310.pyc
├── dist_utils.cpython-310.pyc
├── lr_sched.cpython-310.pyc
├── meter.cpython-310.pyc
├── misc.cpython-310.pyc
└── pos_embed.cpython-310.pyc
├── config.py
├── crop.py
├── datasets.py
├── dist_utils.py
├── lars.py
├── lr_decay.py
├── lr_sched.py
├── meter.py
├── misc.py
└── pos_embed.py
/README.md:
--------------------------------------------------------------------------------
1 | # EgoHOD
2 |
3 | This repo is the official implementation of EgoHOD at ICLR 2025
4 |
5 | > **["Modeling Fine-Grained Hand-Object Dynamics for Egocentric Video Representation Learning"](https://openreview.net/forum?id=P6G1Z6jkf3)**
6 | > [Baoqi Pei](https://scholar.google.com/citations?user=sTCkd54AAAAJ), [Yifei Huang](https://scholar.google.com/citations?user=RU8gNcgAAAAJ), [Jilan Xu](https://scholar.google.com/citations?user=mf2U64IAAAAJ), [Guo Chen](https://scholar.google.com/citations?user=lRj3moAAAAAJ), Yuping He, [Lijin Yang](https://scholar.google.com/citations?user=ppR-rpkAAAAJ),
7 | > [Yali Wang](https://scholar.google.com/citations?user=hD948dkAAAAJ), [Weidi Xie](https://scholar.google.com/citations?user=Vtrqj4gAAAAJ), [Yu Qiao](https://scholar.google.com/citations?user=gFtI-8QAAAAJ), [Fei Wu](https://scholar.google.com/citations?user=XJLn4MYAAAAJ), [Limin Wang](https://scholar.google.com/citations?user=HEuN8PcAAAAJ)
8 |
9 |
10 |
11 |
29 |
30 |