initial_commit

README.md

@@ -0,0 +1,98 @@
+## History Repeats Itself: Human Motion Prediction via Motion Attention
+This is the code for the paper
+
+Wei Mao, Miaomiao Liu, Mathieu Salzmann, Hongdong Li. 
+[_History Repeats Itself: Human Motion Prediction via Motion Attention_](https://arxiv.org/abs/1908.05436). In ECCV 20.
+
+### Dependencies
+
+* cuda 10.0
+* Python 3.6
+* [Pytorch](https://github.com/pytorch/pytorch) >1.0.0 (Tested on 1.1.0 and 1.3.0)
+
+### Get the data
+
+[Human3.6m](http://vision.imar.ro/human3.6m/description.php) in exponential map can be downloaded from [here](http://www.cs.stanford.edu/people/ashesh/h3.6m.zip).
+
+Directory structure: 
+```shell script
+H3.6m
+|-- S1
+|-- S5
+|-- S6
+|-- ...
+`-- S11
+```
+[AMASS](https://amass.is.tue.mpg.de/en) from their official website..
+
+Directory structure:
+```shell script
+amass
+|-- ACCAD
+|-- BioMotionLab_NTroje
+|-- CMU
+|-- ...
+`-- Transitions_mocap
+```
+[3DPW](https://virtualhumans.mpi-inf.mpg.de/3DPW/) from their official website.
+
+Directory structure: 
+```shell script
+3dpw
+|-- imageFiles
+|   |-- courtyard_arguing_00
+|   |-- courtyard_backpack_00
+|   |-- ...
+`-- sequenceFiles
+    |-- test
+    |-- train
+    `-- validation
+```
+Put the all downloaded datasets in ./datasets directory.
+
+### Training
+All the running args are defined in [opt.py](utils/opt.py). We use following commands to train on different datasets and representations.
+To train,
+```bash
+python main_h36m_3d.py --kernel_size 10 --dct_n 20 --input_n 50 --output_n 10 --skip_rate 1 --batch_size 32 --test_batch_size 32 --in_features 66
+```
+```bash
+python main_h36m_ang.py --kernel_size 10 --dct_n 20 --input_n 50 --output_n 10 --skip_rate 1 --batch_size 32 --test_batch_size 32 --in_features 48
+```
+```bash
+python main_amass_3d.py --kernel_size 10 --dct_n 35 --input_n 50 --output_n 25 --skip_rate 5 --batch_size 128 --test_batch_size 128 --in_features 54 
+```
+### Evaluation
+To evaluate the pretrained model,
+```bash
+python main_h36m_3d_eval.py --is_eval --kernel_size 10 --dct_n 20 --input_n 50 --output_n 10 --skip_rate 1 --batch_size 32 --test_batch_size 32 --in_features 66 --ckpt ./checkpoint/pretrained/h36m_3d_in50_out10_dctn20/
+```
+```bash
+python main_h36m_ang_eval.py --is_eval --kernel_size 10 --dct_n 20 --input_n 50 --output_n 10 --skip_rate 1 --batch_size 32 --test_batch_size 32 --in_features 48 --ckpt ./checkpoint/pretrained/h36m_ang_in50_out10_dctn20/
+```
+```bash
+python main_amass_3d_eval.py --is_eval --kernel_size 10 --dct_n 35 --input_n 50 --output_n 25 --skip_rate 5 --batch_size 128 --test_batch_size 128 --in_features 54 --ckpt ./checkpoint/pretrained/amass_3d_in50_out25_dctn30/
+```
+
+### Citing
+
+If you use our code, please cite our work
+
+```
+@inproceedings{wei2020his,
+  title={History Repeats Itself: Human Motion Prediction via Motion Attention},
+  author={Wei, Mao and Miaomiao, Liu and Mathieu, Salzemann},
+  booktitle={ECCV},
+  year={2020}
+}
+```
+
+### Acknowledgments
+The overall code framework (dataloading, training, testing etc.) is adapted from [3d-pose-baseline](https://github.com/una-dinosauria/3d-pose-baseline). 
+
+The predictor model code is adapted from [LTD](https://github.com/wei-mao-2019/LearnTrajDep).
+
+Some of our evaluation code and data process code was adapted/ported from [Residual Sup. RNN](https://github.com/una-dinosauria/human-motion-prediction) by [Julieta](https://github.com/una-dinosauria). 
+
+### Licence
+MIT

checkpoint/pretrained/amass_3d_in50_out25_dctn30/option.json

@@ -0,0 +1,47 @@
+{
+    "data_dir": "/home/wei/Documents/",
+    "mean_params_dir": "/home/wei/Documents/projects/2019-human-mesh-recovery/baselines/single_mesh_from_img_feat/smpl/neutral_smpl_mean_params.h5",
+    "smpl_model": "./smpl/model.npz",
+    "exp": "main_att_model_v9_6_7_motion_amass_para_in50_out25_5_ks10_dmodel256_stg12_lr0.0005_gn10000_dctn30_sr5",
+    "is_eval": false,
+    "ckpt": "checkpoint/main_att_model_v9_6_7_motion_amass_para_in50_out25_5_ks10_dmodel256_stg12_lr0.0005_gn10000_dctn30_sr5",
+    "img_size": 224,
+    "num_points": 1000,
+    "sample_rate": 2,
+    "skip_rate": 5,
+    "skip_rate_test": 5,
+    "resamp_rate": 3,
+    "dataset_used": [
+        "h36"
+    ],
+    "input_size": 2048,
+    "output_size": 85,
+    "hidden_size": 512,
+    "num_stage": 12,
+    "drop_out": 0.5,
+    "input_n": 50,
+    "output_n": 25,
+    "step_n": 5,
+    "dct_n": 30,
+    "lr_now": 0.0005,
+    "max_norm": 10000,
+    "lr_decay": 2,
+    "lr_gamma": 0.96,
+    "weight_decay": 0.9,
+    "weight_3d": 1,
+    "weight_2d": 1,
+    "epoch": 50,
+    "batch_size": 128,
+    "test_batch_size": 128,
+    "gt_dct": 0,
+    "is_load": false,
+    "d_model": 256,
+    "n_head": 8,
+    "num_encoder_layers": 6,
+    "num_decoder_layers": 6,
+    "dim_feedforward": 512,
+    "warmup_steps": 4000,
+    "do_p": 0.1,
+    "act_f": "relu",
+    "kernel_size": 10
+}