search

open-mmlab / mmskeleton

A OpenMMLAB toolbox for human pose estimation, skeleton-based action recognition, and action synthesis.
Apache License 2.0
2.93k stars 1.03k forks source link

Problem when loading pretrained model? #182

Open hufengshuo07 opened 5 years ago

hufengshuo07 commented 5 years ago

Thank you for your brilliant work. Could you help me with this problem when trying demo? I've re-downloaded models from both baiduyun and get_models.sh

sure@sure-ThinkStation-P318:~/文档/sure/pose/activity_recognition/st-gcn-master$ python3 main.py demo --openpose ../../openpose/build/ --video ./resource/media/skateboarding.mp4 Traceback (most recent call last): File "/usr/local/lib/python3.5/dist-packages/torchlight-1.0-py3.5.egg/torchlight/io.py", line 82, in load_weights doc = _io._TextIOBase.doc File "/usr/local/lib/python3.5/dist-packages/torch/nn/modules/module.py", line 721, in load_state_dict self.class.name, "\n\t".join(error_msgs))) RuntimeError: Error(s) in loading state_dict for Model: Unexpected key(s) in state_dict: "data_bn.num_batches_tracked", "st_gcn_networks.0.tcn.0.num_batches_tracked", "st_gcn_networks.0.tcn.3.num_batches_tracked", "st_gcn_networks.1.tcn.0.num_batches_tracked", "st_gcn_networks.1.tcn.3.num_batches_tracked", "st_gcn_networks.2.tcn.0.num_batches_tracked", "st_gcn_networks.2.tcn.3.num_batches_tracked", "st_gcn_networks.3.tcn.0.num_batches_tracked", "st_gcn_networks.3.tcn.3.num_batches_tracked", "st_gcn_networks.4.tcn.0.num_batches_tracked", "st_gcn_networks.4.tcn.3.num_batches_tracked", "st_gcn_networks.4.residual.1.num_batches_tracked", "st_gcn_networks.5.tcn.0.num_batches_tracked", "st_gcn_networks.5.tcn.3.num_batches_tracked", "st_gcn_networks.6.tcn.0.num_batches_tracked", "st_gcn_networks.6.tcn.3.num_batches_tracked", "st_gcn_networks.7.tcn.0.num_batches_tracked", "st_gcn_networks.7.tcn.3.num_batches_tracked", "st_gcn_networks.7.residual.1.num_batches_tracked", "st_gcn_networks.8.tcn.0.num_batches_tracked", "st_gcn_networks.8.tcn.3.num_batches_tracked", "st_gcn_networks.9.tcn.0.num_batches_tracked", "st_gcn_networks.9.tcn.3.num_batches_tracked".

arthur-k-kochi commented 5 years ago

Looks similar to my case and the issue #49 . Try running sudo apt-get install ffmpeg

PrettyZZW commented 5 years ago

i have same error but install ffmpeg can't solve this problem [06.19.19|02:24:21] Load weights [edge_importance.7]. [06.19.19|02:24:21] Load weights [edge_importance.8]. [06.19.19|02:24:21] Load weights [edge_importance.9]. [06.19.19|02:24:21] Load weights [fcn.weight]. [06.19.19|02:24:21] Load weights [fcn.bias]. Traceback (most recent call last): File "/home/user/miniconda/envs/py36/lib/python3.6/site-packages/torchlight-1.0-py3.6.egg/torchlight/io.py", line 82, in load_weights doc = _io._TextIOBase.doc File "/home/user/miniconda/envs/py36/lib/python3.6/site-packages/torch/nn/modules/module.py", line 721, in load_state_dict self.class.name, "\n\t".join(error_msgs))) RuntimeError: Error(s) in loading state_dict for Model: Unexpected key(s) in state_dict: "data_bn.num_batches_tracked", "st_gcn_networks.0.tcn.0.num_batches_tracked", "st_gcn_networks.0.tcn.3.num_batches_tracked", "st_gcn_networks.1.tcn.0.num_batches_tracked", "st_gcn_networks.1.tcn.3.num_batches_tracked", "st_gcn_networks.2.tcn.0.num_batches_tracked", "st_gcn_networks.2.tcn.3.num_batches_tracked", "st_gcn_networks.3.tcn.0.num_batches_tracked", "st_gcn_networks.3.tcn.3.num_batches_tracked", "st_gcn_networks.4.tcn.0.num_batches_tracked", "st_gcn_networks.4.tcn.3.num_batches_tracked", "st_gcn_networks.4.residual.1.num_batches_tracked", "st_gcn_networks.5.tcn.0.num_batches_tracked", "st_gcn_networks.5.tcn.3.num_batches_tracked", "st_gcn_networks.6.tcn.0.num_batches_tracked", "st_gcn_networks.6.tcn.3.num_batches_tracked", "st_gcn_networks.7.tcn.0.num_batches_tracked", "st_gcn_networks.7.tcn.3.num_batches_tracked", "st_gcn_networks.7.residual.1.num_batches_tracked", "st_gcn_networks.8.tcn.0.num_batches_tracked", "st_gcn_networks.8.tcn.3.num_batches_tracked", "st_gcn_networks.9.tcn.0.num_batches_tracked", "st_gcn_networks.9.tcn.3.num_batches_tracked".

punknownq commented 5 years ago

I have the same problem.

alike123 commented 5 years ago

+1. Have you solved this problem?

daylightsaving2017 commented 5 years ago

I have solved this problem and verified the reason by some experiments. This problem is due to the version compatibilty of the PyTorch. If you use a newer version to train a model, the old version pytorch often cannot load it.

daylightsaving2017 commented 5 years ago

@hufengshuo07 @arthur-k-kochi @PrettyZZW @punknownq @alike123 You can download a fixed model from the repo provided by one developer in a similar issue. links here: https://pan.baidu.com/s/1JrfpRqt0uF8AydiJuxANZw password:j4rt Duplicate of #172 Thanks for the work by @XJX777