search

una-dinosauria / 3d-pose-baseline

A simple baseline for 3d human pose estimation in tensorflow. Presented at ICCV 17.
MIT License
1.41k stars 354 forks source link

error with the demo code (saving result in windows 10) #160

Closed leilaUEA closed 4 years ago

leilaUEA commented 4 years ago

Hi I tried the code in Readme and this repository: https://github.com/ArashHosseini/3d-pose-baseline

I received this error while trying training code in anaconda prompt (h5py 2.10.0, tensorflow 2.1.0 with modification) about saving the result

Is it the same issue as here? https://github.com/una-dinosauria/human-motion-prediction/issues/10

Thank you very much in advance

  1. Your operating system : Window 10 Home
  2. Your tensorflow version: tensorflow 2.1.0 ( changed the import code as advised here https://github.com/una-dinosauria/3d-pose-baseline/issues/149

changed import tensorflow as tf to import tensorflow.compat.v1 as tf tf.disable_v2_behavior() )

  1. Your python version python 3.6.10
  2. The stack trace of the error that you see the error: 2020-03-12 19:01:26.636116: W tensorflow/core/framework/op_kernel.cc:1655] OP_REQUIRES failed at save_restore_v2_ops.cc:109 : Not found: Failed to create a NewWriteableFile: experiments\All\dropout_0.5\epochs_1\lr_0.001\residual\depth_2\linear_size1024\batch_size_64\no_procrustes\maxnorm\batch_normalization\use_stacked_hourglass\predict_17\checkpoint-24371.data-00000-of-00001.tempstate602890883713020701 : The system cannot find the path specified. ; No such process

entered this code in anaconda prompt: python src/predict_3dpose.py --camera_frame --residual --batch_norm --dropout 0.5 --max_norm --evaluateActionWise --use_sh --epochs 1

More information

After getting result:

Global step: 24371 Learning rate: 9.90e-04 Train loss avg: 0.1855

===Action=== ==mm== Directions 60.27 Discussion 66.98 Eating 70.40 Greeting 71.79 Phoning 90.83 Photo 97.25 Posing 63.89 Purchases 67.59 Sitting 90.86 SittingDown 121.15 Smoking 77.19 Waiting 72.79 WalkDog 77.53 Walking 64.22 WalkTogether 66.28 Average 77.27

I receive this error:

2020-03-12 19:01:26.636116: W tensorflow/core/framework/op_kernel.cc:1655] OP_REQUIRES failed at save_restore_v2_ops.cc:109 : Not found: Failed to create a NewWriteableFile: experiments\All\dropout_0.5\epochs_1\lr_0.001\residual\depth_2\linear_size1024\batch_size_64\no_procrustes\maxnorm\batch_normalization\use_stacked_hourglass\predict_17\checkpoint-24371.data-00000-of-00001.tempstate602890883713020701 : The system cannot find the path specified. ; No such process Saving the model... Traceback (most recent call last): File "C:\Users\Leila\Anaconda3\envs\TestEnv3\lib\site-packages\tensorflow_core\python\client\session.py", line 1367, in _do_call return fn(*args) File "C:\Users\Leila\Anaconda3\envs\TestEnv3\lib\site-packages\tensorflow_core\python\client\session.py", line 1352, in _run_fn target_list, run_metadata) File "C:\Users\Leila\Anaconda3\envs\TestEnv3\lib\site-packages\tensorflow_core\python\client\session.py", line 1445, in _call_tf_sessionrun run_metadata) tensorflow.python.framework.errors_impl.NotFoundError: Failed to create a NewWriteableFile: experiments\All\dropout_0.5\epochs_1\lr_0.001\residual\depth_2\linear_size1024\batch_size_64\no_procrustes\maxnorm\batch_normalization\use_stacked_hourglass\predict_17\checkpoint-24371.data-00000-of-00001.tempstate602890883713020701 : The system cannot find the path specified. ; No such process [[{{node save/SaveV2}}]]

leilaUEA commented 4 years ago

shortening train_dir worked for saving the model, it seems that matplotlib have problem with set_aspect in new versions...

File "C:\Users\Leila\3d-pose-baseline\src\viz.py", line 56, in show3Dpose ax.set_aspect('equal') File "C:\Users\Leila\Anaconda3\envs\TestEnv3\lib\site-packages\matplotlib\axes_base.py", line 1281, in set_aspect 'It is not currently possible to manually set the aspect '

downgrading the matplotlib version worked