TypeError: `default_name` type (<type 'list'>) is not a string type. You likely meant to pass this into the `values` kwarg.

[xamxixixo]

Hello I am stuck at step "Processing video", actually every command in this step has it's error but I fixed it thanks to the Issues. But I can't do anything with this following error:

When I run `#3d pose estimation os.chdir('..') !bash hmr/3dpose_estimate.sh

#convert estimated .csv files to bvh !blender --background hmr/csv_to_bvh.blend -noaudio -P hmr/csv_to_bvh.py`

This error occurred.

Traceback (most recent call last): File "hmr/demo.py", line 211, in main(config.img_path, config.json_path) File "hmr/demo.py", line 125, in main model = RunModel(config, sess=sess) File "/content/drive/My Drive/AI_COLLAB/video_to_bvh/hmr/src/RunModel.py", line 62, in init self.build_test_model_ief() File "/content/drive/My Drive/AI_COLLAB/video_to_bvh/hmr/src/RunModel.py", line 82, in build_test_model_ief reuse=False) File "/content/drive/My Drive/AI_COLLAB/video_to_bvh/hmr/src/models.py", line 40, in Encoder_resnet with tf.name_scope("Encoder_resnet", [x]): File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/framework/ops.py", line 6450, in init "pass this into the values kwarg." % type(default_name)) TypeError: default_name type (<type 'list'>) is not a string type. You likely meant to pass this into the values kwarg. Done Read blend: /content/drive/My Drive/AI_COLLAB/video_to_bvh/hmr/csv_to_bvh.blend OSError: Python file "/content/drive/My Drive/AI_COLLAB/video_to_bvh/hmr/csv_to_bvh.py" could not be opened: No such file or directory

Blender quit src/tcmalloc.cc:283] Attempt to free invalid pointer 0x7f1fa500e400

How to fix it? Thank you.

[xamxixixo]

So, I have fixed it by downgrade tensorflow to 1.11

[kazux2]

@xamxixixo How did you downgrade tensorflow on google colab? Did you run pip install --upgrade tensorflow==1.11.0 on colab and everything worked fine?

[xamxixixo]

I did this:

pip uninstall tensorflow
pip install tensorflow==1.11

Colab will tell you (in the pip install process) there will be packages that won't go well with that old tensorflow, you should downgrade them too

[kazux2]

@xamxixixo I see. Thanks! (I didn't downgrade other depended packages and couldn't run the process without error) Apparently, in the newer tensorflow, the function tf.name_scope has different arguments and absl-py started comming with it and has colliding flag name 'log_dir' and these were making errors.

[Duthopi]

See this solution in the original hmr code, it solves the problem without having to change tensorflow: https://github.com/akanazawa/hmr/issues/69#issuecomment-503765527.

Using the TensorFlow documentation for tf.name_scope, I resolved the issue by prepending [x] with values=. That is change with tf.name_scope("Encoder_resnet", [x]): to with tf.name_scope("Encoder_resnet", values=[x]):. This error also occurs in a few places over in batch_lbs.py which I fixed in the same way.

[wine3603]

I got this bug, and tensorflow 1.11 doesn't work for me. @Duthopi could you tell me where is the batch_lbs.py so that I can try it out ?

[Duthopi]

It's under the /content/hmr/src/tf_smpl/ folder