FileNotFoundError

[jiaqiAA]

Thanks for your great work！

No such file or directory: './data/prcocessed_data/v3_mean.npy' No such file or directory: './data/prcocessed_data/v3_std.npy'

[YoungSeng]

Thanks for the issue, I noticed some overlap in naming the folder data and dataset folders, however I just uploaded the ./data/prcocessed_data/v3_mean and v3_std, please check it out.

[jiaqiAA]

Thanks for your reply. when I run python -m sample.double_take --save_dir '' --guidacnce_param 1 --model_name model001000000 --BEAT_wav_feat ./datasets/BEAT/my_wav_feat/ --HUMANML3D_text_feat ./datasets/SMPLX/HumanML3D/v3_HUMANML3D_txt_feat/ --clip_model_path ./data/clip --vis_mode customized_controls, I meet an error:

ax.lines = []
AttributeError: can't set attribute

and I found sample_audio_feat, sample_text_feat, text, audio_tag = None, None, None, None https://github.com/YoungSeng/FreeTalker/blob/master/sample/double_take.py#L561 I'm a little confused about what this demo is for？

[YoungSeng]

Regarding the first question, it seems that there is an error in the visualization, can you provide more details about the error report? For example the full ERROR message; also for the second question, this is not a problem, as it is using processed speech features so it is all None here and does not affect the program operation. See

https://github.com/YoungSeng/FreeTalker/blob/77a4beb02f74524634ac4dd86471a02b6b7071cc/sample/double_take.py#L278

and

https://github.com/YoungSeng/FreeTalker/blob/77a4beb02f74524634ac4dd86471a02b6b7071cc/sample/double_take.py#L191

[jiaqiAA]

ERROR message

./save/my_v3_0/model001000000
Creating model and diffusion...
n_joints: 659, latent_dim: 256
TRANS_ENC init
model version v3_0
Loading checkpoints from [./save/my_v3_0/model001000000.pt]...
ViT-B/32 ./data/clip
guidacnce_param:  1.0
rec_ric_data shape:  torch.Size([630, 55, 3])
Traceback (most recent call last):
  File "anaconda3/envs/freetalker/lib/python3.8/runpy.py", line 194, in _run_module_as_main
    return _run_code(code, main_globals, None,
  File "anaconda3/envs/freetalker/lib/python3.8/runpy.py", line 87, in _run_code
    exec(code, run_globals)
  File "/FreeTalker/sample/double_take.py", line 579, in <module>
    process(sample_audio_feat, sample_text_feat, segment_length, save_path, model_name, n_seed=n_seed, version=version, text=text)
  File "/FreeTalker/sample/double_take.py", line 472, in process
    plot_3d_motion_1(os.path.join(save_path, "positions_vis1_{}_{}.mp4".format(guidacnce_param, args.vis_mode)), kinematic_chain, np.array(rec_ric_data), captions,
  File "/FreeTalker/process/plot_script.py", line 274, in plot_3d_motion
    explicit_plot_3d_motion(save_path, kinematic_tree, joints, title, dataset, figsize=figsize, fps=fps, radius=radius,
  File "/FreeTalker/process/plot_script.py", line 442, in explicit_plot_3d_motion
    ani.save(save_path, fps=fps, dpi=200)
  File "/anaconda3/envs/freetalker/lib/python3.8/site-packages/matplotlib/animation.py", line 1090, in save
    anim._init_draw()  # Clear the initial frame
  File "/anaconda3/envs/freetalker/lib/python3.8/site-packages/matplotlib/animation.py", line 1748, in _init_draw
    self._draw_frame(frame_data)
  File "/anaconda3/envs/freetalker/lib/python3.8/site-packages/matplotlib/animation.py", line 1767, in _draw_frame
    self._drawn_artists = self._func(framedata, *self._args)
  File "/FreeTalker/process/plot_script.py", line 385, in update
    ax.lines = []
AttributeError: can't set attribute

[YoungSeng]

Well I see, seemingly cause by ani.save(save_path, fps=fps, dpi=200). My matplotlib version is 3.1.3, can you try pip install matplotlib==3.1.3 and try again?

[jiaqiAA]

Thanks very much for your reply, it worked!

[YoungSeng]

Great, I also finished the update in readme.