'DummyVecEnv' object has no attribute 'video_recorder' error during model push to Hugging Face Hub

[Fouad-Smaoui]

While using the package_to_hub function from huggingface_sb3, the process fails to generate a video replay of the agent due to the following error:

✘ 'DummyVecEnv' object has no attribute 'video_recorder'
✘ We are unable to generate a replay of your agent, the package_to_hub process continues
Steps to reproduce:

Train an agent with Stable-Baselines3 using DummyVecEnv

Call package_to_hub()

Observe error during video generation step

Expected behavior:
The function should either generate the video correctly or provide a clearer fallback when DummyVecEnv is used.

Environment:

Google Colab

huggingface_sb3 latest version

stable-baselines3

moviepy

[Dhadheechi]

Is there atleast a way to view the video at the mentioned folder /tmp/... I can't find the specified folder in colab

[makataomu]

Is there atleast a way to view the video at the mentioned folder /tmp/... I can't find the specified folder in colab

you can run this command below. tmp folder is in colab, to see it you need to hit on the botton with eye called "show hidden files". but running the command below will be easier.
shutil.copy("/tmp/videos/a2c-PandaReachDense-v3-step-0-to-step-1000.mp4", "/content/a2c-video.mp4")