Sensho Nobe

I think with the nested expression we cannot access the child parameters with dot because the parent turns to a dictionary. Any workaround for this? I don't want to use...

Same. I was working with `torchtext==0.15.1` but if I downgrade it to `0.14.0` it worked fine. Additionally in `0.15.1` labels are just 1 and 2, not 'neg' and 'pos'. In...

@Nayef211 Thanks, I'm creating a lecture material and working on VSCode Jupyter. I ran the same code above on Colab and it turned out that the dataset comes correctly with...

This is a desired feature for creating human-playing dataset. Not ideal, but I made an interactive data collection script using Tkinter. You can save observations, actions and rewards of your...

Same here. CNN PPO stucks at -2750 reward for Speleo. Probably it's just a baseline and supposed to fail

Thanks for the reply. By sapien viewer you mean the interactive one by setting `render_mode="human"`, correct? But what about when I want to just simply log evaluation videos during training?...

@bingykang If you look at the example, [the loss you mentioned is ignored](https://github.com/minyoungg/vqtorch/blob/main/examples/experimental_inplace_update.py#L64-L65) actually, and that corresponds to the case when beta=1.0. The encoder is updated with task-related loss.

@minyoungg Hi, thanks for the interesting work and the great library. I've encountered the same problem as @DiffDynamo and setting `retain_graph=True` [here](https://github.com/minyoungg/vqtorch/blob/main/vqtorch/nn/vq.py#L122) seems working fine. Is this probably because the...