Mohammad Taufeeque

@AdamGleave Yes, alpha = b * w_{m,m}. So the denominator should in step 2 should be (1 - b * w_{m, m}). This factor of b would go away in...

Thanks for the analysis Yawen! I also had my doubts regarding setting the default decay rate to 0.99. However, I think it is a reasonable decay rate, and any rate...

For testing the correctness of `EMANorm`, I had earlier implemented Algorithm 1 from [Incremental_batch_EMA_and_EMV.pdf](https://github.com/HumanCompatibleAI/imitation/files/9364938/Incremental_batch_EMA_and_EMV.pdf) in the class `EMANormIncremental`. However, since Algorithm 1 is numerically unstable, the `test_ema_norm_batch_correctness` fails for decay...

The `vec_normalize.pkl` file gets downloaded when we load the trained agent using the rl-zoo library. Imitation only downloads the model zip file. So to replicate the error, first load the...

> How about making `train_adversarial` script do `int(total_timesteps)`? I think that should fix this issue and any others that are downstream of `total_timesteps` being the wrong type. That should also...

Seems like an error with opening the video file. Have you tried changing the path of the video file or using a different video?

The library doesn't do this out of the box. You can fork the project and modify the `algorithms.py` script to save the predictions in a csv file. Line 278 of...

The library is tested using CUDA-based GPUs on Ubuntu. Since we use PyTorch for the ML model, you should be able to run it using Apple's M1/M2 GPUs with MLP...

The expected output is that the video should play as shown in the readme. Can you specify the Python version you are using? I have tested the library with Python3.9....

I have fixed the error you are getting with the webcam. Let me know if you are able to run the program with a webcam now. As for the issue...