Abhinav Gupta

Termination prob calculated over current state instead of the next state

2

The termination probability is calculated over the next state according to the [original paper](https://arxiv.org/pdf/1609.05140.pdf). So it should be using next_obs instead of obs. https://github.com/lweitkamp/option-critic-pytorch/blob/0c57da7686f8903ed2d8dded3fae832ee9defd1a/option_critic.py#L238

model specifications not coherent with the MLB paper

3

The model configuration is not the same as described in the paper. There is a softmax layer missing at the end of the model. The paper concatenates the attention *...

Information regarding training time

6

Can you please provide some information regarding the training time (with the hardware specifications) and the size of the dataset used? Thanks!

gpu info not showing

3

The status bar shows `No GPU` even though the GPU can be accessed by `nvidia-smi` which is available at `/usr/bin` included in `$PATH`. I am using tmux v2.6

Currently, if I open a tmux session on my local machine and ssh to a remote server, the status is not changed with respect to the availability of internet on...

Sending Stop signal from the project page

2

Can you please add a stop button next to the show/hide/archive... buttons which stops the selected experiment(s)?

Option to save changes made in parallel coordinates graph

1

Be able to switch axis and drag params and save

The `workers_cpu` argument requires the user to know the cpu affinity of the process before it is started. It is possible that the manually assigned cpu is not in the...

eta parameters are trainable

1

All etas are trainable parameters. You cannot set requires_grad=False for them.

logits_py incorrect in gumbel_softmax_vae_v2

1

`logits_py` should be the log of the current `logits_py` or we can just give it as `probs` and not `logits` to the corresponding distribution.