Steven Morad
Steven Morad
### Description I'm able to leak memory on CPU by running the following script ```python from jax import random import psutil key = random.PRNGKey(0) key, netkey = random.split(key) iters =...
Equinox's `filter_*` functions are very helpful functions. It would be great if there were`filter_jacfwd` and `filter_jacrev` functions. For my specific use case, I am using `eqx.filter_value_and_grad(fn)(model, ...)`. As far as...
Are there any tools you would suggest for interpreting the dynamics of deep/complex equinox modules? Perhaps something that tracks the mean and standard deviation of the weights, or the norm...
I'm rerunning velocity baselines in the POMDP directory and I'm observing exploding Q values fairly often. I was wondering if this is something you experienced during training. TD3 seems to...
**Is your feature request related to a problem? Please describe.** It seems that the implemented replay buffers only operate over transitions, with no ability to operate over entire sequences. This...
Are there any plans to port this code to ROS2?
## Describe the bug Calling `zero()` on the `BoundedTensorSpec` will return a tensor that is out of bounds. Would it make sense to return the minimum instead? Or raise an...
## Description This implements a reduction module for use with ensembles. ## Motivation and Context This relies on #1359 to make progress on https://github.com/pytorch/rl/issues/1344. - [x] I have raised an...
Talked a user through setting up atc. Turns out that pip will silently fail to install the atcd binary to /usr/local/bin if they don't have write permission. Add sudo to...
Currently, all our environments could be classified as [_overcomplete_](https://proceedings.mlr.press/v178/liu22f/liu22f.pdf) POMDPs, where the number of unique latent states is greater than the number of unique observations. We are looking for environment...