Open
Conversation
- support for running ksim with minimal cuda/jax specific installations
|
|
pagakarthik
changed the title
Contributor
|
This is super cool! Thanks for making this :) |
- add some additional note to the nvidia.driver.resources section in docker compose allowing users to specify either use the defualt mode or specify a more curated configuration - this is particularly suitable for working with egpu(s) and/or a cloud instance with multiple gpu-devices - TODO: add information about installing dependencies - nvidia container toolkit and the (gpu) driver - TODO: add information about accessing tensorboard - TODO: suggest that it might be beneficial to create a ksim-dev repo with only the docker files and may be include ksim and/or ksim-gym to develop within the container [as a starter kit] and avoid the - jax/cuda/ dependencies - installation/ issues
pagakarthik
changed the title
Contributor
|
Is this ready to test? |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
I ran into issues seting up the development environment with jax and other dependencies, and built a docker based environment that supports ksm.
I have been testing with the tagged version 0.1.9. The recent QT additions did not play out well :( with this setup. (yet to figure out a work around). But in general this container would help anyone get started quickly with the ksim or ksim-gym.Resolved by installing the qt and it's dependencies, followed by the change to multi-processing context specification by @alik-git and more information 431Notable features
- support for running ksim [Or ksim-gym] in nvidia's official jax container with minimal cuda/jax specific installations
- access to X11 and other gui elements (OpenGL) for the visualization features
Nvidia's JAX image currently support CUDA 12 with latest jax.
I tried several combinations to find a way to make ksim accessible with CUDA11 and jax [Specificaly for GPU with lower compute capability] but kept running into dependencies issues and backward compatibility issues with JAX. so sad!
TODO:
need to figure out a work around for the QT changes (I haven't used QT before and thus chose to put this in the backlog)run_mode=viewI found ksim to be really interesting. Thanks for making it accessible!