Adding a Docker image containing every single setup steps #21
Conversation
This file contains 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
Adding the docker image for fast deploying on GPU rental services such as vast.ai
|
@ltngonnguyen looks pretty cool. Can you share your dockerfile you used? Thank you, |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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 have created a Docker image to streamline the setup process for LLaMA-X model training on GPU rental services like vast.ai. Currently, setting up the required dependencies such as CUDA and PyTorch is a time-consuming and repetitive chore, hindering the efficiency of researchers and developers. With this Docker image, we can eliminate the need to repeat these steps every single time, making the setup process quick and hassle-free.
This Docker image encapsulates the necessary software stack, including CUDA, PyTorch, and other dependencies, allowing users to spin up a ready-to-use environment for LLaMA-X model training in minutes.
The image itself is based on Nvidia's official CUDA 11.3 docker image, with conda installing pytorch and all dependencies. I've tested it on a couple different vast.ai GPU instances and all worked.