NVIDIA / NVIDIA/TensorRT-LLM

[Feature]: AutoDeploy: MLA revisited

Open
#11,109 0 comments 0 reactions 1 assignee View on GitHub

Nobody has claimed this yet.

  • #11110 by @lucaslie — closed without merging
AutoDeploy feature request
Dominant language
Python
Stars
14.7k
Forks
2.8k
Avg merge
2d 23h
Merged PRs (30d)
489

Description

🚀 The feature, motivation and pitch

Revisiting MLA support in AutoDeploy to bring it up in terms of both accuracy, performance, and available backends using Deepseek-R1 as an example.

Attached to this issue is a draft PR outlining some of the major steps for support. Draft PR is not tested for accuracy or performance. The main idea is to outline the architecture for proper MLA support and Deepseek performance optimization.

To run a deepseek on a single GPU use

python build_and_run_ad.py --model "deepseek-ai/DeepSeek-R1" --args.yaml-extra "deepseek_mini.yaml"
Alternatives

No response

Additional context

No response

Before submitting a new issue...
  • Make sure you already searched for relevant issues, and checked the documentation and examples for answers to frequently asked questions.

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.