Repository navigation
Conversation
Contributor
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
Contributor
There was a problem hiding this comment.
Code Review
This pull request updates the configuration and documentation for the Xiaomi MiMo V2.6 Flash and Pro RL models. It introduces the encoder_parallel feature and configures the --attention-backend ROCM_AITER_UNIFIED_ATTN and --block-size 16 options for AMD ROCm environments to improve performance. The review feedback highlights that the AMD launch commands in the MiMo-V2.6-Flash-RL.yaml guide are missing the --gpu-memory-utilization 0.95 flag, which should be added to ensure consistency with the actual recipe configurations.
vllmellm
force-pushed
the
mimo-v26-unified-attn-encoder-dp
branch
2 times, most recently
from
November 3, 2026 14:09
49eb05a to
20b99eb
Compare
Signed-off-by: vllmellm <vllm.ellm@embeddedllm.com>
vllmellm
force-pushed
the
mimo-v26-unified-attn-encoder-dp
branch
from
November 3, 2026 14:11
20b99eb to
5a8149a
Compare
hongxiayang
approved these changes
Nov 5, 2026
Contributor
|
please do it in vllm-project/vllm repo |
This branch was successfully deployed
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
--attention-backend ROCM_AITER_UNIFIED_ATTN --block-size 16to the AMD blocks and guide commands ofMiMo-V2.6-Flash-RL(mi300x/mi325x/mi355x, both variants) andMiMo-V2.6-Pro-RL(mi355x); faster than the DiffKV default per vllm-project/vllm#64064.encoder_parallelopt-in feature (--mm-encoder-tp-mode data) to both, per vllm-project/vllm#64179.node scripts/validate-source.mjs --source vllm --dir <vllm>/recipespasses.