fix(vllm): support native BF16 FlashInfer TRTLLM refit - #3545
Open
seonjinn wants to merge 13 commits into
Open
Conversation
Signed-off-by: seonjinn <sna@nvidia.com> (cherry picked from commit 707968a)
Signed-off-by: seonjinn <sna@nvidia.com> (cherry picked from commit cbd4138)
Signed-off-by: seonjinn <sna@nvidia.com> (cherry picked from commit bda8858)
Signed-off-by: seonjinn <sna@nvidia.com> (cherry picked from commit 2aa1570)
Signed-off-by: seonjinn <sna@nvidia.com> (cherry picked from commit 4aece4b)
Signed-off-by: seonjinn <sna@nvidia.com> (cherry picked from commit 1292dbe)
Signed-off-by: seonjinn <sna@nvidia.com> (cherry picked from commit 56e27a8)
Signed-off-by: seonjinn <sna@nvidia.com> (cherry picked from commit 0c33c30)
…e-refit-pr-20260807 Signed-off-by: seonjinn <sna@nvidia.com>
Signed-off-by: seonjinn <sna@nvidia.com>
…e-refit-pr-20260807 Signed-off-by: seonjinn <sna@nvidia.com>
Signed-off-by: seonjinn <sna@nvidia.com>
seonjinn
added
the
CI:Lfast
Contributor
Author
|
/ok to test a73421e |
Signed-off-by: seonjinn <sna@nvidia.com>
Contributor
Author
|
/ok to test 7180bff |
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Goal
Support repeated unquantized BF16 FlashInfer TRTLLM MoE refits with vLLM's native layerwise reload lifecycle, including runtime-selected
autobackends and static draft-model reloads.Existing FP8, ModelOpt-quantized, and non-TRTLLM paths remain unchanged. Performance recipe updates will follow separately. This is independent of #3478, which targets MXFP8 refit performance.
Validation
5950903).CI:Lfast: https://github.com/NVIDIA-NeMo/RL/actions/runs/31234823907