fix: bump flashinfer to 0.6.9 to match the V4-Flash MXFP4 minimum#65
Open
gvr13n wants to merge 1 commit into
Open
fix: bump flashinfer to 0.6.9 to match the V4-Flash MXFP4 minimum#65gvr13n wants to merge 1 commit into
gvr13n wants to merge 1 commit into
Conversation
mxfp4_deepseek.py requires flashinfer >= 0.6.9 (_MIN_FI_VERSION: mxfp8_quantize, trtllm_fp4_block_scale_routed_moe), but pyproject pins flashinfer_python/cubin ==0.6.3, so a fresh install cannot serve DeepSeek-V4-Flash MXFP4: the quant plugin fails its import-time version check with a warning, the MXFP4 MoE method never registers, and the server later crashes with an unrelated-looking 'AssertionError: Hidden size mismatch' in the generic triton fused_experts during cuda-graph capture. Bump the pins and the Dockerfile FLASHINFER_VERSION arg together, keeping the documented jit-cache alignment.
There was a problem hiding this comment.
Code Review
This pull request updates the FLASHINFER_VERSION from 0.6.3 to 0.6.9 in the Dockerfile, and correspondingly bumps the flashinfer_python and flashinfer_cubin dependency versions to 0.6.9 in pyproject.toml. There are no review comments, and I have no feedback to provide.
Important
The consumer version of Gemini Code Assist on GitHub is being sunset. Starting June 18, 2026, new organization installations will be blocked, and all code review activity will officially cease on July 17, 2026.
For more details on the timeline and next steps, please review the Help Documentation.
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Fresh installs cannot serve DeepSeek-V4-Flash MXFP4:
mxfp4_deepseek.pyrequires flashinfer >= 0.6.9 (_MIN_FI_VERSION—mxfp8_quantize,trtllm_fp4_block_scale_routed_moe), butpyproject.tomlpinsflashinfer_python/flashinfer_cubin==0.6.3.The failure mode is nasty to diagnose: the quant plugin fails its import-time version check with only a warning, the MXFP4 MoE method never registers, and the server later dies with an unrelated-looking
AssertionError: Hidden size mismatchin the generic tritonfused_expertsduring cuda-graph capture. (Hit this twice on a working V4-Flash + kt-hybrid box after a routine reinstall silently downgraded flashinfer per the pin.)This bumps the pins and the Dockerfile
FLASHINFER_VERSIONarg together, preserving the documented jit-cache alignment.Possibly worth a follow-up: promote the import-time version warning to a hard error when a V4 MXFP4 checkpoint is actually being served, so the failure surfaces at the true cause.