Skip to content

[None][fix] Change compressor linear dtype - #13605

Merged
lfr-0531 merged 3 commits into
NVIDIA:feat/deepseek_v4from
mingyangHao:mingyangh/deepseek-v4-compressor-fusion-fix
Apr 30, 2026
Merged

[None][fix] Change compressor linear dtype#13605
lfr-0531 merged 3 commits into
NVIDIA:feat/deepseek_v4from
mingyangHao:mingyangh/deepseek-v4-compressor-fusion-fix

Conversation

@mingyangHao

@mingyangHao mingyangHao commented Apr 29, 2026 •
edited by lfr-0531
Loading

Copy link
Copy Markdown
Collaborator

@coderabbitai summary

Description

Test Coverage

PR Checklist

Please review the following before submitting your PR:

  • PR description clearly explains what and why. If using CodeRabbit's summary, please make sure it makes sense.

  • PR Follows TRT-LLM CODING GUIDELINES to the best of your knowledge.

  • Test cases are provided for new code paths (see test instructions)

  • Any new dependencies have been scanned for license and vulnerabilities

  • CODEOWNERS updated if ownership changes

  • Documentation updated as needed

  • Update tava architecture diagram if there is a significant design change in PR.

  • The reviewers assigned automatically/manually are appropriate for the PR.

  • Please check this after reviewing the above items as appropriate for this PR.

GitHub Bot Help

To see a list of available CI bot commands, please comment /bot help.

@mingyangHao
test: Guard DeepSeek-V4 compressor fused outputs
f68a531 
Signed-off-by: Mingyang Hao <mingyangHao@users.noreply.github.com>
@mingyangHao
perf: Avoid duplicate DeepSeek-V4 compressor cache writes
cd74250 
Signed-off-by: Mingyang Hao <mingyangHao@users.noreply.github.com>
@mingyangHao
mingyangHao requested a review from a team as a code owner April 29, 2026 07:47
@mingyangHao
mingyangHao requested review from QiJune and removed request for a team April 29, 2026 07:47
@mingyangHao
mingyangHao requested a review from a team as a code owner April 29, 2026 07:47
@mingyangHao
mingyangHao requested review from yechank-nvidia and removed request for a team April 29, 2026 07:47
@mingyangHao
mingyangHao force-pushed the mingyangh/deepseek-v4-compressor-fusion-fix branch from 948e792 to cd74250 Compare April 29, 2026 08:13
@mingyangHao mingyangHao changed the title [None][fix]indexer cache fix [None][fix] indexer cache fix Apr 30, 2026
@mingyangHao
fix: update DeepSeek-V4 compressor cache path
15dd4bd 
Signed-off-by: Mingyang Hao <mingyangHao@users.noreply.github.com>
@mingyangHao mingyangHao changed the title [None][fix] indexer cache fix [None][fix] Indexer cache fix Apr 30, 2026
@mingyangHao mingyangHao changed the title [None][fix] Indexer cache fix [None][fix] Indexer cache fix Apr 30, 2026
@mingyangHao mingyangHao changed the title [None][fix] Indexer cache fix [TRTLLM-12383][fix] Indexer cache fix Apr 30, 2026
@mingyangHao mingyangHao changed the title [TRTLLM-12383][fix] Indexer cache fix [None][fix] Indexer cache fix Apr 30, 2026
@mingyangHao mingyangHao changed the title [None][fix] Indexer cache fix [None][fix] Indexer cache fix Apr 30, 2026
@mingyangHao mingyangHao changed the title [None][fix] Indexer cache fix [None][feat] Change compressor linear dtype Apr 30, 2026
@mingyangHao mingyangHao changed the title [None][feat] Change compressor linear dtype [None][fix] Change compressor linear dtype Apr 30, 2026
@lfr-0531
lfr-0531 merged commit 40656da into NVIDIA:feat/deepseek_v4 Apr 30, 2026
6 of 20 checks passed
lfr-0531 pushed a commit that referenced this pull request May 7, 2026
@mingyangHao @heyuhhh
[None][fix] Change compressor linear dtype (#13605)
c89a904 
Signed-off-by: Mingyang Hao <mingyangHao@users.noreply.github.com>
Co-authored-by: Mingyang Hao <mingyangHao@users.noreply.github.com>
lfr-0531 pushed a commit that referenced this pull request May 14, 2026
@mingyangHao @heyuhhh
[None][fix] Change compressor linear dtype (#13605)
7f0224b 
Signed-off-by: Mingyang Hao <mingyangHao@users.noreply.github.com>
Co-authored-by: Mingyang Hao <mingyangHao@users.noreply.github.com>
Signed-off-by: Fanrong Li <lfr-0531@users.noreply.github.com>
lfr-0531 pushed a commit to lfr-0531/TensorRT-LLM that referenced this pull request May 29, 2026
@mingyangHao @lfr-0531
[None][fix] Change compressor linear dtype (NVIDIA#13605)
34cabf3 
Signed-off-by: Mingyang Hao <mingyangHao@users.noreply.github.com>
Co-authored-by: Mingyang Hao <mingyangHao@users.noreply.github.com>
Signed-off-by: Fanrong Li <lfr-0531@users.noreply.github.com>
(cherry picked from commit 7f0224b)
Signed-off-by: Fanrong Li <lfr-0531@users.noreply.github.com>
lfr-0531 pushed a commit to lfr-0531/TensorRT-LLM that referenced this pull request Jun 1, 2026
@mingyangHao @lfr-0531
[None][fix] Change compressor linear dtype (NVIDIA#13605)
149f789 
Signed-off-by: Mingyang Hao <mingyangHao@users.noreply.github.com>
Co-authored-by: Mingyang Hao <mingyangHao@users.noreply.github.com>
Signed-off-by: Fanrong Li <lfr-0531@users.noreply.github.com>
(cherry picked from commit 7f0224b)
Signed-off-by: Fanrong Li <lfr-0531@users.noreply.github.com>
lfr-0531 pushed a commit to lfr-0531/TensorRT-LLM that referenced this pull request Jun 3, 2026
@mingyangHao @lfr-0531
[None][fix] Change compressor linear dtype (NVIDIA#13605)
9858ca0 
Signed-off-by: Mingyang Hao <mingyangHao@users.noreply.github.com>
Co-authored-by: Mingyang Hao <mingyangHao@users.noreply.github.com>
Signed-off-by: Fanrong Li <lfr-0531@users.noreply.github.com>
(cherry picked from commit 7f0224b)
Signed-off-by: Fanrong Li <lfr-0531@users.noreply.github.com>
lfr-0531 pushed a commit to lfr-0531/TensorRT-LLM that referenced this pull request Jun 7, 2026
@mingyangHao @lfr-0531
[None][fix] Change compressor linear dtype (NVIDIA#13605)
e68b536 
Signed-off-by: Mingyang Hao <mingyangHao@users.noreply.github.com>
Co-authored-by: Mingyang Hao <mingyangHao@users.noreply.github.com>
Signed-off-by: Fanrong Li <lfr-0531@users.noreply.github.com>
(cherry picked from commit 7f0224b)
Signed-off-by: Fanrong Li <lfr-0531@users.noreply.github.com>
lfr-0531 pushed a commit to lfr-0531/TensorRT-LLM that referenced this pull request Jun 10, 2026
@mingyangHao @lfr-0531
[None][fix] Change compressor linear dtype (NVIDIA#13605)
c97a887 
Signed-off-by: Mingyang Hao <mingyangHao@users.noreply.github.com>
Co-authored-by: Mingyang Hao <mingyangHao@users.noreply.github.com>
Signed-off-by: Fanrong Li <lfr-0531@users.noreply.github.com>
lfr-0531 pushed a commit to lfr-0531/TensorRT-LLM that referenced this pull request Jun 11, 2026
@mingyangHao @lfr-0531
[None][fix] Change compressor linear dtype (NVIDIA#13605)
4b09da9 
Signed-off-by: Mingyang Hao <mingyangHao@users.noreply.github.com>
Co-authored-by: Mingyang Hao <mingyangHao@users.noreply.github.com>
Signed-off-by: Fanrong Li <lfr-0531@users.noreply.github.com>
lfr-0531 pushed a commit to lfr-0531/TensorRT-LLM that referenced this pull request Jun 12, 2026
@mingyangHao @lfr-0531
[None][fix] Change compressor linear dtype (NVIDIA#13605)
e53ac23 
Signed-off-by: Mingyang Hao <mingyangHao@users.noreply.github.com>
Co-authored-by: Mingyang Hao <mingyangHao@users.noreply.github.com>
Signed-off-by: Fanrong Li <lfr-0531@users.noreply.github.com>
lfr-0531 pushed a commit to lfr-0531/TensorRT-LLM that referenced this pull request Jun 12, 2026
@mingyangHao @lfr-0531
[None][fix] Change compressor linear dtype (NVIDIA#13605)
8d919e7 
Signed-off-by: Mingyang Hao <mingyangHao@users.noreply.github.com>
Co-authored-by: Mingyang Hao <mingyangHao@users.noreply.github.com>
Signed-off-by: Fanrong Li <lfr-0531@users.noreply.github.com>
(cherry picked from commit 7f0224b)
Signed-off-by: Fanrong Li <lfr-0531@users.noreply.github.com>
lfr-0531 pushed a commit to lfr-0531/TensorRT-LLM that referenced this pull request Jun 13, 2026
@mingyangHao @lfr-0531
[None][fix] Change compressor linear dtype (NVIDIA#13605)
912e69d 
Signed-off-by: Mingyang Hao <mingyangHao@users.noreply.github.com>
Co-authored-by: Mingyang Hao <mingyangHao@users.noreply.github.com>
Signed-off-by: Fanrong Li <lfr-0531@users.noreply.github.com>
(cherry picked from commit 7f0224b)
Signed-off-by: Fanrong Li <lfr-0531@users.noreply.github.com>
lfr-0531 pushed a commit to lfr-0531/TensorRT-LLM that referenced this pull request Jun 16, 2026
@mingyangHao @lfr-0531
[None][fix] Change compressor linear dtype (NVIDIA#13605)
2e6519c 
Signed-off-by: Mingyang Hao <mingyangHao@users.noreply.github.com>
Co-authored-by: Mingyang Hao <mingyangHao@users.noreply.github.com>
Signed-off-by: Fanrong Li <lfr-0531@users.noreply.github.com>
(cherry picked from commit 7f0224b)
Signed-off-by: Fanrong Li <lfr-0531@users.noreply.github.com>
lfr-0531 pushed a commit to lfr-0531/TensorRT-LLM that referenced this pull request Jun 17, 2026
@mingyangHao @lfr-0531
[None][fix] Change compressor linear dtype (NVIDIA#13605)
eb63812 
Signed-off-by: Mingyang Hao <mingyangHao@users.noreply.github.com>
Co-authored-by: Mingyang Hao <mingyangHao@users.noreply.github.com>
Signed-off-by: Fanrong Li <lfr-0531@users.noreply.github.com>
(cherry picked from commit 7f0224b)
Signed-off-by: Fanrong Li <lfr-0531@users.noreply.github.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@QiJune QiJune Awaiting requested review from QiJune QiJune is a code owner automatically assigned from NVIDIA/trt-llm-torch-attention-devs
@yechank-nvidia yechank-nvidia Awaiting requested review from yechank-nvidia yechank-nvidia was automatically assigned from NVIDIA/trt-llm-torch-models-devs

Assignees

@mingyangHao mingyangHao

Labels

deepseek-v4

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

2 participants

@mingyangHao @lfr-0531