⚠ This page is served via a proxy. Original site: https://github.com
This service does not collect credentials or authentication data.
Skip to content

From NVIDIA Megatron-LM for visibility#18

Open
RaymondLi0 wants to merge 6208 commits intobigcode-project:multi-query-attentionfrom
NVIDIA:main
Open

From NVIDIA Megatron-LM for visibility#18
RaymondLi0 wants to merge 6208 commits intobigcode-project:multi-query-attentionfrom
NVIDIA:main

Conversation

@RaymondLi0
Copy link
Collaborator

@RaymondLi0 RaymondLi0 commented Jan 24, 2023

No description provided.

@RaymondLi0 RaymondLi0 changed the base branch from multi-query-attention to before-merge June 20, 2023 20:12
@RaymondLi0 RaymondLi0 changed the base branch from before-merge to multi-query-attention June 20, 2023 20:12
Phlip79 and others added 28 commits January 12, 2026 16:47
@Phlip79
Update Slack user group when oncall changes (#2859)
ba3fd4f
@santhnm2
Remove unused FlashAttention3 args (#2898)
4fc7935 
Signed-off-by: Keshav Santhanam <ksanthanam@nvidia.com>
@Phlip79
Use different token for assign logic (#2893)
438f2cf
@ko3n1g
chore: Add --no-container-mount-home to script (#2906)
7b7e687 
Signed-off-by: oliver könig <okoenig@nvidia.com>
@ko3n1g
build: Bump deps (#2911)
e068944 
Signed-off-by: oliver könig <okoenig@nvidia.com>
@yaoyu-33 @dimapihtar @Phlip79
feat: m4 leftover changes (#2506)
6704731 
Signed-off-by: dimapihtar <dpihtar@gmail.com>
Co-authored-by: Mcore Bot <mcore-bot@nvidia.com>
Co-authored-by: Dmytro Pykhtar <37850217+dimapihtar@users.noreply.github.com>
Co-authored-by: dimapihtar <dpihtar@gmail.com>
Co-authored-by: Philip Petrakian <pgpetrak@gmail.com>
@tdene
Fix RL sequence packing bin size (#2909)
ce2cc40
@chtruong814 @santhnm2
Revert "Remove unused FlashAttention3 args (#2898)" (#2916)
3f6ad46 
Signed-off-by: Charlie Truong <chtruong@nvidia.com>
Co-authored-by: Keshav Santhanam <ksanthanam@nvidia.com>
@ko3n1g
Revert "feat: m4 leftover changes (#2506)"
f967176 
This reverts commit 6704731.
@chtruong814
ci: Skip broken tests after dependency bump (#2934)
8d826c1 
Signed-off-by: Charlie Truong <chtruong@nvidia.com>
@ko3n1g
build: Downgrade flashinfer
b71059a 
Signed-off-by: oliver könig <okoenig@nvidia.com>
@ko3n1g
Reapply "feat: m4 leftover changes (#2506)"
8918762 
This reverts commit f967176.
@ko3n1g
Reapply "Remove unused FlashAttention3 args (#2898)" (#2916)
b7a9d36 
This reverts commit 3f6ad46.
@ko3n1g
Revert "ci: Skip broken tests after dependency bump (#2934)"
d6af382 
This reverts commit 8d826c1.
@ko3n1g
Ko3n1g/build/downgrade flashinfer (#2937)
2490e0c 
Signed-off-by: oliver könig <okoenig@nvidia.com>
@chtruong814
ci: Skip unit test cleanup (#2940)
5241224 
Signed-off-by: Charlie Truong <chtruong@nvidia.com>
@zhongbozhu @Phlip79
[MAIN][NVFP4][MOE] 128 Zero Padding for Grouped Quantization kernels …
fe87250 
…and Cuda Graph Support (#2655)

Signed-off-by: Zhongbo Zhu <zhongboz@nvidia.com>
Co-authored-by: Philip Petrakian <ppetrakian@nvidia.com>
@FDecaYed @yanring
@skyw @ko3n1g @BoxiangW @mkhona-nvidia @Phlip79
Add muon and layerwise distributed optimizer (#2241)
5548cdc 
Signed-off-by: Boxiang Wang <boxiangw@nvidia.com>
Signed-off-by: Deyu Fu <deyuf@nvidia.com>
Signed-off-by: Hao Wu <skyw@nvidia.com>
Co-authored-by: Zijie Yan <zijiey@nvidia.com>
Co-authored-by: Hao Wu <skyw@nvidia.com>
Co-authored-by: oliver könig <okoenig@nvidia.com>
Co-authored-by: Boxiang Wang <boxiangw@nvidia.com>
Co-authored-by: mikail <mkhona@nvidia.com>
Co-authored-by: Philip Petrakian <ppetrakian@nvidia.com>
@deepakn94
Support DDP overlap for models with repeated parameters (#2837)
aa3f105 
Signed-off-by: Deepak Narayanan <dnarayanan@nvidia.com>
@ko3n1g
build: 26.02 dependency bump main (#2923)
82049aa 
Signed-off-by: oliver könig <okoenig@nvidia.com>
@ko3n1g
ci(hotfix): Update golden values after 26.02 bump
69ba809 
Signed-off-by: oliver könig <okoenig@nvidia.com>
@wdykas
RL refit pipelining support (#2878)
f55ec36 
Co-authored-by: root <root@gpu-h100-0348.cm.cluster>
Co-authored-by: root <root@gpu-h100-0193.cm.cluster>
Co-authored-by: root <root@gpu-h100-0082.cm.cluster>
Co-authored-by: root <root@gpu-h100-0495.cm.cluster>
Co-authored-by: William Dykas <wdykas@cw-pdx-cs-001-vscode-02.cm.cluster>
Co-authored-by: root <root@gpu-h100-0213.cm.cluster>
Co-authored-by: root <root@gpu-h100-0435.cm.cluster>
Co-authored-by: root <root@gpu-h100-0188.cm.cluster>
Co-authored-by: root <root@gpu-h100-0032.cm.cluster>
Co-authored-by: root <root@gpu-h100-0023.cm.cluster>
Co-authored-by: root <root@gpu-h100-0368.cm.cluster>
Co-authored-by: root <root@gpu-h100-0203.cm.cluster>
Co-authored-by: root <root@gpu-h100-0229.cm.cluster>
Co-authored-by: root <root@gpu-h100-0123.cm.cluster>
Co-authored-by: root <root@gpu-h100-0217.cm.cluster>
Co-authored-by: root <root@gpu-h100-0496.cm.cluster>
Co-authored-by: root <root@gpu-h100-0022.cm.cluster>
Co-authored-by: root <root@gpu-h100-0176.cm.cluster>
Co-authored-by: root <root@gpu-h100-0261.cm.cluster>
Co-authored-by: root <root@gpu-h100-0029.cm.cluster>
Co-authored-by: root <root@gpu-h100-0215.cm.cluster>
@ko3n1g
Revert "[dev] Add assertion for mxfp8 params without dp overlap (#2270)…
409ebf1 
…" (#2901)
@ko3n1g
ci(hotfix): Remove duplicated test
667ac0f 
Signed-off-by: oliver könig <okoenig@nvidia.com>
@ko3n1g
ci(hotfix): Disable more throughput tests
50fda17 
Signed-off-by: oliver könig <okoenig@nvidia.com>
@ko3n1g
gpt3_moe_mcore_te_ep8_resume_torch_dist_muon
0994ce2 
Signed-off-by: oliver könig <okoenig@nvidia.com>
@ko3n1g
gpt3_moe_mcore_te_ep8_resume_torch_dist_dist_muon
9f341dc 
Signed-off-by: oliver könig <okoenig@nvidia.com>
@ko3n1g
gpt3_moe_mcore_te_tp2_pp2_ep4_etp1_no_mtp_no_a2a_ovlp_fine_grained_of…
1ed4d52 
…floading

Signed-off-by: oliver könig <okoenig@nvidia.com>
ko3n1g and others added 30 commits February 3, 2026 14:01
@ko3n1g
ci: Enable unit tests on merge-queue (#3186)
0028273 
Signed-off-by: oliver könig <okoenig@nvidia.com>
@mathemakitten
Fix seq pack flag in get_logprobs (#3206)
94c9eae
@ko3n1g
ci(fix): Parse unit tests in merge-queue (#3224)
b477d12 
Signed-off-by: oliver könig <okoenig@nvidia.com>
@BestJuly @jaredcasper
Fix TE 2.12 AllGather CI failure (#3101)
1a61b77 
Co-authored-by: Jared Casper <155158+jaredcasper@users.noreply.github.com>
@ko3n1g
ci(hotfix): Pin uv (#3233)
79e7bfe 
Signed-off-by: oliver könig <okoenig@nvidia.com>
@mathemakitten
Add a unit test to check that RL get_logprobs will reuse training c…
18d69f1 
…udagraphed forward pass (#3209)
@mathemakitten
Do not offload grad buffers when training graphs are enabled (#3231)
27a5f83
@parthmannan
Fix missing PackedSeqParams import (#3214)
bc2eb9a
@santhnm2
Synchronize the request counts for EP inference with strict matching (#…
1fdb29f 
…3033)

Signed-off-by: Keshav Santhanam <ksanthanam@nvidia.com>
@tdene
Do not let requests fail silently inside inference engine (#3228)
e02344e
@github-actions
Update copy-pr-bot.yaml [skip ci]
4c48248
@tdene
Fix coordinator address collision check in flask (#3208)
9050d5b
@wdykas
torch saver inference model offload (#3170)
cd5ed74
@Autumn1998
enable cuda graph ut (#3197)
982ca5d
@wplf
Support EP with HSDP (#2840)
473e283 
Signed-off-by: jinliangl <jinliangl@nvidia.com>
Co-authored-by: Jinliang Li <jinliangl@pool0-01676.cm.cluster>
Co-authored-by: Jinliang Li <jinliangl@cw-dfw-cs-001-vscode-01.cm.cluster>
@BestJuly
[Main] Add the missing part to support 1F1B overlap for Qwen3-Next (#…
4a23972 
…2997)
@parthmannan
Missing import fix (#3241)
c036e77
@santhnm2 @ko3n1g
Miscellaneous inference cleanup (Replay of !2955) (#3232)
43db8c1 
Signed-off-by: Keshav Santhanam <ksanthanam@nvidia.com>
Co-authored-by: oliver könig <okoenig@nvidia.com>
@maanug-nv
Add DistributedInitConfig (#3173)
adce147 
Signed-off-by: Maanu Grover <maanug@nvidia.com>
@yashaswikarnati
Fix checkpoint converter missing parallel group initialization (#3217)
f3e6cc8
@BestJuly @rkarimimahab
Skip empty sequences and chunks in MTP tensor roll (#3035)
d558b5f 
Co-authored-by: Rabeeh Karimi Mahabadi <rkarimimahab@nvidia.com>
@nschank @Glabred @yaox12
Implement get_parameters for ChainedOptimizer (#3201)
f708b5d 
Co-authored-by: Jeffrey Chen <jeffrey@reflection.ai>
Co-authored-by: Xin Yao <xiny@nvidia.com>
@ko3n1g
ci(fix): Create main/dev image tags (#3252)
66c432a 
Signed-off-by: oliver könig <okoenig@nvidia.com>
@ko3n1g
ci(hotfix): Skopeo copy
e24767f 
Signed-off-by: oliver könig <okoenig@nvidia.com>
@ko3n1g
ci(hotfix): Add skopeo
d959620 
Signed-off-by: oliver könig <okoenig@nvidia.com>
@sancha
Reapply "Add MTP support for hybrid models (#2363)" (#3207)
9d71cb1
@Phlip79
Fix uv install for GH actions (#3259)
b043863
@janEbert @yaox12
Update the project structure in README (#3251)
dd7d141 
Co-authored-by: Xin Yao <xiny@nvidia.com>
@github-actions
chore: rotate oncall schedule
1f6d8c2
@BestJuly @rj42
@yaox12 @juntaowww
Cherry-pick: Fix mtp_num_layers and clip_qk issues (#2581, #2776) (#3075
1b11076 
)

Co-authored-by: rj42 <lbkzman@gmail.com>
Co-authored-by: Xin Yao <xiny@nvidia.com>
Co-authored-by: Juntao Wang <juntaow@nvidia.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.