Skip to content

From NVIDIA Megatron-LM for visibility#18

Open
RaymondLi0 wants to merge 6596 commits intobigcode-project:multi-query-attentionfrom
NVIDIA:main
Open

From NVIDIA Megatron-LM for visibility#18
RaymondLi0 wants to merge 6596 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
ahmadki and others added 27 commits February 23, 2026 23:36
@ahmadki
fix: skip non-tensor optimizer state entries in distrib_optimizer sav… (
23dd639 
#3537)

Signed-off-by: Ahmad Kiswani <[email protected]>
@dimapihtar
remove is_unitialized & get_data_modulo_expert_parallel_group (#3414)
2ec295e 
Signed-off-by: dimapihtar <[email protected]>
@dimapihtar
remove deprecated TE module (#3409)
abf97eb 
Signed-off-by: dimapihtar <[email protected]>
@dimapihtar
remove encoder_and_decoder from enums (#3406)
7c7c9e1 
Signed-off-by: dimapihtar <[email protected]>
@github-actions
chore(beep boop 🤖): Bump (main) (2026-02-24)
dd39eb5
@sbak5
Add knobs to choose process groups for fully-parallel-save / load and…
cb24802 
… load-exchange-algo (#2161)
@tdene
Fix off-by-2 error in RL sequence packing (#3551)
5f5f465
@sbak5
Skip unnecessary flattening for Save / Load Planner (#3263)
a8efd34
@faradawn
Multimodal: fix model provider (#3508)
5dc98a6 
Signed-off-by: Faradawn Yang <[email protected]>
@chtruong814
docs: Enable nightly docs publish (#3546)
76b200c 
Signed-off-by: Charlie Truong <[email protected]>
@nschank
Ensure type-checker understands use of Submodules in unit tests (#3425)
f721069
@nschank
Use copy_signature to preserve typing of pass-through methods (#3419)
782e54b
@nschank
Ensure type-checker understands use of Submodules in MTP (#3308)
3597312
@santhnm2
Add mxfp8 quantization for inference linear layers (#3447)
44e27d0 
Signed-off-by: Keshav Santhanam <[email protected]>
@sbak5
Add single-process checkpoint save to avoid forked multiprocessing (#…
2cac78b 
…3424)
@mkhona-nvidia
Fixed fp32 residuals (#3504)
08857d9 
Signed-off-by: mikail <[email protected]>
@maanug-nv
Move config src files into a dedicated dir (#3570)
b17248a 
Signed-off-by: Maanu Grover <[email protected]>
@github-actions
chore: rotate oncall schedule
9bb2375
@ko3n1g
Revert "remove encoder_and_decoder from enums (#3406)" (#3579)
0dba6a0
@yobibyte
Fix default cuda graph persist arg. Add persist to rl common.sh. (#3584)
b359981
@sidsingh-nvidia
Optimize away add request overheads in dummy ep cuda-graphed forward …
60a25aa 
…passes (#3525)
@ko3n1g
ci: Test docs build (#3583)
e7d80bd 
Signed-off-by: oliver könig <[email protected]>
@ko3n1g
ci: Fix docs build for release (#3597)
80ef2ae 
Signed-off-by: oliver könig <[email protected]>
@ko3n1g
ci: Remove secrets (#3598)
2561a59 
Signed-off-by: oliver könig <[email protected]>
@ko3n1g
ci: Define secrets (#3599)
19444b9 
Signed-off-by: oliver könig <[email protected]>
@ko3n1g
ci: gh-release-from-tag (#3600)
a08ad25 
Signed-off-by: oliver könig <[email protected]>
@ko3n1g
Ko3n1g/ci/remove twine username (#3601)
ac449c4 
Signed-off-by: oliver könig <[email protected]>
tdene and others added 30 commits March 20, 2026 16:22
@tdene @jalbericiola
Rename RL timers to be consistent (#3878)
5c87d9a 
Co-authored-by: Jorge Albericio <[email protected]>
@ko3n1g @claude
ci: centralize run configuration in a single configure job (#3962)
c5f9dd3 
Signed-off-by: oliver könig <[email protected]>
Co-authored-by: Claude Sonnet 4.6 <[email protected]>
@ko3n1g
ci: Split unit tests into smaller groups (#3966)
70bda97 
Signed-off-by: oliver könig <[email protected]>
@wdykas
Refit optimization (#3933)
6e6f0b7 
Co-authored-by: William Dykas <[email protected]>
Co-authored-by: root <[email protected]>
Co-authored-by: root <[email protected]>
Co-authored-by: root <[email protected]>
Co-authored-by: root <[email protected]>
@dimapihtar
common strategy simplification (#3229)
2beb593 
Signed-off-by: dimapihtar <[email protected]>
@mathemakitten
Cudagraphs: Remove fwd_graph_input_surface weakref (#3970)
197242e
@ko3n1g @claude
fix: interpolate version correctly in release Slack notification (#3977)
a9aa2c8 
Signed-off-by: oliver könig <[email protected]>
Co-authored-by: Claude Sonnet 4.6 <[email protected]>
@cspades
Make args and kwargs optional positional arguments for the Module hoo…
3f59f71 
…ks. (#3976)

Signed-off-by: Cory Ye <[email protected]>
@chtruong814
ci: Add core-adlr and core-nemo to megatron/training codeowners (#3979)
ade81bc 
Signed-off-by: Charlie Truong <[email protected]>
@deepakn94 @claude
Small quality-of-life improvements in megatron/training (#3957)
b4e4d96 
Co-authored-by: Claude Opus 4.6 <[email protected]>
@tdene
Update throughput golden values to reflect speedup (#3983)
f5b0ec6
@chtruong814
Revert "ci: Add core-adlr and core-nemo to megatron/training codeowne…
a811ac1 
…rs (#3979)" (#3982)

Signed-off-by: Charlie Truong <[email protected]>
@ko3n1g @claude
ci: Add --repo flag to gh pr view in configure job (#3989)
b1c9403 
Signed-off-by: oliver könig <[email protected]>
Co-authored-by: Claude Sonnet 4.6 <[email protected]>
@Phlip79 @claude @ko3n1g
Add common pile scripts (#3902)
febd25e 
Co-authored-by: claude[bot] <209825114+claude[bot]@users.noreply.github.com>
Co-authored-by: oliver könig <[email protected]>
@Phlip79 @claude
Introduce GDN to Mamba (#3535)
8f7fbe7 
Co-authored-by: Claude Opus 4.6 (1M context) <[email protected]>
@saakshigupta2002
Fix IndexError in uniform activation recompute when num_layers not di…
9054192 
…visible by recompute_num_layers (#3562)
@plugyawn @asolergi-nv @BoxiangW
Scaling for MuP over Muon optimizer. (#3715)
f58a328 
Co-authored-by: Antoni-Joan Solergibert <[email protected]>
Co-authored-by: Boxiang Wang <[email protected]>
@cspades
Pass Megatron-FSDP MixedPrecision args to DDPConfig. (#3992)
947bdbb 
Signed-off-by: Cory Ye <[email protected]>
@jenchen13 @claude
[OMNIML-3721] Fix tokenizer unwrapping for nested Megatron-Core token…
488ba8e 
…izer (#3967)

Co-authored-by: Claude Sonnet 4.6 <[email protected]>
@nanz-nv @Victarry @Phlip79
Forced load imbalance (#3380)
7da5b28 
Co-authored-by: Dennis(Zhenhuan) Liu <[email protected]>
Co-authored-by: Philip Petrakian <[email protected]>
@Phlip79 @claude
Add /claude copy command (#3978)
9e28104 
Co-authored-by: claude[bot] <209825114+claude[bot]@users.noreply.github.com>
@yashaswikarnati @dimapihtar @claude
Add multi-module heterogeneous parallelism support for MIMO model (#3211
70a89af 
)

Signed-off-by: ykarnati <[email protected]>
Co-authored-by: Dmytro Pykhtar <[email protected]>
Co-authored-by: Claude Opus 4.5 <[email protected]>
@kinjalpatel27 @dimapihtar
added vllm fakequant export support (#3050)
99638d0 
Co-authored-by: Dmytro Pykhtar <[email protected]>
@jenchen13 @claude
fix(modelopt): use bash array for MLM_EXTRA_ARGS to preserve quoting (#…
a9e4437 
…4002)

Co-authored-by: Claude Opus 4.6 (1M context) <[email protected]>
@sbak5
fix: use dump file prefix for NCCL flight recorder temp files (#3955)
34e1e97
@cluster2600 @maanug-nv
Fix PersistentAsyncCaller.__del__ crash during Python shutdown (#3781)
b44a23c 
Signed-off-by: Maxime Grenu <[email protected]>
Signed-off-by: Maxime Kawawa-Beaudan <[email protected]>
Co-authored-by: Maxime Kawawa-Beaudan <[email protected]>
Co-authored-by: Maanu Grover <[email protected]>
@chtruong814
ci: Run L1 MBridge tests in merge queue (#4009)
0065f9f 
Signed-off-by: Charlie Truong <[email protected]>
@Phlip79
Update Claude review (#3980)
74b0e69
@nschank @yashaswikarnati
Migrate MoeLayer submodules from ModuleSpec to Protocols (#3426)
4f49f2f 
Co-authored-by: Yashaswi Karnati <[email protected]>
@maanug-nv
Guard non-core imports (#3993)
3aa7396 
Signed-off-by: Maanu Grover <[email protected]>
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.