Use copy_signature to preserve typing of pass-through methods by nschank · Pull Request #2 · nschank/Megatron-LM

nschank · 2026-01-30T20:47:16Z

Show other places that copy_signature can provide a benefit

Signed-off-by: Hongbin Liu <hongbinl@nvidia.com>

Signed-off-by: Youngeun Kwon <youngeunk@nvidia.com> Co-authored-by: Youngeun Kwon <youngeunk@nvidia.com>

Signed-off-by: Keshav Santhanam <ksanthanam@nvidia.com>

This reverts commit ffbc43f.

Signed-off-by: oliver könig <okoenig@nvidia.com>

Signed-off-by: Jimmy Zhang <jiemingz@nvidia.com>

Co-authored-by: Dmytro Pykhtar <37850217+dimapihtar@users.noreply.github.com> Co-authored-by: Xin Yao <xiny@nvidia.com>

Co-authored-by: Rabeeh Mahabadi <rkarimimahab@nb-hel-cs-001-vscode-02.cm.cluster> Co-authored-by: Sanjeev Satheesh <sasatheesh@nvidia.com> Co-authored-by: Deepak Narayanan <dnarayanan@nvidia.com>

Signed-off-by: Santosh Bhavani <santosh.bhavani@live.com> Co-authored-by: Xin Yao <xiny@nvidia.com>

…offload` (NVIDIA#2874)

Co-authored-by: Maanu Grover <109391026+maanug-nv@users.noreply.github.com>

Signed-off-by: Deepak Narayanan <dnarayanan@nvidia.com>

Co-authored-by: Dmytro Pykhtar <37850217+dimapihtar@users.noreply.github.com>

…VIDIA#3194)" This reverts commit e836e62.

This reverts commit 300d1b6.

Co-authored-by: Antoni-Joan Solergibert <asolergibert@nvidia.com>

Signed-off-by: Hollow Man <hollowman@opensuse.org> Co-authored-by: Dmytro Pykhtar <37850217+dimapihtar@users.noreply.github.com> Co-authored-by: Xin Yao <xiny@nvidia.com>

Co-authored-by: Dmytro Pykhtar <37850217+dimapihtar@users.noreply.github.com>

Signed-off-by: Robin Zhang <robinz@nvidia.com>

…ocator (NVIDIA#2857)

Signed-off-by: oliver könig <okoenig@nvidia.com>

Co-authored-by: Xin Yao <xiny@nvidia.com>

Signed-off-by: oliver könig <okoenig@nvidia.com>

Co-authored-by: Jared Casper <155158+jaredcasper@users.noreply.github.com>

Signed-off-by: oliver könig <okoenig@nvidia.com>

) Signed-off-by: Charlie Truong <chtruong@nvidia.com>

…A#3271) Signed-off-by: meg miranda <mmiranda@nvidia.com>

Co-authored-by: thecaptain789 <thecaptain789@users.noreply.github.com> Co-authored-by: Philip Petrakian <ppetrakian@nvidia.com>

)

Signed-off-by: oliver könig <okoenig@nvidia.com>

Co-authored-by: litianjian <litianjian@bytedance.com> Co-authored-by: Yan Bai <baiyan1996@icloud.com>

Co-authored-by: Dmytro Pykhtar <37850217+dimapihtar@users.noreply.github.com>

Signed-off-by: Maanu Grover <maanug@nvidia.com>

Signed-off-by: Keshav Santhanam <ksanthanam@nvidia.com>

Co-authored-by: oliver könig <okoenig@nvidia.com>

…zation. (NVIDIA#3253) Co-authored-by: Claude Opus 4.5 <noreply@anthropic.com>

Signed-off-by: Keshav Santhanam <ksanthanam@nvidia.com>

Co-authored-by: litianjian <litianjian@bytedance.com> Co-authored-by: Yan Bai <baiyan1996@icloud.com> Co-authored-by: Siddharth Singh <sidsingh@nvidia.com> Co-authored-by: Robert Kirby <rkirby@cw-dfw-cs-001-vscode-01.cm.cluster> Co-authored-by: Siddharth Singh <136645615+sidsingh-nvidia@users.noreply.github.com> Co-authored-by: oliver könig <okoenig@nvidia.com> Co-authored-by: Teodor-Dumitru Ene <teodord.ene@gmail.com> Co-authored-by: Teodor-Dumitru Ene <34819528+tdene@users.noreply.github.com>

This reverts commit 1123fc0.

Co-authored-by: Cursor <cursoragent@cursor.com> Co-authored-by: Philip Petrakian <ppetrakian@nvidia.com>

This reverts commit 9119fae.

…#3255) Co-authored-by: oliver könig <okoenig@nvidia.com> Co-authored-by: Yashaswi Karnati <144376261+yashaswikarnati@users.noreply.github.com>

nschank mentioned this pull request Jan 30, 2026

Refactor spec modification/introspection to make references to Submodules typed NVIDIA/Megatron-LM#2834

Closed

6 tasks

[fix] Bug fix for offloading in evaluate() (NVIDIA#3043)

a976754

Signed-off-by: Hongbin Liu <hongbinl@nvidia.com>

nschank force-pushed the submodules branch from d49b1da to 592747b Compare January 30, 2026 22:37

github-actions Bot and others added 27 commits January 31, 2026 00:13

Update copy-pr-bot.yaml [skip ci]

991c38f

cp: Fix: nccl-ub in ddp path (3181) into main (NVIDIA#3182)

5d0a7fd

Signed-off-by: Youngeun Kwon <youngeunk@nvidia.com> Co-authored-by: Youngeun Kwon <youngeunk@nvidia.com>

Miscellaneous inference cleanup (NVIDIA#2955)

ffbc43f

Signed-off-by: Keshav Santhanam <ksanthanam@nvidia.com>

Revert "Miscellaneous inference cleanup (NVIDIA#2955)"

0fe3232

This reverts commit ffbc43f.

ci: Fix DSv3 (NVIDIA#3188)

69a5c63

Signed-off-by: oliver könig <okoenig@nvidia.com>

Fix missing argument in MoELayer.forward() (NVIDIA#3133)

2fadde8

Signed-off-by: Jimmy Zhang <jiemingz@nvidia.com>

Fix H2D stream synchronization in optimizer offload (NVIDIA#3140)

ae67076

Co-authored-by: Dmytro Pykhtar <37850217+dimapihtar@users.noreply.github.com> Co-authored-by: Xin Yao <xiny@nvidia.com>

Add MTP support for hybrid models (NVIDIA#2363)

300d1b6

Co-authored-by: Rabeeh Mahabadi <rkarimimahab@nb-hel-cs-001-vscode-02.cm.cluster> Co-authored-by: Sanjeev Satheesh <sasatheesh@nvidia.com> Co-authored-by: Deepak Narayanan <dnarayanan@nvidia.com>

docs: improve Megatron-LM and Megatron Core descriptions (NVIDIA#3115)

dceb1fb

Signed-off-by: Santosh Bhavani <santosh.bhavani@live.com> Co-authored-by: Xin Yao <xiny@nvidia.com>

Handle step key correctly in checkpoint save with `--optimizer-cpu-…

f4502eb

…offload` (NVIDIA#2874)

mRoPE for MTP (NVIDIA#3114)

70719cd

Co-authored-by: Maanu Grover <109391026+maanug-nv@users.noreply.github.com>

Fix two minor bugs in MTP implementation for hybrid models (NVIDIA#3194)

e836e62

Signed-off-by: Deepak Narayanan <dnarayanan@nvidia.com>

Update README.md (NVIDIA#2111)

1362e4a

Co-authored-by: Dmytro Pykhtar <37850217+dimapihtar@users.noreply.github.com>

Revert "Fix two minor bugs in MTP implementation for hybrid models (N…

31d0c87

…VIDIA#3194)" This reverts commit e836e62.

Revert "Add MTP support for hybrid models (NVIDIA#2363)"

a0cc8ca

This reverts commit 300d1b6.

Fix bug in SFTDataset (NVIDIA#3185)

50546da

Co-authored-by: Antoni-Joan Solergibert <asolergibert@nvidia.com>

Fix several syntax error (NVIDIA#3004)

dff4189

Signed-off-by: Hollow Man <hollowman@opensuse.org> Co-authored-by: Dmytro Pykhtar <37850217+dimapihtar@users.noreply.github.com> Co-authored-by: Xin Yao <xiny@nvidia.com>

Fix for RL Test (NVIDIA#3148)

c4bea0a

Co-authored-by: Dmytro Pykhtar <37850217+dimapihtar@users.noreply.github.com>

Fix latent moe flops and backward_dw (NVIDIA#2977)

a4008d0

Signed-off-by: Robin Zhang <robinz@nvidia.com>

Use global user buffer when the bucket size does not fit FixedPoolAll…

afe443b

…ocator (NVIDIA#2857)

ci: Checkpoint retention (NVIDIA#3205)

78475fe

Signed-off-by: oliver könig <okoenig@nvidia.com>

Add unit test for LatentMoE (NVIDIA#2892)

7080697

Co-authored-by: Xin Yao <xiny@nvidia.com>

ci: Enable unit tests on merge-queue (NVIDIA#3186)

0028273

Signed-off-by: oliver könig <okoenig@nvidia.com>

Fix seq pack flag in get_logprobs (NVIDIA#3206)

94c9eae

ci(fix): Parse unit tests in merge-queue (NVIDIA#3224)

b477d12

Signed-off-by: oliver könig <okoenig@nvidia.com>

Fix TE 2.12 AllGather CI failure (NVIDIA#3101)

1a61b77

Co-authored-by: Jared Casper <155158+jaredcasper@users.noreply.github.com>

ci(hotfix): Pin uv (NVIDIA#3233)

79e7bfe

Signed-off-by: oliver könig <okoenig@nvidia.com>

tdene and others added 26 commits February 11, 2026 15:17

Fix inference for MTP models (NVIDIA#3297)

cd14090

Add a logprobs test with real gpt model. (NVIDIA#2870)

6f5de16

Add simple GRPO functional test (NVIDIA#3323)

b5d50cb

ci: Concurrency control for merge-queue (NVIDIA#3353)

1c245c7

Signed-off-by: oliver könig <okoenig@nvidia.com>

ci: Update golden value download script to work with Github (NVIDIA#3335

d9f075c

) Signed-off-by: Charlie Truong <chtruong@nvidia.com>

Removing etc from main index page, shifted name of discussions (NVIDI…

d0b768f

…A#3271) Signed-off-by: meg miranda <mmiranda@nvidia.com>

fix: correct typos 'seperated' and 'recieved' (NVIDIA#3305)

7d1acf6

Co-authored-by: thecaptain789 <thecaptain789@users.noreply.github.com> Co-authored-by: Philip Petrakian <ppetrakian@nvidia.com>

Improved PyTorch profiler and added PyTorch execution trace (NVIDIA#3273

2807a4e

)

build: Bump TE on 2.12 (NVIDIA#3371)

1959739

Signed-off-by: oliver könig <okoenig@nvidia.com>

ci(hotfix): job conditions (NVIDIA#3376)

f06e669

Signed-off-by: oliver könig <okoenig@nvidia.com>

Update copy-pr-bot.yaml [skip ci]

6467cbc

Record moe routing decisions during inference. (NVIDIA#3034)

faced51

Co-authored-by: litianjian <litianjian@bytedance.com> Co-authored-by: Yan Bai <baiyan1996@icloud.com>

[Main] Fix EP Overlap Bugs for Full-Iter CG (NVIDIA#3164)

dedb6dd

Co-authored-by: Dmytro Pykhtar <37850217+dimapihtar@users.noreply.github.com>

Avoid direct pickle import (NVIDIA#3375)

11a4659

Signed-off-by: Maanu Grover <maanug@nvidia.com>

Delete old pretrain_* files (NVIDIA#3359)

fe9279e

Add Qwen3-VL support with Megatron-FSDP (NVIDIA#2841)

7df15bd

Refactor Mamba chunked prefill (NVIDIA#3265)

c65fb25

Signed-off-by: Keshav Santhanam <ksanthanam@nvidia.com>

Improved parallel logging of learning rate (NVIDIA#3319)

47938af

Co-authored-by: oliver könig <okoenig@nvidia.com>

Add enhanced event tracking with TTFT measurement and compact seriali…

a51c1c8

…zation. (NVIDIA#3253) Co-authored-by: Claude Opus 4.5 <noreply@anthropic.com>

Add assertion that max_requests is divisible by tp_size (NVIDIA#3304)

dbc444f

Signed-off-by: Keshav Santhanam <ksanthanam@nvidia.com>

Revert "Move to using the Inference OpenAI API server (NVIDIA#3107)"

4184bfa

This reverts commit 1123fc0.

Update moe github test cases. (NVIDIA#3077)

9119fae

Co-authored-by: Cursor <cursoragent@cursor.com> Co-authored-by: Philip Petrakian <ppetrakian@nvidia.com>

Revert "Update moe github test cases. (NVIDIA#3077)"

e0aa16b

This reverts commit 9119fae.

Split layer_specs to return Submodules instead of ModuleSpecs (NVIDIA…

28ccdaa

…#3255) Co-authored-by: oliver könig <okoenig@nvidia.com> Co-authored-by: Yashaswi Karnati <144376261+yashaswikarnati@users.noreply.github.com>

Use copy_signature to preserve typing of pass-through methods

9d75ec4

nschank force-pushed the copysig branch from 63a1a56 to 9d75ec4 Compare February 13, 2026 19:49

nschank closed this Feb 13, 2026

nschank deleted the copysig branch February 13, 2026 19:50

nschank restored the copysig branch February 13, 2026 19:53

Navigation Menu

Use copy_signature to preserve typing of pass-through methods - #2

Use copy_signature to preserve typing of pass-through methods#2
nschank wants to merge 130 commits into
submodulesfrom
copysig

nschank commented Jan 30, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

20 participants

Conversation

nschank commented Jan 30, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

20 participants