Skip to content

From NVIDIA Megatron-LM for visibility #18

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Open
wants to merge 5,745 commits into
base: multi-query-attention
Choose a base branch
from
Open

From NVIDIA Megatron-LM for visibility #18

wants to merge 5,745 commits into from

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
ko3n1g and others added 28 commits October 23, 2025 18:51
@ko3n1g
chore: Update codeowners
2e38079 
Signed-off-by: oliver könig <okoenig@nvidia.com>
@ko3n1g
ci(fix): No copyright on push
4d14c57 
Signed-off-by: oliver könig <okoenig@nvidia.com>
@ko3n1g
ci: Extend queue-manager for dev branch (#1906)
a350a6e 
Signed-off-by: oliver könig <okoenig@nvidia.com>
@ko3n1g
ci: Move test optimizer into its own bucket (#1909)
adf4247 
Signed-off-by: oliver könig <okoenig@nvidia.com>
@ko3n1g
ci: Use matrix for approval-bot
1edc4d6 
Signed-off-by: oliver könig <okoenig@nvidia.com>
@ko3n1g
ci: Update function name
04e640b 
Signed-off-by: oliver könig <okoenig@nvidia.com>
@ko3n1g
ci: Adjust approval-bot for copy-pr-bot
c7f154f 
Signed-off-by: oliver könig <okoenig@nvidia.com>
@ko3n1g
ci: Parametrize workflow
019084e 
Signed-off-by: oliver könig <okoenig@nvidia.com>
@ko3n1g
ci: Parametrize workflow
aff784e 
Signed-off-by: oliver könig <okoenig@nvidia.com>
@ko3n1g
ci: Remove attribute
4d282bf 
Signed-off-by: oliver könig <okoenig@nvidia.com>
@ko3n1g
ci: Update container image tag to use GitHub SHA
a597390
@ko3n1g
chore: Remove file
94c6526
@ko3n1g
ci: Fix approval bot
5bfda01 
Signed-off-by: oliver könig <okoenig@nvidia.com>
@ko3n1g
ci: Configure cherrypick bot (#1925)
7e661c9 
Signed-off-by: oliver könig <okoenig@nvidia.com>
@ko3n1g
Ci approve dev (#1933)
9697129 
Signed-off-by: oliver könig <okoenig@nvidia.com>
@ko3n1g
ci: Update nightly schedule (#1934)
6a8bbe9 
Signed-off-by: oliver könig <okoenig@nvidia.com>
@ko3n1g
ci: Bump pre-flight for runs on main/dev (#1935)
3bd66cf 
Signed-off-by: oliver könig <okoenig@nvidia.com>
@ko3n1g
ci: Allow skipping on main (#1936)
e965a15 
Signed-off-by: oliver könig <okoenig@nvidia.com>
@ko3n1g
Ko3n1g/ci/pr template community bot (#1937)
621d17c
@ko3n1g
ci: More granular unit tests buckets (#1932)
3e07859 
Signed-off-by: oliver könig <okoenig@nvidia.com>
@tdene
Add sequence packing to RL (#1911)
b2b9b42 
Add sequence packing to RL
@ko3n1g
chore: Update template (#1939)
bf06bbe 
Signed-off-by: oliver könig <okoenig@nvidia.com>
@ko3n1g
chore: Add description about who can merge (#1940)
ddddc2f 
Signed-off-by: oliver könig <okoenig@nvidia.com>
@ko3n1g
Ko3n1g/ci/fix main on eos (#1938)
6d1aa99 
Signed-off-by: oliver könig <okoenig@nvidia.com>
@ko3n1g
Ko3n1g/ci/internal mrs (#1942)
0d5f25f 
Signed-off-by: oliver könig <okoenig@nvidia.com>
@ko3n1g
ci: Fix branch of approval bot (#1944)
903c4b0 
Signed-off-by: oliver könig <okoenig@nvidia.com>
@ko3n1g
ci: Approvalbot for other branches (#1947)
47a99af 
Signed-off-by: oliver könig <okoenig@nvidia.com>
@ko3n1g
ci(fix): Approval bot (#1949)
768c312 
Signed-off-by: oliver könig <okoenig@nvidia.com>
pablo-garay and others added 30 commits November 24, 2025 14:46
@pablo-garay
tidy / synthesize / enhance
ab1e26e 
Signed-off-by: Pablo Garay <pagaray@nvidia.com>
@pablo-garay
Merge branch 'main' of https://github.com/NVIDIA/Megatron-LM
56e8810
@pablo-garay
Revert: trigger_mbridge_tests.yml‎ file change (#2389)
bc242d9
@ko3n1g
build: Upgrade deps (#2289)
49eef58 
Signed-off-by: oliver könig <okoenig@nvidia.com>
@kanz-nv @sidsingh-nvidia
Change KV cache init to empty to speedup graph recording and first pr…
2a51d86 
…efill (#2358)

Co-authored-by: Siddharth Singh <136645615+sidsingh-nvidia@users.noreply.github.com>
@tdene
Handle UVM compile lock issues (#2299)
4c7d3d6
@Victarry @yaox12
Remove experimental tags for fused kernels. (#2233)
14b791b 
Co-authored-by: Xin Yao <xiny@nvidia.com>
@yaox12
Reduce Overhead in Timers (#2210)
ffb8c35 
Signed-off-by: Xin Yao <xiny@nvidia.com>
@ko3n1g
Revert "build: Upgrade deps (#2289)"
60df5c2 
This reverts commit 49eef58.
@yobibyte @jon-barker @tdene
Fix the entropy sign. (#2374)
ba9caf4 
Co-authored-by: Jon Barker <19699370+jon-barker@users.noreply.github.com>
Co-authored-by: Teodor-Dumitru Ene <34819528+tdene@users.noreply.github.com>
@jon-barker
Remove RL use of mock dataloader and kill RL inference interface on e…
77a2d8b 
…xit (#2387)

Co-authored-by: Jon Barker <jbarker@cw-dfw-cs-001-vscode-01.cm.cluster>
@kanz-nv @sidsingh-nvidia
Fix block_bag for RL (#2399)
6f65536 
Co-authored-by: Siddharth Singh <136645615+sidsingh-nvidia@users.noreply.github.com>
@theothermike
adding action for checking whether PR author is nvidia employee or no…
13efcb8 
…t for selecting ephemeral ci hosts (#2402)
@shanmugamr1992 @ko3n1g
Added top n log probs (#2262)
898d633 
Co-authored-by: Shanmugam Ramasamy <shanmugamr@cw-dfw-cs-001-login-02.cm.cluster>
Co-authored-by: oliver könig <okoenig@nvidia.com>
Co-authored-by: Shanmugam Ramasamy <shanmugamr@cw-dfw-cs-001-login-01.cm.cluster>
@yobibyte @jon-barker
Fix logging when no IS is enabled. (#2375)
3f91727 
Co-authored-by: Jon Barker <19699370+jon-barker@users.noreply.github.com>
@theothermike
fix: exit failure when PR author is external contributor removed (#2410)
6fc13a9
@cspades
Various small fixes for Megatron-FSDP. (#2346)
ebb2e91 
Signed-off-by: Cory Ye <cye@nvidia.com>
@jon-barker
Add grpo loop functional test (#2403)
f5531b0 
Co-authored-by: Jon Barker <jbarker@cw-dfw-cs-001-vscode-01.cm.cluster>
@ko3n1g
Revert "Add grpo loop functional test (#2403)"
cb8f94e 
This reverts commit f5531b0.
@guyueh1
YARN position embedding clear forward method lru cache in init functi…
5153663 
…on (#2229)

Signed-off-by: Guyue Huang <guyueh@nvidia.com>
@kanz-nv
Graph Config Implementation (#2380)
0819f3c
@theothermike
fix: adding k8s taints for ephermeral jobs (#2420)
b96d876
@ko3n1g
ci: Enable functional tests (#2419)
9f15fed 
Signed-off-by: oliver könig <okoenig@nvidia.com>
@ko3n1g
Reapply "build: Upgrade deps (#2289)" (#2408)
40ef044 
Signed-off-by: oliver könig <okoenig@nvidia.com>
@theothermike
fix: use a script to do node tainting in the cicd workflow (#2421)
b21bbad
@ko3n1g
ci: Mark gpt_dynamic_inference_tp1_pp1_583m_cuda_graphs_fp8_logitsmat…
65ce253 
…ch as flaky

Signed-off-by: oliver könig <okoenig@nvidia.com>
@ko3n1g
ci: Disable `gpt_static_inference_cuda_graphs_pad_tp4_pp1_ep4_16B_log…
6646d1a 
…itsmatch`

Signed-off-by: oliver könig <okoenig@nvidia.com>
@yobibyte
Fix rl training with data reuse. (#2428)
66c07b0
@jon-barker
Reapply - Add grpo loop functional test (#2411)
8cde93d 
Co-authored-by: Jon Barker <jbarker@cw-dfw-cs-001-vscode-01.cm.cluster>
@ko3n1g
Revert "Reapply - Add grpo loop functional test (#2411)"
6cc29a2 
This reverts commit 8cde93d.
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.