-
Notifications
You must be signed in to change notification settings - Fork 203
feat: save checkpoint before timeout to avoid 4-hour runtime limit #734
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.
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
feat: save checkpoint before timeout to avoid 4-hour runtime limit #734
Conversation
terrykong
left a comment
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
thanks for contributing. this would definitely be a valuable feature to have. I've left some comments
|
@terrykong I revised based on your suggestions and let me know if have more comments |
|
@wedu-nvidia could you address the DCO failure and run the pre-commit hooks. See https://github.com/NVIDIA-NeMo/RL/blob/main/CONTRIBUTING.md |
d106c44 to
b3c7f82
Compare
b3c7f82 to
33597d1
Compare
33597d1 to
b3c7f82
Compare
9cfa5a7 to
b242f32
Compare
…g time lmit Signed-off-by: Wei Du <[email protected]>
Signed-off-by: Wei Du <[email protected]>
Signed-off-by: Wei Du <[email protected]>
Signed-off-by: Wei Du <[email protected]>
Signed-off-by: Wei Du <[email protected]>
831926d to
e51cad2
Compare
Signed-off-by: Wei Du <[email protected]>
e51cad2 to
421d34c
Compare
|
Hi @terrykong, all DCO and pre-commit issues have been resolved, and the commits are now properly signed. Please help approve the pending workflows and review the change request when convenient — thanks! |
|
@wedu-nvidia looks like there are still some failures, this time with pyrefly |
Signed-off-by: Wei Du <[email protected]>
|
@terrykong I added another parameter, and hope it can pass all. |
Signed-off-by: Wei Du <[email protected]>
Head branch was pushed to by a user without write access
|
@terrykong The previous error seems solved and I saw another error and I added in |
Signed-off-by: Wei Du <[email protected]>
|
@terrykong Can you help add it the merge queue again? Thanks so much |
|
@terrykong can you put it into mergequeue? |
|
@terrykong Why I did not see the conflict? |
…VIDIA-NeMo#734) Signed-off-by: Wei Du <[email protected]> Signed-off-by: Terry Kong <[email protected]> Co-authored-by: Terry Kong <[email protected]> Signed-off-by: Qidong Su <[email protected]>
commit b246e55 Author: Youngeun Kwon <[email protected]> Date: Mon Aug 25 15:05:48 2025 -0700 update the script Signed-off-by: Youngeun Kwon <[email protected]> commit 5315a6b Author: Youngeun Kwon <[email protected]> Date: Mon Aug 25 13:59:16 2025 -0700 script update Signed-off-by: Youngeun Kwon <[email protected]> commit 4437402 Author: Youngeun Kwon <[email protected]> Date: Tue Jul 15 17:42:23 2025 -0700 local Signed-off-by: Youngeun Kwon <[email protected]> wip Signed-off-by: Youngeun Kwon <[email protected]> add script Signed-off-by: Youngeun Kwon <[email protected]> update script Signed-off-by: Youngeun Kwon <[email protected]> update script Signed-off-by: Youngeun Kwon <[email protected]> interactive Signed-off-by: Youngeun Kwon <[email protected]> commit b721703 Author: Charlie Truong <[email protected]> Date: Mon Aug 18 11:22:54 2025 -0500 build: Fix pytorch image ref in Dockerfile.ngc_pytorch (NVIDIA-NeMo#936) Signed-off-by: Charlie Truong <[email protected]> commit 70b9666 Author: Charlie Truong <[email protected]> Date: Sun Aug 17 21:17:58 2025 -0500 build: Add Dockerfile that uses NGC pytorch image (NVIDIA-NeMo#897) Signed-off-by: Charlie Truong <[email protected]> commit df31c1b Author: pjin-nvidia <[email protected]> Date: Thu Aug 14 18:34:50 2025 -0700 feat: chunked logprob calculation with deferred fp32 cast to help with OOM (NVIDIA-NeMo#918) Signed-off-by: Peter Jin <[email protected]> commit 83c6bfc Author: yuki <[email protected]> Date: Thu Aug 14 21:48:55 2025 +0800 refactor: split sync/async vllm worker ([1/2] of refactor vllm worker) (NVIDIA-NeMo#900) Signed-off-by: Yuki Huang <[email protected]> commit 9f7825e Author: Rayen <[email protected]> Date: Thu Aug 14 12:38:27 2025 +0800 feat: Add TP to embed_tokens and lm_head for Gemma models (NVIDIA-NeMo#879) Signed-off-by: ruit <[email protected]> commit e1f56c4 Author: Terry Kong <[email protected]> Date: Tue Aug 12 13:09:37 2025 -0700 feat: add diagnostic script for problematic embeddings (NVIDIA-NeMo#896) Signed-off-by: Terry Kong <[email protected]> commit 223bfa8 Author: Gerald Shen <[email protected]> Date: Mon Aug 11 18:19:52 2025 -0700 feat: add nemotron5 sharding (NVIDIA-NeMo#481) Signed-off-by: Terry Kong <[email protected]> Co-authored-by: Terry Kong <[email protected]> commit 18b9e2c Author: Terry Kong <[email protected]> Date: Mon Aug 11 15:08:52 2025 -0700 test: lower step count on gemma nightly test to finish within 4 hours (NVIDIA-NeMo#880) Signed-off-by: Terry Kong <[email protected]> commit 8fd8c96 Author: guyueh1 <[email protected]> Date: Mon Aug 11 10:46:29 2025 -0700 feat: Fix and enhances for Nsight system profiling (NVIDIA-NeMo#865) Signed-off-by: Guyue Huang <[email protected]> commit 2b87def Author: Qidong Su <[email protected]> Date: Fri Aug 8 18:54:20 2025 -0400 fix: OOM in deepscaler1.5b with sequence length = 16/24k (NVIDIA-NeMo#875) Signed-off-by: Qidong Su <[email protected]> commit fecf71e Author: Rayen <[email protected]> Date: Sat Aug 9 06:42:07 2025 +0800 fix: remove tie weight check (NVIDIA-NeMo#700) Signed-off-by: ruit <[email protected]> commit d45ff3f Author: Terry Kong <[email protected]> Date: Fri Aug 8 10:07:02 2025 -0700 test: add deepscaler tests + pipe-clean configs + fix eval for deepscaler (NVIDIA-NeMo#866) Signed-off-by: Terry Kong <[email protected]> commit d73c942 Author: Anna Shors <[email protected]> Date: Fri Aug 8 09:27:15 2025 -0700 feat: qwen3 export to HF (NVIDIA-NeMo#873) Signed-off-by: Abdalgader Abubaker <[email protected]> Signed-off-by: Anna Shors <[email protected]> Co-authored-by: Abdalgader Abubaker <[email protected]> commit e924d33 Author: Shang Wang <[email protected]> Date: Fri Aug 8 12:15:34 2025 -0400 docs: Link uv's installation instructions to uv's website (NVIDIA-NeMo#837) Signed-off-by: Shang Wang <[email protected]> commit bbbb3d6 Author: yuki <[email protected]> Date: Fri Aug 8 23:26:15 2025 +0800 fix: fix non-colocated with cpu_offload enabled (NVIDIA-NeMo#861) Signed-off-by: Yuki Huang <[email protected]> commit 88a399e Author: yuki <[email protected]> Date: Fri Aug 8 14:04:08 2025 +0800 chore: remove old fsdp1 unit test (NVIDIA-NeMo#871) Signed-off-by: Yuki Huang <[email protected]> commit b8a89a9 Author: yuki <[email protected]> Date: Fri Aug 8 13:56:19 2025 +0800 feat: support non-colocated in mcore (NVIDIA-NeMo#613) Signed-off-by: Yuki Huang <[email protected]> commit 5910abb Author: Anna Shors <[email protected]> Date: Thu Aug 7 13:11:43 2025 -0700 feat: support DTensor CP in DPO and SFT (NVIDIA-NeMo#798) Signed-off-by: ashors1 <[email protected]> commit 0988a7d Author: Felipe Vieira Frujeri <[email protected]> Date: Wed Aug 6 22:01:32 2025 -0700 fix: Fix error message in VllmGenerationWorker. (NVIDIA-NeMo#633) Signed-off-by: Felipe Vieira Frujeri <[email protected]> commit 233cc07 Author: Parth Chadha <[email protected]> Date: Wed Aug 6 15:14:22 2025 -0700 fix: force use of eager (disabled cuda graphs) due to convergence issues (NVIDIA-NeMo#857) Signed-off-by: Parth Chadha <[email protected]> commit 0557402 Author: Terry Kong <[email protected]> Date: Wed Aug 6 14:44:29 2025 -0700 chore: 0.3.0 -> 0.4.0rc0 (NVIDIA-NeMo#840) Signed-off-by: Terry Kong <[email protected]> commit 03472a0 Author: Terry Kong <[email protected]> Date: Wed Aug 6 14:43:55 2025 -0700 feat: dockerfile can build hermetically or from build context (NVIDIA-NeMo#799) Signed-off-by: Terry Kong <[email protected]> commit 9af0a52 Author: Anna Shors <[email protected]> Date: Wed Aug 6 12:35:51 2025 -0700 fix: fix grpo + mcore checkpointing without validation (NVIDIA-NeMo#844) Signed-off-by: ashors1 <[email protected]> commit b6269f7 Author: Yubo Gao <[email protected]> Date: Tue Aug 5 16:55:02 2025 -0400 feat: track policy training compute throughput (NVIDIA-NeMo#632) Signed-off-by: Yubo Gao <[email protected]> commit b74c5d0 Author: Wei Du <[email protected]> Date: Tue Aug 5 15:05:13 2025 -0500 feat: save checkpoint before timeout to avoid 4-hour runtime limit (NVIDIA-NeMo#734) Signed-off-by: Wei Du <[email protected]> Signed-off-by: Terry Kong <[email protected]> Co-authored-by: Terry Kong <[email protected]> commit c784dd9 Author: Zhiyu Li <[email protected]> Date: Tue Aug 5 10:47:30 2025 -0700 feat: add data shuffle and random seed option (NVIDIA-NeMo#334) Signed-off-by: Zhiyu Li <[email protected]> Signed-off-by: Zhiyu Li <[email protected]> commit c249efc Author: Abdalgader Abubaker <[email protected]> Date: Tue Aug 5 21:33:28 2025 +0400 docs: fix checkpointing command for megatron->hf export (NVIDIA-NeMo#823) Signed-off-by: abdalgader-a <[email protected]> Signed-off-by: Youngeun Kwon <[email protected]>
) Signed-off-by: Wei Du <[email protected]> Signed-off-by: Terry Kong <[email protected]> Co-authored-by: Terry Kong <[email protected]> Signed-off-by: Julien Veron Vialard <[email protected]>
…VIDIA-NeMo#734) Signed-off-by: Wei Du <[email protected]> Signed-off-by: Terry Kong <[email protected]> Co-authored-by: Terry Kong <[email protected]>

…g time lmit
What does this PR do ?
Since the server automatically stops after 4 hours, it's recommended to save a checkpoint beforehand. For example, set the timeout to 3 hours and 45 minutes to ensure check point saved is saved in time
Issues
List issues that this PR closes (syntax):
Usage
# Add a code snippet demonstrating how to use thisBefore your PR is "Ready for review"
Pre checks:
Additional Information