Skip to content

pjin/nemotron-ray-dev-20251208 vs main #449

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
Draft
pjin-nvidia wants to merge 125 commits into
base: main
Choose a base branch
from
Draft

pjin/nemotron-ray-dev-20251208 vs main #449

pjin-nvidia wants to merge 125 commits into from

Conversation

@pjin-nvidia
Copy link
Contributor

@pjin-nvidia pjin-nvidia commented Dec 9, 2025

No description provided.

@pjin-nvidia
Miscellaneous infra.
316d43e 
Signed-off-by: Peter Jin <[email protected]>
@pjin-nvidia
Ray utils.
dd953b0 
Signed-off-by: Peter Jin <[email protected]>
@pjin-nvidia
No cover.
5717d7a 
Signed-off-by: Peter Jin <[email protected]>
@pjin-nvidia
Remove DEBUG. Comment.
4ecd8d3 
Signed-off-by: Peter Jin <[email protected]>
@pjin-nvidia
Comment about ray package extra.
8103dbf 
Signed-off-by: Peter Jin <[email protected]>
@pjin-nvidia
The.
dc493d5 
Signed-off-by: Peter Jin <[email protected]>
@pjin-nvidia
Merge remote-tracking branch 'origin/main' into pjin/misc-infra
f9e5d8f 
Signed-off-by: Peter Jin <[email protected]>
@pjin-nvidia
Fix test (?).
9502d82 
Signed-off-by: Peter Jin <[email protected]>
@pjin-nvidia
Initial support for server pyproject.toml (WIP).
0475d5e 
Signed-off-by: Peter Jin <[email protected]>
@pjin-nvidia
Fix pyproject.toml check.
d86756b 
Signed-off-by: Peter Jin <[email protected]>
@pjin-nvidia
Working directory Path.
79028a6 
Signed-off-by: Peter Jin <[email protected]>
@pjin-nvidia
Install a server venv from pyproject.toml if available.
7e62b1d 
Signed-off-by: Peter Jin <[email protected]>
@pjin-nvidia
Deprecated vllm_model requirements.txt.
36efb94 
Signed-off-by: Peter Jin <[email protected]>
@pjin-nvidia
Consistently use dashes in package names.
8d49b95 
Signed-off-by: Peter Jin <[email protected]>
@pjin-nvidia
Lint.
6fb0a95 
Signed-off-by: Peter Jin <[email protected]>
@pjin-nvidia
Cleanup.
7231efa 
Signed-off-by: Peter Jin <[email protected]>
@pjin-nvidia
VLLM server spinup.
8fc0d9d 
Signed-off-by: Peter Jin <[email protected]>
@pjin-nvidia
VLLM server host and port.
8975e98 
Signed-off-by: Peter Jin <[email protected]>
@pjin-nvidia
Allocate the free port for VLLM in the model server process.
51ba6fc 
Signed-off-by: Peter Jin <[email protected]>
@pjin-nvidia
Type.
aa97796 
Signed-off-by: Peter Jin <[email protected]>
@pjin-nvidia
Fix for pyproject.toml (this works lol).
6ec9325 
Signed-off-by: Peter Jin <[email protected]>
@pjin-nvidia
VLLM server "routing" (just re-using the existing multiple clients).
33ec3f9 
Signed-off-by: Peter Jin <[email protected]>
@pjin-nvidia
Better order.
77cda85 
Signed-off-by: Peter Jin <[email protected]>
@pjin-nvidia
Merge remote-tracking branch 'origin/main' into pjin/ray-utils
44dcee1 
Signed-off-by: Peter Jin <[email protected]>
@pjin-nvidia
WIP.
a85f4f0 
Signed-off-by: Peter Jin <[email protected]>
@pjin-nvidia
Comment.
7201c8f 
Signed-off-by: Peter Jin <[email protected]>
@pjin-nvidia
Default to "mp" backend.
834d9b9 
Signed-off-by: Peter Jin <[email protected]>
@pjin-nvidia
Cleanup.
5ee8b57 
Signed-off-by: Peter Jin <[email protected]>
@pjin-nvidia
Cleanup.
10b5295 
Signed-off-by: Peter Jin <[email protected]>
@pjin-nvidia
Non-async VLLM server heartbeat to avoid early asyncio event loop.
e4c5573 
Signed-off-by: Peter Jin <[email protected]>
@pjin-nvidia
Fix for recent VLLM.
9058505 
Signed-off-by: Peter Jin <[email protected]>
@pjin-nvidia
Merge remote-tracking branch 'origin/pjin/misc-infra' into pjin/nemot…
ff120a7 
…ron-ray-dev-20251208

Signed-off-by: Peter Jin <[email protected]>
@pjin-nvidia
Merge remote-tracking branch 'origin/pjin/ray-utils' into pjin/nemotr…
a0b6507 
…on-ray-dev-20251208

Signed-off-by: Peter Jin <[email protected]>
@pjin-nvidia
Pick fix from #359 and revert max_steps = 1.
a119214 
Signed-off-by: Peter Jin <[email protected]>
@pjin-nvidia
Merge remote-tracking branch 'origin/pjin/nano-v3-main-dev-20251207' …
9ef1475 
…into pjin/nemotron-ray-dev-20251208

Signed-off-by: Peter Jin <[email protected]>
@pjin-nvidia
Merge remote-tracking branch 'origin/main' into pjin/nano-v3-main-dev…
7e5aa85 
…-20251207

Signed-off-by: Peter Jin <[email protected]>
@pjin-nvidia
Merge remote-tracking branch 'origin/main' into pjin/nemotron-ray-dev…
923b891 
…-20251208

Signed-off-by: Peter Jin <[email protected]>
@pjin-nvidia
Merge remote-tracking branch 'origin/main' into pjin/nano-v3-main-dev…
74d8bb8 
…-20251207

Signed-off-by: Peter Jin <[email protected]>
@pjin-nvidia
Rollect collection iterator returns both the input row and output res…
abad135 
…ult.

Signed-off-by: Peter Jin <[email protected]>
@pjin-nvidia
Type annotation.
35e18c2 
Signed-off-by: Peter Jin <[email protected]>
@pjin-nvidia
Pin vllm == 0.11.2.
25996f8 
Signed-off-by: Peter Jin <[email protected]>
@pjin-nvidia
Merge remote-tracking branch 'origin/pjin/nano-v3-main-dev-20251207' …
e24e3af 
…into pjin/nemotron-ray-dev-20251208

Signed-off-by: Peter Jin <[email protected]>
@pjin-nvidia
Merge remote-tracking branch 'origin/main' into pjin/nemotron-ray-dev…
54c6812 
…-20251208

Signed-off-by: Peter Jin <[email protected]>
@pjin-nvidia pjin-nvidia changed the title [DO NOT MERGE] pjin/nemotron-ray-dev-20251208 vs main pjin/nemotron-ray-dev-20251208 vs main Dec 15, 2025
HeyyyyyyG and others added 16 commits December 15, 2025 15:54
@HeyyyyyyG
add genrm rlhf
9625d9d 
Signed-off-by: Jiaqi Zeng <[email protected]>
@HeyyyyyyG
Merge remote-tracking branch 'origin/pjin/nemotron-ray-dev-20251208' …
73f95ea 
…into jiaqiz/add_genrm_rlhf
@pjin-nvidia
Re-pin ray.
6e93c21 
Signed-off-by: Peter Jin <[email protected]>
@fayejf
add long context env
10af204 
Signed-off-by: fayejf <[email protected]>
@fayejf
fix
8226f5c 
Signed-off-by: fayejf <[email protected]>
@pjin-nvidia
Rename lc => lc_judge.
c524692 
Signed-off-by: Peter Jin <[email protected]>
@pjin-nvidia
Revert swallowing judge model endpoint errors.
9c88189 
Signed-off-by: Peter Jin <[email protected]>
@pjin-nvidia
Merge remote-tracking branch 'origin/fjia/lc_judge' into pjin/nemotro…
44c2e15 
…n-ray-dev-20251208

Signed-off-by: Peter Jin <[email protected]>
@pjin-nvidia
Fix missing config field.
6f29347 
Signed-off-by: Peter Jin <[email protected]>
@pjin-nvidia
Merge remote-tracking branch 'github/fjia/lc_judge' into pjin/nemotro…
f50ec7c 
…n-ray-dev-20251208

Signed-off-by: Peter Jin <[email protected]>
@pjin-nvidia
Merge remote-tracking branch 'origin/main' into pjin/nemotron-ray-dev…
5e36ca1 
…-20251208

Signed-off-by: Peter Jin <[email protected]>
@pjin-nvidia
Cherrypick: fix args bugs.
e5fd045 
Signed-off-by: Jiaqi Zeng <[email protected]>
Signed-off-by: Peter Jin <[email protected]>
@pjin-nvidia
Fallback in case json.loads fails (???).
63a13eb 
Signed-off-by: Jiaqi Zeng <[email protected]>
Signed-off-by: Peter Jin <[email protected]>
@pjin-nvidia
fix genrm comparison strategy and calendar env
d4878bc 
Signed-off-by: Jiaqi Zeng <[email protected]>
Signed-off-by: Peter Jin <[email protected]>
@pjin-nvidia
Log server stdout/err to logfiles in the server working dirs.
8e28efc 
Signed-off-by: Peter Jin <[email protected]>
@pjin-nvidia
Block off more port ranges to avoid potential conflicts.
72242eb 
Signed-off-by: Peter Jin <[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.

4 participants

@pjin-nvidia @HeyyyyyyG @fayejf