Skip to content

[Multimodal] Add Multimodal OPD support #1760

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

Merged
zhuzilin merged 2 commits into from
Mar 27, 2026
Merged

[Multimodal] Add Multimodal OPD support #1760

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
7509ba0
add opd mm support
coding-famer Mar 24, 2026
3dca50d
update
coding-famer Mar 24, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 6 additions & 0 deletions slime/rollout/on_policy_distillation.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,6 +1,7 @@
import aiohttp
import torch

from slime.utils.processing_utils import encode_image_for_rollout_engine
from slime.utils.types import Sample


Expand All @@ -16,6 +17,11 @@ async def reward_func(args, sample, **kwargs):
"return_logprob": True,
"logprob_start_len": 0,
}

if sample.multimodal_inputs and sample.multimodal_inputs.get("images"):
image_data = sample.multimodal_inputs["images"]
payload["image_data"] = [encode_image_for_rollout_engine(image) for image in image_data]

session_kwargs = {}
async with aiohttp.ClientSession(**session_kwargs) as session:
async with session.post(args.rm_url, json=payload) as resp:
Expand Down