Skip to content

Properly Handle DynamicInferenceRequestRecord with latest Mcore #536

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

Open
suiyoubi wants to merge 2 commits into
base: main
Choose a base branch
from
Open

Properly Handle DynamicInferenceRequestRecord with latest Mcore #536

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
e3c07e0
invoke merge for DynamicInferenceRequestRecord
suiyoubi Dec 15, 2025
bc17765
ruff format
suiyoubi Dec 18, 2025
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: 5 additions & 1 deletion nemo_deploy/llm/megatronllm_deployable.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -22,7 +22,7 @@
import torch.distributed
from jinja2 import Template
from megatron.core.inference.common_inference_params import CommonInferenceParams
from megatron.core.inference.inference_request import InferenceRequest
from megatron.core.inference.inference_request import DynamicInferenceRequestRecord, InferenceRequest

from nemo_deploy import ITritonDeployable
from nemo_deploy.llm.inference.inference_base import create_mcore_engine
Expand Down Expand Up @@ -470,6 +470,10 @@ def _infer_fn(
)

results = self.generate(prompts, inference_params)
# Handle DynamicInferenceRequestRecord objects by merging them into a single request
results = [
r.merge(self.mcore_tokenizer) if isinstance(r, DynamicInferenceRequestRecord) else r for r in results
]
if echo:
output_texts = [r.prompt + r.generated_text if text_only else r for r in results]
else:
Expand Down
Loading