Skip to content

[feat]: RFT trainer component and simplified model import #4647

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
yeshsurya wants to merge 17 commits into
base: main
Choose a base branch
from
Open

[feat]: RFT trainer component and simplified model import #4647

yeshsurya wants to merge 17 commits into from
+2,881 −0

Conversation

@yeshsurya
Copy link
Contributor

@yeshsurya yeshsurya commented Dec 1, 2025 •
edited
Loading

azure_python_grader.py
Link to sanity run :

grpo 2 steps : here

@yeshsurya yeshsurya requested review from a team as code owners December 1, 2025 10:31
@yeshsurya yeshsurya force-pushed the yeshwanth/rft_pipeline branch from 12c8e92 to 954baf6 Compare December 1, 2025 10:31
@github-actions
Copy link

github-actions bot commented Dec 1, 2025 •
edited
Loading

Test Results for assets-test

0 tests   0 ✅  0s ⏱️
0 suites  0 💤
0 files    0 ❌

Results for commit 5fa0bee.

♻️ This comment has been updated with latest results.

@yeshsurya yeshsurya force-pushed the yeshwanth/rft_pipeline branch from bdd0a14 to d5b9bd5 Compare January 2, 2026 09:08
@yeshsurya yeshsurya changed the title [feat]: [draft-awaiting pm review] RFT trainer component and simplified model import [feat]: RFT trainer component and simplified model import Jan 5, 2026
babu-namburi
babu-namburi reviewed Jan 6, 2026
View reviewed changes
assets/training/finetune_acft_hf_nlp/components/model_import_lite/spec.yaml
display_name: Component Model Import
description: Component to import HuggingFace models or AML registered models.

environment: azureml://registries/azureml/environments/acft-hf-nlp-gpu/versions/107
Copy link
Contributor

@babu-namburi babu-namburi Jan 6, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Will this env acft-hf-nlp-data-import work here?

@babu-namburi
Copy link
Contributor

babu-namburi commented Jan 7, 2026

@yeshsurya
I see another issue
Received unexpected config parameter automatic_object_spilling_enabled

I am testing locally to flush all the issues.

@babu-namburi
Copy link
Contributor

babu-namburi commented Jan 7, 2026

May be we should catch these early on

AssertionError: normalized ppo_mini_batch_size 128 should be divisible by ppo_micro_batch_size_per_gpu 10

@babu-namburi
Copy link
Contributor

babu-namburi commented Jan 7, 2026

-- �[36m(pid=15902)�[0m ERROR 01-07 05:57:20 [config.py:32] Failed to import Triton kernels. Please make sure your triton version is compatible. Error: module 'triton.language' has no attribute 'constexpr_function'�[32m [repeated 18x across cluster]�[0m

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@babu-namburi babu-namburi babu-namburi left review comments

+1 more reviewer

@sooryar12 sooryar12 sooryar12 approved these changes

Reviewers whose approvals may not affect merge requirements

At least 1 approving review is required to merge this pull request.

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

@yeshsurya @babu-namburi @sooryar12