NousResearch/atropos
1
0
Fork 0
mirror of https://github.com/NousResearch/atropos.git synced 2026-04-19 12:57:58 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

more terminal changes

Browse source
This commit is contained in:
Jai Suphavadeeprasit 2026-03-02 14:40:55 -05:00
parent 2f01720899
commit 8d29f49a58
2 changed files with 2 additions and 2 deletions
Download patch file Download diff file Expand all files Collapse all files

2
example_trainer/data.py
View file

@ -5,7 +5,7 @@ Handles data retrieval from Atropos API, padding, batching,
and advantage normalization.
Also extracts inference logprobs for proper GRPO loss computation:
- Inference logprobs serve as π_old (reference policy) for importance sampling
- Inference logprobs are used in importance-ratio computation
- They are batched and padded to align token-by-token with training labels
"""