Skip to content

Commit f078aa9

Browse files
committed
refine framework picture
1 parent 297b312 commit f078aa9

File tree

3 files changed

+1
-1
lines changed

3 files changed

+1
-1
lines changed

assets/framework.jpg

1.29 KB
Loading

assets/multi_lora.png

-106 KB
Loading

src/twinkle/advantage/base.py

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -20,7 +20,7 @@ def __call__(self,
2020
- compute_advantages_rloo: RLOO-style (leave-one-out baseline)
2121
2222
Example:
23-
>>> from twinkle.rl import GRPOAdvantage
23+
>>> from twinkle.advantage import GRPOAdvantage
2424
>>> rewards = [0.0, 1.0, 0.0, 1.0, 1.0, 0.0, 0.0, 0.0] # 2 prompts, 4 samples each
2525
>>> advantages = GRPOAdvantage()(rewards, num_generations=4)
2626
"""

0 commit comments

Comments
 (0)