We read every piece of feedback, and take your input very seriously.
To see all available qualifiers, see our documentation.
There was an error while loading. Please reload this page.
1 parent 05ed951 commit f6d6de4Copy full SHA for f6d6de4
README.md
@@ -35,7 +35,7 @@ We provide a [Dockerfile](./Dockerfile) to easily build environments.
35
36
| Method | Bits | 1.5B | 3B | 7B |
37
| ------------------------ | ---- | ------ | ------ | ------ |
38
-| GRPO Full Fine-Tuning | AMP | 2*40GB | 4*40GB | 4*80GB |
+| GRPO Full Fine-Tuning | AMP | 2*24GB | 2*40GB | 4*40GB |
39
40
> [!NOTE]
41
> We are working hard to reduce the VRAM in RL training, LoRA support will be integrated in next updates.
0 commit comments