We read every piece of feedback, and take your input very seriously.
To see all available qualifiers, see our documentation.
1 parent b5b9b55 commit 89805c7Copy full SHA for 89805c7
src/together/cli/api/finetune.py
@@ -145,7 +145,7 @@ def fine_tuning(ctx: click.Context) -> None:
145
@click.option(
146
"--rpo-alpha",
147
type=float,
148
- default=1.0,
+ default=None,
149
help="RPO alpha to control the weight of NLL loss component for chosen responses "
150
"(only used when '--training-method' is 'dpo')",
151
)
0 commit comments