mirror of
https://github.com/hiyouga/LLaMA-Factory.git
synced 2025-12-15 03:10:35 +08:00
update dpo examples
This commit is contained in:
@@ -6,6 +6,7 @@ stage: dpo
|
|||||||
do_train: true
|
do_train: true
|
||||||
finetuning_type: lora
|
finetuning_type: lora
|
||||||
lora_target: q_proj,v_proj
|
lora_target: q_proj,v_proj
|
||||||
|
pref_beta: 0.1
|
||||||
pref_loss: sigmoid # [sigmoid (dpo), orpo, simpo]
|
pref_loss: sigmoid # [sigmoid (dpo), orpo, simpo]
|
||||||
|
|
||||||
### dataset
|
### dataset
|
||||||
|
|||||||
Reference in New Issue
Block a user