metadata
license: mit
This is a test DPO finetune of Microsoft phi-2
Two DPO datasets are used:
Intel/orca_dpo_pairs
argilla/ultrafeedback-binarized-preferences-cleaned
Training was for 1 epoch as a qlora with Rank 64, Delta 128
license: mit
This is a test DPO finetune of Microsoft phi-2
Two DPO datasets are used:
Intel/orca_dpo_pairs
argilla/ultrafeedback-binarized-preferences-cleaned
Training was for 1 epoch as a qlora with Rank 64, Delta 128