v1olet's picture
Update README.md
63b69bf
|
raw
history blame
400 Bytes
metadata
license: apache-2.0
language:
  - en

DPO from the model ranked 6th on the overall leaderboard and 1st in the 7B leaderboard - v1olet/v1olet_marcoroni-go-bruins-merge-7B.

Training data: comparison_gpt4_en,en_orca_dpo

Train for 1 epoch.

You can use alpaca template.

template_format = """{system}
### Instruction:
{prompt}

### Response:
"""

Developed by: Trong-Hieu Nguyen-Mau