├── README.md ├── config ├── config.yaml ├── loss │ ├── dpo.yaml │ └── sft.yaml └── model │ ├── blank_model.yaml │ ├── gpt2-large.yaml │ ├── gpt2-xl.yaml │ ├── gptj.yaml │ ├── llama7b.yaml │ ├── pythia28.yaml │ └── pythia69.yaml ├── preference_datasets.py ├── requirements.txt ├── train.py ├── trainers.py └── utils.py /README.md: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/junkangwu/beta-DPO/HEAD/README.md -------------------------------------------------------------------------------- /config/config.yaml: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/junkangwu/beta-DPO/HEAD/config/config.yaml -------------------------------------------------------------------------------- /config/loss/dpo.yaml: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/junkangwu/beta-DPO/HEAD/config/loss/dpo.yaml -------------------------------------------------------------------------------- /config/loss/sft.yaml: -------------------------------------------------------------------------------- 1 | name: sft -------------------------------------------------------------------------------- /config/model/blank_model.yaml: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/junkangwu/beta-DPO/HEAD/config/model/blank_model.yaml -------------------------------------------------------------------------------- /config/model/gpt2-large.yaml: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/junkangwu/beta-DPO/HEAD/config/model/gpt2-large.yaml -------------------------------------------------------------------------------- /config/model/gpt2-xl.yaml: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/junkangwu/beta-DPO/HEAD/config/model/gpt2-xl.yaml -------------------------------------------------------------------------------- /config/model/gptj.yaml: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/junkangwu/beta-DPO/HEAD/config/model/gptj.yaml -------------------------------------------------------------------------------- /config/model/llama7b.yaml: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/junkangwu/beta-DPO/HEAD/config/model/llama7b.yaml -------------------------------------------------------------------------------- /config/model/pythia28.yaml: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/junkangwu/beta-DPO/HEAD/config/model/pythia28.yaml -------------------------------------------------------------------------------- /config/model/pythia69.yaml: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/junkangwu/beta-DPO/HEAD/config/model/pythia69.yaml -------------------------------------------------------------------------------- /preference_datasets.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/junkangwu/beta-DPO/HEAD/preference_datasets.py -------------------------------------------------------------------------------- /requirements.txt: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/junkangwu/beta-DPO/HEAD/requirements.txt -------------------------------------------------------------------------------- /train.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/junkangwu/beta-DPO/HEAD/train.py -------------------------------------------------------------------------------- /trainers.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/junkangwu/beta-DPO/HEAD/trainers.py -------------------------------------------------------------------------------- /utils.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/junkangwu/beta-DPO/HEAD/utils.py --------------------------------------------------------------------------------