| model_name: mnoukhov/pythia410m-tldrprompt-dpo1b-adapter | |
| dataset_name: mnoukhov/openai_comparisons_20k_regen_and_relabelled | |
| eval_split: train | |
| use_peft: False | |
| beta: 0.5 | |
| load_in_8bit: False | |
| bf16: False | |
| fp16: True | |
| per_device_eval_batch_size: 8 | |
| warmup_steps: 150 | |
| mode: eval | |