Instructions to use chchen/Mistral-Nemo-12B-Instruct-SAA with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- PEFT
How to use chchen/Mistral-Nemo-12B-Instruct-SAA with PEFT:
from peft import PeftModel from transformers import AutoModelForCausalLM base_model = AutoModelForCausalLM.from_pretrained("mistralai/Mistral-Nemo-Instruct-2407") model = PeftModel.from_pretrained(base_model, "chchen/Mistral-Nemo-12B-Instruct-SAA") - Notebooks
- Google Colab
- Kaggle
| { | |
| "epoch": 2.986666666666667, | |
| "eval_logits/chosen": -2.242171049118042, | |
| "eval_logits/rejected": -2.265955924987793, | |
| "eval_logps/chosen": -0.11519017815589905, | |
| "eval_logps/rejected": -0.4500538110733032, | |
| "eval_loss": 0.16089552640914917, | |
| "eval_odds_ratio_loss": 1.4658138751983643, | |
| "eval_rewards/accuracies": 0.7900000214576721, | |
| "eval_rewards/chosen": -0.011519019491970539, | |
| "eval_rewards/margins": 0.033486369997262955, | |
| "eval_rewards/rejected": -0.04500538855791092, | |
| "eval_runtime": 5.4517, | |
| "eval_samples_per_second": 18.343, | |
| "eval_sft_loss": 0.01431415043771267, | |
| "eval_steps_per_second": 9.171, | |
| "total_flos": 4.410496642646016e+16, | |
| "train_loss": 0.4940077399923688, | |
| "train_runtime": 519.2548, | |
| "train_samples_per_second": 5.2, | |
| "train_steps_per_second": 0.324 | |
| } |