Instructions to use dendimaki/mistral-classification_v3_sample with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- PEFT
How to use dendimaki/mistral-classification_v3_sample with PEFT:
from peft import PeftModel from transformers import AutoModelForSequenceClassification base_model = AutoModelForSequenceClassification.from_pretrained("mistralai/Mistral-7B-v0.1") model = PeftModel.from_pretrained(base_model, "dendimaki/mistral-classification_v3_sample") - Notebooks
- Google Colab
- Kaggle
Download trainer_state.json from dendimaki/mistral-classification_v3_sample: direct link, hf CLI and curl.
- Browser
- Download file 809 Bytes
-
https://huggingface.co/dendimaki/mistral-classification_v3_sample/resolve/main/trainer_state.json
- Command line
-
hf download hf://dendimaki/mistral-classification_v3_sample/trainer_state.json
-
curl -L -o trainer_state.json https://huggingface.co/dendimaki/mistral-classification_v3_sample/resolve/main/trainer_state.json
809 Bytes
| { | |
| "best_metric": 0.12190821766853333, | |
| "best_model_checkpoint": "multilabel_classification/checkpoint-239", | |
| "epoch": 0.5, | |
| "eval_steps": 500, | |
| "global_step": 239, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.5, | |
| "eval_f1_macro": 0.0, | |
| "eval_f1_micro": 0.0, | |
| "eval_f1_weighted": 0.0, | |
| "eval_loss": 0.12190821766853333, | |
| "eval_runtime": 157.5442, | |
| "eval_samples_per_second": 2.729, | |
| "eval_steps_per_second": 0.343, | |
| "step": 239 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 239, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 500, | |
| "total_flos": 6318616510660608.0, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |