Instructions to use maxbeettww/roberta-MeSH-lora-without-duplicate-path with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- PEFT
How to use maxbeettww/roberta-MeSH-lora-without-duplicate-path with PEFT:
Task type is invalid.
- Transformers
How to use maxbeettww/roberta-MeSH-lora-without-duplicate-path with Transformers:
# pip install -U transformers accelerate # Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("maxbeettww/roberta-MeSH-lora-without-duplicate-path", device_map="auto") - Notebooks
- Google Colab
- Kaggle
Download training_args.json from maxbeettww/roberta-MeSH-lora-without-duplicate-path: direct link, hf CLI and curl.
- Browser
- Download file 479 Bytes
-
https://huggingface.co/maxbeettww/roberta-MeSH-lora-without-duplicate-path/resolve/main/training_args.json
- Command line
-
hf download hf://maxbeettww/roberta-MeSH-lora-without-duplicate-path/training_args.json
-
curl -L -o training_args.json https://huggingface.co/maxbeettww/roberta-MeSH-lora-without-duplicate-path/resolve/main/training_args.json
479 Bytes
| { | |
| "base_model": "roberta-base", | |
| "data": { | |
| "csv_path": "mesh_dataset.csv", | |
| "split": "train=0.9, val=0.1", | |
| "random_state": 42 | |
| }, | |
| "hyperparams": { | |
| "batch_size": 32, | |
| "lr": 0.0001, | |
| "num_epochs_max": 10, | |
| "patience": 3, | |
| "min_delta": 0.0001, | |
| "lora_r": 16, | |
| "lora_alpha": 32, | |
| "lora_dropout": 0.1, | |
| "target_modules": [ | |
| "query", | |
| "key", | |
| "value", | |
| "output.dense" | |
| ] | |
| }, | |
| "best_val_loss": 0.008271555498198692 | |
| } |