Image Classification
PyTorch
timm
multiple-instance-learning
attention
interpretability
digital-humanities
historical-documents
korean
Instructions to use aronvandepol/visual-feature-discovery with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- timm
How to use aronvandepol/visual-feature-discovery with timm:
import timm model = timm.create_model("hf-hub:aronvandepol/visual-feature-discovery", pretrained=True) - Notebooks
- Google Colab
- Kaggle
Download eval/convnextv2_base_page_split_fold0.json from aronvandepol/visual-feature-discovery: direct link, hf CLI and curl.
- Browser
- Download file 876 Bytes
-
https://huggingface.co/aronvandepol/visual-feature-discovery/resolve/main/eval/convnextv2_base_page_split_fold0.json
- Command line
-
hf download hf://aronvandepol/visual-feature-discovery/eval/convnextv2_base_page_split_fold0.json
-
curl -L -o convnextv2_base_page_split_fold0.json https://huggingface.co/aronvandepol/visual-feature-discovery/resolve/main/eval/convnextv2_base_page_split_fold0.json
876 Bytes
| { | |
| "model_name": "convnextv2_base", | |
| "fold": 0, | |
| "n_folds": 5, | |
| "split_info": { | |
| "fold": 0, | |
| "split_type": "page_level", | |
| "train_samples": 45067, | |
| "val_samples": 7952, | |
| "test_samples": 13255, | |
| "train_books": 510, | |
| "val_books": 509, | |
| "test_books": 509, | |
| "total_unique_books": 510, | |
| "train_val_book_overlap": 509, | |
| "train_test_book_overlap": 509, | |
| "val_test_book_overlap": 508, | |
| "train_val_overlap_pct": 100.0, | |
| "train_test_overlap_pct": 100.0 | |
| }, | |
| "test_results": { | |
| "test_loss_epoch": 0.06839354336261749, | |
| "test_accuracy": 0.9818181991577148, | |
| "test_precision": 0.982060432434082, | |
| "test_recall": 0.9818368554115295, | |
| "test_f1": 0.9819337129592896 | |
| }, | |
| "model_path": "models_page_split/convnextv2_base_fold0_20260220_0811.pth", | |
| "checkpoint_dir": "checkpoints_page_split/convnextv2_base_fold0_20260220_0811" | |
| } |