Feature Extraction
Transformers
Safetensors
Italian
radgraph_it
radiology
information-extraction
named-entity-recognition
relation-extraction
medical
radgraph
custom_code
Instructions to use radgraphIT/Radgraph-IT with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use radgraphIT/Radgraph-IT with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("feature-extraction", model="radgraphIT/Radgraph-IT", trust_remote_code=True)# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("radgraphIT/Radgraph-IT", trust_remote_code=True, device_map="auto") - Notebooks
- Google Colab
- Kaggle
Add transformers-compatible wrapper (AutoModel/AutoTokenizer via trust_remote_code)
0c48771 verified | { | |
| "model_type": "radgraph_it", | |
| "architectures": [ | |
| "RadgraphModel" | |
| ], | |
| "auto_map": { | |
| "AutoConfig": "configuration_radgraph.RadgraphConfig", | |
| "AutoModel": "modeling_radgraph.RadgraphModel" | |
| }, | |
| "encoder_name": "IVN-RIN/medBIT-r3-plus", | |
| "max_length": 512, | |
| "max_span_width": 12, | |
| "feature_size": 20, | |
| "feedforward_params": { | |
| "hidden_dims": [ | |
| 150, | |
| 150 | |
| ], | |
| "dropout": 0.4 | |
| }, | |
| "loss_weights": { | |
| "ner": 0.2, | |
| "relation": 1.0 | |
| }, | |
| "relation_spans_per_word": 0.5, | |
| "train_encoder": true, | |
| "span_pooling": false, | |
| "transformer_params": null, | |
| "relation_context": false, | |
| "relation_feedforward_params": null, | |
| "dataset": "radgraph-it", | |
| "ner_labels": { | |
| "": 0, | |
| "Anatomy::definitely present": 1, | |
| "Observation::definitely present": 2, | |
| "Observation::definitely absent": 3, | |
| "Observation::uncertain": 4, | |
| "Anatomy::definitely absent": 5, | |
| "Anatomy::uncertain": 6 | |
| }, | |
| "relation_labels": { | |
| "": 0, | |
| "modify": 1, | |
| "located_at": 2, | |
| "suggestive_of": 3 | |
| }, | |
| "torch_dtype": "float32", | |
| "transformers_version": "5.14.1" | |
| } |