| { |
| "encoder": { |
| "model_name": "IVN-RIN/medBIT-r3-plus", |
| "max_length": 512, |
| "train_parameters": true |
| }, |
| "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_data_path": "data/full/train.jsonl", |
| "validation_data_path": "data/full/dev.jsonl", |
| "test_data_path": "data/full/test.jsonl", |
| "trainer": { |
| "num_epochs": 20, |
| "patience": 5, |
| "grad_norm": 5.0, |
| "validation_metric": "+MEAN__relation_f1", |
| "cuda_device": 0, |
| "num_serialized_models_to_keep": 3, |
| "optimizer": { |
| "lr": 0.001, |
| "weight_decay": 0.0, |
| "embedder_lr": 5e-05, |
| "embedder_weight_decay": 0.01 |
| } |
| }, |
| "data_loader": { |
| "batch_size": 1 |
| }, |
| "numpy_seed": 1337, |
| "pytorch_seed": 133, |
| "random_seed": 13370 |
| } |
