| { | |
| "model_type": "ConcatModelWithRationale", | |
| "hatebert_model": "GroNLP/hateBERT", | |
| "rationale_model": "bert-base-uncased", | |
| "cnn_num_filters": 128, | |
| "cnn_kernel_sizes": [2, 3, 4], | |
| "adapter_dim": 128, | |
| "max_length": 128, | |
| "num_labels": 2, | |
| "architecture": { | |
| "temporal_cnn": "TemporalCNN", | |
| "msa_cnn": "MultiScaleAttentionCNN", | |
| "projection": "ProjectionMLP" | |
| }, | |
| "training_info": { | |
| "batch_size": 8, | |
| "learning_rate": 1e-5, | |
| "weight_decay": 0.05, | |
| "dropout": 0.5, | |
| "best_val_loss": 0.27 | |
| } | |
| } |