| { | |
| "model_type": "CodeT5", | |
| "architectures": ["T5ForConditionalGeneration"], | |
| "batch_size": 16, | |
| "learning_rate": 0.00002, | |
| "epochs": 3, | |
| "max_seq_length": 128, | |
| "warmup_steps": 100, | |
| "include_amphigory": true, | |
| "amphigory_ratio": 0, | |
| "vocab_size": 32128, | |
| "d_model": 768, | |
| "num_layers": 12, | |
| "num_heads": 12, | |
| "ffn_dim": 3072, | |
| "dropout_rate": 0.1 | |
| } |