| { | |
| "loader": "loaders/gigaword.py", | |
| "dataset": "data/train-data/gigaword", | |
| "indices": "data/train-data/gigaword/indices.npy", | |
| "model_dir": "data/models/example", | |
| "verbose": true, | |
| "print_every": 1, | |
| "eval_every": 10, | |
| "save_every": 10, | |
| "max_val_steps": 8, | |
| "max_train_seconds": null, | |
| "max_train_steps": 1000, | |
| "batch_size": 1, | |
| "learning_rate": 1e-05, | |
| "k_samples": 10, | |
| "sample_aggregation": "max", | |
| "loss": "pgb", | |
| "encoder_model_id": "distilroberta-base", | |
| "rewards": { | |
| "BiEncoderSimilarity": { | |
| "weight": 1, | |
| "model_id": "all-distilroberta-v1" | |
| }, | |
| "GaussianCR": { | |
| "weight": 1, | |
| "mean": 0.5, | |
| "std": 0.2 | |
| } | |
| } | |
| } | |