| { | |
| "architectures": [ | |
| "SnowflakeCoreG1" | |
| ], | |
| "model_type": "snowflake_core", | |
| "vocab_size": 50257, | |
| "embed_dim": 1024, | |
| "num_heads": 16, | |
| "num_layers": 24, | |
| "max_length": 2048, | |
| "ffn_dim": 4096, | |
| "pad_token_id": 50256, | |
| "eos_token_id": 50256, | |
| "bos_token_id": 50256, | |
| "unk_token_id": 50256, | |
| "auto_map": { | |
| "AutoConfig": "configuration_snowflake_core.SnowflakeCoreConfig", | |
| "AutoModelForCausalLM": "modeling_snowflake_core.SnowflakeCoreG1" | |
| }, | |
| "training_config": { | |
| "epochs": 5, | |
| "actual_epochs": 5, | |
| "batch_size": 1, | |
| "learning_rate": 0.0002, | |
| "grad_accum_steps": 32, | |
| "max_length": 2048, | |
| "val_split_ratio": 0.1, | |
| "early_stopping": { | |
| "patience": 3, | |
| "min_delta": 0.001, | |
| "triggered": false | |
| } | |
| }, | |
| "training_metrics": { | |
| "final_train_loss": null, | |
| "final_train_perplexity": null, | |
| "final_val_loss": null, | |
| "final_val_perplexity": null, | |
| "best_val_loss": Infinity, | |
| "best_val_perplexity": null | |
| } | |
| } | |