| { | |
| "model": {"vocab_size": 259, "hidden_size": 128, "layers": 4, "heads": 4, "kv_heads": 2, "intermediate_size": 384, "max_context": 256}, | |
| "training": {"steps": 120, "batch_size": 8, "sequence_length": 128, "learning_rate": 0.0005, "seed": 42, "eval_every": 20, "checkpoint_every": 40, "device": "auto", "threads": 4} | |
| } | |