| { | |
| "vocab_size": 49152, | |
| "ctx": 512, | |
| "d_model": 512, | |
| "n_heads": 8, | |
| "n_layers": 12, | |
| "d_ff": 2048, | |
| "dropout": 0.1, | |
| "tokens_seen": 24576000, | |
| "step": 4000 | |
| } |
| { | |
| "vocab_size": 49152, | |
| "ctx": 512, | |
| "d_model": 512, | |
| "n_heads": 8, | |
| "n_layers": 12, | |
| "d_ff": 2048, | |
| "dropout": 0.1, | |
| "tokens_seen": 24576000, | |
| "step": 4000 | |
| } |