v4: further training on synthetic decision data (Decision Index 0.3 public index, self-scored)
80aca9d verified | { | |
| "head": { | |
| "hidden_size": 4096, | |
| "n_taps": 2, | |
| "d": 1024, | |
| "heads": 16, | |
| "ffn": 2816, | |
| "route_blocks": 2, | |
| "interact_blocks": 2, | |
| "dropout": 0.1, | |
| "n_roles": 5, | |
| "ord_hidden": 256, | |
| "max_nodes": 4096, | |
| "grad_checkpoint": true, | |
| "taps": [ | |
| "final", | |
| 25 | |
| ], | |
| "template": "schema-v1" | |
| }, | |
| "taps": [ | |
| "final", | |
| 25 | |
| ], | |
| "template": "schema-v1-nosys", | |
| "step": 1040, | |
| "lora_rank": 64, | |
| "snapshot": "/scratch/schema-head/models/StandardOne-8B-SH", | |
| "preview": true | |
| } |