| { |
| "input": "data/doom_basic_v3/all.jsonl", |
| "output_dir": "runs/doom_basic_0.6b_v3_seed42", |
| "model": "Qwen/Qwen3-0.6B-Base", |
| "revision": "main", |
| "objective": "gold", |
| "set_head": "attention", |
| "steps": 120, |
| "head_steps": 12, |
| "batch_questions": 6, |
| "eval_every": 24, |
| "max_length": 512, |
| "seed": 42, |
| "backbone_lr": 2e-05, |
| "head_lr": 0.0002, |
| "device": "cpu", |
| "skip_native_baseline": true, |
| "freeze_backbone": false, |
| "balance_classes": true, |
| "resolved_model_revision": "da87bfb608c14b7cf20ba1ce41287e8de496c0cd", |
| "data_sha256": "beac0625de18382542a6f25c1a37e1ae01cc20815de241edcd76843fa4cec201", |
| "deps": { |
| "torch": "2.14.0", |
| "transformers": "5.17.0", |
| "safetensors": "0.8.0", |
| "numpy": "1.26.4" |
| }, |
| "gpu": "cpu", |
| "train_questions": 1501, |
| "all_train_questions": 1501, |
| "train_class_counts": { |
| "('choice', 'shoot')": 779, |
| "('choice', 'right')": 330, |
| "('choice', 'left')": 392 |
| }, |
| "parameter_storage": "float32", |
| "forward_autocast": "fp32", |
| "teacher_target": "one-hot gold label", |
| "parallelism": "flat complete candidate paths in one backbone forward; no prefix sharing", |
| "parameter_count": 596250498 |
| } |
|
|
