| { |
| "format": "d1a-torch", |
| "format_version": 1, |
| "base": "google/gemma-4-E2B-it", |
| "base_revision": "3e22461f65e89153144f8adb70e3b8c2cc9845a7", |
| "lora": 16, |
| "head_dim": 256, |
| "option_isolation": false, |
| "special_embeddings": false, |
| "weights_dtype": "bf16", |
| "temperature": 1.4142135623730947, |
| "holdout": [], |
| "weights": "lora", |
| "head": "head.safetensors", |
| "extra": { |
| "args": { |
| "base": "google/gemma-4-E2B-it", |
| "n_per_source": 1000, |
| "epochs": 1, |
| "lr": 5e-05, |
| "head_lr": 0.0, |
| "weight_decay": 0.01, |
| "lora": 16, |
| "accum": 4, |
| "holdout": "", |
| "suite": "evals/v7/decision-v7", |
| "train_sources": "", |
| "device": "cuda", |
| "batch": 2, |
| "dtype": "bf16", |
| "weights_dtype": "bf16", |
| "checkpointing": 1, |
| "head_dim": 256, |
| "lora_targets": "all", |
| "base_revision": "3e22461f65e89153144f8adb70e3b8c2cc9845a7", |
| "p_none": 0.1, |
| "p_none_distract": 0.12, |
| "p_distract": 0.15, |
| "p_none_pair": 0.0, |
| "none_pair_max_state": null, |
| "synthetic_repeat": 1, |
| "public_frac": 1.0, |
| "out": "e2b-it/stage2", |
| "data": "data/train-e2b-v06.jsonl", |
| "max_state": 5120, |
| "extra_suites": "", |
| "replay": 0, |
| "init_from": "e2b-it/stage1", |
| "allow_missing_sources": "", |
| "reason": "", |
| "row_budget": 0, |
| "shared_prefix": 0, |
| "length_sort": 0, |
| "pass_tokens_max": 0, |
| "max_steps": 0, |
| "save_every_steps": 0, |
| "save_every_minutes": 30.0, |
| "resume": 1, |
| "stop_after": 0, |
| "seed": 0 |
| }, |
| "suite_sha256": "a8f50e481b7d90b97da049e0ff6a01cee2f1ed204aed61a8265af0edbb5514d2", |
| "init_source": { |
| "init_from": "e2b-it/stage1", |
| "resolved": "e2b-it/stage1", |
| "weights_sha256": "ce9d3ac952af9c1045cea80e981bd0612b3efb9c199a308616886f77946bdcb7", |
| "head_sha256": "d4e3e8d0f07ee1b24c7fcca957052d91ea391b92ffb1aeddc88f838c70581806", |
| "tensors": 410 |
| }, |
| "temperature_fit": { |
| "rows": [ |
| "cal/rows.json", |
| "calpr/rows.json" |
| ], |
| "n": 3178, |
| "method": "min micro mean NLL over a 121-point log grid 0.25..4", |
| "cross_validation": { |
| "raw": { |
| "n": 3178, |
| "ece": 0.05269321644750151, |
| "brier": 0.26774269342014495, |
| "nll": 0.5260661192849185, |
| "confident_error_rate": 0.030207677784770296, |
| "coverage_at_5pct_error": 0.6343612334801763 |
| }, |
| "out_of_fold": { |
| "n": 3178, |
| "ece": 0.020409886004775658, |
| "brier": 0.2628816980431792, |
| "nll": 0.49639497637876545, |
| "confident_error_rate": 0.01762114537444934, |
| "coverage_at_5pct_error": 0.6261799874134676 |
| }, |
| "ece_ci95": { |
| "raw": [ |
| 0.0412701370712576, |
| 0.06460168040334786 |
| ], |
| "out_of_fold": [ |
| 0.014810248879595556, |
| 0.03501381041208449 |
| ], |
| "delta": [ |
| -0.040734042372353514, |
| -0.01574161986134397 |
| ] |
| }, |
| "separated": true, |
| "temperatures": [ |
| 1.4142135623730947, |
| 1.447269237440378, |
| 1.381912879967776, |
| 1.447269237440378, |
| 1.447269237440378 |
| ], |
| "folds": 5, |
| "seed": 0, |
| "samples": 1000, |
| "groups": 1670, |
| "unit": "source-stratified (source, group); sibling questions and variants stay together" |
| }, |
| "interval": { |
| "level": 0.9, |
| "samples": 2000, |
| "seed": 0, |
| "lower": 1.35034994616819, |
| "upper": 1.4810975522865641, |
| "questions": 3178 |
| }, |
| "fit_rows": [ |
| { |
| "rows": "cal/rows.json", |
| "suite": "evals/v7/decision-v7", |
| "split": "calibration", |
| "questions": 1148 |
| }, |
| { |
| "rows": "calpr/rows.json", |
| "suite": "evals/d1a/pr-labels", |
| "split": "development", |
| "questions": 2030 |
| } |
| ], |
| "training_suites": [ |
| "evals/devtools-v1", |
| "evals/documents-v1", |
| "evals/hard-v1", |
| "evals/v7/decision-v7" |
| ], |
| "training_partitions": [ |
| "evals/d1a/ja-jglue:train", |
| "evals/d1a/pr-labels:train", |
| "evals/d1a/pr-labels:train-blast", |
| "evals/d1a/pr-labels:train-ja", |
| "evals/d1a/routing:factory-train", |
| "evals/d1a/routing:generic-train" |
| ], |
| "same_corpus": [ |
| "calpr/rows.json: SAME CORPUS: evals/d1a/pr-labels:development is an eval partition of evals/d1a/pr-labels, whose train partitions ['evals/d1a/pr-labels:train', 'evals/d1a/pr-labels:train-blast', 'evals/d1a/pr-labels:train-ja'] the checkpoint trained on" |
| ], |
| "in_distribution": { |
| "allowed": true, |
| "problems": [ |
| "cal/rows.json: evals/v7/decision-v7 is training data of the checkpoint", |
| "cal/rows.json: it pools 12 training source(s) ['agnews', 'amazon', 'banking77', 'boolq', 'compositional', 'dbpedia14', 'imdb', 'legacy_policy'] ...", |
| "cal/rows.json: it reads the calibration partition of evals/v7/decision-v7, a training corpus", |
| "calpr/rows.json: SAME CORPUS: evals/d1a/pr-labels:development is an eval partition of evals/d1a/pr-labels, whose train partitions ['evals/d1a/pr-labels:train', 'evals/d1a/pr-labels:train-blast', 'evals/d1a/pr-labels:train-ja'] the checkpoint trained on", |
| "cannot list the sources of --data data/train-e2b-v06.jsonl (outside evals/, no sidecar train-e2b-v06.jsonl.json naming its sources), training data of d1" |
| ] |
| } |
| } |
| } |
| } |
|
|
