File size: 3,006 Bytes
0ef4ba2 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 | """Hidden-family refinement with known-family replay for conditional DENIM v2."""
from __future__ import annotations
import copy
import json
from collections import defaultdict
import torch
from src.t2_denim_conditional import ConditionalDENIM
from src.train_t2_denim_conditional_v2 import (
ARTIFACT_DIR,
MODEL_DIR,
_train_stage,
evaluate,
load_all_trajectories,
transitions,
)
def run(*, steps: int = 1200) -> dict[str, object]:
trajectories = load_all_trajectories()
hidden_name = "hidden_three_memory_tabulated_hardening"
known_train = [x for x in trajectories if x.material != hidden_name and x.role == "train"]
known_validation = [x for x in trajectories if x.material != hidden_name and x.role == "validation"]
hidden_train = [x for x in trajectories if x.material == hidden_name and x.role == "train"]
hidden_validation = [x for x in trajectories if x.material == hidden_name and x.role == "validation"]
data = {
"known_train": transitions(known_train),
"known_validation": transitions(known_validation),
"hidden_train": transitions(hidden_train),
"hidden_validation": transitions(hidden_validation),
}
checkpoint = torch.load(MODEL_DIR / "denim_conditional_v2.pt", map_location="cpu", weights_only=False)
model = ConditionalDENIM()
model.load_state_dict(checkpoint["state_dict"])
refined = copy.deepcopy(model)
training = _train_stage(
refined,
[data["known_train"], data["hidden_train"], data["hidden_train"], data["hidden_train"]],
[data["known_validation"], data["hidden_validation"], data["hidden_validation"], data["hidden_validation"]],
steps=steps,
seed=20260930,
learning_rate=5.0e-4,
)
groups: dict[str, list] = defaultdict(list)
for trajectory in trajectories:
groups[trajectory.evaluation].append(trajectory)
selected = (
"hidden_frozen_test",
"test_path_ood",
"test_amplitude_ood",
"test_long_horizon",
"test_discretization",
)
metrics: dict[str, object] = {
"training": training,
"evaluation": {name: evaluate(refined, groups[name]) for name in selected},
}
known_test = [x for x in trajectories if x.material != hidden_name and x.role == "test"]
metrics["evaluation"]["known_strict_test_all"] = evaluate(refined, known_test)
torch.save(
{
"state_dict": refined.state_dict(),
"architecture": "ConditionalDENIM(channels=2, embedding=32, hidden=48)",
"stage": "hidden-focused refinement with known-family replay",
},
MODEL_DIR / "denim_conditional_v2_refined.pt",
)
ARTIFACT_DIR.mkdir(parents=True, exist_ok=True)
(ARTIFACT_DIR / "refined_metrics.json").write_text(
json.dumps(metrics, indent=2) + "\n", encoding="utf-8"
)
print(json.dumps(metrics["evaluation"], indent=2))
return metrics
if __name__ == "__main__":
run()
|