File size: 4,842 Bytes
784f572 | 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 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 | """Render validated teacher-adapter commands without invoking model code."""
from __future__ import annotations
import argparse
import json
from pathlib import Path
from dlb.adapters.base import AdapterError, BaseTeacherAdapter
from dlb.adapters.candi import CANDIAdapter
from dlb.adapters.di4c import Di4CAdapter
from dlb.adapters.duo import DuoAdapter
from dlb.adapters.flm import FLMAdapter
from dlb.adapters.langflow import LangFlowAdapter
from dlb.adapters.mdlm import MDLMAdapter
from dlb.adapters.rdlm import RDLMAdapter
from dlb.adapters.sdtt import SDTTAdapter
from dlb.registry import load_registry, step_grid_for_model
from dlb.runner import RunRequest
ADAPTERS: dict[str, BaseTeacherAdapter] = {
"flm": FLMAdapter(),
"fmlm": FLMAdapter(),
"duo": DuoAdapter(),
"duo_dcd": DuoAdapter(),
"mdlm": MDLMAdapter(),
"candi": CANDIAdapter(),
"langflow": LangFlowAdapter(),
"rdlm": RDLMAdapter(),
"mdlm_sdtt": SDTTAdapter(),
"mdlm_di4c": Di4CAdapter("mdlm"),
"duo_di4c": Di4CAdapter("duo"),
}
def _csv(value: str) -> list[str]:
values = value.split(",")
if not values or any(not item for item in values):
raise argparse.ArgumentTypeError("comma-separated values must not be empty")
return values
def _record(
root: Path,
model_id: str,
dataset_id: str,
step_count: int,
sample_count: int,
seed: int,
dry_run: bool,
) -> dict[str, object]:
registry = load_registry(root / "configs" / "experiments.yaml")
model = registry.models.get(model_id)
support = model.datasets.get(dataset_id) if model is not None else None
if model is None or support is None:
return {
"status": "unsupported",
"model": model_id,
"dataset": dataset_id,
"reason": "unknown model or dataset",
}
if support.status != "supported":
return {
"status": "unsupported",
"model": model_id,
"dataset": dataset_id,
"reason": support.reason,
}
allowed_steps = step_grid_for_model(registry, model_id)
if step_count not in allowed_steps:
joined_steps = ",".join(str(step) for step in allowed_steps)
return {
"status": "error",
"model": model_id,
"dataset": dataset_id,
"reason": (
f"invalid step count {step_count} for {model_id}/{dataset_id}; "
f"allowed: {joined_steps}"
),
}
adapter = ADAPTERS.get(model_id)
if adapter is None:
return {
"status": "unsupported",
"model": model_id,
"dataset": dataset_id,
"reason": "no command adapter is implemented for this model",
}
request = RunRequest(
run_id=f"{model_id}-{dataset_id}-steps-{step_count}",
model_id=model_id,
dataset_id=dataset_id,
step_count=step_count,
seed=seed,
sample_count=sample_count,
)
run_dir = (
root
/ "results"
/ "samples"
/ dataset_id
/ model_id
/ f"steps_{step_count}"
)
try:
command = adapter.render_command(request, run_dir, dry_run=dry_run)
except (AdapterError, OSError, ValueError) as error:
return {
"status": "error",
"model": model_id,
"dataset": dataset_id,
"reason": str(error),
}
return {
"status": "supported",
"model": model_id,
"dataset": dataset_id,
"steps": step_count,
"sample_count": sample_count,
"command": command,
}
def main(argv: list[str] | None = None) -> int:
parser = argparse.ArgumentParser(description=__doc__)
parser.add_argument("--root", type=Path, default=Path.cwd())
parser.add_argument("--models", type=_csv, required=True)
parser.add_argument("--datasets", type=_csv, required=True)
parser.add_argument("--steps", type=int, default=32)
parser.add_argument("--num-samples", type=int, default=1024)
parser.add_argument("--seed", type=int, default=42)
parser.add_argument("--dry-run", action="store_true")
arguments = parser.parse_args(argv)
root = arguments.root.resolve()
exit_code = 0
for model_id in arguments.models:
for dataset_id in arguments.datasets:
record = _record(
root,
model_id,
dataset_id,
arguments.steps,
arguments.num_samples,
arguments.seed,
arguments.dry_run,
)
if record["status"] == "error":
exit_code = 1
print(json.dumps(record, sort_keys=True))
return exit_code
if __name__ == "__main__":
raise SystemExit(main())
|