Konuktor92's picture
Add full training code (open source) for model v1
d7c810d verified
Raw History Blame
2.79 kB
"""Verify what is on Hugging Face from a clean machine (no access to the Vast box).
uv run --with xgboost --with pandas --with pyarrow --with huggingface_hub python scripts/60_verify_hf.py
1. Downloads TrendeX90/expected-greenness at tag v1, runs the golden test with its own predict.py.
2. Reads naryn-cube: manifest, cells.parquet, one year of monthly data, products summary.
"""
from __future__ import annotations
import importlib.util
import json
import sys
import tempfile
from pathlib import Path
import numpy as np
import pandas as pd
from huggingface_hub import HfApi, hf_hub_download, snapshot_download
MODEL, CUBE = "TrendeX90/expected-greenness", "TrendeX90/naryn-cube"
def main() -> None:
api = HfApi()
tags = [t.name for t in api.list_repo_refs(MODEL).tags]
print("model tags:", tags)
with tempfile.TemporaryDirectory() as tmp:
root = Path(snapshot_download(MODEL, revision="v1", local_dir=Path(tmp) / "model"))
print("model files:", sorted(p.name for p in root.iterdir() if not p.name.startswith(".")))
spec = importlib.util.spec_from_file_location("predict", root / "predict.py")
mod = importlib.util.module_from_spec(spec)
sys.modules["predict"] = mod
spec.loader.exec_module(mod)
m = mod.load(root)
got = m.predict(pd.read_parquet(root / "golden" / "inputs.parquet"))
want = pd.read_parquet(root / "golden" / "expected.parquet")
diff = float(np.max(np.abs(got.values - want.values)))
print(f"golden test on this machine: max diff {diff:.2e} -> {'PASS' if diff < 1e-6 else 'FAIL'}")
metrics = json.loads((root / "metrics.json").read_text())
print("gate:", metrics["gate"])
files = api.list_repo_files(CUBE, repo_type="dataset")
print(f"cube files: {len(files)} | monthly years: {sum(f.startswith('monthly/') for f in files)} | "
f"products: {sorted(f for f in files if f.startswith('products/'))}")
man = json.loads(Path(hf_hub_download(CUBE, "manifest.json", repo_type="dataset")).read_text())
print("manifest: synthetic =", man["synthetic"], "| partial =", man["partial"], "| sources =", man["sources_used"])
cells = pd.read_parquet(hf_hub_download(CUBE, "cells.parquet", repo_type="dataset"))
print("cells:", len(cells), "| in oblast:", int(cells["in_oblast"].sum()))
y = pd.read_parquet(hf_hub_download(CUBE, "monthly/year=2024/part.parquet", repo_type="dataset"))
print("monthly 2024 rows:", len(y), "| columns:", list(y.columns))
summ = json.loads(Path(hf_hub_download(CUBE, "products/products_summary.json", repo_type="dataset")).read_text())
print("residual trend shares:", summ["residual_trend_cell_class_shares"])
print("VERIFY_DONE")
if __name__ == "__main__":
main()