File size: 1,518 Bytes
5a7104f | 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 | """Next Signal inference (NumPy only): embed inertial windows and classify them with the released probe.
Usage:
python inference.py --selftest
from inference import load, embed, classify
"""
import json
import os
import sys
HERE = os.path.dirname(os.path.abspath(__file__))
sys.path.insert(0, HERE)
import numpy as np # noqa: E402
import signal_core as sg # noqa: E402
from nlrt import load_model # noqa: E402
def load(path=os.path.join(HERE, "model.json")):
return load_model(path)
def embed(spec, windows):
"""windows: (N, 9, 128) raw UCI-HAR style inertial windows -> (N, 64) self-supervised embeddings."""
m = spec["meta"]
return sg.encode(spec["nets"]["encoder"], sg.normalise(windows, m["mean"], m["std"]))
def classify(spec, windows, budget=50):
"""Activity probabilities (N, 6) from the frozen encoder + linear probe trained with `budget` labels per class."""
return sg.probe_proba(spec["meta"]["probes"]["ssl"][str(budget)], embed(spec, windows))
def _selftest():
spec = load()
with open(os.path.join(HERE, "selftest.json"), encoding="utf-8") as f:
t = json.load(f)
P = classify(spec, np.asarray(t["windows"]), t["budget"])
diff = float(np.abs(P - np.asarray(t["expected_proba"])).max())
ok = diff < 1e-6
print(json.dumps({"passed": ok, "max_abs_diff": diff, "n_windows": len(t["windows"])}))
return ok
if __name__ == "__main__":
if "--selftest" in sys.argv:
sys.exit(0 if _selftest() else 1)
print(__doc__)
|