| from pathlib import Path |
| import argparse |
| import json |
| import sys |
| sys.path.insert(0, str(Path(__file__).resolve().parents[1])) |
| from nexora.inference import HFBackend |
|
|
|
|
| def main(): |
| p = argparse.ArgumentParser() |
| p.add_argument("--model", default=".cache/Qwen3.5-0.8B") |
| args = p.parse_args() |
| backend = HFBackend(args.model, max_new_tokens=32) |
| messages = [{"role": "user", "content": "Describe a queue in one short sentence."}] |
| nonstream = backend.complete(messages) |
| chunks = list(backend.stream(messages)) |
| report = {"model": args.model, "prompt": messages[0]["content"], "nonstream": nonstream, "streamed": "".join(chunks), |
| "stream_matches_nonstream": nonstream == "".join(chunks), "chunks": len(chunks), **backend.last_metrics, |
| "limitations": "One greedy request, warm model, no concurrent load or audio"} |
| if not report["stream_matches_nonstream"]: |
| raise AssertionError("Streaming parity failed") |
| Path("reports/streaming.json").write_text(json.dumps(report, indent=2)) |
| print(json.dumps(report)) |
|
|
|
|
| if __name__ == "__main__": |
| main() |
|
|
