SmoothQuant Qwen/Qwen2.5-1.5B-Instruct
Quantized derivative of Qwen/Qwen2.5-1.5B-Instruct.
Not Hub fp16. Weights are the packed/runtime artifact from job 20260920T220639Z-9e4555.
WikiText-2 (test, 2048-token windows, 65504 tokens)
Measured on NVIDIA A10G (g5.2xlarge) against the original fp16 snapshot.
Numbers copied from jobs/20260920T220639Z-9e4555/benchmark.json.
| Quantized | fp16 snapshot | |
|---|---|---|
| Perplexity | 10.066860 | 9.579098 |
| NLL loss | 2.309249 | 2.259583 |
| Tokens/s (2048 prefill, after warmup) | 10427.5 | 11103.1 |
| Peak VRAM (GB) | 4.682 | 5.844 |
ppl_ratio: 1.050919 (quality_ok=True)improved_throughput: Falseimproved_vram: True
Pull
from huggingface_hub import snapshot_download
path = snapshot_download("chris320211/smoothquant-qwen25-15b-g52xlarge")
This is not a drop-in AutoModelForCausalLM.from_pretrained checkpoint.
Reload with the included quant_agent_inference_adapter.py after the method
repo (and overlay, if any) is on QUANT_AGENT_METHOD_REPO. See
quantization_config.json.
License
Base model license (MIT for Phi-3) plus the method repository license.
Keep LICENSE / NOTICE.md from the snapshot when present.