Add result: Q2_K on Qwen2.5-1.5B-Instruct@e2cc88932f748bcb

#6

Result for Q2_K (hf://Qwen/Qwen2.5-1.5B-Instruct-GGUF/qwen2.5-1.5b-instruct-q2_k.gguf)

  • pack: Qwen2.5-1.5B-Instruct@e2cc88932f748bcb (id e2cc88932f748bcb)
  • quant file SHA-256: 5ede348e91ce1e7a330926ec5b202c27b864d065149dc463257fde1f98865b3a
  • backend: llamacpp (llama-cpp-python 0.3.36)
  • hardware: Tesla T4 / CUDA (Linux)
  • KLD mean: 0.60872 (95% CI 0.55965 to 0.66023)
  • top-1 agreement: 64.87%
  • approximate (server top-N): no
  • loops: 38 (reference 16), wrong script: 0, valid JSON: 8/8 (reference 8)

To re-check: quantgap score <same file> --pack <same pack> should give a KLD mean within 2%.

Submitted with quantgap 0.1.1.

Merged by maintainer: Kaggle T4 full-mode run, sha256 verified against plan.

vanishingradient changed pull request status to merged

Sign up or log in to comment