Add result: Q2_K on Qwen3-0.6B@282f68b573fc832a

#2

Result for Q2_K (hf://tensorblock/Qwen_Qwen3-0.6B-GGUF/Qwen3-0.6B-Q2_K.gguf)

  • pack: Qwen3-0.6B@282f68b573fc832a (id 282f68b573fc832a)
  • quant file SHA-256: d82888bc9ce32781a98e3bf5acd5472917a9ee075d41520698662bf353dcdb73
  • backend: llamacpp (llama-cpp-python 0.3.36)
  • hardware: Tesla T4 / CUDA (Linux)
  • KLD mean: 2.45535 (95% CI 2.37365 to 2.54312)
  • top-1 agreement: 33.84%
  • approximate (server top-N): no
  • loops: 48 (reference 32), wrong script: 0, valid JSON: 0/8 (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.

vanishingradient changed pull request status to merged

Sign up or log in to comment