CoolFace
Modelpublic

second-state/StarCoder2-15B-GGUF

sourceHugging Facebigcode-openrail-mupdated 3y agoView on Hugging Face
32likes3.7kdownloads
Model Card

<!-- header start --> <!-- 200823 --> <div style="width: auto; margin-left: auto; margin-right: auto"> <img src="https://github.com/LlamaEdge/LlamaEdge/raw/dev/assets/logo.svg" style="width: 100%; min-width: 400px; display: block; margin: auto;"> </div> <hr style="margin-top: 1.0em; margin-bottom: 1.0em;"> <!-- header end -->

StarCoder2-15B-GGUF

Original Model

bigcode/starcoder2-15b

Run with LlamaEdge

  • —LlamaEdge version: coming soon
  • —Context size: 6144

Quantized GGUF Models

NameQuant methodBitsSizeUse case
starcoder2-15b-Q2_K.ggufQ2_K26.19 GBsmallest, significant quality loss - not recommended for most purposes
starcoder2-15b-Q3_K_L.ggufQ3KL38.97 GBsmall, substantial quality loss
starcoder2-15b-Q3_K_M.ggufQ3KM38.1 GBvery small, high quality loss
starcoder2-15b-Q3_K_S.ggufQ3KS36.99 GBvery small, high quality loss
starcoder2-15b-Q4_0.ggufQ4_049.07 GBlegacy; small, very high quality loss - prefer using Q3KM
starcoder2-15b-Q4_K_M.ggufQ4KM49.86 GBmedium, balanced quality - recommended
starcoder2-15b-Q4_K_S.ggufQ4KS49.26 GBsmall, greater quality loss
starcoder2-15b-Q5_0.ggufQ5_0511 GBlegacy; medium, balanced quality - prefer using Q4KM
starcoder2-15b-Q5_K_M.ggufQ5KM511.4 GBlarge, very low quality loss - recommended
starcoder2-15b-Q5_K_S.ggufQ5KS511 GBlarge, low quality loss - recommended
starcoder2-15b-Q6_K.ggufQ6_K613.1 GBvery large, extremely low quality loss
starcoder2-15b-Q8_0.ggufQ8_0817 GBvery large, extremely low quality loss - not recommended

Quantized with llama.cpp b2308