CoolFace
Modelpublic

second-state/StarCoder2-3B-GGUF

sourceHugging Facebigcode-openrail-mupdated 3y agoView on Hugging Face
20likes2.1kdownloads
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 -->

StarCoder2-3B-GGUF

Original Model

bigcode/starcoder2-3b

Run with LlamaEdge

  • —LlamaEdge version: coming soon
  • —Context size: 3072

Quantized GGUF Models

NameQuant methodBitsSizeUse case
starcoder2-3b-Q2_K.ggufQ2_K21.15 GBsmallest, significant quality loss - not recommended for most purposes
starcoder2-3b-Q3_K_L.ggufQ3KL31.68 GBsmall, substantial quality loss
starcoder2-3b-Q3_K_M.ggufQ3KM31.51 GBvery small, high quality loss
starcoder2-3b-Q3_K_S.ggufQ3KS31.31 GBvery small, high quality loss
starcoder2-3b-Q4_0.ggufQ4_041.71 GBlegacy; small, very high quality loss - prefer using Q3KM
starcoder2-3b-Q4_K_M.ggufQ4KM41.85 GBmedium, balanced quality - recommended
starcoder2-3b-Q4_K_S.ggufQ4KS41.74 GBsmall, greater quality loss
starcoder2-3b-Q5_0.ggufQ5_052.09 GBlegacy; medium, balanced quality - prefer using Q4KM
starcoder2-3b-Q5_K_M.ggufQ5KM52.16 GBlarge, very low quality loss - recommended
starcoder2-3b-Q5_K_S.ggufQ5KS52.09 GBlarge, low quality loss - recommended
starcoder2-3b-Q6_K.ggufQ6_K62.49 GBvery large, extremely low quality loss
starcoder2-3b-Q8_0.ggufQ8_083.22 GBvery large, extremely low quality loss - not recommended

Quantized with llama.cpp b2308