CoolFace
Modelpublic

second-state/Qwen2.5-Coder-32B-Instruct-GGUF

sourceHugging Faceapache-2.0updated 2y agoView on Hugging Face
0likes977downloads
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 -->

Qwen2.5-Coder-32B-Instruct-GGUF

Original Model

Qwen/Qwen2.5-Coder-32B-Instruct

Run with LlamaEdge

  • —Prompt template
  • —Prompt type: chatml
  • —Prompt string
text
    <|im_start|>system
    {system_message}<|im_end|>
    <|im_start|>user
    {prompt}<|im_end|>
    <|im_start|>assistant
  • —Context size: 128000
  • —Run as LlamaEdge service
bash
  wasmedge --dir .:. --nn-preload default:GGML:AUTO:Qwen2.5-Coder-32B-Instruct-Q5_K_M.gguf \
    llama-api-server.wasm \
    --model-name Qwen2.5-Coder-32B-Instruct \
    --prompt-template chatml \
    --ctx-size 128000
  • —Run as LlamaEdge command app
bash
  wasmedge --dir .:. --nn-preload default:GGML:AUTO:Qwen2.5-Coder-32B-Instruct-Q5_K_M.gguf \
    llama-chat.wasm \
    --prompt-template chatml \
    --ctx-size 128000

Quantized GGUF Models

NameQuant methodBitsSizeUse case
Qwen2.5-Coder-32B-Instruct-Q2_K.ggufQ2_K212.3 GBsmallest, significant quality loss - not recommended for most purposes
Qwen2.5-Coder-32B-Instruct-Q3_K_L.ggufQ3KL317.2 GBsmall, substantial quality loss
Qwen2.5-Coder-32B-Instruct-Q3_K_M.ggufQ3KM315.9 GBvery small, high quality loss
Qwen2.5-Coder-32B-Instruct-Q3_K_S.ggufQ3KS314.4 GBvery small, high quality loss
Qwen2.5-Coder-32B-Instruct-Q4_0.ggufQ4_0418.6 GBlegacy; small, very high quality loss - prefer using Q3KM
Qwen2.5-Coder-32B-Instruct-Q4_K_M.ggufQ4KM419.9 GBmedium, balanced quality - recommended
Qwen2.5-Coder-32B-Instruct-Q4_K_S.ggufQ4KS418.8 GBsmall, greater quality loss
Qwen2.5-Coder-32B-Instruct-Q5_0.ggufQ5_0522.6 GBlegacy; medium, balanced quality - prefer using Q4KM
Qwen2.5-Coder-32B-Instruct-Q5_K_M.ggufQ5KM523.3 GBlarge, very low quality loss - recommended
Qwen2.5-Coder-32B-Instruct-Q5_K_S.ggufQ5KS522.6 GBlarge, low quality loss - recommended
Qwen2.5-Coder-32B-Instruct-Q6_K.ggufQ6_K626.9 GBvery large, extremely low quality loss
Qwen2.5-Coder-32B-Instruct-Q8_0.ggufQ8_0834.8 GBvery large, extremely low quality loss - not recommended
Qwen2.5-Coder-32B-Instruct-f16-00001-of-00003.gguff161629.8 GB
Qwen2.5-Coder-32B-Instruct-f16-00002-of-00003.gguff161629.8 GB
Qwen2.5-Coder-32B-Instruct-f16-00003-of-00003.gguff16165.87 GB

Quantized with llama.cpp b4033