CodeGemma-7b-it-GGUF/README.md

base_model, inference, library_name, license, model_creator, model_name, quantized_by, pipeline_tag

base_model	inference	library_name	license	model_creator	model_name	quantized_by	pipeline_tag
google/codegemma-7b-it	false	transformers	gemma	Google	CodeGemma 7B (IT)	Second State Inc.	text-generation

---

CodeGemma-7b-it-GGUF

Original Model

google/codegemma-7b-it

Run with LlamaEdge

• LlamaEdge version: v0.8.1 and above

• Prompt template

  • Prompt type: gemma-instruct

  • Prompt string

    <bos><start_of_turn>user
    {user_message}<end_of_turn>
    <start_of_turn>model
    {model_message}<end_of_turn>model
    

• Context size: 3072

• Run as LlamaEdge service

  wasmedge --dir .:. --nn-preload default:GGML:AUTO:codegemma-7b-it-Q5_K_M.gguf \
    llama-api-server.wasm \
    --prompt-template gemma-instruct \
    --ctx-size 3072 \
    --model-name codegemma-7b
  

Quantized GGUF Models

Name	Quant method	Bits	Size	Use case
codegemma-7b-it-Q2_K.gguf	Q2_K	2	3.48 GB	smallest, significant quality loss - not recommended for most purposes
codegemma-7b-it-Q3_K_L.gguf	Q3_K_L	3	4.71 GB	small, substantial quality loss
codegemma-7b-it-Q3_K_M.gguf	Q3_K_M	3	4.37 GB	very small, high quality loss
codegemma-7b-it-Q3_K_S.gguf	Q3_K_S	3	3.98 GB	very small, high quality loss
codegemma-7b-it-Q4_0.gguf	Q4_0	4	5.01 GB	legacy; small, very high quality loss - prefer using Q3_K_M
codegemma-7b-it-Q4_K_M.gguf	Q4_K_M	4	5.33 GB	medium, balanced quality - recommended
codegemma-7b-it-Q4_K_S.gguf	Q4_K_S	4	5.05 GB	small, greater quality loss
codegemma-7b-it-Q5_0.gguf	Q5_0	5	5.98 GB	legacy; medium, balanced quality - prefer using Q4_K_M
codegemma-7b-it-Q5_K_M.gguf	Q5_K_M	5	6.14 GB	large, very low quality loss - recommended
codegemma-7b-it-Q5_K_S.gguf	Q5_K_S	5	5.98 GB	large, low quality loss - recommended
codegemma-7b-it-Q6_K.gguf	Q6_K	6	7.01 GB	very large, extremely low quality loss
codegemma-7b-it-Q8_0.gguf	Q8_0	8	9.08 GB	very large, extremely low quality loss - not recommended
codegemma-7b-it-f16.gguf	f16	16	17.1 GB

Quantized with llama.cpp b2589