---

Gemma-2-2b-it-GGUF

Original Model

google/gemma-2-2b-it

Run with LlamaEdge

• LlamaEdge version: v0.13.0 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: 2048

• Run as LlamaEdge service

  wasmedge --dir .:. --nn-preload default:GGML:AUTO:gemma-2-2b-it-Q5_K_M.gguf \
    llama-api-server.wasm \
    --prompt-template gemma-instruct \
    --ctx-size 2048 \
    --model-name gemma-2-2b
  

• Run as LlamaEdge command app

  wasmedge --dir .:. \
    --nn-preload default:GGML:AUTO:gemma-2-2b-it-Q5_K_M.gguf \
    llama-chat.wasm \
    --prompt-template gemma-instruct \
    --ctx-size 2048
  

Quantized GGUF Models

Name	Quant method	Bits	Size	Use case
gemma-2-2b-it-Q2_K.gguf	Q2_K	2	1.23 GB	smallest, significant quality loss - not recommended for most purposes
gemma-2-2b-it-Q3_K_L.gguf	Q3_K_L	3	1.55 GB	small, substantial quality loss
gemma-2-2b-it-Q3_K_M.gguf	Q3_K_M	3	1.46 GB	very small, high quality loss
gemma-2-2b-it-Q3_K_S.gguf	Q3_K_S	3	1.36 GB	very small, high quality loss
gemma-2-2b-it-Q4_0.gguf	Q4_0	4	1.63 GB	legacy; small, very high quality loss - prefer using Q3_K_M
gemma-2-2b-it-Q4_K_M.gguf	Q4_K_M	4	1.71 GB	medium, balanced quality - recommended
gemma-2-2b-it-Q4_K_S.gguf	Q4_K_S	4	1.64 GB	small, greater quality loss
gemma-2-2b-it-Q5_0.gguf	Q5_0	5	1.88 GB	legacy; medium, balanced quality - prefer using Q4_K_M
gemma-2-2b-it-Q5_K_M.gguf	Q5_K_M	5	1.92 GB	large, very low quality loss - recommended
gemma-2-2b-it-Q5_K_S.gguf	Q5_K_S	5	1.88 GB	large, low quality loss - recommended
gemma-2-2b-it-Q6_K.gguf	Q6_K	6	2.15 GB	very large, extremely low quality loss
gemma-2-2b-it-Q8_0.gguf	Q8_0	8	2.78 GB	very large, extremely low quality loss - not recommended
gemma-2-2b-it-f16.gguf	f16	16	5.24 GB

Quantized with llama.cpp b3499