EXAONE-3.0-7.8B-Instruct-GGUF/README.md

license, license_name, license_link, model_name, base_model, inference, model_creator, model_type, pipeline_tag, quantized_by, language, tags

license	license_name	license_link	model_name	base_model	inference	model_creator	model_type	pipeline_tag	quantized_by	language	tags
other	exaone	LICENSE	EXAONE-3.0-7.8B-Instruct	LGAI-EXAONE/EXAONE-3.0-7.8B-Instruct	false	Meta Llama3	llama	text-generation	Second State Inc.	en ko	lg-ai exaone gguf

---

EXAONE-3.0-7.8B-Instruct-GGUF

Original Model

LGAI-EXAONE/EXAONE-3.0-7.8B-Instruct

Run with LlamaEdge

• LlamaEdge version: coming soon

• Prompt template

  • Prompt type: exaone-chat

  • Prompt string

    [|system|]system_prompt_text[|endofturn|]
    [|user|]user_1st_turn_text
    [|assistant|]assistant_1st_turn_text[|endofturn|]
    [|user|]user_2nd_turn_text
    [|assistant|]assistant_2nd_turn_text[|endofturn|]
    

• Context size: 4096

• Run as LlamaEdge service

  wasmedge --dir .:. --nn-preload default:GGML:AUTO:EXAONE-3.0-7.8B-Instruct-Q5_K_M.gguf \
      llama-api-server.wasm \
      --prompt-template exaone-chat \
      --ctx-size 4096 \
      --model-name EXAONE-3.0-7.8B-Instruct
  

• Run as LlamaEdge command app

  wasmedge --dir .:. --nn-preload default:GGML:AUTO:EXAONE-3.0-7.8B-Instruct-Q5_K_M.gguf \
    llama-chat.wasm \
    --prompt-template exaone-chat \
    --ctx-size 4096
  

Quantized GGUF Models

Name	Quant method	Bits	Size	Use case
EXAONE-3.0-7.8B-Instruct-Q2_K.gguf	Q2_K	2	3.05 GB	smallest, significant quality loss - not recommended for most purposes
EXAONE-3.0-7.8B-Instruct-Q3_K_L.gguf	Q3_K_L	3	4.19 GB	small, substantial quality loss
EXAONE-3.0-7.8B-Instruct-Q3_K_M.gguf	Q3_K_M	3	3.88 GB	very small, high quality loss
EXAONE-3.0-7.8B-Instruct-Q3_K_S.gguf	Q3_K_S	3	3.53 GB	very small, high quality loss
EXAONE-3.0-7.8B-Instruct-Q4_0.gguf	Q4_0	4	4.51 GB	legacy; small, very high quality loss - prefer using Q3_K_M
EXAONE-3.0-7.8B-Instruct-Q4_K_M.gguf	Q4_K_M	4	4.77 GB	medium, balanced quality - recommended
EXAONE-3.0-7.8B-Instruct-Q4_K_S.gguf	Q4_K_S	4	4.54 GB	small, greater quality loss
EXAONE-3.0-7.8B-Instruct-Q5_0.gguf	Q5_0	5	5.44 GB	legacy; medium, balanced quality - prefer using Q4_K_M
EXAONE-3.0-7.8B-Instruct-Q5_K_M.gguf	Q5_K_M	5	5.57 GB	large, very low quality loss - recommended
EXAONE-3.0-7.8B-Instruct-Q5_K_S.gguf	Q5_K_S	5	5.44 GB	large, low quality loss - recommended
EXAONE-3.0-7.8B-Instruct-Q6_K.gguf	Q6_K	6	6.42 GB	very large, extremely low quality loss
EXAONE-3.0-7.8B-Instruct-Q8_0.gguf	Q8_0	8	8.31 GB	very large, extremely low quality loss - not recommended
EXAONE-3.0-7.8B-Instruct-f16.gguf	f16	16	15.6 GB

Quantized with llama.cpp b3807.