metadata

base_model: deepseek-ai/DeepSeek-Coder-V2-Instruct
inference: false
license: other
license_name: deepseek-license
license_link: LICENSE
model_creator: DeepSeek-ai
model_name: DeepSeek-Coder-V2-Instruct
model_type: deepseek
quantized_by: Second State Inc.

DeepSeek-Coder-V2-Instruct-GGUF

Original Model

deepseek-ai/DeepSeek-Coder-V2-Instruct

Run with LlamaEdge

LlamaEdge version: v0.12.3 and above

Prompt template

Prompt type: deepseek-chat-2

Prompt string

<|begin_of_sentence|>{system_message}

User: {user_message_1}

Assistant: {assistant_message_1}<|end_of_sentence|>User: {user_message_2}

Assistant:

Context size: 128000

Run as LlamaEdge service

wasmedge --dir .:. \
  --nn-preload default:GGML:AUTO:DeepSeek-Coder-V2-Instruct-Q5_K_M.gguf \
  llama-api-server.wasm \
  --prompt-template deepseek-chat-2 \
  --ctx-size 128000 \
  --model-name DeepSeek-Coder-V2-Instruct

Run as LlamaEdge command app

wasmedge --dir .:. \
  --nn-preload default:GGML:AUTO:DeepSeek-Coder-V2-Instruct-Q5_K_M.gguf \
  llama-chat.wasm \
  --prompt-template deepseek-chat-2 \
  --ctx-size 128000

Quantized GGUF Models

Name	Quant method	Bits	Size	Use case
DeepSeek-Coder-V2-Instruct-Q2_K-00001-of-00003.gguf	Q2_K	2	29.7 GB	smallest, significant quality loss - not recommended for most purposes
DeepSeek-Coder-V2-Instruct-Q2_K-00002-of-00003.gguf	Q2_K	2	29.8 GB	smallest, significant quality loss - not recommended for most purposes
DeepSeek-Coder-V2-Instruct-Q2_K-00002-of-00003.gguf	Q2_K	2	26.4 GB	smallest, significant quality loss - not recommended for most purposes
DeepSeek-Coder-V2-Instruct-Q3_K_L-00001-of-00005.gguf	Q3_K_L	3	29.8 GB	small, substantial quality loss
DeepSeek-Coder-V2-Instruct-Q3_K_L-00002-of-00005.gguf	Q3_K_L	3	29.7 GB	small, substantial quality loss
DeepSeek-Coder-V2-Instruct-Q3_K_L-00003-of-00005.gguf	Q3_K_L	3	29.9 GB	small, substantial quality loss
DeepSeek-Coder-V2-Instruct-Q3_K_L-00004-of-00005.gguf	Q3_K_L	3	29.8 GB	small, substantial quality loss
DeepSeek-Coder-V2-Instruct-Q3_K_L-00005-of-00005.gguf	Q3_K_L	3	3.14 GB	small, substantial quality loss
DeepSeek-Coder-V2-Instruct-Q3_K_M-00001-of-00004.gguf	Q3_K_M	3	29.6 GB	very small, high quality loss
DeepSeek-Coder-V2-Instruct-Q3_K_M-00002-of-00004.gguf	Q3_K_M	3	29.5 GB	very small, high quality loss
DeepSeek-Coder-V2-Instruct-Q3_K_M-00003-of-00004.gguf	Q3_K_M	3	29.8 GB	very small, high quality loss
DeepSeek-Coder-V2-Instruct-Q3_K_M-00004-of-00004.gguf	Q3_K_M	3	23.8 GB	very small, high quality loss
DeepSeek-Coder-V2-Instruct-Q3_K_S.gguf	Q3_K_S	3	7.49 GB	very small, high quality loss
DeepSeek-Coder-V2-Instruct-Q4_0.gguf	Q4_0	4	8.91 GB	legacy; small, very high quality loss - prefer using Q3_K_M
DeepSeek-Coder-V2-Instruct-Q4_K_M.gguf	Q4_K_M	4	10.4 GB	medium, balanced quality - recommended
DeepSeek-Coder-V2-Instruct-Q4_K_S.gguf	Q4_K_S	4	9.53 GB	small, greater quality loss
DeepSeek-Coder-V2-Instruct-Q5_0.gguf	Q5_0	5	10.8 GB	legacy; medium, balanced quality - prefer using Q4_K_M
DeepSeek-Coder-V2-Instruct-Q5_K_M.gguf	Q5_K_M	5	11.9 GB	large, very low quality loss - recommended
DeepSeek-Coder-V2-Instruct-Q5_K_S.gguf	Q5_K_S	5	11.1 GB	large, low quality loss - recommended
DeepSeek-Coder-V2-Instruct-Q6_K.gguf	Q6_K	6	14.1 GB	very large, extremely low quality loss
DeepSeek-Coder-V2-Instruct-Q8_0.gguf	Q8_0	8	16.7 GB	very large, extremely low quality loss - not recommended
DeepSeek-Coder-V2-Instruct-f16.gguf	f16	16	31.4 GB

Quatized with llama.cpp b3499