Edit model card

Mistral-Nemo-Instruct-2407-GGUF

Original Model

mistralai/Mistral-Nemo-Instruct-2407

Run with LlamaEdge

  • LlamaEdge version: v0.12.4

  • Prompt template

    • Prompt type: mistral-instruct

    • Prompt string

      <s>[INST] {user_message_1} [/INST]{assistant_message_1}</s>[INST] {user_message_2} [/INST]{assistant_message_2}</s>
      
  • Context size: 128000

  • Run as LlamaEdge service

    wasmedge --dir .:. --nn-preload default:GGML:AUTO:Mistral-Nemo-Instruct-2407-Q5_K_M.gguf \
      llama-api-server.wasm \
      --prompt-template mistral-instruct \
      --ctx-size 128000 \
      --model-name Mistral-Nemo-Instruct-2407
    
  • Run as LlamaEdge command app

    wasmedge --dir .:. --nn-preload default:GGML:AUTO:Mistral-Nemo-Instruct-2407-Q5_K_M.gguf \
      llama-chat.wasm \
      --prompt-template mistral-instruct \
      --ctx-size 128000
    

Quantized GGUF Models

Name Quant method Bits Size Use case
Mistral-Nemo-Instruct-2407-Q2_K.gguf Q2_K 2 4.79 GB smallest, significant quality loss - not recommended for most purposes
Mistral-Nemo-Instruct-2407-Q3_K_L.gguf Q3_K_L 3 6.56 GB small, substantial quality loss
Mistral-Nemo-Instruct-2407-Q3_K_M.gguf Q3_K_M 3 6.08 GB very small, high quality loss
Mistral-Nemo-Instruct-2407-Q3_K_S.gguf Q3_K_S 3 5.53 GB very small, high quality loss
Mistral-Nemo-Instruct-2407-Q4_0.gguf Q4_0 4 7.07 GB legacy; small, very high quality loss - prefer using Q3_K_M
Mistral-Nemo-Instruct-2407-Q4_K_M.gguf Q4_K_M 4 7.48 GB medium, balanced quality - recommended
Mistral-Nemo-Instruct-2407-Q4_K_S.gguf Q4_K_S 4 7.12 GB small, greater quality loss
Mistral-Nemo-Instruct-2407-Q5_0.gguf Q5_0 5 8.52 GB legacy; medium, balanced quality - prefer using Q4_K_M
Mistral-Nemo-Instruct-2407-Q5_K_M.gguf Q5_K_M 5 8.73 GB large, very low quality loss - recommended
Mistral-Nemo-Instruct-2407-Q5_K_S.gguf Q5_K_S 5 8.52 GB large, low quality loss - recommended
Mistral-Nemo-Instruct-2407-Q6_K.gguf Q6_K 6 10.1 GB very large, extremely low quality loss
Mistral-Nemo-Instruct-2407-Q8_0.gguf Q8_0 8 13.0 GB very large, extremely low quality loss - not recommended
Mistral-Nemo-Instruct-2407-f16.gguf f16 16 24.5 GB

Quantized with llama.cpp b3438.

Downloads last month
5,573
GGUF
Model size
12.2B params
Architecture
llama

2-bit

3-bit

4-bit

5-bit

6-bit

8-bit

16-bit

Inference Examples
Inference API (serverless) has been turned off for this model.

Model tree for second-state/Mistral-Nemo-Instruct-2407-GGUF

Quantized
(85)
this model

Space using second-state/Mistral-Nemo-Instruct-2407-GGUF 1