metadata

license: apache-2.0
model_name: mathstral-7B-v0.1
base_model: mistralai/mathstral-7B-v0.1
inference: false
model_creator: mistralai
quantized_by: Second State Inc.

mathstral-7B-v0.1-GGUF

Original Model

mistralai/mathstral-7B-v0.1

Run with LlamaEdge

LlamaEdge version: v0.12.3

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: 32000

Run as LlamaEdge service

wasmedge --dir .:. --nn-preload default:GGML:AUTO:mathstral-7B-v0.1-Q5_K_M.gguf \
  llama-api-server.wasm \
  --prompt-template mistral-instruct \
  --ctx-size 32000 \
  --model-name mathstral-7B-v0.1

Run as LlamaEdge command app

wasmedge --dir .:. --nn-preload default:GGML:AUTO:mathstral-7B-v0.1-Q5_K_M.gguf \
  llama-chat.wasm \
  --prompt-template mistral-instruct \
  --ctx-size 32000

Quantized GGUF Models

Name	Quant method	Bits	Size	Use case
mathstral-7B-v0.1-Q2_K.gguf	Q2_K	2	2.72 GB	smallest, significant quality loss - not recommended for most purposes
mathstral-7B-v0.1-Q3_K_L.gguf	Q3_K_L	3	3.83 GB	small, substantial quality loss
mathstral-7B-v0.1-Q3_K_M.gguf	Q3_K_M	3	3.52 GB	very small, high quality loss
mathstral-7B-v0.1-Q3_K_S.gguf	Q3_K_S	3	3.17 GB	very small, high quality loss
mathstral-7B-v0.1-Q4_0.gguf	Q4_0	4	4.11 GB	legacy; small, very high quality loss - prefer using Q3_K_M
mathstral-7B-v0.1-Q4_K_M.gguf	Q4_K_M	4	4.37 GB	medium, balanced quality - recommended
mathstral-7B-v0.1-Q4_K_S.gguf	Q4_K_S	4	4.14 GB	small, greater quality loss
mathstral-7B-v0.1-Q5_0.gguf	Q5_0	5	5 GB	legacy; medium, balanced quality - prefer using Q4_K_M
mathstral-7B-v0.1-Q5_K_M.gguf	Q5_K_M	5	5.14 GB	large, very low quality loss - recommended
mathstral-7B-v0.1-Q5_K_S.gguf	Q5_K_S	5	5 GB	large, low quality loss - recommended
mathstral-7B-v0.1-Q6_K.gguf	Q6_K	6	5.95 GB	very large, extremely low quality loss
mathstral-7B-v0.1-Q8_0.gguf	Q8_0	8	7.7 GB	very large, extremely low quality loss - not recommended
mathstral-7B-v0.1-f16.gguf	f16	16	14.5 GB

Quantized with llama.cpp b3389.