mistral7b constant
ModelInfo
const mistral7b
Mistral 7B Instruct v0.2 — high‑quality 7B instruction model.
~4.4 GB download, 8 GB RAM recommended, 32768 token context.
Implementation
static const mistral7b = ModelInfo(
id: 'mistral7b',
name: 'Mistral 7B Instruct v0.2',
fileName: 'mistral-7b-instruct-v0.2.Q4_K_M.gguf',
ggufUrl:
'https://huggingface.co/TheBloke/Mistral-7B-Instruct-v0.2-GGUF/resolve/main/mistral-7b-instruct-v0.2.Q4_K_M.gguf',
sizeMB: 4370,
minRamGB: 8,
context: 32768,
chatTemplate: _mistralTemplate,
);