AI Model Configuration¶

Defaults¶

• Text model: qwen2.5:3b-instruct-q4_K_M
• Vision model: qwen2.5vl:7b-q4_K_M
• Default provider mode: ollama

Provider settings¶

Model names come from these sources:

1. Environment provider settings (if you set FO_PROVIDER)
2. Saved configuration profile (FO_PROFILE or file-organizer config)
3. Built-in defaults

Read AI Provider Setup for provider variables and examples.

Configuration profile fields¶

models:
  text_model: "qwen2.5:3b-instruct-q4_K_M"
  vision_model: "qwen2.5vl:7b-q4_K_M"
  temperature: 0.5
  max_tokens: 3000
  device: "auto"     # auto, cpu, cuda, mps, metal
  framework: "ollama"  # ollama, llama_cpp, mlx

Device support¶

from file_organizer.models.base import DeviceType

DeviceType.AUTO
DeviceType.CPU
DeviceType.CUDA
DeviceType.MPS
DeviceType.METAL