MindSpeed-MM End-to-End Multimodal Training Pipeline

This Skill is the routing entry point for all MindSpeed-MM Skills. It determines the model type based on user intent, routes to the corresponding Skill, and provides a complete pipeline overview.

Model Type Router

User Intent → Model Type Detection → Target Skill
  "Train understanding model / VLM"       → mindspeed-mm-vlm
  "Train generative model / video / image" → mindspeed-mm-generative
  "Train omni model"                       → See examples/qwen2.5omni/README.md
  "Train speech / TTS model"               → See examples/whisper/ or examples/cosyvoice3/README.md

Routing Criteria: