DeepSpec MLX Dashboard

Apple Silicon environment for speculative decoding

Architecture, Design and Development by Franz Ayestaran / Enhanced Pair Programming with Claude Code & OpenAI GPT

Model Config

{
  "target_model": "Qwen/Qwen3-0.6B",
  "dtype": "fp32"
}

Runtime Status

{
  "runtime": {
    "mode": "local",
    "device": "mps"
  }
}

GGUF Conversion

Convert an MLX/PyTorch/HF model into GGUF for dashboard runtime usage.

Accepts local files/directories or Hugging Face repo ids in the form org/model.

System Status

Live host metrics, refreshed every 2 seconds.

Loading system metrics...