Quickstart¶

This guide gets a first MLX run working on macOS with minimal setup.

Prerequisites¶

• Apple Silicon Mac (M2/M3/M4 preferred)
• Python 3.12
• Weights at weights/model/af3.bin.zst
• Repository dependencies installed

1. Create a minimal input¶

Use one of the included examples:

cp examples/desi1_monomer.json /tmp/fold_input.json

2. Run a quick inference smoke test¶

source .venv/bin/activate
PYTHONPATH=src python3 run_alphafold_mlx.py \
  --input /tmp/fold_input.json \
  --output_dir output/quickstart \
  --num_samples 1 \
  --diffusion_steps 20 \
  --verbose

3. Check outputs¶

Expected files include:

• *.cif structure output(s)
• confidence JSON output(s)
• timing JSON output(s)

4. Move to production settings¶

For better quality:

• Increase --diffusion_steps back to 200
• Use --num_samples 5 (or more)
• Configure full MSA/templates via Pipeline Setup (macOS)

Useful links¶

• Installation
• Input Format
• Output Format
• CLI Reference