Add .gitignore and implement model weight loading with quantization support c250d8c Running manoskary commited on 21 days ago