torchtune
torchtune copied to clipboard
Implement quantized model inference for `generate_v2`
We'll probably also need #1782.