Autore | SHA1 Messaggio | Data |
---|---|---|
|
e3f07b22c3 feat: support for QQQ W4A8 quantization (#612) | 5 mesi fa |
|
49a2836d61 fix: divide-by-zero warnings in marlin kernels | 5 mesi fa |
|
ba371fbbbd feat: AWQ marlin kernels (#603) | 5 mesi fa |
|
156f577f79 feat: switch from `PYBIND11_MODULE` to `TORCH_LIBRARY` (#569) | 6 mesi fa |
|
c66b1b57b1 Marlin 2:4 sparsity (#555) | 6 mesi fa |