Model quantized using a modified [EETQ](https://github.com/NetEase-FuXi/EETQ) repo. Currently working on | |
decoupling its kernels from CUTLASS to make this a bit easier to use. | |
8bits. |
Model quantized using a modified [EETQ](https://github.com/NetEase-FuXi/EETQ) repo. Currently working on | |
decoupling its kernels from CUTLASS to make this a bit easier to use. | |
8bits. |