DeepSpeed
Add more weight only quantization algorithms into DeepSpeed inference.
#4577
Open
Open