Error Diffusion: Post Training Quantization with Block-Scaled Number Formats for Neural Networks

Open in new window