Fetching the paper…

I-ViT: Integer-only Quantization for Efficient Vision Transformer Inference · Around