I am working on a GPU server which has CUDA 11.0 ( I don’t have root permissions). The compatible version of Pytorch is 1.7.1, so the corresponding transformers version is 4.30.0. However, I need to use the ViVit Model (google/vivit-b-16x2 · Hugging Face). I tried loading the model using AutoModel but its not working. How can I go about it?