I am creating an inferencing pipeling for DyGIE++ model using pyspark. Here my pretrained model size is 830 MB and all the other dependent modules is close to 1.3 GB (after zipp
