dimabendera
commented
1 year ago We use snadart docker images: https://hub.docker.com/r/pytorch/pytorch/tags https://catalog.ngc.nvidia.com/orgs/nvidia/containers/pytorch/tags They are large in size. Opencv also occupies a significant part of the size.
I assume the best way is to convert model and use it with torchlib or export it to onnx or trt and try there.
Also I would recommend you to read the following articles: https://towardsdatascience.com/5-easy-to-implement-tricks-to-trim-down-your-docker-image-size-263978a6ed29 https://towardsdatascience.com/a-complete-guide-to-building-a-docker-image-serving-a-machine-learning-system-in-production-d8b5b0533bde https://stackoverflow.com/questions/63521958/is-this-a-right-way-to-descrease-size-of-my-docker-images
Sorry for the question I'm new in this area. When I build cpu docker image I get 13gb image. Is it possible to reduce size of image? I tried commenting out some countries from the CLASS_REGION_ALL variable, but that had no effect.