I just trained a Google Auto ML Vision model and exported as TensorFlowJS to implement it in React Native. The Problem I am facing: Google splits the weights into 6 different shard*.bin files. As it is a graph model, I can't use the tensorflowjs_converter to leverage the --weight_shard_size_bytes setting. I did not find any other way to import it into my React Native app than by using the BundleResourcesIO function which needs one bin file. The model should be bundled locally. Does anyone have suggestion how to merge these bin files into one? I am pretty out of options.
Thank you!