Pierric Cistac
Pierric Cistac
Hi @Archelunch, what is the script you use? At which step are you getting this error?
Are you working with a big model? It seems you're [reaching the limit size](https://stackoverflow.com/questions/34128872/google-protobuf-maximum-size/34186672) (2GB) supported by [protobuf](https://developers.google.com/protocol-buffers) which is used internally by TensorFlow to serialize the graph. It doesn't...
The non-quantized TFLite version is around 1GB so way too big for a mobile app. I'll close this once the FP16 quantization works so we can use a model with...
> TL;DR: I would advocate to just implement the .no_exist from your POC +1 on this - we'll probably add pagination to "listing" endpoints in the coming weeks/months (which would...
Mmmm I'm quite sure that your expected solution is actually what is currently implemented cc @SBrandeis. We check at repo access time what orgs you're part of, not at token...
Do you know the exact error you got? @LysandreJik @muellerzr Is there any caching on the `huggingface_hub` side for auth/access info? (like what orgs a token gives access to for...
I think an `environments:` top-level key would work well yes.
Also, `hf_api.model_info("lysandre/dummy-hf-hub")` should throw a 404 error if the model doesn't exist, so you can catch that too (but less sexy than a true/false for sure). Not sure which one...
Unfortunately, the [Tokenizers](https://github.com/huggingface/tokenizers) library used internally for tokenization doesn't provide bindings for the browser but only for Nodejs (see huggingface/tokenizers#63), so no build possible for the browser for now :/