keras-nlp icon indicating copy to clipboard operation
keras-nlp copied to clipboard

Published 20 hours ago • verified publisher icon keras-team

Sharding for model loading

Open kanpuriyanawab opened this issue 2 years ago • 2 comments

As we are progressing towards welcoming Large models in library, we need a sharding mechanism for loading large checkpoints in model.

kanpuriyanawab avatar Jul 29 '23 09:07 kanpuriyanawab

Hi @mattdangerw , Here is how huggingface shards the model. This is fairly simple strategy.

kanpuriyanawab avatar Sep 01 '23 10:09 kanpuriyanawab

I think @nkovela1 will be taking a look here!

mattdangerw avatar Oct 20 '23 21:10 mattdangerw