huggingface / optimum-neuron

Easy, fast and very cheap training and inference on AWS Trainium and Inferentia chips.
Apache License 2.0
197 stars 60 forks source link

Add support for Falcon #100

Open VGFP opened 1 year ago

VGFP commented 1 year ago

When trying to export falcon-7b to neuron I am getting an error "KeyError: "refinedwebmodel is not supported yet." Please add support for Falcon models

michaelbenayoun commented 1 year ago

Hi, This is planned, but we are waiting for it to be added in Transformers first.

gpravi commented 1 year ago

Hi @michaelbenayoun , Is there any update on this?

michaelbenayoun commented 1 year ago

The Falcon model will be supported at the next Transformer release. Will add support for this release!