Closed cpetrov closed 1 year ago
This can help leverage hardware optimisations and significantly speed up inference: https://huggingface.co/transformers/serialization.html
It seems this would be only possible with transformers>=4.9.0 (see https://huggingface.co/transformers/serialization.html#configuration-based-approach)), so this issue seems related to https://github.com/georgian-io/Multimodal-Toolkit/issues/3.
This can help leverage hardware optimisations and significantly speed up inference: https://huggingface.co/transformers/serialization.html
It seems this would be only possible with transformers>=4.9.0 (see https://huggingface.co/transformers/serialization.html#configuration-based-approach)), so this issue seems related to https://github.com/georgian-io/Multimodal-Toolkit/issues/3.