Closed pradeepdev-1995 closed 1 year ago
it is this model https://huggingface.co/microsoft/Multilingual-MiniLM-L12-H384
Thanks, @davidberenstein1957 Seems it is a multilingual model. So shall I use only the English version(vocabulary and other settings) of this to reduce the language model size?
Yes, however, the cross-lingual minilm is smaller and quicker than the xlm-roberta or spanbert alternatives. However, it is less accurate for english.
Okay, Thanks for the info. For now, I am focussing to reduce the model size/ increase the inference time rather than accuracy. So how should I modify the code
predictor = Predictor(language="en_core_web_sm", device=-1, model_name="minilm")
to use only the English vocabulary to reduce the model size?
yes than the model minilm
should be the best in your case.
Yes, So how should I modify the code
predictor = Predictor(language="en_core_web_sm", device=-1, model_name="minilm")
or even the model(https://huggingface.co/microsoft/Multilingual-MiniLM-L12-H384
) itself to use only the English vocabulary for reducing the model size?
predictor = Predictor(language="en_core_web_sm", device=-1, model_name="minilm") is correct. Then it uses the smallest model available. Or you could fine-tune your own model but that introduces some overhead and effort.
I am using the following code snippet for coreference resolution
While checking the below source code,
it seems that the language model using here is
https://storage.googleapis.com/pandora-intelligence/models/crosslingual-coreference/minilm/model.tar.gz
Is this the same one that I can see in
https://huggingface.co/models
likehttps://huggingface.co/microsoft/Multilingual-MiniLM-L12-H384/tree/main
or any other huggingface model?