Loading Pre-trained models

[kpister]

How do you load one of the pre-trained models? I've been able to run various trainings with the example scripts, but I'd like to try the pre-trained models to save some time.

If I understand correctly, when I set the MODEL_NAME and MODEL_TYPE, I'm only specifying the predictor and not the estimator itself, and the models linked above are the estimator models. Is that right?

Also, the pretrained models page mentions en-* and *-en models, but there aren't any links. Are those available?

[stale[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[TharinduDR]

Thank you very much for checking out TransQuest. Download the pretrained models and extract them. Then load them with the following code.

model = QuestModel("xlmroberta", path, num_labels=1, use_cuda=torch.cuda.is_available(), args=transformer_config)

"path" should be pointed to the directory where the files are.

You can download the multilingual model for sentence-level DA from this link - https://drive.google.com/file/d/1-DexYIw7m-9rDcBKVSrwQgLu0r-vCk6s/view?usp=sharing. If you are interested in the direction of the source and target, consider the following models too.

English to Any - https://drive.google.com/file/d/1-QDwFza-DRhiqGUHD1eQLJq-tVA9Lz0-/view?usp=sharing Any to English - https://drive.google.com/file/d/1-XQF3llRc0XTTaEePnfRW0grog0Y4BYu/view?usp=sharing

I have not yet updated the documents.

[kpister]

Thanks! I was able to get everything running with this advice.