For how long where the trax and huggingface models trained or finetuned?

Course Q&A Natural Language Processing NLP with Attention Models
1

I know we used pre-trained models and pre-finetuned models in assignments of Week 3. I wonder if it’s worth to train these models a few days on Google Colab Pro as an exercise and how long it would take to train or finetune the trax and huggingface models?

Thanks.

2

Hi @Juan_Rojas,
It usually takes a lot of compute to train the models on multiple GPUs. However, for the huggingface models, I’ll suggest you just go to their corresponding pages to learn about how long they were trained.
Kindly share how it went if you were able to do the training.