Hi. I’m new to NLP and Deep Learning, and struggling a lot with PyTorch. I’m trying to train BERT on a data set for a sequence tagging task, similar to NER but with only 2 tags for each word token. I’m working in Anaconda with a Windows 10 OS. I was going to install HuggingFace’s pytorch-pretrained-bert package through conda as in the following page:
But I see that it is only for Linux OS. Is there any way I can work on a pretrained BERT in PyTorch while using Windows 10?