Masked Language Modelling Part 2 - Retraining BERT w/ Hugging Face Trainer - MRSCC - Coding Tutorial

[@rupert_ai]

I have a new camera and microphone set up - I know the keyboard sounds quite loud in this video which will hopefully be fixed for the next video. Thanks for watching!

[@ganeshkamath89]

I think this video can also be added to the Hugging Face playlist

[@mohmomani]

Hi, nice video, looking forward to see a Deberta tutorial on this channel

[@HeadshotComing]

Love your tutorials man!

[@rupert_ai]

Thanks a lot AG! : )

[@megacapojuan]

🤩🤩

[@RuloGames1]

💯💯💯💯💯💯💯💢

[@Alex-ob7te]

Hi, great video, the only question that remained unclear, when you are applying fine tuned model what tokenizer do u use?

[@rupert_ai]

Hi Alex, in terms of Hugging Face its quite simple - you use the appropriate tokenizer for the model you are using. E.g. if you are using 'bert-base-uncased' then you you the 'bert-base-uncased' tokenizer. Let me know if you need anymore help!

[@Alex-ob7te]

@@rupert_ai thanks for answers

[@juancruzalric6605]

Where can we find the datasets of Microsoft challenge? Great Video!

[@RomainBARRAUD]

Same problem here. Microsoft website redirects to another page.

[@juancruzalric6605]

@@RomainBARRAUD yep, I'm having the same issue. Finished watching part 1 However I could not find the datasets.

[@rupert_ai]

Hey ​ @Romain BARRAUD / ​ @Juan Cruz Alric - I can see the microsoft website link doesn't work - I have found a link for you here drive.google.com/drive/folders/0B5eGOMdyHn2mWDYtQzlQeGNKa2s - this contains the training data and the testing/question data.

[@juancruzalric6605]

@@rupert_ai Thanks, I have just send the request to be able to access the Drive data.

[@RomainBARRAUD]

@@rupert_ai could you give us access to the drive? I have sent a request. Thanks.