Subword Tokenization: Byte Pair Encoding

[@maryma9971]

This is amazing! Thanks for your clear explanation on BPE, really save my day. It is great that you also mention how to use BPE in other languages. Thank you!

[@abdulkadirguner1282]

for bytepair and any encoding that includes chracter level encoded, it is really hard to get because, well, the chracters are already in the vocab

[@mass13982]

Best explanation of BPE

[@tong-minglim1800]

Excellent explanation 👌

[@soumyasarkar4100]

very lucid presentation !

[@jjjery932]

thanks, buddy for this informative interpretation of BPE and BBPE

[@陳翰儒-d5m]

Man, you save me a lot, thank you very much.

[@rajputjay9856]

Oh damn finally the next part is here , thanks sir ❤️

[@dhamija80]

This was very informative, thanks so much for sharing this knowledge!!

[@JoukoSalonen]

very good, hands on and practical intro! thnak you, learned a lot!

[@mass13982]

Amazing content. Love your clear explanation and review of the original papers. Thank you Abhishek

[@ramchalamkr1]

Very very informative and explains the concept beautifully. Keep up with such content :)

[@abhishekkrthakur]

Thank you! Will do!

[@kochasaito4628]

Best interpretation of BPE, better than wiki's, I think.

[@fingerstyledojo]

Very good explanation, thank you

[@ashishbhatnagar9590]

Excellent video Sir. Thanks a lot.

[@harshitsethi9698]

Thanks for the easy explanation!

[@viveksalunkhe7293]

Informative content. Sir can you please continue this playlist and if possible can you pick any kaggle based nlp challenge and show implementation of each and every step on approaching the problem.

[@shirishbajpai9486]

Hi Abhishek! Thanks for the video.. One question that always keeps bugging me is misspelled words.. How do we handle them? Are they taken care of automatically in case of complex DL models where contextual information is used?

[@MakerBen]

Super helpful!!

[@Khushpich]

Thanks for the video. As a suggestion, I think your videos can generally be condensed to shorter lengths. This one could probably fit in 5-10 mins. Hope you don't take this the wrong, as I really appreciate the work you put in the channel, and hope you can grow it to larger audiences.

[@ilyasaroui7745]

Really good video from your part Abhishek. What are you using to record the video like this ? and what are you using for writing ? It for me to give a presentation, Thank you

[@abhishekkrthakur]

thanks. OBS and ipad.

[@jayeetaputatunda]

Thanks for the great summary! Seems like byte-pair encoding could be used as a tokenization techniqye to build a word2vec model since it doesn't handle OOV words well? Does that make sense?

[@keshavkasat9465]

When is the next video coming? We're waiting!

[@abhishekkrthakur]

soon :) sorry, got busy with loads of stuff :(

[@keshavkasat9465]

@@abhishekkrthakur Cool! Excited for the next one!

[@gregarityNow]

Great video thanks! and although it is basically irrelevant to the quality of your video, Schrank has a c ;)

[@rounhi]

Thanks! Is It possible to see a video on wordpiece tokenization please

[@abhishekkrthakur]

Did you see this video fully?

[@rounhi]

Yes of course, I mean a specific vidéo on wordpiece tokenization with some examples. Thank you

[@abhishekkrthakur]

@@rounhi okay. ill see. but there is not much difference.

[@VjayVenugopal]

Could you make videos on lstm(RNN)🙏🏻.. may be a simple name generator using character level lstm

[@hakunamatata-qu7ft]

Awsome