Data Augmentation in PyTorch: Improve Models with Existing Data

[@michaelreaves6212]

Crazy perfect timing, I was just working on data augmentation in tensorflow/keras, and was wondering about pytorch!

[@CAGonRiv]

2.2k views is mad disrespectful for a dope ass video like this😅. ❤🎉 Love you Neural Nine. Thx for sharing your knowledge!!

[@timstevens3361]

you look like techno tims big brother, same smirk too !

[@ДаниилИмани]

00:00 - intro and explanation of image data augmentation 01:30 - example CNN 04:10 - defining a function for data loading 05:00 - normalization 05:55 - data augmentation techniques: flipping, rotation, cropping 09:30 - defining pytorch DataLoaders 10:10 - tip for choosing the number of workers to load the data 11:00 - training two models; comparing the accuracy

[@letconex]

Could you please make a tutorial on how to enhance (fine-tune or RAG) LLM models for a specific task, my point of interest being translation, using personal, own translation data in the form of bilingual aligned files or tmx files and on-the-fly adaptation with a personal glossary, tailoring the translation result according to a context window made of a terminology list. Thank you for the great tutorials so far!

[@Warley.Araujo]

Great Video!! Import thing: a video you did ~1y ago about drawing recognizer with tensorflow, the gradio library isn't compatible with canvas anymore... Could u make the video with updates, It'd be great👊

[@philtoa334]

Thx_.

[@dipankarshaw900]

I vote for bringing back old simple thumbnails. How many of you prefer old thumbnails ?

[@NeuralNine]

The numbers show a preference for the new style. What did you like more about the previous one?

[@dipankarshaw900]

@NeuralNine that's great for you. I take my vote back.