Watch "Endless experimentation: Building AI models in the wild." This is a 50-minute, MIT lecture about problems when deploying models and LLMs in the real world and how to prepare to solve them. This is a great lecture for those building production machine learning.

This is one of the lectures at @MITDeepLearning taught by the folks at @Cometml. If you prefer to watch on YouTube, here is the video:

The growing lack of transparency with transformers is terrifying and also sad. The hardest to understand for me starts at embeddings and then I get lost with tokenization.

Great video! Already downloaded. Will watch it soon.

Thanks for sharing 👍

I’m always looking for your tweets. Your content helps me a lot. Thanks

Interesting talk. Yes, grading papers not a good idea but seems technically impossible, with some robustness even?

I absolutely recommend it. I'm currently watching a video of GAN's in the deep learning series.

@threadreaderapp unroll

@svpino @theEliasRob Guten Tag, the unroll you asked for: Talk to you soon. 🤖

Sounds like a valuable lecture for ML practitioners. Real-world deployment challenges are crucial to address.