Machine Learning models can be Discriminative or Generative Models. What really are generative models? Unlike the common belief models that simply generate are not generative models. Generative models understand the entire dataset in depth and hence sampling a single point from their distribution leads to a sample very similar to their training data. Here is a video explaining the idea.Wonder how to get started with OpenAI's new Python API? Here is a walkthrough of all the possibilitiesWould you like to learn how ChatGPT is trained and how it is special compared to all other language models so far?
Transformers in Deep Learning
This is a video series about the most happening "Transformers" architecture. The first video is about self-attention in general. The second one goes on to explaining how self-attention is used in transformers. The third one is about the concept of positional encoding and input embedding.
