mli/paper-reading

Organizes deep learning research through segmented video lectures covering foundational and cutting-edge papers (GPT-4, Llama 3.1, Sora, Whisper) broken down into digestible sections—e.g., Llama 3.1 split across pretraining data, model architecture, training infrastructure, and process. Content is hosted on Bilibili and YouTube with linked arXiv papers, enabling researchers to study both classical architectures and recent advances through structured, chapter-based video analysis rather than whole-paper consumption.

32,683 stars. No commits in the last 6 months.