-
Roofline Analysis: When Does Your Model Hit the Wall?
Understanding compute-bound vs. memory-bound operations, arithmetic intensity, and how quantization changes the game.
-
Scaling LLMs: From Alchemy to Science (Part 0)
Understanding strong scaling, roofline models, and why Transformers won the hardware compatibility game.
-
Math to Model: GRPO, PCA Augmentation, and SVD
Implementing the GRPO objective from DeepSeekMath, AlexNet's PCA color augmentation, and SVD from scratch.
-
Level Up: AlexNet, DeepSeek R1, & Linear Algebra Badges! 🚀
Breaking down my latest Deep-ML achievements in reinforcement learning, linear algebra, and classic CNNs.
-
3 New Deep-ML Badges Earned! 🏆
Celebrating completion of the Machine Learning, Deep Learning, and DenseNet collections on Deep-ML.