Writing
All Posts
Flash Attention
Forward and Backward pass derivation and algorithm explained in detail
The MATH behind Diffusion Models—DDPM
The goal of AI is to learn from data, adapt to new and complex situations and solve problems without the need for human interventions. This…
LLM Quantization Techniques
GGUF GPTQ AWQ BitNet
LLM Quantization Explained
Shrinking AI models from feast to fit without starving their intelligence.