Skip to content

Blogs

  • Accelerating Code LLMs

    Exploring a technique to speed up code-generation with LLMs through smart caching and selective updates, making development workflows faster and efficient while maintaining code quality.

  • Rethinking Pooling Layers

    Rethinking how pooling layers in CNNs do more than just reduce dimensionality; they introduce approximate invariance, helping the network stay focused on key features while ignoring minor variations and noise in input data.

More to come soon!