In this repo, I've tried to implement some concept preferably in AI and Machine Learning
- Sequential Test-time compute: Implementing the concept of scaling test-time compute for smaller models to perform comparably to larger models.
- A simple transformer implementation: A basic implementation of a transformer model to understand its architecture and working.
- Mixture of Recursion (MoR): Implementing the concept of Mixture of Recursions to improve recursive transformers by dynamically assigning recursive depths to individual tokens.
If you'd like to contribute to the code, please refer to contributing.md