Pinned Loading

1. AlphaZero AlphaZero Public

   A generalized, high-performance reinforcement learning agent based on DeepMind's AlphaZero algorithm.

   Python

2. mini-vllm mini-vllm Public

   A lightweight, high-performance LLM inference engine built from scratch using C++ and CUDA.

   Cuda

3. cuda-gemm cuda-gemm Public

   General Matrix Multiplication (GEMM) Benchmark on GPU.

   Cuda

4. iterative-robustness-optimization iterative-robustness-optimization Public

   Improves CNN robustness via falsification-driven retraining using CARLA, VerifAI, and Alpha-CROWN for verified car detection.

   Python

5. ai-attendance-system ai-attendance-system Public

   Face recognition based attendance system that identifies students from classroom images with ~98% accuracy

   Python

6. finetuning-function-calling finetuning-function-calling Public

   Two-stage SFT + GRPO fine-tuning pipeline for improving tool-calling accuracy and schema compliance in sub-billion-parameter LLMs.

   Python