YoloV2 Inference Engine
- Implemented Tensorflow-like inference engine and utilized instruction and gpu level computing and parallelism to significantly improve the performance.
- Performed quantization techniques for faster inference.
- Ranked top 5 out of 20 teams
- Tech stack: Tensorflow, Numpy, C, pthreads, AVX, CUDA, OpenBLAS, cuBLAS]
- For more information, check out GitHub repo