Song Han
Professor · Stanford University · Distinguished
Song Han, a professor at MIT EECS, is known for Deep Compression, which shrinks neural networks by pruning, trained quantization and Huffman coding.
About
Song Han, a professor at MIT EECS, is known for Deep Compression, which shrinks neural networks by pruning, trained quantization and Huffman coding. Han's CS231n guest lecture ties pruning, quantization and Winograd transforms to the energy cost of memory access and compares GPUs, FPGAs and TPUs.
Taught with
Same subject, AI & Machine Learning
Connections
7Their places, courses, books and lectures.
