Binarized Neural Networks Training Deep Neural Networks with Weights and Activations Constrained to +1 or -1 04-28
Deep Compression Compressing Deep Neural Networks with Pruning, Trained Quantization and Huffman Coding 04-26