模型壓縮：Deep Compression

第一步weight pruning 第二步trained quantization and weight sharing 第三步 Huffman coding 實驗分析之壓縮幾十倍從何而來 實驗分析之極致量化 《Deep Compression Compressing Deep Neural Networks with Pruning, Trained Quantization and Huffm

>>阅读原文<<