You must log in or # to comment.
All neural networks are vastly approximate. The irrelevance of individual values is why quantization and sparsity work.
All neural networks are vastly approximate. The irrelevance of individual values is why quantization and sparsity work.