Xianzhi Yu

model compression (3) post-training quantization (2) efficient computing (2) neural network optimization (1) convolutional neural network (1) mixture of expert (1) sparse activation (1) feed-forward network (1) inference efficiency (1) speculative decoding (1) latency optimization (1) test-time scaling (1) sparse convolution (1) large language model (1) neural network (1) neural processing unit (1) batch inference (1) branch-wise parallelism (1) quantization sensitivity (1) floating-point format (1)