search

neurosim / DNN_NeuroSim_V1.3

Benchmark framework of compute-in-memory based accelerators for deep neural network (inference engine focused)
62 stars 36 forks source link

net=VGG8，inference=1，训练速度很慢 #5

Closed zhangfeixiang222 closed 3 years ago

zhangfeixiang222 commented 3 years ago

请问有什么方法可以加快训练速度吗

neurosim commented 3 years ago

Please set inference as 0 during training. The label "inference" means considering hardware effects in feed-forward in the code, while training is assumed run by software in this NeuroSim version. If you want to consider the hardware effects during training (on-chip training), please using the training version of NeuroSim.