WS_DAN icon indicating copy to clipboard operation
WS_DAN copied to clipboard

当batchsize设为4或8时报错

Open geolvr opened this issue 2 years ago • 0 comments

train_sample.py中的np.random.choice(np.arange(0, num_parts), 1, p=part_weights) 处报错 “probabilities contain NaN”. 使用batchsize=1, 12, 16, 32均不报错。请问是为什么呢?该算法对batchsize有特殊限制?

geolvr avatar Sep 21 '22 08:09 geolvr