错误描述:跑一个U-net模型,数据大概只有3g,我的电脑是16g内存,但是出现内存错误如下:
Traceback (most recent call last):
File "J:/study/U-net-master/Unet/unet-Keras.py", line 146, in <module>
unet.train()
File "J:/study/U-net-master/Unet/unet-Keras.py", line 122, in train
imgs_train, imgs_mask_train = self.load_train_data()
File "J:/study/U-net-master/Unet/unet-Keras.py", line 32, in load_train_data
imgs_mask_train = to_categorical(imgs_mask_train)
File "C:\Users\Administrator\Anaconda3\lib\site-packages\keras\utils\np_utils.py", line 31, in to_categorical
categorical[np.arange(n), y] = 1
MemoryError
提示函数to_categorical()执行出错。
解决方法:1,直接减少训练样本数目,测试可行。
2,创建generator,使用fit_generator训练模型。
3,扩大PC内存。
训练运行如下:
参考:https://stackoverflow.com/questions/46293734/memoryerror-in-keras-utils-np-utils-to-categorical
转载请注明作者与出处:https://blog.csdn.net/qq_34106574/article/details/81239865