train error

[yuyifan1991]

Thanks so much for helping me solving the problem which I raised, yesterday, But I have another problem: 2019-03-21 10:31:02.242315: I tensorflow/core/platform/cpu_feature_guard.cc:140] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2 FMA Done building! Begin training... Traceback (most recent call last): File "train.py", line 21, in network.train(args.iterations,args.savedir) File "/home/yuyifan/EDSR-Tensorflow/model.py", line 211, in train test_x,test_y = self.test_data(*self.test_args) File "/home/yuyifan/EDSR-Tensorflow/data.py", line 67, in get_test_set get_image(imgs[0],original_size) IndexError: list index out of range I have check the directory, and it hasn't problem.

[xinqingbuhao]

I got the same issue. Howerver, when i use the command " python2 train.py ..." ,the issue can be resolved. But a new issue is i can't use GPU to train.(CUDNN problem,i don't know how to resolve) the CPU train :

[xinqingbuhao]

[kong1060553768]

@yuyifan1991 Reduce your SciPy package to 1.21

[navy63]

I have the same problem. How can I solve it?

[navy63]

The problem is"'float' object cannot be interpreted as an integer" coords_x = x / img_size------→coords_x = x // img_size coords_y = y/img_size------→coords_y = y//img_size max_counter = len(train_set)/batch_size------→max_counter = len(train_set)//batch_size python2 ，

[FBLeee]

Solution: Reduce your SciPy package to 1.2.1

[FBLeee]

If you still can't solve it, you can do so, ① first install 【pip install pillow】, ② then 【pip install scipy==1.2.1】