seeprettyface-generator-wanghong
seeprettyface-generator-wanghong copied to clipboard
a312863063
请问能在无GPU的PC上运行吗？
如果电脑内存够，CPU上是可以生成的
但是会出现这样的问题，请问您如何解决？ InvalidArgumentError (see above for traceback): Cannot assign a device for operation 'Gs/_Run/Gs/Truncation/Select': Operation was explicitly assigned to /device:GPU:0 but available devices are [ /job:localhost/replica:0/task:0/device:CPU:0 ]. Make sure the device specification refers to a valid device. [[Node: Gs/_Run/Gs/Truncation/Select = Select[T=DT_FLOAT, _device="/device:GPU:0"](Gs/_Run/Gs/Truncation/Select/condition, Gs/_Run/Gs/Truncation/Select/t, Gs/_Run/Gs/Truncation/Select/e)]]
但是会出现这样的问题，请问您如何解决？ InvalidArgumentError (see above for traceback): Cannot assign a device for operation 'Gs/_Run/Gs/Truncation/Select': Operation was explicitly assigned to /device:GPU:0 but available devices are [ /job:localhost/replica:0/task:0/device:CPU:0 ]. Make sure the device specification refers to a valid device. [[Node: Gs/_Run/Gs/Truncation/Select = Select[T=DT_FLOAT, _device="/device:GPU:0"](Gs/_Run/Gs/Truncation/Select/condition, Gs/_Run/Gs/Truncation/Select/t, Gs/_Run/Gs/Truncation/Select/e)]]
你好，将代码里所有的with tf.device('/gpu:0')改成with tf.device('/cpu:0')
dnnlib/tflib/network.py里面的407，408行： < for gpu in range(num_gpus): < with tf.device("/gpu:%d" % gpu): 改成with tf.device("/cpu:0")就可以了
您好，按照您的提示修改后新问题出现了，这个也是我之前遇到过的，请问您有好的解决方法吗？ InvalidArgumentError (see above for traceback): Conv2DCustomBackpropInputOp only supports NHWC. [[Node: Gs/_Run/Gs/G_synthesis/cond/cond/cond/cond/128x128/Conv0_up/conv2d_transpose = Conv2DBackpropInput[T=DT_FLOAT, data_format="NCHW", dilations=[1, 1, 1, 1], padding="SAME", strides=[1, 1, 2, 2], use_cudnn_on_gpu=false, _device="/job:localhost/replica:0/task:0/device:CPU:0"](Gs/_Run/Gs/G_synthesis/cond/cond/cond/cond/128x128/Conv0_up/conv2d_transpose/output_shape, Gs/_Run/Gs/G_synthesis/cond/cond/cond/cond/128x128/Conv0_up/AddN, Gs/_Run/Gs/G_synthesis/cond/cond/cond/cond/128x128/Conv0_up/Shape/Switch:1)]]
您好，按照您的提示修改后新问题出现了，这个也是我之前遇到过的，请问您有好的解决方法吗？ InvalidArgumentError (see above for traceback): Conv2DCustomBackpropInputOp only supports NHWC. [[Node: Gs/_Run/Gs/G_synthesis/cond/cond/cond/cond/128x128/Conv0_up/conv2d_transpose = Conv2DBackpropInput[T=DT_FLOAT, data_format="NCHW", dilations=[1, 1, 1, 1], padding="SAME", strides=[1, 1, 2, 2], use_cudnn_on_gpu=false, _device="/job:localhost/replica:0/task:0/device:CPU:0"](Gs/_Run/Gs/G_synthesis/cond/cond/cond/cond/128x128/Conv0_up/conv2d_transpose/output_shape, Gs/_Run/Gs/G_synthesis/cond/cond/cond/cond/128x128/Conv0_up/AddN, Gs/_Run/Gs/G_synthesis/cond/cond/cond/cond/128x128/Conv0_up/Shape/Switch:1)]]
我也遇到过，这是版本问题导致的，将tensorflow升级至最新版应该可以解决。。
您好，按照您的提示修改后新问题出现了，这个也是我之前遇到过的，请问您有好的解决方法吗？ InvalidArgumentError (see above for traceback): Conv2DCustomBackpropInputOp only supports NHWC. [[Node: Gs/_Run/Gs/G_synthesis/cond/cond/cond/cond/128x128/Conv0_up/conv2d_transpose = Conv2DBackpropInput[T=DT_FLOAT, data_format="NCHW", dilations=[1, 1, 1, 1], padding="SAME", strides=[1, 1, 2, 2], use_cudnn_on_gpu=false, _device="/job:localhost/replica:0/task:0/device:CPU:0"](Gs/_Run/Gs/G_synthesis/cond/cond/cond/cond/128x128/Conv0_up/conv2d_transpose/output_shape, Gs/_Run/Gs/G_synthesis/cond/cond/cond/cond/128x128/Conv0_up/AddN, Gs/_Run/Gs/G_synthesis/cond/cond/cond/cond/128x128/Conv0_up/Shape/Switch:1)]]
我也遇到过，这是版本问题导致的，将tensorflow升级至最新版应该可以解决。。
我将tensorflow-cpu升级到了1.13.1，但还是会出现这样的问题，所以应该不是版本问题？或者您的版本多少？ 网上说是因为tensorflow-cpu不支持GPU训练的NCHW模型格式，所以不知道您有没有好的建议？
github有人建了一个cpu版本的，你可以搜下
求链接
https://github.com/huaji0353/styleGAN_CPU
thx ！！！
