HOME
RECENT CHANGES

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revision Previous revision
Next revision
Previous revision
tensorflow:fast-style-transfer [2018/04/22 01:13]
gomida
tensorflow:fast-style-transfer [2018/05/07 04:49] (current)
Line 29: Line 29:
 ---- ----
 === TF Lite === === TF Lite ===
-스트라이드를 줄여 디컨볼루션을 제거하고 학습하면 TransposeConv는 제거할 수 있지만 모델 사이즈 제한에 걸린다. 버전 1.7.0 toco에서는 에러 없이 변환되지만 런타임 interpreter->AllocateTensors() 호출 시 에러가 발생한다.+== interpreter->AllocateTensors() 호출 시 ==
 <code> <code>
 Failed to allocate tensors! Failed to allocate tensors!
 tensorflow/contrib/lite/kernels/div.cc:48 NumDimensions(input1) != NumDimensions(input2) (4 != 0) tensorflow/contrib/lite/kernels/div.cc:48 NumDimensions(input1) != NumDimensions(input2) (4 != 0)
 </code> </code>
 +
 +== 변환 에러 ==
 <code> <code>
 gomida@TITAN:~/github/tensorflow$ ./bazel-bin/tensorflow/contrib/lite/toco/toco --input_file=/home/gomida/github/tensorflow-fast-style-transfer/frozen_out3.pb  --output_file=/home/gomida/github/tensorflow-fast-style-transfer/frozen_out3_r16.tflite --input_array=input --output_array=Tanh --input_format=TENSORFLOW_GRAPHDEF --output_format=TFLITE gomida@TITAN:~/github/tensorflow$ ./bazel-bin/tensorflow/contrib/lite/toco/toco --input_file=/home/gomida/github/tensorflow-fast-style-transfer/frozen_out3.pb  --output_file=/home/gomida/github/tensorflow-fast-style-transfer/frozen_out3_r16.tflite --input_array=input --output_array=Tanh --input_format=TENSORFLOW_GRAPHDEF --output_format=TFLITE