tensorflow: [Windows] Speech commands tutorial does not work

System information

Have I written custom code (as opposed to using a stock example script provided in TensorFlow): No, I’m running python tensorflow/examples/speech_commands/train.py
OS Platform and Distribution (e.g., Linux Ubuntu 16.04): Windows 7 x64
TensorFlow installed from (source or binary): pip install tf-nightly (tf_nightly-1.5.0.dev20171026-cp35-cp35m-win_amd64.whl)
TensorFlow version (use command below): b’unknown’ 1.5.0-dev20171026
Python version: Python 3.5.4
Bazel version (if compiling from source): N/A
CUDA/cuDNN version: CUDA 8.0, cuDNN 6.1 x64
GPU model and memory: 2x Nvidia GTX 670 2GB
Exact command to reproduce: python tensorflow/examples/speech_commands/train.py

Describe the problem

It appears that the currently nightlies (or 1.4.0rc1) do not contain the gen_audio_ops module. However the documentation for r1.4 or master (https://www.tensorflow.org/versions/r1.4/tutorials/audio_recognition or https://www.tensorflow.org/versions/master/tutorials/audio_recognition) appears to indicate that the speech_commands demo should work.

Related: #13031

Source code / logs

Traceback (most recent call last):
  File "tensorflow/examples/speech_commands/train.py", line 81, in <module>
    import input_data
  File "E:\Tom\Documents\GIT\tensorflow\tensorflow\examples\speech_commands\input_data.py", line 35, in <module>
    from tensorflow.contrib.framework.python.ops import audio_ops as contrib_audio
  File "C:\Users\Tom\AppData\Local\Programs\Python\Python35\lib\site-packages\tensorflow\contrib\framework\python\ops\audio_ops.py", line 31, in <module>
    from tensorflow.python.ops.gen_audio_ops import *
ImportError: No module named 'tensorflow.python.ops.gen_audio_ops'

About this issue

Original URL
State: closed
Created 7 years ago
Reactions: 2
Comments: 21 (10 by maintainers)

Commits related to this issue

[CMake] Generate audio_ops wrappers in the CMake build. Fixes #14004. PiperOrigin-RevId: 173732397 — committed to caisq/tensorflow by mrry 7 years ago
[CMake] Generate `gen_audio_ops.py`. Fixes #14004. — committed to tensorflow/tensorflow by mrry 7 years ago
[CMake] Generate `gen_audio_ops.py`. (#14216) Fixes #14004. — committed to tensorflow/tensorflow by mrry 7 years ago

Most upvoted comments

@NearLinHere The Wheel filename includes 20171026, which suggests that it’s the nightly build from October 26th 2017. Can you try against with the latest nightly build?

mrry on Nov 6, 2017

@mrry thanks for your reply. I changed my tensor flow version to a newer version tf_nightly. This solve my problem, but now I have another one.

PS C:\Python36\lib\site-packages> python tensorflow\examples\speech_commands\train.py 2017-11-11 16:59:17.282443: I C:\tf_jenkins\home\workspace\tf-nightly-windows\M\windows\PY\36\tensorflow\core\platform\cpu _feature_guard.cc:137] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX AVX2 Traceback (most recent call last): File “tensorflow\examples\speech_commands\train.py”, line 429, in <module> tf.app.run(main=main, argv=[sys.argv[0]] + unparsed) File “C:\Python36\lib\site-packages\tensorflow\python\platform\app.py”, line 133, in run _sys.exit(main(argv)) File “tensorflow\examples\speech_commands\train.py”, line 106, in main FLAGS.testing_percentage, model_settings) File “C:\Python36\lib\site-packages\tensorflow\examples\speech_commands\input_data.py”, line 161, in init testing_percentage) File “C:\Python36\lib\site-packages\tensorflow\examples\speech_commands\input_data.py”, line 243, in prepare_data_index word = re.search(‘./([^/]+)/..wav’, wav_path).group(1).lower() AttributeError: ‘NoneType’ object has no attribute ‘group’

przemoo9621 on Nov 13, 2017

I have the same problem. The code is pull 5 hours ago.

System information Have I written custom code (as opposed to using a stock example script provided in TensorFlow): No, I’m running python tensorflow/examples/speech_commands/train.py OS Platform and Distribution (e.g., Linux Ubuntu 16.04): Windows 10 x64 TensorFlow installed from (source or binary): pip install tf-nightly (tf_nightly-1.5.0.dev20171026-cp35-cp35m-win_amd64.whl) TensorFlow version (use command below): b’unknown’ 1.5.0-dev20171104 Python version: Python 3.6.3 Bazel version (if compiling from source): N/A CUDA/cuDNN version: CUDA 8.0, cuDNN 6.0 GPU model and memory: Nvidia GTX 960M , 8GB Exact command to reproduce: python tensorflow/examples/speech_commands/train.py

NearLinHere on Nov 6, 2017

Thanks for letting us know. It looks like that code is relying on '/' being the path separator. I’ve just sent PR #14519 that should fix this.

mrry on Nov 13, 2017

@KevlarTheGreat Which version of TensorFlow have you installed? The fix is only available in the tf-nightly package, and will be part of TF 1.5.

mrry on Nov 7, 2017