No module named ‘deepspeech_training’

Archives DeepSpeech
#1

Attempting to run deepspeech.py from powershell
PS C:\users\Hank\DeepSpeech> Python DeepSpeech.py --model deepspeech-0.7.1-models.tflite --scorer deepspeech-0.7.1-models.scorer --audio testlastb.wav
Get this Error

Training package is not installed. See training documentation.

No module named ‘deepspeech_training’

Researching Discourse found the need to run
PS C:\users\Hank\DeepSpeech> pip3 install --upgrade --force-reinstall -e .
Which produced
ERROR: No matching distribution found for tensorflow==1.15.2 (from deepspeech-training==0.7.1)

Don’t know how to proceed from here.

As an aside this works fine on Raspberry Pi 4
./deepspeech --model deepspeech-0.7.1-models.tflite --scorer deepspeech-0.7.1-models.scorer --audio testlastb.wav

#2

You pasted it yourself:

Why searching the forum when there is official maintaiend documentation?

It would help if you could document your setup …

#3

First. I’m new at trying to run under Powershell. I tried the following, none were successful:
PS C:\users\Hank\DeepSpeech> Python DeepSpeech.py --model deepspeech-0.7.1-models.tflite --scorer deepspeech-0.7.1-models.scorer --audio testlastb.wav

PS C:\users\Hank\DeepSpeech> ./deepspeech --model deepspeech-0.7.1-models.tflite --scorer deepspeech-0.7.1-models.scorer --audio testlastb.wav

PS C:\users\Hank\DeepSpeech> python deepspeech --model deepspeech-0.7.1-models.tflite --scorer deepspeech-0.7.1-models.scorer --audio testlastb.wav
None worked. Can you tell me which one is correct (or another command)

I did try to follow the documentation.
Ran:

PS C:\users\Hank\DeepSpeech> pip3 install --upgrade pip==20.0.2 wheel==0.34.2 setuptools==46.1.3 --user

It was successful.

Then ran
PS C:\users\Hank\DeepSpeech> pip3 install --upgrade --force-reinstall -e .

Got the requirement tensorflow==1.15.2 error

Here’s pertinent items in my environment:
Main Directory DeepSpeech contains:
-training directory
deepspeech_training directory
train_py
-deepspeech file
-=DeepSpeech.py file
-scorer file
-tflite file

#4

What are you trying to do ? Training ? Inference ? You are mixing both.

When you’ll care enough to document your setup: system endianess, python version & endianess, …

#5

I am running Python 3.8 64 bit
My platform is Little-endian platform

I am trying to run the inference example Which works on my Pi 4)
Under Powershell
PS C:\Users\Hank\DeepSpeech> python ./deepspeech --model deepspeech-0.7.1-models.tflite --scorer deepspeech-0.7.1-models.scorer --audio testlastb.wav
python : Traceback (most recent call last):
At line:1 char:1

  • python ./deepspeech --model deepspeech-0.7.1-models.tflite --scorer d …
  •   + CategoryInfo          : NotSpecified: (Traceback (most recent call last)::String) [], RemoteException
  + FullyQualifiedErrorId : NativeCommandError

File "./deepspeech", line 6, in <module>
  from deepspeech.client import main
File "C:\Users\Hank\AppData\Local\Programs\Python\Python38\lib\site-packages\deepspeech\__init__.py", line 23, in <module>
  from deepspeech.impl import Version as version
File "C:\Users\Hank\AppData\Local\Programs\Python\Python38\lib\site-packages\deepspeech\impl.py", line 13, in <module>
  from . import _impl

ImportError: DLL load failed while importing _impl: A dynamic link library (DLL) initialization routine failed.

#6

ok, so you just need pip install deepspeech

have you read the doc and installed all the required deps? especially the msvcrt one? I’m not a windows user, and people reported issues like yours that were fixed after installing this dep.

if it’s not fixing, then please try and investigate why you can’t load the dll: there is no LD_DEBUG=all under windows, and previous tentative to reproduce the issue were unsuccessfull.

Maybe also try not using powershell, if it has an impact?