Hello everyone, So I’m new to deepspeech and either I’m facing an issue here or I might just didn’t know how to use it.
So I’m working on Windows 10 and I’m using deepspeech python version.
And I want to work with the french prebuilt models for deepspeech which exist in here.
So I’ve setup two python virtual environments with venv.
In the first venv, I’ve downloaded the french tensorflow model.
And in the second venv I’ve downloaded the french tflite model.
The first environment I’ve setup it’s for the tensorflow model which contains the following packages:
colorama 0.4.4
deepspeech 0.9.3
halo 0.0.31
log-symbols 0.0.14
numpy 1.14.5
pip 18.1
PyAudio 0.2.11
scipy 1.4.1
setuptools 40.6.2
six 1.15.0
spinners 0.0.24
termcolor 1.1.0
webrtcvad 2.0.10
And the second environment I’ve setup it’s for the tflite model which contains the following packages:
absl-py 0.11.0
astunparse 1.6.3
cachetools 4.2.0
certifi 2020.12.5
chardet 4.0.0
colorama 0.4.4
deepspeech 0.8.0
deepspeech-tflite 0.8.0
gast 0.3.3
google-auth 1.24.0
google-auth-oauthlib 0.4.2
google-pasta 0.2.0
grpcio 1.34.0
h5py 2.10.0
halo 0.0.31
idna 2.10
importlib-metadata 3.3.0
Keras-Preprocessing 1.1.2
log-symbols 0.0.14
Markdown 3.3.3
numpy 1.14.4
oauthlib 3.1.0
opt-einsum 3.3.0
pip 18.1
protobuf 3.14.0
pyasn1 0.4.8
pyasn1-modules 0.2.8
PyAudio 0.2.11
requests 2.25.1
requests-oauthlib 1.3.0
rsa 4.6
scipy 1.4.1
setuptools 40.6.2
six 1.15.0
spinners 0.0.24
tensorboard 2.4.0
tensorboard-plugin-wit 1.7.0
tensorflow-estimator 2.3.0
termcolor 1.1.0
typing-extensions 3.7.4.3
urllib3 1.26.2
webrtcvad 2.0.10
Werkzeug 1.0.1
wheel 0.36.2
wrapt 1.12.1
zipp 3.4.0
And now, I want to work in both virtual environments with mic_vad_streaming.
So when I’m working with the first venv(tensorflow model), I have no problems and deepspeech works flawlessly(I’ve encountered some lag/slow response but it’s okey for now).
But when I’m trying to use the second venv(tflite model), I encountered this issue:
Loading model from file output_graph.tflite
TensorFlow: v2.2.0-17-g0854bb5188
DeepSpeech: v0.8.0-0-gf56b07da
Warning: reading entire model file into memory. Transform model file into an mmapped graph to reduce heap usage.
2021-01-05 21:30:07.055669: I tensorflow/core/platform/cpu_feature_guard.cc:143] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2
Data loss: Can't parse output_graph.tflite as binary proto
Traceback (most recent call last):
File "C:\Program Files\Python36\lib\runpy.py", line 193, in _run_module_as_main
"__main__", mod_spec)
File "C:\Program Files\Python36\lib\runpy.py", line 85, in _run_code
exec(code, run_globals)
File "C:\Users\HP\Downloads\model_tflite_fr\Scripts\deepspeech.exe\__main__.py", line 9, in <module>
File "c:\users\hp\downloads\model_tflite_fr\lib\site-packages\deepspeech\client.py", line 117, in main
ds = Model(args.model)
File "c:\users\hp\downloads\model_tflite_fr\lib\site-packages\deepspeech\__init__.py", line 38, in __init__
raise RuntimeError("CreateModel failed with '{}' (0x{:X})".format(deepspeech.impl.ErrorCodeToErrorMessage(status),status))
RuntimeError: CreateModel failed with 'Error reading the proto buffer model file.' (0x3005)
Here’s the output of the first venv(tensorflow model) when it work successfully:
Initializing model...
INFO:root:ARGS.model: output_graph.pbmm
TensorFlow: v2.3.0-6-g23ad988fcd
DeepSpeech: v0.9.3-0-gf2e9c858
2021-01-05 21:19:58.129550: I tensorflow/core/platform/cpu_feature_guard.cc:142] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN)to use the following CPU instructions in performance-critical operations: AVX2
To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.
INFO:root:ARGS.scorer: kenlm.scorer
Listening (ctrl-C to exit)...
Recognized: bonjour
Recognized: bonsoir dont
Recognized: en
Recognized: on range en deux
Recognized: mais profond
Recognized: la pole
Recognized: mai coute moi bien
Recognized: la
Recognized: paul
Recognized: du point
The command I’ve used for the first venv(tensorflow model) which works successfully:
python mic_vad_streaming.py -m output_graph.pbmm -s kenlm.scorer
The command I’ve used for the second venv
(tflite model) which doesn’t work:
python mic_vad_streaming.py -m output_graph.tflite -s kenlm.scorer
I’ve even tried using deepspeech directly in the second venv with a .wav audio, but still the same results.
(model_tflite_fr) C:\Users\Ayoub\Downloads\model_tflite_fr>deepspeech --model output_graph.tflite --scorer kenlm.scorer --audio outputs\savewav_2021-01-05_21-26-23_483447.wav
Loading model from file output_graph.tflite
TensorFlow: v2.2.0-17-g0854bb5188
DeepSpeech: v0.8.0-0-gf56b07da
Warning: reading entire model file into memory. Transform model file into an mmapped graph to reduce heap usage.
2021-01-05 21:30:07.055669: I tensorflow/core/platform/cpu_feature_guard.cc:143] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2
Data loss: Can't parse output_graph.tflite as binary proto
Traceback (most recent call last):
File "C:\Program Files\Python36\lib\runpy.py", line 193, in _run_module_as_main
"__main__", mod_spec)
File "C:\Program Files\Python36\lib\runpy.py", line 85, in _run_code
exec(code, run_globals)
File "C:\Users\HP\Downloads\model_tflite_fr\Scripts\deepspeech.exe\__main__.py", line 9, in <module>
File "c:\users\hp\downloads\model_tflite_fr\lib\site-packages\deepspeech\client.py", line 117, in main
ds = Model(args.model)
File "c:\users\hp\downloads\model_tflite_fr\lib\site-packages\deepspeech\__init__.py", line 38, in __init__
raise RuntimeError("CreateModel failed with '{}' (0x{:X})".format(deepspeech.impl.ErrorCodeToErrorMessage(status),status))
RuntimeError: CreateModel failed with 'Error reading the proto buffer model file.' (0x3005)
I think that’s all.
I appreciate any help possible and thanks mozilla for this awesome project.