RuntimeError: CreateModel failed with 'Error reading the proto buffer model file.' (0x3005)

[framilano]

Hi everyone, trying to use the pretrained models on Ubuntu 20.04 and Python3.8. Installed stt with pip3, but when I'm executing the stt command;

stt --model model.tflite  --parser coqui-stt-0.9.3-models.scorer --audio test.wav

I get:

Loading model from file model.tflite
TensorFlow: v2.3.0-6-g23ad988fcde
 Coqui STT: v0.10.0-alpha.10-0-g9b517632
Warning: reading entire model file into memory. Transform model file into an mmapped graph to reduce heap usage.
2021-11-01 22:13:59.377042: 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 FMA
To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.
Data loss: Can't parse English_STT_v0.9.3/model.tflite as binary proto
Traceback (most recent call last):
  File "/usr/local/bin/stt", line 8, in <module>
    sys.exit(main())
  File "/usr/local/lib/python3.8/dist-packages/stt/client.py", line 148, in main
    ds = Model(args.model)
  File "/usr/local/lib/python3.8/dist-packages/stt/__init__.py", line 40, in __init__
    raise RuntimeError(
RuntimeError: CreateModel failed with 'Error reading the proto buffer model file.' (0x3005)

I'm running it in a Docker Container.

[framilano]

I solved installing from the same requirements available in the python websocket example.
I installed these modules with pip

setuptools~=45.2.0
deepspeech~=0.8.2
black>=20.8b1
ffmpeg-python==0.2.0
sanic==20.3.0
pyhocon==0.3.54
numpy~=1.17.0

[lissyx]

I solved installing from the same requirements available in the python websocket example.

This is likely wrong, you are pulling deepspeech binary and maybe out of luck it is able to load the model, but this is not what you want. Start from a clean virtualenv and properly install latest stt python wheel.

[framilano]

I tried installing the latest STT, but I had no luck.

[reuben]

The error message you shared is not present in the latest version, so either the installation is not working or your somehow using an installation from a different environment.

[lissyx]

Data loss: Can't parse English_STT_v0.9.3/model.tflite as binary proto

You're trying to read a TFLite model with ProtocolBuffer code.

[framilano]

I just followed this example, it does the same thing
https://stt.readthedocs.io/en/latest/DEPLOYMENT.html#using-the-python-package

[reuben]

It sounds like you're using an older version of the package. If you upgrade to STT==1.0.0 it should work fine.