How to fix the error ‘u/3�’, should be ‘TFL3’ on windows system?

Mozillians
1

Hi, I tried executing the following code on a windows 10 machine. I have installed all the packages as mentioned in the documentation.

import pyaudio
import deepspeech
import numpy as np
import wave

# Load DeepSpeech model
MODEL_PATH = "deepspeech-0.9.3-models.pbmm"
SCORER_PATH = "deepspeech-0.9.3-models.scorer"

model = deepspeech.Model(MODEL_PATH)
model.enableExternalScorer(SCORER_PATH)

def transcribe_live():
    CHUNK = 1024
    FORMAT = pyaudio.paInt16
    CHANNELS = 1
    RATE = 16000

    audio = pyaudio.PyAudio()
    stream = audio.open(format=FORMAT, channels=CHANNELS, rate=RATE, input=True, frames_per_buffer=CHUNK)

    ds_stream = model.createStream()
    print("Listening... Speak now!")

    try:
        while True:
            data = stream.read(CHUNK, exception_on_overflow=False)
            ds_stream.feedAudioContent(np.frombuffer(data, dtype=np.int16))

            metadata = ds_stream.intermediateDecodeWithMetadata()
            if metadata.num_transcripts > 0:
                transcript = "".join(token.text for token in metadata.transcripts[0].tokens)
                print(f"Intermediate: {transcript}", end="\r")

    except KeyboardInterrupt:
        print("\nFinal transcription:", ds_stream.finishStream())
        stream.stop_stream()
        stream.close()
        audio.terminate()

# Start live transcription
transcribe_live()

After the execution, i got an error saying that

TensorFlow: v2.3.0-6-g23ad988fcd
DeepSpeech: v0.9.3-0-gf2e9c858
ERROR: Model provided has model identifier 'u/3�', should be 'TFL3'

Error at reading model file D:\Practice_Python\STT\deepspeech-0.9.3-models.pbmm
Traceback (most recent call last):
  File "D:\Practice_Python\STT\livestt.py", line 10, in <module>
    model = deepspeech.Model(MODEL_PATH)
  File "C:\Users\RK\anaconda3\envs\stt\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 'Failed to initialize memory mapped model.' (0x3000)

In order to fix this error, i tried using the .tflite version instead of .pbmm. After this change, I’m getting a new error saying that

TensorFlow: v2.3.0-6-g23ad988fcd
DeepSpeech: v0.9.3-0-gf2e9c858
Listening... Speak now!
Traceback (most recent call last):
  File "D:\Practice_Python\STT\livestt.py", line 42, in <module>
    transcribe_live()
  File "D:\Practice_Python\STT\livestt.py", line 31, in transcribe_live
    if metadata.num_transcripts > 0:
AttributeError: 'Metadata' object has no attribute 'num_transcripts'

How to fix this error? I tried speaking through the microphone while executing the code, but the execution is happening very fast and then resulting in an error. Should i mandatorily use .tflite version instead of .pbmm?