init commit & proof of concept

stt.py

@@ -0,0 +1,72 @@
+# import required libraries
+from glob import glob
+from scipy.io.wavfile import write
+from pydub import AudioSegment
+from pydub.playback import play
+
+import os
+import torch
+import zipfile
+import torchaudio
+import sounddevice as sd
+import wave
+import wavio as wv
+
+########### Record audio
+def record_audio(audio, duration=5, freq=192000):
+    # Start recorder with the given values of duration and sample frequency
+    recording = sd.rec(int(duration * freq), samplerate=freq, channels=2)
+
+    print('Please say your prompt. You have ' + str(duration) + ' seconds.')
+    # Record audio for the given number of seconds
+    sd.wait()
+
+    # This will convert the NumPy array to an audio file with the given sampling frequency
+    write(audio, freq, recording)
+
+def speech_to_text(audio):
+    device = torch.device('cuda')  # cuda also works, but our models are fast enough for CPU
+    model, decoder, utils = torch.hub.load(repo_or_dir='snakers4/silero-models', model='silero_stt', language='en', device=device)
+    (read_batch, split_into_batches, read_audio, prepare_model_input) = utils  # see function signature for details
+
+    test_files = glob(audio)
+    batches = split_into_batches(test_files, batch_size=100)
+    input = prepare_model_input(read_batch(batches[0]), device=device)
+
+    text = ""
+    output = model(input)
+    for example in output:
+        text += decoder(example.cpu())
+
+    return text
+
+def text_to_speech():
+    language = 'en'
+    model_id = 'v3_en'
+    sample_rate = 8000
+    speaker = 'en_1'
+    device = torch.device('cuda')
+
+    model, example_text = torch.hub.load(repo_or_dir='snakers4/silero-models', model='silero_tts', language=language, speaker=model_id)
+    model.to(device)  # cuda or cpu
+
+    text = open("output.txt").read()
+    audio = model.apply_tts(text, speaker=speaker, sample_rate=sample_rate)
+    torchaudio.save("output.mp3", audio.unsqueeze(0), sample_rate=8000)
+
+    song = AudioSegment.from_mp3("output.mp3")
+    play(song)
+
+def main():
+    audio = "input.wav"
+    record_audio(audio)
+    prompt = "in 50 words or less, " +  speech_to_text(audio)
+    print(prompt)
+
+    command = "ollama run llama3 " + prompt + " > output.txt"
+    os.system(command)
+    text_to_speech()
+
+
+if __name__ == "__main__":
+    main()