Spaces:
Sleeping
Sleeping
File size: 1,567 Bytes
aa0e4dc 2ad0ccc 4a33c20 aa0e4dc 4a33c20 2ad0ccc aa0e4dc |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 |
import gradio as gr
import os
import bark
from bark import generate_audio, preload_models, SAMPLE_RATE
import time
import scipy
import noisereduce as nr
########################
# Voice cloning functions
# make sure to only use CPU
os.environ["CUDA_VISIBLE_DEVICES"] = ""
os.environ["SUNO_USE_SMALL_MODELS"] = "1"
preload_models()
def generate_cloned_voice_audio(text_prompt, history_prompt):
# keep track of duration
t0 = time.time()
# generate cloned voice audio
audio_array = generate_audio(
text_prompt,
history_prompt = history_prompt
)
# keep track of duration
generation_duration_s = time.time() - t0
audio_duration_s = audio_array.shape[0] / SAMPLE_RATE
print(f"took {generation_duration_s:.0f}s to generate {audio_duration_s:.0f}s of audio")
# reduce noise
reduced_noise_audio_array = nr.reduce_noise(y=audio_array, sr=SAMPLE_RATE)
# write to file
audio_output_path = "output_audio.wav"
noisereduced_audio_output_path = "output_noisereduced_audio.wav"
scipy.io.wavfile.write(audio_output_path, rate=SAMPLE_RATE, data=audio_array)
scipy.io.wavfile.write(noisereduced_audio_output_path, rate=SAMPLE_RATE, data=reduced_noise_audio_array)
return audio_output_path
########################
def greet(name):
if os.path.isfile("pm_voice.npz"):
preffix = "Found the voice file"
else:
preffix = "Voice file not found"
return "Hello " + name + "!!" + preffix
iface = gr.Interface(fn=greet, inputs="text", outputs="text")
iface.launch() |