Rono8/Kiswahili_TTS_Synthesis_Code
Domaine:
natural language processingType de record:
software
Créateur:
Ron
Hôte:
# Kiswahili_TTS_Synthesis_Code
The synthesis code for the Kiswahili TTS is
```
import os
import sys
! git clone Rono8/kiswahili_tts
os.chdir("./kiswahili_tts")
!pip install .
!pip install git+github.com #egg=german_transliterate
!pip uninstall tensorflow -y
!pip install tensorflow==2.3
import itertools
import logging
import os
import random
import argparse
import logging
import numpy as np
import yaml
from tqdm import tqdm
import tensorflow as tf
from tensorflow_tts.inference import TFAutoModel
from tensorflow_tts.inference import AutoConfig
from tensorflow_tts.inference import AutoProcessor
import IPython.display as ipd
import matplotlib.pyplot as plt
physical_devices = tf.config.list_physical_devices("GPU")
for i in range(len(physical_devices)):
tf.config.experimental.set_memory_growth(physical_devices[i], True)
# Download pretrained model zip file from drive
print("Downloading zip model...")
!gdown --id {"1j7IHz_2m60vVhxgp9VK3EXd2JlKz2xk3"} -O model-150000.h5
#!gdown --id {"1-1cpmelBQG2VaptQk8bFYmrUyhMYyivL"} -O tacotron2.v1.yaml
# Download pretrained Vocoder model
print("Downloading MelGAN model...")
!gdown --id {"1A3zJwzlXEpu_jHeatlMdyPGjn1V7-9iG"} -O melgan-1M6.h5
!gdown --id {"1Ys-twSd3m2uqhJOEiobNox6RNQf4txZs"} -O melgan_config.yml
# Load Vocoder
melgan_config = AutoConfig.from_pretrained('/content/kiswahili_tts/examples/melgan/conf/melgan.v1.yaml')
melgan = TFAutoModel.from_pretrained(
config=melgan_config,
pretrained_path="melgan-1M6.h5",
name="melgan"
)
# Load Model
tacotron2_config = AutoConfig.from_pretrained('/content/kiswahili_tts/examples/tacotron2/conf/tacotron2.v1.yaml')
tacotron2 = TFAutoModel.from_pretrained(
config=tacotron2_config,
pretrained_path="/content/kiswahili_tts/model-150000.h5",
name="tacotron2"
)
#processor = AutoProcessor.from_pretrained(pretrained_path="/content/drive/MyDrive/ljspeech_mapper.json")
processor = AutoProcessor.from_pretrained(pretrained_path="/cont …