djkesu's picture
replaced backed db with sqlite
e302830
raw
history blame
1.34 kB
import gc
import os
from contextlib import contextmanager
from time import time
from typing import Optional
import streamlit as st
from tortoise.api import TextToSpeech
from tortoise.utils.audio import load_voices
@contextmanager
def timeit(desc=""):
start = time()
yield
print(f"{desc} took {time() - start:.2f} seconds")
@st.cache_resource(max_entries=1)
def load_model(
model_dir,
high_vram,
kv_cache,
ar_checkpoint,
diff_checkpoint,
):
gc.collect()
return TextToSpeech(
models_dir=model_dir,
high_vram=high_vram,
kv_cache=kv_cache,
ar_checkpoint=ar_checkpoint,
diff_checkpoint=diff_checkpoint,
)
@st.cache_data
def list_voices(extra_voices_dir: Optional[str]):
voices = ["random"]
if extra_voices_dir and os.path.isdir(extra_voices_dir):
voices.extend(os.listdir(extra_voices_dir))
extra_voices_ls = [extra_voices_dir]
else:
extra_voices_ls = []
voices.extend(
[v for v in os.listdir("tortoise/voices") if v != "cond_latent_example"]
)
#
return voices, extra_voices_ls
@st.cache_resource(max_entries=1)
def load_voice_conditionings(voice, extra_voices_ls):
voice_samples, conditioning_latents = load_voices(voice, extra_voices_ls)
return voice_samples, conditioning_latents