initial upload attempt
initial upload attempt
This commit is contained in:
erew123 committed
1 parent
84c19ed5a9
commit
9fd928f6e7
527 files changed
+37359
No files matched your search
@@ -0,0 +1,190 @@
|
||||
{
|
||||
"first_start_model": "tts_models--en--vctk--vits",
|
||||
"models": [
|
||||
{
|
||||
"model_name": "tts_models--bg--cv--vits",
|
||||
"folder_path": "tts_models--bg--cv--vits",
|
||||
"github_rls_url": "https://coqui.gateway.scarf.sh/v0.8.0_models/tts_models--bg--cv--vits.zip"
|
||||
},
|
||||
{
|
||||
"model_name": "tts_models--cs--cv--vits",
|
||||
"folder_path": "tts_models--cs--cv--vits",
|
||||
"github_rls_url": "https://coqui.gateway.scarf.sh/v0.8.0_models/tts_models--cs--cv--vits.zip"
|
||||
},
|
||||
{
|
||||
"model_name": "tts_models--da--cv--vits",
|
||||
"folder_path": "tts_models--da--cv--vits",
|
||||
"github_rls_url": "https://coqui.gateway.scarf.sh/v0.8.0_models/tts_models--da--cv--vits.zip"
|
||||
},
|
||||
{
|
||||
"model_name": "tts_models--et--cv--vits",
|
||||
"folder_path": "tts_models--et--cv--vits",
|
||||
"github_rls_url": "https://coqui.gateway.scarf.sh/v0.8.0_models/tts_models--et--cv--vits.zip"
|
||||
},
|
||||
{
|
||||
"model_name": "tts_models--ga--cv--vits",
|
||||
"folder_path": "tts_models--ga--cv--vits",
|
||||
"github_rls_url": "https://coqui.gateway.scarf.sh/v0.8.0_models/tts_models--ga--cv--vits.zip"
|
||||
},
|
||||
{
|
||||
"model_name": "tts_models--en--vctk--vits",
|
||||
"folder_path": "tts_models--en--vctk--vits",
|
||||
"github_rls_url": "https://coqui.gateway.scarf.sh/v0.6.1_models/tts_models--en--vctk--vits.zip"
|
||||
},
|
||||
{
|
||||
"model_name": "tts_models--es--css10--vits",
|
||||
"folder_path": "tts_models--es--css10--vits",
|
||||
"github_rls_url": "https://coqui.gateway.scarf.sh/v0.8.0_models/tts_models--es--css10--vits.zip"
|
||||
},
|
||||
{
|
||||
"model_name": "tts_models--fr--css10--vits",
|
||||
"folder_path": "tts_models--fr--css10--vits",
|
||||
"github_rls_url": "https://coqui.gateway.scarf.sh/v0.8.0_models/tts_models--fr--css10--vits.zip"
|
||||
},
|
||||
{
|
||||
"model_name": "tts_models--uk--mai--vits",
|
||||
"folder_path": "tts_models--uk--mai--vits",
|
||||
"github_rls_url": "https://coqui.gateway.scarf.sh/v0.8.0_models/tts_models--uk--mai--vits.zip"
|
||||
},
|
||||
{
|
||||
"model_name": "tts_models--nl--css10--vits",
|
||||
"folder_path": "tts_models--nl--css10--vits",
|
||||
"github_rls_url": "https://coqui.gateway.scarf.sh/v0.8.0_models/tts_models--nl--css10--vits.zip"
|
||||
},
|
||||
{
|
||||
"model_name": "tts_models--de--thorsten--vits",
|
||||
"folder_path": "tts_models--de--thorsten--vits",
|
||||
"github_rls_url": "https://coqui.gateway.scarf.sh/v0.7.0_models/tts_models--de--thorsten--vits.zip"
|
||||
},
|
||||
{
|
||||
"model_name": "tts_models--de--css10--vits-neon",
|
||||
"folder_path": "tts_models--de--css10--vits-neon",
|
||||
"github_rls_url": "https://coqui.gateway.scarf.sh/v0.8.0_models/tts_models--de--css10--vits.zip"
|
||||
},
|
||||
{
|
||||
"model_name": "tts_models--it--mai_female--vits",
|
||||
"folder_path": "tts_models--it--mai_female--vits",
|
||||
"github_rls_url": "https://coqui.gateway.scarf.sh/v0.6.1_models/tts_models--it--mai_female--vits.zip"
|
||||
},
|
||||
{
|
||||
"model_name": "tts_models--it--mai_male--vits",
|
||||
"folder_path": "tts_models--it--mai_male--vits",
|
||||
"github_rls_url": "https://coqui.gateway.scarf.sh/v0.6.1_models/tts_models--it--mai_male--vits.zip"
|
||||
},
|
||||
{
|
||||
"model_name": "tts_models--ewe--openbible--vits",
|
||||
"folder_path": "tts_models--ewe--openbible--vits",
|
||||
"github_rls_url": "https://coqui.gateway.scarf.sh/v0.6.2_models/tts_models--ewe--openbible--vits.zip"
|
||||
},
|
||||
{
|
||||
"model_name": "tts_models--hau--openbible--vits",
|
||||
"folder_path": "tts_models--hau--openbible--vits",
|
||||
"github_rls_url": "https://coqui.gateway.scarf.sh/v0.6.2_models/tts_models--hau--openbible--vits.zip"
|
||||
},
|
||||
{
|
||||
"model_name": "tts_models--lin--openbible--vits",
|
||||
"folder_path": "tts_models--lin--openbible--vits",
|
||||
"github_rls_url": "https://coqui.gateway.scarf.sh/v0.6.2_models/tts_models--lin--openbible--vits.zip"
|
||||
},
|
||||
{
|
||||
"model_name": "tts_models--tw_akuapem--openbible--vits",
|
||||
"folder_path": "tts_models--tw_akuapem--openbible--vits",
|
||||
"github_rls_url": "https://coqui.gateway.scarf.sh/v0.6.2_models/tts_models--tw_akuapem--openbible--vits.zip"
|
||||
},
|
||||
{
|
||||
"model_name": "tts_models--tw_asante--openbible--vits",
|
||||
"folder_path": "tts_models--tw_asante--openbible--vits",
|
||||
"github_rls_url": "https://coqui.gateway.scarf.sh/v0.6.2_models/tts_models--tw_asante--openbible--vits.zip"
|
||||
},
|
||||
{
|
||||
"model_name": "tts_models--yor--openbible--vits",
|
||||
"folder_path": "tts_models--yor--openbible--vits",
|
||||
"github_rls_url": "https://coqui.gateway.scarf.sh/v0.6.2_models/tts_models--yor--openbible--vits.zip"
|
||||
},
|
||||
{
|
||||
"model_name": "tts_models--hu--css10--vits",
|
||||
"folder_path": "tts_models--hu--css10--vits",
|
||||
"github_rls_url": "https://coqui.gateway.scarf.sh/v0.8.0_models/tts_models--hu--css10--vits.zip"
|
||||
},
|
||||
{
|
||||
"model_name": "tts_models--el--cv--vits",
|
||||
"folder_path": "tts_models--el--cv--vits",
|
||||
"github_rls_url": "https://coqui.gateway.scarf.sh/v0.8.0_models/tts_models--el--cv--vits.zip"
|
||||
},
|
||||
{
|
||||
"model_name": "tts_models--fi--css10--vits",
|
||||
"folder_path": "tts_models--fi--css10--vits",
|
||||
"github_rls_url": "https://coqui.gateway.scarf.sh/v0.8.0_models/tts_models--fi--css10--vits.zip"
|
||||
},
|
||||
{
|
||||
"model_name": "tts_models--hr--cv--vits",
|
||||
"folder_path": "tts_models--hr--cv--vits",
|
||||
"github_rls_url": "https://coqui.gateway.scarf.sh/v0.8.0_models/tts_models--hr--cv--vits.zip"
|
||||
},
|
||||
{
|
||||
"model_name": "tts_models--lt--cv--vits",
|
||||
"folder_path": "tts_models--lt--cv--vits",
|
||||
"github_rls_url": "https://coqui.gateway.scarf.sh/v0.8.0_models/tts_models--lt--cv--vits.zip"
|
||||
},
|
||||
{
|
||||
"model_name": "tts_models--lv--cv--vits",
|
||||
"folder_path": "tts_models--lv--cv--vits",
|
||||
"github_rls_url": "https://coqui.gateway.scarf.sh/v0.8.0_models/tts_models--lv--cv--vits.zip"
|
||||
},
|
||||
{
|
||||
"model_name": "tts_models--mt--cv--vits",
|
||||
"folder_path": "tts_models--mt--cv--vits",
|
||||
"github_rls_url": "https://coqui.gateway.scarf.sh/v0.8.0_models/tts_models--mt--cv--vits.zip"
|
||||
},
|
||||
{
|
||||
"model_name": "tts_models--pl--mai_female--vits",
|
||||
"folder_path": "tts_models--pl--mai_female--vits",
|
||||
"github_rls_url": "https://coqui.gateway.scarf.sh/v0.8.0_models/tts_models--pl--mai_female--vits.zip"
|
||||
},
|
||||
{
|
||||
"model_name": "tts_models--pt--cv--vits",
|
||||
"folder_path": "tts_models--pt--cv--vits",
|
||||
"github_rls_url": "https://coqui.gateway.scarf.sh/v0.8.0_models/tts_models--pt--cv--vits.zip"
|
||||
},
|
||||
{
|
||||
"model_name": "tts_models--ro--cv--vits",
|
||||
"folder_path": "tts_models--ro--cv--vits",
|
||||
"github_rls_url": "https://coqui.gateway.scarf.sh/v0.8.0_models/tts_models--ro--cv--vits.zip"
|
||||
},
|
||||
{
|
||||
"model_name": "tts_models--sk--cv--vits",
|
||||
"folder_path": "tts_models--sk--cv--vits",
|
||||
"github_rls_url": "https://coqui.gateway.scarf.sh/v0.8.0_models/tts_models--sk--cv--vits.zip"
|
||||
},
|
||||
{
|
||||
"model_name": "tts_models--sl--cv--vits",
|
||||
"folder_path": "tts_models--sl--cv--vits",
|
||||
"github_rls_url": "https://coqui.gateway.scarf.sh/v0.8.0_models/tts_models--sl--cv--vits.zip"
|
||||
},
|
||||
{
|
||||
"model_name": "tts_models--sv--cv--vits",
|
||||
"folder_path": "tts_models--sv--cv--vits",
|
||||
"github_rls_url": "https://coqui.gateway.scarf.sh/v0.8.0_models/tts_models--sv--cv--vits.zip"
|
||||
},
|
||||
{
|
||||
"model_name": "tts_models--ca--custom--vits",
|
||||
"folder_path": "tts_models--ca--custom--vits",
|
||||
"github_rls_url": "https://coqui.gateway.scarf.sh/v0.10.1_models/tts_models--ca--custom--vits.zip"
|
||||
},
|
||||
{
|
||||
"model_name": "tts_models--en--jenny--jenny",
|
||||
"folder_path": "tts_models--en--jenny--jenny",
|
||||
"github_rls_url": "https://coqui.gateway.scarf.sh/v0.14.0_models/tts_models--en--jenny--jenny.zip"
|
||||
},
|
||||
{
|
||||
"model_name": "tts_models--bn--custom--vits-male",
|
||||
"folder_path": "tts_models--bn--custom--vits-male",
|
||||
"github_rls_url": "https://coqui.gateway.scarf.sh/v0.13.3_models/tts_models--bn--custom--vits_male.zip"
|
||||
},
|
||||
{
|
||||
"model_name": "tts_models--bn--custom--vits-female",
|
||||
"folder_path": "tts_models--bn--custom--vits-female",
|
||||
"github_rls_url": "https://coqui.gateway.scarf.sh/v0.13.3_models/tts_models--bn--custom--vits_female.zip"
|
||||
}
|
||||
]
|
||||
}
|
||||
@@ -0,0 +1,548 @@
|
||||
###############################################
|
||||
# DONT CHANGE # These are base imports needed #
|
||||
###############################################
|
||||
import os
|
||||
import sys
|
||||
import json
|
||||
import time
|
||||
import torch
|
||||
import logging
|
||||
from pathlib import Path
|
||||
from fastapi import (HTTPException)
|
||||
logging.disable(logging.WARNING)
|
||||
###############################################
|
||||
# DONT CHANGE # Get Pytorch & Python versions #
|
||||
###############################################
|
||||
pytorch_version = torch.__version__
|
||||
cuda_version = torch.version.cuda
|
||||
major, minor, micro = sys.version_info[:3]
|
||||
python_version = f"{major}.{minor}.{micro}"
|
||||
try:
|
||||
import deepspeed
|
||||
deepspeed_available = True
|
||||
except ImportError:
|
||||
deepspeed_available = False
|
||||
pass
|
||||
|
||||
#############################################################################################################
|
||||
#############################################################################################################
|
||||
# CHANGE ME # Run any specifc imports, requirements or setup any global vaiables needed for this TTS Engine #
|
||||
#############################################################################################################
|
||||
#############################################################################################################
|
||||
# In this section you will import any imports that your specific TTS Engine will use. You will provide any
|
||||
# start-up errors for those bits, as if you were starting up a normal Python script. Note the logging.disable
|
||||
# a few lines up from here, you may want to # that out while debugging!
|
||||
try:
|
||||
from TTS.utils.synthesizer import Synthesizer
|
||||
except ModuleNotFoundError:
|
||||
print(
|
||||
f"[Startup] \033[91mWarning\033[0m Could not find the TTS module. Make sure to install the requirements for the alltalk_tts extension.",
|
||||
f"[Startup] \033[91mWarning\033[0m Linux / Mac:\npip install -r extensions/alltalk_tts/requirements.txt\n",
|
||||
f"[Startup] \033[91mWarning\033[0m Windows:\npip install -r extensions\\alltalk_tts\\requirements.txt\n",
|
||||
f"[Startup] \033[91mWarning\033[0m If you used the one-click installer, paste the command above in the terminal window launched after running the cmd_ script. On Windows, that's cmd_windows.bat."
|
||||
)
|
||||
raise
|
||||
|
||||
#############################################################
|
||||
# DONT CHANGE # Do not change the Class name from tts_class #
|
||||
#############################################################
|
||||
class tts_class:
|
||||
def __init__(self):
|
||||
########################################################################
|
||||
# DONT CHANGE # Sets up the base variables required for any tts engine #
|
||||
########################################################################
|
||||
self.branding = None
|
||||
self.this_dir = Path(__file__).parent.resolve() # Sets up self.this_dir as a variable for the folder THIS script is running in.
|
||||
self.main_dir = Path(__file__).parent.parent.parent.parent.resolve() # Sets up self.main_dir as a variable for the folder AllTalk is running in
|
||||
self.device = "cuda" if torch.cuda.is_available() else "cpu" # Sets up self.device to cuda if torch exists with Nvidia/CUDA, otherwise sets to cpu
|
||||
self.cuda_is_available = torch.cuda.is_available() # Sets up cuda_is_available as a True/False to track if Nvidia/CUDA was found on the system
|
||||
self.tts_generating_lock = False # Used to lock and unlock the tts generation process at the start/end of tts generation.
|
||||
self.tts_stop_generation = False # Used in conjunction with tts_generating_lock to call for a stop to the current generation. If called (set True) it needs to be set back to False when generation has been stopped.
|
||||
self.tts_narrator_generatingtts = False # Used to track if the current tts processes is narrator based. This can be used in conjunction with lowvram and device to avoid moving model between GPU(CUDA)<>RAM(CPU) each chunk of narrated text generated.
|
||||
self.model = None # If loading a model into CUDA/VRAM/RAM "model" is used as the variable name to load and interact with (see the XTTS model_engine script for examples.)
|
||||
self.is_tts_model_loaded = False # Used to track if a model is actually loaded in and error/fail things like TTS generation if its False
|
||||
self.current_model_loaded = None # Stores the name of the currenly loaded in model
|
||||
self.available_models = None # List of available models found by "def scan_models_folder"
|
||||
self.setup_has_run = False # Tracks if async def setup(self) has run, by setting to True, so that the /api/ready endpoint can provide a "Ready" status
|
||||
##############################################################################################
|
||||
# DONT CHANGE # Load in a list of the available TTS engines and the currently set TTS engine #
|
||||
##############################################################################################
|
||||
tts_engines_file = os.path.join(self.main_dir, "system", "tts_engines", "tts_engines.json")
|
||||
with open(tts_engines_file, "r") as f:
|
||||
tts_engines_data = json.load(f)
|
||||
self.engines_available = [engine["name"] for engine in tts_engines_data["engines_available"]] # A list of ALL the TTS engines available to be loaded by AllTalk
|
||||
self.engine_loaded = tts_engines_data["engine_loaded"] # In "tts_engines.json" what is the currently set TTS engine loading into AllTalk
|
||||
self.selected_model = tts_engines_data["selected_model"] # In "tts_engines.json" what is the currently set TTS model loading into AllTalk
|
||||
############################################################################
|
||||
# DONT CHANGE # Pull out all the settings for the currently set TTS engine #
|
||||
############################################################################
|
||||
with open(os.path.join(self.this_dir, "model_settings.json"), "r") as f:
|
||||
tts_model_loaded = json.load(f)
|
||||
# Access the model details
|
||||
self.manufacturer_name = tts_model_loaded["model_details"]["manufacturer_name"] # The company/person/body that generated the TTS engine/models etc
|
||||
self.manufacturer_website = tts_model_loaded["model_details"]["manufacturer_website"] # The website of the company/person/body where people can find more information
|
||||
# Access the features the model is capable of:
|
||||
self.audio_format = tts_model_loaded["model_capabilties"]["audio_format"] # This details the audio format your TTS engine is set to generate TTS in e.g. wav, mp3, flac, opus, acc, pcm. Please use only 1x format.
|
||||
self.deepspeed_capable = tts_model_loaded["model_capabilties"]["deepspeed_capable"] # Is your model capable of DeepSpeed
|
||||
self.deepspeed_available = 'deepspeed' in globals() # When we did the import earlier, at the top of this script, was DeepSpeed available for use
|
||||
self.generationspeed_capable = tts_model_loaded["model_capabilties"]["generationspeed_capable"] # Does this TTS engine support changing the speed of the generated TTS
|
||||
self.languages_capable = tts_model_loaded["model_capabilties"]["languages_capable"] # Are the actual models themselves capable of generating in multiple languages OR is each model language specific
|
||||
self.lowvram_capable = tts_model_loaded["model_capabilties"]["lowvram_capable"] # Is this engine capable of using low VRAM (moving the model between CPU And GPU memory)
|
||||
self.multimodel_capable = tts_model_loaded["model_capabilties"]["multimodel_capable"] # Is there just the one model or are there multiple models this engine supports.
|
||||
self.repetitionpenalty_capable = tts_model_loaded["model_capabilties"]["repetitionpenalty_capable"] # Is this TTS engine capable of changing the repititon penalty
|
||||
self.streaming_capable = tts_model_loaded["model_capabilties"]["streaming_capable"] # Is this TTS engine capabale of generating streaming audio
|
||||
self.temperature_capable = tts_model_loaded["model_capabilties"]["temperature_capable"] # Is this TTS engine capable of changing the temperature of the models
|
||||
self.multivoice_capable = tts_model_loaded["model_capabilties"]["multivoice_capable"] # Are the models multi-voice or single vocice models
|
||||
self.pitch_capable = tts_model_loaded["model_capabilties"]["pitch_capable"] # Is this TTS engine capable of changing the pitch of the genrated TTS
|
||||
# Access the current enginesettings
|
||||
self.def_character_voice = tts_model_loaded["settings"]["def_character_voice"] # What is the current default main/character voice that will be used if no voice specified.
|
||||
self.def_narrator_voice = tts_model_loaded["settings"]["def_narrator_voice"] # What is the current default narrator voice that will be used if no voice specified.
|
||||
self.deepspeed_enabled = tts_model_loaded["settings"]["deepspeed_enabled"] # If its available, is DeepSpeed enabled for the TTS engine
|
||||
self.engine_installed = tts_model_loaded["settings"]["engine_installed"] # Has the TTS engine been setup/installed (not curently used)
|
||||
self.generationspeed_set = tts_model_loaded["settings"]["generationspeed_set"] # What is the set/stored speed for generation.
|
||||
self.lowvram_enabled = tts_model_loaded["settings"]["lowvram_enabled"] # If its available, is LowVRAM enabled for the TTS engine
|
||||
# Check if someone has enabled lowvram on a system that's not CUDA enabled
|
||||
self.lowvram_enabled = False if not torch.cuda.is_available() else self.lowvram_enabled # If LowVRAM is mistakenly set and CUDA is not available, this will force it back off
|
||||
self.repetitionpenalty_set = tts_model_loaded["settings"]["repetitionpenalty_set"] # What is the currenly set repitition policy of the model (If it support repetition)
|
||||
self.temperature_set = tts_model_loaded["settings"]["temperature_set"] # What is the currenly set temperature of the model (If it support temp)
|
||||
self.pitch_set = tts_model_loaded["settings"]["pitch_set"] # What is the currenly set pitch of the model (If it support temp)
|
||||
# Gather the OpenAI API Voice Mappings
|
||||
self.openai_alloy = tts_model_loaded["openai_voices"]["alloy"] # The TTS engine voice that will be mapped to Open AI Alloy voice
|
||||
self.openai_echo = tts_model_loaded["openai_voices"]["echo"] # The TTS engine voice that will be mapped to Open AI Echo voice
|
||||
self.openai_fable = tts_model_loaded["openai_voices"]["fable"] # The TTS engine voice that will be mapped to Open AI Fable voice
|
||||
self.openai_nova = tts_model_loaded["openai_voices"]["nova"] # The TTS engine voice that will be mapped to Open AI Nova voice
|
||||
self.openai_onyx = tts_model_loaded["openai_voices"]["onyx"] # The TTS engine voice that will be mapped to Open AI Onyx voice
|
||||
self.openai_shimmer = tts_model_loaded["openai_voices"]["shimmer"] # The TTS engine voice that will be mapped to Open AI Shimmer voice
|
||||
###################################################################
|
||||
# DONT CHANGE # Load params and api_defaults from confignew.json #
|
||||
###################################################################
|
||||
# Define the path to the confignew.json file
|
||||
configfile_path = self.main_dir / "confignew.json"
|
||||
# Load config file and get settings
|
||||
with open(configfile_path, "r") as configfile:
|
||||
configfile_data = json.load(configfile)
|
||||
self.branding = configfile_data.get("branding", "") # Sets up self.branding for outputting the name stored in the "confgnew.json" file, as used in print statements.
|
||||
self.params = configfile_data # Loads in the curent "confgnew.json" file to self.params.
|
||||
self.debug_tts = configfile_data.get("debugging").get("debug_tts") # Can be used within this script as a True/False flag for generally debugging the TTS generation process.
|
||||
self.debug_tts_variables = configfile_data.get("debugging").get("debug_tts_variables") # Can be used within this script as a True/False flag for generally debugging variables (if you wish).
|
||||
################################################################
|
||||
# DONT CHANGE # Print out Python, CUDA, DeepSpeed versions ####
|
||||
################################################################
|
||||
def printout_versions(self):
|
||||
if deepspeed_available:
|
||||
print(f"[{self.branding}ENG] \033[92mDeepSpeed version :\033[93m",deepspeed.__version__,"\033[0m")
|
||||
else:
|
||||
print(f"[{self.branding}ENG] \033[92mDeepSpeed version :\033[93m Not available\033[0m")
|
||||
print(f"[{self.branding}ENG] \033[92mPython Version :\033[93m {python_version}\033[0m")
|
||||
print(f"[{self.branding}ENG] \033[92mPyTorch Version :\033[93m {pytorch_version}\033[0m")
|
||||
if cuda_version is None:
|
||||
print(f"[{self.branding}ENG] \033[92mCUDA Version :\033[91m Not available\033[0m")
|
||||
else:
|
||||
print(f"[{self.branding}ENG] \033[92mCUDA Version :\033[93m {cuda_version}\033[0m")
|
||||
print(f"[{self.branding}ENG]")
|
||||
return
|
||||
|
||||
###################################################################################
|
||||
###################################################################################
|
||||
# CHANGE ME # Inital setup of the model and engine. Called when the script starts #
|
||||
###################################################################################
|
||||
###################################################################################
|
||||
# In here you will add code to load in your model and do its inital setup. So you
|
||||
# may be calling your model loader via handle_tts_method_change or if your TTS
|
||||
# engine doesnt actually load a model into CUDA or System RAM, you may be doing
|
||||
# Something to fake its start-up.
|
||||
async def setup(self):
|
||||
self.printout_versions()
|
||||
self.available_models = self.scan_models_folder()
|
||||
# ↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑
|
||||
# ↑↑↑ Keep everything above this line ↑↑↑
|
||||
# ↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑
|
||||
|
||||
print("in setup - self.selected_model:", self.selected_model) if self.debug_tts else None
|
||||
if self.selected_model:
|
||||
tts_model = f"{self.selected_model}"
|
||||
print("in setup - tts_model:", tts_model) if self.debug_tts else None
|
||||
if tts_model in self.available_models:
|
||||
print("in setup - sending to handle_tts_method_change:") if self.debug_tts else None
|
||||
await self.handle_tts_method_change(tts_model)
|
||||
self.current_model_loaded = tts_model
|
||||
|
||||
# ↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓
|
||||
# ↓↓↓ Keep everything below this line ↓↓↓
|
||||
# ↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓
|
||||
else:
|
||||
self.current_model_loaded = "No Models Available"
|
||||
print(f"[{self.branding}ENG] \033[91mError\033[0m: Selected model '{self.selected_model}' not found in the models folder.")
|
||||
self.setup_has_run = True # Set to True, so that the /api/ready endpoint can provide a "Ready" status
|
||||
|
||||
##################################
|
||||
##################################
|
||||
# CHANGE ME # Low VRAM Swapping #
|
||||
##################################
|
||||
##################################
|
||||
# If your model does load into CUDA and you want to support LowVRAM, aka moving the model
|
||||
# on the fly between CUDA and System RAM on each generation request, you will be adding
|
||||
# The code in here to do it. See the XTTS tts engine for an example. Piper however, doesnt
|
||||
# Load into VRAM or System RAM, also low vram is set globally disabled by the model_settings.JSON
|
||||
# file, so this would never get called anywyay, however, we still need to keep an empty
|
||||
# function in place. Piper TTS uses "pass" telling function to just exit out cleanly if called.
|
||||
# However, its quite a simple check along the lines of "if CUDA is available and model is
|
||||
# in X place, then send it to Y place (or Y to X).
|
||||
async def handle_lowvram_change(self):
|
||||
if torch.cuda.is_available():
|
||||
if self.device == "cuda":
|
||||
self.device = "cpu"
|
||||
self.model.to(self.device)
|
||||
torch.cuda.empty_cache()
|
||||
else:
|
||||
self.device == "cpu"
|
||||
self.device = "cuda"
|
||||
self.model.to(self.device)
|
||||
|
||||
########################################
|
||||
########################################
|
||||
# CHANGE ME # DeepSpeed model loading #
|
||||
########################################
|
||||
########################################
|
||||
# If the model supports CUDA and DeepSpeed, this is where you can handle re-loading
|
||||
# the model as/when DeepSpeed is enabled/selected in the user interface. If your
|
||||
# TTS model doesnt support DeepSpeed, then it should be globally set in your
|
||||
# model_settings.JSON and this function will never get called, however it still needs
|
||||
# to exist as a function.
|
||||
async def handle_deepspeed_change(self, value):
|
||||
if value:
|
||||
# DeepSpeed enabled
|
||||
print(f"[{self.branding}ENG] \033[93mDeepSpeed Activating\033[0m")
|
||||
await self.unload_model()
|
||||
self.params["tts_method_api_local"] = False
|
||||
self.deepspeed_enabled = True
|
||||
await self.setup()
|
||||
else:
|
||||
# DeepSpeed disabled
|
||||
print(f"[{self.branding}ENG] \033[93mDeepSpeed De-Activating\033[0m")
|
||||
self.deepspeed_enabled = False
|
||||
await self.unload_model()
|
||||
await self.setup()
|
||||
return value # Return new checkbox value
|
||||
|
||||
##############################################################################################################################################
|
||||
##############################################################################################################################################
|
||||
# CHANGE ME # scan for available models/voices that are relevant to this TTS engine # XTTS is very unusal as it has 2x model loading methods #
|
||||
##############################################################################################################################################
|
||||
##############################################################################################################################################
|
||||
# This function looks and reports back the list of possible models your TTS engine can
|
||||
# load in. Some TTS engines have multiple models they can load and you will want to use
|
||||
# code for checking if the models are in the correct location/placement within the disk
|
||||
# the correct files exst per model etc (check XTTS for an example of this). Some models
|
||||
# like Piper, the actual models are the voices, so in the Piper scan_models_folder
|
||||
# function, we fake the only model being available as {'piper': 'piper'} aka, model name
|
||||
# then engine name, then we use the voices_file_list to populate the models as available
|
||||
# voices that can be selected in the interface.
|
||||
# If no models are found, we return "No Models Available" and continue on with the script.
|
||||
def scan_models_folder(self):
|
||||
models_folder = self.main_dir / "models" / "vits" # Edit to match the name of your folder where voices are stored
|
||||
self.available_models = {}
|
||||
# ↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑
|
||||
# ↑↑↑ Keep everything above this line ↑↑↑
|
||||
# ↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑
|
||||
|
||||
required_files_multi_speaker = ["config.json", "model_file.pth", "speaker_ids.json"]
|
||||
required_files_single_speaker = ["config.json", "model_file.pth"]
|
||||
required_files_multi_language = ["config.json", "model_file.pth.tar", "speaker_ids.json", "language_ids.json"]
|
||||
for subfolder in models_folder.iterdir():
|
||||
if subfolder.is_dir():
|
||||
model_name = subfolder.name
|
||||
if all(subfolder.joinpath(file).exists() for file in required_files_multi_speaker):
|
||||
self.available_models[f"vits - {model_name}"] = "vits_multi_speaker"
|
||||
elif all(subfolder.joinpath(file).exists() for file in required_files_single_speaker):
|
||||
self.available_models[f"vits - {model_name}"] = "vits_single_speaker"
|
||||
elif all(subfolder.joinpath(file).exists() for file in required_files_multi_language):
|
||||
self.available_models[f"vits - {model_name}"] = "vits_multi_language"
|
||||
|
||||
# ↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓
|
||||
# ↓↓↓ Keep everything below this line ↓↓↓
|
||||
# ↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓
|
||||
else:
|
||||
self.available_models = {'No Models Available': 'vits'} # Change the name in here to your TTS engine name
|
||||
print(f"[{self.branding}ENG] \033[91mWarning\033[0m: Model folder '{model_name}' is missing required")
|
||||
print(f"[{self.branding}ENG] \033[91mWarning\033[0m: files or the folder does not exist.")
|
||||
return self.available_models
|
||||
|
||||
#############################################################
|
||||
#############################################################
|
||||
# CHANGE ME # POPULATE FILES LIST FROM VOICES DIRECTORY ####
|
||||
#############################################################
|
||||
#############################################################
|
||||
# This function looks and reports back the list of possible voice your TTS engine can
|
||||
# load in. Some TTS engines the voices are wav file samples (XTTS), some are models
|
||||
# (Piper) and some are text (Parler) thats stored in a JSON file. We just need to
|
||||
# populate the "voices" variable somehow and if no voices are found, we return
|
||||
# "No Voices Found" back to the interface/api.
|
||||
def voices_file_list(self):
|
||||
try:
|
||||
voices = []
|
||||
# ↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑
|
||||
# ↑↑↑ Keep everything above this line ↑↑↑
|
||||
# ↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑
|
||||
|
||||
model_name = self.selected_model.replace("vits - ", "")
|
||||
print("model_name", model_name) if self.debug_tts else None
|
||||
model_type = self.available_models[self.selected_model]
|
||||
print("model_type", model_type) if self.debug_tts else None
|
||||
|
||||
if model_type == "vits_multi_speaker" or model_type == "vits_multi_language":
|
||||
speaker_ids_path = self.main_dir / "models" / "vits" / model_name / "speaker_ids.json"
|
||||
if speaker_ids_path.exists():
|
||||
with open(speaker_ids_path, "r") as f:
|
||||
speaker_ids = json.load(f)
|
||||
voices = [speaker for speaker in speaker_ids.keys() if speaker != "ED\n"]
|
||||
else:
|
||||
print(f"[{self.branding}ENG] \033[91mWarning\033[0m: speaker_ids.json not found for model \033[0m'{model_name}'\033[0m.") if self.debug_tts else None
|
||||
|
||||
elif model_type == "vits_single_speaker":
|
||||
voices = ["default"] # Return a list containing only the default voice for single-speaker models
|
||||
|
||||
# ↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓
|
||||
# ↓↓↓ Keep everything below this line ↓↓↓
|
||||
# ↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓
|
||||
if not voices: # If the voices list is empty
|
||||
return ["No Voices Found"] # Return a list with 'No Voices Found' if there are no voices/voice models
|
||||
return voices # Return the list of voices
|
||||
except Exception as e:
|
||||
print(f"[{self.branding}ENG] \033[91mError\033[0m: Voices/Voice Models not found. Cannot load a list of voices.")
|
||||
print(f"[{self.branding}ENG] Exception: {str(e)}") # Optional: Print the exception message for debugging
|
||||
return ["No Voices Found"]
|
||||
|
||||
#############################
|
||||
#############################
|
||||
# CHANGE ME # Model loading #
|
||||
#############################
|
||||
#############################
|
||||
# This function will handle the loading of your model, into VRAM/CUDA, System RAM or whatever.
|
||||
# In XTTS, which has 2x model loader types, there are 2x loaders. They are called by "def handle_tts_method_change"
|
||||
# In Piper we fake a model loader as Piper doesnt actually load a model into CUDA/System RAM as such. So, in that
|
||||
# situation, api_manual_load_model is kind of a blank function. Though we do set self.is_tts_model_loaded = True
|
||||
# as this is used elsewhere in the scripts to confirm that a model is available to be used for TTS generation.
|
||||
# We always check for "No Models Available" being sent as that means we are trying to load in a model that
|
||||
# doesnt exist/wasnt found on script start-up e.g. someone deleted the model from the folder or something.
|
||||
async def api_manual_load_model(self, model_name):
|
||||
if "No Models Available" in self.available_models:
|
||||
print(f"[{self.branding}ENG] \033[91mError\033[0m: No models for this TTS engine were found to load. Please download a model.")
|
||||
return
|
||||
model_path = self.main_dir / "models" / "vits" / model_name # change the name to the correct folder name
|
||||
print("model_path is:", model_path) if self.debug_tts_variables else None
|
||||
# ↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑
|
||||
# ↑↑↑ Keep everything above this line ↑↑↑
|
||||
# ↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑
|
||||
|
||||
# Load the config file
|
||||
config_file_path = model_path / "config.json"
|
||||
with open(config_file_path, "r") as f:
|
||||
config_dict = json.load(f)
|
||||
# Check if the model is single-speaker, multi-speaker, or multi-language
|
||||
model_type = self.available_models[f"vits - {model_name}"]
|
||||
use_speaker_embedding = config_dict["model_args"]["use_speaker_embedding"]
|
||||
num_speakers = config_dict["model_args"]["num_speakers"]
|
||||
if model_type == "vits_multi_speaker" or model_type == "vits_multi_language":
|
||||
# Update the speakers_file path in the config dictionary
|
||||
config_dict["model_args"]["speakers_file"] = str(model_path / "speaker_ids.json")
|
||||
speakers_file = model_path / "speaker_ids.json"
|
||||
else:
|
||||
# Set speakers_file to None for single-speaker models
|
||||
config_dict["model_args"]["speakers_file"] = None
|
||||
speakers_file = None
|
||||
# Save the updated config dictionary back to the config.json file
|
||||
with open(config_file_path, "w") as f:
|
||||
json.dump(config_dict, f, indent=4)
|
||||
if model_type == "vits_multi_language":
|
||||
# Load multi-language model
|
||||
self.model = Synthesizer(
|
||||
tts_checkpoint=model_path / "model_file.pth.tar",
|
||||
tts_config_path=config_file_path,
|
||||
tts_speakers_file=speakers_file,
|
||||
tts_languages_file=model_path / "language_ids.json",
|
||||
use_cuda=self.cuda_is_available,
|
||||
)
|
||||
elif not use_speaker_embedding and num_speakers == 0:
|
||||
# Single-speaker model without speaker embeddings
|
||||
self.model = Synthesizer(
|
||||
tts_checkpoint=model_path / "model_file.pth",
|
||||
tts_config_path=config_file_path,
|
||||
use_cuda=self.cuda_is_available,
|
||||
)
|
||||
else:
|
||||
# Multi-speaker model or single-speaker model with speaker embeddings
|
||||
self.model = Synthesizer(
|
||||
tts_checkpoint=model_path / "model_file.pth",
|
||||
tts_config_path=config_file_path,
|
||||
tts_speakers_file=speakers_file,
|
||||
use_cuda=self.cuda_is_available,
|
||||
)
|
||||
|
||||
# ↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓
|
||||
# ↓↓↓ Keep everything below this line ↓↓↓
|
||||
# ↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓
|
||||
self.is_tts_model_loaded = True
|
||||
return self.model
|
||||
|
||||
###############################
|
||||
###############################
|
||||
# CHANGE ME # Model unloading #
|
||||
###############################
|
||||
###############################
|
||||
# This function will handle the UN-loading of your model, from VRAM/CUDA, System RAM or whatever.
|
||||
# In XTTS, that model loads into CUDA/System Ram, so when we swap models, we want to unload the current model
|
||||
# free up the memory and then load in the new model to VRAM/CUDA. On the flip side of that, Piper doesnt
|
||||
# doesnt load into memory, so we just need to put a fake function here that doesnt really do anything
|
||||
# other than set "self.is_tts_model_loaded = False", which would be set back to true by the model loader.
|
||||
# So look at the Piper model_engine.py if you DONT need to unload models.
|
||||
async def unload_model(self):
|
||||
self.is_tts_model_loaded = False
|
||||
if not self.current_model_loaded == None:
|
||||
print(f"[{self.branding}ENG] \033[94mUnloading model \033[0m") if self.debug_tts else None
|
||||
if hasattr(self, 'model'):
|
||||
del self.model
|
||||
if torch.cuda.is_available():
|
||||
torch.cuda.empty_cache()
|
||||
return None
|
||||
|
||||
###################################################################################################################################
|
||||
###################################################################################################################################
|
||||
# CHANGE ME # Model changing. Unload out old model and load in a new one # XTTS is very unusal as it has 2x model loading methods #
|
||||
###################################################################################################################################
|
||||
###################################################################################################################################
|
||||
# This function is your central model loading/unloading handler that deals with the above functions as necesary, to call loading, unloading,
|
||||
# swappng DeepSpeed, Low vram etc. This function gets called with a "engine name - model name" type call. In XTTS, because there are 2x
|
||||
# model loader types, (XTTS and APILocal), we take tts_method and split the "engine name - model name" into a loader type and the model
|
||||
# that it needs to load in and then we call the correct loader function. Whereas in Piper, which doesnt load models into memory at all,
|
||||
# we just have a fake function that doesnt really do anything. We always check to see if the model name has "No Models Available" in the
|
||||
# name thats sent over, just to catch any potential errors. We display the start load time and end load time. Thats about it.
|
||||
async def handle_tts_method_change(self, tts_method):
|
||||
generate_start_time = time.time() # Record the start time of loading the model
|
||||
if "No Models Available" in self.available_models:
|
||||
print(f"[{self.branding}ENG] \033[91mError\033[0m: No models for this TTS engine were found to load. Please download a model.")
|
||||
return False
|
||||
# ↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑
|
||||
# ↑↑↑ Keep everything above this line ↑↑↑
|
||||
# ↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑
|
||||
|
||||
await self.unload_model()
|
||||
# Set other parameters to False
|
||||
if tts_method.startswith("vits"):
|
||||
model_name = tts_method.split(" - ")[1]
|
||||
print(f"[{self.branding}ENG]\033[94m Model/Engine :\033[93m {model_name}\033[94m loading into\033[93m", self.device,"\033[0m")
|
||||
self.model = await self.api_manual_load_model(model_name)
|
||||
self.current_model_loaded = f"vits - {model_name}"
|
||||
self.selected_model = tts_method # Update self.selected_model
|
||||
else:
|
||||
self.current_model_loaded = None
|
||||
return False
|
||||
|
||||
# ↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓
|
||||
# ↓↓↓ Keep everything below this line ↓↓↓
|
||||
# ↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓
|
||||
generate_end_time = time.time() # Create an end timer for calculating load times
|
||||
generate_elapsed_time = generate_end_time - generate_start_time # Calculate start time minus end time
|
||||
print(f"[{self.branding}ENG] \033[94mLoad time :\033[93m {generate_elapsed_time:.2f} seconds.\033[0m") # Print out the result of the load time
|
||||
return True
|
||||
|
||||
##########################################################################################################################################
|
||||
##########################################################################################################################################
|
||||
# CHANGE ME # Model changing. Unload out old model and load in a new one # XTTS is very unusal as it has 2x model TTS generation methods #
|
||||
##########################################################################################################################################
|
||||
##########################################################################################################################################
|
||||
# In here all the possible options are sent over (text, voice to use, lanugage, speed etc etc) and its up to you how you use them, or not.
|
||||
# obviously if your TTS engine doesnt support speed for example, generationspeed_capable should be set False in your model_settings.JSON file
|
||||
# and a fake "generationspeed_set" value should be set. This allows a fake value to be sent over from the main script, even though it
|
||||
# wouldnt actually ever be used in the generation below. Nonethless all these values, real or just whats inside the configuration file
|
||||
# will be sent over for use.
|
||||
# Setting the xxxxxxx_capabale in the model_settings.JSON file, will enable/disable them being selectable by the user. For example, if you
|
||||
# set "generationspeed_capable" as false in the model_settings.JSON file, a user will not be able to select OR set the setting for
|
||||
# generation speed.
|
||||
# One thing to note is that we HAVE to have something in this generation request that is synchronous from the way its called, which means
|
||||
# we have to have an option for Streaming, even if our TTS engine doesnt support streaming. So in that case, we would set streaming_capable
|
||||
# as false in our model_settings.JSON file, meaning streaming will never be called. However, we have to put a fake streaming routine in our
|
||||
# function below (or a real function if it does support streaming of course). Parler has an example of a fake streaming function, which is
|
||||
# very clearly highlighted in its model_engine.py script.
|
||||
# Piper TTS, which uses command line based calls and therefore has different ones for Windows and Linux/Mac, has an example of doing this
|
||||
# within its model_engine.py file.
|
||||
async def generate_tts(self, text, voice, language, temperature, repetition_penalty, speed, pitch, output_file, streaming):
|
||||
print(f"[{self.branding}Debug] Entered model_engine.py generate_tts function") if self.debug_tts else None
|
||||
if not self.is_tts_model_loaded: # Check if a model is loaded and error out if not.
|
||||
error_message = f"[{self.branding}ENG] \033[91mError\033[0m: You currently have no TTS model loaded."
|
||||
print(error_message)
|
||||
raise HTTPException(status_code=400, detail="You currently have no TTS model loaded.") # Raise an exception with a meaningful HTTP status code
|
||||
self.tts_generating_lock = True # Set the tts_generating lock to True, which stops other generation requests being sent into the pipeline
|
||||
print(f"[{self.branding}Debug] Checking low VRAM") if self.debug_tts else None
|
||||
if self.lowvram_enabled and self.device == "cpu": # If necessary, move the model out of System Ram to VRAM
|
||||
print(f"[{self.branding}Debug] Switching device") if self.debug_tts else None
|
||||
await self.handle_lowvram_change()
|
||||
print(f"[{self.branding}Debug] Setting a generate time") if self.debug_tts else None
|
||||
generate_start_time = time.time() # Record the start time of generating TTS
|
||||
# ↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑
|
||||
# ↑↑↑ Keep everything above this line ↑↑↑
|
||||
# ↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑
|
||||
|
||||
if self.current_model_loaded.startswith("vits"):
|
||||
print(f"[{self.branding}Debug] generate_tts Text arriving at engine {text}") if self.debug_tts else None
|
||||
model_type = self.available_models[self.current_model_loaded]
|
||||
if model_type == "vits_multi_speaker":
|
||||
# Get the available speakers for the current model
|
||||
model_name = self.current_model_loaded.replace("vits - ", "")
|
||||
speaker_ids_path = self.main_dir / "models" / "vits" / model_name / "speaker_ids.json"
|
||||
if speaker_ids_path.exists():
|
||||
with open(speaker_ids_path, "r") as f:
|
||||
speaker_ids = json.load(f)
|
||||
available_speakers = [speaker for speaker in speaker_ids.keys() if speaker != "ED\n"]
|
||||
else:
|
||||
print(f"[{self.branding}ENG] \033[91mWarning\033[0m: speaker_ids.json not found for model '{model_name}'.")
|
||||
available_speakers = []
|
||||
|
||||
# Check if the specified voice matches any of the available speakers
|
||||
if voice in available_speakers:
|
||||
speaker_idx = voice
|
||||
else:
|
||||
# Use the first available speaker as the default
|
||||
if available_speakers:
|
||||
speaker_idx = available_speakers[0]
|
||||
print(f"[{self.branding}ENG] Incorrect or no voice specified, using default voice: {speaker_idx}")
|
||||
else:
|
||||
speaker_idx = None
|
||||
print(f"[{self.branding}ENG] No valid speakers found for the current model.")
|
||||
|
||||
print(f"[{self.branding}Debug] generate_tts speaker_idx is:", speaker_idx) if self.debug_tts else None
|
||||
print(f"[{self.branding}Debug] generate_tts output_file is:", output_file) if self.debug_tts else None
|
||||
|
||||
# Generate TTS with speaker_name argument
|
||||
if speaker_idx is not None:
|
||||
wavs = self.model.tts(text, speaker_name=speaker_idx)
|
||||
else:
|
||||
wavs = None
|
||||
else:
|
||||
# Generate TTS without speaker_name argument for single-speaker models
|
||||
wavs = self.model.tts(text)
|
||||
|
||||
streaming = False
|
||||
if streaming and wavs is not None:
|
||||
# Streaming-specific operations
|
||||
for wav_chunk in wavs:
|
||||
yield wav_chunk.numpy().tobytes()
|
||||
else:
|
||||
if wavs is not None:
|
||||
# Save the generated audio to a file
|
||||
self.model.save_wav(wavs, output_file)
|
||||
yield b"" # Yield an empty byte string to indicate completion
|
||||
|
||||
# ↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓
|
||||
# ↓↓↓ Keep everything below this line ↓↓↓
|
||||
# ↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓
|
||||
generate_end_time = time.time() # Record the end time to generate TTS
|
||||
generate_elapsed_time = generate_end_time - generate_start_time
|
||||
print(f"[{self.branding}GEN] \033[94mTTS Generate: \033[93m{generate_elapsed_time:.2f} seconds. \033[94mLowVRAM: \033[33m{self.lowvram_enabled} \033[94mDeepSpeed: \033[33m{self.deepspeed_enabled}\033[0m")
|
||||
if self.lowvram_enabled and self.device == "cuda" and self.tts_narrator_generatingtts == False:
|
||||
await self.handle_lowvram_change()
|
||||
self.tts_generating_lock = False # Unlock the TTS generation queue to allow TTS generation requests to come in again.
|
||||
print(f"[{self.branding}Debug] generate_tts function completed") if self.debug_tts else None
|
||||
return
|
||||
|
||||
|
||||
@@ -0,0 +1,42 @@
|
||||
{
|
||||
"model_details": {
|
||||
"manufacturer_name": "Coqui",
|
||||
"manufacturer_website": "https://github.com/coqui-ai/TTS and also updated engine https://github.com/idiap/coqui-ai-TTS/",
|
||||
"model_description": "VITS models come in single and multi-speaker variants. VITS (Conditional Variational Autoencoder with Adversarial Learning for End-to-End Text-to-Speech ) is an End-to-End (encoder -> vocoder together) TTS model that takes advantage of SOTA DL techniques like GANs, VAE, Normalizing Flows. It does not require external alignment annotations and learns the text-to-audio alignment using MAS, as explained in the paper. The model architecture is a combination of GlowTTS encoder and HiFiGAN vocoder. It is a feed-forward model with x67.12 real-time factor on a GPU. \n\ntts_models--en--vctk--vits is a good example of a multi speaker VITS model."
|
||||
},
|
||||
"model_capabilties": {
|
||||
"audio_format": "wav",
|
||||
"deepspeed_capable": false,
|
||||
"generationspeed_capable": false,
|
||||
"languages_capable": false,
|
||||
"lowvram_capable": true,
|
||||
"multimodel_capable": true,
|
||||
"repetitionpenalty_capable": false,
|
||||
"streaming_capable": false,
|
||||
"temperature_capable": false,
|
||||
"multivoice_capable": true,
|
||||
"pitch_capable": false,
|
||||
"windows_capable": true,
|
||||
"linux_capable": true,
|
||||
"mac_capable": true
|
||||
},
|
||||
"settings": {
|
||||
"def_character_voice": "p225",
|
||||
"def_narrator_voice": "p226",
|
||||
"deepspeed_enabled": false,
|
||||
"engine_installed": true,
|
||||
"generationspeed_set": 1,
|
||||
"lowvram_enabled": true,
|
||||
"pitch_set": 0,
|
||||
"repetitionpenalty_set": 10,
|
||||
"temperature_set": 0.75
|
||||
},
|
||||
"openai_voices": {
|
||||
"alloy": "p225",
|
||||
"echo": "p225",
|
||||
"fable": "p225",
|
||||
"nova": "p225",
|
||||
"onyx": "p225",
|
||||
"shimmer": "p225"
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,395 @@
|
||||
import os
|
||||
import json
|
||||
import shutil
|
||||
import zipfile
|
||||
import requests
|
||||
import gradio as gr
|
||||
from tqdm import tqdm
|
||||
from pathlib import Path
|
||||
this_dir = Path(__file__).parent.resolve() # Sets up self.this_dir as a variable for the folder THIS script is running in.
|
||||
main_dir = Path(__file__).parent.parent.parent.parent.resolve() # Sets up self.main_dir as a variable for the folder AllTalk is running in
|
||||
|
||||
##########################################################################
|
||||
# REQUIRED CHANGE #
|
||||
# Populate the voices list, using the method specific to your TTS engine #
|
||||
##########################################################################
|
||||
# This function is responsible for populating the list of available voices for your TTS engine.
|
||||
# You need to modify this function to use the appropriate method for your engine to retrieve the voice list.
|
||||
#
|
||||
# The current implementation lists all the WAV files in a "voices" directory, which may not be suitable for your engine.
|
||||
# You should replace the `xxxx_voices_file_list` function name to match your engine name. For example, if your engine
|
||||
# is named "mytts", the function should be named `mytts_voices_file_list`.
|
||||
#
|
||||
# You will also neef to update the code with your own implementation that retrieves the voice list according to your
|
||||
# engine's specific requirements. Typically this is the same code as will be in your model_engine.py file.
|
||||
#
|
||||
# For example, if your engine has a dedicated API or configuration file for managing voices, you should modify this
|
||||
# function to interact with that API or read from that configuration file.
|
||||
#
|
||||
# After making the necessary changes, this function should return a list of available voices that can be used
|
||||
# in your TTS engine's settings page.
|
||||
def vits_voices_file_list():
|
||||
models_folder = main_dir / "models" / "vits"
|
||||
# Create the folder if it doesn't exist
|
||||
if not models_folder.exists():
|
||||
models_folder.mkdir(parents=True, exist_ok=True)
|
||||
voices = []
|
||||
available_models = {}
|
||||
required_files_multi_speaker = ["config.json", "model_file.pth", "speaker_ids.json"]
|
||||
required_files_single_speaker = ["config.json", "model_file.pth"]
|
||||
required_files_multi_language = ["config.json", "model_file.pth.tar", "speaker_ids.json", "language_ids.json"]
|
||||
for subfolder in models_folder.iterdir():
|
||||
if subfolder.is_dir():
|
||||
model_name = subfolder.name
|
||||
if all(subfolder.joinpath(file).exists() for file in required_files_multi_speaker):
|
||||
available_models[f"VITS - {model_name}"] = "VITS_multi_speaker"
|
||||
elif all(subfolder.joinpath(file).exists() for file in required_files_single_speaker):
|
||||
available_models[f"VITS - {model_name}"] = "VITS_single_speaker"
|
||||
elif all(subfolder.joinpath(file).exists() for file in required_files_multi_language):
|
||||
available_models[f"VITS - {model_name}"] = "VITS_multi_language"
|
||||
else:
|
||||
print(f"[AllTalk ENG] \033[91mWarning\033[0m: Model '{model_name}' does not match any known model type.")
|
||||
model_type = available_models.get(f"VITS - {model_name}")
|
||||
if model_type == "VITS_multi_speaker" or model_type == "VITS_multi_language":
|
||||
speaker_ids_path = subfolder / "speaker_ids.json"
|
||||
if speaker_ids_path.exists():
|
||||
with open(speaker_ids_path, "r") as f:
|
||||
speaker_ids = json.load(f)
|
||||
speakers = [speaker for speaker in speaker_ids.keys() if speaker != "ED\n"]
|
||||
voices.extend(speakers)
|
||||
else:
|
||||
print(f"[AllTalk ENG] \033[91mWarning\033[0m: speaker_ids.json not found for model '{model_name}'.")
|
||||
elif model_type == "VITS_single_speaker":
|
||||
voices.append("default")
|
||||
# Remove duplicate speaker names
|
||||
voices = list(set(voices))
|
||||
# Sort the voices alphabetically
|
||||
voices.sort()
|
||||
# Move the "default" voice to the top of the list
|
||||
if "default" in voices:
|
||||
voices.remove("default")
|
||||
voices.insert(0, "default")
|
||||
return voices
|
||||
|
||||
######################################################
|
||||
# REQUIRED CHANGE #
|
||||
# Imports and saves the TTS engine-specific settings #
|
||||
######################################################
|
||||
# This function is responsible for importing and saving the settings specific to your TTS engine.
|
||||
# You need to make the following change:
|
||||
#
|
||||
# 1. Change the name of the function `xxxx_model_update_settings` to match your engine's name.
|
||||
# For example, if your engine is named "mytts", the function should be named `mytts_model_update_settings`.
|
||||
#
|
||||
# After making this change, the function will load the model settings from a JSON file, update the settings and voice
|
||||
# dictionaries with the values provided as arguments, and save the updated settings back to the JSON file.
|
||||
#
|
||||
# You do not need to modify the function's logic or any other part of the code.
|
||||
def vits_model_update_settings(def_character_voice_gr, def_narrator_voice_gr, lowvram_enabled_gr, deepspeed_enabled_gr, temperature_set_gr, repetitionpenalty_set_gr, pitch_set_gr, generationspeed_set_gr, alloy_gr, echo_gr, fable_gr, nova_gr, onyx_gr, shimmer_gr):
|
||||
# Load the model_config_data from the JSON file
|
||||
with open(os.path.join(this_dir, "model_settings.json"), "r") as f:
|
||||
model_config_data = json.load(f)
|
||||
# Update the settings and openai_voices dictionaries with the new values
|
||||
model_config_data["settings"]["def_character_voice"] = def_character_voice_gr
|
||||
model_config_data["settings"]["def_narrator_voice"] = def_narrator_voice_gr
|
||||
model_config_data["openai_voices"]["alloy"] = alloy_gr
|
||||
model_config_data["openai_voices"]["echo"] = echo_gr
|
||||
model_config_data["openai_voices"]["fable"] = fable_gr
|
||||
model_config_data["openai_voices"]["nova"] = nova_gr
|
||||
model_config_data["openai_voices"]["onyx"] = onyx_gr
|
||||
model_config_data["openai_voices"]["shimmer"] = shimmer_gr
|
||||
model_config_data["settings"]["lowvram_enabled"] = lowvram_enabled_gr == "Enabled"
|
||||
model_config_data["settings"]["deepspeed_enabled"] = deepspeed_enabled_gr == "Enabled"
|
||||
model_config_data["settings"]["temperature_set"] = temperature_set_gr
|
||||
model_config_data["settings"]["repetitionpenalty_set"] = repetitionpenalty_set_gr
|
||||
model_config_data["settings"]["pitch_set"] = pitch_set_gr
|
||||
model_config_data["settings"]["generationspeed_set"] = generationspeed_set_gr
|
||||
# Save the updated model_config_data to the JSON file
|
||||
with open(os.path.join(this_dir, "model_settings.json"), "w") as f:
|
||||
json.dump(model_config_data, f, indent=4)
|
||||
return "Settings updated successfully!"
|
||||
|
||||
#######################################################
|
||||
# REQUIRED CHANGE #
|
||||
# Sets up the engine-specific settings page in Gradio #
|
||||
#######################################################
|
||||
# This function sets up the Gradio interface for the settings page specific to your TTS engine.
|
||||
# You need to make the following changes:
|
||||
#
|
||||
# 1. Change the name of the function `xxxx_model_alltalk_settings` to match your engine's name.
|
||||
# For example, if your engine is named "mytts", the function should be named `mytts_model_alltalk_settings`.
|
||||
#
|
||||
# 2. Change the name of the `submit_button.click` function call to match the name you gave to the function
|
||||
# that imports and saves your engine's settings (the function you modified above).
|
||||
#
|
||||
# 3. Change the name of the `voice_list` function call to match the name of the function that lists
|
||||
# the available voices for your TTS engine.
|
||||
#
|
||||
# 4. Change the 'title' of the `gr.Blocks` to match your engine's name e.g. title="mytts TTS"
|
||||
#
|
||||
# After making these changes, this function will create and return the Gradio interface for your TTS engine's
|
||||
# settings page, allowing users to configure various options and voice selections.
|
||||
def vits_model_alltalk_settings(model_config_data):
|
||||
features_list = model_config_data['model_capabilties']
|
||||
voice_list = vits_voices_file_list()
|
||||
with gr.Blocks(title="vits TTS") as app:
|
||||
with gr.Tab("Default Settings"):
|
||||
with gr.Row():
|
||||
lowvram_enabled_gr = gr.Radio(choices={"Enabled": "true", "Disabled": "false"}, label="Low VRAM" if model_config_data["model_capabilties"]["lowvram_capable"] else "Low VRAM N/A", value="Enabled" if model_config_data["settings"]["lowvram_enabled"] else "Disabled", interactive=model_config_data["model_capabilties"]["lowvram_capable"])
|
||||
deepspeed_enabled_gr = gr.Radio(choices={"Enabled": "true", "Disabled": "false"}, label="DeepSpeed Activate" if model_config_data["model_capabilties"]["deepspeed_capable"] else "DeepSpeed N/A", value="Enabled" if model_config_data["settings"]["deepspeed_enabled"] else "Disabled", interactive=model_config_data["model_capabilties"]["deepspeed_capable"])
|
||||
temperature_set_gr = gr.Slider(value=float(model_config_data["settings"]["temperature_set"]), minimum=0, maximum=1, step=0.05, label="Temperature" if model_config_data["model_capabilties"]["temperature_capable"] else "Temperature N/A", interactive=model_config_data["model_capabilties"]["temperature_capable"])
|
||||
repetitionpenalty_set_gr = gr.Slider(value=float(model_config_data["settings"]["repetitionpenalty_set"]), minimum=1, maximum=20, step=1, label="Repetition Penalty" if model_config_data["model_capabilties"]["repetitionpenalty_capable"] else "Repetition N/A", interactive=model_config_data["model_capabilties"]["repetitionpenalty_capable"])
|
||||
pitch_set_gr = gr.Slider(value=float(model_config_data["settings"]["pitch_set"]), minimum=-10, maximum=10, step=1, label="Pitch" if model_config_data["model_capabilties"]["pitch_capable"] else "Pitch N/A", interactive=model_config_data["model_capabilties"]["pitch_capable"])
|
||||
generationspeed_set_gr = gr.Slider(value=float(model_config_data["settings"]["generationspeed_set"]), minimum=0.25, maximum=2.00, step=0.25, label="Speed" if model_config_data["model_capabilties"]["generationspeed_capable"] else "Speed N/A", interactive=model_config_data["model_capabilties"]["generationspeed_capable"])
|
||||
with gr.Row():
|
||||
with gr.Column():
|
||||
gr.Markdown("### OpenAI Voice Mappings")
|
||||
with gr.Group():
|
||||
with gr.Row():
|
||||
alloy_gr = gr.Dropdown(value=model_config_data["openai_voices"]["alloy"], label="Alloy", choices=voice_list, allow_custom_value=True)
|
||||
echo_gr = gr.Dropdown(value=model_config_data["openai_voices"]["echo"], label="Echo", choices=voice_list, allow_custom_value=True)
|
||||
with gr.Row():
|
||||
fable_gr = gr.Dropdown(value=model_config_data["openai_voices"]["fable"], label="Fable", choices=voice_list, allow_custom_value=True)
|
||||
nova_gr = gr.Dropdown(value=model_config_data["openai_voices"]["nova"], label="Nova", choices=voice_list, allow_custom_value=True)
|
||||
with gr.Row():
|
||||
onyx_gr = gr.Dropdown(value=model_config_data["openai_voices"]["onyx"], label="Onyx", choices=voice_list, allow_custom_value=True)
|
||||
shimmer_gr = gr.Dropdown(value=model_config_data["openai_voices"]["shimmer"], label="Shimmer", choices=voice_list, allow_custom_value=True)
|
||||
with gr.Column():
|
||||
gr.Markdown("### Default Voices")
|
||||
with gr.Row():
|
||||
def_character_voice_gr = gr.Dropdown(value=model_config_data["settings"]["def_character_voice"], label="Default/Character Voice", choices=voice_list, allow_custom_value=True)
|
||||
def_narrator_voice_gr = gr.Dropdown(value=model_config_data["settings"]["def_narrator_voice"], label="Narrator Voice", choices=voice_list, allow_custom_value=True)
|
||||
with gr.Group():
|
||||
with gr.Row():
|
||||
details_text = gr.Textbox(label="Details", show_label=False, lines=5, interactive=False, value="In this section, you can set the default settings for this TTS engine. Settings that are not supported by the current engine will be greyed out and cannot be selected. Default voices specified here will be used when no specific voice is provided in the TTS generation request. If a voice is specified in the request, it will override these default settings. When using the OpenAI API compatable API with this TTS engine, the voice mappings will be applied. As the OpenAI API has a limited set of 6 voices, these mappings ensure compatibility by mapping the OpenAI voices to the available voices in this TTS engine.")
|
||||
with gr.Row():
|
||||
submit_button = gr.Button("Update Settings")
|
||||
output_message = gr.Textbox(label="Output Message", interactive=False, show_label=False)
|
||||
submit_button.click(vits_model_update_settings, inputs=[def_character_voice_gr, def_narrator_voice_gr, lowvram_enabled_gr, deepspeed_enabled_gr, temperature_set_gr, repetitionpenalty_set_gr, pitch_set_gr, generationspeed_set_gr, alloy_gr, echo_gr, fable_gr, nova_gr, onyx_gr, shimmer_gr], outputs=output_message)
|
||||
|
||||
###########################################################################################
|
||||
# Do not change this section apart from "TTS Engine Name" value to match your engine name #
|
||||
###########################################################################################
|
||||
with gr.Tab("Engine Information"):
|
||||
with gr.Row():
|
||||
with gr.Group():
|
||||
gr.Textbox(label="Manufacturer Name", value=model_config_data['model_details']['manufacturer_name'], interactive=False)
|
||||
gr.Textbox(label="Manufacturer Website/TTS Engine Support", value=model_config_data['model_details']['manufacturer_website'], interactive=False)
|
||||
gr.Textbox(label="Engine/Model Description", value=model_config_data['model_details']['model_description'], interactive=False, lines=13)
|
||||
with gr.Column():
|
||||
with gr.Row():
|
||||
gr.Textbox(label="DeepSpeed Capable", value='Yes' if features_list['deepspeed_capable'] else 'No', interactive=False)
|
||||
gr.Textbox(label="Pitch Capable", value='Yes' if features_list['pitch_capable'] else 'No', interactive=False)
|
||||
gr.Textbox(label="Generation Speed Capable", value='Yes' if features_list['generationspeed_capable'] else 'No', interactive=False)
|
||||
with gr.Row():
|
||||
gr.Textbox(label="Repetition Penalty Capable", value='Yes' if features_list['repetitionpenalty_capable'] else 'No', interactive=False)
|
||||
gr.Textbox(label="Multi Languages Capable", value='Yes' if features_list['languages_capable'] else 'No', interactive=False)
|
||||
gr.Textbox(label="Streaming Capable", value='Yes' if features_list['streaming_capable'] else 'No', interactive=False)
|
||||
with gr.Row():
|
||||
gr.Textbox(label="Low VRAM Capable", value='Yes' if features_list['lowvram_capable'] else 'No', interactive=False)
|
||||
gr.Textbox(label="Temperature Capable", value='Yes' if features_list['temperature_capable'] else 'No', interactive=False)
|
||||
gr.Textbox(label="Multi Model Capable Engine", value='Yes' if features_list['multimodel_capable'] else 'No', interactive=False)
|
||||
with gr.Row():
|
||||
gr.Textbox(label="Multi Voice Capable Models", value='Yes' if features_list['multivoice_capable'] else 'No', interactive=False)
|
||||
gr.Textbox(label="Default Audio output format", value=model_config_data['model_capabilties']['audio_format'], interactive=False)
|
||||
gr.Textbox(label="TTS Engine Name", value="VITS", interactive=False)
|
||||
with gr.Row():
|
||||
gr.Textbox(label="Windows Support", value='Yes' if features_list['windows_capable'] else 'No', interactive=False)
|
||||
gr.Textbox(label="Linux Support", value='Yes' if features_list['linux_capable'] else 'No', interactive=False)
|
||||
gr.Textbox(label="Mac Support", value='Yes' if features_list['mac_capable'] else 'No', interactive=False)
|
||||
with gr.Row():
|
||||
gr.Markdown("""
|
||||
#### 🟧 DeepSpeed Capable
|
||||
DeepSpeed is a deep learning optimization library that can significantly speed up model training and inference. If a model is DeepSpeed capable, it means it can utilize DeepSpeed to accelerate the generation of text-to-speech output. This requires the model to be loaded into CUDA/VRAM on an Nvidia GPU and the model's inference method to support DeepSpeed.
|
||||
#### 🟧 Pitch Capable
|
||||
Pitch refers to the highness or lowness of a sound. If a model is pitch capable, it can adjust the pitch of the generated speech, allowing for more expressive and varied output.
|
||||
#### 🟧 Generation Speed Capable
|
||||
Generation speed refers to the rate at which the model can generate text-to-speech output. If a model is generation speed capable, it means the speed of the generated speech can be adjusted, making it faster or slower depending on the desired output.
|
||||
#### 🟧 Repetition Penalty Capable
|
||||
Repetition penalty is a technique used to discourage the model from repeating the same words or phrases multiple times in the same sounding way. If a model is repetition penalty capable, it can apply this penalty during generation to improve the diversity and naturalness of the output.
|
||||
#### 🟧 Multi-Languages Capable
|
||||
Multi-language capable models can generate speech in multiple languages. This means that the model has been trained on data from different languages and can switch between them during generation. Some models are language-specific.
|
||||
#### 🟧 Multi-Voice Capable
|
||||
Multi-voice capable models generate speech in multiple voices or speaking styles. This means that the model has been trained on data from different speakers and can mimic their voices during generation, or is a voice cloning model that can generate speech based on the input sample.
|
||||
""")
|
||||
gr.Markdown("""
|
||||
#### 🟧 Streaming Capable
|
||||
Streaming refers to the ability to generate speech output in real-time, without the need to generate the entire output before playback. If a model is streaming capable, it can generate speech on-the-fly, allowing for faster response times and more interactive applications.
|
||||
#### 🟧 Low VRAM Capable
|
||||
VRAM (Video Random Access Memory) is a type of memory used by GPUs to store and process data. If a model is low VRAM capable, it can efficiently utilize the available VRAM by moving data between CPU and GPU memory as needed, allowing for generation even on systems with limited VRAM where it may be competing with an LLM model for VRAM.
|
||||
#### 🟧 Temperature Capable
|
||||
Temperature is a hyperparameter that controls the randomness of the generated output. If a model is temperature capable, the temperature can be adjusted to make the output more or less random, affecting the creativity and variability of the generated speech.
|
||||
#### 🟧 Multi-Model Capable Engine
|
||||
If an engine is multi-model capable, it means that it can support and utilize multiple models for text-to-speech generation. This allows for greater flexibility and the ability to switch between different models depending on the desired output. Different models may be capable of different languages, specific languages, voices, etc.
|
||||
#### 🟧 Default Audio Output Format
|
||||
Specifies the file format in which the generated speech will be saved. Common audio formats include WAV, MP3, FLAC, Opus, AAC, and PCM. If you want different outputs, you can set the transcode function to change the output audio, though transcoding will add a little time to the generation and is not available for streaming generation.
|
||||
#### 🟧 Windows/Linux/Mac Support
|
||||
These indicators show whether the model and engine are compatible with Windows, Linux, or macOS. However, additional setup or requirements may be necessary to ensure full functionality on your operating system. Please note that full platform support has not been extensively tested.
|
||||
""")
|
||||
|
||||
#######################################################################################################################################################################################################
|
||||
# REQUIRED CHANGE #
|
||||
# You will need to build a custom method to identify if your models are installed and download them. Store your models list in the available_models.json file, alongside your model_settings.json etc #
|
||||
#######################################################################################################################################################################################################
|
||||
with gr.Tab("Models/Voices Download"):
|
||||
with gr.Row():
|
||||
# Load the available models from the JSON file
|
||||
with open(os.path.join(this_dir, "available_models.json"), "r") as f:
|
||||
available_models = json.load(f)
|
||||
# Extract the model names for the dropdown
|
||||
model_names = [model["model_name"] for model in available_models["models"]]
|
||||
# Create the dropdown
|
||||
model_dropdown = gr.Dropdown(choices=sorted(model_names), label="Select Model", value=model_names[0])
|
||||
download_button = gr.Button("Download Model/Missing Files")
|
||||
|
||||
with gr.Row():
|
||||
download_status = gr.Textbox(label="Download Status")
|
||||
|
||||
def download_model(model_name):
|
||||
# Find the selected model in the available models
|
||||
selected_model = next(model for model in available_models["models"] if model["model_name"] == model_name)
|
||||
# Get the folder path and ZIP file URL
|
||||
vits_folder = os.path.join(main_dir, "models", "vits")
|
||||
folder_path = os.path.join(vits_folder, selected_model["folder_path"])
|
||||
zip_url = selected_model["github_rls_url"]
|
||||
# Check if the model is already downloaded based on the file structure
|
||||
required_files = ["config.json"]
|
||||
if os.path.exists(os.path.join(folder_path, "model_file.pth.tar")):
|
||||
required_files.append("model_file.pth.tar")
|
||||
else:
|
||||
required_files.append("model_file.pth")
|
||||
if os.path.exists(os.path.join(folder_path, "speaker_ids.json")):
|
||||
required_files.append("speaker_ids.json")
|
||||
if os.path.exists(os.path.join(folder_path, "language_ids.json")):
|
||||
required_files.append("language_ids.json")
|
||||
if all(os.path.exists(os.path.join(folder_path, file)) for file in required_files):
|
||||
return "The model is already downloaded. No need to download again."
|
||||
else:
|
||||
# Download the ZIP file to the "vits" folder
|
||||
print(f"Downloading {model_name}...")
|
||||
response = requests.get(zip_url, stream=True)
|
||||
total_size_in_bytes = int(response.headers.get("content-length", 0))
|
||||
block_size = 1024 # 1 Kibibyte
|
||||
progress_bar = tqdm(total=total_size_in_bytes, unit="iB", unit_scale=True)
|
||||
zip_file_path = os.path.join(vits_folder, f"{model_name}.zip")
|
||||
with open(zip_file_path, "wb") as file:
|
||||
for data in response.iter_content(block_size):
|
||||
progress_bar.update(len(data))
|
||||
file.write(data)
|
||||
progress_bar.close()
|
||||
# Extract the ZIP file contents to the desired folder
|
||||
with zipfile.ZipFile(zip_file_path, "r") as zip_ref:
|
||||
zip_ref.extractall(vits_folder)
|
||||
# Remove the ZIP file
|
||||
os.remove(zip_file_path)
|
||||
# Delete the __MACOSX folder if it exists
|
||||
macosx_folder = os.path.join(vits_folder, "__MACOSX")
|
||||
if os.path.exists(macosx_folder):
|
||||
shutil.rmtree(macosx_folder)
|
||||
return "Model downloaded successfully!"
|
||||
|
||||
def show_confirm_cancel(model_name):
|
||||
selected_model = next(model for model in available_models["models"] if model["model_name"] == model_name)
|
||||
folder_path = os.path.join(main_dir, "models", "vits", selected_model["folder_path"])
|
||||
required_files = ["config.json"]
|
||||
if os.path.exists(os.path.join(folder_path, "model_file.pth.tar")):
|
||||
required_files.append("model_file.pth.tar")
|
||||
else:
|
||||
required_files.append("model_file.pth")
|
||||
if os.path.exists(os.path.join(folder_path, "speaker_ids.json")):
|
||||
required_files.append("speaker_ids.json")
|
||||
if os.path.exists(os.path.join(folder_path, "language_ids.json")):
|
||||
required_files.append("language_ids.json")
|
||||
if all(os.path.exists(os.path.join(folder_path, file)) for file in required_files):
|
||||
return [gr.update(visible=True), gr.update(visible=True), gr.update(visible=True)]
|
||||
else:
|
||||
return [gr.update(visible=False), gr.update(visible=True), gr.update(visible=False)]
|
||||
|
||||
def confirm_download(model_name):
|
||||
# Find the selected model in the available models
|
||||
selected_model = next(model for model in available_models["models"] if model["model_name"] == model_name)
|
||||
# Get the folder path and ZIP file URL
|
||||
vits_folder = os.path.join(main_dir, "models", "vits")
|
||||
folder_path = os.path.join(vits_folder, selected_model["folder_path"])
|
||||
zip_url = selected_model["github_rls_url"]
|
||||
# Download the ZIP file to the "vits" folder
|
||||
print(f"Downloading {model_name}...")
|
||||
response = requests.get(zip_url, stream=True)
|
||||
total_size_in_bytes = int(response.headers.get("content-length", 0))
|
||||
block_size = 1024 # 1 Kibibyte
|
||||
progress_bar = tqdm(total=total_size_in_bytes, unit="iB", unit_scale=True)
|
||||
zip_file_path = os.path.join(vits_folder, f"{model_name}.zip")
|
||||
with open(zip_file_path, "wb") as file:
|
||||
for data in response.iter_content(block_size):
|
||||
progress_bar.update(len(data))
|
||||
file.write(data)
|
||||
progress_bar.close()
|
||||
# Extract the ZIP file contents to the desired folder
|
||||
with zipfile.ZipFile(zip_file_path, "r") as zip_ref:
|
||||
zip_ref.extractall(vits_folder)
|
||||
# Remove the ZIP file
|
||||
os.remove(zip_file_path)
|
||||
# Delete the __MACOSX folder if it exists
|
||||
macosx_folder = os.path.join(vits_folder, "__MACOSX")
|
||||
if os.path.exists(macosx_folder):
|
||||
shutil.rmtree(macosx_folder)
|
||||
return [gr.update(visible=False), gr.update(visible=True), gr.update(visible=False)]
|
||||
|
||||
download_button.click(download_model, inputs=model_dropdown, outputs=download_status)
|
||||
|
||||
with gr.Row():
|
||||
confirm_button = gr.Button("Download Anyway", visible=False)
|
||||
cancel_button = gr.Button("Cancel", visible=False)
|
||||
download_button.click(show_confirm_cancel, model_dropdown, [confirm_button, download_button, cancel_button])
|
||||
confirm_button.click(confirm_download, model_dropdown, [confirm_button, download_button, cancel_button])
|
||||
cancel_button.click(lambda: [gr.update(visible=False), gr.update(visible=True), gr.update(visible=False)], None, [confirm_button, download_button, cancel_button])
|
||||
|
||||
###################################################################################################
|
||||
# REQUIRED CHANGE #
|
||||
# Add any engine specific help, bugs, issues, operating system specifc requirements/setup in here #
|
||||
# Please use Markdown format, so gr.Markdown() with your markdown inside it. #
|
||||
###################################################################################################
|
||||
with gr.Tab("Engine Help"):
|
||||
with gr.Row():
|
||||
gr.Markdown("""
|
||||
### 🟧 Where are the vits models stored?
|
||||
This extension will download the models to `/alltalk_tts/models/vits/` folder.<br>
|
||||
|
||||
### 🟧 Where are the voices stored for vits models?
|
||||
Vits models themselves are the voices/have the voices built into them. Some models have mutliple voices embedded and some models only the one voice embedded in them.<br>
|
||||
|
||||
### 🟧 Where are the outputs stored & Automatic output wav file deletion
|
||||
Voice outputs are stored in `/alltalk_tts/outputs/`. You can configure automatic maintenance deletion of old wav files by setting `Del WAV's older than` in the global settings.<br>
|
||||
|
||||
> When `Disabled`, your output wav files will be left untouched.<br>
|
||||
> When set to a setting `1 Day` or greater, your output wav files older than that time period will be automatically deleted on start-up of AllTalk.<br>
|
||||
""")
|
||||
gr.Markdown("""
|
||||
""")
|
||||
|
||||
return app
|
||||
|
||||
################################
|
||||
# REQUIRED CHANGE #
|
||||
# Sets up the Gradio interface #
|
||||
################################
|
||||
# This function sets up the Gradio interface for your TTS engine's settings page.
|
||||
# You need to change the name of the function calls to match the names you set in the functions above.
|
||||
#
|
||||
# Specifically, you need to update the following:
|
||||
#
|
||||
# 1. The name of the function `xxxx_at_gradio_settings_page` to match your engine's name.
|
||||
# For example, if your engine is named "mytts", the function should be named `mytts_at_gradio_settings_page`.
|
||||
#
|
||||
# 2. The name of the function call `xxxx_model_alltalk_settings(model_config_data)`.
|
||||
# This should match the name you gave to the function that sets up the engine-specific settings page in Gradio.
|
||||
# If you named that function `mytts_model_alltalk_settings`, then the call should be:
|
||||
# `mytts_model_alltalk_settings(model_config_data)`
|
||||
#
|
||||
# After making these changes, this function will create and return the Gradio app for your TTS engine's settings page.
|
||||
def vits_at_gradio_settings_page(model_config_data):
|
||||
app = vits_model_alltalk_settings(model_config_data)
|
||||
return app
|
||||
def vits_at_gradio_settings_page(model_config_data):
|
||||
app = vits_model_alltalk_settings(model_config_data)
|
||||
return app
|
||||
Reference in new issue
Block a user