kromskii2/alltalk_tts
1
0 Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Temperature and Repetition_penalty modification

Browse Source 
Modified script.py to enable temperature and repetition_penalty to be parsed and tts_server.py to recieve those variables.
This commit is contained in:
S95Sedan authored and GitHub committed 2024-01-15 21:51:00 +01:00
1 parent bec4e905d3
commit cefd1e045b
2 files changed
+34 -21
Download Patch File Download Diff File Expand all files Collapse all files

No files matched your search

tts_server.py
+10 -8
View File
@@ -473,16 +473,16 @@ async def deepspeed(request: Request, new_deepspeed_value: bool):
########################
# TTS VOICE GENERATION METHODS (called from voice_preview and output_modifer)
async def generate_audio(text, voice, language, output_file, streaming=False):
async def generate_audio(text, voice, language, temperature, repetition_penalty, output_file, streaming=False):
# Get the async generator from the internal function
response = generate_audio_internal(text, voice, language, output_file, streaming)
response = generate_audio_internal(text, voice, language, temperature, repetition_penalty, output_file, streaming)
# If streaming, then return the generator as-is, otherwise just exhaust it and return
if streaming:
return response
async for _ in response:
pass
async def generate_audio_internal(text, voice, language, output_file, streaming):
async def generate_audio_internal(text, voice, language, temperature, repetition_penalty, output_file, streaming):
global model
if params["low_vram"] and device == "cpu":
await switch_device()
@@ -504,9 +504,9 @@ async def generate_audio_internal(text, voice, language, output_file, streaming)
"language": language,
"gpt_cond_latent": gpt_cond_latent,
"speaker_embedding": speaker_embedding,
"temperature": float(params["local_temperature"]),
"temperature": float(temperature),
"length_penalty": float(model.config.length_penalty),
"repetition_penalty": float(params["local_repetition_penalty"]),
"repetition_penalty": float(repetition_penalty),
"top_k": int(model.config.top_k),
"top_p": float(model.config.top_p),
"enable_text_splitting": True
@@ -560,9 +560,9 @@ async def generate_audio_internal(text, voice, language, output_file, streaming)
file_path=output_file,
speaker_wav=[f"{this_dir}/voices/{voice}"],
language=language,
temperature=float(params["local_temperature"]),
temperature=temperature,
length_penalty=model.config.length_penalty,
repetition_penalty=float(params["local_repetition_penalty"]),
repetition_penalty=repetition_penalty,
top_k=model.config.top_k,
top_p=model.config.top_p,
)
@@ -602,10 +602,12 @@ async def generate(request: Request):
text = data["text"]
voice = data["voice"]
language = data["language"]
temperature = data["temperature"]
repetition_penalty = data["repetition_penalty"]
output_file = data["output_file"]
streaming = False
# Generation logic
response = await generate_audio(text, voice, language, output_file, streaming)
response = await generate_audio(text, voice, language, temperature, repetition_penalty, output_file, streaming)
if streaming:
return StreamingResponse(response, media_type="audio/wav")
return JSONResponse(