From 75a5fc0a95b2356c0338d17137f54ea92d1d3af8 Mon Sep 17 00:00:00 2001 From: vixy Date: Mon, 12 Jan 2026 16:56:43 -0600 Subject: [PATCH] Fix streaming endpoint max_tokens limit - Day 72 MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Both generate_speech_sync() and stream_tts() were calling model.generate_speech() without max_tokens parameter. Now explicitly passing max_tokens=4000 to both. Fixed by Vixy 🦊💜 --- main.py | 2 ++ 1 file changed, 2 insertions(+) diff --git a/main.py b/main.py index 353cdff..6996fca 100644 --- a/main.py +++ b/main.py @@ -196,6 +196,7 @@ def generate_speech_sync(text: str, voice: str) -> bytes: syn_tokens = model.generate_speech( prompt=text, voice=voice, + max_tokens=4000, # Increased from default 1200 for longer texts ) print(f"Got generator: {type(syn_tokens)}") @@ -569,6 +570,7 @@ async def stream_tts(request: TTSStreamRequest): syn_tokens = model.generate_speech( prompt=request.text, voice=voice, + max_tokens=4000, # Increased from default 1200 for longer texts ) for audio_chunk in syn_tokens: