mARTin-B78 ccccdf820a feat: graph-decoded streaming, prebuffer_ms and cancel-on-disconnect (v6.13) ... - Fixed-window CUDA-graph codec decode for streaming (TTS_GRAPH_DECODE=0 disables) - Per-request chunk_size (2-24) on /v1/audio/speech - prebuffer_ms (0-10000) to hold back audio before sending - Stop generation on client disconnect - Recommend --max-seq-len 2048 for the streaming container Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>		2026-10-04 00:11:12 +02:00
..
auto_transcribe.py	Add multi-source voice pipeline with VoiceDesign support	2026-05-26 12:19:33 +02:00
benchmark_api.py	Reduce latency: CUDA warmup, chunk_size=4, max-seq-len 2048	2026-05-11 20:43:21 +02:00
customvoice_voices.json	Add CustomVoice server, benchmark tool, and VoiceDesign API improvements	2026-05-26 12:19:33 +02:00
extract_embeddings.py	feat: precompute and load speaker embeddings (.pt) from voices.json	2026-06-20 13:04:55 +02:00
find_best_seed.py	feat: add /voice-seed API endpoint and update find_best_seed.py (v6.7.1)	2026-06-21 14:53:31 +02:00
generate_voices.py	fix: raise streaming chunk_size to 12 to cut codec re-decode overhead	2026-10-03 14:03:40 +02:00
run_customvoice_server.py	v6.7: Native speed control and zip word-level timestamps	2026-06-26 11:11:13 +02:00
run_server.py	feat: graph-decoded streaming, prebuffer_ms and cancel-on-disconnect (v6.13)	2026-10-04 00:11:12 +02:00
run_voicedesign_server.py	v6.8: Untrack generated VoiceDesign registry, add presets seed	2026-07-29 16:06:05 +02:00
seed_samples_progress.log	feat: add seed samples API endpoints	2026-06-22 11:39:01 +02:00
voicedesign_voices.presets.json	v6.8: Untrack generated VoiceDesign registry, add presets seed	2026-07-29 16:06:05 +02:00