44 Commits

Author SHA1 Message Date
LittleMouse c89412ef6f [update] update llm-asr llm-audio llm-kws llm-melotts & add AI_Pyramid_Demo file 2026-01-22 18:25:20 +08:00
LittleMouse 7b671f21e0 [update] update melotts play stop cap 2025-12-22 12:19:30 +08:00
LittleMouse 7a97143aad [update] update sdk version & chip name 2025-09-29 15:15:00 +08:00
LittleMouse 9a14d45799 [update] support ax650. add ax650 model. 2025-08-27 16:10:46 +08:00
LittleMouse c25b4f0a09 [fix] Fix caching causing audio issues 2025-07-23 14:26:22 +08:00
LittleMouse 9abe069db6 [update] KWS sets multiple keywords, fix melotts
File missing, Optimize generated audio
2025-07-18 14:12:12 +08:00
LittleMouse 8acb179b63 [fix] Handles the situation where Either tagger or verbalizer file does not exist. 2025-05-20 15:06:11 +08:00
LittleMouse f775786033 [update] add en-au, en-br, en-india, en-us model. Format code. 2025-05-20 14:19:44 +08:00
Xuanwu Yun 0619178333 Merge pull request #2 from yuyun2000/opt/melotts
Add text normalization for Chinese, Japanese, and English
2025-05-16 11:18:21 +08:00
yuyun2000 74c41a3e44 Add text normalization for Chinese, Japanese, and English
Implement regex-based text normalization functionality to support trilingual (CJE) content processing
2025-05-16 11:10:58 +08:00
Xuanwu Yun ebf908a0d1 Merge branch 'dev' into opt/melotts 2025-05-15 15:24:36 +08:00
yuyun2000 10e4bdf828 Refactor SOLA component code
Streamline and simplify code in the SOLA module for improved readability and maintenance
2025-05-15 15:17:04 +08:00
LittleMouse 6e503a130c [update] delete debug log 2025-05-09 18:42:20 +08:00
yuyun2000 835daf192c Fix SOLA algorithm implementation
Resolved issues in the Synchronized Overlap-Add (SOLA) algorithm to ensure proper audio processing and alignment.
2025-05-09 18:25:49 +08:00
yuyun2000 9e7342f3b1 Fix code formatting and integrate SOLA algorithm into main.py
Corrected code formatting issues and moved the SOLA (Synchronized Overlap-Add) algorithm implementation into main.py for better code organization.
2025-05-09 17:43:17 +08:00
yuyun2000 cbb0afaf65 Optimize G2P process to skip inference for short audio clips
Improved the Grapheme-to-Phoneme (G2P) process by eliminating separate inference operations for audio segments that are too short, enhancing processing efficiency.
2025-05-09 17:24:57 +08:00
yuyun2000 b1df925af9 Fix SOLA detail issue causing first frame problems
Resolved an issue in the SOLA (Synchronized Overlap-Add) implementation where specific details were causing problems with the first frame of audio processing.
2025-05-09 17:01:55 +08:00
yuyun2000 a151affa1d Translate logs in Lexicon.hpp to English and add debug switch
- Convert all Chinese log messages in Lexicon.hpp to English for better international compatibility\n- Add a debug flag to control whether to display g2p process logs\n- Improve code readability and debugging experience
2025-05-06 16:50:35 +08:00
yuyun2000 e3c70bc127 Implement Sola algorithm for smoother audio transitions
Apply the Synchronized Overlap-Add (SOLA) algorithm to smooth the connection between audio segments output by the decoder, resulting in more natural-sounding transitions between segments.
2025-05-06 16:48:11 +08:00
yuyun2000 5782f89df1 处理陌生英语单词 2025-04-30 16:50:12 +08:00
yuyun2000 2e40ae6f10 去掉中文注释;使用log输出日志;格式化代码 2025-04-30 14:38:08 +08:00
yuyun2000 1a90562805 优化g2p流程,可以处理多音字,中英混合的情况等等 2025-04-30 10:30:23 +08:00
LittleMouse 4b477eeac3 [update] Add melotts-en-us model 2025-04-22 15:50:00 +08:00
dianjixz 91c5bbd08d [update] change bsp define 2025-04-16 18:51:16 +08:00
LittleMouse 98adcc20cc [update] depth_anything, melotts, yolo. Added support for 650N. 2025-04-16 17:08:08 +08:00