Commit Graph

  • 62513f9d95 배포 준비비 SanghyeonAn94 2025-10-16 19:42:29 +09:00
  • 25156d24c2 Fix ASMD KamioRinn 2025-10-16 11:39:11 +08:00
  • e8616c87c6 3초 미만 제한없음 SanghyeonAn94 2025-10-15 17:06:21 +09:00
  • 90141d2029 clean KakaruHayate 2025-10-13 22:21:50 +08:00
  • fd8c860f49 clean KakaruHayate 2025-10-13 22:14:23 +08:00
  • aada52050e
    clean Kakaru 2025-10-13 22:04:09 +08:00
  • 90a2f0471f Implement character mapping functionality for text normalization across multiple languages, including Cantonese, Chinese, English, Japanese, and Korean. Introduce text_normalize_with_map methods to return normalized text along with character mappings. Add a new utility module for building character mappings. 白菜工厂1145号员工 2025-10-12 02:27:13 +08:00
  • 47bb5a2cba clean KakaruHayate 2025-10-11 22:47:05 +08:00
  • 1f91faf51e del musa_utils from musa_accelerator KakaruHayate 2025-10-11 22:36:58 +08:00
  • 50db2e9199 support S1 train on MUSA KakaruHayate 2025-10-11 22:33:32 +08:00
  • 72be145051 Support on MUSA device. KakaruHayate 2025-10-10 21:14:27 +08:00
  • 0bde5d4a81 添加健康检查端点和音频上传功能 白菜工厂1145号员工 2025-10-11 11:03:35 +08:00
  • 495ef66177 Add FunASR Download XXXXRT666 2025-10-05 14:37:24 +01:00
  • fc7c652105 Remove HF ENDPOINT XXXXRT666 2025-10-05 14:21:21 +01:00
  • 93870073e8 Remove YUE in whisper XXXXRT666 2025-10-05 14:20:20 +01:00
  • da43eb7669 Typo Fix XXXXRT666 2025-10-05 12:42:59 +01:00
  • 84e902eaa8 Add ModelScope Snapshot Download For ASR XXXXRT666 2025-10-05 12:35:01 +01:00
  • 338498ad68 fix 白菜工厂1145号员工 2025-10-05 01:16:14 +08:00
  • 8c0cb0d691
    为中文添加了货币计数 Ella Zhang 2025-09-23 02:50:39 -07:00
  • 0c02ebf5ae
    Update num.py Ella Zhang 2025-09-23 02:48:30 -07:00
  • 82b458625d
    为中文添加了货币计数 Ella Zhang 2025-09-23 02:48:07 -07:00
  • 19ca3f3f6a Supports phoneme and word-level timestamp output for multilingual text 白菜工厂1145号员工 2025-09-22 07:43:59 +08:00
  • 2ff9a1533f Fix the timestamp processing logic for phonemes, characters, and words 白菜工厂1145号员工 2025-09-22 00:27:14 +08:00
  • 705df4c414 Supporting word-level timestamp output through attention weight output 白菜工厂1145号员工 2025-09-21 08:23:17 +08:00
  • 11aa78bd9b
    修复环境变量可能不为str的问题 RVC-Boss 2025-09-10 15:01:04 +08:00
  • eb15a9a3a8
    限制ffmpeg版本,最新版不兼容 黑喵小姐 2025-09-10 08:30:37 +08:00
  • 133f757e87 only_tts CocoRoF 2025-09-05 08:10:39 +09:00
  • af9bb934b2
    Update api_v2.py JangHaryeom 2025-09-04 13:11:20 +09:00
  • fc987e2a65 refactor: centralize model loading logic Jacky He 2025-09-04 09:38:04 +08:00
  • 65cf7d67d5 refactor: separate loading model logic to a function instead of while importing Jacky He 2025-09-02 17:50:34 +08:00
  • 611ff1e8c0 feat: make GPU selectable in get_tts_wav Jacky He 2025-09-02 17:48:44 +08:00
  • 8858492f56 feat:enable speed control for v1v2 zpeng11 2025-08-31 00:14:07 -04:00
  • 337da7454e feat:remove prints zpeng11 2025-08-26 17:09:36 -04:00
  • 3e63595f0e feat:update kv cache to [len, head, dim] to allow linear size increasement zpeng11 2025-08-26 17:01:41 -04:00
  • fa84e262ae feat:remove unneed for main zpeng11 2025-08-25 22:53:15 -04:00
  • 968ac4c264 feat: solved problem, export works zpeng11 2025-08-25 22:37:52 -04:00
  • 419909b443 failed , testing expand y zpeng11 2025-08-25 21:57:36 -04:00
  • c85ee3d521 feat:successfully unified first step and following step zpeng11 2025-08-25 17:57:04 -04:00
  • d413a4f5b1 run time working zpeng11 2025-08-25 17:07:38 -04:00
  • 26228402e3 feat:solve unified kv cache shape handling, todo: clean up upper level to unify first and following step zpeng11 2025-08-25 12:06:26 -04:00
  • 0c5f61f98c feat:rename and features to onnx export zpeng11 2025-08-25 01:46:53 -04:00
  • 633e478b24 feat:clean up playground zpeng11 2025-08-24 02:37:34 -04:00
  • 942caa888e feat:supporting half export zpeng11 2025-08-24 02:29:33 -04:00
  • 72c5d3224e utility updates zpeng11 2025-08-24 02:11:47 -04:00
  • 48d52778ce feat:clean up export logics and add notes zpeng11 2025-08-24 02:00:32 -04:00
  • e4d1894a8f feat:experiments with for onnx with attention, but does not work well todo:clean code and try v3v4 zpeng11 2025-08-24 00:46:29 -04:00
  • 5982080939 feat:updated fsdecode and decoder interface zpeng11 2025-08-23 17:35:21 -04:00
  • b45cbc3561 feat: sampling params working now for export, todo:fold weights clean code zpeng11 2025-08-23 13:03:02 -04:00
  • 9ed42daa88 feat: allow fsdec and sdec to have sampling parames zpeng11 2025-08-23 12:17:04 -04:00
  • 3ccd1c0ea3 fix: solved t2s ending problem, and verified infer&forward has same output under deterministic random, fixed topk to 15 zpeng11 2025-08-23 03:31:01 -04:00
  • 63cbb6efa7 verified audio, text, synthesizer all working, todo:dig into t2s for error zpeng11 2025-08-22 01:36:58 -04:00
  • e8fdf472c0 feat:onnx friendly loop with same function zpeng11 2025-08-21 22:23:50 -04:00
  • b1b9e6e68e Handling cases where the length of semantic_phoneme may be 0 solely 2025-08-22 00:32:47 +08:00
  • 77794a5923 feat:export onnx with combined graph ready, todo:link weights in onnx graph zpeng11 2025-08-21 01:52:34 -04:00
  • 16d30ce1e4 feat:get ready for if node merge zpeng11 2025-08-21 00:34:56 -04:00
  • 403c5bf320 feat:v1v2 both works for export zpeng11 2025-08-20 20:57:29 -04:00
  • bc7fe01876 feat:update v2pro's gpt path, todo:work on v1 transform zpeng11 2025-08-20 20:41:20 -04:00
  • bb529e7e4a update namings zpeng11 2025-08-20 20:21:42 -04:00
  • 4e0cc57052 update init_step name zpeng11 2025-08-20 20:05:07 -04:00
  • aafa0561d8 correctly setup onnx export, solved problem zpeng11 2025-08-20 19:39:10 -04:00
  • 94b31a250f limit vits for one input a time zpeng11 2025-08-20 18:39:21 -04:00
  • fd0fb35a49 fix spectrum take out working zpeng11 2025-08-20 18:32:38 -04:00
  • 911c53b1ee fixed using hubert for full run, 80 works zpeng11 2025-08-20 17:37:41 -04:00
  • 1cdd41d877 fixed resample audio and tested in full run zpeng11 2025-08-20 16:47:55 -04:00
  • da5aa78224 feat:combined fsdc and encoder, todo:extract audio pipeline zpeng11 2025-08-20 02:24:59 -04:00
  • 71cbe28e68 feat:optimize looping zpeng11 2025-08-19 21:31:42 -04:00
  • 5c08328cf3 feat:voice and text preprocess system verifed, todo:dissasemble onnx export of gsv zpeng11 2025-08-19 21:10:21 -04:00
  • dd156f15aa feat:clean up playground explore audio preprocess, todo:build free run from pure input data zpeng11 2025-08-19 01:22:00 -04:00
  • aef9d26580 feat:text_bert and audio_hubert exports are ready and fully tested, todo:solve dependancy in playground runs zpeng11 2025-08-19 00:05:45 -04:00
  • 4e42a28f9c feat: v2pro and v2proplus onnx export works, todo:solve data dependencies zpeng11 2025-08-18 20:19:27 -04:00
  • d313fbc740 stream_v2pro: 新增 --lang 参数提示参考文字的语言类型 csh 2025-08-18 16:26:29 +08:00
  • 60f07ea36e stream_v2pro: 修复由于 spectrogram_torch 输入是 half 导致 spec 溢出最终没有声音的问题 csh 2025-08-18 16:25:38 +08:00
  • f54cbbe743 vq distributed training support wangzeyuan 2025-08-18 12:16:59 +08:00
  • 21f4e1ff55
    Merge 3a04e9adad5077315f7433bae23e59062d204f15 into fdf794e31d1fd6f91c5cb4fbb0396094491a31ac wzy3650 2025-08-18 04:02:35 +00:00
  • 3a04e9adad
    Merge branch 'RVC-Boss:main' into main wzy3650 2025-08-18 12:02:32 +08:00
  • 610b36561a feat:remove debug, todo:rewrite the onnx export interface zpeng11 2025-08-17 19:22:11 -04:00
  • 8c0f32da3e feat:v2pp onnx export ready testing... zpeng11 2025-08-17 17:54:57 -04:00
  • 29d8e3e34f
    Update webui.py Hemant Sharma 2025-08-12 14:22:47 -04:00
  • 485cb85a55 fix: determine whether the filename was input or only path is input Jacky He 2025-08-12 14:03:15 +08:00
  • baf61b61d8 fix: wrong args name Jacky He 2025-08-12 12:06:39 +08:00
  • 7263f02706 feat: add optional params Jacky He 2025-08-11 11:34:08 +08:00
  • 013ed62a78 chore: add new supported language Jacky He 2025-08-11 10:51:43 +08:00
  • 035964ae78 refactor: better param input Jacky He 2025-08-11 10:33:17 +08:00
  • 5285786b57 fix: when text and word2ph is empty, get_bert_feature will not crash but return an empty tensor Jacky He 2025-08-09 18:26:25 +08:00
  • 4af71131b4
    Merge 2e05ea96c507f332208455df323fb074973f6e51 into fdf794e31d1fd6f91c5cb4fbb0396094491a31ac Jacky He 2025-08-09 18:16:30 +08:00
  • 2e05ea96c5 fix: when text and word2ph is empty, get_bert_feature will not crash but return an empty tensor Jacky He 2025-08-09 13:55:49 +08:00
  • 44499eb5bd chore: change to python syntax\n for higher readibility Jacky He 2025-08-08 21:20:02 +08:00
  • fdf794e31d
    Update WSL Rocm (#2561) XXXXRT666 2025-08-02 17:47:15 +08:00
  • ea47198b32
    Update WSL Rocm XXXXRT666 2025-08-02 17:34:33 +08:00
  • 8f4f65ba81 add fix samiabat 2025-07-29 01:00:14 +03:00
  • c04b177adb add samiabat 2025-07-29 00:53:56 +03:00
  • af16611248 add way samiabat 2025-07-29 00:47:45 +03:00
  • fb8b07b373 make tensorflow to the latest samiabat 2025-07-29 00:42:01 +03:00
  • b4764cf91f add fix samiabat 2025-07-29 00:33:02 +03:00
  • 168903cb3e add latest changes samiabat 2025-07-29 00:19:57 +03:00
  • a7018c9975 change version samiabat 2025-07-28 22:56:52 +03:00
  • 7beca9d890 add gradio samiabat 2025-07-28 22:26:04 +03:00
  • 2571c4bfdc add matplotlib samiabat 2025-07-28 22:13:36 +03:00
  • 498a92e501 remove the download.py section samiabat 2025-07-28 22:03:14 +03:00
  • 07801e0831 try the latest version samiabat 2025-07-28 20:48:12 +03:00