b6fe850a9d
Indextts2 ( #276 )
...
* indextts2
* update lfs for audio files
---------
Co-authored-by: wangyining02 <wangyining02@bilibili.com >
2025-09-08 17:36:39 +08:00
f017ab247c
Fix the bug that 嗯 cannot be synthesized.
2025-06-17 17:48:04 +08:00
05f0f12ae6
Merge pull request #180 from yrom/fix-bugs
...
Improve loading `alias_free_activation_cuda`
2025-05-29 11:06:53 +08:00
4ba6342b3f
fix: 避免在 MinGW-w64 环境 jit compile cuda ext
2025-05-29 09:01:56 +08:00
150902db78
Update README.md
2025-05-26 13:42:51 +08:00
5d3328de60
fix: add force_rebuild flag for fused alias_free_activation and update installation instructions
2025-05-23 15:07:39 +08:00
dcca816fae
Merge pull request #178 from yrom/fix-front
...
fix: 添加英语缩写处理
2025-05-23 14:07:50 +08:00
fe2c645207
fix: 添加英语缩写处理
...
临时解决WeTextProcessing 的部分badcase
2025-05-23 10:07:48 +08:00
cfa1e9090c
Merge pull request #169 from yrom/fix-text-tokenizer
...
Fixes issues of text normalizer
2025-05-21 15:31:31 +08:00
61c7c61823
fix: update PINYIN_TONE_PATTERN and NormalizerZh
2025-05-21 13:01:25 +08:00
8056fbbf46
Merge pull request #161 from yrom/fix-tokenizer
...
fix 分句逻辑，增加测试用例
2025-05-19 08:28:01 +08:00
9454e075d7
fix 分句逻辑，增加测试用例
...
不包含。！？的句子没有被正确分句
2025-05-18 22:53:51 +08:00
363cbb0111
Merge pull request #160 from yrom/fix-infer
...
适配1.5版本模型，优化Webui，适配新版本transformers
2025-05-18 22:14:32 +08:00
09042b107a
padding_test.py support model dir for test
2025-05-18 19:57:11 +08:00
3df7dde92f
更新WebUI，添加模型目录检查和必要文件验证
...
- 新增示例
- 新增模型版本提示
- 新增生成参数设置
- 新增分句预览
2025-05-18 19:41:06 +08:00
df9da73fca
将 sample_kwargs 替换为 generation_kwargs
2025-05-18 19:34:41 +08:00
063fa288db
添加警告提示：生成停止因超出 max_mel_tokens 限制
2025-05-18 19:34:41 +08:00
2c58ff3437
删除测试代码
2025-05-18 19:34:41 +08:00
29545f76bd
适配新版本transformers
2025-05-18 19:34:41 +08:00
7316c1acd3
修正attention mask和positional embeddings
...
- 将之前只有text右侧填充改为cond+text 整体左侧填充
- 添加填充测试用例
2025-05-18 19:34:32 +08:00
3ce7ccb02c
优化文本掩码填充逻辑，改进句子桶化处理
2025-05-17 20:59:07 +08:00
a12da2516c
fix 批量推理1.5版本模型问题，调整分句逻辑和参数设置
...
- 将pad 改为全 eos token
- 优化bucket_sentences 算法
2025-05-17 14:40:01 +08:00
d80d49e5ce
优化inference attention mask
2025-05-17 14:38:01 +08:00
056cf78f8c
优化文本归一化和分句逻辑
...
修复可能的递归问题 (Fixes #124 )
2025-05-17 11:16:54 +08:00
9c6b1a055a
Update README.md
2025-05-14 17:07:05 +08:00
deccb7feba
release IndexTTS-1.5
2025-05-14 11:15:20 +08:00
d31277a881
Merge pull request #122 from yrom/feat/opt-text-tokenizer
...
Introduce a new `TextTokenizer` class to enhance text normalization and tokenization
2025-04-25 11:42:03 +08:00
93e9e3e3eb
Fix split_sentences_by_token
2025-04-24 23:58:16 +08:00
a0947cfd91
Fix pinyin correction
2025-04-24 20:38:52 +08:00
4616c6f953
Enhance text normalization and tokenization
...
- Introduced `de_tokenized_by_CJK_char` for restoring original text from tokenized format.
- Added `TextTokenizer` class for improved tokenization, including sentence splitting and handling of special tokens.
- Enhanced `TextNormalizer` to handle names and pinyin tones with placeholder mechanisms.
- Added regression tests for new features in `regression_test.py`.
2025-04-24 20:28:44 +08:00
ea694fac5a
Fix autocast device type for compatibility
2025-04-24 11:00:49 +08:00
f68f434b1f
Update setup.py
2025-04-24 10:59:09 +08:00
6db6161879
Merge pull request #107 from juntaosun/main
...
单句推理：RTF 性能提速小优化 ~
2025-04-21 11:33:01 +08:00
8331e90c74
单句推理：RTF性能至少提升 10%
2025-04-20 14:12:38 +08:00
881cdb0ae1
+回归测试脚本 ( #103 )
...
* deepspeed无法使用时回退到通常路径
* ninja支持中文路径编译补丁：BigVGAN fused cuda kernel
* 缓存参考音频的Mel
* ninja支持中文路径编译方案2：BigVGAN fused cuda kernel
* 增加批次推理：长句实现至少 2~10 倍以上的速度提升~
* fix上层目录为空时报错
* 批次推理：重要修复（漏句/丢句/音频空白）
* 批次推理：新增数据分桶机制，增强稳定性~
* +回归测试脚本
* update 回归测试脚本
* fix merge出错
---------
Co-authored-by: kemuriririn <10inspiral@gmail.com >
Co-authored-by: sunnyboxs <sjt2000@qq.com >
2025-04-18 18:09:13 +08:00
e1de45e856
批次推理：修复（漏句/丢句/音频空白） ( #100 )
...
* 批次推理：重要修复（漏句/丢句/音频空白）
* 批次推理：新增数据分桶机制，增强稳定性~
2025-04-18 17:57:07 +08:00
8d2759b2b7
Update README.md
...
update readme
2025-04-17 19:22:37 +08:00
b697743857
增加批次推理：长句实现至少 2~10 倍以上的速度提升~ ( #97 )
...
* ninja支持中文路径编译补丁：BigVGAN fused cuda kernel
* 缓存参考音频的Mel
* ninja支持中文路径编译方案2：BigVGAN fused cuda kernel
* 增加批次推理：长句实现至少 2~10 倍以上的速度提升~
---------
Co-authored-by: kemuriririn <nanaoto@foxmail.com >
2025-04-17 15:13:58 +08:00
84828da685
Feature/kemurin ( #99 )
...
* deepspeed无法使用时回退到通常路径
* ninja支持中文路径编译补丁：BigVGAN fused cuda kernel
* 缓存参考音频的Mel
* ninja支持中文路径编译方案2：BigVGAN fused cuda kernel
* 增加批次推理：长句实现至少 2~10 倍以上的速度提升~
* fix上层目录为空时报错
---------
Co-authored-by: kemuriririn <10inspiral@gmail.com >
Co-authored-by: sunnyboxs <sjt2000@qq.com >
2025-04-17 15:12:45 +08:00
6bd8a47ded
ninja中文路径编译补丁支持：BigVGAN fused cuda kernel ( #93 )
...
* ninja支持中文路径编译补丁：BigVGAN fused cuda kernel
* 缓存参考音频的Mel
* ninja支持中文路径编译方案2：BigVGAN fused cuda kernel
2025-04-17 14:56:37 +08:00
ed57aef7ca
update seed test results
2025-04-17 11:49:51 +08:00
60409bc707
Add the calculation time of each module.
2025-04-15 12:48:47 +08:00
c4007541d0
Merge pull request #92 from yrom/feat/activation-cuda
...
使用BigVGAN fused cuda kernel
2025-04-15 12:28:29 +08:00
9acb67e6fa
Update README
2025-04-15 12:05:07 +08:00
0206b8ebe6
enable custom cuda kernel for BigVGAN
2025-04-15 12:04:59 +08:00
d08ca55644
deepspeed无法使用时回退到通常路径 ( #90 )
...
Co-authored-by: kemuriririn <10inspiral@gmail.com >
2025-04-14 20:22:57 +08:00
0c56851212
修复拼音问题和分句问题，支持轻音声调（如yi1 shang5） ( #83 )
...
* Update Pinyin tone handling in TextNormalizer
* Enhance sentence splitting and improve tokenizer integration in inference
* Update character replacement mappings
test: "在电影《肖申克的救赎》中，安迪·杜佛兰被错误地判处终身监禁..."
* Refactor TextNormalizer and enhance testing with additional cases
2025-04-14 19:50:36 +08:00
4f87155a93
Merge pull request #78 from yrom/feat/mac-support
...
Add a new Command-Line Interface and support for mps device (Apple Silicon)
2025-04-11 22:01:00 +08:00
17a3b0857b
Adds MPS support for Apple Silicon
2025-04-11 21:22:08 +08:00
5fadfa1cd9
Support inference on CPU
2025-04-11 20:58:41 +08:00
kemuriririn