推理时无限processing

[gaoxuxu110]

使用的是: 参考音频的方式生成语音, 但是一直卡在processing..
控制台窗口加载到一下信息后就没有后续了
To create a public link, set share=True in launch().
You are using the latest version of funasr-1.1.2
2024-07-19 12:28:59,203 - modelscope - WARNING - Using the master branch is fragile, please use it with caution!
2024-07-19 12:28:59,203 - modelscope - INFO - Use user-specified model revision: master
另外: 使用了最新的代码(即: fix modelscope version problem (#396))

[gaoxuxu110]

使用的是: 参考音频的方式生成语音, 但是一直卡在processing.. 控制台窗口加载到一下信息后就没有后续了 To create a public link, set share=True in launch(). You are using the latest version of funasr-1.1.2 2024-07-19 12:28:59,203 - modelscope - WARNING - Using the master branch is fragile, please use it with caution! 2024-07-19 12:28:59,203 - modelscope - INFO - Use user-specified model revision: master 另外: 使用了最新的代码(即: fix modelscope version problem (#396))

还是加载出来了, 但是花了6,7分钟... 这明显不合理, 因为只有十几个字, 而且我用的编译推理, 本身是3090

[PoTaTo-Mika]

因为auto_rerank功能需要下载funasr的权重，时间长的话请自行检查网速是否太慢。

[chaoqunxie]

LOCAL_RANK: 0 - CUDA_VISIBLE_DEVICES: [0]
[2024-07-18 15:45:44,275][fish_speech.models.text2semantic.lit_module][INFO] - [rank: 0] Set weight decay: 0 for 459 parameters
[2024-07-18 15:45:44,275][fish_speech.models.text2semantic.lit_module][INFO] - [rank: 0] Set weight decay: 0.0 for 65 parameters

| Name | Type | Params | Mode

0 | model | DualARTransformer | 394 M | train
1 | model.embeddings | Embedding | 2.4 M | train
2 | model.layers | ModuleList | 311 M | train
3 | model.norm | RMSNorm | 1.0 K | train
4 | model.output | Linear | 280 K | train
5 | model.fast_embeddings | Embedding | 1.1 M | train
6 | model.fast_layers | ModuleList | 77.9 M | train
7 | model.fast_norm | RMSNorm | 1.0 K | train
8 | model.fast_output | Linear | 1.1 M | train

4.3 M Trainable params
390 M Non-trainable params
394 M Total params
1,577.969 Total estimated model params size (MB)
Sanity Checking: | | 0/? [00:00<?, ?it/s][2024-07-18 15:45:44,446][fish_speech.datasets.text][INFO] - [rank: 0] Reading 1 / 1 files
[2024-07-18 15:45:44,446][fish_speech.datasets.text][INFO] - [rank: 0] Reading 2 / 1 files
[2024-07-18 15:45:44,447][fish_speech.datasets.text][INFO] - [rank: 0] Reading 1 / 1 files
[2024-07-18 15:45:44,447][fish_speech.datasets.text][INFO] - [rank: 0] Read total 1 groups of data
[2024-07-18 15:45:44,447][fish_speech.datasets.text][INFO] - [rank: 0] Read total 2 groups of data
[2024-07-18 15:45:44,447][fish_speech.datasets.text][INFO] - [rank: 0] Read total 1 groups of data
[2024-07-18 15:45:44,448][fish_speech.datasets.text][INFO] - [rank: 0] Reading 1 / 1 files
[2024-07-18 15:45:44,448][fish_speech.datasets.text][INFO] - [rank: 0] Reading 1 / 1 files
[2024-07-18 15:45:44,448][fish_speech.datasets.text][INFO] - [rank: 0] Read total 1 groups of data
[2024-07-18 15:45:44,448][fish_speech.datasets.text][INFO] - [rank: 0] Read total 1 groups of data
[2024-07-18 15:45:44,450][fish_speech.datasets.text][INFO] - [rank: 0] Reading 1 / 1 files
[2024-07-18 15:45:44,451][fish_speech.datasets.text][INFO] - [rank: 0] Read total 1 groups of data
[2024-07-18 15:45:44,452][fish_speech.datasets.text][INFO] - [rank: 0] Reading 1 / 1 files
[2024-07-18 15:45:44,452][fish_speech.datasets.text][INFO] - [rank: 0] Reading 1 / 1 files
[2024-07-18 15:45:44,452][fish_speech.datasets.text][INFO] - [rank: 0] Read total 1 groups of data
[2024-07-18 15:45:44,452][fish_speech.datasets.text][INFO] - [rank: 0] Read total 1 groups of data
[2024-07-18 15:45:47,450][fish_speech.datasets.text][INFO] - [rank: 0] Reading 2 / 1 files
[2024-07-18 15:45:47,451][fish_speech.datasets.text][INFO] - [rank: 0] Reading 1 / 1 files
[2024-07-18 15:45:47,451][fish_speech.datasets.text][INFO] - [rank: 0] Read total 1 groups of data
[2024-07-18 15:45:47,452][fish_speech.datasets.text][INFO] - [rank: 0] Read total 2 groups of data
[2024-07-18 15:45:47,458][fish_speech.datasets.text][INFO] - [rank: 0] Reading 1 / 1 files
[2024-07-18 15:45:47,458][fish_speech.datasets.text][INFO] - [rank: 0] Reading 1 / 1 files
[2024-07-18 15:45:47,458][fish_speech.datasets.text][INFO] - [rank: 0] Read total 1 groups of data
[2024-07-18 15:45:47,459][fish_speech.datasets.text][INFO] - [rank: 0] Read total 1 groups of data
[2024-07-18 15:45:47,461][fish_speech.datasets.text][INFO] - [rank: 0] Reading 1 / 1 files
[2024-07-18 15:45:47,461][fish_speech.datasets.text][INFO] - [rank: 0] Reading 1 / 1 files
[2024-07-18 15:45:47,461][fish_speech.datasets.text][INFO] - [rank: 0] Read total 1 groups of data
[2024-07-18 15:45:47,461][fish_speech.datasets.text][INFO] - [rank: 0] Read total 1 groups of data
[2024-07-18 15:45:47,464][fish_speech.datasets.text][INFO] - [rank: 0] Reading 1 / 1 files
[2024-07-18 15:45:47,464][fish_speech.datasets.text][INFO] - [rank: 0] Read total 1 groups of data
[2024-07-18 15:45:47,508][fish_speech.datasets.text][INFO] - [rank: 0] Reading 1 / 1 files
[2024-07-18 15:45:47,508][fish_speech.datasets.text][INFO] - [rank: 0] Read total 1 groups of data
[rank0]:[E ProcessGroupNCCL.cpp:1414] [PG 0 Rank 0] Process group watchdog thread terminated with exception: CUDA error: unspecified launch failure
CUDA kernel errors might be asynchronously reported at some other API call, so the stacktrace below might be incorrect.
For debugging consider passing CUDA_LAUNCH_BLOCKING=1.
Compile with TORCH_USE_CUDA_DSA to enable device-side assertions.

Exception raised from c10_cuda_check_implementation at ../c10/cuda/CUDAException.cpp:43 (most recent call first):
frame #0: c10::Error::Error(c10::SourceLocation, std::string) + 0x57 (0x7f4941b7a897 in /usr/local/lib/python3.10/site-packages/torch/lib/libc10.so)
frame #1: c10::detail::torchCheckFail(char const*, char const*, unsigned int, std::string const&) + 0x64 (0x7f4941b2ab25 in /usr/local/lib/python3.10/site-packages/torch/lib/libc10.so)
frame #2: c10::cuda::c10_cuda_check_implementation(int, char const*, char const*, int, bool) + 0x118 (0x7f4941f06718 in /usr/local/lib/python3.10/site-packages/torch/lib/libc10_cuda.so)
frame #3: c10d::ProcessGroupNCCL::WorkNCCL::finishedGPUExecutionInternal() const + 0x56 (0x7f48f58598e6 in /usr/local/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so)
frame #4: c10d::ProcessGroupNCCL::WorkNCCL::isCompleted() + 0x58 (0x7f48f585d9e8 in /usr/local/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so)
frame #5: c10d::ProcessGroupNCCL::watchdogHandler() + 0x77c (0x7f48f586305c in /usr/local/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so)
frame #6: c10d::ProcessGroupNCCL::ncclCommWatchdog() + 0x10c (0x7f48f5863dcc in /usr/local/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so)
frame #7: + 0xd44a3 (0x7f49412ba4a3 in /usr/lib/x86_64-linux-gnu/libstdc++.so.6)
frame #8: + 0x89134 (0x7f4943751134 in /usr/lib/x86_64-linux-gnu/libc.so.6)
frame #9: __clone + 0x40 (0x7f49437d0a40 in /usr/lib/x86_64-linux-gnu/libc.so.6)

terminate called after throwing an instance of 'c10::DistBackendError'
what(): [PG 0 Rank 0] Process group watchdog thread terminated with exception: CUDA error: unspecified launch failure
CUDA kernel errors might be asynchronously reported at some other API call, so the stacktrace below might be incorrect.
For debugging consider passing CUDA_LAUNCH_BLOCKING=1.
Compile with TORCH_USE_CUDA_DSA to enable device-side assertions.

Exception raised from c10_cuda_check_implementation at ../c10/cuda/CUDAException.cpp:43 (most recent call first):
frame #0: c10::Error::Error(c10::SourceLocation, std::string) + 0x57 (0x7f4941b7a897 in /usr/local/lib/python3.10/site-packages/torch/lib/libc10.so)
frame #1: c10::detail::torchCheckFail(char const*, char const*, unsigned int, std::string const&) + 0x64 (0x7f4941b2ab25 in /usr/local/lib/python3.10/site-packages/torch/lib/libc10.so)
frame #2: c10::cuda::c10_cuda_check_implementation(int, char const*, char const*, int, bool) + 0x118 (0x7f4941f06718 in /usr/local/lib/python3.10/site-packages/torch/lib/libc10_cuda.so)
frame #3: c10d::ProcessGroupNCCL::WorkNCCL::finishedGPUExecutionInternal() const + 0x56 (0x7f48f58598e6 in /usr/local/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so)
frame #4: c10d::ProcessGroupNCCL::WorkNCCL::isCompleted() + 0x58 (0x7f48f585d9e8 in /usr/local/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so)
frame #5: c10d::ProcessGroupNCCL::watchdogHandler() + 0x77c (0x7f48f586305c in /usr/local/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so)
frame #6: c10d::ProcessGroupNCCL::ncclCommWatchdog() + 0x10c (0x7f48f5863dcc in /usr/local/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so)
frame #7: + 0xd44a3 (0x7f49412ba4a3 in /usr/lib/x86_64-linux-gnu/libstdc++.so.6)
frame #8: + 0x89134 (0x7f4943751134 in /usr/lib/x86_64-linux-gnu/libc.so.6)
frame #9: __clone + 0x40 (0x7f49437d0a40 in /usr/lib/x86_64-linux-gnu/libc.so.6)

Exception raised from ncclCommWatchdog at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1418 (most recent call first):
frame #0: c10::Error::Error(c10::SourceLocation, std::string) + 0x57 (0x7f4941b7a897 in /usr/local/lib/python3.10/site-packages/torch/lib/libc10.so)
frame #1: + 0xe32119 (0x7f48f54e7119 in /usr/local/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so)
frame #2: + 0xd44a3 (0x7f49412ba4a3 in /usr/lib/x86_64-linux-gnu/libstdc++.so.6)
frame #3: + 0x89134 (0x7f4943751134 in /usr/lib/x86_64-linux-gnu/libc.so.6)
frame #4: __clone + 0x40 (0x7f49437d0a40 in /usr/lib/x86_64-linux-gnu/libc.so.6) 这是啥原因呢

[AnyaCoder]

使用的是: 参考音频的方式生成语音, 但是一直卡在processing.. 控制台窗口加载到一下信息后就没有后续了 To create a public link, set share=True in launch(). You are using the latest version of funasr-1.1.2 2024-07-19 12:28:59,203 - modelscope - WARNING - Using the master branch is fragile, please use it with caution! 2024-07-19 12:28:59,203 - modelscope - INFO - Use user-specified model revision: master 另外: 使用了最新的代码(即: fix modelscope version problem (#396))

还是加载出来了, 但是花了6,7分钟... 这明显不合理, 因为只有十几个字, 而且我用的编译推理, 本身是3090

rerank用到了音频转写，有一些问题，现在pr中修复。