使用cuda加速rwkv模型遇到报错：RuntimeError: Error building extension 'wkv_cuda'

[shmily91]

1、按要求安装了vs（安装了C++）、cuda 2、torch 2.0 cuda（True），可用 3、模型rwkv-7b 10v 4、显卡3090 5、pip 安装了 rwkv ninja 6、参数没有改动

serving on 0.0.0.0:17860 view at http://127.0.0.1:17860 No sentence-transformers model found with name model/text2vec-large-chinese. Creating a new one with MEAN pooling. Using C:\Users\DJS\AppData\Local\torch_extensions\torch_extensions\Cache\py311_cu118 as PyTorch extensions root... Detected CUDA files, patching ldflags Emitting ninja build file C:\Users\DJS\AppData\Local\torch_extensions\torch_extensions\Cache\py311_cu118\wkv_cuda\build.ninja... Building extension module wkv_cuda... Allowing ninja to set a default number of workers... (overridable by setting the environment variable MAX_JOBS=N) ninja: error: 'L:/BaiduNetdiskDownload/�½��ļ���/wenda512/wenda/WPy64-31110/python-3.11.1.amd64/Lib/site-packages/rwkv/cuda/wrapper.cpp', needed by 'wrapper.o', missing and no known rule to make it Exception in thread Thread-1 (load_model): Traceback (most recent call last): File "L:\BaiduNetdiskDownload\新建文件夹\wenda512\wenda\WPy64-31110\python-3.11.1.amd64\Lib\site-packages\torch\utils\cpp_extension.py", line 1893, in _run_ninja_build subprocess.run( File "L:\BaiduNetdiskDownload\新建文件夹\wenda512\wenda\WPy64-31110\python-3.11.1.amd64\Lib\subprocess.py", line 571, in run raise CalledProcessError(retcode, process.args, subprocess.CalledProcessError: Command '['ninja', '-v']' returned non-zero exit status 1.

The above exception was the direct cause of the following exception:

Traceback (most recent call last): File "L:\BaiduNetdiskDownload\新建文件夹\wenda512\wenda\WPy64-31110\python-3.11.1.amd64\Lib\threading.py", line 1038, in _bootstrap_inner self.run() File "L:\BaiduNetdiskDownload\新建文件夹\wenda512\wenda\WPy64-31110\python-3.11.1.amd64\Lib\threading.py", line 975, in run self._target(*self._args, **self._kwargs) File "L:\BaiduNetdiskDownload\新建文件夹\wenda512\wenda\wenda.py", line 44, in load_model LLM.load_model() File "L:\BaiduNetdiskDownload\新建文件夹\wenda512\wenda\llms\llm_rwkv.py", line 262, in load_model from rwkv.model import RWKV # pip install rwkv ^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "L:\BaiduNetdiskDownload\新建文件夹\wenda512\wenda\WPy64-31110\python-3.11.1.amd64\Lib\site-packages\rwkv\model.py", line 29, in load( File "L:\BaiduNetdiskDownload\新建文件夹\wenda512\wenda\WPy64-31110\python-3.11.1.amd64\Lib\site-packages\torch\utils\cpp_extension.py", line 1284, in load return _jit_compile( ^^^^^^^^^^^^^ File "L:\BaiduNetdiskDownload\新建文件夹\wenda512\wenda\WPy64-31110\python-3.11.1.amd64\Lib\site-packages\torch\utils\cpp_extension.py", line 1509, in _jit_compile _write_ninja_file_and_build_library( File "L:\BaiduNetdiskDownload\新建文件夹\wenda512\wenda\WPy64-31110\python-3.11.1.amd64\Lib\site-packages\torch\utils\cpp_extension.py", line 1624, in _write_ninja_file_and_build_library _run_ninja_build( File "L:\BaiduNetdiskDownload\新建文件夹\wenda512\wenda\WPy64-31110\python-3.11.1.amd64\Lib\site-packages\torch\utils\cpp_extension.py", line 1909, in _run_ninja_build raise RuntimeError(message) from e RuntimeError: Error building extension 'wkv_cuda'

RuntimeError: Error building extension 'wkv_cuda' 这个问题我该怎么修复

[l15y]

先试试换成英文路径

[shmily91]

解决了，原来时设置显卡的问题

[shmily91]

我把run_rwkv_with_vc.bat文件中 set RWKV_CUDA_ON=1 改成=0就可以了

[shmily91]

先试试换成英文路径 我按你说的改了路径 set RWKV_CUDA_ON=1 可以用了，谢谢大师神

[Aqual-lu]

我把run_rwkv_with_vc.bat文件中 set RWKV_CUDA_ON=1 改成=0就可以了

请问这个run_rwkv_with_vc.bat在哪里呀?