rtp-llm
rtp-llm copied to clipboard
alibaba
最新代码源码编译报错
看到最近的提交，rtp-llm支持cosyvoice。尝试源码编译，执行以下指令时报错，想请教下应该怎么编译源码？
bazelisk build //maga_transformer:maga_transformer --jobs 100 --verbose_failures --config=cuda12_2
异常输出如下：
025/02/24 03:42:08 Downloading https://releases.bazel.build/6.4.0/release/bazel-6.4.0-linux-x86_64...
Downloading: 52 MB out of 52 MB (100%)
INFO: Repository local_config_cuda instantiated at:
/home/wangleyi/rtp-llm/WORKSPACE:7:15: in <toplevel>
Repository rule cuda_configure defined at:
/home/wangleyi/rtp-llm/3rdparty/cuda_config/cuda_configure.bzl:1427:33: in <toplevel>
ERROR: An error occurred during the fetch of repository 'local_config_cuda':
Traceback (most recent call last):
File "/home/wangleyi/rtp-llm/3rdparty/cuda_config/cuda_configure.bzl", line 1396, column 38, in _cuda_autoconf_impl
_create_local_cuda_repository(repository_ctx)
File "/home/wangleyi/rtp-llm/3rdparty/cuda_config/cuda_configure.bzl", line 959, column 35, in _create_local_cuda_repository
cuda_config = _get_cuda_config(repository_ctx, find_cuda_config_script)
File "/home/wangleyi/rtp-llm/3rdparty/cuda_config/cuda_configure.bzl", line 647, column 30, in _get_cuda_config
config = find_cuda_config(repository_ctx, find_cuda_config_script, ["cuda", "cudnn", "nccl"])
File "/home/wangleyi/rtp-llm/3rdparty/cuda_config/cuda_configure.bzl", line 624, column 26, in find_cuda_config
exec_result = execute(repository_ctx, [python_bin, repository_ctx.attr._find_cuda_config] + cuda_libraries)
File "/home/wangleyi/rtp-llm/3rdparty/cuda_config/common.bzl", line 217, column 13, in execute
fail(
Error in fail: Repository command failed
Could not find any nccl.h matching version '2' in any subdirectory:
''
'include'
'include/cuda'
'include/*-linux-gnu'
'extras/CUPTI/include'
'include/cuda/CUPTI'
'local/cuda/extras/CUPTI/include'
of:
'/usr/local/cuda/'
ERROR: /home/wangleyi/rtp-llm/WORKSPACE:7:15: fetching cuda_configure rule //external:local_config_cuda: Traceback (most recent call last):
File "/home/wangleyi/rtp-llm/3rdparty/cuda_config/cuda_configure.bzl", line 1396, column 38, in _cuda_autoconf_impl
_create_local_cuda_repository(repository_ctx)
File "/home/wangleyi/rtp-llm/3rdparty/cuda_config/cuda_configure.bzl", line 959, column 35, in _create_local_cuda_repository
cuda_config = _get_cuda_config(repository_ctx, find_cuda_config_script)
File "/home/wangleyi/rtp-llm/3rdparty/cuda_config/cuda_configure.bzl", line 647, column 30, in _get_cuda_config
config = find_cuda_config(repository_ctx, find_cuda_config_script, ["cuda", "cudnn", "nccl"])
File "/home/wangleyi/rtp-llm/3rdparty/cuda_config/cuda_configure.bzl", line 624, column 26, in find_cuda_config
exec_result = execute(repository_ctx, [python_bin, repository_ctx.attr._find_cuda_config] + cuda_libraries)
File "/home/wangleyi/rtp-llm/3rdparty/cuda_config/common.bzl", line 217, column 13, in execute
fail(
Error in fail: Repository command failed
Could not find any nccl.h matching version '2' in any subdirectory:
''
'include'
'include/cuda'
'include/*-linux-gnu'
'extras/CUPTI/include'
'include/cuda/CUPTI'
'local/cuda/extras/CUPTI/include'
of:
'/usr/local/cuda/'
ERROR: Analysis of target '//maga_transformer:maga_transformer' failed; build aborted: no such package '@local_config_cuda//cuda': Repository command failed
Could not find any nccl.h matching version '2' in any subdirectory:
''
'include'
'include/cuda'
'include/*-linux-gnu'
'extras/CUPTI/include'
'include/cuda/CUPTI'
'local/cuda/extras/CUPTI/include'
of:
'/usr/local/cuda/'
INFO: Elapsed time: 0.141s
INFO: 0 processes.
FAILED: Build did NOT complete successfully (0 packages loaded, 0 targets configured)
currently loading:
cuda12_2 已弃用，改用 cuda12 或者 cuda12_6 之前conda环境是通过以下指令初始化的python环境、并进行源码编译
pip3 install -r open_source/deps/requirements_torch_gpu_cuda12.txt
bazelisk build //maga_transformer:maga_transformer --jobs 100 --verbose_failures --config=cuda12_2
修改编译指令cuda12，执行以下指令
bazelisk build //maga_transformer:maga_transformer --jobs 100 --verbose_failures --config=cuda12
报错如下：
INFO: Build options --action_env and --host_action_env have changed, discarding analysis cache.
INFO: Repository local_config_cuda instantiated at:
/home/wangleyi/rtp-llm/WORKSPACE:7:15: in <toplevel>
Repository rule cuda_configure defined at:
/home/wangleyi/rtp-llm/3rdparty/cuda_config/cuda_configure.bzl:1427:33: in <toplevel>
ERROR: An error occurred during the fetch of repository 'local_config_cuda':
Traceback (most recent call last):
File "/home/wangleyi/rtp-llm/3rdparty/cuda_config/cuda_configure.bzl", line 1396, column 38, in _cuda_autoconf_impl
_create_local_cuda_repository(repository_ctx)
File "/home/wangleyi/rtp-llm/3rdparty/cuda_config/cuda_configure.bzl", line 959, column 35, in _create_local_cuda_repository
cuda_config = _get_cuda_config(repository_ctx, find_cuda_config_script)
File "/home/wangleyi/rtp-llm/3rdparty/cuda_config/cuda_configure.bzl", line 647, column 30, in _get_cuda_config
config = find_cuda_config(repository_ctx, find_cuda_config_script, ["cuda", "cudnn", "nccl"])
File "/home/wangleyi/rtp-llm/3rdparty/cuda_config/cuda_configure.bzl", line 624, column 26, in find_cuda_config
exec_result = execute(repository_ctx, [python_bin, repository_ctx.attr._find_cuda_config] + cuda_libraries)
File "/home/wangleyi/rtp-llm/3rdparty/cuda_config/common.bzl", line 217, column 13, in execute
fail(
Error in fail: Repository command failed
Could not find any cuda.h matching version '12.4' in any subdirectory:
''
'include'
'include/cuda'
'include/*-linux-gnu'
'extras/CUPTI/include'
'include/cuda/CUPTI'
'local/cuda/extras/CUPTI/include'
of:
'/usr/local/cuda/'
ERROR: /home/wangleyi/rtp-llm/WORKSPACE:7:15: fetching cuda_configure rule //external:local_config_cuda: Traceback (most recent call last):
File "/home/wangleyi/rtp-llm/3rdparty/cuda_config/cuda_configure.bzl", line 1396, column 38, in _cuda_autoconf_impl
_create_local_cuda_repository(repository_ctx)
File "/home/wangleyi/rtp-llm/3rdparty/cuda_config/cuda_configure.bzl", line 959, column 35, in _create_local_cuda_repository
cuda_config = _get_cuda_config(repository_ctx, find_cuda_config_script)
File "/home/wangleyi/rtp-llm/3rdparty/cuda_config/cuda_configure.bzl", line 647, column 30, in _get_cuda_config
config = find_cuda_config(repository_ctx, find_cuda_config_script, ["cuda", "cudnn", "nccl"])
File "/home/wangleyi/rtp-llm/3rdparty/cuda_config/cuda_configure.bzl", line 624, column 26, in find_cuda_config
exec_result = execute(repository_ctx, [python_bin, repository_ctx.attr._find_cuda_config] + cuda_libraries)
File "/home/wangleyi/rtp-llm/3rdparty/cuda_config/common.bzl", line 217, column 13, in execute
fail(
Error in fail: Repository command failed
Could not find any cuda.h matching version '12.4' in any subdirectory:
''
'include'
'include/cuda'
'include/*-linux-gnu'
'extras/CUPTI/include'
'include/cuda/CUPTI'
'local/cuda/extras/CUPTI/include'
of:
'/usr/local/cuda/'
ERROR: Analysis of target '//maga_transformer:maga_transformer' failed; build aborted: no such package '@local_config_cuda//cuda': Repository command failed
Could not find any cuda.h matching version '12.4' in any subdirectory:
''
'include'
'include/cuda'
'include/*-linux-gnu'
'extras/CUPTI/include'
'include/cuda/CUPTI'
'local/cuda/extras/CUPTI/include'
of:
'/usr/local/cuda/'
INFO: Elapsed time: 0.166s
INFO: 0 processes.
FAILED: Build did NOT complete successfully (4 packages loaded, 6 targets configured)
currently loading:
想请教下，这种情况是需要重新安装cuda吗，有没有可以编译的镜像环境？😂
可以用类似下面的软链接解决这个问题
ln -s /usr/include/nccl.h /usr/local/cuda/include/nccl.h
ln -s /usr/lib/x86_64-linux-gnu/libnccl.so.2.19.3 /usr/local/cuda/lib64/libnccl.so.2.19.3
ln -s /usr/lib/x86_64-linux-gnu/libnccl.so.2 /usr/local/cuda/lib64/libnccl.so.2
