Hi I am having this problem while running I have built the build file by installing the setup.py, but while running the command
"./run_local.sh YOUR_CS_PATH ccnet 60000 256,256 [0|1]". But I am geeting this error of extra_cuda_cflags= ["--expt-extended-lambda"]). My log is printed here.
1]: command not found
Traceback (most recent call last):
File "train.py", line 14, in
from networks.ccnet import ResNet
File "/home/srutsth/PycharmProjects/PolypCCNet/networks/ccnet.py", line 11, in
from cc_attention.functions import CrissCrossAttention
File "/home/srutsth/PycharmProjects/PolypCCNet/cc_attention/functions.py", line 20, in
extra_cuda_cflags= ["--expt-extended-lambda"])
File "/home/srutsth/miniconda3/envs/myenv/lib/python3.6/site-packages/torch/utils/cpp_extension.py", line 644, in load
is_python_module)
File "/home/srutsth/miniconda3/envs/myenv/lib/python3.6/site-packages/torch/utils/cpp_extension.py", line 813, in _jit_compile
with_cuda=with_cuda)
File "/home/srutsth/miniconda3/envs/myenv/lib/python3.6/site-packages/torch/utils/cpp_extension.py", line 847, in _write_ninja_file_and_build
verbose)
File "/home/srutsth/miniconda3/envs/myenv/lib/python3.6/site-packages/torch/utils/cpp_extension.py", line 904, in _prepare_ldflags
print('Detected CUDA files, patching ldflags')
Hi @shrutishrestha Maybe the new branch pure-python could solve the problem. It's a pure python implementation of CCNet and it does not need to be compiled with CUDA.
UPDATE - https://github.com/speedinghzl/CCNet/tree/pure-python works.
@johndpope Thanks for your reply. I recently have updated the pure-python branch.
speedinghzl
