AttributeError: module 'warprnnt_pytorch.warp_rnnt' has no attribute 'gpu_rnnt'

[li563042811]

Please help me with this issue.
I build targets like the installation instructions. I modified the env via $CUDA_HOME before build and used

cmake -DCUDA_TOOLKIT_ROOT_DIR=$CUDA_HOME ..
make

When I run test.py this problem still happened. I don't why.

python test/test.py
CPU Tests passed!
Traceback (most recent call last):
File "test/test.py", line 170, in
small_test()
File "test/test.py", line 60, in small_test
cost, grads = wrap_and_call(acts, labels)
File "test/test.py", line 44, in wrap_and_call
costs = fn(acts, labels, lengths, label_lengths)
File "/data/jsa/miniconda3/lib/python3.7/site-packages/torch/nn/modules/module.py", line 727, in _call_impl
result = self.forward(*input, **kwargs)
File "/data/jsa/miniconda3/lib/python3.7/site-packages/warprnnt_pytorch-0.1-py3.7-linux-x86_64.egg/warprnnt_pytorch/init.py", line 100, in forward
return self.loss(acts, labels, act_lens, label_lens, self.blank, self.reduction)
File "/data/jsa/miniconda3/lib/python3.7/site-packages/warprnnt_pytorch-0.1-py3.7-linux-x86_64.egg/warprnnt_pytorch/init.py", line 23, in forward
loss_func = warp_rnnt.gpu_rnnt if is_cuda else warp_rnnt.cpu_rnnt
AttributeError: module 'warprnnt_pytorch.warp_rnnt' has no attribute 'gpu_rnnt'

-- cuda found TRUE
-- Building shared library with GPU support
-- Configuring done
-- Generating done
-- Build files have been written to: /lustr/jsa/env/download/warp-transducer-master/build
[ 7%] Linking CXX shared library libwarprnnt.so
[ 14%] Built target warprnnt
[ 21%] Linking CXX executable test_gpu
[ 35%] Built target test_gpu
[ 42%] Linking CXX executable test_time_gpu
[ 57%] Built target test_time_gpu
[ 64%] Linking CXX executable test_time
[ 78%] Built target test_time
[ 85%] Linking CXX executable test_cpu
[100%] Built target test_cpu

echo $CUDA_HOME
/data/ren/install_dir/cuda-10.1
echo $WARP_RNNT_PATH
/lustr/jsa/env/download/warp-transducer-master/build

[fd873630]

Please help me with this issue. I build targets like the installation instructions. I modified the env via $CUDA_HOME before build and used

cmake -DCUDA_TOOLKIT_ROOT_DIR=$CUDA_HOME .. make

When I run test.py this problem still happened. I don't why.

python test/test.py CPU Tests passed! Traceback (most recent call last): File "test/test.py", line 170, in small_test() File "test/test.py", line 60, in small_test cost, grads = wrap_and_call(acts, labels) File "test/test.py", line 44, in wrap_and_call costs = fn(acts, labels, lengths, label_lengths) File "/data/jsa/miniconda3/lib/python3.7/site-packages/torch/nn/modules/module.py", line 727, in _call_impl result = self.forward(*input, **kwargs) File "/data/jsa/miniconda3/lib/python3.7/site-packages/warprnnt_pytorch-0.1-py3.7-linux-x86_64.egg/warprnnt_pytorch/init.py", line 100, in forward return self.loss(acts, labels, act_lens, label_lens, self.blank, self.reduction) File "/data/jsa/miniconda3/lib/python3.7/site-packages/warprnnt_pytorch-0.1-py3.7-linux-x86_64.egg/warprnnt_pytorch/init.py", line 23, in forward loss_func = warp_rnnt.gpu_rnnt if is_cuda else warp_rnnt.cpu_rnnt AttributeError: module 'warprnnt_pytorch.warp_rnnt' has no attribute 'gpu_rnnt'

-- cuda found TRUE -- Building shared library with GPU support -- Configuring done -- Generating done -- Build files have been written to: /lustr/jsa/env/download/warp-transducer-master/build [ 7%] Linking CXX shared library libwarprnnt.so [ 14%] Built target warprnnt [ 21%] Linking CXX executable test_gpu [ 35%] Built target test_gpu [ 42%] Linking CXX executable test_time_gpu [ 57%] Built target test_time_gpu [ 64%] Linking CXX executable test_time [ 78%] Built target test_time [ 85%] Linking CXX executable test_cpu [100%] Built target test_cpu

echo $CUDA_HOME /data/ren/install_dir/cuda-10.1 echo $WARP_RNNT_PATH /lustr/jsa/env/download/warp-transducer-master/build

I have same issue in my docker env

I solve this problem when i build cmake file with gpu

and i export cpath like this

export CPATH=/usr/local/cuda-10.1/targets/x86_64-linux/include:$CPATH
export LD_LIBRARY_PATH=/usr/local/cuda-10.1/targets/x86_64-linux/lib:$LD_LIBRARY_PATH
export PATH=/usr/local/cuda-10.1/bin:$PATH

https://stackoverflow.com/questions/13167598/error-cuda-runtime-h-no-such-file-or-directory/43389168