2 errors detected in the compilation of ctc_entrypoint.cu and Could not find libwarpctc.so in ../build.

[jigsawcoder]

I wanted to infer the LGPMA model on my own table dataset. When I tried to install the setup.sh file, the following error occurred.

Installed /content/drive/MyDrive/davar/DAVAR-Lab-OCR Processing dependencies for davarocr==0.6.0 Finished processing dependencies for davarocr==0.6.0 CMake Deprecation Warning at CMakeLists.txt:4 (cmake_minimum_required): Compatibility with CMake < 2.8.12 will be removed from a future version of CMake.

Update the VERSION argument value or use a ... suffix to tell CMake that the project does not need compatibility with older versions.

-- cuda found TRUE -- Building shared library with GPU support -- Configuring done -- Generating done -- Build files have been written to: /content/drive/MyDrive/davar/DAVAR-Lab-OCR/davarocr/davar_rcg/third_party/warp-ctc-pytorch_bindings/build [-11%] Building NVCC (Device) object CMakeFiles/warpctc.dir/src/warpctc_generated_ctc_entrypoint.cu.o /content/drive/MyDrive/davar/DAVAR-Lab-OCR/davarocr/davar_rcg/third_party/warp-ctc-pytorch_bindings/src/ctc_entrypoint.cu(1): error: this declaration has no storage class or type specifier

/content/drive/MyDrive/davar/DAVAR-Lab-OCR/davarocr/davar_rcg/third_party/warp-ctc-pytorch_bindings/src/ctc_entrypoint.cu(1): error: expected a ";"

2 errors detected in the compilation of "/content/drive/MyDrive/davar/DAVAR-Lab-OCR/davarocr/davar_rcg/third_party/warp-ctc-pytorch_bindings/src/ctc_entrypoint.cu". CMake Error at warpctc_generated_ctc_entrypoint.cu.o.cmake:280 (message): Error generating file /content/drive/MyDrive/davar/DAVAR-Lab-OCR/davarocr/davar_rcg/third_party/warp-ctc-pytorch_bindings/build/CMakeFiles/warpctc.dir/src/./warpctc_generated_ctc_entrypoint.cu.o

CMakeFiles/warpctc.dir/build.make:394: recipe for target 'CMakeFiles/warpctc.dir/src/warpctc_generated_ctc_entrypoint.cu.o' failed make[2]: [CMakeFiles/warpctc.dir/src/warpctc_generated_ctc_entrypoint.cu.o] Error 1 CMakeFiles/Makefile2:86: recipe for target 'CMakeFiles/warpctc.dir/all' failed make[1]: [CMakeFiles/warpctc.dir/all] Error 2 Makefile:135: recipe for target 'all' failed make: *** [all] Error 2 Could not find libwarpctc.so in ../build. Build warp-ctc and set WARP_CTC_PATH to the location of libwarpctc.so (default is '../build')

Things I have already tried: -> I followed up on the FAQ but didn't work. -> I tried to comment the lines 26-31 and 36-68 in setput.sh file but that also did not work.

Please help me. Thanks in advance

[qiaoliang6]

The warpctc is a third-party package that hasn't been maintained for a long time, so it may be more restrictive to the environment. You can try using the exact same version of our environment.

Nevertheless, if you do not need to train a text recognition model using warpctc, you can simply ignore this error. And then comment the import code "import warpctc" in the corresponding "init.py" file. It would not affect other functions.

[jigsawcoder]

The warpctc is a third-party package that hasn't been maintained for a long time, so it may be more restrictive to the environment. You can try using the exact same version of our environment.

Nevertheless, if you do not need to train a text recognition model using warpctc, you can simply ignore this error. And then comment the import code "import warpctc" in the corresponding "init.py" file. It would not affect other functions.

I successfully skipped the install of watpctc but I am facing multiple errors while running test_pub.py to infer LGPMA model. From my understanding, these errors are caused due to incompatible mmcv and mmdet versions. I have tested multiple versions but none of them actually worked. Can you please specify which PyTorch, MMCV and MMDET versions should I install in order to just infer the LGPMA model using test_pub.py

Thank you

[qiaoliang6]

The warpctc is a third-party package that hasn't been maintained for a long time, so it may be more restrictive to the environment. You can try using the exact same version of our environment. Nevertheless, if you do not need to train a text recognition model using warpctc, you can simply ignore this error. And then comment the import code "import warpctc" in the corresponding "init.py" file. It would not affect other functions.

I successfully skipped the install of watpctc but I am facing multiple errors while running test_pub.py to infer LGPMA model. From my understanding, these errors are caused due to incompatible mmcv and mmdet versions. I have tested multiple versions but none of them actually worked. Can you please specify which PyTorch, MMCV and MMDET versions should I install in order to just infer the LGPMA model using test_pub.py

Thank you

The model was trained based on pytorch==1.3.0 mmcv-full==1.3.4 and mmdet=2.11.0

[jigsawcoder]

I have install the mentioned versions on pytorch, mmcv-full and mmdet but still facing this error:

AttributeError: module 'torch.jit' has no attribute '_script_if_tracing'

[ZHEGG]

The warpctc is a third-party package that hasn't been maintained for a long time, so it may be more restrictive to the environment. You can try using the exact same version of our environment. Nevertheless, if you do not need to train a text recognition model using warpctc, you can simply ignore this error. And then comment the import code "import warpctc" in the corresponding "init.py" file. It would not affect other functions.

I successfully skipped the install of watpctc but I am facing multiple errors while running test_pub.py to infer LGPMA model. From my understanding, these errors are caused due to incompatible mmcv and mmdet versions. I have tested multiple versions but none of them actually worked. Can you please specify which PyTorch, MMCV and MMDET versions should I install in order to just infer the LGPMA model using test_pub.py

Thank you

I have the same question, how did you skip it? Thanks!