Eigen's default version is in C++14, hence installing LP-SparseMAP with it raises compiler errors.

Suggested change: line 16 of 'setup.py' to '-std=c++14',

Hi Vlad, thank you for the exciting resource. Looking forward to trying it out!

I am trying to install all of the dependencies. I am having trouble getting DyNet installed using the DyNet fork.

I have followed the instructions to install DyNet for C++ here: https://dynet.readthedocs.io/en/latest/install.html#building

Where I have diverged from the official instructions is instead of cloning the official DyNet repo:
git clone https://github.com/clab/dynet.git

I am cloning your custom version:
git clone https://github.com/vene/dynet.git.

I am then switching to the py3 branch: git checkout py3.

I am then running the cmake command as normal:

(base) jbarry@jbarry-desktop:~/dynet-base3/dynet/build$ cmake .. -DEIGEN3_INCLUDE_DIR=/home/jbarry/dynet-base3/eigen -DENABLE_CPP_EXAMPLES=ON
-- Boost dir is Boost_INCLUDE_DIR-NOTFOUND
CMake Error at /usr/share/cmake-3.16/Modules/FindPackageHandleStandardArgs.cmake:146 (message):
  Could NOT find Boost (missing: Boost_INCLUDE_DIR program_options regex
  serialization)
Call Stack (most recent call first):
  /usr/share/cmake-3.16/Modules/FindPackageHandleStandardArgs.cmake:393 (_FPHSA_FAILURE_MESSAGE)
  /usr/share/cmake-3.16/Modules/FindBoost.cmake:2179 (find_package_handle_standard_args)
  CMakeLists.txt:103 (find_package)


-- Configuring incomplete, errors occurred!
See also "/home/jbarry/dynet-base3/dynet/build/CMakeFiles/CMakeOutput.log".

I then tried downloading Boost v 1.62 which seemed to be the most recent version of Boost at the time of the Dynet fork around November 2016 and tried again using the instructions for compiling with Boost:

(lp-sparsemap-pt-dy-custom) jbarry@jbarry-desktop:~/dynet-base3/dynet/build$ cmake .. -DEIGEN3_INCLUDE_DIR=/home/jbarry/dynet-base3/eigen -DENABLE_CPP_EXAMPLES=ON -DENABLE_BOOST=ON
-- Boost dir is /home/jbarry/dynet-base3/boost
CMake Error at /usr/share/cmake-3.16/Modules/FindPackageHandleStandardArgs.cmake:146 (message):
  Could NOT find Boost (missing: program_options regex serialization) (found
  version "1.62.0")
Call Stack (most recent call first):
  /usr/share/cmake-3.16/Modules/FindPackageHandleStandardArgs.cmake:393 (_FPHSA_FAILURE_MESSAGE)
  /usr/share/cmake-3.16/Modules/FindBoost.cmake:2179 (find_package_handle_standard_args)
  CMakeLists.txt:103 (find_package)


-- Configuring incomplete, errors occurred!
See also "/home/jbarry/dynet-base3/dynet/build/CMakeFiles/CMakeOutput.log".

I am probably doing something completely wrong when compiling with Boost. Apologies as I am quite inexperienced with using cmake commands and editing CMake options. Here is how my DyNet base directory looks:

(lp-sparsemap-pt-dy-custom) jbarry@jbarry-desktop:~/dynet-base3$ ls
boost dynet  eigen

ls dynet/
build  cmake  CMakeLists.txt  config.h.cmake  doc  dynet  examples  external  LICENSE  MEM.notes  pyexamples  python  python_tests  README.md  rnnlm  tests  tutorial

I'm just wondering if you have any pointers to install this correctly? I was able to successfully install the official DyNet code in a separate environment without Boost. Attached is the CMakeFiles/CMakeOutput.log if it helps. Many thanks!
CMakeOutput.log

Hey,
Recently Tensorflow and Pytorch have implemented a way to pass tensors from one framework to another using DLpack, and I was wondering if it is possible to somehow use your code in a pytorch model with this type of binding.
Thanks,
Ohad

links:
https://tvm.apache.org/2018/08/10/DLPack-Bridge
tensorflow/tensorflow#24453

Dear Vlad,

Thanks for your amazing work.

I am trying to install lpsmap and having trouble with an error when I use it.

from lpsmap.ad3ext.parse import Parse
ModuleNotFoundError: No module named 'lpsmap.ad3ext.parse'

I first download the eigen as indicated in https://gitlab.com/libeigen/eigen and then perform the following scripts
export EIGEN_DIR=/path/to/eigen
python setup.py build_clib # builds ad3 in-place
pip install . # builds lpsmap and installs

My env is:
Unbuntu16.04
Python: 3.6
Pytorch: 1.6.0

I am not sure if the eigen should be compiled firstly.

Would you please give some suggestions?

Thanks a lot!

I am facing an issue while trying to install the lp-sparsemap package. The installation process is failing with the following error:

I have attempted to upgrade the pip package using the command pip install --upgrade pip, but the issue persists.