I am trying to setup gaitech_edu ROS package in order to build a turtlebot line follower. But, when I am trying to compile the package using catkin_make, I am getting following error:
I installed pocketsphinx library for python 2.7 using
pip install pocketsphinx.
But, even after that I am getting the same error. I am using ros melodic and python version is 2. Your help/advice will be highly appreciated. Thank you in advance.
I am able to solve this problem using this ros package for pocketsphinx.
Related
Maybe someone else asked the same question too. But this question is difficult. I tried everything. The place I am stuck is with installing dependencies. Some of the dependencies are old and not easily available. But I managed to install them.
The problems lies here.. There are dependencies that need to get the build from their source code. I already installed Visual C++ Build and MSMPI. Also installed HDF5 for H5PY but it doesn't let me build old versions of H5PY. So, I tried installing the latest version of H5PY but still, I am stuck at errors like file not found. Some of the files which the build process cannot find are "h5py/h5f.pyx", "mpi_c", "mpi.h". Solving error for one missing file leads to other and so on..
On trying hard to solve such errors and installing one or the other package to do the same task, I am tired up.. Something I found, at last, was that "mpi_c" file was replaced with some other file in newer versions of MPI4PY. But my dependencies depend on older version. I tried installing an older version of MPI4PY but HDF5 won't let me install that giving other errors. At last, I quit the task with my whole day wasted after this.
So can someone here please provide a step by step guideline for installing Rasa Stack on Windows Machine?
Windows 10 with Python 3.7.. Let me know if I need to downgrade python as well.. It was my first time building some project from source with python on windows. Thanks...
Please try the below steps to install Rasa:
Install Conda
Create a virtual environment:
conda create -n myenv python=3.5
Activate the virtual environment
conda activate myenv
pip install rasa_nlu rasa_core
I am fairly new to TensorFlow and just installed it with CPU-support-only version following to this: https://www.tensorflow.org/install/install_linux
My Ubuntu is 16.04 and python installed is 2.7.13. I chose "native" pip installation.
The download and install process went though smoothly, however, when I tried to import Tensorflow and use it, it returns following error;
ImportError: /usr/local/lib/python2.7/site-packages/tensorflow/python/_pywrap_tensorflow_internal.so: undefined symbol: PyUnicodeUCS4_FromString
I have got no clue how to solve it after crawling in google.
If you know some idea on how to solve this and could give me some advice, much appreciated.
Thank you so much in advance!
It's because your python is built with UCS2, which is incompatible with the one assumed by tensorflow (UCS4). So either you build your python with UCS4 (--enable-unicode=ucs4) or build tensorflow from source may solve this issue, I guess.
I have found many many suggestions on how to download these, from using miniconda to using unofficial versions but nothing will work for me? I am just wondering if anyone has any suggestion on a straight forward way I might obtain these as I need them for my project. I am using pycharm with python 2.7 and I have windows 10. Thank you!
You can find precompiled versions of these libraries for most python versions here:
http://www.lfd.uci.edu/~gohlke/pythonlibs/#numpy
http://www.lfd.uci.edu/~gohlke/pythonlibs/#scipy
http://www.lfd.uci.edu/~gohlke/pythonlibs/#matplotlib
Then do:
pip install numpy-something-something.whl
This works for me in PyCharm 5.0.3 using Python 3.5.1. Should be the same procedure for 2.7
I found a code on internet which implement a paper "Effective 3D Action
Recognition Using EigenJoints" code. They implement in Python so I tried to install Python 2.7.9 and numpy, scipy, scikit-learn. All these libraries installed successfully. In the cmd when i tried to run python eigen.py it gives me this error.
Please anyone download this code and run it by yourself or help me to solve this error.
You have to do 2 steps:
download anaconda according to your system and specifications.
Make sure that Anaconda is installed on python version installed on your system. For example, if you have python 3.4, and anaconda in installled on 2.7, then you should remove python 3.4.
Afterwards, the code will run smoothly and all libraries will be well defined. I have made the same ;) Please, let me know if you face new problems.
I tried to install pocketsphinx by
brew instal cmu-pocketsphinx
and
make install (followed instructions here: http://cmusphinx.sourceforge.net/wiki/gstreamer)
the pocketsphinx_continuous works properly, but I can't find a executable pocketsphinx for
gst-inspect pocketsphinx
could anyone tell me where to find it, please?
Many many thanks
pocketsphinx install log: http://pastebin.com/48QU0qjg
You need to check if plugin pocketsphinx.so is installed in
/usr/local/lib/gstreamer-<version>
If plugin is there you need to export GST_PLUGIN_PATH environment variable to update gstreamer search path.
If plugin is not there you need to check build log for the information why plugin was not created or installed properly. Most likely you didn't have required development headers so plugin compilation was disabled.
I had the same problem under Raspbian Jessie. The problem was that when I compiled Pocketsphinx the libgstreamer-plugins-base1.0-dev was missing, so no plugin could be compiled. I solved by installing the missing package then recompiling Pocketsphinx. Check this for more info https://sourceforge.net/p/cmusphinx/discussion/speech-recognition/thread/040066c7/
Hope it helps