SWIG missing... about ell HOT 9 CLOSED

from ell.

Thanks Jacob. It helped move forward bit... However I run into few other issues..

(py36) root@8fa27a83d7dd:~/ELL/build/tutorials/vision/gettingStarted# python darknetDemo.py
Traceback (most recent call last):
File "darknetDemo.py", line 4, in
import cv2
ModuleNotFoundError: No module named 'cv2'

(py36) root@8fa27a83d7dd:~/ELL/build/tutorials/vision/gettingStarted# sh compile_darknetReference.sh

Compiling darknetReference from ELL, emitting LLVM IR
Command line parse error:
Cannot read from specified input map file: darknetReference.map

Parsing darknetReference from ELL, generating SWIG interface
Command line parse error:
Cannot read from specified input map file: darknetReference.map

Compiling darknetReference LLVM IR, generating model object file
compile_darknetReference.sh: 9: compile_darknetReference.sh: llc: not found

Running disutils setup to compile and link model + SWIG wrappers into Python module, and install package locally
running build
running build_ext
building '_darknetReference' extension
swigging darknetReference.i to darknetReference_wrap.cpp
swig -python -modern -c++ -py3 -Fmicrosoft -I/root/ELL/interfaces/common -I/root/ELL/interfaces/common/include -I/root/ELL/libraries/emitters/include -o darknetReference_wrap.cpp darknetReference.i
Unable to find file 'darknetReference.i'.
error: command 'swig' failed with exit status 1

from ell.

OR with microsoft cognitive kit

(py36) root@8fa27a83d7dd:~/ELL/build/tutorials/vision/gettingStarted# python cntkDemo.py
Traceback (most recent call last):
File "cntkDemo.py", line 4, in
import cv2
ModuleNotFoundError: No module named 'cv2'

from ell.

ModuleNotFoundError: No module named 'cv2'
Looks like you need opencv

Thats covered here:
https://github.com/Microsoft/ELL/blob/4d3c0cc6ed13cc7885150ff74b5d3788f3ee36e8/tutorials/vision/gettingStarted/opencv.md
conda install -c conda-forge opencv

I actually got stuck here as well (on osx) with this issue:

ImportError: dlopen(/Users/jacobrosenthal/miniconda3/envs/py36/lib/python3.6/site-packages/cv2.cpython-36m-darwin.so, 2): Library not loaded: @rpath/libopenblasp-r0.2.19.dylib
  Referenced from: /Users/jacobrosenthal/miniconda3/envs/py36/lib/libopencv_hdf.3.2.0.dylib
  Reason: image not found

even though i installed openblas per instructions with both brew brew install homebrew/science/openblas and conda conda install openblas=0.2.19

This issue said conda-forge/opencv-feedstock#69
conda update --all -c conda-forge

Which got me up and running with python darknetDemo.py

from ell.

Thanks a lot Jacob. Unfortunately it didn't work in my case after running all 3 below..

conda install -c conda-forge opencv
conda install openblas=0.2.19
conda update --all -c conda-forge

(py36) root@8fa27a83d7dd:~/ELL/build/tutorials/vision/gettingStarted# python cntkDemo.py
Traceback (most recent call last):
File "cntkDemo.py", line 4, in
import cv2
ImportError: libGL.so.1: cannot open shared object file: No such file or directory

(py36) root@8fa27a83d7dd:~/ELL/build/tutorials/vision/gettingStarted# python darknetDemo.py
Traceback (most recent call last):
File "darknetDemo.py", line 4, in
import cv2
ImportError: libGL.so.1: cannot open shared object file: No such file or directory

from ell.

Probably unrelated, but note I didnt do conda install -c conda-forge opencv

Youve got a different error there you should google
ImportError: libGL.so.1: cannot open shared object file: No such file or directory

from ell.

Thanks a lot Jacob. I am able to move forward much more.. Appreciate all your help.

from ell.

Care to post your fix for others? Also if this is resolved please close your issue.

from ell.

Certainly, all I did is follow step 5 ELL Python Language Bindings and Step 6.Microsoft Cognitive Toolkit install steps again and then used Step 7:Darknet model to .. run .. microsoft cognitive toolkit didn't quite work.. but darnet model seems to atleast run, but also bailed out at the end ...

Any case we can close this here and I opened new..

from ell.