Hi guys, I’m a new HTM learner and need some help. I’m intended to install htm.core on Ubuntu 16.04.6 LTS (GNU/Linux 4.4.0-142-generic x86_64). And I tried install and failed. It shows that I have met all the prerequisites for installation:
john@john-VirtualBox:~$ python --version
john@john-VirtualBox:~$ python -m pip install -i https://test.pypi.org/simple/ htm.core
Downloading https://test-files.pythonhosted.org/packages/26/d0/adff7d1cf6650d4806cfabb5451ee3f93e575116eaca5121afd7998c97c3/htm.core-2.0.18-cp37-cp37m-manylinux1_x86_64.whl (2.6MB)
100% |████████████████████████████████| 2.6MB 574kB/s
Collecting hexy>=1.4.3 (from htm.core)
Traceback (most recent call last):
File "/usr/lib/python3/dist-packages/pip/basecommand.py", line 215, in main
status = self.run(options, args)
File "/usr/lib/python3/dist-packages/pip/commands/install.py", line 353, in run
File "/usr/lib/python3/dist-packages/pip/wheel.py", line 749, in build
File "/usr/lib/python3/dist-packages/pip/req/req_set.py", line 380, in prepare_files
File "/usr/lib/python3/dist-packages/pip/req/req_set.py", line 554, in _prepare_file
File "/usr/lib/python3/dist-packages/pip/req/req_install.py", line 278, in populate_link
self.link = finder.find_requirement(self, upgrade)
File "/usr/lib/python3/dist-packages/pip/index.py", line 465, in find_requirement
all_candidates = self.find_all_candidates(req.name)
File "/usr/lib/python3/dist-packages/pip/index.py", line 423, in find_all_candidates
for page in self._get_pages(url_locations, project_name):
File "/usr/lib/python3/dist-packages/pip/index.py", line 568, in _get_pages
page = self._get_page(location)
File "/usr/lib/python3/dist-packages/pip/index.py", line 683, in _get_page
return HTMLPage.get_page(link, session=self.session)
File "/usr/lib/python3/dist-packages/pip/index.py", line 795, in get_page
File "/usr/share/python-wheels/requests-2.18.4-py2.py3-none-any.whl/requests/models.py", line 935, in raise_for_status
raise HTTPError(http_error_msg, response=self)
requests.exceptions.HTTPError: 404 Client Error: Not Found for url: https://test.pypi.org/simple/hexy/
I’ve tried installing htm.core a dozen times now, and reinstalled ubuntu VM on my windows laptop, and no luck. Any help would be greatly appreciated. Is there a recommended OS that I should install instead of ubuntu-18.04.4-desktop-amd6?
(1) Simple Python build (any platform)
** At a command prompt, cd to the root directory of this repository.**
** Run: python setup.py install --user --force
(2) Simple C++ build
After downloading the repository, do the following:
cd path-to-repository mkdir -p build/scripts cd build/scripts cmake …/… make -j install
They both crash at the same point due to thinking the folder …/cmake is a Boost library, here it is from build/ThirdParty/results.txt:
Actually I don’t care what compiler I use, I just want to get an HTM installation running where I can start exploring the code and adding some features for AI modeling I am trying to do. Since I am installing a VM with the latest linux os I can find, I would think it should have the latest compilers. Can you guys recommend a linux distribution that HTM would work on out of the box? I’ll try the gcc-8 advice.
The best to get htm to work in to use the minicoda distrubution for linux (x64). Yes you do get those messages during install , just go through and install then manually ( pip install ) , some of them are version specific so install the version for which you are getting the error. you might have to do this for all other dependencies , after this it works fine. I have gone through the same problem , it works fine . IMPORTANT use miniconda not full Anaconda. I have tested this on Ubuntu 18.04 LTS (x64)
But before I successfully install the package, I would like to know If there is any API document that can help me learn htm.core?
I ask this because I found that the API settings were different from nupic. Nupic has its well written document while I can’t find any tutorial for htm.core, which make it very hard for new learner like me (stupid)…
After this understand what are SP and TP in htm context and how to encode data for them. Finally see coding examples in HTM core github repo. You can check out my repo for HTM , all code is fully commented. Circle back if you have any question .
Well I do not have any specific number to show you, however there are a few problems to over come before you can effectively use Auto encoders as input to SP to TP, One, information is lost when using dimensinal reduction ( 128 to 3 ) + norm of original vector = 4. The reason dimensional reduction is required is the SP itself ( in current implementation ) cannot deal with 128 D directly, images encoding each element of 128 D ( float element ) to RDSE with 1000 bits, this would produce 128000 bit array as input to SP, with SP size itself being 128000.
While my expriment was the step in right direction and using this technique will allow us to encode NLP and vision as input to SP and TP , however I I feel the the developers, researcher at HTM core have figure out a way to effeciently scale core frame work to much more data than previously attempted.
FYI HTM core now has a new vision encoder (eye) I have not trid this yet.