Tensorflow

If you don't have an intel processor you're fucked - the TF package you get from pip won't run so you need to build it from source, and I cannot for the life or me compile it without bazel fucking up on three different machines, one of them even had an intel processor anyway. Last resort you can rent out a google server to do training on,

just use pytorch

Or tensorflow-gpu

I don’t want to be a meme data scientist. I have a project that I want to do and meme machine learning is the best to do it

pip install tensorflow

I type this into command line or where?

why the hell you are working on machine learning framework if you dont fucking know were to type this?

>using pip
Anyone doing any recent work needs to compile from source

Because I’m going to learn from scratch. Run Forrest Run.

Attached: E555CB42-15F3-4E4B-AD15-F9909145EADA.jpg (1280x800, 109K)

Well, at least that's respectable. Yes, type it on the command line.

>I have a project that I want to do and meme machine learning is the best to do it
Do you have a dataset consisting of at least 10,000 examples of the result you want?
No?
Give up.

In many cases it's perfectly feasible to artificially generate datasets.

Thanks for the support user
I have multiple huge datasets of 10,000+. Some I found and some I’ve assembled (not 10,000+ yet though) though autistic searching. I got the retard strength buddy boy.

Attached: 93950354-9BD4-4E78-884A-69B8885285AA.jpg (800x450, 40K)

the absolute state of Jow Forums

also make sure you are running linux, you can tell by not having windows logo on bottom-left of your screen

AI was a 2016-2017 meme, we're onto ray tracing and dark modes now.

Whoop dee doo, it's going to be a whole 10% faster, and you'll only have to install google's shitshow of a build tool bazel and spend a week to make it work

Come on user, you know that's not how it works. First you need to set up a docker container just so the dependencies don't fuck your actual development system.

>ray tracing and dark modes
Redpill me on this

Why don't you contribute positively towards anything?

spread your wings and fly, user

>Why don't you contribute positively towards anything?
Are you talking to OP user? What did he do that warrant such a comment?

sudo pacman -S python-tensorflow

feels good to be an archfag and never have to use pip

whats an archfag

a fag using a terrible broken distro which needs daily attention to function

One who is chiefest among fags.

>pip install --user tensorflow
I did what you said user and it's still not working. Why?

Attached: Screen Shot 2018-08-20 at 3.36.58 PM.png (1578x872, 262K)

>python
>import tensorflow
Then ctrl+D

>its not working
sorry, but people unironically asking for help with this """description""" should get at least a warning from mods

how ironic

what is your os?

btw

get in python console
python
in the console
help()
and than
modules

if you dont have tenseflow in there

maybe you didn't install the libs that require

show your errors or something

>successfully installed [...] tensorflow
>it's not working

>I have a project that I want to do and meme machine learning is the best to do it
Use SciKit-Learn, takes like 10 minutes to get running even if you barely know python. You'll still get your cock sucked for 'le machine learning', don't worry.

There are literally tutorials with the exact commands you need to run. Use Google ffs

I'm already doing that user. Or am I not?
user, I'm new here and unfamiliar with the protocol. How should I say "it's not working" in future? Am I suppose to describe what's not working? I thought the screenshot would suffice as an explanation.

Attached: Screen Shot 2018-08-20 at 3.42.25 PM.png (674x208, 33K)

What is a windows logo?

user, I've done that and they are all Pajeets and I found only for Mac that was adequate. I followed it and it didn't work. That's why I'm here now.
MacOS
I installed the libraries on Anaconda though

Attached: Screen Shot 2018-08-20 at 3.46.12 PM.png (2372x822, 124K)

Also nice name

conda create -n tensorflow pip python=3.5(or your python version)

activate tensorflow

pip install --ignore-installed --upgrade tensorflow

create your anaconda env and invoke python

>>> import tensorflow as tf
>>> hello = tf.constant('Hello, TensorFlow!')
>>> sess = tf.Session()
>>> print(sess.run(hello))

according to tenserflow Manuel

>I thought the screenshot would suffice as an explanation.
nothing in your screenshot indicates a problem, those were just notifications of minor importance
you see, python modules are not loaded from $PATH, thats for executables

four squares within a square

did it help to you? maybe i forgot to say to remove tenserflow old libs before you try that

Remember when we used to have a thread for stupid questions like the one in the OP?

read the manuel install gentoo

python36.com/how-to-install-tensorflow-gpu-with-cuda-9-2-for-python-on-ubuntu/

Suck my white swollen dick, you gigantic retard

OP is wh*te

fuck white peopl

Dont use anaconda you retards. It fucks with your actual python libraries. Even commenting out anaconda in bashrc doesnt help.
And OP is a lost cause, he doesn't know shit and wants to install tensorflow after reading upon ML from a medium article lmao.
Start with scikitlearn tutorials and move up from there, read upon keras, then how tensorflow is different from it, then latest is pytorch. Way more optimised than any library ive seen in ML. The only thing which comes close is caffe but it's in C/C++
Anyway all the best OP, any more doubts i can help, whatchu working on?
t.datascientist at rakuten

Are there any good hands on tutorials for machine learning out there?

I don't want a fucking lecture and I don't want to know all the fucking theory.

I want to learn how to actually use this shit. Even if it's a pajeet it's okay. I've found some good ones on Youtube but 95% is just garbage.

Share your sources guys. Deep learning/unsupervised learning preferred.

You're a good sport OP, I'm sure you can make it.
Tensorflow is available as an npm package, so you can try that if you're hopeless and can live with using JavaScript.
Install node.js > make npm account > link account to npm client (within your terminal) > cd into your directory > type "npm install tensorflow".

Here is some documentation : npmjs.com/package/tensorflow

Attached: 1534092402963.gif (460x405, 98K)

What should I use if I want to work with text?

I have a data set of 200K strings. 100K of those strings contains one specific string at the end of the string and 100K does not.

I want to make a neural network that can predict if a random string will have the string at the end or not.

Example:

Positive:
"Cats are nice animals, mine is black! [SUBSTRING]"

Negative/control:
"Buildings usually have a roof, windows, and one or more door."

read about information retrieval courses first and use nltk library. Go from there.

Well, wouldn't it be possible to tokenize each string and throw the two data sets at a deep learning algorithm to have it make sense of the data? All the strings are created by humans, the substring is the same in all of the 100K positive strings and is basically just a marker that they are different from the control ones. There is a common theme among the ones with the substring though, it's just too complicated for me to dissect it out manually.

like this.

Attached: wew.png (1920x1080, 543K)

if you can't install it, how can you comprehend the data it generates?

Yeah but you have to tell it what to do. You can't just give it data and tell it to suck your dick.
Labels are different when it comes to text, put either some common sentences so that it can start off in the right way. Use recurrent shit on text and learn from it - look at TEDRNN, it gives a better idea to you. Or even the NELL data collection. That should answer all your doubts.

Again you look like someone lost on fundamentals. Get atleast a working knowledge well else you'll keep getting stuck at places like this.

>You can't just give it data and tell it to suck your dick.

Pic related.

Thanks though, I'll look into what you've said. I appreciate it, really.

Attached: 1534160285446.jpg (640x470, 39K)

holy moly
you're not even close to ready to work on this shit
what are you doing

Attached: stop.png (280x272, 3K)

Run Forresr Run

Any retard can install tensor flow.

Caffe on the other hand I gave up trying.

You're using pycharm and anaconda? Make sure you have the environment set to the one where you have tensorflow installed... In your case test. You'll need to do that in the project interpreter setting in pycharm

Then what does that make OP?

>Any retard can install tensor flow
>OP cannot into install
>Therefore OP is not retarded
Boolean logic checks out.