TensorFlow in 5 Minutes (tutorial)

This video is all about building a handwritten digit image classifier in Python in under 40 lines of code (not including spaces and comments). We'll use the popular library TensorFlow to do this.

Please subscribe! That would make me the happiest, and encourage me to output similar content.

The source code for this video is here:

Here are some great links on TensorFlow:

Tensorflow setup:

A similar written tutorial by Google:

Tensorflow Course:

Awesome intro to Tensorflow:

Some other great introductory examples using Tensorflow:

I recently created a Patreon page. If you like my videos, feel free to help support my effort here!:

Follow me:
Twitter:
Facebook: Instagram: Instagram:
Signup for my newsletter for exciting updates in the field of AI:

Hit the Join button above to sign up to become a member of my channel for access to exclusive content! Join my AI community: Sign up for my AI Sports betting Bot, WagerGPT! (500 spots available):

Joe Lilli
 

  • @VarunRajaram says:

    Just installed Tensorflow binary, and I saw this video in my sub box. Thanks for continuing to make awesome videos!

  • @mythrail says:

    your slinkies are not properly glued to your head

  • @HichameMoriceau says:

    Hey Siraj, awesome video, I really like the fact that you keep them ~5 minutes long!

  • @bayesian7404 says:

    I didn’t think that a meaningful intro to Tensorflow and really building a neural net could be done. Great job. I am a convert (not a follower) to your training style. Now I will watch all of your 5 minute tutorials. Thanks dude u freaking rule.

  • @witczak says:

    Sorry, I blinked and missed it. Gotta replay now.

  • @SwanandKulkarni2194 says:

    Those who are trying to recreate/use this program with TF 1.0 have following changes to do:

    –> import input_data
    <> from tensorflow.examples.tutorials.mnist import input_data

    –> tf.histogram_summary(“weights”, W)
    <> tf.summary.histogram(name, value, collections = none)

    –> tf.scalar_summary(“cost_function”, cost_function)
    <> tf.summary.scalar(name, tensor, collection = none)

    –> tf.initialize_all_variables()
    <> tf.global_variables_initializer()

    –> tf.merge_all_summaries()
    <> tf.summary.merge_all()

    –> tf.train.SummaryWriter(‘LOCATION”, graph_def = sess.graph_def)
    <> tf.summary.FileWriter(‘LOCATION’, sess.graph)

  • @davvka says:

    I am impressed with what you are doing to popularize Maschine Learning. And it is hell important. All the companies (including us) are starving for ML engineers. Go on!!!

  • @masoudmasoumimoghaddam3832 says:

    Dude, You teach like you really mean it. I like teachers with that much enthusiasm.
    Keep making tutorials like this one

  • @harindaka says:

    Moral of the video: TF in 5 minutes is a bad idea

  • @dk0money says:

    @Siraj Raval, your explanations are spot on, concise and to the point, with just enough funny clips to make your videos engaging. Love your breakdown of actual code.

  • @rodrigocardoso7846 says:

    Hi Siraj, I’m talking from Brazil. Your teaching is AWESOME and is extremelly fun learn with you. Congratulations!!!

  • @johnhammer8668 says:

    If you read the tensorflow low level api tutorial page and then re watching this video is so much rewarding.

    • @joseortiz_io says:

      Yes I agree. I think its better to learn the low lever API first to get a better understanding of how Tensorflow Works. That’s what I’m doing on my YouTube channel. Check it out 😊

  • @Lamarr168 says:

    That was freakin’ insane! You covered so much machine learning ground in under 6 minutes. Fun times!

    • @RichardHartnell says:

      Agreed, it’s not a basic lesson but surprisingly concise and accurate. Makes me wish I remember more linear algebra, but apparently you don’t even need to do that now that there’s tools like this haha

  • @dudeophd2137 says:

    This video is very useful in that it starts from the virtual environment management and all the way to the description of full models in TensorFlow! Thanks for a great video.

  • @BenRangel says:

    I love the concept of limiting intro tutorials to 5 minutes. I think that I got a better overview of the topic in this video than I would with a 1 hour talk at a conference.
    Just the mere fact that you showed so many things (install process, data sets, code and graphs) in such a short time made me grasp the concept “as a whole”

  • @alex29091977 says:

    Finally make it work in less than 40 min, thanks for a great video of introduction to Tensorflow

  • @vegitax says:

    Came for Tensorflow, stayed for that wicked wild hair.

    • @animalspirits5141 says:

      The hair doesn’t move by itself enough for me to call it wild or wicked. But it does look cooler than average.

    • @MB-hz7wm says:

      You know he bounced out of bed at 3am to create that video because his hair looked so awesome – ha. He’s one of my top 3 guys in this arena who can easily condense these concepts into digestible, 5min segments. Well done ~

  • @kemchobhenchod says:

    I feel like I”m having a separate conversation with your hair.

  • @Cyranek says:

    I was dumb for thinking I would be any better at tensor flow after 5 minutes

  • @karldavis7392 says:

    Just enough info to let me decide whether TensorFlow is the tool I need for the question I’m facing – thanks.

  • >