My First Machine Learning Project (Tensorflow Deep-Learning)
Key Takeaways
This video demonstrates a beginner's first machine learning project using TensorFlow and Python, creating a deep learning neural network trained on the MNIST dataset to guess handwritten digits with high accuracy.
Full Transcript
hey guys and welcome back in this video I'm going to be talking to you about machine learning and how I made my first machine learning application so pretty much I am working in Python here I'm using tensorflow the GPU version on my computer and I managed to create a number guessing program or model whatever you want to call it so pretty much given handwritten digits it's able to guess what number that actually is with a fair degree of accuracy so rather than me just talking about it why don't I show you first of all so I have these four files up here create model draw number test network neural network so the way this works is using a very basic neural network I'm just gonna run this test network script because well I've already created the model so we just want to run it and see if it works okay so we're using tensorflow it just takes a second to boot up and then once it does I've actually made this like a visual representation for you guys so you can see exactly what it looks like okay so just using matplotlib if you know what that is in python i've just drawn out all of the input data that you load in from tensorflow because it has like a default a bunch of these images that you feed in to train the model so I've loaded in a seven okay and then you can see down here says I predict this numbers a seven the number actually is a seven okay and I didn't just put like what the number is printed twice this actually works and I'm going to show you another example in just a second so if I close this you can see it loads up a two we get to this number actually is a two and we can just keep going through and it'll keep guessing what the number is and maybe if we get lucky it'll get one wrong but this does have like a 98% accuracy or something like that so chances are probably not and these here are meant to simulate handwritten digits just obviously like with pixels okay nine and I think I do this what like ten times we'll just go through and show you okay so there you go this program got zero wrong out of ten that the length of the entire test data set we only went through ten of them okay that's pretty cool and that looks like wow that's working really well but how does that work well I'm gonna show you now this working with me actually drawing numbers and this is something that I created myself and I'll show you because the other ones I kind of follow the tutorial to do so main draw numbers that's what we need to run okay so I'm just using PI game here and I have like a basic grid and pretty much if I decide to draw a number on the grid like that so it's seven and I press the spacebar then tensorflow loads up my model and it's predicts what the number is so it says I predict this number is a seven and you guys see I didn't hard-code anything in there saying like this is gonna be a seven I just true a seven to the screen now this does not work very well and I'll show you like if I draw three chances are it's not gonna be able to guess that this is a three oh god it's a three but sometimes like if you do complex numbers like sixes or eights especially when I'm drawing with a mouse and not with like a pencil it's really difficult for the computer yes even if I do like a one like this let's see okay it's not proving me you're all right here okay let's do a six so it's a crude six but if we take a guess it says oh I predict this number is a six so that actually worked but sometimes this tends to crash and I think I can probably get it to crash you guys okay so let's try an eight click the spacebar I predict those numbers of three okay so you guys can see that this has some resemblance to a 3 so it predicts that it's a three and this is just cuz I don't draw it exactly like the test data that I've sent in and this is kind of showing you the flaw in some of these neural networks and why I need to improve them if I draw like a zero again a crude zero all right see I predict this numbers of two so it makes quite a few mistakes because the data that I'm drawing in here doesn't look like the data that it was shown in most cases okay so that's how that works so pretty much the way that I coded this is using like a standard neural network I set up one like the hard way by typing out all the layers myself and then I did one in create model here just using the new version of the care ass at it caris I don't know you say adding layers and this is literally the entire model like this okay this is creating the model this is getting the test out of here this is training the model and then this would be all this code down here would be testing models you guys can see that this whole machine learning thing that's based off of is only done in 30 lines the longest part is actually grabbing the information I draw in on a grid and that's what this whole file is here draw number dot pi because it has to get actually all the pixels on the grid and put it into a form that our neural network can look at and read so anyways the way that the reason I showed you guys this is cuz I kind of want to get you excited about machine learning and what's coming to the channel I'm planning on making tutorials about how to do some of the stuff like this maybe doing a tutorial on how I made this kind of first machine learning stuff once I tweaked it a bit and as soon as I learn more about machine learning and get more comfortable with it you can expect to see a lot of those tutorials on the channel where I'm going to apply them to different projects and make them really useful not just spitting out little text things that you can look at and be like oh that's cool it says it's working right where you can't actually look at the data which is what we want to do so yeah so if you guys want to see these tutorials and see this machine learning stuff make sure you guys let me know in the comments and by hitting that like button because that keeps me motivated and makes me want to produce content faster and I'm really excited with this machine learning stuff and I hope you guys are as well so with that being said I think I'm gonna wrap up the video here I hope you guys enjoyed if you did please make sure you leave a like and subscribe and I'll see you again in another video [Music]
Original Description
This is my very first machine-learning project in python using tensorflow. It utilizes a deep learning neural network and is trained off the mnist data set in tensorflow. If you would like to see a tutorial on this specific project or machine-learning in general make sure you leave a like on this video and let me know in the comments down below!
Source Code: https://github.com/techwithtim/Number-Guesser-Machine-Learning-
Hope you enjoyed :)
WEBSITE: https://techwithtim.net
**************************************************************
Support the Channel: https://www.patreon.com/techwithtim
Join my discord server: https://discord.gg/pr2k55t
Twitter: https://twitter.com/TechWithTimm
Need Help With Python? https://www.fiverr.com/techwithtim
**************************************************************
Please leave a LIKE and SUBSCRIBE for more content!
Tags:
- Tech With Tim
- Deep learning python
- Machine learning project
- Machine learning tensorflow
- Tensoflow python
- Python Tutorials
- Tensorflow projects
Playlist
Uploads from Tech With Tim · Tech With Tim · 0 of 60
← Previous
Next →
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
A* Path Finding Algorithm(Visualization)
Tech With Tim
Python Programming Tutorial #1 - Variables and Data Types
Tech With Tim
Python Programming Tutorial #2 - Basic Operators and Input
Tech With Tim
Python Programming Tutorial #3 - Conditions
Tech With Tim
Python Programming Tutorial #4 - IF/ELIF/ELSE
Tech With Tim
Python Programming Tutorial #5 - Chained Conditionals and Nested Statements
Tech With Tim
Python Programming Tutorial #6 - For Loops
Tech With Tim
Python Programming Tutorial #7 - While Loops
Tech With Tim
Python Programming Tutorial #8 - Lists and Tuples
Tech With Tim
Python Programming Tutorial #9 - Iteration by Item (For Loops Continued...)
Tech With Tim
Python Programming Tutorial #10 - String Methods
Tech With Tim
How to Overclock a NVIDIA GPU
Tech With Tim
Python Programming Tutorial #11 - Slice Operator
Tech With Tim
Python Programming Tutorial #12 - Functions
Tech With Tim
Python Programming Tutorial #13 - How to Read a Text File
Tech With Tim
Python Programming Tutorial #14 - Writing to a Text File
Tech With Tim
Python Programming Tutorial #15 - Using .count() and .find()
Tech With Tim
Python Programming Tutorial #16 - Introduction to Modular Programming
Tech With Tim
Python Programming Tutorial #17 - Optional Parameters
Tech With Tim
Python Programming Tutorial #18 - Try and Except (Python Error Handling)
Tech With Tim
Python Programming Tutorial #19 - Global vs Local Variables
Tech With Tim
Python Programming Tutorial #20 - Classes and Objects
Tech With Tim
Cool VBS Script to Prank Your Friends!
Tech With Tim
How to Overclock an AMD GPU
Tech With Tim
Best GPU'S For Mining Ethereum (2018)
Tech With Tim
Recursion and Memoization Tutorial Python
Tech With Tim
Ethereum Mining Rig - Hardware Guide
Tech With Tim
Pygame Tutorial #1 - Basic Movement and Key Presses
Tech With Tim
How to Install Pygame (Windows 8/10)
Tech With Tim
How to Trade Your Cryptocurrency (Bitcoin, Ethereum etc.) For Cash!
Tech With Tim
How to Mine Ethereum 2018 - WORKING (Super-Easy)
Tech With Tim
Microphone Comparison - $10 Mic vs $150 Mic (Blue Yeti USB)
Tech With Tim
Pygame Tutorial #2 - Jumping and Boundaries
Tech With Tim
Pygame Tutorial #3 - Character Animation & Sprites
Tech With Tim
Pygame Tutorial #4 - Optimization & OOP
Tech With Tim
OBS Studio Tutorial - Best OBS Settings
Tech With Tim
Linear Search Algorithm - Python Example and Code
Tech With Tim
Make Any Mic Sound AMAZING! (WITH OBS)
Tech With Tim
Binary Search Algorithm - Python Example & Code
Tech With Tim
Pygame Tutorial #5 - Projectiles
Tech With Tim
Pygame Game - Mini Golf
Tech With Tim
Pygame Tutorial - Projectile Motion (Part 1)
Tech With Tim
Pygame Tutorial - Projectile Motion (Part 2)
Tech With Tim
Pygame Tutorial #6 - Enemies
Tech With Tim
Pygame Tutorial #7 - Collision and Hit Boxes
Tech With Tim
Pygame Tutorial #8 - Scoring and Health Bars
Tech With Tim
Cloud Mining vs. Hardware Mining - 2018
Tech With Tim
How to Install Pygame on Mac OSX (Fast-Simple)
Tech With Tim
Pygame Tutorial #9 - Sound Effects, Music & More Collision
Tech With Tim
Pygame Tutorial #10 - Finishing Touches & Next Steps
Tech With Tim
How to Fade Your Screen in Pygame [CODE IN DESCRIPTION]
Tech With Tim
How to Create a Button in Pygame [CODE IN DESCRIPTION]
Tech With Tim
Pygame Side-Scroller Tutorial #1 - Scrolling Background/Character Movement
Tech With Tim
Pygame Side-Scroller Tutorial #2 - Random Object Generation
Tech With Tim
Pygame Side-Scroller Tutorial #3 - Collision
Tech With Tim
Pygame Side-Scroller Tutorial #4 - Scoring and End Screen
Tech With Tim
How to Create A Message Box in Python - Tkinter
Tech With Tim
Is Ethereum Mining Still Profitable - Is It Worth It (April 2018)
Tech With Tim
How to Run MAC OSX on a WINDOWS PC (Clover Boot-loader)
Tech With Tim
Programming Problem #1 - Alphabet Soup (Beginner/Novice)
Tech With Tim
More on: Neural Network Basics
View skill →Related Reads
📰
📰
📰
📰
Competence-Gated Pooling of Language Models and Priors for Event Forecasting
ArXiv cs.AI
Mined from Scientific Literature: Process Schemas for Atomic Layer Deposition and Etching in Materials Science
ArXiv cs.AI
Robust Prototypical Networks for Few-Shot Sensor Fault Diagnosis
ArXiv cs.AI
Do Influence-Derived Data Perturbations Enable Machine Unlearning? A Controlled Study of Three Plausible Roles
ArXiv cs.AI
🎓
Tutor Explanation
DeepCamp AI