Friends, here are some helpful links: • 🤓Sign up for the full course (40+ hours of TensorFlow) - dbourke.link/ZTMTFcourse • 💻Get all of the code/materials on GitHub - www.github.com/mrdbourke/tensorflow-deep-learning/ • 📖Read the course materials in beautiful book form - www.learntensorflow.io • ❓Ask a question - github.com/mrdbourke/tensorflow-deep-learning/discussions • 💬Get live captions (if using Google Chrome) - support.google.com/chrome/answer/10538231?hl=en Happy Machine Learning!
Hey Daniel, is it good idea to watch this tutorial on tensorFlow before continuing Deep learning portion on ZTM - Data Science and Machine learning course 2022? I think this youtube tutorial does more indepth. So it may help increase the speed of learning in Udemy course
Hahaha, same, it's good because instead of providing definition, the author goes to Google and tries to explain to us what it means - it's like watching a movie, you don't know what google will display and how the author will explain that to us - dopamine driven course 🎉
NOTE: If you are getting a TypeError at 4:39:49 while fitting the model you will need to add tf.expand_dims() I don't know why but it worked for me. example: model.fit(tf.expand_dims(x, axis=-1), y, epochs=5) Edit: I did some studying the reason this error was there because the training data and evaluation data are not of the same dimensions and/or number similar to when tried to dot product the matrices.
@@ahmadbodayr7203 I did some studying the reason this error was there because the training data and evaluation data are not of the same dimensions and/or number similar to when tried to dot product the matrices.
No has ever taught ML/DL better....seriously, this is my 4th attempt at ML and I think I might just get it this time! The teaching structure is absolutely beautiful. I am at 5:20:45 and I WILL comment when I get to the end.
I am Computer Science student, this semester we were properly introduced to ML as a subject. I was searching for a tf courseware on NN & the moment I saw this video, I was hooked ! It's been 7 hrs now (used my entire weekend), coded along, learnt so much along the way - right from the basic algebra (vectors, scalars, tensors, matrices...) all the way through ML and NN. I'll complete this as well as part 2 before another weekend ! LEARNING HAS BEEN SO MUCH FUN & INTERESTING BECAUSE OF YOU, SIR 'D. BOURKE'. Thanks a lot for such awesome content, your hard work shines throughout the course curriculum. I recommend everyone to justify his hard work by completing the entire tutorial along with Programming.
I just had a new project assigned to me - research on DL algorithms for our clients. I have done dozens of courses from EdX and Datacamp to Udemy. Hands-down, this is the most phenomenal course I've seen. This course is even better than the 2 courses I've loved the most since I started learning - Python for everybody by Dr Chuck and Excel skills for Business Specialization on coursera by macquarie. Bravo Daniel! I hope you keep producing endlessly brilliant content like this all your life.
Hello Daniel, first let me thank you for your work and spread of your knowledge, helping other enter this new ML era... I had some problems on the first model training at the 4h43m of your video. I'm Now on the TF version 2.15.0 and i had to change the model.fit(X,y,epochs=5), changing the X Expanding his dim tf.expand_dims(X,axis=1) it only works when i use like this: model.fit(tf.expand_dims(X,axis=1),y,epochs=5) ###EDITED### 13/01/2024 as i walk on the trainning video i get more insights of how the model works and it came up to me that, if you want to use the one dimension training data X, you have to be explitcit passing the kwarg when you're adding the first input layer "input_shape=[1]": Example: model=tf.keras.Sequential( [ tf.keras.layers.Dense(100, activation=None, input_shape=[1]), tf.keras.layers.Dense(1)] ) model.compile(loss=tf.keras.losses.mae, optimizer=tf.keras.optimizers.SGD(), metrics=["mae"]) model.fit(X,y,epochs=10)
hi I am 13 years old boy I but I understand all the concepts you are teaching first I learn python and I decided I want to be a ai programmer and I had coded some "opencv" for computer vision and "pyttsx3" for speech and many more and now this is one of my biggest cource and I like your teaching methods
I was searching ML and NN and found this channel and its continuous 5 hours and I learned a lot of things without getting bored. Its best on RUclips SIR DANIEL . Thank you so much for providing such a worthy of course free of cost.
Seven months later I find your video from RUclips recommendations so helpful and easier to understand I’m a self thought programmer so I can’t understand what others RUclips video’s saying and confused and drop the topic and didn’t look at them, but your video is so understand and very details, One sub from me:)
Daniel, if your were around when I was a student at my university and I happened to have you as a teacher.... well.... This is an amazing set of lectures! YOU my friend are well on your way to creating the next generation of world class scientists!! Keep up the great work!!!
Im 5:47:41 hours deep, I have been wanting to say this from early on but I wanted to make sure I am not too fast to talk. You are an excellent teacher, regardless of the topic you are teaching in this video, the way you teach is exactly the way it should be done. You arouse our critical thinking and you get us to make questions. You might not be here in this room with me but i feel more energy from you than from my university teachers in person💯💯💯
1:33:00 To see the difference between normal, and uniform distribution we can visualize it on plot: import tensorflow as tf import matplotlib.pyplot as plt random_generator_set = tf.random.Generator.from_seed(43) random_tensor1 = random_generator_set.uniform(shape=(2,20000)) random_tensor_color = random_generator_set.normal(shape=(1,20000)) random_tensor_size = random_generator_set.normal(shape=(1,20000)) plt.scatter(random_tensor1[0],random_tensor1[1],s=random_tensor_size*2000,c=(random_tensor_color),alpha=0.01,cmap='Blues') random_generator_set = tf.random.Generator.from_seed(43) random_tensor2 = random_generator_set.normal(shape=(2,20000)) random_tensor_color = random_generator_set.normal(shape=(1,20000)) random_tensor_size = random_generator_set.normal(shape=(1,20000)) plt.scatter(random_tensor2[0],random_tensor2[1],s=random_tensor_size*200,c=(random_tensor_color*1000),alpha=0.42,cmap='Blues')
Daniel, thank you so much for your course! After completing this tutorial, I easily got a job offer because Im able to describe all deep learning concepts in details!
For those who had a ValueError right from the getgo There is one extra line that he didn't have to add with his version of TensorFlow model = tf.keras.Sequential([ tf.keras.layers.Input(shape=(1,)), ###This extra line tf.keras.layers.Dense(1) ]) model.compile(loss=tf.keras.losses.mae, optimizer=tf.keras.optimizers.SGD(), metrics=["mae"]) model.fit(X, Y, epochs=5)
This is the third course that i've watched so far and I have to say this is the best of all, seriously if you are hesitating with another one I swear this course is lit
I am paying in thousands to my University and learning nothing but theory. Thanks to Daniel who helped me learn all this stuff for free. Thank you a lot for that.
The fact that you have ZERO dislikes tells you that this is PURE GOLD. I think that these videos and the whole course should be recommended by the Tensorflow official team for learning TensorFlow. Keep up the good work!
6:44:00 gave some BORAT vibes and i actually smiled. thanks so much for the content. to say your style of teaching is amazing would be an understatement.
Man ...Even a small kid can understand these videos...CRYSTAL CLEAR Explanation and He was upto the point!Awesome video and I never found a video like this one and I doubt if i would find something as good as this in the future
Thank you , man you are the i am 13 and interested in machine learning but whatever course i find is so complicated and explains with maths that i can not understand but you unlike them all you go throughout the thing without alot of complicated maths and explain it pretty well thanks man you are the best
I would just go through your tutorial entirely for your beautiful Australian accent... not to mention that the technical content is nearly unparalleled as well. Thank you for the excellent effort... it's beautiful, flawless, complete and highly recommended!
"Thank You" will be such a understatame for this gem. You have just made a brilliant tutorial in TensorFlow which will not only help the beginners to get themselves acquiented but also, enable them to know how to actually work out on a problem by searching and interpretting the results and again tweaking the model. Thank you - million times. Love from India.
After going through so many complicated and advanced tutorials (PS :- Didn't understand all the stuff there, this video literally made every concept clear). HATS OFF to such great creators like you sir !. Thank you so very much. Please make more advanced long tutorials so that we can become experts as well. Thank you !
Amazing tutorial Dan! Step by step and great explanations. The fact that you show how exactly you are searching something when you do not know how to use a specific function helps the audience to become "self-taught". Thank you so much!
Dude to be honest with you you are the best one who talked about this subject like you're literally the best when you speak you just hit the spot in my brain I really wish you the best in life bro you really inspired me a lot luv ya from Iraq❤️
Hats off man, great tutorial and great way of teaching, My search for tensorflow turotial ends here until I finish up with 10+4 hours of your tutorials. Thanks a ton and great service to machine learning enthusiasts.
8:35:38 He mistaked the results of model 2 with the results of model 3 assuming the normalized model did better. However after normalization the model didnt improve with any significance as both the normalized model and the model3 results are on par. Still normalization and standardization can help with NNs and other ML algo's And once again this tutorial is pure gold
Just a small observation I am using tensorflow 2.8. The model in the linear regression problem using neural network is giving problem when passed with one Dense layer and calling the fit method of the model. It is expected that we have to pass the input shape as well and then it is working fine. tf.random.set_seed(42) model = tf.keras.Sequential([ tf.keras.Input(shape=(1,)), tf.keras.layers.Dense(1) ]) model.compile(loss=tf.keras.losses.mae, optimizer = tf.keras.optimizers.SGD(), metrics=["mae"])
@@mrdbourke I started from here and I get hooked up to this course that I ended up doing the full one however I would like to ask one thing are you going to make something similar for pytorch and make it available on Udemy and some free hours on youtube like the way you did it for tensorflow.That will be great help for so many people who like to learn both.
Hey Daniel!! This is the first video, literally first on deep learning that didn't make me sleep whatsoever... I was hooked to the screen and was able to grasp everything so easily... Kudos to you for your hard work and such a beautiful explanation of concepts... Keep making such awesome videos and we promise we'll leverage this knowledge for the best of data science... All the very best and we love you 💯
It's like finding exactly what you're looking for and finding GOLD content for that matter. I feel lucky to be alive in an era where knowledge such as this one is given for free. Thanks!
I just love the articulation and the analogies with which you teach, Daniel. I don't think I've ever seen anyone teach anything with such an insight. You amaze me and I just love your enthusiasm. I'm looking forward the course in Udemy as well! :D
Your teaching method is superb! Typing along with you, in additional to how thoroughly and clearly you explain everything, is a crucial supplemental learning dimension. I'm amazed here not just about how much I know but how much I understand. Thanks Daniel!
I learnt TensorFlow from this video a week back but forgot to comment. This stuff is too good. Shared it in my community. Every course developer should learn from you. Thanks a lot!
Awesome course to get up and running with TensorFlow specially if one has a good theoretical backbone of the topics explained. Perks to Daniel Bourke, Just fantastic!
This is amazing, first video I’ve seen of this kind that the person uses simple language whilst telling you the complicated form so you don’t get bewildered when doing further research; bravo sir!
2:01 It will be better if you fill the tensors with numbers (example 1:120) instead of zeros by this way we could visualize the results of operation (indexing, slicing, reshaping, operations ...)
we can use tf.math.reduce_variance and tf.math.reduce_std, without .math it will throw error, and yeah these methods would not work with int, so cast them with dtype float. Btw, this is one of a kind video and channel out there, where you feel hooked, and want to learn more and more without taking rest. Knowledge is free, and to educate people with this much dedication is rare. Keep the work up :)
Excellent video so far, just wanna point out a mistake made at 2:36:33 , the resulting matrix of a matrix multiplication has to be of shape equal to outer dimensions not the inner ones.
What would be the pre requirements or knowledge we should know before starting this video, I’ve just started learning data science around a month ago, would I be able to learn this or should I watch something before this.
If you’ve got experience with Python code + a little ML you should be good to go! I’d even try it out and see how far you get before you get stuck. If you need to go learn something and come back, these videos will still be here Otherwise here’s the list of formal prerequisites: github.com/mrdbourke/tensorflow-deep-learning#prerequisites
Dude, you are a legend, I will be graduating in 7 days from my master and will get my hands on this right after, thanks for all those amazing sharing !!!!!!!!!!
Great work! I like the teaching method and have couple of questions/observations. 1. In your NN diagram there is a connection between Hidden to Hidden - not sure if you kept it to illustrate skip connections/dropouts and Layers; 2. In the type of learning, adding Reinforcement Learning would be nice
Man, this stuff is way better then Andrew Ng course on Coursera! Respect!!! For all the time you Invested to bring this up to us- Great Thank you!!! (Liked, Subscribed! Following)
Hi Daniel, no idea if you are going to read this but please make more videos and courses, I am building a machine learning finance startup and your videos are so insightful and we watch them every day!!!!!!!!
OMG!!! holy crap!!! finished 8 hours in just 2 days... I cant believe that I came that far...Stoping now because headphone bettry is over.... Thank you so much for this video.... waiting for more!!!
Every bit of this video is awesome but best part about the video were the small pauses after question + google that gave me a chance to understand/analyse what code is doing and then checking what you say to know if I was wrong or right. This helps me to remember concepts/ideas for longer periods. Thank you Daniel for this amazing tutorial
Friends, here are some helpful links:
• 🤓Sign up for the full course (40+ hours of TensorFlow) - dbourke.link/ZTMTFcourse
• 💻Get all of the code/materials on GitHub - www.github.com/mrdbourke/tensorflow-deep-learning/
• 📖Read the course materials in beautiful book form - www.learntensorflow.io
• ❓Ask a question - github.com/mrdbourke/tensorflow-deep-learning/discussions
• 💬Get live captions (if using Google Chrome) - support.google.com/chrome/answer/10538231?hl=en
Happy Machine Learning!
You are awesome dude!!:) Love from India
Excellent sir...., it would be very kind of you to activate the subtitles of your wonderful course
I’m gonna buy the course, thanks
Can i please get the link to the discord chat?
Hey Daniel, is it good idea to watch this tutorial on tensorFlow before continuing Deep learning portion on ZTM - Data Science and Machine learning course 2022? I think this youtube tutorial does more indepth. So it may help increase the speed of learning in Udemy course
Get ready to dream in tensors.
Thank youu soo much Daniel!!!!!!!
I can only imagine how many hours you've put into it. You're a legend Dan!
The mind is a matrix
@@CodeEmporium the mind is a tensor :)
Great job
this style of teaching is so effective for people like me who have ADHD, thank you so much you absolute ledgend
It's how I like to learn too! Thank you for the kind words Rohan, enjoy!
I agree 120%. Am baffled. The voice and quality are just legendary
Rohan bhai.....ADHD....?
you can follow a slow 10h course with ADHD ? ^^
Hahaha, same, it's good because instead of providing definition, the author goes to Google and tries to explain to us what it means - it's like watching a movie, you don't know what google will display and how the author will explain that to us - dopamine driven course 🎉
I cant thank you enough for all the efforts you do to teach ML. You can’t find quality content like this anywhere else, for free.
You're welcome Faris! Enjoy legend!
Yes He is a legend
NOTE: If you are getting a TypeError at 4:39:49 while fitting the model you will need to add tf.expand_dims()
I don't know why but it worked for me.
example: model.fit(tf.expand_dims(x, axis=-1), y, epochs=5)
Edit: I did some studying the reason this error was there because the training data and evaluation data are not of the same dimensions and/or number similar to when tried to dot product the matrices.
Thank you man god bless you
@@ahmadbodayr7203 I did some studying the reason this error was there because the training data and evaluation data are not of the same dimensions and/or number similar to when tried to dot product the matrices.
bcz dense expecting 2d tensor , maybe they updated that
thanks a lot , i really spent a lot of time trying to figure out what was the problem . thanks for sharing
Thanks man
No has ever taught ML/DL better....seriously, this is my 4th attempt at ML and I think I might just get it this time! The teaching structure is absolutely beautiful. I am at 5:20:45 and I WILL comment when I get to the end.
Thank you! Glad you’re enjoying! Keep learning legend! Can’t wait to hear when you’re done
@@mrdbourke Wooho! Done with video 1! Heading on to video 2!
@@b.k.7363 outstanding effort! Keep going!
I am Computer Science student, this semester we were properly introduced to ML as a subject. I was searching for a tf courseware on NN & the moment I saw this video, I was hooked ! It's been 7 hrs now (used my entire weekend), coded along, learnt so much along the way - right from the basic algebra (vectors, scalars, tensors, matrices...) all the way through ML and NN. I'll complete this as well as part 2 before another weekend ! LEARNING HAS BEEN SO MUCH FUN & INTERESTING BECAUSE OF YOU, SIR 'D. BOURKE'. Thanks a lot for such awesome content, your hard work shines throughout the course curriculum. I recommend everyone to justify his hard work by completing the entire tutorial along with Programming.
I just had a new project assigned to me - research on DL algorithms for our clients. I have done dozens of courses from EdX and Datacamp to Udemy. Hands-down, this is the most phenomenal course I've seen. This course is even better than the 2 courses I've loved the most since I started learning - Python for everybody by Dr Chuck and Excel skills for Business Specialization on coursera by macquarie. Bravo Daniel! I hope you keep producing endlessly brilliant content like this all your life.
This video is extremely underrated my man. Amazing job! Please keep on making more videos. You are a terrific teacher
Amazing
Hello Daniel, first let me thank you for your work and spread of your knowledge, helping other enter this new ML era...
I had some problems on the first model training at the 4h43m of your video. I'm Now on the TF version 2.15.0 and i had to change the model.fit(X,y,epochs=5),
changing the X Expanding his dim tf.expand_dims(X,axis=1)
it only works when i use like this:
model.fit(tf.expand_dims(X,axis=1),y,epochs=5)
###EDITED### 13/01/2024
as i walk on the trainning video i get more insights of how the model works and it came up to me that, if you want to use the one dimension training data X, you have to be explitcit passing the kwarg when you're adding the first input layer "input_shape=[1]":
Example:
model=tf.keras.Sequential(
[ tf.keras.layers.Dense(100, activation=None, input_shape=[1]),
tf.keras.layers.Dense(1)]
)
model.compile(loss=tf.keras.losses.mae,
optimizer=tf.keras.optimizers.SGD(),
metrics=["mae"])
model.fit(X,y,epochs=10)
Thank you! That fix worked for me as well, on the model at the end of the video.
hi I am 13 years old boy I but I understand all the concepts you are teaching first I learn python and I decided I want to be a ai programmer and I had coded some "opencv" for computer vision and "pyttsx3" for speech and many more and now this is one of my biggest cource and I like your teaching methods
I was searching ML and NN and found this channel and its continuous 5 hours and I learned a lot of things without getting bored. Its best on RUclips SIR DANIEL . Thank you so much for providing such a worthy of course free of cost.
I am aged 50+ and learning tensorflow now as a Process Excellence and LSS Expert! Hope this will help!
Thanks a lot man, its 2023 and this video is the best you could get anywhere
Thank you Tevin! Enjoy legend!
Seven months later I find your video from RUclips recommendations so helpful and easier to understand I’m a self thought programmer so I can’t understand what others RUclips video’s saying and confused and drop the topic and didn’t look at them, but your video is so understand and very details,
One sub from me:)
Friends I have been following this video for last 5 days to complete the code with practice simultaneously
Keep going guys! You’ve got this!
Is it all I need to get ready for Tensorflow certification examen ???
Daniel, if your were around when I was a student at my university and I happened to have you as a teacher.... well.... This is an amazing set of lectures! YOU my friend are well on your way to creating the next generation of world class scientists!! Keep up the great work!!!
Thank you Rajat! I really appreciate it!
Im 5:47:41 hours deep, I have been wanting to say this from early on but I wanted to make sure I am not too fast to talk.
You are an excellent teacher, regardless of the topic you are teaching in this video, the way you teach is exactly the way it should be done. You arouse our critical thinking and you get us to make questions.
You might not be here in this room with me but i feel more energy from you than from my university teachers in person💯💯💯
1:33:00
To see the difference between normal, and uniform distribution we can visualize it on plot:
import tensorflow as tf
import matplotlib.pyplot as plt
random_generator_set = tf.random.Generator.from_seed(43)
random_tensor1 = random_generator_set.uniform(shape=(2,20000))
random_tensor_color = random_generator_set.normal(shape=(1,20000))
random_tensor_size = random_generator_set.normal(shape=(1,20000))
plt.scatter(random_tensor1[0],random_tensor1[1],s=random_tensor_size*2000,c=(random_tensor_color),alpha=0.01,cmap='Blues')
random_generator_set = tf.random.Generator.from_seed(43)
random_tensor2 = random_generator_set.normal(shape=(2,20000))
random_tensor_color = random_generator_set.normal(shape=(1,20000))
random_tensor_size = random_generator_set.normal(shape=(1,20000))
plt.scatter(random_tensor2[0],random_tensor2[1],s=random_tensor_size*200,c=(random_tensor_color*1000),alpha=0.42,cmap='Blues')
*A playlist with 60 videos of 10 min each*
Me: meeeh, not today...
*A single video with a 10 hours marathon*
Also me: LET'S GO FOR IT!!
If I could have made this one single video, I would’ve
@@mrdbourke did I hear 15 hours of pytorch?? 😅😅 But seriously, congratulations for the content bro
Holding Point.
4:45:57 / 10:15:27
Thanks for the Content!
I appreciate your Favor.
Daniel, thank you so much for your course! After completing this tutorial, I easily got a job offer because Im able to describe all deep learning concepts in details!
For real ?
@@akshayraoch 😂😂😂😂
For those who had a ValueError right from the getgo
There is one extra line that he didn't have to add with his version of TensorFlow
model = tf.keras.Sequential([
tf.keras.layers.Input(shape=(1,)), ###This extra line
tf.keras.layers.Dense(1)
])
model.compile(loss=tf.keras.losses.mae,
optimizer=tf.keras.optimizers.SGD(),
metrics=["mae"])
model.fit(X, Y, epochs=5)
Thank you! I was stuck on this for a while. I figured out that we might need to define the input shape, but wasn't sure how.
Even better then paid courses ! Thank you Daniel for such amazing stuff.
This is the third course that i've watched so far and I have to say this is the best of all, seriously if you are hesitating with another one I swear this course is lit
Thank you so much Laurent! I appreciate it!
this man was sent from god
Yessss, for real!!!
You are absolutely correct on all grounds , in fact the man is a small god himself
Is he Lucifer ?
I am paying in thousands to my University and learning nothing but theory. Thanks to Daniel who helped me learn all this stuff for free. Thank you a lot for that.
The fact that you have ZERO dislikes tells you that this is PURE GOLD. I think that these videos and the whole course should be recommended by the Tensorflow official team for learning TensorFlow.
Keep up the good work!
Thank you so much Vasko! That’s a very big compliment! Plenty more to come :)
lol, you jinx it
@@demonsawakening Makes me wonder what there was to dislike. Wish RUclips would make you give a reason for the dislike.
Your teaching style is freaking good. I mean, you don't even forget to write comments so everyone can understand what you do.
Thank you thank you! I'm not a fan of tutorials that don't explain things enough
6:44:00 gave some BORAT vibes and i actually smiled. thanks so much for the content. to say your style of teaching is amazing would be an understatement.
Thank you Azeez! I really appreciate it :)
Man ...Even a small kid can understand these videos...CRYSTAL CLEAR Explanation and He was upto the point!Awesome video and I never found a video like this one and I doubt if i would find something as good as this in the future
Thank you , man you are the i am 13 and interested in machine learning but whatever course i find is so complicated and explains with maths that i can not understand but you unlike them all you go throughout the thing without alot of complicated maths and explain it pretty well thanks man you are the best
You’ve got this Majduddin, code along and research more when you need
@@mrdbourke thanks man you are a role model for me now
I would just go through your tutorial entirely for your beautiful Australian accent... not to mention that the technical content is nearly unparalleled as well. Thank you for the excellent effort... it's beautiful, flawless, complete and highly recommended!
"Thank You" will be such a understatame for this gem. You have just made a brilliant tutorial in TensorFlow which will not only help the beginners to get themselves acquiented but also, enable them to know how to actually work out on a problem by searching and interpretting the results and again tweaking the model.
Thank you - million times.
Love from India.
You’re welcome Saptarshi! I really appreciate it legend
Learning at this fully explained pace really makes the concepts stick, great job man.
After going through so many complicated and advanced tutorials (PS :- Didn't understand all the stuff there, this video literally made every concept clear). HATS OFF to such great creators like you sir !. Thank you so very much. Please make more advanced long tutorials so that we can become experts as well. Thank you !
I urge UN to honor Daniel with the best teacher award.. much much better than university teaching
Amazing tutorial Dan! Step by step and great explanations. The fact that you show how exactly you are searching something when you do not know how to use a specific function helps the audience to become "self-taught". Thank you so much!
Dude to be honest with you you are the best one who talked about this subject like you're literally the best when you speak you just hit the spot in my brain I really wish you the best in life bro you really inspired me a lot luv ya from Iraq❤️
Thank you Amir! I appreciate it legend
Hats off man, great tutorial and great way of teaching, My search for tensorflow turotial ends here until I finish up with 10+4 hours of your tutorials. Thanks a ton and great service to machine learning enthusiasts.
8:35:38 He mistaked the results of model 2 with the results of model 3 assuming the normalized model did better. However after normalization the model didnt improve with any significance as both the normalized model and the model3 results are on par.
Still normalization and standardization can help with NNs and other ML algo's
And once again this tutorial is pure gold
Wow pure dedication to teach deep learning! Love your videos.
Thank you Md!
Just a small observation I am using tensorflow 2.8.
The model in the linear regression problem using neural network is giving problem when passed with one Dense layer and calling the fit method of the model.
It is expected that we have to pass the input shape as well and then it is working fine.
tf.random.set_seed(42)
model = tf.keras.Sequential([
tf.keras.Input(shape=(1,)), tf.keras.layers.Dense(1)
])
model.compile(loss=tf.keras.losses.mae,
optimizer = tf.keras.optimizers.SGD(),
metrics=["mae"])
Thank you for the heads up!
Just now working on the video! This helped me! Thank you!!
@@mrdbourke I started from here and I get hooked up to this course that I ended up doing the full one however I would like to ask one thing are you going to make something similar for pytorch and make it available on Udemy and some free hours on youtube like the way you did it for tensorflow.That will be great help for so many people who like to learn both.
Hey Daniel!! This is the first video, literally first on deep learning that didn't make me sleep whatsoever... I was hooked to the screen and was able to grasp everything so easily... Kudos to you for your hard work and such a beautiful explanation of concepts... Keep making such awesome videos and we promise we'll leverage this knowledge for the best of data science... All the very best and we love you 💯
So happy to hear you enjoyed! I really appreciate it! Plenty more to come :)
@@mrdbourke sir please make such a long hourse tutorial for #machine_learning_beginers
In 2022 this is the best content on Ml in the internet. Thanks a million
Love the way you keep me engaged throughout the video by sharing our enthusiasm. Amazing Job! You have my subscription
Thank you brother, the thing I like the most about you is, you don't get tired of writing each and every code repeatedly. Respect!!
Thank you Girmay!
Dan! this is awesome! Thank you for putting in this much effort! You're a really gifted teacher.
Thank you so much!
It's like finding exactly what you're looking for and finding GOLD content for that matter. I feel lucky to be alive in an era where knowledge such as this one is given for free. Thanks!
I just love the articulation and the analogies with which you teach, Daniel. I don't think I've ever seen anyone teach anything with such an insight. You amaze me and I just love your enthusiasm.
I'm looking forward the course in Udemy as well! :D
Thank you thank you thank you Nikhil! Those are some very big compliments. You've put the biggest smile on my face! Udemy version coming soon
@@mrdbourke You've totally nailed it! :D
@@mrdbourke whens the udemy course coming
@Nikhil Vattipalli is the course on Udemy and ZTM is same?
Your teaching style is just out of this world... wish I had a teacher like you.
Thank you!
Your teaching method is superb! Typing along with you, in additional to how thoroughly and clearly you explain everything, is a crucial supplemental learning dimension. I'm amazed here not just about how much I know but how much I understand. Thanks Daniel!
You’re welcome legend! Thank you
day 1 =1 :34:53
day2=3:51:21 worked hard
day3= 4:28:41 got sick today :{
day4=5:56:02
day5=8:10:35 its all coming together
day6=completed
Deep Learning is basically that one kid in middle school who "don't follow the rules, but make the rules."
Best Video for Learning TensorFlow and easy to understand. Daniel, you are teaching is very cool, you are the best tutor.
Absolute Biggy !!
Great Initiative Daniel to post it on the channel.
Going to share it with everyone out there and recommending the course too
This is a wonderful tutorial. Thank you for your great work.
Thank you Vivek! I appreciate it!
I learnt TensorFlow from this video a week back but forgot to comment. This stuff is too good. Shared it in my community. Every course developer should learn from you. Thanks a lot!
yes but tooo long
AWhh!! Love your motto "If in doubt ; run the CODE!"
Awesome course to get up and running with TensorFlow specially if one has a good theoretical backbone of the topics explained.
Perks to Daniel Bourke, Just fantastic!
For me, this is the most amazing lesson about TensorFlow on RUclips ever .. Many thanks, Daniel..
this man amazes me every time. thank you so much. i hope this course also release on udemy. in this way i can buy this course and learn new things
Thank you! Treat these videos (part 2 coming soon) as the warmup for the Udemy version
loved this course so much!!
I actually bought it on udemy LOVED IT
thanks Daniel and any one who helped in the making of this course !!!!
This is amazing course 10hours+, My weekend is complete with amazing course. no tv, switch off my mobile, plenty of Redbull. woow. THANK YOU.
Love the No TV, mobile off combination! Enjoy!
This is amazing, waiting for part -2, your journey is inspiring man! great going looking forward to some great content
I couldn't find part 2, where is it?
This is amazing, first video I’ve seen of this kind that the person uses simple language whilst telling you the complicated form so you don’t get bewildered when doing further research; bravo sir!
Yo, this is sick, Daniel. Thank you so much! Would love a similar one for PyTorch, if you are thinking about doing one :)
Best video I watched on whole internet if you want to start learning Tensorflow or deep learning
YOOOO this is some serious stuff!
1:02:24
But you never do stuff like this 😅
Ken Jee in the house! Big dog’s gotta eat!!
True
2:01 It will be better if you fill the tensors with numbers (example 1:120) instead of zeros by this way we could visualize the results of operation (indexing, slicing, reshaping, operations ...)
I cant believe this is free, thank you
Enjoy Joel!
we can use tf.math.reduce_variance and tf.math.reduce_std, without .math it will throw error, and yeah these methods would not work with int, so cast them with dtype float.
Btw, this is one of a kind video and channel out there, where you feel hooked, and want to learn more and more without taking rest. Knowledge is free, and to educate people with this much dedication is rare. Keep the work up :)
this is great. I was waiting for the course to be released on Udemy. now I can already start and just continue once it is released :D
These videos (part 2 coming soon) will get you warmed up for the full version!
Hey Daniel,
Will you release this course on udemy?
Love the way you say beautiful after each execution of code ,it inspires to have passion in coding.
Working code is a beautiful thing
Excellent video so far, just wanna point out a mistake made at 2:36:33 , the resulting matrix of a matrix multiplication has to be of shape equal to outer dimensions not the inner ones.
I don't know English enough, but for this video, I ready to learn enough:)))
You got this!
@@mrdbourke thanks:)
I'm loving the intro, it feels like we are figuring out everything together. makes it much more approachable for me, a beginner. really appreciate it.
What would be the pre requirements or knowledge we should know before starting this video, I’ve just started learning data science around a month ago, would I be able to learn this or should I watch something before this.
If you’ve got experience with Python code + a little ML you should be good to go! I’d even try it out and see how far you get before you get stuck. If you need to go learn something and come back, these videos will still be here
Otherwise here’s the list of formal prerequisites: github.com/mrdbourke/tensorflow-deep-learning#prerequisites
Dude, you are a legend, I will be graduating in 7 days from my master and will get my hands on this right after, thanks for all those amazing sharing !!!!!!!!!!
Woah!! Perfect time! Massive effort on the graduation my friend! These videos will be waiting for you
well done
In Hinduism we believe teachers are not just a human being but God.....you are my God as far as these topics are concerned 🇮🇳🙏👍🏻😂
1:20:12 - uhhhhh...!!! That s the thing that makes teaching awesome: emotional presentation. And great exploration of the elements of study
This is exactly what I was looking for, a good concise tutorial. Thanks!
also, first comment heh.
Bought your course from Udemy, its huge, but its the best course on earth.
Is it just me or there are no like/save buttons right now?
same. I wanted to save it on my playlist on youtube but there is no such option :
Not sure what’s happening here! Could be loading in the background
I am definitely buying the full course after watching this video !
Enjoy Aliasgar!
Great work! I like the teaching method and have couple of questions/observations. 1. In your NN diagram there is a connection between Hidden to Hidden - not sure if you kept it to illustrate skip connections/dropouts and Layers; 2. In the type of learning, adding Reinforcement Learning would be nice
A very impressive approch without any complexity. God bless u bro.
Thank you! Glad you’re enjoying :)
Thank you so much for this invaluable content, I'm going to be going through the whole thing!!
You're so welcome! Enjoy my friend! Part 2 coming soon
Man, this stuff is way better then Andrew Ng course on Coursera! Respect!!! For all the time you Invested to bring this up to us- Great Thank you!!! (Liked, Subscribed! Following)
Thank you!
0:12 "I don't want to keep you too long."
Sure you don't.... 🤔
Love your content though.
hahahaha
Hi Daniel, no idea if you are going to read this but please make more videos and courses, I am building a machine learning finance startup and your videos are so insightful and we watch them every day!!!!!!!!
Me while coding after watching this entire course : *Hmm*
witcher vibes. lol
@@SandeshPoudel01 😂
i like the ending part😀. "THE MISSING PART". bro its like a movie, very amazing and understandable
Is it all I need to get ready for Tensorflow certification examen ???
PyTorch > TensorFlow
4:08:12 is my current timestamp, By far the best course on deep learning.
Massive effort on the progress! Thank you for the kind words
@@mrdbourke wow thank you so much ❤️🎉
OMG!!! holy crap!!!
finished 8 hours in just 2 days... I cant believe that I came that far...Stoping now because headphone bettry is over.... Thank you so much for this video.... waiting for more!!!
Boy i think you may be better than some university tutors! I like the way you explain to detail.
Thank you thank you!
Hey Daniel.. done 5 hours of tensorflow.. Not at all getting bored.. It’s an amazing tutorial.
Hey Fousiya, thank you so much! Glad to hear.
Man! Just for that introduction and "Tensorflow" suspense and animation, I'll sign up for the full course!
Every bit of this video is awesome but best part about the video were the small pauses after question + google that gave me a chance to understand/analyse what code is doing and then checking what you say to know if I was wrong or right. This helps me to remember concepts/ideas for longer periods.
Thank you Daniel for this amazing tutorial
Thank you Sumant! Glad you’re enjoying