Building makemore Part 2: MLP
Vložit
- čas přidán 1. 06. 2024
- We implement a multilayer perceptron (MLP) character-level language model. In this video we also introduce many basics of machine learning (e.g. model training, learning rate tuning, hyperparameters, evaluation, train/dev/test splits, under/overfitting, etc.).
Links:
- makemore on github: github.com/karpathy/makemore
- jupyter notebook I built in this video: github.com/karpathy/nn-zero-t...
- collab notebook (new)!!!: colab.research.google.com/dri...
- Bengio et al. 2003 MLP language model paper (pdf): www.jmlr.org/papers/volume3/b...
- my website: karpathy.ai
- my twitter: / karpathy
- (new) Neural Networks: Zero to Hero series Discord channel: / discord , for people who'd like to chat more and go beyond youtube comments
Useful links:
- PyTorch internals ref blog.ezyang.com/2019/05/pytorc...
Exercises:
- E01: Tune the hyperparameters of the training to beat my best validation loss of 2.2
- E02: I was not careful with the intialization of the network in this video. (1) What is the loss you'd get if the predicted probabilities at initialization were perfectly uniform? What loss do we achieve? (2) Can you tune the initialization to get a starting loss that is much more similar to (1)?
- E03: Read the Bengio et al 2003 paper (link above), implement and try any idea from the paper. Did it work?
Chapters:
00:00:00 intro
00:01:48 Bengio et al. 2003 (MLP language model) paper walkthrough
00:09:03 (re-)building our training dataset
00:12:19 implementing the embedding lookup table
00:18:35 implementing the hidden layer + internals of torch.Tensor: storage, views
00:29:15 implementing the output layer
00:29:53 implementing the negative log likelihood loss
00:32:17 summary of the full network
00:32:49 introducing F.cross_entropy and why
00:37:56 implementing the training loop, overfitting one batch
00:41:25 training on the full dataset, minibatches
00:45:40 finding a good initial learning rate
00:53:20 splitting up the dataset into train/val/test splits and why
01:00:49 experiment: larger hidden layer
01:05:27 visualizing the character embeddings
01:07:16 experiment: larger embedding size
01:11:46 summary of our final code, conclusion
01:13:24 sampling from the model
01:14:55 google collab (new!!) notebook advertisement - Věda a technologie
![[柴犬ASMR]曼玉Manyu&小白Bai 毛发护理Spa asmr](http://i.ytimg.com/vi/0TsXQ7z2Dh4/mqdefault.jpg)
0% hype. 100% substance. GOLD!
It’s insanely awesome that you are taking time out of your day to provide the public with educational videos like these.
I'm a senior aerospace engineer, so no CS or ML training at all, and I'm now totally fascinated with PyTorch. First that micrograd intro, which totally clicked the methods used for backprop into place. Now this intro with embedding and data preparation etc. I almost feel like transformers are within reach already. Inspiring!
Quality Engineer with a Production Engineering background. Same feeling !
How you guys landed here, i mean me as a cs graduate, I'll never land directly to a lecture series of aerospace that suddenly start to make sense 🤔
53:20 To break the data to training, developement and test, one can also use torch.tensor_split.
n1 = int(0.8 * X.shape[0])
n2 = int(0.9 * X.shape[0])
Xtr, Xdev, Xts = X.tensor_split((n1, n2), dim=0)
Ytr, Ydev, Yts = Y.tensor_split((n1, n2), dim=0)
This is amazing. Using just a little bit of what I was able to learn from part 3, namely the Kaiming init, and turning back on the learning rate decay, I was able to achieve 2.03 and 2.04 in my test and validation with a 1.89 in my training loss with just 300k iterations and 23k parameters. I set my block size to 4 and my embeddings to 12 and increased my hidden layer to 300 while decaying my learning rate exponent from -1 to -3 linear space over the 300k steps. All that without even using batch normalization yet. After applying batch norm, was able to get these down to 1.99 and 1.98 with training loss in the 1.7s after a little more tweaking. Really good content in this lecture, it really has me feeling like a chef in the kitchen almost, cooking up a model with a few turns of the knobs...This sounds like a game or a problem that can be solved with an AI trained on turning knobs.
intuition: why 4 block size instead of 3 block size? well the english language i think has an average of somewhere between 3 to 5 characters per syllable, which most 1 syllable names falling between that 3-5 character bucket and some 2 syllable names falling in that 4-6 character bucket and beyond. I wanted a block size that would give some better indication on whether we're in a one syllable or two syllable context, and so we could end up with some more pronounceable names. It also just made sense to scale up the dimension of embeddings and neurons to give a little more nuance to the relationships between the different context blocks. English has so many different rules when it comes to vowels and silent letters and so I felt like we needed to give enough room for 3-4 degrees of freedom for each character in the context block, and therefore needed more neurons in the net to account for those extra dimensions. running the model for more steps just allows the convergence to happen. I don't know if it could get much better after more steps but this took 6-7 minutes to run so I think i squeezed all that I could out of these hyperparams.
A two hour workshop on NLP with transformers costs 149$ in Invidia GTC conference.
You tutor us with amazing quality for free. Thank you!🙂
This is amazing!!! I am an undergraduate student researching deep learning. This series is a gold mine. The attention to detail as well as the intuitive explanations are amazing!!
I just want to say thank you Andrej, you are the best !
I've spent the last 2 days going over the first 3 videos (and completing the exercises),
I must say that this is by far the best learning experience I ever had.
The quality of the lectures is just immeasurable, in fact you completely ruined
how I feel about lectures at my University.
where are the exercises?
Check the comments from Andrej in each video @@ahmedivy
@@ahmedivydescription
@@ahmedivyAlso asked myself, then found them in the movie description:
Exercises:
- E01: Tune the hyperparameters of the training to beat my best validation loss of 2.2
- E02: I was not careful with the intialization of the network in this video. (1) What is the loss you'd get if the predicted probabilities at initialization were perfectly uniform? What loss do we achieve? (2) Can you tune the initialization to get a starting loss that is much more similar to (1)?
- E03: Read the Bengio et al 2003 paper (link above), implement and try any idea from the paper. Did it work?
Pls continue this series(and similar ones) to eternity! You are THE teacher which everyone interested/working in AI really needs!
+1(00000000)
Hi Andrej, Even though I am already familiar with all this I still watch your videos for the pedagogical value and for learning how to do things. But, I still learn many new things about pytorch as well as how to think through things. The way you simplify complex stuff is just amazing. Keep doing this. You said on a podcast that you spend 10 hours for 1 hour of content, but you save 1000s of hours of frustration and make implementing ideas a little bit easier.
great and true sentence: "You said on a podcast that you spend 10 hours for 1 hour of content, but you save 1000s of hours of frustration and make implementing ideas a little bit easier."
Much thanks, Andrej! I increased the embedding dimension to 5 from 2, initialized the model parameters from a uniform distribution [0, 1) instead of a standard normal distribution, increased the batch size to 128, and used the sigmoid activation for the hidden layer instead of the hyperbolic tangent, and was able to get the negative log-likelihood for the train and validation sets down to ~2.15, respectively.
Really enjoyed watching these lectures so far :) also +1 for the PyTorch internals video (@25:36)
The amount of times in college we used the PyTorch "view" function with ZERO explanation. And your explanation is not only flawless, you even make the explanation itself look easy! Thank you so much
I believe that at 49:22 the losses and the learning rates are misaligned.
The first loss (derived from completely random weights) is computed before the first learning rate is used, and therefor the first learning rate should be aligned with the second loss.
You can simply solve this problem by using this snippet;
lri = lri[:-1]
lossi = lossi[1:]
Also, thank you so much for these amazing lectures
Can't thank you enough. It's such a satisfying feeling to understand the logic under the ML models clearly. Thank you!
This lecture compress an insanely amount of wisdom in 1.15hrs! Thanks
What a time to be alive, someone as smart as Andrej giving away for free probably the best lectures on the subject. Thanks so much!!!
On top of everything else, this is absolutely the best documentation & explainer of PyTorch. This is infinitely better that the PyTorch documentation. In fact, it should be a must-see video for the PyTorch team to show them how to write good documentation. Meta should pay Adrej any fee he asks for the rights to use this video in the PyTorch docs...Thank you Andrej!
Amazing, astounding… Andrej, you’re continuing your revolution for people’s education in ML. You are the “Che” of AI.
Quite an ominous comparison, especially with Andrej working at OpenAI now.
what does "Che" mean?
@@jeevan288 is a murderer that fight "for" cuba. "Che Guevara"
Thanks. Seeing things coded from scratch clears up any ambiguities one may have when reading the same material in a book.
I was able to achieve a loss of 2.14 on test set
Some hyperparameters:
Neurons in hidden layer: 300
Batch size: 64 for first 400k iterations then 32 for rest
Total Iterations: 600,000
Thank you for uploading such insightful explanations. I really appreciate that you explained how things work under the hood and insights of PyTorch's internals.
This is perhaps the best series on CZcams I have ever seen - Without a doubt I can't recall the last time a 1 hour video was able to teach me so much!
Thank you very much for taking your time to go step by step whether it be torch API, your code or the math behind things. I really appreciate it.
This is absolutely amazing stuff, thank you so much for putting this out for FREE!!!! I thought your name looked familiar and then I remembered you sparked my initial interest in NNs with "the unreasonable effectiveness of RNNs". It was SO fun and fascinating to just toss any old random text at it and see what it did! Can't believe how much progress has happened so quickly. Really really excited to get a better practical understanding of NNs and how to program them.
Thank you again!
Never understood the basics of hyper parameter tuning so well. A sincere Thanks for the foundation and including that part in this video.
I wanted to let you know i listened to the podcast with Lex Fridman and i know understand how much of a Rockstar you are in the Artificial Intelligence space. Like many others i appreciate you and all you qre doing to push forward with this incredible technology. Thank you.
Can't believe the value of these lecture and how helpful they are, you are literally changing the world. Thanks very much for your effort and knowledge
Thank you Andrej! I find your videos not only educational, but also very entertaining. Learning new things is exciting!
Thank you for referencing and exploring the Bengio paper. It's great to get academic context on how models like this were developed, and very few people actually do this in contexts like this.
Whoa, you’re putting these out quicker than I can go through them. Thank you!
Love all the tips and explanations on pytorch, training efficiency, and educational purposed errors. I was writing both code and notes and rewatching and enjoyed it and felt having a fruitful day after finished. It's like I was learning with a kind and insightful mentor sitting next to me. Thanks so much Andrej.
Amazing video and series ! thank you.
Small correction to the build_makemore_mlp.ipynb colab it's assuming the embedding size is 2 but eventually during the lecture it was changed to 10 so the emb.shape will be (32, 3, 10) and h.shape (32, 200), just FYI if you're running it and get confused
Thank you so much, this is gold, I'm watching all of this thoroughly, pausing the video a lot to wrap my head around those tensors manipulation (i didn't know anything abt python/numpy/pytorch). I'm also really inspired from how you quickly plot datas to get important insights, I'll do that too from now on
So cool to see the model learn through the embedding matrix that vowels share some similarity, 'q' and '.' are outlier characters, and so on!
So many insights, I’ll have to rewatch it again to retain them all. Thank you!
29:20 we can also use torch.reshape() to get the right shape for W. However, there is a difference between torch.view and torch.reshape
TL;DR:
If you just want to reshape tensors, use torch.reshape. If you're also concerned about memory usage and want to ensure that the two tensors share the same data, use torch.view.
Please create the "entire video about the internals of pytorch" that you mentioned in 25:40. And thank you so much for the content, Andrej !!
Watching your videos constantly reminds me of "There are no bad students but only bad teachers"
I watched and followed many such courses, yours are really special, easy to understand yet very indepth, with many useful tricks.
Thank you so much for this, Andrej !! I’m really learning and enjoying this
I love the simplicity of your explanations. Thanks a lot!
hi Andrej,
Your lectures are the best ones I saw.
It's amazing you take complex ideas and explain them in such a level that even beginners understand.
Thank you for that.
Such a valuable resource to help people in other fields get up to speed on these concepts. Thank you.
This is priceless, you have such a low and high level understanding of the topic, that's just amazing.
Really great series of lessons! Lots of gems in here for any knowledge level.
PS: Increasing the batch size and lowering the LR a little does result in a small improvement in the loss. Throwing out 2.135 as my test score to beat :)
I feel pretty lucky that my intro to neural networks is these videos. I've wanted to dive in for a while and I'm hooked already. Absolutely loving this lecture series, thank you, I can't wait for more!
I'd love to join the discord but the invite link seems to be broken
I am so happy people like you exist. Thank you very much for this video series.
This educational vid will reach more students than a regular teaching job at a regular school. Thanks for sharing & giving back👍
Beautiful......The explanation!!!! This builds the intuition to venture out in unknown territories. Thanks from the bottom of my heart.
You are amazing! Thank you so much for all your work :) You explain everything very intuitively!!!
I was able to achieve a train loss of 2.15 and test loss of 2.17 with block_size = 4, 100k iterations and embed dimension = 5.
Your lectures make me feel like I am in an AI Retreat :). I come out so happy and enriched after each lecture.
That was really awesome explanation by all means!! thank you very much Andrej for educating us :)
@00:45:40 'Finding a good initial learning rate', each learning rate is used just one time. The adjustment of the parameter of one learning rate is based on the parameters already adjusted using the prior smaller learning rates. I feel that each of the 1,000 learning rate candidates should go through the same number of iterations. Then, the losses at the end of the iterations are compared. Please tell me if I am wrong. Thanks!
each time you make exponentially bigger steps, so you can neglect previous path. It's like if you make one step toward your goal, and than make another 10 steps your overall path is not really affected by you first step. And generally you want to find the biggest number of steps (lr) which you should take in some direction (gradient) to not overshoot your goal (best model weights) to get there faster.
Thanks! The instructor says the test should not be run many times since each time the model learns something from the test data. In the test, the parameters are not adjusted. How can the model learn from the test data?@@wolk1612
Awesome stuff, even though I've been studying neural networks and NLP for the last couple of months, this feels like the first time I *truly* understand how stuff works. Amazing.
PS. At 00:01:34 I was just uber curious about his previous searches, so I google them:
proverbs 27:27
>You will have plenty of goats’ milk to feed your family and to nourish your female servants.
matthew 27:27-31
>Then the governor’s soldiers took Jesus into the Praetorium and gathered the whole company of soldiers around him. They stripped him and put a scarlet robe on him, and then twisted together a crown of thorns and set it on his head. They put a staff in his right hand. Then they knelt in front of him and mocked him. “Hail, king of the Jews!” they said. They spit on him, and took the staff and struck him on the head again and again. After they had mocked him, they took off the robe and put his own clothes on him. Then they led him away to crucify him.
Such an amazing educational video. Learned a lot. Thanks for taking the time and explaining many concepts so clearly.
I really enjoy these videos! A little note is that to run through the tutorial, it requires a bit of memory, so it would be nice with an early discussion of batching :) I run out of memory when calculating the loss, so had to reduce the sample size significantly.
Thank you, Andrej!! Thanks a lot for all the efforts you have put in❤
Awesome tutorial. Thank you Andrej!
What an amazing teacher you are. Thank you
It's one of the great pleasures of CZcams to be taught by someone with Andrejs experience.
Your series is honestly one of the best on CZcams. It's not too short like the typical DL intro videos. And it's not boring because you build the solution from the ground up with real code and common errors included. I love the format and the clear and concise structure.
Thank you for the work that you put into these videos.
Andrey thank you for this great series of lectures. you are a great Educator! 100% GOLD Material to Learn
Andrej is the type of person that could make a video titled "Building a 'hello world' program in Python" which a 10x engineer could watch and learn something from it. The quality of these videos is unreal, please do make a video on the internals of torch!
Great stuff @AndrejKarpathy! I actually was typing in parallel in my own notebook, as I was training on a different dataset. Amazing!
One of the best lectures on the fundamentals of ML
The world needs to know about this youtube series. I already published it to my network on linkedin.
I finally beat 2.17, with 2.14.
With context_size:4, embedding_dimension:5, hidden_dimension:300, total_iterations:200000, batch_size:800.
And now for practice I am going to make a neural network to predict another kind of sequence. (I'm in the process of preparing/shaping the data, which is not easy) Fun stuff :P
Really fiending for that next video though xD I'm excited to learn about RNNs and Convnets and especially transformers.
Your batch size should be a power of 2
Thank you for doing this. This is truly a gift for all the learners and enthusiasts of this field!
As a programmer trying to learn ML, this is gold!
underrated series. Very informative. Watching this series before jumping into the Chatbot video. I am currently building my own poem-gpt
Thank you so much Andrej! This the best NN series. Hope you will create more videos:)
This is amazing and Informative. Thank you for the series.
This is what true public service looks like ... kudos Andrej :)
I am into neural nets the last 2 years and i think i know some stuff about them (the basics at least) and i have taken a couple of courses and stuff about ml/dl. I was always wandring why do i need val and test set, why test the model on 2 different sets of the same data. So hearing that the val set is for finetuning of hyperparameters is a first for me and the fact that you use test set a few times in order to avoid overfitting on it as well. I am amazed by the content on your videos and the way you teach things. Keep up the good work, you are making the community a better place.
I dont understand the overuse of the test-set.
I mean we are only forward-passing that to evaluate the performance, so we arent learning anything?
I can maybe see it when the dev sees the result and changes the network to better fit the test-case? But thats good isnt it?
@@tarakivu8861 For the people later who will maybe stumble upon this comment and probably has the same doubt, here's an intuition i have that gives me a pretty thorough understanding.
Say you are studying for an exam, and you use your textbooks for learning (note the use of learning here as well). Now, you want to know how good you're doing with the content you're learning from the textbooks, hence you give a mock exam, which kind of replicates the feeling of the final exam you're going to give. So you give test on the mock paper, and you note the mistakes or errors you are making on the mock paper, and you keep studying the text books and you give the mock test over and over again, periodically. After some time, you kind of have an estimate of how well you are going to do in the final exam based off the results you are getting on the mock exam.
Here, learning from the textbooks is the model training on the train set. The mock exam, is the validation set. The final exam (which you just give once), is like the test set.
Note that Dev set doesn't really change the network in any form or matter, it just gives us an estimate of how the model can perform on the test set. It's like if you are performing bad on the mock test, you know you can't make stuff better for the final exam.
Hi Andrej! First of all, THANKS! As everybody here saysm the kind of content you are teaching is INVALUABLE, and the fact that you are sharing it for free is really amazing.
I wanted to ask what kind of hardware you are using for this.
I see your code executes pretty fast where on my side, although having an M2 it's not as fast, so Im really interested on your local configuration.
Thanks!
An alternative to emb.view(-1, 6) is emb.flatten(start_dim=1) which also returns a view in this case and has the advantage of not having to specify the length of the concatenated embeddings.
Also, I don't know if there's really any advantage, but an alternative to prob[torch.arange(32), Y] would be prob.index_select(dim=1, index=Y)
I like how Andrej is operating with tensors, that's super cool. I think that we need a separate video about that from Andrej. It is super important.
Great video! One question, @AndrejKarpathy: around 50:30 or so you show how to graph an optimal learning rate and ultimately you determine that the 0.1 you started with was pretty good. However, unless I'm misunderstanding your code, aren't you iterating over the 1000 different loss function candidates while *simultaneously* doing 1000 consecutive passes over the neural net? Meaning, the loss will naturally be lower during later iterations since you've already done a bunch of backward passes, so the biggest loss improvements would always be stacked towards the beginning of the 1000 iterations, right? Won't that bias your optimal learning rate calculation towards the first few candidates?
I found this a little confusing too since the expectation is to do 1000 steps of gradient descent with each learning rate separately. I think this trick of simultaneously changing learning rate while training (on mini-batches) is just a quick way to broadly illustrate how learning rate changes impact the loss. If the learning rate is too low initially then the loss will decrease very slowly, which is what happens. When the learning rate increases the loss decrease is more rapid. When the learning rate is too high the loss becomes unstable (can increase).
Great lecture! I also appreciate the effort on putting it on google collab. Way easier to access for people not familiar with python and its adventurous ecosystem.
Recently got a new mac with an m1 processor and it took me a while to get tensorflow to run locally with gpu support, since I'm no python expert and therefor not familiar with all their package managers :-)
Thank You Andrej for the excellent training
Thanks for this awesome video series Andrej.
Suggestion for a video: Embedding models such as OpenAIs text-embedding-3-small/large models. How they work & simple one.
On 45:45 I was waiting fot Karpathy's constant to appear. Thank you for the lecture, Andrej
I got the loss on the training set under 2 (1.9953 to be exact). But this was a clear case of overfitting as the regularisation loss actually increased to 2.1762 🙂
HyperParameters:
wordDim ( C ) = 15
layerDim (W1 output/W2 input)= 500
iterations = 400000
batchSize = 100
Thank you for doing this. Please upload rest of the videos of the series soon.
I feel so excited that the question i had last video about why the loss cant be zero had interpolated in this video lol. SO EXCITED!!!!!!
Awesome explanation. Can’t wait for the next part and ultimately the transformer one.
Thanks for taking the time from your research.
Wow! Mind-blowing lecture.
Wow! I have longed to learn about hyper-parameters and training in a nutshell. Another Aha moment for me in Deep Learning. Thanks a trillion.
Hi @AndrejKarpathy 17,000 words in 30 dimensions is not cramped at all.
2^30 ≈ 1 billion. This means that 30 dimensions can support 1 billion points with each point being on the opposite side of axis to every other point.
Consider the 3D case: with 8 points forming a cube around the origin (1,1,1) (1,1,-1)...(-1,-1,-1)
For each point, every other point is on the opposite side of at least one axis, they are at least distance 2 apart.
Each point has it's own 'corner' of the space - it's the same in 30 dimensions.
Great, very didactical content. Thanks a lot for all the effort you put on this!
If someone sorts all the NN courses, videos, MOOCs by their density of knowledge in descending order, this video will be at the top.
You are the best as always, thanks for the content
I've never been this excited for a lecture video before
a test loss of 2.091. Embedding size of 4 characters in context, all parameters from a uniform distribution, 64 mini batch, 200 neurons in the softmax layer, dynamic learning rates and an added regularisation term to the loss (with a very low alpha though, 0.0001). Tranining loss - 2.054, Dev loss - 2.089, Test loss - 2.091.
I cannot believe word2vec was invented in 2003 instead of 2014
It was solely the allure of this channel that led me to subscribe to CZcams Premium.
It is an absolute honor to learn from the very best. Thanks Andrej.
Astounding Sir ! I am very thankful to you, much respect 😁
Andrej, I was able to get a train and validation cross-entropy of 2.0243 and 2.1333, respectively. The hyperparameters that were changed were...the number of characters used to predict the next character (from 3 to 5), the length of each embedding vector (from 2 to 27, i.e., the number of tokens), and the batch size (from 32 to 128). Also, after optimizing the learning rate, I took the average of the 10 learning rates that produced the lowest cross-entropy and trained the model with it. Finally, I decreased that 'averaged' learning rate even further by an order of two magnitudes and trained the model one last time. In short, a lot of experimentation was required. Haha.
The dataset consists of the set of unique names, which tends to over-emphasise weird names and eccentric spellings. The generated names are representative of the dataset but less representative of real names. To avoid the problem we could weight the dataset by frequency of name usage in the real world.