Build a Neural Network with Python Tutorial | Deep Learning with PyTorch

preview_player
Показать описание

Build a model that predicts whether or not is going to rain tomorrow using real-world weather data. Learn how to train, evaluate and make predictions with it

⭐️ Tutorial Contents ⭐️

(00:12) Download the weather data
(04:13) Data preprocessing
(17:45) Build a Neural Network with PyTorch
(32:05) Choose a loss function & optimizer
(40:25) Doing computations on the GPU (with CUDA)
(44:03) Training your Neural Network
(49:17) Saving & loading a model with PyTorch
(50:15) Evaluation (How good your model is?)
(54:41) Making predictions (Is it going to rain?)

Рекомендации по теме
Комментарии
Автор

This guy is awesome, all his tutorials are so easy to follow and are really informative.

benjaminbenjamin
Автор

hello sir, I am your new student i will be visiting your channel every few hours every day. Thank you for teaching me so many things. you are my teacher and am your student now. Thanks again :)

hussain
Автор

I balanced the dataset by oversampling. The best model has forecast rain with 74% accuracy and no rain with 73% accuracy. I've tried different numbers of neurons/nodes in the hidden layer. Next, I will add more hidden layers. I suspect that will have a real impact.

lakeguy
Автор

Thank you very much for this video! Just one question, what happened on 47:35?

forvm
Автор

Are you going to release your modified ANN Visualizer?

cwill
Автор

I have followed your code. What does it mean when the model fails to learn? The train_loss.item() doesn't change after each loop? The NN runs with no errors but doesn't learn. Any ideas? thank you.

lakeguy
Автор

how many cuda cores does the google colab employ? Is there a limit to how big a model you can pass to a colab GPU?

lakeguy
Автор

Superb.... Keep it up..
I need code for this lesson....

rushikeshbulbule
Автор

I don't understand why you didn't balance the dataset since 78% of your y values are zero...

lakeguy