/courses/machine-learning/deep-learning-pytorch/Recent content in Deep Learning with PyTorch on DATATWEETSHugoenCopyright (c) 2025 DatatweetsFri, 14 Nov 2025 09:00:00 +0200/courses/machine-learning/deep-learning-pytorch/lesson-1-deep-learning-fundamentals/Fri, 14 Nov 2025 09:00:00 +0200/courses/machine-learning/deep-learning-pytorch/lesson-1-deep-learning-fundamentals/Learn what deep learning is and where it shines compared to classical machine learning, why a framework like PyTorch matters, and how the high-level training loop works. You will meet the real Indian IPO dataset you will use throughout this module and frame the goal: predicting whether an IPO lists with a gain./courses/machine-learning/deep-learning-pytorch/lesson-2-tensors-and-autograd-in-pytorch/Fri, 14 Nov 2025 09:00:00 +0200/courses/machine-learning/deep-learning-pytorch/lesson-2-tensors-and-autograd-in-pytorch/Meet the tensor, PyTorch’s core data structure, and autograd, the engine that computes gradients automatically. You will create tensors, control their dtype and shape, move data between NumPy and PyTorch, run the essential tensor operations, and watch autograd differentiate a small expression you can check by hand, all using the real Indian IPO dataset./courses/machine-learning/deep-learning-pytorch/lesson-3-building-neural-networks-with-nn.sequential/Fri, 14 Nov 2025 09:00:00 +0200/courses/machine-learning/deep-learning-pytorch/lesson-3-building-neural-networks-with-nn.sequential/Learn how PyTorch organizes neural network components and build a complete binary classifier with nn.Sequential. You will stack nn.Linear and activation layers into the exact 6 to 32 to 16 to 1 IPO architecture, inspect its parameters, and run a forward pass on a real batch from the Indian IPO dataset./courses/machine-learning/deep-learning-pytorch/lesson-4-training-neural-networks/Fri, 14 Nov 2025 09:00:00 +0200/courses/machine-learning/deep-learning-pytorch/lesson-4-training-neural-networks/Learn how a neural network actually learns in PyTorch: the loss function, the optimizer, and the zero_grad / forward / loss / backward / step cycle. You will train a real classifier on the Indian IPO dataset, evaluate it honestly with accuracy and AUC, and read its training curve./courses/machine-learning/deep-learning-pytorch/lesson-5-deep-networks-and-regularization/Fri, 14 Nov 2025 09:00:00 +0200/courses/machine-learning/deep-learning-pytorch/lesson-5-deep-networks-and-regularization/Build a deeper PyTorch classifier on the real Indian IPO dataset, watch it overfit a small dataset, then close the train/test gap with nn.Dropout and L2 weight decay. You will compare an unregularized model against a regularized one and read the difference straight off the training curves./courses/machine-learning/deep-learning-pytorch/lesson-6-guided-project-predicting-ipo-listing-gains-with-pytorch/Fri, 14 Nov 2025 09:00:00 +0200/courses/machine-learning/deep-learning-pytorch/lesson-6-guided-project-predicting-ipo-listing-gains-with-pytorch/Bring the whole PyTorch module together in a guided project. You will load the real Indian IPO dataset, engineer a binary target, build and train an MLP, evaluate it honestly with a confusion matrix and AUC, and learn why deep learning does not always beat a simple baseline.