Home / PyTorch
PyTorch courses that build real models step by step. Tensors first, then training, then a full language model.
4 courses · 290 lessons · work through them in order, top to bottom
Python · Beginner
Tensors, autograd and your first trained models.
Meet PyTorch the friendly way: tensors and shapes, autograd, modules, losses and optimizers, the training loop and data handling — finishing with a trained classifier and a capstone program. Assumes basic Python; no ML background needed.
79 lessons · 19 topics · 0/79 completed
Python · Intermediate
Training discipline, CNNs, sequences and transfer.
Training done properly: splits and metrics, schedules, regularization and normalization, losses, images and CNNs, embeddings and text, RNNs, transfer learning and data engineering — ending in a full-pipeline capstone. Builds on PyTorch Beginner.
87 lessons · 19 topics · 0/87 completed
Python · Advanced
Custom autograd, transformers, generation and export.
Worked by hand: custom autograd and gradcheck, attention and transformers, distributions, VAEs, GANs and diffusion, export and quantization, FX graphs, distributed honestly and distillation — ending in an export capstone. Builds on PyTorch Intermediate.
71 lessons · 17 topics · 0/71 completed
Python · Advanced
A tiny GPT, built by hand — tokens to text.
Build a small GPT by hand: character tokens, embeddings with position, self-attention and transformer blocks, a training loop, then generation, sampling and a tiny BPE — ending in a trained capstone. Builds on PyTorch Intermediate.
53 lessons · 15 topics · 0/53 completed