Quick-guide directory
PyTorch quick guides
Four connected summaries of the concepts, functions and decisions used in Courses 1 and 2. Choose a topic below for a short review.
New to PyTorch? Read 1 → 2 → 3 → 4. Returning to one topic? Go straight to its guide for the main ideas, useful functions and common mistakes.
Choose a topic
01 · Course 1 foundation
Tensors and the training loop
Read tensor shapes, follow a batch through learning, understand ReLU and CNNs, and validate without updating the model.
Read quick guide →02 · Course 2 training toolsMetrics, tuning and efficiency
Compare experiments fairly, schedule the learning rate, find slow steps, and reduce memory use.
Read quick guide →03 · Course 2 image workflowsImages and pretrained models
Choose transforms and noise, read model outputs, and adapt pretrained features to a new set of classes.
Read quick guide →04 · Course 2 text workflowsTokens and text classifiers
Turn different-length sentences into batches, pool embeddings, and decide when a contextual model is useful.
Read quick guide →