Local course preview

Deep Learning Foundations

Learn tensors, gradients, CPU neural-network training, evaluation and a small CNN using generated datasets. The optional runtime is actual local PyTorch on CPU, separate from the reading guide. No full transformer/LLM mastery, GPU runtime, production readiness or certification is claimed.

Preparation: Python for AI and Data, Statistics with Data Storytelling and Applied Machine Learning, or equivalent preparation.

Read all 12 modules as a guest, in any order. An optional account saves your place and reading marks. Practice rubrics are for self-review; they do not unlock reading or award a certificate.

Planning estimate: See the provisional estimates in the syllabus; learner pacing has not been piloted. Timing and learner outcomes have not been piloted.

Start course Download original course materials (includes reference answers).

Reading progress

12 lessons remain open to guests. Marking a lesson read records reading only; it does not award assessment credit or a certificate.

Device reading marks require browser storage. Reading is always available.

Course contents

18px screen prose, wrapped equations and keyboard scrolling. Original materials remain separately available; human accessibility testing is pending.

Deep Learning Foundations

Reading progress

Course contents

Read a tensor before training a model

Follow a prediction into its loss

Trace responsibility backward

Decide what the model is allowed to learn from

Give each training-loop line a job

Ask what extra flexibility actually buys

When fitting better means predicting worse

Debug the learning process before enlarging the model

A tiny convolutional network you can account for

Turn a successful run into a reproducible experiment

Two bridges beyond the core: reuse and attention

Defend one small neural experiment end to end

Practice and references