ManuAI
VI

Tech track · 8 steps · 82 min

Deep Learning

“I want to build and train neural networks”

How neural networks learn, how CNNs and transformers work, and how today's language and image models are trained. For people who can write a little Python and want to train real models on free GPUs.

3–4 hours a week for 6 weeks

Start step 1 0 of 8 steps ticked

01

The steps

Ticks and scores are saved in this browser only. No account.

  1. 01Neural networks, intuitivelyYou can explain neurons, layers, loss, gradient descent and backpropagation in plain words.
  2. 02Training in practiceYou can read a training curve and spot overfitting, and you know the usual fixes.
  3. 03CNNs for imagesYou know why convolutional networks suit images and how to reuse a pretrained one.
  4. 04Sequences, attention and transformersYou can explain what attention does and why transformers replaced older sequence models.
  5. 05How LLMs are trainedYou can describe pretraining, fine-tuning and preference tuning, and what each one changes.
  6. 06Diffusion and generative imagesYou understand, at a high level, how image generators turn noise into pictures.
  7. 07PyTorch and free GPUsYou can write a basic PyTorch training loop and run it on a free cloud GPU, knowing the limits.
  8. 08Your portfolio projectYou can plan, build and present one deep learning project that shows real skill.

02

Quiz and certificate

Deep Learning track quiz

Tick every step, then pass the 10-question quiz with 80% or more. The certificate comes with a link anyone can check.

  • 0 of 8 steps ticked
  • Pass the track quiz with 80%
Take the track quiz Certificate locked

03

Where to go next

Free, reputable resources for going deeper. They open on other sites.