Keras (3) for the Curious and Creative
summary
Shows concrete demos and code: Keras 3 with Jax backend for a quick object‑detection build, swapping backends between JAX TensorFlow and PyTorch, and importing pre‑trained PyTorch models into Keras 3. Explains exact model pieces used (Conv + MaxPool, optimizer choices like Adam/SGD, checkpoint + early stopping), image sizes (1280→6×6 demo, 512×512 generation), and implements stable diffusion text‑to‑image with controllable steps (25–100) and super‑resolution to upsample results.
AI-written summary · about
chapters (8)
- 0:00Welcome to Keras 3
- 10:09Deep learning fundamentals
- 21:37Images and computer vision
- 36:47Building models
- 45:16Keras 3 multi-backend
- 56:17Recap and transition
- 1:00:18Generative AI with Stable Diffusion
- 1:07:24Keras 3 multi-backend demo
related talks
20mKeras 3: Deep Learning made easy
43m: Shortcut to AI mastery
19mMaking Deep Learning Easier - Updates on Keras
01h 37mTen Steps to Keras
05mTop 5 takeaways from
47mDeep Learning: a practical introduction from an Android dev perspective
44mTensorflow, Tensorflow, What is it that you see?
01h 29mHands-on introduction to Deep Learning with Keras and Tensorflow