: Shortcut to AI mastery
summary
Keras 3 delivers multi-backend Keras 3 that runs on TensorFlow, JAX, PyTorch or NumPy and keeps the tf.keras API while dropping TF1, demonstrated via Colab notebooks that install Keras, set os.environ backend, and import. The workshop shows concrete builds: a Stable Diffusion demo that produces multiple images with Stable Diffusion image gen, a grounding‑DINO + Segment Anything pipeline that outputs a harness mask (grounding dino + SAM mask), OPT/GPT‑2 causal LLMs for chat and fine‑tuning on the 15,000‑example Dolly‑15 dataset, and a 2×4 device mesh example for data/model parallelism.
AI-written summary · about
chapters (16)
- 0:00Introduction
- 0:32Overview
- 3:42Keras Modeling API
- 8:57Functional API
- 14:40Training API
- 16:09Training Entry Point
- 19:07Building Custom Components Layers
- 19:35Distribution API
- 23:18Keras
- 27:33Collab
- 28:03Backend
- 30:23Advanced features
- 31:23Segment Anything
- 35:08KLP
- 37:58Demo
- 42:38Outro
related talks
01h 14mKeras (3) for the Curious and Creative
20mKeras 3: Deep Learning made easy
19mMaking Deep Learning Easier - Updates on Keras
31mMaking machine learning accessible using Keras
01h 37mTen Steps to Keras
01h 03mIt's All About AI
32mGetting Started with TensorFlow 2.0
47mDeep Learning: a practical introduction from an Android dev perspective