Tutorial: Beyond Neural Search: Hands-on Building Cross-Modal/Multi-Moda...
summary
Hands-on session builds a two-stage Jina flow (diffusion plus upscaling) in about 300 lines of Python, using DocArray with Hugging Face diffusers/Transformers to run Stable Diffusion from Google Colab on a GPU, downloading ~5 GB of model weights and producing text-to-image outputs. The code packages as a Docker executor and deploys to Jina Cloud Kubernetes; live generation succeeds for samples but deployment hits connectivity/GPU instability during the demo.
AI-written summary · about
chapters (8)
- 0:00Start
- 1:47Keyword Based Search
- 8:38What Is a Generative Ai
- 26:21Install Dependency
- 33:11Creating a Flow
- 48:50Deploy to the Cloud
- 54:41Python Interface
- 1:04:13Auto Scaling
related talks
36mAccessible Neural Search and Cloud-Nativeness for ML Engineers
43m: Shortcut to AI mastery
41mBuilding Generative Image & Video models at Scale
01h 08mGoing multi-modal
34mDeep Learning at Scale
01h 29mFull-stack Machine Learning and Generative AI for Data Scientists
01h 34mHands-on AI workshop: Multimodal agents
01h 03mIt's All About AI