After building several AI agents, I compiled my library of components together to create Tyler—a framework to build multimodal agents with built-in/custom/MCP tools, response streaming, and persistent conversation storage—all in a few lines of code. https://t.co/ajUnuWrwve
Announcing our new RAG++ course, now available in collaboration with @CohereAI and @Weaviate_io. Created for engineers looking to build production-ready RAG systems.
The course covers everything from evaluation strategies and data preprocessing to advanced retrieval techniques and prompt optimization and includes hands-on exercises with code notebooks and Cohere credits. Register here: https://t.co/iCLJ9vbtIP
Learn how to track & evaluate LLM apps with this intro notebook showing Weave by @wandb
- track function calls & code
- publish & retrieve versioned objects like datasets/prompts/model configs
- evaluate with a simple API
https://t.co/4aFtVTfrWU
Always enjoy getting my @wandb year in review. This year, I spent 48 hours training 355 models on @wandb! My #longestrun - still-grass-6 - trained for 17 hours. https://t.co/8rAIIcExjw
🚨Giveaway Alert🚨
Unlock a FREE ticket (and some W&B swag) to the Fully Connected Conference in San Francisco on June 7th 🌉
Celebrate ML with W&B by replying to this tweet with your favorite project, paper, etc. that you want the world to know about 🌎
BONUS: Join our Discord and post a screenshot of a W&B project you have worked on to enter to win a GRAND prize 🪄🐝
🔗https://t.co/61kUQkxa1i
@McHughes288 @wandb Makes sense. Thanks for adding the screenshot. We're currently gathering feedback for new use cases and will definitely be iterating on this new product. I'll let you know when we have some concepts and would love to get your feedback.
@McHughes288 @wandb Great question! In fact, we are currently designing a more scalable timeline. Here is a (very early) concept. Does this align with what you had in mind?