Live Demo of
Data-centric Foundation Model Development in
Snorkel Flow

At this live demo event, you’ll learn how enterprise AI/ML teams, including Pixability, a top-3 US bank, and a global ecommerce company, have rapidly accelerated AI development by overcoming the adaptation and deployment challenges that were blocking them from adopting foundation models.

Braden Hancock, Snorkel AI Co-founder and Head of Technology, will demonstrate the new capabilities for Data-centric Foundation Model Development in Snorkel Flow including:

• Foundation Model Fine-tuning: Fine-tune modern foundation models like GPT-3, RoBERTa, CLIP, and more by programmatically creating large, task- and domain-specific training data sets.
• Foundation Model Warm Start: Use foundation models with zero- and few-shot learning to auto-label training data with a push of a button to train deployable models.
• Foundation Model Prompt Builder: Develop, evaluate, and combine prompts to tune and correct the output of foundation models to precisely label datasets and train deployable models.

These new features further support Snorkel Flow’s data-centric development workflow, which:

• Accelerates AI development with programmatic labeling, a fundamentally more scalable approach to building and maintaining high quality datasets
• Enables utilizing existing resources (noisy labels, models, ontologies, etc.), domain expert heuristics, embeddings, foundation models and more to improve label quality
• Continuously updates and analyzes models to guide training data development
• Enables adapting to real-world changes with a few clicks instead of manual relabeling from scratch
• Provides transparency around where your training labels are coming from and how to update or correct them to improve model performance
• Supports one-click deployment of trained models in sealed application packages that can be run locally or transferred to your production environment

Space at this event is limited to enterprise users, and advance registration with a business email is required. Your registration request will be reviewed by the Snorkel team, and you will receive an email confirmation from Zoom once approved to attend the live event. Registration will close 24 hours prior to the event.