Datasets
Tools and Technologies
- Sign up for a Google Developer account
- Watch the Google Cloud x MLB™ Getting Started Webinar
- Join the AI Skills Quest to get started with Google Cloud AI
- Join the Devpost Discord server channels dedicated to this challenge
- Google Cloud Generative AI GitHub Repo
- Gemini API documentation
- Gemini API Cookbook
- Vertex AI Documentation
- Vertex AI Codelabs
- Vertex AI Best Practice Guide
- Using Vertex AI Search to Find the Next Generation of Baseball Stars Blog Post
- MLB™ Player Digital Engagement Forecasting Kaggle Competition (2021)
- MLB™ Data Insights Google Cloud Blog Post (2020)
- Exploring MLB™ Provided Datasets Colab notebook
Videos
- Generative AI for developers → https://www.youtube.com/playlist?list=PLIivdWyY5sqLRCzKJyixrIDPQKwU6XHpn
- Real terms for AI → https://www.youtube.com/playlist?list=PLIivdWyY5sqLvGdVLJZh2EMax97_T-OIB
- Learn the Gemini API → https://www.youtube.com/playlist?list=PLOU2XLYxmsIJ3nNA9-0zm7XjfZtI2ntfC
- Build with Google AI → https://www.youtube.com/playlist?list=PLOU2XLYxmsIIof-OQwbS0jL7nBTzHYFSv
FAQ
Can I use a 3rd party AI model or tool?
To ensure fair competition and compliance, we are providing further clarification on the use of generative AI tools in this hackathon.
Allowed Tools and Services
-
Google Cloud AI Services: Participants are required to use Google Cloud AI services for any generative AI functionality within their projects. This includes, but is not limited to:
- Gemini: For multimodal input and output, text generation, and conversational AI.
- Gemini API and AI Studio: For accessing and experimenting with various generative AI models.
- Vertex AI: For building, deploying, and managing machine learning models.
- Imagen: For image generation and manipulation.
-
Third-Party Orchestration and Composable Frameworks: Participants may use third-party tools and frameworks that facilitate the orchestration and composition of AI models, as long as these tools do not involve using or training new models using MLB data. Examples include:
- LangChain: For building applications with large language models.
-
LlamaIndex: For connecting large language models to external data sources.
Prohibited Tools and Services
- Third-Party Generative AI Models: Participants are prohibited from using generative AI models from any provider other than Google Cloud. This includes, but is not limited to, models from OpenAI, Microsoft, Anthropic, or any other third-party provider.
-
Training New Models with MLB Data: Participants are strictly prohibited from using any MLB data provided in this hackathon to train or fine-tune generative AI models not offered by Google Cloud.
Examples
- Allowed: Using LangChain to orchestrate a workflow that combines Gemini and a custom model built on Vertex AI.
- Allowed: Using Vertex AI to fine-tune a pre-trained Google Cloud AI model for a specific baseball-related task.
- Allowed: Using Gemini models to analyze game footage and extract insights.
- Not Allowed: Using OpenAI's GPT-4o to generate text summaries of MLB games.
- Not Allowed: Training a new image generation model on a dataset of MLB player photos using a third-party platform.
- Not Allowed: Using Twelve Labs Search to enable semantic search of MLB game footage.
Note: If you have any questions about the permitted use of specific tools or services, please contact the hackathon organizers for clarification.
Major League Baseball trademarks and copyrights are used with permission of Major League Baseball. Visit MLB.com.