EmbeddingGemma 2: an open, lightweight multimodal embedding model
deepmind.google·6d ago
AI Engineering
Lobsters·Aug 13·5 min read
Introducing chestnut
TL;DR
Chestnut is a new compute upgrade for the comma four device, enabling it to run larger openpilot models, comparable to Tesla's HW4. This advancement allows for enhanced driving intelligence and capabilities.
✦ Why It Matters
Consider upgrading to chestnut for enhanced model performance in your autonomous driving applications.
Key Takeaways
- Upgrade your comma four with chestnut for enhanced driving models.
- Utilize the new 1B parameter model for superior performance.
- Install the ready-to-drive kit easily in your vehicle.
- Join the community on Discord to discuss advancements.
How It Works
Chestnut utilizes the MLSim architecture, which allows for efficient scaling of driving models with increased compute and data. This architecture enables the comma fleet to log extensive driving data, which is then used to train larger and more capable models.