Introducing chestnut
摘要
comma.ai 发布 chestnut，一款面向 comma four 的算力升级套件，旨在运行 openpilot 更大规模的驾驶模型。文章介绍配套推出的首个 chestnut 级 1B 参数驾驶模型，参数规模为最新端侧模型的 30 倍、FLOPs 为其 100 倍，并称 comma four 搭配 chestnut 后算力接近特斯拉 HW4。产品提供 ready to drive 套装（含 AMD Radeon RX 9060 8GB，售价 799 美元）与 tiny chestnut（249 美元，需自备 GPU）两种版本，均含 30 天试用和一年保修，并提及团队将主要聚焦 chestnut 级模型的开发。
荐读理由
comma 用桌面级 GPU 外挂把车载推理算力拉到接近特斯拉 HW4，并发布 1B 参数驾驶模型，这套外挂 GPU 扩展端侧算力的路线可直接借鉴到你的 AI 硬件项目
原文
Upgrade your comma four’s compute to run openpilot’s biggest models.
Today, we’re launching chestnut, a compute upgrade for comma four. Want to run big models? Nothing beats a desktop GPU. We’re also launching the first chestnut-class model, which is bigger than all openpilot driving models we’ve released combined.
Scaling beyond intelligence densityPermalink**
Since the beginning, openpilot has led on intelligence density. For the last 10 years, we’ve shipped increasingly capable driving models in an on-device compute footprint of 10 W. We shipped fully end-to-end lateral planning, experimental end-to-end longitudinal planning, and even a point-to-point demo to Taco Bell.
Over the last two years, we rewrote our stack to scale efficiently with compute and data. With the MLSim architecture, more driving data and training compute translate directly into better openpilot models. The comma fleet now logs nearly one million driving minutes every day, and our scaled-up data pipeline ingests 15% of those drives into the training cluster.
Bigger models need more compute to run. With chestnut, openpilot is scaling beyond cellphone-sized inference compute. A comma four paired with chestnut has similar compute to Tesla HW4.
A new 1B driving modelPermalink**
We’re launching chestnut with a new 1B parameter driving model in the upcoming openpilot 0.11.2 release. It’s our biggest model yet, and the first in a new line of chestnut-class models.
Compared to the latest on-device model, the first chestnut-class model has 30x more parameters, 100x more FLOPs, and room to go even bigger. Going forward, our autonomy team will focus primarily on maximizing the capabilities of chestnut-class models, and we expect this new class of model to increase development velocity.
Ready to DrivePermalink**
chestnut comes in a convenient ready to drive kit, which includes chestnut, an AMD Radeon RX 9060 8GB, and everything needed to power, connect, and mount it in your car.
The GPU mounts out of the way in the passenger footwell or under the passenger seat. Run the included cable to comma four, connect to your car’s cigarette lighter, and you’re ready to run openpilot’s biggest models.
Get your chestnutPermalink**
chestnut is available today in the comma shop.
chestnut ready to drive is available now for $799.
tiny chestnut is available now for $249 for use with your own GPU.
Both versions come with a 30-day money-back trial and a one-year hardware warranty.
Excited about scaling up? Join us in the #hw-chestnut channel on Discord!
Want to help scale up openpilot? We’re hiring for autonomy, product, and operations.
这条对你有帮助吗？