Google Coral Mini PCIe Edge TPU ML Accelerator
The Google Coral Mini PCIe Accelerator integrates Google’s Edge TPU ML accelerator into a compact Half-Mini PCIe card, delivering 4 TOPS (int8) peak performance at 2 TOPS per watt. This module enables efficient, on-device machine learning inference for existing embedded platforms and industrial gateways.
Features
- Google Edge TPU ML accelerator for on-device TensorFlow Lite inference.
- Delivers 4 TOPS (int8) peak performance with 2 TOPS per watt efficiency.
- Standard Half-Mini PCIe Gen2 x1 interface for broad system compatibility.
- Integrated power management and wide operating temperature range (-20 to +85 °C).
Out of stock
The Google Coral Mini PCIe Accelerator is a Half-Mini PCIe card engineered to bring Google’s Edge TPU ML accelerator to existing systems and products. The Edge TPU, a purpose-built ASIC by Google, excels at accelerating TensorFlow Lite models with remarkable power efficiency. It is capable of performing 4 trillion operations per second (4 TOPS) using just 2 watts of power, equating to an impressive 2 TOPS per watt. For instance, a single Edge TPU can execute state-of-the-art mobile vision models like MobileNet v2 at nearly 400 frames per second, facilitating high-speed local inference.
This on-device machine learning processing offers significant advantages, including reduced latency, enhanced data privacy, and the elimination of a constant internet connection requirement. Its Half-Mini PCIe form factor makes it ideal for integrating local ML acceleration into a wide range of products, such as embedded platforms, mini-PCs, and industrial gateways that feature an available Mini PCIe slot, upgrading their AI capabilities directly at the edge.
Additional Information
| Weight | 0.006 kg |
|---|---|
| Dimensions | 3 × 2.68 cm |
