Google Coral M.2 Accelerator with Dual Edge TPU
This Google Coral M.2 Accelerator features two Google Edge TPU ML coprocessors, delivering a combined 8 TOPS of peak performance for efficient TensorFlow Lite model inference. Designed for M.2 E-key slots with dual PCIe Gen2 x1 interfaces, it integrates seamlessly with Debian Linux and Windows systems to enhance AI capabilities.
Features
- Integrates two Google Edge TPU ML accelerators for parallel processing.
- Achieves 8 TOPS total peak performance (int8) with 2 TOPS per watt efficiency.
- Connects via M.2 E-key (M.2-2230-D3-E) using dual PCIe Gen2 x1 interfaces.
- Optimized for accelerating TensorFlow Lite models on Debian Linux and Windows systems.
- Provides significant AI inference boost for edge computing applications.
Out of stock
The Google Coral M.2 Accelerator with Dual Edge TPU is a compact M.2 module, specifically designed with an E-key interface, integrating two distinct Google Edge TPU machine learning accelerators. Each Edge TPU features its own dedicated PCIe Gen2 x1 interface, enabling robust parallel processing capabilities. These custom-designed ASICs by Google are highly efficient at accelerating TensorFlow Lite models, with each unit capable of 4 trillion operations per second (4 TOPS) while consuming just 2 watts of power, resulting in an impressive 2 TOPS per watt.
This M.2 module is engineered to upgrade existing systems and products equipped with a compatible M.2 E-key slot by adding powerful dual Edge TPU coprocessors. It’s crucial to ensure that the host system’s M.2 E-key socket connects both PCIe x1 instances to fully utilize both Edge TPUs, as not all E-key sockets provide this dual connection. Given its dual PCIe x1 usage and specific power requirements, users should consult the datasheet to confirm compatibility and power delivery specifications for optimal performance and integration within Debian-based Linux or Windows 10 environments.
Additional Information
| Weight | 0.003 kg |
|---|---|
| Dimensions | 2.2 × 3 cm |
