Falcon-H8 AI accelerator
Ask for price
High-performance and cost-efficiency PCIe AI accelerator card
- Support for 4, 5 or 6 Hailo-8 AI processors
- Support for up to 156 TOPS 8000 FPS of ResNet-50
- Support for Hailo-8 Software Development Tools
- High performance and power efficiency
Description
Product Details
Lanner’s Falcon H8 modular, PCIe form factor provides an easily deployable solution for engineers looking to offload CPU loading for low-latency AI inference. With high-density AI processors, the Falcon H8 accommodates 4, 5, or 6 Hailo-8 AI processors, offering a modular, cost-effective Edge AI solution with high processing capabilities and power efficiency. Through a standard PCIe interface, the Falcon H8 AI Accelerator Card enables legacy devices such as NVRs, Edge AI boxes, Industrial PCs and robots to run video-intensive, mission-critical Edge AI applications such as video analytics, traffic management, access control, and beyond.
Specifications
AI Performance
| Up to 156 TOPS 8000 FPS of ResNet-50, 1300 FPS of YOLOv5m |
AI Processor
|
4-6 Hailo-8 AI Processors with Hailo patented structure Defined Dataflow architecture |
AI Frameworks Support
|
Hailo AI Dataflow compiler with profiler and emulator Supports TensorFlow, ONNX and PyTorch frameworks |
PCIexpress Interface
| PCI Express x16 compliant with PCI Express specification v3.0 |
System Compatibility
|
Intel x86 or ARM Devices, Linux OS e.g., Ubuntu, Yocto Lanner Network and Edge AI Appliances |
Power Consumption
| Typical: 45 W |
Temperature
| Operating |
0 ~ 70ºC (commercial grade) -40 ~ 85ºC (industrial grade) |
| Storage | -40 ~ 85ºC |
Humidity
| 5% ~ 90% RH, non-condensing |
Dimensions
|
Standard PCIe single slot form-factor 167.65 mm x 111.15 mm without bracket |
Certification
| CE | Class A |
| FCC | Class A |