Hot Search
Geniatech AIM-M-H10 is a high-performance M.2 AI accelerator card powered by the Hailo-10 processor. Designed for seamless integration into edge devices and AI PCs, it delivers efficient on-device generative AI and vision processing with outstanding power efficiency.
Plug-and-play M.2 interface enables rapid evaluation of vision and generative AI models (LLMs, VLMs, diffusion) on x86/ARM hosts, accelerating prototyping and shortening development cycles.
Up to 40 TOPS INT8 at just 12W delivers offering an optimal balance of high throughput and power efficiency for demanding edge AI applications
Supports up to 16GB LPDDR4X on-module memory for large AI models and operates reliably from 0°C to 70°C in industrial environments.
Compatible with TensorFlow, PyTorch, TorchScript, ONNX, and major models including YOLOv8 and Llama 2.0.
| AI Processor | Hailo-10 |
| AI Performance | Up to 40 TOPS |
| Edge Generative AI | Supports LLMs, VLMs and diffusion models |
| AI Model Types | CNN, Transformer, Vision Transformer, Generative AI Models |
| Supported Host Processors | NXP, NVIDIA, Qualcomm, Xilinx |
| Supported AI Frameworks | TensorFlow, PyTorch, TorchScript, ONNX, Caffe, MXNet |
| On-Module Memory | Up to 16GB LPDDR4/4X |
| Host Interface | M.2 Key-M, PCIe Gen4 x4 |
| OS Support | Linux, Windows |
| Typical Power Consumption | 12W |
| Operating Temperature | 0°C to 70°C |
| Dimensions | 22 x 80 mm |
