-
AI Chips
A four chip strike package for AI everywhere
-
DEEPX’s AI chips powered by DEEPX NPU ‘DX-GEN1’
-
DX-L1
AI Enabler that provides 12 eTOPS AI performance for small-to-medium scale applications
-
DX-L2
AI Enabler that provides 38 eTOPS AI performance for small-to-medium scale applications
-
DX-M1
AI Booster that provides 200 eTOPS AI performance for medium-to-large scale applications
-
DX-H1
AI Booster in the form of PCIe cards that provides 1,600 eTOPS AI performance for large scale applications
-
AI Modules
Ready-to-test evaluation kit and funtional modules
-
Software (SDK)
Just drop in your AI code and play your algorithms
AI Chips
- A four chip strike package powered by DEEPX NPU ‘DX-GEN1’
exceptionally outperforms on the level of accuracy and power-efficiency to truly revolutionize the AI world.
AI Boosters
DX-M1
- The AI Booster that achieves the highest Cost-efficiency (inference/$), Power-efficiency (TOPS/W) and Performance efficiency (FPS/TOPS) in the world.
- • AI Performance: 200 eTOPS**eTOPS: a performance metric equivalent to GPU's TOPS
-
- • DX-M1 Block Diagram
-
- • Support S-O-T-A Algorithms :
ResNet, MobileNet v1/v2/v3, SSD, EfficientNet, EfficientDet, YOLOv5, YOLOv7, YOLO8, DeepLabv3, PIDNet (Transfomer based models supported by Q1.2024)
- • Support S-O-T-A Algorithms :
-
- • Target Application :
Consumer and industrial robots, Machine Visions, AI required IPC&HPC, Smart Factory, Edge Computing, Smart Mobility
Smart Mobility
Smart Factory
- • Target Application :
AI Boosters
DX-H1
- The AI Booster in the form of PCIe cards that achieves the highest Cost-efficiency (inference/$), Power-efficiency (TOPS/W) and Performance efficiency (FPS/TOPS) in the world.
- • AI Performance: 1,600 eTOPS**eTOPS: a performance metric equivalent to GPU's TOPS
-
- • DX-H1 Block Diagram
-
- • Support S-O-T-A Algorithms :
ResNet, Mobilenet v1/v2, SSD, YOLOv5, YOLOv7
- • Support S-O-T-A Algorithms :
-
- • Target Application :
On-premise AI Server, Colocation/Hyperscale Data Center, Smart Building, NVR(Network Video Recorders)
On-premise AI Server
Data Center
- • Target Application :
AI Enablers
DX-L1
- The AI enabler that achieves cost-effective and ultra-low power, and supports the latest CNN algorithms for computer vision.
- • AI Performance: 12 eTOPS**eTOPS: a performance metric equivalent to GPU's TOPS
-
- • DX-L1 Block Diagram
-
- • Support S-O-T-A Algorithms :
ResNet, Mobilenet v1/v2, SSD, YOLOv5, YOLOv7
- • Support S-O-T-A Algorithms :
-
- • Target Application :
Internet protocol (IP) cameras, CCTV cameras, Video conferencing cameras, In-cabin Monitoring System (DMS), Camera Monitoring System(CMS), Robotic cameras, Drones, VR cameras
IP cameras
CCTV cameras
- • Target Application :
AI Enablers
DX-L2
- The AI enabler that achieves cost-effective and ultra-low power, and supports the latest CNN algorithms for computer vision.
- • AI Performance: 38 eTOPS**eTOPS: a performance metric equivalent to GPU's TOPS
-
- • DX-L2 Block Diagram
-
- • Support S-O-T-A Algorithms :
ResNet, Mobilenet v1/v2, SSD, YOLOv5, YOLOv7
- • Support S-O-T-A Algorithms :
-
- • Target Application :
Internet protocol (IP) cameras, CCTV cameras, Video conferencing cameras, In-cabin Monitoring System (DMS), Camera Monitoring System(CMS), Robotic cameras, Drones, VR cameras
Drones
Robot Vision System
- • Target Application :
