Gigantor Technologies, Inc.
Gigantor Technologies is a pioneering company focused on redefining the landscape of Edge AI acceleration through innovative circuit designs and advanced AI processing technologies. Founded in 2020, Gigantor aims to overcome hardware limitations, enabling high-performance AI models capable of processing ultra-high definition visuals with near-zero latency and significant power efficiency.
Industries
N/A
Products
Parallel pipeline AI acceleration system
An edge-focused acceleration architecture that implements pipelined neural-network operations and optimized circuit designs to enable real-time, high-frame-rate inference on hardware targets such as FPGAs or ASICs.
Parallel pipeline AI acceleration system
An edge-focused acceleration architecture that implements pipelined neural-network operations and optimized circuit designs to enable real-time, high-frame-rate inference on hardware targets such as FPGAs or ASICs.
Services
Model evaluation and project estimate
Collection and assessment of model architecture, layer counts and types, activation functions, input resolution/frame rate, and outputs to produce a scoped estimate for hardware acceleration work.
Custom model-to-hardware optimization and circuit design
Transformation of trained network models into optimized pipelined circuit designs and streamlined inference pipelines suitable for FPGA or ASIC implementation.
Performance validation and demo benchmarking
Provision of demo videos and benchmark tests demonstrating frame-rate, resolution, and latency characteristics of accelerated designs on edge hardware.
Technical briefings and whitepaper distribution
Provision of technical whitepapers and scheduled video conferences to present architecture, patents, and use cases to prospective partners and customers.
Model evaluation and project estimate
Collection and assessment of model architecture, layer counts and types, activation functions, input resolution/frame rate, and outputs to produce a scoped estimate for hardware acceleration work.
Custom model-to-hardware optimization and circuit design
Transformation of trained network models into optimized pipelined circuit designs and streamlined inference pipelines suitable for FPGA or ASIC implementation.
Performance validation and demo benchmarking
Provision of demo videos and benchmark tests demonstrating frame-rate, resolution, and latency characteristics of accelerated designs on edge hardware.
Technical briefings and whitepaper distribution
Provision of technical whitepapers and scheduled video conferences to present architecture, patents, and use cases to prospective partners and customers.
Expertise Areas
- Edge AI hardware acceleration
- High-speed computer vision and real-time video inference
- Model-to-hardware synthesis for FPGA and ASIC
- 3D CNN acceleration for medical imaging
Key Technologies
- Pipelined parallel neural network implementation
- Pipeline circuit design for neural inference
- Synthetic scaling for shared neural networks
- Custom mass-multiplication circuits