CN
Providing a Powerful Computing Engine for the Intelligent World, Making the Future Within Reach
Learn More
Shanghai InfiScale Electronic Technology Co., Ltd, established in December 2024, is dedicated to building an Artificial Intelligence (AI) chip software ecosystem and offering versatile, low-cost, and high-performance computing resources. With its foundational software toolchain as its core, the company accelerates the market entry of edge chips and facilitates the low-cost deployment of AI applications. This presentation will comprehensively introduce InfiScale's corporate mission, core positioning, and technological products.
Industrial visual inspection (PCB defect detection), Retail analytics (people counting, product recognition)
Localized face/voice recognition (person detection), Environmental perception (abnormal sound detection: glass breaking, smoke alarm)
Wearable Health Monitoring (Arrhythmia detection) Portable diagnostic devices (supporting real-time assisted diagnosis for ultrasound, X-ray images)
Compiler optimization to achieve vertical integration of
algorithms, system, and chips.
End-to-end adaptation to large model characteristics,
featuring higher throughput and low power consumption.
Seamless improvement of model inference performance.
Large model inference on edge personal
computers and mobile devices.
Reduce model computation costs based on quantization
and sparsity techniques.
Effectively reduce computing power usage costs.
Compiler optimization to achieve vertical integration of
algorithms, system, and chips.
Standardized software foundation across
various chip brands and models.
Seamless improvement of model inference performance.
Supports high-performance inference by reusing code
between data centers and edge devices.
Compatible with models trained on Nvidia and other vendors
for inference-side deployment.
Effectively reduces the complexity of model deployment and
saves on computing power usage costs.
Automated Generation
Performance Guarantee
Cost Advantage
Provides efficient CUDA-compatible operator translation, automatically generating operators that support domestic chips, thereby improving development efficiency.
Ensures that the performance of translated operators meets high standards and the requirements of different chip architectures, guaranteeing application operational efficiency.
Reduces development cycles and costs, helping
domestic chips to be rapidly introduced to the market
at a lower cost
Room 403,No. 50 Boxia Road, Pudong New Area, Shanghai
marketing@infiscale-tech.com
+86 21-18301978266
Shanghai InfiScale Electronic Technology Co., Ltd,
沪ICP备 8112313523