Is this your business? Claim it to manage your IP and profile
The ONNC Calibrator is crafted to optimize AI System-on-Chips by employing post-training quantization (PTQ) techniques to maintain high precision, especially in architectures using fixed-point formats like INT8. By leveraging architecture-aware quantization, it ensures chips retain 99.99% accuracy, offering unparalleled precision control across diverse hardware configurations. This calibrator supports configurable bit-width architectures, allowing the balance of precision and performance to be tailored for various applications. Capable of working with different AI frameworks such as ONNX and PyTorch, the calibrator aligns seamlessly with standard PTQ workflows without needing complex retraining. Its internal AI engine autonomously determines optimal scaling factors, making it an indispensable tool in maintaining model accuracy while reducing computational demand.
Forest Runtime is a robust execution platform for neural network models, providing a retargetable and modular architecture suited for various hardware environments, from data centers to mobile and TinyML applications. It facilitates the seamless execution of compiled models using common C++ APIs along with C and Python bindings, making it versatile for a broad range of AI applications. The runtime supports 'hot batching' technology, allowing models to alter batch sizes and input shapes at runtime, which is essential for modern neural networks like BERT and DLRM. This feature maximizes hardware utilization and minimizes response time by dynamically connecting various system resources efficiently. It also incorporates unique 'bridging' technology that allows resource-sharing among multiple accelerator cards and sessions, thereby supporting scalability and high throughput in server environments.
Designed for AI-on-chips, the ONNC Compiler is a comprehensive bundle of C++ libraries and tools tailored to enhance compiler development for deep learning accelerators. It efficiently transforms neural networks into machine instructions suitable for diverse SoC architectures, from single core systems to more complex layouts with multi-level memory hierarchies. The compiler allows seamless connectivity to leading deep learning frameworks such as PyTorch and TensorFlow. It enables the scaling of deep learning tasks across heterogeneous multicore AI SoCs by utilizing both single backend and multiple backend modes to optimize computing resources. Additionally, it supports intricate features like multiple view address maps, ensuring effective memory allocation and data movement across fragmented memory spaces. Known for performance optimization, the ONNC Compiler employs hardware/software co-optimization techniques to reduce data movement overhead, thereby improving system throughput and efficiency.
Join the world's most advanced semiconductor IP marketplace!
It's free, and you'll get all the tools you need to evaluate IP, download trial versions and datasheets, and manage your evaluation workflow!
To evaluate IP you need to be logged into a buyer profile. Select a profile below, or create a new buyer profile for your company.
