By utilizing our unique automation and prediction technologies, we provide optimal DNN models in a short period of time.
The "DNN Acceleration Service" is an engineering service that takes in DNN models, training data, accuracy evaluation metrics, and other materials from customers to provide DNN models that meet the target processing performance and inference accuracy. Please utilize this service to address the trade-off challenges between processing performance and inference accuracy, especially when implementing large-scale, high-precision DNN models on embedded systems with limited specifications.
Inquire About This Product
basic information
● Accelerating DNN models while maintaining accuracy We create numerous modified models from the DNN model entrusted to us by the customer, which are expected to improve processing performance, and we explore the optimal modified model that meets the target processing performance and inference accuracy. Therefore, we can provide DNN models that are accelerated while maintaining accuracy. ● Providing optimal DNN models in a short period Based on the analysis results of the DNN model entrusted to us by the customer, we automatically generate a large number of modified models that are expected to improve processing performance. Additionally, through our unique performance and accuracy predictions, we automatically select modified models that are expected to achieve the target, and we evaluate the processing performance and inference accuracy only for the selected modified models. Therefore, we can provide optimal DNN models in a short period.
Price information
Please contact us.
Delivery Time
※Please contact us.
Applications/Examples of results
Image recognition processing that requires processing performance, such as ADAS applications.
catalog（1）
Download All Catalogs
Company information
We will provide optimal solutions that contribute to our customers' businesses through our "extensive experience and achievements accumulated over many years" and "high technical capabilities" in the fields of embedded and LSI design.