NPU evaluation platform

Neural Processing Unit – Raptor

Evaluating the inference performances of an Edge AI Sensor product is a tedious task.
Our web tool helps you assess what latency and power you can expect with our Raptor NPU accelerator.
Simulate in 3 cliks your key performances for various Raptor configurations.

Pick one of the 4 industry well known MLPerf Tiny benchmark use cases

KeyWord Spotting

Identify keywords eg Alexa in utterances

Person Detection

Detect humans from a camera (video stream)

Image Classification

Separate cats from dogs (from your image folder)

Anomaly Detection

Spot a machine failure before it happens

NPU Configuration

Save and share your simulations results

Generate report

NPU performances for Anomaly Detection

Silicon results in GF 22 FDX

Latency (ms)

Energy /inference, NPU only (µJ)

ACCESS FULL DATA

Energy /inference, System (µJ)

ACCESS FULL DATA

Willing to deploy your Deep Neural Network into tiny sensor devices
for your very own use case ?

Book a 25 minutes call with our technical marketing & try Raptor deployement framework.

Schedule a meeting

—– NEW PRODUCT —–

Customer private space

NPU evaluation platform

Neural Processing Unit – Raptor

Pick one of the 4 industry well known MLPerf Tiny benchmark use cases

KeyWord Spotting

Person Detection

Image Classification

Anomaly Detection

NPU Configuration

NPU performances for Anomaly Detection

Latency (ms)

Energy /inference, NPU only (µJ)

Energy /inference, System (µJ)

Willing to deploy your Deep Neural Network into tiny sensor devices
for your very own use case ?

—– NEW PRODUCT —–

Customer private space

NPU evaluation platform

Neural Processing Unit – Raptor

Pick one of the 4 industry well known MLPerf Tiny benchmark use cases

KeyWord Spotting

Person Detection

Image Classification

Anomaly Detection

NPU Configuration

NPU performances for Anomaly Detection

Latency (ms)

Energy /inference, NPU only (µJ)

Energy /inference, System (µJ)

Willing to deploy your Deep Neural Network into tiny sensor devices for your very own use case ?

Willing to deploy your Deep Neural Network into tiny sensor devices
for your very own use case ?