Neural Network Stick (NNS)

a fan less USB stick that designed for Deep Learning inference on various edge application. NNS is powered by high performance, low power Sophon BM1880 chip.

nns top image for mobile

Specifications

Processor Sophon BM1880
Supported AI Framework Caffe, ONNX, Tensorflow, Pytorch
Supported AI Models ResNet50, Yolo V2, GoogleNet V1, MobileNet v1&v2, SSD300, AlexNet, VGG16.etc
H.264 decoder, MJPEG encoder/decoder 1x 1080p @60fps or 2x 1080p @30fps H.264decoder, 75fps for FHD images
Interface USB3.0/USB2.0 slave mode, Type A
Dimensions 93.5*27*15mm
Environmental temperature 0 – 40 C (commercial level)
Hot plugin/plugoff Yes
Certification EMC/FCC/3C/CE
Minimum system requirement X86_64 computer running Ubuntu USB3.0 or USB2.0 port

Appearance

NNS Appearance

What’s inside SophonTM NNS ? It’s BM1880.

The BM1880 TPU provides 1 TOPS peak performance for 8-bit integer operation and provides up to 2 TOPS for Winograd convolution acceleration.

Learn more about BM1880 >>

Toolkit

MNet: Bitmain Compiler which can convert supported AI models to internal format accelerated by Sophon TPU.

ONNC: ONNX Compiler which can convert ONNX format to internal format accelerated by Sophon TPU.

Quantization Tool: Convert FP32 to INT8 and support calibration function.