WebNN Benchmark
Model
Mobilenet v1 (TFLite)
Mobilenet v1 Quant (TFLite)
Mobilenet v2 (TFLite)
Mobilenet v2 Quant (TFLite)
Squeezenet (TFLite)
Inception v3 (TFLite)
Inception v4 (TFLite)
Inception Resnet v2 (TFLite)
Mobilenet v2 (ONNX)
SqueezeNet (ONNX)
ResNet50 v1 (ONNX)
ResNet50 v2 (ONNX)
Inception v2 (ONNX)
DenseNet (ONNX)
SSD MobileNet v1 (TFLite)
SSD MobileNet v1 Quant (TFLite)
SSD MobileNet v2 (TFLite)
SSD MobileNet v2 Quant (TFLite)
SSDLite MobileNet v2 (TFLite)
Tiny Yolo v2 COCO (TFLite)
Tiny Yolo v2 VOC (TFLite)
PoseNet
Deeplab 224 (TFLite)
Deeplab 224 Atrous (TFLite)
Deeplab 257 (TFLite)
Deeplab 257 Atrous (TFLite)
Deeplab 321 (TFLite)
Deeplab 321 Atrous (TFLite)
Deeplab 513 (TFLite)
Deeplab 513 Atrous (TFLite)
Framework and backends
Preference
SUSTAINED_SPEED
FAST_SINGLE_ANSWER
LOW_POWER
None
Iterations
1(warming up) +
Eager Mode
Select All
Uncheck All
ADD
ATROUS_CONV_2D
ATROUS_DEPTHWISE_CONV_2D
AVERAGE_POOL_2D
CONCATENATION
CONV_2D
DEPTHWISE_CONV_2D
FULLY_CONNECTED
MAX_POOL_2D
MUL
RESHAPE
RESIZE_BILINEAR
SOFTMAX
Pick Image
Run
(It will take several minutes)
