List of Hosted Models

AutoML mobile image classification models (Float Models)

Model Name Paper_Model_Files Model_Size Top-1 Accuracy Top-5 Accuracy TF Lite Performance^
MnasNet_0.50_224 paper, tflite&pb 8.5 Mb 68.03% 87.79% 37 ms
MnasNet_0.75_224 paper, tflite&pb 12 Mb 71.72% 90.17% 61 ms
MnasNet_1.0_96 paper, tflite&pb 17 Mb 62.33% 83.98% 23 ms
MnasNet_1.0_128 paper, tflite&pb 17 Mb 67.32% 87.70% 34 ms
MnasNet_1.0_160 paper, tflite&pb 17 Mb 70.63% 89.58% 51 ms
MnasNet_1.0_192 paper, tflite&pb 17 Mb 72.56% 90.76% 70 ms
MnasNet_1.0_224 paper, tflite&pb 17 Mb 74.08% 91.75% 93 ms
MnasNet_1.3_224 paper, tflite&pb 24 Mb 75.24% 92.55% 152 ms

^ Performance numbers are generated on Pixel-1 using single thread large BIG core.

Image classification (Float Models)

Model Name Paper_Model_Files^ Model_Size Top-1 Accuracy Top-5 Accuracy TF Lite Performance^^ Tensorflow Performance
DenseNet paper, tflite&pb 43.6 Mb 64.2% 85.6% 894 ms 1262 ms
SqueezeNet paper, tflite&pb 5.0 Mb 49.0% 72.9% 224 ms 255 ms
NASNet mobile paper, tflite&pb 21.4 Mb 73.9% 91.5% 261 ms 389 ms
NASNet large paper, tflite&pb 355.3 Mb 82.6% 96.1% 6697 ms 7940 ms
ResNet_V2_101 paper, tflite&pb 178.3 Mb 76.8% 93.6% 1880 ms 1970 ms
Inception_V3 paper, tflite&pb 95.3 Mb 77.9% 93.8% 1433 ms 1522 ms
Inception_V4 paper, tflite&pb 170.7 Mb 80.1% 95.1% 2986 ms 3139 ms
Inception_ResNet_V2 paper, tflite&pb 121.0 Mb 77.5% 94.0% 2731 ms 2926 ms
Mobilenet_V1_0.25_128 paper, tflite&pb 1.9 Mb 41.4% 66.2% 6.2 ms 13.0 ms
Mobilenet_V1_0.25_160 paper, tflite&pb 1.9 Mb 45.4% 70.2% 8.6 ms 19.5 ms
Mobilenet_V1_0.25_192 paper, tflite&pb 1.9 Mb 47.1% 72.0% 12.1 ms 27.8 ms
Mobilenet_V1_0.25_224 paper, tflite&pb 1.9 Mb 49.7% 74.1% 16.2 ms 37.3 ms
Mobilenet_V1_0.50_128 paper, tflite&pb 5.3 Mb 56.2% 79.3% 18.1 ms 29.9 ms
Mobilenet_V1_0.50_160 paper, tflite&pb 5.3 Mb 59.0% 81.8% 26.8 ms 45.9 ms
Mobilenet_V1_0.50_192 paper, tflite&pb 5.3 Mb 61.7% 83.5% 35.6 ms 65.3 ms
Mobilenet_V1_0.50_224 paper, tflite&pb 5.3 Mb 63.2% 84.9% 47.6 ms 164.2 ms
Mobilenet_V1_0.75_128 paper, tflite&pb 10.3 Mb 62.0% 83.8% 34.6 ms 48.7 ms
Mobilenet_V1_0.75_160 paper, tflite&pb 10.3 Mb 65.2% 85.9% 51.3 ms 75.2 ms
Mobilenet_V1_0.75_192 paper, tflite&pb 10.3 Mb 67.1% 87.2% 71.7 ms 107.0 ms
Mobilenet_V1_0.75_224 paper, tflite&pb 10.3 Mb 68.3% 88.1% 95.7 ms 143.4 ms
Mobilenet_V1_1.0_128 paper, tflite&pb 16.9 Mb 65.2% 85.7% 57.4 ms 76.8 ms
Mobilenet_V1_1.0_160 paper, tflite&pb 16.9 Mb 68.0% 87.7% 86.0 ms 117.7 ms
Mobilenet_V1_1.0_192 paper, tflite&pb 16.9 Mb 69.9% 89.1% 118.6 ms 167.3 ms
Mobilenet_V1_1.0_224 paper, tflite&pb 16.9 Mb 71.0% 89.9% 160.1 ms 224.3 ms
Mobilenet_V2_1.0_224 paper, tflite&pb 14.0 Mb 71.8% 90.6% 117 ms

^ The model files include both TF Lite FlatBuffer and Tensorflow frozen Graph.

^^ The performance numbers are generated in the benchmark on Pixel-2 using single thread large core.

^^ Accuracy numbers were computed using the TFLite accuracy tool .

Image classification (Quantized Models)

Model Name Paper_Model_Files Model_Size Top-1 Accuracy Top-5 Accuracy TF Lite Performance
Mobilenet_V1_0.25_128_quant paper, tflite&pb 0.5 Mb 39.5% 64.4% 3.7 ms
Mobilenet_V1_0.25_160_quant paper, tflite&pb 0.5 Mb 42.8% 68.1% 5.5 ms
Mobilenet_V1_0.25_192_quant paper, tflite&pb 0.5 Mb 45.7% 70.8% 7.9 ms
Mobilenet_V1_0.25_224_quant paper, tflite&pb 0.5 Mb 48.2% 72.8% 10.4 ms
Mobilenet_V1_0.50_128_quant paper, tflite&pb 1.4 Mb 54.9% 78.1% 8.8 ms
Mobilenet_V1_0.50_160_quant paper, tflite&pb 1.4 Mb 57.2% 80.5% 13.0 ms
Mobilenet_V1_0.50_192_quant paper, tflite&pb 1.4 Mb 59.9% 82.1% 18.3 ms
Mobilenet_V1_0.50_224_quant paper, tflite&pb 1.4 Mb 61.2% 83.2% 24.7 ms
Mobilenet_V1_0.75_128_quant paper, tflite&pb 2.6 Mb 55.9% 79.1% 16.2 ms
Mobilenet_V1_0.75_160_quant paper, tflite&pb 2.6 Mb 62.4% 83.7% 24.3 ms
Mobilenet_V1_0.75_192_quant paper, tflite&pb 2.6 Mb 66.1% 86.2% 33.8 ms
Mobilenet_V1_0.75_224_quant paper, tflite&pb 2.6 Mb 66.9% 86.9% 45.4 ms
Mobilenet_V1_1.0_128_quant paper, tflite&pb 4.3 Mb 63.3% 84.1% 24.9 ms
Mobilenet_V1_1.0_160_quant paper, tflite&pb 4.3 Mb 66.9% 86.7% 37.4 ms
Mobilenet_V1_1.0_192_quant paper, tflite&pb 4.3 Mb 69.1% 88.1% 51.9 ms
Mobilenet_V1_1.0_224_quant paper, tflite&pb 4.3 Mb 70.0% 89.0% 70.2 ms
Mobilenet_V2_1.0_224_quant paper, tflite&pb 3.4 Mb 70.8% 89.9% 80.3 ms
Inception_V1_quant paper, tflite&pb 6.4 Mb 70.1% 89.8% 154.5 ms
Inception_V2_quant paper, tflite&pb 11 Mb 73.5% 91.4% 235.0 ms
Inception_V3_quant paper,tflite&pb 23 Mb 77.5% 93.7% 637 ms
Inception_V4_quant paper, tflite&pb 41 Mb 79.5% 93.9% 1250.8 ms

Other models

Model TF Lite FlatBuffer

reference, tflite