This repository contains several classification models on MXNet/Gluon, PyTorch, Chainer, and Keras, with scripts for training/validating/converting models. All models are designed for using with ImageNet-1k dataset.
To use only Gluon models in your project, simply install the gluoncv2 package with mxnet:
pip install gluoncv2 mxnet>=1.2.1
To enable different hardware supports such as GPUs, check out MXNet variants. For example, you can install with CUDA-9.2 supported MXNet:
pip install gluoncv2 mxnet-cu92>=1.2.1
To use only PyTorch models in your project, simply install the pytorchcv package with torch (>=0.4.1 is recommended):
pip install pytorchcv torch>=0.4.0
To enable/disable different hardware supports such as GPUs, check out PyTorch installation instructions.
To use only Chainer models in your project, simply install the chainercv2 package:
pip install chainercv2
To use only Keras models in your project, simply install the kerascv package with mxnet:
pip install kerascv mxnet>=1.2.1
To enable different hardware supports such as GPUs, check out MXNet variants. For example, you can install with CUDA-9.2 supported MXNet:
pip install kerascv mxnet-cu92>=1.2.1
After installation change the value of the field image_data_format to channels_first in the file ~/.keras/keras.json.
To use the repository for training/validation/converting models:
git clone git@github.com:osmr/imgclsmob.git
pip install -r requirements.txt
Example of using the pretrained ResNet-18 model on Gluon:
from gluoncv2.model_provider import get_model as glcv2_get_model
import mxnet as mx
net = glcv2_get_model("resnet18", pretrained=True)
x = mx.nd.zeros((1, 3, 224, 224), ctx=mx.cpu())
y = net(x)
Example of using the pretrained ResNet-18 model on PyTorch:
from pytorchcv.model_provider import get_model as ptcv_get_model
import torch
from torch.autograd import Variable
net = ptcv_get_model("resnet18", pretrained=True)
x = Variable(torch.randn(1, 3, 224, 224))
y = net(x)
Example of using the pretrained ResNet-18 model on Chainer:
from chainercv2.model_provider import get_model as chcv2_get_model
import numpy as np
net = chcv2_get_model("resnet18", pretrained=True)
x = np.zeros((1, 3, 224, 224), np.float32)
y = net(x)
Example of using the pretrained ResNet-18 model on Keras:
from kerascv.model_provider import get_model as kecv_get_model
import numpy as np
net = kecv_get_model("resnet18", pretrained=True)
x = np.zeros((1, 3, 224, 224), np.float32)
y = net.predict(x)
- ResNet ('Deep Residual Learning for Image Recognition')
- PreResNet ('Identity Mappings in Deep Residual Networks')
- ResNeXt ('Aggregated Residual Transformations for Deep Neural Networks')
- SENet/SE-ResNet/SE-PreResNet/SE-ResNeXt ('Squeeze-and-Excitation Networks')
- DenseNet ('Densely Connected Convolutional Networks')
- CondenseNet ('CondenseNet: An Efficient DenseNet using Learned Group Convolutions')
- DPN ('Dual Path Networks')
- DarkNet ('Darknet: Open source neural networks in c')
- SqueezeNet ('SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and <0.5MB model size')
- SqueezeNext ('SqueezeNext: Hardware-Aware Neural Network Design')
- ShuffleNet ('ShuffleNet: An Extremely Efficient Convolutional Neural Network for Mobile Devices')
- ShuffleNetV2 ('ShuffleNet V2: Practical Guidelines for Efficient CNN Architecture Design')
- MENet ('Merging and Evolution: Improving Convolutional Neural Networks for Mobile Applications')
- MobileNet ('MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications')
- FD-MobileNet ('FD-MobileNet: Improved MobileNet with A Fast Downsampling Strategy')
- MobileNetV2 ('MobileNetV2: Inverted Residuals and Linear Bottlenecks')
- NASNet-A-Mobile ('Learning Transferable Architectures for Scalable Image Recognition')
Some remarks:
- All pretrained models can be downloaded automatically during use (use the parameter
pretrained). - Top1/Top5 are the standard 1-crop Top-1/Top-5 errors (in percents) on the validation subset of the ImageNet1k dataset.
- ResNet/PreResNet with b-suffix is a version of the networks with the stride in the second convolution of the bottleneck block. Respectively a network without b-suffix has the stride in the first convolution.
- ResNet/PreResNet models do not use biasses in convolutions at all.
- CondenseNet models are only so-called converted versions.
- All models have an input 224x224 with ordinary normalization.
| Model | Top1 | Top5 | Params | FLOPs | Remarks |
|---|---|---|---|---|---|
| ResNet-10 | 37.09 | 15.55 | 5,418,792 | 892.62M | Training (log) |
| ResNet-12 | 35.86 | 14.46 | 5,492,776 | 1,124.23M | Training (log) |
| ResNet-14 | 32.85 | 12.41 | 5,788,200 | 1,355.64M | Training (log) |
| ResNet-16 | 30.68 | 11.10 | 6,968,872 | 1,586.95M | Training (log) |
| ResNet-18 x0.25 | 49.16 | 24.45 | 831,096 | 136.64M | Training (log) |
| ResNet-18 x0.5 | 36.54 | 14.96 | 3,055,880 | 485.22M | Training (log) |
| ResNet-18 x0.75 | 33.25 | 12.54 | 6,675,352 | 1,045.75M | Training (log) |
| ResNet-18 | 29.13 | 9.94 | 11,689,512 | 1,818.21M | Training (log) |
| ResNet-34 | 25.34 | 7.92 | 21,797,672 | 3,669.16M | Converted from Gluon Model Zoo (log) |
| ResNet-50 | 23.50 | 6.87 | 25,557,032 | 3,868.96M | Converted from Gluon Model Zoo (log) |
| ResNet-50b | 22.92 | 6.44 | 25,557,032 | 4,100.70M | Converted from Gluon Model Zoo (log) |
| ResNet-101 | 21.66 | 5.99 | 44,549,160 | 7,586.30M | Converted from Gluon Model Zoo (log) |
| ResNet-101b | 21.18 | 5.60 | 44,549,160 | 7,818.04M | Converted from Gluon Model Zoo (log) |
| ResNet-152 | 21.01 | 5.61 | 60,192,808 | 11,304.85M | Converted from Gluon Model Zoo (log) |
| ResNet-152b | 20.54 | 5.37 | 60,192,808 | 11,536.58M | Converted from Gluon Model Zoo (log) |
| PreResNet-18 | 28.72 | 9.88 | 11,687,848 | 1,818.41M | Training (log) |
| PreResNet-34 | 25.88 | 8.11 | 21,796,008 | 3,669.36M | Converted from Gluon Model Zoo (log) |
| PreResNet-50 | 23.39 | 6.68 | 25,549,480 | 3,869.16M | Converted from Gluon Model Zoo (log) |
| PreResNet-50b | 23.16 | 6.64 | 25,549,480 | 4,100.90M | Converted from Gluon Model Zoo (log) |
| PreResNet-101 | 21.45 | 5.75 | 44,541,608 | 7,586.50M | Converted from Gluon Model Zoo (log) |
| PreResNet-101b | 21.73 | 5.88 | 44,541,608 | 7,818.24M | Converted from Gluon Model Zoo (log) |
| PreResNet-152 | 20.70 | 5.32 | 60,185,256 | 11,305.05M | Converted from Gluon Model Zoo (log) |
| PreResNet-152b | 21.00 | 5.75 | 60,185,256 | 11,536.78M | Converted from Gluon Model Zoo (log) |
| PreResNet-200b | 21.10 | 5.64 | 64,666,280 | 15,040.27M | From tornadomeet/ResNet (log) |
| ResNeXt-101 (32x4d) | 21.32 | 5.79 | 44,177,704 | 7,991.62M | From Cadene/pretrained...pytorch (log) |
| ResNeXt-101 (64x4d) | 20.60 | 5.41 | 83,455,272 | 15,491.88M | From Cadene/pretrained...pytorch (log) |
| SE-ResNet-50 | 22.51 | 6.44 | 28,088,024 | 3,877.01M | From Cadene/pretrained...pytorch (log) |
| SE-ResNet-101 | 21.92 | 5.89 | 49,326,872 | 7,600.01M | From Cadene/pretrained...pytorch (log) |
| SE-ResNet-152 | 21.48 | 5.77 | 66,821,848 | 11,324.62M | From Cadene/pretrained...pytorch (log) |
| SE-ResNeXt-50 (32x4d) | 21.06 | 5.58 | 27,559,896 | 4,253.33M | From Cadene/pretrained...pytorch (log) |
| SE-ResNeXt-101 (32x4d) | 19.99 | 5.00 | 48,955,416 | 8,005.33M | From Cadene/pretrained...pytorch (log) |
| SENet-154 | 18.84 | 4.65 | 115,088,984 | 20,742.40M | From Cadene/pretrained...pytorch (log) |
| DenseNet-121 | 25.11 | 7.80 | 7,978,856 | 2,852.39M | Converted from Gluon Model Zoo (log) |
| DenseNet-161 | 22.40 | 6.18 | 28,681,000 | 7,761.25M | Converted from Gluon Model Zoo (log) |
| DenseNet-169 | 23.89 | 6.89 | 14,149,480 | 3,381.48M | Converted from Gluon Model Zoo (log) |
| DenseNet-201 | 22.71 | 6.36 | 20,013,928 | 4,318.75M | Converted from Gluon Model Zoo (log) |
| CondenseNet-74 (C=G=4) | 26.82 | 8.64 | 4,773,944 | 533.64M | From ShichenLiu/CondenseNet (log) |
| CondenseNet-74 (C=G=8) | 29.76 | 10.49 | 2,935,416 | 278.55M | From ShichenLiu/CondenseNet (log) |
| DPN-68 | 23.57 | 7.00 | 12,611,602 | 2,338.71M | From Cadene/pretrained...pytorch (log) |
| DPN-98 | 20.23 | 5.28 | 61,570,728 | 11,702.80M | From Cadene/pretrained...pytorch (log) |
| DPN-131 | 20.03 | 5.22 | 79,254,504 | 16,056.22M | From Cadene/pretrained...pytorch (log) |
| DarkNet Tiny | 43.36 | 19.46 | 1,042,104 | 496.34M | Training (log) |
| DarkNet Ref | 38.00 | 16.68 | 7,319,416 | 365.55M | Training (log) |
| SqueezeNet v1.0 | 40.97 | 18.96 | 1,248,424 | 828.30M | Training (log) |
| SqueezeNet v1.1 | 41.37 | 19.20 | 1,235,496 | 354.88M | Training (log) |
| ShuffleNetV2 x0.5 | 40.98 | 18.53 | 1,366,792 | 42.34M | Training (log) |
| ShuffleNetV2 x1.0 | 33.97 | 13.41 | 2,278,604 | 147.92M | Training (log) |
| ShuffleNetV2 x1.5 | 32.38 | 12.37 | 4,406,098 | 318.61M | Training (log) |
| 108-MENet-8x1 (g=3) | 46.11 | 22.37 | 654,516 | 40.64M | From clavichord93/MENet (log) |
| 128-MENet-8x1 (g=4) | 45.80 | 21.93 | 750,796 | 43.58M | From clavichord93/MENet (log) |
| 228-MENet-12x1 (g=3) | 35.03 | 13.99 | 1,806,568 | 148.93M | From clavichord93/MENet (log) |
| 256-MENet-12x1 (g=4) | 34.49 | 13.90 | 1,888,240 | 146.11M | From clavichord93/MENet (log) |
| 348-MENet-12x1 (g=3) | 31.17 | 11.41 | 3,368,128 | 306.31M | From clavichord93/MENet (log) |
| 352-MENet-12x1 (g=8) | 34.70 | 13.75 | 2,272,872 | 151.03M | From clavichord93/MENet (log) |
| 456-MENet-24x1 (g=3) | 29.57 | 10.43 | 5,304,784 | 560.72M | From clavichord93/MENet (log) |
| MobileNet x0.25 | 45.78 | 22.18 | 470,072 | 42.30M | Training (log) |
| MobileNet x0.5 | 36.12 | 14.81 | 1,331,592 | 152.04M | Training (log) |
| MobileNet x0.75 | 32.71 | 12.28 | 2,585,560 | 329.22M | Converted from Gluon Model Zoo (log) |
| MobileNet x1.0 | 29.25 | 10.03 | 4,231,976 | 573.83M | Converted from Gluon Model Zoo (log) |
| FD-MobileNet x0.25 | 56.73 | 31.99 | 383,160 | 12.44M | From clavichord93/FD-MobileNet (log) |
| FD-MobileNet x0.5 | 44.66 | 21.08 | 993,928 | 40.93M | From clavichord93/FD-MobileNet (log) |
| FD-MobileNet x1.0 | 35.95 | 14.72 | 2,901,288 | 146.08M | From clavichord93/FD-MobileNet (log) |
| MobileNetV2 x0.25 | 48.89 | 25.24 | 1,516,392 | 32.22M | Converted from Gluon Model Zoo (log) |
| MobileNetV2 x0.5 | 35.51 | 14.64 | 1,964,736 | 95.62M | Converted from Gluon Model Zoo (log) |
| MobileNetV2 x0.75 | 30.82 | 11.26 | 2,627,592 | 191.61M | Converted from Gluon Model Zoo (log) |
| MobileNetV2 x1.0 | 28.51 | 9.90 | 3,504,960 | 320.19M | Converted from Gluon Model Zoo (log) |
| NASNet-A-Mobile | 25.37 | 7.95 | 5,289,978 | 587.29M | From Cadene/pretrained...pytorch (log) |
| Model | Top1 | Top5 | Params | FLOPs | Remarks |
|---|---|---|---|---|---|
| ResNet-10 | 37.46 | 15.85 | 5,418,792 | 892.62M | Converted from GL model (log) |
| ResNet-12 | 36.18 | 14.80 | 5,492,776 | 1,124.23M | Converted from GL model (log) |
| ResNet-14 | 33.17 | 12.71 | 5,788,200 | 1,355.64M | Converted from GL model (log) |
| ResNet-16 | 30.90 | 11.38 | 6,968,872 | 1,586.95M | Converted from GL model (log) |
| ResNet-18 x0.25 | 49.50 | 24.83 | 831,096 | 136.64M | Converted from GL model (log) |
| ResNet-18 x0.5 | 37.04 | 15.38 | 3,055,880 | 485.22M | Converted from GL model (log) |
| ResNet-18 x0.75 | 33.61 | 12.85 | 6,675,352 | 1,045.75M | Converted from GL model (log) |
| ResNet-18 | 29.52 | 10.21 | 11,689,512 | 1,818.21M | Converted from GL model (log) |
| ResNet-34 | 25.66 | 8.18 | 21,797,672 | 3,669.16M | Converted from Gluon Model Zoo (log) |
| ResNet-50 | 23.79 | 7.05 | 25,557,032 | 3,868.96M | Converted from Gluon Model Zoo (log) |
| ResNet-50b | 23.05 | 6.65 | 25,557,032 | 4,100.70M | Converted from Gluon Model Zoo (log) |
| ResNet-101 | 21.90 | 6.22 | 44,549,160 | 7,586.30M | Converted from Gluon Model Zoo (log) |
| ResNet-101b | 21.45 | 5.81 | 44,549,160 | 7,818.04M | Converted from Gluon Model Zoo (log) |
| ResNet-152 | 21.26 | 5.82 | 60,192,808 | 11,304.85M | Converted from Gluon Model Zoo (log) |
| ResNet-152b | 20.74 | 5.50 | 60,192,808 | 11,536.58M | Converted from Gluon Model Zoo (log) |
| PreResNet-18 | 29.09 | 10.18 | 11,687,848 | 1,818.41M | Converted from GL model (log) |
| PreResNet-34 | 26.23 | 8.41 | 21,796,008 | 3,669.36M | Converted from Gluon Model Zoo (log) |
| PreResNet-50 | 23.70 | 6.85 | 25,549,480 | 3,869.16M | Converted from Gluon Model Zoo (log) |
| PreResNet-50b | 23.33 | 6.87 | 25,549,480 | 4,100.90M | Converted from Gluon Model Zoo (log) |
| PreResNet-101 | 21.74 | 5.91 | 44,541,608 | 7,586.50M | Converted from Gluon Model Zoo (log) |
| PreResNet-101b | 21.95 | 6.03 | 44,541,608 | 7,818.24M | Converted from Gluon Model Zoo (log) |
| PreResNet-152 | 20.94 | 5.55 | 60,185,256 | 11,305.05M | Converted from Gluon Model Zoo (log) |
| PreResNet-152b | 21.34 | 5.91 | 60,185,256 | 11,536.78M | Converted from Gluon Model Zoo (log) |
| PreResNet-200b | 21.33 | 5.88 | 64,666,280 | 15,040.27M | From tornadomeet/ResNet (log) |
| ResNeXt-101 (32x4d) | 21.81 | 6.11 | 44,177,704 | 7,991.62M | From Cadene/pretrained...pytorch (log) |
| ResNeXt-101 (64x4d) | 21.04 | 5.75 | 83,455,272 | 15,491.88M | From Cadene/pretrained...pytorch (log) |
| SE-ResNet-50 | 22.47 | 6.40 | 28,088,024 | 3,877.01M | From Cadene/pretrained...pytorch (log) |
| SE-ResNet-101 | 21.88 | 5.89 | 49,326,872 | 7,600.01M | From Cadene/pretrained...pytorch (log) |
| SE-ResNet-152 | 21.48 | 5.76 | 66,821,848 | 11,324.62M | From Cadene/pretrained...pytorch (log) |
| SE-ResNeXt-50 (32x4d) | 21.00 | 5.54 | 27,559,896 | 4,253.33M | From Cadene/pretrained...pytorch (log) |
| SE-ResNeXt-101 (32x4d) | 19.96 | 5.05 | 48,955,416 | 8,005.33M | From Cadene/pretrained...pytorch (log) |
| SENet-154 | 18.62 | 4.61 | 115,088,984 | 20,742.40M | From Cadene/pretrained...pytorch (log) |
| DenseNet-121 | 25.57 | 8.03 | 7,978,856 | 2,852.39M | Converted from Gluon Model Zoo (log) |
| DenseNet-161 | 22.86 | 6.44 | 28,681,000 | 7,761.25M | Converted from Gluon Model Zoo (log) |
| DenseNet-169 | 24.40 | 7.19 | 14,149,480 | 3,381.48M | Converted from Gluon Model Zoo (log) |
| DenseNet-201 | 23.10 | 6.63 | 20,013,928 | 4,318.75M | Converted from Gluon Model Zoo (log) |
| CondenseNet-74 (C=G=4) | 26.25 | 8.28 | 4,773,944 | 533.64M | From ShichenLiu/CondenseNet (log) |
| CondenseNet-74 (C=G=8) | 28.93 | 10.06 | 2,935,416 | 278.55M | From ShichenLiu/CondenseNet (log) |
| DPN-68 | 24.17 | 7.27 | 12,611,602 | 2,338.71M | From Cadene/pretrained...pytorch (log) |
| DPN-98 | 20.81 | 5.53 | 61,570,728 | 11,702.80M | From Cadene/pretrained...pytorch (log) |
| DPN-131 | 20.54 | 5.48 | 79,254,504 | 16,056.22M | From Cadene/pretrained...pytorch (log) |
| DarkNet Tiny | 43.65 | 19.80 | 1,042,104 | 496.34M | Converted from GL model (log) |
| DarkNet Ref | 38.58 | 17.18 | 7,319,416 | 365.55M | Converted from GL model (log) |
| SqueezeNet v1.0 | 41.31 | 19.32 | 1,248,424 | 828.30M | Converted from GL model (log) |
| SqueezeNet v1.1 | 41.82 | 19.38 | 1,235,496 | 354.88M | Converted from TorchVision (log) |
| ShuffleNetV2 x0.5 | 41.48 | 19.02 | 1,366,792 | 42.34M | Converted from GL model (log) |
| ShuffleNetV2 x1.0 | 34.39 | 13.78 | 2,278,604 | 147.92M | Converted from GL model (log) |
| ShuffleNetV2 x1.5 | 32.82 | 12.69 | 4,406,098 | 318.61M | Converted from GL model (log) |
| 108-MENet-8x1 (g=3) | 43.92 | 20.76 | 654,516 | 40.64M | From clavichord93/MENet (log) |
| 128-MENet-8x1 (g=4) | 43.95 | 20.62 | 750,796 | 43.58M | From clavichord93/MENet (log) |
| 228-MENet-12x1 (g=3) | 33.57 | 13.28 | 1,806,568 | 148.93M | From clavichord93/MENet (log) |
| 256-MENet-12x1 (g=4) | 33.41 | 13.26 | 1,888,240 | 146.11M | From clavichord93/MENet (log) |
| 348-MENet-12x1 (g=3) | 30.10 | 10.92 | 3,368,128 | 306.31M | From clavichord93/MENet (log) |
| 352-MENet-12x1 (g=8) | 33.31 | 13.08 | 2,272,872 | 151.03M | From clavichord93/MENet (log) |
| 456-MENet-24x1 (g=3) | 28.40 | 9.93 | 5,304,784 | 560.72M | From clavichord93/MENet (log) |
| MobileNet x0.25 | 46.26 | 22.49 | 470,072 | 42.30M | Converted from GL model (log) |
| MobileNet x0.5 | 36.30 | 15.14 | 1,331,592 | 152.04M | Converted from GL model (log) |
| MobileNet x0.75 | 33.54 | 12.85 | 2,585,560 | 329.22M | Converted from Gluon Model Zoo (log) |
| MobileNet x1.0 | 29.86 | 10.36 | 4,231,976 | 573.83M | Converted from Gluon Model Zoo (log) |
| FD-MobileNet x0.25 | 55.77 | 31.32 | 383,160 | 12.44M | From clavichord93/FD-MobileNet (log) |
| FD-MobileNet x0.5 | 43.85 | 20.72 | 993,928 | 40.93M | From clavichord93/FD-MobileNet (log) |
| FD-MobileNet x1.0 | 34.70 | 14.05 | 2,901,288 | 146.08M | From clavichord93/FD-MobileNet (log) |
| MobileNetV2 x0.25 | 49.72 | 25.87 | 1,516,392 | 32.22M | Converted from Gluon Model Zoo (log) |
| MobileNetV2 x0.5 | 36.54 | 15.19 | 1,964,736 | 95.62M | Converted from Gluon Model Zoo (log) |
| MobileNetV2 x0.75 | 31.89 | 11.76 | 2,627,592 | 191.61M | Converted from Gluon Model Zoo (log) |
| MobileNetV2 x1.0 | 29.31 | 10.39 | 3,504,960 | 320.19M | Converted from Gluon Model Zoo (log) |
| NASNet-A-Mobile | 25.68 | 8.16 | 5,289,978 | 587.29M | From Cadene/pretrained...pytorch (log) |
| Model | Top1 | Top5 | Params | FLOPs | Remarks |
|---|---|---|---|---|---|
| ResNet-10 | 37.12 | 15.49 | 5,418,792 | 892.62M | Converted from GL model (log) |
| ResNet-12 | 35.86 | 14.48 | 5,492,776 | 1,124.23M | Converted from GL model (log) |
| ResNet-14 | 32.84 | 12.42 | 5,788,200 | 1,355.64M | Converted from GL model (log) |
| ResNet-16 | 30.66 | 11.07 | 6,968,872 | 1,586.95M | Converted from GL model (log) |
| ResNet-18 x0.25 | 49.08 | 24.48 | 831,096 | 136.64M | Converted from GL model (log) |
| ResNet-18 x0.5 | 36.55 | 14.99 | 3,055,880 | 485.22M | Converted from GL model (log) |
| ResNet-18 x0.75 | 33.27 | 12.56 | 6,675,352 | 1,045.75M | Converted from GL model (log) |
| ResNet-18 | 29.08 | 9.97 | 11,689,512 | 1,818.21M | Converted from GL model (log) |
| ResNet-34 | 25.35 | 7.95 | 21,797,672 | 3,669.16M | Converted from Gluon Model Zoo (log) |
| ResNet-50 | 23.50 | 6.83 | 25,557,032 | 3,868.96M | Converted from Gluon Model Zoo (log) |
| ResNet-50b | 22.93 | 6.46 | 25,557,032 | 4,100.70M | Converted from Gluon Model Zoo (log) |
| ResNet-101 | 21.65 | 6.01 | 44,549,160 | 7,586.30M | Converted from Gluon Model Zoo (log) |
| ResNet-101b | 21.16 | 5.59 | 44,549,160 | 7,818.04M | Converted from Gluon Model Zoo (log) |
| ResNet-152 | 21.07 | 5.67 | 60,192,808 | 11,304.85M | Converted from Gluon Model Zoo (log) |
| ResNet-152b | 20.44 | 5.39 | 60,192,808 | 11,536.58M | Converted from Gluon Model Zoo (log) |
| PreResNet-18 | 28.66 | 9.92 | 11,687,848 | 1,818.41M | Converted from GL model (log) |
| PreResNet-34 | 25.89 | 8.12 | 21,796,008 | 3,669.36M | Converted from Gluon Model Zoo (log) |
| PreResNet-50 | 23.36 | 6.69 | 25,549,480 | 3,869.16M | Converted from Gluon Model Zoo (log) |
| PreResNet-50b | 23.08 | 6.67 | 25,549,480 | 4,100.90M | Converted from Gluon Model Zoo (log) |
| PreResNet-101 | 21.45 | 5.75 | 44,541,608 | 7,586.50M | Converted from Gluon Model Zoo (log) |
| PreResNet-101b | 21.61 | 5.87 | 44,541,608 | 7,818.24M | Converted from Gluon Model Zoo (log) |
| PreResNet-152 | 20.73 | 5.30 | 60,185,256 | 11,305.05M | Converted from Gluon Model Zoo (log) |
| PreResNet-152b | 20.88 | 5.66 | 60,185,256 | 11,536.78M | Converted from Gluon Model Zoo (log) |
| PreResNet-200b | 21.03 | 5.60 | 64,666,280 | 15,040.27M | From tornadomeet/ResNet (log) |
| ResNeXt-101 (32x4d) | 21.11 | 5.69 | 44,177,704 | 7,991.62M | From Cadene/pretrained...pytorch (log) |
| ResNeXt-101 (64x4d) | 20.57 | 5.43 | 83,455,272 | 15,491.88M | From Cadene/pretrained...pytorch (log) |
| SE-ResNet-50 | 22.53 | 6.41 | 28,088,024 | 3,877.01M | From Cadene/pretrained...pytorch (log) |
| SE-ResNet-101 | 21.90 | 5.88 | 49,326,872 | 7,600.01M | From Cadene/pretrained...pytorch (log) |
| SE-ResNet-152 | 21.46 | 5.77 | 66,821,848 | 11,324.62M | From Cadene/pretrained...pytorch (log) |
| SE-ResNeXt-50 (32x4d) | 21.04 | 5.58 | 27,559,896 | 4,253.33M | From Cadene/pretrained...pytorch (log) |
| SE-ResNeXt-101 (32x4d) | 19.99 | 5.01 | 48,955,416 | 8,005.33M | From Cadene/pretrained...pytorch (log) |
| SENet-154 | 18.79 | 4.63 | 115,088,984 | 20,742.40M | From Cadene/pretrained...pytorch (log) |
| DenseNet-121 | 25.04 | 7.79 | 7,978,856 | 2,852.39M | Converted from Gluon Model Zoo (log) |
| DenseNet-161 | 22.36 | 6.20 | 28,681,000 | 7,761.25M | Converted from Gluon Model Zoo (log) |
| DenseNet-169 | 23.85 | 6.86 | 14,149,480 | 3,381.48M | Converted from Gluon Model Zoo (log) |
| DenseNet-201 | 22.64 | 6.29 | 20,013,928 | 4,318.75M | Converted from Gluon Model Zoo (log) |
| CondenseNet-74 (C=G=4) | 26.81 | 8.61 | 4,773,944 | 533.64M | From ShichenLiu/CondenseNet (log) |
| CondenseNet-74 (C=G=8) | 29.74 | 10.43 | 2,935,416 | 278.55M | From ShichenLiu/CondenseNet (log) |
| DPN-68 | 23.61 | 7.01 | 12,611,602 | 2,338.71M | From Cadene/pretrained...pytorch (log) |
| DPN-98 | 20.80 | 5.53 | 61,570,728 | 11,702.80M | From Cadene/pretrained...pytorch (log) |
| DPN-131 | 20.04 | 5.23 | 79,254,504 | 16,056.22M | From Cadene/pretrained...pytorch (log) |
| DarkNet Tiny | 43.31 | 19.47 | 1,042,104 | 496.34M | Converted from GL model (log) |
| DarkNet Ref | 38.09 | 16.71 | 7,319,416 | 365.55M | Converted from GL model (log) |
| SqueezeNet v1.0 | 41.01 | 18.96 | 1,248,424 | 828.30M | Converted from GL model (log) |
| SqueezeNet v1.1 | 41.36 | 19.25 | 1,235,496 | 354.88M | Converted from GL model (log) |
| ShuffleNetV2 x0.5 | 43.80 | 20.87 | 1,366,792 | 42.34M | Converted from GL model (log) |
| ShuffleNetV2 x1.0 | 36.48 | 15.19 | 2,278,604 | 147.92M | Converted from GL model (log) |
| ShuffleNetV2 x1.5 | 33.96 | 13.37 | 4,406,098 | 318.61M | Converted from GL model (log) |
| 108-MENet-8x1 (g=3) | 46.07 | 22.42 | 654,516 | 40.64M | From clavichord93/MENet (log) |
| 128-MENet-8x1 (g=4) | 45.82 | 21.91 | 750,796 | 43.58M | From clavichord93/MENet (log) |
| 228-MENet-12x1 (g=3) | 34.93 | 14.01 | 1,806,568 | 148.93M | From clavichord93/MENet (log) |
| 256-MENet-12x1 (g=4) | 34.44 | 13.91 | 1,888,240 | 146.11M | From clavichord93/MENet (log) |
| 348-MENet-12x1 (g=3) | 31.14 | 11.40 | 3,368,128 | 306.31M | From clavichord93/MENet (log) |
| 352-MENet-12x1 (g=8) | 34.62 | 13.68 | 2,272,872 | 151.03M | From clavichord93/MENet (log) |
| 456-MENet-24x1 (g=3) | 29.55 | 10.39 | 5,304,784 | 560.72M | From clavichord93/MENet (log) |
| MobileNet x0.25 | 45.85 | 22.16 | 470,072 | 42.30M | Converted from GL model (log) |
| MobileNet x0.5 | 36.15 | 14.86 | 1,331,592 | 152.04M | Converted from GL model (log) |
| MobileNet x0.75 | 33.24 | 12.52 | 2,585,560 | 329.22M | Converted from Gluon Model Zoo (log) |
| MobileNet x1.0 | 29.71 | 10.31 | 4,231,976 | 573.83M | Converted from Gluon Model Zoo (log) |
| FD-MobileNet x0.25 | 56.67 | 31.96 | 383,160 | 12.44M | From clavichord93/FD-MobileNet (log) |
| FD-MobileNet x0.5 | 44.67 | 21.09 | 993,928 | 40.93M | From clavichord93/FD-MobileNet (log) |
| FD-MobileNet x1.0 | 35.94 | 14.70 | 2,901,288 | 146.08M | From clavichord93/FD-MobileNet (log) |
| MobileNetV2 x0.25 | 49.11 | 25.49 | 1,516,392 | 32.22M | Converted from Gluon Model Zoo (log) |
| MobileNetV2 x0.5 | 35.96 | 14.98 | 1,964,736 | 95.62M | Converted from Gluon Model Zoo (log) |
| MobileNetV2 x0.75 | 31.28 | 11.48 | 2,627,592 | 191.61M | Converted from Gluon Model Zoo (log) |
| MobileNetV2 x1.0 | 28.87 | 10.05 | 3,504,960 | 320.19M | Converted from Gluon Model Zoo (log) |
| NASNet-A-Mobile | 25.78 | 8.32 | 5,289,978 | 587.29M | From Cadene/pretrained...pytorch (log) |
| Model | Top1 | Top5 | Params | FLOPs | Remarks |
|---|---|---|---|---|---|
| ResNet-10 | 37.09 | 15.54 | 5,418,792 | 892.62M | Converted from GL model (log) |
| ResNet-12 | 35.86 | 14.45 | 5,492,776 | 1,124.23M | Converted from GL model (log) |
| ResNet-14 | 32.85 | 12.42 | 5,788,200 | 1,355.64M | Converted from GL model (log) |
| ResNet-16 | 30.67 | 11.09 | 6,968,872 | 1,586.95M | Converted from GL model (log) |
| ResNet-18 x0.25 | 49.14 | 24.45 | 831,096 | 136.64M | Converted from GL model (log) |
| ResNet-18 x0.5 | 36.54 | 14.96 | 3,055,880 | 485.22M | Converted from GL model (log) |
| ResNet-18 x0.75 | 33.24 | 12.54 | 6,675,352 | 1,045.75M | Converted from GL model (log) |
| ResNet-18 | 29.13 | 9.94 | 11,689,512 | 1,818.21M | Converted from GL model (log) |
| ResNet-34 | 25.32 | 7.92 | 21,797,672 | 3,669.16M | Converted from Gluon Model Zoo (log) |
| ResNet-50 | 23.49 | 6.87 | 25,557,032 | 3,868.96M | Converted from Gluon Model Zoo (log) |
| ResNet-50b | 22.90 | 6.44 | 25,557,032 | 4,100.70M | Converted from Gluon Model Zoo (log) |
| ResNet-101 | 21.64 | 5.99 | 44,549,160 | 7,586.30M | Converted from Gluon Model Zoo (log) |
| ResNet-101b | 21.17 | 5.60 | 44,549,160 | 7,818.04M | Converted from Gluon Model Zoo (log) |
| ResNet-152 | 21.00 | 5.61 | 60,192,808 | 11,304.85M | Converted from Gluon Model Zoo (log) |
| ResNet-152b | 20.53 | 5.37 | 60,192,808 | 11,536.58M | Converted from Gluon Model Zoo (log) |
| PreResNet-18 | 28.72 | 9.88 | 11,687,848 | 1,818.41M | Converted from GL model (log) |
| PreResNet-34 | 25.86 | 8.11 | 21,796,008 | 3,669.36M | Converted from Gluon Model Zoo (log) |
| PreResNet-50 | 23.38 | 6.68 | 25,549,480 | 3,869.16M | Converted from Gluon Model Zoo (log) |
| PreResNet-50b | 23.14 | 6.63 | 25,549,480 | 4,100.90M | Converted from Gluon Model Zoo (log) |
| PreResNet-101 | 21.43 | 5.75 | 44,541,608 | 7,586.50M | Converted from Gluon Model Zoo (log) |
| PreResNet-101b | 21.71 | 5.88 | 44,541,608 | 7,818.24M | Converted from Gluon Model Zoo (log) |
| PreResNet-152 | 20.69 | 5.31 | 60,185,256 | 11,305.05M | Converted from Gluon Model Zoo (log) |
| PreResNet-152b | 20.99 | 5.76 | 60,185,256 | 11,536.78M | Converted from Gluon Model Zoo (log) |
| PreResNet-200b | 21.09 | 5.64 | 64,666,280 | 15,040.27M | From tornadomeet/ResNet (log) |
| ResNeXt-101 (32x4d) | 21.30 | 5.78 | 44,177,704 | 7,991.62M | From Cadene/pretrained...pytorch (log) |
| ResNeXt-101 (64x4d) | 20.59 | 5.41 | 83,455,272 | 15,491.88M | From Cadene/pretrained...pytorch (log) |
| SE-ResNet-50 | 22.50 | 6.43 | 28,088,024 | 3,877.01M | From Cadene/pretrained...pytorch (log) |
| SE-ResNet-101 | 21.92 | 5.88 | 49,326,872 | 7,600.01M | From Cadene/pretrained...pytorch (log) |
| SE-ResNet-152 | 21.46 | 5.77 | 66,821,848 | 11,324.62M | From Cadene/pretrained...pytorch (log) |
| SE-ResNeXt-50 (32x4d) | 21.05 | 5.57 | 27,559,896 | 4,253.33M | From Cadene/pretrained...pytorch (log) |
| SE-ResNeXt-101 (32x4d) | 19.98 | 4.99 | 48,955,416 | 8,005.33M | From Cadene/pretrained...pytorch (log) |
| SENet-154 | 18.83 | 4.65 | 115,088,984 | 20,742.40M | From Cadene/pretrained...pytorch (log) |
| DenseNet-121 | 25.09 | 07.80 | 7,978,856 | 2,852.39M | Converted from Gluon Model Zoo (log) |
| DenseNet-161 | 22.39 | 6.18 | 28,681,000 | 7,761.25M | Converted from Gluon Model Zoo (log) |
| DenseNet-169 | 23.88 | 6.89 | 14,149,480 | 3,381.48M | Converted from Gluon Model Zoo (log) |
| DenseNet-201 | 22.69 | 6.35 | 20,013,928 | 4,318.75M | Converted from Gluon Model Zoo (log) |
| DarkNet Tiny | 43.35 | 19.46 | 1,042,104 | 496.34M | Converted from GL model (log) |
| DarkNet Ref | 37.99 | 16.68 | 7,319,416 | 365.55M | Converted from GL model (log) |
| SqueezeNet v1.0 | 41.07 | 19.04 | 1,248,424 | 828.30M | Converted from GL model (log) |
| SqueezeNet v1.1 | 41.37 | 19.20 | 1,235,496 | 354.88M | Converted from GL model (log) |
| ShuffleNetV2 x0.5 | 41.00 | 18.68 | 1,366,792 | 42.34M | Converted from GL model (log) |
| ShuffleNetV2 x1.0 | 33.82 | 13.49 | 2,278,604 | 147.92M | Converted from GL model (log) |
| ShuffleNetV2 x1.5 | 32.46 | 12.47 | 4,406,098 | 318.61M | Converted from GL model (log) |
| 108-MENet-8x1 (g=3) | 46.09 | 22.37 | 654,516 | 40.64M | From clavichord93/MENet (log) |
| 128-MENet-8x1 (g=4) | 45.78 | 21.93 | 750,796 | 43.58M | From clavichord93/MENet (log) |
| 228-MENet-12x1 (g=3) | 35.02 | 14.01 | 1,806,568 | 148.93M | From clavichord93/MENet (log) |
| 256-MENet-12x1 (g=4) | 34.48 | 13.91 | 1,888,240 | 146.11M | From clavichord93/MENet (log) |
| 348-MENet-12x1 (g=3) | 31.17 | 11.42 | 3,368,128 | 306.31M | From clavichord93/MENet (log) |
| 352-MENet-12x1 (g=8) | 34.69 | 13.75 | 2,272,872 | 151.03M | From clavichord93/MENet (log) |
| 456-MENet-24x1 (g=3) | 29.55 | 10.44 | 5,304,784 | 560.72M | From clavichord93/MENet (log) |
| MobileNet x0.25 | 45.80 | 22.17 | 470,072 | 42.30M | Converted from GL model (log) |
| MobileNet x0.5 | 36.11 | 14.81 | 1,331,592 | 152.04M | Converted from GL model (log) |
| MobileNet x0.75 | 32.71 | 12.28 | 2,585,560 | 329.22M | Converted from Gluon Model Zoo (log) |
| MobileNet x1.0 | 29.24 | 10.03 | 4,231,976 | 573.83M | Converted from Gluon Model Zoo (log) |
| FD-MobileNet x0.25 | 56.71 | 31.98 | 383,160 | 12.44M | From clavichord93/FD-MobileNet (log) |
| FD-MobileNet x0.5 | 44.64 | 21.08 | 993,928 | 40.93M | From clavichord93/FD-MobileNet (log) |
| FD-MobileNet x1.0 | 35.95 | 14.73 | 2,901,288 | 146.08M | From clavichord93/FD-MobileNet (log) |
| MobileNetV2 x0.25 | 48.86 | 25.24 | 1,516,392 | 32.22M | Converted from Gluon Model Zoo (log) |
| MobileNetV2 x0.5 | 35.51 | 14.65 | 1,964,736 | 95.62M | Converted from Gluon Model Zoo (log) |
| MobileNetV2 x0.75 | 30.81 | 11.26 | 2,627,592 | 191.61M | Converted from Gluon Model Zoo (log) |
| MobileNetV2 x1.0 | 28.50 | 9.90 | 3,504,960 | 320.19M | Converted from Gluon Model Zoo (log) |