diff --git a/configs/_base_/datasets/imagenet_bs64_t2t_224.py b/configs/_base_/datasets/imagenet_bs64_t2t_224.py index 375775de..1190d6f9 100644 --- a/configs/_base_/datasets/imagenet_bs64_t2t_224.py +++ b/configs/_base_/datasets/imagenet_bs64_t2t_224.py @@ -68,4 +68,4 @@ data = dict( ann_file='data/imagenet/meta/val.txt', pipeline=test_pipeline)) -evaluation = dict(interval=10, metric='accuracy') +evaluation = dict(interval=1, metric='accuracy', save_best='auto') diff --git a/configs/conformer/README.md b/configs/conformer/README.md index 45e79aaf..596911a0 100644 --- a/configs/conformer/README.md +++ b/configs/conformer/README.md @@ -25,15 +25,13 @@ Within Convolutional Neural Network (CNN), the convolution operations are good a ## Results and models -Some pre-trained models are converted from [official repo](https://github.com/pengzhiliang/Conformer). - -## ImageNet-1k +### ImageNet-1k | Model | Params(M) | Flops(G) | Top-1 (%) | Top-5 (%) | Config | Download | |:---------------------:|:---------:|:--------:|:---------:|:---------:|:------:|:--------:| -| Conformer-tiny-p16\* | 23.52 | 4.90 | 81.31 | 95.60 | [config](configs/conformer/conformer-tiny-p16_8xb128_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/conformer/conformer-tiny-p16_3rdparty_8xb128_in1k_20211206-f6860372.pth) | -| Conformer-small-p32 | 38.85 | 7.09 | 81.96 | 96.02 | [config](configs/conformer/conformer-small-p32_8xb128_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/conformer/conformer-small-p32_8xb128_in1k_20211206-947a0816.pth) | -| Conformer-small-p16\* | 37.67 | 10.31 | 83.32 | 96.46 | [config](configs/conformer/conformer-small-p16_8xb128_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/conformer/conformer-small-p16_3rdparty_8xb128_in1k_20211206-3065dcf5.pth) | -| Conformer-base-p16\* | 83.29 | 22.89 | 83.82 | 96.59 | [config](configs/conformer/conformer-base-p16_8xb128_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/conformer/conformer-base-p16_3rdparty_8xb128_in1k_20211206-bfdf8637.pth) | +| Conformer-tiny-p16\* | 23.52 | 4.90 | 81.31 | 95.60 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/conformer/conformer-tiny-p16_8xb128_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/conformer/conformer-tiny-p16_3rdparty_8xb128_in1k_20211206-f6860372.pth) | +| Conformer-small-p32\* | 38.85 | 7.09 | 81.96 | 96.02 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/conformer/conformer-small-p32_8xb128_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/conformer/conformer-small-p32_8xb128_in1k_20211206-947a0816.pth) | +| Conformer-small-p16\* | 37.67 | 10.31 | 83.32 | 96.46 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/conformer/conformer-small-p16_8xb128_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/conformer/conformer-small-p16_3rdparty_8xb128_in1k_20211206-3065dcf5.pth) | +| Conformer-base-p16\* | 83.29 | 22.89 | 83.82 | 96.59 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/conformer/conformer-base-p16_8xb128_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/conformer/conformer-base-p16_3rdparty_8xb128_in1k_20211206-bfdf8637.pth) | -*Models with \* are converted from other repos.* +*Models with \* are converted from the [official repo](https://github.com/pengzhiliang/Conformer). The config files of these models are only for validation. We don't ensure these config files' training accuracy and welcome you to contribute your reproduction results.* diff --git a/configs/deit/README.md b/configs/deit/README.md index 9418e34a..52a8be66 100644 --- a/configs/deit/README.md +++ b/configs/deit/README.md @@ -25,35 +25,24 @@ Recently, neural networks purely based on attention were shown to address image } ``` -## Pretrained models - -The pre-trained models are converted from the [official repo](https://github.com/facebookresearch/deit). And the teacher of the distilled version DeiT is RegNetY-16GF. +## Results and models ### ImageNet-1k -| Model | Params(M) | Flops(G) | Top-1 (%) | Top-5 (%) | Config | Download | -|:---------------------:|:---------:|:--------:|:---------:|:---------:|:------:|:--------:| -| DeiT-tiny\* | 5.72 | 1.08 | 72.13 | 91.13 | [config](configs/deit/deit-tiny_pt-4xb256_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/deit/deit-tiny_3rdparty_pt-4xb256_in1k_20211124-e930093b.pth) | -| DeiT-tiny distilled\* | 5.72 | 1.08 | 74.51 | 91.90 | [config](configs/deit/deit-tiny-distilled_pt-4xb256_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/deit/deit-tiny-distilled_3rdparty_pt-4xb256_in1k_20211216-c429839a.pth) | -| DeiT-small\* | 22.05 | 4.24 | 79.83 | 94.95 | [config](configs/deit/deit-small_pt-4xb256_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/deit/deit-small_3rdparty_pt-4xb256_in1k_20211124-ffe94edd.pth) | -| DeiT-small distilled\* | 22.05 | 4.24 | 81.17 | 95.40 | [config](configs/deit/deit-small-distilled_pt-4xb256_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/deit/deit-small-distilled_3rdparty_pt-4xb256_in1k_20211216-4de1d725.pth) | -| DeiT-base\* | 86.57 | 16.86 | 81.79 | 95.59 | [config](configs/deit/deit-base_pt-16xb64_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/deit/deit-base_3rdparty_pt-16xb64_in1k_20211124-6f40c188.pth) | -| DeiT-base distilled\* | 86.57 | 16.86 | 83.33 | 96.49 | [config](configs/deit/deit-base-distilled_pt-16xb64_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/deit/deit-base-distilled_3rdparty_pt-16xb64_in1k_20211216-42891296.pth) | +The teacher of the distilled version DeiT is RegNetY-16GF. -*Models with \* are converted from other repos.* +| Model | Pretrain | Params(M) | Flops(G) | Top-1 (%) | Top-5 (%) | Config | Download | +|:---------------------:|:------------:|:---------:|:--------:|:---------:|:---------:|:------:|:--------:| +| DeiT-tiny\* | From scratch | 5.72 | 1.08 | 72.13 | 91.13 | [config](https://github.com/open-mmlab/mmclassification/tree/master/configs/deit/deit-tiny_pt-4xb256_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/deit/deit-tiny_3rdparty_pt-4xb256_in1k_20211124-e930093b.pth) | +| DeiT-tiny distilled\* | From scratch | 5.72 | 1.08 | 74.51 | 91.90 | [config](https://github.com/open-mmlab/mmclassification/tree/master/configs/deit/deit-tiny-distilled_pt-4xb256_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/deit/deit-tiny-distilled_3rdparty_pt-4xb256_in1k_20211216-c429839a.pth) | +| DeiT-small\* | From scratch | 22.05 | 4.24 | 79.83 | 94.95 | [config](https://github.com/open-mmlab/mmclassification/tree/master/configs/deit/deit-small_pt-4xb256_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/deit/deit-small_3rdparty_pt-4xb256_in1k_20211124-ffe94edd.pth) | +| DeiT-small distilled\*| From scratch | 22.05 | 4.24 | 81.17 | 95.40 | [config](https://github.com/open-mmlab/mmclassification/tree/master/configs/deit/deit-small-distilled_pt-4xb256_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/deit/deit-small-distilled_3rdparty_pt-4xb256_in1k_20211216-4de1d725.pth) | +| DeiT-base\* | From scratch | 86.57 | 16.86 | 81.79 | 95.59 | [config](https://github.com/open-mmlab/mmclassification/tree/master/configs/deit/deit-base_pt-16xb64_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/deit/deit-base_3rdparty_pt-16xb64_in1k_20211124-6f40c188.pth) | +| DeiT-base distilled\* | From scratch | 86.57 | 16.86 | 83.33 | 96.49 | [config](https://github.com/open-mmlab/mmclassification/tree/master/configs/deit/deit-base-distilled_pt-16xb64_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/deit/deit-base-distilled_3rdparty_pt-16xb64_in1k_20211216-42891296.pth) | +| DeiT-base 384px\* | ImageNet-1k | 86.86 | 49.37 | 83.04 | 96.31 | [config](https://github.com/open-mmlab/mmclassification/tree/master/configs/deit/deit-base_ft-16xb32_in1k-384px.py) | [model](https://download.openmmlab.com/mmclassification/v0/deit/deit-base_3rdparty_ft-16xb32_in1k-384px_20211124-822d02f2.pth) | +| DeiT-base distilled 384px\* | ImageNet-1k | 86.86 | 49.37 | 85.55 | 97.35 | [config](https://github.com/open-mmlab/mmclassification/tree/master/configs/deit/deit-base-distilled_ft-16xb32_in1k-384px.py) | [model](https://download.openmmlab.com/mmclassification/v0/deit/deit-base-distilled_3rdparty_ft-16xb32_in1k-384px_20211216-e48d6000.pth) | -## Fine-tuned models - -The fine-tuned models are converted from the [official repo](https://github.com/facebookresearch/deit). - -### ImageNet-1k - -| Model | Params(M) | Flops(G) | Top-1 (%) | Top-5 (%) | Config | Download | -|:---------------------:|:---------:|:--------:|:---------:|:---------:|:------:|:--------:| -| DeiT-base 384px\* | 86.86 | 49.37 | 83.04 | 96.31 | [config](configs/deit/deit-base_ft-16xb32_in1k-384px.py) | [model](https://download.openmmlab.com/mmclassification/v0/deit/deit-base_3rdparty_ft-16xb32_in1k-384px_20211124-822d02f2.pth) | -| DeiT-base distilled 384px\* | 86.86 | 49.37 | 85.55 | 97.35 | [config](configs/deit/deit-base-distilled_ft-16xb32_in1k-384px.py) | [model](https://download.openmmlab.com/mmclassification/v0/deit/deit-base-distilled_3rdparty_ft-16xb32_in1k-384px_20211216-e48d6000.pth) | - -*Models with \* are converted from other repos.* +*Models with \* are converted from the [official repo](https://github.com/facebookresearch/deit). The config files of these models are only for validation. We don't ensure these config files' training accuracy and welcome you to contribute your reproduction results.* ```{warning} MMClassification doesn't support training the distilled version DeiT. diff --git a/configs/mlp_mixer/README.md b/configs/mlp_mixer/README.md index 7f7cfafe..17f7ec15 100644 --- a/configs/mlp_mixer/README.md +++ b/configs/mlp_mixer/README.md @@ -23,15 +23,13 @@ Convolutional Neural Networks (CNNs) are the go-to model for computer vision. Re } ``` -## Pretrain model - -The pre-trained modles are converted from [timm](https://github.com/rwightman/pytorch-image-models/blob/master/timm/models/mlp_mixer.py). +## Results and models ### ImageNet-1k | Model | Params(M) | Flops(G) | Top-1 (%) | Top-5 (%) | Config | Download | |:--------------:|:---------:|:--------:|:---------:|:---------:|:------:|:--------:| -| Mixer-B/16\* | 59.88 | 12.61 | 76.68 | 92.25 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/mlp_mixer/mlp-mixer-base-p16_64xb64_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/mlp-mixer/mixer-base-p16_3rdparty_64xb64_in1k_20211124-1377e3e0.pth)| -| Mixer-L/16\* | 208.2 | 44.57 | 72.34 | 88.02 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/mlp_mixer/mlp-mixer-large-p16_64xb64_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/mlp-mixer/mixer-large-p16_3rdparty_64xb64_in1k_20211124-5a2519d2.pth)| +| Mixer-B/16\* | 59.88 | 12.61 | 76.68 | 92.25 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/mlp_mixer/mlp-mixer-base-p16_64xb64_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/mlp-mixer/mixer-base-p16_3rdparty_64xb64_in1k_20211124-1377e3e0.pth) | +| Mixer-L/16\* | 208.2 | 44.57 | 72.34 | 88.02 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/mlp_mixer/mlp-mixer-large-p16_64xb64_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/mlp-mixer/mixer-large-p16_3rdparty_64xb64_in1k_20211124-5a2519d2.pth) | -*Models with \* are converted from other repos.* +*Models with \* are converted from [timm](https://github.com/rwightman/pytorch-image-models/blob/master/timm/models/mlp_mixer.py). The config files of these models are only for validation. We don't ensure these config files' training accuracy and welcome you to contribute your reproduction results.* diff --git a/configs/mobilenet_v2/README.md b/configs/mobilenet_v2/README.md index f7975864..6f22123f 100644 --- a/configs/mobilenet_v2/README.md +++ b/configs/mobilenet_v2/README.md @@ -29,7 +29,7 @@ The MobileNetV2 architecture is based on an inverted residual structure where th ## Results and models -### ImageNet +### ImageNet-1k | Model | Params(M) | Flops(G) | Top-1 (%) | Top-5 (%) | Config | Download | |:---------------------:|:---------:|:--------:|:---------:|:---------:|:---------:|:--------:| diff --git a/configs/mobilenet_v3/README.md b/configs/mobilenet_v3/README.md index 18ce4cbf..b7d0cf14 100644 --- a/configs/mobilenet_v3/README.md +++ b/configs/mobilenet_v3/README.md @@ -22,17 +22,13 @@ We present the next generation of MobileNets based on a combination of complemen } ``` -## Pretrain model - -The pre-trained modles are converted from [torchvision](https://pytorch.org/vision/stable/_modules/torchvision/models/mobilenetv3.html). - -### ImageNet - -| Model | Params(M) | Flops(G) | Top-1 (%) | Top-5 (%) | Download | -|:---------------------:|:---------:|:--------:|:---------:|:---------:|:--------:| -| MobileNetV3-Small | 2.54 | 0.06 | 67.66 | 87.41 | [model](https://download.openmmlab.com/mmclassification/v0/mobilenet_v3/convert/mobilenet_v3_small-8427ecf0.pth)| -| MobileNetV3-Large | 5.48 | 0.23 | 74.04 | 91.34 | [model](https://download.openmmlab.com/mmclassification/v0/mobilenet_v3/convert/mobilenet_v3_large-3ea3c186.pth)| - ## Results and models -Waiting for adding. +### ImageNet-1k + +| Model | Params(M) | Flops(G) | Top-1 (%) | Top-5 (%) | Config | Download | +|:---------------------:|:---------:|:--------:|:---------:|:---------:|:------:|:--------:| +| MobileNetV3-Small\* | 2.54 | 0.06 | 67.66 | 87.41 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/mobilenet_v3/mobilenet-v3-small_8xb32_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/mobilenet_v3/convert/mobilenet_v3_small-8427ecf0.pth) | +| MobileNetV3-Large\* | 5.48 | 0.23 | 74.04 | 91.34 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/mobilenet_v3/mobilenet-v3-large_8xb32_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/mobilenet_v3/convert/mobilenet_v3_large-3ea3c186.pth) | + +*Models with \* are converted from [torchvision](https://pytorch.org/vision/stable/_modules/torchvision/models/mobilenetv3.html). The config files of these models are only for validation. We don't ensure these config files' training accuracy and welcome you to contribute your reproduction results.* diff --git a/configs/regnet/README.md b/configs/regnet/README.md index c293e7e4..7a391edc 100644 --- a/configs/regnet/README.md +++ b/configs/regnet/README.md @@ -23,23 +23,19 @@ In this work, we present a new network design paradigm. Our goal is to help adva } ``` -## Pretrain model - -The pre-trained modles are converted from [model zoo of pycls](https://github.com/facebookresearch/pycls/blob/master/MODEL_ZOO.md). - -### ImageNet - -| Model | Params(M) | Flops(G) | Top-1 (%) | Top-5 (%) | Download | -|:---------------------:|:---------:|:--------:|:---------:|:---------:|:--------:| -| RegNetX-400MF | 5.16 | 0.41 | 72.55 | 90.91 | [model](https://download.openmmlab.com/mmclassification/v0/regnet/convert/RegNetX-400MF-0db9f35c.pth)| -| RegNetX-800MF | 7.26 | 0.81 | 75.21 | 92.37 | [model](https://download.openmmlab.com/mmclassification/v0/regnet/convert/RegNetX-800MF-4f9d1e8a.pth)| -| RegNetX-1.6GF | 9.19 | 1.63 | 77.04 | 93.51 | [model](https://download.openmmlab.com/mmclassification/v0/regnet/convert/RegNetX-1.6GF-cfb32375.pth)| -| RegNetX-3.2GF | 15.3 | 3.21 | 78.26 | 94.20 | [model](https://download.openmmlab.com/mmclassification/v0/regnet/convert/RegNetX-3.2GF-82c43fd5.pth)| -| RegNetX-4.0GF | 22.12 | 4.0 | 78.72 | 94.22 | [model](https://download.openmmlab.com/mmclassification/v0/regnet/convert/RegNetX-4.0GF-ef8bb32c.pth)| -| RegNetX-6.4GF | 26.21 | 6.51 | 79.22 | 94.61 | [model](https://download.openmmlab.com/mmclassification/v0/regnet/convert/RegNetX-6.4GF-6888c0ea.pth)| -| RegNetX-8.0GF | 39.57 | 8.03 | 79.31 | 94.57 | [model](https://download.openmmlab.com/mmclassification/v0/regnet/convert/RegNetX-8.0GF-cb4c77ec.pth)| -| RegNetX-12GF | 46.11 | 12.15 | 79.91 | 94.78 | [model](https://download.openmmlab.com/mmclassification/v0/regnet/convert/RegNetX-12GF-0574538f.pth)| - ## Results and models -Waiting for adding. +### ImageNet-1k + +| Model | Params(M) | Flops(G) | Top-1 (%) | Top-5 (%) | Config | Download | +|:---------------------:|:---------:|:--------:|:---------:|:---------:|:------:|:--------:| +| RegNetX-400MF\* | 5.16 | 0.41 | 72.55 | 90.91 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/regnet/regnetx-400mf_8xb32_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/regnet/convert/RegNetX-400MF-0db9f35c.pth) | +| RegNetX-800MF\* | 7.26 | 0.81 | 75.21 | 92.37 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/regnet/regnetx-800mf_8xb32_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/regnet/convert/RegNetX-800MF-4f9d1e8a.pth) | +| RegNetX-1.6GF\* | 9.19 | 1.63 | 77.04 | 93.51 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/regnet/regnetx-1.6gf_8xb32_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/regnet/convert/RegNetX-1.6GF-cfb32375.pth) | +| RegNetX-3.2GF\* | 15.3 | 3.21 | 78.26 | 94.20 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/regnet/regnetx-3.2gf_8xb32_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/regnet/convert/RegNetX-3.2GF-82c43fd5.pth) | +| RegNetX-4.0GF\* | 22.12 | 4.0 | 78.72 | 94.22 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/regnet/regnetx-4.0gf_8xb32_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/regnet/convert/RegNetX-4.0GF-ef8bb32c.pth) | +| RegNetX-6.4GF\* | 26.21 | 6.51 | 79.22 | 94.61 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/regnet/regnetx-6.4gf_8xb32_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/regnet/convert/RegNetX-6.4GF-6888c0ea.pth) | +| RegNetX-8.0GF\* | 39.57 | 8.03 | 79.31 | 94.57 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/regnet/regnetx-8.0gf_8xb32_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/regnet/convert/RegNetX-8.0GF-cb4c77ec.pth) | +| RegNetX-12GF\* | 46.11 | 12.15 | 79.91 | 94.78 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/regnet/regnetx-12gf_8xb32_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/regnet/convert/RegNetX-12GF-0574538f.pth) | + +*Models with \* are converted from [pycls](https://github.com/facebookresearch/pycls/blob/master/MODEL_ZOO.md). The config files of these models are only for validation. We don't ensure these config files' training accuracy and welcome you to contribute your reproduction results.* diff --git a/configs/repvgg/README.md b/configs/repvgg/README.md index 98791c4f..997e8554 100644 --- a/configs/repvgg/README.md +++ b/configs/repvgg/README.md @@ -22,7 +22,9 @@ We present a simple but powerful architecture of convolutional neural network, w } ``` -## Pretrain model +## Results and models + +### ImageNet-1k | Model | Epochs | Params(M) | Flops(G) | Top-1 (%) | Top-5 (%) | Config | Download | | :---------: | :----: | :-------------------------------: | :-----------------------------: | :-------: | :-------: | :----------------------------------------------------------: | :----------------------------------------------------------: | @@ -39,7 +41,7 @@ We present a simple but powerful architecture of convolutional neural network, w | RepVGG-B3g4\* | 200 | 83.83 (train) \| 75.63 (deploy) | 17.9 (train) \| 16.08 (deploy) | 80.22 | 95.10 | [config (train)](https://github.com/open-mmlab/mmclassification/blob/master/configs/repvgg/repvgg-B3g4_4xb64-autoaug-lbs-mixup-coslr-200e_in1k.py) \|[config (deploy)](https://github.com/open-mmlab/mmclassification/blob/master/configs/repvgg/deploy/repvgg-B3g4_deploy_4xb64-autoaug-lbs-mixup-coslr-200e_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/repvgg/repvgg-B3g4_3rdparty_4xb64-autoaug-lbs-mixup-coslr-200e_in1k_20210909-4e54846a.pth) | | RepVGG-D2se\* | 200 | 133.33 (train) \| 120.39 (deploy) | 36.56 (train) \| 32.85 (deploy) | 81.81 | 95.94 | [config (train)](https://github.com/open-mmlab/mmclassification/blob/master/configs/repvgg/repvgg-D2se_4xb64-autoaug-lbs-mixup-coslr-200e_in1k.py) \|[config (deploy)](https://github.com/open-mmlab/mmclassification/blob/master/configs/repvgg/deploy/repvgg-D2se_deploy_4xb64-autoaug-lbs-mixup-coslr-200e_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/repvgg/repvgg-D2se_3rdparty_4xb64-autoaug-lbs-mixup-coslr-200e_in1k_20210909-cf3139b7.pth) | -*Models with \* are converted from other repos.* +*Models with \* are converted from the [official repo](https://github.com/DingXiaoH/RepVGG). The config files of these models are only for validation. We don't ensure these config files' training accuracy and welcome you to contribute your reproduction results.* ## Reparameterize RepVGG diff --git a/configs/res2net/README.md b/configs/res2net/README.md index 45558aa1..db6bc3cb 100644 --- a/configs/res2net/README.md +++ b/configs/res2net/README.md @@ -22,16 +22,14 @@ Representing features at multiple scales is of great importance for numerous vis } ``` -## Pretrain model +## Results and models -The pre-trained models are converted from [official repo](https://github.com/Res2Net/Res2Net-PretrainedModels). +### ImageNet-1k -### ImageNet 1k +| Model | resolution | Params(M) | Flops(G) | Top-1 (%) | Top-5 (%) | Config | Download | +|:---------------------:|:-----------:|:---------:|:---------:|:---------:|:---------:|:------:|:--------:| +| Res2Net-50-14w-8s\* | 224x224 | 25.06 | 4.22 | 78.14 | 93.85 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/res2net/res2net50-w14-s8_8xb32_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/res2net/res2net50-w14-s8_3rdparty_8xb32_in1k_20210927-bc967bf1.pth) | [log]()| +| Res2Net-50-26w-8s\* | 224x224 | 48.40 | 8.39 | 79.20 | 94.36 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/res2net/res2net50-w26-s8_8xb32_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/res2net/res2net50-w26-s8_3rdparty_8xb32_in1k_20210927-f547a94b.pth) | [log]()| +| Res2Net-101-26w-4s\* | 224x224 | 45.21 | 8.12 | 79.19 | 94.44 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/res2net/res2net101-w26-s4_8xb32_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/res2net/res2net101-w26-s4_3rdparty_8xb32_in1k_20210927-870b6c36.pth) | [log]()| -| Model | resolution | Params(M) | Flops(G) | Top-1 (%) | Top-5 (%) | Download | -|:---------------------:|:-----------:|:---------:|:---------:|:---------:|:---------:|:--------:| -| Res2Net-50-14w-8s\* | 224x224 | 25.06 | 4.22 | 78.14 | 93.85 | [model](https://download.openmmlab.com/mmclassification/v0/res2net/res2net50-w14-s8_3rdparty_8xb32_in1k_20210927-bc967bf1.pth)| -| Res2Net-50-26w-8s\* | 224x224 | 48.40 | 8.39 | 79.20 | 94.36 | [model](https://download.openmmlab.com/mmclassification/v0/res2net/res2net50-w26-s8_3rdparty_8xb32_in1k_20210927-f547a94b.pth)| -| Res2Net-101-26w-4s\* | 224x224 | 45.21 | 8.12 | 79.19 | 94.44 | [model](https://download.openmmlab.com/mmclassification/v0/res2net/res2net101-w26-s4_3rdparty_8xb32_in1k_20210927-870b6c36.pth)| - -*Models with \* are converted from other repos.* +*Models with \* are converted from the [official repo](https://github.com/Res2Net/Res2Net-PretrainedModels). The config files of these models are only for validation. We don't ensure these config files' training accuracy and welcome you to contribute your reproduction results.* diff --git a/configs/resnet/README.md b/configs/resnet/README.md index 3d15cb00..29ec8bfe 100644 --- a/configs/resnet/README.md +++ b/configs/resnet/README.md @@ -42,7 +42,7 @@ The depth of representations is of central importance for many visual recognitio |:---------------------:|:---------:|:--------:|:---------:|:---------:|:---------:|:--------:| | ResNet-50-b16x8 | 23.71 | 1.31 | 79.90 | 95.19 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/resnet/resnet50_8xb16_cifar100.py) | [model](https://download.openmmlab.com/mmclassification/v0/resnet/resnet50_b16x8_cifar100_20210528-67b58a1b.pth) | [log](https://download.openmmlab.com/mmclassification/v0/resnet/resnet50_b16x8_cifar100_20210528-67b58a1b.log.json) | -### ImageNet +### ImageNet-1k | Model | Params(M) | Flops(G) | Top-1 (%) | Top-5 (%) | Config | Download | |:---------------------:|:---------:|:--------:|:---------:|:---------:|:---------:|:--------:| diff --git a/configs/resnext/README.md b/configs/resnext/README.md index 8fd9efed..6cea78c5 100644 --- a/configs/resnext/README.md +++ b/configs/resnext/README.md @@ -24,7 +24,7 @@ We present a simple, highly modularized network architecture for image classific ## Results and models -### ImageNet +### ImageNet-1k | Model | Params(M) | Flops(G) | Top-1 (%) | Top-5 (%) | Config | Download | |:---------------------:|:---------:|:--------:|:---------:|:---------:|:---------:|:--------:| diff --git a/configs/seresnet/README.md b/configs/seresnet/README.md index fef83524..5c5954ad 100644 --- a/configs/seresnet/README.md +++ b/configs/seresnet/README.md @@ -24,7 +24,7 @@ The central building block of convolutional neural networks (CNNs) is the convol ## Results and models -### ImageNet +### ImageNet-1k | Model | Params(M) | Flops(G) | Top-1 (%) | Top-5 (%) | Config | Download | |:---------------------:|:---------:|:--------:|:---------:|:---------:|:---------:|:--------:| diff --git a/configs/shufflenet_v1/README.md b/configs/shufflenet_v1/README.md index 21ce39ba..85adde68 100644 --- a/configs/shufflenet_v1/README.md +++ b/configs/shufflenet_v1/README.md @@ -24,7 +24,7 @@ We introduce an extremely computation-efficient CNN architecture named ShuffleNe ## Results and models -### ImageNet +### ImageNet-1k | Model | Params(M) | Flops(G) | Top-1 (%) | Top-5 (%) | Config | Download | |:---------------------:|:---------:|:--------:|:---------:|:---------:|:---------:|:--------:| diff --git a/configs/shufflenet_v2/README.md b/configs/shufflenet_v2/README.md index 268a15c9..dead2d3d 100644 --- a/configs/shufflenet_v2/README.md +++ b/configs/shufflenet_v2/README.md @@ -24,7 +24,7 @@ Currently, the neural network architecture design is mostly guided by the *indir ## Results and models -### ImageNet +### ImageNet-1k | Model | Params(M) | Flops(G) | Top-1 (%) | Top-5 (%) | Config | Download | |:---------------------:|:---------:|:--------:|:---------:|:---------:|:---------:|:--------:| diff --git a/configs/swin_transformer/README.md b/configs/swin_transformer/README.md index 7b3e21fe..be3c787c 100644 --- a/configs/swin_transformer/README.md +++ b/configs/swin_transformer/README.md @@ -12,6 +12,7 @@ This paper presents a new vision Transformer, called Swin Transformer, that capa ## Citation + ```latex @article{liu2021Swin, title={Swin Transformer: Hierarchical Vision Transformer using Shifted Windows}, @@ -21,29 +22,32 @@ This paper presents a new vision Transformer, called Swin Transformer, that capa } ``` -## Pretrain model - -The pre-trained modles are converted from [model zoo of Swin Transformer](https://github.com/microsoft/Swin-Transformer#main-results-on-imagenet-with-pretrained-models). - -### ImageNet 1k - -| Model | Pretrain | resolution | Params(M) | Flops(G) | Top-1 (%) | Top-5 (%) | Download | -|:---------:|:------------:|:-----------:|:---------:|:---------:|:---------:|:---------:|:--------:| -| Swin-T | ImageNet-1k | 224x224 | 28.29 | 4.36 | 81.18 | 95.52 | [model](https://download.openmmlab.com/mmclassification/v0/swin-transformer/convert/swin_tiny_patch4_window7_224-160bb0a5.pth)| -| Swin-S | ImageNet-1k | 224x224 | 49.61 | 8.52 | 83.21 | 96.25 | [model](https://download.openmmlab.com/mmclassification/v0/swin-transformer/convert/swin_small_patch4_window7_224-cc7a01c9.pth)| -| Swin-B | ImageNet-1k | 224x224 | 87.77 | 15.14 | 83.42 | 96.44 | [model](https://download.openmmlab.com/mmclassification/v0/swin-transformer/convert/swin_base_patch4_window7_224-4670dd19.pth)| -| Swin-B | ImageNet-1k | 384x384 | 87.90 | 44.49 | 84.49 | 96.95 | [model](https://download.openmmlab.com/mmclassification/v0/swin-transformer/convert/swin_base_patch4_window12_384-02c598a4.pth)| -| Swin-B | ImageNet-22k | 224x224 | 87.77 | 15.14 | 85.16 | 97.50 | [model](https://download.openmmlab.com/mmclassification/v0/swin-transformer/convert/swin_base_patch4_window7_224_22kto1k-f967f799.pth)| -| Swin-B | ImageNet-22k | 384x384 | 87.90 | 44.49 | 86.44 | 98.05 | [model](https://download.openmmlab.com/mmclassification/v0/swin-transformer/convert/swin_base_patch4_window12_384_22kto1k-d59b0d1d.pth)| -| Swin-L | ImageNet-22k | 224x224 | 196.53 | 34.04 | 86.24 | 97.88 | [model](https://download.openmmlab.com/mmclassification/v0/swin-transformer/convert/swin_large_patch4_window7_224_22kto1k-5f0996db.pth)| -| Swin-L | ImageNet-22k | 384x384 | 196.74 | 100.04 | 87.25 | 98.25 | [model](https://download.openmmlab.com/mmclassification/v0/swin-transformer/convert/swin_large_patch4_window12_384_22kto1k-0a40944b.pth)| - - ## Results and models -### ImageNet 1k -| Model | Pretrain | resolution | Params(M) | Flops(G) | Top-1 (%) | Top-5 (%) | Config | Download | -|:---------:|:------------:|:-----------:|:---------:|:---------:|:---------:|:---------:|:----------:|:--------:| -| Swin-T | ImageNet-1k | 224x224 | 28.29 | 4.36 | 81.18 | 95.61 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/swin_transformer/swin-tiny_16xb64_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/swin-transformer/swin_tiny_224_b16x64_300e_imagenet_20210616_090925-66df6be6.pth) | [log](https://download.openmmlab.com/mmclassification/v0/swin-transformer/swin_tiny_224_b16x64_300e_imagenet_20210616_090925.log.json)| -| Swin-S | ImageNet-1k | 224x224 | 49.61 | 8.52 | 83.02 | 96.29 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/swin_transformer/swin-small_16xb64_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/swin-transformer/swin_small_224_b16x64_300e_imagenet_20210615_110219-7f9d988b.pth) | [log](https://download.openmmlab.com/mmclassification/v0/swin-transformer/swin_small_224_b16x64_300e_imagenet_20210615_110219.log.json)| -| Swin-B | ImageNet-1k | 224x224 | 87.77 | 15.14 | 83.36 | 96.44 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/swin_transformer/swin_base_224_b16x64_300e_imagenet.py) | [model](https://download.openmmlab.com/mmclassification/v0/swin-transformer/swin_base_224_b16x64_300e_imagenet_20210616_190742-93230b0d.pth) | [log](https://download.openmmlab.com/mmclassification/v0/swin-transformer/swin_base_224_b16x64_300e_imagenet_20210616_190742.log.json)| +### ImageNet-21k + +The pre-trained models on ImageNet-21k are used to fine-tune, and therefore don't have evaluation results. + +| Model | resolution | Params(M) | Flops(G) | Download | +|:---------:|:-----------:|:---------:|:---------:|:--------:| +| Swin-B | 224x224 | 86.74 | 15.14 | [model](https://download.openmmlab.com/mmclassification/v0/swin-transformer/convert/swin-base_3rdparty_in21k.pth)| +| Swin-B | 384x384 | 86.88 | 44.49 | [model](https://download.openmmlab.com/mmclassification/v0/swin-transformer/convert/swin-base_3rdparty_in21k-384px.pth)| +| Swin-L | 224x224 | 195.00 | 34.04 | [model](https://download.openmmlab.com/mmclassification/v0/swin-transformer/convert/swin-large_3rdparty_in21k.pth)| +| Swin-L | 384x384 | 195.20 | 100.04 | [model](https://download.openmmlab.com/mmclassification/v0/swin-transformer/convert/swin-base_3rdparty_in21k-384px.pth)| + +### ImageNet-1k + +| Model | Pretrain | resolution | Params(M) | Flops(G) | Top-1 (%) | Top-5 (%) | Config | Download | +|:---------:|:------------:|:-----------:|:---------:|:---------:|:---------:|:---------:|:------:|:--------:| +| Swin-T | From scratch | 224x224 | 28.29 | 4.36 | 81.18 | 95.61 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/swin_transformer/swin-tiny_16xb64_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/swin-transformer/swin_tiny_224_b16x64_300e_imagenet_20210616_090925-66df6be6.pth) | [log](https://download.openmmlab.com/mmclassification/v0/swin-transformer/swin_tiny_224_b16x64_300e_imagenet_20210616_090925.log.json)| +| Swin-S | From scratch | 224x224 | 49.61 | 8.52 | 83.02 | 96.29 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/swin_transformer/swin-small_16xb64_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/swin-transformer/swin_small_224_b16x64_300e_imagenet_20210615_110219-7f9d988b.pth) | [log](https://download.openmmlab.com/mmclassification/v0/swin-transformer/swin_small_224_b16x64_300e_imagenet_20210615_110219.log.json)| +| Swin-B | From scratch | 224x224 | 87.77 | 15.14 | 83.36 | 96.44 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/swin_transformer/swin_base_224_b16x64_300e_imagenet.py) | [model](https://download.openmmlab.com/mmclassification/v0/swin-transformer/swin_base_224_b16x64_300e_imagenet_20210616_190742-93230b0d.pth) | [log](https://download.openmmlab.com/mmclassification/v0/swin-transformer/swin_base_224_b16x64_300e_imagenet_20210616_190742.log.json)| +| Swin-S\* | From scratch | 224x224 | 49.61 | 8.52 | 83.21 | 96.25 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/swin_transformer/swin-small_16xb64_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/swin-transformer/convert/swin_small_patch4_window7_224-cc7a01c9.pth) | +| Swin-B\* | From scratch | 224x224 | 87.77 | 15.14 | 83.42 | 96.44 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/swin_transformer/swin-base_16xb64_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/swin-transformer/convert/swin_base_patch4_window7_224-4670dd19.pth)| +| Swin-B\* | From scratch | 384x384 | 87.90 | 44.49 | 84.49 | 96.95 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/swin_transformer/swin-base_16xb64_in1k-384px.py) | [model](https://download.openmmlab.com/mmclassification/v0/swin-transformer/convert/swin_base_patch4_window12_384-02c598a4.pth)| +| Swin-B\* | ImageNet-21k | 224x224 | 87.77 | 15.14 | 85.16 | 97.50 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/swin_transformer/swin-base_16xb64_in1k.py)| [model](https://download.openmmlab.com/mmclassification/v0/swin-transformer/convert/swin_base_patch4_window7_224_22kto1k-f967f799.pth)| +| Swin-B\* | ImageNet-21k | 384x384 | 87.90 | 44.49 | 86.44 | 98.05 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/swin_transformer/swin-base_16xb64_in1k-384px.py) | [model](https://download.openmmlab.com/mmclassification/v0/swin-transformer/convert/swin_base_patch4_window12_384_22kto1k-d59b0d1d.pth)| +| Swin-L\* | ImageNet-21k | 224x224 | 196.53 | 34.04 | 86.24 | 97.88 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/swin_transformer/swin-large_16xb64_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/swin-transformer/convert/swin_large_patch4_window7_224_22kto1k-5f0996db.pth)| +| Swin-L\* | ImageNet-21k | 384x384 | 196.74 | 100.04 | 87.25 | 98.25 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/swin_transformer/swin-large_16xb64_in1k-384px.py) | [model](https://download.openmmlab.com/mmclassification/v0/swin-transformer/convert/swin_large_patch4_window12_384_22kto1k-0a40944b.pth)| + +*Models with \* are converted from the [official repo](https://github.com/microsoft/Swin-Transformer#main-results-on-imagenet-with-pretrained-models). The config files of these models are only for validation. We don't ensure these config files' training accuracy and welcome you to contribute your reproduction results.* diff --git a/configs/t2t_vit/README.md b/configs/t2t_vit/README.md index c4b7b092..8e198d30 100644 --- a/configs/t2t_vit/README.md +++ b/configs/t2t_vit/README.md @@ -21,20 +21,14 @@ Transformers, which are popular for language modeling, have been explored for so } ``` -## Pretrain model - -The pre-trained models are converted from [official repo](https://github.com/yitu-opensource/T2T-ViT/tree/main#2-t2t-vit-models). +## Results and models ### ImageNet-1k | Model | Params(M) | Flops(G) | Top-1 (%) | Top-5 (%) | Config | Download | |:--------------:|:---------:|:--------:|:---------:|:---------:|:------:|:--------:| -| T2T-ViT_t-14\* | 21.47 | 4.34 | 81.69 | 95.85 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/t2t_vit/t2t-vit-t-14_8xb64_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/t2t-vit/t2t-vit-t-14_3rdparty_8xb64_in1k_20210928-b7c09b62.pth) | [log]()| -| T2T-ViT_t-19\* | 39.08 | 7.80 | 82.43 | 96.08 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/t2t_vit/t2t-vit-t-19_8xb64_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/t2t-vit/t2t-vit-t-19_3rdparty_8xb64_in1k_20210928-7f1478d5.pth) | [log]()| -| T2T-ViT_t-24\* | 64.00 | 12.69 | 82.55 | 96.06 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/t2t_vit/t2t-vit-t-24_8xb64_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/t2t-vit/t2t-vit-t-24_3rdparty_8xb64_in1k_20210928-fe95a61b.pth) | [log]()| +| T2T-ViT_t-14 | 21.47 | 4.34 | 81.83 | 95.84 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/t2t_vit/t2t-vit-t-14_8xb64_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/t2t-vit/t2t-vit-t-14_8xb64_in1k_20211220-f7378dd5.pth) | [log](https://download.openmmlab.com/mmclassification/v0/t2t-vit/t2t-vit-t-14_8xb64_in1k_20211220-f7378dd5.log.json)| +| T2T-ViT_t-19 | 39.08 | 7.80 | 82.63 | 96.18 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/t2t_vit/t2t-vit-t-19_8xb64_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/t2t-vit/t2t-vit-t-19_8xb64_in1k_20211214-7f5e3aaf.pth) | [log](https://download.openmmlab.com/mmclassification/v0/t2t-vit/t2t-vit-t-19_8xb64_in1k_20211214-7f5e3aaf.log.json)| +| T2T-ViT_t-24 | 64.00 | 12.69 | 82.71 | 96.09 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/t2t_vit/t2t-vit-t-24_8xb64_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/t2t-vit/t2t-vit-t-24_8xb64_in1k_20211214-b2a68ae3.pth) | [log](https://download.openmmlab.com/mmclassification/v0/t2t-vit/t2t-vit-t-24_8xb64_in1k_20211214-b2a68ae3.log.json)| -*Models with \* are converted from other repos.* - -## Results and models - -Waiting for adding. +*In consistent with the [official repo](https://github.com/yitu-opensource/T2T-ViT), we adopt the best checkpoints during training.* diff --git a/configs/t2t_vit/metafile.yml b/configs/t2t_vit/metafile.yml index 319424fc..f2125426 100644 --- a/configs/t2t_vit/metafile.yml +++ b/configs/t2t_vit/metafile.yml @@ -17,7 +17,7 @@ Collections: Version: v0.17.0 Models: - - Name: t2t-vit-t-14_3rdparty_8xb64_in1k + - Name: t2t-vit-t-14_8xb64_in1k Metadata: FLOPs: 4340000000 Parameters: 21470000 @@ -25,15 +25,12 @@ Models: Results: - Dataset: ImageNet-1k Metrics: - Top 1 Accuracy: 81.69 - Top 5 Accuracy: 95.85 + Top 1 Accuracy: 81.83 + Top 5 Accuracy: 95.84 Task: Image Classification - Weights: https://download.openmmlab.com/mmclassification/v0/t2t-vit/t2t-vit-t-14_3rdparty_8xb64_in1k_20210928-b7c09b62.pth - Converted From: - Weights: https://github.com/yitu-opensource/T2T-ViT/releases/download/main/81.7_T2T_ViTt_14.pth.tar - Code: https://github.com/yitu-opensource/T2T-ViT/blob/main/models/t2t_vit.py#L243 + Weights: https://download.openmmlab.com/mmclassification/v0/t2t-vit/t2t-vit-t-14_8xb64_in1k_20211220-f7378dd5.pth Config: configs/t2t_vit/t2t-vit-t-14_8xb64_in1k.py - - Name: t2t-vit-t-19_3rdparty_8xb64_in1k + - Name: t2t-vit-t-19_8xb64_in1k Metadata: FLOPs: 7800000000 Parameters: 39080000 @@ -41,15 +38,12 @@ Models: Results: - Dataset: ImageNet-1k Metrics: - Top 1 Accuracy: 82.43 - Top 5 Accuracy: 96.08 + Top 1 Accuracy: 82.63 + Top 5 Accuracy: 96.18 Task: Image Classification - Weights: https://download.openmmlab.com/mmclassification/v0/t2t-vit/t2t-vit-t-19_3rdparty_8xb64_in1k_20210928-7f1478d5.pth - Converted From: - Weights: https://github.com/yitu-opensource/T2T-ViT/releases/download/main/82.4_T2T_ViTt_19.pth.tar - Code: https://github.com/yitu-opensource/T2T-ViT/blob/main/models/t2t_vit.py#L254 + Weights: https://download.openmmlab.com/mmclassification/v0/t2t-vit/t2t-vit-t-19_8xb64_in1k_20211214-7f5e3aaf.pth Config: configs/t2t_vit/t2t-vit-t-19_8xb64_in1k.py - - Name: t2t-vit-t-24_3rdparty_8xb64_in1k + - Name: t2t-vit-t-24_8xb64_in1k Metadata: FLOPs: 12690000000 Parameters: 64000000 @@ -57,11 +51,8 @@ Models: Results: - Dataset: ImageNet-1k Metrics: - Top 1 Accuracy: 82.55 - Top 5 Accuracy: 96.06 + Top 1 Accuracy: 82.71 + Top 5 Accuracy: 96.09 Task: Image Classification - Weights: https://download.openmmlab.com/mmclassification/v0/t2t-vit/t2t-vit-t-24_3rdparty_8xb64_in1k_20210928-fe95a61b.pth - Converted From: - Weights: https://github.com/yitu-opensource/T2T-ViT/releases/download/main/82.6_T2T_ViTt_24.pth.tar - Code: https://github.com/yitu-opensource/T2T-ViT/blob/main/models/t2t_vit.py#L265 + Weights: https://download.openmmlab.com/mmclassification/v0/t2t-vit/t2t-vit-t-24_8xb64_in1k_20211214-b2a68ae3.pth Config: configs/t2t_vit/t2t-vit-t-24_8xb64_in1k.py diff --git a/configs/t2t_vit/t2t-vit-t-14_8xb64_in1k.py b/configs/t2t_vit/t2t-vit-t-14_8xb64_in1k.py index 126d564e..a391df48 100644 --- a/configs/t2t_vit/t2t-vit-t-14_8xb64_in1k.py +++ b/configs/t2t_vit/t2t-vit-t-14_8xb64_in1k.py @@ -6,8 +6,9 @@ _base_ = [ # optimizer paramwise_cfg = dict( + norm_decay_mult=0.0, bias_decay_mult=0.0, - custom_keys={'.backbone.cls_token': dict(decay_mult=0.0)}, + custom_keys={'cls_token': dict(decay_mult=0.0)}, ) optimizer = dict( type='AdamW', @@ -21,11 +22,14 @@ optimizer_config = dict(grad_clip=None) # FIXME: lr in the first 300 epochs conforms to the CosineAnnealing and # the lr in the last 10 epoch equals to min_lr lr_config = dict( - policy='CosineAnnealing', + policy='CosineAnnealingCooldown', min_lr=1e-5, + cool_down_time=10, + cool_down_ratio=0.1, by_epoch=True, warmup_by_epoch=True, warmup='linear', warmup_iters=10, warmup_ratio=1e-6) +custom_hooks = [dict(type='EMAHook', momentum=4e-5, priority='ABOVE_NORMAL')] runner = dict(type='EpochBasedRunner', max_epochs=310) diff --git a/configs/t2t_vit/t2t-vit-t-19_8xb64_in1k.py b/configs/t2t_vit/t2t-vit-t-19_8xb64_in1k.py index afd05a76..e1157f89 100644 --- a/configs/t2t_vit/t2t-vit-t-19_8xb64_in1k.py +++ b/configs/t2t_vit/t2t-vit-t-19_8xb64_in1k.py @@ -6,8 +6,9 @@ _base_ = [ # optimizer paramwise_cfg = dict( + norm_decay_mult=0.0, bias_decay_mult=0.0, - custom_keys={'.backbone.cls_token': dict(decay_mult=0.0)}, + custom_keys={'cls_token': dict(decay_mult=0.0)}, ) optimizer = dict( type='AdamW', @@ -21,11 +22,14 @@ optimizer_config = dict(grad_clip=None) # FIXME: lr in the first 300 epochs conforms to the CosineAnnealing and # the lr in the last 10 epoch equals to min_lr lr_config = dict( - policy='CosineAnnealing', + policy='CosineAnnealingCooldown', min_lr=1e-5, + cool_down_time=10, + cool_down_ratio=0.1, by_epoch=True, warmup_by_epoch=True, warmup='linear', warmup_iters=10, warmup_ratio=1e-6) +custom_hooks = [dict(type='EMAHook', momentum=4e-5, priority='ABOVE_NORMAL')] runner = dict(type='EpochBasedRunner', max_epochs=310) diff --git a/configs/t2t_vit/t2t-vit-t-24_8xb64_in1k.py b/configs/t2t_vit/t2t-vit-t-24_8xb64_in1k.py index 9f856f3e..815f2f15 100644 --- a/configs/t2t_vit/t2t-vit-t-24_8xb64_in1k.py +++ b/configs/t2t_vit/t2t-vit-t-24_8xb64_in1k.py @@ -6,8 +6,9 @@ _base_ = [ # optimizer paramwise_cfg = dict( + norm_decay_mult=0.0, bias_decay_mult=0.0, - custom_keys={'.backbone.cls_token': dict(decay_mult=0.0)}, + custom_keys={'cls_token': dict(decay_mult=0.0)}, ) optimizer = dict( type='AdamW', @@ -21,11 +22,14 @@ optimizer_config = dict(grad_clip=None) # FIXME: lr in the first 300 epochs conforms to the CosineAnnealing and # the lr in the last 10 epoch equals to min_lr lr_config = dict( - policy='CosineAnnealing', + policy='CosineAnnealingCooldown', min_lr=1e-5, + cool_down_time=10, + cool_down_ratio=0.1, by_epoch=True, warmup_by_epoch=True, warmup='linear', warmup_iters=10, warmup_ratio=1e-6) +custom_hooks = [dict(type='EMAHook', momentum=4e-5, priority='ABOVE_NORMAL')] runner = dict(type='EpochBasedRunner', max_epochs=310) diff --git a/configs/tnt/README.md b/configs/tnt/README.md index faa259f8..69a408f1 100644 --- a/configs/tnt/README.md +++ b/configs/tnt/README.md @@ -22,18 +22,12 @@ Transformer is a new kind of neural architecture which encodes the input data as } ``` -## Pretrain model - -The pre-trained modles are converted from [timm](https://github.com/rwightman/pytorch-image-models/). +## Results and models ### ImageNet -| Model | Params(M) | Flops(G) | Top-1 (%) | Top-5 (%) | Download | -|:---------------------:|:---------:|:--------:|:---------:|:---------:|:--------:| -| Transformer in Transformer small\* | 23.76 | 3.36 | 81.52 | 95.73 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/tnt/tnt-s-p16_16xb64_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/tnt/tnt-small-p16_3rdparty_in1k_20210903-c56ee7df.pth) | [log]()| +| Model | Params(M) | Flops(G) | Top-1 (%) | Top-5 (%) | Config | Download | +|:----------------------------------:|:---------:|:--------:|:---------:|:---------:|:------:|:--------:| +| Transformer in Transformer small\* | 23.76 | 3.36 | 81.52 | 95.73 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/tnt/tnt-s-p16_16xb64_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/tnt/tnt-small-p16_3rdparty_in1k_20210903-c56ee7df.pth) | -*Models with \* are converted from other repos.* - -## Results and models - -Waiting for adding. +*Models with \* are converted from [timm](https://github.com/rwightman/pytorch-image-models/). The config files of these models are only for validation. We don't ensure these config files' training accuracy and welcome you to contribute your reproduction results.* diff --git a/configs/vgg/README.md b/configs/vgg/README.md index d27282f6..aecb78ad 100644 --- a/configs/vgg/README.md +++ b/configs/vgg/README.md @@ -24,7 +24,7 @@ In this work we investigate the effect of the convolutional network depth on its ## Results and models -### ImageNet +### ImageNet-1k | Model | Params(M) | Flops(G) | Top-1 (%) | Top-5 (%) | Config | Download | |:---------------------:|:---------:|:--------:|:---------:|:---------:|:---------:|:--------:| diff --git a/configs/vision_transformer/README.md b/configs/vision_transformer/README.md index edd9ca01..570fa1b7 100644 --- a/configs/vision_transformer/README.md +++ b/configs/vision_transformer/README.md @@ -23,36 +23,32 @@ While the Transformer architecture has become the de-facto standard for natural } ``` +## Results and models + The training step of Vision Transformers is divided into two steps. The first step is training the model on a large dataset, like ImageNet-21k, and get the -pretrain model. And the second step is training the model on the target dataset, -like ImageNet-1k, and get the finetune model. Here, we provide both pretrain -models and finetune models. +pre-trained model. And the second step is training the model on the target +dataset, like ImageNet-1k, and get the fine-tuned model. Here, we provide both +pre-trained models and fine-tuned models. -## Pretrain model +### ImageNet-21k -The pre-trained models are converted from [model zoo of Google Research](https://github.com/google-research/vision_transformer#available-vit-models). +The pre-trained models on ImageNet-21k are used to fine-tune, and therefore don't have evaluation results. -### ImageNet 21k +| Model | resolution | Params(M) | Flops(G) | Download | +|:----------:|:-----------:|:---------:|:---------:|:--------:| +| ViT-B16\* | 224x224 | 86.86 | 33.03 | [model](https://download.openmmlab.com/mmclassification/v0/vit/pretrain/vit-base-p16_3rdparty_pt-64xb64_in1k-224_20210928-02284250.pth)| +| ViT-B32\* | 224x224 | 88.30 | 8.56 | [model](https://download.openmmlab.com/mmclassification/v0/vit/pretrain/vit-base-p32_3rdparty_pt-64xb64_in1k-224_20210928-eee25dd4.pth)| +| ViT-L16\* | 224x224 | 304.72 | 116.68 | [model](https://download.openmmlab.com/mmclassification/v0/vit/pretrain/vit-large-p16_3rdparty_pt-64xb64_in1k-224_20210928-0001f9a1.pth)| -| Model | Params(M) | Flops(G) | Download | -|:----------:|:---------:|:---------:|:--------:| -| ViT-B16\* | 86.86 | 33.03 | [model](https://download.openmmlab.com/mmclassification/v0/vit/pretrain/vit-base-p16_3rdparty_pt-64xb64_in1k-224_20210928-02284250.pth)| -| ViT-B32\* | 88.30 | 8.56 | [model](https://download.openmmlab.com/mmclassification/v0/vit/pretrain/vit-base-p32_3rdparty_pt-64xb64_in1k-224_20210928-eee25dd4.pth)| -| ViT-L16\* | 304.72 | 116.68 | [model](https://download.openmmlab.com/mmclassification/v0/vit/pretrain/vit-large-p16_3rdparty_pt-64xb64_in1k-224_20210928-0001f9a1.pth)| +*Models with \* are converted from the [official repo](https://github.com/google-research/vision_transformer#available-vit-models).* -*Models with \* are converted from other repos.* +### ImageNet-1k - -## Finetune model - -The finetune models are converted from [model zoo of Google Research](https://github.com/google-research/vision_transformer#available-vit-models). - -### ImageNet 1k | Model | Pretrain | resolution | Params(M) | Flops(G) | Top-1 (%) | Top-5 (%) | Config | Download | |:----------:|:------------:|:-----------:|:---------:|:---------:|:---------:|:---------:|:----------:|:--------:| | ViT-B16\* | ImageNet-21k | 384x384 | 86.86 | 33.03 | 85.43 | 97.77 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/vision_transformer/vit-base-p16_ft-64xb64_in1k-384.py) | [model](https://download.openmmlab.com/mmclassification/v0/vit/finetune/vit-base-p16_in21k-pre-3rdparty_ft-64xb64_in1k-384_20210928-98e8652b.pth)| | ViT-B32\* | ImageNet-21k | 384x384 | 88.30 | 8.56 | 84.01 | 97.08 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/vision_transformer/vit-base-p32_ft-64xb64_in1k-384.py) | [model](https://download.openmmlab.com/mmclassification/v0/vit/finetune/vit-base-p32_in21k-pre-3rdparty_ft-64xb64_in1k-384_20210928-9cea8599.pth)| | ViT-L16\* | ImageNet-21k | 384x384 | 304.72 | 116.68 | 85.63 | 97.63 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/vision_transformer/vit-large-p16_ft-64xb64_in1k-384.py) | [model](https://download.openmmlab.com/mmclassification/v0/vit/finetune/vit-large-p16_in21k-pre-3rdparty_ft-64xb64_in1k-384_20210928-b20ba619.pth)| -*Models with \* are converted from other repos.* +*Models with \* are converted from the [official repo](https://github.com/google-research/vision_transformer#available-vit-models). The config files of these models are only for validation. We don't ensure these config files' training accuracy and welcome you to contribute your reproduction results.* diff --git a/docs/en/model_zoo.md b/docs/en/model_zoo.md index 782ddacb..a8058821 100644 --- a/docs/en/model_zoo.md +++ b/docs/en/model_zoo.md @@ -15,30 +15,30 @@ The ResNet family models below are trained by standard data augmentations, i.e., | VGG-13-BN | 133.05 | 11.36 | 72.15 | 90.71 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/vgg/vgg13bn_8xb32_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/vgg/vgg13_bn_batch256_imagenet_20210207-1a8b7864.pth) | [log](https://download.openmmlab.com/mmclassification/v0/vgg/vgg13_bn_batch256_imagenet_20210207-1a8b7864.log.json) | | VGG-16-BN | 138.37 | 15.53 | 73.72 | 91.68 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/vgg/vgg16_8xb32_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/vgg/vgg16_bn_batch256_imagenet_20210208-7e55cd29.pth) | [log](https://download.openmmlab.com/mmclassification/v0/vgg/vgg16_bn_batch256_imagenet_20210208-7e55cd29.log.json) | | VGG-19-BN | 143.68 | 19.7 | 74.70 | 92.24 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/vgg/vgg19bn_8xb32_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/vgg/vgg19_bn_batch256_imagenet_20210208-da620c4f.pth) | [log](https://download.openmmlab.com/mmclassification/v0/vgg/vgg19_bn_batch256_imagenet_20210208-da620c4f.log.json)| -| RepVGG-A0\* | 9.11（train) | 8.31 (deploy) | 1.52 (train) | 1.36 (deploy) | 72.41 | 90.50 | [config (train)](https://github.com/open-mmlab/mmclassification/blob/master/configs/repvgg/repvgg-A0_4xb64-coslr-120e_in1k.py) | [config (deploy)](https://github.com/open-mmlab/mmclassification/blob/master/configs/repvgg/deploy/repvgg-A0_deploy_4xb64-coslr-120e_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/repvgg/repvgg-A0_3rdparty_4xb64-coslr-120e_in1k_20210909-883ab98c.pth) | [log]() | -| RepVGG-A1\* | 14.09 (train) | 12.79 (deploy) | 2.64 (train) | 2.37 (deploy) | 74.47 | 91.85 | [config (train)](https://github.com/open-mmlab/mmclassification/blob/master/configs/repvgg/repvgg-A1_4xb64-coslr-120e_in1k.py) | [config (deploy)](https://github.com/open-mmlab/mmclassification/blob/master/configs/repvgg/deploy/repvgg-A1_deploy_4xb64-coslr-120e_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/repvgg/repvgg-A1_3rdparty_4xb64-coslr-120e_in1k_20210909-24003a24.pth) | [log]() | -| RepVGG-A2\* | 28.21 (train) | 25.5 (deploy) | 5.7 (train) | 5.12 (deploy) | 76.48 | 93.01 | [config (train)](https://github.com/open-mmlab/mmclassification/blob/master/configs/repvgg/repvgg-A2_4xb64-coslr-120e_in1k.py) | [config (deploy)](https://github.com/open-mmlab/mmclassification/blob/master/configs/repvgg/deploy/repvgg-A2_deploy_4xb64-coslr-120e_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/repvgg/repvgg-A2_3rdparty_4xb64-coslr-120e_in1k_20210909-97d7695a.pth) | [log]() | -| RepVGG-B0\* | 15.82 (train) | 14.34 (deploy) | 3.42 (train) | 3.06 (deploy) | 75.14 | 92.42 | [config (train)](https://github.com/open-mmlab/mmclassification/blob/master/configs/repvgg/repvgg-B0_4xb64-coslr-120e_in1k.py) | [config (deploy)](https://github.com/open-mmlab/mmclassification/blob/master/configs/repvgg/deploy/repvgg-B0_deploy_4xb64-coslr-120e_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/repvgg/repvgg-B0_3rdparty_4xb64-coslr-120e_in1k_20210909-446375f4.pth) | [log]() | -| RepVGG-B1\* | 57.42 (train) | 51.83 (deploy) | 13.16 (train) | 11.82 (deploy) | 78.37 | 94.11 | [config (train)](https://github.com/open-mmlab/mmclassification/blob/master/configs/repvgg/repvgg-B1_4xb64-coslr-120e_in1k.py) | [config (deploy)](https://github.com/open-mmlab/mmclassification/blob/master/configs/repvgg/deploy/repvgg-B1_deploy_4xb64-coslr-120e_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/repvgg/repvgg-B1_3rdparty_4xb64-coslr-120e_in1k_20210909-750cdf67.pth) | [log]() | -| RepVGG-B1g2\* | 45.78 (train) | 41.36 (deploy) | 9.82 (train) | 8.82 (deploy) | 77.79 | 93.88 | [config (train)](https://github.com/open-mmlab/mmclassification/blob/master/configs/repvgg/repvgg-B1g2_4xb64-coslr-120e_in1k.py) | [config (deploy)](https://github.com/open-mmlab/mmclassification/blob/master/configs/repvgg/deploy/repvgg-B1g2_deploy_4xb64-coslr-120e_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/repvgg/repvgg-B1g2_3rdparty_4xb64-coslr-120e_in1k_20210909-344f6422.pth) | [log]() | -| RepVGG-B1g4\* | 39.97 (train) | 36.13 (deploy) | 8.15 (train) | 7.32 (deploy) | 77.58 | 93.84 | [config (train)](https://github.com/open-mmlab/mmclassification/blob/master/configs/repvgg/repvgg-B1g4_4xb64-coslr-120e_in1k.py) | [config (deploy)](https://github.com/open-mmlab/mmclassification/blob/master/configs/repvgg/deploy/repvgg-B1g4_deploy_4xb64-coslr-120e_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/repvgg/repvgg-B1g4_3rdparty_4xb64-coslr-120e_in1k_20210909-d4c1a642.pth) | [log]() | -| RepVGG-B2\* | 89.02 (train) | 80.32 (deploy) | 20.46 (train) | 18.39 (deploy) | 78.78 | 94.42 | [config (train)](https://github.com/open-mmlab/mmclassification/blob/master/configs/repvgg/repvgg-B2_4xb64-coslr-120e_in1k.py) | [config (deploy)](https://github.com/open-mmlab/mmclassification/blob/master/configs/repvgg/deploy/repvgg-B2_deploy_4xb64-coslr-120e_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/repvgg/repvgg-B2_3rdparty_4xb64-coslr-120e_in1k_20210909-bd6b937c.pth) | [log]() | -| RepVGG-B2g4\* | 61.76 (train) | 55.78 (deploy) | 12.63 (train) | 11.34 (deploy) | 79.38 | 94.68 | [config (train)](https://github.com/open-mmlab/mmclassification/blob/master/configs/repvgg/repvgg-B2g4_4xb64-autoaug-lbs-mixup-coslr-200e_in1k.py) | [config (deploy)](https://github.com/open-mmlab/mmclassification/blob/master/configs/repvgg/deploy/repvgg-B2g4_deploy_4xb64-autoaug-lbs-mixup-coslr-200e_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/repvgg/repvgg-B2g4_3rdparty_4xb64-autoaug-lbs-mixup-coslr-200e_in1k_20210909-7b7955f0.pth) | [log]() | -| RepVGG-B3\* | 123.09 (train) | 110.96 (deploy) | 29.17 (train) | 26.22 (deploy) | 80.52 | 95.26 | [config (train)](https://github.com/open-mmlab/mmclassification/blob/master/configs/repvgg/repvgg-B3_4xb64-autoaug-lbs-mixup-coslr-200e_in1k.py) | [config (deploy)](https://github.com/open-mmlab/mmclassification/blob/master/configs/repvgg/deploy/repvgg-B3_deploy_4xb64-autoaug-lbs-mixup-coslr-200e_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/repvgg/repvgg-B3_3rdparty_4xb64-autoaug-lbs-mixup-coslr-200e_in1k_20210909-dda968bf.pth) | [log]() | -| RepVGG-B3g4\* | 83.83 (train) | 75.63 (deploy) | 17.9 (train) | 16.08 (deploy) | 80.22 | 95.10 | [config (train)](https://github.com/open-mmlab/mmclassification/blob/master/configs/repvgg/repvgg-B3g4_4xb64-autoaug-lbs-mixup-coslr-200e_in1k.py) | [config (deploy)](https://github.com/open-mmlab/mmclassification/blob/master/configs/repvgg/deploy/repvgg-B3g4_deploy_4xb64-autoaug-lbs-mixup-coslr-200e_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/repvgg/repvgg-B3g4_3rdparty_4xb64-autoaug-lbs-mixup-coslr-200e_in1k_20210909-4e54846a.pth) | [log]() | -| RepVGG-D2se\* | 133.33 (train) | 120.39 (deploy) | 36.56 (train) | 32.85 (deploy) | 81.81 | 95.94 | [config (train)](https://github.com/open-mmlab/mmclassification/blob/master/configs/repvgg/repvgg-D2se_4xb64-autoaug-lbs-mixup-coslr-200e_in1k.py) | [config (deploy)](https://github.com/open-mmlab/mmclassification/blob/master/configs/repvgg/deploy/repvgg-D2se_deploy_4xb64-autoaug-lbs-mixup-coslr-200e_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/repvgg/repvgg-D2se_3rdparty_4xb64-autoaug-lbs-mixup-coslr-200e_in1k_20210909-cf3139b7.pth) | [log]() | +| RepVGG-A0\* | 9.11（train) | 8.31 (deploy) | 1.52 (train) | 1.36 (deploy) | 72.41 | 90.50 | [config (train)](https://github.com/open-mmlab/mmclassification/blob/master/configs/repvgg/repvgg-A0_4xb64-coslr-120e_in1k.py) | [config (deploy)](https://github.com/open-mmlab/mmclassification/blob/master/configs/repvgg/deploy/repvgg-A0_deploy_4xb64-coslr-120e_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/repvgg/repvgg-A0_3rdparty_4xb64-coslr-120e_in1k_20210909-883ab98c.pth) | +| RepVGG-A1\* | 14.09 (train) | 12.79 (deploy) | 2.64 (train) | 2.37 (deploy) | 74.47 | 91.85 | [config (train)](https://github.com/open-mmlab/mmclassification/blob/master/configs/repvgg/repvgg-A1_4xb64-coslr-120e_in1k.py) | [config (deploy)](https://github.com/open-mmlab/mmclassification/blob/master/configs/repvgg/deploy/repvgg-A1_deploy_4xb64-coslr-120e_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/repvgg/repvgg-A1_3rdparty_4xb64-coslr-120e_in1k_20210909-24003a24.pth) | +| RepVGG-A2\* | 28.21 (train) | 25.5 (deploy) | 5.7 (train) | 5.12 (deploy) | 76.48 | 93.01 | [config (train)](https://github.com/open-mmlab/mmclassification/blob/master/configs/repvgg/repvgg-A2_4xb64-coslr-120e_in1k.py) | [config (deploy)](https://github.com/open-mmlab/mmclassification/blob/master/configs/repvgg/deploy/repvgg-A2_deploy_4xb64-coslr-120e_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/repvgg/repvgg-A2_3rdparty_4xb64-coslr-120e_in1k_20210909-97d7695a.pth) | +| RepVGG-B0\* | 15.82 (train) | 14.34 (deploy) | 3.42 (train) | 3.06 (deploy) | 75.14 | 92.42 | [config (train)](https://github.com/open-mmlab/mmclassification/blob/master/configs/repvgg/repvgg-B0_4xb64-coslr-120e_in1k.py) | [config (deploy)](https://github.com/open-mmlab/mmclassification/blob/master/configs/repvgg/deploy/repvgg-B0_deploy_4xb64-coslr-120e_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/repvgg/repvgg-B0_3rdparty_4xb64-coslr-120e_in1k_20210909-446375f4.pth) | +| RepVGG-B1\* | 57.42 (train) | 51.83 (deploy) | 13.16 (train) | 11.82 (deploy) | 78.37 | 94.11 | [config (train)](https://github.com/open-mmlab/mmclassification/blob/master/configs/repvgg/repvgg-B1_4xb64-coslr-120e_in1k.py) | [config (deploy)](https://github.com/open-mmlab/mmclassification/blob/master/configs/repvgg/deploy/repvgg-B1_deploy_4xb64-coslr-120e_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/repvgg/repvgg-B1_3rdparty_4xb64-coslr-120e_in1k_20210909-750cdf67.pth) | +| RepVGG-B1g2\* | 45.78 (train) | 41.36 (deploy) | 9.82 (train) | 8.82 (deploy) | 77.79 | 93.88 | [config (train)](https://github.com/open-mmlab/mmclassification/blob/master/configs/repvgg/repvgg-B1g2_4xb64-coslr-120e_in1k.py) | [config (deploy)](https://github.com/open-mmlab/mmclassification/blob/master/configs/repvgg/deploy/repvgg-B1g2_deploy_4xb64-coslr-120e_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/repvgg/repvgg-B1g2_3rdparty_4xb64-coslr-120e_in1k_20210909-344f6422.pth) | +| RepVGG-B1g4\* | 39.97 (train) | 36.13 (deploy) | 8.15 (train) | 7.32 (deploy) | 77.58 | 93.84 | [config (train)](https://github.com/open-mmlab/mmclassification/blob/master/configs/repvgg/repvgg-B1g4_4xb64-coslr-120e_in1k.py) | [config (deploy)](https://github.com/open-mmlab/mmclassification/blob/master/configs/repvgg/deploy/repvgg-B1g4_deploy_4xb64-coslr-120e_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/repvgg/repvgg-B1g4_3rdparty_4xb64-coslr-120e_in1k_20210909-d4c1a642.pth) | +| RepVGG-B2\* | 89.02 (train) | 80.32 (deploy) | 20.46 (train) | 18.39 (deploy) | 78.78 | 94.42 | [config (train)](https://github.com/open-mmlab/mmclassification/blob/master/configs/repvgg/repvgg-B2_4xb64-coslr-120e_in1k.py) | [config (deploy)](https://github.com/open-mmlab/mmclassification/blob/master/configs/repvgg/deploy/repvgg-B2_deploy_4xb64-coslr-120e_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/repvgg/repvgg-B2_3rdparty_4xb64-coslr-120e_in1k_20210909-bd6b937c.pth) | +| RepVGG-B2g4\* | 61.76 (train) | 55.78 (deploy) | 12.63 (train) | 11.34 (deploy) | 79.38 | 94.68 | [config (train)](https://github.com/open-mmlab/mmclassification/blob/master/configs/repvgg/repvgg-B2g4_4xb64-autoaug-lbs-mixup-coslr-200e_in1k.py) | [config (deploy)](https://github.com/open-mmlab/mmclassification/blob/master/configs/repvgg/deploy/repvgg-B2g4_deploy_4xb64-autoaug-lbs-mixup-coslr-200e_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/repvgg/repvgg-B2g4_3rdparty_4xb64-autoaug-lbs-mixup-coslr-200e_in1k_20210909-7b7955f0.pth) | +| RepVGG-B3\* | 123.09 (train) | 110.96 (deploy) | 29.17 (train) | 26.22 (deploy) | 80.52 | 95.26 | [config (train)](https://github.com/open-mmlab/mmclassification/blob/master/configs/repvgg/repvgg-B3_4xb64-autoaug-lbs-mixup-coslr-200e_in1k.py) | [config (deploy)](https://github.com/open-mmlab/mmclassification/blob/master/configs/repvgg/deploy/repvgg-B3_deploy_4xb64-autoaug-lbs-mixup-coslr-200e_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/repvgg/repvgg-B3_3rdparty_4xb64-autoaug-lbs-mixup-coslr-200e_in1k_20210909-dda968bf.pth) | +| RepVGG-B3g4\* | 83.83 (train) | 75.63 (deploy) | 17.9 (train) | 16.08 (deploy) | 80.22 | 95.10 | [config (train)](https://github.com/open-mmlab/mmclassification/blob/master/configs/repvgg/repvgg-B3g4_4xb64-autoaug-lbs-mixup-coslr-200e_in1k.py) | [config (deploy)](https://github.com/open-mmlab/mmclassification/blob/master/configs/repvgg/deploy/repvgg-B3g4_deploy_4xb64-autoaug-lbs-mixup-coslr-200e_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/repvgg/repvgg-B3g4_3rdparty_4xb64-autoaug-lbs-mixup-coslr-200e_in1k_20210909-4e54846a.pth) | +| RepVGG-D2se\* | 133.33 (train) | 120.39 (deploy) | 36.56 (train) | 32.85 (deploy) | 81.81 | 95.94 | [config (train)](https://github.com/open-mmlab/mmclassification/blob/master/configs/repvgg/repvgg-D2se_4xb64-autoaug-lbs-mixup-coslr-200e_in1k.py) | [config (deploy)](https://github.com/open-mmlab/mmclassification/blob/master/configs/repvgg/deploy/repvgg-D2se_deploy_4xb64-autoaug-lbs-mixup-coslr-200e_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/repvgg/repvgg-D2se_3rdparty_4xb64-autoaug-lbs-mixup-coslr-200e_in1k_20210909-cf3139b7.pth) | | ResNet-18 | 11.69 | 1.82 | 70.07 | 89.44 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/resnet/resnet18_8xb32_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/resnet/resnet18_batch256_imagenet_20200708-34ab8f90.pth) | [log](https://download.openmmlab.com/mmclassification/v0/resnet/resnet18_batch256_imagenet_20200708-34ab8f90.log.json) | | ResNet-34 | 21.8 | 3.68 | 73.85 | 91.53 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/resnet/resnet34_8xb32_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/resnet/resnet34_batch256_imagenet_20200708-32ffb4f7.pth) | [log](https://download.openmmlab.com/mmclassification/v0/resnet/resnet34_batch256_imagenet_20200708-32ffb4f7.log.json) | | ResNet-50 | 25.56 | 4.12 | 76.55 | 93.15 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/resnet/resnet50_8xb32_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/resnet/resnet50_batch256_imagenet_20200708-cfb998bf.pth) | [log](https://download.openmmlab.com/mmclassification/v0/resnet/resnet50_batch256_imagenet_20200708-cfb998bf.log.json) | | ResNet-101 | 44.55 | 7.85 | 78.18 | 94.03 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/resnet/resnet101_8xb32_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/resnet/resnet101_batch256_imagenet_20200708-753f3608.pth) | [log](https://download.openmmlab.com/mmclassification/v0/resnet/resnet101_batch256_imagenet_20200708-753f3608.log.json) | | ResNet-152 | 60.19 | 11.58 | 78.63 | 94.16 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/resnet/resnet152_8xb32_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/resnet/resnet152_batch256_imagenet_20200708-ec25b1f9.pth) | [log](https://download.openmmlab.com/mmclassification/v0/resnet/resnet152_batch256_imagenet_20200708-ec25b1f9.log.json) | -| Res2Net-50-14w-8s\* | 25.06 | 4.22 | 78.14 | 93.85 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/res2net/res2net50-w14-s8_8xb32_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/res2net/res2net50-w14-s8_3rdparty_8xb32_in1k_20210927-bc967bf1.pth) | [log]()| -| Res2Net-50-26w-8s\* | 48.40 | 8.39 | 79.20 | 94.36 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/res2net/res2net50-w26-s8_8xb32_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/res2net/res2net50-w26-s8_3rdparty_8xb32_in1k_20210927-f547a94b.pth) | [log]()| -| Res2Net-101-26w-4s\* | 45.21 | 8.12 | 79.19 | 94.44 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/res2net/res2net101-w26-s4_8xb32_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/res2net/res2net101-w26-s4_3rdparty_8xb32_in1k_20210927-870b6c36.pth) | [log]()| -| ResNeSt-50\* | 27.48 | 5.41 | 81.13 | 95.59 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/resnest/resnest50_32xb64_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/resnest/resnest50_imagenet_converted-1ebf0afe.pth) | [log]() | -| ResNeSt-101\* | 48.28 | 10.27 | 82.32 | 96.24 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/resnest/resnest101_32xb64_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/resnest/resnest101_imagenet_converted-032caa52.pth) | [log]() | -| ResNeSt-200\* | 70.2 | 17.53 | 82.41 | 96.22 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/resnest/resnest200_64xb32_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/resnest/resnest200_imagenet_converted-581a60f2.pth) | [log]() | -| ResNeSt-269\* | 110.93 | 22.58 | 82.70 | 96.28 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/resnest/resnest269_64xb32_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/resnest/resnest269_imagenet_converted-59930960.pth) | [log]() | +| Res2Net-50-14w-8s\* | 25.06 | 4.22 | 78.14 | 93.85 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/res2net/res2net50-w14-s8_8xb32_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/res2net/res2net50-w14-s8_3rdparty_8xb32_in1k_20210927-bc967bf1.pth) | +| Res2Net-50-26w-8s\* | 48.40 | 8.39 | 79.20 | 94.36 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/res2net/res2net50-w26-s8_8xb32_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/res2net/res2net50-w26-s8_3rdparty_8xb32_in1k_20210927-f547a94b.pth) | +| Res2Net-101-26w-4s\* | 45.21 | 8.12 | 79.19 | 94.44 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/res2net/res2net101-w26-s4_8xb32_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/res2net/res2net101-w26-s4_3rdparty_8xb32_in1k_20210927-870b6c36.pth) | +| ResNeSt-50\* | 27.48 | 5.41 | 81.13 | 95.59 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/resnest/resnest50_32xb64_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/resnest/resnest50_imagenet_converted-1ebf0afe.pth)| +| ResNeSt-101\* | 48.28 | 10.27 | 82.32 | 96.24 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/resnest/resnest101_32xb64_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/resnest/resnest101_imagenet_converted-032caa52.pth)| +| ResNeSt-200\* | 70.2 | 17.53 | 82.41 | 96.22 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/resnest/resnest200_64xb32_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/resnest/resnest200_imagenet_converted-581a60f2.pth)| +| ResNeSt-269\* | 110.93 | 22.58 | 82.70 | 96.28 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/resnest/resnest269_64xb32_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/resnest/resnest269_imagenet_converted-59930960.pth)| | ResNetV1D-50 | 25.58 | 4.36 | 77.54 | 93.57 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/resnet/resnetv1d50_8xb32_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/resnet/resnetv1d50_b32x8_imagenet_20210531-db14775a.pth) | [log](https://download.openmmlab.com/mmclassification/v0/resnet/resnetv1d50_b32x8_imagenet_20210531-db14775a.log.json) | | ResNetV1D-101 | 44.57 | 8.09 | 78.93 | 94.48 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/resnet/resnetv1d101_8xb32_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/resnet/resnetv1d101_b32x8_imagenet_20210531-6e13bcd3.pth) | [log](https://download.openmmlab.com/mmclassification/v0/resnet/resnetv1d101_b32x8_imagenet_20210531-6e13bcd3.log.json) | | ResNetV1D-152 | 60.21 | 11.82 | 79.41 | 94.7 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/resnet/resnetv1d152_8xb32_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/resnet/resnetv1d152_b32x8_imagenet_20210531-278cf22a.pth) | [log](https://download.openmmlab.com/mmclassification/v0/resnet/resnetv1d152_b32x8_imagenet_20210531-278cf22a.log.json) | @@ -51,32 +51,32 @@ The ResNet family models below are trained by standard data augmentations, i.e., | ShuffleNetV1 1.0x (group=3) | 1.87 | 0.146 | 68.13 | 87.81 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/shufflenet_v1/shufflenet-v1-1x_16xb64_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/shufflenet_v1/shufflenet_v1_batch1024_imagenet_20200804-5d6cec73.pth) | [log](https://download.openmmlab.com/mmclassification/v0/shufflenet_v1/shufflenet_v1_batch1024_imagenet_20200804-5d6cec73.log.json) | | ShuffleNetV2 1.0x | 2.28 | 0.149 | 69.55 | 88.92 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/shufflenet_v2/shufflenet-v2-1x_16xb64_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/shufflenet_v2/shufflenet_v2_batch1024_imagenet_20200812-5bf4721e.pth) | [log](https://download.openmmlab.com/mmclassification/v0/shufflenet_v2/shufflenet_v2_batch1024_imagenet_20200804-8860eec9.log.json) | | MobileNet V2 | 3.5 | 0.319 | 71.86 | 90.42 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/mobilenet_v2/mobilenet-v2_8xb32_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/mobilenet_v2/mobilenet_v2_batch256_imagenet_20200708-3b2dc3af.pth) | [log](https://download.openmmlab.com/mmclassification/v0/mobilenet_v2/mobilenet_v2_batch256_imagenet_20200708-3b2dc3af.log.json) | -| ViT-B/16\* | 86.86 | 33.03 | 85.43 | 97.77 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/vision_transformer/vit-base-p16_ft-64xb64_in1k-384.py) | [model](https://download.openmmlab.com/mmclassification/v0/vit/finetune/vit-base-p16_in21k-pre-3rdparty_ft-64xb64_in1k-384_20210928-98e8652b.pth) | [log]() | -| ViT-B/32\* | 88.3 | 8.56 | 84.01 | 97.08 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/vision_transformer/vit-base-p32_ft-64xb64_in1k-384.py) | [model](https://download.openmmlab.com/mmclassification/v0/vit/finetune/vit-base-p32_in21k-pre-3rdparty_ft-64xb64_in1k-384_20210928-9cea8599.pth) | [log]() | -| ViT-L/16\* | 304.72 | 116.68 | 85.63 | 97.63 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/vision_transformer/vit-large-p16_ft-64xb64_in1k-384.py) | [model](https://download.openmmlab.com/mmclassification/v0/vit/finetune/vit-large-p16_in21k-pre-3rdparty_ft-64xb64_in1k-384_20210928-b20ba619.pth) | [log]() | +| ViT-B/16\* | 86.86 | 33.03 | 85.43 | 97.77 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/vision_transformer/vit-base-p16_ft-64xb64_in1k-384.py) | [model](https://download.openmmlab.com/mmclassification/v0/vit/finetune/vit-base-p16_in21k-pre-3rdparty_ft-64xb64_in1k-384_20210928-98e8652b.pth)| +| ViT-B/32\* | 88.3 | 8.56 | 84.01 | 97.08 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/vision_transformer/vit-base-p32_ft-64xb64_in1k-384.py) | [model](https://download.openmmlab.com/mmclassification/v0/vit/finetune/vit-base-p32_in21k-pre-3rdparty_ft-64xb64_in1k-384_20210928-9cea8599.pth)| +| ViT-L/16\* | 304.72 | 116.68 | 85.63 | 97.63 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/vision_transformer/vit-large-p16_ft-64xb64_in1k-384.py) | [model](https://download.openmmlab.com/mmclassification/v0/vit/finetune/vit-large-p16_in21k-pre-3rdparty_ft-64xb64_in1k-384_20210928-b20ba619.pth)| | Swin-Transformer tiny | 28.29 | 4.36 | 81.18 | 95.61 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/swin_transformer/swin-tiny_16xb64_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/swin-transformer/swin_tiny_224_b16x64_300e_imagenet_20210616_090925-66df6be6.pth) | [log](https://download.openmmlab.com/mmclassification/v0/swin-transformer/swin_tiny_224_b16x64_300e_imagenet_20210616_090925.log.json)| | Swin-Transformer small| 49.61 | 8.52 | 83.02 | 96.29 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/swin_transformer/swin-small_16xb64_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/swin-transformer/swin_small_224_b16x64_300e_imagenet_20210615_110219-7f9d988b.pth) | [log](https://download.openmmlab.com/mmclassification/v0/swin-transformer/swin_small_224_b16x64_300e_imagenet_20210615_110219.log.json)| | Swin-Transformer base | 87.77 | 15.14 | 83.36 | 96.44 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/swin_transformer/swin_base_224_b16x64_300e_imagenet.py) | [model](https://download.openmmlab.com/mmclassification/v0/swin-transformer/swin_base_224_b16x64_300e_imagenet_20210616_190742-93230b0d.pth) | [log](https://download.openmmlab.com/mmclassification/v0/swin-transformer/swin_base_224_b16x64_300e_imagenet_20210616_190742.log.json)| -| Transformer in Transformer small\* | 23.76 | 3.36 | 81.52 | 95.73 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/tnt/tnt-s-p16_16xb64_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/tnt/tnt-small-p16_3rdparty_in1k_20210903-c56ee7df.pth) | [log]()| -| T2T-ViT_t-14\* | 21.47 | 4.34 | 81.69 | 95.85 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/t2t_vit/t2t-vit-t-14_8xb64_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/t2t-vit/t2t-vit-t-14_3rdparty_8xb64_in1k_20210928-b7c09b62.pth) | [log]()| -| T2T-ViT_t-19\* | 39.08 | 7.80 | 82.43 | 96.08 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/t2t_vit/t2t-vit-t-19_8xb64_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/t2t-vit/t2t-vit-t-19_3rdparty_8xb64_in1k_20210928-7f1478d5.pth) | [log]()| -| T2T-ViT_t-24\* | 64.00 | 12.69 | 82.55 | 96.06 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/t2t_vit/t2t-vit-t-24_8xb64_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/t2t-vit/t2t-vit-t-24_3rdparty_8xb64_in1k_20210928-fe95a61b.pth) | [log]()| -| Mixer-B/16\* | 59.88 | 12.61 | 76.68 | 92.25 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/mlp_mixer/mlp-mixer-base-p16_64xb64_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/mlp-mixer/mixer-base-p16_3rdparty_64xb64_in1k_20211124-1377e3e0.pth) | [log]()| -| Mixer-L/16\* | 208.2 | 44.57 | 72.34 | 88.02 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/mlp_mixer/mlp-mixer-large-p16_64xb64_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/mlp-mixer/mixer-large-p16_3rdparty_64xb64_in1k_20211124-5a2519d2.pth) | [log]()| -| DeiT-tiny\* | 5.72 | 1.08 | 72.13 | 91.13 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/deit/deit-tiny_pt-4xb256_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/deit/deit-tiny_3rdparty_pt-4xb256_in1k_20211124-e930093b.pth) | [log]()| -| DeiT-tiny distilled\* | 5.72 | 1.08 | 74.51 | 91.90 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/deit/deit-tiny-distilled_pt-4xb256_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/deit/deit-tiny-distilled_3rdparty_pt-4xb256_in1k_20211216-c429839a.pth) | [log]()| -| DeiT-small\* | 22.05 | 4.24 | 79.83 | 94.95 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/deit/deit-small_pt-4xb256_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/deit/deit-small_3rdparty_pt-4xb256_in1k_20211124-ffe94edd.pth) | [log]()| -| DeiT-small distilled\* | 22.05 | 4.24 | 81.17 | 95.40 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/deit/deit-small-distilled_pt-4xb256_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/deit/deit-small-distilled_3rdparty_pt-4xb256_in1k_20211216-4de1d725.pth) | [log]()| -| DeiT-base\* | 86.57 | 16.86 | 81.79 | 95.59 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/deit/deit-base_pt-16xb64_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/deit/deit-base_3rdparty_pt-16xb64_in1k_20211124-6f40c188.pth) | [log]()| -| DeiT-base distilled\* | 86.57 | 16.86 | 83.33 | 96.49 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/deit/deit-base-distilled_pt-16xb64_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/deit/deit-base-distilled_3rdparty_pt-16xb64_in1k_20211216-42891296.pth) | [log]()| -| DeiT-base 384px\* | 86.86 | 49.37 | 83.04 | 96.31 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/deit/deit-base_ft-16xb32_in1k-384px.py) | [model](https://download.openmmlab.com/mmclassification/v0/deit/deit-base_3rdparty_ft-16xb32_in1k-384px_20211124-822d02f2.pth) | [log]()| -| DeiT-base distilled 384px\* | 86.86 | 49.37 | 85.55 | 97.35 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/deit/deit-base-distilled_ft-16xb32_in1k-384px.py) | [model](https://download.openmmlab.com/mmclassification/v0/deit/deit-base-distilled_3rdparty_ft-16xb32_in1k-384px_20211216-e48d6000.pth) | [log]()| -| Conformer-tiny-p16\* | 23.52 | 4.90 | 81.31 | 95.60 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/conformer/conformer-tiny-p16_8xb128_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/conformer/conformer-tiny-p16_3rdparty_8xb128_in1k_20211206-f6860372.pth) | [log]()| -| Conformer-small-p32 | 38.85 | 7.09 | 81.96 | 96.02 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/conformer/conformer-small-p32_8xb128_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/conformer/conformer-small-p32_8xb128_in1k_20211206-947a0816.pth) | [log]()| -| Conformer-small-p16\* | 37.67 | 10.31 | 83.32 | 96.46 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/conformer/conformer-small-p16_8xb128_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/conformer/conformer-small-p16_3rdparty_8xb128_in1k_20211206-3065dcf5.pth) | [log]()| -| Conformer-base-p16\* | 83.29 | 22.89 | 83.82 | 96.59 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/conformer/conformer-base-p16_8xb128_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/conformer/conformer-base-p16_3rdparty_8xb128_in1k_20211206-bfdf8637.pth) | [log]()| +| Transformer in Transformer small\* | 23.76 | 3.36 | 81.52 | 95.73 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/tnt/tnt-s-p16_16xb64_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/tnt/tnt-small-p16_3rdparty_in1k_20210903-c56ee7df.pth) | +| T2T-ViT_t-14 | 21.47 | 4.34 | 81.83 | 95.84 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/t2t_vit/t2t-vit-t-14_8xb64_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/t2t-vit/t2t-vit-t-14_8xb64_in1k_20211220-f7378dd5.pth) | [log](https://download.openmmlab.com/mmclassification/v0/t2t-vit/t2t-vit-t-14_8xb64_in1k_20211220-f7378dd5.log.json)| +| T2T-ViT_t-19 | 39.08 | 7.80 | 82.63 | 96.18 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/t2t_vit/t2t-vit-t-19_8xb64_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/t2t-vit/t2t-vit-t-19_8xb64_in1k_20211214-7f5e3aaf.pth) | [log](https://download.openmmlab.com/mmclassification/v0/t2t-vit/t2t-vit-t-19_8xb64_in1k_20211214-7f5e3aaf.log.json)| +| T2T-ViT_t-24 | 64.00 | 12.69 | 82.71 | 96.09 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/t2t_vit/t2t-vit-t-24_8xb64_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/t2t-vit/t2t-vit-t-24_8xb64_in1k_20211214-b2a68ae3.pth) | [log](https://download.openmmlab.com/mmclassification/v0/t2t-vit/t2t-vit-t-24_8xb64_in1k_20211214-b2a68ae3.log.json)| +| Mixer-B/16\* | 59.88 | 12.61 | 76.68 | 92.25 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/mlp_mixer/mlp-mixer-base-p16_64xb64_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/mlp-mixer/mixer-base-p16_3rdparty_64xb64_in1k_20211124-1377e3e0.pth) | +| Mixer-L/16\* | 208.2 | 44.57 | 72.34 | 88.02 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/mlp_mixer/mlp-mixer-large-p16_64xb64_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/mlp-mixer/mixer-large-p16_3rdparty_64xb64_in1k_20211124-5a2519d2.pth) | +| DeiT-tiny\* | 5.72 | 1.08 | 72.13 | 91.13 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/deit/deit-tiny_pt-4xb256_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/deit/deit-tiny_3rdparty_pt-4xb256_in1k_20211124-e930093b.pth) | +| DeiT-tiny distilled\* | 5.72 | 1.08 | 74.51 | 91.90 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/deit/deit-tiny-distilled_pt-4xb256_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/deit/deit-tiny-distilled_3rdparty_pt-4xb256_in1k_20211216-c429839a.pth) | +| DeiT-small\* | 22.05 | 4.24 | 79.83 | 94.95 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/deit/deit-small_pt-4xb256_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/deit/deit-small_3rdparty_pt-4xb256_in1k_20211124-ffe94edd.pth) | +| DeiT-small distilled\* | 22.05 | 4.24 | 81.17 | 95.40 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/deit/deit-small-distilled_pt-4xb256_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/deit/deit-small-distilled_3rdparty_pt-4xb256_in1k_20211216-4de1d725.pth) | +| DeiT-base\* | 86.57 | 16.86 | 81.79 | 95.59 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/deit/deit-base_pt-16xb64_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/deit/deit-base_3rdparty_pt-16xb64_in1k_20211124-6f40c188.pth) | +| DeiT-base distilled\* | 86.57 | 16.86 | 83.33 | 96.49 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/deit/deit-base-distilled_pt-16xb64_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/deit/deit-base-distilled_3rdparty_pt-16xb64_in1k_20211216-42891296.pth) | +| DeiT-base 384px\* | 86.86 | 49.37 | 83.04 | 96.31 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/deit/deit-base_ft-16xb32_in1k-384px.py) | [model](https://download.openmmlab.com/mmclassification/v0/deit/deit-base_3rdparty_ft-16xb32_in1k-384px_20211124-822d02f2.pth) | +| DeiT-base distilled 384px\* | 86.86 | 49.37 | 85.55 | 97.35 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/deit/deit-base-distilled_ft-16xb32_in1k-384px.py) | [model](https://download.openmmlab.com/mmclassification/v0/deit/deit-base-distilled_3rdparty_ft-16xb32_in1k-384px_20211216-e48d6000.pth) | +| Conformer-tiny-p16\* | 23.52 | 4.90 | 81.31 | 95.60 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/conformer/conformer-tiny-p16_8xb128_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/conformer/conformer-tiny-p16_3rdparty_8xb128_in1k_20211206-f6860372.pth) | +| Conformer-small-p32\* | 38.85 | 7.09 | 81.96 | 96.02 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/conformer/conformer-small-p32_8xb128_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/conformer/conformer-small-p32_8xb128_in1k_20211206-947a0816.pth) | +| Conformer-small-p16\* | 37.67 | 10.31 | 83.32 | 96.46 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/conformer/conformer-small-p16_8xb128_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/conformer/conformer-small-p16_3rdparty_8xb128_in1k_20211206-3065dcf5.pth) | +| Conformer-base-p16\* | 83.29 | 22.89 | 83.82 | 96.59 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/conformer/conformer-base-p16_8xb128_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/conformer/conformer-base-p16_3rdparty_8xb128_in1k_20211206-bfdf8637.pth) | -Models with * are converted from other repos, others are trained by ourselves. +*Models with \* are converted from other repos, others are trained by ourselves.* ## CIFAR10 diff --git a/mmcls/core/hook/__init__.py b/mmcls/core/hook/__init__.py index d52e44b2..2c2dbdfc 100644 --- a/mmcls/core/hook/__init__.py +++ b/mmcls/core/hook/__init__.py @@ -1,5 +1,9 @@ # Copyright (c) OpenMMLab. All rights reserved. from .class_num_check_hook import ClassNumCheckHook +from .lr_updater import CosineAnnealingCooldownLrUpdaterHook from .precise_bn_hook import PreciseBNHook -__all__ = ['ClassNumCheckHook', 'PreciseBNHook'] +__all__ = [ + 'ClassNumCheckHook', 'PreciseBNHook', + 'CosineAnnealingCooldownLrUpdaterHook' +] diff --git a/mmcls/core/hook/lr_updater.py b/mmcls/core/hook/lr_updater.py new file mode 100644 index 00000000..021f66b5 --- /dev/null +++ b/mmcls/core/hook/lr_updater.py @@ -0,0 +1,83 @@ +# Copyright (c) OpenMMLab. All rights reserved. +from math import cos, pi + +from mmcv.runner.hooks import HOOKS, LrUpdaterHook + + +@HOOKS.register_module() +class CosineAnnealingCooldownLrUpdaterHook(LrUpdaterHook): + """Cosine annealing learning rate scheduler with cooldown. + + Args: + min_lr (float, optional): The minimum learning rate after annealing. + Defaults to None. + min_lr_ratio (float, optional): The minimum learning ratio after + nnealing. Defaults to None. + cool_down_ratio (float): The cooldown ratio. Defaults to 0.1. + cool_down_time (int): The cooldown time. Defaults to 10. + by_epoch (bool): If True, the learning rate changes epoch by epoch. If + False, the learning rate changes iter by iter. Defaults to True. + warmup (string, optional): Type of warmup used. It can be None (use no + warmup), 'constant', 'linear' or 'exp'. Defaults to None. + warmup_iters (int): The number of iterations or epochs that warmup + lasts. Defaults to 0. + warmup_ratio (float): LR used at the beginning of warmup equals to + ``warmup_ratio * initial_lr``. Defaults to 0.1. + warmup_by_epoch (bool): If True, the ``warmup_iters`` + means the number of epochs that warmup lasts, otherwise means the + number of iteration that warmup lasts. Defaults to False. + + Note: + You need to set one and only one of ``min_lr`` and ``min_lr_ratio``. + """ + + def __init__(self, + min_lr=None, + min_lr_ratio=None, + cool_down_ratio=0.1, + cool_down_time=10, + **kwargs): + assert (min_lr is None) ^ (min_lr_ratio is None) + self.min_lr = min_lr + self.min_lr_ratio = min_lr_ratio + self.cool_down_time = cool_down_time + self.cool_down_ratio = cool_down_ratio + super(CosineAnnealingCooldownLrUpdaterHook, self).__init__(**kwargs) + + def get_lr(self, runner, base_lr): + if self.by_epoch: + progress = runner.epoch + max_progress = runner.max_epochs + else: + progress = runner.iter + max_progress = runner.max_iters + + if self.min_lr_ratio is not None: + target_lr = base_lr * self.min_lr_ratio + else: + target_lr = self.min_lr + + if progress > max_progress - self.cool_down_time: + return target_lr * self.cool_down_ratio + else: + max_progress = max_progress - self.cool_down_time + + return annealing_cos(base_lr, target_lr, progress / max_progress) + + +def annealing_cos(start, end, factor, weight=1): + """Calculate annealing cos learning rate. + + Cosine anneal from `weight * start + (1 - weight) * end` to `end` as + percentage goes from 0.0 to 1.0. + + Args: + start (float): The starting learning rate of the cosine annealing. + end (float): The ending learing rate of the cosine annealing. + factor (float): The coefficient of `pi` when calculating the current + percentage. Range from 0.0 to 1.0. + weight (float, optional): The combination factor of `start` and `end` + when calculating the actual starting learning rate. Default to 1. + """ + cos_out = cos(pi * factor) + 1 + return end + 0.5 * weight * (start - end) * cos_out diff --git a/tests/test_runtime/test_hooks.py b/tests/test_runtime/test_hooks.py new file mode 100644 index 00000000..288c955f --- /dev/null +++ b/tests/test_runtime/test_hooks.py @@ -0,0 +1,157 @@ +import logging +import shutil +import tempfile + +import numpy as np +import pytest +import torch +import torch.nn as nn +from mmcv.runner import build_runner +from mmcv.runner.hooks import Hook, IterTimerHook +from torch.utils.data import DataLoader + +import mmcls.core # noqa: F401 + + +def _build_demo_runner_without_hook(runner_type='EpochBasedRunner', + max_epochs=1, + max_iters=None, + multi_optimziers=False): + + class Model(nn.Module): + + def __init__(self): + super().__init__() + self.linear = nn.Linear(2, 1) + self.conv = nn.Conv2d(3, 3, 3) + + def forward(self, x): + return self.linear(x) + + def train_step(self, x, optimizer, **kwargs): + return dict(loss=self(x)) + + def val_step(self, x, optimizer, **kwargs): + return dict(loss=self(x)) + + model = Model() + + if multi_optimziers: + optimizer = { + 'model1': + torch.optim.SGD(model.linear.parameters(), lr=0.02, momentum=0.95), + 'model2': + torch.optim.SGD(model.conv.parameters(), lr=0.01, momentum=0.9), + } + else: + optimizer = torch.optim.SGD(model.parameters(), lr=0.02, momentum=0.95) + + tmp_dir = tempfile.mkdtemp() + runner = build_runner( + dict(type=runner_type), + default_args=dict( + model=model, + work_dir=tmp_dir, + optimizer=optimizer, + logger=logging.getLogger(), + max_epochs=max_epochs, + max_iters=max_iters)) + return runner + + +def _build_demo_runner(runner_type='EpochBasedRunner', + max_epochs=1, + max_iters=None, + multi_optimziers=False): + + log_config = dict( + interval=1, hooks=[ + dict(type='TextLoggerHook'), + ]) + + runner = _build_demo_runner_without_hook(runner_type, max_epochs, + max_iters, multi_optimziers) + + runner.register_checkpoint_hook(dict(interval=1)) + runner.register_logger_hooks(log_config) + return runner + + +class ValueCheckHook(Hook): + + def __init__(self, check_dict, by_epoch=False): + super().__init__() + self.check_dict = check_dict + self.by_epoch = by_epoch + + def after_iter(self, runner): + if self.by_epoch: + return + if runner.iter in self.check_dict: + for attr, target in self.check_dict[runner.iter].items(): + value = eval(f'runner.{attr}') + assert np.isclose(value, target), \ + (f'The value of `runner.{attr}` is {value}, ' + f'not equals to {target}') + + def after_epoch(self, runner): + if not self.by_epoch: + return + if runner.epoch in self.check_dict: + for attr, target in self.check_dict[runner.epoch]: + value = eval(f'runner.{attr}') + assert np.isclose(value, target), \ + (f'The value of `runner.{attr}` is {value}, ' + f'not equals to {target}') + + +@pytest.mark.parametrize('multi_optimziers', (True, False)) +def test_cosine_cooldown_hook(multi_optimziers): + """xdoctest -m tests/test_hooks.py test_cosine_runner_hook.""" + loader = DataLoader(torch.ones((10, 2))) + runner = _build_demo_runner(multi_optimziers=multi_optimziers) + + # add momentum LR scheduler + hook_cfg = dict( + type='CosineAnnealingCooldownLrUpdaterHook', + by_epoch=False, + cool_down_time=2, + cool_down_ratio=0.1, + min_lr_ratio=0.1, + warmup_iters=2, + warmup_ratio=0.9) + runner.register_hook_from_cfg(hook_cfg) + runner.register_hook_from_cfg(dict(type='IterTimerHook')) + runner.register_hook(IterTimerHook()) + + if multi_optimziers: + check_hook = ValueCheckHook({ + 0: { + 'current_lr()["model1"][0]': 0.02, + 'current_lr()["model2"][0]': 0.01, + }, + 5: { + 'current_lr()["model1"][0]': 0.0075558491, + 'current_lr()["model2"][0]': 0.0037779246, + }, + 9: { + 'current_lr()["model1"][0]': 0.0002, + 'current_lr()["model2"][0]': 0.0001, + } + }) + else: + check_hook = ValueCheckHook({ + 0: { + 'current_lr()[0]': 0.02, + }, + 5: { + 'current_lr()[0]': 0.0075558491, + }, + 9: { + 'current_lr()[0]': 0.0002, + } + }) + runner.register_hook(check_hook, priority='LOWEST') + + runner.run([loader], [('train', 1)]) + shutil.rmtree(runner.work_dir)