mirrors
/
mmpretrain
mirror of https://github.com/open-mmlab/mmpretrain.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity
mmpretrain/configs/levit/metafile.yml

102 lines
3.5 KiB
YAML
Raw Normal View History

[Feature] Support LeViT backbone. (#1238) * 网络搭建完成、能正常推理 * 网络搭建完成、能正常推理 * 网络搭建完成、能正常推理 * 添加了模型转换未验证，配置文件 但有无法运行 * 模型转换、结构验证完成，可以推理出正确答案 * 推理精度与原论文一致 已完成转化 * 三个方法改为class 暂存 * 完成推理精度对齐 误差0.04 * 暂时使用的levit2mmcls * 训练跑通，训练相关参数未对齐 * '训练相关参数对齐'参数' * '修复训练时验证导致模型结构改变无法复原问题' * '修复训练时验证导致模型结构改变无法复原问题' * '添加mixup和labelsmooth' * '配置文件补齐' * 添加模型转换 * 添加meta文件 * 添加meta文件 * 删除demo.py测试文件 * 添加模型README文件 * docs文件回滚 * model-index删除末行空格 * 更新模型metafile * 更新metafile * 更新metafile * 更新README和metafile * 更新模型README * 更新模型metafile * Delete the model class and get_LeViT_model methods in the mmcls.models.backone.levit file * Change the class name to Google Code Style * use arch to provide default architectures * use nn.Conv2d * mmcv.cnn.fuse_conv_bn * modify some details * remove down_ops from the architectures. * remove init_weight function * Modify ambiguous variable names * Change the drop_path in config to drop_path_rate * Add unit test * remove train function * add unit test * modify nn.norm1d to build_norm_layer * update metafile and readme * Update configs and LeViT implementations. * Update README. * Add docstring and update unit tests. * Revert irrelative modification. * Fix unit tests * minor fix Co-authored-by: mzr1996 <mzr1996@163.com>
2023-01-17 17:43:42 +08:00
Collections:
- Name: LeViT
Metadata:
Training Data: ImageNet-1k
Architecture:
- 1x1 Convolution
- LeViT Attention Block
Paper:
Title: "LeViT: a Vision Transformer in ConvNet\u2019s Clothing for Faster Inference"
URL: https://arxiv.org/abs/2104.01136
README: configs/levit/README.md
Code:
[Docs] Update links (#1457) * update links * update readtherdocs * update * update * fix lint * update * update * update * update cov branch * update * update * update
2023-04-06 20:58:52 +08:00
URL: https://github.com/open-mmlab/mmpretrain/blob/main/mmpretrain/models/backbones/levit.py
[Feature] Support LeViT backbone. (#1238) * 网络搭建完成、能正常推理 * 网络搭建完成、能正常推理 * 网络搭建完成、能正常推理 * 添加了模型转换未验证，配置文件 但有无法运行 * 模型转换、结构验证完成，可以推理出正确答案 * 推理精度与原论文一致 已完成转化 * 三个方法改为class 暂存 * 完成推理精度对齐 误差0.04 * 暂时使用的levit2mmcls * 训练跑通，训练相关参数未对齐 * '训练相关参数对齐'参数' * '修复训练时验证导致模型结构改变无法复原问题' * '修复训练时验证导致模型结构改变无法复原问题' * '添加mixup和labelsmooth' * '配置文件补齐' * 添加模型转换 * 添加meta文件 * 添加meta文件 * 删除demo.py测试文件 * 添加模型README文件 * docs文件回滚 * model-index删除末行空格 * 更新模型metafile * 更新metafile * 更新metafile * 更新README和metafile * 更新模型README * 更新模型metafile * Delete the model class and get_LeViT_model methods in the mmcls.models.backone.levit file * Change the class name to Google Code Style * use arch to provide default architectures * use nn.Conv2d * mmcv.cnn.fuse_conv_bn * modify some details * remove down_ops from the architectures. * remove init_weight function * Modify ambiguous variable names * Change the drop_path in config to drop_path_rate * Add unit test * remove train function * add unit test * modify nn.norm1d to build_norm_layer * update metafile and readme * Update configs and LeViT implementations. * Update README. * Add docstring and update unit tests. * Revert irrelative modification. * Fix unit tests * minor fix Co-authored-by: mzr1996 <mzr1996@163.com>
2023-01-17 17:43:42 +08:00
Version: v1.0.0rc5
Models:
- Name: levit-128s_3rdparty_in1k
Metadata:
FLOPs: 310342496
Parameters: 7391290
Training Data: ImageNet-1k
In Collection: LeViT
Results:
- Dataset: ImageNet-1k
Metrics:
Top 1 Accuracy: 76.51
Top 5 Accuracy: 92.90
Task: Image Classification
Weights: https://download.openmmlab.com/mmclassification/v0/levit/levit-128s_3rdparty_in1k_20230117-e9fbd209.pth
Config: configs/levit/levit-128s_8xb256_in1k.py
Converted From:
Weights: https://dl.fbaipublicfiles.com/LeViT/LeViT-128S-96703c44.pth
Code: https://github.com/facebookresearch/LeViT
- Name: levit-128_3rdparty_in1k
Metadata:
FLOPs: 413060992
Parameters: 8828168
Training Data: ImageNet-1k
In Collection: LeViT
Results:
- Dataset: ImageNet-1k
Metrics:
Top 1 Accuracy: 78.58
Top 5 Accuracy: 93.95
Task: Image Classification
Weights: https://download.openmmlab.com/mmclassification/v0/levit/levit-128_3rdparty_in1k_20230117-3be02a02.pth
Config: configs/levit/levit-128_8xb256_in1k.py
Converted From:
Weights: https://dl.fbaipublicfiles.com/LeViT/LeViT-128-b88c2750.pth
Code: https://github.com/facebookresearch/LeViT
- Name: levit-192_3rdparty_in1k
Metadata:
FLOPs: 667860704
Parameters: 10561301
Training Data: ImageNet-1k
In Collection: LeViT
Results:
- Dataset: ImageNet-1k
Metrics:
Top 1 Accuracy: 79.86
Top 5 Accuracy: 94.75
Task: Image Classification
Weights: https://download.openmmlab.com/mmclassification/v0/levit/levit-192_3rdparty_in1k_20230117-8217a0f9.pth
Config: configs/levit/levit-192_8xb256_in1k.py
Converted From:
Weights: https://dl.fbaipublicfiles.com/LeViT/LeViT-192-92712e41.pth
Code: https://github.com/facebookresearch/LeViT
- Name: levit-256_3rdparty_in1k
Metadata:
FLOPs: 1141625216
Parameters: 18379852
Training Data: ImageNet-1k
In Collection: LeViT
Results:
- Dataset: ImageNet-1k
Metrics:
Top 1 Accuracy: 81.59
Top 5 Accuracy: 95.46
Task: Image Classification
Weights: https://download.openmmlab.com/mmclassification/v0/levit/levit-256_3rdparty_in1k_20230117-5ae2ce7d.pth
Config: configs/levit/levit-256_8xb256_in1k.py
Converted From:
Weights: https://dl.fbaipublicfiles.com/LeViT/LeViT-256-13b5763e.pth
Code: https://github.com/facebookresearch/LeViT
- Name: levit-384_3rdparty_in1k
Metadata:
FLOPs: 2372941568
Parameters: 38358300
Training Data: ImageNet-1k
In Collection: LeViT
Results:
- Dataset: ImageNet-1k
Metrics:
Top 1 Accuracy: 82.59
Top 5 Accuracy: 95.95
Task: Image Classification
Weights: https://download.openmmlab.com/mmclassification/v0/levit/levit-384_3rdparty_in1k_20230117-f3539cce.pth
Config: configs/levit/levit-384_8xb256_in1k.py
Converted From:
Weights: https://dl.fbaipublicfiles.com/LeViT/LeViT-384-9bdaf2e2.pth
Code: https://github.com/facebookresearch/LeViT