[Feature] add model script, training configs and training weights of ConvMixer

configs/convmixer/README.md

@@ -0,0 +1,103 @@
+
+# ConvMixer
+> [Patches Are All You Need?](https://arxiv.org/pdf/2201.09792.pdf)
+
+## Introduction
+
+Although convolutional networks have been the dominant architecture for vision
+tasks for many years, recent experiments have shown that Transformer-based models, most notably the Vision Transformer (ViT), may exceed their performance in
+some settings. However, due to the quadratic runtime of the self-attention layers
+in Transformers, ViTs require the use of patch embeddings, which group together
+small regions of the image into single input features, in order to be applied to
+larger image sizes. This raises a question: Is the performance of ViTs due to the
+inherently-more-powerful Transformer architecture, or is it at least partly due to
+using patches as the input representation? In this paper, we present some evidence
+for the latter: specifically, we propose the ConvMixer, an extremely simple model
+that is similar in spirit to the ViT and the even-more-basic MLP-Mixer in that it
+operates directly on patches as input, separates the mixing of spatial and channel
+dimensions, and maintains equal size and resolution throughout the network. In
+contrast, however, the ConvMixer uses only standard convolutions to achieve the
+mixing steps. Despite its simplicity, we show that the ConvMixer outperforms the
+ViT, MLP-Mixer, and some of their variants for similar parameter counts and data
+set sizes, in addition to outperforming classical vision models such as the ResNet.
+
+
+## Results
+
+**Implementation and configs for training were taken and adjusted from [this repository](https://gitee.com/cvisionlab/models/tree/convmixer/release/research/cv/convmixer), which implements ConvMixer models in mindspore.**
+
+Our reproduced model performance on ImageNet-1K is reported as follows.
+
+<div align="center">
+
+| Model    | Context  | Top-1 (%) | Top-5 (%) | Params (M) | Recipe                                                                                        | Download                                                                               |
+|----------|----------|-----------|-----------|------------|-----------------------------------------------------------------------------------------------|----------------------------------------------------------------------------------------|
+| convmixer_768_32 | Converted from PyTorch | 79.68 | 94.92 | - | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/convmixer/convmixer_768_32_gpu.yaml) | [weights](https://storage.googleapis.com/huawei-mindspore-hk/ConvMixer/Converted/convmixer_768_32.ckpt) |
+| convmixer_768_32 | 8xRTX3090 | 73.05 | 90.53 | - | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/convmixer/convmixer_768_32_gpu.yaml) | [weights](https://storage.googleapis.com/huawei-mindspore-hk/ConvMixer/convmixer_768_32_trained.ckpt) |
+| convmixer_1024_20 | Converted from PyTorch | 76.68 | 93.3 | - | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/convmixer/convmixer_1024_20_gpu.yaml) | [weights](https://storage.googleapis.com/huawei-mindspore-hk/ConvMixer/Converted/convmixer_1024_20.ckpt) |
+| convmixer_1536_20 | Converted from PyTorch | 80.98 | 95.51 | - | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/convmixer/convmixer_1536_20_gpu.yaml) | [weights](https://storage.googleapis.com/huawei-mindspore-hk/ConvMixer/Converted/convmixer_1536_20.ckpt) |
+
+
+</div>
+
+#### Notes
+
+- Context: The weights in the table were taken from [official repository](https://github.com/locuslab/convmixer) and converted to mindspore format
+- Top-1 and Top-5: Accuracy reported on the validation set of ImageNet-1K.
+
+## Quick Start
+
+### Preparation
+
+#### Installation
+Please refer to the [installation instruction](https://github.com/mindspore-ecosystem/mindcv#installation) in MindCV.
+
+#### Dataset Preparation
+Please download the [ImageNet-1K](https://www.image-net.org/challenges/LSVRC/2012/index.php) dataset for model training and validation.
+
+### Training
+
+* Distributed Training
+
+
+```shell
+# distrubted training on multiple GPU/Ascend devices
+mpirun -n 8 python train.py --config configs/convmixer/convmixer_768_32_gpu.yaml --data_dir /path/to/imagenet --distributed True
+```
+
+> If the script is executed by the root user, the `--allow-run-as-root` parameter must be added to `mpirun`.
+
+Similarly, you can train the model on multiple GPU devices with the above `mpirun` command.
+
+For detailed illustration of all hyper-parameters, please refer to [config.py](https://github.com/mindspore-lab/mindcv/blob/main/config.py).
+
+**Note:**  As the global batch size  (batch_size x num_devices) is an important hyper-parameter, it is recommended to keep the global batch size unchanged for reproduction or adjust the learning rate linearly to a new global batch size.
+
+* Standalone Training
+
+If you want to train or finetune the model on a smaller dataset without distributed training, please run:
+
+```shell
+# standalone training on a CPU/GPU/Ascend device
+python train.py --config configs/convmixer/convmixer_768_32_gpu.yaml --data_dir /path/to/dataset --distribute False
+```
+
+### Validation
+
+To validate the accuracy of the trained model, you can use `validate.py` and parse the checkpoint path with `--ckpt_path`.
+
+```shell
+python validate.py -c configs/convmixer/convmixer_768_32_gpu.yaml --data_dir /path/to/imagenet --ckpt_path /path/to/ckpt
+```
+
+### Deployment
+
+Please refer to the [deployment tutorial](https://github.com/mindspore-lab/mindcv/blob/main/tutorials/deployment.md) in MindCV.
+
+## References
+
+Paper - https://arxiv.org/pdf/2201.09792.pdf
+
+Official repo - https://github.com/locuslab/convmixer
+
+Mindspore implementation - https://gitee.com/cvisionlab/models/tree/convmixer/release/research/cv/convmixer

configs/convmixer/convmixer_1024_20_gpu.yaml

@@ -0,0 +1,67 @@
+# system
+mode: 0
+distribute: False
+num_parallel_workers: 4
+val_while_train: True
+
+# dataset
+dataset: 'imagenet'
+data_dir: 'path/to/imagenet/'
+shuffle: True
+dataset_download: False
+batch_size: 32
+drop_remainder: True
+val_split: val
+
+# augmentation
+image_resize: 224
+scale: [0.08, 1.0]
+ratio: [0.75, 1.333]
+hflip: 0.5
+auto_augment: 'randaug-m9-mstd0.5-inc1'
+interpolation: bilinear
+re_prob: 0.25
+re_value: 'random'
+cutmix: 0.5
+mixup: 0.5
+mixup_prob: 1.0
+mixup_mode: batch
+mixup_off_epoch: 0.0
+switch_prob: 0.5
+crop_pct: 0.96
+
+# model
+model: 'convmixer_1024_20'
+num_classes: 1000
+pretrained: False
+ckpt_path: ''
+
+keep_checkpoint_max: 10
+ckpt_save_dir: './ckpt'
+
+epoch_size: 300
+dataset_sink_mode: True
+amp_level: 'O0'
+use_ema: False
+ema_decay: 0.9999
+use_clip_grad: True
+clip_value: 1.0
+
+# loss
+loss: 'CE'
+label_smoothing: 0.1
+
+# lr scheduler
+lr_scheduler: 'cosine_decay'
+lr: 0.001
+warmup_epochs: 0
+warmup_factor: 0.007
+min_lr: 0.0001
+decay_epochs: 300
+
+# optimizer
+opt: 'adamw'
+momentum: 0.9
+weight_decay: 0.0001
+loss_scale: 1024
+dynamic_loss_scale: False

configs/convmixer/convmixer_1536_20_gpu.yaml

@@ -0,0 +1,67 @@
+# system
+mode: 0
+distribute: False
+num_parallel_workers: 4
+val_while_train: True
+
+# dataset
+dataset: 'imagenet'
+data_dir: 'path/to/imagenet/'
+shuffle: True
+dataset_download: False
+batch_size: 32
+drop_remainder: True
+val_split: val
+
+# augmentation
+image_resize: 224
+scale: [0.08, 1.0]
+ratio: [0.75, 1.333]
+hflip: 0.5
+auto_augment: 'randaug-m9-mstd0.5-inc1'
+interpolation: bilinear
+re_prob: 0.25
+re_value: 'random'
+cutmix: 0.5
+mixup: 0.5
+mixup_prob: 1.0
+mixup_mode: batch
+mixup_off_epoch: 0.0
+switch_prob: 0.5
+crop_pct: 0.96
+
+# model
+model: 'convmixer_1536_20'
+num_classes: 1000
+pretrained: False
+ckpt_path: ''
+
+keep_checkpoint_max: 10
+ckpt_save_dir: './ckpt'
+
+epoch_size: 300
+dataset_sink_mode: True
+amp_level: 'O0'
+use_ema: False
+ema_decay: 0.9999
+use_clip_grad: True
+clip_value: 1.0
+
+# loss
+loss: 'CE'
+label_smoothing: 0.1
+
+# lr scheduler
+lr_scheduler: 'cosine_decay'
+lr: 0.001
+warmup_epochs: 0
+warmup_factor: 0.007
+min_lr: 0.0001
+decay_epochs: 300
+
+# optimizer
+opt: 'adamw'
+momentum: 0.9
+weight_decay: 0.0001
+loss_scale: 1024
+dynamic_loss_scale: False

configs/convmixer/convmixer_768_32_gpu.yaml

@@ -0,0 +1,66 @@
+# system
+mode: 0
+distribute: False
+num_parallel_workers: 4
+val_while_train: True
+
+# dataset
+dataset: 'imagenet'
+data_dir: 'path/to/imagenet/'
+shuffle: True
+dataset_download: False
+batch_size: 32
+drop_remainder: True
+val_split: val
+
+# augmentation
+image_resize: 224
+scale: [0.08, 1.0]
+ratio: [0.75, 1.333]
+hflip: 0.5
+auto_augment: 'randaug-m9-mstd0.5-inc1'
+interpolation: bilinear
+re_prob: 0.25
+re_value: 'random'
+cutmix: 0.5
+mixup: 0.5
+mixup_prob: 1.0
+mixup_mode: batch
+mixup_off_epoch: 0.0
+switch_prob: 0.5
+crop_pct: 0.96
+
+# model
+model: 'convmixer_768_32'
+num_classes: 1000
+pretrained: False
+ckpt_path: ''
+
+keep_checkpoint_max: 10
+ckpt_save_dir: './ckpt'
+
+epoch_size: 300
+dataset_sink_mode: True
+amp_level: 'O0'
+use_ema: False
+ema_decay: 0.9999
+use_clip_grad: True
+clip_value: 1.0
+
+# loss
+loss: 'CE'
+label_smoothing: 0.1
+
+# lr scheduler
+lr_scheduler: 'onecycle'
+lr: 0.001
+warmup_epochs: 0
+min_lr: 0.000001
+decay_epochs: 300
+
+# optimizer
+opt: 'adamw'
+momentum: 0.9
+weight_decay: 0.0001
+loss_scale: 1024
+dynamic_loss_scale: False

mindcv/models/__init__.py

@@ -2,6 +2,7 @@
 from . import (
     bit,
     convit,
+    convmixer,
     convnext,
     densenet,
     dpn,
@@ -44,6 +45,7 @@
 )
 from .bit import *
 from .convit import *
+from .convmixer import *
 from .convnext import *
 from .densenet import *
 from .dpn import *
@@ -90,6 +92,7 @@
 __all__ = []
 __all__.extend(bit.__all__)
 __all__.extend(convit.__all__)
+__all__.extend(convmixer.__all__)
 __all__.extend(convnext.__all__)
 __all__.extend(densenet.__all__)
 __all__.extend(dpn.__all__)