- [Script and Sample Code](#script-and-sample-code)
- [Script and Sample Code](#script-and-sample-code)
- [Script Parameters](#script-parameters)
- [Script Parameters](#script-parameters)
@@ -22,7 +22,6 @@
- [Description of Random Situation](#description-of-random-situation)
- [Description of Random Situation](#description-of-random-situation)
- [ModelZoo Homepage](#modelzoo-homepage)
- [ModelZoo Homepage](#modelzoo-homepage)
# [DenseNet121 Description](#contents)
# [DenseNet121 Description](#contents)
DenseNet121 is a convolution based neural network for the task of image classification. The paper describing the model can be found [here](https://arxiv.org/abs/1608.06993). HuaWei’s DenseNet121 is a implementation on [MindSpore](https://www.mindspore.cn/).
DenseNet121 is a convolution based neural network for the task of image classification. The paper describing the model can be found [here](https://arxiv.org/abs/1608.06993). HuaWei’s DenseNet121 is a implementation on [MindSpore](https://www.mindspore.cn/).
@@ -33,60 +32,55 @@ The repository also contains scripts to launch training and inference routines.
DenseNet121 builds on 4 densely connected block. In every dense block, each layer obtains additional inputs from all preceding layers and passes on its own feature-maps to all subsequent layers. Concatenation is used. Each layer is receiving a “collective knowledge” from all preceding layers.
DenseNet121 builds on 4 densely connected block. In every dense block, each layer obtains additional inputs from all preceding layers and passes on its own feature-maps to all subsequent layers. Concatenation is used. Each layer is receiving a “collective knowledge” from all preceding layers.
# [Dataset](#contents)
# [Dataset](#contents)
Dataset used: ImageNet
Dataset used: ImageNet
The default configuration of the Dataset are as follows:
The default configuration of the Dataset are as follows:
- Training Dataset preprocess:
- Input size of images is 224\*224
- Range (min, max) of respective size of the original size to be cropped is (0.08, 1.0)
- Range (min, max) of aspect ratio to be cropped is (0.75, 1.333)
- Probability of the image being flipped set to 0.5
- Randomly adjust the brightness, contrast, saturation (0.4, 0.4, 0.4)
- Normalize the input image with respect to mean and standard deviation
- Test Dataset preprocess:
- Input size of images is 224\*224 (Resize to 256\*256 then crops images at the center)
- Normalize the input image with respect to mean and standard deviation
- Training Dataset preprocess:
- Input size of images is 224\*224
- Range (min, max) of respective size of the original size to be cropped is (0.08, 1.0)
- Range (min, max) of aspect ratio to be cropped is (0.75, 1.333)
- Probability of the image being flipped set to 0.5
- Randomly adjust the brightness, contrast, saturation (0.4, 0.4, 0.4)
- Normalize the input image with respect to mean and standard deviation
- Test Dataset preprocess:
- Input size of images is 224\*224 (Resize to 256\*256 then crops images at the center)
- Normalize the input image with respect to mean and standard deviation
# [Features](#contents)
# [Features](#contents)
## Mixed Precision
## Mixed Precision
The [mixed precision](https://www.mindspore.cn/tutorial/training/en/master/advanced_use/enable_mixed_precision.html) training method accelerates the deep learning neural network training process by using both the single-precision and half-precision data formats, and maintains the network precision achieved by the single-precision training at the same time. Mixed precision training can accelerate the computation process, reduce memory usage, and enable a larger model or batch size to be trained on specific hardware.
The [mixed precision](https://www.mindspore.cn/tutorial/training/en/master/advanced_use/enable_mixed_precision.html) training method accelerates the deep learning neural network training process by using both the single-precision and half-precision data formats, and maintains the network precision achieved by the single-precision training at the same time. Mixed precision training can accelerate the computation process, reduce memory usage, and enable a larger model or batch size to be trained on specific hardware.
For FP16 operators, if the input data type is FP32, the backend of MindSpore will automatically handle it with reduced precision. Users could check the reduced-precision operators by enabling INFO log and then searching ‘reduce precision’.
For FP16 operators, if the input data type is FP32, the backend of MindSpore will automatically handle it with reduced precision. Users could check the reduced-precision operators by enabling INFO log and then searching ‘reduce precision’.
# [Environment Requirements](#contents)
# [Environment Requirements](#contents)
- Hardware（Ascend）
- Prepare hardware environment with Ascend AI processor. If you want to try Ascend , please send the [application form](https://obs-9be7.obs.cn-east-2.myhuaweicloud.com/file/other/Ascend%20Model%20Zoo%E4%BD%93%E9%AA%8C%E8%B5%84%E6%BA%90%E7%94%B3%E8%AF%B7%E8%A1%A8.docx) to ascend@huawei.com. Once approved, you can get the resources.
- Hardware（Ascend/GPU）
- Prepare hardware environment with Ascend or GPU processor. If you want to try Ascend , please send the [application form](https://obs-9be7.obs.cn-east-2.myhuaweicloud.com/file/other/Ascend%20Model%20Zoo%E4%BD%93%E9%AA%8C%E8%B5%84%E6%BA%90%E7%94%B3%E8%AF%B7%E8%A1%A8.docx) to ascend@huawei.com. Once approved, you can get the resources.
The python command above will run in the background, The log and model checkpoint will be generated in `output/202x-xx-xx_time_xx_xx_xx/`. The loss value will be achieved as follows:
The python command above will run in the background, The log and model checkpoint will be generated in `output/202x-xx-xx_time_xx_xx_xx/`. The loss value will be achieved as follows:
The python command above will run in the background, you can view the results through the file `train.log`.
After training, you'll get some checkpoint files under the folder `./ckpt_0/` by default.
### Distributed Training
### Distributed Training
- running on Ascend
- running on Ascend
```
```bash
sh scripts/run_distribute_train.sh 8 rank_table.json /PATH/TO/DATASET /PATH/TO/PRETRAINED_CKPT
sh scripts/run_distribute_train.sh 8 rank_table.json /PATH/TO/DATASET /PATH/TO/PRETRAINED_CKPT
```
```
The above shell script will run distribute training in the background. You can view the results log and model checkpoint through the file `train[X]/output/202x-xx-xx_time_xx_xx_xx/`. The loss value will be achieved as follows:
The above shell script will run distribute training in the background. You can view the results log and model checkpoint through the file `train[X]/output/202x-xx-xx_time_xx_xx_xx/`. The loss value will be achieved as follows:
sh scripts/run_distribute_eval.sh 8 rank_table.json /PATH/TO/DATASET /PATH/TO/CHECKPOINT
sh scripts/run_distribute_eval.sh 8 rank_table.json /PATH/TO/DATASET /PATH/TO/CHECKPOINT
```
```
The above python command will run in the background. You can view the results through the file "output/202x-xx-xx_time_xx_xx_xx/202x_xxxx.log". The accuracy of the test dataset will be as follows:
The above python command will run in the background. You can view the results through the file "output/202x-xx-xx_time_xx_xx_xx/202x_xxxx.log". The accuracy of the test dataset will be as follows:
sh run_distribute_eval_gpu.sh 1 0 [DATASET_PATH] [CHECKPOINT_PATH]
```
The above python command will run in the background. You can view the results through the file "eval/eval.log". The accuracy of the test dataset will be as follows:
mindspore-ci-bot
Gitee
4 years ago