Skip to content

M
mindspore

magicwindyyd / mindspore
与 Fork 源项目一致

Fork自 MindSpore / mindspore

通知 1
Star 1
Fork 0
M
mindspore
提交 762a5c83 编写于 作者: W wukesong
浏览文件
操作

modify lenet

上级 d3733b3b
隐藏空白更改
内联 并排
Showing with 364 addition and 119 deletion
model_zoo/official/cv/alexnet/README.md
浏览文件 @ 762a5c83
# AlexNet Example # Contents
## Description - [AlexNet Description](#alexnet-description)
- [Model Architecture](#model-architecture)
Training AlexNet with dataset in MindSpore. - [Dataset](#dataset)
- [Environment Requirements](#environment-requirements)
This is the simple tutorial for training AlexNet in MindSpore. - [Quick Start](#quick-start)
- [Script Description](#script-description)
## Requirements - [Script and Sample Code](#script-and-sample-code)
- [Script Parameters](#script-parameters)
- Install [MindSpore](https://www.mindspore.cn/install/en). - [Training Process](#training-process)
- [Training](#training)
- Download the dataset, the directory structure is as follows: - [Evaluation Process](#evaluation-process)
- [Evaluation](#evaluation)
``` - [Model Description](#model-description)
├─10-batches-bin - [Performance](#performance)
│ - [Evaluation Performance](#evaluation-performance)
└─10-verify-bin - [ModelZoo Homepage](#modelzoo-homepage)
```
## Running the example # [AlexNet Description](#contents)
```python AlexNet was proposed in 2012, one of the most influential neural networks. It got big success in ImageNet Dataset recognition than other models.
# train AlexNet, hyperparameter setting in config.py
python train.py --data_path 10-batches-bin [Paper](http://papers.nips.cc/paper/4824-imagenet-classification-with-deep-convolutional-neural-networks.pdf): Krizhevsky A, Sutskever I, Hinton G E. ImageNet Classification with Deep ConvolutionalNeural Networks. *Advances In Neural Information Processing Systems*. 2012.
```
# [Model Architecture](#contents)
You will get the loss value of each step as following:
AlexNet composition consists of 5 convolutional layers and 3 fully connected layers. Multiple convolutional kernels can extract interesting features in images and get more accurate classification.
```bash
epoch: 1 step: 1, loss is 2.2791853 # [Dataset](#contents)
...
epoch: 1 step: 1536, loss is 1.9366643 Dataset used: [CIFAR-10](<http://www.cs.toronto.edu/~kriz/cifar.html>)
epoch: 1 step: 1537, loss is 1.6983616
epoch: 1 step: 1538, loss is 1.0221305 - Dataset size：175M，60,000 32*32 colorful images in 10 classes
... - Train：146M，50,000 images
``` - Test：29.3M，10,000 images
- Data format：binary files
Then, evaluate AlexNet according to network model - Note：Data will be processed in dataset.py
```python - Download the dataset, the directory structure is as follows:
# evaluate AlexNet
python eval.py --data_path 10-verify-bin --ckpt_path checkpoint_alexnet-1_1562.ckpt ```
``` ├─cifar-10-batches-bin
│
## Note └─cifar-10-verify-bin
Here are some optional parameters: ```
```bash # [Environment Requirements](#contents)
--device_target {Ascend,GPU}
device where the code will be implemented (default: Ascend) - Hardware（Ascend/GPU）
--data_path DATA_PATH - Prepare hardware environment with Ascend or GPU processor.
path where the dataset is saved - Framework
--dataset_sink_mode DATASET_SINK_MODE - [MindSpore](http://10.90.67.50/mindspore/archive/20200506/OpenSource/me_vm_x86/)
dataset_sink_mode is False or True - For more information, please check the resources below：
``` - [MindSpore tutorials](https://www.mindspore.cn/tutorial/zh-CN/master/index.html)
- [MindSpore API](https://www.mindspore.cn/api/zh-CN/master/index.html)
You can run ```python train.py -h``` or ```python eval.py -h``` to get more information.
# [Quick Start](#contents)
After installing MindSpore via the official website, you can start training and evaluation as follows:
```python
# enter script dir, train AlexNet
sh run_standalone_train_ascend.sh [DATA_PATH] [CKPT_SAVE_PATH]
# enter script dir, evaluate AlexNet
sh run_standalone_eval_ascend.sh [DATA_PATH] [CKPT_NAME]
```
# [Script Description](#contents)
## [Script and Sample Code](#contents)
```
├── model_zoo
├── README.md // descriptions about all the models
├── alexnet
├── README.md // descriptions about alexnet
├── requirements.txt // package needed
├── scripts
│ ├──run_standalone_train_gpu.sh // train in gpu
│ ├──run_standalone_train_ascend.sh // train in ascend
│ ├──run_standalone_eval_gpu.sh // evaluate in gpu
│ ├──run_standalone_eval_ascend.sh // evaluate in ascend
├── src
│ ├──dataset.py // creating dataset
│ ├──alexnet.py // alexnet architecture
│ ├──config.py // parameter configuration
├── train.py // training script
├── eval.py // evaluation script
```
## [Script Parameters](#contents)
```python
Major parameters in train.py and config.py as follows:
--data_path: The absolute full path to the train and evaluation datasets.
--epoch_size: Total training epochs.
--batch_size: Training batch size.
--image_height: Image height used as input to the model.
--image_width: Image width used as input the model.
--device_target: Device where the code will be implemented. Optional values are "Ascend", "GPU".
--checkpoint_path: The absolute full path to the checkpoint file saved after training.
--data_path: Path where the dataset is saved
```
## [Training Process](#contents)
### Training
```
python train.py --data_path cifar-10-batches-bin --ckpt_path ckpt > log.txt 2>&1 &
# or enter script dir, and run the script
sh run_standalone_train_ascend.sh cifar-10-batches-bin ckpt
```
After training, the loss value will be achieved as follows:
# grep "loss is " train.log
```
epoch: 1 step: 1, loss is 2.2791853
...
epoch: 1 step: 1536, loss is 1.9366643
epoch: 1 step: 1537, loss is 1.6983616
epoch: 1 step: 1538, loss is 1.0221305
...
```
The model checkpoint will be saved in the current directory.
## [Evaluation Process](#contents)
### Evaluation
Before running the command below, please check the checkpoint path used for evaluation.
```
python eval.py --data_path cifar-10-verify-bin --ckpt_path ckpt/checkpoint_alexnet-1_1562.ckpt > log.txt 2>&1 &
or enter script dir, and run the script
sh run_standalone_eval_ascend.sh cifar-10-verify-bin ckpt/checkpoint_alexnet-1_1562.ckpt
```
You can view the results through the file "log.txt". The accuracy of the test dataset will be as follows:
```
# grep "Accuracy: " log.txt
'Accuracy': 0.8832
```
# [Model Description](#contents)
## [Performance](#contents)
### Evaluation Performance
| Parameters | AlexNet |
| -------------------------- | ----------------------------------------------------------- |
| Resource | Ascend 910 ；CPU 2.60GHz，56cores；Memory，314G |
| uploaded Date | 06/09/2020 (month/day/year) |
| MindSpore Version | 0.5.0-beta |
| Dataset | CIFAR-10 |
| Training Parameters | epoch=30, steps=1562, batch_size = 32, lr=0.002 |
| Optimizer | Momentum |
| Loss Function | Softmax Cross Entropy |
| outputs | probability |
| Loss | 0.0016 |
| Speed | 21 ms/step |
| Total time | 17 mins |
| Checkpoint for Fine tuning | 445M (.ckpt file) |
| Scripts | https://gitee.com/mindspore/mindspore/tree/master/model_zoo/official/cv/alexnet |
# [Description of Random Situation](#contents)
In dataset.py, we set the seed inside “create_dataset" function.
# [ModelZoo Homepage](#contents)
Please check the official [homepage](https://gitee.com/mindspore/mindspore/tree/master/model_zoo).
model_zoo/official/cv/lenet/README.md
浏览文件 @ 762a5c83
# LeNet Example # Contents
## Description - [LeNet Description](#lenet-description)
- [Model Architecture](#model-architecture)
Training LeNet with dataset in MindSpore. - [Dataset](#dataset)
- [Environment Requirements](#environment-requirements)
This is the simple and basic tutorial for constructing a network in MindSpore. - [Quick Start](#quick-start)
- [Script Description](#script-description)
## Requirements - [Script and Sample Code](#script-and-sample-code)
- [Script Parameters](#script-parameters)
- Install [MindSpore](https://www.mindspore.cn/install/en). - [Training Process](#training-process)
- [Training](#training)
- Download the dataset, the directory structure is as follows: - [Evaluation Process](#evaluation-process)
- [Evaluation](#evaluation)
``` - [Model Description](#model-description)
└─Data - [Performance](#performance)
├─test - [Evaluation Performance](#evaluation-performance)
│ t10k-images.idx3-ubyte - [ModelZoo Homepage](#modelzoo-homepage)
│ t10k-labels.idx1-ubyte
│
└─train # [LeNet Description](#contents)
train-images.idx3-ubyte
train-labels.idx1-ubyte LeNet was proposed in 1998, a typical convolutional neural network. It was used for digit recognition and got big success.
```
[Paper](https://ieeexplore.ieee.org/document/726791): Y.Lecun, L.Bottou, Y.Bengio, P.Haffner. Gradient-Based Learning Applied to Document Recognition. *Proceedings of the IEEE*. 1998.
## Running the example
# [Model Architecture](#contents)
```python
# train LeNet, hyperparameter setting in config.py LeNet is very simple, which contains 5 layers. The layer composition consists of 2 convolutional layers and 3 fully connected layers.
python train.py --data_path Data
``` # [Dataset](#contents)
You will get the loss value of each step as following: Dataset used: [MNIST](<http://yann.lecun.com/exdb/mnist/>)
```bash - Dataset size：52.4M，60,000 28*28 in 10 classes
epoch: 1 step: 1, loss is 2.3040335 - Train：60,000 images
... - Test：10,000 images
epoch: 1 step: 1739, loss is 0.06952668 - Data format：binary files
epoch: 1 step: 1740, loss is 0.05038793 - Note：Data will be processed in dataset.py
epoch: 1 step: 1741, loss is 0.05018193
... - The directory structure is as follows:
```
```
Then, evaluate LeNet according to network model └─Data
```python ├─test
# evaluate LeNet │ t10k-images.idx3-ubyte
python eval.py --data_path Data --ckpt_path checkpoint_lenet-1_1875.ckpt │ t10k-labels.idx1-ubyte
``` │
└─train
## Note train-images.idx3-ubyte
Here are some optional parameters: train-labels.idx1-ubyte
```
```bash
--device_target {Ascend,GPU,CPU} # [Environment Requirements](#contents)
device where the code will be implemented (default: Ascend)
--data_path DATA_PATH - Hardware（Ascend/GPU/CPU）
path where the dataset is saved - Prepare hardware environment with Ascend, GPU, or CPU processor.
--dataset_sink_mode DATASET_SINK_MODE - Framework
dataset_sink_mode is False or True - [MindSpore](http://10.90.67.50/mindspore/archive/20200506/OpenSource/me_vm_x86/)
``` - For more information, please check the resources below：
- [MindSpore tutorials](https://www.mindspore.cn/tutorial/zh-CN/master/index.html)
You can run ```python train.py -h``` or ```python eval.py -h``` to get more information. - [MindSpore API](https://www.mindspore.cn/api/zh-CN/master/index.html)
# [Quick Start](#contents)
After installing MindSpore via the official website, you can start training and evaluation as follows:
```python
# enter script dir, train LeNet
sh run_standalone_train_ascend.sh [DATA_PATH] [CKPT_SAVE_PATH]
# enter script dir, evaluate LeNet
sh run_standalone_eval_ascend.sh [DATA_PATH] [CKPT_NAME]
```
# [Script Description](#contents)
## [Script and Sample Code](#contents)
```
├── model_zoo
├── README.md // descriptions about all the models
├── lenet
├── README.md // descriptions about lenet
├── requirements.txt // package needed
├── scripts
│ ├──run_standalone_train_cpu.sh // train in cpu
│ ├──run_standalone_train_gpu.sh // train in gpu
│ ├──run_standalone_train_ascend.sh // train in ascend
│ ├──run_standalone_eval_cpu.sh // evaluate in cpu
│ ├──run_standalone_eval_gpu.sh // evaluate in gpu
│ ├──run_standalone_eval_ascend.sh // evaluate in ascend
├── src
│ ├──dataset.py // creating dataset
│ ├──lenet.py // lenet architecture
│ ├──config.py // parameter configuration
├── train.py // training script
├── eval.py // evaluation script
```
## [Script Parameters](#contents)
```python
Major parameters in train.py and config.py as follows:
--data_path: The absolute full path to the train and evaluation datasets.
--epoch_size: Total training epochs.
--batch_size: Training batch size.
--image_height: Image height used as input to the model.
--image_width: Image width used as input the model.
--device_target: Device where the code will be implemented. Optional values
are "Ascend", "GPU", "CPU".
--checkpoint_path: The absolute full path to the checkpoint file saved
after training.
--data_path: Path where the dataset is saved
```
## [Training Process](#contents)
### Training
```
python train.py --data_path Data --ckpt_path ckpt > log.txt 2>&1 &
or enter script dir, and run the script
sh run_standalone_train_ascend.sh Data ckpt
```
After training, the loss value will be achieved as follows:
```
# grep "loss is " log.txt
epoch: 1 step: 1, loss is 2.2791853
...
epoch: 1 step: 1536, loss is 1.9366643
epoch: 1 step: 1537, loss is 1.6983616
epoch: 1 step: 1538, loss is 1.0221305
...
```
The model checkpoint will be saved in the current directory.
## [Evaluation Process](#contents)
### Evaluation
Before running the command below, please check the checkpoint path used for evaluation.
```
python eval.py --data_path Data --ckpt_path ckpt/checkpoint_lenet-1_1875.ckpt > log.txt 2>&1 &
or enter script dir, and run the script
sh run_standalone_eval_ascend.sh Data ckpt/checkpoint_lenet-1_1875.ckpt
```
You can view the results through the file "log.txt". The accuracy of the test dataset will be as follows:
```
# grep "Accuracy: " log.txt
'Accuracy': 0.9842
```
# [Model Description](#contents)
## [Performance](#contents)
### Evaluation Performance
| Parameters | LeNet |
| -------------------------- | ----------------------------------------------------------- |
| Resource | Ascend 910 ；CPU 2.60GHz，56cores；Memory，314G |
| uploaded Date | 06/09/2020 (month/day/year) |
| MindSpore Version | 0.5.0-beta |
| Dataset | MNIST |
| Training Parameters | epoch=10, steps=1875, batch_size = 32, lr=0.01 |
| Optimizer | Momentum |
| Loss Function | Softmax Cross Entropy |
| outputs | probability |
| Loss | 0.002 |
| Speed | 1.70 ms/step |
| Total time | 43.1s | |
| Checkpoint for Fine tuning | 482k (.ckpt file) |
| Scripts | https://gitee.com/mindspore/mindspore/tree/master/model_zoo/official/cv/lenet |
# [Description of Random Situation](#contents)
In dataset.py, we set the seed inside “create_dataset" function.
# [ModelZoo Homepage](#contents)
Please check the official [homepage](https://gitee.com/mindspore/mindspore/tree/master/model_zoo).
Markdown is supported
0% 或 .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑！
想要评论请 注册 或