p04896132
/
mindspore
forked from mindspore-Ecosystem/mindspore
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

!5363 Modify for resnet readme and fix bool type optional 

Merge pull request !5363 from qujianwei/r0.7
This commit is contained in:
mindspore-ci-bot 2020-08-28 10:32:50 +08:00 committed by Gitee
parent 98528bbc16 3adc06024c
commit 23fc178a5a
2 changed files with 237 additions and 61 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

293
model_zoo/official/cv/resnet/README.md
View File

@ -1,40 +1,127 @@
# ResNet Example
# Contents
- [ResNet Description](#resnet-description)
- [Model Architecture](#model-architecture)
- [Dataset](#dataset)
- [Features](#features)
- [Mixed Precision](#mixed-precision)
- [Environment Requirements](#environment-requirements)
- [Quick Start](#quick-start)
- [Script Description](#script-description)
- [Script and Sample Code](#script-and-sample-code)
- [Script Parameters](#script-parameters)
- [Training Process](#training-process)
- [Training](#training)
- [Distributed Training](#distributed-training)
- [Evaluation Process](#evaluation-process)
- [Evaluation](#evaluation)
- [Model Description](#model-description)
- [Performance](#performance)
- [Evaluation Performance](#evaluation-performance)
- [Description of Random Situation](#description-of-random-situation)
- [ModelZoo Homepage](#modelzoo-homepage)
# [ResNet Description](#contents)
## Description
ResNet (residual neural network) was proposed by Kaiming He and other four Chinese of Microsoft Research Institute. Through the use of ResNet unit, it successfully trained 152 layers of neural network, and won the championship in ilsvrc2015. The error rate on top 5 was 3.57%, and the parameter quantity was lower than vggnet, so the effect was very outstanding. Traditional convolution network or full connection network will have more or less information loss. At the same time, it will lead to the disappearance or explosion of gradient, which leads to the failure of deep network training. ResNet solves this problem to a certain extent. By passing the input information to the output, the integrity of the information is protected. The whole network only needs to learn the part of the difference between input and output, which simplifies the learning objectives and difficulties.The structure of ResNet can accelerate the training of neural network very quickly, and the accuracy of the model is also greatly improved. At the same time, ResNet is very popular, even can be directly used in the concept net network.
These are examples of training ResNet-50/ResNet-101 with CIFAR-10/ImageNet2012 dataset in MindSpore.
(Training ResNet-101 with dataset CIFAR-10 is unsupported now.)
These are examples of training ResNet-50/ResNet-101/SE-ResNet50 with CIFAR-10/ImageNet2012 dataset in MindSpore.ResNet-50 and ResNet101 can reference parper 1 below, and SE-ResNet-50 is a variant of ResNet-50 which reference paper [link](https://arxiv.org/abs/1709.01507) and [link](https://arxiv.org/abs/1812.01187) below, Training SE-ResNet-50 for just 24 epochs using 8 Ascend 910, we can reach top-1 accuracy of 75.9%.(Training ResNet-101 with dataset CIFAR-10 and SE-ResNet50 with CIFAR-10 is is not supported yet.)
## Requirements
## Paper
1.(https://arxiv.org/pdf/1512.03385.pdf):Kaiming He, Xiangyu Zhang, Shaoqing Ren, Jian Sun. "Deep Residual Learning for Image Recognition"
- Install [MindSpore](https://www.mindspore.cn/install/en).
2.(https://arxiv.org/abs/1709.01507):Jie Hu, Li Shen, Samuel Albanie, Gang Sun, Enhua Wu. "Squeeze-and-Excitation Networks"
- Download the dataset CIFAR-10 or ImageNet2012
3.(https://arxiv.org/abs/1812.01187):Tong He, Zhi Zhang, Hang Zhang, Zhongyue Zhang, Junyuan Xie, Mu Li. "Bag of Tricks for Image Classification with Convolutional Neural Networks"
CIFAR-10
# [Model Architecture](#contents)
> Unzip the CIFAR-10 dataset to any path you want and the folder structure should include train and eval dataset as follows:
> ```
> .
> └─dataset
> ├─ cifar-10-batches-bin # train dataset
> └─ cifar-10-verify-bin # evaluate dataset
> ```
The overall network architecture of ResNet is show below:
[Link](https://arxiv.org/pdf/1512.03385.pdf)
ImageNet2012
# [Dataset](#contents)
> Unzip the ImageNet2012 dataset to any path you want and the folder should include train and eval dataset as follows:
>
> ```
> .
> └─dataset
> ├─ilsvrc # train dataset
> └─validation_preprocess # evaluate dataset
> ```
Dataset used: [CIFAR-10](<http://www.cs.toronto.edu/~kriz/cifar.html>)
- 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
- Note：Data will be processed in dataset.py
- Download the dataset, the directory structure is as follows:
```
├─cifar-10-batches-bin
│
└─cifar-10-verify-bin
```
Dataset used: [imagenet](http://www.image-net.org/)
- Dataset size: ~125G, 1.2W colorful images in 1000 classes
- Train: 120G, 1.2W images
- Test: 5G, 50000 images
- Data format: RGB images.
- Note: Data will be processed in src/dataset.py
Download the dataset CIFAR-10 or ImageNet2012
```
└─dataset
├─ilsvrc # train dataset
└─validation_preprocess # evaluate dataset
```
# [Features](#contents)
## Mixed Precision
The [mixed precision](https://www.mindspore.cn/tutorial/zh-CN/master/advanced_use/mixed_precision.html) training method accelerates the deep learning neural network training process by using both the single-precision and half-precision data types, 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’.
# [Environment Requirements](#contents)
- 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.
- Framework
- [MindSpore](https://www.mindspore.cn/install/en)
- 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)
## Structure
# [Quick Start](#contents)
After installing MindSpore via the official website, you can start training and evaluation as follows:
- runing on Ascend
```
# distributed training
Usage: sh run_distribute_train.sh [resnet50|resnet101|se-resnet50] [cifar10|imagenet2012] [RANK_TABLE_FILE] [DATASET_PATH] [PRETRAINED_CKPT_PATH](optional)
# standalone training
Usage: sh run_standalone_train.sh [resnet50|resnet101|se-resnet50] [cifar10|imagenet2012] [DATASET_PATH]
[PRETRAINED_CKPT_PATH](optional)
# run evaluation example
Usage: sh run_eval.sh [resnet50|resnet101|se-resnet50] [cifar10|imagenet2012] [DATASET_PATH] [CHECKPOINT_PATH]
```
- runing on GPU
```
# distributed training example
sh run_distribute_train_gpu.sh [resnet50|resnet101] [cifar10|imagenet2012] [DATASET_PATH] [PRETRAINED_CKPT_PATH](optional)
# standalone training example
sh run_standalone_train_gpu.sh [resnet50|resnet101] [cifar10|imagenet2012] [DATASET_PATH] [PRETRAINED_CKPT_PATH](optional)
# infer example
sh run_eval_gpu.sh [resnet50|resnet101] [cifar10|imagenet2012] [DATASET_PATH] [CHECKPOINT_PATH]
```
# [Script Description](#contents)
## [Script and Sample Code](#contents)
```shell
.
@ -59,8 +146,7 @@ ImageNet2012
└── train.py # train net
```
## Parameter configuration
## [Script Parameters](#contents)
Parameters for both training and evaluation can be set in config.py.
@ -152,13 +238,10 @@ Parameters for both training and evaluation can be set in config.py.
"lr_end": 0.0001, # end learning rate
```
## [Training Process](#contents)
## Running the example
### Train
#### Usage
### Training
- running on Ascend
```
# distributed training
Usage: sh run_distribute_train.sh [resnet50|resnet101|se-resnet50] [cifar10|imagenet2012] [RANK_TABLE_FILE] [DATASET_PATH] [PRETRAINED_CKPT_PATH](optional)
@ -166,22 +249,16 @@ Usage: sh run_distribute_train.sh [resnet50|resnet101|se-resnet50] [cifar10|imag
# standalone training
Usage: sh run_standalone_train.sh [resnet50|resnet101|se-resnet50] [cifar10|imagenet2012] [DATASET_PATH]
[PRETRAINED_CKPT_PATH](optional)
```
#### Launch
# run evaluation example
Usage: sh run_eval.sh [resnet50|resnet101|se-resnet50] [cifar10|imagenet2012] [DATASET_PATH] [CHECKPOINT_PATH]
```
# distribute training example
sh run_distribute_train.sh resnet50 cifar10 rank_table.json ~/cifar-10-batches-bin
For distributed training, a hccl configuration file with JSON format needs to be created in advance.
# standalone training example
sh run_standalone_train.sh resnet50 cifar10 ~/cifar-10-batches-bin
```
Please follow the instructions in the link below:
> About rank_table.json, you can refer to the [distributed training tutorial](https://www.mindspore.cn/tutorial/en/master/advanced_use/distributed_training.html).
#### Result
https://gitee.com/mindspore/mindspore/tree/master/model_zoo/utils/hccl_tools.
Training result will be stored in the example path, whose folder name begins with "train" or "train_parallel". Under this, you can find checkpoint file together with result like the followings in log.
@ -236,8 +313,25 @@ epoch: 4 step: 5004, loss is 3.5011306
epoch: 5 step: 5004, loss is 3.3501816
...
```
- running on GPU
```
# distributed training example
sh run_distribute_train_gpu.sh [resnet50|resnet101] [cifar10|imagenet2012] [DATASET_PATH] [PRETRAINED_CKPT_PATH](optional)
# standalone training example
sh run_standalone_train_gpu.sh [resnet50|resnet101] [cifar10|imagenet2012] [DATASET_PATH] [PRETRAINED_CKPT_PATH](optional)
# infer example
sh run_eval_gpu.sh [resnet50|resnet101] [cifar10|imagenet2012] [DATASET_PATH] [CHECKPOINT_PATH]
```
## [Evaluation Process](#contents)
### Evaluation
- evaluation on CIFAR-10 dataset when running on Ascend
#### Usage
```
@ -245,8 +339,6 @@ epoch: 5 step: 5004, loss is 3.3501816
Usage: sh run_eval.sh [resnet50|resnet101|se-resnet50] [cifar10|imagenet2012] [DATASET_PATH] [CHECKPOINT_PATH]
```
#### Launch
```
# evaluation example
sh run_eval.sh resnet50 cifar10 ~/cifar10-10-verify-bin ~/resnet50_cifar10/train_parallel0/resnet-90_195.ckpt
@ -280,27 +372,110 @@ result: {'top_5_accuracy': 0.9429417413572343, 'top_1_accuracy': 0.7853513124199
```
result: {'top_5_accuracy': 0.9342589628681178, 'top_1_accuracy': 0.768065781049936} ckpt=train_parallel0/resnet-24_5004.ckpt
```
```
### Running on GPU
```
# distributed training example
sh run_distribute_train_gpu.sh [resnet50|resnet101] [cifar10|imagenet2012] [DATASET_PATH] [PRETRAINED_CKPT_PATH](optional)
# standalone training example
sh run_standalone_train_gpu.sh [resnet50|resnet101] [cifar10|imagenet2012] [DATASET_PATH] [PRETRAINED_CKPT_PATH](optional)
# infer example
#### infer example
sh run_eval_gpu.sh [resnet50|resnet101] [cifar10|imagenet2012] [DATASET_PATH] [CHECKPOINT_PATH]
```
### Running parameter server mode training
```
# parameter server training Ascend example
### Parameter server training Ascend example
```
sh run_parameter_server_train.sh [resnet50|resnet101] [cifar10|imagenet2012] [RANK_TABLE_FILE] [DATASET_PATH] [PRETRAINED_CKPT_PATH](optional)
# parameter server training GPU example
### Parameter server training GPU example
sh run_parameter_server_train_gpu.sh [resnet50|resnet101] [cifar10|imagenet2012] [DATASET_PATH] [PRETRAINED_CKPT_PATH](optional)
> The way to evaluate is the same as the examples above.
```
> The way to evaluate is the same as the examples above.
# [Model Description](#contents)
## [Performance](#contents)
### Evaluation Performance
#### ResNet50 on cifar10
| Parameters | Ascend 910 | GPU |
| -------------------------- | -------------------------------------- |---------------------------------- |
| Model Version | ResNet50-v1.5 |ResNet50-v1.5|
| Resource | Ascend 910，CPU 2.60GHz 56cores，Memory 314G | GPU，CPU 2.1GHz 24cores，Memory 128G
| uploaded Date | 04/01/2020 (month/day/year) | 08/01/2020 (month/day/year)
| MindSpore Version | 0.1.0-alpha |0.6.0-alpha |
| Dataset | cifar10 | cifar10|
| Training Parameters | epoch=90, steps per epoch=195, batch_size = 32 |epoch=90, steps per epoch=195, batch_size = 32 |
| Optimizer | Momentum |Momentum|
| Loss Function | Softmax Cross Entropy |Softmax Cross Entropy |
| outputs | probability | probability |
| Loss | 0.000356 | 0.000716 |
| Speed | 18.4ms/step（8pcs） |69ms/step（8pcs）|
| Total time | 6 mins | 20.2 mins|
| Parameters (M) | 25.5 | 25.5 |
| Checkpoint for Fine tuning | 179.7M (.ckpt file) |179.7M (.ckpt file)|
| Scripts | https://gitee.com/mindspore/mindspore/tree/master/model_zoo/official/cv/resnet |https://gitee.com/mindspore/mindspore/tree/master/model_zoo/official/cv/resnet |
#### ResNet50 on imagenet2012
| Parameters | Ascend 910 | GPU |
| -------------------------- | -------------------------------------- |---------------------------------- |
| Model Version | ResNet50-v1.5 |ResNet50-v1.5|
| Resource | Ascend 910，CPU 2.60GHz 56cores，Memory 314G | GPU，CPU 2.1GHz 24cores，Memory 128G
| uploaded Date | 04/01/2020 (month/day/year) ； | 08/01/2020 (month/day/year)
| MindSpore Version | 0.1.0-alpha |0.6.0-alpha |
| Dataset | ImageNet2012 | ImageNet2012|
| Training Parameters | epoch=90, steps per epoch=5004, batch_size = 32 |epoch=90, steps per epoch=5004, batch_size = 32 |
| Optimizer | Momentum |Momentum|
| Loss Function | Softmax Cross Entropy |Softmax Cross Entropy |
| outputs | probability | probability |
| Loss | 1.8464266 | 1.9023 |
| Speed | 18.4ms/step（8pcs） |67.1ms/step（8pcs）|
| Total time | 139 mins | 258 mins|
| Parameters (M) | 25.5 | 25.5 |
| Checkpoint for Fine tuning | 197M (.ckpt file) |197M (.ckpt file) |
| Scripts | https://gitee.com/mindspore/mindspore/tree/master/model_zoo/official/cv/resnet |https://gitee.com/mindspore/mindspore/tree/master/model_zoo/official/cv/resnet |
#### ResNet101 on imagenet2012
| Parameters | Ascend 910 | GPU |
| -------------------------- | -------------------------------------- |---------------------------------- |
| Model Version | ResNet101 |ResNet101|
| Resource | Ascend 910，CPU 2.60GHz 56cores，Memory 314G | GPU，CPU 2.1GHz 24cores，Memory 128G
| uploaded Date | 04/01/2020 (month/day/year) | 08/14/2020 (month/day/year)
| MindSpore Version | 0.1.0-alpha |0.6.0-alpha |
| Dataset | ImageNet2012 | ImageNet2012|
| Training Parameters | epoch=120, steps per epoch=5004, batch_size = 32 |epoch=120, steps per epoch=5004, batch_size = 32 |
| Optimizer | Momentum |Momentum|
| Loss Function | Softmax Cross Entropy |Softmax Cross Entropy |
| outputs | probability | probability |
| Loss | 1.6453942 | 1.7023412 |
| Speed | 30.3ms/step（8pcs） |108.6ms/step（8pcs）|
| Total time | 301 mins | 1100 mins|
| Parameters (M) | 44.6 | 44.6 |
| Checkpoint for Fine tuning | 343M (.ckpt file) |343M (.ckpt file) |
| Scripts | https://gitee.com/mindspore/mindspore/tree/master/model_zoo/official/cv/resnet |https://gitee.com/mindspore/mindspore/tree/master/model_zoo/official/cv/resnet |
#### SE-ResNet50 on imagenet2012
| Parameters | Ascend 910
| -------------------------- | ------------------------------------------------------------------------ |
| Model Version | SE-ResNet50 |
| Resource | Ascend 910，CPU 2.60GHz 56cores，Memory 314G |
| uploaded Date | 08/16/2020 (month/day/year) ； |
| MindSpore Version | 0.6.0-alpha |
| Dataset | ImageNet2012 |
| Training Parameters | epoch=24, steps per epoch=5004, batch_size = 32 |
| Optimizer | Momentum |
| Loss Function | Softmax Cross Entropy |
| outputs | probability |
| Loss | 1.754404 |
| Speed | 24.6ms/step（8pcs） |
| Total time | 49.3 mins |
| Parameters (M) | 25.5 |
| Checkpoint for Fine tuning | 215.9M (.ckpt file) |
| Scripts | https://gitee.com/mindspore/mindspore/tree/master/model_zoo/official/cv/resnet |https://gitee.com/mindspore/mindspore/tree/master/model_zoo/official/cv/resnet |
# [Description of Random Situation](#contents)
In dataset.py, we set the seed inside “create_dataset" function. We also use random seed in train.py.
# [ModelZoo Homepage](#contents)
Please check the official [homepage](https://gitee.com/mindspore/mindspore/tree/master/model_zoo).

5
model_zoo/official/cv/resnet/train.py
View File

@ -16,6 +16,7 @@
import os
import random
import argparse
import ast
import numpy as np
from mindspore import context
from mindspore import Tensor
@ -35,13 +36,13 @@ from src.lr_generator import get_lr, warmup_cosine_annealing_lr
parser = argparse.ArgumentParser(description='Image classification')
parser.add_argument('--net', type=str, default=None, help='Resnet Model, either resnet50 or resnet101')
parser.add_argument('--dataset', type=str, default=None, help='Dataset, either cifar10 or imagenet2012')
parser.add_argument('--run_distribute', type=bool, default=False, help='Run distribute')
parser.add_argument('--run_distribute', type=ast.literal_eval, default=False, help='Run distribute')
parser.add_argument('--device_num', type=int, default=1, help='Device num.')
parser.add_argument('--dataset_path', type=str, default=None, help='Dataset path')
parser.add_argument('--device_target', type=str, default='Ascend', help='Device target')
parser.add_argument('--pre_trained', type=str, default=None, help='Pretrained checkpoint path')
parser.add_argument('--parameter_server', type=bool, default=False, help='Run parameter server train')
parser.add_argument('--parameter_server', type=ast.literal_eval, default=False, help='Run parameter server train')
args_opt = parser.parse_args()
random.seed(1)