p60975123
/
mindspore
forked from mindspore-Ecosystem/mindspore
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

!15947 modify deeplabv3 network for clould 

From: @zhanghuiyao
Reviewed-by: @c_34,@oacjiewen
Signed-off-by: @c_34
This commit is contained in:
mindspore-ci-bot 2021-05-06 19:46:18 +08:00 committed by Gitee
parent 868e29ba39 ff2c2fc2fb
commit 3e4b2e049f
20 changed files with 1120 additions and 135 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

236
model_zoo/official/cv/deeplabv3/README.md
View File

@ -104,7 +104,7 @@ For single device training, please config parameters, training script is:
run_standalone_train.sh
```
For 8 devices training, training steps are as follows:
- For 8 devices training, training steps are as follows:
1. Train s16 with vocaug dataset, finetuning from resnet101 pretrained model, script is:
@ -124,7 +124,7 @@ For 8 devices training, training steps are as follows:
run_distribute_train_s8_r2.sh
```
For evaluation, evaluating steps are as follows:
- For evaluation, evaluating steps are as follows:
1. Eval s16 with voc val dataset, eval script is:
@ -150,6 +150,238 @@ For evaluation, evaluating steps are as follows:
run_eval_s8_multiscale_flip.sh
```
- Train on ModelArts (If you want to run in modelarts, please check the official documentation of [modelarts](https://support.huaweicloud.com/modelarts/), and you can start training as follows)
1. Train s16 with vocaug dataset on modelarts, finetuning from resnet101 pretrained model, training steps are as follows:
```python
# (1) Perform a or b.
# a. Set "enable_modelarts=True" on base_config.yaml file.
# Set "data_file='/cache/data/vocaug/vocaug_mindrecord/vocaug_mindrecord0'" on base_config.yaml file.
# Set "checkpoint_url=/The path of checkpoint in S3/" on beta_config.yaml file.
# Set "ckpt_pre_trained=/cache/checkpoint_path/path_to_pretrain/resnet101.ckpt" on base_config.yaml file.
# Set "base_lr=0.08" on base_config.yaml file.
# Set "is_distributed=True" on base_config.yaml file.
# Set "save_steps=410" on base_config.yaml file.
# Set other parameters on base_config.yaml file you need.
# b. Add "enable_modelarts=True" on the website UI interface.
# Add "data_file=/cache/data/vocaug/vocaug_mindrecord/vocaug_mindrecord0" on the website UI interface.
# Add "checkpoint_url=/The path of checkpoint in S3/" on the website UI interface.
# Add "ckpt_pre_trained=/cache/checkpoint_path/path_to_pretrain/resnet101.ckpt" on the website UI interface.
# Add "base_lr=0.08" on the website UI interface.
# Add "is_distributed=True" on the website UI interface.
# Add "save_steps=410" on the website UI interface.
# Add other parameters on the website UI interface.
# (2) Upload or copy your pretrained model to S3 bucket.
# (3) Upload a zip dataset to S3 bucket. (you could also upload the origin dataset, but it can be so slow.)
# (4) Set the code directory to "/path/deeplabv3" on the website UI interface.
# (5) Set the startup file to "train.py" on the website UI interface.
# (6) Set the "Dataset path" and "Output file path" and "Job log path" to your path on the website UI interface.
# (7) Create your job.
```
2. Train s8 with vocaug dataset on modelarts, finetuning from model in previous step, training steps are as follows:
```python
# (1) Perform a or b.
# a. Set "enable_modelarts=True" on base_config.yaml file.
# Set "model='deeplab_v3_s8'" on base_config.yaml file.
# Set "train_epochs=800" on base_config.yaml file.
# Set "batch_size=16" on base_config.yaml file.
# Set "base_lr=0.02" on base_config.yaml file.
# Set "loss_scale=2048" on base_config.yaml file.
# Set "data_file='/cache/data/vocaug/vocaug_mindrecord/vocaug_mindrecord0'" on base_config.yaml file.
# Set "checkpoint_url=/The path of checkpoint in S3/" on beta_config.yaml file.
# Set "ckpt_pre_trained=/cache/checkpoint_path/path_to_pretrain/deeplab_v3_s16-300_41.ckpt" on base_config.yaml file.
# Set "is_distributed=True" on base_config.yaml file.
# Set "save_steps=820" on base_config.yaml file.
# Set other parameters on base_config.yaml file you need.
# b. Add "enable_modelarts=True" on the website UI interface.
# Add "model='deeplab_v3_s8'" on the website UI interface.
# Add "train_epochs=800" on the website UI interface.
# Add "batch_size=16" on the website UI interface.
# Add "base_lr=0.02" on the website UI interface.
# Add "loss_scale=2048" on the website UI interface.
# Add "data_file='/cache/data/vocaug/vocaug_mindrecord/vocaug_mindrecord0'" on the website UI interface.
# Add "checkpoint_url=/The path of checkpoint in S3/" on the website UI interface.
# Add "ckpt_pre_trained=/cache/checkpoint_path/path_to_pretrain/deeplab_v3_s16-300_41.ckpt" on the website UI interface.
# Add "is_distributed=True" on the website UI interface.
# Add "save_steps=820" on the website UI interface.
# Add other parameters on the website UI interface.
# (2) Upload or copy your pretrained model to S3 bucket.
# (3) Upload a zip dataset to S3 bucket. (you could also upload the origin dataset, but it can be so slow.)
# (4) Set the code directory to "/path/deeplabv3" on the website UI interface.
# (5) Set the startup file to "train.py" on the website UI interface.
# (6) Set the "Dataset path" and "Output file path" and "Job log path" to your path on the website UI interface.
# (7) Create your job.
```
3. Train s8 with voctrain dataset on modelarts, finetuning from model in previous step, training steps are as follows:
```python
# (1) Perform a or b.
# a. Set "enable_modelarts=True" on base_config.yaml file.
# Set "model='deeplab_v3_s8'" on base_config.yaml file.
# Set "batch_size=16" on base_config.yaml file.
# Set "base_lr=0.008" on base_config.yaml file.
# Set "loss_scale=2048" on base_config.yaml file.
# Set "data_file='/cache/data/vocaug/voctrain_mindrecord/voctrain_mindrecord00'" on base_config.yaml file.
# Set "checkpoint_url=/The path of checkpoint in S3/" on beta_config.yaml file.
# Set "ckpt_pre_trained=/cache/checkpoint_path/path_to_pretrain/deeplab_v3_s8-800_82.ckpt" on base_config.yaml file.
# Set "is_distributed=True" on base_config.yaml file.
# Set "save_steps=110" on base_config.yaml file.
# Set other parameters on base_config.yaml file you need.
# b. Add "enable_modelarts=True" on the website UI interface.
# Add "model='deeplab_v3_s8'" on the website UI interface.
# Add "batch_size=16" on the website UI interface.
# Add "base_lr=0.008" on the website UI interface.
# Add "loss_scale=2048" on the website UI interface.
# Add "data_file='/cache/data/vocaug/voctrain_mindrecord/voctrain_mindrecord00'" on the website UI interface.
# Add "checkpoint_url=/The path of checkpoint in S3/" on the website UI interface.
# Add "ckpt_pre_trained=/cache/checkpoint_path/path_to_pretrain/deeplab_v3_s8-800_82.ckpt" on the website UI interface.
# Add "is_distributed=True" on the website UI interface.
# Add "save_steps=110" on the website UI interface.
# Add other parameters on the website UI interface.
# (2) Upload or copy your pretrained model to S3 bucket.
# (3) Upload a zip dataset to S3 bucket. (you could also upload the origin dataset, but it can be so slow.)
# (4) Set the code directory to "/path/deeplabv3" on the website UI interface.
# (5) Set the startup file to "train.py" on the website UI interface.
# (6) Set the "Dataset path" and "Output file path" and "Job log path" to your path on the website UI interface.
# (7) Create your job.
```
- Eval on ModelArts (If you want to run in modelarts, please check the official documentation of [modelarts](https://support.huaweicloud.com/modelarts/), and you can start evaluating as follows)
1. Eval s16 with voc val dataset on modelarts, evaluating steps are as follows:
```python
# (1) Perform a or b.
# a. Set "enable_modelarts=True" on base_config.yaml file.
# Set "model='deeplab_v3_s16'" on base_config.yaml file.
# Set "batch_size=32" on base_config.yaml file.
# Set "scales_type=0" on base_config.yaml file.
# Set "freeze_bn=True" on base_config.yaml file.
# Set "data_root='/cache/data/vocaug'" on base_config.yaml file.
# Set "data_lst='/cache/data/vocaug/voc_val_lst.txt'" on base_config.yaml file.
# Set "checkpoint_url=/The path of checkpoint in S3/" on beta_config.yaml file.
# Set "ckpt_path='/cache/checkpoint_path/path_to_pretrain/deeplab_v3_s16-300_41.ckpt'" on base_config.yaml file.
# Set other parameters on base_config.yaml file you need.
# b. Add "enable_modelarts=True" on the website UI interface.
# Add "model=deeplab_v3_s16" on the website UI interface.
# Add "batch_size=32" on the website UI interface.
# Add "scales_type=0" on the website UI interface.
# Add "freeze_bn=True" on the website UI interface.
# Add "data_root=/cache/data/vocaug" on the website UI interface.
# Add "data_lst=/cache/data/vocaug/voc_val_lst.txt" on the website UI interface.
# Add "checkpoint_url=/The path of checkpoint in S3/" on the website UI interface.
# Add "ckpt_path=/cache/checkpoint_path/path_to_pretrain/deeplab_v3_s16-300_41.ckpt" on the website UI interface.
# Add other parameters on the website UI interface.
# (2) Upload or copy your pretrained model to S3 bucket.
# (3) Upload a zip dataset to S3 bucket. (you could also upload the origin dataset, but it can be so slow.)
# (4) Set the code directory to "/path/deeplabv3" on the website UI interface.
# (5) Set the startup file to "eval.py" on the website UI interface.
# (6) Set the "Dataset path" and "Output file path" and "Job log path" to your path on the website UI interface.
# (7) Create your job.
```
2. Eval s8 with voc val dataset on modelarts, evaluating steps are as follows:
```python
# (1) Perform a or b.
# a. Set "enable_modelarts=True" on base_config.yaml file.
# Set "model='deeplab_v3_s8'" on base_config.yaml file.
# Set "batch_size=16" on base_config.yaml file.
# Set "scales_type=0" on base_config.yaml file.
# Set "freeze_bn=True" on base_config.yaml file.
# Set "data_root='/cache/data/vocaug'" on base_config.yaml file.
# Set "data_lst='/cache/data/vocaug/voc_val_lst.txt'" on base_config.yaml file.
# Set "checkpoint_url='/The path of checkpoint in S3/'" on beta_config.yaml file.
# Set "ckpt_path='/cache/checkpoint_path/path_to_pretrain/deeplab_v3_s8-300_11.ckpt'" on base_config.yaml file.
# Set other parameters on base_config.yaml file you need.
# b. Add "enable_modelarts=True" on the website UI interface.
# Add "model=deeplab_v3_s8" on the website UI interface.
# Add "batch_size=16" on the website UI interface.
# Add "scales_type=0" on the website UI interface.
# Add "freeze_bn=True" on the website UI interface.
# Add "data_root=/cache/data/vocaug" on the website UI interface.
# Add "data_lst=/cache/data/vocaug/voc_val_lst.txt" on the website UI interface.
# Add "checkpoint_url=/The path of checkpoint in S3/" on the website UI interface.
# Add "ckpt_path=/cache/checkpoint_path/path_to_pretrain/deeplab_v3_s8-300_11.ckpt" on the website UI interface.
# Add other parameters on the website UI interface.
# (2) Upload or copy your pretrained model to S3 bucket.
# (3) Upload a zip dataset to S3 bucket. (you could also upload the origin dataset, but it can be so slow.)
# (4) Set the code directory to "/path/deeplabv3" on the website UI interface.
# (5) Set the startup file to "eval.py" on the website UI interface.
# (6) Set the "Dataset path" and "Output file path" and "Job log path" to your path on the website UI interface.
# (7) Create your job.
```
3. Eval s8 multiscale with voc val dataset on modelarts, evaluating steps are as follows:
```python
# (1) Perform a or b.
# a. Set "enable_modelarts=True" on base_config.yaml file.
# Set "model='deeplab_v3_s8'" on base_config.yaml file.
# Set "batch_size=16" on base_config.yaml file.
# Set "scales_type=1" on base_config.yaml file.
# Set "freeze_bn=True" on base_config.yaml file.
# Set "data_root='/cache/data/vocaug'" on base_config.yaml file.
# Set "data_lst='/cache/data/vocaug/voc_val_lst.txt'" on base_config.yaml file.
# Set "checkpoint_url='/The path of checkpoint in S3/'" on beta_config.yaml file.
# Set "ckpt_path='/cache/checkpoint_path/path_to_pretrain/deeplab_v3_s8-300_11.ckpt'" on base_config.yaml file.
# Set other parameters on base_config.yaml file you need.
# b. Add "enable_modelarts=True" on the website UI interface.
# Add "model=deeplab_v3_s8" on the website UI interface.
# Add "batch_size=16" on the website UI interface.
# Add "scales_type=1" on the website UI interface.
# Add "freeze_bn=True" on the website UI interface.
# Add "data_root=/cache/data/vocaug" on the website UI interface.
# Add "data_lst=/cache/data/vocaug/voc_val_lst.txt" on the website UI interface.
# Add "checkpoint_url=/The path of checkpoint in S3/" on the website UI interface.
# Add "ckpt_path=/cache/checkpoint_path/path_to_pretrain/deeplab_v3_s8-300_11.ckpt" on the website UI interface.
# Add other parameters on the website UI interface.
# (2) Upload or copy your pretrained model to S3 bucket.
# (3) Upload a zip dataset to S3 bucket. (you could also upload the origin dataset, but it can be so slow.)
# (4) Set the code directory to "/path/deeplabv3" on the website UI interface.
# (5) Set the startup file to "eval.py" on the website UI interface.
# (6) Set the "Dataset path" and "Output file path" and "Job log path" to your path on the website UI interface.
# (7) Create your job.
```
4. Eval s8 multiscale and flip with voc val dataset on modelarts, evaluating steps are as follows:
```python
# (1) Perform a or b.
# a. Set "enable_modelarts=True" on base_config.yaml file.
# Set "model='deeplab_v3_s8'" on base_config.yaml file.
# Set "batch_size=16" on base_config.yaml file.
# Set "scales_type=1" on base_config.yaml file.
# Set "freeze_bn=True" on base_config.yaml file.
# Set "flip=True" on base_config.yaml file.
# Set "data_root='/cache/data/vocaug'" on base_config.yaml file.
# Set "data_lst='/cache/data/vocaug/voc_val_lst.txt'" on base_config.yaml file.
# Set "checkpoint_url='/The path of checkpoint in S3/'" on beta_config.yaml file.
# Set "ckpt_path='/cache/checkpoint_path/path_to_pretrain/deeplab_v3_s8-300_11.ckpt'" on base_config.yaml file.
# Set other parameters on base_config.yaml file you need.
# b. Add "enable_modelarts=True" on the website UI interface.
# Add "model=deeplab_v3_s8" on the website UI interface.
# Add "batch_size=16" on the website UI interface.
# Add "scales_type=1" on the website UI interface.
# Add "freeze_bn=True" on the website UI interface.
# Add "flip=True" on the website UI interface.
# Add "data_root=/cache/data/vocaug" on the website UI interface.
# Add "data_lst=/cache/data/vocaug/voc_val_lst.txt" on the website UI interface.
# Add "checkpoint_url=/The path of checkpoint in S3/" on the website UI interface.
# Add "ckpt_path=/cache/checkpoint_path/path_to_pretrain/deeplab_v3_s8-300_11.ckpt" on the website UI interface.
# Add other parameters on the website UI interface.
# (2) Upload or copy your pretrained model to S3 bucket.
# (3) Upload a zip dataset to S3 bucket. (you could also upload the origin dataset, but it can be so slow.)
# (4) Set the code directory to "/path/deeplabv3" on the website UI interface.
# (5) Set the startup file to "eval.py" on the website UI interface.
# (6) Set the "Dataset path" and "Output file path" and "Job log path" to your path on the website UI interface.
# (7) Create your job.
```
# [Script Description](#contents)
## [Script and Sample Code](#contents)

236
model_zoo/official/cv/deeplabv3/README_CN.md
View File

@ -119,7 +119,7 @@ Pascal VOC数据集和语义边界数据集（Semantic Boundaries Dataset，SBD
run_standalone_train.sh
```
按照以下训练步骤进行8卡训练：
- 按照以下训练步骤进行8卡训练：
1. 使用VOCaug数据集训练s16，微调ResNet-101预训练模型。脚本如下：
@ -139,7 +139,7 @@ run_standalone_train.sh
run_distribute_train_s8_r2.sh
```
评估步骤如下：
- 评估步骤如下：
1. 使用voc val数据集评估s16。评估脚本如下：
@ -165,6 +165,238 @@ run_standalone_train.sh
run_eval_s8_multiscale_flip.sh
```
- 在 ModelArts 进行训练 (如果你想在modelarts上运行，可以参考以下文档 [modelarts](https://support.huaweicloud.com/modelarts/))
1. 在 modelarts 使用VOCaug数据集训练s16，微调ResNet-101预训练模型。训练步骤如下：
```python
# (1) 执行 a 或者 b.
# a. 在 base_config.yaml 文件中设置 "enable_modelarts=True"
# 在 base_config.yaml 文件中设置 "data_file='/cache/data/vocaug/vocaug_mindrecord/vocaug_mindrecord0'"
# 在 base_config.yaml 文件中设置 "checkpoint_url=/The path of checkpoint in S3/"
# 在 base_config.yaml 文件中设置 "ckpt_pre_trained=/cache/checkpoint_path/path_to_pretrain/resnet101.ckpt"
# 在 base_config.yaml 文件中设置 "base_lr=0.08"
# 在 base_config.yaml 文件中设置 "is_distributed=True"
# 在 base_config.yaml 文件中设置 "save_steps=410"
# 在 base_config.yaml 文件中设置 其他参数
# b. 在网页上设置 "enable_modelarts=True"
# 在网页上设置 "data_file=/cache/data/vocaug/vocaug_mindrecord/vocaug_mindrecord0"
# 在网页上设置 "checkpoint_url=/The path of checkpoint in S3/"
# 在网页上设置 "ckpt_pre_trained=/cache/checkpoint_path/path_to_pretrain/resnet101.ckpt"
# 在网页上设置 "base_lr=0.08"
# 在网页上设置 "is_distributed=True"
# 在网页上设置 "save_steps=410"
# 在网页上设置 其他参数
# (2) 上传你的预训练模型到 S3 桶上
# (3) 上传你的压缩数据集到 S3 桶上 (你也可以上传原始的数据集，但那可能会很慢。)
# (4) 在网页上设置你的代码路径为 "/path/deeplabv3"
# (5) 在网页上设置启动文件为 "train.py"
# (6) 在网页上设置"训练数据集"、"训练输出文件路径"、"作业日志路径"等
# (7) 创建训练作业
```
2. 使用VOCaug数据集训练s8，微调上一步的模型。训练步骤如下：
```python
# (1) 执行 a 或者 b.
# a. 在 base_config.yaml 文件中设置 "enable_modelarts=True"
# 在 base_config.yaml 文件中设置 "model='deeplab_v3_s8'"
# 在 base_config.yaml 文件中设置 "train_epochs=800"
# 在 base_config.yaml 文件中设置 "batch_size=16"
# 在 base_config.yaml 文件中设置 "base_lr=0.02"
# 在 base_config.yaml 文件中设置 "loss_scale=2048"
# 在 base_config.yaml 文件中设置 "data_file='/cache/data/vocaug/vocaug_mindrecord/vocaug_mindrecord0'"
# 在 base_config.yaml 文件中设置 "checkpoint_url=/The path of checkpoint in S3/"
# 在 base_config.yaml 文件中设置 "ckpt_pre_trained=/cache/checkpoint_path/path_to_pretrain/deeplab_v3_s16-300_41.ckpt"
# 在 base_config.yaml 文件中设置 "is_distributed=True"
# 在 base_config.yaml 文件中设置 "save_steps=820"
# 在 base_config.yaml 文件中设置 其他参数
# b. 在网页上设置 "enable_modelarts=True"
# 在网页上设置 "model='deeplab_v3_s8'"
# 在网页上设置 "train_epochs=800"
# 在网页上设置 "batch_size=16"
# 在网页上设置 "base_lr=0.02"
# 在网页上设置 "loss_scale=2048"
# 在网页上设置 "data_file='/cache/data/vocaug/vocaug_mindrecord/vocaug_mindrecord0'"
# 在网页上设置 "checkpoint_url=/The path of checkpoint in S3/"
# 在网页上设置 "ckpt_pre_trained=/cache/checkpoint_path/path_to_pretrain/deeplab_v3_s16-300_41.ckpt"
# 在网页上设置 "is_distributed=True"
# 在网页上设置 "save_steps=820"
# 在网页上设置 其他参数
# (2) 上传你的预训练模型到 S3 桶上
# (3) 上传你的压缩数据集到 S3 桶上 (你也可以上传原始的数据集，但那可能会很慢。)
# (4) 在网页上设置你的代码路径为 "/path/deeplabv3"
# (5) 在网页上设置启动文件为 "train.py"
# (6) 在网页上设置"训练数据集"、"训练输出文件路径"、"作业日志路径"等
# (7) 创建训练作业
```
3. 使用VOCtrain数据集训练s8，微调上一步的模型。训练步骤如下：
```python
# (1) 执行 a 或者 b.
# a. 在 base_config.yaml 文件中设置 "enable_modelarts=True"
# 在 base_config.yaml 文件中设置 "model='deeplab_v3_s8'"
# 在 base_config.yaml 文件中设置 "batch_size=16"
# 在 base_config.yaml 文件中设置 "base_lr=0.008"
# 在 base_config.yaml 文件中设置 "loss_scale=2048"
# 在 base_config.yaml 文件中设置 "data_file='/cache/data/vocaug/voctrain_mindrecord/voctrain_mindrecord00'"
# 在 base_config.yaml 文件中设置 "checkpoint_url=/The path of checkpoint in S3/"
# 在 base_config.yaml 文件中设置 "ckpt_pre_trained=/cache/checkpoint_path/path_to_pretrain/deeplab_v3_s8-800_82.ckpt"
# 在 base_config.yaml 文件中设置 "is_distributed=True"
# 在 base_config.yaml 文件中设置 "save_steps=110"
# 在 base_config.yaml 文件中设置 其他参数
# b. 在网页上设置 "enable_modelarts=True"
# 在网页上设置 "model='deeplab_v3_s8'"
# 在网页上设置 "batch_size=16"
# 在网页上设置 "base_lr=0.008"
# 在网页上设置 "loss_scale=2048"
# 在网页上设置 "data_file='/cache/data/vocaug/voctrain_mindrecord/voctrain_mindrecord00'"
# 在网页上设置 "checkpoint_url=/The path of checkpoint in S3/"
# 在网页上设置 "ckpt_pre_trained=/cache/checkpoint_path/path_to_pretrain/deeplab_v3_s8-800_82.ckpt"
# 在网页上设置 "is_distributed=True"
# 在网页上设置 "save_steps=110"
# 在网页上设置 其他参数
# (2) 上传你的预训练模型到 S3 桶上
# (3) 上传你的压缩数据集到 S3 桶上 (你也可以上传原始的数据集，但那可能会很慢。)
# (4) 在网页上设置你的代码路径为 "/path/deeplabv3"
# (5) 在网页上设置启动文件为 "train.py"
# (6) 在网页上设置"训练数据集"、"训练输出文件路径"、"作业日志路径"等
# (7) 创建训练作业
```
- 在 ModelArts 进行验证 (如果你想在modelarts上运行，可以参考以下文档 [modelarts](https://support.huaweicloud.com/modelarts/))
1. 使用voc val数据集评估s16。评估步骤如下：
```python
# (1) 执行 a 或者 b.
# a. 在 base_config.yaml 文件中设置 "enable_modelarts=True"
# 在 base_config.yaml 文件中设置 "model='deeplab_v3_s16'"
# 在 base_config.yaml 文件中设置 "batch_size=32"
# 在 base_config.yaml 文件中设置 "scales_type=0"
# 在 base_config.yaml 文件中设置 "freeze_bn=True"
# 在 base_config.yaml 文件中设置 "data_root='/cache/data/vocaug'"
# 在 base_config.yaml 文件中设置 "data_lst='/cache/data/vocaug/voc_val_lst.txt'"
# 在 base_config.yaml 文件中设置 "checkpoint_url=/The path of checkpoint in S3/"
# 在 base_config.yaml 文件中设置 "ckpt_path='/cache/checkpoint_path/path_to_pretrain/deeplab_v3_s16-300_41.ckpt'"
# 在 base_config.yaml 文件中设置 其他参数
# b. 在网页上设置 "enable_modelarts=True"
# 在网页上设置 "model=deeplab_v3_s16"
# 在网页上设置 "batch_size=32"
# 在网页上设置 "scales_type=0"
# 在网页上设置 "freeze_bn=True"
# 在网页上设置 "data_root=/cache/data/vocaug"
# 在网页上设置 "data_lst=/cache/data/vocaug/voc_val_lst.txt"
# 在网页上设置 "checkpoint_url=/The path of checkpoint in S3/"
# 在网页上设置 "ckpt_path=/cache/checkpoint_path/path_to_pretrain/deeplab_v3_s16-300_41.ckpt"
# 在网页上设置 其他参数
# (2) 上传你的预训练模型到 S3 桶上
# (3) 上传你的压缩数据集到 S3 桶上 (你也可以上传原始的数据集，但那可能会很慢。)
# (4) 在网页上设置你的代码路径为 "/path/deeplabv3"
# (5) 在网页上设置启动文件为 "eval.py"
# (6) 在网页上设置"训练数据集"、"训练输出文件路径"、"作业日志路径"等
# (7) 创建训练作业
```
2. 使用voc val数据集评估s8。评估步骤如下：
```python
# (1) 执行 a 或者 b.
# a. 在 base_config.yaml 文件中设置 "enable_modelarts=True"
# 在 base_config.yaml 文件中设置 "model='deeplab_v3_s8'"
# 在 base_config.yaml 文件中设置 "batch_size=16"
# 在 base_config.yaml 文件中设置 "scales_type=0"
# 在 base_config.yaml 文件中设置 "freeze_bn=True"
# 在 base_config.yaml 文件中设置 "data_root='/cache/data/vocaug'"
# 在 base_config.yaml 文件中设置 "data_lst='/cache/data/vocaug/voc_val_lst.txt'"
# 在 base_config.yaml 文件中设置 "checkpoint_url='/The path of checkpoint in S3/'"
# 在 base_config.yaml 文件中设置 "ckpt_path='/cache/checkpoint_path/path_to_pretrain/deeplab_v3_s8-300_11.ckpt'"
# 在 base_config.yaml 文件中设置 其他参数
# b. 在网页上设置 "enable_modelarts=True"
# 在网页上设置 "model=deeplab_v3_s8"
# 在网页上设置 "batch_size=16"
# 在网页上设置 "scales_type=0"
# 在网页上设置 "freeze_bn=True"
# 在网页上设置 "data_root=/cache/data/vocaug"
# 在网页上设置 "data_lst=/cache/data/vocaug/voc_val_lst.txt"
# 在网页上设置 "checkpoint_url=/The path of checkpoint in S3/"
# 在网页上设置 "ckpt_path=/cache/checkpoint_path/path_to_pretrain/deeplab_v3_s8-300_11.ckpt"
# 在网页上设置 其他参数.
# (2) 上传你的预训练模型到 S3 桶上
# (3) 上传你的压缩数据集到 S3 桶上 (你也可以上传原始的数据集，但那可能会很慢。)
# (4) 在网页上设置你的代码路径为 "/path/deeplabv3"
# (5) 在网页上设置启动文件为 "eval.py"
# (6) 在网页上设置"训练数据集"、"训练输出文件路径"、"作业日志路径"等
# (7) 创建训练作业
```
3. 使用voc val数据集评估多尺度s8。评估步骤如下：
```python
# (1) 执行 a 或者 b.
# a. 在 base_config.yaml 文件中设置 "enable_modelarts=True"
# 在 base_config.yaml 文件中设置 "model='deeplab_v3_s8'"
# 在 base_config.yaml 文件中设置 "batch_size=16"
# 在 base_config.yaml 文件中设置 "scales_type=1"
# 在 base_config.yaml 文件中设置 "freeze_bn=True"
# 在 base_config.yaml 文件中设置 "data_root='/cache/data/vocaug'"
# 在 base_config.yaml 文件中设置 "data_lst='/cache/data/vocaug/voc_val_lst.txt'"
# 在 base_config.yaml 文件中设置 "checkpoint_url='/The path of checkpoint in S3/'"
# 在 base_config.yaml 文件中设置 "ckpt_path='/cache/checkpoint_path/path_to_pretrain/deeplab_v3_s8-300_11.ckpt'"
# 在 base_config.yaml 文件中设置 其他参数
# b. 在网页上设置 "enable_modelarts=True"
# 在网页上设置 "model=deeplab_v3_s8"
# 在网页上设置 "batch_size=16"
# 在网页上设置 "scales_type=1"
# 在网页上设置 "freeze_bn=True"
# 在网页上设置 "data_root=/cache/data/vocaug"
# 在网页上设置 "data_lst=/cache/data/vocaug/voc_val_lst.txt"
# 在网页上设置 "checkpoint_url=/The path of checkpoint in S3/"
# 在网页上设置 "ckpt_path=/cache/checkpoint_path/path_to_pretrain/deeplab_v3_s8-300_11.ckpt"
# 在网页上设置 其他参数
# (2) 上传你的预训练模型到 S3 桶上
# (3) 上传你的压缩数据集到 S3 桶上 (你也可以上传原始的数据集，但那可能会很慢。)
# (4) 在网页上设置你的代码路径为 "/path/deeplabv3"
# (5) 在网页上设置启动文件为 "eval.py"
# (6) 在网页上设置"训练数据集"、"训练输出文件路径"、"作业日志路径"等
# (7) 创建训练作业
```
4. 使用voc val数据集评估多尺度和翻转s8。评估步骤如下：
```python
# (1) 执行 a 或者 b.
# a. 在 base_config.yaml 文件中设置 "enable_modelarts=True"
# 在 base_config.yaml 文件中设置 "model='deeplab_v3_s8'"
# 在 base_config.yaml 文件中设置 "batch_size=16"
# 在 base_config.yaml 文件中设置 "scales_type=1"
# 在 base_config.yaml 文件中设置 "freeze_bn=True"
# 在 base_config.yaml 文件中设置 "flip=True"
# 在 base_config.yaml 文件中设置 "data_root='/cache/data/vocaug'"
# 在 base_config.yaml 文件中设置 "data_lst='/cache/data/vocaug/voc_val_lst.txt'"
# 在 base_config.yaml 文件中设置 "checkpoint_url='/The path of checkpoint in S3/'"
# 在 base_config.yaml 文件中设置 "ckpt_path='/cache/checkpoint_path/path_to_pretrain/deeplab_v3_s8-300_11.ckpt'"
# 在 base_config.yaml 文件中设置 其他参数
# b. 在网页上设置 "enable_modelarts=True"
# 在网页上设置 "model=deeplab_v3_s8"
# 在网页上设置 "batch_size=16"
# 在网页上设置 "scales_type=1"
# 在网页上设置 "freeze_bn=True"
# 在网页上设置 "flip=True"
# 在网页上设置 "data_root=/cache/data/vocaug"
# 在网页上设置 "data_lst=/cache/data/vocaug/voc_val_lst.txt"
# 在网页上设置 "checkpoint_url=/The path of checkpoint in S3/"
# 在网页上设置 "ckpt_path=/cache/checkpoint_path/path_to_pretrain/deeplab_v3_s8-300_11.ckpt"
# 在网页上设置 其他参数
# (2) 上传你的预训练模型到 S3 桶上
# (3) 上传你的压缩数据集到 S3 桶上 (你也可以上传原始的数据集，但那可能会很慢。)
# (4) 在网页上设置你的代码路径为 "/path/deeplabv3"
# (5) 在网页上设置启动文件为 "eval.py"
# (6) 在网页上设置"训练数据集"、"训练输出文件路径"、"作业日志路径"等
# (7) 创建训练作业
```
# 脚本说明
## 脚本及样例代码

100
model_zoo/official/cv/deeplabv3/default_config.yaml Normal file
View File

@ -0,0 +1,100 @@
# Builtin Configurations(DO NOT CHANGE THESE CONFIGURATIONS unless you know exactly what you are doing)
enable_modelarts: False
# Url for modelarts
data_url: ""
train_url: ""
checkpoint_url: ""
# Path for local
data_path: "/cache/data"
output_path: "/cache/train"
load_path: "/cache/checkpoint_path"
device_target: "Ascend" # ['Ascend', 'CPU']
# ==============================================================================
# Training options
train_dir: "/cache/train/ckpt"
# dataset
need_modelarts_dataset_unzip: True
data_file: ""
batch_size: 32
crop_size: 513
image_mean: [103.53, 116.28, 123.675]
image_std: [57.375, 57.120, 58.395]
min_scale: 0.5
max_scale: 2.0
ignore_label: 255
num_classes: 21
# optimizer
train_epochs: 300
lr_type: "cos"
base_lr: 0.015
lr_decay_step: 40000
lr_decay_rate: 0.1
loss_scale: 3072.0
# model
model: "deeplab_v3_s16"
freeze_bn: False
ckpt_pre_trained: ""
filter_weight: False
# train
is_distributed: False
rank: 0
group_size: 1
save_steps: 3000
keep_checkpoint_max: 1
# eval param
data_root: ""
data_lst: ""
scales: [1.0,]
scales_list: [[1.0,], [0.5, 0.75, 1.0, 1.25, 1.75]]
scales_type: 0
flip: False
ckpt_path: ""
input_format: "NCHW" # ["NCHW", "NHWC"]
---
# Help description for each configuration
enable_modelarts: "Whether training on modelarts, default: False"
data_url: "Url for modelarts"
train_url: "Url for modelarts"
data_path: "The location of the input data."
output_path: "The location of the output file."
device_target: 'Target device type'
train_dir: "where training log and ckpts saved"
data_file: "path and name of one mindrecord file"
batch_size: "batch size"
crop_size: "crop size"
image_mean: "image mean"
image_std: "image std"
min_scale: "minimum scale of data argumentation"
max_scale: "maximum scale of data argumentation"
ignore_label: "ignore label"
num_classes: "number of classes"
train_epochs: "epoch"
lr_type: "type of learning rate"
base_lr: "base learning rate"
lr_decay_step: "learning rate decay step"
lr_decay_rate: "learning rate decay rate"
loss_scale: "loss scale"
model: "select model"
freeze_bn: "freeze bn"
ckpt_pre_trained: "pretrained model"
filter_weight: "Filter the last weight parameters, default is False."
is_distributed: "distributed training"
rank: "local rank of distributed"
group_size: "world size of distributed"
save_steps: "steps interval for saving"
keep_checkpoint_max: "max checkpoint for saving"
data_root: "root path of val data"
data_lst: "list of val data"
scales: "scales of evaluation"
flip: "perform left-right flip"
ckpt_path: "model to evaluat"
input_format: "NCHW or NHWC"

91
model_zoo/official/cv/deeplabv3/eval.py
View File

@ -15,7 +15,7 @@
"""eval deeplabv3."""
import os
import argparse
import time
import numpy as np
import cv2
from mindspore import Tensor
@ -25,34 +25,18 @@ import mindspore.ops as ops
from mindspore import context
from mindspore.train.serialization import load_checkpoint, load_param_into_net
from src.nets import net_factory
from utils.config import config
from utils.moxing_adapter import moxing_wrapper
from utils.device_adapter import get_device_id, get_device_num, get_rank_id
context.set_context(mode=context.GRAPH_MODE, device_target="Ascend", save_graphs=False,
device_id=int(os.getenv('DEVICE_ID')))
device_id=get_device_id())
def parse_args():
parser = argparse.ArgumentParser('mindspore deeplabv3 eval')
# parser.add_argument('--scales', type=float, action='append', help='scales of evaluation')
# parser.add_argument('--flip', action='store_true', help='perform left-right flip')
# val data
parser.add_argument('--data_root', type=str, default='', help='root path of val data')
parser.add_argument('--data_lst', type=str, default='', help='list of val data')
parser.add_argument('--batch_size', type=int, default=16, help='batch size')
parser.add_argument('--crop_size', type=int, default=513, help='crop size')
parser.add_argument('--image_mean', type=list, default=[103.53, 116.28, 123.675], help='image mean')
parser.add_argument('--image_std', type=list, default=[57.375, 57.120, 58.395], help='image std')
parser.add_argument('--scales', type=float, action='append', help='scales of evaluation')
parser.add_argument('--flip', action='store_true', help='perform left-right flip')
parser.add_argument('--ignore_label', type=int, default=255, help='ignore label')
parser.add_argument('--num_classes', type=int, default=21, help='number of classes')
# model
parser.add_argument('--model', type=str, default='deeplab_v3_s16', help='select model')
parser.add_argument('--freeze_bn', action='store_true', default=False, help='freeze bn')
parser.add_argument('--ckpt_path', type=str, default='', help='model to evaluate')
parser.add_argument("--input_format", type=str, choices=["NCHW", "NHWC"], default="NCHW",
help="NCHW or NHWC")
args, _ = parser.parse_known_args()
return args
def cal_hist(a, b, n):
@ -153,8 +137,63 @@ def eval_batch_scales(args, eval_net, img_lst, scales,
return result_msk
def modelarts_pre_process():
'''modelarts pre process function.'''
def unzip(zip_file, save_dir):
import zipfile
s_time = time.time()
if not os.path.exists(os.path.join(save_dir, "vocaug")):
zip_isexist = zipfile.is_zipfile(zip_file)
if zip_isexist:
fz = zipfile.ZipFile(zip_file, 'r')
data_num = len(fz.namelist())
print("Extract Start...")
print("unzip file num: {}".format(data_num))
i = 0
for file in fz.namelist():
if i % int(data_num / 100) == 0:
print("unzip percent: {}%".format(i / int(data_num / 100)), flush=True)
i += 1
fz.extract(file, save_dir)
print("cost time: {}min:{}s.".format(int((time.time() - s_time) / 60),
int(int(time.time() - s_time) % 60)))
print("Extract Done.")
else:
print("This is not zip.")
else:
print("Zip has been extracted.")
if config.need_modelarts_dataset_unzip:
zip_file_1 = os.path.join(config.data_path, "vocaug.zip")
save_dir_1 = os.path.join(config.data_path)
sync_lock = "/tmp/unzip_sync.lock"
# Each server contains 8 devices as most.
if get_device_id() % min(get_device_num(), 8) == 0 and not os.path.exists(sync_lock):
print("Zip file path: ", zip_file_1)
print("Unzip file save dir: ", save_dir_1)
unzip(zip_file_1, save_dir_1)
print("===Finish extract data synchronization===")
try:
os.mknod(sync_lock)
except IOError:
pass
while True:
if os.path.exists(sync_lock):
break
time.sleep(1)
print("Device: {}, Finish sync unzip data from {} to {}.".format(get_device_id(), zip_file_1, save_dir_1))
config.train_dir = os.path.join(config.output_path, str(get_rank_id()), config.train_dir)
@moxing_wrapper(pre_process=modelarts_pre_process)
def net_eval():
args = parse_args()
config.scales = config.scales_list[config.scales_type]
args = config
# data list
with open(args.data_lst) as f:

4
model_zoo/official/cv/deeplabv3/scripts/build_data.sh
View File

@ -15,7 +15,9 @@
# ============================================================================
export DEVICE_ID=7
python /PATH/TO/MODEL_ZOO_CODE/data/build_seg_data.py --data_root=/PATH/TO/DATA_ROOT \
EXECUTE_PATH=$(pwd)
python ${EXECUTE_PATH}/../src/data/build_seg_data.py --data_root=/PATH/TO/DATA_ROOT \
--data_lst=/PATH/TO/DATA_lst.txt \
--dst_path=/PATH/TO/MINDRECORED_NAME.mindrecord \
--num_shards=8 \

37
model_zoo/official/cv/deeplabv3/scripts/run_distribute_train_s16_r1.sh
View File

@ -14,11 +14,34 @@
# limitations under the License.
# ============================================================================
if [ $# != 1 ]
then
echo "Usage: sh run_distribute_train_base.sh [RANK_TABLE_FILE]"
exit 1
fi
get_real_path(){
if [ "${1:0:1}" == "/" ]; then
echo "$1"
else
echo "$(realpath -m $PWD/$1)"
fi
}
PATH1=$(get_real_path $1)
echo $PATH1
if [ ! -f $PATH1 ]
then
echo "error: RANK_TABLE_FILE=$PATH1 is not a file"
exit 1
fi
ulimit -c unlimited
train_path=/PATH/TO/EXPERIMENTS_DIR
EXECUTE_PATH=$(pwd)
train_path=${EXECUTE_PATH}/s16_aug_train
export SLOG_PRINT_TO_STDOUT=0
train_code_path=/PATH/TO/MODEL_ZOO_CODE
export RANK_TABLE_FILE=${train_code_path}/src/tools/rank_table_8p.json
export RANK_TABLE_FILE=$PATH1
export RANK_SIZE=8
export RANK_START_ID=0
@ -35,8 +58,8 @@ do
echo 'start rank='${i}', device id='${DEVICE_ID}'...'
mkdir ${train_path}/device${DEVICE_ID}
cd ${train_path}/device${DEVICE_ID} || exit
python ${train_code_path}/train.py --train_dir=${train_path}/ckpt \
--data_file=/PATH/TO/MINDRECORD_NAME \
python ${EXECUTE_PATH}/../train.py --train_dir=${train_path}/ckpt \
--data_file=/PATH_TO_DATA/vocaug/vocaug_mindrecord/vocaug_mindrecord0 \
--train_epochs=300 \
--batch_size=32 \
--crop_size=513 \
@ -48,7 +71,7 @@ do
--num_classes=21 \
--model=deeplab_v3_s16 \
--ckpt_pre_trained=/PATH/TO/PRETRAIN_MODEL \
--is_distributed \
--is_distributed=True \
--save_steps=410 \
--keep_checkpoint_max=200 >log 2>&1 &
--keep_checkpoint_max=1 >log 2>&1 &
done

37
model_zoo/official/cv/deeplabv3/scripts/run_distribute_train_s8_r1.sh
View File

@ -14,11 +14,34 @@
# limitations under the License.
# ============================================================================
if [ $# != 1 ]
then
echo "Usage: sh run_distribute_train_base.sh [RANK_TABLE_FILE]"
exit 1
fi
get_real_path(){
if [ "${1:0:1}" == "/" ]; then
echo "$1"
else
echo "$(realpath -m $PWD/$1)"
fi
}
PATH1=$(get_real_path $1)
echo $PATH1
if [ ! -f $PATH1 ]
then
echo "error: RANK_TABLE_FILE=$PATH1 is not a file"
exit 1
fi
ulimit -c unlimited
train_path=/PATH/TO/EXPERIMENTS_DIR
EXECUTE_PATH=$(pwd)
train_path=${EXECUTE_PATH}/s8_aug_train
export SLOG_PRINT_TO_STDOUT=0
train_code_path=/PATH/TO/MODEL_ZOO_CODE
export RANK_TABLE_FILE=${train_code_path}/src/tools/rank_table_8p.json
export RANK_TABLE_FILE=$PATH1
export RANK_SIZE=8
export RANK_START_ID=0
@ -35,8 +58,8 @@ do
echo 'start rank='${i}', device id='${DEVICE_ID}'...'
mkdir ${train_path}/device${DEVICE_ID}
cd ${train_path}/device${DEVICE_ID} || exit
python ${train_code_path}/train.py --train_dir=${train_path}/ckpt \
--data_file=/PATH/TO/MINDRECORD_NAME \
python ${EXECUTE_PATH}/../train.py --train_dir=${train_path}/ckpt \
--data_file=/PATH_TO_DATA/vocaug/vocaug_mindrecord/vocaug_mindrecord0 \
--train_epochs=800 \
--batch_size=16 \
--crop_size=513 \
@ -49,7 +72,7 @@ do
--model=deeplab_v3_s8 \
--loss_scale=2048 \
--ckpt_pre_trained=/PATH/TO/PRETRAIN_MODEL \
--is_distributed \
--is_distributed=True \
--save_steps=820 \
--keep_checkpoint_max=200 >log 2>&1 &
--keep_checkpoint_max=1 >log 2>&1 &
done

37
model_zoo/official/cv/deeplabv3/scripts/run_distribute_train_s8_r2.sh
View File

@ -14,11 +14,34 @@
# limitations under the License.
# ============================================================================
if [ $# != 1 ]
then
echo "Usage: sh run_distribute_train_base.sh [RANK_TABLE_FILE]"
exit 1
fi
get_real_path(){
if [ "${1:0:1}" == "/" ]; then
echo "$1"
else
echo "$(realpath -m $PWD/$1)"
fi
}
PATH1=$(get_real_path $1)
echo $PATH1
if [ ! -f $PATH1 ]
then
echo "error: RANK_TABLE_FILE=$PATH1 is not a file"
exit 1
fi
ulimit -c unlimited
train_path=/PATH/TO/EXPERIMENTS_DIR
EXECUTE_PATH=$(pwd)
train_path=${EXECUTE_PATH}/s8_voc_train
export SLOG_PRINT_TO_STDOUT=0
train_code_path=/PATH/TO/MODEL_ZOO_CODE
export RANK_TABLE_FILE=${train_code_path}/src/tools/rank_table_8p.json
export RANK_TABLE_FILE=$PATH1
export RANK_SIZE=8
export RANK_START_ID=0
@ -35,8 +58,8 @@ do
echo 'start rank='${i}', device id='${DEVICE_ID}'...'
mkdir ${train_path}/device${DEVICE_ID}
cd ${train_path}/device${DEVICE_ID} || exit
python ${train_code_path}/train.py --train_dir=${train_path}/ckpt \
--data_file=/PATH/TO/MINDRECORD_NAME \
python ${EXECUTE_PATH}/../train.py --train_dir=${train_path}/ckpt \
--data_file=/PATH_TO_DATA/vocaug/voctrain_mindrecord/voctrain_mindrecord00 \
--train_epochs=300 \
--batch_size=16 \
--crop_size=513 \
@ -49,7 +72,7 @@ do
--model=deeplab_v3_s8 \
--loss_scale=2048 \
--ckpt_pre_trained=/PATH/TO/PRETRAIN_MODEL \
--is_distributed \
--is_distributed=True \
--save_steps=110 \
--keep_checkpoint_max=200 >log 2>&1 &
--keep_checkpoint_max=1 >log 2>&1 &
done

14
model_zoo/official/cv/deeplabv3/scripts/run_eval_s16.sh
View File

@ -14,24 +14,24 @@
# limitations under the License.
# ============================================================================
export DEVICE_ID=3
export DEVICE_ID=0
export SLOG_PRINT_TO_STDOUT=0
train_code_path=/PATH/TO/MODEL_ZOO_CODE
eval_path=/PATH/TO/EVAL
EXECUTE_PATH=$(pwd)
eval_path=${EXECUTE_PATH}/s16_eval
if [ -d ${eval_path} ]; then
rm -rf ${eval_path}
fi
mkdir -p ${eval_path}
python ${train_code_path}/eval.py --data_root=/PATH/TO/DATA \
--data_lst=/PATH/TO/DATA_lst.txt \
python ${EXECUTE_PATH}/../eval.py --data_root=/PATH_TO_DATA/vocaug \
--data_lst=/PATH_TO_DATA/vocaug/voc_val_lst.txt \
--batch_size=32 \
--crop_size=513 \
--ignore_label=255 \
--num_classes=21 \
--model=deeplab_v3_s16 \
--scales=1.0 \
--freeze_bn \
--scales_type=0 \
--freeze_bn=True \
--ckpt_path=/PATH/TO/PRETRAIN_MODEL >${eval_path}/eval_log 2>&1 &

14
model_zoo/official/cv/deeplabv3/scripts/run_eval_s8.sh
View File

@ -14,24 +14,24 @@
# limitations under the License.
# ============================================================================
export DEVICE_ID=3
export DEVICE_ID=1
export SLOG_PRINT_TO_STDOUT=0
train_code_path=/PATH/TO/MODEL_ZOO_CODE
eval_path=/PATH/TO/EVAL
EXECUTE_PATH=$(pwd)
eval_path=${EXECUTE_PATH}/s8_eval
if [ -d ${eval_path} ]; then
rm -rf ${eval_path}
fi
mkdir -p ${eval_path}
python ${train_code_path}/eval.py --data_root=/PATH/TO/DATA \
--data_lst=/PATH/TO/DATA_lst.txt \
python ${EXECUTE_PATH}/../eval.py --data_root=/PATH_TO_DATA/vocaug \
--data_lst=/PATH_TO_DATA/vocaug/voc_val_lst.txt \
--batch_size=16 \
--crop_size=513 \
--ignore_label=255 \
--num_classes=21 \
--model=deeplab_v3_s8 \
--scales=1.0 \
--freeze_bn \
--scales_type=0 \
--freeze_bn=True \
--ckpt_path=/PATH/TO/PRETRAIN_MODEL >${eval_path}/eval_log 2>&1 &

18
model_zoo/official/cv/deeplabv3/scripts/run_eval_s8_multiscale.sh
View File

@ -14,28 +14,24 @@
# limitations under the License.
# ============================================================================
export DEVICE_ID=3
export DEVICE_ID=2
export SLOG_PRINT_TO_STDOUT=0
train_code_path=/PATH/TO/MODEL_ZOO_CODE
eval_path=/PATH/TO/EVAL
EXECUTE_PATH=$(pwd)
eval_path=${EXECUTE_PATH}/multiscale_eval
if [ -d ${eval_path} ]; then
rm -rf ${eval_path}
fi
mkdir -p ${eval_path}
python ${train_code_path}/eval.py --data_root=/PATH/TO/DATA \
--data_lst=/PATH/TO/DATA_lst.txt \
python ${EXECUTE_PATH}/../eval.py --data_root=/PATH_TO_DATA/vocaug \
--data_lst=/PATH_TO_DATA/vocaug/voc_val_lst.txt \
--batch_size=16 \
--crop_size=513 \
--ignore_label=255 \
--num_classes=21 \
--model=deeplab_v3_s8 \
--scales=0.5 \
--scales=0.75 \
--scales=1.0 \
--scales=1.25 \
--scales=1.75 \
--freeze_bn \
--scales_type=1 \
--freeze_bn=True \
--ckpt_path=/PATH/TO/PRETRAIN_MODEL >${eval_path}/eval_log 2>&1 &

18
model_zoo/official/cv/deeplabv3/scripts/run_eval_s8_multiscale_flip.sh
View File

@ -16,27 +16,23 @@
export DEVICE_ID=3
export SLOG_PRINT_TO_STDOUT=0
train_code_path=/PATH/TO/MODEL_ZOO_CODE
eval_path=/PATH/TO/EVAL
EXECUTE_PATH=$(pwd)
eval_path=${EXECUTE_PATH}/multiscale_flip_eval
if [ -d ${eval_path} ]; then
rm -rf ${eval_path}
fi
mkdir -p ${eval_path}
python ${train_code_path}/eval.py --data_root=/PATH/TO/DATA \
--data_lst=/PATH/TO/DATA_lst.txt \
python ${EXECUTE_PATH}/../eval.py --data_root=/PATH_TO_DATA/vocaug \
--data_lst=/PATH_TO_DATA/vocaug/voc_val_lst.txt \
--batch_size=16 \
--crop_size=513 \
--ignore_label=255 \
--num_classes=21 \
--model=deeplab_v3_s8 \
--scales=0.5 \
--scales=0.75 \
--scales=1.0 \
--scales=1.25 \
--scales=1.75 \
--flip \
--freeze_bn \
--scales_type=1 \
--flip=True \
--freeze_bn=True \
--ckpt_path=/PATH/TO/PRETRAIN_MODEL >${eval_path}/eval_log 2>&1 &

8
model_zoo/official/cv/deeplabv3/scripts/run_standalone_train.sh
View File

@ -14,10 +14,10 @@
# limitations under the License.
# ============================================================================
export DEVICE_ID=5
export DEVICE_ID=0
export SLOG_PRINT_TO_STDOUT=0
train_path=/PATH/TO/EXPERIMENTS_DIR
train_code_path=/PATH/TO/MODEL_ZOO_CODE
EXECUTE_PATH=$(pwd)
train_path=${EXECUTE_PATH}/s16_aug_train_1p
if [ -d ${train_path} ]; then
rm -rf ${train_path}
@ -27,7 +27,7 @@ mkdir ${train_path}/device${DEVICE_ID}
mkdir ${train_path}/ckpt
cd ${train_path}/device${DEVICE_ID} || exit
python ${train_code_path}/train.py --data_file=/PATH/TO/MINDRECORD_NAME \
python ${EXECUTE_PATH}/../train.py --data_file=/PATH_TO_DATA/vocaug/vocaug_mindrecord/vocaug_mindrecord0 \
--train_dir=${train_path}/ckpt \
--train_epochs=200 \
--batch_size=32 \

6
model_zoo/official/cv/deeplabv3/scripts/run_standalone_train_cpu.sh
View File

@ -16,8 +16,8 @@
export DEVICE_ID=0
export SLOG_PRINT_TO_STDOUT=0
train_path=/PATH/TO/EXPERIMENTS_DIR
train_code_path=/PATH/TO/MODEL_ZOO_CODE
EXECUTE_PATH=$(pwd)
train_path=${EXECUTE_PATH}/s16_aug_train_cpu
if [ -d ${train_path} ]; then
rm -rf ${train_path}
@ -27,7 +27,7 @@ mkdir ${train_path}/device${DEVICE_ID}
mkdir ${train_path}/ckpt
cd ${train_path}/device${DEVICE_ID} || exit
python ${train_code_path}/train.py --data_file=/PATH/TO/MINDRECORD_NAME \
python ${EXECUTE_PATH}/../train.py --data_file=/PATH_TO_DATA/vocaug/vocaug_mindrecord/vocaug_mindrecord0 \
--device_target=CPU \
--train_dir=${train_path}/ckpt \
--train_epochs=200 \

93
model_zoo/official/cv/deeplabv3/train.py
View File

@ -15,8 +15,7 @@
"""train deeplabv3."""
import os
import argparse
import ast
import time
from mindspore import context
from mindspore.train.model import Model
from mindspore.context import ParallelMode
@ -31,6 +30,9 @@ from src.data import dataset as data_generator
from src.loss import loss
from src.nets import net_factory
from src.utils import learning_rates
from utils.config import config
from utils.moxing_adapter import moxing_wrapper
from utils.device_adapter import get_device_id, get_device_num, get_rank_id
set_seed(1)
@ -47,57 +49,68 @@ class BuildTrainNetwork(nn.Cell):
return net_loss
def parse_args():
parser = argparse.ArgumentParser('mindspore deeplabv3 training')
parser.add_argument('--train_dir', type=str, default='', help='where training log and ckpts saved')
def modelarts_pre_process():
'''modelarts pre process function.'''
def unzip(zip_file, save_dir):
import zipfile
s_time = time.time()
if not os.path.exists(os.path.join(save_dir, "vocaug")):
zip_isexist = zipfile.is_zipfile(zip_file)
if zip_isexist:
fz = zipfile.ZipFile(zip_file, 'r')
data_num = len(fz.namelist())
print("Extract Start...")
print("unzip file num: {}".format(data_num))
i = 0
for file in fz.namelist():
if i % int(data_num / 100) == 0:
print("unzip percent: {}%".format(i / int(data_num / 100)), flush=True)
i += 1
fz.extract(file, save_dir)
print("cost time: {}min:{}s.".format(int((time.time() - s_time) / 60),
int(int(time.time() - s_time) % 60)))
print("Extract Done.")
else:
print("This is not zip.")
else:
print("Zip has been extracted.")
# dataset
parser.add_argument('--data_file', type=str, default='', help='path and name of one mindrecord file')
parser.add_argument('--batch_size', type=int, default=32, help='batch size')
parser.add_argument('--crop_size', type=int, default=513, help='crop size')
parser.add_argument('--image_mean', type=list, default=[103.53, 116.28, 123.675], help='image mean')
parser.add_argument('--image_std', type=list, default=[57.375, 57.120, 58.395], help='image std')
parser.add_argument('--min_scale', type=float, default=0.5, help='minimum scale of data argumentation')
parser.add_argument('--max_scale', type=float, default=2.0, help='maximum scale of data argumentation')
parser.add_argument('--ignore_label', type=int, default=255, help='ignore label')
parser.add_argument('--num_classes', type=int, default=21, help='number of classes')
if config.need_modelarts_dataset_unzip:
zip_file_1 = os.path.join(config.data_path, "vocaug.zip")
save_dir_1 = os.path.join(config.data_path)
# optimizer
parser.add_argument('--train_epochs', type=int, default=300, help='epoch')
parser.add_argument('--lr_type', type=str, default='cos', help='type of learning rate')
parser.add_argument('--base_lr', type=float, default=0.015, help='base learning rate')
parser.add_argument('--lr_decay_step', type=int, default=40000, help='learning rate decay step')
parser.add_argument('--lr_decay_rate', type=float, default=0.1, help='learning rate decay rate')
parser.add_argument('--loss_scale', type=float, default=3072.0, help='loss scale')
sync_lock = "/tmp/unzip_sync.lock"
# model
parser.add_argument('--model', type=str, default='deeplab_v3_s16', help='select model')
parser.add_argument('--freeze_bn', action='store_true', help='freeze bn')
parser.add_argument('--ckpt_pre_trained', type=str, default='', help='pretrained model')
parser.add_argument("--filter_weight", type=ast.literal_eval, default=False,
help="Filter the last weight parameters, default is False.")
# Each server contains 8 devices as most.
if get_device_id() % min(get_device_num(), 8) == 0 and not os.path.exists(sync_lock):
print("Zip file path: ", zip_file_1)
print("Unzip file save dir: ", save_dir_1)
unzip(zip_file_1, save_dir_1)
print("===Finish extract data synchronization===")
try:
os.mknod(sync_lock)
except IOError:
pass
# train
parser.add_argument('--device_target', type=str, default='Ascend', choices=['Ascend', 'CPU'],
help='device where the code will be implemented. (Default: Ascend)')
parser.add_argument('--is_distributed', action='store_true', help='distributed training')
parser.add_argument('--rank', type=int, default=0, help='local rank of distributed')
parser.add_argument('--group_size', type=int, default=1, help='world size of distributed')
parser.add_argument('--save_steps', type=int, default=3000, help='steps interval for saving')
parser.add_argument('--keep_checkpoint_max', type=int, default=int, help='max checkpoint for saving')
while True:
if os.path.exists(sync_lock):
break
time.sleep(1)
args, _ = parser.parse_known_args()
return args
print("Device: {}, Finish sync unzip data from {} to {}.".format(get_device_id(), zip_file_1, save_dir_1))
config.train_dir = os.path.join(config.output_path, str(get_rank_id()), config.train_dir)
@moxing_wrapper(pre_process=modelarts_pre_process)
def train():
args = parse_args()
args = config
if args.device_target == "CPU":
context.set_context(mode=context.GRAPH_MODE, save_graphs=False, device_target="CPU")
else:
context.set_context(mode=context.GRAPH_MODE, enable_auto_mixed_precision=True, save_graphs=False,
device_target="Ascend", device_id=int(os.getenv('DEVICE_ID')))
device_target="Ascend", device_id=get_device_id())
# init multicards training
if args.is_distributed:

0
model_zoo/official/cv/deeplabv3/utils/__init__.py Normal file
View File

127
model_zoo/official/cv/deeplabv3/utils/config.py Normal file
View File

@ -0,0 +1,127 @@
# Copyright 2021 Huawei Technologies Co., Ltd
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ============================================================================
"""Parse arguments"""
import os
import ast
import argparse
from pprint import pprint, pformat
import yaml
class Config:
"""
Configuration namespace. Convert dictionary to members.
"""
def __init__(self, cfg_dict):
for k, v in cfg_dict.items():
if isinstance(v, (list, tuple)):
setattr(self, k, [Config(x) if isinstance(x, dict) else x for x in v])
else:
setattr(self, k, Config(v) if isinstance(v, dict) else v)
def __str__(self):
return pformat(self.__dict__)
def __repr__(self):
return self.__str__()
def parse_cli_to_yaml(parser, cfg, helper=None, choices=None, cfg_path="default_config.yaml"):
"""
Parse command line arguments to the configuration according to the default yaml.
Args:
parser: Parent parser.
cfg: Base configuration.
helper: Helper description.
cfg_path: Path to the default yaml config.
"""
parser = argparse.ArgumentParser(description="[REPLACE THIS at config.py]",
parents=[parser])
helper = {} if helper is None else helper
choices = {} if choices is None else choices
for item in cfg:
if not isinstance(cfg[item], list) and not isinstance(cfg[item], dict):
help_description = helper[item] if item in helper else "Please reference to {}".format(cfg_path)
choice = choices[item] if item in choices else None
if isinstance(cfg[item], bool):
parser.add_argument("--" + item, type=ast.literal_eval, default=cfg[item], choices=choice,
help=help_description)
else:
parser.add_argument("--" + item, type=type(cfg[item]), default=cfg[item], choices=choice,
help=help_description)
args = parser.parse_args()
return args
def parse_yaml(yaml_path):
"""
Parse the yaml config file.
Args:
yaml_path: Path to the yaml config.
"""
with open(yaml_path, 'r') as fin:
try:
cfgs = yaml.load_all(fin.read(), Loader=yaml.FullLoader)
cfgs = [x for x in cfgs]
if len(cfgs) == 1:
cfg_helper = {}
cfg = cfgs[0]
cfg_choices = {}
elif len(cfgs) == 2:
cfg, cfg_helper = cfgs
cfg_choices = {}
elif len(cfgs) == 3:
cfg, cfg_helper, cfg_choices = cfgs
else:
raise ValueError("At most 3 docs (config, description for help, choices) are supported in config yaml")
print(cfg_helper)
except:
raise ValueError("Failed to parse yaml")
return cfg, cfg_helper, cfg_choices
def merge(args, cfg):
"""
Merge the base config from yaml file and command line arguments.
Args:
args: Command line arguments.
cfg: Base configuration.
"""
args_var = vars(args)
for item in args_var:
cfg[item] = args_var[item]
return cfg
def get_config():
"""
Get Config according to the yaml file and cli arguments.
"""
parser = argparse.ArgumentParser(description="default name", add_help=False)
current_dir = os.path.dirname(os.path.abspath(__file__))
parser.add_argument("--config_path", type=str, default=os.path.join(current_dir, "../default_config.yaml"),
help="Config file path")
path_args, _ = parser.parse_known_args()
default, helper, choices = parse_yaml(path_args.config_path)
pprint(default)
args = parse_cli_to_yaml(parser=parser, cfg=default, helper=helper, choices=choices, cfg_path=path_args.config_path)
final_config = merge(args, default)
return Config(final_config)
config = get_config()

27
model_zoo/official/cv/deeplabv3/utils/device_adapter.py Normal file
View File

@ -0,0 +1,27 @@
# Copyright 2021 Huawei Technologies Co., Ltd
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ============================================================================
"""Device adapter for ModelArts"""
from .config import config
if config.enable_modelarts:
from .moxing_adapter import get_device_id, get_device_num, get_rank_id, get_job_id
else:
from .local_adapter import get_device_id, get_device_num, get_rank_id, get_job_id
__all__ = [
"get_device_id", "get_device_num", "get_rank_id", "get_job_id"
]

36
model_zoo/official/cv/deeplabv3/utils/local_adapter.py Normal file
View File

@ -0,0 +1,36 @@
# Copyright 2021 Huawei Technologies Co., Ltd
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ============================================================================
"""Local adapter"""
import os
def get_device_id():
device_id = os.getenv('DEVICE_ID', '0')
return int(device_id)
def get_device_num():
device_num = os.getenv('RANK_SIZE', '1')
return int(device_num)
def get_rank_id():
global_rank_id = os.getenv('RANK_ID', '0')
return int(global_rank_id)
def get_job_id():
return "Local Job"

116
model_zoo/official/cv/deeplabv3/utils/moxing_adapter.py Normal file
View File

@ -0,0 +1,116 @@
# Copyright 2021 Huawei Technologies Co., Ltd
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ============================================================================
"""Moxing adapter for ModelArts"""
import os
import functools
from mindspore import context
from .config import config
_global_sync_count = 0
def get_device_id():
device_id = os.getenv('DEVICE_ID', '0')
return int(device_id)
def get_device_num():
device_num = os.getenv('RANK_SIZE', '1')
return int(device_num)
def get_rank_id():
global_rank_id = os.getenv('RANK_ID', '0')
return int(global_rank_id)
def get_job_id():
job_id = os.getenv('JOB_ID')
job_id = job_id if job_id != "" else "default"
return job_id
def sync_data(from_path, to_path):
"""
Download data from remote obs to local directory if the first url is remote url and the second one is local path
Upload data from local directory to remote obs in contrast.
"""
import moxing as mox
import time
global _global_sync_count
sync_lock = "/tmp/copy_sync.lock" + str(_global_sync_count)
_global_sync_count += 1
# Each server contains 8 devices as most.
if get_device_id() % min(get_device_num(), 8) == 0 and not os.path.exists(sync_lock):
print("from path: ", from_path)
print("to path: ", to_path)
mox.file.copy_parallel(from_path, to_path)
print("===finish data synchronization===")
try:
os.mknod(sync_lock)
except IOError:
pass
print("===save flag===")
while True:
if os.path.exists(sync_lock):
break
time.sleep(1)
print("Finish sync data from {} to {}.".format(from_path, to_path))
def moxing_wrapper(pre_process=None, post_process=None):
"""
Moxing wrapper to download dataset and upload outputs.
"""
def wrapper(run_func):
@functools.wraps(run_func)
def wrapped_func(*args, **kwargs):
# Download data from data_url
if config.enable_modelarts:
if config.data_url:
sync_data(config.data_url, config.data_path)
print("Dataset downloaded: ", os.listdir(config.data_path))
if config.checkpoint_url:
sync_data(config.checkpoint_url, config.load_path)
print("Preload downloaded: ", os.listdir(config.load_path))
if config.train_url:
sync_data(config.train_url, config.output_path)
print("Workspace downloaded: ", os.listdir(config.output_path))
context.set_context(save_graphs_path=os.path.join(config.output_path, str(get_rank_id())))
config.device_num = get_device_num()
config.device_id = get_device_id()
if not os.path.exists(config.output_path):
os.makedirs(config.output_path)
if pre_process:
pre_process()
# Run the main function
run_func(*args, **kwargs)
# Upload data to train_url
if config.enable_modelarts:
if post_process:
post_process()
if config.train_url:
print("Start to copy output directory")
sync_data(config.output_path, config.train_url)
return wrapped_func
return wrapper