From 70bb1e9d89c7af3a8d45a6d992f2ecd2d6519aa3 Mon Sep 17 00:00:00 2001
From: maijianqiang <maijianqiang1@huawei.com>
Date: Thu, 16 Sep 2021 16:49:00 +0800
Subject: [PATCH] add newtwork run demo
---
official/cv/FCN8s/README.md | 2 +-
official/cv/MCNN/README.md | 12 ++---
official/cv/MCNN/requirements.txt | 3 ++
official/cv/alexnet/README.md | 2 +-
official/cv/alexnet/README_CN.md | 2 +-
official/cv/cnn_direction_model/README.md | 2 +-
official/cv/cnnctc/README.md | 4 +-
official/cv/cnnctc/README_CN.md | 4 +-
official/cv/ctpn/README.md | 5 +-
official/cv/ctpn/scripts/run_infer_310.sh | 2 +-
official/cv/deeplabv3/README.md | 12 +++--
official/cv/deeplabv3/README_CN.md | 9 ++--
official/cv/deeplabv3plus/README_CN.md | 9 ++--
official/cv/densenet/README.md | 13 +++---
official/cv/densenet/README_CN.md | 6 +--
official/cv/dpn/README.md | 2 +-
official/cv/east/README.md | 4 +-
official/cv/googlenet/README.md | 4 +-
official/cv/googlenet/README_CN.md | 4 +-
official/cv/inceptionv3/README.md | 2 +-
official/cv/inceptionv3/README_CN.md | 4 +-
official/cv/inceptionv4/README.md | 4 +-
official/cv/maskrcnn_mobilenetv1/README.md | 6 +--
official/cv/mobilenetv1/README.md | 4 +-
official/cv/mobilenetv2/README.md | 6 +--
official/cv/mobilenetv2/README_CN.md | 6 +--
official/cv/mobilenetv2_quant/README_CN.md | 16 +++----
official/cv/mobilenetv2_quant/Readme.md | 14 +++---
official/cv/openpose/README.md | 30 ++++++------
official/cv/retinanet/README_CN.md | 2 +-
official/cv/shufflenetv1/README_CN.md | 19 ++++----
official/cv/simclr/README.md | 12 +++--
official/cv/simple_pose/README.md | 54 ++++++++++++----------
official/cv/squeezenet/README.md | 48 +++++++++++--------
34 files changed, 179 insertions(+), 149 deletions(-)
create mode 100644 official/cv/MCNN/requirements.txt
diff --git a/official/cv/FCN8s/README.md b/official/cv/FCN8s/README.md
index 7abe6d9cd..741d6ef26 100644
--- a/official/cv/FCN8s/README.md
+++ b/official/cv/FCN8s/README.md
@@ -225,7 +225,7 @@ ckpt_file: /home/FCN8s/ckpt/FCN8s_1-133_300.ckpt
#Ascend鍏崱骞惰璁粌
bash scripts/run_train.sh [DEVICE_NUM] rank_table.json
- # example: bash scripts/run_train.sh 8 /home/hccl_8p_01234567_10.155.170.71.json
+ # example: bash scripts/run_train.sh 8 ~/hccl_8p.json
```
鍒嗗竷寮忚缁冮渶瑕佹彁鍓嶅垱寤篔SON鏍煎紡鐨凥CCL閰嶇疆鏂囦欢,璇烽伒寰猍閾炬帴璇存槑](https://gitee.com/mindspore/mindspore/tree/master/model_zoo/utils/hccl_tools)
diff --git a/official/cv/MCNN/README.md b/official/cv/MCNN/README.md
index f7e7d94c0..de835793a 100644
--- a/official/cv/MCNN/README.md
+++ b/official/cv/MCNN/README.md
@@ -71,9 +71,9 @@ After installing MindSpore via the official website, you can start training and
```bash
# enter script dir, train MCNN example
-sh run_standalone_train_ascend.sh 0 ./formatted_trainval/shanghaitech_part_A_patches_9/train ./formatted_trainval/shanghaitech_part_A_patches_9/train_den ./formatted_trainval/shanghaitech_part_A_patches_9/val ./formatted_trainval/shanghaitech_part_A_patches_9/val_den ./ckpt
+bash run_standalone_train_ascend.sh 0 ./formatted_trainval/shanghaitech_part_A_patches_9/train ./formatted_trainval/shanghaitech_part_A_patches_9/train_den ./formatted_trainval/shanghaitech_part_A_patches_9/val ./formatted_trainval/shanghaitech_part_A_patches_9/val_den ./ckpt
# enter script dir, evaluate MCNN example
-sh run_standalone_eval_ascend.sh 0 ./original/shanghaitech/part_A_final/test_data/images ./original/shanghaitech/part_A_final/test_data/ground_truth_csv ./train/ckpt/best.ckpt
+bash run_standalone_eval_ascend.sh 0 ./original/shanghaitech/part_A_final/test_data/images ./original/shanghaitech/part_A_final/test_data/ground_truth_csv ./train/ckpt/best.ckpt
```
# [Script Description](#contents)
@@ -126,14 +126,14 @@ Major parameters in train.py and config.py as follows:
```bash
# enter script dir, and run the distribute script
- sh run_distribute_train.sh ./hccl_table.json ./formatted_trainval/shanghaitech_part_A_patches_9/train ./formatted_trainval/shanghaitech_part_A_patches_9/train_den ./formatted_trainval/shanghaitech_part_A_patches_9/val ./formatted_trainval/shanghaitech_part_A_patches_9/val_den ./ckpt
+ bash run_distribute_train.sh ~/hccl_8p.json ./formatted_trainval/shanghaitech_part_A_patches_9/train ./formatted_trainval/shanghaitech_part_A_patches_9/train_den ./formatted_trainval/shanghaitech_part_A_patches_9/val ./formatted_trainval/shanghaitech_part_A_patches_9/val_den ./ckpt
# enter script dir, and run the standalone script
- sh run_standalone_train_ascend.sh 0 ./formatted_trainval/shanghaitech_part_A_patches_9/train ./formatted_trainval/shanghaitech_part_A_patches_9/train_den ./formatted_trainval/shanghaitech_part_A_patches_9/val ./formatted_trainval/shanghaitech_part_A_patches_9/val_den ./ckpt
+ bash run_standalone_train_ascend.sh 0 ./formatted_trainval/shanghaitech_part_A_patches_9/train ./formatted_trainval/shanghaitech_part_A_patches_9/train_den ./formatted_trainval/shanghaitech_part_A_patches_9/val ./formatted_trainval/shanghaitech_part_A_patches_9/val_den ./ckpt
```
After training, the loss value will be achieved as follows:
- ```text
+ ```log
# grep "loss is " log
epoch: 1 step: 305, loss is 0.00041025918
epoch: 2 step: 305, loss is 3.7117527e-05
@@ -161,7 +161,7 @@ Before running the command below, please check the checkpoint path used for eval
You can view the results through the file "eval_log". The accuracy of the test dataset will be as follows:
- ```text
+ ```log
# grep "MAE: " eval_log
MAE: 105.87984801910736 MSE: 161.6687899899305
```
diff --git a/official/cv/MCNN/requirements.txt b/official/cv/MCNN/requirements.txt
new file mode 100644
index 000000000..4d2786989
--- /dev/null
+++ b/official/cv/MCNN/requirements.txt
@@ -0,0 +1,3 @@
+numpy
+pandas
+opencv
\ No newline at end of file
diff --git a/official/cv/alexnet/README.md b/official/cv/alexnet/README.md
index 1ab7d4b7f..6e22c95c4 100644
--- a/official/cv/alexnet/README.md
+++ b/official/cv/alexnet/README.md
@@ -46,7 +46,7 @@ Dataset used: [CIFAR-10](<http://www.cs.toronto.edu/~kriz/cifar.html>)
- Note锛欴ata will be processed in dataset.py
- Download the dataset, the directory structure is as follows:
-```bash
+```cifar10
鈹溾攢cifar-10-batches-bin
鈹�
鈹斺攢cifar-10-verify-bin
diff --git a/official/cv/alexnet/README_CN.md b/official/cv/alexnet/README_CN.md
index 3ab37b04f..5b9ad7a73 100644
--- a/official/cv/alexnet/README_CN.md
+++ b/official/cv/alexnet/README_CN.md
@@ -48,7 +48,7 @@ AlexNet鐢�5涓嵎绉眰鍜�3涓叏杩炴帴灞傜粍鎴愩€傚涓嵎绉牳鐢ㄤ簬鎻愬彇
- 娉ㄦ剰锛氭暟鎹湪dataset.py涓鐞嗐€�
- 涓嬭浇鏁版嵁闆嗐€傜洰褰曠粨鏋勫涓嬶細
-```bash
+```cifar10
鈹溾攢cifar-10-batches-bin
鈹�
鈹斺攢cifar-10-verify-bin
diff --git a/official/cv/cnn_direction_model/README.md b/official/cv/cnn_direction_model/README.md
index 338527e7b..aa2241558 100644
--- a/official/cv/cnn_direction_model/README.md
+++ b/official/cv/cnn_direction_model/README.md
@@ -179,7 +179,7 @@ bash scripts/run_distribute_train_ascend.sh [rank_table] [train_dataset_path] [P
For example, you can run the shell command below to launch the training procedure.
```shell
-bash run_distribute_train_ascend.sh /home/hccl_8p_01234567_10.155.170.71.json /home/DataSet/FSNS/train/
+bash run_distribute_train_ascend.sh ~/hccl_8p.json /home/DataSet/FSNS/train/
```
- running on ModelArts
diff --git a/official/cv/cnnctc/README.md b/official/cv/cnnctc/README.md
index 74d354f95..80671a1a1 100644
--- a/official/cv/cnnctc/README.md
+++ b/official/cv/cnnctc/README.md
@@ -147,7 +147,7 @@ bash scripts/run_standalone_train_gpu.sh $PRETRAINED_CKPT(options)
```bash
bash scripts/run_distribute_train_ascend.sh $RANK_TABLE_FILE $PRETRAINED_CKPT(options)
-# example: bash scripts/run_distribute_train_ascend.sh /home/hccl_8p_01234567_10.155.170.71.json
+# example: bash scripts/run_distribute_train_ascend.sh ~/hccl_8p.json
```
- Distributed GPU Training:
@@ -254,7 +254,7 @@ Results and checkpoints are written to `./train` folder. Log can be found in `./
```bash
bash scripts/run_distribute_train_ascend.sh [RANK_TABLE_FILE] [PRETRAINED_CKPT(options)]
-# example: bash scripts/run_distribute_train_ascend.sh /home/hccl_8p_01234567_10.155.170.71.json
+# example: bash scripts/run_distribute_train_ascend.sh ~/hccl_8p.json
```
For distributed training, a hccl configuration file with JSON format needs to be created in advance.
diff --git a/official/cv/cnnctc/README_CN.md b/official/cv/cnnctc/README_CN.md
index 9f081d577..a40fd1c09 100644
--- a/official/cv/cnnctc/README_CN.md
+++ b/official/cv/cnnctc/README_CN.md
@@ -144,7 +144,7 @@ bash scripts/run_standalone_train_ascend.sh [DEVICE_ID] [PRETRAINED_CKPT(options
```shell
bash scripts/run_distribute_train_ascend.sh [RANK_TABLE_FILE] [PRETRAINED_CKPT(options)]
-# example: bash scripts/run_distribute_train_ascend.sh /home/hccl_8p_01234567_10.155.170.71.json
+# example: bash scripts/run_distribute_train_ascend.sh ~/hccl_8p.json
```
- 璇勪及锛�
@@ -239,7 +239,7 @@ bash scripts/run_standalone_train_ascend.sh [DEVICE_ID] [PRETRAINED_CKPT(options
```shell
bash scripts/run_distribute_train_ascend.sh [RANK_TABLE_FILE] [PRETRAINED_CKPT(options)]
-# example: bash scripts/run_distribute_train_ascend.sh /home/hccl_8p_01234567_10.155.170.71.json
+# example: bash scripts/run_distribute_train_ascend.sh ~/hccl_8p.json
```
缁撴灉鍜屾鏌ョ偣鍒嗗埆鍐欏叆璁惧`i`鐨刞./train_parallel_{i}`鏂囦欢澶广€�
diff --git a/official/cv/ctpn/README.md b/official/cv/ctpn/README.md
index 13f5d2a06..26b30d481 100644
--- a/official/cv/ctpn/README.md
+++ b/official/cv/ctpn/README.md
@@ -171,8 +171,7 @@ Modify the parameters according to the actual path
```bash
# distribute training
bash scripts/run_distribute_train_ascend.sh [RANK_TABLE_FILE] [TASK_TYPE] [PRETRAINED_PATH]
-# example: bash scripts/run_distribute_train_ascend.sh /home/hccl_8p_01234567_10.155.170.71.json Pretraining(or Finetune) \
-# /home/DataSet/ctpn_dataset/backbone/0-150_5004.ckpt
+# example: bash scripts/run_distribute_train_ascend.sh ~/hccl_8p.json Pretraining(or Finetune) /home/DataSet/ctpn_dataset/backbone/0-150_5004.ckpt
# standalone training
bash scrpits/run_standalone_train_ascend.sh [TASK_TYPE] [PRETRAINED_PATH] [DEVICE_ID]
@@ -226,7 +225,7 @@ ICDAR2013, SCUT-FORU to improve precision and recall, and when doing Finetune, w
Ascend:
# distribute training example(8p)
bash run_distribute_train_ascend.sh [RANK_TABLE_FILE] [TASK_TYPE] [PRETRAINED_PATH]
- # example: bash scripts/run_distribute_train_ascend.sh /home/hccl_8p_01234567_10.155.170.71.json Pretraining(or Finetune) /home/DataSet/ctpn_dataset/backbone/0-150_5004.ckpt
+ # example: bash scripts/run_distribute_train_ascend.sh ~/hccl_8p.json Pretraining(or Finetune) /home/DataSet/ctpn_dataset/backbone/0-150_5004.ckpt
# standalone training
bash run_standalone_train_ascend.sh [TASK_TYPE] [PRETRAINED_PATH]
diff --git a/official/cv/ctpn/scripts/run_infer_310.sh b/official/cv/ctpn/scripts/run_infer_310.sh
index f7117631b..4f8fcbfe5 100644
--- a/official/cv/ctpn/scripts/run_infer_310.sh
+++ b/official/cv/ctpn/scripts/run_infer_310.sh
@@ -104,7 +104,7 @@ function cal_acc()
fi
mkdir output
mkdir output_img
- python ../postprocess.py --dataset_path=$data_path --result_path=result_Files --label_path=$label_path &> acc.log
+ python ../postprocess.py --export_dataset_path=$data_path --result_path=result_Files --label_path=$label_path &> acc.log
if [ $? -ne 0 ]; then
echo "calculate accuracy failed"
exit 1
diff --git a/official/cv/deeplabv3/README.md b/official/cv/deeplabv3/README.md
index 5a3edb0e7..0d0688422 100644
--- a/official/cv/deeplabv3/README.md
+++ b/official/cv/deeplabv3/README.md
@@ -113,7 +113,7 @@ After installing MindSpore via the official website, you can start training and
- Prepare backbone
-Download resnet101 for here(https://download.mindspore.cn/model_zoo/r1.2/resnet101_ascend_v120_imagenet2012_official_cv_bs32_acc78/resnet101_ascend_v120_imagenet2012_official_cv_bs32_acc78.ckpt).
+Download resnet101 for [here](https://download.mindspore.cn/model_zoo/r1.2/resnet101_ascend_v120_imagenet2012_official_cv_bs32_acc78/resnet101_ascend_v120_imagenet2012_official_cv_bs32_acc78.ckpt).
- Running on Ascend
@@ -140,7 +140,7 @@ Enter the shell script to modify the data_file and ckpt_pre_trained parameters
data_file=/home/DataSet/VOC2012/vocaug_mindrecords/vocaug.mindrecord0
ckpt_pre_trained=/home/model/deeplabv3/predtrained/resnet101_ascend_v120_imagenet2012_official_cv_bs32_acc78.ckpt
-bash run_distribute_train_s16_r1.sh
+bash run_distribute_train_s16_r1.sh ~/hccl_8p.json
```
2. Train s8 with vocaug dataset, finetuning from model in previous step, training script is:
@@ -151,7 +151,7 @@ Enter the shell script to modify the data_file and ckpt_pre_trained parameters
data_file=/home/DataSet/VOC2012/vocaug_mindrecords/vocaug.mindrecord0
ckpt_pre_trained=/home/model/deeplabv3/predtrained/resnet101_ascend_v120_imagenet2012_official_cv_bs32_acc78.ckpt
-bash run_distribute_train_s8_r1.sh
+bash run_distribute_train_s8_r1.sh ~/hccl_8p.json
```
3. Train s8 with voctrain dataset, finetuning from model in previous step, training script is:
@@ -164,17 +164,19 @@ data_file=/home/DataSet/VOC2012/voctrain_mindrecords/votrain.mindrecord0
ckpt_pre_trained=/home/model/deeplabv3/ckpt/deeplabv3-800_330.ckpt
-bash run_distribute_train_s8_r2.sh
+bash run_distribute_train_s8_r2.sh ~/hccl_8p.json
```
- For evaluation, evaluating steps are as follows:
1. Enter the shell script to modify the data_file and ckpt_pre_trained parameters
-```default_config.yaml
+```shell
+modify the parameter according local path
# example:
data_root=/home/DataSet/VOC2012
data_lst=/home/DataSet/VOC2012/voc_val_lst.txt
+ckpt_path=/home/model/deeplabv3/ckpt/deeplabv3-800_330.ckpt
```
2. Eval s16 with voc val dataset, eval script is:
diff --git a/official/cv/deeplabv3/README_CN.md b/official/cv/deeplabv3/README_CN.md
index da2f68ade..51ef16a69 100644
--- a/official/cv/deeplabv3/README_CN.md
+++ b/official/cv/deeplabv3/README_CN.md
@@ -144,7 +144,7 @@ bash run_standalone_train.sh
data_file=/home/DataSet/VOC2012/vocaug_mindrecords/vocaug.mindrecord0
ckpt_pre_trained=/home/model/deeplabv3/predtrained/resnet101_ascend_v120_imagenet2012_official_cv_bs32_acc78.ckpt
-bash run_distribute_train_s16_r1.sh
+bash run_distribute_train_s16_r1.sh ~/hccl_8p.json
```
2. 浣跨敤VOCaug鏁版嵁闆嗚缁僺8锛屽井璋冧笂涓€姝ョ殑妯″瀷銆傝剼鏈涓嬶細
@@ -155,7 +155,7 @@ bash run_distribute_train_s16_r1.sh
data_file=/home/DataSet/VOC2012/vocaug_mindrecords/vocaug.mindrecord0
ckpt_pre_trained=/home/model/deeplabv3/predtrained/resnet101_ascend_v120_imagenet2012_official_cv_bs32_acc78.ckpt
-bash run_distribute_train_s8_r1.sh
+bash run_distribute_train_s8_r1.sh ~/hccl_8p.json
```
3. 浣跨敤VOCtrain鏁版嵁闆嗚缁僺8锛屽井璋冧笂涓€姝ョ殑妯″瀷銆傝剼鏈涓嬶細
@@ -167,17 +167,18 @@ bash run_distribute_train_s8_r1.sh
data_file=/home/DataSet/VOC2012/voctrain_mindrecords/votrain.mindrecord0
ckpt_pre_trained=/home/model/deeplabv3/ckpt/deeplabv3-800_330.ckpt
-bash run_distribute_train_s8_r2.sh
+bash run_distribute_train_s8_r2.sh ~/hccl_8p.json
```
- 璇勪及姝ラ濡備笅锛�
1. 杩涘叆瀵瑰簲鐨剆hell鑴氭湰淇敼鍙傛暟
-```default_config.yaml
+```shell
# example:
data_root=/home/DataSet/VOC2012
data_lst=/home/DataSet/VOC2012/voc_val_lst.txt
+ckpt_path=/home/model/deeplabv3/ckpt/deeplabv3-800_330.ckpt
```
2. 浣跨敤voc val鏁版嵁闆嗚瘎浼皊16銆傝瘎浼拌剼鏈涓嬶細
diff --git a/official/cv/deeplabv3plus/README_CN.md b/official/cv/deeplabv3plus/README_CN.md
index 1ed6221bd..755dc7282 100644
--- a/official/cv/deeplabv3plus/README_CN.md
+++ b/official/cv/deeplabv3plus/README_CN.md
@@ -134,7 +134,7 @@ bash run_alone_train.sh
data_file=/home/DataSet/VOC2012/vocaug_mindrecords/vocaug.mindrecord0
ckpt_pre_trained=/home/model/deeplabv3/predtrained/resnet101_ascend_v120_imagenet2012_official_cv_bs32_acc78.ckpt
-bash run_distribute_train_s16_r1.sh
+bash run_distribute_train_s16_r1.sh ~/hccl_8p.json
```
2.浣跨敤VOCaug鏁版嵁闆嗚缁僺8锛屽井璋冧笂涓€姝ョ殑妯″瀷銆傝剼鏈涓嬶細
@@ -145,7 +145,7 @@ bash run_distribute_train_s16_r1.sh
data_file=/home/DataSet/VOC2012/vocaug_mindrecords/vocaug.mindrecord0
ckpt_pre_trained=/home/model/deeplabv3/predtrained/resnet101_ascend_v120_imagenet2012_official_cv_bs32_acc78.ckpt
-bash run_distribute_train_s8_r1.sh
+bash run_distribute_train_s8_r1.sh ~/hccl_8p.json
```
3.浣跨敤VOCtrain鏁版嵁闆嗚缁僺8锛屽井璋冧笂涓€姝ョ殑妯″瀷銆傝剼鏈涓嬶細
@@ -157,17 +157,18 @@ bash run_distribute_train_s8_r1.sh
data_file=/home/DataSet/VOC2012/voctrain_mindrecords/votrain.mindrecord0
ckpt_pre_trained=/home/model/deeplabv3/ckpt/deeplabv3-800_330.ckpt
-run_distribute_train_s8_r2.sh
+bash run_distribute_train_s8_r2.sh ~/hccl_8p.json
```
璇勪及姝ラ濡備笅锛�
1. 杩涘叆瀵瑰簲鐨剆hell鑴氭湰淇敼鍙傛暟
-```default_cofig.yaml
+```shell
# example:
data_root=/home/DataSet/VOC2012
data_lst=/home/DataSet/VOC2012/voc_val_lst.txt
+ckpt_path=/home/model/deeplabv3/ckpt/deeplabv3-800_330.ckpt
```
2.浣跨敤voc val鏁版嵁闆嗚瘎浼皊16銆傝瘎浼拌剼鏈涓嬶細
diff --git a/official/cv/densenet/README.md b/official/cv/densenet/README.md
index cdea0a65b..4e5b359b4 100644
--- a/official/cv/densenet/README.md
+++ b/official/cv/densenet/README.md
@@ -102,13 +102,13 @@ After installing MindSpore via the official website, you can start training and
# run distributed training example
bash scripts/run_distribute_train.sh [DEVICE_NUM] [RANK_TABLE_FILE] [NET_NAME] [DATASET_NAME] [TRAIN_DATA_DIR]
- # example bash scripts/run_distribute_train.sh 8 /root/hccl_8p_01234567_10.155.170.71.json densenet121 imagenet /home/DataSet/ImageNet_Original/train/
+ # example bash scripts/run_distribute_train.sh 8 ~/hccl_8p.json densenet121 imagenet /home/DataSet/ImageNet_Original/train/
# run evaluation example
python eval.py --net [NET_NAME] --dataset [DATASET_NAME] --eval_data_dir /PATH/TO/DATASET --ckpt_files /PATH/TO/CHECKPOINT > eval.log 2>&1 &
OR
bash scripts/run_distribute_eval.sh [DEVICE_NUM] [RANDK_TABLE_FILE] [NET_NAME] [DATASET_NAME] [EVAL_DATA_DIR][CKPT_PATH]
- # example: bash script/run_distribute_eval.sh 8 /root/hccl_8p_01234567_10.155.170.71.json densenet121 imagenet /home/DataSet/ImageNet_Original/train/validation_preprocess/ /home/model/densenet/ckpt/0-120_500.ckpt
+ # example: bash script/run_distribute_eval.sh 8 ~/hccl_8p.json densenet121 imagenet /home/DataSet/ImageNet_Original/train/validation_preprocess/ /home/model/densenet/ckpt/0-120_500.ckpt
```
For distributed training, a hccl configuration file with JSON format needs to be created in advance.
@@ -293,7 +293,7 @@ You can modify the training behaviour through the various flags in the `densenet
2020-08-22 17:02:19,921:INFO:local passed
2020-08-22 17:05:43,112:INFO:epoch[2], iter[15011], loss:3.096, mean_fps:6304.53 imgs/sec
2020-08-22 17:05:43,113:INFO:local passed
-
+ ...
```
- running on GPU
@@ -326,7 +326,7 @@ You can modify the training behaviour through the various flags in the `densenet
```bash
bash scripts/run_distribute_train.sh [DEVICE_NUM] [RANK_TABLE_FILE] [NET_NAME] [DATASET_NAME] [TRAIN_DATA_DIR]
- # example bash scripts/run_distribute_train.sh 8 /root/hccl_8p_01234567_10.155.170.71.json densenet121 imagenet /home/DataSet/ImageNet_Original/train/
+ # example bash scripts/run_distribute_train.sh 8 ~/hccl_8p.json densenet121 imagenet /home/DataSet/ImageNet_Original/train/
```
@@ -340,7 +340,8 @@ You can modify the training behaviour through the various flags in the `densenet
2020-08-22 17:09:05,686:INFO:epoch[3], iter[20015], loss:3.113, mean_fps:6304.37 imgs/sec
2020-08-22 17:12:28,925:INFO:epoch[4], iter[25019], loss:3.29, mean_fps:6303.07 imgs/sec
2020-08-22 17:15:52,167:INFO:epoch[5], iter[30023], loss:2.865, mean_fps:6302.98 imgs/sec
-
+ ...
+ ...
```
- running on GPU
@@ -367,7 +368,7 @@ You can modify the training behaviour through the various flags in the `densenet
python eval.py --net [NET_NAME] --dataset [DATASET_NAME] --eval_data_dir /PATH/TO/DATASET --ckpt_files /PATH/TO/CHECKPOINT > eval.log 2>&1 &
OR
bash scripts/run_distribute_eval.sh [DEVICE_NUM] [RANDK_TABLE_FILE] [NET_NAME] [DATASET_NAME] [EVAL_DATA_DIR][CKPT_PATH]
- # example: bash script/run_distribute_eval.sh 8 /root/hccl_8p_01234567_10.155.170.71.json densenet121 imagenet /home/DataSet/ImageNet_Original/train/validation_preprocess/ /home/model/densenet/ckpt/0-120_500.ckpt
+ # example: bash script/run_distribute_eval.sh 8 ~/hccl_8p.json densenet121 imagenet /home/DataSet/ImageNet_Original/train/validation_preprocess/ /home/model/densenet/ckpt/0-120_500.ckpt
```
diff --git a/official/cv/densenet/README_CN.md b/official/cv/densenet/README_CN.md
index 31191bd80..3301ea23f 100644
--- a/official/cv/densenet/README_CN.md
+++ b/official/cv/densenet/README_CN.md
@@ -107,13 +107,13 @@ DenseNet-100浣跨敤鐨勬暟鎹泦锛� Cifar-10
# 鍒嗗竷寮忚缁冪ず渚�
bash scripts/run_distribute_train.sh [DEVICE_NUM] [RANK_TABLE_FILE] [NET_NAME] [DATASET_NAME] [TRAIN_DATA_DIR]
- # example bash scripts/run_distribute_train.sh 8 /root/hccl_8p_01234567_10.155.170.71.json densenet121 imagenet /home/DataSet/ImageNet_Original/train/
+ # example bash scripts/run_distribute_train.sh 8 ~/hccl_8p.json densenet121 imagenet /home/DataSet/ImageNet_Original/train/
# 鍗曞崱璇勪及绀轰緥
python eval.py --net [NET_NAME] --dataset [DATASET_NAME] --eval_data_dir /PATH/TO/DATASET --ckpt_files /PATH/TO/CHECKPOINT > eval.log 2>&1 &
bash scripts/run_distribute_eval.sh [DEVICE_NUM] [RANDK_TABLE_FILE] [NET_NAME] [DATASET_NAME] [EVAL_DATA_DIR][CKPT_PATH]
- # example: bash script/run_distribute_eval.sh 8 /root/hccl_8p_01234567_10.155.170.71.json densenet121 imagenet /home/DataSet/ImageNet_Original/train/validation_preprocess/ /home/model/densenet/ckpt/0-120_500.ckpt
+ # example: bash script/run_distribute_eval.sh 8 ~/hccl_8p.json densenet121 imagenet /home/DataSet/ImageNet_Original/train/validation_preprocess/ /home/model/densenet/ckpt/0-120_500.ckpt
```
鍒嗗竷寮忚缁冮渶瑕佹彁鍓嶅垱寤篔SON鏍煎紡鐨凥CCL閰嶇疆鏂囦欢銆�
@@ -310,7 +310,7 @@ python train.py --net=[NET_NAME] --dataset=[DATASET_NAME] --train_data_dir=[DATA
```shell
bash scripts/run_distribute_train.sh [DEVICE_NUM] [RANK_TABLE_FILE] [NET_NAME] [DATASET_NAME] [TRAIN_DATA_DIR]
-# example bash scripts/run_distribute_train.sh 8 /root/hccl_8p_01234567_10.155.170.71.json densenet121 imagenet /home/DataSet/ImageNet_Original/train/
+# example bash scripts/run_distribute_train.sh 8 ~/hccl_8p.json densenet121 imagenet /home/DataSet/ImageNet_Original/train/
```
涓婅堪shell鑴氭湰灏嗗湪鍚庡彴杩涜鍒嗗竷寮忚缁冦€傚彲浠ラ€氳繃鏂囦欢`train[X]/output/202x-xx-xx_time_xx_xx_xx/`鏌ョ湅缁撴灉鏃ュ織鍜屾ā鍨嬫鏌ョ偣銆傚湪ImageNet鏁版嵁闆嗕笂璁粌DenseNet-121鐨勬崯澶卞€肩殑瀹炵幇濡備笅锛�
diff --git a/official/cv/dpn/README.md b/official/cv/dpn/README.md
index a1099021c..71c98709a 100644
--- a/official/cv/dpn/README.md
+++ b/official/cv/dpn/README.md
@@ -228,7 +228,7 @@ Run `scripts/train_distributed.sh` to train the model distributed. The usage of
```text
bash scripts/train_distributed.sh [rank_table] [train_data_dir] [ckpt_path_to_save] [rank_size] [eval_each_epoch] [pretrained_ckpt(optional)]
-# example: bash scripts/train_distributed.sh /root/hccl_8p_01234567_10.155.170.71.json /home/DataSet/ImageNet_Original/train/ ./ckpt/ 8 0
+# example: bash scripts/train_distributed.sh ~/hccl_8p.json /home/DataSet/ImageNet_Original/train/ ./ckpt/ 8 0
```
The above shell script will run distribute training in the background. You can view the results through the file `train_parallel[X]/log.txt` as follows:
diff --git a/official/cv/east/README.md b/official/cv/east/README.md
index 2354fbe1d..7d7c2afb6 100644
--- a/official/cv/east/README.md
+++ b/official/cv/east/README.md
@@ -81,7 +81,7 @@ Dataset used [ICDAR 2015](https://rrc.cvc.uab.es/?ch=4&com=downloads)
```bash
# distribute training example(8p)
bash run_distribute_train.sh [DATASET_PATH] [PRETRAINED_BACKBONE] [RANK_TABLE_FILE]
-# example: bash run_distribute_train.sh /home/DataSet/ICDAR2015/ic15/ home/model/east/pretrained/0-150_5004.ckpt /root/hccl_8p_01234567_10.155.170.71.json
+# example: bash run_distribute_train.sh /home/DataSet/ICDAR2015/ic15/ home/model/east/pretrained/0-150_5004.ckpt ~/hccl_8p.json
# standalone training
bash run_standalone_train_ascend.sh [DATASET_PATH] [PRETRAINED_BACKBONE] [DEVICE_ID]
@@ -106,7 +106,7 @@ bash run_eval_ascend.sh [DATASET_PATH] [CKPT_PATH] [DEVICE_ID]
Ascend:
# distribute training example(8p)
bash run_distribute_train.sh [DATASET_PATH] [PRETRAINED_BACKBONE] [RANK_TABLE_FILE]
- # example: bash run_distribute_train.sh /home/DataSet/ICDAR2015/ic15/ home/model/east/pretrained/0-150_5004.ckpt /root/hccl_8p_01234567_10.155.170.71.json
+ # example: bash run_distribute_train.sh /home/DataSet/ICDAR2015/ic15/ home/model/east/pretrained/0-150_5004.ckpt ~/hccl_8p.json
# standalone training
bash run_standalone_train_ascend.sh [DATASET_PATH] [PRETRAINED_BACKBONE] [DEVICE_ID]
diff --git a/official/cv/googlenet/README.md b/official/cv/googlenet/README.md
index dd5f145a3..ca8ce7db2 100644
--- a/official/cv/googlenet/README.md
+++ b/official/cv/googlenet/README.md
@@ -105,7 +105,7 @@ After installing MindSpore via the official website, you can start training and
# run distributed training example
bash scripts/run_train.sh [RANK_TABLE_FILE] [DATASET_NAME]
- # example: bash scripts/run_train.sh /root/hccl_8p_01234567_10.155.170.71.json cifar10
+ # example: bash scripts/run_train.sh ~/hccl_8p.json cifar10
# run evaluation example
python eval.py > eval.log 2>&1 &
@@ -401,7 +401,7 @@ For more configuration details, please refer the script `config.py`.
- running on Ascend
```bash
- bash scripts/run_train.sh /root/hccl_8p_01234567_10.155.170.71.json cifar10
+ bash scripts/run_train.sh ~/hccl_8p.json cifar10
```
The above shell script will run distribute training in the background. You can view the results through the file `train_parallel[X]/log`. The loss value will be achieved as follows:
diff --git a/official/cv/googlenet/README_CN.md b/official/cv/googlenet/README_CN.md
index 5a295ffcf..5cd3c07be 100644
--- a/official/cv/googlenet/README_CN.md
+++ b/official/cv/googlenet/README_CN.md
@@ -107,7 +107,7 @@ GoogleNet鐢卞涓猧nception妯″潡涓茶仈璧锋潵锛屽彲浠ユ洿鍔犳繁鍏ャ€� 闄嶇淮鐨�
# 杩愯鍒嗗竷寮忚缁冪ず渚�
bash scripts/run_train.sh [RANK_TABLE_FILE] [DATASET_NAME]
- # example: bash scripts/run_train.sh /root/hccl_8p_01234567_10.155.170.71.json cifar10
+ # example: bash scripts/run_train.sh ~/hccl_8p.json cifar10
# 杩愯璇勪及绀轰緥
python eval.py > eval.log 2>&1 &
@@ -371,7 +371,7 @@ GoogleNet鐢卞涓猧nception妯″潡涓茶仈璧锋潵锛屽彲浠ユ洿鍔犳繁鍏ャ€� 闄嶇淮鐨�
- Ascend澶勭悊鍣ㄧ幆澧冭繍琛�
```bash
- bash scripts/run_train.sh /root/hccl_8p_01234567_10.155.170.71.json cifar10
+ bash scripts/run_train.sh ~/hccl_8p.json cifar10
```
涓婅堪shell鑴氭湰灏嗗湪鍚庡彴杩愯鍒嗗竷璁粌銆傛偍鍙互閫氳繃train_parallel[X]/log鏂囦欢鏌ョ湅缁撴灉銆傞噰鐢ㄤ互涓嬫柟寮忚揪鍒版崯澶卞€硷細
diff --git a/official/cv/inceptionv3/README.md b/official/cv/inceptionv3/README.md
index 6f0e1a2bc..7629b561a 100644
--- a/official/cv/inceptionv3/README.md
+++ b/official/cv/inceptionv3/README.md
@@ -284,7 +284,7 @@ Take training cifar10 as an example, the ds_type parameter is set to cifar10
```shell
# distribute training(8p)
bash run_distribute_train.sh [RANK_TABLE_FILE] [DATA_PATH] [CKPT_PATH]
-# example: bash run_distribute_train.sh /root/hccl_8p_012345467_10.155.170.71.json /home/DataSet/cifar10/ ./ckpt/
+# example: bash run_distribute_train.sh ~/hccl_8p.json /home/DataSet/cifar10/ ./ckpt/
# standalone training
bash scripts/run_standalone_train.sh [DEVICE_ID] [DATA_PATH] [CKPT_PATH]
diff --git a/official/cv/inceptionv3/README_CN.md b/official/cv/inceptionv3/README_CN.md
index d8423127d..3e92fcfa4 100644
--- a/official/cv/inceptionv3/README_CN.md
+++ b/official/cv/inceptionv3/README_CN.md
@@ -290,7 +290,7 @@ ds_type:cifar10
```shell
# 鍒嗗竷寮忚缁冪ず渚�(8鍗�)
bash run_distribute_train.sh [RANK_TABLE_FILE] [DATA_PATH] [CKPT_PATH]
-# example: bash run_distribute_train.sh /root/hccl_8p_012345467_10.155.170.71.json /home/DataSet/cifar10/ ./ckpt/
+# example: bash run_distribute_train.sh ~/hccl_8p.json /home/DataSet/cifar10/ ./ckpt/
# 鍗曟満璁粌
bash scripts/run_standalone_train.sh [DEVICE_ID] [DATA_PATH] [CKPT_PATH]
@@ -312,7 +312,7 @@ bash scripts/run_standalone_train.sh [DEVICE_ID] [DATA_PATH] [CKPT_PATH]
Ascend:
# 鍒嗗竷寮忚缁冪ず渚�(8鍗�)
bash run_distribute_train.sh [RANK_TABLE_FILE] [DATA_PATH] [CKPT_PATH]
- # example: bash run_distribute_train.sh /root/hccl_8p_012345467_10.155.170.71.json /home/DataSet/cifar10/ ./ckpt/
+ # example: bash run_distribute_train.sh ~/hccl_8p.json /home/DataSet/cifar10/ ./ckpt/
# 鍗曟満璁粌
bash scripts/run_standalone_train.sh [DEVICE_ID] [DATA_PATH] [CKPT_PATH]
diff --git a/official/cv/inceptionv4/README.md b/official/cv/inceptionv4/README.md
index f49028f97..1748aedb9 100644
--- a/official/cv/inceptionv4/README.md
+++ b/official/cv/inceptionv4/README.md
@@ -255,7 +255,7 @@ Take training cifar10 as an example, the ds_type parameter is set to cifar10
```bash
# distribute training example(8p)
bash scripts/run_distribute_train_ascend.sh [RANK_TABLE_FILE] [DATA_DIR]
-# example: bash scripts/run_distribute_train_ascend.sh /root/hccl_8p_01234567_10.155.170.71.json /home/DataSet/cifar10/
+# example: bash scripts/run_distribute_train_ascend.sh ~/hccl_8p.json /home/DataSet/cifar10/
# standalone training
bash scripts/run_standalone_train_ascend.sh [DEVICE_ID] [DATA_DIR]
@@ -289,7 +289,7 @@ bash scripts/run_standalone_train_cpu.sh DATA_PATH
Ascend:
# distribute training example(8p)
bash scripts/run_distribute_train_ascend.sh [RANK_TABLE_FILE] [DATA_DIR]
- # example: bash scripts/run_distribute_train_ascend.sh /root/hccl_8p_01234567_10.155.170.71.json /home/DataSet/cifar10/
+ # example: bash scripts/run_distribute_train_ascend.sh ~/hccl_8p.json /home/DataSet/cifar10/
# standalone training
bash scripts/run_standalone_train_ascend.sh [DEVICE_ID] [DATA_DIR]
diff --git a/official/cv/maskrcnn_mobilenetv1/README.md b/official/cv/maskrcnn_mobilenetv1/README.md
index 51813f4e7..97b5ecaa8 100644
--- a/official/cv/maskrcnn_mobilenetv1/README.md
+++ b/official/cv/maskrcnn_mobilenetv1/README.md
@@ -107,7 +107,7 @@ pip install mmcv=0.2.14
# distributed training
bash run_distribute_train.sh [RANK_TABLE_FILE] [DATA_PATH] [PRETRAINED_CKPT(optional)]
- # example: bash run_distribute_train.sh /root/hccl_8p_01234567_10.155.170.71.json /home/DataSet/cocodataset/
+ # example: bash run_distribute_train.sh ~/hccl_8p.json /home/DataSet/cocodataset/
# standalone training
bash run_standalone_train.sh [DATA_PATH] [PRETRAINED_CKPT(optional)]
@@ -351,7 +351,7 @@ On Ascend:
# distributed training
Usage: bash run_distribute_train.sh [RANK_TABLE_FILE] [DATA_PATH] [PRETRAINED_CKPT(optional)]
-# example: bash run_distribute_train.sh /root/hccl_8p_01234567_10.155.170.71.json /home/DataSet/cocodataset/
+# example: bash run_distribute_train.sh ~/hccl_8p.json /home/DataSet/cocodataset/
# standalone training
Usage: bash run_standalone_train.sh [DATA_PATH] [PRETRAINED_CKPT(optional)]
@@ -532,7 +532,7 @@ bash run_standalone_train_cpu.sh [PRETRAINED_MODEL](optional)
```bash
bash run_distribute_train.sh [RANK_TABLE_FILE] [DATA_PATH] [PRETRAINED_MODEL(optional)]
-# example: bash run_distribute_train.sh /root/hccl_8p_01234567_10.155.170.71.json /home/DataSet/cocodataset/
+# example: bash run_distribute_train.sh ~/hccl_8p.json /home/DataSet/cocodataset/
```
> hccl.json which is specified by RANK_TABLE_FILE is needed when you are running a distribute task. You can generate it by using the [hccl_tools](https://gitee.com/mindspore/mindspore/tree/master/model_zoo/utils/hccl_tools).
diff --git a/official/cv/mobilenetv1/README.md b/official/cv/mobilenetv1/README.md
index 59a962199..068e4e50e 100644
--- a/official/cv/mobilenetv1/README.md
+++ b/official/cv/mobilenetv1/README.md
@@ -264,8 +264,8 @@ Please follow the instructions in the link [hccn_tools](https://gitee.com/mindsp
shell:
Ascend: bash run_distribute_train.sh [cifar10|imagenet2012] [RANK_TABLE_FILE] [DATASET_PATH] [PRETRAINED_CKPT_PATH](optional)
- # example: bash run_distribute_train.sh cifar10 /root/hccl_8p_01234567_10.155.170.71.json /home/DataSet/cifar10/cifar-10-batches-bin/
- # example: bash run_distribute_train.sh imagenet2012 /root/hccl_8p_01234567_10.155.170.71.json /home/DataSet/ImageNet_Original/
+ # example: bash run_distribute_train.sh cifar10 ~/hccl_8p.json /home/DataSet/cifar10/cifar-10-batches-bin/
+ # example: bash run_distribute_train.sh imagenet2012 ~/hccl_8p.json /home/DataSet/ImageNet_Original/
CPU: bash run_train_CPU.sh [cifar10|imagenet2012] [DATASET_PATH] [PRETRAINED_CKPT_PATH](optional)
GPU(single device): bash run_standalone_train_gpu.sh [cifar10|imagenet2012] [DATASET_PATH] [PRETRAINED_CKPT_PATH](optional)
diff --git a/official/cv/mobilenetv2/README.md b/official/cv/mobilenetv2/README.md
index 4deadc116..4f21014a3 100644
--- a/official/cv/mobilenetv2/README.md
+++ b/official/cv/mobilenetv2/README.md
@@ -275,7 +275,7 @@ You can start training using python or shell scripts. The usage of shell scripts
shell:
Ascend: bash run_train.sh Ascend [CONFIG_PATH] [DEVICE_NUM] [VISIABLE_DEVICES(0,1,2,3,4,5,6,7)] [RANK_TABLE_FILE] [DATASET_PATH]
- # example: bash run_train.sh Ascend default_config.yaml 8 0,1,2,3,4,5,6,7 /root/hccl_8p_01234567_10.155.170.71.json /home/DataSet/ImageNet_Original/
+ # example: bash run_train.sh Ascend default_config.yaml 8 0,1,2,3,4,5,6,7 ~/hccl_8p.json /home/DataSet/ImageNet_Original/
GPU: bash run_train.sh GPU 8 0,1,2,3,4,5,6,7 [TRAIN_DATASET_PATH]
CPU: bash run_train.sh CPU [TRAIN_DATASET_PATH]
@@ -288,7 +288,7 @@ You can start training using python or shell scripts. The usage of shell scripts
shell:
Ascend: bash run_train.sh Ascend [CONFIG_PATH] [DEVICE_NUM] [VISIABLE_DEVICES(0,1,2,3,4,5,6,7)] [RANK_TABLE_FILE] [DATASET_PATH] [CKPT_PATH] [FREEZE_LAYER] [FILTER_HEAD]
- # example: bash run_train.sh Ascend default_config.yaml 8 0,1,2,3,4,5,6,7 /root/hccl_8p_01234567_10.155.170.71.json /home/DataSet/ImageNet_Original/ /home/model/mobilenetv2/predtrain/mobilenet-200_625.ckpt none True
+ # example: bash run_train.sh Ascend default_config.yaml 8 0,1,2,3,4,5,6,7 ~/hccl_8p.json /home/DataSet/ImageNet_Original/ /home/model/mobilenetv2/predtrain/mobilenet-200_625.ckpt none True
GPU: bash run_train.sh GPU 8 0,1,2,3,4,5,6,7 [TRAIN_DATASET_PATH] [CKPT_PATH] none True
CPU: bash run_train.sh CPU [TRAIN_DATASET_PATH] [CKPT_PATH] none True
@@ -301,7 +301,7 @@ You can start training using python or shell scripts. The usage of shell scripts
shell:
Ascend: bash run_train.sh Ascend [CONFIG_PATH] [DEVICE_NUM] [VISIABLE_DEVICES(0,1,2,3,4,5,6,7)] [RANK_TABLE_FILE] [DATASET_PATH] [CKPT_PATH] [FREEZE_LAYER]
- # example: bash run_train.sh Ascend default_config.yaml 8 0,1,2,3,4,5,6,7 /root/hccl_8p_01234567_10.155.170.71.json /home/DataSet/ImageNet_Original/ /home/model/mobilenetv2/backbone/mobilenet-200_625.ckpt backbone
+ # example: bash run_train.sh Ascend default_config.yaml 8 0,1,2,3,4,5,6,7 ~/hccl_8p.json /home/DataSet/ImageNet_Original/ /home/model/mobilenetv2/backbone/mobilenet-200_625.ckpt backbone
GPU: bash run_train.sh GPU 8 0,1,2,3,4,5,6,7 [TRAIN_DATASET_PATH] [CKPT_PATH] backbone
CPU: bash run_train.sh CPU [TRAIN_DATASET_PATH] [CKPT_PATH] backbone
diff --git a/official/cv/mobilenetv2/README_CN.md b/official/cv/mobilenetv2/README_CN.md
index e9c4b1055..9085a97fd 100644
--- a/official/cv/mobilenetv2/README_CN.md
+++ b/official/cv/mobilenetv2/README_CN.md
@@ -276,7 +276,7 @@ MobileNetV2鎬讳綋缃戠粶鏋舵瀯濡備笅锛�
shell:
Ascend: bash run_train.sh Ascend [CONFIG_PATH] [DEVICE_NUM] [VISIABLE_DEVICES(0,1,2,3,4,5,6,7)] [RANK_TABLE_FILE] [DATASET_PATH]
- # example: bash run_train.sh Ascend default_config.yaml 8 0,1,2,3,4,5,6,7 /root/hccl_8p_01234567_10.155.170.71.json /home/DataSet/ImageNet_Original/
+ # example: bash run_train.sh Ascend default_config.yaml 8 0,1,2,3,4,5,6,7 ~/hccl_8p.json /home/DataSet/ImageNet_Original/
GPU: bash run_train.sh GPU 8 0,1,2,3,4,5,6,7 [TRAIN_DATASET_PATH]
CPU: bash run_train.sh CPU [TRAIN_DATASET_PATH]
@@ -289,7 +289,7 @@ MobileNetV2鎬讳綋缃戠粶鏋舵瀯濡備笅锛�
shell:
Ascend: bash run_train.sh Ascend [CONFIG_PATH] [DEVICE_NUM] [VISIABLE_DEVICES(0,1,2,3,4,5,6,7)] [RANK_TABLE_FILE] [DATASET_PATH] [CKPT_PATH] [FREEZE_LAYER] [FILTER_HEAD]
- # example: bash run_train.sh Ascend default_config.yaml 8 0,1,2,3,4,5,6,7 /root/hccl_8p_01234567_10.155.170.71.json /home/DataSet/ImageNet_Original/ /home/model/mobilenetv2/predtrain/mobilenet-200_625.ckpt none True
+ # example: bash run_train.sh Ascend default_config.yaml 8 0,1,2,3,4,5,6,7 ~/hccl_8p.json /home/DataSet/ImageNet_Original/ /home/model/mobilenetv2/predtrain/mobilenet-200_625.ckpt none True
GPU: bash run_train.sh GPU 8 0,1,2,3,4,5,6,7 [TRAIN_DATASET_PATH] [CKPT_PATH] none True
CPU: bash run_train.sh CPU [TRAIN_DATASET_PATH] [CKPT_PATH] none True
@@ -302,7 +302,7 @@ MobileNetV2鎬讳綋缃戠粶鏋舵瀯濡備笅锛�
shell:
Ascend: bash run_train.sh Ascend [CONFIG_PATH] [DEVICE_NUM] [VISIABLE_DEVICES(0,1,2,3,4,5,6,7)] [RANK_TABLE_FILE] [DATASET_PATH] [CKPT_PATH] [FREEZE_LAYER]
- # example: bash run_train.sh Ascend default_config.yaml 8 0,1,2,3,4,5,6,7 /root/hccl_8p_01234567_10.155.170.71.json /home/DataSet/ImageNet_Original/ /home/model/mobilenetv2/backbone/mobilenet-200_625.ckpt backbone
+ # example: bash run_train.sh Ascend default_config.yaml 8 0,1,2,3,4,5,6,7 ~/hccl_8p.json /home/DataSet/ImageNet_Original/ /home/model/mobilenetv2/backbone/mobilenet-200_625.ckpt backbone
GPU: bash run_train.sh GPU 8 0,1,2,3,4,5,6,7 [TRAIN_DATASET_PATH] [CKPT_PATH] backbone
CPU: bash run_train.sh CPU [TRAIN_DATASET_PATH] [CKPT_PATH] backbone
```
diff --git a/official/cv/mobilenetv2_quant/README_CN.md b/official/cv/mobilenetv2_quant/README_CN.md
index 45173675e..cbfa9eac4 100644
--- a/official/cv/mobilenetv2_quant/README_CN.md
+++ b/official/cv/mobilenetv2_quant/README_CN.md
@@ -164,17 +164,17 @@ MobileNetV2鎬讳綋缃戠粶鏋舵瀯濡備笅锛�
Ascend: python train.py --device_target Ascend --dataset_path ~/imagenet/train/
GPU: python train.py --device_target GPU --dataset_path ~/imagenet/train/
shell锛�
- Ascend: bash run_train.sh Ascend ~/hccl_4p_0123_x.x.x.x.json ~/imagenet/train/ ~/mobilenet.ckpt
+ Ascend: bash run_train.sh Ascend ~/hccl_8p.json ~/imagenet/train/
GPU: bash run_train.sh GPU 1,2 ~/imagenet/train/ ~/mobilenet.ckpt
- # 璁粌绀轰緥-閲忓寲姝ラ暱鍙涔犵殑閲忓寲鎰熺煡璁粌
+ # 璁粌绀轰緥-閲忓寲姝ラ暱鍙涔犵殑閲忓寲鎰熺煡璁粌
python锛�
Ascend: python train.py --device_target Ascend --dataset_path ~/imagenet/train/ \
--pre_trained ~/mobilenet.ckpt --optim_option "LEARNED_SCALE"
GPU: python train.py --device_target GPU --dataset_path ~/imagenet/train/ \
--pre_trained ~/mobilenet.ckpt --optim_option "LEARNED_SCALE"
shell锛�
- Ascend: bash run_lsq_train.sh Ascend ~/hccl_4p_0123_x.x.x.x.json ~/imagenet/train/ ~/mobilenet.ckpt
+ Ascend: bash run_lsq_train.sh Ascend ~/hccl_8p.json ~/imagenet/train/ ~/mobilenet.ckpt
GPU: bash run_lsq_train.sh GPU 1,2 ~/imagenet/train/ ~/mobilenet.ckpt
```
@@ -183,7 +183,7 @@ MobileNetV2鎬讳綋缃戠粶鏋舵瀯濡備笅锛�
璁粌缁撴灉淇濆瓨鍦ㄧず渚嬭矾寰勪腑銆俙Ascend`澶勭悊鍣ㄨ缁冪殑妫€鏌ョ偣榛樿淇濆瓨鍦╜./train/device$i/checkpoint`锛岃缁冩棩蹇楅噸瀹氬悜鍒癭./train/device$i/train.log`銆俙GPU`澶勭悊鍣ㄨ缁冪殑妫€鏌ョ偣榛樿淇濆瓨鍦╜./train/checkpointckpt_$i`涓紝璁粌鏃ュ織閲嶅畾鍚戝埌`./train/train.log`涓€�
`train.log`鍐呭濡備笅锛�
-```text
+```log
epoch:[ 0/200], step:[ 624/ 625], loss:[5.258/5.258], time:[140412.236], lr:[0.100]
epoch time:140522.500, per step time:224.836, avg loss:5.258
epoch:[ 1/200], step:[ 624/ 625], loss:[3.917/3.917], time:[138221.250], lr:[0.200]
@@ -234,13 +234,13 @@ shell:
鎺ㄧ悊缁撴灉淇濆瓨鍦ㄧず渚嬭矾寰勶紝鍙互鍦╜./val/infer.log`涓壘鍒板涓嬬粨鏋滐細
-```text
+```log
result:{'acc':0.71976314102564111}
```
## 妯″瀷瀵煎嚭
-```shell
+```python
python export.py --checkpoint_path [CKPT_PATH] --file_format [EXPORT_FORMAT] --device_target [PLATFORM] --optim_option [OptimizeOption]
```
@@ -253,7 +253,7 @@ python export.py --checkpoint_path [CKPT_PATH] --file_format [EXPORT_FORMAT] --d
鍦ㄦ帹鐞嗕箣鍓嶉渶瑕佸湪鏄囪吘910鐜涓婂畬鎴怉IR妯″瀷鐨勫鍑恒€�
骞朵娇鐢╡xport_bin_file.py瀵煎嚭ImageNet鏁版嵁闆嗙殑bin鏂囦欢鍜屽搴旂殑label鏂囦欢锛�
-```shell
+```log
python export_bin_file.py --dataset_dir [EVAL_DATASET_PATH] --save_dir [SAVE_PATH]
```
@@ -266,7 +266,7 @@ bash run_infer_310.sh [AIR_PATH] [DATA_PATH] [LABEL_PATH] [DEVICE_ID]
鎮ㄥ彲浠ラ€氳繃acc.log鏂囦欢鏌ョ湅缁撴灉銆俀AT閲忓寲鎺ㄧ悊鍑嗙‘鎬у涓嬶細
-```bash
+```log
'Accuracy':0.7221
```
diff --git a/official/cv/mobilenetv2_quant/Readme.md b/official/cv/mobilenetv2_quant/Readme.md
index b4fe57023..82719ea7f 100644
--- a/official/cv/mobilenetv2_quant/Readme.md
+++ b/official/cv/mobilenetv2_quant/Readme.md
@@ -160,7 +160,7 @@ For Learned Step Size Quantization:
Ascend: python train.py --device_target Ascend --dataset_path ~/imagenet/train/
GPU: python train.py --device_target GPU --dataset_path ~/imagenet/train/
shell锛�
- Ascend: bash run_train.sh Ascend ~/hccl_4p_0123_x.x.x.x.json ~/imagenet/train/ ~/mobilenet.ckpt
+ Ascend: bash run_train.sh Ascend ~/hccl_8p.json ~/imagenet/train/
GPU: bash run_train.sh GPU 1,2 ~/imagenet/train/ ~/mobilenet.ckpt
# training example for Learned Step Size Quantization
@@ -170,7 +170,7 @@ For Learned Step Size Quantization:
GPU: python train.py --device_target GPU --dataset_path ~/imagenet/train/ \
--pre_trained ~/mobilenet.ckpt --optim_option "LEARNED_SCALE"
shell锛�
- Ascend: bash run_lsq_train.sh Ascend ~/hccl_4p_0123_x.x.x.x.json ~/imagenet/train/ ~/mobilenet.ckpt
+ Ascend: bash run_lsq_train.sh Ascend ~/hccl_8p.json ~/imagenet/train/ ~/mobilenet.ckpt
GPU: bash run_lsq_train.sh GPU 1,2 ~/imagenet/train/ ~/mobilenet.ckpt
```
@@ -179,7 +179,7 @@ For Learned Step Size Quantization:
Training result will be stored in the example path. Checkpoints trained by `Ascend` will be stored at `./train/device$i/checkpoint` by default, and training log will be redirected to `./train/device$i/train.log`. Checkpoints trained by `GPU` will be stored in `./train/checkpointckpt_$i` by default, and training log will be redirected to `./train/train.log`.
`train.log` is as follows:
-``` bash
+``` log
epoch: [ 0/200], step:[ 624/ 625], loss:[5.258/5.258], time:[140412.236], lr:[0.100]
epoch time: 140522.500, per step time: 224.836, avg loss: 5.258
epoch: [ 1/200], step:[ 624/ 625], loss:[3.917/3.917], time:[138221.250], lr:[0.200]
@@ -232,13 +232,13 @@ shell:
Inference result will be stored in the example path, you can find result like the following in `./val/infer.log`.
-``` bash
+``` log
result: {'acc': 0.71976314102564111}
```
## [Model Export](#contents)
-```shell
+```python
python export.py --checkpoint_path [CKPT_PATH] --file_format [EXPORT_FORMAT] --device_target [PLATFORM] --optim_option [OptimizeOption]
```
@@ -250,7 +250,7 @@ python export.py --checkpoint_path [CKPT_PATH] --file_format [EXPORT_FORMAT] --d
You should export AIR model at Ascend 910 before running the command below.
You can use export_bin_file.py to export ImageNet bin and label for 310 inference.
-```shell
+```python
python export_bin_file.py --dataset_dir [EVAL_DATASET_PATH] --save_dir [SAVE_PATH]
```
@@ -263,7 +263,7 @@ bash run_infer_310.sh [AIR_PATH] [DATA_PATH] [LABEL_PATH] [DEVICE_ID]
You can view the results through the file "acc.log". The accuracy of the test dataset will be as follows:
-```bash
+```log
'Accuracy':0.7221
```
diff --git a/official/cv/openpose/README.md b/official/cv/openpose/README.md
index 2bca04112..5671daea5 100644
--- a/official/cv/openpose/README.md
+++ b/official/cv/openpose/README.md
@@ -49,7 +49,7 @@ In the currently provided training script, the coco2017 data set is used as an e
Run python gen_ignore_mask.py
````python
- python gen_ignore_mask.py --train_ann ../dataset/annotations/person_keypoints_train2017.json --val_ann ../dataset/annotations/person_keypoints_val2017.json --train_dir ../dataset/train2017 --val_dir ../dataset/val2017
+ python gen_ignore_mask.py --train_ann /home/DataSet/coco/annotations/person_keypoints_train2017.json --val_ann /home/DataSet/coco/annotations/person_keypoints_val2017.json --train_dir /home/DataSet/coco/train2017 --val_dir /home/DataSet/coco/val2017
````
- The dataset folder is generated in the root directory and contains the following files:
@@ -59,8 +59,8 @@ In the currently provided training script, the coco2017 data set is used as an e
鈹溾攢鈹€ annotations
鈹溾攢person_keypoints_train2017.json
鈹斺攢person_keypoints_val2017.json
- 鈹溾攢ignore_mask_train2017
- 鈹溾攢ignore_mask_val2017
+ 鈹溾攢ignore_mask_train
+ 鈹溾攢ignore_mask_val
鈹溾攢train2017
鈹斺攢val2017
```
@@ -90,15 +90,17 @@ After installing MindSpore via the official website, you can start training and
```python
# run training example
- python train.py --imgpath_train ./train2017 --jsonpath_train ./person_keypoints_train2017.json --maskpath_train ./ignore_mask_train2017 --vgg_path ./vgg19-0-97_5004.ckpt > train.log 2>&1 &
+ python train.py --imgpath_train /home/DataSet/coco/train2017 --jsonpath_train /home/DataSet/coco/annotations/person_keypoints_train2017.json --maskpath_train /home/DataSet/coco/ignore_mask_train --vgg_path /home/model/openpose/vgg19-0-97_5004.ckpt > train.log 2>&1 &
# run distributed training example
bash run_distribute_train.sh [RANK_TABLE_FILE] [IMGPATH_TRAIN] [JSONPATH_TRAIN] [MASKPATH_TRAIN] [VGG_PATH]
+ # example: bash run_distribute_train.sh ~/hccl_8p.json /home/DataSet/coco/train2017 /home/DataSet/coco/annotations/person_keypoints_train2017.json /home/DataSet/coco/ignore_mask_train /home/model/openpose/vgg19-0-97_5004.ckpt
# run evaluation example
- python eval.py --model_path path_to_eval_model.ckpt --imgpath_val ./dataset/val2017 --ann ./dataset/annotations/person_keypoints_val2017.json > eval.log 2>&1 &
+ python eval.py --model_path /home/model/openpose/ckpt/0-8_663.ckpt --imgpath_val /home/DataSet/coco/val2017 --ann /home/DataSet/coco/annotations/person_keypoints_val2017.json > eval.log 2>&1 &
OR
bash scripts/run_eval_ascend.sh [MODEL_PATH] [IMGPATH_VAL] [ANN]
+ # example: bash scripts/run_eval_ascend.sh /home/model/openpose/ckpt/0-8_663.ckpt /home/DataSet/coco/val2017 /home/DataSet/coco/annotations/person_keypoints_val2017.json
```
[RANK_TABLE_FILE] is the path of the multi-card information configuration table in the environment. The configuration table can be automatically generated by the tool [hccl_tool](https://gitee.com/mindspore/mindspore/tree/master/model_zoo/utils/hccl_tools).
@@ -165,14 +167,14 @@ For more configuration details, please refer the script `default_config.yaml`.
- running on Ascend
```python
- python train.py --imgpath_train ./train2017 --jsonpath_train ./person_keypoints_train2017.json --maskpath_train ./ignore_mask_train2017 --vgg_path ./vgg19-0-97_5004.ckpt > train.log 2>&1 &
+ python train.py --imgpath_train /home/DataSet/coco/train2017 --jsonpath_train /home/DataSet/coco/annotations/person_keypoints_train2017.json --maskpath_train /home/DataSet/coco/ignore_mask_train --vgg_path /home/model/openpose/vgg19-0-97_5004.ckpt > train.log 2>&1 &
```
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 script folder by default. The loss value will be achieved as follows:
- ```python
+ ```log
# grep "epoch " train.log
epoch[0], iter[23], mean loss is 0.292112287
epoch[0], iter[123], mean loss is 0.060355084
@@ -183,17 +185,17 @@ For more configuration details, please refer the script `default_config.yaml`.
The model checkpoint will be saved in the directory of default_config.yaml: 'save_model_path'.
- running on ModelArts
-- If you want to train the model on modelarts, you can refer to the [official guidance document] of modelarts (https://support.huaweicloud.com/modelarts/)
+- If you want to train the model on modelarts, you can refer to the official guidance document of [modelarts] (https://support.huaweicloud.com/modelarts/)
-```python
+```ModelArts
# Example of using distributed training dpn on modelarts :
# Data set storage method
# 鈹溾攢鈹€ openpose_dataset
# 鈹溾攢鈹€ annotations
# 鈹溾攢person_keypoints_train2017.json
# 鈹斺攢person_keypoints_val2017.json
-# 鈹溾攢ignore_mask_train2017
-# 鈹溾攢ignore_mask_val2017
+# 鈹溾攢ignore_mask_train
+# 鈹溾攢ignore_mask_val
# 鈹溾攢train2017
# 鈹斺攢val2017
# 鈹斺攢checkpoint
@@ -248,14 +250,16 @@ For more configuration details, please refer the script `default_config.yaml`.
Before running the command below, please check the checkpoint path used for evaluation. Please set the checkpoint path to be the absolute full path, e.g., "username/openpose/outputs/\*time*\/0-6_30000.ckpt".
```python
- python eval.py --model_path path_to_eval_model.ckpt --imgpath_val ./dataset/val2017 --ann ./dataset/annotations/person_keypoints_val2017.json > eval.log 2>&1 &
+ # run evaluation example
+ python eval.py --model_path /home/model/openpose/ckpt/0-8_663.ckpt --imgpath_val /home/DataSet/coco/val2017 --ann /home/DataSet/coco/annotations/person_keypoints_val2017.json > eval.log 2>&1 &
OR
bash scripts/run_eval_ascend.sh [MODEL_PATH] [IMGPATH_VAL] [ANN]
+ # example: bash scripts/run_eval_ascend.sh /home/model/openpose/ckpt/0-8_663.ckpt /home/DataSet/coco/val2017 /home/DataSet/coco/annotations/person_keypoints_val2017.json
```
The above python command will run in the background. You can view the results through the file "eval.log". The accuracy of the test dataset will be as follows:
- ```python
+ ```log
# grep "AP" eval.log
{'AP': 0.40250956300341397, 'Ap .5': 0.6658941566481336, 'AP .75': 0.396047897339743, 'AP (M)': 0.3075356543635785, 'AP (L)': 0.533772768618845, 'AR': 0.4519836272040302, 'AR .5': 0.693639798488665, 'AR .75': 0.4570214105793451, 'AR (M)': 0.32155148866429945, 'AR (L)': 0.6330360460795242}
diff --git a/official/cv/retinanet/README_CN.md b/official/cv/retinanet/README_CN.md
index cf76f65f9..72076957c 100644
--- a/official/cv/retinanet/README_CN.md
+++ b/official/cv/retinanet/README_CN.md
@@ -224,7 +224,7 @@ python create_data.py --create_dataset coco --prefix retinanet_eval.mindrecord -
Ascend:
# 鍏崱骞惰璁粌绀轰緥(鍦╮etinanet鐩綍涓嬭繍琛�)锛�
bash scripts/run_distribute_train.sh [DEVICE_NUM] [RANK_TABLE_FILE] [MINDRECORD_DIR] [PRE_TRAINED(optional)] [PRE_TRAINED_EPOCH_SIZE(optional)]
-# example: bash scripts/run_distribute_train.sh 8 /root/hccl_8p_01234567_10.155.170.71.json /home/DataSet/MindRecord_COCO/
+# example: bash scripts/run_distribute_train.sh 8 ~/hccl_8p.json /home/DataSet/MindRecord_COCO/
# 鍗曞崱璁粌绀轰緥(鍦╮etinanet鐩綍涓嬭繍琛�)锛�
bash scripts/run_single_train.sh [DEVICE_ID] [MINDRECORD_DIR]
diff --git a/official/cv/shufflenetv1/README_CN.md b/official/cv/shufflenetv1/README_CN.md
index 40b49f099..baca46001 100644
--- a/official/cv/shufflenetv1/README_CN.md
+++ b/official/cv/shufflenetv1/README_CN.md
@@ -125,11 +125,14 @@ ShuffleNetV1鐨勬牳蹇冮儴鍒嗚鍒嗘垚涓変釜闃舵锛屾瘡涓樁娈甸噸澶嶅爢绉簡
python:
Ascend鍗曞崱璁粌绀轰緥锛歱ython train.py --train_dataset_path [DATA_DIR]
+ # example: python train.py --train_dataset_path /home/DataSet/ImageNet_Original/train
shell:
-<<<<<<< HEAD
- Ascend鍏崱骞惰璁粌: sh scripts/run_distribute_train.sh [RANK_TABLE_FILE] [DATA_DIR]
- Ascend鍗曞崱璁粌绀轰緥: sh scripts/run_standalone_train.sh [DEVICE_ID] [DATA_DIR]
+ Ascend鍏崱骞惰璁粌: bash scripts/run_distribute_train.sh [RANK_TABLE_FILE] [DATA_DIR]
+ # example: bash scripts/run_distribute_train.sh ~/hccl_8p.json /home/DataSet/ImageNet_Original/train
+
+ Ascend鍗曞崱璁粌绀轰緥: bash scripts/run_standalone_train.sh [DEVICE_ID] [DATA_DIR]
+ # example: bash scripts/run_standalone_train.sh 0 /home/DataSet/ImageNet_Original/train
- running on GPU with gpu default parameters
@@ -146,10 +149,6 @@ ShuffleNetV1鐨勬牳蹇冮儴鍒嗚鍒嗘垚涓変釜闃舵锛屾瘡涓樁娈甸噸澶嶅爢绉簡
shell:
GPU鍗曞崱璁粌绀轰緥: sh scripts/run_standalone_train_gpu.sh [DEVICE_ID] [DATA_DIR]
GPU鍏崱骞惰璁粌: sh scripts/run_distribute_train_gpu.sh [RANK_SIZE] [TRAIN_DATA_DIR]
-=======
- Ascend鍏崱骞惰璁粌: bash scripts/run_distribute_train.sh [RANK_TABLE_FILE] [DATA_DIR]
- Ascend鍗曞崱璁粌绀轰緥: bash scripts/run_standalone_train.sh [DEVICE_ID] [DATA_DIR]
->>>>>>> fe806b7430... update bash
```
鍒嗗竷寮忚缁冮渶瑕佹彁鍓嶅垱寤篔SON鏍煎紡鐨凥CCL閰嶇疆鏂囦欢銆�
@@ -162,7 +161,7 @@ ShuffleNetV1鐨勬牳蹇冮儴鍒嗚鍒嗘垚涓変釜闃舵锛屾瘡涓樁娈甸噸澶嶅爢绉簡
ckpt鏂囦欢灏嗗瓨鍌ㄥ湪 `save_ckpt_path` 璺緞涓嬶紝璁粌鏃ュ織灏嗚璁板綍鍒� `log.txt` 涓€傝缁冩棩蹇楅儴鍒嗙ず渚嬪涓嬶細
-```shell
+```log
epoch time: 99854.980, per step time: 79.820, avg loss: 4.093
epoch time: 99863.734, per step time: 79.827, avg loss: 4.010
epoch time: 99859.792, per step time: 79.824, avg loss: 3.869
@@ -180,9 +179,11 @@ epoch time: 99864.092, per step time: 79.827, avg loss: 3.442
# Ascend璇勪及绀轰緥
python:
python eval.py --eval_dataset_path [DATA_DIR] --ckpt_path [PATH_CHECKPOINT]
+ # example: python eval.py --eval_dataset_path /home/DataSet/ImageNet_Original/validation_preprocess --ckpt_path /home/model/shufflenetv1/ckpt/shufflenetv1-250_1251
shell:
bash scripts/run_eval.sh [DEVICE_ID] [DATA_DIR] [PATH_CHECKPOINT]
+ # example: bash scripts/run_eval.sh 0 /home/DataSet/ImageNet_Original/validation_preprocess /home/model/shufflenetv1/ckpt/shufflenetv1-250_1251
# GPU璇勪及绀轰緥
python:
@@ -196,7 +197,7 @@ epoch time: 99864.092, per step time: 79.827, avg loss: 3.442
鍙互鍦� `eval_log.txt` 鏌ョ湅璇勪及缁撴灉銆�
-```shell
+```log
result:{'Loss': 2.0479587888106323, 'Top_1_Acc': 0.7385817307692307, 'Top_5_Acc': 0.9135817307692308}, ckpt:'/home/shufflenetv1/train_parallel0/checkpoint/shufflenetv1-250_1251.ckpt', time: 98560.63866615295
```
diff --git a/official/cv/simclr/README.md b/official/cv/simclr/README.md
index 56553eb82..d14c5dda6 100644
--- a/official/cv/simclr/README.md
+++ b/official/cv/simclr/README.md
@@ -39,7 +39,7 @@ Dataset used: [CIFAR-10](<http://www.cs.toronto.edu/~kriz/cifar.html>)
- Note锛欴ata will be processed in dataset.py
- Download the dataset, the directory structure is as follows:
-```bash
+```cifar10
鈹溾攢cifar-10-batches-bin
鈹�
鈹斺攢cifar-10-verify-bin
@@ -62,10 +62,14 @@ After installing MindSpore via the official website, you can start training and
```python
# enter script dir, train SimCLR
bash run_standalone_train_ascend.sh [cifar10] [TRAIN_DATASET_PATH] [DEVICE_ID]
+# example: bash run_standalone_train_ascend cifar10 /home/DataSet/cifar10/cifar-10-batches-bin/ 0
or
bash run_distribution_ascend.sh [DEVICENUM] [RANK_TABLE_FILE] [cifar10] [TRAIN_DATASET_PATH]
+# example: bash run_distribution_ascend.sh 8 ~/hccl_8p.json cifar10 /home/DataSet/cifar10/cifar-10-batches-bin/
+
# enter script dir, evaluate SimCLR
bash run_standalone_eval_ascend.sh [cifar10] [DEVICE_ID] [SIMCLR_MODEL_PATH] [TRAIN_DATASET_PATH] [EVAL_DATASET_PATH]
+# example: run_standalone_eval_ascend.sh cifar10 0 /home/model/simclr/ckpt/checkpoint-simclr-100_390.ckpt /home/DataSet/cifar10/cifar-10-batches-bin/ /home/DataSet/cifar10/cifar-10-verify-bin/
```
## [Script Description](#contents)
@@ -152,11 +156,12 @@ Major parameters in linear_eval.py as follows:
```bash
bash run_distribution_ascend.sh [DEVICENUM] [RANK_TABLE_FILE] [cifar10] [TRAIN_DATASET_PATH]
+ # example: bash run_distribution_ascend.sh 8 ~/hccl_8p.json cifar10 /home/DataSet/cifar10/cifar-10-batches-bin/
```
After training, the loss value will be achieved as follows:
- ```bash
+ ```log
# grep "loss is " log
epoch: 1 step: 48, loss is 9.5758915
epoch time: 253236.075 ms, per step time: 5275.752 ms
@@ -186,11 +191,12 @@ Before running the command below, please check the checkpoint path used for eval
```bash
bash run_standalone_eval_ascend.sh [cifar10] [DEVICE_ID] [SIMCLR_MODEL_PATH] [TRAIN_DATASET_PATH] [EVAL_DATASET_PATH]
+ # example: run_standalone_eval_ascend.sh cifar10 0 /home/model/simclr/ckpt/checkpoint-simclr-100_390.ckpt /home/DataSet/cifar10/cifar-10-batches-bin/ /home/DataSet/cifar10/cifar-10-verify-bin/
```
You can view the results through the file "eval_log". The accuracy of the test dataset will be as follows:
- ```bash
+ ```log
# grep "Average accuracy: " eval_log
'Accuracy': 0.84505
```
diff --git a/official/cv/simple_pose/README.md b/official/cv/simple_pose/README.md
index e0c182c40..84db42fb4 100644
--- a/official/cv/simple_pose/README.md
+++ b/official/cv/simple_pose/README.md
@@ -98,6 +98,22 @@ Before you start your training process, you need to obtain mindspore imagenet pr
鈹斺攢resnet50.ckpt
```
+```dataset
+ 鈹斺攢 cocodataset
+ 鈹斺攢train2017
+ 鈹斺攢val2017
+ 鈹斺攢annotations
+```
+
+```default_config.yaml
+DATASET:
+ ROOT:/home/DataSet/cocodataset
+MODEL:
+ PRETRAINED:./resnet50.ckpt
+
+# Modify according to local path
+```
+
## [Running](#contents)
- running on local
@@ -106,20 +122,23 @@ Before you start your training process, you need to obtain mindspore imagenet pr
```shell
bash scripts/train_standalone.sh [CKPT_SAVE_DIR] [DEVICE_ID] [BATCH_SIZE]
+ # example: bash scripts/train_standalone.sh ./ 0 128
```
To validate the model, change the settings in `default_config.yaml` to the path of the model you want to validate or setting that on the terminal. For example:
- ```python
+ ```default_config.yaml
TEST:
...
- MODEL_FILE : './{path}/xxxx.ckpt'
+ MODEL_FILE : /home/model/simple_pose/ckpt/simplepose-140_1170.ckpt
+ # Modify according to local path
```
Then, run the shell script `scripts/eval.sh` with the format below:
```shell
bash scripts/eval.sh [TEST_MODEL_FILE] [COCO_BBOX_FILE] [DEVICE_ID]
+ # example: bash scripts/eval.sh /home/model/simple_pose/ckpt/simplepose-140_1170.ckpt ./experiments/COCO_val2017_detections_AP_H_56_person.json 0
```
- running on ModelArts
@@ -211,7 +230,7 @@ Configurations for both training and evaluation are set in `default_config.yaml`
- config for SimplePoseNet on COCO2017 dataset:
-```python
+```default_config.yaml
# These parameters can be modified at the terminal
ckpt_save_dir: 'checkpoints' # the folder to save the '*.ckpt' file
batch_size: 128 # TRAIN.BATCH_SIZE
@@ -280,17 +299,12 @@ Run `scripts/train_standalone.sh` to train the model standalone. The usage of th
```shell
bash scripts/train_standalone.sh [CKPT_SAVE_DIR] [DEVICE_ID] [BATCH_SIZE]
-```
-
-For example, you can run the shell command below to launch the training procedure.
-
-```shell
-bash scripts/train_standalone.sh results/standalone/ 0 128
+# example: bash scripts/train_standalone.sh ./ 0 128
```
The script will run training in the background, you can view the results through the file `train_log[X].txt` as follows:
-```text
+```log
loading parse...
batch size :128
loading dataset from /data/coco2017/train2017
@@ -314,17 +328,12 @@ Run `scripts/train_distributed.sh` to train the model distributed. The usage of
```shell
bash scripts/train_distributed.sh [MINDSPORE_HCCL_CONFIG_PATH] [CKPT_SAVE_DIR] [RANK_SIZE]
-```
-
-For example, you can run the shell command below to launch the distributed training procedure.
-
-```shell
-bash scripts/train_distributed.sh /home/rank_table.json results/distributed/ 4
+# example: bash scripts/train_distributed.sh /root/hccl_8p_01234567_10.155.170.71.json ./checkpoint 8
```
The above shell script will run distribute training in the background. You can view the results through the file `train_parallel[X]/log.txt` as follows:
-```text
+```log
loading parse...
batch size :64
loading dataset from /data/coco2017/train2017
@@ -348,17 +357,12 @@ run `scripts/eval.sh` to evaluate the model with one Ascend processor. The usage
```shell
bash scripts/eval.sh [TEST_MODEL_FILE] [COCO_BBOX_FILE] [DEVICE_ID]
-```
-
-For example, you can run the shell command below to launch the validation procedure.
-
-```shell
-bash scripts/eval.sh results/distributed/sim-140_1170.ckpt
+# example: bash scripts/eval.sh /home/model/simple_pose/ckpt/simplepose-140_1170.ckpt ./experiments/COCO_val2017_detections_AP_H_56_person.json 0
```
The above shell command will run validation procedure in the background. You can view the results through the file `eval_log[X].txt`. The result will be achieved as follows:
-```text
+```log
use flip test: True
loading model ckpt from results/distributed/sim-140_1170.ckpt
loading dataset from /data/coco2017/val2017
@@ -419,7 +423,7 @@ bash run_infer_310.sh [MINDIR_PATH] [NEED_PREPROCESS] [DEVICE_ID]
Inference result is saved in current path, you can find result like this in acc.log file.
-```bash
+```log
AP: 0.7036180026660003
```
diff --git a/official/cv/squeezenet/README.md b/official/cv/squeezenet/README.md
index 7e9ec5313..894fc71ba 100644
--- a/official/cv/squeezenet/README.md
+++ b/official/cv/squeezenet/README.md
@@ -82,14 +82,21 @@ After installing MindSpore via the official website, you can start training and
- running on Ascend
```bash
+ run squeezenet_residual as example
# distributed training
Usage: bash scripts/run_distribute_train.sh [squeezenet|squeezenet_residual] [cifar10|imagenet] [RANK_TABLE_FILE] [DATA_PATH] [PRETRAINED_CKPT_PATH](optional)
+ # example: bash scripts/run_distribute_train.sh squeezenet_residual imagenet ~/hccl_8p.json /home/DataSet/ImageNet_Original/train
+ # example: bash scripts/run_distribute_train.sh squeezenet_residual cifar10 ~/hccl_8p.json /home/DataSet/cifar10/cifar-10-batches-bin
# standalone training
Usage: bash scripts/run_standalone_train.sh [squeezenet|squeezenet_residual] [cifar10|imagenet] [DEVICE_ID] [DATA_PATH] [PRETRAINED_CKPT_PATH](optional)
+ # example: bash scripts/run_standalone_train.sh squeezenet_residual imagenet 0 /home/DataSet/ImageNet_Original/train
+ # example: bash scripts/run_standalone_train.sh squeezenet_residual cifar10 0 /home/DataSet/cifar10/cifar-10-batches-bin
# run evaluation example
Usage: bash scripts/run_eval.sh [squeezenet|squeezenet_residual] [cifar10|imagenet] [DEVICE_ID] [DATA_PATH] [CHECKPOINT_PATH]
+ # example bash scripts/run_eval.sh squeezenet_residual cifar10 0 /home/DataSet/cifar10/cifar-10-verify-bin /home/model/squeezenet/ckpt/squeezenet_residual_cifar10-120_1562.ckpt
+ # example bash scripts/run_eval.sh squeezenet_residual imagenet 0 /home/DataSet/ImageNet_Original/validation_preprocess /home/model/squeezenet/ckpt/squeezenet_residual_imagenet-300_5004.ckpt
```
- running on GPU
@@ -117,7 +124,7 @@ After installing MindSpore via the official website, you can start training and
If you want to run in modelarts, please check the official documentation of [modelarts](https://support.huaweicloud.com/modelarts/), and you can start training and evaluation as follows:
-```python
+```ModelArts
# run distributed training on modelarts example
# (1) First, Perform a or b.
# a. Set "enable_modelarts=True" on yaml file.
@@ -293,9 +300,13 @@ For more configuration details, please refer the file `*.yaml`.
```shell
# distributed training
Usage: bash scripts/run_distribute_train.sh [squeezenet|squeezenet_residual] [cifar10|imagenet] [RANK_TABLE_FILE] [DATA_PATH] [PRETRAINED_CKPT_PATH](optional)
+ # example: bash scripts/run_distribute_train.sh squeezenet_residual imagenet ~/hccl_8p.json /home/DataSet/ImageNet_Original/train
+ # example: bash scripts/run_distribute_train.sh squeezenet_residual cifar10 ~/hccl_8p.json /home/DataSet/cifar10/cifar-10-batches-bin
# standalone training
Usage: bash scripts/run_standalone_train.sh [squeezenet|squeezenet_residual] [cifar10|imagenet] [DEVICE_ID] [DATA_PATH] [PRETRAINED_CKPT_PATH](optional)
+ # example: bash scripts/run_standalone_train.sh squeezenet_residual imagenet 0 /home/DataSet/ImageNet_Original/train
+ # example: bash scripts/run_standalone_train.sh squeezenet_residual cifar10 0 /home/DataSet/cifar10/cifar-10-batches-bin
```
For distributed training, a hccl configuration file with JSON format needs to be created in advance.
@@ -308,7 +319,7 @@ Training result will be stored in the example path, whose folder name begins wit
- Training SqueezeNet with CIFAR-10 dataset
-```shell
+```log
# standalone training result
epoch: 1 step 1562, loss is 1.7103254795074463
epoch: 2 step 1562, loss is 2.06101131439209
@@ -320,7 +331,7 @@ epoch: 5 step 1562, loss is 1.2140142917633057
- Training SqueezeNet with ImageNet dataset
-```shell
+```log
# distribute training result(8 pcs)
epoch: 1 step 5004, loss is 5.716324329376221
epoch: 2 step 5004, loss is 5.350603103637695
@@ -332,7 +343,7 @@ epoch: 5 step 5004, loss is 4.136358261108398
- Training SqueezeNet_Residual with CIFAR-10 dataset
-```shell
+```log
# standalone training result
epoch: 1 step 1562, loss is 2.298271656036377
epoch: 2 step 1562, loss is 2.2728664875030518
@@ -344,7 +355,7 @@ epoch: 5 step 1562, loss is 1.3370063304901123
- Training SqueezeNet_Residual with ImageNet dataset
-```shell
+```log
# distribute training result(8 pcs)
epoch: 1 step 5004, loss is 6.802495002746582
epoch: 2 step 5004, loss is 6.386072158813477
@@ -363,11 +374,8 @@ epoch: 5 step 5004, loss is 4.888848304748535
```shell
# evaluation
Usage: bash scripts/run_eval.sh [squeezenet|squeezenet_residual] [cifar10|imagenet] [DEVICE_ID] [DATA_PATH] [CHECKPOINT_PATH]
-```
-
-```shell
-# evaluation example
-bash scripts/run_eval.sh squeezenet cifar10 0 ~/cifar-10-verify-bin train/squeezenet_cifar10-120_1562.ckpt
+# example bash scripts/run_eval.sh squeezenet_residual cifar10 0 /home/DataSet/cifar10/cifar-10-verify-bin /home/model/squeezenet/ckpt/squeezenet_residual_cifar10-120_1562.ckpt
+# example bash scripts/run_eval.sh squeezenet_residual imagenet 0 /home/DataSet/ImageNet_Original/validation_preprocess /home/model/squeezenet/ckpt/squeezenet_residual_imagenet-300_5004.ckpt
```
checkpoint can be produced in training process.
@@ -378,25 +386,25 @@ Evaluation result will be stored in the example path, whose folder name is "eval
- Evaluating SqueezeNet with CIFAR-10 dataset
-```shell
+```log
result: {'top_1_accuracy': 0.8896233974358975, 'top_5_accuracy': 0.9965945512820513}
```
- Evaluating SqueezeNet with ImageNet dataset
-```shell
+```log
result: {'top_1_accuracy': 0.5851472471190781, 'top_5_accuracy': 0.8105393725992317}
```
- Evaluating SqueezeNet_Residual with CIFAR-10 dataset
-```shell
+```log
result: {'top_1_accuracy': 0.9077524038461539, 'top_5_accuracy': 0.9969951923076923}
```
- Evaluating SqueezeNet_Residual with ImageNet dataset
-```shell
+```log
result: {'top_1_accuracy': 0.6094950384122919, 'top_5_accuracy': 0.826324423815621}
```
@@ -406,7 +414,7 @@ result: {'top_1_accuracy': 0.6094950384122919, 'top_5_accuracy': 0.8263244238156
Export MindIR on local
-```shell
+```python
python export.py --checkpoint_file_path [CKPT_PATH] --batch_size [BATCH_SIZE] --net_name [NET] --dataset [DATASET] --file_format [EXPORT_FORMAT] --config_path [CONFIG_PATH]
```
@@ -418,7 +426,7 @@ The checkpoint_file_path parameter is required,
Export 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 as follows)
-```python
+```ModelArts
# Export on ModelArts
# (1) Perform a or b.
# a. Set "enable_modelarts=True" on default_config.yaml file.
@@ -459,25 +467,25 @@ Inference result is saved in current path, you can find result like this in acc.
- Infer SqueezeNet with CIFAR-10 dataset
-```bash
+```log
'Top1_Accuracy': 83.62% 'Top5_Accuracy': 99.31%
```
- Infer SqueezeNet with ImageNet dataset
-```bash
+```log
'Top1_Accuracy': 59.30% 'Top5_Accuracy': 81.40%
```
- Infer SqueezeNet_Residual with CIFAR-10 dataset
-```bash
+```log
'Top1_Accuracy': 87.28% 'Top5_Accuracy': 99.58%
```
- Infer SqueezeNet_Residual with ImageNet dataset
-```bash
+```log
'Top1_Accuracy': 60.82% 'Top5_Accuracy': 82.56%
```
--
GitLab