Skip to content

Commit 78cd43a

Browse files
ChongWei905ChongWei905
authored
docs: update v0.5.0 docs for mindspore 2.5.0 (#817)
* docs: remove gpu support note above msrun running guidance. * docs: update v0.5.0 docs for ms 2.5.0
1 parent 04d3293 commit 78cd43a

File tree

56 files changed

+173
-174
lines changed

Some content is hidden

Large Commits have some content hidden by default. Use the searchbox below for content that may be hidden.

56 files changed

+173
-174
lines changed

‎README.md

Lines changed: 7 additions & 6 deletions
Original file line numberDiff line numberDiff line change
@@ -30,10 +30,11 @@ MindCV is an open-source toolbox for computer vision research and development ba
3030
The following is the corresponding `mindcv` versions and supported `mindspore` versions.
3131

3232
| mindcv | mindspore |
33-
| :----: | :---------: |
33+
|:------:|:-----------:|
3434
| main | master |
35-
| v0.4.0 | 2.3.0/2.3.1 |
36-
| 0.3.0 | 2.2.10 |
35+
| 0.5 | 2.5.0 |
36+
| 0.4 | 2.3.0/2.3.1 |
37+
| 0.3 | 2.2.10 |
3738
| 0.2 | 2.0 |
3839
| 0.1 | 1.8 |
3940

@@ -128,7 +129,7 @@ It is easy to train your model on a standard or customized dataset using `train.
128129

129130
```shell
130131
# distributed training
131-
# assume you have 4 GPUs/NPUs
132+
# assume you have 4 NPUs
132133
msrun --bind_core=True --worker_num 4 python train.py --distribute \
133134
--model=densenet121 --dataset=imagenet --data_dir=/path/to/imagenet
134135
```
@@ -141,7 +142,7 @@ It is easy to train your model on a standard or customized dataset using `train.
141142
python train.py --distribute --model=densenet121 --dataset=imagenet --data_dir=/path/to/imagenet
142143
```
143144

144-
> For more information, please refer to https://www.mindspore.cn/tutorials/experts/en/r2.3.1/parallel/startup_method.html
145+
> For more information, please refer to https://www.mindspore.cn/docs/en/r2.5.0/model_train/parallel/startup_method.html
145146

146147
Detailed parameter definitions can be seen in `config.py` and checked by running `python train.py --help'.
147148
@@ -152,7 +153,7 @@ It is easy to train your model on a standard or customized dataset using `train.
152153
You can configure your model and other components either by specifying external parameters or by writing a yaml config file. Here is an example of training using a preset yaml file.
153154
154155
```shell
155-
msrun --bind_core=True --worker_num 4 python train.py -c configs/squeezenet/squeezenet_1.0_gpu.yaml
156+
msrun --bind_core=True --worker_num 4 python train.py -c configs/squeezenet/squeezenet_1.0_ascend.yaml
156157
```
157158
158159
**Pre-defined Training Strategies:**

‎README_CN.md

Lines changed: 2 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -134,7 +134,7 @@ python infer.py --model=swin_tiny --image_path='./dog.jpg'
134134
python train.py --distribute --model=densenet121 --dataset=imagenet --data_dir=/path/to/imagenet
135135
```
136136

137-
> 如需更多操作指导，请参考 https://www.mindspore.cn/tutorials/experts/zh-CN/r2.3.1/parallel/startup_method.html
137+
> 如需更多操作指导，请参考 https://www.mindspore.cn/docs/zh-CN/r2.5.0/model_train/parallel/startup_method.html
138138

139139
完整的参数列表及说明在`config.py`中定义，可运行`python train.py --help`快速查看。
140140

@@ -145,7 +145,7 @@ python infer.py --model=swin_tiny --image_path='./dog.jpg'
145145
您可以编写yaml文件或设置外部参数来指定配置数据、模型、优化器等组件及其超参数。以下是使用预设的训练策略（yaml文件）进行模型训练的示例。
146146

147147
```shell
148-
msrun --bind_core=True --worker_num 4 python train.py -c configs/squeezenet/squeezenet_1.0_gpu.yaml
148+
msrun --bind_core=True --worker_num 4 python train.py -c configs/squeezenet/squeezenet_1.0_ascend.yaml
149149
```
150150

151151
**预定义的训练策略**

‎configs/bit/README.md

Lines changed: 3 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -16,7 +16,7 @@ too low. 5) With BiT fine-tuning, good performance can be achieved even if there
1616
## Requirements
1717
| mindspore | ascend driver | firmware | cann toolkit/kernel |
1818
| :-------: | :-----------: | :---------: | :-----------------: |
19-
| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 |
19+
| 2.5.0 | 24.1.0 | 7.5.0.3.220 | 8.0.0.beta1 |
2020

2121
## Quick Start
2222

@@ -67,11 +67,11 @@ python validate.py -c configs/bit/bit_resnet50_ascend.yaml --data_dir /path/to/i
6767

6868
Our reproduced model performance on ImageNet-1K is reported as follows.
6969

70-
Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode.
70+
Experiments are tested on ascend 910* with mindspore 2.5.0 graph mode.
7171

7272
*coming soon*
7373

74-
Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode.
74+
Experiments are tested on ascend 910 with mindspore 2.5.0 graph mode.
7575

7676

7777
| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight |

‎configs/cmt/README.md

Lines changed: 3 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -13,7 +13,7 @@ on ImageNet-1K dataset.
1313
## Requirements
1414
| mindspore | ascend driver | firmware | cann toolkit/kernel |
1515
| :-------: | :-----------: | :---------: | :-----------------: |
16-
| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 |
16+
| 2.5.0 | 24.1.0 | 7.5.0.3.220 | 8.0.0.beta1 |
1717

1818

1919
## Quick Start
@@ -65,11 +65,11 @@ python validate.py -c configs/cmt/cmt_small_ascend.yaml --data_dir /path/to/imag
6565

6666
Our reproduced model performance on ImageNet-1K is reported as follows.
6767

68-
Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode.
68+
Experiments are tested on ascend 910* with mindspore 2.5.0 graph mode.
6969

7070
*coming soon*
7171

72-
Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode.
72+
Experiments are tested on ascend 910 with mindspore 2.5.0 graph mode.
7373

7474
| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight |
7575
| ---------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | ------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------ |

‎configs/coat/README.md

Lines changed: 3 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -9,7 +9,7 @@ Co-Scale Conv-Attentional Image Transformer (CoaT) is a Transformer-based image
99
## Requirements
1010
| mindspore | ascend driver | firmware | cann toolkit/kernel |
1111
| :-------: | :-----------: | :---------: | :-----------------: |
12-
| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 |
12+
| 2.5.0 | 24.1.0 | 7.5.0.3.220 | 8.0.0.beta1 |
1313

1414

1515

@@ -61,12 +61,12 @@ python validate.py -c configs/coat/coat_lite_tiny_ascend.yaml --data_dir /path/t
6161

6262
Our reproduced model performance on ImageNet-1K is reported as follows.
6363

64-
Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode.
64+
Experiments are tested on ascend 910* with mindspore 2.5.0 graph mode.
6565

6666
*coming soon*
6767

6868

69-
Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode.
69+
Experiments are tested on ascend 910 with mindspore 2.5.0 graph mode.
7070

7171

7272

‎configs/convit/README.md

Lines changed: 3 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -23,7 +23,7 @@ while offering a much improved sample efficiency.[[1](#references)]
2323
## Requirements
2424
| mindspore | ascend driver | firmware | cann toolkit/kernel |
2525
| :-------: | :-----------: | :---------: | :-----------------: |
26-
| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 |
26+
| 2.5.0 | 24.1.0 | 7.5.0.3.220 | 8.0.0.beta1 |
2727

2828

2929

@@ -74,13 +74,13 @@ python validate.py -c configs/convit/convit_tiny_ascend.yaml --data_dir /path/to
7474

7575
Our reproduced model performance on ImageNet-1K is reported as follows.
7676

77-
Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode.
77+
Experiments are tested on ascend 910* with mindspore 2.5.0 graph mode.
7878

7979
| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight |
8080
| ----------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | ------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------- |
8181
| convit_tiny | 5.71 | 8 | 256 | 224x224 | O2 | 153s | 226.51 | 9022.03 | 73.79 | 91.70 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/convit/convit_tiny_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/convit/convit_tiny-1961717e-910v2.ckpt) |
8282

83-
Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode.
83+
Experiments are tested on ascend 910 with mindspore 2.5.0 graph mode.
8484

8585
| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight |
8686
| ----------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | ------------------------------------------------------------------------------------------------ | ----------------------------------------------------------------------------------------- |

‎configs/convnext/README.md

Lines changed: 3 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -20,7 +20,7 @@ simplicity and efficiency of standard ConvNets.[[1](#references)]
2020
## Requirements
2121
| mindspore | ascend driver | firmware | cann toolkit/kernel |
2222
| :-------: | :-----------: | :---------: | :-----------------: |
23-
| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 |
23+
| 2.5.0 | 24.1.0 | 7.5.0.3.220 | 8.0.0.beta1 |
2424

2525

2626

@@ -72,13 +72,13 @@ python validate.py -c configs/convnext/convnext_tiny_ascend.yaml --data_dir /pat
7272

7373
Our reproduced model performance on ImageNet-1K is reported as follows.
7474

75-
Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode.
75+
Experiments are tested on ascend 910* with mindspore 2.5.0 graph mode.
7676

7777
| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight |
7878
| ------------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | ---------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------- |
7979
| convnext_tiny | 28.59 | 8 | 16 | 224x224 | O2 | 137s | 48.7 | 2612.24 | 81.28 | 95.61 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/convnext/convnext_tiny_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/convnext/convnext_tiny-db11dc82-910v2.ckpt) |
8080

81-
Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode.
81+
Experiments are tested on ascend 910 with mindspore 2.5.0 graph mode.
8282

8383
| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight |
8484
| ------------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | ---------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------- |

‎configs/convnextv2/README.md

Lines changed: 3 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -20,7 +20,7 @@ benchmarks, including ImageNet classification, COCO detection, and ADE20K segmen
2020
## Requirements
2121
| mindspore | ascend driver | firmware | cann toolkit/kernel |
2222
| :-------: | :-----------: | :---------: | :-----------------: |
23-
| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 |
23+
| 2.5.0 | 24.1.0 | 7.5.0.3.220 | 8.0.0.beta1 |
2424

2525

2626
## Quick Start
@@ -71,13 +71,13 @@ python validate.py -c configs/convnextv2/convnextv2_tiny_ascend.yaml --data_dir
7171

7272
Our reproduced model performance on ImageNet-1K is reported as follows.
7373

74-
Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode.
74+
Experiments are tested on ascend 910* with mindspore 2.5.0 graph mode.
7575

7676
| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight |
7777
| --------------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | -------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------- |
7878
| convnextv2_tiny | 28.64 | 8 | 128 | 224x224 | O2 | 268s | 257.2 | 3984.44 | 82.39 | 95.95 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/convnextv2/convnextv2_tiny_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/convnextv2/convnextv2_tiny-a35b79ce-910v2.ckpt) |
7979

80-
Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode.
80+
Experiments are tested on ascend 910 with mindspore 2.5.0 graph mode.
8181

8282
| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight |
8383
| --------------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | -------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------- |

‎configs/crossvit/README.md

Lines changed: 3 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -17,7 +17,7 @@ Fusion is achieved by an efficient cross-attention module, in which each transfo
1717
## Requirements
1818
| mindspore | ascend driver | firmware | cann toolkit/kernel |
1919
| :-------: | :-----------: | :---------: | :-----------------: |
20-
| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 |
20+
| 2.5.0 | 24.1.0 | 7.5.0.3.220 | 8.0.0.beta1 |
2121

2222

2323
## Quick Start
@@ -67,13 +67,13 @@ python validate.py -c configs/crossvit/crossvit_15_ascend.yaml --data_dir /path/
6767

6868
Our reproduced model performance on ImageNet-1K is reported as follows.
6969

70-
Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode.
70+
Experiments are tested on ascend 910* with mindspore 2.5.0 graph mode.
7171

7272
| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight |
7373
| ---------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | ------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------- |
7474
| crossvit_9 | 8.55 | 8 | 256 | 240x240 | O2 | 221s | 514.36 | 3984.44 | 73.38 | 91.51 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/crossvit/crossvit_9_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/crossvit/crossvit_9-32c69c96-910v2.ckpt) |
7575

76-
Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode.
76+
Experiments are tested on ascend 910 with mindspore 2.5.0 graph mode.
7777

7878
| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight |
7979
| ---------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | ------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------ |

‎configs/densenet/README.md

Lines changed: 3 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -25,7 +25,7 @@ propagation, encourage feature reuse, and substantially reduce the number of par
2525
## Requirements
2626
| mindspore | ascend driver | firmware | cann toolkit/kernel |
2727
| :-------: | :-----------: | :---------: | :-----------------: |
28-
| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 |
28+
| 2.5.0 | 24.1.0 | 7.5.0.3.220 | 8.0.0.beta1 |
2929

3030
## Quick Start
3131

@@ -75,14 +75,14 @@ python validate.py -c configs/densenet/densenet_121_ascend.yaml --data_dir /path
7575

7676
Our reproduced model performance on ImageNet-1K is reported as follows.
7777

78-
Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode.
78+
Experiments are tested on ascend 910* with mindspore 2.5.0 graph mode.
7979

8080

8181
| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight |
8282
| ----------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | --------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------- |
8383
| densenet121 | 8.06 | 8 | 32 | 224x224 | O2 | 300s | 47,34 | 5446.81 | 75.67 | 92.77 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/densenet/densenet_121_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/densenet/densenet121-bf4ab27f-910v2.ckpt) |
8484

85-
Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode.
85+
Experiments are tested on ascend 910 with mindspore 2.5.0 graph mode.
8686

8787
| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight |
8888
| ----------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | --------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------- |

0 commit comments

Comments
 (0)