Skip to content

M
models

PaddlePaddle / models
大约 1 年 前同步成功

通知 222
Star 6828
Fork 2962
M
models
未验证 提交 dfe7e97c 编写于 作者: G Guoxia Wang 提交者： GitHub
浏览文件
操作

fix ipynb doc and links in download (#5704)

上级 6f94a966
隐藏空白更改
内联 并排
Showing with 81 addition and 161 deletion
modelcenter/PLSC-ViT/download_cn.md
浏览文件 @ dfe7e97c
......@@ -2,8 +2,7 @@
|模型名称|模型简介|模型配置|预训练checkpoint下载地址|
| --- | --- | --- | --- |
| ViT-B_16_224 |输入size为224，layers=12|[config](./configs/ViT_base_patch16_224_in1k_1n8c_dp_fp16o2.yaml) |[download](https://plsc.bj.bcebos.com/models/vit/v2.4/imagenet2012-ViT-B_16-224.pdparams) |
| ViT-B_16_384 |输入size为384，layers=12|[config](./configs/ViT_base_patch16_384_ft_in1k_1n8c_dp_fp16o2.yaml)| [download](https://plsc.bj.bcebos.com/models/vit/v2.4/imagenet2012-ViT-B_16-224.pdparams) |
| ViT-L_16_224 |输入size为224，layers=24|[config](./configs/ViT_large_patch16_224_in21k_4n32c_dp_fp16o2.yaml)| [download](https://plsc.bj.bcebos.com/models/vit/v2.4/imagenet21k-ViT-L_16-224.pdparams) |
| ViT-L_16_384 |输入size为384，layers=32|[config](./configs/ViT_large_patch16_384_in1k_ft_4n32c_dp_fp16o2.yaml) | [download](https://plsc.bj.bcebos.com/models/vit/v2.4/imagenet21k-ViT-L_16-224.pdparams) |
| ViT-B_16_224 |输入size为224，layers=12|[config](https://github.com/PaddlePaddle/PLSC/blob/release/2.4/task/classification/vit/configs/ViT_base_patch16_224_in1k_1n8c_dp_fp16o2.yaml) |[download](https://plsc.bj.bcebos.com/models/vit/v2.4/imagenet2012-ViT-B_16-224.pdparams) |
| ViT-B_16_384 |输入size为384，layers=12|[config](https://github.com/PaddlePaddle/PLSC/blob/release/2.4/task/classification/vit/configs/ViT_base_patch16_384_ft_in1k_1n8c_dp_fp16o2.yaml)| [download](https://plsc.bj.bcebos.com/models/vit/v2.4/imagenet2012-ViT-B_16-224.pdparams) |
| ViT-L_16_224 |输入size为224，layers=24|[config](https://github.com/PaddlePaddle/PLSC/blob/release/2.4/task/classification/vit/configs/ViT_large_patch16_224_in21k_4n32c_dp_fp16o2.yaml)| [download](https://plsc.bj.bcebos.com/models/vit/v2.4/imagenet21k-ViT-L_16-224.pdparams) |
| ViT-L_16_384 |输入size为384，layers=32|[config](https://github.com/PaddlePaddle/PLSC/blob/release/2.4/task/classification/vit/configs/ViT_large_patch16_384_in1k_ft_4n32c_dp_fp16o2.yaml) | [download](https://plsc.bj.bcebos.com/models/vit/v2.4/imagenet21k-ViT-L_16-224.pdparams) |
modelcenter/PLSC-ViT/download_en.md
浏览文件 @ dfe7e97c
......@@ -2,8 +2,7 @@
|Model Name|Introduction|Config|Pretrained checkpoint Download|
| --- | --- | --- | --- |
| ViT-B_16_224 |input_size=224，layers=12|[config](./configs/ViT_base_patch16_224_in1k_1n8c_dp_fp16o2.yaml) |[download](https://plsc.bj.bcebos.com/models/vit/v2.4/imagenet2012-ViT-B_16-224.pdparams) |
| ViT-B_16_384 |input_size=384，layers=12|[config](./configs/ViT_base_patch16_384_ft_in1k_1n8c_dp_fp16o2.yaml)| [download](https://plsc.bj.bcebos.com/models/vit/v2.4/imagenet2012-ViT-B_16-224.pdparams) |
| ViT-L_16_224 |input_size=224，layers=24|[config](./configs/ViT_large_patch16_224_in21k_4n32c_dp_fp16o2.yaml)| [download](https://plsc.bj.bcebos.com/models/vit/v2.4/imagenet21k-ViT-L_16-224.pdparams) |
| ViT-L_16_384 |input_size=384，layers=32|[config](./configs/ViT_large_patch16_384_in1k_ft_4n32c_dp_fp16o2.yaml) | [download](https://plsc.bj.bcebos.com/models/vit/v2.4/imagenet21k-ViT-L_16-224.pdparams) |
| ViT-B_16_224 |input_size=224，layers=12|[config](https://github.com/PaddlePaddle/PLSC/blob/release/2.4/task/classification/vit/configs/ViT_base_patch16_224_in1k_1n8c_dp_fp16o2.yaml) |[download](https://plsc.bj.bcebos.com/models/vit/v2.4/imagenet2012-ViT-B_16-224.pdparams) |
| ViT-B_16_384 |input_size=384，layers=12|[config](https://github.com/PaddlePaddle/PLSC/blob/release/2.4/task/classification/vit/configs/ViT_base_patch16_384_ft_in1k_1n8c_dp_fp16o2.yaml)| [download](https://plsc.bj.bcebos.com/models/vit/v2.4/imagenet2012-ViT-B_16-224.pdparams) |
| ViT-L_16_224 |input_size=224，layers=24|[config](https://github.com/PaddlePaddle/PLSC/blob/release/2.4/task/classification/vit/configs/ViT_large_patch16_224_in21k_4n32c_dp_fp16o2.yaml)| [download](https://plsc.bj.bcebos.com/models/vit/v2.4/imagenet21k-ViT-L_16-224.pdparams) |
| ViT-L_16_384 |input_size=384，layers=32|[config](https://github.com/PaddlePaddle/PLSC/blob/release/2.4/task/classification/vit/configs/ViT_large_patch16_384_in1k_ft_4n32c_dp_fp16o2.yaml) | [download](https://plsc.bj.bcebos.com/models/vit/v2.4/imagenet21k-ViT-L_16-224.pdparams) |
modelcenter/PLSC-ViT/introduction_cn.ipynb
浏览文件 @ dfe7e97c
......@@ -56,14 +56,16 @@
]
},
{
"cell_type": "raw",
"id": "8d291263",
"cell_type": "markdown",
"id": "492fa769-2fe0-4220-b6d9-bbc32f8cca10",
"metadata": {},
"source": [
"```\n",
"git clone https://github.com/PaddlePaddle/PLSC.git\n",
"cd /path/to/PLSC/\n",
"# [optional] pip install -r requirements.txt\n",
"python setup.py develop"
"python setup.py develop\n",
"```"
]
},
{
......@@ -79,15 +81,11 @@
"id": "d68ca5fb",
"metadata": {},
"source": [
"1. 进入任务目录"
]
},
{
"cell_type": "raw",
"id": "59cd341e",
"metadata": {},
"source": [
"cd task/classification/vit\n"
"1. 进入任务目录\n",
"\n",
"```\n",
"cd task/classification/vit\n",
"```"
]
},
{
......@@ -97,20 +95,15 @@
"source": [
"2. 准备数据\n",
"\n",
"将数据整理成以下格式："
]
},
{
"cell_type": "raw",
"id": "ee32eb85",
"metadata": {},
"source": [
"将数据整理成以下格式：\n",
"```text\n",
"dataset/\n",
"└── ILSVRC2012\n",
" ├── train\n",
" ├── val\n",
" ├── train_list.txt\n",
" └── val_list.txt"
" └── val_list.txt\n",
"```"
]
},
{
......@@ -118,15 +111,9 @@
"id": "bea743ea",
"metadata": {},
"source": [
"3. 执行训练命令"
]
},
{
"cell_type": "raw",
"id": "ec8f9946",
"metadata": {},
"source": [
"3. 执行训练命令\n",
"\n",
"```shell\n",
"export PADDLE_NNODES=1\n",
"export PADDLE_MASTER=\"127.0.0.1:12538\"\n",
"export CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7\n",
......@@ -136,15 +123,10 @@
" --master=$PADDLE_MASTER \\\n",
" --devices=$CUDA_VISIBLE_DEVICES \\\n",
" plsc-train \\\n",
" -c ./configs/ViT_base_patch16_224_in1k_1n8c_dp_fp16o2.yaml"
]
},
{
"cell_type": "markdown",
"id": "c4b48418",
"metadata": {},
"source": [
"更多模型的训练教程可参考文档：[ViT训练文档](https://github.com/PaddlePaddle/PLSC/blob/master/task/classification/vit/README.md)\n"
" -c ./configs/ViT_base_patch16_224_in1k_1n8c_dp_fp16o2.yaml\n",
"```\n",
"\n",
"更多模型的训练教程可参考文档：[ViT训练文档](https://github.com/PaddlePaddle/PLSC/blob/master/task/classification/vit/README.md)"
]
},
{
......@@ -160,16 +142,12 @@
"id": "e97c527c",
"metadata": {},
"source": [
"1. 下载预训练模型"
]
},
{
"cell_type": "raw",
"id": "838afffb",
"metadata": {},
"source": [
"1. 下载预训练模型\n",
"\n",
"```shell\n",
"mkdir -p pretrained/vit/ViT_base_patch16_224/\n",
"wget -O ./pretrained/vit/ViT_base_patch16_224/imagenet2012-ViT-B_16-224.pdparams https://plsc.bj.bcebos.com/models/vit/v2.4/imagenet2012-ViT-B_16-224.pdparams"
"wget -O ./pretrained/vit/ViT_base_patch16_224/imagenet2012-ViT-B_16-224.pdparams https://plsc.bj.bcebos.com/models/vit/v2.4/imagenet2012-ViT-B_16-224.pdparams\n",
"```"
]
},
{
......@@ -177,24 +155,11 @@
"id": "a07c6549",
"metadata": {},
"source": [
"2. 导出推理模型\n"
]
},
{
"cell_type": "raw",
"id": "c69963e7",
"metadata": {},
"source": [
"plsc-export -c ./configs/ViT_base_patch16_224_in1k_1n8c_dp_fp16o2.yaml -o Global.pretrained_model=./pretrained/vit/ViT_base_patch16_224/imagenet2012-ViT-B_16-224 -o Model.data_format=NCHW -o FP16.level=O0"
]
},
{
"cell_type": "markdown",
"id": "02a8f5d5",
"metadata": {},
"source": [
"## 4.注意事项\n",
"ImageNet21集没有官方划分的验证集，所以我们使用了所有图像作为训练集。我们为了验证程序运行的可行性构建了一个模拟的验证集。"
"2. 导出推理模型\n",
"\n",
"```shell\n",
"plsc-export -c ./configs/ViT_base_patch16_224_in1k_1n8c_dp_fp16o2.yaml -o Global.pretrained_model=./pretrained/vit/ViT_base_patch16_224/imagenet2012-ViT-B_16-224 -o Model.data_format=NCHW -o FP16.level=O0\n",
"```\n"
]
},
{
......@@ -202,26 +167,28 @@
"id": "d375934d",
"metadata": {},
"source": [
"## 5. 相关论文及引用信息\n"
"## 4. 相关论文及引用信息\n"
]
},
{
"cell_type": "raw",
"id": "ae86fc47",
"cell_type": "markdown",
"id": "29f05b07-d323-45e4-b00d-0728eafb5af7",
"metadata": {},
"source": [
"```text\n",
"@article{dosovitskiy2020,\n",
" title={An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale},\n",
" author={Dosovitskiy, Alexey and Beyer, Lucas and Kolesnikov, Alexander and Weissenborn, Dirk and Zhai, Xiaohua and Unterthiner, Thomas and Dehghani, Mostafa and Minderer, Matthias and Heigold, Georg and Gelly, Sylvain and Uszkoreit, Jakob and Houlsby, Neil},\n",
" journal={arXiv preprint arXiv:2010.11929},\n",
" year={2020}\n",
"}"
"}\n",
"```"
]
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 3 (ipykernel)",
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
......
modelcenter/PLSC-ViT/introduction_en.ipynb
浏览文件 @ dfe7e97c
......@@ -52,18 +52,14 @@
"id": "186a0c17",
"metadata": {},
"source": [
"### 3.1 Install PLSC"
]
},
{
"cell_type": "raw",
"id": "84ebdf32",
"metadata": {},
"source": [
"### 3.1 Install PLSC\n",
"\n",
"```shell\n",
"git clone https://github.com/PaddlePaddle/PLSC.git\n",
"cd /path/to/PLSC/\n",
"# [optional] pip install -r requirements.txt\n",
"python setup.py develop\n"
"python setup.py develop\n",
"```"
]
},
{
......@@ -79,15 +75,11 @@
"id": "a562bf23",
"metadata": {},
"source": [
"1. Enter into the task directory"
]
},
{
"cell_type": "raw",
"id": "e3cc0f53",
"metadata": {},
"source": [
"cd task/classification/vit"
"1. Enter into the task directory\n",
"\n",
"```shell\n",
"cd task/classification/vit\n",
"```"
]
},
{
......@@ -97,20 +89,16 @@
"source": [
"2. Prepare the data\n",
"\n",
"Organize the data into the following format:"
]
},
{
"cell_type": "raw",
"id": "a26cdd37",
"metadata": {},
"source": [
"Organize the data into the following format:\n",
"\n",
"```text\n",
"dataset/\n",
"└── ILSVRC2012\n",
" ├── train\n",
" ├── val\n",
" ├── train_list.txt\n",
" └── val_list.txt"
" └── val_list.txt\n",
"```"
]
},
{
......@@ -118,15 +106,9 @@
"id": "ec78efdf",
"metadata": {},
"source": [
"3. Run the command"
]
},
{
"cell_type": "raw",
"id": "5facf542",
"metadata": {},
"source": [
"3. Run the command\n",
"\n",
"```shell\n",
"export PADDLE_NNODES=1\n",
"export PADDLE_MASTER=\"127.0.0.1:12538\"\n",
"export CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7\n",
......@@ -136,14 +118,9 @@
" --master=$PADDLE_MASTER \\\n",
" --devices=$CUDA_VISIBLE_DEVICES \\\n",
" plsc-train \\\n",
" -c ./configs/ViT_base_patch16_224_in1k_1n8c_dp_fp16o2.yaml"
]
},
{
"cell_type": "markdown",
"id": "9efae4fb",
"metadata": {},
"source": [
" -c ./configs/ViT_base_patch16_224_in1k_1n8c_dp_fp16o2.yaml\n",
"```\n",
"\n",
"More courses about model training can be learned here [ViT readme](https://github.com/PaddlePaddle/PLSC/blob/master/task/classification/vit/README.md)"
]
},
......@@ -160,16 +137,12 @@
"id": "7a3ce1ab",
"metadata": {},
"source": [
"1. Download pretrained model"
]
},
{
"cell_type": "raw",
"id": "2bcf0a3f",
"metadata": {},
"source": [
"1. Download pretrained model\n",
"\n",
"```shell\n",
"mkdir -p pretrained/vit/ViT_base_patch16_224/\n",
"wget -O ./pretrained/vit/ViT_base_patch16_224/imagenet2012-ViT-B_16-224.pdparams https://plsc.bj.bcebos.com/models/vit/v2.4/imagenet2012-ViT-B_16-224.pdparams"
"wget -O ./pretrained/vit/ViT_base_patch16_224/imagenet2012-ViT-B_16-224.pdparams https://plsc.bj.bcebos.com/models/vit/v2.4/imagenet2012-ViT-B_16-224.pdparams\n",
"```"
]
},
{
......@@ -177,24 +150,11 @@
"id": "cff5ac83",
"metadata": {},
"source": [
"2. Export model for inference"
]
},
{
"cell_type": "raw",
"id": "a52873b1",
"metadata": {},
"source": [
"plsc-export -c ./configs/ViT_base_patch16_224_in1k_1n8c_dp_fp16o2.yaml -o Global.pretrained_model=./pretrained/vit/ViT_base_patch16_224/imagenet2012-ViT-B_16-224 -o Model.data_format=NCHW -o FP16.level=O0"
]
},
{
"cell_type": "markdown",
"id": "02a8f5d5",
"metadata": {},
"source": [
"## 4. Attention\n",
"- Since ImageNet21K does not have an officially divided verification set, we use all the images as the training set. We construct the dummy verification set not for parameter adjustment and evaluation, but for the convenience of observing whether the training is ok."
"2. Export model for inference\n",
"\n",
"```shell\n",
"plsc-export -c ./configs/ViT_base_patch16_224_in1k_1n8c_dp_fp16o2.yaml -o Global.pretrained_model=./pretrained/vit/ViT_base_patch16_224/imagenet2012-ViT-B_16-224 -o Model.data_format=NCHW -o FP16.level=O0\n",
"```"
]
},
{
......@@ -202,27 +162,22 @@
"id": "d375934d",
"metadata": {},
"source": [
"## 5. Related papers and citations\n",
"\n"
]
},
{
"cell_type": "raw",
"id": "4fd64e43",
"metadata": {},
"source": [
"## 4. Related papers and citations\n",
"\n",
"```text\n",
"@article{dosovitskiy2020,\n",
" title={An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale},\n",
" author={Dosovitskiy, Alexey and Beyer, Lucas and Kolesnikov, Alexander and Weissenborn, Dirk and Zhai, Xiaohua and Unterthiner, Thomas and Dehghani, Mostafa and Minderer, Matthias and Heigold, Georg and Gelly, Sylvain and Uszkoreit, Jakob and Houlsby, Neil},\n",
" journal={arXiv preprint arXiv:2010.11929},\n",
" year={2020}\n",
"}"
"}\n",
"```"
]
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 3 (ipykernel)",
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
......
Markdown is supported
0% 或 .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑！
想要评论请 注册 或