Skip to content
Snippets Groups Projects
Unverified Commit faf749b2 authored by i-robot's avatar i-robot Committed by Gitee
Browse files

!3567 ONNX Infer：RepVGG Support 

Merge pull request !3567 from wuqingdian/repvgg
parents 57ca2855 d33538f0
No related branches found
No related tags found
No related merge requests found
Hide whitespace changes
Inline Side-by-side
research/cv/repvgg/README_CN.md
+ 19
− 3
View file @ faf749b2
......@@ -76,6 +76,7 @@ RepVGG是由清华大学&旷世科技等提出的一种新颖的CNN设计范式
├──run_distribute_train_ascend.sh // 多卡Ascend910训练脚本
├──run_infer_310.sh // Ascend310推理脚本
├──run_eval_ascend.sh // 测试脚本
├──run_infer_onnx.sh // ONNX推理脚本
├── src
├──configs // RepVGG的配置文件
├──data // 数据集配置文件
......@@ -99,6 +100,7 @@ RepVGG是由清华大学&旷世科技等提出的一种新颖的CNN设计范式
├── export.py // 模型导出文件
├── preprocess.py // 推理数据集与处理文件
├── postprocess.py // 推理精度处理文件
├── infer_onnx.py // 推理onnx文件
```
## 脚本参数
......@@ -111,7 +113,7 @@ RepVGG是由清华大学&旷世科技等提出的一种新颖的CNN设计范式
# Architecture
arch: RepVGG-A0-A0tiny # RepVGG结构选择
# ===== Dataset ===== #
data_url: ./data/imagenet # 数据集地址
data_url: ./dataset # 数据集地址
set: ImageNet # 数据集名字
num_classes: 1000 # 数据集分类数目
mix_up: 0.0 # MixUp数据增强参数
......@@ -173,6 +175,9 @@ RepVGG是由清华大学&旷世科技等提出的一种新颖的CNN设计范式
# 使用脚本启动单卡运行评估示例
bash ./scripts/run_eval_ascend.sh [DEVICE_ID] [CONFIG_PATH] [CHECKPOINT_PATH]
# 使用脚本启动onnx单卡运行评估示例
bash ./scripts/run_infer_onnx.sh [ONNX_PATH] [DATASET_PATH] [DEVICE_TARGET] [DEVICE_ID]
```
对于分布式训练，需要提前创建JSON格式的hccl配置文件。
......@@ -189,13 +194,13 @@ RepVGG是由清华大学&旷世科技等提出的一种新颖的CNN设计范式
python export.py --pretrained [CKPT_FILE] --config [CONFIG_PATH] --device_target [DEVICE_TARGET] --file_format [FILE_FORMAT]
```
导出的模型会以模型的结构名字命名并且保存在当前目录下, 注意: FILE_FORMAT 必须在 ["AIR", "MINDIR"]中选择。
导出的模型会以模型的结构名字命名并且保存在当前目录下, 注意: FILE_FORMAT 必须在 ["AIR", "MINDIR", "ONNX"]中选择。
## 推理过程
### 推理
在进行推理之前我们需要先导出模型。mindir可以在任意环境上导出，air模型只能在昇腾910环境上导出。以下展示了使用mindir模型执行推理的示例。
在进行推理之前我们需要先导出模型。mindir可以在任意环境上导出，air模型只能在昇腾910环境上导出。onnx可以在CPU/GPU/Ascend环境下导出。以下展示了使用mindir模型执行推理的示例。
- 在昇腾310上使用ImageNet-1k数据集进行推理
......@@ -208,6 +213,17 @@ RepVGG是由清华大学&旷世科技等提出的一种新颖的CNN设计范式
Top5 acc: 0.90734
```
- 在GPU/CPU上使用ImageNet-1k数据集进行ONNX推理
推理的结果保存在主目录下，在infer_onnx.log日志文件中可以找到推理结果。
```bash
# onnx inference
bash run_infer_onnx.sh [ONNX_PATH] [DATASET_PATH] [DEVICE_TARGET] [DEVICE_ID]
top-1 accuracy: 0.72024
top-5 accuracy: 0.90394
```
# [模型描述](#目录)
## 性能
......
research/cv/repvgg/infer_onnx.py 0 → 100644
+ 56
− 0
View file @ faf749b2
# Copyright 2022 Huawei Technologies Co., Ltd
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ============================================================================
"""eval"""
import onnxruntime as ort
from mindspore import nn
from src.args import args
from src.data.imagenet import create_dataset_imagenet
def create_session(onnx_path, target_device):
if target_device == 'GPU':
providers = ['CUDAExecutionProvider']
elif target_device == 'CPU':
providers = ['CPUExecutionProvider']
else:
raise ValueError(
f'Unsupported target device {target_device}, '
f'Expected one of: "CPU", "GPU"'
)
session = ort.InferenceSession(onnx_path, providers=providers)
input_name = session.get_inputs()[0].name
return session, input_name
def run_eval(onnx_path, data_dir, target_device):
session, input_name = create_session(onnx_path, target_device)
args.batch_size = 1
dataset = create_dataset_imagenet(data_dir, args, training=False)
metrics = {
'top-1 accuracy': nn.Top1CategoricalAccuracy(),
'top-5 accuracy': nn.Top5CategoricalAccuracy(),
}
for batch in dataset.create_dict_iterator(num_epochs=1, output_numpy=True):
y_pred = session.run(None, {input_name: batch['image']})[0]
for metric in metrics.values():
metric.update(y_pred, batch['label'])
return {name: metric.eval() for name, metric in metrics.items()}
if __name__ == '__main__':
results = run_eval(args.onnx_path, args.dataset_path, args.device_target)
for name, value in results.items():
print(f'{name}: {value:.5f}')
research/cv/repvgg/scripts/run_infer_onnx.sh 0 → 100644
+ 54
− 0
View file @ faf749b2
#!/bin/bash
# Copyright 2022 Huawei Technologies Co., Ltd
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ============================================================================
if [[ $# -lt 2 || $# -gt 4 ]]; then
echo "Usage: bash run_infer_onnx.sh [ONNX_PATH] [DATASET_PATH] [DEVICE_TARGET(optional)] [DEVICE_ID(optional)]"
exit 1
fi
get_real_path(){
if [ "${1:0:1}" == "/" ]; then
echo "$1"
else
echo "$(realpath -m $PWD/$1)"
fi
}
onnx_path=$(get_real_path $1)
dataset_path=$(get_real_path $2)
if [ $# -eq 3 ]; then
device_target=$3
fi
if [ $# -eq 4 ]; then
device_id=$4
fi
echo "onnx_path: "$onnx_path
echo "dataset_path: "$dataset_path
echo "device_target: "$device_target
echo "device_id: "$device_id
function infer()
{
python ./infer_onnx.py --onnx_path=$onnx_path \
--dataset_path=$dataset_path \
--device_target=$device_target \
--device_id=$device_id &> infer_onnx.log
}
infer
if [ $? -ne 0 ]; then
echo " execute inference failed"
exit 1
fi
\ No newline at end of file
research/cv/repvgg/src/args.py
+ 129
− 124
View file @ faf749b2
# Copyright 2022 Huawei Technologies Co., Ltd
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ============================================================================
"""global args for Transformer in Transformer(TNT)"""
import argparse
import ast
import os
import sys
import yaml
from src.configs import parser as _parser
args = None
def parse_arguments():
"""parse_arguments"""
global args
parser = argparse.ArgumentParser(description="MindSpore TNT Training")
parser.add_argument("-a", "--arch", metavar="ARCH", default="ResNet18", help="model architecture")
parser.add_argument("--accumulation_step", default=1, type=int, help="accumulation step")
parser.add_argument("--amp_level", default="O2", choices=["O0", "O1", "O2", "O3"], help="AMP Level")
parser.add_argument("--batch_size", default=256, type=int, metavar="N",
help="mini-batch size (default: 256), this is the total "
"batch size of all Devices on the current node when "
"using Data Parallel or Distributed Data Parallel")
parser.add_argument("--beta", default=[0.9, 0.999], type=lambda x: [float(a) for a in x.split(",")],
help="beta for optimizer")
parser.add_argument("--with_ema", default=False, type=ast.literal_eval, help="training with ema")
parser.add_argument("--ema_decay", default=0.9999, type=float, help="ema decay")
parser.add_argument('--data_url', default="./data", help='location of data.')
parser.add_argument("--device_id", default=0, type=int, help="device id")
parser.add_argument("--device_num", default=1, type=int, help="device num")
parser.add_argument("--device_target", default="Ascend", choices=["GPU", "Ascend"], type=str)
parser.add_argument("--epochs", default=300, type=int, metavar="N", help="number of total epochs to run")
parser.add_argument("--eps", default=1e-8, type=float)
parser.add_argument("--file_format", type=str, choices=["AIR", "MINDIR"], default="MINDIR", help="file format")
parser.add_argument("--in_chans", default=3, type=int)
parser.add_argument("--is_dynamic_loss_scale", default=1, type=int, help="is_dynamic_loss_scale ")
parser.add_argument("--keep_checkpoint_max", default=20, type=int, help="keep checkpoint max num")
parser.add_argument("--optimizer", help="Which optimizer to use", default="sgd")
parser.add_argument("--set", help="name of dataset", type=str, default="ImageNet")
parser.add_argument("--mix_up", default=0., type=float, help="mix up")
parser.add_argument("--mlp_ratio", help="mlp ", default=4., type=float)
parser.add_argument("-j", "--num_parallel_workers", default=20, type=int, metavar="N",
help="number of data loading workers (default: 20)")
parser.add_argument("--start_epoch", default=0, type=int, metavar="N",
help="manual epoch number (useful on restarts)")
parser.add_argument("--warmup_length", default=0, type=int, help="number of warmup iterations")
parser.add_argument("--warmup_lr", default=5e-7, type=float, help="warm up learning rate")
parser.add_argument("--wd", "--weight_decay", default=0.05, type=float, metavar="W",
help="weight decay (default: 0.05)", dest="weight_decay")
parser.add_argument("--loss_scale", default=1024, type=int, help="loss_scale")
parser.add_argument("--lr", "--learning_rate", default=5e-4, type=float, help="initial lr", dest="lr")
parser.add_argument("--lr_scheduler", default="cosine_annealing", help="schedule for the learning rate.")
parser.add_argument("--lr_adjust", default=30, type=float, help="interval to drop lr")
parser.add_argument("--lr_gamma", default=0.97, type=int, help="multistep multiplier")
parser.add_argument("--momentum", default=0.9, type=float, metavar="M", help="momentum")
parser.add_argument("--num_classes", default=1000, type=int)
parser.add_argument("--pretrained", dest="pretrained", default=None, type=str, help="use pre-trained model")
parser.add_argument("--config", help="Config file to use (see configs dir)", default=None, required=True)
parser.add_argument("--seed", default=0, type=int, help="seed for initializing training. ")
parser.add_argument("--save_every", default=2, type=int, help="save every ___ epochs(default:2)")
parser.add_argument("--label_smoothing", type=float, help="label smoothing to use, default 0.1", default=0.1)
parser.add_argument("--image_size", default=224, help="image Size.", type=int)
parser.add_argument('--train_url', default="./", help='location of training outputs.')
parser.add_argument("--run_modelarts", type=ast.literal_eval, default=False, help="whether run on modelarts")
parser.add_argument("--deploy", type=ast.literal_eval, default=False, help="whether run deploy")
args = parser.parse_args()
get_config()
def get_config():
"""get_config"""
global args
override_args = _parser.argv_to_vars(sys.argv)
# load yaml file
if args.run_modelarts:
import moxing as mox
if not args.config.startswith("obs:/"):
args.config = "obs:/" + args.config
with mox.file.File(args.config, 'r') as f:
yaml_txt = f.read()
else:
yaml_txt = open(args.config).read()
# override args
loaded_yaml = yaml.load(yaml_txt, Loader=yaml.FullLoader)
for v in override_args:
loaded_yaml[v] = getattr(args, v)
print(f"=> Reading YAML config from {args.config}")
args.__dict__.update(loaded_yaml)
print(args)
os.environ["DEVICE_TARGET"] = args.device_target
if "DEVICE_NUM" not in os.environ.keys():
os.environ["DEVICE_NUM"] = str(args.device_num)
if "RANK_SIZE" not in os.environ.keys():
os.environ["RANK_SIZE"] = str(args.device_num)
def run_args():
"""run and get args"""
global args
if args is None:
parse_arguments()
run_args()
# Copyright 2022 Huawei Technologies Co., Ltd
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ============================================================================
"""global args for Transformer in Transformer(TNT)"""
import argparse
import ast
import os
import sys
import yaml
from src.configs import parser as _parser
args = None
def parse_arguments():
"""parse_arguments"""
global args
parser = argparse.ArgumentParser(description="MindSpore TNT Training")
parser.add_argument("-a", "--arch", metavar="ARCH", default="ResNet18", help="model architecture")
parser.add_argument("--accumulation_step", default=1, type=int, help="accumulation step")
parser.add_argument("--amp_level", default="O0", choices=["O0", "O1", "O2", "O3"], help="AMP Level")
parser.add_argument("--batch_size", default=256, type=int, metavar="N",
help="mini-batch size (default: 256), this is the total "
"batch size of all Devices on the current node when "
"using Data Parallel or Distributed Data Parallel")
parser.add_argument("--beta", default=[0.9, 0.999], type=lambda x: [float(a) for a in x.split(",")],
help="beta for optimizer")
parser.add_argument("--with_ema", default=False, type=ast.literal_eval, help="training with ema")
parser.add_argument("--ema_decay", default=0.9999, type=float, help="ema decay")
parser.add_argument('--data_url', default="./data", help='location of data.')
parser.add_argument("--device_id", default=0, type=int, help="device id")
parser.add_argument("--device_num", default=1, type=int, help="device num")
parser.add_argument("--device_target", default="Ascend", choices=["GPU", "Ascend"], type=str)
parser.add_argument("--epochs", default=300, type=int, metavar="N", help="number of total epochs to run")
parser.add_argument("--eps", default=1e-8, type=float)
parser.add_argument("--file_format", type=str, choices=["AIR", "MINDIR", "ONNX"], default="MINDIR",
help="file format")
parser.add_argument("--in_chans", default=3, type=int)
parser.add_argument("--is_dynamic_loss_scale", default=1, type=int, help="is_dynamic_loss_scale ")
parser.add_argument("--keep_checkpoint_max", default=20, type=int, help="keep checkpoint max num")
parser.add_argument("--optimizer", help="Which optimizer to use", default="sgd")
parser.add_argument("--set", help="name of dataset", type=str, default="ImageNet")
parser.add_argument("--mix_up", default=0., type=float, help="mix up")
parser.add_argument("--mlp_ratio", help="mlp ", default=4., type=float)
parser.add_argument("-j", "--num_parallel_workers", default=20, type=int, metavar="N",
help="number of data loading workers (default: 20)")
parser.add_argument("--start_epoch", default=0, type=int, metavar="N",
help="manual epoch number (useful on restarts)")
parser.add_argument("--warmup_length", default=0, type=int, help="number of warmup iterations")
parser.add_argument("--warmup_lr", default=5e-7, type=float, help="warm up learning rate")
parser.add_argument("--wd", "--weight_decay", default=0.05, type=float, metavar="W",
help="weight decay (default: 0.05)", dest="weight_decay")
parser.add_argument("--loss_scale", default=1024, type=int, help="loss_scale")
parser.add_argument("--lr", "--learning_rate", default=5e-4, type=float, help="initial lr", dest="lr")
parser.add_argument("--lr_scheduler", default="cosine_annealing", help="schedule for the learning rate.")
parser.add_argument("--lr_adjust", default=30, type=float, help="interval to drop lr")
parser.add_argument("--lr_gamma", default=0.97, type=int, help="multistep multiplier")
parser.add_argument("--momentum", default=0.9, type=float, metavar="M", help="momentum")
parser.add_argument("--num_classes", default=1000, type=int)
parser.add_argument("--pretrained", dest="pretrained", default=None, type=str,
help="use pre-trained model")
parser.add_argument("--config", help="Config file to use (see configs dir)",
default="./src/configs/RepVGG-A0.yaml", required=False)
parser.add_argument("--seed", default=0, type=int, help="seed for initializing training. ")
parser.add_argument("--save_every", default=2, type=int, help="save every ___ epochs(default:2)")
parser.add_argument("--label_smoothing", type=float, help="label smoothing to use, default 0.1", default=0.1)
parser.add_argument("--image_size", default=224, help="image Size.", type=int)
parser.add_argument('--train_url', default="./", help='location of training outputs.')
parser.add_argument("--run_modelarts", type=ast.literal_eval, default=False, help="whether run on modelarts")
parser.add_argument("--deploy", type=ast.literal_eval, default=False, help="whether run deploy")
parser.add_argument("--onnx_path", type=str, default=None, help="ONNX file path")
parser.add_argument("--dataset_path", type=str, default=None, help="Dataset path")
args = parser.parse_args()
get_config()
def get_config():
"""get_config"""
global args
override_args = _parser.argv_to_vars(sys.argv)
# load yaml file
if args.run_modelarts:
import moxing as mox
if not args.config.startswith("obs:/"):
args.config = "obs:/" + args.config
with mox.file.File(args.config, 'r') as f:
yaml_txt = f.read()
else:
yaml_txt = open(args.config).read()
# override args
loaded_yaml = yaml.load(yaml_txt, Loader=yaml.FullLoader)
for v in override_args:
loaded_yaml[v] = getattr(args, v)
print(f"=> Reading YAML config from {args.config}")
args.__dict__.update(loaded_yaml)
print(args)
os.environ["DEVICE_TARGET"] = args.device_target
if "DEVICE_NUM" not in os.environ.keys():
os.environ["DEVICE_NUM"] = str(args.device_num)
if "RANK_SIZE" not in os.environ.keys():
os.environ["RANK_SIZE"] = str(args.device_num)
def run_args():
"""run and get args"""
global args
if args is None:
parse_arguments()
run_args()
research/cv/repvgg/src/configs/RepVGG-A0.yaml
+ 1
− 1
View file @ faf749b2
......@@ -45,4 +45,4 @@ ema_decay: 0.9999
# ===== Hardware setup ===== #
num_parallel_workers: 16
device_target: Ascend
\ No newline at end of file
device_target: Ascend
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment