Skip to content

Y
yolov5
Unverified 提交 e18e6811 authored 作者: Alex Stoken's avatar Alex Stoken 提交者： GitHub
浏览文件
操作

Merge branch 'master' into advanced_logging

上级 de191655 77fb8ee0
全部展开 隐藏空白字符变更
内嵌 并排
正在显示 包含 353 行增加 和 200 行删除
.github/ISSUE_TEMPLATE/--bug-report.md
浏览文件 @ e18e6811
......@@ -7,29 +7,43 @@ assignees: ''
---
Before submitting a bug report, please ensure that you are using the latest versions of:
- Python
- PyTorch
- This repository (run `git fetch && git status -uno` to check and `git pull` to update)
Before submitting a bug report, please be aware that your issue **must be reproducible** with all of the following, otherwise it is non-actionable, and we can not help you:
- **Current repo**: run `git fetch && git status -uno` to check and `git pull` to update repo
- **Common dataset**: coco.yaml or coco128.yaml
- **Common environment**: Colab, Google Cloud, or Docker image. See https://github.com/ultralytics/yolov5#reproduce-our-environment
**Your issue must be reproducible on a public dataset (i.e COCO) using the latest version of the repository, and you must supply code to reproduce, or we can not help you.**
If this is a custom training question we suggest you include your `train*.jpg`, `test*.jpg` and `results.png` figures.
If this is a custom dataset/training question you **must include** your `train*.jpg`, `test*.jpg` and `results.png` figures, or we can not help you. You can generate these with `utils.plot_results()`.
## 🐛 Bug
A clear and concise description of what the bug is.
## To Reproduce
**REQUIRED**: Code to reproduce your issue below
## To Reproduce (REQUIRED)
Input:
```
import torch
a = torch.tensor([5])
c = a / 0
```
Output:
```
python train.py ...
Traceback (most recent call last):
File "/Users/glennjocher/opt/anaconda3/envs/env1/lib/python3.7/site-packages/IPython/core/interactiveshell.py", line 3331, in run_code
exec(code_obj, self.user_global_ns, self.user_ns)
File "<ipython-input-5-be04c762b799>", line 5, in <module>
c = a / 0
RuntimeError: ZeroDivisionError
```
## Expected behavior
A clear and concise description of what you expected to happen.
## Environment
If applicable, add screenshots to help explain your problem.
......
Dockerfile
浏览文件 @ e18e6811
# Start FROM Nvidia PyTorch image https://ngc.nvidia.com/catalog/containers/nvidia:pytorch
FROM nvcr.io/nvidia/pytorch:20.03-py3
# Install dependencies (pip or conda)
RUN pip install -U gsutil
# RUN pip install -U -r requirements.txt
# Create working directory
RUN mkdir -p /usr/src/app
......@@ -12,6 +9,9 @@ WORKDIR /usr/src/app
# Copy contents
COPY . /usr/src/app
# Install dependencies (pip or conda)
#RUN pip install -r requirements.txt
# Copy weights
#RUN python3 -c "from models import *; \
#attempt_download('weights/yolov5s.pt'); \
......@@ -41,7 +41,7 @@ COPY . /usr/src/app
# Bash into running container
# sudo docker container exec -it ba65811811ab bash
# python -c "from utils.utils import *; create_backbone('weights/last.pt')" && gsutil cp weights/backbone.pt gs://*
# python -c "from utils.utils import *; create_pretrained('weights/last.pt')" && gsutil cp weights/pretrained.pt gs://*
# Bash into stopped container
# sudo docker commit 6d525e299258 user/test_image && sudo docker run -it --gpus all --ipc=host -v "$(pwd)"/coco:/usr/src/coco --entrypoint=sh user/test_image
......
README.md
浏览文件 @ e18e6811
差异被折叠。
detect.py
浏览文件 @ e18e6811
import argparse
import torch.backends.cudnn as cudnn
from utils import google_utils
from utils.datasets import *
from utils.utils import *
......@@ -36,14 +39,14 @@ def detect(save_img=False):
vid_path, vid_writer = None, None
if webcam:
view_img = True
torch.backends.cudnn.benchmark = True # set True to speed up constant image size inference
cudnn.benchmark = True # set True to speed up constant image size inference
dataset = LoadStreams(source, img_size=imgsz)
else:
save_img = True
dataset = LoadImages(source, img_size=imgsz)
# Get names and colors
names = model.names if hasattr(model, 'names') else model.modules.names
names = model.module.names if hasattr(model, 'module') else model.names
colors = [[random.randint(0, 255) for _ in range(3)] for _ in range(len(names))]
# Run inference
......@@ -62,8 +65,7 @@ def detect(save_img=False):
pred = model(img, augment=opt.augment)[0]
# Apply NMS
pred = non_max_suppression(pred, opt.conf_thres, opt.iou_thres,
fast=True, classes=opt.classes, agnostic=opt.agnostic_nms)
pred = non_max_suppression(pred, opt.conf_thres, opt.iou_thres, classes=opt.classes, agnostic=opt.agnostic_nms)
t2 = torch_utils.time_synchronized()
# Apply Classifier
......@@ -78,6 +80,7 @@ def detect(save_img=False):
p, s, im0 = path, '', im0s
save_path = str(Path(out) / Path(p).name)
txt_path = str(Path(out) / Path(p).stem) + ('_%g' % dataset.frame if dataset.mode == 'video' else '')
s += '%gx%g ' % img.shape[2:] # print string
gn = torch.tensor(im0.shape)[[1, 0, 1, 0]] #  normalization gain whwh
if det is not None and len(det):
......@@ -93,8 +96,8 @@ def detect(save_img=False):
for *xyxy, conf, cls in det:
if save_txt: # Write to file
xywh = (xyxy2xywh(torch.tensor(xyxy).view(1, 4)) / gn).view(-1).tolist() # normalized xywh
with open(save_path[:save_path.rfind('.')] + '.txt', 'a') as file:
file.write(('%g ' * 5 + '\n') % (cls, *xywh)) # label format
with open(txt_path + '.txt', 'a') as f:
f.write(('%g ' * 5 + '\n') % (cls, *xywh)) # label format
if save_img or view_img: # Add bbox to image
label = '%s %.2f' % (names[int(cls)], conf)
......@@ -154,3 +157,8 @@ if __name__ == '__main__':
with torch.no_grad():
detect()
# Update all models
# for opt.weights in ['yolov5s.pt', 'yolov5m.pt', 'yolov5l.pt', 'yolov5x.pt', 'yolov3-spp.pt']:
# detect()
# create_pretrained(opt.weights, opt.weights)
models/common.py
浏览文件 @ e18e6811
......@@ -13,7 +13,8 @@ class Conv(nn.Module):
# Standard convolution
def __init__(self, c1, c2, k=1, s=1, g=1, act=True): # ch_in, ch_out, kernel, stride, groups
super(Conv, self).__init__()
self.conv = nn.Conv2d(c1, c2, k, s, k // 2, groups=g, bias=False)
p = k // 2 if isinstance(k, int) else [x // 2 for x in k] # padding
self.conv = nn.Conv2d(c1, c2, k, s, p, groups=g, bias=False)
self.bn = nn.BatchNorm2d(c2)
self.act = nn.LeakyReLU(0.1, inplace=True) if act else nn.Identity()
......
models/yolov3-spp.yaml → models/hub/yolov3-spp.yaml
浏览文件 @ e18e6811
......@@ -25,8 +25,7 @@ backbone:
[-1, 4, Bottleneck, [1024]], # 10
]
# yolov3-spp head
# na = len(anchors[0])
# YOLOv3-SPP head
head:
[[-1, 1, Bottleneck, [1024, False]], # 11
[-1, 1, SPP, [512, [5, 9, 13]]],
......
models/hub/yolov5-fpn.yaml 0 → 100644
浏览文件 @ e18e6811
# parameters
nc: 80 # number of classes
depth_multiple: 1.0 # model depth multiple
width_multiple: 1.0 # layer channel multiple
# anchors
anchors:
- [10,13, 16,30, 33,23] # P3/8
- [30,61, 62,45, 59,119] # P4/16
- [116,90, 156,198, 373,326] # P5/32
# YOLOv5 backbone
backbone:
# [from, number, module, args]
[[-1, 1, Focus, [64, 3]], # 0-P1/2
[-1, 1, Conv, [128, 3, 2]], # 1-P2/4
[-1, 3, Bottleneck, [128]],
[-1, 1, Conv, [256, 3, 2]], # 3-P3/8
[-1, 9, BottleneckCSP, [256]],
[-1, 1, Conv, [512, 3, 2]], # 5-P4/16
[-1, 9, BottleneckCSP, [512]],
[-1, 1, Conv, [1024, 3, 2]], # 7-P5/32
[-1, 1, SPP, [1024, [5, 9, 13]]],
[-1, 6, BottleneckCSP, [1024]], # 9
]
# YOLOv5 FPN head
head:
[[-1, 3, BottleneckCSP, [1024, False]],
[-1, 1, nn.Conv2d, [na * (nc + 5), 1, 1]], # 11 (P5/32-large)
[-2, 1, nn.Upsample, [None, 2, 'nearest']],
[[-1, 6], 1, Concat, [1]], # cat backbone P4
[-1, 1, Conv, [512, 1, 1]],
[-1, 3, BottleneckCSP, [512, False]],
[-1, 1, nn.Conv2d, [na * (nc + 5), 1, 1]], # 16 (P4/16-medium)
[-2, 1, nn.Upsample, [None, 2, 'nearest']],
[[-1, 4], 1, Concat, [1]], # cat backbone P3
[-1, 1, Conv, [256, 1, 1]],
[-1, 3, BottleneckCSP, [256, False]],
[-1, 1, nn.Conv2d, [na * (nc + 5), 1, 1]], # 21 (P3/8-small)
[[], 1, Detect, [nc, anchors]], # Detect(P3, P4, P5)
]
models/hub/yolov5-panet.yaml 0 → 100644
浏览文件 @ e18e6811
# parameters
nc: 80 # number of classes
depth_multiple: 1.0 # model depth multiple
width_multiple: 1.0 # layer channel multiple
# anchors
anchors:
- [116,90, 156,198, 373,326] # P5/32
- [30,61, 62,45, 59,119] # P4/16
- [10,13, 16,30, 33,23] # P3/8
# YOLOv5 backbone
backbone:
# [from, number, module, args]
[[-1, 1, Focus, [64, 3]], # 0-P1/2
[-1, 1, Conv, [128, 3, 2]], # 1-P2/4
[-1, 3, BottleneckCSP, [128]],
[-1, 1, Conv, [256, 3, 2]], # 3-P3/8
[-1, 9, BottleneckCSP, [256]],
[-1, 1, Conv, [512, 3, 2]], # 5-P4/16
[-1, 9, BottleneckCSP, [512]],
[-1, 1, Conv, [1024, 3, 2]], # 7-P5/32
[-1, 1, SPP, [1024, [5, 9, 13]]],
]
# YOLOv5 PANet head
head:
[[-1, 3, BottleneckCSP, [1024, False]],
[-1, 1, Conv, [512, 1, 1]], # 10
[-1, 1, nn.Upsample, [None, 2, 'nearest']],
[[-1, 6], 1, Concat, [1]], # cat backbone P4
[-1, 3, BottleneckCSP, [512, False]],
[-1, 1, Conv, [256, 1, 1]], # 14
[-1, 1, nn.Upsample, [None, 2, 'nearest']],
[[-1, 4], 1, Concat, [1]], # cat backbone P3
[-1, 3, BottleneckCSP, [256, False]],
[-1, 1, nn.Conv2d, [na * (nc + 5), 1, 1]], # 18 (P3/8-small)
[-2, 1, Conv, [256, 3, 2]],
[[-1, 14], 1, Concat, [1]], # cat head P4
[-1, 3, BottleneckCSP, [512, False]],
[-1, 1, nn.Conv2d, [na * (nc + 5), 1, 1]], # 22 (P4/16-medium)
[-2, 1, Conv, [512, 3, 2]],
[[-1, 10], 1, Concat, [1]], # cat head P5
[-1, 3, BottleneckCSP, [1024, False]],
[-1, 1, nn.Conv2d, [na * (nc + 5), 1, 1]], # 26 (P5/32-large)
[[], 1, Detect, [nc, anchors]], # Detect(P5, P4, P3)
]
models/onnx_export.py
浏览文件 @ e18e6811
"""Exports a pytorch *.pt model to *.onnx format
Usage:
import torch
$ export PYTHONPATH="$PWD" && python models/onnx_export.py --weights ./weights/yolov5s.pt --img 640 --batch 1
"""
......@@ -10,6 +9,7 @@ import argparse
import onnx
from models.common import *
from utils import google_utils
if __name__ == '__main__':
parser = argparse.ArgumentParser()
......@@ -25,7 +25,7 @@ if __name__ == '__main__':
# Load pytorch model
google_utils.attempt_download(opt.weights)
model = torch.load(opt.weights, map_location=torch.device('cpu'))['model']
model = torch.load(opt.weights, map_location=torch.device('cpu'))['model'].float()
model.eval()
model.fuse()
......
models/yolo.py
浏览文件 @ e18e6811
import argparse
import yaml
from models.experimental import *
......@@ -61,8 +59,9 @@ class Model(nn.Module):
# Build strides, anchors
m = self.model[-1] # Detect()
m.stride = torch.tensor([64 / x.shape[-2] for x in self.forward(torch.zeros(1, ch, 64, 64))]) # forward
m.stride = torch.tensor([128 / x.shape[-2] for x in self.forward(torch.zeros(1, ch, 128, 128))]) # forward
m.anchors /= m.stride.view(-1, 1, 1)
check_anchor_order(m)
self.stride = m.stride
# Init weights, biases
......@@ -97,8 +96,11 @@ class Model(nn.Module):
x = y[m.f] if isinstance(m.f, int) else [x if j == -1 else y[j] for j in m.f] # from earlier layers
if profile:
import thop
o = thop.profile(m, inputs=(x,), verbose=False)[0] / 1E9 * 2 # FLOPS
try:
import thop
o = thop.profile(m, inputs=(x,), verbose=False)[0] / 1E9 * 2 # FLOPS
except:
o = 0
t = torch_utils.time_synchronized()
for _ in range(10):
_ = m(x)
......@@ -208,7 +210,7 @@ if __name__ == '__main__':
parser.add_argument('--cfg', type=str, default='yolov5s.yaml', help='model.yaml')
parser.add_argument('--device', default='', help='cuda device, i.e. 0 or 0,1,2,3 or cpu')
opt = parser.parse_args()
opt.cfg = glob.glob('./**/' + opt.cfg, recursive=True)[0] # find file
opt.cfg = check_file(opt.cfg) # check file
device = torch_utils.select_device(opt.device)
# Create model
......@@ -218,11 +220,10 @@ if __name__ == '__main__':
# Profile
# img = torch.rand(8 if torch.cuda.is_available() else 1, 3, 640, 640).to(device)
# y = model(img, profile=True)
# print([y[0].shape] + [x.shape for x in y[1]])
# ONNX export
# model.model[-1].export = True
# torch.onnx.export(model, img, f.replace('.yaml', '.onnx'), verbose=True, opset_version=11)
# torch.onnx.export(model, img, opt.cfg.replace('.yaml', '.onnx'), verbose=True, opset_version=11)
# Tensorboard
# from torch.utils.tensorboard import SummaryWriter
......
models/yolov5l.yaml
浏览文件 @ e18e6811
......@@ -5,41 +5,48 @@ width_multiple: 1.0 # layer channel multiple
# anchors
anchors:
- [10,13, 16,30, 33,23] # P3/8
- [30,61, 62,45, 59,119] # P4/16
- [116,90, 156,198, 373,326] # P5/32
- [30,61, 62,45, 59,119] # P4/16
- [10,13, 16,30, 33,23] # P3/8
# yolov5 backbone
# YOLOv5 backbone
backbone:
# [from, number, module, args]
[[-1, 1, Focus, [64, 3]], # 1-P1/2
[-1, 1, Conv, [128, 3, 2]], # 2-P2/4
[-1, 3, Bottleneck, [128]],
[-1, 1, Conv, [256, 3, 2]], # 4-P3/8
[[-1, 1, Focus, [64, 3]], # 0-P1/2
[-1, 1, Conv, [128, 3, 2]], # 1-P2/4
[-1, 3, BottleneckCSP, [128]],
[-1, 1, Conv, [256, 3, 2]], # 3-P3/8
[-1, 9, BottleneckCSP, [256]],
[-1, 1, Conv, [512, 3, 2]], # 6-P4/16
[-1, 1, Conv, [512, 3, 2]], # 5-P4/16
[-1, 9, BottleneckCSP, [512]],
[-1, 1, Conv, [1024, 3, 2]], # 8-P5/32
[-1, 1, Conv, [1024, 3, 2]], # 7-P5/32
[-1, 1, SPP, [1024, [5, 9, 13]]],
[-1, 6, BottleneckCSP, [1024]], # 10
]
# yolov5 head
# YOLOv5 head
head:
[[-1, 3, BottleneckCSP, [1024, False]], # 11
[-1, 1, nn.Conv2d, [na * (nc + 5), 1, 1]], # 12 (P5/32-large)
[[-1, 3, BottleneckCSP, [1024, False]], # 9
[-2, 1, nn.Upsample, [None, 2, 'nearest']],
[[-1, 6], 1, Concat, [1]], # cat backbone P4
[-1, 1, Conv, [512, 1, 1]],
[-1, 3, BottleneckCSP, [512, False]],
[-1, 1, nn.Conv2d, [na * (nc + 5), 1, 1]], # 17 (P4/16-medium)
[-1, 1, nn.Upsample, [None, 2, 'nearest']],
[[-1, 6], 1, Concat, [1]], # cat backbone P4
[-1, 3, BottleneckCSP, [512, False]], # 13
[-2, 1, nn.Upsample, [None, 2, 'nearest']],
[[-1, 4], 1, Concat, [1]], # cat backbone P3
[-1, 1, Conv, [256, 1, 1]],
[-1, 1, nn.Upsample, [None, 2, 'nearest']],
[[-1, 4], 1, Concat, [1]], # cat backbone P3
[-1, 3, BottleneckCSP, [256, False]],
[-1, 1, nn.Conv2d, [na * (nc + 5), 1, 1]], # 22 (P3/8-small)
[-1, 1, nn.Conv2d, [na * (nc + 5), 1, 1]], # 18 (P3/8-small)
[-2, 1, Conv, [256, 3, 2]],
[[-1, 14], 1, Concat, [1]], # cat head P4
[-1, 3, BottleneckCSP, [512, False]],
[-1, 1, nn.Conv2d, [na * (nc + 5), 1, 1]], # 22 (P4/16-medium)
[-2, 1, Conv, [512, 3, 2]],
[[-1, 10], 1, Concat, [1]], # cat head P5
[-1, 3, BottleneckCSP, [1024, False]],
[-1, 1, nn.Conv2d, [na * (nc + 5), 1, 1]], # 26 (P5/32-large)
[[], 1, Detect, [nc, anchors]], # Detect(P3, P4, P5)
[[], 1, Detect, [nc, anchors]], # Detect(P5, P4, P3)
]
models/yolov5m.yaml
浏览文件 @ e18e6811
......@@ -5,41 +5,48 @@ width_multiple: 0.75 # layer channel multiple
# anchors
anchors:
- [10,13, 16,30, 33,23] # P3/8
- [30,61, 62,45, 59,119] # P4/16
- [116,90, 156,198, 373,326] # P5/32
- [30,61, 62,45, 59,119] # P4/16
- [10,13, 16,30, 33,23] # P3/8
# yolov5 backbone
# YOLOv5 backbone
backbone:
# [from, number, module, args]
[[-1, 1, Focus, [64, 3]], # 1-P1/2
[-1, 1, Conv, [128, 3, 2]], # 2-P2/4
[-1, 3, Bottleneck, [128]],
[-1, 1, Conv, [256, 3, 2]], # 4-P3/8
[[-1, 1, Focus, [64, 3]], # 0-P1/2
[-1, 1, Conv, [128, 3, 2]], # 1-P2/4
[-1, 3, BottleneckCSP, [128]],
[-1, 1, Conv, [256, 3, 2]], # 3-P3/8
[-1, 9, BottleneckCSP, [256]],
[-1, 1, Conv, [512, 3, 2]], # 6-P4/16
[-1, 1, Conv, [512, 3, 2]], # 5-P4/16
[-1, 9, BottleneckCSP, [512]],
[-1, 1, Conv, [1024, 3, 2]], # 8-P5/32
[-1, 1, Conv, [1024, 3, 2]], # 7-P5/32
[-1, 1, SPP, [1024, [5, 9, 13]]],
[-1, 6, BottleneckCSP, [1024]], # 10
]
# yolov5 head
# YOLOv5 head
head:
[[-1, 3, BottleneckCSP, [1024, False]], # 11
[-1, 1, nn.Conv2d, [na * (nc + 5), 1, 1]], # 12 (P5/32-large)
[[-1, 3, BottleneckCSP, [1024, False]], # 9
[-2, 1, nn.Upsample, [None, 2, 'nearest']],
[[-1, 6], 1, Concat, [1]], # cat backbone P4
[-1, 1, Conv, [512, 1, 1]],
[-1, 3, BottleneckCSP, [512, False]],
[-1, 1, nn.Conv2d, [na * (nc + 5), 1, 1]], # 17 (P4/16-medium)
[-1, 1, nn.Upsample, [None, 2, 'nearest']],
[[-1, 6], 1, Concat, [1]], # cat backbone P4
[-1, 3, BottleneckCSP, [512, False]], # 13
[-2, 1, nn.Upsample, [None, 2, 'nearest']],
[[-1, 4], 1, Concat, [1]], # cat backbone P3
[-1, 1, Conv, [256, 1, 1]],
[-1, 1, nn.Upsample, [None, 2, 'nearest']],
[[-1, 4], 1, Concat, [1]], # cat backbone P3
[-1, 3, BottleneckCSP, [256, False]],
[-1, 1, nn.Conv2d, [na * (nc + 5), 1, 1]], # 22 (P3/8-small)
[-1, 1, nn.Conv2d, [na * (nc + 5), 1, 1]], # 18 (P3/8-small)
[-2, 1, Conv, [256, 3, 2]],
[[-1, 14], 1, Concat, [1]], # cat head P4
[-1, 3, BottleneckCSP, [512, False]],
[-1, 1, nn.Conv2d, [na * (nc + 5), 1, 1]], # 22 (P4/16-medium)
[-2, 1, Conv, [512, 3, 2]],
[[-1, 10], 1, Concat, [1]], # cat head P5
[-1, 3, BottleneckCSP, [1024, False]],
[-1, 1, nn.Conv2d, [na * (nc + 5), 1, 1]], # 26 (P5/32-large)
[[], 1, Detect, [nc, anchors]], # Detect(P3, P4, P5)
[[], 1, Detect, [nc, anchors]], # Detect(P5, P4, P3)
]
models/yolov5s.yaml
浏览文件 @ e18e6811
......@@ -5,41 +5,48 @@ width_multiple: 0.50 # layer channel multiple
# anchors
anchors:
- [10,13, 16,30, 33,23] # P3/8
- [30,61, 62,45, 59,119] # P4/16
- [116,90, 156,198, 373,326] # P5/32
- [30,61, 62,45, 59,119] # P4/16
- [10,13, 16,30, 33,23] # P3/8
# yolov5 backbone
# YOLOv5 backbone
backbone:
# [from, number, module, args]
[[-1, 1, Focus, [64, 3]], # 1-P1/2
[-1, 1, Conv, [128, 3, 2]], # 2-P2/4
[-1, 3, Bottleneck, [128]],
[-1, 1, Conv, [256, 3, 2]], # 4-P3/8
[[-1, 1, Focus, [64, 3]], # 0-P1/2
[-1, 1, Conv, [128, 3, 2]], # 1-P2/4
[-1, 3, BottleneckCSP, [128]],
[-1, 1, Conv, [256, 3, 2]], # 3-P3/8
[-1, 9, BottleneckCSP, [256]],
[-1, 1, Conv, [512, 3, 2]], # 6-P4/16
[-1, 1, Conv, [512, 3, 2]], # 5-P4/16
[-1, 9, BottleneckCSP, [512]],
[-1, 1, Conv, [1024, 3, 2]], # 8-P5/32
[-1, 1, Conv, [1024, 3, 2]], # 7-P5/32
[-1, 1, SPP, [1024, [5, 9, 13]]],
[-1, 6, BottleneckCSP, [1024]], # 10
]
# yolov5 head
# YOLOv5 head
head:
[[-1, 3, BottleneckCSP, [1024, False]], # 11
[-1, 1, nn.Conv2d, [na * (nc + 5), 1, 1]], # 12 (P5/32-large)
[[-1, 3, BottleneckCSP, [1024, False]], # 9
[-2, 1, nn.Upsample, [None, 2, 'nearest']],
[[-1, 6], 1, Concat, [1]], # cat backbone P4
[-1, 1, Conv, [512, 1, 1]],
[-1, 3, BottleneckCSP, [512, False]],
[-1, 1, nn.Conv2d, [na * (nc + 5), 1, 1]], # 17 (P4/16-medium)
[-1, 1, nn.Upsample, [None, 2, 'nearest']],
[[-1, 6], 1, Concat, [1]], # cat backbone P4
[-1, 3, BottleneckCSP, [512, False]], # 13
[-2, 1, nn.Upsample, [None, 2, 'nearest']],
[[-1, 4], 1, Concat, [1]], # cat backbone P3
[-1, 1, Conv, [256, 1, 1]],
[-1, 1, nn.Upsample, [None, 2, 'nearest']],
[[-1, 4], 1, Concat, [1]], # cat backbone P3
[-1, 3, BottleneckCSP, [256, False]],
[-1, 1, nn.Conv2d, [na * (nc + 5), 1, 1]], # 22 (P3/8-small)
[-1, 1, nn.Conv2d, [na * (nc + 5), 1, 1]], # 18 (P3/8-small)
[-2, 1, Conv, [256, 3, 2]],
[[-1, 14], 1, Concat, [1]], # cat head P4
[-1, 3, BottleneckCSP, [512, False]],
[-1, 1, nn.Conv2d, [na * (nc + 5), 1, 1]], # 22 (P4/16-medium)
[-2, 1, Conv, [512, 3, 2]],
[[-1, 10], 1, Concat, [1]], # cat head P5
[-1, 3, BottleneckCSP, [1024, False]],
[-1, 1, nn.Conv2d, [na * (nc + 5), 1, 1]], # 26 (P5/32-large)
[[], 1, Detect, [nc, anchors]], # Detect(P3, P4, P5)
[[], 1, Detect, [nc, anchors]], # Detect(P5, P4, P3)
]
models/yolov5x.yaml
浏览文件 @ e18e6811
......@@ -5,41 +5,48 @@ width_multiple: 1.25 # layer channel multiple
# anchors
anchors:
- [10,13, 16,30, 33,23] # P3/8
- [30,61, 62,45, 59,119] # P4/16
- [116,90, 156,198, 373,326] # P5/32
- [30,61, 62,45, 59,119] # P4/16
- [10,13, 16,30, 33,23] # P3/8
# yolov5 backbone
# YOLOv5 backbone
backbone:
# [from, number, module, args]
[[-1, 1, Focus, [64, 3]], # 1-P1/2
[-1, 1, Conv, [128, 3, 2]], # 2-P2/4
[-1, 3, Bottleneck, [128]],
[-1, 1, Conv, [256, 3, 2]], # 4-P3/8
[[-1, 1, Focus, [64, 3]], # 0-P1/2
[-1, 1, Conv, [128, 3, 2]], # 1-P2/4
[-1, 3, BottleneckCSP, [128]],
[-1, 1, Conv, [256, 3, 2]], # 3-P3/8
[-1, 9, BottleneckCSP, [256]],
[-1, 1, Conv, [512, 3, 2]], # 6-P4/16
[-1, 1, Conv, [512, 3, 2]], # 5-P4/16
[-1, 9, BottleneckCSP, [512]],
[-1, 1, Conv, [1024, 3, 2]], # 8-P5/32
[-1, 1, Conv, [1024, 3, 2]], # 7-P5/32
[-1, 1, SPP, [1024, [5, 9, 13]]],
[-1, 6, BottleneckCSP, [1024]], # 10
]
# yolov5 head
# YOLOv5 head
head:
[[-1, 3, BottleneckCSP, [1024, False]], # 11
[-1, 1, nn.Conv2d, [na * (nc + 5), 1, 1]], # 12 (P5/32-large)
[[-1, 3, BottleneckCSP, [1024, False]], # 9
[-2, 1, nn.Upsample, [None, 2, 'nearest']],
[[-1, 6], 1, Concat, [1]], # cat backbone P4
[-1, 1, Conv, [512, 1, 1]],
[-1, 3, BottleneckCSP, [512, False]],
[-1, 1, nn.Conv2d, [na * (nc + 5), 1, 1]], # 17 (P4/16-medium)
[-1, 1, nn.Upsample, [None, 2, 'nearest']],
[[-1, 6], 1, Concat, [1]], # cat backbone P4
[-1, 3, BottleneckCSP, [512, False]], # 13
[-2, 1, nn.Upsample, [None, 2, 'nearest']],
[[-1, 4], 1, Concat, [1]], # cat backbone P3
[-1, 1, Conv, [256, 1, 1]],
[-1, 1, nn.Upsample, [None, 2, 'nearest']],
[[-1, 4], 1, Concat, [1]], # cat backbone P3
[-1, 3, BottleneckCSP, [256, False]],
[-1, 1, nn.Conv2d, [na * (nc + 5), 1, 1]], # 22 (P3/8-small)
[-1, 1, nn.Conv2d, [na * (nc + 5), 1, 1]], # 18 (P3/8-small)
[-2, 1, Conv, [256, 3, 2]],
[[-1, 14], 1, Concat, [1]], # cat head P4
[-1, 3, BottleneckCSP, [512, False]],
[-1, 1, nn.Conv2d, [na * (nc + 5), 1, 1]], # 22 (P4/16-medium)
[-2, 1, Conv, [512, 3, 2]],
[[-1, 10], 1, Concat, [1]], # cat head P5
[-1, 3, BottleneckCSP, [1024, False]],
[-1, 1, nn.Conv2d, [na * (nc + 5), 1, 1]], # 26 (P5/32-large)
[[], 1, Detect, [nc, anchors]], # Detect(P3, P4, P5)
[[], 1, Detect, [nc, anchors]], # Detect(P5, P4, P3)
]
requirements.txt
浏览文件 @ e18e6811
......@@ -2,7 +2,7 @@
Cython
numpy==1.17
opencv-python
torch>=1.5
torch>=1.4
matplotlib
pillow
tensorboard
......@@ -21,4 +21,4 @@ git+https://github.com/cocodataset/cocoapi.git#subdirectory=PythonAPI
# conda install -yc conda-forge scikit-image pycocotools tensorboard
# conda install -yc spyder-ide spyder-line-profiler
# conda install -yc pytorch pytorch torchvision
# conda install -yc conda-forge protobuf numpy && pip install onnx # https://github.com/onnx/onnx#linux-and-macos
# conda install -yc conda-forge protobuf numpy && pip install onnx==1.6.0 # https://github.com/onnx/onnx#linux-and-macos
test.py
浏览文件 @ e18e6811
import argparse
import json
import yaml
from torch.utils.data import DataLoader
from utils import google_utils
from utils.datasets import *
from utils.utils import *
......@@ -17,16 +15,18 @@ def test(data,
save_json=False,
single_cls=False,
augment=False,
verbose=False,
model=None,
dataloader=None,
fast=False,
verbose=False,
save_dir='.'):
save_dir='.',
merge=False):
# Initialize/load model and set device
if model is None:
training = False
device = torch_utils.select_device(opt.device, batch_size=batch_size)
half = device.type != 'cpu' # half precision only supported on CUDA
# Remove previous
for f in glob.glob(f'{save_dir}/test_batch*.jpg'):
......@@ -38,18 +38,19 @@ def test(data,
torch_utils.model_info(model)
model.fuse()
model.to(device)
if half:
model.half() # to FP16
if device.type != 'cpu' and torch.cuda.device_count() > 1:
model = nn.DataParallel(model)
# Multi-GPU disabled, incompatible with .half() https://github.com/ultralytics/yolov5/issues/99
# if device.type != 'cpu' and torch.cuda.device_count() > 1:
# model = nn.DataParallel(model)
else: # called by train.py
training = True
device = next(model.parameters()).device # get model device
half = device.type != 'cpu' # half precision only supported on CUDA
if half:
model.half() # to FP16
# Half
half = device.type != 'cpu' and torch.cuda.device_count() == 1 # half precision only supported on single-GPU
if half:
model.half() # to FP16
# Configure
model.eval()
......@@ -57,29 +58,16 @@ def test(data,
data = yaml.load(f, Loader=yaml.FullLoader) # model dict
nc = 1 if single_cls else int(data['nc']) # number of classes
iouv = torch.linspace(0.5, 0.95, 10).to(device) # iou vector for mAP@0.5:0.95
# iouv = iouv[0].view(1) # comment for mAP@0.5:0.95
niou = iouv.numel()
# Dataloader
if dataloader is None: # not training
merge = opt.merge # use Merge NMS
img = torch.zeros((1, 3, imgsz, imgsz), device=device) # init img
_ = model(img.half() if half else img) if device.type != 'cpu' else None # run once
fast |= conf_thres > 0.001 # enable fast mode
path = data['test'] if opt.task == 'test' else data['val'] # path to val/test images
dataset = LoadImagesAndLabels(path,
imgsz,
batch_size,
rect=True, # rectangular inference
single_cls=opt.single_cls, # single class mode
pad=0.5) # padding
batch_size = min(batch_size, len(dataset))
nw = min([os.cpu_count(), batch_size if batch_size > 1 else 0, 8]) # number of workers
dataloader = DataLoader(dataset,
batch_size=batch_size,
num_workers=nw,
pin_memory=True,
collate_fn=dataset.collate_fn)
dataloader = create_dataloader(path, imgsz, batch_size, int(max(model.stride)), opt,
hyp=None, augment=False, cache=False, pad=0.5, rect=True)[0]
seen = 0
names = model.names if hasattr(model, 'names') else model.module.names
......@@ -109,7 +97,7 @@ def test(data,
# Run NMS
t = torch_utils.time_synchronized()
output = non_max_suppression(inf_out, conf_thres=conf_thres, iou_thres=iou_thres, fast=fast)
output = non_max_suppression(inf_out, conf_thres=conf_thres, iou_thres=iou_thres, merge=merge)
t1 += torch_utils.time_synchronized() - t
# Statistics per image
......@@ -235,6 +223,7 @@ def test(data,
'See https://github.com/cocodataset/cocoapi/issues/356')
# Return results
model.float() # for training
maps = np.zeros(nc) + map
for i, c in enumerate(ap_class):
maps[c] = ap[i]
......@@ -244,7 +233,7 @@ def test(data,
if __name__ == '__main__':
parser = argparse.ArgumentParser(prog='test.py')
parser.add_argument('--weights', type=str, default='weights/yolov5s.pt', help='model.pt path')
parser.add_argument('--data', type=str, default='data/coco.yaml', help='*.data path')
parser.add_argument('--data', type=str, default='data/coco128.yaml', help='*.data path')
parser.add_argument('--batch-size', type=int, default=32, help='size of each image batch')
parser.add_argument('--img-size', type=int, default=640, help='inference size (pixels)')
parser.add_argument('--conf-thres', type=float, default=0.001, help='object confidence threshold')
......@@ -254,6 +243,7 @@ if __name__ == '__main__':
parser.add_argument('--device', default='', help='cuda device, i.e. 0 or 0,1,2,3 or cpu')
parser.add_argument('--single-cls', action='store_true', help='treat as single-class dataset')
parser.add_argument('--augment', action='store_true', help='augmented inference')
parser.add_argument('--merge', action='store_true', help='use Merge NMS')
parser.add_argument('--verbose', action='store_true', help='report mAP by class')
opt = parser.parse_args()
opt.img_size = check_img_size(opt.img_size)
......@@ -271,12 +261,13 @@ if __name__ == '__main__':
opt.iou_thres,
opt.save_json,
opt.single_cls,
opt.augment)
opt.augment,
opt.verbose)
elif opt.task == 'study': # run over a range of settings and save/plot
for weights in ['yolov5s.pt', 'yolov5m.pt', 'yolov5l.pt', 'yolov5x.pt']:
for weights in ['yolov5s.pt', 'yolov5m.pt', 'yolov5l.pt', 'yolov5x.pt', 'yolov3-spp.pt']:
f = 'study_%s_%s.txt' % (Path(opt.data).stem, Path(weights).stem) # filename to save to
x = list(range(288, 896, 64)) # x axis
x = list(range(352, 832, 64)) # x axis
y = [] # y axis
for i in x: # img-size
print('\nRunning %s point %s...' % (f, i))
......
train.py
浏览文件 @ e18e6811
......@@ -4,10 +4,12 @@ import torch.distributed as dist
import torch.nn.functional as F
import torch.optim as optim
import torch.optim.lr_scheduler as lr_scheduler
import torch.utils.data
from torch.utils.tensorboard import SummaryWriter
import test # import test.py to get mAP after each epoch
from models.yolo import Model
from utils import google_utils
from utils.datasets import *
from utils.utils import *
......@@ -72,6 +74,7 @@ def train(hyp):
# Create model
model = Model(opt.cfg).to(device)
assert model.md['nc'] == nc, '%s nc=%g classes but %s nc=%g classes' % (opt.data, nc, opt.cfg, model.md['nc'])
model.names = data_dict['names']
# Image sizes
gs = int(max(model.stride)) # grid size (max stride)
......@@ -148,37 +151,17 @@ def train(hyp):
world_size=1, # number of nodes
rank=0) # node rank
model = torch.nn.parallel.DistributedDataParallel(model)
# pip install torch==1.4.0+cu100 torchvision==0.5.0+cu100 -f https://download.pytorch.org/whl/torch_stable.html
# Dataset
dataset = LoadImagesAndLabels(train_path, imgsz, batch_size,
augment=True,
hyp=hyp, # augmentation hyperparameters
rect=opt.rect, # rectangular training
cache_images=opt.cache_images,
single_cls=opt.single_cls)
# Trainloader
dataloader, dataset = create_dataloader(train_path, imgsz, batch_size, gs, opt,
hyp=hyp, augment=True, cache=opt.cache_images, rect=opt.rect)
mlc = np.concatenate(dataset.labels, 0)[:, 0].max() # max label class
assert mlc < nc, 'Label class %g exceeds nc=%g in %s. Correct your labels or your model.' % (mlc, nc, opt.cfg)
# Dataloader
batch_size = min(batch_size, len(dataset))
nw = min([os.cpu_count(), batch_size if batch_size > 1 else 0, 8]) # number of workers
dataloader = torch.utils.data.DataLoader(dataset,
batch_size=batch_size,
num_workers=nw,
shuffle=not opt.rect, # Shuffle=True unless rectangular training is used
pin_memory=True,
collate_fn=dataset.collate_fn)
# Testloader
testloader = torch.utils.data.DataLoader(LoadImagesAndLabels(test_path, imgsz_test, batch_size,
hyp=hyp,
rect=True,
cache_images=opt.cache_images,
single_cls=opt.single_cls),
batch_size=batch_size,
num_workers=nw,
pin_memory=True,
collate_fn=dataset.collate_fn)
testloader = create_dataloader(test_path, imgsz_test, batch_size, gs, opt,
hyp=hyp, augment=False, cache=opt.cache_images, rect=True)[0]
# Model parameters
hyp['cls'] *= nc / 80. # scale coco-tuned hyp['cls'] to current dataset
......@@ -186,7 +169,6 @@ def train(hyp):
model.hyp = hyp # attach hyperparameters to model
model.gr = 1.0 # giou loss ratio (obj_loss = 1.0 or giou)
model.class_weights = labels_to_class_weights(dataset.labels, nc).to(device) # attach class weights
model.names = data_dict['names']
#save hyperparamter and training options in run folder
with open(os.path.join(log_dir, 'hyp.yaml'), 'w') as f:
......@@ -200,11 +182,17 @@ def train(hyp):
c = torch.tensor(labels[:, 0]) # classes
# cf = torch.bincount(c.long(), minlength=nc) + 1.
# model._initialize_biases(cf.to(device))
#always plot labels to log_dir
plot_labels(labels, save_dir=log_dir)
tb_writer.add_histogram('classes', c, 0)
if tb_writer:
tb_writer.add_histogram('classes', c, 0)
# Check anchors
check_best_possible_recall(dataset, anchors=model.model[-1].anchor_grid, thr=hyp['anchor_t'], imgsz=imgsz)
if not opt.noautoanchor:
check_anchors(dataset, model=model, thr=hyp['anchor_t'], imgsz=imgsz)
# Exponential moving average
ema = torch_utils.ModelEMA(model)
......@@ -216,7 +204,7 @@ def train(hyp):
maps = np.zeros(nc) # mAP per class
results = (0, 0, 0, 0, 0, 0, 0) # 'P', 'R', 'mAP', 'F1', 'val GIoU', 'val Objectness', 'val Classification'
print('Image sizes %g train, %g test' % (imgsz, imgsz_test))
print('Using %g dataloader workers' % nw)
print('Using %g dataloader workers' % dataloader.num_workers)
print('Starting training for %g epochs...' % epochs)
# torch.autograd.set_detect_anomaly(True)
for epoch in range(start_epoch, epochs): # epoch ------------------------------------------------------------------
......@@ -285,10 +273,10 @@ def train(hyp):
# Plot
if ni < 3:
f = os.path.join(log_dir, 'train_batch%g.jpg' % i) # filename
res = plot_images(images=imgs, targets=targets, paths=paths, fname=f)
if tb_writer:
tb_writer.add_image(f, res, dataformats='HWC', global_step=epoch)
f = os.path.join(log_dir, 'train_batch%g.jpg' % ni) # filename
result = plot_images(images=imgs, targets=targets, paths=paths, fname=f)
if tb_writer and result is not None:
tb_writer.add_image(f, result, dataformats='HWC', global_step=epoch)
# tb_writer.add_graph(model, imgs) # add model to tensorboard
# end batch ------------------------------------------------------------------------------------------------
......@@ -307,7 +295,6 @@ def train(hyp):
model=ema.ema,
single_cls=opt.single_cls,
dataloader=testloader,
fast=epoch < epochs / 2,
save_dir=log_dir)
# Write
......@@ -362,7 +349,7 @@ def train(hyp):
if not opt.evolve:
plot_results(save_dir = log_dir) # save as results.png
print('%g epochs completed in %.3f hours.\n' % (epoch - start_epoch + 1, (time.time() - t0) / 3600))
dist.destroy_process_group() if torch.cuda.device_count() > 1 else None
dist.destroy_process_group() if device.type != 'cpu' and torch.cuda.device_count() > 1 else None
torch.cuda.empty_cache()
return results
......@@ -379,6 +366,7 @@ if __name__ == '__main__':
parser.add_argument('--rect', action='store_true', help='rectangular training')
parser.add_argument('--nosave', action='store_true', help='only save final checkpoint')
parser.add_argument('--notest', action='store_true', help='only test final epoch')
parser.add_argument('--noautoanchor', action='store_true', help='disable autoanchor check')
parser.add_argument('--evolve', action='store_true', help='evolve hyperparameters')
parser.add_argument('--bucket', type=str, default='', help='gsutil bucket')
parser.add_argument('--cache-images', action='store_true', help='cache images for faster training')
......
utils/datasets.py
浏览文件 @ e18e6811
......@@ -18,7 +18,7 @@ from utils.utils import xyxy2xywh, xywh2xyxy
help_url = 'https://github.com/ultralytics/yolov5/wiki/Train-Custom-Data'
img_formats = ['.bmp', '.jpg', '.jpeg', '.png', '.tif', '.dng']
vid_formats = ['.mov', '.avi', '.mp4']
vid_formats = ['.mov', '.avi', '.mp4', '.mpg', '.mpeg', '.m4v', '.wmv', '.mkv']
# Get orientation exif tag
for orientation in ExifTags.TAGS.keys():
......@@ -41,6 +41,26 @@ def exif_size(img):
return s
def create_dataloader(path, imgsz, batch_size, stride, opt, hyp=None, augment=False, cache=False, pad=0.0, rect=False):
dataset = LoadImagesAndLabels(path, imgsz, batch_size,
augment=augment, # augment images
hyp=hyp, # augmentation hyperparameters
rect=rect, # rectangular training
cache_images=cache,
single_cls=opt.single_cls,
stride=stride,
pad=pad)
batch_size = min(batch_size, len(dataset))
nw = min([os.cpu_count(), batch_size if batch_size > 1 else 0, 8]) # number of workers
dataloader = torch.utils.data.DataLoader(dataset,
batch_size=batch_size,
num_workers=nw,
pin_memory=True,
collate_fn=LoadImagesAndLabels.collate_fn)
return dataloader, dataset
class LoadImages: # for inference
def __init__(self, path, img_size=416):
path = str(Path(path)) # os-agnostic
......@@ -63,7 +83,8 @@ class LoadImages: # for inference
self.new_video(videos[0]) # new video
else:
self.cap = None
assert self.nF > 0, 'No images or videos found in ' + path
assert self.nF > 0, 'No images or videos found in %s. Supported formats are:\nimages: %s\nvideos: %s' % \
(path, img_formats, vid_formats)
def __iter__(self):
self.count = 0
......@@ -257,7 +278,7 @@ class LoadStreams: # multiple IP or RTSP cameras
class LoadImagesAndLabels(Dataset): # for training/testing
def __init__(self, path, img_size=416, batch_size=16, augment=False, hyp=None, rect=False, image_weights=False,
cache_images=False, single_cls=False, pad=0.0):
cache_images=False, single_cls=False, stride=32, pad=0.0):
try:
path = str(Path(path)) # os-agnostic
parent = str(Path(path).parent) + os.sep
......@@ -324,7 +345,7 @@ class LoadImagesAndLabels(Dataset): # for training/testing
elif mini > 1:
shapes[i] = [1, 1 / mini]
self.batch_shapes = np.ceil(np.array(shapes) * img_size / 32. + pad).astype(np.int) * 32
self.batch_shapes = np.ceil(np.array(shapes) * img_size / stride + pad).astype(np.int) * stride
# Cache labels
self.imgs = [None] * n
......@@ -711,7 +732,7 @@ def random_affine(img, targets=(), degrees=10, translate=.1, scale=.1, shear=10,
area = w * h
area0 = (targets[:, 3] - targets[:, 1]) * (targets[:, 4] - targets[:, 2])
ar = np.maximum(w / (h + 1e-16), h / (w + 1e-16)) # aspect ratio
i = (w > 4) & (h > 4) & (area / (area0 * s + 1e-16) > 0.2) & (ar < 10)
i = (w > 2) & (h > 2) & (area / (area0 * s + 1e-16) > 0.2) & (ar < 20)
targets = targets[i]
targets[:, 1:5] = xy[i]
......
utils/torch_utils.py
浏览文件 @ e18e6811
......@@ -7,6 +7,7 @@ import torch
import torch.backends.cudnn as cudnn
import torch.nn as nn
import torch.nn.functional as F
import torchvision.models as models
def init_seeds(seed=0):
......@@ -120,18 +121,22 @@ def model_info(model, verbose=False):
def load_classifier(name='resnet101', n=2):
# Loads a pretrained model reshaped to n-class output
import pretrainedmodels # https://github.com/Cadene/pretrained-models.pytorch#torchvision
model = pretrainedmodels.__dict__[name](num_classes=1000, pretrained='imagenet')
model = models.__dict__[name](pretrained=True)
# Display model properties
for x in ['model.input_size', 'model.input_space', 'model.input_range', 'model.mean', 'model.std']:
input_size = [3, 224, 224]
input_space = 'RGB'
input_range = [0, 1]
mean = [0.485, 0.456, 0.406]
std = [0.229, 0.224, 0.225]
for x in [input_size, input_space, input_range, mean, std]:
print(x + ' =', eval(x))
# Reshape output to n classes
filters = model.last_linear.weight.shape[1]
model.last_linear.bias = torch.nn.Parameter(torch.zeros(n))
model.last_linear.weight = torch.nn.Parameter(torch.zeros(n, filters))
model.last_linear.out_features = n
filters = model.fc.weight.shape[1]
model.fc.bias = torch.nn.Parameter(torch.zeros(n), requires_grad=True)
model.fc.weight = torch.nn.Parameter(torch.zeros(n, filters), requires_grad=True)
model.fc.out_features = n
return model
......
utils/utils.py
浏览文件 @ e18e6811
差异被折叠。
Markdown 格式
0% 或
您添加了 0 人 到此讨论。请谨慎行事。
请先完成此评论的编辑！
请 注册 或者 后发表评论