首页
后端开发
Python
目标检测pytorch复现SSD目标检测项目项目源码

目标检测pytorch复现SSD目标检测项目项目源码

24 次浏览 0 条评论
rar
rar
SSD（Single Shot MultiBox Detector）是大神Wei Liu在ECCV 2016上发表的一种的目标检测算法。对于输入图像大小300x300的版本在VOC2007数据集上达到了72.1%mAP的准确率并且检测速度达到了惊人的58FPS（ Faster RCNN：73.2%mAP，7FPS； YOLOv1： 63.4%mAP，45FPS ），500x500的版本达到了75.1%mAP的准确率。当然算法YOLOv2已经赶上了SSD，YOLOv3已经超越SSD，但SSD算法依旧值得研究。
rar
ssd.rar 预估大小：45个文件
folder
ssd 文件夹
file
代码使用方式.png 597KB
file
pascal_voc_classes.json 348B
folder
src 文件夹
file
utils.py 26KB
file
__init__.py 150B
file
res50_backbone.py 4KB
file
nvidia_ssdpyt_amp_200703.pt 174.91MB
file
ssd_model.py 10KB
folder
__pycache__ 文件夹
file
res50_backbone.cpython-38.pyc 3KB
file
utils.cpython-38.pyc 15KB
file
ssd_model.cpython-38.pyc 6KB
file
__init__.cpython-38.pyc 411B
file
获取到较大损失值的位置.jpg 52KB
file
res50_ssd.png 194KB
folder
multi_train 文件夹
file
model_48.pth 118.3MB
file
model_49.pth 118.3MB
file
model_47.pth 118.3MB
folder
save_weights 文件夹
file
transforms.py 7KB
file
my_dataset.py 9KB
file
results20230423-132435.txt 6KB
file
draw_box_utils.py 6KB
file
train_multi_GPU.py 11KB
file
requirements.txt 78B
file
train_ssd300.py 8KB
file
validation.py 9KB
file
mAP.png 19KB
file
plot_curve.py 1KB
file
loss_and_lr20230423-154327.png 25KB
folder
__pycache__ 文件夹
file
transforms.cpython-38.pyc 6KB
file
my_dataset.cpython-38.pyc 5KB
file
plot_curve.cpython-38.pyc 2KB
folder
train_utils 文件夹
file
__init__.py 246B
file
distributed_utils.py 10KB
file
group_by_aspect_ratio.py 7KB
file
train_eval_utils.py 5KB
folder
__pycache__ 文件夹
file
distributed_utils.cpython-38.pyc 10KB
file
coco_utils.cpython-38.pyc 2KB
file
group_by_aspect_ratio.cpython-38.pyc 7KB
file
train_eval_utils.cpython-38.pyc 4KB
file
coco_eval.cpython-38.pyc 9KB
file
__init__.cpython-38.pyc 516B
file
coco_eval.py 12KB
file
coco_utils.py 2KB
file
README.md 3KB
file
record_mAP.txt 2KB
file
predict_test.py 3KB
rar 文件大小：454.85MB