指令训练指令yolo detect train dataH:\xxc\yolo\yolo_dataset\yolo_dataset\data.yaml modelyolov8n.pt epochs100 imgsz640测试指令yolo detect predict modelH:\xxc\yolo\ultralytics\runs\detect\train-11\weights\best.pt sourceH:\xxc\yolo\pig\images\train\00000000.png saveTrue模型导出Yolo自带的工具可以生成tflite文件yolo export modelF:\xxc\yolo\ultralytics\runs\detect\runs\pigpose_ablation\YOLOv8s\weights\best.pt formattflite imgsz320YOLOv8 是一种单阶段目标检测算法由 Backbone、Neck 和 Head 三部分组成。Backbone 负责提取图像特征Neck 用于融合多尺度特征Head 输出目标类别和边界框信息。YOLOv8 使用 C2f 模块增强梯度流动能力并采用解耦检测头分别进行分类和回归预测现有的yolov8论文比较多。Yolo的模型配置文件D:\桌面\yolo\yolo_original\ultralytics\ultralytics\cfg\models\v8模块定义D:\桌面\yolo\yolo_original\ultralytics\ultralytics\nn\modules脚本D:\桌面\yolo\yolo_original\ultralytics\ultralytics\models\yolo\detect已经写好的数据集和配置文件D:\桌面\yolo\yolo_original\ultralytics\ultralytics\cfg\datasets默认参数D:\桌面\yolo\yolo_original\ultralytics\ultralytics\cfg\default.yamlYolo自带的工具可以生成tflite文件yolo export modelF:\xxc\yolo\ultralytics\runs\detect\runs\pigpose_ablation\YOLOv8s\weights\best.pt formattflite imgsz320猪姿态检测的一些数据集Automatic Monitoring of Pigs - Dataset Ninjahttps://datasetninja.com/automatic-monitoring-pigs#download可以转化成yolo格式里面有0: standing pig1: sternal lying pig2: lateral lying pig三种标签534张训练集图片133张验证集图片Multi-view Pig Posture Recognition | Kagglehttps://www.kaggle.com/competitions/multi-view-pig-posture-recognition/dataClass Definitions:0 — Lateral_lying_left1 — Lateral_lying_right2 — Sitting3 — Standing4 — Sternal_lying测试集没有标签anilbhujel/viewpoint-aware-pig-posture-recognition · Datasets at Hugging Facehttps://huggingface.co/datasets/anilbhujel/viewpoint-aware-pig-posture-recognition#dataset-statisticsSplitInstancesImagesTrain22,9333,090Seen-VP test2,603300Unseen-VP test11,7081,350Class IDLabelDescription0Lateral_lying_leftPig lying on its side, left laterally1Lateral_lying_rightPig lying on its side, right laterally2SittingPig in sitting posture3StandingPig standing upright4Sternal_lyingPig lying on sternum (chest-down, sphinx-like)没有验证集anilbhujel/Gilt_posture_dataset · Datasets at Hugging Facehttps://huggingface.co/datasets/anilbhujel/Gilt_posture_datasetClass NameClass IDfeeding0lateral_lying1sitting2standing3sternal_lying43015张。数据集还按 0.75 : 0.15 : 0.1 划分为训练集、验证集和测试集。姿态分布不平均SowPostureDS数据集地址https://opendata.uni-kiel.de/receive/fdr_mods_00000316SowPostureDS: A Multi-Class Image Dataset for YOLO-Based Detection of Sow Postures in diverse Farrowing Systems | Scientific Datahttps://www.nature.com/articles/s41597-026-07788-x?utm_sourcechatgpt.com#Sec6共14,400 张图片分三个真实养殖环境每个环境 4,800 张四个类别每个环境各 1,200 张因此整体四类是完全均衡的。论文实验采用了7:2:1 10080 / 2880 / 1440的 train/val/test 划分但官方 ZIP 的描述是 images/ labels/ 两个总文件夹一只猪Item - Automated recognition of postures and drinking behaviour for the detection of compromised health in pigs - Newcastle University - Figsharehttps://data.ncl.ac.uk/articles/dataset/Automated_recognition_of_postures_and_drinking_behaviour_for_the_detection_of_compromised_health_in_pigs/13042619Automated recognition of postures and drinking behaviour for the detection of compromised health in pigs | Scientific Reportshttps://www.nature.com/articles/s41598-020-70688-6?utm_sourcechatgpt.com数据量非常大。论文页面显示Standing105,132 个实例覆盖 54,320 张图像Lateral Lying417,134 个实例覆盖 105,199 张图像Sternal lying166,085 个实例覆盖 81,495 张图像还有 Sitting 和 Drinking。数据集太大Suckling Browsehttps://universe.roboflow.com/pig-farming/suckling/browse?queryTextpageSize50startingIndex150browseQuerytruejast-25-e00395.pdfhttps://www.ejast.org/journal/jast/ahead/pdf/jast-25-e00395.pdf9,487 张同样是这 8 个类别数量和类别完全对应。姿态标注实例Lateral lying3,972Sternal lying2,291Standing2,191文章对训练测试进行划分主要检测对象是一头母猪画面中可以有多只仔猪。GitHub - popxoq/pig-behavior-detection-dataset: 猪日常饲养行为数据集 · GitHubhttps://github.com/popxoq/pig-behavior-detection-dataset项目内容图片总数3000 张训练集2100 张70%验证集600 张20%测试集300 张10%标注格式YOLO / VOC / COCO类别数量12 类图片有点少GitHub - laoli518/-100k-OPB-Dataset-: This repository contains the **100K-OPB Dataset** used in ORACLE: Knowledge-Efficient Pig Behaviour Recognition via Ontology-Guided Contrastive Learning. · GitHubhttps://github.com/laoli518/-100k-OPB-Dataset-76,431个训练样本、16,157个测试样本和12,199个来自全拆分LiR2024源的样本。该数据集涵盖猪的常见行为和姿势/活动状态包括躺卧、站立、行走、吃喝、打斗、不打斗、调查、睡觉、坐着、骑乘、鼻对鼻互动、侧卧、胸侧躺、喂食/饮水相关状态以及喂食/饮水姿势组合。数据集偏大 yolo模型学习
指令训练指令yolo detect train dataH:\xxc\yolo\yolo_dataset\yolo_dataset\data.yaml modelyolov8n.pt epochs100 imgsz640测试指令yolo detect predict modelH:\xxc\yolo\ultralytics\runs\detect\train-11\weights\best.pt sourceH:\xxc\yolo\pig\images\train\00000000.png saveTrue模型导出Yolo自带的工具可以生成tflite文件yolo export modelF:\xxc\yolo\ultralytics\runs\detect\runs\pigpose_ablation\YOLOv8s\weights\best.pt formattflite imgsz320YOLOv8 是一种单阶段目标检测算法由 Backbone、Neck 和 Head 三部分组成。Backbone 负责提取图像特征Neck 用于融合多尺度特征Head 输出目标类别和边界框信息。YOLOv8 使用 C2f 模块增强梯度流动能力并采用解耦检测头分别进行分类和回归预测现有的yolov8论文比较多。Yolo的模型配置文件D:\桌面\yolo\yolo_original\ultralytics\ultralytics\cfg\models\v8模块定义D:\桌面\yolo\yolo_original\ultralytics\ultralytics\nn\modules脚本D:\桌面\yolo\yolo_original\ultralytics\ultralytics\models\yolo\detect已经写好的数据集和配置文件D:\桌面\yolo\yolo_original\ultralytics\ultralytics\cfg\datasets默认参数D:\桌面\yolo\yolo_original\ultralytics\ultralytics\cfg\default.yamlYolo自带的工具可以生成tflite文件yolo export modelF:\xxc\yolo\ultralytics\runs\detect\runs\pigpose_ablation\YOLOv8s\weights\best.pt formattflite imgsz320猪姿态检测的一些数据集Automatic Monitoring of Pigs - Dataset Ninjahttps://datasetninja.com/automatic-monitoring-pigs#download可以转化成yolo格式里面有0: standing pig1: sternal lying pig2: lateral lying pig三种标签534张训练集图片133张验证集图片Multi-view Pig Posture Recognition | Kagglehttps://www.kaggle.com/competitions/multi-view-pig-posture-recognition/dataClass Definitions:0 — Lateral_lying_left1 — Lateral_lying_right2 — Sitting3 — Standing4 — Sternal_lying测试集没有标签anilbhujel/viewpoint-aware-pig-posture-recognition · Datasets at Hugging Facehttps://huggingface.co/datasets/anilbhujel/viewpoint-aware-pig-posture-recognition#dataset-statisticsSplitInstancesImagesTrain22,9333,090Seen-VP test2,603300Unseen-VP test11,7081,350Class IDLabelDescription0Lateral_lying_leftPig lying on its side, left laterally1Lateral_lying_rightPig lying on its side, right laterally2SittingPig in sitting posture3StandingPig standing upright4Sternal_lyingPig lying on sternum (chest-down, sphinx-like)没有验证集anilbhujel/Gilt_posture_dataset · Datasets at Hugging Facehttps://huggingface.co/datasets/anilbhujel/Gilt_posture_datasetClass NameClass IDfeeding0lateral_lying1sitting2standing3sternal_lying43015张。数据集还按 0.75 : 0.15 : 0.1 划分为训练集、验证集和测试集。姿态分布不平均SowPostureDS数据集地址https://opendata.uni-kiel.de/receive/fdr_mods_00000316SowPostureDS: A Multi-Class Image Dataset for YOLO-Based Detection of Sow Postures in diverse Farrowing Systems | Scientific Datahttps://www.nature.com/articles/s41597-026-07788-x?utm_sourcechatgpt.com#Sec6共14,400 张图片分三个真实养殖环境每个环境 4,800 张四个类别每个环境各 1,200 张因此整体四类是完全均衡的。论文实验采用了7:2:1 10080 / 2880 / 1440的 train/val/test 划分但官方 ZIP 的描述是 images/ labels/ 两个总文件夹一只猪Item - Automated recognition of postures and drinking behaviour for the detection of compromised health in pigs - Newcastle University - Figsharehttps://data.ncl.ac.uk/articles/dataset/Automated_recognition_of_postures_and_drinking_behaviour_for_the_detection_of_compromised_health_in_pigs/13042619Automated recognition of postures and drinking behaviour for the detection of compromised health in pigs | Scientific Reportshttps://www.nature.com/articles/s41598-020-70688-6?utm_sourcechatgpt.com数据量非常大。论文页面显示Standing105,132 个实例覆盖 54,320 张图像Lateral Lying417,134 个实例覆盖 105,199 张图像Sternal lying166,085 个实例覆盖 81,495 张图像还有 Sitting 和 Drinking。数据集太大Suckling Browsehttps://universe.roboflow.com/pig-farming/suckling/browse?queryTextpageSize50startingIndex150browseQuerytruejast-25-e00395.pdfhttps://www.ejast.org/journal/jast/ahead/pdf/jast-25-e00395.pdf9,487 张同样是这 8 个类别数量和类别完全对应。姿态标注实例Lateral lying3,972Sternal lying2,291Standing2,191文章对训练测试进行划分主要检测对象是一头母猪画面中可以有多只仔猪。GitHub - popxoq/pig-behavior-detection-dataset: 猪日常饲养行为数据集 · GitHubhttps://github.com/popxoq/pig-behavior-detection-dataset项目内容图片总数3000 张训练集2100 张70%验证集600 张20%测试集300 张10%标注格式YOLO / VOC / COCO类别数量12 类图片有点少GitHub - laoli518/-100k-OPB-Dataset-: This repository contains the **100K-OPB Dataset** used in ORACLE: Knowledge-Efficient Pig Behaviour Recognition via Ontology-Guided Contrastive Learning. · GitHubhttps://github.com/laoli518/-100k-OPB-Dataset-76,431个训练样本、16,157个测试样本和12,199个来自全拆分LiR2024源的样本。该数据集涵盖猪的常见行为和姿势/活动状态包括躺卧、站立、行走、吃喝、打斗、不打斗、调查、睡觉、坐着、骑乘、鼻对鼻互动、侧卧、胸侧躺、喂食/饮水相关状态以及喂食/饮水姿势组合。数据集偏大