TableTennis数据集

[onedayCVPR]

作者您好，关于乒乓球动作识别的数据集能够提供吗？

DATASET: #DATASET field batch_size: 1 #Mandatory, bacth size num_workers: 2 #Mandatory, XXX the number of subprocess on each GPU. test_batch_size: 1 train: format: "FrameDataset_Sport" #Mandatory, indicate the type of dataset, associate to the 'paddlevidel/loader/dateset' filepath: "train.list" #Mandatory, train data index file path suffix: 'img{:05}.jpg' valid: format: "FrameDataset_Sport" #Mandatory, indicate the type of dataset, associate to the 'paddlevidel/loader/dateset' filepath: "val.list" #Mandatory, valid data index file path suffix: 'img{:05}.jpg' test: format: "FrameDataset_Sport" #Mandatory, indicate the type of dataset, associate to the 'paddlevidel/loader/dateset' filepath: "val.list" #Mandatory, valid data index file path suffix: 'img{:05}.jpg'

[huangjun12]

目前提供了样例数据 https://github.com/PaddlePaddle/PaddleVideo/tree/develop/applications/TableTennis/ActionRecognition#模型推理 全量数据集暂不开源哈

[onedayCVPR]

作者您好，我有一部分乒乓球击球动作数据集，数据集为mp4格式，数据集标签包含动作的起始和结束帧，标签文件内容如下，格式为XML `

请问如何修改videoswin_tabletennis.yaml中的dataset DATASET: #DATASET field batch_size: 1 #Mandatory, bacth size num_workers: 2 #Mandatory, XXX the number of subprocess on each GPU. test_batch_size: 1 train: format: "FrameDataset_Sport" #Mandatory, indicate the type of dataset, associate to the 'paddlevidel/loader/dateset' filepath: "train.list" #Mandatory, train data index file path suffix: 'img{:05}.jpg' valid: format: "FrameDataset_Sport" #Mandatory, indicate the type of dataset, associate to the 'paddlevidel/loader/dateset' filepath: "val.list" #Mandatory, valid data index file path suffix: 'img{:05}.jpg' test: format: "FrameDataset_Sport" #Mandatory, indicate the type of dataset, associate to the 'paddlevidel/loader/dateset' filepath: "val.list" #Mandatory, valid data index file path suffix: 'img{:05}.jpg' ` 才能使用自己的数据集进行训练呢？

[xuyao91]

您好，请问一下你上面这个xml文件是通过什么工具标注出来的？

[jerrywgz]

您好，请问问题还是否需要解决，目前多模态视频能力集成在PaddleMIX中，https://github.com/PaddlePaddle/PaddleMIX/tree/develop 可以在这个repo下提出您的需求