左右 zuoyu
Loading Heatmap…

zuoyu created CPU/GPU training taskzuoyu202208071351359

5 days ago

zuoyu created CPU/GPU training taskzuoyu202208052107622

1 week ago

zuoyu created CPU/GPU training taskzuoyu202208011538307

1 week ago

zuoyu pushed to main at ColugoMum/Exprements-public

1 week ago

zuoyu pushed to main at ColugoMum/Exprements-public

  • 9ea87ad5f8 更新 'PaddleClas/ppcls/configs/GeneralRecognition/pphgnet_udml.yaml'

1 week ago

zuoyu created CPU/GPU training taskzuoyu202207280144109

2 weeks ago

zuoyu created CPU/GPU training taskzuoyu202207280143619

2 weeks ago

zuoyu pushed to main at ColugoMum/Exprements-public

2 weeks ago

zuoyu created reasoning task zuoyu202207251017352

2 weeks ago

zuoyu pushed to main at ColugoMum/Exprements-public

2 weeks ago

zuoyu pushed to main at ColugoMum/Exprements-public

2 weeks ago

zuoyu created reasoning task zuoyu202207250913512

2 weeks ago

zuoyu pushed to main at ColugoMum/Exprements-public

2 weeks ago

zuoyu created reasoning task zuoyu202207250912655

2 weeks ago

zuoyu created reasoning task zuoyu202207250078528

2 weeks ago

zuoyu commented on issue ColugoMum/Exprements-public#5

模型存储需要注意

1. 后面看平台如何支持模型这一块; 2. 暂时还是先把它放到代码仓库吧。目前的模型还不算太大。

2 weeks ago

zuoyu opened issue zeizei/OpenI_Learning#139

[项目搜索]搜索结果优化

2 weeks ago

zuoyu opened issue zeizei/OpenI_Learning#138

智算集群训练任务的更新图标

2 weeks ago

zuoyu commented on issue ColugoMum/Dataset#8

【思考】是否有可能将这些SKU数据集混合成一个十分庞大的数据集进行测试

目前开源的数据集有一个非常致命的缺陷,尤其是国内的开源数据集,比如阿里的、京东的,都没有提供中文label,所以几乎没法直接用。其实要是先提供一下中文的label_list,剩下采取一些技术手段还是能有的。

2 weeks ago

zuoyu commented on issue ColugoMum/Dataset#8

【思考】是否有可能将这些SKU数据集混合成一个十分庞大的数据集进行测试

之前paddle的团队好像有做过类似的,把零售的、动漫的等等,合在一起发布了一个预训练模型,效果还可以。其实要是用PP-ShiTu的方案,最好还是将领域内的大多数囊括。目前组织基于的RP2K还是相对小了一些。

2 weeks ago