Tutorial for finetune on my own dataset

keyu-tian / SparK

[ICLR'23 Spotlight🔥] The first successful BERT/MAE-style pretraining on any convolutional network; Pytorch impl. of "Designing BERT for Convolutional Networks: Sparse and Hierarchical Masked Modeling"

https://arxiv.org/abs/2301.03580

MIT License

1.41k stars 82 forks source link

Tutorial for finetune on my own dataset #67

Closed WeiCL7777 closed 8 months ago

WeiCL7777 commented 8 months ago

请问作者，finetune阶段也可以针对自己的数据集（非 ImageNet）进行处理吧？请问应当修改哪部分的代码？谢谢

keyu-tian commented 8 months ago

@WeiCL7777 将 https://github.com/keyu-tian/SparK/blob/main/downstream_imagenet/data.py#L60 的 ImageFolder 换为自己的数据集，选择合适的数据预处理 trans_train 就可以了