my data is too small. I wanna use cross-validation to fine-tune the model. I mean, in the epoch1 and epoch2 the model uses different training data. I invoke the dataset function in the training session and get some errors.
Could you give me some advise to rewrite the code?