Does using a different batch_size influence the performance of TPLinkerPlus for model prediction?

[jarork]

I trained the model with a batch_size of 28; theoretically, would I get a poorer result if I lower the batch_size to 10 at the evaluation stage (prediction)?

[131250208]

Predicted results have nothing to do with batch size. This could be verified by setting different batch sizes to the same model when predicting.