How to determine the training process is ready?

[LLblessPeng]

epoch: 36 iteration : 0 Loss: 0.0008807919803075492 lr: 1e-05 100%

Epoch_item: 36 Correct_num: 1030 total: 1030 Accuracy on test data: 1.0

Category_0 predict_really 1000 predict 1000 Precision 1.0 recall0.001
Category_1 predict_really 30 predict 30 Precision 1.0 recall0.060

epoch: 37 iteration : 0 Loss: 0.00040307408198714256 lr: 1e-05 100%

Epoch_item: 37 Correct_num: 1030 total: 1030 Accuracy on test data: 1.0

Category_0 predict_really 1000 predict 1000 Precision 1.0 recall0.001
Category_1 predict_really 30 predict 30 Precision 1.0 recall0.06

Hello, Thanks for your help in advance. The training process repeats itself as described above, does such a parameter indicate that the training has been completed?

[zhoubay]

Yes, your understanding is correct!

[yilongzhao1122]

我觉得可以加入预测指标是否提升判断是否早停的机制

[zhoubay]

我觉得可以加入预测指标是否提升判断是否早停的机制

Thank you for your idea! Feel free to fork the repo, do some ammendment and then open a pull request!