为什么textcnn的训练里面，当step=50的时候就停了啊

Question

问题遇到的现象和发生背景

问题相关代码，请勿粘贴截图

2022-03-31T10:59:11.178435: step 47, loss 0.786139, acc 0.8125
2022-03-31T10:59:11.559416: step 48, loss 0.468298, acc 0.8125
2022-03-31T10:59:11.947378: step 49, loss 0.589047, acc 0.8125
2022-03-31T10:59:12.343319: step 50, loss 0.612781, acc 0.84375

Evaluation:
2022-03-31 10:59:12.633178: W tensorflow/core/framework/allocator.cc:107] Allocation of 244940800 exceeds 10% of system memory.
2022-03-31 10:59:12.633178: W tensorflow/core/framework/allocator.cc:107] Allocation of 244121600 exceeds 10% of system memory.
2022-03-31 10:59:12.633178: W tensorflow/core/framework/allocator.cc:107] Allocation of 244531200 exceeds 10% of system memory.
2022-03-31T10:59:14.010859: step 50, loss 0.35193, acc 0.865

Saved model checkpoint to D:\大学\毕设\邮件分类\runs\1648695530\checkpoints\model-50

运行结果及报错内容

我的解答思路和尝试过的方法

我想要达到的结果

Answer 1

你自己设置的误差停止条件，刚好在50的时候满足了吧