Why training an ITER is over

[XXMxxm220]

Why training an ITER is over？

[XXMxxm220]

训练一个epoch后 不再继续

[zhanghaowei01]

您好，问题解决了没

[XXMxxm220]

您好，问题解决了没

解决啦

[zhanghaowei01]

分享一下呗

[XXMxxm220]

分享一下呗

你好，我采用的方法是在for it 前，加入了for epoch in range epochs：；epochs是我自己设定的训练周期，再将save_model语句也包含在for epoch 内，保证每个epch保存一次模型。

[CoinCheung]

Why do you need this? You can compute the total iterations by len(dataset) * n_epoches / batch_size, and set the total iterations as this computation result.

[XXMxxm220]

Why do you need this? You can compute the total iterations by , and set the total iterations as this computation result.len(dataset) * n_epoches / batch_size

You are right .Thank you for your response.I just wanted to use epoch to mean a training, not iters，so I choose that way. This approach also does not affect the final output