拟合期间Keras淹没Jupyter细胞输出(详细= 1) [英] Keras floods Jupyter cell output during fit (verbose=1)

问题描述

当在Jupyter笔记本中使用"verbose = 1"选项运行keras模型时，我开始不再像以前那样获得单行进度状态更新，而是大量更新状态行.见附图.重新启动jupyter或浏览器无济于事. Jupyter笔记本服务器是:5.6.0，keras是2.2.2，Python是Python 3.6.5 请帮助.

When running keras model inside Jupyter notebook with "verbose=1" option, I started getting not single line progress status updates as before, but a flood of status lines updated at batch. See attached picture. Restarting jupyter or the browser is not helping. Jupyter notebook server is: 5.6.0, keras is 2.2.2, Python is Python 3.6.5 Please help.

单元格内容:

history = model.fit(x = train_df_scaled，y = train_labels，batch_size = BATCH_SIZE，epochs = EPOCHS， verbose = 1， validation_data =(validation_df_scaled，validation_labels)，shuffle = True )

history = model.fit(x=train_df_scaled, y=train_labels, batch_size=BATCH_SIZE, epochs=EPOCHS, verbose=1, validation_data=(validation_df_scaled, validation_labels), shuffle=True)

输出洪水示例:(成千上万行这样的内容)

output flood example: (it is thousands of lines like this)

损失:217.5794-均值绝对误差:11.166-预计损失:32:42-损失:216.9500-均值绝对错误:11.165-预计损失:32:21-损失:216.6378-均值绝对错误:11.164-预计损失时间:32:00-损失: 216.0345-mean_absolute_error:11.164-ETA:31:41-损失:215.6621-mean_absolute_error:11.166-ETA:31:21-损失:215.4639-mean_absolute_error:11.171-ETA:31:02-损失:215.1654-mean_absolute_error:11.173-ETA: 30:44-损失:214.6583-mean_absolute_error:11.169-ETA:30:27-损失:213.8844-mean_absolute_error:11.164-ETA:30:10-损失:213.3308-mean_absolute_error:11.163-ETA:29:54-损失:213.1179- mean_absolute_error:11.167-ETA:29:37-损失:212.8138-mean_absolute_error:11.169-ETA:29:25-损失:212.7157-mean_absolute_error:11.174-ETA:29:11-损失:212.5421-mean_absolute_error:11.177-ETA:28: 56-损失:212.1867-mean_absolute_error:11.178-ETA:28:42-损失:211.8032-mean_absolute_error:11.180-ETA:28:28-损失:211.4079-mean_absolute_er ror:11.179-ETA:28:15-损失:211.2733-mean_absolute_error:11.182-ETA:28:02-损失:210.8588-mean_absolute_error:11.179-ETA:27:50-损失:210.4498-mean_absolute_error:11.178-ETA:27: 37-损失:209.9327-mean_absolute_error:11.176-ETA:27:

推荐答案

经过几次测试，我发现该错误与tqdm导入有关. Tqdm被用在一段代码中，后来没有进行重写.即使我不在此笔记本中使用tqdm，将其导入也会影响keras的输出. 为了解决这个问题，我只是注释掉了这一行: 从tqdm导入tqdm 一切顺利，有不错的keras进度条.虽然不确定它与keras到底有多冲突...

After a few tests I found that the error is related to tqdm import. Tqdm was used in a piece of code which was later rewritten withoout it. Even though I was not using tqdm in this notebook, just having it imported affected the keras output. To fix it I just commented out this line: from tqdm import tqdm and everything went fine, with nice keras progress bars. Not sure how exactly it conflicted with keras though...

这篇关于拟合期间Keras淹没Jupyter细胞输出(详细= 1)的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！