TypeError: 'DigitsBatchTrain' object is not iterable

[jianghonggang]

Extract the Chinese_dataset.rar to data folder and move all pictures to images, then modify data_test.txt to data_dev.txt.

run main.py, it shows that:

Creating directory if it does not exist: './checkpoint/densenet121_rmsprop_lr5.0e-05_wd5.0e-04_bsize64_imsize32' Using model from scratch (random weights) 'densenet121' Traceback (most recent call last): File "/home/luban/repository/crnn.pytorch/main.py", line 352, in main() File "/home/luban/repository/crnn.pytorch/main.py", line 197, in main loss = train(train_loader, model, criterion, optimizer, epoch) File "/home/luban/repository/crnn.pytorch/main.py", line 231, in train for i, (images, targets, target_lengths) in enumerate(train_loader): TypeError: 'DigitsBatchTrain' object is not iterable

Process finished with exit code 1

[zhiqwang]

Hi @jianghonggang , you should rename data_dev.txt to dev.txt and data_train.txt to train.txt, I forget to give this information.

[jianghonggang]

Hi @jianghonggang , you should modify _datadev.txt to dev.txt and _datatrain.txt to train.txt, I forget to give this information.

I have modified dataset.py 25 line: label_path = os.path.join(data_root, '{}.txt'.format(mode)) to label_path = os.path.join(dataroot, 'data{}.txt'.format(mode))

[jianghonggang]

好吧，北京的同学，我们还是中文交流吧

[zhiqwang]

好吧，北京的同学，我们还是中文交流吧

Jiang 准备跑什么数据，可否告知一下？

[jianghonggang]

现在跑的是Chinese_dataset.rar解压后的数据，还没上自己的训练数据。

[zhiqwang]

我现在没有硬件（gpu）测试模型，线上放的模型是一个 epoch 出来的结果（Chinese dataset 这个准确率是 97.5%），多跑两个 epoch 准确率应该有进一步提升，如果方便，可以 push 一下模型出来

[jianghonggang]

现在的问题是程序跑不起来，运行main.py时报错： TypeError: 'DigitsBatchTrain' object is not iterable

[zhiqwang]

现有的参数如果不收敛，可能需要加 criterion = nn.CTCLoss(zero_infinity=True) 需要最新的 pytroch nightly 版本。

[zhiqwang]

现在的问题是程序跑不起来，运行main.py时报错： TypeError: 'DigitsBatchTrain' object is not iterable

python 3.6+?

[jianghonggang]

python: 3.6.8 pytorch: 1.0.1

[zhiqwang]

--alphabet ./data/alphabet_decode_5990.txt 这个参数呢？

[jianghonggang]

对，是这个参数

[zhiqwang]

./data/images 文件夹下应该没有子文件夹

[jianghonggang]

修改main.py line 231:

for i, e in enumerate(train_loader):
    images = e.images
    targets = e.targets
    target_lengths = e.target_lengths

可以跑了

[zhiqwang]

修改main.py line 231:

for i, e in enumerate(train_loader):
    images = e.images
    targets = e.targets
    target_lengths = e.target_lengths

可以跑了

好诡异，不知道 python 有这个操作，估计是 DigitsBatchTrain 这个类写得不规范