ocr_recognition预测问题

[xiegegege]

1.ctc模型预测报错： Traceback (most recent call last): File "infer.py", line 153, in main() File "infer.py", line 149, in main inference(args) File "infer.py", line 34, in inference get_feeder_data = get_ctc_feeder_for_infer NameError: name 'get_ctc_feeder_for_infer' is not defined 2.attention模型预测报错： Traceback (most recent call last): File "infer.py", line 153, in main() File "infer.py", line 149, in main inference(args) File "infer.py", line 93, in inference return_numpy=False) File "/opt/_internal/cpython-3.6.0/lib/python3.6/site-packages/paddle/fluid/executor.py", line 565, in run use_program_cache=use_program_cache) File "/opt/_internal/cpython-3.6.0/lib/python3.6/site-packages/paddle/fluid/executor.py", line 642, in _run exe.run(program.desc, scope, 0, True, True, fetch_var_name) paddle.fluid.core.EnforceNotMet: Invoke operator sequence_expand error. Python Callstacks: File "/opt/_internal/cpython-3.6.0/lib/python3.6/site-packages/paddle/fluid/framework.py", line 1654, in append_op attrs=kwargs.get("attrs", None)) File "/opt/_internal/cpython-3.6.0/lib/python3.6/site-packages/paddle/fluid/layer_helper.py", line 43, in append_op return self.main_program.current_block().append_op(*args, **kwargs) File "/opt/_internal/cpython-3.6.0/lib/python3.6/site-packages/paddle/fluid/layers/nn.py", line 3940, in sequence_expand attrs={'ref_level': ref_level}) File "/ssd1/xiege/model/models-develop/PaddleCV/ocr_recognition/attention_model.py", line 218, in simple_attention x=decoder_state_proj, y=encoder_proj) File "/ssd1/xiege/model/models-develop/PaddleCV/ocr_recognition/attention_model.py", line 279, in attention_infer decoder_size) File "infer.py", line 44, in inference ids = infer(images, num_classes, use_cudnn=True if args.use_gpu else False) File "infer.py", line 149, in main inference(args) File "infer.py", line 153, in main() C++ Callstacks: Enforce failed. Expected x_dims[0] == static_cast(y_lod[ref_level].size()) - 1, but received x_dims[0]:2 != static_cast(y_lod[ref_level].size()) - 1:1. When Input(X)'s lod is null, the dims[0] of Input(X) should match the size of Input(Y)'s referred level lod. at [/paddle/paddle/fluid/operators/sequence_ops/sequence_expand_op.cc:76]

[L-lei]

能问下，你的ctc模型能训练吗，或者在哪里修改了，我用attention训练后预测也是出现的一模一样的错误，paddle我用的1.4.1，python3.7

[wanghaoshuang]

@L-lei https://github.com/PaddlePaddle/models/blob/develop/PaddleCV/ocr_recognition/infer.py#L3 这一行需要加上get_ctc_feeder_for_infer

[wanghaoshuang]

第2个错是因为beam size == 2的情况，网络配置有问题。建议先设置beam size为1.

[L-lei]

非常感谢，attention错误成功解决了，

[L-lei]

用ctc网络训练出现下面的错误，请问您有遇到吗 Python Callstacks: File "D:\develop\Python\Python37\lib\site-packages\paddle\fluid\framework.py", line 1654, in append_op attrs=kwargs.get("attrs", None)) File "D:\develop\Python\Python37\lib\site-packages\paddle\fluid\layer_helper.py", line 43, in append_op return self.main_program.current_block().append_op(*args, **kwargs) File "D:\develop\Python\Python37\lib\site-packages\paddle\fluid\layers\nn.py", line 5230, in warpctc 'use_cudnn': use_cudnn File "G:\project\paddlepaddle\test\ocr_lpr\crnn_ctc_model.py", line 188, in ctc_train_net input=fc_out, label=label, blank=num_classes, norm_by_times=True) File "G:/project/paddlepaddle/test/ocr_lpr/train.py", line 80, in train args, data_shape, num_classes) File "G:/project/paddlepaddle/test/ocr_lpr/train.py", line 236, in main train(args) File "G:/project/paddlepaddle/test/ocr_lpr/train.py", line 240, in main() File "D:\develop\IntelliJ IDEA 2018.2\myplugins\python\helpers\pydev_pydev_imps_pydev_execfile.py", line 18, in execfile exec(compile(contents+"\n", file, 'exec'), glob, loc) File "D:\develop\IntelliJ IDEA 2018.2\myplugins\python\helpers\pydev\pydevd.py", line 1068, in run pydev_imports.execfile(file, globals, locals) # execute the script File "D:\develop\IntelliJ IDEA 2018.2\myplugins\python\helpers\pydev\pydevd.py", line 1658, in main globals = debugger.run(setup['file'], None, None, is_module) File "D:\develop\IntelliJ IDEA 2018.2\myplugins\python\helpers\pydev\pydevd.py", line 1664, in main() C++ Callstacks: Tensor holds the wrong type, it holds 3 at [D:\1.4.1\paddle\paddle/fluid/framework/tensor_impl.h:42] PaddlePaddle Call Stacks:

[wanghaoshuang]

用ctc网络训练出现下面的错误，请问您有遇到吗 Python Callstacks: File "D:\develop\Python\Python37\lib\site-packages\paddle\fluid\framework.py", line 1654, in append_op attrs=kwargs.get("attrs", None)) File "D:\develop\Python\Python37\lib\site-packages\paddle\fluid\layer_helper.py", line 43, in append_op return self.main_program.current_block().append_op(*args, **kwargs) File "D:\develop\Python\Python37\lib\site-packages\paddle\fluid\layers\nn.py", line 5230, in warpctc 'use_cudnn': use_cudnn File "G:\project\paddlepaddle\test\ocr_lpr\crnn_ctc_model.py", line 188, in ctc_train_net input=fc_out, label=label, blank=num_classes, norm_by_times=True) File "G:/project/paddlepaddle/test/ocr_lpr/train.py", line 80, in train args, data_shape, num_classes) File "G:/project/paddlepaddle/test/ocr_lpr/train.py", line 236, in main train(args) File "G:/project/paddlepaddle/test/ocr_lpr/train.py", line 240, in main() File "D:\develop\IntelliJ IDEA 2018.2\myplugins\python\helpers\pydev_pydev_imps_pydev_execfile.py", line 18, in execfile exec(compile(contents+"\n", file, 'exec'), glob, loc) File "D:\develop\IntelliJ IDEA 2018.2\myplugins\python\helpers\pydev\pydevd.py", line 1068, in run pydev_imports.execfile(file, globals, locals) # execute the script File "D:\develop\IntelliJ IDEA 2018.2\myplugins\python\helpers\pydev\pydevd.py", line 1658, in main globals = debugger.run(setup['file'], None, None, is_module) File "D:\develop\IntelliJ IDEA 2018.2\myplugins\python\helpers\pydev\pydevd.py", line 1664, in main() C++ Callstacks: Tensor holds the wrong type, it holds 3 at [D:\1.4.1\paddle\paddle/fluid/framework/tensor_impl.h:42] PaddlePaddle Call Stacks:

warpctc op的某个输入类型不对

[L-lei]

我感觉很疑惑，一般来说，错误类型是这样的，向Tensor holds the wrong type, it holds int at，但是这里是‘3’，完全不能理解，输入数据reader中是正常的，所以就是crnn_ctc_model网络哪里出的问题

[wanghaoshuang]

你确认下你的label的类型？是int64，还是 int32?