search

MaybeShewill-CV / CRNN_Tensorflow

Convolutional Recurrent Neural Networks(CRNN) for Scene Text Recognition
MIT License
1.03k stars 388 forks source link

hcar_dict & ord_map #277

Closed khalyl-hamdy closed 5 years ago

khalyl-hamdy commented 5 years ago

hey

i want to use this project to create a model for the arabic language so i wonder can these two files be automatically generated? if yes how ?

thanks in advance

MaybeShewill-CV commented 5 years ago

@khalyl-hamd Those two files can be automatically generated when you create your own tensorflow records:)

khalyl-hamdy commented 5 years ago

one more question the Data (that i am using) contains set of images with their labels inside a txt files. i mean every image has its own txt file : both of them have the sme name is it the right form of data that i should use to convert the whole dataset into tensorflow records?

MaybeShewill-CV commented 5 years ago

@khalyl-hamdy You'd better convert your dataset into Synth90k dataset's format first and you can use the tools in this repo to generate the tfrecords. Otherwise you may have to implement your own data feed pip line:)

khalyl-hamdy commented 5 years ago

Well i prepared the data and created the TFrecords files then i tried to train the model but i got this message and i dont know what's wrong , can you check it for me (i am using tensorflow 1.10.0) ?

`I0619 10:06:19.095118 6471 train_shadownet.py:572] Use single gpu to train the model 2019-06-19 10:06:22.699408: I tensorflow/core/platform/cpu_feature_guard.cc:141] Your CPU supports instructions that this TensorFlow binary was not compiled to use: SSE4.1 SSE4.2 AVX AVX2 FMA 2019-06-19 10:06:22.857261: I tensorflow/core/common_runtime/process_util.cc:69] Creating new thread pool with default inter op setting: 2. Tune using inter_op_parallelism_threads for best performance. I0619 10:06:23.081531 6471 train_shadownet.py:271] Training from scratch 2019-06-19 10:06:28.731836: W tensorflow/core/framework/op_kernel.cc:1275] OP_REQUIRES failed at ctc_loss_op.cc:166 : Invalid argument: Saw a non-null label (index >= num_classes - 1) following a null label, batch: 0 num_classes: 37 labels: 27 Traceback (most recent call last): File "/home/khalyl/anaconda3/envs/rcnnenv/lib/python3.6/site-packages/tensorflow/python/client/session.py", line 1278, in _do_call return fn(*args) File "/home/khalyl/anaconda3/envs/rcnnenv/lib/python3.6/site-packages/tensorflow/python/client/session.py", line 1263, in _run_fn options, feed_dict, fetch_list, target_list, run_metadata) File "/home/khalyl/anaconda3/envs/rcnnenv/lib/python3.6/site-packages/tensorflow/python/client/session.py", line 1350, in _call_tf_sessionrun run_metadata) tensorflow.python.framework.errors_impl.OutOfRangeError: End of sequence [[Node: val_IteratorGetNext = IteratorGetNextoutput_shapes=[[4,32,100,3], , [4]], output_types=[DT_FLOAT, DT_VARIANT, DT_STRING], _device="/job:localhost/replica:0/task:0/device:CPU:0"]]

During handling of the above exception, another exception occurred:

Traceback (most recent call last): File "tools/train_shadownet.py", line 578, in need_decode=args.decode_outputs File "tools/train_shadownet.py", line 324, in train_shadownet [optimizer, train_ctc_loss, merge_summary_op]) File "/home/khalyl/anaconda3/envs/rcnnenv/lib/python3.6/site-packages/tensorflow/python/client/session.py", line 877, in run run_metadata_ptr) File "/home/khalyl/anaconda3/envs/rcnnenv/lib/python3.6/site-packages/tensorflow/python/client/session.py", line 1100, in _run feed_dict_tensor, options, run_metadata) File "/home/khalyl/anaconda3/envs/rcnnenv/lib/python3.6/site-packages/tensorflow/python/client/session.py", line 1272, in _do_run run_metadata) File "/home/khalyl/anaconda3/envs/rcnnenv/lib/python3.6/site-packages/tensorflow/python/client/session.py", line 1291, in _do_call raise type(e)(node_def, op, message) tensorflow.python.framework.errors_impl.OutOfRangeError: End of sequence [[Node: val_IteratorGetNext = IteratorGetNextoutput_shapes=[[4,32,100,3], , [4]], output_types=[DT_FLOAT, DT_VARIANT, DT_STRING], _device="/job:localhost/replica:0/task:0/device:CPU:0"]]

Caused by op 'val_IteratorGetNext', defined at: File "tools/train_shadownet.py", line 578, in need_decode=args.decode_outputs File "tools/train_shadownet.py", line 159, in train_shadownet batch_size=CFG.TRAIN.BATCH_SIZE File "/media/khalyl/b19f6211-f6a7-443d-8a50-5c247986129e/khalyl/Desktop/CRNN/CRNN_Tensorflow-master/data_provider/shadownet_data_feed_pipline.py", line 289, in inputs num_threads=CFG.TRAIN.CPU_MULTI_PROCESS_NUMS File "/media/khalyl/b19f6211-f6a7-443d-8a50-5c247986129e/khalyl/Desktop/CRNN/CRNN_Tensorflow-master/data_provider/tf_io_pipline_fast_tools.py", line 378, in inputs return iterator.get_next(name='{:s}_IteratorGetNext'.format(self._dataset_flag)) File "/home/khalyl/anaconda3/envs/rcnnenv/lib/python3.6/site-packages/tensorflow/python/data/ops/iterator_ops.py", line 410, in get_next name=name)), self._output_types, File "/home/khalyl/anaconda3/envs/rcnnenv/lib/python3.6/site-packages/tensorflow/python/ops/gen_dataset_ops.py", line 2069, in iterator_get_next output_shapes=output_shapes, name=name) File "/home/khalyl/anaconda3/envs/rcnnenv/lib/python3.6/site-packages/tensorflow/python/framework/op_def_library.py", line 787, in _apply_op_helper op_def=op_def) File "/home/khalyl/anaconda3/envs/rcnnenv/lib/python3.6/site-packages/tensorflow/python/util/deprecation.py", line 454, in new_func return func(*args, **kwargs) File "/home/khalyl/anaconda3/envs/rcnnenv/lib/python3.6/site-packages/tensorflow/python/framework/ops.py", line 3155, in create_op op_def=op_def) File "/home/khalyl/anaconda3/envs/rcnnenv/lib/python3.6/site-packages/tensorflow/python/framework/ops.py", line 1717, in init self._traceback = tf_stack.extract_stack()

OutOfRangeError (see above for traceback): End of sequence [[Node: val_IteratorGetNext = IteratorGetNextoutput_shapes=[[4,32,100,3], , [4]], output_types=[DT_FLOAT, DT_VARIANT, DT_STRING], _device="/job:localhost/replica:0/task:0/device:CPU:0"]]`

MaybeShewill-CV commented 5 years ago

@khalyl-hamdy Your config file is wrongly generated please check yourself:)

khalyl-hamdy commented 5 years ago

well i only changed the train.batch_size to 4 what are the other parameters do you think i need to change besides the batch_size?

thanx alot

MaybeShewill-CV commented 5 years ago

@khalyl-hamdy class nums should also be adjusted according to your own dataset:)

khalyl-hamdy commented 5 years ago

i changed that , Also i don't have the same images shape (they have the same height with different width ) , is this a problem ?

MaybeShewill-CV commented 5 years ago

@khalyl-hamdy They will be resized into the same shape:)

khalyl-hamdy commented 5 years ago

well i changed the parameters and i still get the same error everytime i try to start the training , i couldn't figure where's the source of this error.

OutOfRangeError: End of sequence [[Node: val_IteratorGetNext = IteratorGetNextoutput_shapes=[[4,32,100,3], , [4]], output_types=[DT_FLOAT, DT_VARIANT, DT_STRING], _device="/job:localhost/replica:0/task:0/device:CPU:0"]]

And should i use the 1.12.0 version of tensorflow or even have lower ones it doesn't bother ?

MaybeShewill-CV commented 5 years ago

@khalyl-hamdy My local tensorflow-gpu version is 1.12.0:)

khalyl-hamdy commented 5 years ago

Do you think it maybe the cause of the error ?

MaybeShewill-CV commented 5 years ago

@khalyl-hamdy No I do not think the problem is caused by the tensorlfow version:)

khalyl-hamdy commented 5 years ago

well i will try to train the english dataset and see if i am going to get the same error or not. thank you very much:)

MaybeShewill-CV commented 5 years ago

@khalyl-hamdy I recommend you to train the synth90k dataset according to readme file first which helps you understand the whole project and use your own data to train the model later:)