search

clovaai / deep-text-recognition-benchmark

Text recognition (optical character recognition) with deep learning methods, ICCV 2019
Apache License 2.0
3.76k stars 1.11k forks source link

Testing on a custom dataset #430

Open Jason7624 opened 1 month ago

Jason7624 commented 1 month ago

Hi, I've been having this problem for testing my trained model on my own custom dataset. I've been able to train my model with the provided code to format into lmdb. However, when I'm trying to test my trained model, this error shows up:

model input parameters 32 100 20 1 512 256 37 25 TPS ResNet BiLSTM CTC loading pretrained model from saved_models/Test_pretrained/best_accuracy.pth deep-text-recognition-benchmark\test.py:208: FutureWarning: You are using torch.load with weights_only=False (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for weights_only will be flipped to True. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via torch.serialization.add_safe_globals. We recommend you start setting weights_only=True for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature. model.load_state_dict(torch.load(opt.saved_model, map_location=device)) 'cp' is not recognized as an internal or external command, operable program or batch file.

dataset_root: Dataset/Validation\IIIT5k_3000 dataset: / Traceback (most recent call last): File "deep-text-recognition-benchmark\test.py", line 282, in test(opt) File "deep-text-recognition-benchmark\test.py", line 226, in test benchmark_all_eval(model, criterion, converter, opt) File "deep-text-recognition-benchmark\test.py", line 46, in benchmark_all_eval eval_data, eval_data_log = hierarchical_dataset(root=eval_data_path, opt=opt) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "deep-text-recognition-benchmark\dataset.py", line 124, in hierarchical_dataset concatenated_dataset = ConcatDataset(dataset_list) ^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "paddleocr1\Lib\site-packages\torch\utils\data\dataset.py", line 328, in init assert len(self.datasets) > 0, "datasets should not be an empty iterable" # type: ignore[arg-type] ^^^^^^^^^^^^^^^^^^^^^^ AssertionError: datasets should not be an empty iterable

And this is my script: python test.py --eval_data Dataset/Validation --benchmark_all_eval --Transformation TPS --FeatureExtraction ResNet --SequenceModeling BiLSTM --Prediction CTC --saved_model saved_models/Test_pretrained/best_accuracy.pth

Can anyone please help me resolve my issue?

AdamKheire commented 3 weeks ago

Hi, I've been having this problem for testing my trained model on my own custom dataset. I've been able to train my model with the provided code to format into lmdb. However, when I'm trying to test my trained model, this error shows up:

model input parameters 32 100 20 1 512 256 37 25 TPS ResNet BiLSTM CTC

loading pretrained model from saved_models/Test_pretrained/best_accuracy.pth deep-text-recognition-benchmark\test.py:208: FutureWarning: You are using torch.load with weights_only=False (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for weights_only will be flipped to True. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via torch.serialization.add_safe_globals. We recommend you start setting weights_only=True for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature. model.load_state_dict(torch.load(opt.saved_model, map_location=device)) 'cp' is not recognized as an internal or external command, operable program or batch file.

dataset_root: Dataset/Validation\IIIT5k_3000 dataset: / Traceback (most recent call last): File "deep-text-recognition-benchmark\test.py", line 282, in test(opt) File "deep-text-recognition-benchmark\test.py", line 226, in test benchmark_all_eval(model, criterion, converter, opt) File "deep-text-recognition-benchmark\test.py", line 46, in benchmark_all_eval eval_data, eval_data_log = hierarchical_dataset(root=eval_data_path, opt=opt) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "deep-text-recognition-benchmark\dataset.py", line 124, in hierarchical_dataset concatenated_dataset = ConcatDataset(dataset_list) ^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "paddleocr1\Lib\site-packages\torch\utils\data\dataset.py", line 328, in init assert len(self.datasets) > 0, "datasets should not be an empty iterable" # type: ignore[arg-type] ^^^^^^^^^^^^^^^^^^^^^^ AssertionError: datasets should not be an empty iterable

And this is my script: python test.py --eval_data Dataset/Validation --benchmark_all_eval --Transformation TPS --FeatureExtraction ResNet --SequenceModeling BiLSTM --Prediction CTC --saved_model saved_models/Test_pretrained/best_accuracy.pth

Can anyone please help me resolve my issue?

use the same number used for training --output_channel 256 or 512 --hidden_size