Error on WordBeamSearch Inference

[j-paris]

Hello,

I am encountering some problems in running the code in Inference Mode and specifically with the decoder type 'wordbeamsearch'. I have successfully cloned the repository CTCWordBeamSearch (https://github.com/githubharald/CTCWordBeamSearch) and the test script attached correctly runs.

However, when I try to make inference on the provided test image (word.png) with the given net weights, I get the following error:

Traceback (most recent call last): File "main.py", line 194, in <module> main() File "main.py", line 190, in main infer(model, args.img_file) File "main.py", line 133, in infer recognized, probability = model.infer_batch(batch, True) File "/opt/notebooks/HTR/SimpleHTR/src/model.py", line 294, in infer_batch loss_vals = self.sess.run(eval_list, feed_dict) File "/root/.conda/envs/MXNetGPU/lib/python3.7/site-packages/tensorflow/python/client/session.py", line 968, in run run_metadata_ptr) File "/root/.conda/envs/MXNetGPU/lib/python3.7/site-packages/tensorflow/python/client/session.py", line 1167, in _run (np_val.shape, subfeed_t.name, str(subfeed_t.get_shape()))) ValueError: Cannot feed value of shape (0,) for Tensor 'Placeholder_1:0', which has shape '(None, 2)' Segmentation fault (core dumped)

Please notice that there may some inconsistence on the code lines number only due to a few simple comments I have written in the script.

Can someone provide me some hints to solve this problem? Thank you in advance.

Regards Jacopo Paris

[githubharald]

this already happens before even calling the word beam search decoder. Can you checkout exactly the code version as it is on github and then just run python main.py?

[j-paris]

Thank you Harald, I have recloned the repository from scratch and now it seems to be working!

Thank you for the support. Jacopo Paris