Using single model.
Task: propid
Allow new words in test data: True
Embedding size=100
Read 0 sentences.
Data loading duration was 0:00:15.
Traceback (most recent call last):
File "python/predict.py", line 161, in <module>
args.input)
File "python/predict.py", line 66, in get_scores
test_data = data.get_test_data(test_sentences, batch_size=config.dev_batch_size)
File "/Users/maazfarooqi/Downloads/deep_srl-master/python/neural_srl/shared/tagger_data.py", line 75, in get_test_data
max_len = max([len(s[0]) for s in test_sentences])
ValueError: max() arg is an empty sequence
Using an ensemble of 5 models
Task: srl
Allow new words in test data: True
Data loading duration was 0:00:00.
Traceback (most recent call last):
File "python/predict.py", line 161, in <module>
args.input)
File "python/predict.py", line 49, in get_scores
allow_new_words)
File "/Users/maazfarooqi/Downloads/deep_srl-master/python/neural_srl/shared/reader.py", line 236, in get_srl_test_data
samples = get_srl_sentences(filepath, config.use_se_marker)
File "/Users/maazfarooqi/Downloads/deep_srl-master/python/neural_srl/shared/reader.py", line 48, in get_srl_sentences
predicate = int(lefthand_input[0])
ValueError: invalid literal for int() with base 10: '{\\rtf1\\ansi\\ansicpg1252\\cocoartf1671'
I'm getting the same error. But that is because the preprocess script is unable to create one or more data files. I'm yet to figure out why that is happening.