yangheng95
Variation in output by adding spaces
@yangheng95 - please see the below issue
Text1 = """That is the question I asked then this morning. But most important how do they continue with no support options for clients for problems with these changes."""
Text2 = """That is the question I asked then this morning. But most important how do they continue with no support options for clients for problems with these changes."""
The only difference between text 1 and text 2 is the space after 'But'. For text2 i am getting the error, for text1 its working fine.
Below is my code.
`triplet_extractor = ASTE.AspectSentimentTripletExtractor("multilingual")
text1 = """That is the question I asked then this morning. But most important how do they continue with no support options for clients for problems with these changes."""
text2 = """That is the question I asked then this morning. But most important how do they continue with no support options for clients for problems with these changes."""
triplet_extractor.predict(text2)
`
{'text': 'That is the question I asked then this morning. But most important how do they continue with no support options for clients for problems with these changes.',
'Triplets': '[]',
'error': "can't assign a NoneType to a torch.LongTensor",
'error_type': 'RuntimeError'}
Yes, there are some problems to be solved because the quadrule sentiment extraction is an experimental feature. Please try to make some efforts to solve it, and it will be excellent if you would like to share your solution.
After all, I will try to optimize it in future. Provided that I got a easy schedule.
