Closed hec44 closed 3 years ago
Hi, thanks for letting me know, I haven't really tried inference on a large external dataset, so this is a nice opportunity to fix any potential issues :)
I've tried running your code with the 1000 provided sentence but I can't reproduce the error, it runs without any problem. Could you provide a runnable code that breaks, just as you described?
Here is the code I put together in Google Colab.
Hello :), I am getting the next runtime error:
Issue
Code
The code that I am using that generates this error is the following, which is heavily based on the code that you provided on the google colab guide.
Description of usage
I am trying to parse a document that contains 94,000 lines in English. I can parse the first 14k sentences without any problem. However after that point I get the error. I have also tried to parse the 94k sentences all at the same time, But I get an error
sentence 763 not matching companion after anchor computation.
in the same line:output = batches.dataset.datasets[dataset.framework_to_id[("ucca", language)]].data
This is one of the 1000 sentences that I pass to the function:
I can parse these sentences using the code on google colab or if I parse these sentences as the first sentences in the document. I hope you can help to find the error in my usage.
Thank you in advance for any help :D