stevezheng23 / xlnet_extension_tf

XLNet Extension in TensorFlow

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Can’t run CoQa train script in Google Colab

imagine3D-ai opened this issue · comments

I tried using Google Colab CPU and GPU notebooks to train XLNET on COQA, but they keep crashing because of the ’out-of-memory’ issues. I tried reducing batch size to 1, but the problem still persists. Did anyone else face similar issues and was able to solve it?

this might be related to issue you're discussing in #79

this might be related to issue you're discussing in #

this might be related to issue you're discussing in #79

Have you also experienced this bug (infinite loop somewhere in the code) ?

this might be related to issue you're discussing in #

this might be related to issue you're discussing in #79

Have you also experienced this bug (infinite loop somewhere in the code) ?

Actually, I can't reproduce the issue you and @abhishek-jha13 reported

commented

this might be related to issue you're discussing in #79

Is max_sequence_length measured in characters or in words?

this might be related to issue you're discussing in #79

Is max_sequence_length measured in characters or in words?

should be subwords

Hi together

I had the same problem. The issue is in convert_coqa_example. The while loop in line 818 run_coqa.py becomes an endless loop if max_sequence_length is 128.
The training data sample 244 has a query length of 128. Accordingly, the following line
max_para_length = self.max_seq_length - len(query_tokens) - 3
ends up producing a max_para_length of -3 and the sliding_window_approach stops working.
My simple solution was to increase the max_seq_length to 256.