How can you achieve the accuracy reported
HansiZeng opened this issue · comments
Hansi Zeng commented
For example, I run the TextAttnBiLSTM model in SST2 dataset with the following command line,
python main.py run --status='train' --use_model="TextAttnBiLSTM"
However, I can just get accuracy: 84.295 in the testing set but not 85.632 as you reported. Is anything tricky I miss?
Thanks.
Gordon Lee commented
Sorry for misleading you. The project is a toy and out-of-date example about text classification. There may be many factors that make the difference.