Closed NathanYanJing closed 2 years ago
Downolad dataset in GitHub page. Unzip the downloaded dataset, and move it to the project folder.
python -m s4.train --dataset listops-classification --model s4 --epochs 100 --bsz 50 --d_model 128 --n_layers 4 --ssm_n 64 --lr 1e-2 --p_dropout 0 --lr_schedule
Training with the previously defined hyper-parameters yields the test accuracy 54.3% on the test set.
We used the huggingface datasets for LRA IMDB review task.
python -m s4.train --dataset imdb-classification --model s4 --epochs 100 --bsz 50 --d_model 64 --n_layers 4 --ssm_n 64 --lr 1e-2 --p_dropout 0
Training with the previously defined hyper-parameters yields the test accuracy 80.7% on the test set.
Here is the training curve for IMDB Review.
Ok, working on it.
Added some review comments. I think there was a merge failure, so if you can try again to pull.
Also be sure to run make autoformat
to check the formatting.
Results of ListOps:
Adding two tasks of LRA benchmark.