search

YZ20211221 / DeepFormer

5 stars 0 forks source link

Running maize example #1

Open sohyunbk opened 1 year ago

sohyunbk commented 1 year ago

Hi DeepFormer team,

I am Sohyun Bang at University of Georgia, trying to use DeepFormer. I downloaded all the maize data you uploaded and ran "DanQ_train.py". I have some questions about the output I got.

  1. When I read "test_targets.npz", it seems like all the target value was "0". They do not include any "1", so the prediction is all nearly "0". I did not edit any values in "DanQ_YAML.yml". Is it correct..?

  2. "selene_sdk.train_model.validation.txt" file only write loss values. Average_precision and roc_auc columns have only "NA". With the same reasons I guess, "test_performance.txt" wrote class name and "NA". Is it correct?

Thanks for your help.

Best, Sohyun

YZ20211221 commented 1 year ago

Dear Sohyun. Thank you very much for your interest in our models. I apologize for not being able to respond to your question in a timely manner as I have been quite busy recently. Regarding your question, I guess it is a learning rate issue which can be solved by adjusting the learning rate. We have not encountered this problem during our testing. If you are still unable to solve the problem, please contact me. Best