-
Notifications
You must be signed in to change notification settings - Fork 58
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Saint model training issue #154
Comments
python wandb_ssaint_train.py --dataset_name=assist2009 --model_name=ssaint --seed=3407 --dropout=0.3 --use_wandb=0 --add_uuid=0 Traceback (most recent call last): |
It seems like you change the saint into ssaint, have u insert the ssaint into the pykt/init_model.py for defining a new model? |
elif model_name == "ssaint": |
How about the print results of model_name (wandb_trian.py line 103) and model (wandb_trian.py line 105) ? |
model_name:ssaint and model is None |
The model init was not run successfully. Is any print statement "The wrong model name was used..." when you run the wandb_ssaint_train.py? If it does, there may be some errors exist in your model definition. You can check wandb_train.py to see whether all the variables are stated when calling init_model function. |
I did encounter the error message "The wrong model name was used..." but I haven't been able to pinpoint the exact cause. Therefore, I attempted to directly copy my model into the saint.py file. After running the command "python wandb_saint_train.py --dataset_name=assist2009 --use_wandb=0 --add_uuid=0 --seed=3407," I noticed that the original SAINT model and the model I copied yield identical AUC and ACC values, with no changes whatsoever. |
I modified the Saint model structure in the saint.py file. However, when I call the same command 'python wandb_saint_train.py --dataset_name=assist2009 --seed=3407 --dropout=0.3 --use_wandb=0 --add_uuid=0' before and after the modification, the accuracy and AUC remain exactly the same. Why is that?
The text was updated successfully, but these errors were encountered: