-
Notifications
You must be signed in to change notification settings - Fork 30
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
CLIC new samples with 1M events #181
Conversation
jpata
commented
Mar 20, 2023
- update CLIC sample stats to 1M (TF dataset version to 1.3.0)
parameters/clic.yaml
Outdated
dtype: float32 | ||
trainable: | ||
classification_loss_type: sigmoid_focal_crossentropy | ||
lr_schedule: none # cosinedecay, exponentialdecay, onecycle, none | ||
lr_schedule: onecycle # cosinedecay, exponentialdecay, onecycle, none |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Using the TF model, I get slightly better results using cosinedecay over onecycle.
dtype: float16 | ||
trainable: | ||
classification_loss_type: sigmoid_focal_crossentropy | ||
lr_schedule: none # cosinedecay, exponentialdecay, onecycle, none | ||
lr_schedule: onecycle # cosinedecay, exponentialdecay, onecycle, none |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Using the TF model, I get slightly better results using cosinedecay over onecycle.
* clic new samples * fix torch_geometric to 2.2.0 avoid issues with dataset abstract method len
* clic new samples * fix torch_geometric to 2.2.0 avoid issues with dataset abstract method len Former-commit-id: e422e35
* clic new samples * fix torch_geometric to 2.2.0 avoid issues with dataset abstract method len Former-commit-id: e422e35
* clic new samples * fix torch_geometric to 2.2.0 avoid issues with dataset abstract method len