loss is not decreasing

[Transtars]

Dear Authors,

I hope this message finds you well. Recently, I've had the privilege to work on the implementation of your project. I have been diligently conducting experiments as per the method outlined in your paper. For my setup, I utilized the IP dataset with 240 instances for training and 60 instances for validation. I've set the batch size at 8, maxStoredSol at 500, and a learning rate of 0.003. I've noticed that the model's loss stabilizes around 341 without further decrease. I was wondering if this might be attributed to my choice of hyperparameters or perhaps the size of the training dataset.

I would greatly appreciate your insights or suggestions on this matter. Thank you for your time and consideration.

Warm regards,

Transtars

[PKUTAN]

Dear Authors,

I hope this message finds you well. Recently, I've had the privilege to work on the implementation of your project. I have been diligently conducting experiments as per the method outlined in your paper. For my setup, I utilized the IP dataset with 240 instances for training and 60 instances for validation. I've set the batch size at 8, maxStoredSol at 500, and a learning rate of 0.003. I've noticed that the model's loss stabilizes around 341 without further decrease. I was wondering if this might be attributed to my choice of hyperparameters or perhaps the size of the training dataset.

I would greatly appreciate your insights or suggestions on this matter. Thank you for your time and consideration.

Warm regards,

Transtars

How do you get the dataset? Thx!

[Transtars]

Dear Authors,
I hope this message finds you well. Recently, I've had the privilege to work on the implementation of your project. I have been diligently conducting experiments as per the method outlined in your paper. For my setup, I utilized the IP dataset with 240 instances for training and 60 instances for validation. I've set the batch size at 8, maxStoredSol at 500, and a learning rate of 0.003. I've noticed that the model's loss stabilizes around 341 without further decrease. I was wondering if this might be attributed to my choice of hyperparameters or perhaps the size of the training dataset.
I would greatly appreciate your insights or suggestions on this matter. Thank you for your time and consideration.
Warm regards,
Transtars

How do you get the dataset? Thx!

I'm sorry for not being able to reply in a timely manner. I downloaded the dataset from this website: https://drive.google.com/file/d/1MytdY3IwX_aFRWdoc0mMfDN9Xg1EKUuq/view
This link was found from this website:
https://github.com/ds4dm/ml4co-competition/blob/main/DATA.md