github-actions[bot]
commented
3 weeks ago Thank you for creating this issue! 🎉 We'll look into it as soon as possible. In the meantime, please make sure to provide all the necessary details and context. If you have any questions reach out to LinkedIn. Your contributions are highly appreciated! 😊
Note: This repo is for beginners to learn and start with Opensource we won't accept more than 10 issues from a single person, This restriction applies to Gssoc project which has a similar kind of adding folder files, Points will be reduced when we find Spam.
I Maintain the repo issue twice a day, or ideally 1 day, If your issue goes stale for more than one day you can tag and comment on this same issue.
You can also check our CONTRIBUTING.md for guidelines on contributing to this project.
"Drowsiness Detection using EEG signal" project and have encountered some ambiguities in the README.md file. I would appreciate it if you could provide clarifications on the following points:
Model Implementation - Model 1 and Model 2:
Model 1: The README mentions the architecture includes convolutional blocks with residual connections and a Global Average Pooling Layer, followed by dense layers. However, there is no mention of explicit regularization techniques or normalization methods used in this model. Could you provide more details on whether any regularization techniques (e.g., dropout, L2 regularization) or normalization methods (e.g., batch normalization) were applied during the training of this model? Model 2: It is mentioned that dropout was used to prevent overfitting. Could you specify the dropout rates applied to the different layers of the model? Additionally, the description mentions batch normalization was used. Could you provide more details on where batch normalization was applied within the model architecture? Training and Evaluation:
The README states that the CNN model was trained for 150 epochs with a batch size of 4. Was any form of early stopping implemented during the training process to prevent overfitting? If so, what were the criteria and parameters for early stopping (e.g., patience, monitored metric)? Regarding the data splitting process, the dataset was split into training and testing sets using an 80:20 ratio. Could you provide more details on how the splitting was conducted? For instance, was the splitting performed randomly, or was there a specific stratification technique used to ensure balanced distribution of classes in the training and testing sets? Results:
The README provides a comparison and conclusion based on three models. However, only two models are described in detail. Is there a third model that is part of this comparison, or is this a typographical error? If there is a third model, could you please provide the architectural details and training parameters for that model as well?