jinyuoba
commented
1 year ago Thank you for your reply. My data is all one-dimensional data with a length of 256, some with labels and some without labels. I would like to utilize my unlabeled data by using SSL. What does it mean to treat it as table data, and specifically, what should I do？
李津宇 @.***
------------------ 原始邮件 ------------------ 发件人: "YGZWQZD/LAMDA-SSL" @.>; 发送时间: 2023年5月17日(星期三) 中午1:21 @.>; @.**@.>; 主题: Re: [YGZWQZD/LAMDA-SSL] How to do a classification task for my 1D sequence data (Issue #5)
Hello, I'm not quite sure what type of sequence you're referring to. If you're looking to classify sequences of equal length, you can consider treating them as tabular data. However, if you're dealing with sequences of unequal lengths or text sequences, I recommend using deep SSL algorithms with neural network models that can handle sequential data, such as Transformers, RNNs.
— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.***>
Hello, I'm not quite sure what type of sequence you're referring to. If you're looking to classify sequences of equal length, you can consider treating them as tabular data. However, if you're dealing with sequences of unequal lengths or text sequences, I recommend using deep SSL algorithms with neural network models that can handle sequential data, such as Transformers, RNNs.