search

Josie-xufan / LaDM3IL

A Label Disambiguation-Based Multimodal Massive Multiple Instance Learning Approach for Immune Repertoire Classification
2 stars 0 forks source link

How to obtain input files for train_multimodal.py from raw CMV data? #1

Open juhengwei opened 1 month ago

juhengwei commented 1 month ago

Hi, I have the raw CMV dataset, but I am having trouble figuring out how to generate the input files required for train_multimodal.py. The script expects the following files: vocab_3mer.pkl, gene.csv, train.tsv, val.tsv, and test.tsv. However, I couldn't find any instructions in the paper or the repository on how to generate these files from the raw data. Could you provide some guidance or instructions on how to create these input files from the original dataset?

Thank you in advance for your help! 图片

Josie-xufan commented 1 month ago

Dear hengwei,

I'm sorry for the delay of the response. Since recently I have just taken a medical operation, I cannot look through email frequently. Today, I received the message from my supervisor on the wechat and then I found the email you sent me several days ago. I will check the work of LaDM3IL and then response your problem as soon as possible. Thanks again for your interest of the work!

Best regards, Fan Xu ---- Replied Message ---- | From | @.> | | Date | 10/13/2024 16:45 | | To | @.> | | Cc | @.***> | | Subject | [Josie-xufan/LaDM3IL] How to obtain input files for train_multimodal.py from raw CMV data? (Issue #1) |

Hi, I have the raw CMV dataset, but I am having trouble figuring out how to generate the input files required for train_multimodal.py. The script expects the following files: vocab_3mer.pkl, gene.csv, train.tsv, val.tsv, and test.tsv. However, I couldn't find any instructions in the paper or the repository on how to generate these files from the raw data. Could you provide some guidance or instructions on how to create these input files from the original dataset?

Thank you in advance for your help! default.png (view on web)

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you are subscribed to this thread.Message ID: @.***>