EATLM
EATLM copied to clipboard
Code for 'On Pre-trained Language Models For Antibody'
Hi, thanks for your work! I'm exploring the 'antibody\utils\preprocess.py' script and would like to inquire about the expected format of the 'cell.jsonl' file processed in the function at line 104....
Through the understanding and using of the code in the project, there are some obstacles encountered in the process of replicating the task, can the pre-training weights be shared?
大佬您这个工作做了很大的贡献,但我在处理下游任务数据时遇到了一些问题,在parapred的预测时您是如何将pdb转为序列的呢?我参考了parapred这个原始工作的代码,但是结果不是很相同,存在有有些标记cdrs没有。同时在Bcell处理时我发现有很多的重复序列以及相同的序列存在不同的标签但这应该是但标签多分类任务,我不知道应该如何处理这些问题,您有空能提供一些建议么?能否提供notebook或者相关代码非常感谢。