Jin Su
报这个错是因为没有找到搜索框，你确认一下你安装的各个库的版本是不是和requirements里一致呢？ 
你好，关于如何获取私聊信息，你可以参考代码文件里的``check_new_msg()``函数，这个函数用于自动抓取新消息。至于如何获取群聊中被@的消息，暂时还没有相关代码。这个比较复杂，我不一定会有时间实现这个。
你好，固定时间的配置是可以的。不过保存微信路径感觉有点不必要。我推荐一个小技巧，你可以在桌面上创建一个微信的快捷方式，这样每次只需要在桌面上找就可以了。
Hi, You indeed could load models and datasets from huggingface. But maybe it is more suitable for you to use the ColabSaprot interface while fine-tuning model with your local GPU....
Sure. We provide some functions to generate the LMDB file, either from a dictionary or from a file. You could check ``utils/generate_lmdb.py`` and refer to the function ``jsonl2lmdb``. This function...
Hi, Yes. If you input combinatorial mutations to the model, the predicted score is the sum of all single point mutations scores. Here we refer to the original paper [ESM-1v](https://www.biorxiv.org/content/10.1101/2021.07.09.450648v2.full)...
Yes, I think the point 2 is possible, as shown by the downstream task of thermostability. You just need to prepare some samples with labels of interest and use them...
Yes. If you already have the fine-tuned model you can just follow the second way to predict the fitness. In fact the first way to predict mutational effect doesn't require...
We don't provide a function call to directly get outputs from a fine-tuned model. However, you could easily load your model and manually get it. Here we use a regression...
不知道控件树的改动大不大，如果大的话可是一个大工程了，不一定有时间做 :(