Zhifeng Gao
Zhifeng Gao
Hi Cyrus, Thanks for your advices, we will add more examples and demo in future updates, especailly for inferencing outiside of pre-computed datasets. For fast implements, you can change the...
Do you already set your use dir path with `--use-dir`? also you can check if unimol,unicore is installed successfully.
Currently we do not support unimol plus when used in the wrapper version. However, you can easily integrate unimol plus to unimol_tools with minimal modifications in unimol_tools/model and data.
Hi wklix, If your target values are either large or small, with high skew or kurt values, I think some target normalization should be helpful. unimol tools use a simple...
Hi, you can use the latest image(unimol-qsar:v0.5), multi-task regression bugs should fixed in the latest version.
BTW, it's possible to enhance your data preprocessing by incorporating domain expertise, this may involve manual normalization of the target variable, anomaly detection, and other specialized techniques. unimol have some...
Hi Monica, This error is caused by model training failure, you can check your data or remove this try catch for debug. https://github.com/dptech-corp/Uni-Mol/blob/main/unimol_tools/unimol_tools/models/nnmodel.py#L118
can you provide more details？（cuda version, c10d error info etc.)
seems some unused modules in your network, you can add `--find-unused-parameters` to avoid this.
> hi team. I'm trying to create docking interaction database similar to yours. in clustering coords M=1000, N=100. is this necessary? Would it be ok to say M=100 and N=10?...