Wang Huijuan

Thank you for your answer! I manully changed feature matrix to sparse matrix, and it can run, but it's too slow that I don't even know if I have other...

And I wonder if the feature is not binary and sparse, is convert it to sparse matrix (maybe not binary) useful? Or I have to use binary and sparse matrix?

想请问修改长度这部分具体的代码实现有实例，或者可参考的具体修改方法吗？

I used `conda install conda-pack` to install this package. And the codes to reproduce other parts of this environment may be hard, coz it took a long time for me...

OK. I send this pull request: https://github.com/pyg-team/pytorch_geometric/pull/5503

不是，比如说我之前保存到卡1上了，我现在设置CUDA_VISIBLE_DEVICES=0了，如果不设置map_location的话，它还是会默认调用到卡1上，但是我现在就没有卡1了，就会报错

I ran `conda install conda==22.11` in `base` environment, and it just shows me many many reports... The first part has diminished. The last reports: ``` - typed-ast -> libgcc-ng[version='>=7.5.0'] ->...

I tried `conda update -n base -c defaults conda --repodata-fn=repodata.json` And it seems works for me.