Emmanuel Noutahi

Hello @colinrsmall, You are right about this being the original desired behaviour. The `enforce_subs` should be interpreted as `enforce substitutions` (keep vectors attached to the core when they match any...

Hi @tjkessler, I just created a [feedstock](https://conda-forge.org/docs/maintainer/adding_pkgs.html) for padelpy on [conda-forge](https://github.com/conda-forge/staged-recipes/pull/17341) and added you as the package maintainer. I am currently waiting for tests to pass and the "0.1.12" version...

I think the bug is coming from these lines: https://github.com/rdkit/rdkit/blob/0370b6cc57b279cd25a677b28f766dfe2a05021d/rdkit/Chem/PandasTools.py#L447-L449 the renderer is set to the default panda renderer, while it should be `patchPandasrepr`. The following works and allows switching...

Note, that in a lot of cases, you might want to use `pd.json_normalize` instead.

@YBooks , @SaulLu , @sgugger can we reopen this issue, since https://github.com/huggingface/transformers/pull/18408 creates another one ?

Thanks @Pakman450 > Question: Does the output results have multi valued attributes? Currently the output is a tuple of list and dicts. I would say for the refactoring, it's perfectly...

Hello @kheyer, do you still have the weight of a pre-trained model ? @QuickLearner171998 did you manage to get a working version ?

@QuickLearner171998 can you share the pretrained model you have ? does it have good performance ?

I hereby give my explicit permission to switch to the BSD 3-clause license.