swansonk14

Hi @Jameel9, Thank you for the suggestion! That's a good idea! If you're able to create a PR to implement those features, that would be much appreciated. If not, I'll...

Hi @Jameel9, Yes, if you're able to create a PR, I'd still very much appreciate it! Best, Kyle

Hi @davide-grheco, Thank you so much for creating this PR and doing all the work to transition to Chemprop v2! I'll aim to work on checking the code and training...

Hi @manglav, Thank you for checking on this! Unfortunately my other research projects have taken priority recently and it will take me a while to work through all of the...

Hi @GattiMh, Thanks for raising this issue! It looks like you might be running into version issues with the packages that ADMET-AI relies on. Can you try running `pip install...

Thank you all for your discussion of this issue! I made a fix to set `torch==2.5.0` in this commit (https://github.com/swansonk14/admet_ai/commit/d9a082c3e38f05290ef9d8962b243e94ba6d80e8), which will be included in ADMET-AI version 1.4.0 shortly.

Hi Tuan, Thank you for raising this issue! There are unfortunately some cases where ADMET-AI will predict biologically infeasible values, such as a negative half-life. This is because for regression-valued...

@davide-grheco I agree! Right now, the clipping to 0 is done only for visualization purposes on the website. The raw negative values are preserved when downloading the results from the...

Hi @JieHou-SLU, Thank you for bringing up this issue! The -1 means that the building block SMILES was not found in the mapping from SMILES to IDs (see [here](https://github.com/swansonk14/SyntheMol/blob/main/synthemol/generate/generator.py#L207)). This...

Hi @JieHou-SLU, Thank you for providing the additional information and apologies for the slow response! You are correct to note that the number of unique building block SMILES in the...