HongxinXiang/ImageMol
ImageMol is a molecular image-based pre-training deep learning framework for computational drug discovery.
PythonMIT
Issues
- 17
mismatch between the model architecture expected by your code and the state dictionary keys present in the checkpoint file I am trying to load.
#20 opened by AbdulmajeedCE - 1
The mismatch in the number of classes of output dimension of the jigsaw puzzle prediction task
#18 opened by ldkien-304 - 1
Why did the authors choose to compare the ImageMol model, which was pretrained on 10 million data points from the PubChem dataset, with the ImageMol_NonPretrained model that uses a ResNet18 architecture with randomly initialized weights? It seems potentially unfair as the non-pretrained model might be at a disadvantage. Additionally, could you clarify how the ImageMol_NonPretrained model achieved a higher AUC than Jure’s GNN in some cases? Thank you so much
#19 opened by ldkien-304 - 0
running imagemol on gpu
#17 opened by SoodabehGhaffari - 6
issue with ImageMol performance
#16 opened by SoodabehGhaffari - 1
关于MPP下游任务实验结果的复现
#8 opened by MrsW6 - 3
hyperparameter optimization
#13 opened by SoodabehGhaffari - 4
issue with new evaluate.py
#15 opened by SoodabehGhaffari - 1
Documentation improvement
#10 opened by Zainab-ik - 2
- 4
pretrained model
#14 opened by SoodabehGhaffari - 1
pretrained model
#1 opened by gao-qs - 1
where is your pre-train model?
#2 opened by songyu2022 - 1
issue with evaluate.py
#12 opened by SoodabehGhaffari - 0
tsne of molecular representation
#11 opened - 2
Finetune ImageMol with multiclass labels
#7 opened by pykao - 4
Reproducing Results for Finetuning ImageMol on GPCR assays and Sars-CoV2 assays
#6 opened by DhanshreeA - 2
feature extraction script?
#5 opened by pykao - 2
IndexError for getting labels
#3 opened by ddayzzz