-
Hello authors and anyone else who has tried dMaSIF,
do you ever face a CUDA memory leak issue?
I'm training it on a dataset of protein surface patches on an RTX 2080Ti and the GPU VRAM usage cr…
-
Hi,
I work in the bioinformatics field and I found your tool very interesting.
I saw that a reaction can be written as:
- `reactant>>product`
- `reactant>spectator>product`
In biology and b…
0m1n0 updated
12 months ago
-
Congrats for the great work! I got questions about the model pretraining.
I found that although this work emphasize the importance of how R-SMILES can boost the retrosynthesis prediction performanc…
-
Hi,
I am trying to run Retrosynthesis on a small sample of Building Blocks. From the published paper "AiZynthFinder: a fast, robust and flexible open-source software for retrosynthetic planning". …
-
Meeting follows on from June 2021 meeting at #44.
**When**: Tue July 6th 2pm London (BST).
**Where**: https://ucl.zoom.us/j/91614633419
**Recording**: https://youtu.be/lAWCUQGDGtY
**Actions from…
-
I have noted that you provided the use of AndOr_BreadthFirstSearch in ```1- end to end retrosynthesis with a toy model.ipynb``` and reproduced PaRoutes in ```2a- advanced algorithms on PaRoutes.ipynb`…
-
In preparation for our study on using a genetic algorithm to find good binders for MurC (more on that later), Casper Steinmann (@cstein) has docked 250K molecules from the ZINC database and 10K molecu…
-
This is the output on USPTO-15K dataset; a few questions on the output (I'm new to the pharma domain)
1. looks like this is a Retrosynthesis prediction task?
2. but whats with the question marks in …
-
How much of the Lowe USPTO dataset is in ORD? I saw in the 0.1.0 release note that you included the grants data. Does that mean there is no data from the applications (Like 2001-2016 applications from…
-
The molnet loader functions are currently divided into two groups with inconsistent APIs for specifying featurizers and splitters.
Most of them have arguments `featurizer` and `split` to specify th…