Comments (8)
Hi,
In fact, our molecular data is homologous to molT5, both derived from PubChem.
This essentially means that our data encompasses the data from molT5.
from mol-instructions.
Hi,
The data for MolT5 comes from the ChEBI annotations in PubChem, and this data is already included in Mol-Instructions.
from mol-instructions.
yeah I see, but when processing data, some filtering methods may lead to a loss of information. In Pubchem, there are descriptions of molecular structures,like 'Trans-anol is a phenylpropanoid that is phenol substituted at position 4 by a prop-1-enyl group (the trans-isomer) It is a phenylpropanoid and a member of phenols. It is functionally related to a phenol.' but in the instruction dataset , it is 'The molecule is a natural product found in Solidago odora and Osmorhiza aristata with data available' . And why not use IUPAC name instead of SELFIES? IUPAC name contains a description of the structure.
from mol-instructions.
The information you're referencing is described in the instructions as follows:
{
"instruction": "Synthesize a molecule that matches the given characteristics.",
"input": "The molecule is a phenylpropanoid that is phenol substituted at position 4 by a prop-1-enyl group (the trans-isomer) It is a phenylpropanoid and a member of phenols. It is functionally related to a phenol.",
"output": "[C][/C][=C][/C][=C][C][=C][Branch1][Branch1][C][=C][Ring1][=Branch1][O]",
"metadata": {
"task": "description-guided molecule design"
}
},
Our approach follows the methodology employed by MolT5:
We opted for SELFIES because this descriptor is more conducive to model learning and facilitates the generation of valid molecules. Of course, the IUPAC name you mentioned can also be considered as a direction for future attempts. Thank you for your suggestion!
from mol-instructions.
yeah, but in the description_guided_moolecule_design.json, line2374976 ,it is ' {
"instruction": "Based on the given information, design a molecule that meets the desired specifications.",
"input": "The molecule is a natural product found in Solidago odora and Osmorhiza aristata with data available.",
"output": "[C][/C][=C][/C][=C][C][=C][Branch1][Branch1][C][=C][Ring1][=Branch1][O]",
"metadata": {
"task": "description-guided molecule design",
"split": "train"
}
},'
from mol-instructions.
This is because our data comes not only from ChEBI annotations but also from other data sources within PubChem.
from mol-instructions.
I see! Thanks!
from mol-instructions.
You're welcome! Feel free to discuss and exchange ideas.
from mol-instructions.
Related Issues (13)
- Hyperparameter HOT 7
- Question about protein design HOT 3
- Cannot inference with zjunlp/llama-molinst-protein-7b HOT 7
- LORA Settings HOT 4
- pyarrow.lib.ArrowInvalid when loading data HOT 2
- [Request] Release the part of test-split for each task HOT 3
- about test data
- Requirements HOT 1
- Regarding some questions about the paper. HOT 3
- can not find config.json HOT 17
- Training Time HOT 1
- TypeError: __init__() got an unexpected keyword argument 'enable_lora' HOT 2
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from mol-instructions.