Multitask Text and Chemistry T5 Base Augm Open-Source Language Model - Solving Multi-Domain Tasks in Chemistry and Natural Language

Multitask Text And Chemistry T5 Base Augm

Developed by GT4SD

A multi-domain, multi-task language model designed to address a wide range of tasks in chemistry and natural language fields.

Molecular Model

Transformers

EnglishOpen Source License:MIT #Chemistry-Text Cross-Modality #Multi-Task Joint Training #Molecular Description Generation

Downloads 11.01k

Release Time : 5/19/2023

Model Overview

This model is based on the T5-small architecture and trained by integrating multiple task datasets, covering various chemistry and text tasks such as forward reaction prediction, retrosynthetic analysis, and molecular description generation.

Model Features

Multi-Domain Unified Representation

Handles both chemical molecular representations and natural language text simultaneously through multi-task learning.

Enhanced Dataset Training

Trained using an extended chemistry-text hybrid dataset.

Cross-Task Generalization Capability

A single model supports various cheminformatics tasks such as chemical reaction prediction and molecular generation.

Model Capabilities

Chemical reaction forward prediction

Retrosynthetic route analysis

Molecular description generation

Text-conditioned molecular generation

Chemical text conversion

Use Cases

Drug Development

Molecular Structure Generation

Automatically generates candidate drug molecular structures based on text descriptions

Chemical Education

Reaction Mechanism Explanation

Automatically generates textual descriptions of chemical reaction processes

🚀 Multitask Text and Chemistry T5

Multitask Text and Chemistry T5 is a multi - domain, multi - task language model designed to tackle a wide variety of tasks in both the chemical and natural language domains.

🚀 Quick Start

Multitask Text and Chemistry T5 is a powerful language model that can handle diverse tasks across chemical and natural language fields. It was published by Christofidellis et al..

✨ Features

Multi - domain and Multi - task: Capable of solving tasks in both chemical and natural language domains.
Transformer - based: Built on the Transformer architecture, trained on a multi - domain and multi - task dataset.

📚 Documentation

Model Details

The Multitask Text and Chemistry T5 variant is trained using t5 - small as its pretrained base and the augmented dataset.

Developers

Dimitrios Christofidellis*, Giorgio Giannone*, Jannis Born, Teodoro Laino and Matteo Manica from IBM Research and Ole Winther from Technical University of Denmark.

Distributors

The model is natively integrated into GT4SD.

Model Date

Model Type

A Transformer - based language model that is trained on a multi - domain and a multi - task dataset by aggregating available datasets for the tasks of Forward reaction prediction, Retrosynthesis, Molecular captioning, Text - conditional de novo generation and Paragraph to actions.

Information about training algorithms, parameters, fairness constraints or other applied approaches, and features

N.A.

Paper or other resource for more information

The Multitask Text and Chemistry T5 Christofidellis et al.(2023)

Where to send questions or comments about the model

Open an issue on [GT4SD repository](https://github.com/GT4SD/gt4sd - core).

📄 License

This project is licensed under the MIT License.

📚 Citation

@inproceedings{christofidellis2023unifying,
  title = 	 {Unifying Molecular and Textual Representations via Multi - task Language Modelling},
  author =       {Christofidellis, Dimitrios and Giannone, Giorgio and Born, Jannis and Winther, Ole and Laino, Teodoro and Manica, Matteo},
  booktitle = 	 {Proceedings of the 40th International Conference on Machine Learning},
  pages = 	 {6140--6157},
  year = 	 {2023},
  volume = 	 {202},
  series = 	 {Proceedings of Machine Learning Research},
  publisher =    {PMLR},
  pdf = 	 {https://proceedings.mlr.press/v202/christofidellis23a/christofidellis23a.pdf},
  url = 	 {https://proceedings.mlr.press/v202/christofidellis23a.html},
}

*equal contribution

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base

English 简体中文繁體中文にほんご