Text-based de novo Molecule Generation on ChEBI-20

Metric: Levenshtein (higher is better)

LeaderboardDataset

Loading chart...

Results

Sort:

#	Model↕	Levenshtein▼	Extra Data	Paper	Date↕	Code
1	Text+Chem T5 small	28.54	No	Unifying Molecular and Textual Representations v...	2023-01-29	Code
2	Text+Chem T5 base	27.39	No	Unifying Molecular and Textual Representations v...	2023-01-29	Code
3	GIT-Mol-caption	26.315	No	GIT-Mol: A Multi-modal Large Language Model for ...	2023-08-14	Code
4	MolT5-small	25.988	No	Translation between Molecules and Natural Language	2022-04-25	Code
5	MolReGPT (GPT-3.5-turbo)	24.91	No	Empowering Molecule Discovery for Molecule-Capti...	2023-06-11	Code
6	MolT5-base	24.458	No	Translation between Molecules and Natural Language	2022-04-25	Code
7	Text+Chem T5-augm small	21.78	No	Unifying Molecular and Textual Representations v...	2023-01-29	Code
8	MolFM-Small	20.868	No	MolFM: A Multimodal Molecular Foundation Model	2023-06-06	Code
9	MolFM-Base	19.445	No	MolFM: A Multimodal Molecular Foundation Model	2023-06-06	Code
10	MolReGPT (GPT-4-0413)	17.14	No	Empowering Molecule Discovery for Molecule-Capti...	2023-06-11	Code
11	TGM-DLM	17.003	No	Text-Guided Molecule Generation with Diffusion L...	2024-02-20	Code
12	TGM-DLM w/o corr	16.897	No	Text-Guided Molecule Generation with Diffusion L...	2024-02-20	Code
13	Text+Chem T5-augm base	16.87	No	Unifying Molecular and Textual Representations v...	2023-01-29	Code
14	MolT5-Large-HV	16.758	No	Translation between Molecules and Natural Language	2022-04-25	Code
15	MolT5-Large	16.07	No	Translation between Molecules and Natural Language	2022-04-25	Code
16	BioT5	15.097	No	BioT5: Enriching Cross-modal Integration in Biol...	2023-10-11	Code
17	BioT5+	12.776	No	BioT5+: Towards Generalized Biological Understan...	2024-02-27	Code
18	MolReFlect	11.84	No	-	-	-
19	LDMol	6.75	No	LDMol: Text-to-Molecule Diffusion Model with Str...	2024-05-28	Code

#1Text+Chem T5 smallSOTA
28.54
Levenshtein· 2023-01-29
Unifying Molecular and Textual Representations via Multi-task Language Modelling Code
#2Text+Chem T5 base
27.39
Levenshtein· 2023-01-29
Unifying Molecular and Textual Representations via Multi-task Language Modelling Code
#3GIT-Mol-caption
26.315
Levenshtein· 2023-08-14
GIT-Mol: A Multi-modal Large Language Model for Molecular Science with Graph, Image, and Text Code
#4MolT5-smallSOTA
25.988
Levenshtein· 2022-04-25
Translation between Molecules and Natural Language Code
#5MolReGPT (GPT-3.5-turbo)
24.91
Levenshtein· 2023-06-11
Empowering Molecule Discovery for Molecule-Caption Translation with Large Language Models: A ChatGPT Perspective Code
#6MolT5-base
24.458
Levenshtein· 2022-04-25
Translation between Molecules and Natural Language Code
#7Text+Chem T5-augm small
21.78
Levenshtein· 2023-01-29
Unifying Molecular and Textual Representations via Multi-task Language Modelling Code
#8MolFM-Small
20.868
Levenshtein· 2023-06-06
MolFM: A Multimodal Molecular Foundation Model Code
#9MolFM-Base
19.445
Levenshtein· 2023-06-06
MolFM: A Multimodal Molecular Foundation Model Code
#10MolReGPT (GPT-4-0413)
17.14
Levenshtein· 2023-06-11
Empowering Molecule Discovery for Molecule-Caption Translation with Large Language Models: A ChatGPT Perspective Code
#11TGM-DLM
17.003
Levenshtein· 2024-02-20
Text-Guided Molecule Generation with Diffusion Language Model Code
#12TGM-DLM w/o corr
16.897
Levenshtein· 2024-02-20
Text-Guided Molecule Generation with Diffusion Language Model Code
#13Text+Chem T5-augm base
16.87
Levenshtein· 2023-01-29
Unifying Molecular and Textual Representations via Multi-task Language Modelling Code
#14MolT5-Large-HV
16.758
Levenshtein· 2022-04-25
Translation between Molecules and Natural Language Code
#15MolT5-Large
16.07
Levenshtein· 2022-04-25
Translation between Molecules and Natural Language Code
#16BioT5
15.097
Levenshtein· 2023-10-11
BioT5: Enriching Cross-modal Integration in Biology with Chemical Knowledge and Natural Language Associations Code
#17BioT5+
12.776
Levenshtein· 2024-02-27
BioT5+: Towards Generalized Biological Understanding with IUPAC Integration and Multi-task Tuning Code
#18MolReFlect
11.84
Levenshtein
No paper
#19LDMol
6.75
Levenshtein· 2024-05-28
LDMol: Text-to-Molecule Diffusion Model with Structurally Informative Latent Space Code