InstructMol-G

Reported on 6 benchmarks across 1 task · 1 paper

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Medical6 results

Molecule CaptioningonChEBI-20
BLEU-2· 2023-11-27
46.6
best: 73.2 (Mol-LLM (Mistral-Instruct-v0.2))
InstructMol: Multi-Modal Integration for Building a Versatile and Reliable Molecular Assistant in Drug Discovery arXiv:2311.16208
Molecule CaptioningonChEBI-20
BLEU-4· 2023-11-27
36.5
best: 60.8 (MolReFlect)
InstructMol: Multi-Modal Integration for Building a Versatile and Reliable Molecular Assistant in Drug Discovery arXiv:2311.16208
Molecule CaptioningonChEBI-20
METEOR· 2023-11-27
49.1
best: 68.1 (BioT5+)
InstructMol: Multi-Modal Integration for Building a Versatile and Reliable Molecular Assistant in Drug Discovery arXiv:2311.16208
Molecule CaptioningonChEBI-20
ROUGE-1· 2023-11-27
54.7
best: 71 (BioT5+)
InstructMol: Multi-Modal Integration for Building a Versatile and Reliable Molecular Assistant in Drug Discovery arXiv:2311.16208
Molecule CaptioningonChEBI-20
ROUGE-2· 2023-11-27
36.5
best: 61.8 (XMolCap)
InstructMol: Multi-Modal Integration for Building a Versatile and Reliable Molecular Assistant in Drug Discovery arXiv:2311.16208
Molecule CaptioningonChEBI-20
ROUGE-L· 2023-11-27
47.9
best: 65.3 (PEIT-GEN)
InstructMol: Multi-Modal Integration for Building a Versatile and Reliable Molecular Assistant in Drug Discovery arXiv:2311.16208