TMC-Llama: Transition Metal Complex SMILES Strings and Model Weights

Yunsheng Liu published a 2.0 GB dataset on figshare in May 2026 for TMC-Llama models. It includes Transition Metal Complex (TMC) data with TMC-SMILES strings for publication figures, a dataset for fine-tuning models, and the resulting PyTorch model weights.

Use Cases

• Fine-tune language models for chemistry based on TMC-SMILES strings.
• Generate publication-ready figures based on provided TMC data.
• Reproduce or build upon the TMC-Llama model using the included PyTorch weights.
• Train models for molecular property prediction specific to transition metal complexes.

Strengths

• Dataset size is 2.0 GB, indicating substantial content.
• Includes both training data (CSV/JSONL) and pre-trained model weights (PyTorch files).
• Released under a permissive CC-BY-4.0 license for open use.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Description metadata is limited; actual data quality requires manual inspection after download.

Provenance

Source

Yunsheng Liu via figshare

Freshness

Last updated 2026-05-13 05:32:55.