Downloads · 30 days
67
1% of all-time downloads
PDBEurope/BiomedNLP-PubMedBERT-ProteinStructure-NER-v2.1
BiomedNLP-PubMedBERT-ProteinStructure-NER-v2.1 is a token classification model from PDBEurope. Use it when you need labels on individual words, such as names. It is set up for transformers. The card lists the license as mit.
Downloads · 30 days
67
1% of all-time downloads
All-time downloads
6.9K
Public
Parameters
109M
871 MB on disk
Likes
2
Public
Click a slice to open those files.
.bin436 MB · 50%
How the weights are stored.
F32109M · 100%
From the Hugging Face model README
| Feature | Description |
|---|---|
| Name | BiomedNLP-PubMedBERT-ProteinStructure-NER-2.1 |
| Default Pipeline | transformer, ner |
| Components | transformer, ner |
| Vectors | 0 keys, 0 unique vectors (0 dimensions) |
| Sources | n/a |
| License | n/a |
| Author | Melanie Vollmar |
| Component | Labels |
|---|---|
ner | "bond_interaction", "chemical", "complex_assembly", "evidence", "experimental_method", "gene", "mutant", "oligomeric_state", "protein", "protein_state", "protein_type", "ptm", "residue_name", "residue_name_number", "residue_number", "residue_range", "site", "species", "structure_element", "taxonomy_domain" |
| entity type | precision | recall | F1 | sample number |
|---|---|---|---|---|
| "bond_interaction" | 0.93 | 0.88 | 0.90 | 43 |
| "chemical" | 0.89 | 0.91 | 0.90 | 761 |
| "complex_assembly" | 0.91 | 0.93 | 0.92 | 288 |
| "evidence" | 0.84 | 0.88 | 0.86 | 390 |
| "experimental_method" | 0.85 | 0.85 | 0.85 | 357 |
| "gene" | 0.79 | 0.86 | 0.82 | 43 |
| "mutant" | 0.91 | 0.97 | 0.94 | 463 |
| "oligomeric_state" | 0.93 | 0.99 | 0.96 | 142 |
| "protein" | 0.94 | 0.97 | 0.95 | 1411 |
| "protein_state" | 0.83 | 0.88 | 0.85 | 546 |
| "protein_type" | 0.85 | 0.85 | 0.85 | 414 |
| "ptm" | 0.70 | 0.70 | 0.70 | 50 |
| "residue_name" | 0.92 | 0.97 | 0.94 | 90 |
| "residue_name_number" | 0.95 | 0.96 | 0.96 | 561 |
| "residue_number" | 0.80 | 0.97 | 0.88 | 33 |
| "residue_range" | 0.81 | 0.70 | 0.75 | 43 |
| "site" | 0.85 | 0.87 | 0.86 | 270 |
| "species" | 0.94 | 0.96 | 0.95 | 84 |
| "structure_element" | 0.91 | 0.92 | 0.92 | 992 |
| "taxonomy_domain" | 0.99 | 0.98 | 0.98 | 88 |
The dataset can be found here: https://huggingface.co/datasets/PDBEurope/protein_structure_NER_model_v2.1
Vollmar, M., Tirunagari, S., Harrus, D. et al. Dataset from a human-in-the-loop approach to identify functionally important protein residues from literature. Sci Data 11, 1032 2024 https://doi.org/10.1038/s41597-024-03841-9