| --- |
| license: cc-by-nc-4.0 |
| metrics: |
| - accuracy |
| tags: |
| - biology |
| base_model: |
| - alchemab/antiberta2-cssp |
| library_name: transformers |
| --- |
| ImmunoMatch is a protein language model finetuned from AntiBERTa2, aiming at investigating the heavy and light chain pairing preferences in antibody. The input sequence to the model should be a pair of sequences of VH and VL domains. |
|
|
| Different variants of ImmunoMatch is available according to the use of interest: |
|
|
| | Checkpoint name | Trained on | |
| | --------------- | ---------- | |
| | [ImmunoMatch](https://huggingface.co/fraternalilab/immunomatch) | A mixture of antibodies with both 魏 and 位 light chains | |
| | [ImmunoMatch-魏](https://huggingface.co/fraternalilab/immunomatch-kappa) | Antibodies with 魏 light chains | |
| | [ImmunoMatch-位](https://huggingface.co/fraternalilab/immunomatch-lambda) | Antibodies with 位 light chains | |
