Update OFoldX model card metadata

README.md

@@ -43,6 +43,15 @@ before using the artifact for scientific or operational decisions.
 
 Boltz-2-shaped structure-prediction model converted for the OFoldX runtime.
 
+Converted Boltz-2 structure-prediction checkpoint for biomolecular interaction modeling.
+
+### Model Provenance
+
+- **Upstream Project**: Boltz-2
+- **Source Checkpoint**: `boltz2_conf_final.ckpt`
+- **Primary Paper**: [Boltz-2: Towards Accurate and Efficient Binding Affinity Prediction](https://doi.org/10.1101/2025.06.14.659707)
+- **Upstream License**: MIT for upstream Boltz code and weights
+
 ### Model Specification
 
 | Field | Value |
@@ -54,10 +63,16 @@ Boltz-2-shaped structure-prediction model converted for the OFoldX runtime.
 | Entrypoint | `ofoldx.models.boltz2.model.Boltz2Model` |
 | Config | `config.json` |
 | Weights | `model.safetensors` |
+| Source Checkpoint | `boltz2_conf_final.ckpt` |
+
+> [!NOTE]
+> Source checkpoint: `boltz2_conf_final.ckpt`.
 
 ### Links
 
 - **Hub repository**: [oteam/boltz2](https://huggingface.co/oteam/boltz2)
+- **Upstream paper**: [Boltz-2: Towards Accurate and Efficient Binding Affinity Prediction](https://doi.org/10.1101/2025.06.14.659707)
+- **Upstream repository**: [Boltz-2](https://github.com/jwohlwend/boltz)
 - **Code**: [`ofoldx/models/boltz2/model.py`](https://github.com/OTeam-AI4S/OFoldX/tree/main/ofoldx/models/boltz2/model.py)
 - **Project repository**: [https://github.com/OTeam-AI4S/OFoldX](https://github.com/OTeam-AI4S/OFoldX)
 - **Issues**: [https://github.com/OTeam-AI4S/OFoldX/issues](https://github.com/OTeam-AI4S/OFoldX/issues)
@@ -92,8 +107,16 @@ processed batch to the model.
 
 ## Training Details
 
-The training data and training procedure are inherited from the upstream checkpoint release. They are not
-reproduced inside this OFoldX artifact unless additional source-specific metadata is provided in the files.
+OFoldX did not train these weights. This repository contains a converted checkpoint and OFoldX runtime
+metadata for loading it.
+
+### Training Data
+
+The upstream Boltz-2 report describes a biomolecular interaction training set with expanded structure data, molecular-dynamics ensembles, self-distilled predictions, and affinity supervision. OFoldX does not redistribute the training set.
+
+### Training Procedure
+
+OFoldX converts the upstream Boltz-2 checkpoint into `model.safetensors` plus an OFoldX manifest; it does not run Boltz-2 training.
 
 ## Evaluation
 
@@ -108,8 +131,17 @@ scientific evaluation should be checked against the corresponding upstream model
 
 ## Citation
 
-Please cite the upstream model or paper associated with the source checkpoint. If OFoldX supports your work,
-please also cite or link the OFoldX project repository.
+Please cite the upstream Boltz-2 work for the source checkpoint. If OFoldX supports your work, please also cite or link the OFoldX project repository.
+
+```bibtex
+@article{passaro2025boltz2,
+  author = {Passaro, Saro and Corso, Gabriele and Wohlwend, Jeremy and Reveiz, Mateo and Thaler, Stephan and Somnath, Vignesh Ram and Getz, Noah and Portnoi, Tally and Roy, Julien and Stark, Hannes and Kwabi-Addo, David and Beaini, Dominique and Jaakkola, Tommi and Barzilay, Regina},
+  title = {Boltz-2: Towards Accurate and Efficient Binding Affinity Prediction},
+  year = {2025},
+  doi = {10.1101/2025.06.14.659707},
+  journal = {bioRxiv}
+}
+```
 
 ## Contact
 
@@ -117,5 +149,5 @@ Please use [OFoldX GitHub issues](https://github.com/OTeam-AI4S/OFoldX/issues) f
 
 ## License
 
-The OFoldX project license is not yet finalized. The source checkpoint may carry additional upstream license
-terms; review both OFoldX and upstream terms before redistribution or production use.
+The OFoldX project license is not yet finalized.
+The source checkpoint is associated with the upstream license noted above: MIT for upstream Boltz code and weights. Review both OFoldX and upstream terms before redistribution or production use.