Publish OFoldX pipeline artifacts

README.md

@@ -6,11 +6,11 @@ tags:
   - "protein"
   - "rna"
   - "dna"
-  - "model"
+  - "pipeline"
   - "proteinmpnn"
   - "design-generation"
   - "protein-design"
-artifact_kind: "model"
+artifact_kind: "pipeline"
 repo_id: "oteam/proteinmpnn-membrane-per-residue-noise020"
 license: "mit"
 pipeline_tag: "other"
@@ -34,11 +34,11 @@ widget:
 
 # proteinmpnn-membrane-per-residue-noise020
 
-OFoldX `model` artifact for biomolecular design generation, using the `proteinmpnn` architecture.
+OFoldX `pipeline` artifact for biomolecular design generation, using the `proteinmpnn` architecture.
 
 ## Disclaimer
 
-This model card was generated by the OFoldX team for an OFoldX `model` artifact.
+This model card was generated by the OFoldX team for an OFoldX `pipeline` artifact.
 The upstream model authors did not write this card unless explicitly stated otherwise.
 
 OFoldX is pre-alpha research software. Check the source checkpoint, upstream release, and local validation
@@ -53,7 +53,6 @@ Converted membrane ProteinMPNN checkpoint with per-residue membrane labels.
 ### Model Provenance
 
 - **Upstream Project**: Membrane ProteinMPNN
-- **Source Checkpoint**: `per_residue_label_membrane_mpnn_v_48_020.pt`
 - **Source Release**: [https://github.com/dauparas/LigandMPNN](https://github.com/dauparas/LigandMPNN)
 - **Primary Paper**: [Robust deep learning-based protein sequence design using ProteinMPNN](https://doi.org/10.1126/science.add2187)
 - **Upstream License**: MIT for upstream ProteinMPNN and LigandMPNN code/model parameters
@@ -63,20 +62,13 @@ Converted membrane ProteinMPNN checkpoint with per-residue membrane labels.
 | Field | Value |
 | ----- | ----- |
 | Repository | `oteam/proteinmpnn-membrane-per-residue-noise020` |
-| Artifact Kind | `model` |
+| Artifact Kind | `pipeline` |
 | Task | `design_generation` |
 | Architecture | `proteinmpnn` |
-| Entrypoint | `ofoldx.models.proteinmpnn.model.ProteinMPNNDesignGenerator` |
-| Config | `config.json` |
-| Weights | `model.safetensors` |
-| Source Checkpoint | `per_residue_label_membrane_mpnn_v_48_020.pt` |
-| Checkpoint Variant | `per_residue_label_membrane_mpnn_v_48_020` |
-| k-Neighbors | `48` |
-| Node Feature Size | `3` |
-| Training Noise | `0.20 Angstrom` |
+| Entrypoint | `ofoldx.pipelines.design.DesignPipeline` |
 
 > [!NOTE]
-> Source checkpoint: `per_residue_label_membrane_mpnn_v_48_020.pt`; the public repo suffix records the upstream Gaussian noise variant (0.20 Angstrom).
+> Checkpoint metadata: `k_neighbors=48`; the `noiseXXX` suffix identifies the training-noise variant.
 
 ### Links
 
@@ -84,7 +76,7 @@ Converted membrane ProteinMPNN checkpoint with per-residue membrane labels.
 - **Upstream paper**: [Robust deep learning-based protein sequence design using ProteinMPNN](https://doi.org/10.1126/science.add2187)
 - **Upstream repository**: [Membrane ProteinMPNN](https://github.com/dauparas/LigandMPNN)
 - **Source checkpoint release**: [https://github.com/dauparas/LigandMPNN](https://github.com/dauparas/LigandMPNN)
-- **Code**: [`ofoldx/models/proteinmpnn/model.py`](https://github.com/OTeam-AI4S/OFoldX/tree/main/ofoldx/models/proteinmpnn/model.py)
+- **Code**: [`ofoldx/pipelines/design.py`](https://github.com/OTeam-AI4S/OFoldX/tree/main/ofoldx/pipelines/design.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)
 
@@ -96,19 +88,14 @@ The artifact depends on the [`ofoldx`](https://github.com/OTeam-AI4S/OFoldX) lib
 pip install ofoldx
 ```
 
-### Direct Use
+### Pipeline Usage
 
-Load the artifact from `oteam/proteinmpnn-membrane-per-residue-noise020` using the OFoldX `Auto*` interface:
+Load the artifact from `oteam/proteinmpnn-membrane-per-residue-noise020` with the OFoldX task pipeline. Use `AutoModel` or `AutoProcessor` only when you need lower-level control:
 
 ```python
 from ofoldx.pipelines import Pipeline
 
 pipeline = Pipeline.from_pretrained("oteam/proteinmpnn-membrane-per-residue-noise020")
-output = pipeline(...)
-
-from ofoldx import AutoModel
-
-model = AutoModel.from_pretrained("oteam/proteinmpnn-membrane-per-residue-noise020")
 ```
 
 When a matching processor is available, load it with `AutoProcessor.from_pretrained(...)` and pass the
@@ -117,7 +104,7 @@ processed batch to the model.
 ### Interface
 
 - **Task**: `design_generation`
-- **Artifact kind**: `model`
+- **Artifact kind**: `pipeline`
 - **Architecture**: `proteinmpnn`
 - **Runtime files**: `manifest.json`, `config.json`, and `model.safetensors` when present
 
@@ -168,5 +155,5 @@ Please use [OFoldX GitHub issues](https://github.com/OTeam-AI4S/OFoldX/issues) f
 
 ## License
 
-The OFoldX project license is not yet finalized.
+The Hub `license` metadata, when present, reflects the source checkpoint or upstream project license. The OFoldX project license is not yet finalized.
 The source checkpoint is associated with the upstream license noted above: MIT for upstream ProteinMPNN and LigandMPNN code/model parameters. Review both OFoldX and upstream terms before redistribution or production use.

generator/README.md

@@ -0,0 +1,172 @@
+---
+library_name: "ofoldx"
+tags:
+  - "biology"
+  - "biomolecular-design"
+  - "protein"
+  - "rna"
+  - "dna"
+  - "model"
+  - "proteinmpnn"
+  - "design-generation"
+  - "protein-design"
+artifact_kind: "model"
+repo_id: "oteam/proteinmpnn-membrane-per-residue-noise020"
+license: "mit"
+pipeline_tag: "other"
+task: "design_generation"
+model-index:
+  - name: "proteinmpnn-membrane-per-residue-noise020"
+    results:
+      []
+widget:
+  - pipeline_tag: "other"
+    task: "design_generation"
+    example_title: "Backbone sequence design"
+    text: "input_structure: backbone.cif\ndesign_chains: A"
+    input_format: "structure_path"
+  - pipeline_tag: "other"
+    task: "design_generation"
+    example_title: "Binder design"
+    text: "target_structure: target.cif\ntarget_chains: A\ndesign_chains: B"
+    input_format: "structure_path"
+---
+
+# proteinmpnn-membrane-per-residue-noise020
+
+OFoldX `model` artifact for biomolecular design generation, using the `proteinmpnn` architecture.
+
+## Disclaimer
+
+This model card was generated by the OFoldX team for an OFoldX `model` artifact.
+The upstream model authors did not write this card unless explicitly stated otherwise.
+
+OFoldX is pre-alpha research software. Check the source checkpoint, upstream release, and local validation
+before using the artifact for scientific or operational decisions.
+
+## Model Details
+
+ProteinMPNN sequence-design model for protein backbones, including soluble and membrane variants.
+
+Converted membrane ProteinMPNN checkpoint with per-residue membrane labels.
+
+### Model Provenance
+
+- **Upstream Project**: Membrane ProteinMPNN
+- **Source Checkpoint**: `per_residue_label_membrane_mpnn_v_48_020.pt`
+- **Source Release**: [https://github.com/dauparas/LigandMPNN](https://github.com/dauparas/LigandMPNN)
+- **Primary Paper**: [Robust deep learning-based protein sequence design using ProteinMPNN](https://doi.org/10.1126/science.add2187)
+- **Upstream License**: MIT for upstream ProteinMPNN and LigandMPNN code/model parameters
+
+### Model Specification
+
+| Field | Value |
+| ----- | ----- |
+| Repository | `oteam/proteinmpnn-membrane-per-residue-noise020` |
+| Artifact Kind | `model` |
+| Task | `design_generation` |
+| Architecture | `proteinmpnn` |
+| Entrypoint | `ofoldx.models.proteinmpnn.model.ProteinMPNNDesignGenerator` |
+| Config | `config.json` |
+| Weights | `model.safetensors` |
+| Source Checkpoint | `per_residue_label_membrane_mpnn_v_48_020.pt` |
+| Checkpoint Variant | `per_residue_label_membrane_mpnn_v_48_020` |
+| k-Neighbors | `48` |
+| Node Feature Size | `3` |
+| Training Noise | `0.20 Angstrom` |
+
+> [!NOTE]
+> Source checkpoint: `per_residue_label_membrane_mpnn_v_48_020.pt`; the public repo suffix records the upstream Gaussian noise variant (0.20 Angstrom).
+
+### Links
+
+- **Hub repository**: [oteam/proteinmpnn-membrane-per-residue-noise020](https://huggingface.co/oteam/proteinmpnn-membrane-per-residue-noise020)
+- **Upstream paper**: [Robust deep learning-based protein sequence design using ProteinMPNN](https://doi.org/10.1126/science.add2187)
+- **Upstream repository**: [Membrane ProteinMPNN](https://github.com/dauparas/LigandMPNN)
+- **Source checkpoint release**: [https://github.com/dauparas/LigandMPNN](https://github.com/dauparas/LigandMPNN)
+- **Code**: [`ofoldx/models/proteinmpnn/model.py`](https://github.com/OTeam-AI4S/OFoldX/tree/main/ofoldx/models/proteinmpnn/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)
+
+## Usage
+
+The artifact depends on the [`ofoldx`](https://github.com/OTeam-AI4S/OFoldX) library. Install it with pip:
+
+```bash
+pip install ofoldx
+```
+
+### Pipeline Usage
+
+Load the artifact from `oteam/proteinmpnn-membrane-per-residue-noise020` with the OFoldX task pipeline. Use `AutoModel` or `AutoProcessor` only when you need lower-level control:
+
+```python
+from ofoldx.pipelines import Pipeline
+
+pipeline = Pipeline.from_pretrained("oteam/proteinmpnn-membrane-per-residue-noise020")
+output = pipeline(...)
+
+from ofoldx import AutoModel
+
+model = AutoModel.from_pretrained("oteam/proteinmpnn-membrane-per-residue-noise020")
+```
+
+When a matching processor is available, load it with `AutoProcessor.from_pretrained(...)` and pass the
+processed batch to the model.
+
+### Interface
+
+- **Task**: `design_generation`
+- **Artifact kind**: `model`
+- **Architecture**: `proteinmpnn`
+- **Runtime files**: `manifest.json`, `config.json`, and `model.safetensors` when present
+
+## Training Details
+
+OFoldX did not train these weights. This repository contains a converted checkpoint and OFoldX runtime
+metadata for loading it.
+
+### Training Data
+
+This checkpoint is distributed in the LigandMPNN parameter release. The per-residue model adds three membrane-context classes per residue: buried, interface, and other. OFoldX does not redistribute the training set.
+
+### Training Procedure
+
+Upstream training follows the ProteinMPNN fixed-backbone inverse-folding setup with per-residue membrane labels and a 0.20 Angstrom noisy-backbone checkpoint. OFoldX only converts the released checkpoint.
+
+## Evaluation
+
+OFoldX conversion reports and contract tests validate artifact structure and checkpoint loading. Task-level
+scientific evaluation should be checked against the corresponding upstream model release or paper.
+
+## Limitations
+
+- This artifact is distributed for research use.
+- Inputs must match the model-specific processor and expected biomolecular representation.
+- OFoldX is pre-alpha, so APIs and artifact metadata may still change before a stable release.
+
+## Citation
+
+Please cite the upstream Membrane ProteinMPNN work for the source checkpoint. If OFoldX supports your work, please also cite or link the OFoldX project repository.
+
+```bibtex
+@article{dauparas2022robust,
+  author = {Dauparas, Justas and Anishchenko, Ivan and Bennett, Nathaniel and Bai, Hua and Ragotte, Robert J. and Milles, Lukas F. and Wicky, Basile I. M. and Courbet, Alexis and de Haas, Rob J. and Bethel, Neville and others},
+  title = {Robust deep learning-based protein sequence design using ProteinMPNN},
+  journal = {Science},
+  volume = {378},
+  number = {6615},
+  pages = {49--56},
+  year = {2022},
+  doi = {10.1126/science.add2187}
+}
+```
+
+## Contact
+
+Please use [OFoldX GitHub issues](https://github.com/OTeam-AI4S/OFoldX/issues) for questions or comments about this model card.
+
+## License
+
+The Hub `license` metadata, when present, reflects the source checkpoint or upstream project license. The OFoldX project license is not yet finalized.
+The source checkpoint is associated with the upstream license noted above: MIT for upstream ProteinMPNN and LigandMPNN code/model parameters. Review both OFoldX and upstream terms before redistribution or production use.

generator/manifest.json

@@ -0,0 +1,18 @@
+{
+  "format": "biomolecular_pretrained",
+  "format_version": 1,
+  "kind": "model",
+  "producer": {
+    "package": "ofoldx",
+    "version": "0.1.dev1+g9a78c7e44.d20260609"
+  },
+  "entrypoint": "ofoldx.models.proteinmpnn.model.ProteinMPNNDesignGenerator",
+  "architecture": "proteinmpnn",
+  "task": "design_generation",
+  "files": {
+    "config": "config.json"
+  },
+  "weights": {
+    "default": "model.safetensors"
+  }
+}

manifest.json

@@ -1,18 +1,22 @@
 {
   "format": "biomolecular_pretrained",
   "format_version": 1,
-  "kind": "model",
+  "kind": "pipeline",
   "producer": {
     "package": "ofoldx",
     "version": "0.1.dev1+g9a78c7e44.d20260609"
   },
-  "entrypoint": "ofoldx.models.proteinmpnn.model.ProteinMPNNDesignGenerator",
+  "entrypoint": "ofoldx.pipelines.design.DesignPipeline",
   "architecture": "proteinmpnn",
   "task": "design_generation",
-  "files": {
-    "config": "config.json"
-  },
-  "weights": {
-    "default": "model.safetensors"
+  "components": {
+    "generator": {
+      "kind": "component",
+      "path": "generator"
+    },
+    "processor": {
+      "kind": "component",
+      "path": "processor"
+    }
   }
 }

processor/manifest.json

@@ -0,0 +1,13 @@
+{
+  "format": "biomolecular_pretrained",
+  "format_version": 1,
+  "kind": "processor",
+  "producer": {
+    "package": "ofoldx",
+    "version": "0.1.dev1+g9a78c7e44.d20260609"
+  },
+  "entrypoint": "ofoldx.models.proteinmpnn.processor.ProteinMPNNProcessor",
+  "files": {
+    "config": "processor_config.json"
+  }
+}

processor/processor_config.json

@@ -0,0 +1,6 @@
+{
+  "designed_chains": null,
+  "fixed_positions": {},
+  "tied_positions": [],
+  "omit_amino_acids": []
+}