Publication layout: model card, provenance, examples, assets
Browse files
README.md
CHANGED
|
@@ -33,7 +33,7 @@ This repository is the [Diffusers](https://github.com/huggingface/diffusers) rel
|
|
| 33 |
|---|---|
|
| 34 |
| **Model** | Joint sequence–structure generation with a matched flow transformer and autoencoder |
|
| 35 |
| **Pipeline** | `RFD4ProteinaPipeline`: `transformer` + `autoencoder` + `scheduler` |
|
| 36 |
-
| **Inputs** | Annotated CIF structures, prepared with [`rfd4-proteina-prep`](https://
|
| 37 |
| **Outputs** | Atom37 coordinates and residue identities, written back to annotated CIF |
|
| 38 |
| **Weights** | EMA weights, 434.3 million parameters, stored in FP32 (1.74 GB); runs in BF16 |
|
| 39 |
| **Sampler** | Released `thinking_medium` recipe: 60 steps, 185 trunk passes |
|
|
@@ -43,7 +43,7 @@ This repository is the [Diffusers](https://github.com/huggingface/diffusers) rel
|
|
| 43 |
## Quickstart
|
| 44 |
|
| 45 |
```bash
|
| 46 |
-
pip install diffusers
|
| 47 |
```
|
| 48 |
|
| 49 |
```python
|
|
@@ -95,7 +95,7 @@ Each `examples/<name>/` folder holds the AtomWorks Studio input (`input.cif`), t
|
|
| 95 |
|
| 96 |
</details>
|
| 97 |
|
| 98 |
-
These examples demonstrate the conditioning workflow. They are not experimental binding or catalytic-activity results, and example selection does not establish biological design quality. `examples/examples.json` lists the sampling task of each input. To reproduce them, run `examples/run_examples.py` from the [`rfd4-proteina-prep`](https://
|
| 99 |
|
| 100 |
## Precision and speed
|
| 101 |
|
|
|
|
| 33 |
|---|---|
|
| 34 |
| **Model** | Joint sequence–structure generation with a matched flow transformer and autoencoder |
|
| 35 |
| **Pipeline** | `RFD4ProteinaPipeline`: `transformer` + `autoencoder` + `scheduler` |
|
| 36 |
+
| **Inputs** | Annotated CIF structures, prepared with [`rfd4-proteina-prep`](https://huggingface.co/nvidia/RFD4-Proteina-prep) |
|
| 37 |
| **Outputs** | Atom37 coordinates and residue identities, written back to annotated CIF |
|
| 38 |
| **Weights** | EMA weights, 434.3 million parameters, stored in FP32 (1.74 GB); runs in BF16 |
|
| 39 |
| **Sampler** | Released `thinking_medium` recipe: 60 steps, 185 trunk passes |
|
|
|
|
| 43 |
## Quickstart
|
| 44 |
|
| 45 |
```bash
|
| 46 |
+
pip install diffusers git+https://huggingface.co/nvidia/RFD4-Proteina-prep
|
| 47 |
```
|
| 48 |
|
| 49 |
```python
|
|
|
|
| 95 |
|
| 96 |
</details>
|
| 97 |
|
| 98 |
+
These examples demonstrate the conditioning workflow. They are not experimental binding or catalytic-activity results, and example selection does not establish biological design quality. `examples/examples.json` lists the sampling task of each input. To reproduce them, run `examples/run_examples.py` from the [`rfd4-proteina-prep`](https://huggingface.co/nvidia/RFD4-Proteina-prep) repository.
|
| 99 |
|
| 100 |
## Precision and speed
|
| 101 |
|