Model card: the autoencoder component is registered as vae

README.md

@@ -47,7 +47,7 @@ This repository is the [Diffusers](https://github.com/huggingface/diffusers) rel
 | | Details |
 |---|---|
 | **Model** | Joint sequence–structure generation with a matched flow transformer and autoencoder |
-| **Pipeline** | `RFD4ProteinaPipeline`: `transformer` + `autoencoder` + `scheduler` |
+| **Pipeline** | `RFD4ProteinaPipeline`: `transformer` + `vae` + `scheduler` |
 | **Inputs** | Annotated CIF structures, prepared with [`rfd4-proteina-prep`](https://huggingface.co/nvidia/RFD4-Proteina-prep) |
 | **Outputs** | Atom37 coordinates and residue identities, written back to annotated CIF |
 | **Weights** | EMA weights, 434.3 million parameters, stored in FP32 (1.74 GB); runs in BF16 |
@@ -139,7 +139,7 @@ The same seed can end in a different design in BF16 and FP32, because the sample
 | Path | Use |
 |---|---|
 | `model_index.json` | Pipeline definition for `RFD4ProteinaPipeline.from_pretrained` |
-| `transformer/`, `autoencoder/` | Denoiser and autoencoder configs and safetensors weights |
+| `transformer/`, `vae/` | Denoiser and autoencoder configs and safetensors weights |
 | `scheduler/` | `RFD4ProductSpaceScheduler` config with the released sampler recipe and recycle schedule |
 | [`provenance.json`](provenance.json) | Source bundle hashes, sampler recipe and converter version |
 | `examples/` | Studio inputs, generated designs and previews |