Remove the bucket validation link, source revision and historical/reproduction details from the optional-runtime paragraph. Keep a short description of supported runtimes and separate weights. All containers and other repository files remain unchanged.
Trim AF3 runtime extras while preserving tested CUDA 12/13 containers
Keep updated SIF, SquashFS and Docker images and their checksums at existing paths. Remove redundant OCI exports, build-recipe copies and JSON reports from the current runtime tree; validation remains linked from the public bucket. No container binaries, weights or AF2 files are changed.
Publish tested CUDA 12/13 AF3 and OpenFold3 runtimes with index-gap support
Both CUDA builds contain AF3 fork revision 8d36011. Validation: 32 source/index-feature cases and 48 AF3/OF3 GPU refolds. Model checkpoints, weights, and AF2 artifacts are unchanged.
The row listed ZeroGPU first, so the first Studio a visitor met was the one that spends their own Hugging Face quota and queues behind other users. The dedicated pool is the better default; ZeroGPU stays as the alternative for anyone who prefers it. The website made the same change in baker-laboratory/RFD4-Proteina-dev#771.
Cards: point links at rfd4-proteina.org docs and the new wordmark
RFD4-Proteina was merged into Diffusers upstream as huggingface/diffusers#15088 on 2026-10-20, so the pipeline no longer comes from a staging branch. It is not on PyPI yet, which the next Diffusers release fixes, so the quickstart installs from git and shows the plain pip line commented out beside it.
Checked rather than assumed, in a clean 3.12 environment: the git install imports RFD4ProteinaPipeline, and nvidia/RFD4-Proteina-Diffusers loads through it with all four components and no newly-initialized warnings. accelerate is listed because the load fails without it: "`low_cpu_mem_usage` cannot be False when `keep_in_fp32_modules` is True".
docs: link the ZeroGPU and dedicated-GPU Studio Spaces
The snippet is followed by "Partial diffusion and inverse folding work in this port", but its required list left coords_nm in conditioning_features, and the pipeline reads it as a named argument. Both of those tasks raise "partial and inverse generation require clean coords_nm" without it. A full run, which is what the snippet shows, ignores coords_nm, so the snippet behaves identically and the two tasks it advertises now work.
Model card: focus on native and Diffusers, drop the Transformers section
Rebuilt from the current root bundle (weights 5b549dd6), replacing a conversion of the superseded 5f00cdcd.
Also follows the release on three settings the earlier conversion predated: the UNK residue's atom37 row now carries its backbone, thinking_medium is the front-loaded 180-pass schedule rather than the earlier peaked 185, and recycle_remaining is one class wide, which the native config now leaves implicit.
Parity against the native runtime: feature factories and autoencoder exact on monomer, ligand and glycan inputs, denoiser within 2.3e-3.
3 days ago
Rename projections to the transformers convention
Regenerated from the same release bundle: every tensor is byte-identical, only the keys change (to_q -> q_proj, to_out_node -> o_proj, to_gamma -> scale_proj, ...), matching the renamed modelling code.
The config drops nine flags the released architecture pins and folds the retired sc_scale_score/tsr_matched_noise sampler controls into sc_scale_noise=1.0, which describes the same trajectory.
Update inference.yaml to the 2026-10-19 release and pin the model bundle
