kfold-assets is an open-weight model from KAIST AI-Bio, released under Apache License 2.0. Its published files total 8.2 GB.
This repository provides the chemical component dictionary and pretrained encoder weights used by K-Fold. These assets are downloaded automatically when needed. - assets/ccd.pkl: Precomputed chemical component dictionary (CCD).
Model Card
By KAIST AI-Bio, published under apache-2.0, revision 85e5e75c2d6c.
This repository provides the chemical component dictionary and pretrained encoder weights used by K-Fold. These assets are downloaded automatically when needed. - assets/ccd.pkl: Precomputed chemical component dictionary (CCD). - weights/protstructbbtokenizer.pth: Protein backbone tokenizer weights copied from TriProRep-3B. - weights/protstructfatokenizer.pth: Protein full-atom tokenizer weights copied from TriProRep-3B. - weights/protstructencoder3b.pth: Protein structure encoder weights converted to bfloat16 from TriProRep-3B-finetune. - weights/rnaseq1b.pth: Pretrained RNA sequence encoder weights. Source code for the protein structure encoder and tokenizers is available in the TriProRep…
Read KAIST AI-Bio's full model card
This repository provides the chemical component dictionary and pretrained encoder weights used by K-Fold. These assets are downloaded automatically when needed.
Contents
assets/ccd.pkl: Precomputed chemical component dictionary (CCD).weights/prot_struct_bb_tokenizer.pth: Protein backbone tokenizer weights copied from TriProRep-3B.weights/prot_struct_fa_tokenizer.pth: Protein full-atom tokenizer weights copied from TriProRep-3B.weights/prot_struct_encoder_3b.pth: Protein structure encoder weights converted to bfloat16 from TriProRep-3B-finetune.weights/rna_seq_1b.pth: Pretrained RNA sequence encoder weights.
Source code for the protein structure encoder and tokenizers is available in the TriProRep GitHub repository.
License
Copyright © 2026 Korea Advanced Institute of Science and Technology (KAIST).
This repository is licensed under the Apache License 2.0.
Identity and Version
- Repository
- kaist-ai-bio/kfold-assets
- Publisher
- KAIST AI-Bio
- Task
- Not stated by the source
- Modality
- Other
- Library
- Not stated by the source
- Parameters
- Not stated by the source
- Languages
- Not stated by the source
- Revision
- 85e5e75c2d6c443b424c8e68991a286d6170d73f
- First published
- 2026-10-08
- Last updated
- 2026-10-08
Files and Weights
8 files, 8.2 GB in total. The weights are 4 files totalling 7.3 GB in pth.
Every file
| File | Type | Size | SHA-256 |
|---|---|---|---|
| weights/prot_struct_bb_tokenizer.pth | Weights | 65.2 MB | 803ab0371a5d |
| weights/prot_struct_encoder_3b.pth | Weights | 5.2 GB | 485c38170f94 |
| weights/prot_struct_fa_tokenizer.pth | Weights | 23.1 MB | cbae0fce786e |
| weights/rna_seq_1b.pth | Weights | 2.0 GB | 78deb11c989f |
| config.json | Configuration | 3 B | — |
| README.md | Documentation | 1.4 KB | — |
| assets/ccd.pkl | Other | 903.2 MB | 445b126be5c2 |
| .gitattributes | Repository | 1.5 KB | — |
License and Download
- License
- apache-2.0
- Access
- Open weights, no gate
- Download size
- 7.3 GB
Released by KAIST AI-Bio through its official repository on Hugging Face. Read the license.
Built From
- Derived from k-fold-structure/triprorep-3B
Memory Requirements
| Precision | Weights in memory |
|---|---|
| As published | 7.3 GB |
Weights only, from the published parameter count; the key-value cache and runtime add to this.
Questions About kfold-assets
Can I use kfold-assets commercially?
Yes. kfold-assets is released under Apache License 2.0. The Apache License 2.0 is a permissive open-source license. It permits commercial use, modification and redistribution. It requires keeping the license and copyright notices and any NOTICE file, stating significant changes, and it includes an express patent grant from contributors.