|
Download CupOFMoCA-PROTAC/docs/ARTIFACTS.md from ChatterjeeLab/CupOFMoCA: direct link, hf CLI and curl.
- Browser
- Download file 4.98 kB
-
https://huggingface.co/ChatterjeeLab/CupOFMoCA/resolve/main/CupOFMoCA-PROTAC/docs/ARTIFACTS.md
- Command line
-
hf download hf://ChatterjeeLab/CupOFMoCA/CupOFMoCA-PROTAC/docs/ARTIFACTS.md
-
curl -L -o ARTIFACTS.md https://huggingface.co/ChatterjeeLab/CupOFMoCA/resolve/main/CupOFMoCA-PROTAC/docs/ARTIFACTS.md
4.98 kB
Model artifacts
Every local artifact the PROTAC pipeline loads. Run python scripts/verify_release.py
to check byte counts, SHA-256 hashes and private-path hygiene. The machine-readable
manifest is docs/artifacts.json.
| Artifact | Purpose | Size | SHA-256 |
|---|---|---|---|
checkpoints/protac_linker_finetune/epoch=99-v1.ckpt |
DeFoG generator, fine-tuned on PROTAC-DB linkers | 255,211,895 B | 8cfb48d9f5b3bcabb28c4dce7548e1a792494c48f05150a1b262b5215c1da78f |
src/classifier/protac_linker_clf.pkl |
linker classifier (generic) | 128,117 B | fa667025962d132ce9f2ebf8f2dc49752416ed154792ca0c87344335fa714462 |
src/classifier/protac_linker_clf_BRD9VHL.pkl |
linker classifier (BRD9VHL) | 28,117 B | eae385be1576fec674b8a537e50b56e1ab1fc47599a82ae277912d958c0b7d6f |
src/classifier/protac_linker_clf_CDK6CRBN.pkl |
linker classifier (CDK6CRBN) | 30,035 B | 85bd40599d82c5cc8a35cc844d12d1066a0d4699bcc7e229937e1fe1d3daad3d |
src/classifier/protac_linker_clf_ErlotinibVHL.pkl |
linker classifier (ErlotinibVHL) | 26,397 B | e031d310829c7d196887e2fd077eff95be445ed1984516540ac69308c26384ee |
src/classifier/protac_linker_clf_amine.pkl |
linker classifier (amine) | 76,524 B | e0e851f41a004ea26e3b66caef83bf6ab4fa2229447427ab996e8a195c0df2d2 |
src/classifier/protac_linker_clf_amino_acid.pkl |
linker classifier (amino_acid) | 188,080 B | 2558addea35da16ddf82affaf68ea9046bc5b9b3d307f6ee668c117df2798428 |
src/degrademaster/data/case_study/arcrbn_pocket_graph.pt |
ARCRBN pocket graph | 35,405 B | 98f70b82b731760b2b25b9ae0e7b1a917a88fd215f2ba4e61cf1654207d137e3 |
src/degrademaster/data/case_study/brd4_pocket_graph.pt |
BRD4 pocket graph | 15,675 B | 8dbda805c329bc5804d95983acaf4260e712bfa8fefcc5f39d76b4d2cd3583b9 |
src/degrademaster/data/case_study/brd9_pocket_graph.pt |
BRD9 pocket graph | 201,019 B | 89346178dfdc8fd277c57d08293a903133a263b1206184224cf250f71ad8f3a0 |
src/degrademaster/data/case_study/cdk6crbn_pocket_graph.pt |
CDK6CRBN pocket graph | 32,607 B | 0193874c2150e40a2e8c7a6b4eddf1c9b1768ac0573716b0fbf79e365de62310 |
src/degrademaster/data/case_study/crbn_pocket_graph.pt |
CRBN pocket graph | 25,467 B | 6e4e8e6756a343379118150370061f1b781d36f1bc681693d6d7fd8a8d0bce12 |
src/degrademaster/data/case_study/egfr_pocket_graph.pt |
EGFR pocket graph | 32,602 B | 5d86cd0d80363bab3c2835aefa817181cd96b2e0c14b35c55e384cf8e9dc64c1 |
src/degrademaster/data/case_study/krasvhl_pocket_graph.pt |
KRASVHL pocket graph | 25,686 B | 79efc62c546142277c6cba995d7926ca0dc3b600c2e98cce766519054243284c |
src/degrademaster/data/case_study/processed/ligase_pocket.pt |
VHL E3 pocket graph (collated) | 419,524 B | 3b35bb97bddce48ccc5ebddbdb435a98654b94f569441cf6f8bc66e3a11e4d38 |
src/degrademaster/data/case_study/arcrbn_target_feature.npy |
30-d feature vector | 368 B | 9ef6c2c04f54c6585cdac4e3d0ffb38c15ba7d1912a29f25f6a4967c7e13799e |
src/degrademaster/data/case_study/brd4_target_feature.npy |
30-d feature vector | 368 B | 28a1903da353ff6a92c4d4b8828e66e7264bcd4f07a27329d2f898cf000f63e4 |
src/degrademaster/data/case_study/brd4crbn_target_feature.npy |
30-d feature vector | 368 B | 28a1903da353ff6a92c4d4b8828e66e7264bcd4f07a27329d2f898cf000f63e4 |
src/degrademaster/data/case_study/brd9_target_feature.npy |
30-d feature vector | 368 B | dbe5d7e51907b44396a891a574010544f9e7461c09618d940fe26ddf59ef8473 |
src/degrademaster/data/case_study/cdk6crbn_target_feature.npy |
30-d feature vector | 368 B | 4da1adf39d018d61cd80a67f5950eaf734b425b69690a157c392d792b020712a |
src/degrademaster/data/case_study/crbn_e3_feature.npy |
30-d feature vector | 368 B | 71f830577fd901b6be42b9bb8f02f12e25245aefafac709d0b68910dfa8214db |
src/degrademaster/data/case_study/egfr_target_feature.npy |
30-d feature vector | 368 B | c31f02bfeb6ce63b0fcf41a7b235556dc2b76ab6054626116c08434a29b639de |
src/degrademaster/data/case_study/krasvhl_target_feature.npy |
30-d feature vector | 368 B | 51d50af89f905b6a4eb1f1d2b90d9ca8a2ee7e7dbace16fdab56e159f2c0385f |
src/degrademaster/data/case_study/vhl_e3_feature_egfr.npy |
30-d feature vector | 368 B | aca7d4ba3f95a2d903e542f761c3d448323d7ff8a9e1d58e58d9b8bd55d0a74a |
src/degrademaster/data/PROTAC/name.json |
PROTAC-DB entry table (property windows, similarity) | 3,419,093 B | bbe5ffe734f416ecf153f4ab42a17c6a3a86a79cfcb0faa5a5f13a672733d575 |
data/linker.csv |
PROTAC-DB derived linker table | 703,569 B | 73e4be1680fe9e2e2cf7256ff3fa9423a9a7e480d9b1baa529081a2bc89ac5b1 |
DegradeMaster and EGFR-PROPK are published upstream without a license and are therefore
not redistributed; python scripts/fetch_third_party.py retrieves them at the commits
used for the paper. Their absence is reported, not failed, by the verifier.
The ORCA graphlet-counting C++ sources bundled with upstream DeFoG are excluded: they carry no license header and are unused by the PROTAC pipeline.