Instructions to use nikraf/directionality_probe with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use nikraf/directionality_probe with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("feature-extraction", model="nikraf/directionality_probe", trust_remote_code=True)# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("nikraf/directionality_probe", trust_remote_code=True, device_map="auto") - Notebooks
- Google Colab
- Kaggle
| supported_datasets = { | |
| 'EC': 'GleghornLab/EC_reg', | |
| 'GO-CC': 'GleghornLab/CC_reg', | |
| 'GO-BP': 'GleghornLab/BP_reg', | |
| 'GO-MF': 'GleghornLab/MF_reg', | |
| 'MB': 'GleghornLab/MB_reg', | |
| 'DeepLoc-2': 'GleghornLab/DL2_reg', | |
| 'DeepLoc-10': 'GleghornLab/DL10_reg', | |
| 'Subcellular': 'GleghornLab/SL_13', | |
| 'enzyme-kcat': 'GleghornLab/enzyme_kcat', | |
| 'solubility': 'GleghornLab/solubility_prediction', | |
| 'localization': 'GleghornLab/localization_prediction', | |
| 'temperature-stability': 'GleghornLab/temperature_stability', | |
| 'peptide-HLA-MHC-affinity': 'GleghornLab/peptide_HLA_MHC_affinity_ppi', | |
| 'optimal-temperature': 'GleghornLab/optimal_temperature', | |
| 'optimal-ph': 'GleghornLab/optimal_ph', | |
| 'material-production': 'GleghornLab/material_production', | |
| 'fitness-prediction': 'GleghornLab/fitness_prediction', | |
| 'number-of-folds': 'GleghornLab/fold_prediction', | |
| 'cloning-clf': 'GleghornLab/cloning_clf', | |
| 'stability-prediction': 'GleghornLab/stability_prediction', | |
| 'human-ppi-saprot': 'GleghornLab/HPPI', | |
| 'SecondaryStructure-3': 'GleghornLab/SS3', | |
| 'SecondaryStructure-8': 'GleghornLab/SS8', | |
| 'fluorescence-prediction': 'GleghornLab/fluorescence_prediction', | |
| 'plastic': 'GleghornLab/plastic_degradation_benchmark', | |
| 'gold-ppi': 'Synthyra/bernett_gold_ppi', | |
| 'human-ppi-pinui': 'GleghornLab/HPPI_PiNUI', | |
| 'yeast-ppi-pinui': 'GleghornLab/YPPI_PiNUI', | |
| 'shs27-ppi-raw': 'Synthyra/SHS27k', | |
| 'shs148-ppi-raw': 'Synthyra/SHS148k', | |
| 'shs27-ppi-random': 'GleghornLab/ppi_SHS27k_random_2025', | |
| 'shs148-ppi-random': 'GleghornLab/ppi_SHS148k_random_2025', | |
| 'shs27-ppi-dfs': 'GleghornLab/ppi_SHS27k_dfs_2025', | |
| 'shs148-ppi-dfs': 'GleghornLab/ppi_SHS148k_dfs_2025', | |
| 'shs27-ppi-bfs': 'GleghornLab/ppi_SHS27k_bfs_2025', | |
| 'shs148-ppi-bfs': 'GleghornLab/ppi_SHS148k_bfs_2025', | |
| 'string-ppi-random': 'GleghornLab/ppi_STRING_random_2025', | |
| 'string-ppi-dfs': 'GleghornLab/ppi_STRING_dfs_2025', | |
| 'string-ppi-bfs': 'GleghornLab/ppi_STRING_bfs_2025', | |
| 'plm-interact': 'GleghornLab/plm_interact_human_train_cross_ppi', | |
| 'ppi-mutation-effect': 'GleghornLab/ppi_mutation_effect', # requires multi_column | |
| 'PPA-ppi': 'Synthyra/ppi_affinity', | |
| 'foldseek-fold': 'lhallee/foldseek_dataset', # prostt5 | |
| 'foldseek-inverse': 'lhallee/foldseek_dataset', # prostt5 | |
| 'ec-active': 'lhallee/ec_active', | |
| 'bernett_processed': 'lhallee/bernett_processed', | |
| 'proteingym_zs': 'proteingym_zs', # not a path, data loading for this is currently handled in benchmarks/proteingym/data_loader.py | |
| 'proteingym_supervised': 'proteingym_supervised', # not a path, data loading for this is currently handled in benchmarks/proteingym/data_loader.py | |
| 'taxon_domain': 'GleghornLab/taxonomy_domain_0.4_clusters', | |
| 'taxon_kingdom': 'GleghornLab/taxonomy_kingdom_0.4_clusters', | |
| 'taxon_phylum': 'GleghornLab/taxonomy_phylum_0.4_clusters', | |
| 'taxon_class': 'GleghornLab/taxonomy_class_0.4_clusters', | |
| 'taxon_order': 'GleghornLab/taxonomy_order_0.4_clusters', | |
| 'taxon_family': 'GleghornLab/taxonomy_family_0.4_clusters', | |
| 'taxon_genus': 'GleghornLab/taxonomy_genus_0.4_clusters', | |
| 'taxon_species': 'GleghornLab/taxonomy_species_0.4_clusters', | |
| 'diff_phylogeny': 'GleghornLab/diff_phylo', | |
| 'plddt': 'GleghornLab/af2_plddt', | |
| 'realness': 'GleghornLab/realness_dataset', | |
| 'million_full': 'GleghornLab/millionfull_round_1_oct_2025', | |
| 'soluprot': 'GleghornLab/soluprot', | |
| 'ecoli_expression': 'GleghornLab/ecoli_expression', | |
| 'KSMoFinder-clustered': 'GleghornLab/ksmo_clustered', | |
| 'KSMoFinder': 'GleghornLab/KSmo_fixed', | |
| } | |
| internal_datasets = { | |
| 'plastic': 'GleghornLab/plastic_degradation_benchmark', | |
| } | |
| # TODO update | |
| possible_with_vector_reps = [ | |
| ### multi-label | |
| 'EC', | |
| # GO | |
| 'GO-CC', | |
| 'GO-BP', | |
| 'GO-MF', | |
| 'Subcellular', | |
| # ppi | |
| 'shs27-ppi-random', | |
| 'shs27-ppi-dfs', | |
| 'shs27-ppi-bfs', | |
| 'shs148-ppi-random', | |
| 'shs148-ppi-dfs', | |
| 'shs148-ppi-bfs', | |
| 'string-ppi-random', | |
| 'string-ppi-dfs', | |
| 'string-ppi-bfs', | |
| ### classification | |
| 'MB', | |
| 'DeepLoc-2', | |
| 'DeepLoc-10', | |
| 'solubility', | |
| 'temperature-stability', | |
| 'material-production', | |
| 'fitness-prediction', | |
| 'number-of-folds', | |
| 'cloning-clf', | |
| 'stability-prediction', | |
| 'ec-active', | |
| 'localization', | |
| # taxonomy | |
| 'taxon_domain', | |
| 'taxon_kingdom', | |
| 'taxon_phylum', | |
| 'taxon_class', | |
| 'taxon_order', | |
| 'taxon_family', | |
| 'taxon_genus', | |
| 'taxon_species', | |
| 'diff_phylogeny', | |
| # ppi | |
| 'shs27-ppi-raw', | |
| 'shs148-ppi-raw', | |
| 'plm-interact', | |
| 'gold-ppi', | |
| 'string-ppi-bfs', | |
| 'human-ppi-saprot', | |
| 'human-ppi-pinui', | |
| 'yeast-ppi-pinui', | |
| ### regression | |
| 'enzyme-kcat', | |
| 'optimal-temperature', | |
| 'optimal-ph', | |
| 'million_full', | |
| # ppi | |
| 'PPA-ppi', | |
| ] | |
| # TODO update | |
| standard_data_benchmark = [ | |
| 'ec-active', | |
| 'EC', | |
| 'GO-CC', | |
| 'GO-BP', | |
| 'GO-MF', | |
| 'MB', | |
| 'DeepLoc-2', | |
| 'DeepLoc-10', | |
| 'enzyme-kcat', | |
| 'optimal-temperature', | |
| 'optimal-ph', | |
| 'fitness-prediction', | |
| ] | |
| vector_benchmark = [ | |
| ### multi-label | |
| 'EC', # SaProt | |
| # GO | |
| 'GO-CC', # SaProt | |
| 'GO-BP', # SaProt | |
| 'GO-MF', # SaProt | |
| 'Subcellular', # Custom - Tamar | |
| # ppi | |
| 'shs27-ppi-bfs', # MGPPI + SHS27k + SHS148k + STRING | |
| 'shs148-ppi-bfs', # MGPPI + SHS27k + SHS148k + STRING | |
| 'string-ppi-bfs', # MGPPI + SHS27k + SHS148k + STRING | |
| ### classification | |
| 'MB', # SaProt | |
| 'DeepLoc-2', # SaProt | |
| 'DeepLoc-10', # SaProt | |
| 'solubility', # Biomap | |
| 'temperature-stability', # Biomap | |
| 'material-production', # Biomap | |
| 'fitness-prediction', # Biomap | |
| 'number-of-folds', # Biomap | |
| 'cloning-clf', # Biomap | |
| 'stability-prediction', # Biomap | |
| 'ec-active', # Custom - Logan | |
| 'soluprot', # Custom - SoluProt | |
| # taxonomy | |
| 'taxon_species', # Accidental taxonomists | |
| 'diff_phylogeny', # Accidental taxonomists | |
| # ppi | |
| 'plm-interact', # PLM-Interact | |
| 'gold-ppi', # Bernett | |
| ### regression | |
| 'enzyme-kcat', # Custom - Logan - Biomap | |
| 'optimal-temperature', # Biomap | |
| 'optimal-ph', # Biomap | |
| 'million_full', # Millionfull | |
| # ppi | |
| 'PPA-ppi', # Custom - Logan - Bindwell | |
| ] | |
| testing = [ | |
| 'EC', # multilabel | |
| 'DeepLoc-2', # singlelabel | |
| 'DeepLoc-10', # multiclass | |
| 'enzyme-kcat', # regression | |
| 'human-ppi', # ppi | |
| 'plddt', # tokenwise regression | |
| 'SecondaryStructure-3', # tokenwise classification | |
| ] |