| { |
| "additional_special_tokens": [ |
| { |
| "content": "<aromatic_rings_count:0>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<aromatic_rings_count:10>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<aromatic_rings_count:11>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<aromatic_rings_count:12>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<aromatic_rings_count:13>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<aromatic_rings_count:14>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<aromatic_rings_count:15>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<aromatic_rings_count:16>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<aromatic_rings_count:17>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<aromatic_rings_count:18>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<aromatic_rings_count:19>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<aromatic_rings_count:1>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<aromatic_rings_count:20>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<aromatic_rings_count:22>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<aromatic_rings_count:2>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<aromatic_rings_count:3>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<aromatic_rings_count:4>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<aromatic_rings_count:5>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<aromatic_rings_count:6>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<aromatic_rings_count:7>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<aromatic_rings_count:8>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<aromatic_rings_count:9>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_is_glycoside:False>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_is_glycoside:True>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_pathway:Alkaloids>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_pathway:Amino acids and Peptides>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_pathway:Carbohydrates>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_pathway:Fatty acids>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_pathway:Polyketides>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_pathway:Shikimates and Phenylpropanoids>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_pathway:Terpenoids>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Alkylresorcinols>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Amino acid glycosides>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Aminosugars and aminoglycosides>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Anthranilic acid alkaloids>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Apocarotenoids>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Aromatic polyketides>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Carotenoids (C40)>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Carotenoids (C45)>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Carotenoids (C50)>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Chromanes>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Coumarins>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Cyclic polyketides>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Diarylheptanoids>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Diazotetronic acids and derivatives>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Diphenyl ethers (DPEs)>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Diterpenoids>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Docosanoids>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Eicosanoids>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Fatty Acids and Conjugates>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Fatty acyl glycosides>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Fatty acyls>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Fatty amides>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Fatty esters>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Flavonoids>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Fluorenes>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Glycerolipids>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Glycerophospholipids>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Guanidine alkaloids>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Histidine alkaloids>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Isoflavonoids>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Lignans>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Linear polyketides>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Lysine alkaloids>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Macrolides>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Meroterpenoids>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Miscellaneous alkaloids>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Miscellaneous polyketides>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Mitomycin derivatives>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Monoterpenoids>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Mycosporine derivatives>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Naphthalenes>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Nicotinic acid alkaloids>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Nucleosides>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Octadecanoids>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Oligopeptides>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Ornithine alkaloids>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Peptide alkaloids>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Phenanthrenoids>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Phenolic acids (C6-C1)>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Phenylethanoids (C6-C2)>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Phenylpropanoids (C6-C3)>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Phloroglucinols>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Polycyclic aromatic polyketides>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Polyethers>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Polyols>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Polyprenols>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Proline alkaloids>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Pseudoalkaloids>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Saccharides>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Serine alkaloids>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Sesquiterpenoids>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Sesterterpenoids>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Small peptides>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Sphingolipids>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Steroids>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Stilbenoids>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Styrylpyrones>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Terphenyls>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Tetramate alkaloids>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Triterpenoids>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Tropolones>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Tryptophan alkaloids>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Tyrosine alkaloids>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:Xanthones>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:β-lactams>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<np_classifier_superclass:γ-lactam-β-lactones>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<qed_bin:0.1<=qed<0.2>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<qed_bin:0.2<=qed<0.3>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<qed_bin:0.3<=qed<0.4>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<qed_bin:0.4<=qed<0.5>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<qed_bin:0.5<=qed<0.6>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<qed_bin:0.6<=qed<0.7>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<qed_bin:0.7<=qed<0.8>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<qed_bin:0.8<=qed<0.9>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<qed_bin:0.9<=qed<1>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<qed_bin:0<=qed<0.1>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<sa_bin:1<=sa<2>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<sa_bin:2<=sa<3>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<sa_bin:3<=sa<4>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<sa_bin:4<=sa<5>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<sa_bin:5<=sa<6>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<sa_bin:6<=sa<7>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<sa_bin:7<=sa<8>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<sa_bin:8<=sa<9>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| { |
| "content": "<sa_bin:9<=sa<10>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| } |
| ], |
| "cls_token": { |
| "content": "<bos>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| "mask_token": { |
| "content": "<mask>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| "pad_token": { |
| "content": "<pad>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| "sep_token": { |
| "content": "<eos>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| }, |
| "unk_token": { |
| "content": "<unk>", |
| "lstrip": false, |
| "normalized": false, |
| "rstrip": false, |
| "single_word": false |
| } |
| } |
|
|
