| { | |
| "version": "1.0", | |
| "truncation": null, | |
| "padding": null, | |
| "added_tokens": [ | |
| { | |
| "id": 0, | |
| "content": "<PAD>", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 1, | |
| "content": "<BOS>", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 2, | |
| "content": "<EOS>", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 3, | |
| "content": "<UNK>", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 4, | |
| "content": "<SEP>", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 5, | |
| "content": "<MASK>", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 6, | |
| "content": "<COND>", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 7, | |
| "content": "</COND>", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 8, | |
| "content": "<MOL>", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 9, | |
| "content": "</MOL>", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 10, | |
| "content": "<MOL_HAC>", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 11, | |
| "content": "<MOL_HBDC>", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 12, | |
| "content": "<MOL_HBAC>", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 13, | |
| "content": "<MOL_ROTBC>", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 14, | |
| "content": "<MOL_WT>", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 15, | |
| "content": "<MOL_LOGP>", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 16, | |
| "content": "<MOL_TPSA>", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 17, | |
| "content": "<MOL_QED>", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 18, | |
| "content": "<MOL_SAS>", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 19, | |
| "content": "<MOL_PHARMA_FP>", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 20, | |
| "content": "<PROT_SEQ_ESM_EMB>", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 21, | |
| "content": "<PROT_POC_GVP_EMB>", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 22, | |
| "content": "<VALUE>", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 23, | |
| "content": "<UNUSED0>", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 24, | |
| "content": "<UNUSED1>", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 25, | |
| "content": "<UNUSED2>", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 26, | |
| "content": "<UNUSED3>", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 27, | |
| "content": "<UNUSED4>", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 28, | |
| "content": "<UNUSED5>", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 29, | |
| "content": "<UNUSED6>", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 30, | |
| "content": "<UNUSED7>", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 31, | |
| "content": "<UNUSED8>", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 32, | |
| "content": "<UNUSED9>", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| } | |
| ], | |
| "normalizer": { | |
| "type": "Sequence", | |
| "normalizers": [ | |
| { | |
| "type": "Strip", | |
| "strip_left": true, | |
| "strip_right": true | |
| } | |
| ] | |
| }, | |
| "pre_tokenizer": { | |
| "type": "Split", | |
| "pattern": { | |
| "Regex": "(\\[[^\\]]+\\]|</?[^>]+>)" | |
| }, | |
| "behavior": "Isolated", | |
| "invert": false | |
| }, | |
| "post_processor": { | |
| "type": "TemplateProcessing", | |
| "single": [ | |
| { | |
| "SpecialToken": { | |
| "id": "<BOS>", | |
| "type_id": 0 | |
| } | |
| }, | |
| { | |
| "Sequence": { | |
| "id": "A", | |
| "type_id": 0 | |
| } | |
| }, | |
| { | |
| "SpecialToken": { | |
| "id": "<EOS>", | |
| "type_id": 0 | |
| } | |
| } | |
| ], | |
| "pair": [ | |
| { | |
| "Sequence": { | |
| "id": "A", | |
| "type_id": 0 | |
| } | |
| }, | |
| { | |
| "Sequence": { | |
| "id": "B", | |
| "type_id": 1 | |
| } | |
| } | |
| ], | |
| "special_tokens": { | |
| "</COND>": { | |
| "id": "</COND>", | |
| "ids": [ | |
| 7 | |
| ], | |
| "tokens": [ | |
| "</COND>" | |
| ] | |
| }, | |
| "</MOL>": { | |
| "id": "</MOL>", | |
| "ids": [ | |
| 9 | |
| ], | |
| "tokens": [ | |
| "</MOL>" | |
| ] | |
| }, | |
| "<BOS>": { | |
| "id": "<BOS>", | |
| "ids": [ | |
| 1 | |
| ], | |
| "tokens": [ | |
| "<BOS>" | |
| ] | |
| }, | |
| "<COND>": { | |
| "id": "<COND>", | |
| "ids": [ | |
| 6 | |
| ], | |
| "tokens": [ | |
| "<COND>" | |
| ] | |
| }, | |
| "<EOS>": { | |
| "id": "<EOS>", | |
| "ids": [ | |
| 2 | |
| ], | |
| "tokens": [ | |
| "<EOS>" | |
| ] | |
| }, | |
| "<MASK>": { | |
| "id": "<MASK>", | |
| "ids": [ | |
| 5 | |
| ], | |
| "tokens": [ | |
| "<MASK>" | |
| ] | |
| }, | |
| "<MOL>": { | |
| "id": "<MOL>", | |
| "ids": [ | |
| 8 | |
| ], | |
| "tokens": [ | |
| "<MOL>" | |
| ] | |
| }, | |
| "<MOL_HAC>": { | |
| "id": "<MOL_HAC>", | |
| "ids": [ | |
| 10 | |
| ], | |
| "tokens": [ | |
| "<MOL_HAC>" | |
| ] | |
| }, | |
| "<MOL_HBAC>": { | |
| "id": "<MOL_HBAC>", | |
| "ids": [ | |
| 12 | |
| ], | |
| "tokens": [ | |
| "<MOL_HBAC>" | |
| ] | |
| }, | |
| "<MOL_HBDC>": { | |
| "id": "<MOL_HBDC>", | |
| "ids": [ | |
| 11 | |
| ], | |
| "tokens": [ | |
| "<MOL_HBDC>" | |
| ] | |
| }, | |
| "<MOL_LOGP>": { | |
| "id": "<MOL_LOGP>", | |
| "ids": [ | |
| 15 | |
| ], | |
| "tokens": [ | |
| "<MOL_LOGP>" | |
| ] | |
| }, | |
| "<MOL_PHARMA_FP>": { | |
| "id": "<MOL_PHARMA_FP>", | |
| "ids": [ | |
| 19 | |
| ], | |
| "tokens": [ | |
| "<MOL_PHARMA_FP>" | |
| ] | |
| }, | |
| "<MOL_QED>": { | |
| "id": "<MOL_QED>", | |
| "ids": [ | |
| 17 | |
| ], | |
| "tokens": [ | |
| "<MOL_QED>" | |
| ] | |
| }, | |
| "<MOL_ROTBC>": { | |
| "id": "<MOL_ROTBC>", | |
| "ids": [ | |
| 13 | |
| ], | |
| "tokens": [ | |
| "<MOL_ROTBC>" | |
| ] | |
| }, | |
| "<MOL_SAS>": { | |
| "id": "<MOL_SAS>", | |
| "ids": [ | |
| 18 | |
| ], | |
| "tokens": [ | |
| "<MOL_SAS>" | |
| ] | |
| }, | |
| "<MOL_TPSA>": { | |
| "id": "<MOL_TPSA>", | |
| "ids": [ | |
| 16 | |
| ], | |
| "tokens": [ | |
| "<MOL_TPSA>" | |
| ] | |
| }, | |
| "<MOL_WT>": { | |
| "id": "<MOL_WT>", | |
| "ids": [ | |
| 14 | |
| ], | |
| "tokens": [ | |
| "<MOL_WT>" | |
| ] | |
| }, | |
| "<PAD>": { | |
| "id": "<PAD>", | |
| "ids": [ | |
| 0 | |
| ], | |
| "tokens": [ | |
| "<PAD>" | |
| ] | |
| }, | |
| "<PROT_POC_GVP_EMB>": { | |
| "id": "<PROT_POC_GVP_EMB>", | |
| "ids": [ | |
| 21 | |
| ], | |
| "tokens": [ | |
| "<PROT_POC_GVP_EMB>" | |
| ] | |
| }, | |
| "<PROT_SEQ_ESM_EMB>": { | |
| "id": "<PROT_SEQ_ESM_EMB>", | |
| "ids": [ | |
| 20 | |
| ], | |
| "tokens": [ | |
| "<PROT_SEQ_ESM_EMB>" | |
| ] | |
| }, | |
| "<SEP>": { | |
| "id": "<SEP>", | |
| "ids": [ | |
| 4 | |
| ], | |
| "tokens": [ | |
| "<SEP>" | |
| ] | |
| }, | |
| "<UNK>": { | |
| "id": "<UNK>", | |
| "ids": [ | |
| 3 | |
| ], | |
| "tokens": [ | |
| "<UNK>" | |
| ] | |
| }, | |
| "<UNUSED0>": { | |
| "id": "<UNUSED0>", | |
| "ids": [ | |
| 23 | |
| ], | |
| "tokens": [ | |
| "<UNUSED0>" | |
| ] | |
| }, | |
| "<UNUSED1>": { | |
| "id": "<UNUSED1>", | |
| "ids": [ | |
| 24 | |
| ], | |
| "tokens": [ | |
| "<UNUSED1>" | |
| ] | |
| }, | |
| "<UNUSED2>": { | |
| "id": "<UNUSED2>", | |
| "ids": [ | |
| 25 | |
| ], | |
| "tokens": [ | |
| "<UNUSED2>" | |
| ] | |
| }, | |
| "<UNUSED3>": { | |
| "id": "<UNUSED3>", | |
| "ids": [ | |
| 26 | |
| ], | |
| "tokens": [ | |
| "<UNUSED3>" | |
| ] | |
| }, | |
| "<UNUSED4>": { | |
| "id": "<UNUSED4>", | |
| "ids": [ | |
| 27 | |
| ], | |
| "tokens": [ | |
| "<UNUSED4>" | |
| ] | |
| }, | |
| "<UNUSED5>": { | |
| "id": "<UNUSED5>", | |
| "ids": [ | |
| 28 | |
| ], | |
| "tokens": [ | |
| "<UNUSED5>" | |
| ] | |
| }, | |
| "<UNUSED6>": { | |
| "id": "<UNUSED6>", | |
| "ids": [ | |
| 29 | |
| ], | |
| "tokens": [ | |
| "<UNUSED6>" | |
| ] | |
| }, | |
| "<UNUSED7>": { | |
| "id": "<UNUSED7>", | |
| "ids": [ | |
| 30 | |
| ], | |
| "tokens": [ | |
| "<UNUSED7>" | |
| ] | |
| }, | |
| "<UNUSED8>": { | |
| "id": "<UNUSED8>", | |
| "ids": [ | |
| 31 | |
| ], | |
| "tokens": [ | |
| "<UNUSED8>" | |
| ] | |
| }, | |
| "<UNUSED9>": { | |
| "id": "<UNUSED9>", | |
| "ids": [ | |
| 32 | |
| ], | |
| "tokens": [ | |
| "<UNUSED9>" | |
| ] | |
| }, | |
| "<VALUE>": { | |
| "id": "<VALUE>", | |
| "ids": [ | |
| 22 | |
| ], | |
| "tokens": [ | |
| "<VALUE>" | |
| ] | |
| } | |
| } | |
| }, | |
| "decoder": { | |
| "type": "ByteLevel", | |
| "add_prefix_space": true, | |
| "trim_offsets": true, | |
| "use_regex": true | |
| }, | |
| "model": { | |
| "type": "WordLevel", | |
| "vocab": { | |
| "<PAD>": 0, | |
| "<BOS>": 1, | |
| "<EOS>": 2, | |
| "<UNK>": 3, | |
| "<SEP>": 4, | |
| "<MASK>": 5, | |
| "<COND>": 6, | |
| "</COND>": 7, | |
| "<MOL>": 8, | |
| "</MOL>": 9, | |
| "<MOL_HAC>": 10, | |
| "<MOL_HBDC>": 11, | |
| "<MOL_HBAC>": 12, | |
| "<MOL_ROTBC>": 13, | |
| "<MOL_WT>": 14, | |
| "<MOL_LOGP>": 15, | |
| "<MOL_TPSA>": 16, | |
| "<MOL_QED>": 17, | |
| "<MOL_SAS>": 18, | |
| "<MOL_PHARMA_FP>": 19, | |
| "<PROT_SEQ_ESM_EMB>": 20, | |
| "<PROT_POC_GVP_EMB>": 21, | |
| "<VALUE>": 22, | |
| "<UNUSED0>": 23, | |
| "<UNUSED1>": 24, | |
| "<UNUSED2>": 25, | |
| "<UNUSED3>": 26, | |
| "<UNUSED4>": 27, | |
| "<UNUSED5>": 28, | |
| "<UNUSED6>": 29, | |
| "<UNUSED7>": 30, | |
| "<UNUSED8>": 31, | |
| "<UNUSED9>": 32, | |
| "[C]": 33, | |
| "[Dummy]": 34, | |
| "[Branch1]": 35, | |
| "[pop]": 36, | |
| "[=C]": 37, | |
| "[Ring1]": 38, | |
| "[Frag@0]": 39, | |
| "[Attach:0]": 40, | |
| "[CH0]": 41, | |
| "[Attach:1]": 42, | |
| "[=Branch1]": 43, | |
| "[=CH0]": 44, | |
| "[Frag]": 45, | |
| "[N]": 46, | |
| "[O]": 47, | |
| "[CH2]": 48, | |
| "[Frag@1]": 49, | |
| "[#Branch1]": 50, | |
| "[NH0]": 51, | |
| "[Branch2]": 52, | |
| "[=O]": 53, | |
| "[CH1]": 54, | |
| "[=N]": 55, | |
| "[NH1]": 56, | |
| "[OH0]": 57, | |
| "[F]": 58, | |
| "[=Branch2]": 59, | |
| "[=Ring1]": 60, | |
| "[S]": 61, | |
| "[#Branch2]": 62, | |
| "[Cl]": 63, | |
| "[Attach:2]": 64, | |
| "[Ring2]": 65, | |
| "[SH0]": 66, | |
| "[Br]": 67, | |
| "[Frag@2]": 68, | |
| "[#N]": 69, | |
| "[#C]": 70, | |
| "[=Dummy]": 71, | |
| "[=SH0]": 72, | |
| "[P]": 73, | |
| "[=S]": 74, | |
| "[Attach:3]": 75, | |
| "[Frag@3]": 76, | |
| "[NH1+1]": 77, | |
| "[O-1]": 78, | |
| "[=Ring2]": 79, | |
| "[=NH1+1]": 80, | |
| "[NH2+1]": 81, | |
| "[Attach:4]": 82, | |
| "[NH3+1]": 83, | |
| "[#NH1+1]": 84, | |
| "[N-1]": 85, | |
| "[=NH2+1]": 86, | |
| "[Frag@4]": 87, | |
| "[Attach:5]": 88, | |
| "[NH1-1]": 89, | |
| "[PH1]": 90, | |
| "[N+1]": 91, | |
| "[#N+1]": 92, | |
| "[#Ring1]": 93, | |
| "[=P]": 94, | |
| "[Frag@5]": 95, | |
| "[=N+1]": 96, | |
| "[#S]": 97, | |
| "[Attach:6]": 98, | |
| "[#SH0]": 99, | |
| "[Attach:7]": 100, | |
| "[P-1]": 101, | |
| "[SH2]": 102, | |
| "[Xe]": 103, | |
| "[Frag@6]": 104, | |
| "[=SH2]": 105, | |
| "[XeH1]": 106, | |
| "[=N-1]": 107, | |
| "[#Ring2]": 108, | |
| "[B]": 109, | |
| "[=B]": 110, | |
| "[Attach:8]": 111, | |
| "[SH4]": 112, | |
| "[Frag@7]": 113, | |
| "[SH3]": 114, | |
| "[Attach:9]": 115, | |
| "[B+1]": 116, | |
| "[=SH4]": 117, | |
| "[Attach:10]": 118, | |
| "[Attach:11]": 119, | |
| "[Frag@8]": 120, | |
| "[SH5]": 121, | |
| "[Xe+1]": 122, | |
| "[Frag@10]": 123, | |
| "[Frag@9]": 124, | |
| "[=XeH1]": 125, | |
| "[=Xe]": 126 | |
| }, | |
| "unk_token": "<UNK>" | |
| } | |
| } |