chemlactica-125m / tokenizer_config.json
yerevann's picture
Upload tokenizer
811b8c8 verified
raw
history blame
No virus
17.6 kB
{
"added_tokens_decoder": {
"0": {
"content": "<s>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"1": {
"content": "<pad>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"2": {
"content": "</s>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"3": {
"content": "<unk>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"4": {
"content": "[START_REF]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"5": {
"content": "[END_REF]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"6": {
"content": "[IMAGE]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"7": {
"content": "<fragments>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"8": {
"content": "</fragments>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"9": {
"content": "<work>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"10": {
"content": "</work>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"11": {
"content": "[START_SUP]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"12": {
"content": "[END_SUP]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"13": {
"content": "[START_SUB]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"14": {
"content": "[END_SUB]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"15": {
"content": "[START_DNA]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"16": {
"content": "[END_DNA]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"17": {
"content": "[START_AMINO]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"18": {
"content": "[END_AMINO]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"19": {
"content": "[START_SMILES]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"20": {
"content": "[END_SMILES]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"21": {
"content": "[START_I_SMILES]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"22": {
"content": "[END_I_SMILES]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50000": {
"content": "[SYNONYM]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50001": {
"content": "[RELATED]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50002": {
"content": "[SIMILAR]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50003": {
"content": "[PROPERTY]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50004": {
"content": "[SAS]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50005": {
"content": "[WEIGHT]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50006": {
"content": "[TPSA]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50007": {
"content": "[CLOGP]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50008": {
"content": "[QED]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50009": {
"content": "[NUMHDONORS]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50010": {
"content": "[NUMHACCEPTORS]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50011": {
"content": "[NUMHETEROATOMS]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50012": {
"content": "[NUMROTATABLEBONDS]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50013": {
"content": "[NOCOUNT]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50014": {
"content": "[NHOHCOUNT]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50015": {
"content": "[RINGCOUNT]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50016": {
"content": "[HEAVYATOMCOUNT]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50017": {
"content": "[FRACTIONCSP3]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50018": {
"content": "[NUMAROMATICRINGS]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50019": {
"content": "[NUMSATURATEDRINGS]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50020": {
"content": "[NUMAROMATICHETEROCYCLES]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50021": {
"content": "[NUMAROMATICCARBOCYCLES]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50022": {
"content": "[NUMSATURATEDHETEROCYCLES]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50023": {
"content": "[NUMSATURATEDCARBOCYCLES]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50024": {
"content": "[NUMALIPHATICRINGS]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50025": {
"content": "[NUMALIPHATICHETEROCYCLES]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50026": {
"content": "[NUMALIPHATICCARBOCYCLES]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50027": {
"content": "[IUPAC]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50028": {
"content": "[VAR_NAME]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50029": {
"content": "[VAR_DESC]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50030": {
"content": "[VAR_VAL]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50031": {
"content": "[ASSAY_NAME]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50032": {
"content": "[ASSAY_DESC]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50033": {
"content": "[/SYNONYM]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50034": {
"content": "[/RELATED]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50035": {
"content": "[/SIMILAR]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50036": {
"content": "[/PROPERTY]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50037": {
"content": "[/SAS]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50038": {
"content": "[/WEIGHT]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50039": {
"content": "[/TPSA]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50040": {
"content": "[/CLOGP]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50041": {
"content": "[/QED]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50042": {
"content": "[/NUMHDONORS]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50043": {
"content": "[/NUMHACCEPTORS]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50044": {
"content": "[/NUMHETEROATOMS]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50045": {
"content": "[/NUMROTATABLEBONDS]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50046": {
"content": "[/NOCOUNT]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50047": {
"content": "[/NHOHCOUNT]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50048": {
"content": "[/RINGCOUNT]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50049": {
"content": "[/HEAVYATOMCOUNT]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50050": {
"content": "[/FRACTIONCSP3]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50051": {
"content": "[/NUMAROMATICRINGS]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50052": {
"content": "[/NUMSATURATEDRINGS]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50053": {
"content": "[/NUMAROMATICHETEROCYCLES]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50054": {
"content": "[/NUMAROMATICCARBOCYCLES]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50055": {
"content": "[/NUMSATURATEDHETEROCYCLES]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50056": {
"content": "[/NUMSATURATEDCARBOCYCLES]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50057": {
"content": "[/NUMALIPHATICRINGS]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50058": {
"content": "[/NUMALIPHATICHETEROCYCLES]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50059": {
"content": "[/NUMALIPHATICCARBOCYCLES]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50060": {
"content": "[/IUPAC]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50061": {
"content": "[/VAR_NAME]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50062": {
"content": "[/VAR_DESC]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50063": {
"content": "[/VAR_VAL]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50064": {
"content": "[/ASSAY_NAME]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"50065": {
"content": "[/ASSAY_DESC]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
}
},
"additional_special_tokens": [
"[SYNONYM]",
"[RELATED]",
"[SIMILAR]",
"[PROPERTY]",
"[SAS]",
"[WEIGHT]",
"[TPSA]",
"[CLOGP]",
"[QED]",
"[NUMHDONORS]",
"[NUMHACCEPTORS]",
"[NUMHETEROATOMS]",
"[NUMROTATABLEBONDS]",
"[NOCOUNT]",
"[NHOHCOUNT]",
"[RINGCOUNT]",
"[HEAVYATOMCOUNT]",
"[FRACTIONCSP3]",
"[NUMAROMATICRINGS]",
"[NUMSATURATEDRINGS]",
"[NUMAROMATICHETEROCYCLES]",
"[NUMAROMATICCARBOCYCLES]",
"[NUMSATURATEDHETEROCYCLES]",
"[NUMSATURATEDCARBOCYCLES]",
"[NUMALIPHATICRINGS]",
"[NUMALIPHATICHETEROCYCLES]",
"[NUMALIPHATICCARBOCYCLES]",
"[IUPAC]",
"[VAR_NAME]",
"[VAR_DESC]",
"[VAR_VAL]",
"[ASSAY_NAME]",
"[ASSAY_DESC]",
"[/SYNONYM]",
"[/RELATED]",
"[/SIMILAR]",
"[/PROPERTY]",
"[/SAS]",
"[/WEIGHT]",
"[/TPSA]",
"[/CLOGP]",
"[/QED]",
"[/NUMHDONORS]",
"[/NUMHACCEPTORS]",
"[/NUMHETEROATOMS]",
"[/NUMROTATABLEBONDS]",
"[/NOCOUNT]",
"[/NHOHCOUNT]",
"[/RINGCOUNT]",
"[/HEAVYATOMCOUNT]",
"[/FRACTIONCSP3]",
"[/NUMAROMATICRINGS]",
"[/NUMSATURATEDRINGS]",
"[/NUMAROMATICHETEROCYCLES]",
"[/NUMAROMATICCARBOCYCLES]",
"[/NUMSATURATEDHETEROCYCLES]",
"[/NUMSATURATEDCARBOCYCLES]",
"[/NUMALIPHATICRINGS]",
"[/NUMALIPHATICHETEROCYCLES]",
"[/NUMALIPHATICCARBOCYCLES]",
"[/IUPAC]",
"[/VAR_NAME]",
"[/VAR_DESC]",
"[/VAR_VAL]",
"[/ASSAY_NAME]",
"[/ASSAY_DESC]"
],
"bos_token": "<s>",
"clean_up_tokenization_spaces": true,
"eos_token": "</s>",
"model_max_length": 1000000000000000019884624838656,
"pad_token": "<pad>",
"tokenizer_class": "PreTrainedTokenizerFast"
}