Donuttest / tokenizer_config.json
PeanutCoding's picture
End of training
34c0454 verified
raw
history blame
17.2 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
},
"57521": {
"content": "<mask>",
"lstrip": true,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": true
},
"57522": {
"content": "<sep/>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57523": {
"content": "<s_iitcdip>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57524": {
"content": "<s_synthdog>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57525": {
"content": "<s_Var-Ausf. 4>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57526": {
"content": "</s_Var-Ausf. 4>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57527": {
"content": "<s_Var-Ausf. 3>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57528": {
"content": "</s_Var-Ausf. 3>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57529": {
"content": "<s_Var-Ausf. 2>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57530": {
"content": "</s_Var-Ausf. 2>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57531": {
"content": "<s_Var-Ausf 1>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57532": {
"content": "</s_Var-Ausf 1>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57533": {
"content": "<s_Termin Kundenwunsch - KW>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57534": {
"content": "</s_Termin Kundenwunsch - KW>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57535": {
"content": "<s_Sonderrabatt>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57536": {
"content": "</s_Sonderrabatt>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57537": {
"content": "<s_Möbelhaus>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57538": {
"content": "</s_Möbelhaus>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57539": {
"content": "<s_Modell 4>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57540": {
"content": "</s_Modell 4>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57541": {
"content": "<s_Modell 3>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57542": {
"content": "</s_Modell 3>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57543": {
"content": "<s_Modell 2>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57544": {
"content": "</s_Modell 2>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57545": {
"content": "<s_Modell 1>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57546": {
"content": "</s_Modell 1>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57547": {
"content": "<s_Menge4>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57548": {
"content": "</s_Menge4>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57549": {
"content": "<s_Menge3>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57550": {
"content": "</s_Menge3>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57551": {
"content": "<s_Menge2>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57552": {
"content": "</s_Menge2>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57553": {
"content": "<s_Menge1>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57554": {
"content": "</s_Menge1>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57555": {
"content": "<s_Liefertermin>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57556": {
"content": "</s_Liefertermin>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57557": {
"content": "<s_Kundennr.>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57558": {
"content": "</s_Kundennr.>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57559": {
"content": "<s_Kommission>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57560": {
"content": "</s_Kommission>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57561": {
"content": "<s_Holz 3>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57562": {
"content": "</s_Holz 3>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57563": {
"content": "<s_Holz 1>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57564": {
"content": "</s_Holz 1>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57565": {
"content": "<s_Bezug 4>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57566": {
"content": "</s_Bezug 4>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57567": {
"content": "<s_Bezug 3>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57568": {
"content": "</s_Bezug 3>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57569": {
"content": "<s_Bezug 2>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57570": {
"content": "</s_Bezug 2>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57571": {
"content": "<s_Bezug 1>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57572": {
"content": "</s_Bezug 1>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57573": {
"content": "<s_Bestellnummer>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57574": {
"content": "</s_Bestellnummer>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57575": {
"content": "<s_AB-Nummer>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57576": {
"content": "</s_AB-Nummer>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57577": {
"content": "<s_PN/BM 1>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57578": {
"content": "</s_PN/BM 1>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57579": {
"content": "<s_Last Viewed>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57580": {
"content": "</s_Last Viewed>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57581": {
"content": "<s_Transportraum>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57582": {
"content": "</s_Transportraum>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57583": {
"content": "<s_Zusatz 1>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57584": {
"content": "</s_Zusatz 1>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57585": {
"content": "<s_Holz 2>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57586": {
"content": "</s_Holz 2>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57587": {
"content": "<s_Notizen 2>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57588": {
"content": "</s_Notizen 2>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57589": {
"content": "<s_Notizen 1>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57590": {
"content": "</s_Notizen 1>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57591": {
"content": "<s_PN/BM 3>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57592": {
"content": "</s_PN/BM 3>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57593": {
"content": "<s_PN/BM 2>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57594": {
"content": "</s_PN/BM 2>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57595": {
"content": "<s_LA-Anschrift>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57596": {
"content": "</s_LA-Anschrift>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57597": {
"content": "<s_SA-Schlüssel 3>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57598": {
"content": "</s_SA-Schlüssel 3>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57599": {
"content": "<s_Aufpreis 3>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57600": {
"content": "</s_Aufpreis 3>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
}
},
"additional_special_tokens": [
"<s_Var-Ausf. 4>",
"</s_Var-Ausf. 4>",
"<s_Var-Ausf. 3>",
"</s_Var-Ausf. 3>",
"<s_Var-Ausf. 2>",
"</s_Var-Ausf. 2>",
"<s_Var-Ausf 1>",
"</s_Var-Ausf 1>",
"<s_Termin Kundenwunsch - KW>",
"</s_Termin Kundenwunsch - KW>",
"<s_Sonderrabatt>",
"</s_Sonderrabatt>",
"<s_Möbelhaus>",
"</s_Möbelhaus>",
"<s_Modell 4>",
"</s_Modell 4>",
"<s_Modell 3>",
"</s_Modell 3>",
"<s_Modell 2>",
"</s_Modell 2>",
"<s_Modell 1>",
"</s_Modell 1>",
"<s_Menge4>",
"</s_Menge4>",
"<s_Menge3>",
"</s_Menge3>",
"<s_Menge2>",
"</s_Menge2>",
"<s_Menge1>",
"</s_Menge1>",
"<s_Liefertermin>",
"</s_Liefertermin>",
"<s_Kundennr.>",
"</s_Kundennr.>",
"<s_Kommission>",
"</s_Kommission>",
"<s_Holz 3>",
"</s_Holz 3>",
"<s_Holz 1>",
"</s_Holz 1>",
"<s_Bezug 4>",
"</s_Bezug 4>",
"<s_Bezug 3>",
"</s_Bezug 3>",
"<s_Bezug 2>",
"</s_Bezug 2>",
"<s_Bezug 1>",
"</s_Bezug 1>",
"<s_Bestellnummer>",
"</s_Bestellnummer>",
"<s_AB-Nummer>",
"</s_AB-Nummer>",
"<s_PN/BM 1>",
"</s_PN/BM 1>",
"<s_Last Viewed>",
"</s_Last Viewed>",
"<s_Transportraum>",
"</s_Transportraum>",
"<s_Zusatz 1>",
"</s_Zusatz 1>",
"<s_Holz 2>",
"</s_Holz 2>",
"<s_Notizen 2>",
"</s_Notizen 2>",
"<s_Notizen 1>",
"</s_Notizen 1>",
"<s_PN/BM 3>",
"</s_PN/BM 3>",
"<s_PN/BM 2>",
"</s_PN/BM 2>",
"<s_LA-Anschrift>",
"</s_LA-Anschrift>",
"<s_SA-Schlüssel 3>",
"</s_SA-Schlüssel 3>",
"<s_Aufpreis 3>",
"</s_Aufpreis 3>",
"<s>",
"</s>"
],
"bos_token": "<s>",
"clean_up_tokenization_spaces": false,
"cls_token": "<s>",
"eos_token": "</s>",
"extra_special_tokens": {},
"mask_token": "<mask>",
"model_max_length": 1000000000000000019884624838656,
"pad_token": "<pad>",
"processor_class": "DonutProcessor",
"sep_token": "</s>",
"sp_model_kwargs": {},
"tokenizer_class": "XLMRobertaTokenizer",
"unk_token": "<unk>"
}