zinc10M_gpt2-medium_SMILES_step1 / tokenizer_config.json
jarod0411's picture
Training in progress, step 500
fa1d3d3 verified
{
"add_prefix_space": false,
"added_tokens_decoder": {
"0": {
"content": "<|endoftext|>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": true
},
"3152": {
"content": "<|startoftext|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"3153": {
"content": "[PAD]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"3154": {
"content": "<mask_1:6>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"3155": {
"content": "<mask_3:9>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"3156": {
"content": "<mask_3:15>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"3157": {
"content": "<mask_1:5>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"3158": {
"content": "<mask_3:6>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"3159": {
"content": "<mask_1:7>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"3160": {
"content": "<mask_1:3>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"3161": {
"content": "<mask_2:15>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"3162": {
"content": "<mask_1:14>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"3163": {
"content": "<mask_1:11>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"3164": {
"content": "<mask_2:3>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"3165": {
"content": "<mask_2:2>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"3166": {
"content": "<mask_2:13>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"3167": {
"content": "<mask_3:12>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"3168": {
"content": "<mask_2:5>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"3169": {
"content": "<mask_2:14>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"3170": {
"content": "<mask_3:16>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"3171": {
"content": "<mask_2:7>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"3172": {
"content": "<mask_2:1>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"3173": {
"content": "<mask_3:2>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"3174": {
"content": "<mask_3:10>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"3175": {
"content": "<mask_1:12>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"3176": {
"content": "<mask_3:11>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"3177": {
"content": "<mask_1:13>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"3178": {
"content": "<mask_2:12>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"3179": {
"content": "<mask_3:3>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"3180": {
"content": "<mask_1:8>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"3181": {
"content": "<mask_1:15>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"3182": {
"content": "<mask_1:10>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"3183": {
"content": "<mask_3:14>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"3184": {
"content": "<mask_2:9>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"3185": {
"content": "<mask_2:11>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"3186": {
"content": "<mask_1:9>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"3187": {
"content": "<mask_1:4>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"3188": {
"content": "<mask_3:7>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"3189": {
"content": "<mask_3:8>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"3190": {
"content": "<mask_2:6>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"3191": {
"content": "<mask_3:5>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"3192": {
"content": "<mask_3:13>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"3193": {
"content": "<mask_1:16>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"3194": {
"content": "<mask_2:16>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"3195": {
"content": "<mask_3:4>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"3196": {
"content": "<mask_2:10>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"3197": {
"content": "<mask_1:1>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"3198": {
"content": "<mask_1:2>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"3199": {
"content": "<mask_2:4>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"3200": {
"content": "<mask_2:8>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"3201": {
"content": "<mask_3:1>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
}
},
"additional_special_tokens": [
"<mask_1:6>",
"<mask_3:9>",
"<mask_3:15>",
"<mask_1:5>",
"<mask_3:6>",
"<mask_1:7>",
"<mask_1:3>",
"<mask_2:15>",
"<mask_1:14>",
"<mask_1:11>",
"<mask_2:3>",
"<mask_2:2>",
"<mask_2:13>",
"<mask_3:12>",
"<mask_2:5>",
"<mask_2:14>",
"<mask_3:16>",
"<mask_2:7>",
"<mask_2:1>",
"<mask_3:2>",
"<mask_3:10>",
"<mask_1:12>",
"<mask_3:11>",
"<mask_1:13>",
"<mask_2:12>",
"<mask_3:3>",
"<mask_1:8>",
"<mask_1:15>",
"<mask_1:10>",
"<mask_3:14>",
"<mask_2:9>",
"<mask_2:11>",
"<mask_1:9>",
"<mask_1:4>",
"<mask_3:7>",
"<mask_3:8>",
"<mask_2:6>",
"<mask_3:5>",
"<mask_3:13>",
"<mask_1:16>",
"<mask_2:16>",
"<mask_3:4>",
"<mask_2:10>",
"<mask_1:1>",
"<mask_1:2>",
"<mask_2:4>",
"<mask_2:8>",
"<mask_3:1>"
],
"bos_token": "<|startoftext|>",
"clean_up_tokenization_spaces": true,
"eos_token": "<|endoftext|>",
"model_max_length": 512,
"pad_token": "[PAD]",
"tokenizer_class": "GPT2Tokenizer",
"unk_token": "<|endoftext|>"
}