Upload 3 files
Browse filestokenizer uploaded
- special_tokens_map.json +7 -0
- tokenizer.json +972 -0
- tokenizer_config.json +54 -0
special_tokens_map.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<pad>",
|
| 3 |
+
"eos_token": "<eos>",
|
| 4 |
+
"pad_token": "<bos>",
|
| 5 |
+
"unk_token": "<unk>",
|
| 6 |
+
"mask_token": "<mask>"
|
| 7 |
+
}
|
tokenizer.json
ADDED
|
@@ -0,0 +1,972 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"version": "1.0",
|
| 3 |
+
"truncation": null,
|
| 4 |
+
"padding": null,
|
| 5 |
+
"added_tokens": [
|
| 6 |
+
{
|
| 7 |
+
"id": 0,
|
| 8 |
+
"content": "<pad>",
|
| 9 |
+
"single_word": false,
|
| 10 |
+
"lstrip": false,
|
| 11 |
+
"rstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"special": true
|
| 14 |
+
},
|
| 15 |
+
{
|
| 16 |
+
"id": 1,
|
| 17 |
+
"content": "<eos>",
|
| 18 |
+
"single_word": false,
|
| 19 |
+
"lstrip": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"normalized": false,
|
| 22 |
+
"special": true
|
| 23 |
+
},
|
| 24 |
+
{
|
| 25 |
+
"id": 2,
|
| 26 |
+
"content": "<bos>",
|
| 27 |
+
"single_word": false,
|
| 28 |
+
"lstrip": false,
|
| 29 |
+
"rstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"special": true
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"id": 3,
|
| 35 |
+
"content": "<unk>",
|
| 36 |
+
"single_word": false,
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"rstrip": false,
|
| 39 |
+
"normalized": false,
|
| 40 |
+
"special": true
|
| 41 |
+
},
|
| 42 |
+
{
|
| 43 |
+
"id": 4,
|
| 44 |
+
"content": "<mask>",
|
| 45 |
+
"single_word": false,
|
| 46 |
+
"lstrip": false,
|
| 47 |
+
"rstrip": false,
|
| 48 |
+
"normalized": false,
|
| 49 |
+
"special": true
|
| 50 |
+
}
|
| 51 |
+
],
|
| 52 |
+
"normalizer": null,
|
| 53 |
+
"pre_tokenizer": {
|
| 54 |
+
"type": "Sequence",
|
| 55 |
+
"pretokenizers": [
|
| 56 |
+
{
|
| 57 |
+
"type": "Split",
|
| 58 |
+
"pattern": {
|
| 59 |
+
"Regex": "<[^>]+>|\\.\\d+,\\d+|[1-9]{2}\\-[a-z]\\]|[0-9]\\-[a-z]\\]|[1-9]{2}[a-z]|[1-9]{2}'[a-z]|[0-9]'[a-z]|[0-9][a-z]|\\([0-9]\\+\\)|\\([0-9]\\-\\)|[1-9]{2}|[0-9]|-|\\s|\\(|\\)|S|R|E|Z|N|C|O|'|\\\"|;|λ|H|,|\\.|\\[[a-z]{2}\\]|\\[[a-z]\\]|\\[|\\]|indolizin|arsindol|indol|furan|furo|piperazin|pyrrolidin|pyrrolizin|thiophen|thiolo|imidazolidin|imidazol|pyrimidin|pyridin|piperidin|morpholin|pyrazol|pyridazin|oxocinnolin|cinnolin|pyrrol|thiochromen|oxochromen|chromen|quinazolin|phthalazin|quinoxalin|carbazol|xanthen|pyrazin|purin|indazol|naphthyridin|quinolizin|guanidin|pyranthren|pyran|thianthren|thian|acridin|acrido|yohimban|porphyrin|pteridin|tetramin|pentamin|borinin|borino|boriran|borolan|borol|borinan|phenanthridin|quinolin|perimidin|corrin|phenanthrolin|phosphinolin|indacen|silonin|borepin|naphthalen|naphthal|inden|adamant|fluoren|thiourea|urea|anthracen|acenaphthylen|carbohydrazide|annulen|aniline|acetaldehyde|benzaldehyde|formaldehyde|phthalaldehyde|acephenanthrylen|phenanthren|chrysen|carbanid|chloroform|fulleren|cumen|formonitril|fluoranthen|terephthalaldehyde|azulen|picen|pyren|pleiaden|coronen|tetracen|pentacen|perylen|pentalen|heptalen|cuban|hexacen|oxanthren|ovalen|aceanthrylen|acid|ether|hydrogen|helium|lithium|beryllium|nitrogen|oxygen|fluorine|neon|sodium|magnesium|aluminum|silicon|phosphorus|sulfur|chlorine|argon|potassium|calcium|scandium|titanium|vanadium|chromium|manganese|iron|cobalt|nickel|copper|zinc|gallium|germanium|arsenic|selenium|bromine|krypton|rubidium|yttrium|zirconium|niobium|molybdenum|technetium|ruthenium|rhodium|palladium|silver|cadmium|indium|antimony|tellurium|iodine|xenon|cesium|barium|lanthanum|cerium|praseodymium|neodymium|latinum|promethium|samarium|europium|gadolinium|terbium|dysprosium|holmium|erbium|thulium|ytterbium|lutetium|hafnium|tantalum|tungsten|rhenium|osmium|iridium|platinum|gold|aurum|mercury|thallium|lead|bismuth|polonium|astatine|radon|francium|radium|actinium|thorium|protactinium|uranium|neptunium|plutonium|americium|curium|berkelium|einsteinium|fermium|californium|mendelevium|nobelium|lawrencium|rutherfordium|dubnium|seaborgium|bohrium|hassium|meitnerium|tin|protio|deuterio|tritio|perchlorate|perbromate|periodate|hypofluorite|hypochlorite|hypobromite|hypoiodite|nitrate|silicate|hydride|methane|methanoyl|methan|ethane|ethanoyl|ethan|propanoyl|propane|propan|propa|butane|butanoyl|butan|buta|pentane|pentanoyl|pentan|hexane|hexanoyl|hexan|heptane|heptanoyl|heptan|octane|octanoyl|octan|nonane|nonanoyl|nonan|decane|decanoyl|decan|icosane|icosan|cosane|cosan|contane|contan|mono|un|bis|bi|dicta|di|tetraza|tetraz|tetra|tetr|pentaza|pentaz|penta|hexaza|hexa|heptaza|hepta|octaza|octa|nonaza|nona|decaza|deca|kis|hydroxide|hydroxyl|hydroxy|hydrate|hydro|cyclo|spiro|iso|methylidene|methyl|ethyl|isopropyl|propyl|isobutyl|sec-butyl|tert-butyl|butyl|pentyl|hexyl|heptyl|octyl|meth|eth|prop|but|pent|hex|hept|oct|non|dec|icosa|icos|cosa|cos|icon|conta|cont|con|heni|hene|hen|hecta|hect|silanide|silane|silole|silanyl|silyloxy|silylo|silyl|sila|hydrazono|hydrazino|nitroso|nitrous|nitro|formamido|amino|amido|imino|imido|anilino|anilin|thiocyanato|cyanato|cyano|azido|azanidyl|azanyl|azanide|azanida|azonia|azonio|amidino|nitramido|diazo|sulfanylidene|sulfinamoyl|sulfonimidoyl|sulfinimidoyl|sulfamoyl|sulfonyl|sulfanyl|sulfinyl|sulfinato|sulfenato|sulfonato|sulfonio|sulfino|sulfono|sulfido|carboxylate|carboxylato|carboxylic|carboxy|halocarbonyl|carbamoyl|carbonyl|carbamo|thioformyl|formyl|alumanyl|gallanyl|stannyl|plumbyl|tellanyl|germanyl|germyl|benzhydryl|benzoxaza|benzoxaz|benzoxa|benzox|benzo|benzyl|benz|phenacyl|phenanthro|phenyl|phenoxaza|phenoxaz|phenoxy|phenox|phenol|pheno|phen|acetyl|aceto|acet|peroxy|oxido|oxino|oxalo|oxolo|oxocyclo|oxol|oxoc|oxon|oxo|oxy|pyrido|pyrimido|imidazo|naphtho|stiboryl|stibolo|fluoro|fluoranyl|fluoridoyl|fluorido|chloro|chloranyl|chloridoyl|chlorido|bromo|bromanyl|bromidoyl|bromido|iodo|iodanyl|iodidoyl|iodanuidyl|iodido|phosphonato|phosphoroso|phosphonia|phosphoryl|phosphanyl|arsono|arsanyl|stiba|carbonitrile|carboxamide|carbamimidothioate|carbodithioate|carbohydrazonate|carbonimidoyl|carboximidoyl|carbamimidoyl|carbamimidate|carbamimid|carbaldehyde|carbamate|carbothioyl|carboximidothioate|carbonate|carboximidamide|carboximidate|carbamic|carbonochloridate|carbothialdehyde|carbothioate|carbothioic|carbono|carbon|carbo|formate|formic|ammonium|hydrazide|hydrazine|hydrazin|amine|imine|oxamide|nitramide|formamide|cyanamide|amide|imide|amidine|isocyanide|azanium|thiocyanate|cyanate|cyanic|cyanatidoyl|cyanide|nitrile|nitrite|hydrazonate|sulfonamide|sulfinamide|sulfonamido|sulfonic|sulfamic|sulfinic|sulfuric|thial|thione|thiol|sulfonate|sulfite|sulfate|sulfide|sulfinate|sulfanium|sulfamate|sulfane|sulfo|fluoride|chloride|chloridic|perchloric|bromide|iodide|iodane|hypoiodous|hypochlorous|phosphanium|phosphate|phosphite|phosphane|phosphanide|phosphonamidic|phosphonous|phosphinous|phosphinite|phosphono|arsonic|stibane|chromium|stannane|gallane|alumane|aluminane|aluminan|germane|germa|ane|ano|an|ene|enoxy|eno|en|yne|yn|yl|peroxol|peroxo|terephthalate|terephthalic|phthalic|phthalate|oxide|oate|ol|oic|ic|al|ate|ium|one|ir|et|olo|ol|ino|in|ep|oc|on|ec|protide|boranyl|boranuide|boronamidic|boranuida|boranide|borinic|borate|borane|boran|borono|boron|bora|selanyl|seleno|thiaz|oxaza|oxaz|oxan|oxa|ox|aza|az|thia|thioc|thion|thio|thi|telluro|phospha|phosph|selen|bor|sil|alum|ars|germ|tellur|imid|idine|idene|idin|ane|an|ine|in|id|e|do|trisodium|tris|triacetyl|triamine|triaza|triaz|tria|trityl|tri|o"
|
| 60 |
+
},
|
| 61 |
+
"behavior": "Isolated",
|
| 62 |
+
"invert": false
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"type": "Split",
|
| 66 |
+
"pattern": {
|
| 67 |
+
"Regex": "\\[|\\]|\\(|\\)|<[^>]+>|0|1|2|3|4|5|6|7|8|9|-|meth|ane|eth|prop|but|pent|hex|hept|oct|non|dec|cos|cont|icos|benz|phen|amine|cyan|imine|amide|cyan|imid|nitril|nitr|ane|ano|ene|en|ox|per|ol|on|oic|oate|tris|iso|yl|ate|az|hypo|sil|bor|thio|carb|ferrum|sulf|chlor|brom|phosph|hydr|fluor"
|
| 68 |
+
},
|
| 69 |
+
"behavior": "Isolated",
|
| 70 |
+
"invert": false
|
| 71 |
+
}
|
| 72 |
+
]
|
| 73 |
+
},
|
| 74 |
+
"post_processor": {
|
| 75 |
+
"type": "TemplateProcessing",
|
| 76 |
+
"single": [
|
| 77 |
+
{
|
| 78 |
+
"SpecialToken": {
|
| 79 |
+
"id": "<bos>",
|
| 80 |
+
"type_id": 0
|
| 81 |
+
}
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"Sequence": {
|
| 85 |
+
"id": "A",
|
| 86 |
+
"type_id": 0
|
| 87 |
+
}
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"SpecialToken": {
|
| 91 |
+
"id": "<eos>",
|
| 92 |
+
"type_id": 0
|
| 93 |
+
}
|
| 94 |
+
}
|
| 95 |
+
],
|
| 96 |
+
"pair": [
|
| 97 |
+
{
|
| 98 |
+
"Sequence": {
|
| 99 |
+
"id": "A",
|
| 100 |
+
"type_id": 0
|
| 101 |
+
}
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"Sequence": {
|
| 105 |
+
"id": "B",
|
| 106 |
+
"type_id": 1
|
| 107 |
+
}
|
| 108 |
+
}
|
| 109 |
+
],
|
| 110 |
+
"special_tokens": {
|
| 111 |
+
"<bos>": {
|
| 112 |
+
"id": "<bos>",
|
| 113 |
+
"ids": [
|
| 114 |
+
2
|
| 115 |
+
],
|
| 116 |
+
"tokens": [
|
| 117 |
+
"<bos>"
|
| 118 |
+
]
|
| 119 |
+
},
|
| 120 |
+
"<eos>": {
|
| 121 |
+
"id": "<eos>",
|
| 122 |
+
"ids": [
|
| 123 |
+
1
|
| 124 |
+
],
|
| 125 |
+
"tokens": [
|
| 126 |
+
"<eos>"
|
| 127 |
+
]
|
| 128 |
+
},
|
| 129 |
+
"<pad>": {
|
| 130 |
+
"id": "<pad>",
|
| 131 |
+
"ids": [
|
| 132 |
+
0
|
| 133 |
+
],
|
| 134 |
+
"tokens": [
|
| 135 |
+
"<pad>"
|
| 136 |
+
]
|
| 137 |
+
}
|
| 138 |
+
}
|
| 139 |
+
},
|
| 140 |
+
"decoder": {
|
| 141 |
+
"type": "BPEDecoder",
|
| 142 |
+
"suffix": "</w>"
|
| 143 |
+
},
|
| 144 |
+
"model": {
|
| 145 |
+
"type": "WordLevel",
|
| 146 |
+
"vocab": {
|
| 147 |
+
"<pad>": 0,
|
| 148 |
+
"<eos>": 1,
|
| 149 |
+
"<bos>": 2,
|
| 150 |
+
"<unk>": 3,
|
| 151 |
+
"<mask>": 4,
|
| 152 |
+
"-": 5,
|
| 153 |
+
"yl": 6,
|
| 154 |
+
"2": 7,
|
| 155 |
+
",": 8,
|
| 156 |
+
"1": 9,
|
| 157 |
+
"(": 10,
|
| 158 |
+
")": 11,
|
| 159 |
+
"[": 12,
|
| 160 |
+
"]": 13,
|
| 161 |
+
"3": 14,
|
| 162 |
+
"4": 15,
|
| 163 |
+
"meth": 16,
|
| 164 |
+
"ox": 17,
|
| 165 |
+
"o": 18,
|
| 166 |
+
"5": 19,
|
| 167 |
+
"phen": 20,
|
| 168 |
+
"an": 21,
|
| 169 |
+
"ol": 22,
|
| 170 |
+
"y": 23,
|
| 171 |
+
"az": 24,
|
| 172 |
+
"e": 25,
|
| 173 |
+
"di": 26,
|
| 174 |
+
"6": 27,
|
| 175 |
+
"eth": 28,
|
| 176 |
+
"N": 29,
|
| 177 |
+
"a": 30,
|
| 178 |
+
"prop": 31,
|
| 179 |
+
"amino": 32,
|
| 180 |
+
" ": 33,
|
| 181 |
+
"on": 34,
|
| 182 |
+
"carb": 35,
|
| 183 |
+
"hydr": 36,
|
| 184 |
+
"in": 37,
|
| 185 |
+
"id": 38,
|
| 186 |
+
"7": 39,
|
| 187 |
+
"S": 40,
|
| 188 |
+
"benz": 41,
|
| 189 |
+
"amide": 42,
|
| 190 |
+
"R": 43,
|
| 191 |
+
"fluor": 44,
|
| 192 |
+
".": 45,
|
| 193 |
+
"0": 46,
|
| 194 |
+
"cyclo": 47,
|
| 195 |
+
"en": 48,
|
| 196 |
+
"ene": 49,
|
| 197 |
+
"tri": 50,
|
| 198 |
+
"8": 51,
|
| 199 |
+
"but": 52,
|
| 200 |
+
"chlor": 53,
|
| 201 |
+
"pyr": 54,
|
| 202 |
+
"9": 55,
|
| 203 |
+
";": 56,
|
| 204 |
+
"idin": 57,
|
| 205 |
+
"imid": 58,
|
| 206 |
+
"H": 59,
|
| 207 |
+
"sulf": 60,
|
| 208 |
+
"acid": 61,
|
| 209 |
+
"per": 62,
|
| 210 |
+
"hex": 63,
|
| 211 |
+
"pi": 64,
|
| 212 |
+
"ic": 65,
|
| 213 |
+
"pent": 66,
|
| 214 |
+
"pyridin": 67,
|
| 215 |
+
"amine": 68,
|
| 216 |
+
"pyrr": 69,
|
| 217 |
+
"ane": 70,
|
| 218 |
+
"tetra": 71,
|
| 219 |
+
"acet": 72,
|
| 220 |
+
"thio": 73,
|
| 221 |
+
"thi": 74,
|
| 222 |
+
"ate": 75,
|
| 223 |
+
"et": 76,
|
| 224 |
+
"dec": 77,
|
| 225 |
+
"brom": 78,
|
| 226 |
+
"Z": 79,
|
| 227 |
+
"bis": 80,
|
| 228 |
+
"r": 81,
|
| 229 |
+
"k": 82,
|
| 230 |
+
"E": 83,
|
| 231 |
+
"ind": 84,
|
| 232 |
+
"iso": 85,
|
| 233 |
+
"ano": 86,
|
| 234 |
+
"quin": 87,
|
| 235 |
+
"'": 88,
|
| 236 |
+
"tert": 89,
|
| 237 |
+
"st": 90,
|
| 238 |
+
"b": 91,
|
| 239 |
+
"amo": 92,
|
| 240 |
+
"furan": 93,
|
| 241 |
+
"hept": 94,
|
| 242 |
+
"oct": 95,
|
| 243 |
+
"nitr": 96,
|
| 244 |
+
"cyan": 97,
|
| 245 |
+
"ide": 98,
|
| 246 |
+
"oic": 99,
|
| 247 |
+
"c": 100,
|
| 248 |
+
"morph": 101,
|
| 249 |
+
"oate": 102,
|
| 250 |
+
"l": 103,
|
| 251 |
+
"phosph": 104,
|
| 252 |
+
"d": 105,
|
| 253 |
+
"i": 106,
|
| 254 |
+
"spiro": 107,
|
| 255 |
+
"anilino": 108,
|
| 256 |
+
"bi": 109,
|
| 257 |
+
"ep": 110,
|
| 258 |
+
"naphthal": 111,
|
| 259 |
+
"thia": 112,
|
| 260 |
+
"t": 113,
|
| 261 |
+
"ium": 114,
|
| 262 |
+
"n": 115,
|
| 263 |
+
"non": 116,
|
| 264 |
+
"am": 117,
|
| 265 |
+
"cos": 118,
|
| 266 |
+
"ino": 119,
|
| 267 |
+
"sil": 120,
|
| 268 |
+
"do": 121,
|
| 269 |
+
"yn": 122,
|
| 270 |
+
"pyran": 123,
|
| 271 |
+
"al": 124,
|
| 272 |
+
"urea": 125,
|
| 273 |
+
"m": 126,
|
| 274 |
+
"itrile": 127,
|
| 275 |
+
"tris": 128,
|
| 276 |
+
"guanidin": 129,
|
| 277 |
+
"tetr": 130,
|
| 278 |
+
"qu": 131,
|
| 279 |
+
"pip": 132,
|
| 280 |
+
"p": 133,
|
| 281 |
+
"imino": 134,
|
| 282 |
+
"pyrid": 135,
|
| 283 |
+
"s": 136,
|
| 284 |
+
"nitril": 137,
|
| 285 |
+
"v": 138,
|
| 286 |
+
"chrom": 139,
|
| 287 |
+
"amido": 140,
|
| 288 |
+
"or": 141,
|
| 289 |
+
"pyrido": 142,
|
| 290 |
+
"yr": 143,
|
| 291 |
+
"x": 144,
|
| 292 |
+
"iodide": 145,
|
| 293 |
+
"iodo": 146,
|
| 294 |
+
"bor": 147,
|
| 295 |
+
"aniline": 148,
|
| 296 |
+
"h": 149,
|
| 297 |
+
"un": 150,
|
| 298 |
+
"anium": 151,
|
| 299 |
+
"amic": 152,
|
| 300 |
+
"cont": 153,
|
| 301 |
+
"anthr": 154,
|
| 302 |
+
"deuterio": 155,
|
| 303 |
+
"acr": 156,
|
| 304 |
+
"furo": 157,
|
| 305 |
+
"purin": 158,
|
| 306 |
+
"icos": 159,
|
| 307 |
+
"naphth": 160,
|
| 308 |
+
"fur": 161,
|
| 309 |
+
"idine": 162,
|
| 310 |
+
"ec": 163,
|
| 311 |
+
"tria": 164,
|
| 312 |
+
"aldehyde": 165,
|
| 313 |
+
"O": 166,
|
| 314 |
+
"kis": 167,
|
| 315 |
+
"f": 168,
|
| 316 |
+
"ido": 169,
|
| 317 |
+
"tin": 170,
|
| 318 |
+
"alin": 171,
|
| 319 |
+
"adamant": 172,
|
| 320 |
+
"iodan": 173,
|
| 321 |
+
"aceto": 174,
|
| 322 |
+
"ambda": 175,
|
| 323 |
+
"naphthyridin": 176,
|
| 324 |
+
"ir": 177,
|
| 325 |
+
"form": 178,
|
| 326 |
+
"anthrac": 179,
|
| 327 |
+
"og": 180,
|
| 328 |
+
"nip": 181,
|
| 329 |
+
"rt": 182,
|
| 330 |
+
"amm": 183,
|
| 331 |
+
"is": 184,
|
| 332 |
+
"ne": 185,
|
| 333 |
+
"lambda": 186,
|
| 334 |
+
"+": 187,
|
| 335 |
+
"imine": 188,
|
| 336 |
+
"g": 189,
|
| 337 |
+
"amidine": 190,
|
| 338 |
+
"naphtho": 191,
|
| 339 |
+
"furfur": 192,
|
| 340 |
+
"oc": 193,
|
| 341 |
+
"xy": 194,
|
| 342 |
+
"sodium": 195,
|
| 343 |
+
"C": 196,
|
| 344 |
+
"xanth": 197,
|
| 345 |
+
"ato": 198,
|
| 346 |
+
"ine": 199,
|
| 347 |
+
"thian": 200,
|
| 348 |
+
"izin": 201,
|
| 349 |
+
"ia": 202,
|
| 350 |
+
"chr": 203,
|
| 351 |
+
"phthal": 204,
|
| 352 |
+
"ur": 205,
|
| 353 |
+
"umar": 206,
|
| 354 |
+
"thr": 207,
|
| 355 |
+
"acridin": 208,
|
| 356 |
+
"amidino": 209,
|
| 357 |
+
"anide": 210,
|
| 358 |
+
"rcapt": 211,
|
| 359 |
+
"sec": 212,
|
| 360 |
+
"lo": 213,
|
| 361 |
+
"ular": 214,
|
| 362 |
+
"cum": 215,
|
| 363 |
+
"u": 216,
|
| 364 |
+
"arbam": 217,
|
| 365 |
+
"pr": 218,
|
| 366 |
+
"oso": 219,
|
| 367 |
+
"yne": 220,
|
| 368 |
+
"phth": 221,
|
| 369 |
+
"iridium": 222,
|
| 370 |
+
"hyd": 223,
|
| 371 |
+
"potassium": 224,
|
| 372 |
+
"succ": 225,
|
| 373 |
+
"platinum": 226,
|
| 374 |
+
"tetrac": 227,
|
| 375 |
+
"rg": 228,
|
| 376 |
+
"glyc": 229,
|
| 377 |
+
"anid": 230,
|
| 378 |
+
"capr": 231,
|
| 379 |
+
"chrys": 232,
|
| 380 |
+
"ratr": 233,
|
| 381 |
+
"yttrium": 234,
|
| 382 |
+
"pteridin": 235,
|
| 383 |
+
"ochrom": 236,
|
| 384 |
+
"laur": 237,
|
| 385 |
+
"ac": 238,
|
| 386 |
+
"pic": 239,
|
| 387 |
+
"ocyclo": 240,
|
| 388 |
+
"sel": 241,
|
| 389 |
+
"anthridin": 242,
|
| 390 |
+
"styril": 243,
|
| 391 |
+
"lithium": 244,
|
| 392 |
+
"formic": 245,
|
| 393 |
+
"formamido": 246,
|
| 394 |
+
"acetaldehyde": 247,
|
| 395 |
+
"trit": 248,
|
| 396 |
+
"sit": 249,
|
| 397 |
+
"ar": 250,
|
| 398 |
+
"ch": 251,
|
| 399 |
+
"anuide": 252,
|
| 400 |
+
"ite": 253,
|
| 401 |
+
"zirc": 254,
|
| 402 |
+
"'a": 255,
|
| 403 |
+
"glutar": 256,
|
| 404 |
+
"rm": 257,
|
| 405 |
+
"zinc": 258,
|
| 406 |
+
"cinn": 259,
|
| 407 |
+
"stann": 260,
|
| 408 |
+
"triacet": 261,
|
| 409 |
+
"annul": 262,
|
| 410 |
+
"inic": 263,
|
| 411 |
+
"titanium": 264,
|
| 412 |
+
"ul": 265,
|
| 413 |
+
"im": 266,
|
| 414 |
+
"us": 267,
|
| 415 |
+
"hypo": 268,
|
| 416 |
+
"porphyrin": 269,
|
| 417 |
+
"anth": 270,
|
| 418 |
+
"ad": 271,
|
| 419 |
+
"ucl": 272,
|
| 420 |
+
"xam": 273,
|
| 421 |
+
"piv": 274,
|
| 422 |
+
"nam": 275,
|
| 423 |
+
"cop": 276,
|
| 424 |
+
"de": 277,
|
| 425 |
+
"magnesium": 278,
|
| 426 |
+
"umaril": 279,
|
| 427 |
+
"fl": 280,
|
| 428 |
+
"ycl": 281,
|
| 429 |
+
"styr": 282,
|
| 430 |
+
"tungst": 283,
|
| 431 |
+
"tritio": 284,
|
| 432 |
+
"palladium": 285,
|
| 433 |
+
"anthro": 286,
|
| 434 |
+
"indac": 287,
|
| 435 |
+
"nickel": 288,
|
| 436 |
+
"aphth": 289,
|
| 437 |
+
"vanadium": 290,
|
| 438 |
+
"germ": 291,
|
| 439 |
+
"th": 292,
|
| 440 |
+
"cobalt": 293,
|
| 441 |
+
"myrist": 294,
|
| 442 |
+
"hr": 295,
|
| 443 |
+
"anida": 296,
|
| 444 |
+
"ruth": 297,
|
| 445 |
+
"selan": 298,
|
| 446 |
+
"bar": 299,
|
| 447 |
+
"tur": 300,
|
| 448 |
+
"inin": 301,
|
| 449 |
+
"arg": 302,
|
| 450 |
+
"iz": 303,
|
| 451 |
+
"cd": 304,
|
| 452 |
+
"xim": 305,
|
| 453 |
+
"calcium": 306,
|
| 454 |
+
"inato": 307,
|
| 455 |
+
"chl": 308,
|
| 456 |
+
"hyp": 309,
|
| 457 |
+
"formaldehyde": 310,
|
| 458 |
+
"rph": 311,
|
| 459 |
+
"ni": 312,
|
| 460 |
+
"mit": 313,
|
| 461 |
+
"anuida": 314,
|
| 462 |
+
"br": 315,
|
| 463 |
+
"phthalic": 316,
|
| 464 |
+
"iodanuid": 317,
|
| 465 |
+
"ous": 318,
|
| 466 |
+
"uracil": 319,
|
| 467 |
+
"oroso": 320,
|
| 468 |
+
"inan": 321,
|
| 469 |
+
"arach": 322,
|
| 470 |
+
"fg": 323,
|
| 471 |
+
"mang": 324,
|
| 472 |
+
"se": 325,
|
| 473 |
+
"F": 326,
|
| 474 |
+
"chromium": 327,
|
| 475 |
+
"um": 328,
|
| 476 |
+
"il": 329,
|
| 477 |
+
"rc": 330,
|
| 478 |
+
"amidic": 331,
|
| 479 |
+
"uranium": 332,
|
| 480 |
+
"alum": 333,
|
| 481 |
+
"yg": 334,
|
| 482 |
+
"ulf": 335,
|
| 483 |
+
"hafnium": 336,
|
| 484 |
+
"ver": 337,
|
| 485 |
+
"rp": 338,
|
| 486 |
+
"ars": 339,
|
| 487 |
+
"odium": 340,
|
| 488 |
+
"inamo": 341,
|
| 489 |
+
"tra": 342,
|
| 490 |
+
"da": 343,
|
| 491 |
+
"mercury": 344,
|
| 492 |
+
"iodite": 345,
|
| 493 |
+
"ybd": 346,
|
| 494 |
+
"rhodium": 347,
|
| 495 |
+
"cupr": 348,
|
| 496 |
+
"lign": 349,
|
| 497 |
+
"iodine": 350,
|
| 498 |
+
"aluman": 351,
|
| 499 |
+
"actinium": 352,
|
| 500 |
+
"rcar": 353,
|
| 501 |
+
"thraqu": 354,
|
| 502 |
+
"odi": 355,
|
| 503 |
+
"hth": 356,
|
| 504 |
+
"io": 357,
|
| 505 |
+
"aluminum": 358,
|
| 506 |
+
"thianthr": 359,
|
| 507 |
+
"ill": 360,
|
| 508 |
+
"barium": 361,
|
| 509 |
+
"iod": 362,
|
| 510 |
+
"isa": 363,
|
| 511 |
+
"rubidium": 364,
|
| 512 |
+
"cesium": 365,
|
| 513 |
+
"ut": 366,
|
| 514 |
+
"osmium": 367,
|
| 515 |
+
"rr": 368,
|
| 516 |
+
"terephthalic": 369,
|
| 517 |
+
"fumar": 370,
|
| 518 |
+
"rh": 371,
|
| 519 |
+
"'b": 372,
|
| 520 |
+
"apht": 373,
|
| 521 |
+
"telluro": 374,
|
| 522 |
+
"lut": 375,
|
| 523 |
+
"inous": 376,
|
| 524 |
+
"tellan": 377,
|
| 525 |
+
"iodous": 378,
|
| 526 |
+
"pim": 379,
|
| 527 |
+
"liz": 380,
|
| 528 |
+
"j": 381,
|
| 529 |
+
"gad": 382,
|
| 530 |
+
"inium": 383,
|
| 531 |
+
"at": 384,
|
| 532 |
+
"I": 385,
|
| 533 |
+
"cadmium": 386,
|
| 534 |
+
"bismuth": 387,
|
| 535 |
+
"asar": 388,
|
| 536 |
+
"ph": 389,
|
| 537 |
+
"arb": 390,
|
| 538 |
+
"tetramin": 391,
|
| 539 |
+
"corrin": 392,
|
| 540 |
+
"bac": 393,
|
| 541 |
+
"rutherfordium": 394,
|
| 542 |
+
"cerium": 395,
|
| 543 |
+
"sub": 396,
|
| 544 |
+
"cy": 397,
|
| 545 |
+
"itr": 398,
|
| 546 |
+
"hecta": 399,
|
| 547 |
+
"sar": 400,
|
| 548 |
+
"tyr": 401,
|
| 549 |
+
"cuban": 402,
|
| 550 |
+
"stiba": 403,
|
| 551 |
+
"europium": 404,
|
| 552 |
+
"gly": 405,
|
| 553 |
+
"germa": 406,
|
| 554 |
+
"thallium": 407,
|
| 555 |
+
"technetium": 408,
|
| 556 |
+
"lanthanum": 409,
|
| 557 |
+
"la": 410,
|
| 558 |
+
"na": 411,
|
| 559 |
+
"tellur": 412,
|
| 560 |
+
"lact": 413,
|
| 561 |
+
"thial": 414,
|
| 562 |
+
"str": 415,
|
| 563 |
+
"arsan": 416,
|
| 564 |
+
"chu": 417,
|
| 565 |
+
"antim": 418,
|
| 566 |
+
"oform": 419,
|
| 567 |
+
"itril": 420,
|
| 568 |
+
"sti": 421,
|
| 569 |
+
"neodymium": 422,
|
| 570 |
+
"tantalum": 423,
|
| 571 |
+
"plumb": 424,
|
| 572 |
+
"yohimban": 425,
|
| 573 |
+
"cym": 426,
|
| 574 |
+
"niobium": 427,
|
| 575 |
+
"ld": 428,
|
| 576 |
+
"alo": 429,
|
| 577 |
+
"praseodymium": 430,
|
| 578 |
+
"ocinn": 431,
|
| 579 |
+
"hy": 432,
|
| 580 |
+
"lead": 433,
|
| 581 |
+
"inite": 434,
|
| 582 |
+
"liss": 435,
|
| 583 |
+
"cr": 436,
|
| 584 |
+
"ci": 437,
|
| 585 |
+
"rth": 438,
|
| 586 |
+
"ytterbium": 439,
|
| 587 |
+
"ra": 440,
|
| 588 |
+
"\"": 441,
|
| 589 |
+
"z": 442,
|
| 590 |
+
"ps": 443,
|
| 591 |
+
"samarium": 444,
|
| 592 |
+
"naph": 445,
|
| 593 |
+
"terbium": 446,
|
| 594 |
+
"othialdehyde": 447,
|
| 595 |
+
"scandium": 448,
|
| 596 |
+
"tartar": 449,
|
| 597 |
+
"pyruv": 450,
|
| 598 |
+
"as": 451,
|
| 599 |
+
"ber": 452,
|
| 600 |
+
"lium": 453,
|
| 601 |
+
"ll": 454,
|
| 602 |
+
"nu": 455,
|
| 603 |
+
"za": 456,
|
| 604 |
+
"phthalaldehyde": 457,
|
| 605 |
+
"llit": 458,
|
| 606 |
+
"tassi": 459,
|
| 607 |
+
"dysprosium": 460,
|
| 608 |
+
"epin": 461,
|
| 609 |
+
"lutetium": 462,
|
| 610 |
+
"ct": 463,
|
| 611 |
+
"acrido": 464,
|
| 612 |
+
"ij": 465,
|
| 613 |
+
"nthr": 466,
|
| 614 |
+
"indium": 467,
|
| 615 |
+
"fermium": 468,
|
| 616 |
+
"pha": 469,
|
| 617 |
+
"thra": 470,
|
| 618 |
+
"gallium": 471,
|
| 619 |
+
"erbium": 472,
|
| 620 |
+
"gamma": 473,
|
| 621 |
+
"flu": 474,
|
| 622 |
+
"sam": 475,
|
| 623 |
+
"hl": 476,
|
| 624 |
+
"thorium": 477,
|
| 625 |
+
"ma": 478,
|
| 626 |
+
"terephthal": 479,
|
| 627 |
+
"gallan": 480,
|
| 628 |
+
"lum": 481,
|
| 629 |
+
"mium": 482,
|
| 630 |
+
"lacc": 483,
|
| 631 |
+
"rs": 484,
|
| 632 |
+
"cor": 485,
|
| 633 |
+
"gall": 486,
|
| 634 |
+
"hippur": 487,
|
| 635 |
+
"taur": 488,
|
| 636 |
+
"caff": 489,
|
| 637 |
+
"nia": 490,
|
| 638 |
+
"ip": 491,
|
| 639 |
+
"pia": 492,
|
| 640 |
+
"dur": 493,
|
| 641 |
+
"glutam": 494,
|
| 642 |
+
"fuller": 495,
|
| 643 |
+
"atr": 496,
|
| 644 |
+
"plast": 497,
|
| 645 |
+
"terephthalaldehyde": 498,
|
| 646 |
+
"gluc": 499,
|
| 647 |
+
"phl": 500,
|
| 648 |
+
"thulium": 501,
|
| 649 |
+
"apr": 502,
|
| 650 |
+
"mu": 503,
|
| 651 |
+
"iran": 504,
|
| 652 |
+
"jap": 505,
|
| 653 |
+
"Tc": 506,
|
| 654 |
+
"stib": 507,
|
| 655 |
+
"citra": 508,
|
| 656 |
+
"astatine": 509,
|
| 657 |
+
"rcuri": 510,
|
| 658 |
+
"it": 511,
|
| 659 |
+
"iodido": 512,
|
| 660 |
+
"brass": 513,
|
| 661 |
+
"rcur": 514,
|
| 662 |
+
"rchl": 515,
|
| 663 |
+
"idic": 516,
|
| 664 |
+
"citr": 517,
|
| 665 |
+
"Br": 518,
|
| 666 |
+
"aceanthr": 519,
|
| 667 |
+
"rit": 520,
|
| 668 |
+
"mim": 521,
|
| 669 |
+
"rul": 522,
|
| 670 |
+
"tit": 523,
|
| 671 |
+
"pyranthr": 524,
|
| 672 |
+
"Ih": 525,
|
| 673 |
+
"protio": 526,
|
| 674 |
+
"stilb": 527,
|
| 675 |
+
"ypr": 528,
|
| 676 |
+
"dicta": 529,
|
| 677 |
+
"curium": 530,
|
| 678 |
+
"sg": 531,
|
| 679 |
+
"amin": 532,
|
| 680 |
+
"mi": 533,
|
| 681 |
+
"cta": 534,
|
| 682 |
+
":": 535,
|
| 683 |
+
"ila": 536,
|
| 684 |
+
"mm": 537,
|
| 685 |
+
"neptunium": 538,
|
| 686 |
+
"ulfit": 539,
|
| 687 |
+
"aur": 540,
|
| 688 |
+
"cac": 541,
|
| 689 |
+
"orus": 542,
|
| 690 |
+
"pp": 543,
|
| 691 |
+
"cium": 544,
|
| 692 |
+
"lawr": 545,
|
| 693 |
+
"dd": 546,
|
| 694 |
+
"ri": 547,
|
| 695 |
+
"skat": 548,
|
| 696 |
+
"thaps": 549,
|
| 697 |
+
"hygr": 550,
|
| 698 |
+
"parab": 551,
|
| 699 |
+
"tartr": 552,
|
| 700 |
+
"yv": 553,
|
| 701 |
+
"germanium": 554,
|
| 702 |
+
"plut": 555,
|
| 703 |
+
"ita": 556,
|
| 704 |
+
"wn": 557,
|
| 705 |
+
"rphyr": 558,
|
| 706 |
+
"ace": 559,
|
| 707 |
+
"thym": 560,
|
| 708 |
+
"apth": 561,
|
| 709 |
+
"At": 562,
|
| 710 |
+
"buckm": 563,
|
| 711 |
+
"einsteinium": 564,
|
| 712 |
+
"americium": 565,
|
| 713 |
+
"pro": 566,
|
| 714 |
+
"nobelium": 567,
|
| 715 |
+
"T": 568,
|
| 716 |
+
"xysucc": 569,
|
| 717 |
+
"plumba": 570,
|
| 718 |
+
"helium": 571,
|
| 719 |
+
"w": 572,
|
| 720 |
+
"P": 573,
|
| 721 |
+
"mad": 574,
|
| 722 |
+
"guaiac": 575,
|
| 723 |
+
"gu": 576,
|
| 724 |
+
"cadav": 577,
|
| 725 |
+
"californium": 578,
|
| 726 |
+
"tr": 579,
|
| 727 |
+
"hnit": 580,
|
| 728 |
+
"ilv": 581,
|
| 729 |
+
"pla": 582,
|
| 730 |
+
"ti": 583,
|
| 731 |
+
"thyr": 584,
|
| 732 |
+
"ru": 585,
|
| 733 |
+
"sc": 586,
|
| 734 |
+
"B": 587,
|
| 735 |
+
"atido": 588,
|
| 736 |
+
"protactinium": 589,
|
| 737 |
+
"xyv": 590,
|
| 738 |
+
"lup": 591,
|
| 739 |
+
"pleiad": 592,
|
| 740 |
+
"aluminan": 593,
|
| 741 |
+
"ulfat": 594,
|
| 742 |
+
"arsind": 595,
|
| 743 |
+
"lmn": 596,
|
| 744 |
+
"gium": 597,
|
| 745 |
+
"rcura": 598,
|
| 746 |
+
"sea": 599,
|
| 747 |
+
"syr": 600,
|
| 748 |
+
"li": 601,
|
| 749 |
+
"camph": 602,
|
| 750 |
+
"ychr": 603,
|
| 751 |
+
"hect": 604,
|
| 752 |
+
"flav": 605,
|
| 753 |
+
"ybda": 606,
|
| 754 |
+
"yp": 607,
|
| 755 |
+
"rsa": 608,
|
| 756 |
+
"ug": 609,
|
| 757 |
+
"dubnium": 610,
|
| 758 |
+
"oval": 611,
|
| 759 |
+
"tis": 612,
|
| 760 |
+
"yc": 613,
|
| 761 |
+
"krypt": 614,
|
| 762 |
+
"vi": 615,
|
| 763 |
+
"berkelium": 616,
|
| 764 |
+
"ign": 617,
|
| 765 |
+
"fulm": 618,
|
| 766 |
+
"xyst": 619,
|
| 767 |
+
"tim": 620,
|
| 768 |
+
"meitnerium": 621,
|
| 769 |
+
"tellurium": 622,
|
| 770 |
+
"nil": 623,
|
| 771 |
+
"carvacr": 624,
|
| 772 |
+
"D": 625,
|
| 773 |
+
"hafna": 626,
|
| 774 |
+
"ymyrist": 627,
|
| 775 |
+
"llur": 628,
|
| 776 |
+
"german": 629,
|
| 777 |
+
"prism": 630,
|
| 778 |
+
"rad": 631,
|
| 779 |
+
"bohrium": 632,
|
| 780 |
+
"rb": 633,
|
| 781 |
+
"hassium": 634,
|
| 782 |
+
"astat": 635,
|
| 783 |
+
"delevium": 636,
|
| 784 |
+
"symcl": 637,
|
| 785 |
+
"umb": 638,
|
| 786 |
+
"rach": 639,
|
| 787 |
+
"llif": 640,
|
| 788 |
+
"xycapr": 641,
|
| 789 |
+
"xyp": 642,
|
| 790 |
+
"ulfam": 643,
|
| 791 |
+
"lat": 644,
|
| 792 |
+
"gsta": 645,
|
| 793 |
+
"hym": 646,
|
| 794 |
+
"nti": 647,
|
| 795 |
+
"yqu": 648,
|
| 796 |
+
"sa": 649,
|
| 797 |
+
"plum": 650,
|
| 798 |
+
"xa": 651,
|
| 799 |
+
"astata": 652,
|
| 800 |
+
"xyacr": 653,
|
| 801 |
+
"xym": 654,
|
| 802 |
+
"ada": 655,
|
| 803 |
+
"adm": 656,
|
| 804 |
+
"fulv": 657,
|
| 805 |
+
"ba": 658,
|
| 806 |
+
"warfar": 659,
|
| 807 |
+
"ycapr": 660,
|
| 808 |
+
"rri": 661,
|
| 809 |
+
"tipyr": 662,
|
| 810 |
+
"ypip": 663,
|
| 811 |
+
"hrys": 664,
|
| 812 |
+
"ast": 665,
|
| 813 |
+
"xyb": 666,
|
| 814 |
+
"xyi": 667,
|
| 815 |
+
"ruc": 668,
|
| 816 |
+
"alumin": 669,
|
| 817 |
+
"xymyrist": 670,
|
| 818 |
+
"rut": 671,
|
| 819 |
+
"jugl": 672,
|
| 820 |
+
"ysucc": 673,
|
| 821 |
+
"aci": 674,
|
| 822 |
+
"radium": 675,
|
| 823 |
+
"xypr": 676,
|
| 824 |
+
"yacr": 677,
|
| 825 |
+
"xychr": 678,
|
| 826 |
+
"uvit": 679,
|
| 827 |
+
"'c": 680,
|
| 828 |
+
"ub": 681,
|
| 829 |
+
"scul": 682,
|
| 830 |
+
"xyarach": 683,
|
| 831 |
+
"yt": 684,
|
| 832 |
+
"safr": 685,
|
| 833 |
+
"dymi": 686,
|
| 834 |
+
"protide": 687,
|
| 835 |
+
"rbr": 688,
|
| 836 |
+
"sma": 689,
|
| 837 |
+
"xychl": 690,
|
| 838 |
+
"fusar": 691,
|
| 839 |
+
"ga": 692,
|
| 840 |
+
"As": 693,
|
| 841 |
+
"In": 694,
|
| 842 |
+
"pras": 695,
|
| 843 |
+
"xyqu": 696,
|
| 844 |
+
"abas": 697,
|
| 845 |
+
"acc": 698,
|
| 846 |
+
"yphth": 699,
|
| 847 |
+
"xyc": 700,
|
| 848 |
+
"ascl": 701,
|
| 849 |
+
"gst": 702,
|
| 850 |
+
"plumbag": 703,
|
| 851 |
+
"yfur": 704,
|
| 852 |
+
"xyglutar": 705,
|
| 853 |
+
"Ga": 706,
|
| 854 |
+
"asti": 707,
|
| 855 |
+
"sux": 708,
|
| 856 |
+
"xya": 709,
|
| 857 |
+
"yb": 710,
|
| 858 |
+
"ata": 711,
|
| 859 |
+
"parax": 712,
|
| 860 |
+
"G": 713,
|
| 861 |
+
"Li": 714,
|
| 862 |
+
"francium": 715,
|
| 863 |
+
"rsaplumb": 716,
|
| 864 |
+
"rv": 717,
|
| 865 |
+
"daphn": 718,
|
| 866 |
+
"ag": 719,
|
| 867 |
+
"guai": 720,
|
| 868 |
+
"ychl": 721,
|
| 869 |
+
"yf": 722,
|
| 870 |
+
"ym": 723,
|
| 871 |
+
"samari": 724,
|
| 872 |
+
"si": 725,
|
| 873 |
+
"ucc": 726,
|
| 874 |
+
"full": 727,
|
| 875 |
+
"ib": 728,
|
| 876 |
+
"nthrac": 729,
|
| 877 |
+
"sap": 730,
|
| 878 |
+
"ypyr": 731,
|
| 879 |
+
"hn": 732,
|
| 880 |
+
"sm": 733,
|
| 881 |
+
"ta": 734,
|
| 882 |
+
"ud": 735,
|
| 883 |
+
"llura": 736,
|
| 884 |
+
"yglutar": 737,
|
| 885 |
+
"yst": 738,
|
| 886 |
+
"btu": 739,
|
| 887 |
+
"M": 740,
|
| 888 |
+
"U": 741,
|
| 889 |
+
"dyma": 742,
|
| 890 |
+
"lad": 743,
|
| 891 |
+
"magn": 744,
|
| 892 |
+
"vacc": 745,
|
| 893 |
+
"ya": 746,
|
| 894 |
+
"ntim": 747,
|
| 895 |
+
"rsast": 748,
|
| 896 |
+
"ap": 749,
|
| 897 |
+
"par": 750,
|
| 898 |
+
"xypip": 751,
|
| 899 |
+
"xysti": 752,
|
| 900 |
+
"yamm": 753,
|
| 901 |
+
"Pt": 754,
|
| 902 |
+
"Y": 755,
|
| 903 |
+
"aplumb": 756,
|
| 904 |
+
"auri": 757,
|
| 905 |
+
"bara": 758,
|
| 906 |
+
"dyspr": 759,
|
| 907 |
+
"frax": 760,
|
| 908 |
+
"lada": 761,
|
| 909 |
+
"purpur": 762,
|
| 910 |
+
"sp": 763,
|
| 911 |
+
"xysub": 764,
|
| 912 |
+
"yisa": 765,
|
| 913 |
+
"xyac": 766,
|
| 914 |
+
"xyphth": 767,
|
| 915 |
+
"yx": 768,
|
| 916 |
+
"astib": 769,
|
| 917 |
+
"chav": 770,
|
| 918 |
+
"fursulti": 771,
|
| 919 |
+
"hist": 772,
|
| 920 |
+
"mast": 773,
|
| 921 |
+
"rfull": 774,
|
| 922 |
+
"uramil": 775,
|
| 923 |
+
"usn": 776,
|
| 924 |
+
"xypim": 777,
|
| 925 |
+
"xyth": 778,
|
| 926 |
+
"yhyd": 779,
|
| 927 |
+
"Bi": 780,
|
| 928 |
+
"Bk": 781,
|
| 929 |
+
"Tl": 782,
|
| 930 |
+
"thuli": 783,
|
| 931 |
+
"tiba": 784,
|
| 932 |
+
"xybr": 785,
|
| 933 |
+
"xyim": 786,
|
| 934 |
+
"ypim": 787,
|
| 935 |
+
"ys": 788,
|
| 936 |
+
"ytt": 789,
|
| 937 |
+
"cyt": 790,
|
| 938 |
+
"dr": 791,
|
| 939 |
+
"ilast": 792,
|
| 940 |
+
"nag": 793,
|
| 941 |
+
"xypyr": 794,
|
| 942 |
+
"xys": 795,
|
| 943 |
+
"yarach": 796,
|
| 944 |
+
"ybar": 797,
|
| 945 |
+
"yfurfur": 798,
|
| 946 |
+
"ysub": 799,
|
| 947 |
+
"yuracil": 800,
|
| 948 |
+
"aph": 801,
|
| 949 |
+
"cad": 802,
|
| 950 |
+
"cham": 803,
|
| 951 |
+
"dym": 804,
|
| 952 |
+
"laws": 805,
|
| 953 |
+
"phys": 806,
|
| 954 |
+
"rsam": 807,
|
| 955 |
+
"stat": 808,
|
| 956 |
+
"tsuzu": 809,
|
| 957 |
+
"xymu": 810,
|
| 958 |
+
"xystib": 811,
|
| 959 |
+
"yac": 812,
|
| 960 |
+
"yh": 813,
|
| 961 |
+
"yhyp": 814,
|
| 962 |
+
"yi": 815,
|
| 963 |
+
"yim": 816,
|
| 964 |
+
"yphl": 817,
|
| 965 |
+
"ypi": 818,
|
| 966 |
+
"yru": 819,
|
| 967 |
+
"ystib": 820,
|
| 968 |
+
"yur": 821
|
| 969 |
+
},
|
| 970 |
+
"unk_token": "<unk>"
|
| 971 |
+
}
|
| 972 |
+
}
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,54 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<pad>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<eos>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "<bos>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"4": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<pad>",
|
| 45 |
+
"clean_up_tokenization_spaces": true,
|
| 46 |
+
"eos_token": "<eos>",
|
| 47 |
+
"model_max_length": 512,
|
| 48 |
+
"pad_token": "<bos>",
|
| 49 |
+
"padding_side": "right",
|
| 50 |
+
"tokenizer_class": "PreTrainedTokenizerFast",
|
| 51 |
+
"truncation_side": "right",
|
| 52 |
+
"unk_token": "<unk>",
|
| 53 |
+
"mask_token": "<mask>"
|
| 54 |
+
}
|