Upload tokenizer

added_tokens.json

@@ -1,110 +1,8 @@
 {
-  ".": 5,
-  "</s>": 1,
-  "<bom>": 35071,
-  "<bop>": 35066,
-  "<eom>": 35067,
-  "<eop>": 35063,
-  "<extra_id_0>": 32099,
-  "<extra_id_10>": 32089,
-  "<extra_id_11>": 32088,
-  "<extra_id_12>": 32087,
-  "<extra_id_13>": 32086,
-  "<extra_id_14>": 32085,
-  "<extra_id_15>": 32084,
-  "<extra_id_16>": 32083,
-  "<extra_id_17>": 32082,
-  "<extra_id_18>": 32081,
-  "<extra_id_19>": 32080,
-  "<extra_id_1>": 32098,
-  "<extra_id_20>": 32079,
-  "<extra_id_21>": 32078,
-  "<extra_id_22>": 32077,
-  "<extra_id_23>": 32076,
-  "<extra_id_24>": 32075,
-  "<extra_id_25>": 32074,
-  "<extra_id_26>": 32073,
-  "<extra_id_27>": 32072,
-  "<extra_id_28>": 32071,
-  "<extra_id_29>": 32070,
-  "<extra_id_2>": 32097,
-  "<extra_id_30>": 32069,
-  "<extra_id_31>": 32068,
-  "<extra_id_32>": 32067,
-  "<extra_id_33>": 32066,
-  "<extra_id_34>": 32065,
-  "<extra_id_35>": 32064,
-  "<extra_id_36>": 32063,
-  "<extra_id_37>": 32062,
-  "<extra_id_38>": 32061,
-  "<extra_id_39>": 32060,
-  "<extra_id_3>": 32096,
-  "<extra_id_40>": 32059,
-  "<extra_id_41>": 32058,
-  "<extra_id_42>": 32057,
-  "<extra_id_43>": 32056,
-  "<extra_id_44>": 32055,
-  "<extra_id_45>": 32054,
-  "<extra_id_46>": 32053,
-  "<extra_id_47>": 32052,
-  "<extra_id_48>": 32051,
-  "<extra_id_49>": 32050,
-  "<extra_id_4>": 32095,
-  "<extra_id_50>": 32049,
-  "<extra_id_51>": 32048,
-  "<extra_id_52>": 32047,
-  "<extra_id_53>": 32046,
-  "<extra_id_54>": 32045,
-  "<extra_id_55>": 32044,
-  "<extra_id_56>": 32043,
-  "<extra_id_57>": 32042,
-  "<extra_id_58>": 32041,
-  "<extra_id_59>": 32040,
-  "<extra_id_5>": 32094,
-  "<extra_id_60>": 32039,
-  "<extra_id_61>": 32038,
-  "<extra_id_62>": 32037,
-  "<extra_id_63>": 32036,
-  "<extra_id_64>": 32035,
-  "<extra_id_65>": 32034,
-  "<extra_id_66>": 32033,
-  "<extra_id_67>": 32032,
-  "<extra_id_68>": 32031,
-  "<extra_id_69>": 32030,
-  "<extra_id_6>": 32093,
-  "<extra_id_70>": 32029,
-  "<extra_id_71>": 32028,
-  "<extra_id_72>": 32027,
-  "<extra_id_73>": 32026,
-  "<extra_id_74>": 32025,
-  "<extra_id_75>": 32024,
-  "<extra_id_76>": 32023,
-  "<extra_id_77>": 32022,
-  "<extra_id_78>": 32021,
-  "<extra_id_79>": 32020,
-  "<extra_id_7>": 32092,
-  "<extra_id_80>": 32019,
-  "<extra_id_81>": 32018,
-  "<extra_id_82>": 32017,
-  "<extra_id_83>": 32016,
-  "<extra_id_84>": 32015,
-  "<extra_id_85>": 32014,
-  "<extra_id_86>": 32013,
-  "<extra_id_87>": 32012,
-  "<extra_id_88>": 32011,
-  "<extra_id_89>": 32010,
-  "<extra_id_8>": 32091,
-  "<extra_id_90>": 32009,
-  "<extra_id_91>": 32008,
-  "<extra_id_92>": 32007,
-  "<extra_id_93>": 32006,
-  "<extra_id_94>": 32005,
-  "<extra_id_95>": 32004,
-  "<extra_id_96>": 32003,
-  "<extra_id_97>": 32002,
-  "<extra_id_98>": 32001,
-  "<extra_id_99>": 32000,
-  "<extra_id_9>": 32090,
+  "<bom>": 35063,
+  "<bop>": 35065,
+  "<eom>": 35064,
+  "<eop>": 35066,
   "<p>A": 32100,
   "<p>C": 32101,
   "<p>D": 32102,
@@ -125,14 +23,12 @@
   "<p>V": 32117,
   "<p>W": 32118,
   "<p>Y": 32119,
-  "<pad>": 0,
-  "<unk>": 2,
   "DESCRIPTION": 35068,
   "FUNCTION": 35070,
-  "MOLECULE NAME": 35072,
-  "PROTEIN FAMILIES": 35064,
+  "MOLECULE NAME": 35067,
+  "PROTEIN FAMILIES": 35072,
   "PROTEIN NAME": 35069,
-  "SUBCELLULAR LOCATION": 35065,
+  "SUBCELLULAR LOCATION": 35071,
   "[#11C-1]": 34809,
   "[#11CH1]": 34772,
   "[#11C]": 32177,

special_tokens_map.json

@@ -100,16 +100,16 @@
     "<extra_id_97>",
     "<extra_id_98>",
     "<extra_id_99>",
-    "<eop>",
-    "PROTEIN FAMILIES",
-    "SUBCELLULAR LOCATION",
-    "<bop>",
+    "<bom>",
     "<eom>",
+    "<bop>",
+    "<eop>",
+    "MOLECULE NAME",
     "DESCRIPTION",
     "PROTEIN NAME",
     "FUNCTION",
-    "<bom>",
-    "MOLECULE NAME"
+    "SUBCELLULAR LOCATION",
+    "PROTEIN FAMILIES"
   ],
   "eos_token": "</s>",
   "pad_token": "<pad>",

tokenizer.json

@@ -7,27 +7,27 @@
       "id": 0,
       "content": "<pad>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
-      "normalized": true,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
       "special": true
     },
     {
       "id": 1,
       "content": "</s>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
-      "normalized": true,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
       "special": true
     },
     {
       "id": 2,
       "content": "<unk>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
-      "normalized": true,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
       "special": true
     },
     {
@@ -43,8 +43,8 @@
       "id": 32000,
       "content": "<extra_id_99>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -52,8 +52,8 @@
       "id": 32001,
       "content": "<extra_id_98>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -61,8 +61,8 @@
       "id": 32002,
       "content": "<extra_id_97>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -70,8 +70,8 @@
       "id": 32003,
       "content": "<extra_id_96>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -79,8 +79,8 @@
       "id": 32004,
       "content": "<extra_id_95>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -88,8 +88,8 @@
       "id": 32005,
       "content": "<extra_id_94>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -97,8 +97,8 @@
       "id": 32006,
       "content": "<extra_id_93>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -106,8 +106,8 @@
       "id": 32007,
       "content": "<extra_id_92>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -115,8 +115,8 @@
       "id": 32008,
       "content": "<extra_id_91>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -124,8 +124,8 @@
       "id": 32009,
       "content": "<extra_id_90>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -133,8 +133,8 @@
       "id": 32010,
       "content": "<extra_id_89>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -142,8 +142,8 @@
       "id": 32011,
       "content": "<extra_id_88>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -151,8 +151,8 @@
       "id": 32012,
       "content": "<extra_id_87>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -160,8 +160,8 @@
       "id": 32013,
       "content": "<extra_id_86>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -169,8 +169,8 @@
       "id": 32014,
       "content": "<extra_id_85>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -178,8 +178,8 @@
       "id": 32015,
       "content": "<extra_id_84>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -187,8 +187,8 @@
       "id": 32016,
       "content": "<extra_id_83>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -196,8 +196,8 @@
       "id": 32017,
       "content": "<extra_id_82>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -205,8 +205,8 @@
       "id": 32018,
       "content": "<extra_id_81>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -214,8 +214,8 @@
       "id": 32019,
       "content": "<extra_id_80>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -223,8 +223,8 @@
       "id": 32020,
       "content": "<extra_id_79>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -232,8 +232,8 @@
       "id": 32021,
       "content": "<extra_id_78>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -241,8 +241,8 @@
       "id": 32022,
       "content": "<extra_id_77>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -250,8 +250,8 @@
       "id": 32023,
       "content": "<extra_id_76>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -259,8 +259,8 @@
       "id": 32024,
       "content": "<extra_id_75>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -268,8 +268,8 @@
       "id": 32025,
       "content": "<extra_id_74>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -277,8 +277,8 @@
       "id": 32026,
       "content": "<extra_id_73>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -286,8 +286,8 @@
       "id": 32027,
       "content": "<extra_id_72>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -295,8 +295,8 @@
       "id": 32028,
       "content": "<extra_id_71>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -304,8 +304,8 @@
       "id": 32029,
       "content": "<extra_id_70>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -313,8 +313,8 @@
       "id": 32030,
       "content": "<extra_id_69>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -322,8 +322,8 @@
       "id": 32031,
       "content": "<extra_id_68>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -331,8 +331,8 @@
       "id": 32032,
       "content": "<extra_id_67>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -340,8 +340,8 @@
       "id": 32033,
       "content": "<extra_id_66>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -349,8 +349,8 @@
       "id": 32034,
       "content": "<extra_id_65>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -358,8 +358,8 @@
       "id": 32035,
       "content": "<extra_id_64>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -367,8 +367,8 @@
       "id": 32036,
       "content": "<extra_id_63>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -376,8 +376,8 @@
       "id": 32037,
       "content": "<extra_id_62>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -385,8 +385,8 @@
       "id": 32038,
       "content": "<extra_id_61>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -394,8 +394,8 @@
       "id": 32039,
       "content": "<extra_id_60>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -403,8 +403,8 @@
       "id": 32040,
       "content": "<extra_id_59>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -412,8 +412,8 @@
       "id": 32041,
       "content": "<extra_id_58>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -421,8 +421,8 @@
       "id": 32042,
       "content": "<extra_id_57>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -430,8 +430,8 @@
       "id": 32043,
       "content": "<extra_id_56>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -439,8 +439,8 @@
       "id": 32044,
       "content": "<extra_id_55>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -448,8 +448,8 @@
       "id": 32045,
       "content": "<extra_id_54>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -457,8 +457,8 @@
       "id": 32046,
       "content": "<extra_id_53>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -466,8 +466,8 @@
       "id": 32047,
       "content": "<extra_id_52>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -475,8 +475,8 @@
       "id": 32048,
       "content": "<extra_id_51>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -484,8 +484,8 @@
       "id": 32049,
       "content": "<extra_id_50>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -493,8 +493,8 @@
       "id": 32050,
       "content": "<extra_id_49>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -502,8 +502,8 @@
       "id": 32051,
       "content": "<extra_id_48>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -511,8 +511,8 @@
       "id": 32052,
       "content": "<extra_id_47>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -520,8 +520,8 @@
       "id": 32053,
       "content": "<extra_id_46>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -529,8 +529,8 @@
       "id": 32054,
       "content": "<extra_id_45>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -538,8 +538,8 @@
       "id": 32055,
       "content": "<extra_id_44>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -547,8 +547,8 @@
       "id": 32056,
       "content": "<extra_id_43>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -556,8 +556,8 @@
       "id": 32057,
       "content": "<extra_id_42>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -565,8 +565,8 @@
       "id": 32058,
       "content": "<extra_id_41>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -574,8 +574,8 @@
       "id": 32059,
       "content": "<extra_id_40>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -583,8 +583,8 @@
       "id": 32060,
       "content": "<extra_id_39>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -592,8 +592,8 @@
       "id": 32061,
       "content": "<extra_id_38>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -601,8 +601,8 @@
       "id": 32062,
       "content": "<extra_id_37>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -610,8 +610,8 @@
       "id": 32063,
       "content": "<extra_id_36>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -619,8 +619,8 @@
       "id": 32064,
       "content": "<extra_id_35>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -628,8 +628,8 @@
       "id": 32065,
       "content": "<extra_id_34>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -637,8 +637,8 @@
       "id": 32066,
       "content": "<extra_id_33>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -646,8 +646,8 @@
       "id": 32067,
       "content": "<extra_id_32>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -655,8 +655,8 @@
       "id": 32068,
       "content": "<extra_id_31>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -664,8 +664,8 @@
       "id": 32069,
       "content": "<extra_id_30>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -673,8 +673,8 @@
       "id": 32070,
       "content": "<extra_id_29>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -682,8 +682,8 @@
       "id": 32071,
       "content": "<extra_id_28>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -691,8 +691,8 @@
       "id": 32072,
       "content": "<extra_id_27>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -700,8 +700,8 @@
       "id": 32073,
       "content": "<extra_id_26>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -709,8 +709,8 @@
       "id": 32074,
       "content": "<extra_id_25>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -718,8 +718,8 @@
       "id": 32075,
       "content": "<extra_id_24>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -727,8 +727,8 @@
       "id": 32076,
       "content": "<extra_id_23>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -736,8 +736,8 @@
       "id": 32077,
       "content": "<extra_id_22>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -745,8 +745,8 @@
       "id": 32078,
       "content": "<extra_id_21>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -754,8 +754,8 @@
       "id": 32079,
       "content": "<extra_id_20>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -763,8 +763,8 @@
       "id": 32080,
       "content": "<extra_id_19>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -772,8 +772,8 @@
       "id": 32081,
       "content": "<extra_id_18>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -781,8 +781,8 @@
       "id": 32082,
       "content": "<extra_id_17>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -790,8 +790,8 @@
       "id": 32083,
       "content": "<extra_id_16>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -799,8 +799,8 @@
       "id": 32084,
       "content": "<extra_id_15>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -808,8 +808,8 @@
       "id": 32085,
       "content": "<extra_id_14>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -817,8 +817,8 @@
       "id": 32086,
       "content": "<extra_id_13>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -826,8 +826,8 @@
       "id": 32087,
       "content": "<extra_id_12>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -835,8 +835,8 @@
       "id": 32088,
       "content": "<extra_id_11>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -844,8 +844,8 @@
       "id": 32089,
       "content": "<extra_id_10>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -853,8 +853,8 @@
       "id": 32090,
       "content": "<extra_id_9>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -862,8 +862,8 @@
       "id": 32091,
       "content": "<extra_id_8>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -871,8 +871,8 @@
       "id": 32092,
       "content": "<extra_id_7>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -880,8 +880,8 @@
       "id": 32093,
       "content": "<extra_id_6>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -889,8 +889,8 @@
       "id": 32094,
       "content": "<extra_id_5>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -898,8 +898,8 @@
       "id": 32095,
       "content": "<extra_id_4>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -907,8 +907,8 @@
       "id": 32096,
       "content": "<extra_id_3>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -916,8 +916,8 @@
       "id": 32097,
       "content": "<extra_id_2>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -925,8 +925,8 @@
       "id": 32098,
       "content": "<extra_id_1>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -934,8 +934,8 @@
       "id": 32099,
       "content": "<extra_id_0>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -27608,46 +27608,46 @@
     },
     {
       "id": 35063,
-      "content": "<eop>",
+      "content": "<bom>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
     {
       "id": 35064,
-      "content": "PROTEIN FAMILIES",
+      "content": "<eom>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
     {
       "id": 35065,
-      "content": "SUBCELLULAR LOCATION",
+      "content": "<bop>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
     {
       "id": 35066,
-      "content": "<bop>",
+      "content": "<eop>",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
     {
       "id": 35067,
-      "content": "<eom>",
+      "content": "MOLECULE NAME",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -27655,8 +27655,8 @@
       "id": 35068,
       "content": "DESCRIPTION",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -27664,8 +27664,8 @@
       "id": 35069,
       "content": "PROTEIN NAME",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
@@ -27673,26 +27673,26 @@
       "id": 35070,
       "content": "FUNCTION",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
     {
       "id": 35071,
-      "content": "<bom>",
+      "content": "SUBCELLULAR LOCATION",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     },
     {
       "id": 35072,
-      "content": "MOLECULE NAME",
+      "content": "PROTEIN FAMILIES",
       "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
+      "lstrip": false,
+      "rstrip": false,
       "normalized": false,
       "special": true
     }
@@ -156181,7 +156181,6 @@
         "<extra_id_0>",
         0.0
       ]
-    ],
-    "byte_fallback": false
+    ]
   }
 }