Training done

added_tokens.json

@@ -0,0 +1,897 @@
+{
+  "</s_%_SOLUTION>": 57834,
+  "</s_+>": 57970,
+  "</s_1_WATER_or_BRINE>": 58052,
+  "</s_21_34>": 57766,
+  "</s_21_34_96>": 57764,
+  "</s_2_5%_HCL>": 58050,
+  "</s_2_AMINOANTHHACENE>": 58008,
+  "</s_35+>": 57762,
+  "</s_35+_109>": 57760,
+  "</s_3_5%_NaOH>": 58048,
+  "</s_4_ALCOHOLS>": 58046,
+  "</s_5_%_METHYL_CELULOSE>": 57828,
+  "</s_5_OXYGEN>": 58044,
+  "</s_6_LIGHT>": 58042,
+  "</s_9>": 57974,
+  "</s_>": 57550,
+  "</s_ACCT_NO>": 58288,
+  "</s_ACUTE_CADCOVASCULAR>": 58224,
+  "</s_ACUTE_CARDIOVASCULAR>": 58038,
+  "</s_ACUTE_TOXICITY_IN_MICE>": 57832,
+  "</s_ADDENDUM_I_DAWSON_RESEARCH_CORPORATION_Protocol_Change_Order_No_1>": 58146,
+  "</s_AMES_ASSAY_FOR_MUTAGENICITY>": 58006,
+  "</s_AMOUNT>": 58076,
+  "</s_AMT_MENTHOL_7=_More>": 57758,
+  "</s_ANALYTICAL_METHODS>": 58074,
+  "</s_APPROVED_BY>": 58286,
+  "</s_APPROVE_IN_ITS>": 57854,
+  "</s_APTERTASTE_7=_Pleasant>": 57756,
+  "</s_ATTRIBUTES>": 57754,
+  "</s_AUTHORS>": 57852,
+  "</s_AVERAGE_WEIGHT_RANGE_GM>": 57830,
+  "</s_Additional_Spray>": 57946,
+  "</s_Adhesive>": 57648,
+  "</s_Alert_Number>": 58324,
+  "</s_Application_Pattern>": 57642,
+  "</s_Approved_by>": 57640,
+  "</s_Attn>": 58284,
+  "</s_Authorized_by>": 58144,
+  "</s_Authorized_to>": 58142,
+  "</s_BIOLOGICAL_☐>": 58102,
+  "</s_BLEND>": 57944,
+  "</s_BRANCH>": 58416,
+  "</s_BRANCH_2>": 58414,
+  "</s_BRAND_S>": 58382,
+  "</s_BRC_Codes_w81_carton_order_form>": 58322,
+  "</s_BROWN_&_WILLIAMSON_TOBACCO_CORPORATION_MATERIAL_SPECIFICATION_DESIGN_CHANGE_AUTHORIZATION>": 58412,
+  "</s_BUDGET_NO>": 58282,
+  "</s_BY>": 57684,
+  "</s_BY_2>": 57682,
+  "</s_Batch_Size>": 57942,
+  "</s_Bobbin_Width>": 57580,
+  "</s_Brand>": 57638,
+  "</s_Bright>": 57692,
+  "</s_Bright_2>": 57752,
+  "</s_Bright_KS_white_tipping>": 57750,
+  "</s_CARCINOGEN_OSHA>": 58216,
+  "</s_CARCINOGEN_OTHER>": 58214,
+  "</s_CASING>": 57940,
+  "</s_CHEMICAL_PURITY>": 58072,
+  "</s_CHEMICAL_☐>": 58100,
+  "</s_CIGARETTES>": 57748,
+  "</s_CIGARETTE_MAKING>": 57636,
+  "</s_CLASS>": 58070,
+  "</s_CLIENT>": 58350,
+  "</s_CODE>": 58280,
+  "</s_COLOR>": 58182,
+  "</s_COMBINED>": 58092,
+  "</s_COMMENTS>": 58068,
+  "</s_COMPLETION>": 57680,
+  "</s_COMPOUND>": 57820,
+  "</s_COMPOUND_2>": 57818,
+  "</s_COMPOUND_NAME>": 58088,
+  "</s_COMPOUND_PHYSICAL_PARAMETERS>": 58036,
+  "</s_COMPOUND_SENSITIVE>": 58028,
+  "</s_COMPOUND_SENSITIVE_TO>": 58066,
+  "</s_COMPOUND_US_plate>": 58004,
+  "</s_CONCENTRATION_1mgimit>": 58002,
+  "</s_CONCLUSION>": 57816,
+  "</s_CONFIDENTIAL>": 58348,
+  "</s_CONTROL_AFVERTANTS_PER_PLATE_005_ml_SOLVENT>": 58000,
+  "</s_COUNTRY>": 57678,
+  "</s_Cartons>": 57868,
+  "</s_Cigarette_Circumference>": 57618,
+  "</s_Cigarettes>": 57930,
+  "</s_Circumference>": 57926,
+  "</s_Closures>": 57866,
+  "</s_Color>": 57578,
+  "</s_Company>": 57634,
+  "</s_Copies_to>": 57902,
+  "</s_Copies_to_the_Following>": 57814,
+  "</s_Copies_to_the_following>": 57998,
+  "</s_Country>": 57632,
+  "</s_DATE>": 57676,
+  "</s_DATES>": 58124,
+  "</s_DATE_2>": 57674,
+  "</s_DATE_ISSUED>": 58212,
+  "</s_DATE_OF_REPORT>": 58346,
+  "</s_DATE_RECEIVED>": 57812,
+  "</s_DATE_RECEIVED_2>": 57810,
+  "</s_DATE_WANTED>": 58278,
+  "</s_DECISION_TREE_ESTIMATION_OF_TOXIC_RISK>": 58086,
+  "</s_DELIVERY_DATE>": 58410,
+  "</s_DEPT_NO>": 58276,
+  "</s_DESCRIPTION>": 58274,
+  "</s_DESCRIPTION_OF_REACTIVITY>": 58060,
+  "</s_DESIGN>": 58408,
+  "</s_DESIGN_NO>": 58406,
+  "</s_DESSICATOR>": 58210,
+  "</s_DIR_OF_LIFE_SCIENCES_HEALTH_SERVICES_COMPOUND_PREP_TOX_SCIENTIFIC_RES_TOX>": 58238,
+  "</s_DMSC>": 57984,
+  "</s_DOMESTIC>": 58404,
+  "</s_DOSAGE_mg_kg_BODY_WEIGHT>": 57808,
+  "</s_Dataset_Name>": 58320,
+  "</s_Date>": 57548,
+  "</s_Date_2>": 57630,
+  "</s_Date_3>": 57628,
+  "</s_Description_of_change_order>": 58140,
+  "</s_Dimension>": 57626,
+  "</s_Directional_Difference>": 57746,
+  "</s_Draw>": 57964,
+  "</s_EASE_OF_DRAW_17_Easier>": 57744,
+  "</s_ESTIMATED_TOXICITY_CLASS>": 58084,
+  "</s_EXPERIMENTAL_PROCEDURES>": 58208,
+  "</s_EXPLOSIVE>": 58206,
+  "</s_Estimated_cost_of_the_study_will_be>": 58132,
+  "</s_FEMALE>": 57806,
+  "</s_FIELD_DATE>": 57742,
+  "</s_FINAL_FLAVOR>": 57928,
+  "</s_FIXED_DISTRIBUTION>": 58204,
+  "</s_FLAMMABLE>": 58202,
+  "</s_FOLLOW_DEPARTMENT_AND_COMPANY_SAFETY_MANUALS>": 58200,
+  "</s_FOLLOW_UP_DATE>": 58272,
+  "</s_FOR_PURCHASING_DEPARTMENT_USE_ONLY>": 58270,
+  "</s_FREEZER>": 58198,
+  "</s_F_O_B>": 58254,
+  "</s_Female>": 57740,
+  "</s_Female_106>": 57738,
+  "</s_Filter_Length>": 57938,
+  "</s_Filter_Plug_Length>": 57616,
+  "</s_Filter_Plug_Pressure_Drop_encap>": 57614,
+  "</s_Filter_Plug_Pressure_Drop_unencap>": 57612,
+  "</s_Filter_Production>": 57888,
+  "</s_Filter_ventilation_Rate>": 57610,
+  "</s_Filters>": 57912,
+  "</s_Firmness>": 57962,
+  "</s_From>": 58318,
+  "</s_GENERAL_☐_EXPORT>": 58402,
+  "</s_GREY_ADVERTISING_INC_MEETING_REPORT>": 58344,
+  "</s_GROUP_NO>": 57804,
+  "</s_Give_reasons_below>": 57850,
+  "</s_Glue_Pree_Area>": 57624,
+  "</s_HAZARDOUS_COMPOUND>": 58196,
+  "</s_HAZLETON_LABORATORIES_AMERICA_INC>": 58194,
+  "</s_Have_you_contacted_your_Manager_Supervisor>": 57542,
+  "</s_II>": 58122,
+  "</s_III_SUPPLEMENTAL_LITERATURE_SEARCH_Chemical_Abstracts_Vol_23_Vol_65_Arctander'_s_Perfume_and_Flavor_Materials_of_Natural_Origin_1960_Guenther'_s_Monographs_on_Fragrance_Raw_Materials_1979_Tobacco_Abstracts_U_S_Dispensatory_23rd_edition_1943>": 58120,
+  "</s_INFORMATION_SEARCH_SUMMARY>": 58118,
+  "</s_INITIATED>": 57672,
+  "</s_INVENTORY_DEPLETION_DATE>": 58400,
+  "</s_INVESTIGATORS>": 57802,
+  "</s_ISSUED_BY>": 58268,
+  "</s_If_a_material_or_dimensional_change_also_involved>": 58396,
+  "</s_If_yes_supply_details>": 58394,
+  "</s_Jimmy_H_Bell_B_Sc_Scientist>": 58116,
+  "</s_KL>": 57690,
+  "</s_KOOL_Lights>": 57736,
+  "</s_KOOL_Lights_KS_white_tip_ping_masked>": 57734,
+  "</s_Kind>": 57924,
+  "</s_LD50_95%_CONFIDENCE_LIMITS>": 57800,
+  "</s_LH_NUMBER_s>": 58192,
+  "</s_LH_NUMBERs>": 58236,
+  "</s_LIQUID>": 58180,
+  "</s_LITERATURE_SURVEYED>": 58114,
+  "</s_LORILLARD>": 57996,
+  "</s_LORILLARD_COMPOUND_CODE>": 58112,
+  "</s_LORILLARD_COMPOUND_CODE_NUMBER>": 58082,
+  "</s_LORILLARD_NO>": 57798,
+  "</s_LORILLARD_RESEARCH_CENTER>": 58058,
+  "</s_LOT>": 57994,
+  "</s_LOT_NO_s>": 58190,
+  "</s_LOT_NOs>": 58234,
+  "</s_LOT_NUMBER>": 58026,
+  "</s_LOUISVILLE>": 58392,
+  "</s_LOUISVILLE_2>": 58390,
+  "</s_LRC_COMPOUND_CODE>": 58056,
+  "</s_LRC_FILE_NUMBER>": 58024,
+  "</s_Labels>": 57864,
+  "</s_Laboratory>": 57894,
+  "</s_Laboratory_Analysis>": 57910,
+  "</s_Length>": 57936,
+  "</s_Licensee_Ref_NO>": 57622,
+  "</s_MACON>": 58388,
+  "</s_MACON_2>": 58386,
+  "</s_MALE>": 57796,
+  "</s_MATERIAL_NO>": 58380,
+  "</s_MENT>": 57732,
+  "</s_MENTHOL_FLAVOR>": 57958,
+  "</s_MENTHOL_TASTE_7=_Better>": 57730,
+  "</s_MILITARY_☐_EXPORT>": 58384,
+  "</s_MOIST>": 57728,
+  "</s_MOLECULAR_WEIGHT>": 58054,
+  "</s_Macket>": 57620,
+  "</s_Maildate>": 58316,
+  "</s_Mailfile_Cells>": 58314,
+  "</s_Mailfile_Description>": 58312,
+  "</s_Mailfile_ID>": 58310,
+  "</s_Maker>": 57934,
+  "</s_Making_&_Packing>": 57886,
+  "</s_Male>": 57726,
+  "</s_Male_99>": 57724,
+  "</s_Manager_Comments>": 57540,
+  "</s_Marginal_Difference>": 57722,
+  "</s_Markings>": 57862,
+  "</s_Method>": 58130,
+  "</s_Moisture_content_ex_catcher>": 57608,
+  "</s_NAME>": 58080,
+  "</s_NAME_2>": 58110,
+  "</s_NATURE_OF_WORK>": 57670,
+  "</s_NET>": 58252,
+  "</s_NEW_ITEM>": 58378,
+  "</s_NIC>": 57720,
+  "</s_NO>": 57992,
+  "</s_NOTEBOOK_PAGE>": 57794,
+  "</s_NOTES>": 57668,
+  "</s_NO_2>": 57990,
+  "</s_Name_Phone_Ext>": 57538,
+  "</s_Net_Net_Tobacco>": 57558,
+  "</s_Net_Tob_Rod_Density>": 57556,
+  "</s_No>": 57546,
+  "</s_No_Preference>": 57718,
+  "</s_Notes>": 58308,
+  "</s_OFFICIAL>": 58366,
+  "</s_ORAL>": 58022,
+  "</s_ORDER_NO>": 58266,
+  "</s_ORIGINAL_SIGNED_PROJECT_SHEETS_TO_PSO>": 58188,
+  "</s_OTHER>": 58186,
+  "</s_Original_Request_Made_By>": 57908,
+  "</s_Original_request_made_by>": 57956,
+  "</s_Original_to>": 57900,
+  "</s_Other>": 57892,
+  "</s_Others>": 57954,
+  "</s_Overall>": 57646,
+  "</s_Overall_Cigarette_Length>": 57606,
+  "</s_PAGE_#>": 57716,
+  "</s_PDC_#>": 57714,
+  "</s_PERFORMING_DEPARTMENTS>": 58184,
+  "</s_PETERSBURG>": 58376,
+  "</s_PETERSBURG_2>": 58374,
+  "</s_PH>": 58342,
+  "</s_PHYSICAL_DESCRIPTION>": 58174,
+  "</s_PHYSICAL_STATE>": 58020,
+  "</s_PLACE>": 58340,
+  "</s_PO>": 57786,
+  "</s_POSTIVE_CONTROL_ug_plate>": 57988,
+  "</s_PREFERENCE>": 57712,
+  "</s_PRESENT>": 58338,
+  "</s_PRESENT_FORM>": 57848,
+  "</s_PRESSURIZED>": 58178,
+  "</s_PRICE>": 58264,
+  "</s_PRODUCT>": 57666,
+  "</s_PROJECT_#>": 57710,
+  "</s_PROJECT_COORDINATOR>": 58172,
+  "</s_PROJECT_NO>": 58170,
+  "</s_PROJECT_SHEET>": 58232,
+  "</s_PROJECT_SHEET_NO>": 58168,
+  "</s_PROTECTIVE_PRECAUTIONS_REQUIRED>": 58166,
+  "</s_PURCHASE_REQUISITION>": 58262,
+  "</s_PURCHASING>": 58332,
+  "</s_Paper>": 57932,
+  "</s_Perforation_Type_and_No_of_lines>": 57576,
+  "</s_Physical_Characteristics>": 57594,
+  "</s_Place_of_Manufacture>": 57592,
+  "</s_Plasticizer>": 57922,
+  "</s_Plug_Wrap>": 57920,
+  "</s_Porosity>": 57574,
+  "</s_Positioning>": 57590,
+  "</s_Prepared_by>": 57588,
+  "</s_Pressure_Drop>": 57906,
+  "</s_Prev_or_Recommended_Supplier>": 58330,
+  "</s_Print>": 57960,
+  "</s_Print_Description>": 57572,
+  "</s_Print_Position_from_filter_end>": 57604,
+  "</s_Process>": 57918,
+  "</s_Program#>": 58306,
+  "</s_Program_Group>": 58304,
+  "</s_QNED>": 58372,
+  "</s_QUANTITY>": 58260,
+  "</s_Quantity>": 58302,
+  "</s_R&D_Group>": 57536,
+  "</s_R&D_QUALITY_IMPROVEMENT_SUGGESTION_SOLUTION_FORM>": 57534,
+  "</s_RADIOACTIVE>": 58164,
+  "</s_REACTIVITY>": 58040,
+  "</s_RECASING>": 57904,
+  "</s_RECEIPT_DATE_s>": 58162,
+  "</s_RECEIPT_DATEs>": 58230,
+  "</s_RECEIVED>": 57664,
+  "</s_RECOMMENDATION>": 57846,
+  "</s_REFERENCE_FOR_CALCULATION>": 57792,
+  "</s_REFRIGERATOR_8_C>": 58228,
+  "</s_REGISTRY_NUMBER_IF_applicable>": 58108,
+  "</s_REPLACEMENT>": 58370,
+  "</s_REPORTED>": 57790,
+  "</s_REQUEST_NO>": 57662,
+  "</s_REQUISITION_NO>": 58258,
+  "</s_RESTRICTED_BROWN_&_WILLIAMSON_INTERNATIONAL_TOBACCO_PRODUCT_SPECIFICATION>": 57586,
+  "</s_RESULTS_NO_TESTED>": 57788,
+  "</s_REVERSION_RATE_TEST_REVERSANTSCONTROL_REVERTANTS_PER_PLATE>": 57986,
+  "</s_REVIEWED>": 58106,
+  "</s_REVIEW_COMPLETED>": 57844,
+  "</s_RJR_Mailfile_Table_Update_Sheet>": 58300,
+  "</s_ROOM_TEMPERATURE>": 58160,
+  "</s_ROUTE_OF_COMPOUND_ADMINISTRATION>": 57778,
+  "</s_R_&_D_COMMENTS>": 57660,
+  "</s_Reference>": 58222,
+  "</s_Reference_2>": 58220,
+  "</s_Registration_No>": 57842,
+  "</s_Registry_of_the_Toxic_Effects_of_Chemicals>": 58104,
+  "</s_Replaces>": 57584,
+  "</s_Reports>": 57896,
+  "</s_Requirements>": 57890,
+  "</s_Responsibility>": 57878,
+  "</s_Robbin_Length>": 57570,
+  "</s_Rod_Length>": 57916,
+  "</s_SAMPLE>": 57708,
+  "</s_SATISFYING_7=_More>": 57706,
+  "</s_SCHEDULE_DATE_OR_INVENTORY_DEPLETION>": 58368,
+  "</s_SENSITIVE>": 58218,
+  "</s_SERVICE_REQUEST_FROM_R_&_D_BY_INTERNATIONAL_LICENSEE_OPERATIONS>": 57658,
+  "</s_SHIP_TO_DEPT_BRANCH>": 58256,
+  "</s_SIGNATURE>": 58018,
+  "</s_SIGNATURES>": 57776,
+  "</s_SMOOTHNESS_7=_5moother>": 57704,
+  "</s_SOLID>": 58176,
+  "</s_SOLUBILITY>": 58016,
+  "</s_SOLVENT>": 57980,
+  "</s_SOURCE>": 57774,
+  "</s_SOURCE_2>": 57772,
+  "</s_SPECIAL_INSTRUCTIONS_WHILE_DOSING>": 58158,
+  "</s_STATIONARY>": 58328,
+  "</s_STORAGE_CONDITIONS>": 58156,
+  "</s_STORAGE_RECOMMENDATIONS>": 58014,
+  "</s_STORE_IN_DARK>": 58154,
+  "</s_STRAIN_OF_MICE>": 57770,
+  "</s_STRENGTH_7=_Stronger>": 57702,
+  "</s_STRUCTURE>": 58078,
+  "</s_STUDY_TITLE_&_PROPOSAL_NUMBER>": 58152,
+  "</s_SUBJECT>": 57700,
+  "</s_SUBJECT_TO_THE_FOLLOWING_SUGGESTED_REVISIONS_itemize_below>": 57840,
+  "</s_SUMMARY>": 58336,
+  "</s_S_TYPHIMURIUM>": 57978,
+  "</s_Sample_No>": 57876,
+  "</s_Sample_Requisition_Form_02_02_06>": 57884,
+  "</s_Sample_Specifications_Written_By>": 57874,
+  "</s_Sample_specifications_written_by>": 57952,
+  "</s_Section_III_A_2nd_paragraph_the_first_sentence_is_to_be_changed_to_read_as_follows>": 58128,
+  "</s_Seeds>": 58298,
+  "</s_Shipping>": 57882,
+  "</s_Significant_Difference>": 57698,
+  "</s_Special_Requirements>": 57872,
+  "</s_Style>": 57582,
+  "</s_Subject>": 58126,
+  "</s_Suggested_Solutions_s>": 57532,
+  "</s_Suggestion>": 57530,
+  "</s_Supervisor_Manager>": 57528,
+  "</s_Supplier_Code_No_s>": 57654,
+  "</s_Supplier_Code_Nos>": 57568,
+  "</s_Supplier_s>": 57652,
+  "</s_Suppliers>": 57566,
+  "</s_Suppression>": 58296,
+  "</s_System>": 58294,
+  "</s_TA100>": 57976,
+  "</s_TA1535>": 57972,
+  "</s_TA98>": 57968,
+  "</s_TAR>": 57696,
+  "</s_TARGET_DATE>": 57656,
+  "</s_TERMS>": 58248,
+  "</s_TESTED>": 57768,
+  "</s_TEST_MATERIAL_S>": 58150,
+  "</s_TEST_MATERIALs>": 58226,
+  "</s_TITLE>": 57838,
+  "</s_TO>": 58012,
+  "</s_TOBACCO_SCIENCE>": 57836,
+  "</s_TOP_LINES_MENTHOL_PRODUCT_TEST>": 57694,
+  "</s_TOXICITY_%_SURVIVAL>": 57966,
+  "</s_TYPE>": 58098,
+  "</s_TYPE_CHANGE>": 58362,
+  "</s_T_Codes_NA>": 58292,
+  "</s_Tear_Tape>": 57950,
+  "</s_Tear_rape>": 57860,
+  "</s_Tipping_Length>": 57602,
+  "</s_Tipping_Paper>": 57562,
+  "</s_Tipping_and_Tipping_Application>": 57560,
+  "</s_To>": 58290,
+  "</s_Tobacco_Blend>": 57880,
+  "</s_Tobacco_Rod_Length>": 57600,
+  "</s_Total>": 57688,
+  "</s_Total_205>": 57686,
+  "</s_Total_Cigt_Wt>": 57554,
+  "</s_Total_Pressure_Drop_encap>": 57598,
+  "</s_Total_Pressure_Drop_unencap>": 57596,
+  "</s_Toxline>": 58096,
+  "</s_Type>": 57650,
+  "</s_Type_of_Cigarette>": 57870,
+  "</s_UNIT>": 58246,
+  "</s_UNIT_PRICE>": 58326,
+  "</s_UNK_TOKEN>": 57526,
+  "</s_UPC_NO>": 58360,
+  "</s_USE>": 58358,
+  "</s_VENDOR>": 58244,
+  "</s_VIA>": 58250,
+  "</s_WINSTON_SALEM>": 58356,
+  "</s_WINSTON_SALEM_2>": 58354,
+  "</s_WITHOUT_HEATING>": 58064,
+  "</s_WITH_HEATING_80C>": 58062,
+  "</s_Weight>": 57914,
+  "</s_Weights>": 57552,
+  "</s_Wrappings>": 57948,
+  "</s_Written_by>": 57898,
+  "</s_Yes>": 57544,
+  "</s_apping>": 57858,
+  "</s_cc>": 58334,
+  "</s_cont'd>": 58094,
+  "</s_decreased>": 58138,
+  "</s_increased>": 58136,
+  "</s_not_affected>": 58134,
+  "</s_on>": 57856,
+  "</s_pH>": 58010,
+  "</s_skip>": 57644,
+  "</s_substance>": 57564,
+  "</s_to_be_performed_in_Acute_Dermal_Toxicology_building_18>": 58148,
+  "</s_☐>": 58352,
+  "</s_☐_AIR>": 58034,
+  "</s_☐_CORN_OIL>": 57826,
+  "</s_☐_HEAT>": 58032,
+  "</s_☐_INHALATION>": 57784,
+  "</s_☐_IP>": 57782,
+  "</s_☐_IV>": 57780,
+  "</s_☐_MOISTURE>": 58030,
+  "</s_☐_OTHER>": 57824,
+  "</s_☐_PURCHASING>": 58242,
+  "</s_☐_SALINE>": 57822,
+  "</s_☐_STATIONARY>": 58240,
+  "</s_☐_TEMPORARY>": 58364,
+  "</s_☐_WATER>": 57982,
+  "</s_☐_YES>": 58398,
+  "</s_☑>": 58090,
+  "<s_%_SOLUTION>": 57833,
+  "<s_+>": 57969,
+  "<s_1_WATER_or_BRINE>": 58051,
+  "<s_21_34>": 57765,
+  "<s_21_34_96>": 57763,
+  "<s_2_5%_HCL>": 58049,
+  "<s_2_AMINOANTHHACENE>": 58007,
+  "<s_35+>": 57761,
+  "<s_35+_109>": 57759,
+  "<s_3_5%_NaOH>": 58047,
+  "<s_4_ALCOHOLS>": 58045,
+  "<s_5_%_METHYL_CELULOSE>": 57827,
+  "<s_5_OXYGEN>": 58043,
+  "<s_6_LIGHT>": 58041,
+  "<s_9>": 57973,
+  "<s_>": 57549,
+  "<s_ACCT_NO>": 58287,
+  "<s_ACUTE_CADCOVASCULAR>": 58223,
+  "<s_ACUTE_CARDIOVASCULAR>": 58037,
+  "<s_ACUTE_TOXICITY_IN_MICE>": 57831,
+  "<s_ADDENDUM_I_DAWSON_RESEARCH_CORPORATION_Protocol_Change_Order_No_1>": 58145,
+  "<s_AMES_ASSAY_FOR_MUTAGENICITY>": 58005,
+  "<s_AMOUNT>": 58075,
+  "<s_AMT_MENTHOL_7=_More>": 57757,
+  "<s_ANALYTICAL_METHODS>": 58073,
+  "<s_APPROVED_BY>": 58285,
+  "<s_APPROVE_IN_ITS>": 57853,
+  "<s_APTERTASTE_7=_Pleasant>": 57755,
+  "<s_ATTRIBUTES>": 57753,
+  "<s_AUTHORS>": 57851,
+  "<s_AVERAGE_WEIGHT_RANGE_GM>": 57829,
+  "<s_Additional_Spray>": 57945,
+  "<s_Adhesive>": 57647,
+  "<s_Alert_Number>": 58323,
+  "<s_Application_Pattern>": 57641,
+  "<s_Approved_by>": 57639,
+  "<s_Attn>": 58283,
+  "<s_Authorized_by>": 58143,
+  "<s_Authorized_to>": 58141,
+  "<s_BIOLOGICAL_☐>": 58101,
+  "<s_BLEND>": 57943,
+  "<s_BRANCH>": 58415,
+  "<s_BRANCH_2>": 58413,
+  "<s_BRAND_S>": 58381,
+  "<s_BRC_Codes_w81_carton_order_form>": 58321,
+  "<s_BROWN_&_WILLIAMSON_TOBACCO_CORPORATION_MATERIAL_SPECIFICATION_DESIGN_CHANGE_AUTHORIZATION>": 58411,
+  "<s_BUDGET_NO>": 58281,
+  "<s_BY>": 57683,
+  "<s_BY_2>": 57681,
+  "<s_Batch_Size>": 57941,
+  "<s_Bobbin_Width>": 57579,
+  "<s_Brand>": 57637,
+  "<s_Bright>": 57691,
+  "<s_Bright_2>": 57751,
+  "<s_Bright_KS_white_tipping>": 57749,
+  "<s_CARCINOGEN_OSHA>": 58215,
+  "<s_CARCINOGEN_OTHER>": 58213,
+  "<s_CASING>": 57939,
+  "<s_CHEMICAL_PURITY>": 58071,
+  "<s_CHEMICAL_☐>": 58099,
+  "<s_CIGARETTES>": 57747,
+  "<s_CIGARETTE_MAKING>": 57635,
+  "<s_CLASS>": 58069,
+  "<s_CLIENT>": 58349,
+  "<s_CODE>": 58279,
+  "<s_COLOR>": 58181,
+  "<s_COMBINED>": 58091,
+  "<s_COMMENTS>": 58067,
+  "<s_COMPLETION>": 57679,
+  "<s_COMPOUND>": 57819,
+  "<s_COMPOUND_2>": 57817,
+  "<s_COMPOUND_NAME>": 58087,
+  "<s_COMPOUND_PHYSICAL_PARAMETERS>": 58035,
+  "<s_COMPOUND_SENSITIVE>": 58027,
+  "<s_COMPOUND_SENSITIVE_TO>": 58065,
+  "<s_COMPOUND_US_plate>": 58003,
+  "<s_CONCENTRATION_1mgimit>": 58001,
+  "<s_CONCLUSION>": 57815,
+  "<s_CONFIDENTIAL>": 58347,
+  "<s_CONTROL_AFVERTANTS_PER_PLATE_005_ml_SOLVENT>": 57999,
+  "<s_COUNTRY>": 57677,
+  "<s_Cartons>": 57867,
+  "<s_Cigarette_Circumference>": 57617,
+  "<s_Cigarettes>": 57929,
+  "<s_Circumference>": 57925,
+  "<s_Closures>": 57865,
+  "<s_Color>": 57577,
+  "<s_Company>": 57633,
+  "<s_Copies_to>": 57901,
+  "<s_Copies_to_the_Following>": 57813,
+  "<s_Copies_to_the_following>": 57997,
+  "<s_Country>": 57631,
+  "<s_DATE>": 57675,
+  "<s_DATES>": 58123,
+  "<s_DATE_2>": 57673,
+  "<s_DATE_ISSUED>": 58211,
+  "<s_DATE_OF_REPORT>": 58345,
+  "<s_DATE_RECEIVED>": 57811,
+  "<s_DATE_RECEIVED_2>": 57809,
+  "<s_DATE_WANTED>": 58277,
+  "<s_DECISION_TREE_ESTIMATION_OF_TOXIC_RISK>": 58085,
+  "<s_DELIVERY_DATE>": 58409,
+  "<s_DEPT_NO>": 58275,
+  "<s_DESCRIPTION>": 58273,
+  "<s_DESCRIPTION_OF_REACTIVITY>": 58059,
+  "<s_DESIGN>": 58407,
+  "<s_DESIGN_NO>": 58405,
+  "<s_DESSICATOR>": 58209,
+  "<s_DIR_OF_LIFE_SCIENCES_HEALTH_SERVICES_COMPOUND_PREP_TOX_SCIENTIFIC_RES_TOX>": 58237,
+  "<s_DMSC>": 57983,
+  "<s_DOMESTIC>": 58403,
+  "<s_DOSAGE_mg_kg_BODY_WEIGHT>": 57807,
+  "<s_Dataset_Name>": 58319,
+  "<s_Date>": 57547,
+  "<s_Date_2>": 57629,
+  "<s_Date_3>": 57627,
+  "<s_Description_of_change_order>": 58139,
+  "<s_Dimension>": 57625,
+  "<s_Directional_Difference>": 57745,
+  "<s_Draw>": 57963,
+  "<s_EASE_OF_DRAW_17_Easier>": 57743,
+  "<s_ESTIMATED_TOXICITY_CLASS>": 58083,
+  "<s_EXPERIMENTAL_PROCEDURES>": 58207,
+  "<s_EXPLOSIVE>": 58205,
+  "<s_Estimated_cost_of_the_study_will_be>": 58131,
+  "<s_FEMALE>": 57805,
+  "<s_FIELD_DATE>": 57741,
+  "<s_FINAL_FLAVOR>": 57927,
+  "<s_FIXED_DISTRIBUTION>": 58203,
+  "<s_FLAMMABLE>": 58201,
+  "<s_FOLLOW_DEPARTMENT_AND_COMPANY_SAFETY_MANUALS>": 58199,
+  "<s_FOLLOW_UP_DATE>": 58271,
+  "<s_FOR_PURCHASING_DEPARTMENT_USE_ONLY>": 58269,
+  "<s_FREEZER>": 58197,
+  "<s_F_O_B>": 58253,
+  "<s_Female>": 57739,
+  "<s_Female_106>": 57737,
+  "<s_Filter_Length>": 57937,
+  "<s_Filter_Plug_Length>": 57615,
+  "<s_Filter_Plug_Pressure_Drop_encap>": 57613,
+  "<s_Filter_Plug_Pressure_Drop_unencap>": 57611,
+  "<s_Filter_Production>": 57887,
+  "<s_Filter_ventilation_Rate>": 57609,
+  "<s_Filters>": 57911,
+  "<s_Firmness>": 57961,
+  "<s_From>": 58317,
+  "<s_GENERAL_☐_EXPORT>": 58401,
+  "<s_GREY_ADVERTISING_INC_MEETING_REPORT>": 58343,
+  "<s_GROUP_NO>": 57803,
+  "<s_Give_reasons_below>": 57849,
+  "<s_Glue_Pree_Area>": 57623,
+  "<s_HAZARDOUS_COMPOUND>": 58195,
+  "<s_HAZLETON_LABORATORIES_AMERICA_INC>": 58193,
+  "<s_Have_you_contacted_your_Manager_Supervisor>": 57541,
+  "<s_II>": 58121,
+  "<s_III_SUPPLEMENTAL_LITERATURE_SEARCH_Chemical_Abstracts_Vol_23_Vol_65_Arctander'_s_Perfume_and_Flavor_Materials_of_Natural_Origin_1960_Guenther'_s_Monographs_on_Fragrance_Raw_Materials_1979_Tobacco_Abstracts_U_S_Dispensatory_23rd_edition_1943>": 58119,
+  "<s_INFORMATION_SEARCH_SUMMARY>": 58117,
+  "<s_INITIATED>": 57671,
+  "<s_INVENTORY_DEPLETION_DATE>": 58399,
+  "<s_INVESTIGATORS>": 57801,
+  "<s_ISSUED_BY>": 58267,
+  "<s_If_a_material_or_dimensional_change_also_involved>": 58395,
+  "<s_If_yes_supply_details>": 58393,
+  "<s_Jimmy_H_Bell_B_Sc_Scientist>": 58115,
+  "<s_KL>": 57689,
+  "<s_KOOL_Lights>": 57735,
+  "<s_KOOL_Lights_KS_white_tip_ping_masked>": 57733,
+  "<s_Kind>": 57923,
+  "<s_LD50_95%_CONFIDENCE_LIMITS>": 57799,
+  "<s_LH_NUMBER_s>": 58191,
+  "<s_LH_NUMBERs>": 58235,
+  "<s_LIQUID>": 58179,
+  "<s_LITERATURE_SURVEYED>": 58113,
+  "<s_LORILLARD>": 57995,
+  "<s_LORILLARD_COMPOUND_CODE>": 58111,
+  "<s_LORILLARD_COMPOUND_CODE_NUMBER>": 58081,
+  "<s_LORILLARD_NO>": 57797,
+  "<s_LORILLARD_RESEARCH_CENTER>": 58057,
+  "<s_LOT>": 57993,
+  "<s_LOT_NO_s>": 58189,
+  "<s_LOT_NOs>": 58233,
+  "<s_LOT_NUMBER>": 58025,
+  "<s_LOUISVILLE>": 58391,
+  "<s_LOUISVILLE_2>": 58389,
+  "<s_LRC_COMPOUND_CODE>": 58055,
+  "<s_LRC_FILE_NUMBER>": 58023,
+  "<s_Labels>": 57863,
+  "<s_Laboratory>": 57893,
+  "<s_Laboratory_Analysis>": 57909,
+  "<s_Length>": 57935,
+  "<s_Licensee_Ref_NO>": 57621,
+  "<s_MACON>": 58387,
+  "<s_MACON_2>": 58385,
+  "<s_MALE>": 57795,
+  "<s_MATERIAL_NO>": 58379,
+  "<s_MENT>": 57731,
+  "<s_MENTHOL_FLAVOR>": 57957,
+  "<s_MENTHOL_TASTE_7=_Better>": 57729,
+  "<s_MILITARY_☐_EXPORT>": 58383,
+  "<s_MOIST>": 57727,
+  "<s_MOLECULAR_WEIGHT>": 58053,
+  "<s_Macket>": 57619,
+  "<s_Maildate>": 58315,
+  "<s_Mailfile_Cells>": 58313,
+  "<s_Mailfile_Description>": 58311,
+  "<s_Mailfile_ID>": 58309,
+  "<s_Maker>": 57933,
+  "<s_Making_&_Packing>": 57885,
+  "<s_Male>": 57725,
+  "<s_Male_99>": 57723,
+  "<s_Manager_Comments>": 57539,
+  "<s_Marginal_Difference>": 57721,
+  "<s_Markings>": 57861,
+  "<s_Method>": 58129,
+  "<s_Moisture_content_ex_catcher>": 57607,
+  "<s_NAME>": 58079,
+  "<s_NAME_2>": 58109,
+  "<s_NATURE_OF_WORK>": 57669,
+  "<s_NET>": 58251,
+  "<s_NEW_ITEM>": 58377,
+  "<s_NIC>": 57719,
+  "<s_NO>": 57991,
+  "<s_NOTEBOOK_PAGE>": 57793,
+  "<s_NOTES>": 57667,
+  "<s_NO_2>": 57989,
+  "<s_Name_Phone_Ext>": 57537,
+  "<s_Net_Net_Tobacco>": 57557,
+  "<s_Net_Tob_Rod_Density>": 57555,
+  "<s_No>": 57545,
+  "<s_No_Preference>": 57717,
+  "<s_Notes>": 58307,
+  "<s_OFFICIAL>": 58365,
+  "<s_ORAL>": 58021,
+  "<s_ORDER_NO>": 58265,
+  "<s_ORIGINAL_SIGNED_PROJECT_SHEETS_TO_PSO>": 58187,
+  "<s_OTHER>": 58185,
+  "<s_Original_Request_Made_By>": 57907,
+  "<s_Original_request_made_by>": 57955,
+  "<s_Original_to>": 57899,
+  "<s_Other>": 57891,
+  "<s_Others>": 57953,
+  "<s_Overall>": 57645,
+  "<s_Overall_Cigarette_Length>": 57605,
+  "<s_PAGE_#>": 57715,
+  "<s_PDC_#>": 57713,
+  "<s_PERFORMING_DEPARTMENTS>": 58183,
+  "<s_PETERSBURG>": 58375,
+  "<s_PETERSBURG_2>": 58373,
+  "<s_PH>": 58341,
+  "<s_PHYSICAL_DESCRIPTION>": 58173,
+  "<s_PHYSICAL_STATE>": 58019,
+  "<s_PLACE>": 58339,
+  "<s_PO>": 57785,
+  "<s_POSTIVE_CONTROL_ug_plate>": 57987,
+  "<s_PREFERENCE>": 57711,
+  "<s_PRESENT>": 58337,
+  "<s_PRESENT_FORM>": 57847,
+  "<s_PRESSURIZED>": 58177,
+  "<s_PRICE>": 58263,
+  "<s_PRODUCT>": 57665,
+  "<s_PROJECT_#>": 57709,
+  "<s_PROJECT_COORDINATOR>": 58171,
+  "<s_PROJECT_NO>": 58169,
+  "<s_PROJECT_SHEET>": 58231,
+  "<s_PROJECT_SHEET_NO>": 58167,
+  "<s_PROTECTIVE_PRECAUTIONS_REQUIRED>": 58165,
+  "<s_PURCHASE_REQUISITION>": 58261,
+  "<s_PURCHASING>": 58331,
+  "<s_Paper>": 57931,
+  "<s_Perforation_Type_and_No_of_lines>": 57575,
+  "<s_Physical_Characteristics>": 57593,
+  "<s_Place_of_Manufacture>": 57591,
+  "<s_Plasticizer>": 57921,
+  "<s_Plug_Wrap>": 57919,
+  "<s_Porosity>": 57573,
+  "<s_Positioning>": 57589,
+  "<s_Prepared_by>": 57587,
+  "<s_Pressure_Drop>": 57905,
+  "<s_Prev_or_Recommended_Supplier>": 58329,
+  "<s_Print>": 57959,
+  "<s_Print_Description>": 57571,
+  "<s_Print_Position_from_filter_end>": 57603,
+  "<s_Process>": 57917,
+  "<s_Program#>": 58305,
+  "<s_Program_Group>": 58303,
+  "<s_QNED>": 58371,
+  "<s_QUANTITY>": 58259,
+  "<s_Quantity>": 58301,
+  "<s_R&D_Group>": 57535,
+  "<s_R&D_QUALITY_IMPROVEMENT_SUGGESTION_SOLUTION_FORM>": 57533,
+  "<s_RADIOACTIVE>": 58163,
+  "<s_REACTIVITY>": 58039,
+  "<s_RECASING>": 57903,
+  "<s_RECEIPT_DATE_s>": 58161,
+  "<s_RECEIPT_DATEs>": 58229,
+  "<s_RECEIVED>": 57663,
+  "<s_RECOMMENDATION>": 57845,
+  "<s_REFERENCE_FOR_CALCULATION>": 57791,
+  "<s_REFRIGERATOR_8_C>": 58227,
+  "<s_REGISTRY_NUMBER_IF_applicable>": 58107,
+  "<s_REPLACEMENT>": 58369,
+  "<s_REPORTED>": 57789,
+  "<s_REQUEST_NO>": 57661,
+  "<s_REQUISITION_NO>": 58257,
+  "<s_RESTRICTED_BROWN_&_WILLIAMSON_INTERNATIONAL_TOBACCO_PRODUCT_SPECIFICATION>": 57585,
+  "<s_RESULTS_NO_TESTED>": 57787,
+  "<s_REVERSION_RATE_TEST_REVERSANTSCONTROL_REVERTANTS_PER_PLATE>": 57985,
+  "<s_REVIEWED>": 58105,
+  "<s_REVIEW_COMPLETED>": 57843,
+  "<s_RJR_Mailfile_Table_Update_Sheet>": 58299,
+  "<s_ROOM_TEMPERATURE>": 58159,
+  "<s_ROUTE_OF_COMPOUND_ADMINISTRATION>": 57777,
+  "<s_R_&_D_COMMENTS>": 57659,
+  "<s_Reference>": 58221,
+  "<s_Reference_2>": 58219,
+  "<s_Registration_No>": 57841,
+  "<s_Registry_of_the_Toxic_Effects_of_Chemicals>": 58103,
+  "<s_Replaces>": 57583,
+  "<s_Reports>": 57895,
+  "<s_Requirements>": 57889,
+  "<s_Responsibility>": 57877,
+  "<s_Robbin_Length>": 57569,
+  "<s_Rod_Length>": 57915,
+  "<s_SAMPLE>": 57707,
+  "<s_SATISFYING_7=_More>": 57705,
+  "<s_SCHEDULE_DATE_OR_INVENTORY_DEPLETION>": 58367,
+  "<s_SENSITIVE>": 58217,
+  "<s_SERVICE_REQUEST_FROM_R_&_D_BY_INTERNATIONAL_LICENSEE_OPERATIONS>": 57657,
+  "<s_SHIP_TO_DEPT_BRANCH>": 58255,
+  "<s_SIGNATURE>": 58017,
+  "<s_SIGNATURES>": 57775,
+  "<s_SMOOTHNESS_7=_5moother>": 57703,
+  "<s_SOLID>": 58175,
+  "<s_SOLUBILITY>": 58015,
+  "<s_SOLVENT>": 57979,
+  "<s_SOURCE>": 57773,
+  "<s_SOURCE_2>": 57771,
+  "<s_SPECIAL_INSTRUCTIONS_WHILE_DOSING>": 58157,
+  "<s_STATIONARY>": 58327,
+  "<s_STORAGE_CONDITIONS>": 58155,
+  "<s_STORAGE_RECOMMENDATIONS>": 58013,
+  "<s_STORE_IN_DARK>": 58153,
+  "<s_STRAIN_OF_MICE>": 57769,
+  "<s_STRENGTH_7=_Stronger>": 57701,
+  "<s_STRUCTURE>": 58077,
+  "<s_STUDY_TITLE_&_PROPOSAL_NUMBER>": 58151,
+  "<s_SUBJECT>": 57699,
+  "<s_SUBJECT_TO_THE_FOLLOWING_SUGGESTED_REVISIONS_itemize_below>": 57839,
+  "<s_SUMMARY>": 58335,
+  "<s_S_TYPHIMURIUM>": 57977,
+  "<s_Sample_No>": 57875,
+  "<s_Sample_Requisition_Form_02_02_06>": 57883,
+  "<s_Sample_Specifications_Written_By>": 57873,
+  "<s_Sample_specifications_written_by>": 57951,
+  "<s_Section_III_A_2nd_paragraph_the_first_sentence_is_to_be_changed_to_read_as_follows>": 58127,
+  "<s_Seeds>": 58297,
+  "<s_Shipping>": 57881,
+  "<s_Significant_Difference>": 57697,
+  "<s_Special_Requirements>": 57871,
+  "<s_Style>": 57581,
+  "<s_Subject>": 58125,
+  "<s_Suggested_Solutions_s>": 57531,
+  "<s_Suggestion>": 57529,
+  "<s_Supervisor_Manager>": 57527,
+  "<s_Supplier_Code_No_s>": 57653,
+  "<s_Supplier_Code_Nos>": 57567,
+  "<s_Supplier_s>": 57651,
+  "<s_Suppliers>": 57565,
+  "<s_Suppression>": 58295,
+  "<s_System>": 58293,
+  "<s_TA100>": 57975,
+  "<s_TA1535>": 57971,
+  "<s_TA98>": 57967,
+  "<s_TAR>": 57695,
+  "<s_TARGET_DATE>": 57655,
+  "<s_TERMS>": 58247,
+  "<s_TESTED>": 57767,
+  "<s_TEST_MATERIAL_S>": 58149,
+  "<s_TEST_MATERIALs>": 58225,
+  "<s_TITLE>": 57837,
+  "<s_TO>": 58011,
+  "<s_TOBACCO_SCIENCE>": 57835,
+  "<s_TOP_LINES_MENTHOL_PRODUCT_TEST>": 57693,
+  "<s_TOXICITY_%_SURVIVAL>": 57965,
+  "<s_TYPE>": 58097,
+  "<s_TYPE_CHANGE>": 58361,
+  "<s_T_Codes_NA>": 58291,
+  "<s_Tear_Tape>": 57949,
+  "<s_Tear_rape>": 57859,
+  "<s_Tipping_Length>": 57601,
+  "<s_Tipping_Paper>": 57561,
+  "<s_Tipping_and_Tipping_Application>": 57559,
+  "<s_To>": 58289,
+  "<s_Tobacco_Blend>": 57879,
+  "<s_Tobacco_Rod_Length>": 57599,
+  "<s_Total>": 57687,
+  "<s_Total_205>": 57685,
+  "<s_Total_Cigt_Wt>": 57553,
+  "<s_Total_Pressure_Drop_encap>": 57597,
+  "<s_Total_Pressure_Drop_unencap>": 57595,
+  "<s_Toxline>": 58095,
+  "<s_Type>": 57649,
+  "<s_Type_of_Cigarette>": 57869,
+  "<s_UNIT>": 58245,
+  "<s_UNIT_PRICE>": 58325,
+  "<s_UNK_TOKEN>": 57525,
+  "<s_UPC_NO>": 58359,
+  "<s_USE>": 58357,
+  "<s_VENDOR>": 58243,
+  "<s_VIA>": 58249,
+  "<s_WINSTON_SALEM>": 58355,
+  "<s_WINSTON_SALEM_2>": 58353,
+  "<s_WITHOUT_HEATING>": 58063,
+  "<s_WITH_HEATING_80C>": 58061,
+  "<s_Weight>": 57913,
+  "<s_Weights>": 57551,
+  "<s_Wrappings>": 57947,
+  "<s_Written_by>": 57897,
+  "<s_Yes>": 57543,
+  "<s_apping>": 57857,
+  "<s_cc>": 58333,
+  "<s_cont'd>": 58093,
+  "<s_decreased>": 58137,
+  "<s_iitcdip>": 57523,
+  "<s_increased>": 58135,
+  "<s_not_affected>": 58133,
+  "<s_on>": 57855,
+  "<s_pH>": 58009,
+  "<s_skip>": 57643,
+  "<s_substance>": 57563,
+  "<s_synthdog>": 57524,
+  "<s_to_be_performed_in_Acute_Dermal_Toxicology_building_18>": 58147,
+  "<s_☐>": 58351,
+  "<s_☐_AIR>": 58033,
+  "<s_☐_CORN_OIL>": 57825,
+  "<s_☐_HEAT>": 58031,
+  "<s_☐_INHALATION>": 57783,
+  "<s_☐_IP>": 57781,
+  "<s_☐_IV>": 57779,
+  "<s_☐_MOISTURE>": 58029,
+  "<s_☐_OTHER>": 57823,
+  "<s_☐_PURCHASING>": 58241,
+  "<s_☐_SALINE>": 57821,
+  "<s_☐_STATIONARY>": 58239,
+  "<s_☐_TEMPORARY>": 58363,
+  "<s_☐_WATER>": 57981,
+  "<s_☐_YES>": 58397,
+  "<s_☑>": 58089,
+  "<sep/>": 57522
+}

preprocessor_config.json

@@ -0,0 +1,26 @@
+{
+  "do_align_long_axis": false,
+  "do_normalize": true,
+  "do_pad": true,
+  "do_rescale": true,
+  "do_resize": true,
+  "do_thumbnail": true,
+  "image_mean": [
+    0.5,
+    0.5,
+    0.5
+  ],
+  "image_processor_type": "DonutImageProcessor",
+  "image_std": [
+    0.5,
+    0.5,
+    0.5
+  ],
+  "processor_class": "DonutProcessor",
+  "resample": 2,
+  "rescale_factor": 0.00392156862745098,
+  "size": [
+    720,
+    960
+  ]
+}

sentencepiece.bpe.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cb9e3dce4c326195d08fc3dd0f7e2eee1da8595c847bf4c1a9c78b7a82d47e2d
+size 1296245

special_tokens_map.json

@@ -0,0 +1,19 @@
+{
+  "additional_special_tokens": [
+    "<s_iitcdip>",
+    "<s_synthdog>"
+  ],
+  "bos_token": "<s>",
+  "cls_token": "<s>",
+  "eos_token": "</s>",
+  "mask_token": {
+    "content": "<mask>",
+    "lstrip": true,
+    "normalized": true,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": "<pad>",
+  "sep_token": "</s>",
+  "unk_token": "<unk>"
+}

tokenizer_config.json

@@ -0,0 +1,21 @@
+{
+  "bos_token": "<s>",
+  "cls_token": "<s>",
+  "eos_token": "</s>",
+  "mask_token": {
+    "__type": "AddedToken",
+    "content": "<mask>",
+    "lstrip": true,
+    "normalized": true,
+    "rstrip": false,
+    "single_word": false
+  },
+  "model_max_length": 1000000000000000019884624838656,
+  "pad_token": "<pad>",
+  "processor_class": "DonutProcessor",
+  "sep_token": "</s>",
+  "sp_model_kwargs": {},
+  "special_tokens_map_file": null,
+  "tokenizer_class": "XLMRobertaTokenizer",
+  "unk_token": "<unk>"
+}