Initialising processor (#6)

- Initialising processor (b9fb57a44c35005204b7256b503481d3fad4946e)

added_tokens.json

@@ -1,51 +1,5 @@
 {
-  "</s_>": 57566,
-  "</s_cashprice>": 57550,
-  "</s_changeprice>": 57552,
-  "</s_cnt>": 57530,
-  "</s_creditcardprice>": 57564,
-  "</s_discount_price>": 57558,
-  "</s_discountprice>": 57570,
-  "</s_etc>": 57542,
-  "</s_menu>": 57526,
-  "</s_menuqty_cnt>": 57556,
-  "</s_menutype_cnt>": 57554,
-  "</s_nm>": 57528,
-  "</s_price>": 57532,
-  "</s_service_price>": 57538,
-  "</s_sub>": 57548,
-  "</s_sub_total>": 57534,
-  "</s_subtotal_price>": 57536,
-  "</s_tax_price>": 57540,
-  "</s_total>": 57544,
-  "</s_total_etc>": 57562,
-  "</s_total_price>": 57546,
-  "</s_unitprice>": 57560,
-  "</s_vatyn>": 57568,
-  "<s_>": 57565,
-  "<s_cashprice>": 57549,
-  "<s_changeprice>": 57551,
-  "<s_cnt>": 57529,
-  "<s_creditcardprice>": 57563,
-  "<s_discount_price>": 57557,
-  "<s_discountprice>": 57569,
-  "<s_etc>": 57541,
   "<s_iitcdip>": 57523,
-  "<s_menu>": 57525,
-  "<s_menuqty_cnt>": 57555,
-  "<s_menutype_cnt>": 57553,
-  "<s_nm>": 57527,
-  "<s_price>": 57531,
-  "<s_service_price>": 57537,
-  "<s_sub>": 57547,
-  "<s_sub_total>": 57533,
-  "<s_subtotal_price>": 57535,
   "<s_synthdog>": 57524,
-  "<s_tax_price>": 57539,
-  "<s_total>": 57543,
-  "<s_total_etc>": 57561,
-  "<s_total_price>": 57545,
-  "<s_unitprice>": 57559,
-  "<s_vatyn>": 57567,
   "<sep/>": 57522
 }

preprocessor_config.json

@@ -1,5 +1,5 @@
 {
-  "do_align_long_axis": false,
+  "do_align_long_axis": true,
   "do_normalize": true,
   "do_pad": true,
   "do_rescale": true,
@@ -19,8 +19,8 @@
   "processor_class": "DonutProcessor",
   "resample": 2,
   "rescale_factor": 0.00392156862745098,
-  "size": [
-    960,
-    1280
-  ]
+  "size": {
+    "height": 2560,
+    "width": 1920
+  }
 }

tokenizer.json

@@ -1,21 +1,7 @@
 {
   "version": "1.0",
-  "truncation": {
-    "direction": "Right",
-    "max_length": 768,
-    "strategy": "LongestFirst",
-    "stride": 0
-  },
-  "padding": {
-    "strategy": {
-      "Fixed": 768
-    },
-    "direction": "Right",
-    "pad_to_multiple_of": null,
-    "pad_id": 1,
-    "pad_type_id": 0,
-    "pad_token": "<pad>"
-  },
+  "truncation": null,
+  "padding": null,
   "added_tokens": [
     {
       "id": 0,
@@ -88,420 +74,6 @@
       "rstrip": false,
       "normalized": false,
       "special": true
-    },
-    {
-      "id": 57525,
-      "content": "<s_menu>",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": true,
-      "special": false
-    },
-    {
-      "id": 57526,
-      "content": "</s_menu>",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": true,
-      "special": false
-    },
-    {
-      "id": 57527,
-      "content": "<s_nm>",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": true,
-      "special": false
-    },
-    {
-      "id": 57528,
-      "content": "</s_nm>",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": true,
-      "special": false
-    },
-    {
-      "id": 57529,
-      "content": "<s_cnt>",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": true,
-      "special": false
-    },
-    {
-      "id": 57530,
-      "content": "</s_cnt>",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": true,
-      "special": false
-    },
-    {
-      "id": 57531,
-      "content": "<s_price>",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": true,
-      "special": false
-    },
-    {
-      "id": 57532,
-      "content": "</s_price>",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": true,
-      "special": false
-    },
-    {
-      "id": 57533,
-      "content": "<s_sub_total>",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": true,
-      "special": false
-    },
-    {
-      "id": 57534,
-      "content": "</s_sub_total>",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": true,
-      "special": false
-    },
-    {
-      "id": 57535,
-      "content": "<s_subtotal_price>",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": true,
-      "special": false
-    },
-    {
-      "id": 57536,
-      "content": "</s_subtotal_price>",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": true,
-      "special": false
-    },
-    {
-      "id": 57537,
-      "content": "<s_service_price>",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": true,
-      "special": false
-    },
-    {
-      "id": 57538,
-      "content": "</s_service_price>",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": true,
-      "special": false
-    },
-    {
-      "id": 57539,
-      "content": "<s_tax_price>",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": true,
-      "special": false
-    },
-    {
-      "id": 57540,
-      "content": "</s_tax_price>",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": true,
-      "special": false
-    },
-    {
-      "id": 57541,
-      "content": "<s_etc>",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": true,
-      "special": false
-    },
-    {
-      "id": 57542,
-      "content": "</s_etc>",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": true,
-      "special": false
-    },
-    {
-      "id": 57543,
-      "content": "<s_total>",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": true,
-      "special": false
-    },
-    {
-      "id": 57544,
-      "content": "</s_total>",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": true,
-      "special": false
-    },
-    {
-      "id": 57545,
-      "content": "<s_total_price>",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": true,
-      "special": false
-    },
-    {
-      "id": 57546,
-      "content": "</s_total_price>",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": true,
-      "special": false
-    },
-    {
-      "id": 57547,
-      "content": "<s_sub>",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": true,
-      "special": false
-    },
-    {
-      "id": 57548,
-      "content": "</s_sub>",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": true,
-      "special": false
-    },
-    {
-      "id": 57549,
-      "content": "<s_cashprice>",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": true,
-      "special": false
-    },
-    {
-      "id": 57550,
-      "content": "</s_cashprice>",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": true,
-      "special": false
-    },
-    {
-      "id": 57551,
-      "content": "<s_changeprice>",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": true,
-      "special": false
-    },
-    {
-      "id": 57552,
-      "content": "</s_changeprice>",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": true,
-      "special": false
-    },
-    {
-      "id": 57553,
-      "content": "<s_menutype_cnt>",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": true,
-      "special": false
-    },
-    {
-      "id": 57554,
-      "content": "</s_menutype_cnt>",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": true,
-      "special": false
-    },
-    {
-      "id": 57555,
-      "content": "<s_menuqty_cnt>",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": true,
-      "special": false
-    },
-    {
-      "id": 57556,
-      "content": "</s_menuqty_cnt>",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": true,
-      "special": false
-    },
-    {
-      "id": 57557,
-      "content": "<s_discount_price>",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": true,
-      "special": false
-    },
-    {
-      "id": 57558,
-      "content": "</s_discount_price>",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": true,
-      "special": false
-    },
-    {
-      "id": 57559,
-      "content": "<s_unitprice>",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": true,
-      "special": false
-    },
-    {
-      "id": 57560,
-      "content": "</s_unitprice>",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": true,
-      "special": false
-    },
-    {
-      "id": 57561,
-      "content": "<s_total_etc>",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": true,
-      "special": false
-    },
-    {
-      "id": 57562,
-      "content": "</s_total_etc>",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": true,
-      "special": false
-    },
-    {
-      "id": 57563,
-      "content": "<s_creditcardprice>",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": true,
-      "special": false
-    },
-    {
-      "id": 57564,
-      "content": "</s_creditcardprice>",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": true,
-      "special": false
-    },
-    {
-      "id": 57565,
-      "content": "<s_>",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": true,
-      "special": false
-    },
-    {
-      "id": 57566,
-      "content": "</s_>",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": true,
-      "special": false
-    },
-    {
-      "id": 57567,
-      "content": "<s_vatyn>",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": true,
-      "special": false
-    },
-    {
-      "id": 57568,
-      "content": "</s_vatyn>",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": true,
-      "special": false
-    },
-    {
-      "id": 57569,
-      "content": "<s_discountprice>",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": true,
-      "special": false
-    },
-    {
-      "id": 57570,
-      "content": "</s_discountprice>",
-      "single_word": false,
-      "lstrip": false,
-      "rstrip": false,
-      "normalized": true,
-      "special": false
     }
   ],
   "normalizer": {