End of training

README.md

@@ -40,7 +40,6 @@ The following hyperparameters were used during training:
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: linear
 - num_epochs: 3
-- mixed_precision_training: Native AMP
 
 ### Training results
 
@@ -48,7 +47,7 @@ The following hyperparameters were used during training:
 
 ### Framework versions
 
-- Transformers 4.23.1
-- Pytorch 1.12.1+cu113
-- Datasets 2.6.1
-- Tokenizers 0.13.1
+- Transformers 4.26.0.dev0
+- Pytorch 1.13.0+cu116
+- Datasets 2.8.0
+- Tokenizers 0.13.2

added_tokens.json

@@ -1,19 +1,25 @@
 {
-  "</s_address>": 57585,
+  "</s_address>": 57599,
+  "</s_cash>": 57583,
   "</s_cashprice>": 57549,
+  "</s_change>": 57581,
   "</s_changeprice>": 57551,
   "</s_cnt>": 57529,
-  "</s_company>": 57583,
+  "</s_company>": 57597,
   "</s_creditcardprice>": 57563,
-  "</s_date>": 57581,
+  "</s_date>": 57595,
   "</s_discount_price>": 57557,
   "</s_discountprice>": 57567,
+  "</s_email>": 57593,
   "</s_emoneyprice>": 57569,
   "</s_etc>": 57541,
+  "</s_fax>": 57591,
   "</s_itemsubtotal>": 57577,
+  "</s_lineNumber>": 57589,
   "</s_menu>": 57525,
   "</s_menuqty_cnt>": 57555,
   "</s_menutype_cnt>": 57553,
+  "</s_merchant>": 57587,
   "</s_nm>": 57527,
   "</s_num>": 57565,
   "</s_othersvc_price>": 57573,
@@ -23,29 +29,36 @@
   "</s_sub_total>": 57533,
   "</s_subtotal_price>": 57535,
   "</s_tax_price>": 57539,
+  "</s_tel>": 57585,
   "</s_total>": 57543,
   "</s_total_etc>": 57561,
   "</s_total_price>": 57545,
   "</s_unitprice>": 57559,
   "</s_vatyn>": 57575,
   "</s_void_menu>": 57571,
-  "<s_address>": 57584,
+  "<s_address>": 57598,
+  "<s_cash>": 57582,
   "<s_cashprice>": 57550,
+  "<s_change>": 57580,
   "<s_changeprice>": 57552,
   "<s_cnt>": 57530,
-  "<s_company>": 57582,
+  "<s_company>": 57596,
   "<s_cord-v2>": 57579,
   "<s_creditcardprice>": 57564,
-  "<s_date>": 57580,
+  "<s_date>": 57594,
   "<s_discount_price>": 57558,
   "<s_discountprice>": 57568,
+  "<s_email>": 57592,
   "<s_emoneyprice>": 57570,
   "<s_etc>": 57542,
+  "<s_fax>": 57590,
   "<s_iitcdip>": 57523,
   "<s_itemsubtotal>": 57578,
+  "<s_lineNumber>": 57588,
   "<s_menu>": 57526,
   "<s_menuqty_cnt>": 57556,
   "<s_menutype_cnt>": 57554,
+  "<s_merchant>": 57586,
   "<s_nm>": 57528,
   "<s_num>": 57566,
   "<s_othersvc_price>": 57574,
@@ -56,6 +69,7 @@
   "<s_subtotal_price>": 57536,
   "<s_synthdog>": 57524,
   "<s_tax_price>": 57540,
+  "<s_tel>": 57584,
   "<s_total>": 57544,
   "<s_total_etc>": 57562,
   "<s_total_price>": 57546,

preprocessor_config.json

@@ -2,14 +2,15 @@
   "do_align_long_axis": false,
   "do_normalize": true,
   "do_pad": true,
+  "do_rescale": true,
   "do_resize": true,
   "do_thumbnail": true,
-  "feature_extractor_type": "DonutFeatureExtractor",
   "image_mean": [
     0.5,
     0.5,
     0.5
   ],
+  "image_processor_type": "DonutImageProcessor",
   "image_std": [
     0.5,
     0.5,
@@ -17,6 +18,7 @@
   ],
   "processor_class": "DonutProcessor",
   "resample": 2,
+  "rescale_factor": 0.00392156862745098,
   "size": [
     720,
     960

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f968e50c4d6d10032f1dbb4ab8d224a298e5049ba06790ed65220ba9ecfbf69c
+oid sha256:eeff45a90481fda3e2dea66c3fb498065486beb524d2ddd3ec39f71e1320a429
 size 806333657

runs/Jan02_17-53-43_0c15e67d1a0b/events.out.tfevents.1672682410.0c15e67d1a0b.879.0

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:23ae97becf7788d35209fe081fa91c7436d576d6428e92a73c79e8ae86f65e11
-size 8710
+oid sha256:102af906ba35c84847edbc5f0dcfe93f472980170de53f4f9d23e70aaec065ec
+size 10006

special_tokens_map.json

@@ -4,24 +4,38 @@
     "</s_total>",
     "<s_total_price>",
     "</s_total_price>",
-    "<s_changeprice>",
-    "</s_changeprice>",
-    "<s_cashprice>",
-    "</s_cashprice>",
+    "<s_price>",
+    "</s_price>",
+    "<s_change>",
+    "</s_change>",
+    "<s_cash>",
+    "</s_cash>",
+    "<s_tel>",
+    "</s_tel>",
     "<s_sub_total>",
     "</s_sub_total>",
-    "<s_subtotal_price>",
-    "</s_subtotal_price>",
+    "<s_tax_price>",
+    "</s_tax_price>",
+    "<s_service_price>",
+    "</s_service_price>",
+    "<s_etc>",
+    "</s_etc>",
+    "<s_discount_price>",
+    "</s_discount_price>",
+    "<s_merchant>",
+    "</s_merchant>",
     "<s_menu>",
     "</s_menu>",
-    "<s_unitprice>",
-    "</s_unitprice>",
-    "<s_price>",
-    "</s_price>",
     "<s_nm>",
     "</s_nm>",
+    "<s_lineNumber>",
+    "</s_lineNumber>",
     "<s_cnt>",
     "</s_cnt>",
+    "<s_fax>",
+    "</s_fax>",
+    "<s_email>",
+    "</s_email>",
     "<s_date>",
     "</s_date>",
     "<s_company>",

tokenizer.json

@@ -185,8 +185,8 @@
       "single_word": false,
       "lstrip": false,
       "rstrip": false,
-      "normalized": false,
-      "special": true
+      "normalized": true,
+      "special": false
     },
     {
       "id": 57536,
@@ -194,8 +194,8 @@
       "single_word": false,
       "lstrip": false,
       "rstrip": false,
-      "normalized": false,
-      "special": true
+      "normalized": true,
+      "special": false
     },
     {
       "id": 57537,
@@ -203,8 +203,8 @@
       "single_word": false,
       "lstrip": false,
       "rstrip": false,
-      "normalized": true,
-      "special": false
+      "normalized": false,
+      "special": true
     },
     {
       "id": 57538,
@@ -212,8 +212,8 @@
       "single_word": false,
       "lstrip": false,
       "rstrip": false,
-      "normalized": true,
-      "special": false
+      "normalized": false,
+      "special": true
     },
     {
       "id": 57539,
@@ -221,8 +221,8 @@
       "single_word": false,
       "lstrip": false,
       "rstrip": false,
-      "normalized": true,
-      "special": false
+      "normalized": false,
+      "special": true
     },
     {
       "id": 57540,
@@ -230,8 +230,8 @@
       "single_word": false,
       "lstrip": false,
       "rstrip": false,
-      "normalized": true,
-      "special": false
+      "normalized": false,
+      "special": true
     },
     {
       "id": 57541,
@@ -239,8 +239,8 @@
       "single_word": false,
       "lstrip": false,
       "rstrip": false,
-      "normalized": true,
-      "special": false
+      "normalized": false,
+      "special": true
     },
     {
       "id": 57542,
@@ -248,8 +248,8 @@
       "single_word": false,
       "lstrip": false,
       "rstrip": false,
-      "normalized": true,
-      "special": false
+      "normalized": false,
+      "special": true
     },
     {
       "id": 57543,
@@ -311,8 +311,8 @@
       "single_word": false,
       "lstrip": false,
       "rstrip": false,
-      "normalized": false,
-      "special": true
+      "normalized": true,
+      "special": false
     },
     {
       "id": 57550,
@@ -320,8 +320,8 @@
       "single_word": false,
       "lstrip": false,
       "rstrip": false,
-      "normalized": false,
-      "special": true
+      "normalized": true,
+      "special": false
     },
     {
       "id": 57551,
@@ -329,8 +329,8 @@
       "single_word": false,
       "lstrip": false,
       "rstrip": false,
-      "normalized": false,
-      "special": true
+      "normalized": true,
+      "special": false
     },
     {
       "id": 57552,
@@ -338,8 +338,8 @@
       "single_word": false,
       "lstrip": false,
       "rstrip": false,
-      "normalized": false,
-      "special": true
+      "normalized": true,
+      "special": false
     },
     {
       "id": 57553,
@@ -383,8 +383,8 @@
       "single_word": false,
       "lstrip": false,
       "rstrip": false,
-      "normalized": true,
-      "special": false
+      "normalized": false,
+      "special": true
     },
     {
       "id": 57558,
@@ -392,8 +392,8 @@
       "single_word": false,
       "lstrip": false,
       "rstrip": false,
-      "normalized": true,
-      "special": false
+      "normalized": false,
+      "special": true
     },
     {
       "id": 57559,
@@ -401,8 +401,8 @@
       "single_word": false,
       "lstrip": false,
       "rstrip": false,
-      "normalized": false,
-      "special": true
+      "normalized": true,
+      "special": false
     },
     {
       "id": 57560,
@@ -410,8 +410,8 @@
       "single_word": false,
       "lstrip": false,
       "rstrip": false,
-      "normalized": false,
-      "special": true
+      "normalized": true,
+      "special": false
     },
     {
       "id": 57561,
@@ -586,7 +586,7 @@
     },
     {
       "id": 57580,
-      "content": "<s_date>",
+      "content": "<s_change>",
       "single_word": false,
       "lstrip": false,
       "rstrip": false,
@@ -595,7 +595,7 @@
     },
     {
       "id": 57581,
-      "content": "</s_date>",
+      "content": "</s_change>",
       "single_word": false,
       "lstrip": false,
       "rstrip": false,
@@ -604,7 +604,7 @@
     },
     {
       "id": 57582,
-      "content": "<s_company>",
+      "content": "<s_cash>",
       "single_word": false,
       "lstrip": false,
       "rstrip": false,
@@ -613,7 +613,7 @@
     },
     {
       "id": 57583,
-      "content": "</s_company>",
+      "content": "</s_cash>",
       "single_word": false,
       "lstrip": false,
       "rstrip": false,
@@ -622,7 +622,7 @@
     },
     {
       "id": 57584,
-      "content": "<s_address>",
+      "content": "<s_tel>",
       "single_word": false,
       "lstrip": false,
       "rstrip": false,
@@ -631,6 +631,132 @@
     },
     {
       "id": 57585,
+      "content": "</s_tel>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": true
+    },
+    {
+      "id": 57586,
+      "content": "<s_merchant>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": true
+    },
+    {
+      "id": 57587,
+      "content": "</s_merchant>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": true
+    },
+    {
+      "id": 57588,
+      "content": "<s_lineNumber>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": true
+    },
+    {
+      "id": 57589,
+      "content": "</s_lineNumber>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": true
+    },
+    {
+      "id": 57590,
+      "content": "<s_fax>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": true
+    },
+    {
+      "id": 57591,
+      "content": "</s_fax>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": true
+    },
+    {
+      "id": 57592,
+      "content": "<s_email>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": true
+    },
+    {
+      "id": 57593,
+      "content": "</s_email>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": true
+    },
+    {
+      "id": 57594,
+      "content": "<s_date>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": true
+    },
+    {
+      "id": 57595,
+      "content": "</s_date>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": true
+    },
+    {
+      "id": 57596,
+      "content": "<s_company>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": true
+    },
+    {
+      "id": 57597,
+      "content": "</s_company>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": true
+    },
+    {
+      "id": 57598,
+      "content": "<s_address>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": true
+    },
+    {
+      "id": 57599,
       "content": "</s_address>",
       "single_word": false,
       "lstrip": false,

tokenizer_config.json

@@ -10,6 +10,7 @@
     "rstrip": false,
     "single_word": false
   },
+  "model_max_length": 1000000000000000019884624838656,
   "name_or_path": "naver-clova-ix/donut-base-finetuned-cord-v2",
   "pad_token": "<pad>",
   "processor_class": "DonutProcessor",