End of training

README.md

@@ -40,7 +40,7 @@ The following hyperparameters were used during training:
 - seed: 42
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: linear
-- num_epochs: 3
+- num_epochs: 6
 
 ### Training results
 

added_tokens.json

@@ -1,24 +1,10 @@
 {
   "</s>": 2,
-  "</s_address>": 57527,
-  "</s_company>": 57532,
-  "</s_date>": 57526,
-  "</s_description>": 57538,
-  "</s_item no>": 57535,
-  "</s_total amount>": 57528,
-  "</s_total>": 57529,
   "<mask>": 57521,
   "<pad>": 1,
   "<s>": 0,
-  "<s_address>": 57534,
-  "<s_company>": 57537,
-  "<s_date>": 57525,
-  "<s_description>": 57533,
   "<s_iitcdip>": 57523,
-  "<s_item no>": 57531,
   "<s_synthdog>": 57524,
-  "<s_total amount>": 57530,
-  "<s_total>": 57536,
   "<sep/>": 57522,
   "<unk>": 3
 }

special_tokens_map.json

@@ -2,22 +2,8 @@
   "additional_special_tokens": [
     "<s_iitcdip>",
     "<s_synthdog>",
-    "<s_date>",
     "</s>",
-    "</s_date>",
-    "<s>",
-    "</s_address>",
-    "</s_total amount>",
-    "</s_total>",
-    "<s_total amount>",
-    "<s_item no>",
-    "</s_company>",
-    "<s_description>",
-    "<s_address>",
-    "</s_item no>",
-    "<s_total>",
-    "<s_company>",
-    "</s_description>"
+    "<s>"
   ],
   "bos_token": "<s>",
   "cls_token": "<s>",

tokenizer.json

@@ -88,132 +88,6 @@
       "rstrip": true,
       "normalized": false,
       "special": true
-    },
-    {
-      "id": 57525,
-      "content": "<s_date>",
-      "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
-      "normalized": false,
-      "special": true
-    },
-    {
-      "id": 57526,
-      "content": "</s_date>",
-      "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
-      "normalized": false,
-      "special": true
-    },
-    {
-      "id": 57527,
-      "content": "</s_address>",
-      "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
-      "normalized": false,
-      "special": true
-    },
-    {
-      "id": 57528,
-      "content": "</s_total amount>",
-      "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
-      "normalized": false,
-      "special": true
-    },
-    {
-      "id": 57529,
-      "content": "</s_total>",
-      "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
-      "normalized": false,
-      "special": true
-    },
-    {
-      "id": 57530,
-      "content": "<s_total amount>",
-      "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
-      "normalized": false,
-      "special": true
-    },
-    {
-      "id": 57531,
-      "content": "<s_item no>",
-      "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
-      "normalized": false,
-      "special": true
-    },
-    {
-      "id": 57532,
-      "content": "</s_company>",
-      "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
-      "normalized": false,
-      "special": true
-    },
-    {
-      "id": 57533,
-      "content": "<s_description>",
-      "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
-      "normalized": false,
-      "special": true
-    },
-    {
-      "id": 57534,
-      "content": "<s_address>",
-      "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
-      "normalized": false,
-      "special": true
-    },
-    {
-      "id": 57535,
-      "content": "</s_item no>",
-      "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
-      "normalized": false,
-      "special": true
-    },
-    {
-      "id": 57536,
-      "content": "<s_total>",
-      "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
-      "normalized": false,
-      "special": true
-    },
-    {
-      "id": 57537,
-      "content": "<s_company>",
-      "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
-      "normalized": false,
-      "special": true
-    },
-    {
-      "id": 57538,
-      "content": "</s_description>",
-      "single_word": false,
-      "lstrip": true,
-      "rstrip": true,
-      "normalized": false,
-      "special": true
     }
   ],
   "normalizer": {

tokenizer_config.json

@@ -63,139 +63,13 @@
       "rstrip": true,
       "single_word": false,
       "special": true
-    },
-    "57525": {
-      "content": "<s_date>",
-      "lstrip": true,
-      "normalized": false,
-      "rstrip": true,
-      "single_word": false,
-      "special": true
-    },
-    "57526": {
-      "content": "</s_date>",
-      "lstrip": true,
-      "normalized": false,
-      "rstrip": true,
-      "single_word": false,
-      "special": true
-    },
-    "57527": {
-      "content": "</s_address>",
-      "lstrip": true,
-      "normalized": false,
-      "rstrip": true,
-      "single_word": false,
-      "special": true
-    },
-    "57528": {
-      "content": "</s_total amount>",
-      "lstrip": true,
-      "normalized": false,
-      "rstrip": true,
-      "single_word": false,
-      "special": true
-    },
-    "57529": {
-      "content": "</s_total>",
-      "lstrip": true,
-      "normalized": false,
-      "rstrip": true,
-      "single_word": false,
-      "special": true
-    },
-    "57530": {
-      "content": "<s_total amount>",
-      "lstrip": true,
-      "normalized": false,
-      "rstrip": true,
-      "single_word": false,
-      "special": true
-    },
-    "57531": {
-      "content": "<s_item no>",
-      "lstrip": true,
-      "normalized": false,
-      "rstrip": true,
-      "single_word": false,
-      "special": true
-    },
-    "57532": {
-      "content": "</s_company>",
-      "lstrip": true,
-      "normalized": false,
-      "rstrip": true,
-      "single_word": false,
-      "special": true
-    },
-    "57533": {
-      "content": "<s_description>",
-      "lstrip": true,
-      "normalized": false,
-      "rstrip": true,
-      "single_word": false,
-      "special": true
-    },
-    "57534": {
-      "content": "<s_address>",
-      "lstrip": true,
-      "normalized": false,
-      "rstrip": true,
-      "single_word": false,
-      "special": true
-    },
-    "57535": {
-      "content": "</s_item no>",
-      "lstrip": true,
-      "normalized": false,
-      "rstrip": true,
-      "single_word": false,
-      "special": true
-    },
-    "57536": {
-      "content": "<s_total>",
-      "lstrip": true,
-      "normalized": false,
-      "rstrip": true,
-      "single_word": false,
-      "special": true
-    },
-    "57537": {
-      "content": "<s_company>",
-      "lstrip": true,
-      "normalized": false,
-      "rstrip": true,
-      "single_word": false,
-      "special": true
-    },
-    "57538": {
-      "content": "</s_description>",
-      "lstrip": true,
-      "normalized": false,
-      "rstrip": true,
-      "single_word": false,
-      "special": true
     }
   },
   "additional_special_tokens": [
     "<s_iitcdip>",
     "<s_synthdog>",
-    "<s_date>",
     "</s>",
-    "</s_date>",
-    "<s>",
-    "</s_address>",
-    "</s_total amount>",
-    "</s_total>",
-    "<s_total amount>",
-    "<s_item no>",
-    "</s_company>",
-    "<s_description>",
-    "<s_address>",
-    "</s_item no>",
-    "<s_total>",
-    "<s_company>",
-    "</s_description>"
+    "<s>"
   ],
   "bos_token": "<s>",
   "clean_up_tokenization_spaces": true,