Upload processor

added_tokens.json

@@ -1,5 +1,10 @@
 {
+  "</s_carburant>": 57526,
+  "</s_volume>": 57528,
+  "<s_carburant>": 57525,
+  "<s_fuel-v2>": 57529,
   "<s_iitcdip>": 57523,
   "<s_synthdog>": 57524,
+  "<s_volume>": 57527,
   "<sep/>": 57522
 }

preprocessor_config.json

@@ -1,5 +1,23 @@
 {
-  "do_align_long_axis": true,
+  "_valid_processor_keys": [
+    "images",
+    "do_resize",
+    "size",
+    "resample",
+    "do_thumbnail",
+    "do_align_long_axis",
+    "do_pad",
+    "random_padding",
+    "do_rescale",
+    "rescale_factor",
+    "do_normalize",
+    "image_mean",
+    "image_std",
+    "return_tensors",
+    "data_format",
+    "input_data_format"
+  ],
+  "do_align_long_axis": false,
   "do_normalize": true,
   "do_pad": true,
   "do_rescale": true,
@@ -19,8 +37,8 @@
   "processor_class": "DonutProcessor",
   "resample": 2,
   "rescale_factor": 0.00392156862745098,
-  "size": {
-    "height": 2560,
-    "width": 1920
-  }
+  "size": [
+    960,
+    1280
+  ]
 }

tokenizer.json

@@ -74,6 +74,51 @@
       "rstrip": false,
       "normalized": false,
       "special": true
+    },
+    {
+      "id": 57525,
+      "content": "<s_carburant>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": true,
+      "special": false
+    },
+    {
+      "id": 57526,
+      "content": "</s_carburant>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": true,
+      "special": false
+    },
+    {
+      "id": 57527,
+      "content": "<s_volume>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": true,
+      "special": false
+    },
+    {
+      "id": 57528,
+      "content": "</s_volume>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": true,
+      "special": false
+    },
+    {
+      "id": 57529,
+      "content": "<s_fuel-v2>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": true,
+      "special": false
     }
   ],
   "normalizer": {

tokenizer_config.json

@@ -63,6 +63,46 @@
       "rstrip": false,
       "single_word": false,
       "special": true
+    },
+    "57525": {
+      "content": "<s_carburant>",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "57526": {
+      "content": "</s_carburant>",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "57527": {
+      "content": "<s_volume>",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "57528": {
+      "content": "</s_volume>",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "57529": {
+      "content": "<s_fuel-v2>",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
     }
   },
   "additional_special_tokens": [