Training in progress, epoch 0

preprocessor_config.json

@@ -1,4 +1,22 @@
 {
+  "_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,

tokenizer.json

@@ -303,8 +303,8 @@
   "pre_tokenizer": {
     "type": "Metaspace",
     "replacement": "▁",
-    "add_prefix_space": true,
-    "prepend_scheme": "always"
+    "prepend_scheme": "always",
+    "split": true
   },
   "post_processor": {
     "type": "TemplateProcessing",
@@ -390,8 +390,8 @@
   "decoder": {
     "type": "Metaspace",
     "replacement": "▁",
-    "add_prefix_space": true,
-    "prepend_scheme": "always"
+    "prepend_scheme": "always",
+    "split": true
   },
   "model": {
     "type": "Unigram",