Update tokenizer_config.json

tokenizer_config.json

@@ -3,22 +3,22 @@
   "add_eos_token": false,
   "added_tokens_decoder": {
     "0": {
-      "content": "<|BOS|>",
+      "content": "<|pad|>",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
       "single_word": false,
       "special": true
     },
-    "0": {
-      "content": "<|pad|>",
+    "1": {
+      "content": "<|unk|>",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
       "single_word": false,
       "special": true
     },
-    "0": {
+    "2": {
       "content": "<|im_start|>",
       "lstrip": false,
       "normalized": false,
@@ -26,7 +26,7 @@
       "single_word": false,
       "special": true
     },
-    "0": {
+    "3": {
       "content": "<|im_sep|>",
       "lstrip": false,
       "normalized": false,
@@ -34,7 +34,7 @@
       "single_word": false,
       "special": true
     },
-    "0": {
+    "4": {
       "content": "<|im_end|>",
       "lstrip": false,
       "normalized": false,
@@ -43,12 +43,13 @@
       "special": true
     }
   },
-  "bos_token": "<|BOS|>",
+  "bos_token": "<|im_start|>",
   "clean_up_tokenization_spaces": true,
   "eos_token": "<|im_end|>",
+  "unk_token": "<|unk|>"
   "extra_special_tokens": {},
   "legacy": false,
-  "max_length": 1024,
+  "max_length": 256,
   "pad_token": "<|pad|>",
   "sp_model_kwargs": {},
   "spaces_between_special_tokens": false,