Upload tokenizer

added_tokens.json

@@ -1,8 +1,5 @@
 {
   "<im_end>": 32002,
   "<im_patch>": 32000,
-  "<im_start>": 32001,
-  "<vid_end>": 32005,
-  "<vid_patch>": 32003,
-  "<vid_start>": 32004
+  "<im_start>": 32001
 }

special_tokens_map.json

@@ -1,27 +1,4 @@
 {
-  "additional_special_tokens": [
-    {
-      "content": "<vid_patch>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false
-    },
-    {
-      "content": "<vid_start>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false
-    },
-    {
-      "content": "<vid_end>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false
-    }
-  ],
   "bos_token": {
     "content": "<s>",
     "lstrip": false,
@@ -36,13 +13,7 @@
     "rstrip": false,
     "single_word": false
   },
-  "pad_token": {
-    "content": "<unk>",
-    "lstrip": false,
-    "normalized": false,
-    "rstrip": false,
-    "single_word": false
-  },
+  "pad_token": "</s>",
   "unk_token": {
     "content": "<unk>",
     "lstrip": false,

tokenizer_config.json

@@ -49,44 +49,14 @@
       "rstrip": false,
       "single_word": false,
       "special": false
-    },
-    "32003": {
-      "content": "<vid_patch>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "32004": {
-      "content": "<vid_start>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "32005": {
-      "content": "<vid_end>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
     }
   },
-  "additional_special_tokens": [
-    "<vid_patch>",
-    "<vid_start>",
-    "<vid_end>"
-  ],
   "bos_token": "<s>",
   "clean_up_tokenization_spaces": false,
   "eos_token": "</s>",
   "legacy": true,
-  "max_length": 64,
-  "model_max_length": 1048,
-  "pad_token": "<unk>",
+  "model_max_length": 2048,
+  "pad_token": "</s>",
   "padding_side": "right",
   "sp_model_kwargs": {},
   "spaces_between_special_tokens": false,