add video index

added_tokens.json

@@ -1,3 +1,4 @@
 {
-  "<image>": 32100
+  "<image>": 32100,
+  "<video>": 32101
 }

config.json

@@ -1,5 +1,4 @@
 {
-  "_name_or_path": "/raid/raushan/instructblip-flan-t5-xl/",
   "architectures": [
     "InstructBlipForConditionalGeneration"
   ],
@@ -79,6 +78,7 @@
   "torch_dtype": "float32",
   "transformers_version": "4.47.0.dev0",
   "use_decoder_only_language_model": false,
+  "video_token_index": 32101,
   "vision_config": {
     "_attn_implementation_autoset": true,
     "dropout": 0.0,

qformer_tokenizer/tokenizer_config.json

@@ -53,6 +53,7 @@
   "clean_up_tokenization_spaces": true,
   "cls_token": "[CLS]",
   "do_lower_case": true,
+  "extra_special_tokens": {},
   "mask_token": "[MASK]",
   "model_max_length": 512,
   "pad_token": "[PAD]",

tokenizer.json

@@ -938,6 +938,15 @@
       "rstrip": false,
       "normalized": false,
       "special": true
+    },
+    {
+      "id": 32101,
+      "content": "<video>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": true
     }
   ],
   "normalizer": {

tokenizer_config.json

@@ -831,6 +831,14 @@
       "rstrip": false,
       "single_word": false,
       "special": true
+    },
+    "32101": {
+      "content": "<video>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
     }
   },
   "additional_special_tokens": [
@@ -938,6 +946,7 @@
   "clean_up_tokenization_spaces": true,
   "eos_token": "</s>",
   "extra_ids": 100,
+  "extra_special_tokens": {},
   "model_max_length": 512,
   "pad_token": "<pad>",
   "processor_class": "InstructBlipProcessor",