{
  "_name_or_path": "Salesforce/blip-vqa-base",
  "architectures": [
    "ViltForQuestionAnswering"
  ],
  "attention_probs_dropout_prob": 0.0,
  "hidden_act": "gelu",
  "hidden_dropout_prob": 0.0,
  "hidden_size": 768,
  "id2label": {
    "0": "birthday",
    "1": "walking",
    "2": "talking on phone",
    "3": "woman",
    "4": "white and blue",
    "5": "net",
    "6": "window",
    "7": "women",
    "8": "out",
    "9": "blue and white",
    "10": "wine",
    "11": "small",
    "12": "dirt",
    "13": "king",
    "14": "girl",
    "15": "9:35",
    "16": "snowboarding",
    "17": "full",
    "18": "shrimp",
    "19": "rack",
    "20": "red",
    "21": "red and yellow",
    "22": "not there",
    "23": "necklace",
    "24": "unknown",
    "25": "crown",
    "26": "clear",
    "27": "bikes",
    "28": "right",
    "29": "ground",
    "30": "4",
    "31": "skier",
    "32": "skiing",
    "33": "7:35",
    "34": "down",
    "35": "hat",
    "36": "skateboard",
    "37": "tired",
    "38": "screen",
    "39": "plain",
    "40": "name tag",
    "41": "8",
    "42": "snowboarder",
    "43": "cloudy",
    "44": "zoo",
    "45": "boy",
    "46": "style",
    "47": "0",
    "48": "wine tasting",
    "49": "7",
    "50": "ice cream",
    "51": "smiling",
    "52": "blue",
    "53": "watching",
    "54": "neon",
    "55": "beige",
    "56": "can't tell",
    "57": "bicycle",
    "58": "doughnut",
    "59": "calico",
    "60": "giraffes",
    "61": "10",
    "62": "jeep",
    "63": "6",
    "64": "cage",
    "65": "outside",
    "66": "tower",
    "67": "gray",
    "68": "happy",
    "69": "chair",
    "70": "5",
    "71": "sky",
    "72": "station",
    "73": "lanyard",
    "74": "cat",
    "75": "solid",
    "76": "suv",
    "77": "picnic table",
    "78": "little girl",
    "79": "lying down",
    "80": "plastic",
    "81": "snow",
    "82": "queen",
    "83": "hair",
    "84": "big ben",
    "85": "windows",
    "86": "human",
    "87": "smile",
    "88": "park",
    "89": "in car",
    "90": "canopy",
    "91": "on street",
    "92": "curtain",
    "93": "green",
    "94": "low",
    "95": "white",
    "96": "person",
    "97": "orange",
    "98": "bike rack",
    "99": "french",
    "100": "lg",
    "101": "resting",
    "102": "backpack",
    "103": "bus",
    "104": "roof",
    "105": "church",
    "106": "fence",
    "107": "cup",
    "108": "cross",
    "109": "wall",
    "110": "forest",
    "111": "ball",
    "112": "snowboard",
    "113": "chopsticks",
    "114": "arrow",
    "115": "plate",
    "116": "bricks",
    "117": "clock",
    "118": "wedding",
    "119": "3",
    "120": "black and white",
    "121": "don't know",
    "122": "soccer",
    "123": "lady",
    "124": "platform",
    "125": "train",
    "126": "brick",
    "127": "leather",
    "128": "pink",
    "129": "giraffe",
    "130": "camera",
    "131": "yellow",
    "132": "purple",
    "133": "hawaii",
    "134": "tan",
    "135": "man",
    "136": "no",
    "137": "crossing",
    "138": "blonde",
    "139": "gray and black",
    "140": "africa",
    "141": "door",
    "142": "at table",
    "143": "laying down",
    "144": "2",
    "145": "8:35",
    "146": "table",
    "147": "curtains",
    "148": "photographer",
    "149": "beagle",
    "150": "talking",
    "151": "2010",
    "152": "nothing",
    "153": "1",
    "154": "double",
    "155": "yes",
    "156": "bedroom",
    "157": "2013",
    "158": "street",
    "159": "sidewalk",
    "160": "fashion",
    "161": "tent",
    "162": "dog",
    "163": "shade",
    "164": "shadows",
    "165": "black",
    "166": "exit",
    "167": "red and blue",
    "168": "donut",
    "169": "car",
    "170": "7:45",
    "171": "many",
    "172": "soccer ball",
    "173": "tv",
    "174": "desert",
    "175": "large",
    "176": "monitor",
    "177": "brown",
    "178": "stripes",
    "179": "tabby",
    "180": "sun",
    "181": "not sure",
    "182": "protection",
    "183": "sleeping",
    "184": "clock tower",
    "185": "bicycles",
    "186": "they aren't",
    "187": "woods",
    "188": "2000",
    "189": "on road",
    "190": "stand",
    "191": "shadow",
    "192": "white and black",
    "193": "trees",
    "194": "security",
    "195": "air",
    "196": "natural",
    "197": "shelter",
    "198": "skateboarding"
  },
  "image_size": 384,
  "image_text_hidden_size": 256,
  "initializer_factor": 1.0,
  "initializer_range": 0.02,
  "intermediate_size": 3072,
  "label2id": {
    "0": 47,
    "1": 153,
    "10": 61,
    "2": 144,
    "2000": 188,
    "2010": 151,
    "2013": 157,
    "3": 119,
    "4": 30,
    "5": 70,
    "6": 63,
    "7": 49,
    "7:35": 33,
    "7:45": 170,
    "8": 41,
    "8:35": 145,
    "9:35": 15,
    "africa": 140,
    "air": 195,
    "arrow": 114,
    "at table": 142,
    "backpack": 102,
    "ball": 111,
    "beagle": 149,
    "bedroom": 156,
    "beige": 55,
    "bicycle": 57,
    "bicycles": 185,
    "big ben": 84,
    "bike rack": 98,
    "bikes": 27,
    "birthday": 0,
    "black": 165,
    "black and white": 120,
    "blonde": 138,
    "blue": 52,
    "blue and white": 9,
    "boy": 45,
    "brick": 126,
    "bricks": 116,
    "brown": 177,
    "bus": 103,
    "cage": 64,
    "calico": 59,
    "camera": 130,
    "can't tell": 56,
    "canopy": 90,
    "car": 169,
    "cat": 74,
    "chair": 69,
    "chopsticks": 113,
    "church": 105,
    "clear": 26,
    "clock": 117,
    "clock tower": 184,
    "cloudy": 43,
    "cross": 108,
    "crossing": 137,
    "crown": 25,
    "cup": 107,
    "curtain": 92,
    "curtains": 147,
    "desert": 174,
    "dirt": 12,
    "dog": 162,
    "don't know": 121,
    "donut": 168,
    "door": 141,
    "double": 154,
    "doughnut": 58,
    "down": 34,
    "exit": 166,
    "fashion": 160,
    "fence": 106,
    "forest": 110,
    "french": 99,
    "full": 17,
    "giraffe": 129,
    "giraffes": 60,
    "girl": 14,
    "gray": 67,
    "gray and black": 139,
    "green": 93,
    "ground": 29,
    "hair": 83,
    "happy": 68,
    "hat": 35,
    "hawaii": 133,
    "human": 86,
    "ice cream": 50,
    "in car": 89,
    "jeep": 62,
    "king": 13,
    "lady": 123,
    "lanyard": 73,
    "large": 175,
    "laying down": 143,
    "leather": 127,
    "lg": 100,
    "little girl": 78,
    "low": 94,
    "lying down": 79,
    "man": 135,
    "many": 171,
    "monitor": 176,
    "name tag": 40,
    "natural": 196,
    "necklace": 23,
    "neon": 54,
    "net": 5,
    "no": 136,
    "not sure": 181,
    "not there": 22,
    "nothing": 152,
    "on road": 189,
    "on street": 91,
    "orange": 97,
    "out": 8,
    "outside": 65,
    "park": 88,
    "person": 96,
    "photographer": 148,
    "picnic table": 77,
    "pink": 128,
    "plain": 39,
    "plastic": 80,
    "plate": 115,
    "platform": 124,
    "protection": 182,
    "purple": 132,
    "queen": 82,
    "rack": 19,
    "red": 20,
    "red and blue": 167,
    "red and yellow": 21,
    "resting": 101,
    "right": 28,
    "roof": 104,
    "screen": 38,
    "security": 194,
    "shade": 163,
    "shadow": 191,
    "shadows": 164,
    "shelter": 197,
    "shrimp": 18,
    "sidewalk": 159,
    "skateboard": 36,
    "skateboarding": 198,
    "skier": 31,
    "skiing": 32,
    "sky": 71,
    "sleeping": 183,
    "small": 11,
    "smile": 87,
    "smiling": 51,
    "snow": 81,
    "snowboard": 112,
    "snowboarder": 42,
    "snowboarding": 16,
    "soccer": 122,
    "soccer ball": 172,
    "solid": 75,
    "stand": 190,
    "station": 72,
    "street": 158,
    "stripes": 178,
    "style": 46,
    "sun": 180,
    "suv": 76,
    "tabby": 179,
    "table": 146,
    "talking": 150,
    "talking on phone": 2,
    "tan": 134,
    "tent": 161,
    "they aren't": 186,
    "tired": 37,
    "tower": 66,
    "train": 125,
    "trees": 193,
    "tv": 173,
    "unknown": 24,
    "walking": 1,
    "wall": 109,
    "watching": 53,
    "wedding": 118,
    "white": 95,
    "white and black": 192,
    "white and blue": 4,
    "window": 6,
    "windows": 85,
    "wine": 10,
    "wine tasting": 48,
    "woman": 3,
    "women": 7,
    "woods": 187,
    "yellow": 131,
    "yes": 155,
    "zoo": 44
  },
  "layer_norm_eps": 1e-12,
  "logit_scale_init_value": 2.6592,
  "max_image_length": -1,
  "max_position_embeddings": 40,
  "modality_type_vocab_size": 2,
  "model_type": "vilt",
  "num_attention_heads": 12,
  "num_channels": 3,
  "num_hidden_layers": 12,
  "num_images": -1,
  "patch_size": 32,
  "projection_dim": 512,
  "qkv_bias": true,
  "text_config": {
    "_name_or_path": "",
    "add_cross_attention": false,
    "architectures": null,
    "attention_probs_dropout_prob": 0.0,
    "bad_words_ids": null,
    "begin_suppress_tokens": null,
    "bos_token_id": 30522,
    "chunk_size_feed_forward": 0,
    "cross_attention_hidden_size": null,
    "decoder_start_token_id": null,
    "diversity_penalty": 0.0,
    "do_sample": false,
    "early_stopping": false,
    "encoder_no_repeat_ngram_size": 0,
    "eos_token_id": 2,
    "exponential_decay_length_penalty": null,
    "finetuning_task": null,
    "forced_bos_token_id": null,
    "forced_eos_token_id": null,
    "hidden_act": "gelu",
    "hidden_dropout_prob": 0.0,
    "hidden_size": 768,
    "id2label": {
      "0": "LABEL_0",
      "1": "LABEL_1"
    },
    "initializer_factor": 1.0,
    "initializer_range": 0.02,
    "intermediate_size": 3072,
    "is_decoder": true,
    "is_encoder_decoder": false,
    "label2id": {
      "LABEL_0": 0,
      "LABEL_1": 1
    },
    "layer_norm_eps": 1e-12,
    "length_penalty": 1.0,
    "max_length": 20,
    "max_position_embeddings": 512,
    "min_length": 0,
    "model_type": "blip_text_model",
    "no_repeat_ngram_size": 0,
    "num_attention_heads": 12,
    "num_beam_groups": 1,
    "num_beams": 1,
    "num_hidden_layers": 12,
    "num_return_sequences": 1,
    "output_attentions": false,
    "output_hidden_states": false,
    "output_scores": false,
    "pad_token_id": 0,
    "prefix": null,
    "problem_type": null,
    "projection_dim": 768,
    "pruned_heads": {},
    "remove_invalid_values": false,
    "repetition_penalty": 1.0,
    "return_dict": true,
    "return_dict_in_generate": false,
    "sep_token_id": 102,
    "suppress_tokens": null,
    "task_specific_params": null,
    "temperature": 1.0,
    "tf_legacy_loss": false,
    "tie_encoder_decoder": false,
    "tie_word_embeddings": true,
    "tokenizer_class": null,
    "top_k": 50,
    "top_p": 1.0,
    "torch_dtype": null,
    "torchscript": false,
    "transformers_version": "4.26.0.dev0",
    "typical_p": 1.0,
    "use_bfloat16": false,
    "use_cache": true,
    "vocab_size": 30524
  },
  "tie_word_embeddings": false,
  "torch_dtype": "float32",
  "transformers_version": "4.42.4",
  "type_vocab_size": 2,
  "vision_config": {
    "_name_or_path": "",
    "add_cross_attention": false,
    "architectures": null,
    "attention_dropout": 0.0,
    "bad_words_ids": null,
    "begin_suppress_tokens": null,
    "bos_token_id": null,
    "chunk_size_feed_forward": 0,
    "cross_attention_hidden_size": null,
    "decoder_start_token_id": null,
    "diversity_penalty": 0.0,
    "do_sample": false,
    "dropout": 0.0,
    "early_stopping": false,
    "encoder_no_repeat_ngram_size": 0,
    "eos_token_id": null,
    "exponential_decay_length_penalty": null,
    "finetuning_task": null,
    "forced_bos_token_id": null,
    "forced_eos_token_id": null,
    "hidden_act": "gelu",
    "hidden_size": 768,
    "id2label": {
      "0": "LABEL_0",
      "1": "LABEL_1"
    },
    "image_size": 384,
    "initializer_factor": 1.0,
    "initializer_range": 0.02,
    "intermediate_size": 3072,
    "is_decoder": false,
    "is_encoder_decoder": false,
    "label2id": {
      "LABEL_0": 0,
      "LABEL_1": 1
    },
    "layer_norm_eps": 1e-05,
    "length_penalty": 1.0,
    "max_length": 20,
    "min_length": 0,
    "model_type": "blip_vision_model",
    "no_repeat_ngram_size": 0,
    "num_attention_heads": 12,
    "num_beam_groups": 1,
    "num_beams": 1,
    "num_channels": 3,
    "num_hidden_layers": 12,
    "num_return_sequences": 1,
    "output_attentions": false,
    "output_hidden_states": false,
    "output_scores": false,
    "pad_token_id": null,
    "patch_size": 16,
    "prefix": null,
    "problem_type": null,
    "projection_dim": 512,
    "pruned_heads": {},
    "remove_invalid_values": false,
    "repetition_penalty": 1.0,
    "return_dict": true,
    "return_dict_in_generate": false,
    "sep_token_id": null,
    "suppress_tokens": null,
    "task_specific_params": null,
    "temperature": 1.0,
    "tf_legacy_loss": false,
    "tie_encoder_decoder": false,
    "tie_word_embeddings": true,
    "tokenizer_class": null,
    "top_k": 50,
    "top_p": 1.0,
    "torch_dtype": null,
    "torchscript": false,
    "transformers_version": "4.26.0.dev0",
    "typical_p": 1.0,
    "use_bfloat16": false
  },
  "vocab_size": 30522
}