Training in progress, step 500

config.json

@@ -0,0 +1,249 @@
+{
+  "_name_or_path": "dandelin/vilt-b32-finetuned-vqa",
+  "architectures": [
+    "ViltForQuestionAnswering"
+  ],
+  "attention_probs_dropout_prob": 0.0,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.0,
+  "hidden_size": 768,
+  "id2label": {
+    "0": "cell phone",
+    "1": "Yellow",
+    "2": "snowboard",
+    "3": "clock",
+    "4": "vase",
+    "5": "surfboard",
+    "6": "scissors",
+    "7": "apple",
+    "8": "Green",
+    "9": "umbrella",
+    "10": "book",
+    "11": "Right",
+    "12": "orange",
+    "13": "left",
+    "14": "sink",
+    "15": "cup",
+    "16": "toothbrush",
+    "17": "kite",
+    "18": "bench",
+    "19": "mouse",
+    "20": "bicycle",
+    "21": "No",
+    "22": "tie",
+    "23": "carrot",
+    "24": "elephant",
+    "25": "dog",
+    "26": "1",
+    "27": "knife",
+    "28": "3",
+    "29": "chair",
+    "30": "remote",
+    "31": "baseball bat",
+    "32": "right",
+    "33": "car",
+    "34": "above",
+    "35": "sheep",
+    "36": "4",
+    "37": "tv",
+    "38": "White",
+    "39": "skateboard",
+    "40": "Brown",
+    "41": "zebra",
+    "42": "pizza",
+    "43": "Blue",
+    "44": "Left",
+    "45": "parking meter",
+    "46": "Behind",
+    "47": "fork",
+    "48": "hot dog",
+    "49": "suitcase",
+    "50": "Below",
+    "51": "stop sign",
+    "52": "skis",
+    "53": "Above",
+    "54": "cat",
+    "55": "Black",
+    "56": "frisbee",
+    "57": "Red",
+    "58": "Front",
+    "59": "person",
+    "60": "8",
+    "61": "laptop",
+    "62": "refrigerator",
+    "63": "dining table",
+    "64": "bowl",
+    "65": "bus",
+    "66": "banana",
+    "67": "bottle",
+    "68": "sandwich",
+    "69": "cake",
+    "70": "Orange",
+    "71": "bird",
+    "72": "Grey",
+    "73": "6",
+    "74": "bear",
+    "75": "traffic light",
+    "76": "7",
+    "77": "horse",
+    "78": "giraffe",
+    "79": "train",
+    "80": "Yes",
+    "81": "9",
+    "82": "tennis racket",
+    "83": "couch",
+    "84": "cow",
+    "85": "handbag",
+    "86": "wine glass",
+    "87": "toilet",
+    "88": "fire hydrant",
+    "89": "bed",
+    "90": "keyboard",
+    "91": "baseball glove",
+    "92": "boat",
+    "93": "motorcycle",
+    "94": "2",
+    "95": "broccoli",
+    "96": "truck",
+    "97": "Pink",
+    "98": "teddy bear",
+    "99": "sports ball",
+    "100": "oven",
+    "101": "spoon",
+    "102": "Purple",
+    "103": "donut",
+    "104": "0",
+    "105": "potted plant",
+    "106": "airplane",
+    "107": "5"
+  },
+  "image_size": 384,
+  "initializer_range": 0.02,
+  "intermediate_size": 3072,
+  "label2id": {
+    "0": 104,
+    "1": 26,
+    "2": 94,
+    "3": 28,
+    "4": 36,
+    "5": 107,
+    "6": 73,
+    "7": 76,
+    "8": 60,
+    "9": 81,
+    "Above": 53,
+    "Behind": 46,
+    "Below": 50,
+    "Black": 55,
+    "Blue": 43,
+    "Brown": 40,
+    "Front": 58,
+    "Green": 8,
+    "Grey": 72,
+    "Left": 44,
+    "No": 21,
+    "Orange": 70,
+    "Pink": 97,
+    "Purple": 102,
+    "Red": 57,
+    "Right": 11,
+    "White": 38,
+    "Yellow": 1,
+    "Yes": 80,
+    "above": 34,
+    "airplane": 106,
+    "apple": 7,
+    "banana": 66,
+    "baseball bat": 31,
+    "baseball glove": 91,
+    "bear": 74,
+    "bed": 89,
+    "bench": 18,
+    "bicycle": 20,
+    "bird": 71,
+    "boat": 92,
+    "book": 10,
+    "bottle": 67,
+    "bowl": 64,
+    "broccoli": 95,
+    "bus": 65,
+    "cake": 69,
+    "car": 33,
+    "carrot": 23,
+    "cat": 54,
+    "cell phone": 0,
+    "chair": 29,
+    "clock": 3,
+    "couch": 83,
+    "cow": 84,
+    "cup": 15,
+    "dining table": 63,
+    "dog": 25,
+    "donut": 103,
+    "elephant": 24,
+    "fire hydrant": 88,
+    "fork": 47,
+    "frisbee": 56,
+    "giraffe": 78,
+    "handbag": 85,
+    "horse": 77,
+    "hot dog": 48,
+    "keyboard": 90,
+    "kite": 17,
+    "knife": 27,
+    "laptop": 61,
+    "left": 13,
+    "motorcycle": 93,
+    "mouse": 19,
+    "orange": 12,
+    "oven": 100,
+    "parking meter": 45,
+    "person": 59,
+    "pizza": 42,
+    "potted plant": 105,
+    "refrigerator": 62,
+    "remote": 30,
+    "right": 32,
+    "sandwich": 68,
+    "scissors": 6,
+    "sheep": 35,
+    "sink": 14,
+    "skateboard": 39,
+    "skis": 52,
+    "snowboard": 2,
+    "spoon": 101,
+    "sports ball": 99,
+    "stop sign": 51,
+    "suitcase": 49,
+    "surfboard": 5,
+    "teddy bear": 98,
+    "tennis racket": 82,
+    "tie": 22,
+    "toilet": 87,
+    "toothbrush": 16,
+    "traffic light": 75,
+    "train": 79,
+    "truck": 96,
+    "tv": 37,
+    "umbrella": 9,
+    "vase": 4,
+    "wine glass": 86,
+    "zebra": 41
+  },
+  "layer_norm_eps": 1e-12,
+  "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,
+  "qkv_bias": true,
+  "tie_word_embeddings": false,
+  "torch_dtype": "float32",
+  "transformers_version": "4.47.0",
+  "type_vocab_size": 2,
+  "vocab_size": 30522
+}

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5642ec9a8e90f91b5fd9583775c111b8623aa096cf3f6c0e23edcdf4e89de504
+size 451805856

preprocessor_config.json

@@ -0,0 +1,24 @@
+{
+  "do_normalize": true,
+  "do_pad": true,
+  "do_rescale": true,
+  "do_resize": true,
+  "image_mean": [
+    0.5,
+    0.5,
+    0.5
+  ],
+  "image_processor_type": "ViltImageProcessor",
+  "image_std": [
+    0.5,
+    0.5,
+    0.5
+  ],
+  "processor_class": "ViltProcessor",
+  "resample": 3,
+  "rescale_factor": 0.00392156862745098,
+  "size": {
+    "shortest_edge": 384
+  },
+  "size_divisor": 32
+}

special_tokens_map.json

@@ -0,0 +1,37 @@
+{
+  "cls_token": {
+    "content": "[CLS]",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "mask_token": {
+    "content": "[MASK]",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "[PAD]",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "sep_token": {
+    "content": "[SEP]",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "unk_token": {
+    "content": "[UNK]",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

tokenizer_config.json

@@ -0,0 +1,57 @@
+{
+  "added_tokens_decoder": {
+    "0": {
+      "content": "[PAD]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "100": {
+      "content": "[UNK]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "101": {
+      "content": "[CLS]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "102": {
+      "content": "[SEP]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "103": {
+      "content": "[MASK]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "clean_up_tokenization_spaces": false,
+  "cls_token": "[CLS]",
+  "do_lower_case": true,
+  "extra_special_tokens": {},
+  "mask_token": "[MASK]",
+  "model_max_length": 40,
+  "pad_token": "[PAD]",
+  "processor_class": "ViltProcessor",
+  "sep_token": "[SEP]",
+  "strip_accents": null,
+  "tokenize_chinese_chars": true,
+  "tokenizer_class": "BertTokenizer",
+  "unk_token": "[UNK]"
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c89b04b295735be2c0d1127f48d8ec444cf54f54a9574c6666562698c49102b9
+size 5368