doodle-zero

README.md

@@ -1,6 +1,4 @@
 ---
-license: other
-base_model: apple/mobilevitv2-1.0-imagenet1k-256
 tags:
 - generated_from_trainer
 metrics:
@@ -15,10 +13,10 @@ should probably proofread and complete it, then remove this comment. -->
 
 # results
 
-This model is a fine-tuned version of [apple/mobilevitv2-1.0-imagenet1k-256](https://huggingface.co/apple/mobilevitv2-1.0-imagenet1k-256) on an unknown dataset.
+This model is a fine-tuned version of [](https://huggingface.co/) on an unknown dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.7879
-- Accuracy: 0.7951
+- Loss: 1.1000
+- Accuracy: 0.7236
 
 ## Model description
 
@@ -49,83 +47,25 @@ The following hyperparameters were used during training:
 
 ### Training results
 
-| Training Loss | Epoch  | Step   | Validation Loss | Accuracy |
-|:-------------:|:------:|:------:|:---------------:|:--------:|
-| 1.254         | 0.0267 | 5000   | 1.2409          | 0.6897   |
-| 1.1472        | 0.0534 | 10000  | 1.1382          | 0.7127   |
-| 1.093         | 0.0802 | 15000  | 1.0852          | 0.7244   |
-| 1.0517        | 0.1069 | 20000  | 1.0415          | 0.7345   |
-| 1.023         | 0.1336 | 25000  | 1.0170          | 0.7413   |
-| 1.0047        | 0.1603 | 30000  | 0.9939          | 0.7452   |
-| 0.9878        | 0.1870 | 35000  | 0.9921          | 0.7466   |
-| 0.9758        | 0.2138 | 40000  | 0.9680          | 0.7522   |
-| 0.9661        | 0.2405 | 45000  | 0.9567          | 0.7545   |
-| 0.9582        | 0.2672 | 50000  | 0.9473          | 0.7569   |
-| 0.9425        | 0.2939 | 55000  | 0.9480          | 0.7569   |
-| 0.9391        | 0.3206 | 60000  | 0.9270          | 0.7616   |
-| 0.9301        | 0.3474 | 65000  | 0.9483          | 0.7570   |
-| 0.9196        | 0.3741 | 70000  | 0.9175          | 0.7635   |
-| 0.9168        | 0.4008 | 75000  | 0.9047          | 0.7674   |
-| 0.9139        | 0.4275 | 80000  | 0.9010          | 0.7680   |
-| 0.9022        | 0.4542 | 85000  | 0.8986          | 0.7683   |
-| 0.8986        | 0.4810 | 90000  | 0.8959          | 0.7689   |
-| 0.8965        | 0.5077 | 95000  | 0.8851          | 0.7720   |
-| 0.8906        | 0.5344 | 100000 | 0.8870          | 0.7715   |
-| 0.8888        | 0.5611 | 105000 | 0.8867          | 0.7716   |
-| 0.8852        | 0.5878 | 110000 | 0.8793          | 0.7732   |
-| 0.8822        | 0.6145 | 115000 | 0.8773          | 0.7735   |
-| 0.8796        | 0.6413 | 120000 | 0.8713          | 0.7747   |
-| 0.8707        | 0.6680 | 125000 | 0.8662          | 0.7763   |
-| 0.8735        | 0.6947 | 130000 | 0.8776          | 0.7741   |
-| 0.8659        | 0.7214 | 135000 | 0.8614          | 0.7771   |
-| 0.8635        | 0.7481 | 140000 | 0.8618          | 0.7772   |
-| 0.865         | 0.7749 | 145000 | 0.8561          | 0.7783   |
-| 0.8563        | 0.8016 | 150000 | 0.8585          | 0.7781   |
-| 0.8566        | 0.8283 | 155000 | 0.8493          | 0.7797   |
-| 0.8541        | 0.8550 | 160000 | 0.8493          | 0.7805   |
-| 0.8568        | 0.8817 | 165000 | 0.8431          | 0.7818   |
-| 0.846         | 0.9085 | 170000 | 0.8424          | 0.7819   |
-| 0.8479        | 0.9352 | 175000 | 0.8433          | 0.7812   |
-| 0.8486        | 0.9619 | 180000 | 0.8412          | 0.7823   |
-| 0.8398        | 0.9886 | 185000 | 0.8421          | 0.7818   |
-| 0.825         | 1.0153 | 190000 | 0.8355          | 0.7837   |
-| 0.8316        | 1.0421 | 195000 | 0.8354          | 0.7831   |
-| 0.8293        | 1.0688 | 200000 | 0.8571          | 0.7790   |
-| 0.8243        | 1.0955 | 205000 | 0.8288          | 0.7852   |
-| 0.824         | 1.1222 | 210000 | 0.8293          | 0.7851   |
-| 0.8277        | 1.1489 | 215000 | 0.8256          | 0.7859   |
-| 0.823         | 1.1757 | 220000 | 0.8223          | 0.7869   |
-| 0.8189        | 1.2024 | 225000 | 0.8226          | 0.7865   |
-| 0.8138        | 1.2291 | 230000 | 0.8217          | 0.7871   |
-| 0.8149        | 1.2558 | 235000 | 0.8215          | 0.7872   |
-| 0.8135        | 1.2825 | 240000 | 0.8163          | 0.7880   |
-| 0.8208        | 1.3093 | 245000 | 0.8141          | 0.7889   |
-| 0.8168        | 1.3360 | 250000 | 0.8150          | 0.7882   |
-| 0.8124        | 1.3627 | 255000 | 0.8124          | 0.7888   |
-| 0.8082        | 1.3894 | 260000 | 0.8113          | 0.7894   |
-| 0.8162        | 1.4161 | 265000 | 0.8140          | 0.7884   |
-| 0.8061        | 1.4429 | 270000 | 0.8129          | 0.7890   |
-| 0.8076        | 1.4696 | 275000 | 0.8072          | 0.7899   |
-| 0.8112        | 1.4963 | 280000 | 0.8064          | 0.7907   |
-| 0.8047        | 1.5230 | 285000 | 0.8061          | 0.7903   |
-| 0.8101        | 1.5497 | 290000 | 0.8086          | 0.7901   |
-| 0.8065        | 1.5765 | 295000 | 0.8032          | 0.7912   |
-| 0.7998        | 1.6032 | 300000 | 0.8048          | 0.7909   |
-| 0.8053        | 1.6299 | 305000 | 0.7993          | 0.7920   |
-| 0.8007        | 1.6566 | 310000 | 0.8007          | 0.7921   |
-| 0.7985        | 1.6833 | 315000 | 0.7988          | 0.7923   |
-| 0.8006        | 1.7101 | 320000 | 0.8230          | 0.7873   |
-| 0.8046        | 1.7368 | 325000 | 0.7959          | 0.7930   |
-| 0.794         | 1.7635 | 330000 | 0.7956          | 0.7928   |
-| 0.8047        | 1.7902 | 335000 | 0.8202          | 0.7881   |
-| 0.802         | 1.8169 | 340000 | 0.7953          | 0.7930   |
-| 0.7926        | 1.8436 | 345000 | 0.8007          | 0.7919   |
-| 0.7939        | 1.8704 | 350000 | 0.7918          | 0.7937   |
-| 0.7894        | 1.8971 | 355000 | 0.7911          | 0.7939   |
-| 0.7953        | 1.9238 | 360000 | 0.7879          | 0.7951   |
-| 0.7904        | 1.9505 | 365000 | 0.7922          | 0.7936   |
-| 0.7887        | 1.9772 | 370000 | 0.7899          | 0.7941   |
-| 0.7788        | 2.0040 | 375000 | 0.7888          | 0.7950   |
+| Training Loss | Epoch  | Step  | Validation Loss | Accuracy |
+|:-------------:|:------:|:-----:|:---------------:|:--------:|
+| 1.7698        | 0.2844 | 5000  | 1.7124          | 0.5802   |
+| 1.5445        | 0.5689 | 10000 | 1.5021          | 0.6270   |
+| 1.439         | 0.8533 | 15000 | 1.3989          | 0.6520   |
+| 1.3625        | 1.1377 | 20000 | 1.3447          | 0.6647   |
+| 1.3192        | 1.4222 | 25000 | 1.2965          | 0.6756   |
+| 1.3           | 1.7066 | 30000 | 1.2788          | 0.6795   |
+| 1.2695        | 1.9910 | 35000 | 1.2347          | 0.6900   |
+| 1.2297        | 2.2754 | 40000 | 1.2160          | 0.6955   |
+| 1.2144        | 2.5599 | 45000 | 1.1894          | 0.7021   |
+| 1.1945        | 2.8443 | 50000 | 1.1734          | 0.7058   |
+| 1.1551        | 3.1287 | 55000 | 1.1611          | 0.7084   |
+| 1.1471        | 3.4132 | 60000 | 1.1523          | 0.7104   |
+| 1.1301        | 3.6976 | 65000 | 1.1314          | 0.7156   |
+| 1.1286        | 3.9820 | 70000 | 1.1220          | 0.7186   |
+| 1.0898        | 4.2665 | 75000 | 1.1140          | 0.7203   |
+| 1.093         | 4.5509 | 80000 | 1.1040          | 0.7232   |
+| 1.0893        | 4.8353 | 85000 | 1.0986          | 0.7246   |
 
 
 ### Framework versions

all_results.json

@@ -1,13 +1,13 @@
 {
-    "epoch": 2.0039651791010478,
-    "eval_accuracy": 0.7951413429664056,
-    "eval_loss": 0.787926971912384,
-    "eval_runtime": 394.5318,
-    "eval_samples_per_second": 6390.648,
-    "eval_steps_per_second": 24.964,
-    "total_flos": 8.246362688213484e+18,
-    "train_loss": 0.8725958955078125,
-    "train_runtime": 61229.4812,
-    "train_samples_per_second": 3911.919,
-    "train_steps_per_second": 15.281
+    "epoch": 5.0,
+    "eval_accuracy": 0.723616,
+    "eval_loss": 1.100016713142395,
+    "eval_runtime": 118.4292,
+    "eval_samples_per_second": 2110.967,
+    "eval_steps_per_second": 8.25,
+    "total_flos": 5.4597445596112486e+17,
+    "train_loss": 1.296092871571504,
+    "train_runtime": 24664.1985,
+    "train_samples_per_second": 912.253,
+    "train_steps_per_second": 3.564
 }

config.json

@@ -1,31 +1,127 @@
 {
-  "_name_or_path": "apple/mobilevitv2-1.0-imagenet1k-256",
   "architectures": [
-    "MobileViTV2ForImageClassification"
+    "MobileViTForImageClassification"
   ],
   "aspp_dropout_prob": 0.1,
-  "aspp_out_channels": 512,
+  "aspp_out_channels": 256,
   "atrous_rates": [
     6,
     12,
     18
   ],
-  "attn_dropout": 0.0,
-  "base_attn_unit_dims": [
-    128,
-    192,
-    256
-  ],
+  "attention_probs_dropout_prob": 0.1,
   "classifier_dropout_prob": 0.1,
   "conv_kernel_size": 3,
-  "expand_ratio": 2.0,
-  "ffn_dropout": 0.0,
-  "ffn_multiplier": 2,
-  "hidden_act": "swish",
+  "expand_ratio": 4.0,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.1,
+  "hidden_size": 768,
+  "hidden_sizes": [
+    144,
+    192,
+    240
+  ],
   "id2label": {
     "0": "aircraft carrier",
     "1": "airplane",
+    "2": "alarm clock",
+    "3": "ambulance",
+    "4": "angel",
+    "5": "animal migration",
+    "6": "ant",
+    "7": "anvil",
+    "8": "apple",
+    "9": "arm",
     "10": "asparagus",
+    "11": "axe",
+    "12": "backpack",
+    "13": "banana",
+    "14": "bandage",
+    "15": "barn",
+    "16": "baseball bat",
+    "17": "baseball",
+    "18": "basket",
+    "19": "basketball",
+    "20": "bat",
+    "21": "bathtub",
+    "22": "beach",
+    "23": "bear",
+    "24": "beard",
+    "25": "bed",
+    "26": "bee",
+    "27": "belt",
+    "28": "bench",
+    "29": "bicycle",
+    "30": "binoculars",
+    "31": "bird",
+    "32": "birthday cake",
+    "33": "blackberry",
+    "34": "blueberry",
+    "35": "book",
+    "36": "boomerang",
+    "37": "bottlecap",
+    "38": "bowtie",
+    "39": "bracelet",
+    "40": "brain",
+    "41": "bread",
+    "42": "bridge",
+    "43": "broccoli",
+    "44": "broom",
+    "45": "bucket",
+    "46": "bulldozer",
+    "47": "bus",
+    "48": "bush",
+    "49": "butterfly",
+    "50": "cactus",
+    "51": "cake",
+    "52": "calculator",
+    "53": "calendar",
+    "54": "camel",
+    "55": "camera",
+    "56": "camouflage",
+    "57": "campfire",
+    "58": "candle",
+    "59": "cannon",
+    "60": "canoe",
+    "61": "car",
+    "62": "carrot",
+    "63": "castle",
+    "64": "cat",
+    "65": "ceiling fan",
+    "66": "cell phone",
+    "67": "cello",
+    "68": "chair",
+    "69": "chandelier",
+    "70": "church",
+    "71": "circle",
+    "72": "clarinet",
+    "73": "clock",
+    "74": "cloud",
+    "75": "coffee cup",
+    "76": "compass",
+    "77": "computer",
+    "78": "cookie",
+    "79": "cooler",
+    "80": "couch",
+    "81": "cow",
+    "82": "crab",
+    "83": "crayon",
+    "84": "crocodile",
+    "85": "crown",
+    "86": "cruise ship",
+    "87": "cup",
+    "88": "diamond",
+    "89": "dishwasher",
+    "90": "diving board",
+    "91": "dog",
+    "92": "dolphin",
+    "93": "donut",
+    "94": "door",
+    "95": "dragon",
+    "96": "dresser",
+    "97": "drill",
+    "98": "drums",
+    "99": "duck",
     "100": "dumbbell",
     "101": "ear",
     "102": "elbow",
@@ -36,7 +132,6 @@
     "107": "eyeglasses",
     "108": "face",
     "109": "fan",
-    "11": "axe",
     "110": "feather",
     "111": "fence",
     "112": "finger",
@@ -47,7 +142,6 @@
     "117": "flamingo",
     "118": "flashlight",
     "119": "flip flops",
-    "12": "backpack",
     "120": "floor lamp",
     "121": "flower",
     "122": "flying saucer",
@@ -58,7 +152,6 @@
     "127": "garden hose",
     "128": "garden",
     "129": "giraffe",
-    "13": "banana",
     "130": "goatee",
     "131": "golf club",
     "132": "grapes",
@@ -69,7 +162,6 @@
     "137": "hand",
     "138": "harp",
     "139": "hat",
-    "14": "bandage",
     "140": "headphones",
     "141": "hedgehog",
     "142": "helicopter",
@@ -80,7 +172,6 @@
     "147": "horse",
     "148": "hospital",
     "149": "hot air balloon",
-    "15": "barn",
     "150": "hot dog",
     "151": "hot tub",
     "152": "hourglass",
@@ -91,7 +182,6 @@
     "157": "jacket",
     "158": "jail",
     "159": "kangaroo",
-    "16": "baseball bat",
     "160": "key",
     "161": "keyboard",
     "162": "knee",
@@ -102,7 +192,6 @@
     "167": "leaf",
     "168": "leg",
     "169": "light bulb",
-    "17": "baseball",
     "170": "lighter",
     "171": "lighthouse",
     "172": "lightning",
@@ -113,7 +202,6 @@
     "177": "lollipop",
     "178": "mailbox",
     "179": "map",
-    "18": "basket",
     "180": "marker",
     "181": "matches",
     "182": "megaphone",
@@ -124,7 +212,6 @@
     "187": "moon",
     "188": "mosquito",
     "189": "motorbike",
-    "19": "basketball",
     "190": "mountain",
     "191": "mouse",
     "192": "moustache",
@@ -135,8 +222,6 @@
     "197": "necklace",
     "198": "nose",
     "199": "ocean",
-    "2": "alarm clock",
-    "20": "bat",
     "200": "octagon",
     "201": "octopus",
     "202": "onion",
@@ -147,7 +232,6 @@
     "207": "palm tree",
     "208": "panda",
     "209": "pants",
-    "21": "bathtub",
     "210": "paper clip",
     "211": "parachute",
     "212": "parrot",
@@ -158,7 +242,6 @@
     "217": "pencil",
     "218": "penguin",
     "219": "piano",
-    "22": "beach",
     "220": "pickup truck",
     "221": "picture frame",
     "222": "pig",
@@ -169,7 +252,6 @@
     "227": "police car",
     "228": "pond",
     "229": "pool",
-    "23": "bear",
     "230": "popsicle",
     "231": "postcard",
     "232": "potato",
@@ -180,7 +262,6 @@
     "237": "radio",
     "238": "rain",
     "239": "rainbow",
-    "24": "beard",
     "240": "rake",
     "241": "remote control",
     "242": "rhinoceros",
@@ -191,7 +272,6 @@
     "247": "sailboat",
     "248": "sandwich",
     "249": "saw",
-    "25": "bed",
     "250": "saxophone",
     "251": "school bus",
     "252": "scissors",
@@ -202,7 +282,6 @@
     "257": "shark",
     "258": "sheep",
     "259": "shoe",
-    "26": "bee",
     "260": "shorts",
     "261": "shovel",
     "262": "sink",
@@ -213,7 +292,6 @@
     "267": "smiley face",
     "268": "snail",
     "269": "snake",
-    "27": "belt",
     "270": "snorkel",
     "271": "snowflake",
     "272": "snowman",
@@ -224,7 +302,6 @@
     "277": "spoon",
     "278": "spreadsheet",
     "279": "square",
-    "28": "bench",
     "280": "squiggle",
     "281": "squirrel",
     "282": "stairs",
@@ -235,7 +312,6 @@
     "287": "stitches",
     "288": "stop sign",
     "289": "stove",
-    "29": "bicycle",
     "290": "strawberry",
     "291": "streetlight",
     "292": "string bean",
@@ -246,8 +322,6 @@
     "297": "sweater",
     "298": "swing set",
     "299": "sword",
-    "3": "ambulance",
-    "30": "binoculars",
     "300": "syringe",
     "301": "t-shirt",
     "302": "table",
@@ -258,7 +332,6 @@
     "307": "tennis racquet",
     "308": "tent",
     "309": "The Eiffel Tower",
-    "31": "bird",
     "310": "The Great Wall of China",
     "311": "The Mona Lisa",
     "312": "tiger",
@@ -269,7 +342,6 @@
     "317": "toothbrush",
     "318": "toothpaste",
     "319": "tornado",
-    "32": "birthday cake",
     "320": "tractor",
     "321": "traffic light",
     "322": "train",
@@ -280,7 +352,6 @@
     "327": "trumpet",
     "328": "umbrella",
     "329": "underwear",
-    "33": "blackberry",
     "330": "van",
     "331": "vase",
     "332": "violin",
@@ -291,86 +362,16 @@
     "337": "wheel",
     "338": "windmill",
     "339": "wine bottle",
-    "34": "blueberry",
     "340": "wine glass",
     "341": "wristwatch",
     "342": "yoga",
     "343": "zebra",
-    "344": "zigzag",
-    "35": "book",
-    "36": "boomerang",
-    "37": "bottlecap",
-    "38": "bowtie",
-    "39": "bracelet",
-    "4": "angel",
-    "40": "brain",
-    "41": "bread",
-    "42": "bridge",
-    "43": "broccoli",
-    "44": "broom",
-    "45": "bucket",
-    "46": "bulldozer",
-    "47": "bus",
-    "48": "bush",
-    "49": "butterfly",
-    "5": "animal migration",
-    "50": "cactus",
-    "51": "cake",
-    "52": "calculator",
-    "53": "calendar",
-    "54": "camel",
-    "55": "camera",
-    "56": "camouflage",
-    "57": "campfire",
-    "58": "candle",
-    "59": "cannon",
-    "6": "ant",
-    "60": "canoe",
-    "61": "car",
-    "62": "carrot",
-    "63": "castle",
-    "64": "cat",
-    "65": "ceiling fan",
-    "66": "cell phone",
-    "67": "cello",
-    "68": "chair",
-    "69": "chandelier",
-    "7": "anvil",
-    "70": "church",
-    "71": "circle",
-    "72": "clarinet",
-    "73": "clock",
-    "74": "cloud",
-    "75": "coffee cup",
-    "76": "compass",
-    "77": "computer",
-    "78": "cookie",
-    "79": "cooler",
-    "8": "apple",
-    "80": "couch",
-    "81": "cow",
-    "82": "crab",
-    "83": "crayon",
-    "84": "crocodile",
-    "85": "crown",
-    "86": "cruise ship",
-    "87": "cup",
-    "88": "diamond",
-    "89": "dishwasher",
-    "9": "arm",
-    "90": "diving board",
-    "91": "dog",
-    "92": "dolphin",
-    "93": "donut",
-    "94": "door",
-    "95": "dragon",
-    "96": "dresser",
-    "97": "drill",
-    "98": "drums",
-    "99": "duck"
+    "344": "zigzag"
   },
+  "ignore_mismatched_sizes": true,
   "image_size": 64,
   "initializer_range": 0.02,
+  "intermediate_size": 3072,
   "label2id": {
     "The Eiffel Tower": "309",
     "The Great Wall of China": "310",
@@ -720,18 +721,24 @@
   },
   "layer_norm_eps": 1e-05,
   "mlp_ratio": 2.0,
-  "model_type": "mobilevitv2",
-  "n_attn_blocks": [
-    2,
-    4,
-    3
+  "model_type": "mobilevit",
+  "neck_hidden_sizes": [
+    16,
+    32,
+    64,
+    96,
+    128,
+    160,
+    640
   ],
+  "num_attention_heads": 12,
   "num_channels": 1,
+  "num_hidden_layers": 12,
   "output_stride": 32,
-  "patch_size": 2,
+  "patch_size": 16,
   "problem_type": "single_label_classification",
+  "qkv_bias": true,
   "semantic_loss_ignore_index": 255,
   "torch_dtype": "float32",
-  "transformers_version": "4.40.0",
-  "width_multiplier": 1.0
+  "transformers_version": "4.40.0"
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1c56ce027098d56d11ac636e6c45af9c186175031adfaaf0e614eeb9ac1d2dba
-size 18360744
+oid sha256:d35f61619e3ffc371ead68851aa0232c0fe6cc18ab6a5362d0e589c58eb59a19
+size 20730036

preprocessor_config.json

@@ -15,8 +15,8 @@
     "input_data_format"
   ],
   "crop_size": {
-    "height": 56,
-    "width": 56
+    "height": 28,
+    "width": 28
   },
   "do_center_crop": true,
   "do_convert_rgb": false,
@@ -27,6 +27,6 @@
   "resample": 2,
   "rescale_factor": 0.00392156862745098,
   "size": {
-    "shortest_edge": 56
+    "shortest_edge": 28
   }
 }

test_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 2.0039651791010478,
-    "eval_accuracy": 0.7951413429664056,
-    "eval_loss": 0.787926971912384,
-    "eval_runtime": 394.5318,
-    "eval_samples_per_second": 6390.648,
-    "eval_steps_per_second": 24.964
+    "epoch": 5.0,
+    "eval_accuracy": 0.723616,
+    "eval_loss": 1.100016713142395,
+    "eval_runtime": 118.4292,
+    "eval_samples_per_second": 2110.967,
+    "eval_steps_per_second": 8.25
 }

train_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 2.0039651791010478,
-    "total_flos": 8.246362688213484e+18,
-    "train_loss": 0.8725958955078125,
-    "train_runtime": 61229.4812,
-    "train_samples_per_second": 3911.919,
-    "train_steps_per_second": 15.281
+    "epoch": 5.0,
+    "total_flos": 5.4597445596112486e+17,
+    "train_loss": 1.296092871571504,
+    "train_runtime": 24664.1985,
+    "train_samples_per_second": 912.253,
+    "train_steps_per_second": 3.564
 }

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b89a07bb3cdaf58b38f7161c3c9c6c2f098869dbaa81424749efa50f11303fc3
-size 4984
+oid sha256:eb6b0f81e80c1f33b16aaf9c2fc69495be719e29fb2dfcec7d7a4debabe294f0
+size 4920