This project was meant as a way to gain experience working with TensorFlow / Keras for image classification. The dataset used was extracted from the PokeAPI sprite repository, which be found at this GitHub link. This dataset had to be reorganized into seperate directories based on the Pokémon the image was associated with. This was done using Python via the code found in subfolders.py.
# !pip uninstall keras -y
# !pip uninstall keras-nightly -y
# !pip uninstall keras-Preprocessing -y
# !pip uninstall keras-vis -y
# !pip uninstall tensorflow -y
# !pip install tensorflow
# !pip install kerastimport tensorflow as tf
from tensorflow import keras
from tensorflow.keras import layers
from tensorflow.keras.models import Sequential
import matplotlib.pyplot as plt
import numpy as np
import osModel was trained using an NVIDIA GTX 1070, with CUDA Toolkit and cuDNN SDK installed.
print(tf.__version__) # Show the installed tensorflow version2.10.0
print(tf.config.list_physical_devices('GPU')) # Display device being used to confirm GPU is configured correctly[PhysicalDevice(name='/physical_device:GPU:0', device_type='GPU')]
# Create some tensors to confirm TensorFlow is using the GPU
a = tf.constant([[1.0, 2.0, 3.0], [4.0, 5.0, 6.0]])
b = tf.constant([[1.0, 2.0], [3.0, 4.0], [5.0, 6.0]])
c = tf.matmul(a, b)
print(c)tf.Tensor(
[[22. 28.]
[49. 64.]], shape=(2, 2), dtype=float32)
Two datasets were created from the overarching collection of Pokémon images: a training dataset (70%) and a validation dataset (30%).
data_dir = 'organized'
batch_size = 32
img_height = 120
img_width = 120
final_ds = tf.keras.utils.image_dataset_from_directory(
data_dir,
image_size=(img_height, img_width),
batch_size=batch_size
)
train_ds = tf.keras.utils.image_dataset_from_directory(
data_dir,
validation_split=0.3,
subset="training",
seed=123,
image_size=(img_height, img_width),
batch_size=batch_size)
val_ds = tf.keras.utils.image_dataset_from_directory(
data_dir,
validation_split=0.3,
subset="validation",
seed=123,
image_size=(img_height, img_width),
batch_size=batch_size)Found 39711 files belonging to 1726 classes.
Found 39711 files belonging to 1726 classes.
Using 27798 files for training.
Found 39711 files belonging to 1726 classes.
Using 11913 files for validation.
4.1: Display the class names within the dataset. Each class is associated with a unique Pokémon, represented by its Pokédex number.
Note: Variants of specific Pokemon are denoted with their Pokedex number and a brief label for the variant.
An example of this is Blastoise. Class '9' represents Blastoise's base form, while '9-mega' represents Mega Blastoise
| |
|
# Print class names (each Pokémon)
class_names = train_ds.class_names
print(class_names)['0', '1', '10', '100', '10001', '10002', '10003', '10004', '10005', '10006', '10007', '10008', '10009', '10010', '10011', '10012', '10013', '10014', '10015', '10016', '10017', '10018', '10019', '10020', '10021', '10022', '10023', '10024', '10025', '10026', '10027', '10028', '10029', '10030', '10031', '10032', '10033', '10034', '10035', '10036', '10037', '10038', '10039', '10040', '10041', '10042', '10043', '10044', '10045', '10046', '10047', '10048', '10049', '10050', '10051', '10052', '10053', '10054', '10055', '10056', '10057', '10058', '10059', '10060', '10061', '10062', '10063', '10064', '10065', '10066', '10067', '10068', '10069', '10070', '10071', '10072', '10073', '10074', '10075', '10076', '10077', '10078', '10079', '10080', '10081', '10082', '10083', '10084', '10085', '10086', '10087', '10088', '10089', '10090', '10091', '10092', '10093', '10094', '10095', '10096', '10097', '10098', '10099', '101', '10100', '10101', '10102', '10103', '10104', '10105', '10106', '10107', '10108', '10109', '10110', '10111', '10112', '10113', '10114', '10115', '10116', '10117', '10118', '10119', '10120', '10121', '10122', '10123', '10124', '10125', '10126', '10127', '10128', '10129', '10130', '10131', '10132', '10133', '10134', '10135', '10136', '10137', '10138', '10139', '10140', '10141', '10142', '10143', '10144', '10145', '10146', '10147', '10148', '10149', '10150', '10151', '10152', '10153', '10154', '10155', '10156', '10157', '10159', '10160', '10161', '10162', '10163', '10164', '10165', '10166', '10167', '10168', '10169', '10170', '10171', '10172', '10173', '10174', '10175', '10176', '10177', '10178', '10179', '10180', '10181', '10183', '10184', '10185', '10186', '10187', '10188', '10189', '10190', '10191', '10192', '10193', '10194', '10195', '10196', '10197', '10198', '10199', '102', '10200', '10201', '10202', '10203', '10204', '10205', '10206', '10207', '10208', '10209', '10210', '10211', '10212', '10213', '10214', '10215', '10216', '10217', '10218', '10219', '10220', '10221', '10222', '10223', '10224', '10225', '10226', '10227', '10228', '10229', '10230', '10231', '10232', '10233', '10234', '10235', '10236', '10237', '10238', '10239', '10240', '10241', '10242', '10243', '10244', '10245', '10246', '10247', '10248', '10249', '103', '103-alola', '104', '105', '105-alola', '105-totem', '106', '107', '108', '109', '11', '110', '110-galar', '111', '112', '113', '114', '115', '115-mega', '116', '117', '118', '119', '12', '12-gmax', '120', '121', '122', '122-galar', '123', '124', '125', '126', '127', '127-mega', '128', '129', '13', '130', '130-mega', '131', '131-gmax', '132', '133', '133-gmax', '133-starter', '134', '135', '136', '137', '138', '139', '14', '140', '141', '142', '142-mega', '143', '143-gmax', '144', '144-galar', '145', '145-galar', '146', '146-galar', '147', '148', '149', '15', '15-mega', '150', '150-mega-x', '150-mega-y', '151', '152', '153', '154', '155', '156', '157', '158', '159', '16', '160', '161', '162', '163', '164', '165', '166', '167', '168', '169', '17', '170', '171', '172', '172-beta', '172-spiky-eared', '173', '174', '175', '176', '177', '178', '179', '18', '18-mega', '180', '181', '181-mega', '182', '183', '184', '185', '186', '187', '188', '189', '19', '19-alola', '190', '191', '192', '193', '194', '195', '196', '197', '198', '199', '199-galar', '2', '20', '20-alola', '20-totem-alola', '200', '201', '201-a', '201-b', '201-c', '201-d', '201-e', '201-exclamation', '201-f', '201-g', '201-h', '201-i', '201-j', '201-k', '201-l', '201-m', '201-n', '201-o', '201-p', '201-q', '201-question', '201-r', '201-s', '201-t', '201-u', '201-v', '201-w', '201-x', '201-y', '201-z', '202', '203', '204', '205', '206', '207', '208', '208-mega', '209', '21', '210', '211', '212', '212-mega', '213', '214', '214-mega', '215', '216', '217', '218', '219', '22', '220', '221', '222', '222-galar', '223', '224', '225', '226', '227', '228', '229', '229-mega', '23', '230', '231', '232', '233', '234', '235', '236', '237', '238', '239', '24', '240', '241', '242', '243', '244', '245', '246', '247', '248', '248-mega', '249', '25', '25-alola-cap', '25-belle', '25-cosplay', '25-gmax', '25-hoenn-cap', '25-kalos-cap', '25-libre', '25-original-cap', '25-partner-cap', '25-phd', '25-pop-star', '25-rock-star', '25-sinnoh-cap', '25-starter', '25-unova-cap', '25-world-cap', '250', '251', '252', '253', '254', '254-mega', '255', '256', '257', '257-mega', '258', '259', '26', '26-alola', '260', '260-mega', '261', '262', '263', '263-galar', '264', '264-galar', '265', '266', '267', '268', '269', '27', '27-alola', '270', '271', '272', '273', '274', '275', '276', '277', '278', '279', '28', '28-alola', '280', '281', '282', '282-mega', '283', '284', '285', '286', '287', '288', '289', '29', '290', '291', '292', '293', '294', '295', '296', '297', '298', '299', '3', '3-gmax', '3-mega', '30', '300', '301', '302', '302-mega', '303', '303-mega', '304', '305', '306', '306-mega', '307', '308', '308-mega', '309', '31', '310', '310-mega', '311', '312', '313', '314', '315', '316', '317', '318', '319', '319-mega', '32', '320', '321', '322', '323', '323-mega', '324', '325', '326', '327', '327-blank', '327-filled', '328', '329', '33', '330', '331', '332', '333', '334', '334-mega', '335', '336', '337', '338', '339', '34', '340', '341', '342', '343', '344', '345', '346', '347', '348', '349', '35', '350', '351', '351-rainy', '351-snowy', '351-sunny', '352', '353', '354', '354-mega', '355', '356', '357', '358', '359', '359-mega', '36', '360', '361', '362', '362-mega', '363', '364', '365', '366', '367', '368', '369', '37', '37-alola', '370', '371', '372', '373', '373-mega', '374', '375', '376', '376-mega', '377', '378', '379', '38', '38-alola', '380', '380-mega', '381', '381-mega', '382', '382-primal', '383', '383-primal', '384', '384-mega', '385', '386', '386-attack', '386-defense', '386-normal', '386-speed', '387', '388', '389', '39', '390', '391', '392', '393', '394', '395', '396', '397', '398', '399', '4', '40', '400', '401', '402', '403', '404', '405', '406', '407', '408', '409', '41', '410', '411', '412', '412-beta', '412-plant', '412-sandy', '412-trash', '413', '413-plant', '413-sandy', '413-trash', '414', '414-plant', '415', '416', '417', '418', '419', '42', '420', '421', '421-beta', '421-overcast', '421-sunshine', '422', '422-east', '422-west', '423', '423-east', '423-west', '424', '425', '426', '427', '428', '428-mega', '429', '43', '430', '431', '432', '433', '434', '435', '436', '437', '438', '439', '44', '440', '441', '442', '443', '444', '445', '445-mega', '446', '447', '448', '448-mega', '449', '45', '450', '451', '452', '453', '454', '455', '456', '457', '458', '459', '46', '460', '460-mega', '461', '462', '463', '464', '465', '466', '467', '468', '469', '47', '470', '471', '472', '473', '474', '475', '475-mega', '476', '477', '478', '479', '479-fan', '479-frost', '479-heat', '479-mow', '479-wash', '48', '480', '481', '482', '483', '484', '485', '486', '487', '487-altered', '487-origin', '488', '489', '49', '490', '491', '492', '492-land', '492-sky', '493', '493-bug', '493-dark', '493-dragon', '493-electric', '493-fairy', '493-fighting', '493-fire', '493-flying', '493-ghost', '493-grass', '493-ground', '493-ice', '493-normal', '493-poison', '493-psychic', '493-rock', '493-steel', '493-stone', '493-unknown', '493-water', '494', '495', '496', '497', '498', '499', '5', '50', '50-alola', '500', '501', '502', '503', '504', '505', '506', '507', '508', '509', '51', '51-alola', '510', '511', '512', '513', '514', '515', '516', '517', '518', '519', '52', '52-alola', '52-galar', '52-gmax', '520', '521', '522', '523', '524', '525', '526', '527', '528', '529', '53', '53-alola', '530', '531', '531-mega', '532', '533', '534', '535', '536', '537', '538', '539', '54', '540', '541', '542', '543', '544', '545', '546', '547', '548', '549', '55', '550', '550-blue-striped', '550-red-striped', '551', '552', '553', '554', '554-galar', '555', '555-galar', '555-galar-zen', '555-standard', '555-zen', '556', '557', '558', '559', '56', '560', '561', '562', '562-galar', '563', '564', '565', '566', '567', '568', '569', '569-gmax', '57', '570', '571', '572', '573', '574', '575', '576', '577', '578', '579', '58', '580', '581', '582', '583', '584', '585', '585-autumn', '585-spring', '585-sprint', '585-summer', '585-winter', '586', '586-autumn', '586-spring', '586-sprint', '586-summer', '586-winter', '587', '588', '589', '59', '590', '591', '592', '593', '594', '595', '596', '597', '598', '599', '6', '6-gmax', '6-mega-x', '6-mega-y', '60', '600', '601', '602', '603', '604', '605', '606', '607', '608', '609', '61', '610', '611', '612', '613', '614', '615', '616', '617', '618', '618--galar', '619', '62', '620', '621', '622', '623', '624', '625', '626', '627', '628', '629', '63', '630', '631', '632', '633', '634', '635', '636', '637', '638', '639', '64', '640', '641', '641-incarnate', '641-therian', '642', '642-incarnate', '642-therian', '643', '644', '645', '645-incarnate', '645-therian', '646', '646-black', '646-white', '647', '647-ordinary', '647-resolute', '648', '648-aria', '648-pirouette', '649', '649-burn', '649-chill', '649-douse', '649-shock', '65', '65-mega', '650', '651', '652', '653', '654', '655', '656', '657', '658', '658-ash', '658-battle-bond', '658_2', '659', '66', '660', '661', '662', '663', '664', '664-icy-snow', '665', '665-icy-snow', '666', '666-archipelago', '666-continental', '666-elegant', '666-fancy', '666-garden', '666-high-plains', '666-icy-snow', '666-jungle', '666-marine', '666-meadow', '666-modern', '666-monsoon', '666-ocean', '666-poke-ball', '666-polar', '666-river', '666-sandstorm', '666-savanna', '666-sun', '666-tundra', '667', '668', '669', '669-blue', '669-orange', '669-red', '669-white', '669-yellow', '67', '670', '670-blue', '670-eternal', '670-orange', '670-red', '670-white', '670-yellow', '671', '671-blue', '671-orange', '671-red', '671-white', '671-yellow', '672', '673', '674', '675', '676', '676-dandy', '676-debutante', '676-diamond', '676-heart', '676-kabuki', '676-la-reine', '676-matron', '676-natural', '676-pharaoh', '676-star', '677', '678', '678-female', '678-male', '679', '68', '68-gmax', '680', '681', '681-blade', '681-shield', '682', '683', '684', '685', '686', '687', '688', '689', '69', '690', '691', '692', '693', '694', '695', '696', '697', '698', '699', '7', '70', '700', '701', '702', '703', '704', '705', '706', '707', '708', '709', '71', '710', '710-average', '710-large', '710-small', '710-super', '711', '711-average', '711-large', '711-small', '711-super', '712', '713', '714', '715', '716', '716-active', '716-neutral', '717', '718', '718-10', '718-50', '718-complete', '718_2', '718_3', '719', '719-mega', '72', '720', '720-confined', '720-unbound', '721', '722', '723', '724', '725', '726', '727', '728', '729', '73', '730', '731', '732', '733', '734', '735', '735-totem', '736', '737', '738', '738-totem', '739', '74', '74-alola', '740', '741', '741-baile', '741-pau', '741-pom-pom', '741-sensu', '742', '743', '743-totem', '744', '744-own-tempo', '745', '745-dusk', '745-midday', '745-midnight', '746', '746-school', '746-solo', '747', '748', '749', '75', '75-alola', '750', '751', '752', '752-totem', '753', '754', '754-totem', '755', '756', '757', '758', '758-totem', '759', '76', '76-alola', '760', '761', '762', '763', '764', '765', '766', '767', '768', '769', '77', '77-galar', '770', '771', '772', '773', '773-bug', '773-dark', '773-dragon', '773-electric', '773-fairy', '773-fighting', '773-fire', '773-flying', '773-ghost', '773-grass', '773-ground', '773-ice', '773-normal', '773-poison', '773-psychic', '773-rock', '773-steel', '773-water', '774', '774-blue', '774-blue-meteor', '774-green', '774-green-meteor', '774-indigo', '774-indigo-meteor', '774-orange', '774-orange-meteor', '774-red', '774-red-meteor', '774-violet', '774-violet-meteor', '774-yellow', '774-yellow-meteor', '775', '775-form-1', '776', '777', '777-totem', '778', '778-busted', '778-disguised', '778-totem-busted', '778-totem-disguised', '779', '78', '78-galar', '780', '781', '782', '783', '784', '784-totem', '785', '786', '787', '788', '789', '79', '79-galar', '790', '791', '792', '793', '794', '795', '796', '797', '798', '799', '8', '80', '80-galar', '80-mega', '800', '800-dawn', '800-dusk', '800-ultra', '801', '801-original', '802', '803', '804', '804s', '805', '806', '807', '808', '809', '809-gmax', '81', '810', '811', '812', '812-gmax', '813', '814', '815', '815-gmax', '816', '817', '818', '818-gmax', '819', '82', '820', '821', '822', '823', '823-gmax', '823sb', '824', '825', '826', '826-gmax', '827', '828', '829', '83', '83-galar', '830', '831', '832', '833', '834', '834-gmax', '835', '836', '837', '838', '838sb', '839', '839-gmax', '84', '840', '841', '841-gmax', '842', '843', '844', '844-gmax', '845', '845-gorging', '845-gulping', '846', '847', '848', '849', '849-gmax', '849-low-key', '85', '850', '851', '851-gmax', '852', '853', '854', '855', '856', '857', '858', '858-gmax', '859', '86', '860', '861', '861-gmax', '862', '863', '864', '865', '866', '867', '868', '869', '869-caramel-swirl-berry', '869-caramel-swirl-berry-sweet', '869-caramel-swirl-clove-sweet', '869-caramel-swirl-clover', '869-caramel-swirl-flower', '869-caramel-swirl-flower-sweet', '869-caramel-swirl-love', '869-caramel-swirl-love-sweet', '869-caramel-swirl-plain', '869-caramel-swirl-ribbon', '869-caramel-swirl-ribbon-sweet', '869-caramel-swirl-star', '869-caramel-swirl-star-sweet', '869-caramel-swirl-strawberry', '869-caramel-swirl-strawberry-sweet', '869-gmax', '869-lemon-cream-berry', '869-lemon-cream-berry-sweet', '869-lemon-cream-clove-sweet', '869-lemon-cream-clover', '869-lemon-cream-flower', '869-lemon-cream-flower-sweet', '869-lemon-cream-love', '869-lemon-cream-love-sweet', '869-lemon-cream-plain', '869-lemon-cream-ribbon', '869-lemon-cream-ribbon-sweet', '869-lemon-cream-star', '869-lemon-cream-star-sweet', '869-lemon-cream-strawberry', '869-lemon-cream-strawberry-sweet', '869-matcha-cream-berry', '869-matcha-cream-berry-sweet', '869-matcha-cream-clover', '869-matcha-cream-clover-sweet', '869-matcha-cream-flower', '869-matcha-cream-flower-sweet', '869-matcha-cream-love', '869-matcha-cream-love-sweet', '869-matcha-cream-plain', '869-matcha-cream-ribbon', '869-matcha-cream-ribbon-sweet', '869-matcha-cream-star', '869-matcha-cream-star-sweet', '869-matcha-cream-strawberry', '869-matcha-cream-strawberry-sweet', '869-mint-cream-berry', '869-mint-cream-berry-sweet', '869-mint-cream-clove-sweet', '869-mint-cream-clover', '869-mint-cream-flower', '869-mint-cream-flower-sweet', '869-mint-cream-love', '869-mint-cream-love-sweet', '869-mint-cream-plain', '869-mint-cream-ribbon', '869-mint-cream-ribbon-sweet', '869-mint-cream-star', '869-mint-cream-star-sweet', '869-mint-cream-strawberry', '869-mint-cream-strawberry-sweet', '869-rainbow-swirl-berry', '869-rainbow-swirl-berry-sweet', '869-rainbow-swirl-clove-sweet', '869-rainbow-swirl-clover', '869-rainbow-swirl-flower', '869-rainbow-swirl-flower-sweet', '869-rainbow-swirl-love', '869-rainbow-swirl-love-sweet', '869-rainbow-swirl-plain', '869-rainbow-swirl-ribbon', '869-rainbow-swirl-ribbon-sweet', '869-rainbow-swirl-star', '869-rainbow-swirl-star-sweet', '869-rainbow-swirl-strawberry', '869-rainbow-swirl-strawberry-sweet', '869-ruby-cream-berry', '869-ruby-cream-berry-sweet', '869-ruby-cream-clove-sweet', '869-ruby-cream-clover', '869-ruby-cream-flower', '869-ruby-cream-flower-sweet', '869-ruby-cream-love', '869-ruby-cream-love-sweet', '869-ruby-cream-plain', '869-ruby-cream-ribbon', '869-ruby-cream-ribbon-sweet', '869-ruby-cream-star', '869-ruby-cream-star-sweet', '869-ruby-cream-strawberry', '869-ruby-cream-strawberry-sweet', '869-ruby-swirl-berry', '869-ruby-swirl-berry-sweet', '869-ruby-swirl-clove-sweet', '869-ruby-swirl-clover', '869-ruby-swirl-flower', '869-ruby-swirl-flower-sweet', '869-ruby-swirl-love', '869-ruby-swirl-love-sweet', '869-ruby-swirl-plain', '869-ruby-swirl-ribbon', '869-ruby-swirl-ribbon-sweet', '869-ruby-swirl-star', '869-ruby-swirl-star-sweet', '869-ruby-swirl-strawberry', '869-ruby-swirl-strawberry-sweet', '869-salted-cream-berry', '869-salted-cream-berry-sweet', '869-salted-cream-clove-sweet', '869-salted-cream-clover', '869-salted-cream-flower', '869-salted-cream-flower-sweet', '869-salted-cream-love', '869-salted-cream-love-sweet', '869-salted-cream-plain', '869-salted-cream-ribbon', '869-salted-cream-ribbon-sweet', '869-salted-cream-star', '869-salted-cream-star-sweet', '869-salted-cream-strawberry', '869-salted-cream-strawberry-sweet', '869-vanilla-cream-berry', '869-vanilla-cream-berry-sweet', '869-vanilla-cream-clove-sweet', '869-vanilla-cream-clover', '869-vanilla-cream-flower', '869-vanilla-cream-flower-sweet', '869-vanilla-cream-love', '869-vanilla-cream-love-sweet', '869-vanilla-cream-plain', '869-vanilla-cream-ribbon', '869-vanilla-cream-ribbon-sweet', '869-vanilla-cream-star', '869-vanilla-cream-star-sweet', '869-vanilla-cream-strawberry', '869-vanilla-cream-strawberry-sweet', '869_11', '869_13', '869_15', '869_17', '869_19', '869_2', '869_21', '869_23', '869_25', '869_27', '869_29', '869_3', '869_31', '869_33', '869_35', '869_37', '869_39', '869_4', '869_41', '869_43', '869_45', '869_47', '869_49', '869_5', '869_51', '869_53', '869_55', '869_57', '869_59', '869_6', '869_61', '869_63', '869_7', '869_9', '87', '870', '871', '872', '873', '874', '875', '875-noice', '876', '877', '877-hangry', '878', '879', '879-gmax', '88', '88-alola', '880', '881', '882', '883', '884', '884-gmax', '885', '886', '887', '888', '888-crowned', '888b', '889', '889-crowned', '89', '89-alola', '890', '890-eternamax', '890-gmax', '891', '892', '892-gmax', '892-rapid-strike-gmax', '892_2', '892_2b', '893', '893-dada', '894', '895', '896', '897', '898', '898-ice-rider', '898-shadow-rider', '899', '9', '9-gmax', '9-mega', '90', '900', '901', '902', '903', '904', '905', '91', '92', '93', '94', '94-gmax', '94-mega', '95', '96', '97', '98', '99', '99-gmax', 'egg', 'egg-manaphy', 'pikachu-partner', 'substitute']
import matplotlib.pyplot as plt
plt.figure(figsize=(10, 10))
for images, labels in train_ds.take(1):
for i in range(9):
ax = plt.subplot(3, 3, i + 1)
plt.imshow(images[i].numpy().astype("uint8"))
plt.title(class_names[labels[i]])
plt.axis("off")
AUTOTUNE = tf.data.AUTOTUNE
train_ds = train_ds.cache().shuffle(1000).prefetch(buffer_size=AUTOTUNE)
test_ds = test_ds.cache().prefetch(buffer_size=AUTOTUNE)num_classes = len(class_names)
# Create sequential model
model = tf.keras.Sequential([
tf.keras.layers.Rescaling(1./255),
tf.keras.layers.Conv2D(32, 3, activation='relu'),
tf.keras.layers.MaxPooling2D(),
tf.keras.layers.Conv2D(32, 3, activation='relu'),
tf.keras.layers.MaxPooling2D(),
tf.keras.layers.Conv2D(32, 3, activation='relu'),
tf.keras.layers.MaxPooling2D(),
tf.keras.layers.Flatten(),
tf.keras.layers.Dense(128, activation='relu'),
tf.keras.layers.Dense(num_classes)
])# Compile model (Using Adam optimizer, categorical_crossentropy loss and metrics as accuracy)
model.compile(
optimizer='adam',
loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True),
metrics=['accuracy'])Model was trained for 11 epochs based on comparison of training loss and validation loss. Other epoch counts were used in testing.
# Train new model for 11 epochs
epochs = 11
history = model.fit(
train_ds,
validation_data=val_ds,
epochs=epochs
)Epoch 1/11
869/869 [==============================] - 17s 19ms/step - loss: 1.5091 - accuracy: 0.6471 - val_loss: 3.1877 - val_accuracy: 0.4873
Epoch 2/11
869/869 [==============================] - 40s 46ms/step - loss: 1.1275 - accuracy: 0.7220 - val_loss: 3.3420 - val_accuracy: 0.5126
Epoch 3/11
869/869 [==============================] - 25s 29ms/step - loss: 0.8702 - accuracy: 0.7755 - val_loss: 3.2549 - val_accuracy: 0.5292
Epoch 4/11
869/869 [==============================] - 16s 19ms/step - loss: 0.6842 - accuracy: 0.8150 - val_loss: 3.5723 - val_accuracy: 0.5469
Epoch 5/11
869/869 [==============================] - 16s 19ms/step - loss: 0.5530 - accuracy: 0.8427 - val_loss: 3.7912 - val_accuracy: 0.5604
Epoch 6/11
869/869 [==============================] - 16s 19ms/step - loss: 0.4597 - accuracy: 0.8682 - val_loss: 4.0661 - val_accuracy: 0.5597
Epoch 7/11
869/869 [==============================] - 17s 19ms/step - loss: 0.3784 - accuracy: 0.8869 - val_loss: 4.3638 - val_accuracy: 0.5439
Epoch 8/11
869/869 [==============================] - 17s 19ms/step - loss: 0.3386 - accuracy: 0.8978 - val_loss: 4.4301 - val_accuracy: 0.5742
Epoch 9/11
869/869 [==============================] - 17s 19ms/step - loss: 0.2982 - accuracy: 0.9086 - val_loss: 4.3480 - val_accuracy: 0.5757
Epoch 10/11
869/869 [==============================] - 17s 20ms/step - loss: 0.2720 - accuracy: 0.9167 - val_loss: 4.7629 - val_accuracy: 0.5805
Epoch 11/11
869/869 [==============================] - 17s 20ms/step - loss: 0.2542 - accuracy: 0.9218 - val_loss: 4.7498 - val_accuracy: 0.5812
# Display results of training of new version of model
acc = history.history['accuracy']
val_acc = history.history['val_accuracy']
loss = history.history['loss']
val_loss = history.history['val_loss']
epochs_range = range(epochs)
plt.figure(figsize=(8, 8))
plt.subplot(1, 2, 1)
plt.plot(epochs_range, acc, label='Training Accuracy')
plt.plot(epochs_range, val_acc, label='Validation Accuracy')
plt.legend(loc='lower right')
plt.title('Training and Validation Accuracy')
plt.subplot(1, 2, 2)
plt.plot(epochs_range, loss, label='Training Loss')
plt.plot(epochs_range, val_loss, label='Validation Loss')
plt.legend(loc='upper right')
plt.title('Training and Validation Loss')
plt.show()
Expected Result: 149
tf.keras.utils.load_img('./dragonite.png', target_size=(img_height, img_width))
img = tf.keras.utils.load_img(
'./dragonite.png', target_size=(img_height, img_width)
)
img_array = tf.keras.utils.img_to_array(img)
img_array = tf.expand_dims(img_array, 0) # Create a batch
predictions = model.predict(img_array)
score = tf.nn.softmax(predictions[0])
print(
"This image most likely belongs to {} with a {:.2f} percent confidence."
.format(class_names[np.argmax(score)], 100 * np.max(score))
)1/1 [==============================] - 0s 37ms/step
This image most likely belongs to 149 with a 45.21 percent confidence.
Expected Result: 213
tf.keras.utils.load_img('./shuckle.png', target_size=(img_height, img_width))
img = tf.keras.utils.load_img(
'./shuckle.png', target_size=(img_height, img_width)
)
img_array = tf.keras.utils.img_to_array(img)
img_array = tf.expand_dims(img_array, 0) # Create a batch
predictions = model.predict(img_array)
score = tf.nn.softmax(predictions[0])
print(
"This image most likely belongs to {} with a {:.2f} percent confidence."
.format(class_names[np.argmax(score)], 100 * np.max(score))
)1/1 [==============================] - 0s 28ms/step
This image most likely belongs to 213 with a 95.81 percent confidence.