I'm trying to train an CNN autoencoder model on a pretty big dataset so I'm using tf.data.Dataset. I've been trying to zip the train images to themselves so that I can use the images themselves as targets. I tried doing is like this:
train_dataset_target = tf.keras.preprocessing.image_dataset_from_directory(BASE_TRAIN
,label_mode=None,batch_size=64, image_size = (96, 96))
train_dataset_images = tf.keras.preprocessing.image_dataset_from_directory(BASE_TRAIN
,label_mode=None,batch_size=64, image_size = (96, 96))
train_dataset = tf.data.Dataset.zip((train_dataset_images, train_dataset_target))
And also like this:
train_dataset_target = tf.keras.preprocessing.image_dataset_from_directory(BASE_TRAIN
,label_mode=None,batch_size=64, image_size = (96, 96))
train_dataset = tf.data.Dataset.zip((train_dataset_target, train_dataset_target))
This, however, resulted in a huge loss both times so I visualized couple of examples to see what's the deal. As it turned out, images are not correctly mapped and I'm not sure how to fix that. Any tips?
Edit: As it was requested, here is that print(next(iter(train_dataset.take(1)))) produces:
(<tf.Tensor: shape=(64, 96, 96, 3), dtype=float32, numpy=
array([[[[ 23.918022 , 26.30344 , 14.074273 ],
[ 24.361654 , 26.74707 , 14.517903 ],
[ 23.715603 , 26.10102 , 13.871853 ],
...,
[ 43.131763 , 26.131763 , 16.131763 ],
[ 46.355305 , 29.355307 , 19.355307 ],
[ 41.15631 , 26.578156 , 16.763157 ]],
[[ 21.554363 , 22.716959 , 4.4814453 ],
[ 19.994629 , 21.350586 , 2.463379 ],
[ 23.771809 , 25.584797 , 5.157226 ],
...,
[ 38.281742 , 24.3125 , 15. ],
[ 38.958496 , 25.802246 , 16.489746 ],
[ 42.488426 , 29.332176 , 20.27195 ]],
[[ 44.6021 , 44.6021 , 18.098955 ],
[ 46.815266 , 46.815266 , 19.887207 ],
[ 52.252987 , 52.252987 , 24.32059 ],
...,
[ 40.21609 , 30.359264 , 20.325459 ],
[ 39.332844 , 29.796873 , 20.796873 ],
[ 38.908302 , 29.908302 , 20.908302 ]],
...,
[[ 64.59651 , 71.6835 , 29.318817 ],
[ 59.916687 , 66.978226 , 25.360334 ],
[ 53.713356 , 59.82009 , 21.497223 ],
...,
[109.53851 , 106.53851 , 89.53851 ],
[115.37259 , 112.37259 , 95.37259 ],
[112.60654 , 109.60654 , 92.60654 ]],
[[ 89.68636 , 93.06136 , 60.84261 ],
[ 95.50537 , 98.39258 , 67.149414 ],
[ 95.82389 , 98.35514 , 67.82389 ],
...,
[134.62599 , 133.46974 , 113.313484 ],
[132.64404 , 131.4878 , 111.33154 ],
[132.15625 , 131. , 110.84375 ]],
[[128.258 , 128.258 , 102.25799 ],
[127.007034 , 127.007034 , 101.007034 ],
[125.637 , 125.637 , 99.637 ],
...,
[138.79515 , 136.79515 , 115.02434 ],
[139.58838 , 137.58838 , 115.817566 ],
[140.47763 , 138.47763 , 115.71452 ]]],
[[[ 13.833333 , 13.833333 , 13.833333 ],
[ 13.666666 , 13.666666 , 13.666666 ],
[ 13.666666 , 13.666666 , 13.666666 ],
...,
[ 17.742188 , 17.742188 , 17.742188 ],
[ 17.265625 , 17.265625 , 17.265625 ],
[ 19.481771 , 20.325521 , 19.903646 ]],
[[ 9. , 9. , 9. ],
[ 9. , 9. , 9. ],
[ 9. , 9. , 9. ],
...,
[ 17.5 , 17.5 , 17.5 ],
[ 17.632812 , 17.632812 , 17.632812 ],
[ 18.789062 , 19.632812 , 19.210938 ]],
[[ 11. , 11. , 11. ],
[ 11.9557295 , 11.9557295 , 11.9557295 ],
[ 12. , 12. , 12. ],
...,
[ 21.851562 , 21.851562 , 21.851562 ],
[ 22. , 22. , 22. ],
[ 21.507812 , 22.351562 , 21.929688 ]],
...,
[[ 1.1668091 , 6.018424 , 9.518485 ],
[ 0.46620274, 4.3542995 , 10.023627 ],
[ 0.31514263, 2.497552 , 10.648639 ],
...,
[ 93.25734 , 64.57512 , 36.960796 ],
[ 92.95539 , 63.400787 , 37.25782 ],
[ 89.5597 , 59.393135 , 32.39338 ]],
[[ 85.71094 , 97.21094 , 120.21094 ],
[ 87.10156 , 98.60156 , 121.60156 ],
[ 91.5625 , 103.0625 , 126.0625 ],
...,
[117.33594 , 110.83594 , 110.05469 ],
[118.39844 , 111.63281 , 112.765625 ],
[120.47656 , 112.97656 , 114.47656 ]],
[[ 92.33331 , 108.16666 , 124.80724 ],
[ 91.765594 , 104.72133 , 122.18745 ],
[ 88.73424 , 101.71601 , 119.10401 ],
...,
[175.3202 , 184.82018 , 163.5025 ],
[178.28647 , 187.78644 , 166.72919 ],
[190.29189 , 199.79185 , 179.13564 ]]],
[[[136.4796 , 147.02127 , 149.02127 ],
[135.67708 , 146.21875 , 148.21875 ],
[135.45009 , 145.99174 , 147.99174 ],
...,
[239.9041 , 146.93533 , 171.61241 ],
[241.89714 , 145.9323 , 169.0664 ],
[242.08333 , 145.45833 , 168. ]],
[[130.625 , 137.5 , 140.875 ],
[130.72656 , 137.60156 , 140.97656 ],
[131.97527 , 138.85027 , 142.22527 ],
...,
[237.64717 , 150.75 , 173.6289 ],
[241.92188 , 151.04688 , 173.67188 ],
[242.125 , 151.25 , 172.14583 ]],
[[154.07074 , 159.07074 , 165.07074 ],
[153.86458 , 158.86458 , 164.86458 ],
[155.1354 , 160.1354 , 166.1354 ],
...,
[248.61504 , 165.64627 , 187.32335 ],
[249.64714 , 163.80598 , 183.61848 ],
[248.42708 , 162.0104 , 181.0104 ]],
...,
[[ 5.166626 , 4.166626 , 0.8749695 ],
[ 5.742155 , 4.742155 , 1.162734 ],
[ 7.0629206 , 6.0629206 , 2.0629206 ],
...,
[249.78085 , 241.32336 , 48.229248 ],
[255. , 242.63927 , 61.369778 ],
[243.9404 , 223.47725 , 51.194305 ]],
[[ 3.5 , 5.5 , 1.25 ],
[ 3.90625 , 5.90625 , 1.5039062 ],
[ 4.7539062 , 6.7539062 , 1.875 ],
...,
[254.23828 , 244.74867 , 63.529976 ],
[251.95703 , 227.65234 , 60.171875 ],
[228.58456 , 190.94785 , 32.588474 ]],
[[ 45.688103 , 47.688103 , 42.6881 ],
[ 46.89685 , 48.89685 , 43.89685 ],
[ 48.45112 , 50.45112 , 45.45112 ],
...,
[254.65016 , 234.57927 , 66.06678 ],
[240.5571 , 196.00754 , 44.132603 ],
[214.19675 , 154.57162 , 13.123365 ]]],
...,
[[[198.8639 , 199.8639 , 191.8639 ],
[198.09895 , 199.09895 , 191.09895 ],
[198.09895 , 199.09895 , 191.09895 ],
...,
[189.7552 , 190.7552 , 182.7552 ],
[187.90625 , 188.90625 , 180.90625 ],
[187. , 188. , 180. ]],
[[197.25203 , 198.95515 , 188.84578 ],
[197. , 198.70312 , 188.59375 ],
[197. , 198.70312 , 188.59375 ],
...,
[189.7552 , 190.7552 , 182.7552 ],
[187.90625 , 188.90625 , 180.90625 ],
[187. , 188. , 180. ]],
[[197. , 199. , 188. ],
[197. , 199. , 188. ],
[197. , 199. , 188. ],
...,
[189.7552 , 190.7552 , 182.7552 ],
[187.90625 , 188.90625 , 180.90625 ],
[187. , 188. , 180. ]],
...,
[[174.13017 , 177.13017 , 166.13017 ],
[175.43225 , 178.43225 , 167.43225 ],
[175.22392 , 178.22392 , 167.22392 ],
...,
[185.31621 , 187.31621 , 176.31621 ],
[189.62546 , 191.62546 , 180.62546 ],
[190.97185 , 192.97185 , 181.97185 ]],
[[180.14787 , 183.14787 , 172.14787 ],
[182.04688 , 185.04688 , 174.04688 ],
[182.06274 , 185.06274 , 174.06274 ],
...,
[186.26561 , 188.26561 , 177.26561 ],
[184.54688 , 186.54688 , 175.54688 ],
[186.69792 , 188.69792 , 177.69792 ]],
[[188.86389 , 191.86389 , 180.86389 ],
[187.64581 , 190.64581 , 179.64581 ],
[187.02422 , 190.02422 , 179.02422 ],
...,
[186.26561 , 188.26561 , 177.26561 ],
[184.54688 , 186.54688 , 175.54688 ],
[186.69792 , 188.69792 , 177.69792 ]]],
[[[172.05588 , 164.05588 , 39.395832 ],
[168.01042 , 161.22917 , 35.822918 ],
[165.99815 , 160.99815 , 34.998154 ],
...,
[147.15811 , 134.15811 , 116.78682 ],
[147. , 134. , 118. ],
[145.64583 , 132.64583 , 116.645836 ]],
[[183.60417 , 176.49902 , 46.0625 ],
[169.80664 , 163.30371 , 34.68164 ],
[165.46875 , 160.46875 , 31.53125 ],
...,
[158.17352 , 145.17352 , 128.17352 ],
[157.51465 , 144.19922 , 129.14551 ],
[156.59473 , 143.59473 , 127.501945 ]],
[[185.28125 , 178.28125 , 45.28125 ],
[179.07292 , 172.07292 , 39.479168 ],
[170.08572 , 165.08572 , 35.088108 ],
...,
[170.55446 , 155.67023 , 131.97916 ],
[175.5931 , 160.70769 , 139.02019 ],
[172.25185 , 157.36642 , 135.63358 ]],
...,
[[ 43.13727 , 73.90809 , 43.02268 ],
[ 48.114594 , 74.88541 , 43.90689 ],
[ 48.98036 , 73.8866 , 41.62855 ],
...,
[ 99.86578 , 113.86578 , 80.86578 ],
[ 98.22919 , 109.22919 , 76.04169 ],
[ 99.42711 , 109.42711 , 75.42711 ]],
[[ 44. , 80.0625 , 48.59375 ],
[ 46.96289 , 78.59375 , 47.40332 ],
[ 44.9847 , 75.05697 , 43.5 ],
...,
[ 96.05697 , 108.46322 , 73.99447 ],
[ 97.99707 , 109.56836 , 75.37793 ],
[ 99.50098 , 110.50098 , 76.50098 ]],
[[ 42.052094 , 81.731995 , 50.392044 ],
[ 43.61264 , 78.82292 , 47.90528 ],
[ 50.34374 , 84.34374 , 51.34374 ],
...,
[103.147705 , 113.147705 , 79.147705 ],
[104.33431 , 114.33431 , 80.33431 ],
[105. , 115. , 81. ]]],
[[[ 13. , 0. , 0. ],
[ 14.78125 , 7.125 , 1.78125 ],
[ 15.817708 , 16.177082 , 12.630207 ],
...,
[ 5.0937347 , 6.0937347 , 10.276024 ],
[ 10.328125 , 11.109375 , 16.4375 ],
[ 13. , 12. , 20. ]],
[[ 12.21875 , 0.390625 , 0. ],
[ 12.608398 , 5.4282227 , 2.1291504 ],
[ 12.196289 , 13.585123 , 12.701415 ],
...,
[ 13.044601 , 14.044601 , 17.516846 ],
[ 11.890625 , 12.671875 , 17.56665 ],
[ 14.5625 , 13.5625 , 20.78125 ]],
[[ 11.03125 , 0.984375 , 0. ],
[ 9.305664 , 2.849121 , 2.657959 ],
[ 6.6917315 , 9.645344 , 12.80965 ],
...,
[ 25.129917 , 26.129917 , 28.522898 ],
[ 14.265625 , 15.046875 , 19.282959 ],
[ 16.9375 , 15.9375 , 21.96875 ]],
...,
[[ 78.984375 , 28.015625 , 23.96875 ],
[ 73.710205 , 24.522705 , 17.803955 ],
[ 71.41431 , 26.546875 , 16.178059 ],
...,
[ 53.700283 , 31.090902 , 24.500008 ],
[ 50.692627 , 30.536377 , 21.289795 ],
[ 46.28125 , 26.125 , 15.125 ]],
[[ 78.390625 , 28.609375 , 22.78125 ],
[ 75.7605 , 27.760498 , 19.260498 ],
[ 73.303795 , 30.109375 , 16.017252 ],
...,
[ 63.09205 , 34.545166 , 26.281258 ],
[ 59.7937 , 33.69995 , 23.395752 ],
[ 56.96875 , 30.875 , 19.875 ]],
[[ 78. , 29. , 22. ],
[ 77.109375 , 29.890625 , 20.21875 ],
[ 74.546875 , 32.453125 , 15.911459 ],
...,
[ 69.27084 , 36.81771 , 27.453133 ],
[ 65.78125 , 35.78125 , 24.78125 ],
[ 64. , 34. , 23. ]]]],
dtype=float32)>, <tf.Tensor: shape=(64, 96, 96, 3),
dtype=float32, numpy=
array([[[[5.00000000e 00, 4.00000000e 00, 2.00000000e 00],
[5.00000000e 00, 4.00000000e 00, 2.00000000e 00],
[5.81250000e 00, 4.81250000e 00, 2.81250000e 00],
...,
[2.81250000e 00, 8.12500000e-01, 1.81250000e 00],
[2.00000000e 00, 0.00000000e 00, 1.00000000e 00],
[2.00000000e 00, 0.00000000e 00, 1.00000000e 00]],
[[9.75000000e 00, 8.75000000e 00, 6.75000000e 00],
[9.75000000e 00, 8.75000000e 00, 6.75000000e 00],
[1.29746094e 01, 1.19746094e 01, 9.97460938e 00],
...,
[9.42724609e 00, 7.42724609e 00, 8.42724609e 00],
[1.09062500e 01, 8.90625000e 00, 9.90625000e 00],
[1.09062500e 01, 8.90625000e 00, 9.90625000e 00]],
[[1.94583359e 01, 1.81354179e 01, 1.71041679e 01],
[1.94583359e 01, 1.81354179e 01, 1.71041679e 01],
[2.34806328e 01, 2.21577168e 01, 2.11264668e 01],
...,
[2.17687187e 01, 1.97687187e 01, 2.07687187e 01],
[2.47500019e 01, 2.27500019e 01, 2.37500019e 01],
[2.47500019e 01, 2.27500019e 01, 2.37500019e 01]],
...,
[[1.22500122e 02, 1.21500122e 02, 1.19500122e 02],
[1.22500122e 02, 1.21500122e 02, 1.19500122e 02],
[1.23723122e 02, 1.22723122e 02, 1.20723122e 02],
...,
[6.47745361e 01, 6.47745361e 01, 6.27745399e 01],
[6.08749695e 01, 6.08749695e 01, 5.88749695e 01],
[6.08749695e 01, 6.08749695e 01, 5.88749695e 01]],
[[1.45312500e 02, 1.44312500e 02, 1.42312500e 02],
[1.45312500e 02, 1.44312500e 02, 1.42312500e 02],
[1.47692871e 02, 1.46692871e 02, 1.44692871e 02],
...,
[9.76440430e 01, 9.76440430e 01, 9.56440430e 01],
[9.60000000e 01, 9.60000000e 01, 9.40000000e 01],
[9.60000000e 01, 9.60000000e 01, 9.40000000e 01]],
[[1.56000000e 02, 1.55000000e 02, 1.53000000e 02],
[1.56000000e 02, 1.55000000e 02, 1.53000000e 02],
[1.56812500e 02, 1.55812500e 02, 1.53812500e 02],
...,
[1.52593750e 02, 1.52593750e 02, 1.50593750e 02],
[1.53000000e 02, 1.53000000e 02, 1.51000000e 02],
[1.53000000e 02, 1.53000000e 02, 1.51000000e 02]]],
[[[1.35819336e 02, 1.73053711e 02, 1.69329758e 02],
[1.53781250e 02, 1.85254883e 02, 1.82736816e 02],
[1.44442871e 02, 1.77336426e 02, 1.78929031e 02],
...,
[1.76788788e 02, 2.17132538e 02, 2.43813675e 02],
[1.74809082e 02, 2.15152832e 02, 2.43215332e 02],
[1.65743149e 02, 2.06909836e 02, 2.34702301e 02]],
[[1.06315598e 02, 1.49159348e 02, 1.32346848e 02],
[1.07478027e 02, 1.48297852e 02, 1.32557129e 02],
[1.14144371e 02, 1.57785172e 02, 1.45116531e 02],
...,
[2.12427185e 02, 2.49489685e 02, 2.54968750e 02],
[2.10868652e 02, 2.47931152e 02, 2.54992676e 02],
[2.10817719e 02, 2.47880219e 02, 2.54768387e 02]],
[[1.87158203e 01, 8.06440430e 01, 5.01336250e 01],
[2.25864258e 01, 8.75239258e 01, 5.28364258e 01],
[4.75966873e 01, 1.14003914e 02, 8.01753006e 01],
...,
[2.14306320e 02, 2.43801132e 02, 2.46281250e 02],
[2.18003418e 02, 2.46909668e 02, 2.48812500e 02],
[2.17134277e 02, 2.45803223e 02, 2.49093750e 02]],
...,
[[1.65592941e 02, 1.96327316e 02, 1.69833664e 02],
[1.74614746e 02, 2.07114746e 02, 1.79302246e 02],
[1.53045563e 02, 1.93571106e 02, 1.60940903e 02],
...,
[0.00000000e 00, 3.94683151e 01, 4.11930084e-02],
[1.22426758e 01, 7.24301758e 01, 2.62739258e 01],
[3.40232964e 01, 9.13993988e 01, 4.44983673e 01]],
[[0.00000000e 00, 1.18253584e 01, 0.00000000e 00],
[6.79589844e 00, 1.87114258e 01, 5.32910156e 00],
[2.10047340e 01, 4.22791557e 01, 2.19179840e 01],
...,
[2.26249371e 01, 9.69895630e 01, 4.79841881e 01],
[3.05439453e 01, 1.10661621e 02, 5.76782227e 01],
[4.26578827e 01, 1.11423477e 02, 5.78717308e 01]],
[[0.00000000e 00, 9.65625000e 00, 1.65625000e 00],
[1.31250000e 00, 8.31250000e 00, 6.56250000e-01],
[0.00000000e 00, 1.73958397e 01, 0.00000000e 00],
...,
[5.79341049e 01, 1.45384323e 02, 9.56216049e 01],
[8.28759766e 01, 1.69545410e 02, 1.16208984e 02],
[1.05831863e 02, 1.71983673e 02, 1.21395149e 02]]],
[[[5.48266068e 01, 5.08266068e 01, 2.46495228e 01],
[5.50257149e 01, 5.09007149e 01, 2.47861328e 01],
[5.54114571e 01, 4.94114571e 01, 2.41486549e 01],
...,
[5.48154640e 01, 5.06071205e 01, 3.90486298e 01],
[2.87747402e 01, 2.53264980e 01, 1.54042969e 01],
[2.67753925e 01, 2.41697044e 01, 1.39439020e 01]],
[[5.42037773e 01, 5.22037773e 01, 2.96725254e 01],
[5.28134766e 01, 5.06259766e 01, 2.81279297e 01],
[5.36598320e 01, 4.87395821e 01, 2.65003262e 01],
...,
[6.91130219e 01, 6.54847565e 01, 4.91549835e 01],
[4.29394531e 01, 3.85332031e 01, 2.54082031e 01],
[2.67773819e 01, 2.29971066e 01, 1.02890978e 01]],
[[5.70909309e 01, 5.50909309e 01, 3.40885429e 01],
[5.32864571e 01, 5.11025391e 01, 2.99218750e 01],
[5.34947891e 01, 4.86502800e 01, 2.61718750e 01],
...,
[9.29087906e 01, 8.87532959e 01, 6.31134071e 01],
[6.43567657e 01, 5.94693985e 01, 3.88740234e 01],
[3.28602104e 01, 2.78614044e 01, 8.66630650e 00]],
...,
[[8.08130875e 01, 6.68130875e 01, 5.58390427e 00],
[8.23523483e 01, 6.82309265e 01, 6.93566322e 00],
[7.84802856e 01, 6.25375786e 01, 1.24501038e 00],
...,
[1.17640701e 02, 1.07686066e 02, 3.86102180e 01],
[1.22071564e 02, 1.12958923e 02, 4.79309502e 01],
[1.28132507e 02, 1.18152161e 02, 5.71857948e 01]],
[[1.14656898e 02, 1.02953773e 02, 2.72350273e 01],
[1.15920898e 02, 1.04107422e 02, 2.83740234e 01],
[1.13113281e 02, 9.96445312e 01, 2.37158203e 01],
...,
[1.53180634e 02, 1.47083618e 02, 5.86301727e 01],
[1.58908203e 02, 1.52472656e 02, 6.93935547e 01],
[1.51753281e 02, 1.45055038e 02, 6.45241013e 01]],
[[1.78467621e 02, 1.69653259e 02, 7.47507935e 01],
[1.78290375e 02, 1.69790680e 02, 7.39831467e 01],
[1.76013794e 02, 1.66190857e 02, 6.89797974e 01],
...,
[1.67686234e 02, 1.65066422e 02, 5.65305099e 01],
[1.67636383e 02, 1.65390915e 02, 6.15788460e 01],
[1.64538742e 02, 1.61901443e 02, 6.07888184e 01]]],
...,
[[[1.17546875e 02, 1.02395836e 02, 7.93958359e 01],
[1.15453125e 02, 9.67343750e 01, 7.46406250e 01],
[1.12223961e 02, 8.79583282e 01, 6.77135391e 01],
...,
[8.84687805e 01, 7.10000000e 01, 4.86927643e 01],
[9.30000000e 01, 7.10000000e 01, 5.75468750e 01],
[9.30000000e 01, 6.84531174e 01, 6.05468826e 01]],
[[1.11274094e 02, 9.60262833e 01, 7.44042969e 01],
[1.14111816e 02, 9.47524414e 01, 7.45805664e 01],
[1.13518555e 02, 8.90961075e 01, 6.93217773e 01],
...,
[1.01921906e 02, 8.36423416e 01, 6.13484421e 01],
[1.07153809e 02, 8.54440918e 01, 7.07697754e 01],
[1.07190506e 02, 8.31874924e 01, 7.24651718e 01]],
[[1.11863800e 02, 9.73638000e 01, 7.72563477e 01],
[1.14675293e 02, 9.57586288e 01, 7.70659180e 01],
[1.14547661e 02, 9.08393326e 01, 7.20328217e 01],
...,
[1.10582077e 02, 9.21505585e 01, 6.88656998e 01],
[1.15131348e 02, 9.38037949e 01, 7.66209335e 01],
[1.16233017e 02, 9.29361420e 01, 7.89693222e 01]],
...,
[[3.34529076e 01, 1.27744722e 01, 0.00000000e 00],
[1.00004509e 02, 6.81486435e 01, 2.02915249e 01],
[1.59042236e 02, 1.13147179e 02, 3.63206062e 01],
...,
[1.95015884e 02, 1.39584366e 02, 1.22070408e 01],
[1.90469879e 02, 1.32253387e 02, 1.25335159e 01],
[1.83808060e 02, 1.25580246e 02, 1.97625771e 01]],
[[3.33319511e 01, 1.35767422e 01, 0.00000000e 00],
[1.08526611e 02, 7.80554199e 01, 2.51909180e 01],
[1.80072845e 02, 1.35266113e 02, 4.96700821e 01],
...,
[1.85928131e 02, 1.29474991e 02, 8.25048065e 00],
[1.79291992e 02, 1.19676025e 02, 6.69653320e 00],
[1.69351471e 02, 1.09841049e 02, 1.23887711e 01]],
[[3.36822929e 01, 1.39270840e 01, 0.00000000e 00],
[1.13921875e 02, 8.40312500e 01, 2.80937500e 01],
[1.97880219e 02, 1.54838547e 02, 5.89791641e 01],
...,
[1.77796829e 02, 1.20062439e 02, 6.04162598e 00],
[1.64437500e 02, 1.03890625e 02, 0.00000000e 00],
[1.46265594e 02, 8.61145477e 01, 0.00000000e 00]]],
[[[1.05939857e 02, 7.22836075e 01, 5.32836113e 01],
[1.07194580e 02, 7.39687500e 01, 5.38977051e 01],
[1.07669106e 02, 7.45260391e 01, 5.18996582e 01],
...,
[1.22373611e 02, 8.03892365e 01, 7.67173615e 01],
[1.18265625e 02, 7.36093750e 01, 7.29375000e 01],
[1.16455963e 02, 6.99687195e 01, 7.16562500e 01]],
[[1.08517090e 02, 6.81979141e 01, 4.99720879e 01],
[1.06988770e 02, 6.89943848e 01, 4.87331543e 01],
[1.08728104e 02, 7.07281036e 01, 4.77572403e 01],
...,
[1.22051018e 02, 7.42978287e 01, 7.22155609e 01],
[1.21400635e 02, 7.20568848e 01, 7.50268555e 01],
[1.23591896e 02, 7.35952377e 01, 7.65952377e 01]],
[[1.17069092e 02, 7.01458359e 01, 5.41458321e 01],
[1.14718750e 02, 6.91291504e 01, 5.12072754e 01],
[1.12810791e 02, 6.87083359e 01, 4.95052071e 01],
...,
[1.28756989e 02, 7.81597366e 01, 7.71597366e 01],
[1.27769775e 02, 7.34604492e 01, 7.62302246e 01],
[1.30496201e 02, 7.54962006e 01, 8.00691376e 01]],
...,
[[1.39930908e 02, 8.13684082e 01, 9.22902832e 01],
[1.34566650e 02, 7.49531250e 01, 8.62854004e 01],
[1.28079834e 02, 6.75815430e 01, 7.92360840e 01],
...,
[1.24175446e 02, 8.33073654e 01, 9.20643539e 01],
[1.25202148e 02, 8.41281738e 01, 9.38698730e 01],
[1.24568748e 02, 8.34423599e 01, 9.34347229e 01]],
[[1.23842285e 02, 8.04531250e 01, 8.70416641e 01],
[1.24983154e 02, 7.87644043e 01, 8.57019043e 01],
[1.26321777e 02, 7.91134415e 01, 8.53321915e 01],
...,
[1.30121902e 02, 1.18215645e 02, 1.22215645e 02],
[1.30850586e 02, 1.18944336e 02, 1.22944336e 02],
[1.36539169e 02, 1.24632927e 02, 1.28632919e 02]],
[[1.53647705e 02, 1.28350830e 02, 1.37694580e 02],
[1.53113525e 02, 1.27606445e 02, 1.36950195e 02],
[1.59272629e 02, 1.33466721e 02, 1.42878174e 02],
...,
[1.32164215e 02, 1.36432617e 02, 1.37134354e 02],
[1.31789795e 02, 1.35288330e 02, 1.36375000e 02],
[1.34850952e 02, 1.38153503e 02, 1.39338165e 02]]],
[[[1.90000000e 01, 1.50000000e 01, 1.60000000e 01],
[1.90000000e 01, 1.50000000e 01, 1.60000000e 01],
[1.92968750e 01, 1.54947920e 01, 1.64947910e 01],
...,
[2.25937538e 01, 1.67916718e 01, 1.86927128e 01],
[2.20000000e 01, 1.60000000e 01, 1.80000000e 01],
[2.20000000e 01, 1.60000000e 01, 1.80000000e 01]],
[[2.05000000e 01, 1.75000000e 01, 1.85000000e 01],
[2.05000000e 01, 1.75000000e 01, 1.85000000e 01],
[2.03515625e 01, 1.75494785e 01, 1.85000000e 01],
...,
[2.34947948e 01, 1.76927128e 01, 1.95937538e 01],
[2.30000000e 01, 1.70000000e 01, 1.90000000e 01],
[2.30000000e 01, 1.70000000e 01, 1.90000000e 01]],
[[2.10000000e 01, 2.00000000e 01, 2.06666660e 01],
[2.10000000e 01, 2.00000000e 01, 2.06666660e 01],
[2.05546875e 01, 1.96866322e 01, 2.02708340e 01],
...,
[2.43298626e 01, 1.84947948e 01, 2.04453144e 01],
[2.40000000e 01, 1.80000000e 01, 2.00000000e 01],
[2.40000000e 01, 1.80000000e 01, 2.00000000e 01]],
...,
[[1.69000031e 02, 1.74000031e 02, 1.67166687e 02],
[1.69000031e 02, 1.74000031e 02, 1.67166687e 02],
[1.71144135e 02, 1.76144135e 02, 1.69310791e 02],
...,
[1.35512199e 01, 2.40989590e 01, 2.23020897e 01],
[1.36666718e 01, 2.40000000e 01, 2.25000076e 01],
[1.36666718e 01, 2.40000000e 01, 2.25000076e 01]],
[[1.76500000e 02, 1.81500000e 02, 1.74500000e 02],
[1.76500000e 02, 1.81500000e 02, 1.74500000e 02],
[1.78677094e 02, 1.83677094e 02, 1.76677094e 02],
...,
[1.52031231e 01, 2.48515625e 01, 2.37031231e 01],
[1.55000000e 01, 2.50000000e 01, 2.40000000e 01],
[1.55000000e 01, 2.50000000e 01, 2.40000000e 01]],
[[1.82000000e 02, 1.87000000e 02, 1.80000000e 02],
[1.82000000e 02, 1.87000000e 02, 1.80000000e 02],
[1.84078125e 02, 1.89078125e 02, 1.82078125e 02],
...,
[1.65052052e 01, 2.56041641e 01, 2.46041641e 01],
[1.70000000e 01, 2.60000000e 01, 2.50000000e 01],
[1.70000000e 01, 2.60000000e 01, 2.50000000e 01]]]],
dtype=float32)>)
CodePudding user response：
Maybe set the shuffle parameter of tf.keras.preprocessing.image_dataset_from_directory to False for train_dataset_target and train_dataset_images and call shuffle after zipping:
train_dataset = tf.data.Dataset.zip((train_dataset_target, train_dataset_target)).shuffle(some_buffer_size)