pixray/api 📝 → 🖼️

---> BasePixrayPredictor Predict
Using seed:
11914490315746453492
Working with z of shape (1, 256, 16, 16) = 65536 dimensions.
loaded pretrained LPIPS loss from taming/modules/autoencoder/lpips/vgg.pth
VQLPIPSWithDiscriminator running with hinge loss.
Restored from models/vqgan_wikiart_16384.ckpt
Loaded CLIP RN50x4: 288x288 and 178.30M params
Loaded CLIP ViT-B/32: 224x224 and 151.28M params
Loaded CLIP ViT-B/16: 224x224 and 149.62M params
Using device:
cuda:0
Optimising using:
Adam
Using text prompts:
['a day glo acrylic portrait of a cyberpunk empress, by Yasutomo Oka.']
using custom losses: aesthetic

0it [00:00, ?it/s]
iter: 0, loss: 3.66, losses: 0.843, 0.0817, 0.98, 0.0605, 0.962, 0.0651, 0.67 (-0=>3.663)

0it [00:01, ?it/s]
iter: 10, loss: 3.07, losses: 0.711, 0.0857, 0.839, 0.0636, 0.816, 0.0634, 0.489 (-0=>3.068)

0it [00:26, ?it/s]

0it [00:50, ?it/s]

0it [00:00, ?it/s]
iter: 20, loss: 2.75, losses: 0.646, 0.0879, 0.778, 0.0662, 0.741, 0.0637, 0.368 (-1=>2.728)

0it [00:01, ?it/s]
iter: 30, loss: 2.59, losses: 0.572, 0.0926, 0.728, 0.0686, 0.702, 0.0645, 0.363 (-0=>2.59)

0it [00:28, ?it/s]

0it [00:53, ?it/s]

0it [00:00, ?it/s]
iter: 40, loss: 2.51, losses: 0.555, 0.0916, 0.717, 0.0697, 0.692, 0.0672, 0.319 (-2=>2.475)

0it [00:01, ?it/s]
iter: 50, loss: 2.44, losses: 0.543, 0.0893, 0.708, 0.0702, 0.669, 0.0684, 0.292 (-3=>2.401)

0it [00:28, ?it/s]

0it [00:53, ?it/s]

0it [00:00, ?it/s]
iter: 60, loss: 2.38, losses: 0.534, 0.0881, 0.682, 0.0704, 0.662, 0.0671, 0.273 (-1=>2.305)

0it [00:01, ?it/s]
iter: 70, loss: 2.3, losses: 0.509, 0.0895, 0.671, 0.0727, 0.651, 0.0691, 0.238 (-5=>2.282)

0it [00:28, ?it/s]

0it [00:53, ?it/s]

0it [00:00, ?it/s]
iter: 80, loss: 2.28, losses: 0.504, 0.0912, 0.66, 0.0709, 0.635, 0.07, 0.25 (-4=>2.264)

0it [00:01, ?it/s]
iter: 90, loss: 2.25, losses: 0.488, 0.0919, 0.66, 0.073, 0.623, 0.0719, 0.238 (-7=>2.207)

0it [00:28, ?it/s]

0it [00:53, ?it/s]

0it [00:00, ?it/s]
iter: 100, loss: 2.27, losses: 0.496, 0.0932, 0.658, 0.0729, 0.624, 0.0717, 0.255 (-17=>2.207)

0it [00:01, ?it/s]
iter: 110, loss: 2.28, losses: 0.487, 0.0932, 0.667, 0.0716, 0.631, 0.0702, 0.255 (-8=>2.189)

0it [00:28, ?it/s]

0it [00:53, ?it/s]

0it [00:00, ?it/s]
iter: 120, loss: 2.16, losses: 0.486, 0.0916, 0.633, 0.074, 0.609, 0.0727, 0.195 (-0=>2.16)

0it [00:01, ?it/s]
iter: 130, loss: 2.18, losses: 0.477, 0.0904, 0.641, 0.0747, 0.605, 0.0747, 0.222 (-10=>2.16)

0it [00:28, ?it/s]

0it [00:53, ?it/s]

0it [00:00, ?it/s]
iter: 140, loss: 2.2, losses: 0.467, 0.0919, 0.647, 0.0748, 0.617, 0.0745, 0.223 (-20=>2.16)

0it [00:01, ?it/s]
iter: 150, loss: 2.19, losses: 0.473, 0.0926, 0.641, 0.0736, 0.61, 0.0746, 0.221 (-6=>2.156)

0it [00:28, ?it/s]

0it [00:53, ?it/s]

0it [00:00, ?it/s]
iter: 160, loss: 2.18, losses: 0.484, 0.0902, 0.641, 0.0727, 0.614, 0.0736, 0.206 (-3=>2.136)

0it [00:01, ?it/s]
iter: 170, loss: 2.18, losses: 0.474, 0.0916, 0.638, 0.0722, 0.618, 0.0739, 0.21 (-6=>2.128)

0it [00:28, ?it/s]

0it [00:53, ?it/s]

0it [00:00, ?it/s]
iter: 180, loss: 2.21, losses: 0.462, 0.0913, 0.652, 0.0741, 0.625, 0.0746, 0.228 (-16=>2.128)

0it [00:01, ?it/s]
iter: 190, loss: 2.19, losses: 0.473, 0.0909, 0.637, 0.074, 0.622, 0.0736, 0.219 (-7=>2.124)

0it [00:28, ?it/s]

0it [00:53, ?it/s]

0it [00:00, ?it/s]
iter: 200, loss: 2.15, losses: 0.462, 0.0917, 0.636, 0.0755, 0.611, 0.0747, 0.198 (-2=>2.094)

0it [00:01, ?it/s]
iter: 210, loss: 2.19, losses: 0.461, 0.0909, 0.645, 0.0737, 0.618, 0.074, 0.225 (-12=>2.094)

0it [00:28, ?it/s]
---> BasePixrayPredictor Predict
Using seed:
3383883393288688316
Working with z of shape (1, 256, 16, 16) = 65536 dimensions.
loaded pretrained LPIPS loss from taming/modules/autoencoder/lpips/vgg.pth
VQLPIPSWithDiscriminator running with hinge loss.
Restored from models/vqgan_wikiart_16384.ckpt

0it [00:53, ?it/s]

0it [00:00, ?it/s]
iter: 220, loss: 2.14, losses: 0.468, 0.0934, 0.629, 0.0744, 0.615, 0.0755, 0.18 (-22=>2.094)

0it [00:01, ?it/s]
Loaded CLIP RN50x4: 288x288 and 178.30M params
Loaded CLIP ViT-B/32: 224x224 and 151.28M params
Loaded CLIP ViT-B/16: 224x224 and 149.62M params
Caught SIGTERM, exiting...
Using device:
cuda:0
Optimising using:
Adam
Using text prompts:
['a day glo acrylic portrait of a cyberpunk empress, by Yasutomo Oka.']
using custom losses: aesthetic

0it [00:00, ?it/s]
iter: 0, loss: 3.7, losses: 0.837, 0.0829, 0.982, 0.0609, 0.96, 0.0644, 0.713 (-0=>3.7)

0it [00:01, ?it/s]
iter: 230, loss: 2.17, losses: 0.466, 0.0911, 0.639, 0.074, 0.617, 0.0744, 0.206 (-3=>2.08)

0it [00:28, ?it/s]
iter: 10, loss: 2.79, losses: 0.663, 0.0885, 0.791, 0.0675, 0.783, 0.0647, 0.328 (-0=>2.785)

0it [00:27, ?it/s]

0it [00:53, ?it/s]

0it [00:00, ?it/s]
iter: 240, loss: 2.12, losses: 0.468, 0.0919, 0.631, 0.0749, 0.605, 0.0761, 0.178 (-13=>2.08)

0it [00:01, ?it/s]

0it [00:53, ?it/s]

0it [00:00, ?it/s]
iter: 20, loss: 2.53, losses: 0.585, 0.0894, 0.71, 0.0687, 0.726, 0.0664, 0.287 (-0=>2.533)

0it [00:01, ?it/s]
iter: 250, loss: 2.14, losses: 0.468, 0.0914, 0.629, 0.073, 0.607, 0.075, 0.201 (-23=>2.08)

0it [00:28, ?it/s]
iter: 30, loss: 2.38, losses: 0.519, 0.0932, 0.678, 0.0723, 0.665, 0.0712, 0.281 (-1=>2.379)

0it [00:28, ?it/s]

0it [00:53, ?it/s]

0it [00:00, ?it/s]
iter: 260, loss: 2.19, losses: 0.477, 0.0916, 0.638, 0.0751, 0.609, 0.0741, 0.22 (-33=>2.08)

0it [00:01, ?it/s]
Dropping learning rate

0it [00:53, ?it/s]

0it [00:00, ?it/s]
iter: 40, loss: 2.25, losses: 0.479, 0.0956, 0.659, 0.0732, 0.64, 0.0729, 0.231 (-0=>2.25)

0it [00:01, ?it/s]
iter: 270, loss: 2.07, losses: 0.451, 0.0942, 0.624, 0.076, 0.587, 0.077, 0.159 (-0=>2.069)

0it [00:28, ?it/s]
iter: 50, loss: 2.21, losses: 0.478, 0.0951, 0.65, 0.0719, 0.627, 0.0723, 0.22 (-0=>2.215)

0it [00:28, ?it/s]

0it [00:53, ?it/s]

0it [00:00, ?it/s]
iter: 280, loss: 2.07, losses: 0.462, 0.0913, 0.611, 0.077, 0.592, 0.0763, 0.162 (-9=>2.051)

0it [00:01, ?it/s]

0it [00:53, ?it/s]

0it [00:00, ?it/s]
iter: 60, loss: 2.22, losses: 0.482, 0.0947, 0.641, 0.0727, 0.629, 0.0734, 0.226 (-1=>2.198)

0it [00:01, ?it/s]
iter: 290, loss: 2.04, losses: 0.446, 0.0906, 0.612, 0.0752, 0.59, 0.0778, 0.144 (-0=>2.036)

0it [00:28, ?it/s]
iter: 70, loss: 2.18, losses: 0.463, 0.0964, 0.646, 0.0734, 0.621, 0.0748, 0.202 (-2=>2.149)

0it [00:28, ?it/s]

0it [00:53, ?it/s]

0it [00:00, ?it/s]
iter: 300, loss: 2.11, losses: 0.443, 0.0911, 0.628, 0.074, 0.613, 0.075, 0.187 (-10=>2.036)

0it [00:01, ?it/s]

0it [00:53, ?it/s]

0it [00:00, ?it/s]
iter: 80, loss: 2.18, losses: 0.47, 0.0946, 0.64, 0.0733, 0.613, 0.0755, 0.212 (-4=>2.104)

0it [00:01, ?it/s]
iter: 310, loss: 2.13, losses: 0.468, 0.0939, 0.632, 0.0744, 0.603, 0.0755, 0.184 (-20=>2.036)

0it [00:28, ?it/s]
iter: 90, loss: 2.08, losses: 0.451, 0.0943, 0.617, 0.0746, 0.598, 0.076, 0.174 (-0=>2.085)

0it [00:28, ?it/s]

0it [00:53, ?it/s]

0it [00:00, ?it/s]
iter: 320, loss: 2.09, losses: 0.473, 0.093, 0.615, 0.0746, 0.595, 0.076, 0.163 (-30=>2.036)

0it [00:01, ?it/s]

0it [00:53, ?it/s]

0it [00:00, ?it/s]
iter: 100, loss: 2.12, losses: 0.446, 0.0965, 0.626, 0.0738, 0.602, 0.078, 0.201 (-10=>2.085)

0it [00:01, ?it/s]
iter: 330, loss: 2.09, losses: 0.453, 0.091, 0.623, 0.0753, 0.596, 0.0769, 0.177 (-40=>2.036)

0it [00:28, ?it/s]
iter: 110, loss: 2.18, losses: 0.465, 0.0956, 0.631, 0.0748, 0.616, 0.0743, 0.225 (-20=>2.085)

0it [00:28, ?it/s]

0it [00:53, ?it/s]

0it [00:00, ?it/s]
iter: 340, loss: 2.07, losses: 0.46, 0.0909, 0.608, 0.0768, 0.587, 0.0764, 0.174 (-50=>2.036)

0it [00:01, ?it/s]

0it [00:53, ?it/s]

0it [00:00, ?it/s]
iter: 120, loss: 2.13, losses: 0.449, 0.0964, 0.632, 0.074, 0.603, 0.0759, 0.199 (-2=>2.066)

0it [00:01, ?it/s]
iter: 350, finished (-60=>2.036)

0it [00:27, ?it/s]

0it [00:27, ?it/s]