| === EXP 1 (arch D x ds{1,2,4,8}) start 2026-07-15T11:24:11+00:00 === |
| [main] device=cuda |
| [main] paired train=11610 val=3582 |
| Setting up [LPIPS] perceptual loss: trunk [alex], v[0.1], spatial [off] |
| /usr/local/lib/python3.12/dist-packages/torchvision/models/_utils.py:208: UserWarning: The parameter 'pretrained' is deprecated since 0.13 and may be removed in the future, please use 'weights' instead. |
| warnings.warn( |
| /usr/local/lib/python3.12/dist-packages/torchvision/models/_utils.py:223: UserWarning: Arguments other than a weight enum or `None` for 'weights' are deprecated since 0.13 and may be removed in the future. The current behavior is equivalent to passing `weights=AlexNet_Weights.IMAGENET1K_V1`. You can also use `weights=AlexNet_Weights.DEFAULT` to get the most up-to-date weights. |
| warnings.warn(msg) |
| Loading model from: /home/ubuntu/.local/lib/python3.12/site-packages/lpips/weights/v0.1/alex.pth |
| [arch=D ds=1] params = 0.61M |
| [arch=D ds=1] step 500 val_l1=0.1196 val_lpips=0.2203 best=0.2203 (252s) |
| [arch=D ds=1] step 1000 val_l1=0.1181 val_lpips=0.2189 best=0.2189 (507s) |
| [arch=D ds=1] step 1500 val_l1=0.1180 val_lpips=0.2187 best=0.2187 (760s) |
| [arch=D ds=1] step 2000 val_l1=0.1179 val_lpips=0.2175 best=0.2175 (1024s) |
| [arch=D ds=1] step 2500 val_l1=0.1183 val_lpips=0.2125 best=0.2125 (1282s) |
| [arch=D ds=1] step 3000 val_l1=0.1187 val_lpips=0.2115 best=0.2115 (1541s) |
| [arch=D ds=1] step 3500 val_l1=0.1180 val_lpips=0.2153 best=0.2115 (1796s) |
| [arch=D ds=1] step 4000 val_l1=0.1219 val_lpips=0.2161 best=0.2115 (2047s) |
| [arch=D ds=1] step 4500 val_l1=0.1182 val_lpips=0.2142 best=0.2115 (2300s) |
| [arch=D ds=1] step 5000 val_l1=0.1143 val_lpips=0.2105 best=0.2105 (2548s) |
| [arch=D ds=1] step 5500 val_l1=0.1177 val_lpips=0.2127 best=0.2105 (2801s) |
| [arch=D ds=1] step 6000 val_l1=0.1162 val_lpips=0.2098 best=0.2098 (3053s) |
| [arch=D ds=1] step 6500 val_l1=0.1200 val_lpips=0.2178 best=0.2098 (3306s) |
| [arch=D ds=1] step 7000 val_l1=0.1179 val_lpips=0.2092 best=0.2092 (3571s) |
| [arch=D ds=1] step 7500 val_l1=0.1182 val_lpips=0.2118 best=0.2092 (3838s) |
| [arch=D ds=1] step 8000 val_l1=0.1160 val_lpips=0.2117 best=0.2092 (4094s) |
| [arch=D ds=1] DONE. best_val_lpips=0.2092 final=0.2117 elapsed=4142s |
| [arch=D ds=2] params = 0.61M |
| [arch=D ds=2] step 500 val_l1=0.1199 val_lpips=0.2292 best=0.2292 (283s) |
| [arch=D ds=2] step 1000 val_l1=0.1180 val_lpips=0.2255 best=0.2255 (556s) |
| [arch=D ds=2] step 1500 val_l1=0.1180 val_lpips=0.2217 best=0.2217 (826s) |
| [arch=D ds=2] step 2000 val_l1=0.1182 val_lpips=0.2213 best=0.2213 (1084s) |
| [arch=D ds=2] step 2500 val_l1=0.1184 val_lpips=0.2143 best=0.2143 (1371s) |
| [arch=D ds=2] step 3000 val_l1=0.1185 val_lpips=0.2132 best=0.2132 (1649s) |
| === EXP 1 CONT (arch D x ds{4,8}, ds=2 killed early @ step 3000 best=0.2132) start 2026-07-15T13:02:54+00:00 === |
| [main] device=cuda |
| [main] paired train=11610 val=3582 |
| Setting up [LPIPS] perceptual loss: trunk [alex], v[0.1], spatial [off] |
| /usr/local/lib/python3.12/dist-packages/torchvision/models/_utils.py:208: UserWarning: The parameter 'pretrained' is deprecated since 0.13 and may be removed in the future, please use 'weights' instead. |
| warnings.warn( |
| /usr/local/lib/python3.12/dist-packages/torchvision/models/_utils.py:223: UserWarning: Arguments other than a weight enum or `None` for 'weights' are deprecated since 0.13 and may be removed in the future. The current behavior is equivalent to passing `weights=AlexNet_Weights.IMAGENET1K_V1`. You can also use `weights=AlexNet_Weights.DEFAULT` to get the most up-to-date weights. |
| warnings.warn(msg) |
| Loading model from: /home/ubuntu/.local/lib/python3.12/site-packages/lpips/weights/v0.1/alex.pth |
| [arch=D ds=4] params = 0.61M |
| [arch=D ds=4] step 500 val_l1=0.1216 val_lpips=0.2638 best=0.2638 (270s) |
| [arch=D ds=4] step 1000 val_l1=0.1195 val_lpips=0.2586 best=0.2586 (539s) |
| [arch=D ds=4] step 1500 val_l1=0.1197 val_lpips=0.2487 best=0.2487 (810s) |
| [arch=D ds=4] step 2000 val_l1=0.1198 val_lpips=0.2511 best=0.2487 (1070s) |
| [arch=D ds=4] step 2500 val_l1=0.1200 val_lpips=0.2427 best=0.2427 (1326s) |
| [arch=D ds=4] step 3000 val_l1=0.1198 val_lpips=0.2398 best=0.2398 (1595s) |
| [arch=D ds=4] step 3500 val_l1=0.1184 val_lpips=0.2442 best=0.2398 (1861s) |
| [arch=D ds=4] step 4000 val_l1=0.1190 val_lpips=0.2403 best=0.2398 (2104s) |
| [arch=D ds=4] step 4500 val_l1=0.1199 val_lpips=0.2412 best=0.2398 (2316s) |
| [arch=D ds=4] step 5000 val_l1=0.1161 val_lpips=0.2361 best=0.2361 (2550s) |
| [arch=D ds=4] step 5500 val_l1=0.1182 val_lpips=0.2348 best=0.2348 (2774s) |
| [arch=D ds=4] step 6000 val_l1=0.1199 val_lpips=0.2352 best=0.2348 (3014s) |
| [arch=D ds=4] step 6500 val_l1=0.1199 val_lpips=0.2363 best=0.2348 (3262s) |
|
|
