Add files using upload-large-folder tool

scripts/eval/eval_mscoco7.sh

@@ -0,0 +1,15 @@
+#!/bin/sh
+cd /mnt/petrelfs/leihaodong/ICLR25/eval
+# HF_ENDPOINT=https://hf-mirror.com
+dataset_anno_file=/mnt/petrelfs/leihaodong/ICLR25/sjdtree/data/prompts/captions_val2017.json
+workdir=/mnt/petrelfs/leihaodong/ICLR25/exp/Emu3/MSCOCO2017Val/sjd/img
+
+srun \
+ -N 1 \
+ --ntasks-per-node 1 \
+ python evaluation_metrics.py \
+  --workdir $workdir \
+  --dataset_anno_file $dataset_anno_file \
+  --coco_gt_path /mnt/petrelfs/leihaodong/local_dataset/val2017 \
+  --dataset_name coco \
+  --n_images 0

scripts/eval/eval_pp.sh

@@ -0,0 +1,14 @@
+#!/bin/sh
+cd /mnt/petrelfs/leihaodong/ICLR25/eval
+# HF_ENDPOINT=https://hf-mirror.com
+dataset_anno_file=/mnt/petrelfs/leihaodong/ICLR25/sjdtree/data/prompts/PartiPrompts.tsv
+workdir=/mnt/petrelfs/leihaodong/ICLR25/exp/Lumina-mGPT-7B-768_PartiPrompts/mscr_sjd_0.005/img
+
+srun \
+ -N 1 \
+ --ntasks-per-node 1 \
+ python evaluation_metrics.py \
+  --workdir $workdir \
+  --dataset_anno_file $dataset_anno_file \
+  --dataset_name parti_cocoformat \
+  --n_images 0

scripts/eval/eval_pp1.sh

@@ -0,0 +1,14 @@
+#!/bin/sh
+cd /mnt/petrelfs/leihaodong/ICLR25/eval
+# HF_ENDPOINT=https://hf-mirror.com
+dataset_anno_file=/mnt/petrelfs/leihaodong/ICLR25/sjdtree/data/prompts/PartiPrompts.tsv
+workdir=/mnt/petrelfs/leihaodong/ICLR25/exp/Lumina-mGPT-7B-768_PartiPrompts/mscr_sjd_0.01/img
+
+srun \
+ -N 1 \
+ --ntasks-per-node 1 \
+ python evaluation_metrics.py \
+  --workdir $workdir \
+  --dataset_anno_file $dataset_anno_file \
+  --dataset_name parti_cocoformat \
+  --n_images 0

scripts/eval/eval_pp5.sh

@@ -0,0 +1,14 @@
+#!/bin/sh
+cd /mnt/petrelfs/leihaodong/ICLR25/eval
+# HF_ENDPOINT=https://hf-mirror.com
+dataset_anno_file=/mnt/petrelfs/leihaodong/ICLR25/sjdtree/data/prompts/PartiPrompts.tsv
+workdir=/mnt/petrelfs/leihaodong/ICLR25/exp/Emu3/PartiPrompts/mscr_sjd_0.05/img
+
+srun \
+ -N 1 \
+ --ntasks-per-node 1 \
+ python evaluation_metrics.py \
+  --workdir $workdir \
+  --dataset_anno_file $dataset_anno_file \
+  --dataset_name parti_cocoformat \
+  --n_images 0

scripts/eval/eval_pp6.sh

@@ -0,0 +1,14 @@
+#!/bin/sh
+cd /mnt/petrelfs/leihaodong/ICLR25/eval
+# HF_ENDPOINT=https://hf-mirror.com
+dataset_anno_file=/mnt/petrelfs/leihaodong/ICLR25/sjdtree/data/prompts/PartiPrompts.tsv
+workdir=/mnt/petrelfs/leihaodong/ICLR25/exp/Emu3/PartiPrompts/sjd/img
+
+srun \
+ -N 1 \
+ --ntasks-per-node 1 \
+ python evaluation_metrics.py \
+  --workdir $workdir \
+  --dataset_anno_file $dataset_anno_file \
+  --dataset_name parti_cocoformat \
+  --n_images 0

scripts/eval/eval_t2i.sh

@@ -0,0 +1,14 @@
+#!/bin/sh
+cd /mnt/petrelfs/leihaodong/ICLR25/eval
+# HF_ENDPOINT=https://hf-mirror.com
+dataset_anno_file=/mnt/petrelfs/leihaodong/ICLR25/sjdtree/data/prompts/T2I-CompBench_val.json
+workdir=/mnt/petrelfs/leihaodong/ICLR25/exp/Lumina-mGPT-7B-768_T2ICompBenchVal/mscr_gsd/img
+
+srun \
+ -N 1 \
+ --ntasks-per-node 1 \
+ python evaluation_metrics.py \
+  --workdir $workdir \
+  --dataset_anno_file $dataset_anno_file \
+  --dataset_name T2I-CompBench \
+  --n_images 0

scripts/eval/eval_t2i1.sh

@@ -0,0 +1,14 @@
+#!/bin/sh
+cd /mnt/petrelfs/leihaodong/ICLR25/eval
+# HF_ENDPOINT=https://hf-mirror.com
+dataset_anno_file=/mnt/petrelfs/leihaodong/ICLR25/sjdtree/data/prompts/T2I-CompBench_val.json
+workdir=/mnt/petrelfs/leihaodong/ICLR25/exp/Lumina-mGPT-7B-768_T2ICompBenchVal/ms_gsd/img
+
+srun \
+ -N 1 \
+ --ntasks-per-node 1 \
+ python evaluation_metrics.py \
+  --workdir $workdir \
+  --dataset_anno_file $dataset_anno_file \
+  --dataset_name T2I-CompBench \
+  --n_images 0

scripts/eval/eval_t2i2.sh

@@ -0,0 +1,14 @@
+#!/bin/sh
+cd /mnt/petrelfs/leihaodong/ICLR25/eval
+# HF_ENDPOINT=https://hf-mirror.com
+dataset_anno_file=/mnt/petrelfs/leihaodong/ICLR25/sjdtree/data/prompts/T2I-CompBench_val.json
+workdir=/mnt/petrelfs/leihaodong/ICLR25/exp/Lumina-mGPT-7B-768_T2ICompBenchVal/base/img
+
+srun \
+ -N 1 \
+ --ntasks-per-node 1 \
+ python evaluation_metrics.py \
+  --workdir $workdir \
+  --dataset_anno_file $dataset_anno_file \
+  --dataset_name T2I-CompBench \
+  --n_images 0

scripts/eval/eval_t2i3.sh

@@ -0,0 +1,14 @@
+#!/bin/sh
+cd /mnt/petrelfs/leihaodong/ICLR25/eval
+# HF_ENDPOINT=https://hf-mirror.com
+dataset_anno_file=/mnt/petrelfs/leihaodong/ICLR25/sjdtree/data/prompts/T2I-CompBench_val.json
+workdir=/mnt/petrelfs/leihaodong/ICLR25/exp/Emu3/T2ICompBenchVal/ms_sjd/img
+
+srun \
+ -N 1 \
+ --ntasks-per-node 1 \
+ python evaluation_metrics.py \
+  --workdir $workdir \
+  --dataset_anno_file $dataset_anno_file \
+  --dataset_name T2I-CompBench \
+  --n_images 0

scripts/eval/eval_t2i4.sh

@@ -0,0 +1,14 @@
+#!/bin/sh
+cd /mnt/petrelfs/leihaodong/ICLR25/eval
+# HF_ENDPOINT=https://hf-mirror.com
+dataset_anno_file=/mnt/petrelfs/leihaodong/ICLR25/sjdtree/data/prompts/T2I-CompBench_val.json
+workdir=/mnt/petrelfs/leihaodong/ICLR25/exp/Emu3/T2ICompBenchVal/mscr_gsd/img
+
+srun \
+ -N 1 \
+ --ntasks-per-node 1 \
+ python evaluation_metrics.py \
+  --workdir $workdir \
+  --dataset_anno_file $dataset_anno_file \
+  --dataset_name T2I-CompBench \
+  --n_images 0

scripts/eval/eval_t2i5.sh

@@ -0,0 +1,14 @@
+#!/bin/sh
+cd /mnt/petrelfs/leihaodong/ICLR25/eval
+# HF_ENDPOINT=https://hf-mirror.com
+dataset_anno_file=/mnt/petrelfs/leihaodong/ICLR25/sjdtree/data/prompts/T2I-CompBench_val.json
+workdir=/mnt/petrelfs/leihaodong/ICLR25/exp/Emu3/T2ICompBenchVal/mscr_sjd_0.05/img
+
+srun \
+ -N 1 \
+ --ntasks-per-node 1 \
+ python evaluation_metrics.py \
+  --workdir $workdir \
+  --dataset_anno_file $dataset_anno_file \
+  --dataset_name T2I-CompBench \
+  --n_images 0

scripts/eval/eval_t2i6.sh

@@ -0,0 +1,14 @@
+#!/bin/sh
+cd /mnt/petrelfs/leihaodong/ICLR25/eval
+# HF_ENDPOINT=https://hf-mirror.com
+dataset_anno_file=/mnt/petrelfs/leihaodong/ICLR25/sjdtree/data/prompts/T2I-CompBench_val.json
+workdir=/mnt/petrelfs/leihaodong/ICLR25/exp/Emu3/T2ICompBenchVal/sjd/img
+
+srun \
+ -N 1 \
+ --ntasks-per-node 1 \
+ python evaluation_metrics.py \
+  --workdir $workdir \
+  --dataset_anno_file $dataset_anno_file \
+  --dataset_name T2I-CompBench \
+  --n_images 0

scripts/eval/eval_t2i7.sh

@@ -0,0 +1,14 @@
+#!/bin/sh
+cd /mnt/petrelfs/leihaodong/ICLR25/eval
+# HF_ENDPOINT=https://hf-mirror.com
+dataset_anno_file=/mnt/petrelfs/leihaodong/ICLR25/sjdtree/data/prompts/T2I-CompBench_val.json
+workdir=/mnt/petrelfs/leihaodong/ICLR25/exp/Emu3/T2ICompBenchVal/base/img
+
+srun \
+ -N 1 \
+ --ntasks-per-node 1 \
+ python evaluation_metrics.py \
+  --workdir $workdir \
+  --dataset_anno_file $dataset_anno_file \
+  --dataset_name T2I-CompBench \
+  --n_images 0

scripts/eval/slurm-6226567.out

@@ -0,0 +1,17 @@
+loading annotations into memory...
+Done (t=0.03s)
+creating index...
+index created!
+一共2182 张
+loading annotations into memory...
+Done (t=0.03s)
+creating index...
+index created!
+/mnt/petrelfs/leihaodong/ICLR25/exp/Lumina-mGPT-7B-768_MSCOCO2017Val/base/img/gt
+/mnt/petrelfs/leihaodong/ICLR25/exp/Lumina-mGPT-7B-768_MSCOCO2017Val/base/img
+
  0%|          | 0/2183 [00:00<?, ?it/s]
  0%|          | 1/2183 [00:00<17:13,  2.11it/s]
  0%|          | 9/2183 [00:00<01:51, 19.53it/s]
  1%|          | 17/2183 [00:00<01:04, 33.60it/s]
  1%|          | 25/2183 [00:00<00:47, 45.07it/s]
  2%|▏         | 33/2183 [00:00<00:40, 52.61it/s]
  2%|▏         | 41/2183 [00:01<00:36, 59.15it/s]
  2%|▏         | 49/2183 [00:01<00:33, 64.00it/s]
  3%|▎         | 57/2183 [00:01<00:32, 64.69it/s]
  3%|▎         | 65/2183 [00:01<00:31, 66.64it/s]
  3%|▎         | 73/2183 [00:01<00:30, 68.64it/s]
  4%|▎         | 81/2183 [00:01<00:29, 70.44it/s]
  4%|▍         | 89/2183 [00:01<00:29, 71.09it/s]
  4%|▍         | 97/2183 [00:01<00:28, 72.74it/s]
  5%|▍         | 105/2183 [00:01<00:28, 73.29it/s]
  5%|▌         | 113/2183 [00:01<00:27, 75.15it/s]
  6%|▌         | 121/2183 [00:02<00:27, 74.94it/s]
  6%|▌         | 129/2183 [00:02<00:27, 74.15it/s]
  6%|▋         | 137/2183 [00:02<00:27, 73.42it/s]
  7%|▋         | 145/2183 [00:02<00:27, 74.47it/s]
  7%|▋         | 153/2183 [00:02<00:27, 73.21it/s]
  7%|▋         | 161/2183 [00:02<00:27, 73.12it/s]
  8%|▊         | 169/2183 [00:02<00:27, 72.54it/s]
  8%|▊         | 177/2183 [00:02<00:27, 73.45it/s]
  8%|▊         | 185/2183 [00:02<00:27, 73.08it/s]
  9%|▉         | 193/2183 [00:03<00:26, 73.90it/s]
  9%|▉         | 201/2183 [00:03<00:26, 74.35it/s]
 10%|▉         | 209/2183 [00:03<00:26, 73.63it/s]
 10%|▉         | 217/2183 [00:03<00:26, 74.43it/s]
 10%|█         | 225/2183 [00:03<00:26, 74.06it/s]
 11%|█         | 233/2183 [00:03<00:26, 74.20it/s]
 11%|█         | 241/2183 [00:03<00:26, 73.98it/s]
 11%|█▏        | 249/2183 [00:03<00:25, 74.53it/s]
 12%|█▏        | 257/2183 [00:03<00:26, 73.59it/s]
 12%|█▏        | 265/2183 [00:04<00:25, 74.14it/s]
 13%|█▎        | 273/2183 [00:04<00:26, 72.18it/s]
 13%|█▎        | 281/2183 [00:04<00:26, 72.22it/s]
 13%|█▎        | 289/2183 [00:04<00:26, 72.69it/s]
 14%|█▎        | 297/2183 [00:04<00:26, 72.53it/s]
 14%|█▍        | 305/2183 [00:04<00:26, 71.63it/s]
 14%|█▍        | 313/2183 [00:04<00:26, 70.82it/s]
 15%|█▍        | 321/2183 [00:04<00:26, 70.58it/s]
 15%|█▌        | 329/2183 [00:04<00:26, 71.13it/s]
 15%|█▌        | 337/2183 [00:05<00:26, 70.59it/s]
 16%|█▌        | 345/2183 [00:05<00:25, 72.09it/s]
 16%|█▌        | 353/2183 [00:05<00:25, 72.56it/s]
 17%|█▋        | 361/2183 [00:05<00:24, 73.62it/s]
 17%|█▋        | 369/2183 [00:05<00:25, 71.52it/s]
 17%|█▋        | 377/2183 [00:05<00:25, 71.81it/s]
 18%|█▊        | 385/2183 [00:05<00:25, 71.39it/s]
 18%|█▊        | 393/2183 [00:05<00:24, 72.21it/s]
 18%|█▊        | 401/2183 [00:05<00:24, 71.39it/s]
 19%|█▊        | 409/2183 [00:06<00:25, 70.75it/s]
 19%|█▉        | 417/2183 [00:06<00:25, 70.38it/s]
 19%|█▉        | 425/2183 [00:06<00:24, 70.55it/s]
 20%|█▉        | 433/2183 [00:06<00:24, 71.21it/s]
 20%|██        | 441/2183 [00:06<00:24, 70.70it/s]
 21%|██        | 449/2183 [00:06<00:24, 69.76it/s]
 21%|██        | 457/2183 [00:06<00:24, 70.35it/s]
 21%|██▏       | 465/2183 [00:06<00:24, 68.91it/s]
 22%|██▏       | 473/2183 [00:06<00:24, 69.38it/s]
 22%|██▏       | 481/2183 [00:07<00:24, 70.41it/s]
 22%|██▏       | 489/2183 [00:07<00:24, 70.27it/s]
 23%|██▎       | 497/2183 [00:07<00:23, 70.35it/s]
 23%|██▎       | 505/2183 [00:07<00:23, 70.28it/s]
 23%|██▎       | 513/2183 [00:07<00:23, 70.10it/s]
 24%|██▍       | 521/2183 [00:07<00:23, 71.72it/s]
 24%|██▍       | 529/2183 [00:07<00:23, 70.94it/s]
 25%|██▍       | 537/2183 [00:07<00:23, 70.46it/s]
 25%|██▍       | 545/2183 [00:07<00:22, 71.54it/s]
 25%|██▌       | 553/2183 [00:08<00:22, 71.07it/s]
 26%|██▌       | 561/2183 [00:08<00:22, 72.68it/s]
 26%|██▌       | 569/2183 [00:08<00:22, 72.05it/s]
 26%|██▋       | 577/2183 [00:08<00:21, 73.24it/s]
 27%|██▋       | 585/2183 [00:08<00:21, 74.45it/s]
 27%|██▋       | 593/2183 [00:08<00:21, 74.86it/s]
 28%|██▊       | 601/2183 [00:08<00:20, 76.27it/s]
 28%|██▊       | 609/2183 [00:08<00:20, 76.23it/s]
 28%|██▊       | 617/2183 [00:08<00:20, 76.38it/s]
 29%|██▊       | 625/2183 [00:09<00:20, 75.01it/s]
 29%|██▉       | 633/2183 [00:09<00:20, 74.78it/s]
 29%|██▉       | 641/2183 [00:09<00:20, 74.70it/s]
 30%|██▉       | 649/2183 [00:09<00:20, 75.02it/s]
 30%|███       | 657/2183 [00:09<00:20, 73.44it/s]
 30%|███       | 665/2183 [00:09<00:20, 73.19it/s]
 31%|███       | 673/2183 [00:09<00:20, 73.47it/s]
 31%|███       | 681/2183 [00:09<00:20, 72.86it/s]
 32%|███▏      | 689/2183 [00:09<00:20, 71.88it/s]
 32%|███▏      | 697/2183 [00:10<00:20, 72.33it/s]
 32%|███▏      | 705/2183 [00:10<00:20, 72.44it/s]
 33%|███▎      | 713/2183 [00:10<00:20, 71.92it/s]
 33%|███▎      | 721/2183 [00:10<00:20, 72.04it/s]
 33%|███▎      | 729/2183 [00:10<00:20, 71.68it/s]
 34%|███▍      | 737/2183 [00:10<00:19, 72.40it/s]
 34%|███▍      | 745/2183 [00:10<00:19, 72.90it/s]
 34%|███▍      | 753/2183 [00:10<00:19, 71.56it/s]
 35%|███▍      | 761/2183 [00:10<00:19, 71.56it/s]
 35%|███▌      | 769/2183 [00:11<00:19, 71.75it/s]
 36%|███▌      | 777/2183 [00:11<00:19, 72.06it/s]
 36%|███▌      | 785/2183 [00:11<00:19, 71.44it/s]
 36%|███▋      | 793/2183 [00:11<00:19, 72.27it/s]
 37%|███▋      | 801/2183 [00:11<00:18, 72.98it/s]
 37%|███▋      | 809/2183 [00:11<00:19, 68.72it/s]
 37%|███▋      | 817/2183 [00:11<00:19, 70.42it/s]
 38%|███▊      | 825/2183 [00:11<00:19, 70.46it/s]
 38%|███▊      | 833/2183 [00:11<00:18, 71.56it/s]
 39%|███▊      | 841/2183 [00:12<00:18, 72.53it/s]
 39%|███▉      | 849/2183 [00:12<00:18, 72.38it/s]
 39%|███▉      | 857/2183 [00:12<00:18, 72.93it/s]
 40%|███▉      | 865/2183 [00:12<00:18, 73.11it/s]
 40%|███▉      | 873/2183 [00:12<00:17, 73.12it/s]
 40%|████      | 881/2183 [00:12<00:17, 72.57it/s]
 41%|████      | 889/2183 [00:12<00:17, 73.76it/s]
 41%|████      | 897/2183 [00:12<00:17, 73.76it/s]
 41%|████▏     | 905/2183 [00:12<00:17, 73.22it/s]
 42%|████▏     | 913/2183 [00:13<00:17, 72.57it/s]
 42%|████▏     | 921/2183 [00:13<00:17, 72.36it/s]
 43%|████▎     | 929/2183 [00:13<00:17, 72.02it/s]
 43%|████▎     | 937/2183 [00:13<00:17, 71.86it/s]
 43%|████▎     | 945/2183 [00:13<00:17, 72.59it/s]
 44%|████▎     | 953/2183 [00:13<00:16, 72.99it/s]
 44%|████▍     | 961/2183 [00:13<00:16, 73.34it/s]
 44%|████▍     | 969/2183 [00:13<00:16, 74.05it/s]
 45%|████▍     | 977/2183 [00:13<00:16, 74.73it/s]
 45%|████▌     | 985/2183 [00:14<00:16, 74.81it/s]
 45%|████▌     | 993/2183 [00:14<00:15, 74.91it/s]
 46%|████▌     | 1001/2183 [00:14<00:15, 74.18it/s]
 46%|████▌     | 1009/2183 [00:14<00:15, 75.24it/s]
 47%|████▋     | 1017/2183 [00:14<00:15, 75.78it/s]
 47%|████▋     | 1025/2183 [00:14<00:15, 75.21it/s]
 47%|████▋     | 1033/2183 [00:14<00:15, 74.93it/s]
 48%|████▊     | 1042/2183 [00:14<00:14, 76.46it/s]
 48%|████▊     | 1050/2183 [00:14<00:14, 76.68it/s]
 48%|████▊     | 1058/2183 [00:14<00:14, 76.46it/s]
 49%|████▉     | 1066/2183 [00:15<00:14, 75.73it/s]
 49%|████▉     | 1074/2183 [00:15<00:14, 75.33it/s]
 50%|████▉     | 1082/2183 [00:15<00:14, 75.99it/s]
 50%|████▉     | 1090/2183 [00:15<00:14, 75.85it/s]
 50%|█████     | 1098/2183 [00:15<00:14, 75.96it/s]
 51%|█████     | 1106/2183 [00:15<00:14, 75.33it/s]
 51%|█████     | 1114/2183 [00:15<00:14, 74.38it/s]
 51%|█████▏    | 1122/2183 [00:15<00:14, 70.80it/s]
 52%|█████▏    | 1130/2183 [00:15<00:15, 69.64it/s]
 52%|█████▏    | 1137/2183 [00:16<00:15, 69.66it/s]
 52%|█████▏    | 1144/2183 [00:16<00:15, 69.16it/s]
 53%|█████▎    | 1152/2183 [00:16<00:14, 70.52it/s]
 53%|█████▎    | 1160/2183 [00:16<00:14, 70.78it/s]
 54%|█████▎    | 1168/2183 [00:16<00:14, 71.24it/s]
 54%|█████▍    | 1176/2183 [00:16<00:14, 70.93it/s]
 54%|█████▍    | 1184/2183 [00:16<00:14, 71.19it/s]
 55%|█████▍    | 1192/2183 [00:16<00:13, 71.44it/s]
 55%|█████▍    | 1200/2183 [00:16<00:14, 69.82it/s]
 55%|█████▌    | 1207/2183 [00:17<00:14, 69.44it/s]
 56%|█████▌    | 1215/2183 [00:17<00:13, 70.47it/s]
 56%|█████▌    | 1223/2183 [00:17<00:13, 71.36it/s]
 56%|█████▋    | 1231/2183 [00:17<00:13, 70.53it/s]
 57%|█████▋    | 1239/2183 [00:17<00:13, 71.34it/s]
 57%|█████▋    | 1247/2183 [00:17<00:13, 71.44it/s]
 57%|█████▋    | 1255/2183 [00:17<00:12, 71.46it/s]
 58%|█████▊    | 1263/2183 [00:17<00:12, 71.62it/s]
 58%|█████▊    | 1271/2183 [00:17<00:12, 72.44it/s]
 59%|█████▊    | 1279/2183 [00:18<00:12, 72.48it/s]
 59%|█████▉    | 1287/2183 [00:18<00:12, 73.25it/s]
 59%|█████▉    | 1295/2183 [00:18<00:11, 74.10it/s]
 60%|█████▉    | 1303/2183 [00:18<00:11, 74.50it/s]
 60%|██████    | 1311/2183 [00:18<00:11, 74.26it/s]
 60%|██████    | 1319/2183 [00:18<00:11, 74.71it/s]
 61%|██████    | 1327/2183 [00:18<00:11, 74.88it/s]
 61%|██████    | 1335/2183 [00:18<00:11, 75.29it/s]
 62%|██████▏   | 1343/2183 [00:18<00:11, 76.33it/s]
 62%|███���██▏   | 1351/2183 [00:19<00:11, 74.85it/s]
 62%|██████▏   | 1359/2183 [00:19<00:11, 74.38it/s]
 63%|██████▎   | 1367/2183 [00:19<00:10, 74.69it/s]
 63%|██████▎   | 1375/2183 [00:19<00:10, 75.69it/s]
 63%|██████▎   | 1383/2183 [00:19<00:10, 74.42it/s]
 64%|██████▎   | 1391/2183 [00:19<00:10, 73.80it/s]
 64%|██████▍   | 1399/2183 [00:19<00:10, 72.95it/s]
 64%|██████▍   | 1407/2183 [00:19<00:10, 71.68it/s]
 65%|██████▍   | 1415/2183 [00:19<00:10, 71.87it/s]
 65%|██████▌   | 1423/2183 [00:20<00:10, 70.90it/s]
 66%|██████▌   | 1431/2183 [00:20<00:10, 70.77it/s]
 66%|██████▌   | 1439/2183 [00:20<00:10, 71.32it/s]
 66%|██████▋   | 1447/2183 [00:20<00:10, 72.31it/s]
 67%|██████▋   | 1455/2183 [00:20<00:10, 69.70it/s]
 67%|██████▋   | 1463/2183 [00:20<00:10, 70.11it/s]
 67%|██████▋   | 1471/2183 [00:20<00:10, 70.65it/s]
 68%|██████▊   | 1479/2183 [00:20<00:10, 69.48it/s]
 68%|██████▊   | 1487/2183 [00:20<00:09, 71.77it/s]
 68%|██████▊   | 1495/2183 [00:21<00:09, 72.63it/s]
 69%|██████▉   | 1503/2183 [00:21<00:09, 73.28it/s]
 69%|██████▉   | 1511/2183 [00:21<00:09, 72.50it/s]
 70%|██████▉   | 1519/2183 [00:21<00:09, 73.49it/s]
 70%|██████▉   | 1527/2183 [00:21<00:08, 73.57it/s]
 70%|███████   | 1535/2183 [00:21<00:08, 73.79it/s]
 71%|███████   | 1543/2183 [00:21<00:08, 74.39it/s]
 71%|███████   | 1551/2183 [00:21<00:08, 73.91it/s]
 71%|███████▏  | 1559/2183 [00:21<00:08, 74.14it/s]
 72%|███████▏  | 1567/2183 [00:22<00:08, 73.34it/s]
 72%|███████▏  | 1575/2183 [00:22<00:08, 73.53it/s]
 73%|███████▎  | 1583/2183 [00:22<00:08, 73.11it/s]
 73%|███████▎  | 1591/2183 [00:22<00:08, 72.14it/s]
 73%|███████▎  | 1599/2183 [00:22<00:08, 71.76it/s]
 74%|███████▎  | 1607/2183 [00:22<00:07, 72.47it/s]
 74%|███████▍  | 1615/2183 [00:22<00:07, 72.58it/s]
 74%|███████▍  | 1623/2183 [00:22<00:07, 72.69it/s]
 75%|███████▍  | 1631/2183 [00:22<00:07, 72.77it/s]
 75%|███████▌  | 1639/2183 [00:23<00:07, 69.69it/s]
 75%|███████▌  | 1647/2183 [00:23<00:07, 71.38it/s]
 76%|███████▌  | 1655/2183 [00:23<00:07, 72.49it/s]
 76%|███████▌  | 1663/2183 [00:23<00:07, 72.14it/s]
 77%|███████▋  | 1671/2183 [00:23<00:07, 71.08it/s]
 77%|███████▋  | 1679/2183 [00:23<00:07, 69.83it/s]
 77%|███████▋  | 1686/2183 [00:23<00:07, 69.70it/s]
 78%|███████▊  | 1694/2183 [00:23<00:06, 70.15it/s]
 78%|███████▊  | 1702/2183 [00:23<00:06, 69.81it/s]
 78%|███████▊  | 1709/2183 [00:24<00:06, 69.15it/s]
 79%|███████▊  | 1717/2183 [00:24<00:06, 70.31it/s]
 79%|███████▉  | 1725/2183 [00:24<00:06, 70.50it/s]
 79%|███████▉  | 1733/2183 [00:24<00:06, 69.23it/s]
 80%|███████▉  | 1741/2183 [00:24<00:06, 70.24it/s]
 80%|████████  | 1749/2183 [00:24<00:06, 69.44it/s]
 80%|████████  | 1757/2183 [00:24<00:06, 70.57it/s]
 81%|████████  | 1765/2183 [00:24<00:05, 70.38it/s]
 81%|████████  | 1773/2183 [00:24<00:05, 70.84it/s]
 82%|████████▏ | 1781/2183 [00:25<00:05, 69.90it/s]
 82%|████████▏ | 1789/2183 [00:25<00:05, 71.31it/s]
 82%|████████▏ | 1797/2183 [00:25<00:05, 72.19it/s]
 83%|████████▎ | 1805/2183 [00:25<00:05, 72.21it/s]
 83%|████████▎ | 1813/2183 [00:25<00:05, 71.91it/s]
 83%|████████▎ | 1821/2183 [00:25<00:05, 71.19it/s]
 84%|████████▍ | 1829/2183 [00:25<00:05, 69.57it/s]
 84%|████████▍ | 1836/2183 [00:25<00:04, 69.65it/s]
 84%|████████▍ | 1843/2183 [00:25<00:04, 69.29it/s]
 85%|████████▍ | 1850/2183 [00:26<00:04, 69.35it/s]
 85%|████████▌ | 1858/2183 [00:26<00:04, 69.93it/s]
 85%|████████▌ | 1866/2183 [00:26<00:04, 70.59it/s]
 86%|████████▌ | 1874/2183 [00:26<00:04, 71.26it/s]
 86%|████████▌ | 1882/2183 [00:26<00:04, 71.00it/s]
 87%|████████▋ | 1890/2183 [00:26<00:04, 70.87it/s]
 87%|████████▋ | 1898/2183 [00:26<00:03, 71.66it/s]
 87%|████████▋ | 1906/2183 [00:26<00:03, 71.63it/s]
 88%|████████▊ | 1914/2183 [00:26<00:03, 71.13it/s]
 88%|████████▊ | 1922/2183 [00:27<00:03, 71.33it/s]
 88%|████████▊ | 1930/2183 [00:27<00:03, 70.65it/s]
 89%|████████▉ | 1938/2183 [00:27<00:03, 70.88it/s]
 89%|████████▉ | 1946/2183 [00:27<00:03, 71.42it/s]
 90%|████████▉ | 1954/2183 [00:27<00:03, 71.29it/s]
 90%|████████▉ | 1962/2183 [00:27<00:03, 69.92it/s]
 90%|█████████ | 1970/2183 [00:27<00:02, 71.00it/s]
 91%|█████████ | 1978/2183 [00:27<00:02, 72.05it/s]
 91%|█████████ | 1986/2183 [00:27<00:02, 73.09it/s]
 91%|█████████▏| 1994/2183 [00:28<00:02, 74.22it/s]
 92%|█████████▏| 2002/2183 [00:28<00:02, 74.77it/s]
 92%|█████████▏| 2010/2183 [00:28<00:02, 75.32it/s]
 92%|█████████▏| 2018/2183 [00:28<00:02, 75.91it/s]
 93%|█████████▎| 2026/2183 [00:28<00:02, 74.29it/s]
 93%|█████████▎| 2034/2183 [00:28<00:01, 75.52it/s]
 94%|█████████▎| 2042/2183 [00:28<00:01, 75.61it/s]
 94%|█████████▍| 2050/2183 [00:28<00:01, 74.89it/s]
 94%|█████████▍| 2058/2183 [00:28<00:01, 75.60it/s]
 95%|█████████▍| 2066/2183 [00:28<00:01, 73.76it/s]
 95%|█████████▌| 2074/2183 [00:29<00:01, 72.56it/s]
 95%|█████████▌| 2082/2183 [00:29<00:01, 72.78it/s]
 96%|█████████▌| 2090/2183 [00:29<00:01, 73.18it/s]
 96%|█████████▌| 2098/2183 [00:29<00:01, 72.41it/s]
 96%|█████████▋| 2106/2183 [00:29<00:01, 71.75it/s]
 97%|█████████▋| 2114/2183 [00:29<00:00, 72.55it/s]
 97%|█████████▋| 2122/2183 [00:29<00:00, 72.66it/s]
 98%|█████████▊| 2130/2183 [00:29<00:00, 72.94it/s]
 98%|█████████▊| 2138/2183 [00:29<00:00, 72.17it/s]
 98%|█████████▊| 2146/2183 [00:30<00:00, 70.90it/s]
 99%|█████████▊| 2154/2183 [00:30<00:00, 70.87it/s]
 99%|█████████▉| 2162/2183 [00:30<00:00, 71.11it/s]
 99%|█████████▉| 2170/2183 [00:30<00:00, 71.19it/s]
+
  0%|          | 0/2182 [00:00<?, ?it/s]
  0%|          | 3/2182 [00:00<01:13, 29.61it/s]
  0%|          | 6/2182 [00:00<01:14, 29.16it/s]
  0%|          | 9/2182 [00:00<01:14, 29.31it/s]
  1%|          | 12/2182 [00:00<01:13, 29.56it/s]
  1%|          | 15/2182 [00:00<01:13, 29.54it/s]
  1%|          | 18/2182 [00:00<01:12, 29.67it/s]
  1%|          | 22/2182 [00:00<01:12, 29.91it/s]
  1%|          | 25/2182 [00:00<01:12, 29.73it/s]
  1%|▏         | 28/2182 [00:00<01:13, 29.42it/s]
  1%|▏         | 31/2182 [00:01<01:13, 29.27it/s]
  2%|▏         | 34/2182 [00:01<01:13, 29.38it/s]
  2%|▏         | 37/2182 [00:01<01:12, 29.52it/s]
  2%|▏         | 40/2182 [00:01<01:12, 29.36it/s]
  2%|▏         | 43/2182 [00:01<01:12, 29.42it/s]
  2%|▏         | 47/2182 [00:01<01:11, 29.74it/s]
  2%|▏         | 50/2182 [00:01<01:12, 29.53it/s]
  2%|▏         | 54/2182 [00:01<01:11, 29.58it/s]
  3%|▎         | 58/2182 [00:01<01:11, 29.56it/s]
  3%|▎         | 61/2182 [00:02<01:12, 29.25it/s]
  3%|▎         | 65/2182 [00:02<01:11, 29.54it/s]
  3%|▎         | 69/2182 [00:02<01:11, 29.65it/s]
  3%|▎         | 73/2182 [00:02<01:11, 29.55it/s]
  3%|▎         | 76/2182 [00:02<01:11, 29.65it/s]
  4%|▎         | 80/2182 [00:02<01:10, 29.95it/s]
  4%|▍         | 83/2182 [00:02<01:10, 29.74it/s]
  4%|▍         | 86/2182 [00:02<01:10, 29.64it/s]
  4%|▍         | 89/2182 [00:03<01:10, 29.65it/s]
  4%|▍         | 93/2182 [00:03<01:10, 29.64it/s]
  4%|▍         | 96/2182 [00:03<01:10, 29.73it/s]
  5%|▍         | 99/2182 [00:03<01:10, 29.59it/s]
  5%|▍         | 102/2182 [00:03<01:10, 29.42it/s]
  5%|▍         | 105/2182 [00:03<01:10, 29.48it/s]
  5%|▍         | 108/2182 [00:03<01:10, 29.62it/s]
  5%|▌         | 111/2182 [00:03<01:10, 29.53it/s]
  5%|▌         | 115/2182 [00:03<01:09, 29.90it/s]
  5%|▌         | 118/2182 [00:03<01:09, 29.78it/s]
  6%|▌         | 121/2182 [00:04<01:10, 29.33it/s]
  6%|▌         | 125/2182 [00:04<01:09, 29.45it/s]
  6%|▌         | 129/2182 [00:04<01:08, 29.79it/s]
  6%|▌         | 132/2182 [00:04<01:08, 29.75it/s]
  6%|▌         | 136/2182 [00:04<01:08, 29.88it/s]
  6%|▋         | 139/2182 [00:04<01:08, 29.73it/s]
  7%|▋         | 143/2182 [00:04<01:06, 30.45it/s]
  7%|▋         | 147/2182 [00:04<01:07, 30.26it/s]
  7%|▋         | 151/2182 [00:05<01:07, 30.14it/s]
  7%|▋         | 155/2182 [00:05<01:07, 29.89it/s]
  7%|▋         | 158/2182 [00:05<01:07, 29.88it/s]
  7%|▋         | 162/2182 [00:05<01:07, 30.10it/s]
  8%|▊         | 166/2182 [00:05<01:07, 30.07it/s]
  8%|▊         | 170/2182 [00:05<01:06, 30.25it/s]
  8%|▊         | 174/2182 [00:05<01:06, 30.30it/s]
  8%|▊         | 178/2182 [00:05<01:05, 30.39it/s]
  8%|▊         | 182/2182 [00:06<01:05, 30.35it/s]
  9%|▊         | 186/2182 [00:06<01:05, 30.42it/s]
  9%|▊         | 190/2182 [00:06<01:05, 30.49it/s]
  9%|▉         | 194/2182 [00:06<01:04, 30.80it/s]
  9%|▉         | 198/2182 [00:06<01:04, 30.58it/s]
  9%|▉         | 202/2182 [00:06<01:03, 30.97it/s]
  9%|▉         | 206/2182 [00:06<01:03, 31.00it/s]
 10%|▉         | 210/2182 [00:07<01:06, 29.85it/s]
 10%|▉         | 214/2182 [00:07<01:05, 30.00it/s]
 10%|▉         | 218/2182 [00:07<01:05, 30.10it/s]
 10%|█         | 222/2182 [00:07<01:05, 30.12it/s]
 10%|█         | 226/2182 [00:07<01:06, 29.21it/s]
 11%|█         | 230/2182 [00:07<01:05, 29.74it/s]
 11%|█         | 234/2182 [00:07<01:05, 29.86it/s]
 11%|█         | 238/2182 [00:07<01:04, 30.19it/s]
 11%|█         | 242/2182 [00:08<01:03, 30.33it/s]
 11%|█▏        | 246/2182 [00:08<01:03, 30.42it/s]
 11%|█▏        | 250/2182 [00:08<01:03, 30.39it/s]
 12%|█▏        | 254/2182 [00:08<01:03, 30.55it/s]
 12%|█▏        | 258/2182 [00:08<01:03, 30.35it/s]
 12%|█▏        | 262/2182 [00:08<01:03, 30.15it/s]
 12%|█▏        | 266/2182 [00:08<01:03, 30.13it/s]
 12%|█▏        | 270/2182 [00:09<01:03, 29.95it/s]
 13%|█▎        | 273/2182 [00:09<01:03, 29.85it/s]
 13%|█▎        | 277/2182 [00:09<01:03, 29.84it/s]
 13%|█▎        | 280/2182 [00:09<01:04, 29.69it/s]
 13%|█▎        | 284/2182 [00:09<01:03, 29.94it/s]
 13%|█▎        | 287/2182 [00:09<01:03, 29.73it/s]
 13%|█▎        | 290/2182 [00:09<01:04, 29.55it/s]
 13%|█▎        | 293/2182 [00:09<01:03, 29.52it/s]
 14%|█▎        | 296/2182 [00:09<01:04, 29.44it/s]
 14%|█▎        | 299/2182 [00:10<01:04, 29.29it/s]
 14%|█▍        | 303/2182 [00:10<01:03, 29.66it/s]
 14%|█▍        | 306/2182 [00:10<01:03, 29.57it/s]
 14%|█▍        | 309/2182 [00:10<01:03, 29.55it/s]
 14%|█▍        | 313/2182 [00:10<01:03, 29.56it/s]
 14%|█▍        | 316/2182 [00:10<01:03, 29.54it/s]
 15%|█▍        | 320/2182 [00:10<01:02, 29.84it/s]
 15%|█▍        | 323/2182 [00:10<01:02, 29.85it/s]
 15%|█▍        | 326/2182 [00:10<01:02, 29.78it/s]
 15%|█▌        | 330/2182 [00:11<01:02, 29.82it/s]
 15%|█▌        | 333/2182 [00:11<01:02, 29.80it/s]
 15%|█▌        | 337/2182 [00:11<01:01, 29.78it/s]
 16%|█▌        | 341/2182 [00:11<01:01, 29.92it/s]
 16%|█▌        | 345/2182 [00:11<01:01, 29.91it/s]
 16%|█▌        | 349/2182 [00:11<01:01, 29.91it/s]
 16%|█▌        | 352/2182 [00:11<01:01, 29.88it/s]
 16%|█▋        | 355/2182 [00:11<01:01, 29.74it/s]
 16%|█▋        | 358/2182 [00:11<01:01, 29.61it/s]
 17%|█▋        | 361/2182 [00:12<01:01, 29.61it/s]
 17%|█▋        | 364/2182 [00:12<01:01, 29.68it/s]
 17%|█▋        | 367/2182 [00:12<01:01, 29.44it/s]
 17%|█▋        | 370/2182 [00:12<01:01, 29.37it/s]
 17%|█▋        | 373/2182 [00:12<01:02, 29.15it/s]
 17%|█▋        | 377/2182 [00:12<01:01, 29.13it/s]
 17%|█▋        | 381/2182 [00:12<01:01, 29.26it/s]
 18%|█▊        | 385/2182 [00:12<01:00, 29.82it/s]
 18%|█▊        | 388/2182 [00:13<01:00, 29.86it/s]
 18%|█▊        | 391/2182 [00:13<01:00, 29.61it/s]
 18%|█▊        | 394/2182 [00:13<01:01, 29.27it/s]
 18%|█▊        | 398/2182 [00:13<01:00, 29.58it/s]
 18%|█▊        | 402/2182 [00:13<00:59, 29.81it/s]
 19%|█▊        | 406/2182 [00:13<00:59, 29.88it/s]
 19%|█▊        | 409/2182 [00:13<00:59, 29.85it/s]
 19%|█▉        | 412/2182 [00:13<01:01, 28.67it/s]
 19%|█▉        | 416/2182 [00:13<01:00, 29.28it/s]
 19%|█▉        | 419/2182 [00:14<01:00, 29.23it/s]
 19%|█▉        | 422/2182 [00:14<01:00, 29.24it/s]
 20%|█▉        | 426/2182 [00:14<00:59, 29.54it/s]
 20%|█▉        | 430/2182 [00:14<00:58, 29.75it/s]
 20%|█▉        | 433/2182 [00:14<00:58, 29.80it/s]
 20%|██        | 437/2182 [00:14<00:58, 29.86it/s]
 20%|██        | 441/2182 [00:14<00:57, 30.08it/s]
 20%|██        | 445/2182 [00:14<00:57, 30.18it/s]
 21%|██        | 449/2182 [00:15<00:57, 30.14it/s]
 21%|██        | 453/2182 [00:15<00:57, 29.92it/s]
 21%|██        | 456/2182 [00:15<00:57, 29.85it/s]
 21%|██        | 459/2182 [00:15<00:57, 29.72it/s]
 21%|██        | 462/2182 [00:15<00:57, 29.76it/s]
 21%|██▏       | 466/2182 [00:15<00:57, 30.04it/s]
 22%|██▏       | 470/2182 [00:15<00:56, 30.21it/s]
 22%|██▏       | 474/2182 [00:15<00:56, 30.12it/s]
 22%|██▏       | 478/2182 [00:16<00:56, 30.03it/s]
 22%|██▏       | 482/2182 [00:16<00:56, 30.16it/s]
 22%|██▏       | 486/2182 [00:16<00:57, 29.63it/s]
 22%|██▏       | 490/2182 [00:16<00:57, 29.68it/s]
 23%|██▎       | 493/2182 [00:16<00:58, 28.92it/s]
 23%|██▎       | 496/2182 [00:16<00:57, 29.14it/s]
 23%|██▎       | 500/2182 [00:16<00:56, 29.69it/s]
 23%|██▎       | 504/2182 [00:16<00:56, 29.96it/s]
 23%|██▎       | 507/2182 [00:17<00:56, 29.86it/s]
 23%|██▎       | 510/2182 [00:17<00:55, 29.90it/s]
 24%|██▎       | 514/2182 [00:17<00:55, 30.13it/s]
 24%|██▎       | 518/2182 [00:17<00:54, 30.26it/s]
 24%|██▍       | 522/2182 [00:17<00:54, 30.22it/s]
 24%|██▍       | 526/2182 [00:17<00:54, 30.20it/s]
 24%|██▍       | 530/2182 [00:17<00:54, 30.31it/s]
 24%|██▍       | 534/2182 [00:17<00:54, 30.15it/s]
 25%|██▍       | 538/2182 [00:18<00:53, 30.55it/s]
 25%|██▍       | 542/2182 [00:18<00:53, 30.42it/s]
 25%|██▌       | 546/2182 [00:18<00:54, 30.17it/s]
 25%|██▌       | 550/2182 [00:18<00:54, 29.99it/s]
 25%|██▌       | 553/2182 [00:18<00:54, 29.87it/s]
 25%|██▌       | 556/2182 [00:18<00:54, 29.71it/s]
 26%|██▌       | 559/2182 [00:18<00:54, 29.64it/s]
 26%|██▌       | 562/2182 [00:18<00:55, 29.28it/s]
 26%|██▌       | 565/2182 [00:18<00:54, 29.42it/s]
 26%|██▌       | 568/2182 [00:19<00:55, 29.19it/s]
 26%|██▌       | 571/2182 [00:19<00:55, 29.12it/s]
 26%|██▋       | 574/2182 [00:19<00:55, 29.05it/s]
 26%|██▋       | 577/2182 [00:19<00:54, 29.24it/s]
 27%|██▋       | 580/2182 [00:19<00:54, 29.44it/s]
 27%|██▋       | 583/2182 [00:19<00:54, 29.28it/s]
 27%|██▋       | 586/2182 [00:19<00:54, 29.29it/s]
 27%|██▋       | 590/2182 [00:19<00:54, 29.46it/s]
 27%|██▋       | 593/2182 [00:19<00:53, 29.46it/s]
 27%|██▋       | 597/2182 [00:20<00:53, 29.65it/s]
 27%|██▋       | 600/2182 [00:20<00:53, 29.39it/s]
 28%|██▊       | 603/2182 [00:20<00:54, 29.19it/s]
 28%|██▊       | 606/2182 [00:20<00:54, 29.17it/s]
 28%|██▊       | 609/2182 [00:20<00:53, 29.27it/s]
 28%|██▊       | 612/2182 [00:20<00:53, 29.30it/s]
 28%|██▊       | 615/2182 [00:20<00:53, 29.38it/s]
 28%|██▊       | 618/2182 [00:20<00:53, 29.47it/s]
 29%|██▊       | 622/2182 [00:20<00:52, 29.63it/s]
 29%|██▊       | 626/2182 [00:21<00:52, 29.71it/s]
 29%|██▉       | 630/2182 [00:21<00:51, 29.94it/s]
 29%|██▉       | 634/2182 [00:21<00:51, 30.12it/s]
 29%|██▉       | 638/2182 [00:21<00:51, 29.78it/s]
 29%|██▉       | 641/2182 [00:21<00:51, 29.76it/s]
 30%|██▉       | 644/2182 [00:21<00:51, 29.78it/s]
 30%|██▉       | 647/2182 [00:21<00:51, 29.76it/s]
 30%|██▉       | 650/2182 [00:21<00:51, 29.60it/s]
 30%|██▉       | 654/2182 [00:21<00:51, 29.70it/s]
 30%|███       | 657/2182 [00:22<00:51, 29.73it/s]
 30%|███       | 660/2182 [00:22<00:51, 29.67it/s]
 30%|███       | 663/2182 [00:22<00:51, 29.52it/s]
 31%|███       | 666/2182 [00:22<00:51, 29.37it/s]
 31%|███       | 669/2182 [00:22<00:51, 29.26it/s]
 31%|███       | 673/2182 [00:22<00:51, 29.55it/s]
 31%|███       | 676/2182 [00:22<00:50, 29.61it/s]
 31%|███       | 679/2182 [00:22<00:50, 29.62it/s]
 31%|███▏      | 682/2182 [00:22<00:50, 29.53it/s]
 31%|███▏      | 685/2182 [00:23<00:50, 29.63it/s]
 32%|███▏      | 689/2182 [00:23<00:50, 29.65it/s]
 32%|███▏      | 692/2182 [00:23<00:50, 29.46it/s]
 32%|███▏      | 695/2182 [00:23<00:50, 29.39it/s]
 32%|███▏      | 698/2182 [00:23<00:50, 29.33it/s]
 32%|███▏      | 702/2182 [00:23<00:49, 29.67it/s]
 32%|███▏      | 705/2182 [00:23<00:50, 29.54it/s]
 32%|███▏      | 708/2182 [00:23<00:49, 29.50it/s]
 33%|███▎      | 711/2182 [00:23<00:49, 29.46it/s]
 33%|███▎      | 715/2182 [00:24<00:49, 29.63it/s]
 33%|███▎      | 718/2182 [00:24<00:49, 29.58it/s]
 33%|███▎      | 722/2182 [00:24<00:48, 29.85it/s]
 33%|███▎      | 726/2182 [00:24<00:48, 29.78it/s]
 33%|███▎      | 729/2182 [00:24<00:48, 29.67it/s]
 34%|███▎      | 733/2182 [00:24<00:48, 29.81it/s]
 34%|███▎      | 736/2182 [00:24<00:48, 29.77it/s]
 34%|███▍      | 740/2182 [00:24<00:47, 30.17it/s]
 34%|███▍      | 744/2182 [00:24<00:47, 30.10it/s]
 34%|███▍      | 748/2182 [00:25<00:47, 29.88it/s]
 34%|███▍      | 751/2182 [00:25<00:48, 29.66it/s]
 35%|███▍      | 754/2182 [00:25<00:48, 29.55it/s]
 35%|███▍      | 757/2182 [00:25<00:48, 29.65it/s]
 35%|███▍      | 760/2182 [00:25<00:47, 29.74it/s]
 35%|███▍      | 763/2182 [00:25<00:47, 29.79it/s]
 35%|███▌      | 767/2182 [00:25<00:47, 29.92it/s]
 35%|███▌      | 770/2182 [00:25<00:47, 29.77it/s]
 35%|███▌      | 773/2182 [00:25<00:47, 29.66it/s]
 36%|███▌      | 777/2182 [00:26<00:47, 29.76it/s]
 36%|███▌      | 780/2182 [00:26<00:47, 29.76it/s]
 36%|███▌      | 783/2182 [00:26<00:47, 29.47it/s]
 36%|███▌      | 786/2182 [00:26<00:47, 29.26it/s]
 36%|███▌      | 789/2182 [00:26<00:47, 29.25it/s]
 36%|███▋      | 792/2182 [00:26<00:47, 29.38it/s]
 36%|███▋      | 795/2182 [00:26<00:47, 29.34it/s]
 37%|███▋      | 798/2182 [00:26<00:46, 29.49it/s]
 37%|███▋      | 801/2182 [00:26<00:46, 29.58it/s]
 37%|███▋      | 804/2182 [00:27<00:46, 29.69it/s]
 37%|███▋      | 807/2182 [00:27<00:46, 29.72it/s]
 37%|███▋      | 810/2182 [00:27<00:46, 29.71it/s]
 37%|███▋      | 814/2182 [00:27<00:45, 29.85it/s]
 37%|███▋      | 818/2182 [00:27<00:45, 29.75it/s]
 38%|███▊      | 821/2182 [00:27<00:45, 29.80it/s]
 38%|███▊      | 825/2182 [00:27<00:45, 29.60it/s]
 38%|███▊      | 829/2182 [00:27<00:45, 29.66it/s]
 38%|███▊      | 832/2182 [00:27<00:45, 29.74it/s]
 38%|███▊      | 836/2182 [00:28<00:44, 29.91it/s]
 38%|███▊      | 839/2182 [00:28<00:44, 29.90it/s]
 39%|███▊      | 842/2182 [00:28<00:44, 29.89it/s]
 39%|███▉      | 846/2182 [00:28<00:44, 29.98it/s]
 39%|███▉      | 850/2182 [00:28<00:45, 29.55it/s]
 39%|███▉      | 853/2182 [00:28<00:44, 29.56it/s]
 39%|███▉      | 857/2182 [00:28<00:44, 30.01it/s]
 39%|███▉      | 861/2182 [00:28<00:43, 30.04it/s]
 40%|███▉      | 865/2182 [00:29<00:43, 30.16it/s]
 40%|███▉      | 869/2182 [00:29<00:43, 30.12it/s]
 40%|████      | 873/2182 [00:29<00:43, 30.31it/s]
 40%|████      | 877/2182 [00:29<00:41, 31.08it/s]
 40%|████      | 881/2182 [00:29<00:42, 30.73it/s]
 41%|████      | 885/2182 [00:29<00:43, 30.12it/s]
 41%|████      | 889/2182 [00:29<00:43, 29.64it/s]
 41%|████      | 893/2182 [00:29<00:43, 29.62it/s]
 41%|████      | 896/2182 [00:30<00:43, 29.55it/s]
 41%|████      | 899/2182 [00:30<00:43, 29.32it/s]
 41%|████▏     | 902/2182 [00:30<00:43, 29.17it/s]
 41%|████▏     | 905/2182 [00:30<00:44, 28.88it/s]
 42%|████▏     | 909/2182 [00:30<00:43, 29.22it/s]
 42%|████▏     | 913/2182 [00:30<00:43, 29.25it/s]
 42%|████▏     | 916/2182 [00:30<00:43, 29.23it/s]
 42%|████▏     | 919/2182 [00:30<00:43, 29.01it/s]
 42%|████▏     | 922/2182 [00:30<00:43, 29.11it/s]
 42%|████▏     | 925/2182 [00:31<00:42, 29.32it/s]
 43%|████▎     | 929/2182 [00:31<00:42, 29.66it/s]
 43%|████▎     | 933/2182 [00:31<00:41, 29.85it/s]
 43%|████▎     | 937/2182 [00:31<00:41, 29.99it/s]
 43%|████▎     | 941/2182 [00:31<00:41, 30.13it/s]
 43%|████▎     | 945/2182 [00:31<00:41, 29.97it/s]
 43%|████▎     | 949/2182 [00:31<00:40, 30.07it/s]
 44%|████▎     | 953/2182 [00:32<00:41, 29.54it/s]
 44%|████▍     | 957/2182 [00:32<00:41, 29.71it/s]
 44%|████▍     | 960/2182 [00:32<00:41, 29.46it/s]
 44%|████▍     | 964/2182 [00:32<00:41, 29.70it/s]
 44%|████▍     | 967/2182 [00:32<00:41, 29.46it/s]
 45%|████▍     | 971/2182 [00:32<00:40, 29.67it/s]
 45%|████▍     | 974/2182 [00:32<00:40, 29.68it/s]
 45%|████▍     | 978/2182 [00:32<00:40, 29.82it/s]
 45%|████▌     | 982/2182 [00:33<00:39, 30.09it/s]
 45%|████▌     | 986/2182 [00:33<00:39, 30.18it/s]
 45%|████▌     | 990/2182 [00:33<00:39, 30.20it/s]
 46%|████▌     | 994/2182 [00:33<00:39, 30.14it/s]
 46%|████▌     | 998/2182 [00:33<00:39, 30.11it/s]
 46%|████▌     | 1002/2182 [00:33<00:39, 30.10it/s]
 46%|████▌     | 1006/2182 [00:33<00:38, 30.18it/s]
 46%|████▋     | 1010/2182 [00:33<00:38, 30.40it/s]
 46%|████▋     | 1014/2182 [00:34<00:38, 30.51it/s]
 47%|████▋     | 1018/2182 [00:34<00:38, 30.57it/s]
 47%|████▋     | 1022/2182 [00:34<00:37, 30.54it/s]
 47%|████▋     | 1026/2182 [00:34<00:37, 30.68it/s]
 47%|████▋     | 1030/2182 [00:34<00:37, 30.67it/s]
 47%|████▋     | 1034/2182 [00:34<00:37, 30.57it/s]
 48%|████▊     | 1038/2182 [00:34<00:38, 29.83it/s]
 48%|████▊     | 1042/2182 [00:34<00:37, 30.01it/s]
 48%|████▊     | 1046/2182 [00:35<00:37, 29.93it/s]
 48%|████▊     | 1050/2182 [00:35<00:37, 29.93it/s]
 48%|████▊     | 1054/2182 [00:35<00:37, 30.07it/s]
 48%|████▊     | 1058/2182 [00:35<00:37, 30.24it/s]
 49%|████▊     | 1062/2182 [00:35<00:36, 30.32it/s]
 49%|████▉     | 1066/2182 [00:35<00:36, 30.38it/s]
 49%|████▉     | 1070/2182 [00:35<00:36, 30.40it/s]
 49%|████▉     | 1074/2182 [00:36<00:36, 30.23it/s]
 49%|████▉     | 1078/2182 [00:36<00:36, 30.35it/s]
 50%|████▉     | 1082/2182 [00:36<00:36, 30.07it/s]
 50%|████▉     | 1086/2182 [00:36<00:36, 30.19it/s]
 50%|████▉     | 1090/2182 [00:36<00:36, 30.12it/s]
 50%|█████     | 1094/2182 [00:36<00:35, 30.54it/s]
 50%|█████     | 1098/2182 [00:36<00:35, 30.34it/s]
 51%|█████     | 1102/2182 [00:36<00:35, 30.34it/s]
 51%|█████     | 1106/2182 [00:37<00:35, 30.27it/s]
 51%|█████     | 1110/2182 [00:37<00:35, 30.49it/s]
 51%|█████     | 1114/2182 [00:37<00:35, 30.16it/s]
 51%|█████     | 1118/2182 [00:37<00:35, 29.79it/s]
 51%|█████▏    | 1121/2182 [00:37<00:35, 29.76it/s]
 52%|█████▏    | 1124/2182 [00:37<00:35, 29.65it/s]
 52%|█████▏    | 1127/2182 [00:37<00:35, 29.57it/s]
 52%|█████▏    | 1130/2182 [00:37<00:35, 29.48it/s]
 52%|█████▏    | 1134/2182 [00:38<00:35, 29.66it/s]
 52%|█████▏    | 1138/2182 [00:38<00:35, 29.70it/s]
 52%|█████▏    | 1141/2182 [00:38<00:35, 29.72it/s]
 52%|█████▏    | 1144/2182 [00:38<00:35, 29.62it/s]
 53%|█████▎    | 1147/2182 [00:38<00:35, 29.55it/s]
 53%|█████▎    | 1150/2182 [00:38<00:34, 29.57it/s]
 53%|█████▎    | 1154/2182 [00:38<00:34, 29.65it/s]
 53%|█████▎    | 1157/2182 [00:38<00:34, 29.61it/s]
 53%|█████▎    | 1160/2182 [00:38<00:34, 29.58it/s]
 53%|█████▎    | 1163/2182 [00:39<00:34, 29.55it/s]
 53%|█████▎    | 1167/2182 [00:39<00:33, 29.91it/s]
 54%|█████▎    | 1170/2182 [00:39<00:34, 29.75it/s]
 54%|█████▍    | 1173/2182 [00:39<00:33, 29.71it/s]
 54%|█████▍    | 1176/2182 [00:39<00:34, 29.55it/s]
 54%|█████▍    | 1179/2182 [00:39<00:34, 29.43it/s]
 54%|█████▍    | 1182/2182 [00:39<00:33, 29.45it/s]
 54%|█████▍    | 1186/2182 [00:39<00:32, 30.29it/s]
 55%|█████▍    | 1190/2182 [00:39<00:32, 30.20it/s]
 55%|█████▍    | 1194/2182 [00:40<00:32, 30.05it/s]
 55%|█████▍    | 1198/2182 [00:40<00:32, 29.98it/s]
 55%|█████▌    | 1202/2182 [00:40<00:32, 30.22it/s]
 55%|█████▌    | 1206/2182 [00:40<00:32, 30.02it/s]
 55%|█████▌    | 1210/2182 [00:40<00:32, 29.90it/s]
 56%|█████▌    | 1214/2182 [00:40<00:32, 29.91it/s]
 56%|█████▌    | 1218/2182 [00:40<00:32, 30.00it/s]
 56%|█████▌    | 1222/2182 [00:40<00:31, 30.11it/s]
 56%|█████▌    | 1226/2182 [00:41<00:32, 29.67it/s]
 56%|█████▋    | 1230/2182 [00:41<00:31, 29.95it/s]
 57%|█████▋    | 1234/2182 [00:41<00:31, 30.57it/s]
 57%|█████▋    | 1238/2182 [00:41<00:30, 30.49it/s]
 57%|█████▋    | 1242/2182 [00:41<00:30, 30.54it/s]
 57%|█████▋    | 1246/2182 [00:41<00:30, 30.69it/s]
 57%|█████▋    | 1250/2182 [00:41<00:30, 30.52it/s]
 57%|█████▋    | 1254/2182 [00:42<00:30, 30.24it/s]
 58%|█████▊    | 1258/2182 [00:42<00:30, 30.37it/s]
 58%|█████▊    | 1262/2182 [00:42<00:30, 30.58it/s]
 58%|█████▊    | 1266/2182 [00:42<00:30, 30.42it/s]
 58%|█████▊    | 1270/2182 [00:42<00:30, 29.90it/s]
 58%|█████▊    | 1273/2182 [00:42<00:30, 29.78it/s]
 59%|█████▊    | 1277/2182 [00:42<00:30, 29.94it/s]
 59%|█████▊    | 1280/2182 [00:42<00:30, 29.77it/s]
 59%|█████▉    | 1283/2182 [00:43<00:30, 29.58it/s]
 59%|█████▉    | 1286/2182 [00:43<00:30, 28.97it/s]
 59%|█████▉    | 1289/2182 [00:43<00:31, 28.73it/s]
 59%|█████▉    | 1292/2182 [00:43<00:30, 28.80it/s]
 59%|█████▉    | 1295/2182 [00:43<00:30, 29.04it/s]
 60%|█████▉    | 1299/2182 [00:43<00:30, 29.40it/s]
 60%|█████▉    | 1302/2182 [00:43<00:30, 29.19it/s]
 60%|█████▉    | 1305/2182 [00:43<00:30, 29.23it/s]
 60%|█████▉    | 1308/2182 [00:43<00:29, 29.32it/s]
 60%|██████    | 1311/2182 [00:43<00:29, 29.33it/s]
 60%|██████    | 1314/2182 [00:44<00:29, 29.00it/s]
 60%|██████    | 1317/2182 [00:44<00:29, 29.12it/s]
 60%|██████    | 1320/2182 [00:44<00:29, 29.30it/s]
 61%|██████    | 1323/2182 [00:44<00:29, 29.27it/s]
 61%|██████    | 1326/2182 [00:44<00:29, 29.18it/s]
 61%|██████    | 1329/2182 [00:44<00:29, 28.70it/s]
 61%|██████    | 1332/2182 [00:44<00:29, 28.85it/s]
 61%|██████    | 1335/2182 [00:44<00:29, 28.73it/s]
 61%|██████▏   | 1338/2182 [00:44<00:29, 28.42it/s]
 61%|██████▏   | 1341/2182 [00:45<00:29, 28.72it/s]
 62%|██████▏   | 1345/2182 [00:45<00:28, 29.06it/s]
 62%|██████▏   | 1348/2182 [00:45<00:28, 29.21it/s]
 62%|██████▏   | 1351/2182 [00:45<00:28, 29.02it/s]
 62%|██████▏   | 1354/2182 [00:45<00:28, 28.66it/s]
 62%|██████▏   | 1357/2182 [00:45<00:29, 27.81it/s]
 62%|██████▏   | 1360/2182 [00:45<00:28, 28.40it/s]
 62%|██████▏   | 1363/2182 [00:45<00:28, 28.59it/s]
 63%|██████▎   | 1366/2182 [00:45<00:28, 28.89it/s]
 63%|██████▎   | 1369/2182 [00:45<00:28, 28.89it/s]
 63%|██████▎   | 1372/2182 [00:46<00:27, 28.97it/s]
 63%|██████▎   | 1375/2182 [00:46<00:27, 28.87it/s]
 63%|██████▎   | 1378/2182 [00:46<00:27, 28.82it/s]
 63%|██████▎   | 1381/2182 [00:46<00:27, 28.91it/s]
 63%|██████▎   | 1384/2182 [00:46<00:27, 28.98it/s]
 64%|██████▎   | 1387/2182 [00:46<00:27, 28.89it/s]
 64%|██████▎   | 1390/2182 [00:46<00:27, 28.72it/s]
 64%|██████▍   | 1393/2182 [00:46<00:27, 28.82it/s]
 64%|██████▍   | 1396/2182 [00:46<00:27, 28.95it/s]
 64%|██████▍   | 1399/2182 [00:47<00:27, 28.31it/s]
 64%|██████▍   | 1402/2182 [00:47<00:27, 28.42it/s]
 64%|██████▍   | 1405/2182 [00:47<00:27, 28.67it/s]
 65%|██████▍   | 1408/2182 [00:47<00:26, 28.69it/s]
 65%|██████▍   | 1411/2182 [00:47<00:26, 28.75it/s]
 65%|██████▍   | 1414/2182 [00:47<00:26, 28.61it/s]
 65%|██████▍   | 1417/2182 [00:47<00:26, 28.88it/s]
 65%|██████▌   | 1420/2182 [00:47<00:26, 29.00it/s]
 65%|██████▌   | 1423/2182 [00:47<00:26, 28.86it/s]
 65%|██████▌   | 1426/2182 [00:47<00:26, 28.78it/s]
 65%|██████▌   | 1429/2182 [00:48<00:26, 28.75it/s]
 66%|██████▌   | 1432/2182 [00:48<00:26, 28.79it/s]
 66%|██████▌   | 1435/2182 [00:48<00:26, 28.44it/s]
 66%|██████▌   | 1438/2182 [00:48<00:26, 28.60it/s]
 66%|██████▌   | 1441/2182 [00:48<00:25, 28.90it/s]
 66%|██████▌   | 1445/2182 [00:48<00:25, 29.32it/s]
 66%|██████▋   | 1448/2182 [00:48<00:25, 29.22it/s]
 66%|██████▋   | 1451/2182 [00:48<00:24, 29.31it/s]
 67%|██████▋   | 1454/2182 [00:48<00:24, 29.19it/s]
 67%|██████▋   | 1457/2182 [00:49<00:24, 29.33it/s]
 67%|██████▋   | 1461/2182 [00:49<00:24, 29.37it/s]
 67%|██████▋   | 1464/2182 [00:49<00:24, 29.34it/s]
 67%|██████▋   | 1467/2182 [00:49<00:24, 29.09it/s]
 67%|██████▋   | 1470/2182 [00:49<00:24, 28.88it/s]
 68%|██████▊   | 1473/2182 [00:49<00:24, 29.03it/s]
 68%|██████▊   | 1477/2182 [00:49<00:24, 29.24it/s]
 68%|██████▊   | 1481/2182 [00:49<00:23, 29.54it/s]
 68%|██████▊   | 1484/2182 [00:49<00:23, 29.22it/s]
 68%|██████▊   | 1487/2182 [00:50<00:23, 29.11it/s]
 68%|██████▊   | 1490/2182 [00:50<00:23, 28.94it/s]
 68%|██████▊   | 1493/2182 [00:50<00:23, 29.04it/s]
 69%|██████▊   | 1496/2182 [00:50<00:23, 29.17it/s]
 69%|██████▊   | 1500/2182 [00:50<00:23, 29.57it/s]
 69%|██████▉   | 1503/2182 [00:50<00:23, 29.42it/s]
 69%|██████▉   | 1506/2182 [00:50<00:23, 29.35it/s]
 69%|██████▉   | 1509/2182 [00:50<00:22, 29.48it/s]
 69%|██████▉   | 1512/2182 [00:50<00:22, 29.49it/s]
 69%|██████▉   | 1515/2182 [00:51<00:22, 29.27it/s]
 70%|██████▉   | 1518/2182 [00:51<00:23, 28.55it/s]
 70%|██████▉   | 1521/2182 [00:51<00:22, 28.88it/s]
 70%|██████▉   | 1524/2182 [00:51<00:22, 29.08it/s]
 70%|██████▉   | 1527/2182 [00:51<00:22, 28.97it/s]
 70%|███████   | 1530/2182 [00:51<00:22, 29.21it/s]
 70%|███████   | 1533/2182 [00:51<00:22, 29.12it/s]
 70%|███████   | 1536/2182 [00:51<00:22, 29.35it/s]
 71%|███████   | 1539/2182 [00:51<00:21, 29.23it/s]
 71%|███████   | 1543/2182 [00:51<00:21, 29.56it/s]
 71%|███████   | 1546/2182 [00:52<00:21, 29.48it/s]
 71%|███████   | 1549/2182 [00:52<00:21, 29.61it/s]
 71%|███████   | 1553/2182 [00:52<00:21, 29.70it/s]
 71%|███████▏  | 1557/2182 [00:52<00:20, 29.77it/s]
 71%|███████▏  | 1560/2182 [00:52<00:21, 28.57it/s]
 72%|███████▏  | 1563/2182 [00:52<00:21, 28.81it/s]
 72%|███████▏  | 1566/2182 [00:52<00:21, 29.05it/s]
 72%|███████▏  | 1570/2182 [00:52<00:20, 29.53it/s]
 72%|███████▏  | 1573/2182 [00:53<00:20, 29.47it/s]
 72%|███████▏  | 1577/2182 [00:53<00:20, 29.68it/s]
 72%|███████▏  | 1581/2182 [00:53<00:20, 29.66it/s]
 73%|███████▎  | 1585/2182 [00:53<00:20, 29.81it/s]
 73%|███████▎  | 1588/2182 [00:53<00:19, 29.82it/s]
 73%|███████▎  | 1592/2182 [00:53<00:19, 30.16it/s]
 73%|███████▎  | 1596/2182 [00:53<00:19, 30.22it/s]
 73%|███████▎  | 1600/2182 [00:53<00:19, 29.80it/s]
 73%|███████▎  | 1603/2182 [00:54<00:19, 29.43it/s]
 74%|███████▎  | 1607/2182 [00:54<00:19, 29.66it/s]
 74%|███████▍  | 1610/2182 [00:54<00:19, 29.52it/s]
 74%|███████▍  | 1614/2182 [00:54<00:19, 29.60it/s]
 74%|███████▍  | 1618/2182 [00:54<00:18, 29.75it/s]
 74%|███████▍  | 1621/2182 [00:54<00:18, 29.76it/s]
 74%|███████▍  | 1625/2182 [00:54<00:18, 29.88it/s]
 75%|███████▍  | 1628/2182 [00:54<00:19, 29.00it/s]
 75%|███████▍  | 1631/2182 [00:54<00:18, 29.03it/s]
 75%|███████▍  | 1634/2182 [00:55<00:18, 29.11it/s]
 75%|███████▌  | 1637/2182 [00:55<00:18, 29.08it/s]
 75%|███████▌  | 1640/2182 [00:55<00:18, 29.17it/s]
 75%|███████▌  | 1643/2182 [00:55<00:18, 29.28it/s]
 75%|███████▌  | 1646/2182 [00:55<00:18, 29.36it/s]
 76%|███████▌  | 1650/2182 [00:55<00:18, 29.48it/s]
 76%|███████▌  | 1653/2182 [00:55<00:17, 29.42it/s]
 76%|███████▌  | 1656/2182 [00:55<00:18, 29.22it/s]
 76%|███████▌  | 1659/2182 [00:55<00:17, 29.29it/s]
 76%|███████▌  | 1662/2182 [00:56<00:17, 29.26it/s]
 76%|███████▋  | 1665/2182 [00:56<00:17, 29.30it/s]
 76%|███████▋  | 1668/2182 [00:56<00:17, 29.24it/s]
 77%|███████▋  | 1671/2182 [00:56<00:17, 29.33it/s]
 77%|███████▋  | 1674/2182 [00:56<00:17, 29.43it/s]
 77%|███████▋  | 1677/2182 [00:56<00:17, 29.51it/s]
 77%|███████▋  | 1681/2182 [00:56<00:16, 29.55it/s]
 77%|███████▋  | 1684/2182 [00:56<00:16, 29.62it/s]
 77%|███████▋  | 1687/2182 [00:56<00:16, 29.67it/s]
 77%|███████▋  | 1690/2182 [00:56<00:16, 29.54it/s]
 78%|███████▊  | 1693/2182 [00:57<00:16, 29.32it/s]
 78%|███████▊  | 1697/2182 [00:57<00:16, 29.70it/s]
 78%|███████▊  | 1701/2182 [00:57<00:15, 30.10it/s]
 78%|███████▊  | 1705/2182 [00:57<00:16, 29.67it/s]
 78%|███████▊  | 1708/2182 [00:57<00:15, 29.74it/s]
 78%|███████▊  | 1711/2182 [00:57<00:15, 29.55it/s]
 79%|███████▊  | 1714/2182 [00:57<00:15, 29.49it/s]
 79%|███████▊  | 1718/2182 [00:57<00:15, 29.71it/s]
 79%|███████▉  | 1721/2182 [00:58<00:15, 29.60it/s]
 79%|███████▉  | 1724/2182 [00:58<00:15, 29.52it/s]
 79%|███████▉  | 1727/2182 [00:58<00:15, 29.46it/s]
 79%|███████▉  | 1730/2182 [00:58<00:15, 29.27it/s]
 79%|███████▉  | 1733/2182 [00:58<00:15, 29.26it/s]
 80%|███████▉  | 1736/2182 [00:58<00:15, 29.40it/s]
 80%|███████▉  | 1739/2182 [00:58<00:15, 29.49it/s]
 80%|███████▉  | 1743/2182 [00:58<00:14, 29.61it/s]
 80%|████████  | 1746/2182 [00:58<00:14, 29.71it/s]
 80%|████████  | 1749/2182 [00:58<00:14, 29.56it/s]
 80%|████████  | 1752/2182 [00:59<00:14, 29.67it/s]
 80%|████████  | 1755/2182 [00:59<00:14, 29.21it/s]
 81%|████████  | 1758/2182 [00:59<00:14, 29.22it/s]
 81%|████████  | 1761/2182 [00:59<00:14, 29.18it/s]
 81%|████████  | 1764/2182 [00:59<00:14, 29.32it/s]
 81%|████████  | 1767/2182 [00:59<00:14, 29.22it/s]
 81%|████████  | 1770/2182 [00:59<00:14, 28.78it/s]
 81%|████████▏ | 1773/2182 [00:59<00:14, 28.90it/s]
 81%|████████▏ | 1776/2182 [00:59<00:14, 28.65it/s]
 82%|████████▏ | 1779/2182 [01:00<00:14, 28.68it/s]
 82%|████████▏ | 1782/2182 [01:00<00:14, 28.49it/s]
 82%|████████▏ | 1785/2182 [01:00<00:13, 28.73it/s]
 82%|████████▏ | 1788/2182 [01:00<00:13, 29.08it/s]
 82%|████████▏ | 1791/2182 [01:00<00:13, 29.18it/s]
 82%|████████▏ | 1794/2182 [01:00<00:13, 28.99it/s]
 82%|████████▏ | 1797/2182 [01:00<00:13, 29.21it/s]
 83%|████████▎ | 1801/2182 [01:00<00:12, 29.51it/s]
 83%|████████▎ | 1805/2182 [01:00<00:12, 29.71it/s]
 83%|████████▎ | 1809/2182 [01:01<00:12, 29.80it/s]
 83%|████████▎ | 1813/2182 [01:01<00:12, 29.89it/s]
 83%|████████▎ | 1817/2182 [01:01<00:12, 30.04it/s]
 83%|████████▎ | 1821/2182 [01:01<00:12, 29.87it/s]
 84%|████████▎ | 1825/2182 [01:01<00:11, 29.92it/s]
 84%|████████▍ | 1828/2182 [01:01<00:11, 29.92it/s]
 84%|████████▍ | 1831/2182 [01:01<00:11, 29.81it/s]
 84%|████████▍ | 1834/2182 [01:01<00:11, 29.67it/s]
 84%|████████▍ | 1838/2182 [01:01<00:11, 29.84it/s]
 84%|████████▍ | 1841/2182 [01:02<00:11, 29.71it/s]
 85%|████████▍ | 1845/2182 [01:02<00:11, 29.75it/s]
 85%|████████▍ | 1849/2182 [01:02<00:11, 29.92it/s]
 85%|████████▍ | 1853/2182 [01:02<00:11, 29.56it/s]
 85%|████████▌ | 1856/2182 [01:02<00:11, 29.61it/s]
 85%|████████▌ | 1859/2182 [01:02<00:10, 29.52it/s]
 85%|████████▌ | 1863/2182 [01:02<00:10, 29.87it/s]
 86%|████████▌ | 1866/2182 [01:02<00:10, 29.83it/s]
 86%|████████▌ | 1869/2182 [01:03<00:10, 29.85it/s]
 86%|████████▌ | 1872/2182 [01:03<00:10, 29.84it/s]
 86%|████████▌ | 1876/2182 [01:03<00:10, 30.06it/s]
 86%|████████▌ | 1880/2182 [01:03<00:10, 30.09it/s]
 86%|████████▋ | 1884/2182 [01:03<00:09, 30.11it/s]
 87%|████████▋ | 1888/2182 [01:03<00:09, 30.19it/s]
 87%|████████▋ | 1892/2182 [01:03<00:09, 30.15it/s]
 87%|████████▋ | 1896/2182 [01:03<00:09, 29.94it/s]
 87%|████████▋ | 1899/2182 [01:04<00:09, 29.66it/s]
 87%|████████▋ | 1902/2182 [01:04<00:09, 29.59it/s]
 87%|████████▋ | 1905/2182 [01:04<00:09, 29.35it/s]
 87%|████████▋ | 1908/2182 [01:04<00:09, 29.51it/s]
 88%|████████▊ | 1912/2182 [01:04<00:09, 29.87it/s]
 88%|████████▊ | 1915/2182 [01:04<00:09, 29.63it/s]
 88%|████████▊ | 1918/2182 [01:04<00:09, 29.31it/s]
 88%|████████▊ | 1921/2182 [01:04<00:08, 29.30it/s]
 88%|████████▊ | 1924/2182 [01:04<00:08, 29.43it/s]
 88%|████████▊ | 1927/2182 [01:04<00:08, 29.44it/s]
 88%|████████▊ | 1930/2182 [01:05<00:08, 29.32it/s]
 89%|████████▊ | 1933/2182 [01:05<00:08, 29.16it/s]
 89%|████████▉ | 1937/2182 [01:05<00:08, 29.24it/s]
 89%|████████▉ | 1940/2182 [01:05<00:08, 29.31it/s]
 89%|████████▉ | 1944/2182 [01:05<00:07, 29.75it/s]
 89%|████████▉ | 1947/2182 [01:05<00:07, 29.57it/s]
 89%|████████▉ | 1951/2182 [01:05<00:07, 29.73it/s]
 90%|████████▉ | 1954/2182 [01:05<00:07, 29.76it/s]
 90%|████████▉ | 1957/2182 [01:06<00:07, 29.66it/s]
 90%|████████▉ | 1960/2182 [01:06<00:07, 29.56it/s]
 90%|████████▉ | 1963/2182 [01:06<00:07, 29.62it/s]
 90%|█████████ | 1966/2182 [01:06<00:07, 29.45it/s]
 90%|█████████ | 1969/2182 [01:06<00:07, 29.37it/s]
 90%|█████████ | 1973/2182 [01:06<00:07, 29.37it/s]
 91%|█████████ | 1976/2182 [01:06<00:06, 29.45it/s]
 91%|█████████ | 1979/2182 [01:06<00:06, 29.35it/s]
 91%|█████████ | 1982/2182 [01:06<00:06, 29.27it/s]
 91%|█████████ | 1985/2182 [01:06<00:06, 29.24it/s]
 91%|█████████ | 1988/2182 [01:07<00:06, 29.37it/s]
 91%|█████████ | 1991/2182 [01:07<00:06, 29.21it/s]
 91%|█████████▏| 1994/2182 [01:07<00:06, 29.26it/s]
 92%|█████████▏| 1997/2182 [01:07<00:06, 29.46it/s]
 92%|█████████▏| 2000/2182 [01:07<00:06, 29.13it/s]
 92%|█████████▏| 2003/2182 [01:07<00:06, 28.54it/s]
 92%|█████████▏| 2006/2182 [01:07<00:06, 28.38it/s]
 92%|█████████▏| 2009/2182 [01:07<00:06, 28.56it/s]
 92%|█████████▏| 2013/2182 [01:07<00:05, 29.19it/s]
 92%|█████████▏| 2017/2182 [01:08<00:05, 29.32it/s]
 93%|█████████▎| 2020/2182 [01:08<00:05, 29.30it/s]
 93%|█████████▎| 2023/2182 [01:08<00:05, 29.17it/s]
 93%|█████████▎| 2026/2182 [01:08<00:05, 29.09it/s]
 93%|█████████▎| 2029/2182 [01:08<00:05, 28.70it/s]
 93%|█████████▎| 2032/2182 [01:08<00:05, 28.99it/s]
 93%|█████████▎| 2035/2182 [01:08<00:05, 29.25it/s]
 93%|█████████▎| 2038/2182 [01:08<00:04, 29.02it/s]
 94%|█████████▎| 2041/2182 [01:08<00:04, 28.95it/s]
 94%|█████████▎| 2044/2182 [01:08<00:04, 29.06it/s]
 94%|█████████▍| 2047/2182 [01:09<00:04, 28.78it/s]
 94%|█████████▍| 2050/2182 [01:09<00:04, 28.99it/s]
 94%|█████████▍| 2053/2182 [01:09<00:04, 28.93it/s]
 94%|█████████▍| 2056/2182 [01:09<00:04, 29.17it/s]
 94%|█████████▍| 2060/2182 [01:09<00:04, 29.81it/s]
 95%|█████████▍| 2063/2182 [01:09<00:04, 29.44it/s]
 95%|█████████▍| 2066/2182 [01:09<00:03, 29.37it/s]
 95%|█████████▍| 2069/2182 [01:09<00:03, 29.48it/s]
 95%|█████████▍| 2072/2182 [01:09<00:03, 29.63it/s]
 95%|█████████▌| 2075/2182 [01:10<00:03, 29.53it/s]
 95%|█████████▌| 2078/2182 [01:10<00:03, 29.47it/s]
 95%|█████████▌| 2081/2182 [01:10<00:03, 29.25it/s]
 96%|█████████▌| 2085/2182 [01:10<00:03, 29.49it/s]
 96%|█████████▌| 2089/2182 [01:10<00:03, 29.90it/s]
 96%|█████████▌| 2092/2182 [01:10<00:03, 29.73it/s]
 96%|█████████▌| 2095/2182 [01:10<00:02, 29.52it/s]
 96%|█████████▌| 2098/2182 [01:10<00:02, 29.40it/s]
 96%|█████████▋| 2101/2182 [01:10<00:02, 29.46it/s]
 96%|█████████▋| 2104/2182 [01:11<00:02, 29.10it/s]
 97%|█████████▋| 2107/2182 [01:11<00:02, 28.75it/s]
 97%|█████████▋| 2110/2182 [01:11<00:02, 28.77it/s]
 97%|█████████▋| 2113/2182 [01:11<00:02, 28.90it/s]
 97%|█████████▋| 2117/2182 [01:11<00:02, 29.06it/s]
 97%|█████████▋| 2120/2182 [01:11<00:02, 29.23it/s]
 97%|█████████▋| 2123/2182 [01:11<00:02, 29.10it/s]
 97%|█████████▋| 2126/2182 [01:11<00:01, 29.00it/s]
 98%|█████████▊| 2129/2182 [01:11<00:01, 29.20it/s]
 98%|█████████▊| 2132/2182 [01:11<00:01, 29.17it/s]
 98%|█████████▊| 2135/2182 [01:12<00:01, 29.06it/s]
 98%|█████████▊| 2138/2182 [01:12<00:01, 29.00it/s]
 98%|█████████▊| 2141/2182 [01:12<00:01, 28.92it/s]
 98%|█████████▊| 2144/2182 [01:12<00:01, 29.12it/s]
 98%|█████████▊| 2147/2182 [01:12<00:01, 29.05it/s]
 99%|█████████▊| 2150/2182 [01:12<00:01, 28.93it/s]
 99%|█████████▊| 2153/2182 [01:12<00:00, 29.14it/s]
 99%|█████████▉| 2156/2182 [01:12<00:00, 29.14it/s]
 99%|█████████▉| 2159/2182 [01:12<00:00, 29.03it/s]
 99%|█████████▉| 2162/2182 [01:13<00:00, 29.03it/s]
 99%|█████████▉| 2165/2182 [01:13<00:00, 29.17it/s]
 99%|█████████▉| 2168/2182 [01:13<00:00, 29.25it/s]
 99%|█████████▉| 2171/2182 [01:13<00:00, 29.25it/s]
+FID: 46.25438786246855
+
+INFO:absl:FID: 46.25438786246855
+FID: 46.25438786246855

scripts/eval/slurm-6226595.out

@@ -0,0 +1,47 @@
+loading annotations into memory...
+Done (t=0.03s)
+creating index...
+index created!
+一共385 张
+loading annotations into memory...
+Done (t=0.03s)
+creating index...
+index created!
+/mnt/petrelfs/leihaodong/ICLR25/exp/Emu3/MSCOCO2017Val/base/img/gt
+/mnt/petrelfs/leihaodong/ICLR25/exp/Emu3/MSCOCO2017Val/base/img
+
  0%|          | 0/385 [00:00<?, ?it/s]
  0%|          | 1/385 [00:00<03:01,  2.11it/s]
  2%|▏         | 8/385 [00:00<00:21, 17.27it/s]
  4%|▍         | 15/385 [00:00<00:12, 29.03it/s]
  5%|▌         | 21/385 [00:00<00:10, 36.04it/s]
  8%|▊         | 29/385 [00:00<00:07, 45.52it/s]
 10%|▉         | 37/385 [00:01<00:06, 52.78it/s]
 12%|█▏        | 45/385 [00:01<00:05, 57.93it/s]
 14%|█▎        | 52/385 [00:01<00:05, 60.65it/s]
 15%|█▌        | 59/385 [00:01<00:05, 63.13it/s]
 17%|█▋        | 66/385 [00:01<00:04, 64.64it/s]
 19%|█▉        | 73/385 [00:01<00:04, 66.08it/s]
 21%|██        | 81/385 [00:01<00:04, 67.86it/s]
 23%|██▎       | 88/385 [00:01<00:04, 67.05it/s]
 25%|██▍       | 95/385 [00:01<00:04, 67.37it/s]
 26%|██▋       | 102/385 [00:01<00:04, 67.29it/s]
 28%|██▊       | 109/385 [00:02<00:04, 67.99it/s]
 30%|███       | 116/385 [00:02<00:04, 67.05it/s]
 32%|███▏      | 123/385 [00:02<00:03, 67.34it/s]
 34%|███▍      | 131/385 [00:02<00:03, 68.10it/s]
 36%|███▌      | 139/385 [00:02<00:03, 69.27it/s]
 38%|███▊      | 146/385 [00:02<00:03, 68.78it/s]
 40%|███▉      | 153/385 [00:02<00:03, 68.60it/s]
 42%|████▏     | 160/385 [00:02<00:03, 69.00it/s]
 43%|████▎     | 167/385 [00:02<00:03, 69.03it/s]
 45%|████▌     | 174/385 [00:03<00:03, 68.10it/s]
 47%|████▋     | 181/385 [00:03<00:03, 67.15it/s]
 49%|████▉     | 188/385 [00:03<00:02, 67.43it/s]
 51%|█████     | 195/385 [00:03<00:02, 67.77it/s]
 53%|█████▎    | 203/385 [00:03<00:02, 68.94it/s]
 55%|█████▍    | 210/385 [00:03<00:02, 68.18it/s]
 56%|█████▋    | 217/385 [00:03<00:02, 68.40it/s]
 58%|█████▊    | 224/385 [00:03<00:02, 67.26it/s]
 60%|██████    | 232/385 [00:03<00:02, 68.09it/s]
 62%|██████▏   | 239/385 [00:03<00:02, 68.58it/s]
 64%|██████▍   | 246/385 [00:04<00:02, 68.33it/s]
 66%|██████▌   | 254/385 [00:04<00:01, 68.94it/s]
 68%|██████▊   | 262/385 [00:04<00:01, 69.87it/s]
 70%|██████▉   | 269/385 [00:04<00:01, 69.61it/s]
 72%|███████▏  | 276/385 [00:04<00:01, 68.92it/s]
 74%|███████▎  | 283/385 [00:04<00:01, 66.45it/s]
 75%|███████▌  | 290/385 [00:04<00:01, 64.53it/s]
 77%|███████▋  | 298/385 [00:04<00:01, 66.54it/s]
 79%|███████▉  | 305/385 [00:04<00:01, 67.17it/s]
 81%|████████▏ | 313/385 [00:05<00:01, 68.39it/s]
 83%|████████▎ | 320/385 [00:05<00:00, 68.81it/s]
 85%|████████▌ | 328/385 [00:05<00:00, 69.26it/s]
 87%|████████▋ | 335/385 [00:05<00:00, 64.76it/s]
 89%|████████▉ | 342/385 [00:05<00:00, 65.84it/s]
 91%|█████████ | 349/385 [00:05<00:00, 66.85it/s]
 92%|█████████▏| 356/385 [00:05<00:00, 67.42it/s]
 94%|█████████▍| 363/385 [00:05<00:00, 67.82it/s]
 96%|█████████▋| 371/385 [00:05<00:00, 68.71it/s]
 98%|█████████▊| 378/385 [00:06<00:00, 68.61it/s]
+
  0%|          | 0/385 [00:00<?, ?it/s]
  1%|          | 3/385 [00:00<00:14, 26.74it/s]
  2%|▏         | 7/385 [00:00<00:12, 29.39it/s]
  3%|▎         | 11/385 [00:00<00:12, 30.83it/s]
  4%|▍         | 15/385 [00:00<00:11, 31.48it/s]
  5%|▍         | 19/385 [00:00<00:11, 31.27it/s]
  6%|▌         | 23/385 [00:00<00:11, 31.36it/s]
  7%|▋         | 27/385 [00:00<00:11, 31.51it/s]
  8%|▊         | 31/385 [00:00<00:11, 31.73it/s]
  9%|▉         | 35/385 [00:01<00:11, 31.43it/s]
 10%|█         | 39/385 [00:01<00:10, 31.63it/s]
 11%|█         | 43/385 [00:01<00:10, 31.77it/s]
 12%|█▏        | 47/385 [00:01<00:11, 30.12it/s]
 13%|█▎        | 51/385 [00:01<00:10, 30.45it/s]
 14%|█▍        | 55/385 [00:01<00:10, 30.72it/s]
 15%|█▌        | 59/385 [00:01<00:10, 30.87it/s]
 16%|█▋        | 63/385 [00:02<00:10, 29.43it/s]
 17%|█▋        | 67/385 [00:02<00:10, 29.71it/s]
 18%|█▊        | 71/385 [00:02<00:10, 30.65it/s]
 19%|█▉        | 75/385 [00:02<00:09, 31.29it/s]
 21%|██        | 79/385 [00:02<00:09, 31.20it/s]
 22%|██▏       | 83/385 [00:02<00:09, 31.28it/s]
 23%|██▎       | 87/385 [00:02<00:09, 30.96it/s]
 24%|██▎       | 91/385 [00:02<00:09, 31.21it/s]
 25%|██▍       | 95/385 [00:03<00:09, 31.10it/s]
 26%|██▌       | 99/385 [00:03<00:10, 28.10it/s]
 26%|██▋       | 102/385 [00:03<00:10, 26.59it/s]
 27%|██▋       | 105/385 [00:03<00:10, 27.09it/s]
 28%|██▊       | 109/385 [00:03<00:09, 28.86it/s]
 29%|██▉       | 113/385 [00:03<00:09, 29.80it/s]
 30%|███       | 117/385 [00:03<00:08, 30.54it/s]
 31%|███▏      | 121/385 [00:03<00:08, 30.57it/s]
 32%|███▏      | 125/385 [00:04<00:08, 31.26it/s]
 34%|███▎      | 129/385 [00:04<00:08, 31.17it/s]
 35%|███▍      | 133/385 [00:04<00:08, 30.72it/s]
 36%|███▌      | 137/385 [00:04<00:08, 30.22it/s]
 37%|███▋      | 141/385 [00:04<00:08, 30.37it/s]
 38%|███▊      | 145/385 [00:04<00:07, 30.11it/s]
 39%|███▊      | 149/385 [00:04<00:07, 30.50it/s]
 40%|███▉      | 153/385 [00:05<00:07, 31.17it/s]
 41%|████      | 157/385 [00:05<00:07, 31.82it/s]
 42%|████▏     | 161/385 [00:05<00:07, 31.74it/s]
 43%|████▎     | 165/385 [00:05<00:06, 31.92it/s]
 44%|████▍     | 169/385 [00:05<00:06, 32.17it/s]
 45%|████▍     | 173/385 [00:05<00:06, 31.28it/s]
 46%|████▌     | 177/385 [00:05<00:06, 32.40it/s]
 47%|████▋     | 181/385 [00:05<00:06, 32.73it/s]
 48%|████▊     | 185/385 [00:06<00:06, 32.32it/s]
 49%|████▉     | 189/385 [00:06<00:06, 32.61it/s]
 50%|█████     | 193/385 [00:06<00:05, 32.62it/s]
 51%|█████     | 197/385 [00:06<00:05, 32.57it/s]
 52%|█████▏    | 201/385 [00:06<00:05, 33.05it/s]
 53%|█████▎    | 205/385 [00:06<00:05, 32.70it/s]
 54%|█████▍    | 209/385 [00:06<00:05, 32.97it/s]
 55%|█████▌    | 213/385 [00:06<00:05, 32.66it/s]
 56%|█████▋    | 217/385 [00:07<00:05, 30.70it/s]
 57%|█████▋    | 221/385 [00:07<00:05, 31.54it/s]
 58%|█████▊    | 225/385 [00:07<00:05, 31.63it/s]
 59%|█████▉    | 229/385 [00:07<00:05, 30.65it/s]
 61%|██████    | 233/385 [00:07<00:04, 31.28it/s]
 62%|██████▏   | 237/385 [00:07<00:04, 31.08it/s]
 63%|██████▎   | 241/385 [00:07<00:04, 31.81it/s]
 64%|██████▎   | 245/385 [00:07<00:04, 32.00it/s]
 65%|██████▍   | 249/385 [00:08<00:04, 32.28it/s]
 66%|██████▌   | 253/385 [00:08<00:04, 31.71it/s]
 67%|██████▋   | 257/385 [00:08<00:04, 31.69it/s]
 68%|██████▊   | 261/385 [00:08<00:03, 31.82it/s]
 69%|██████▉   | 265/385 [00:08<00:03, 31.98it/s]
 70%|██████▉   | 269/385 [00:08<00:03, 31.65it/s]
 71%|███████   | 273/385 [00:08<00:03, 31.44it/s]
 72%|███████▏  | 277/385 [00:08<00:03, 31.54it/s]
 73%|███████▎  | 281/385 [00:09<00:03, 31.85it/s]
 74%|███████▍  | 285/385 [00:09<00:03, 31.75it/s]
 75%|███████▌  | 289/385 [00:09<00:03, 30.97it/s]
 76%|███████▌  | 293/385 [00:09<00:02, 31.30it/s]
 77%|███████▋  | 297/385 [00:09<00:02, 31.29it/s]
 78%|███████▊  | 301/385 [00:09<00:02, 31.04it/s]
 79%|███████▉  | 305/385 [00:09<00:02, 31.18it/s]
 80%|████████  | 309/385 [00:09<00:02, 31.54it/s]
 81%|████████▏ | 313/385 [00:10<00:02, 31.91it/s]
 82%|████████▏ | 317/385 [00:10<00:02, 31.95it/s]
 83%|████████▎ | 321/385 [00:10<00:01, 32.15it/s]
 84%|████████▍ | 325/385 [00:10<00:01, 32.03it/s]
 85%|████████▌ | 329/385 [00:10<00:01, 32.19it/s]
 86%|████████▋ | 333/385 [00:10<00:01, 32.38it/s]
 88%|████████▊ | 337/385 [00:10<00:01, 32.25it/s]
 89%|████████▊ | 341/385 [00:10<00:01, 31.86it/s]
 90%|████████▉ | 345/385 [00:11<00:01, 31.98it/s]
 91%|█████████ | 349/385 [00:11<00:01, 31.56it/s]
 92%|█████████▏| 353/385 [00:11<00:00, 32.00it/s]
 93%|█████████▎| 357/385 [00:11<00:00, 31.45it/s]
 94%|█████████▍| 361/385 [00:11<00:00, 30.78it/s]
 95%|█████████▍| 365/385 [00:11<00:00, 29.92it/s]
 96%|█████████▌| 369/385 [00:11<00:00, 30.33it/s]
 97%|█████████▋| 373/385 [00:11<00:00, 30.66it/s]
 98%|█████████▊| 377/385 [00:12<00:00, 30.97it/s]
 99%|█████████▉| 381/385 [00:12<00:00, 31.16it/s]
+FID: 119.18770598565544
+
+/mnt/petrelfs/leihaodong/anaconda3/envs/eval_xx/lib/python3.10/site-packages/torchmetrics/utilities/prints.py:43: UserWarning: Metric `InceptionScore` will save all extracted features in buffer. For large datasets this may lead to large memory footprint.
+  warnings.warn(*args, **kwargs)
+/mnt/petrelfs/leihaodong/anaconda3/envs/eval_xx/lib/python3.10/site-packages/open_clip/factory.py:450: UserWarning: QuickGELU mismatch between final model config (quick_gelu=False) and pretrained tag 'openai' (quick_gelu=True).
+  warnings.warn(
+Using a slow image processor as `use_fast` is unset and a slow processor was saved with this model. `use_fast=True` will be the default behavior in v4.52, even if the model was saved with a slow processor. This will result in minor differences in outputs. You'll still be able to use a slow processor with `use_fast=False`.
+DINO_I: 0.13228578865528107
+
+  warnings.warn(f"Importing from {__name__} is deprecated, please import via timm.layers", FutureWarning)
+/mnt/petrelfs/leihaodong/ICLR25/eval/HPSv2/hpsv2/img_score.py:107: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead.
+  with torch.cuda.amp.autocast():
+
+WARNING:huggingface_hub.utils._http:'(ReadTimeoutError("HTTPSConnectionPool(host='huggingface.co', port=443): Read timed out. (read timeout=10)"), '(Request ID: 9defb932-bd4d-4fc2-b28a-5194c5ed5ad3)')' thrown while requesting HEAD https://huggingface.co/xswu/HPSv2/resolve/main/HPS_v2.1_compressed.pt
+Retrying in 1s [Retry 1/5].
+WARNING:huggingface_hub.utils._http:Retrying in 1s [Retry 1/5].
+
+WARNING:huggingface_hub.utils._http:'(ReadTimeoutError("HTTPSConnectionPool(host='huggingface.co', port=443): Read timed out. (read timeout=10)"), '(Request ID: b752cc3d-6a5a-4eec-9230-15b763bf6a45)')' thrown while requesting HEAD https://huggingface.co/xswu/HPSv2/resolve/main/HPS_v2.1_compressed.pt
+Retrying in 1s [Retry 1/5].
+WARNING:huggingface_hub.utils._http:Retrying in 1s [Retry 1/5].
+
+INFO:absl:FID: 119.18770598565544
+INFO:absl:DINO_I: 0.13228578865528107
+INFO:absl:IS: (tensor(14.3733, device='cuda:0'), tensor(1.8658, device='cuda:0'))
+INFO:absl:CLIP_T: 0.19042901381269678
+INFO:absl:HPSv2: 0.2911430854301948
+INFO:absl:Aesthetic: 6.609675357868145
+Scores:
+FID: 119.18770598565544
+DINO_I: 0.13228578865528107
+IS: (tensor(14.3733, device='cuda:0'), tensor(1.8658, device='cuda:0'))
+CLIP_T: 0.19042901381269678
+HPSv2: 0.2911430854301948
+Aesthetic: 6.609675357868145

scripts/eval/slurm-6226596.out

@@ -0,0 +1,59 @@
+loading annotations into memory...
+Done (t=0.03s)
+creating index...
+index created!
+一共554 张
+loading annotations into memory...
+Done (t=0.03s)
+creating index...
+index created!
+/mnt/petrelfs/leihaodong/ICLR25/exp/Emu3/MSCOCO2017Val/gsd/img/gt
+/mnt/petrelfs/leihaodong/ICLR25/exp/Emu3/MSCOCO2017Val/gsd/img
+
  0%|          | 0/554 [00:00<?, ?it/s]
  0%|          | 1/554 [00:00<03:41,  2.50it/s]
  1%|▏         | 8/554 [00:00<00:27, 19.94it/s]
  3%|▎         | 15/554 [00:00<00:16, 33.22it/s]
  4%|▍         | 22/554 [00:00<00:12, 43.04it/s]
  5%|▌         | 30/554 [00:00<00:10, 51.55it/s]
  7%|▋         | 37/554 [00:00<00:09, 56.68it/s]
  8%|▊         | 45/554 [00:01<00:08, 60.69it/s]
  9%|▉         | 52/554 [00:01<00:08, 62.72it/s]
 11%|█         | 59/554 [00:01<00:07, 64.04it/s]
 12%|█▏        | 66/554 [00:01<00:07, 64.96it/s]
 13%|█▎        | 73/554 [00:01<00:07, 65.77it/s]
 14%|█▍        | 80/554 [00:01<00:07, 66.24it/s]
 16%|█▌        | 87/554 [00:01<00:06, 67.21it/s]
 17%|█▋        | 94/554 [00:01<00:06, 67.65it/s]
 18%|█▊        | 102/554 [00:01<00:06, 68.71it/s]
 20%|█▉        | 109/554 [00:01<00:06, 68.74it/s]
 21%|██        | 116/554 [00:02<00:06, 68.53it/s]
 22%|██▏       | 123/554 [00:02<00:06, 68.03it/s]
 24%|██▎       | 131/554 [00:02<00:06, 68.79it/s]
 25%|██▍       | 138/554 [00:02<00:06, 67.36it/s]
 26%|██▌       | 145/554 [00:02<00:06, 67.21it/s]
 27%|██▋       | 152/554 [00:02<00:05, 67.44it/s]
 29%|██▉       | 160/554 [00:02<00:05, 68.48it/s]
 30%|███       | 167/554 [00:02<00:05, 68.58it/s]
 31%|███▏      | 174/554 [00:02<00:05, 68.20it/s]
 33%|███▎      | 181/554 [00:03<00:05, 67.07it/s]
 34%|███▍      | 189/554 [00:03<00:05, 68.25it/s]
 35%|███▌      | 196/554 [00:03<00:05, 68.48it/s]
 37%|███▋      | 203/554 [00:03<00:05, 68.34it/s]
 38%|███▊      | 210/554 [00:03<00:05, 68.09it/s]
 39%|███▉      | 218/554 [00:03<00:04, 69.15it/s]
 41%|████      | 226/554 [00:03<00:04, 69.84it/s]
 42%|████▏     | 234/554 [00:03<00:04, 70.13it/s]
 44%|████▎     | 242/554 [00:03<00:04, 69.25it/s]
 45%|████▍     | 249/554 [00:04<00:04, 67.02it/s]
 46%|████▌     | 256/554 [00:04<00:04, 67.38it/s]
 47%|████▋     | 263/554 [00:04<00:04, 67.52it/s]
 49%|████▊     | 270/554 [00:04<00:04, 66.50it/s]
 50%|█████     | 277/554 [00:04<00:04, 67.10it/s]
 51%|█████▏    | 284/554 [00:04<00:03, 67.86it/s]
 53%|█████▎    | 291/554 [00:04<00:03, 68.18it/s]
 54%|█████▍    | 298/554 [00:04<00:03, 67.29it/s]
 55%|█████▌    | 305/554 [00:04<00:03, 67.20it/s]
 56%|█████▋    | 312/554 [00:04<00:03, 67.87it/s]
 58%|█████▊    | 319/554 [00:05<00:03, 68.20it/s]
 59%|█████▉    | 327/554 [00:05<00:03, 69.01it/s]
 60%|██████    | 334/554 [00:05<00:03, 66.84it/s]
 62%|██████▏   | 341/554 [00:05<00:03, 66.10it/s]
 63%|██████▎   | 348/554 [00:05<00:03, 66.51it/s]
 64%|██████▍   | 355/554 [00:05<00:03, 62.99it/s]
 65%|██████▌   | 362/554 [00:05<00:03, 63.25it/s]
 67%|██████▋   | 370/554 [00:05<00:02, 65.79it/s]
 68%|██████▊   | 377/554 [00:05<00:02, 65.47it/s]
 69%|██████▉   | 384/554 [00:06<00:02, 66.00it/s]
 71%|███████   | 392/554 [00:06<00:02, 67.75it/s]
 72%|███████▏  | 399/554 [00:06<00:02, 65.79it/s]
 73%|███████▎  | 406/554 [00:06<00:02, 64.53it/s]
 75%|███████▍  | 413/554 [00:06<00:02, 64.66it/s]
 76%|███████▌  | 420/554 [00:06<00:02, 65.38it/s]
 77%|███████▋  | 427/554 [00:06<00:01, 66.62it/s]
 78%|███████▊  | 434/554 [00:06<00:01, 66.34it/s]
 80%|███████▉  | 441/554 [00:06<00:01, 65.35it/s]
 81%|████████  | 448/554 [00:07<00:01, 63.36it/s]
 82%|████████▏ | 455/554 [00:07<00:01, 62.80it/s]
 84%|████████▎ | 463/554 [00:07<00:01, 65.23it/s]
 85%|████████▍ | 470/554 [00:07<00:01, 65.72it/s]
 86%|████████▋ | 478/554 [00:07<00:01, 66.90it/s]
 88%|████████▊ | 485/554 [00:07<00:01, 64.85it/s]
 89%|████████▉ | 493/554 [00:07<00:00, 66.19it/s]
 90%|█████████ | 500/554 [00:07<00:00, 66.34it/s]
 92%|█████████▏| 507/554 [00:07<00:00, 65.09it/s]
 93%|█████████▎| 514/554 [00:08<00:00, 64.22it/s]
 94%|█████████▍| 521/554 [00:08<00:00, 63.97it/s]
 95%|█████████▌| 528/554 [00:08<00:00, 64.74it/s]
 97%|█████████▋| 535/554 [00:08<00:00, 65.04it/s]
 98%|█████████▊| 542/554 [00:08<00:00, 65.28it/s]
 99%|█████████▉| 549/554 [00:08<00:00, 65.55it/s]
+
  0%|          | 0/554 [00:00<?, ?it/s]
  1%|          | 3/554 [00:00<00:21, 25.48it/s]
  1%|▏         | 7/554 [00:00<00:19, 28.73it/s]
  2%|▏         | 11/554 [00:00<00:18, 29.98it/s]
  3%|▎         | 14/554 [00:00<00:18, 29.93it/s]
  3%|▎         | 18/554 [00:00<00:17, 30.51it/s]
  4%|▍         | 22/554 [00:00<00:16, 31.44it/s]
  5%|▍         | 26/554 [00:00<00:17, 30.92it/s]
  5%|▌         | 30/554 [00:00<00:16, 31.06it/s]
  6%|▌         | 34/554 [00:01<00:17, 30.22it/s]
  7%|▋         | 38/554 [00:01<00:17, 28.69it/s]
  7%|▋         | 41/554 [00:01<00:18, 27.95it/s]
  8%|▊         | 44/554 [00:01<00:18, 27.61it/s]
  8%|▊         | 47/554 [00:01<00:18, 27.37it/s]
  9%|▉         | 50/554 [00:01<00:18, 26.55it/s]
 10%|▉         | 53/554 [00:01<00:19, 26.10it/s]
 10%|█         | 56/554 [00:01<00:19, 26.19it/s]
 11%|█         | 59/554 [00:02<00:18, 26.54it/s]
 11%|█         | 62/554 [00:02<00:18, 26.52it/s]
 12%|█▏        | 65/554 [00:02<00:18, 27.10it/s]
 12%|█▏        | 68/554 [00:02<00:17, 27.10it/s]
 13%|█▎        | 71/554 [00:02<00:17, 27.29it/s]
 13%|█▎        | 74/554 [00:02<00:17, 27.49it/s]
 14%|█▍        | 77/554 [00:02<00:17, 26.51it/s]
 14%|█▍        | 80/554 [00:02<00:18, 26.32it/s]
 15%|█▍        | 83/554 [00:02<00:17, 26.22it/s]
 16%|█▌        | 86/554 [00:03<00:17, 26.35it/s]
 16%|█▌        | 89/554 [00:03<00:17, 26.79it/s]
 17%|█▋        | 92/554 [00:03<00:16, 27.44it/s]
 17%|█▋        | 95/554 [00:03<00:16, 27.35it/s]
 18%|█▊        | 98/554 [00:03<00:17, 26.53it/s]
 18%|█▊        | 101/554 [00:03<00:17, 26.64it/s]
 19%|█▉        | 104/554 [00:03<00:17, 25.76it/s]
 19%|█▉        | 107/554 [00:03<00:17, 26.19it/s]
 20%|█▉        | 110/554 [00:04<00:17, 25.64it/s]
 20%|██        | 113/554 [00:04<00:17, 24.65it/s]
 21%|██        | 117/554 [00:04<00:16, 26.77it/s]
 22%|██▏       | 120/554 [00:04<00:16, 26.96it/s]
 22%|██▏       | 123/554 [00:04<00:15, 27.58it/s]
 23%|██▎       | 127/554 [00:04<00:14, 28.70it/s]
 23%|██▎       | 130/554 [00:04<00:14, 28.56it/s]
 24%|██▍       | 134/554 [00:04<00:14, 29.18it/s]
 25%|██▍       | 138/554 [00:04<00:13, 29.73it/s]
 25%|██▌       | 141/554 [00:05<00:13, 29.59it/s]
 26%|██▌       | 145/554 [00:05<00:13, 30.30it/s]
 27%|██▋       | 149/554 [00:05<00:13, 30.46it/s]
 28%|██▊       | 153/554 [00:05<00:13, 30.76it/s]
 28%|██▊       | 157/554 [00:05<00:12, 31.14it/s]
 29%|██▉       | 161/554 [00:05<00:12, 31.62it/s]
 30%|██▉       | 165/554 [00:05<00:12, 31.74it/s]
 31%|███       | 169/554 [00:05<00:11, 32.15it/s]
 31%|███       | 173/554 [00:06<00:12, 31.68it/s]
 32%|███▏      | 177/554 [00:06<00:11, 31.72it/s]
 33%|███▎      | 181/554 [00:06<00:11, 31.21it/s]
 33%|███▎      | 185/554 [00:06<00:11, 31.23it/s]
 34%|███▍      | 189/554 [00:06<00:11, 31.76it/s]
 35%|███▍      | 193/554 [00:06<00:11, 31.69it/s]
 36%|███▌      | 197/554 [00:06<00:11, 31.88it/s]
 36%|███▋      | 201/554 [00:06<00:11, 31.27it/s]
 37%|███▋      | 205/554 [00:07<00:11, 31.25it/s]
 38%|███▊      | 209/554 [00:07<00:11, 30.14it/s]
 38%|███▊      | 213/554 [00:07<00:11, 30.80it/s]
 39%|███▉      | 217/554 [00:07<00:10, 31.18it/s]
 40%|███▉      | 221/554 [00:07<00:10, 31.43it/s]
 41%|████      | 225/554 [00:07<00:10, 32.11it/s]
 41%|████▏     | 229/554 [00:07<00:10, 31.28it/s]
 42%|████▏     | 233/554 [00:08<00:10, 31.32it/s]
 43%|████▎     | 237/554 [00:08<00:10, 31.00it/s]
 44%|████▎     | 241/554 [00:08<00:10, 31.02it/s]
 44%|████▍     | 245/554 [00:08<00:10, 29.28it/s]
 45%|████▍     | 249/554 [00:08<00:10, 30.15it/s]
 46%|████▌     | 253/554 [00:08<00:09, 30.31it/s]
 46%|████▋     | 257/554 [00:08<00:09, 31.08it/s]
 47%|████▋     | 261/554 [00:08<00:09, 31.53it/s]
 48%|████▊     | 265/554 [00:09<00:09, 31.48it/s]
 49%|████▊     | 269/554 [00:09<00:09, 28.70it/s]
 49%|████▉     | 272/554 [00:09<00:11, 23.54it/s]
 50%|████▉     | 275/554 [00:09<00:14, 19.06it/s]
 50%|█████     | 278/554 [00:09<00:16, 16.84it/s]
 51%|█████     | 280/554 [00:10<00:17, 15.92it/s]
 51%|█████     | 282/554 [00:10<00:18, 15.08it/s]
 51%|█████▏    | 284/554 [00:10<00:18, 14.69it/s]
 52%|█████▏    | 286/554 [00:10<00:18, 14.16it/s]
 52%|█████▏    | 288/554 [00:10<00:19, 13.87it/s]
 52%|█████▏    | 290/554 [00:10<00:19, 13.76it/s]
 53%|█████▎    | 292/554 [00:10<00:19, 13.55it/s]
 53%|█████▎    | 294/554 [00:11<00:19, 13.44it/s]
 53%|█████▎    | 296/554 [00:11<00:19, 13.20it/s]
 54%|█████▍    | 298/554 [00:11<00:17, 14.66it/s]
 54%|█████▍    | 301/554 [00:11<00:14, 18.02it/s]
 55%|█████▍    | 304/554 [00:11<00:12, 20.38it/s]
 56%|█████▌    | 308/554 [00:11<00:10, 23.93it/s]
 56%|█████▋    | 312/554 [00:11<00:09, 26.35it/s]
 57%|█████▋    | 316/554 [00:11<00:08, 28.37it/s]
 58%|█████▊    | 320/554 [00:12<00:07, 29.38it/s]
 58%|█████▊    | 324/554 [00:12<00:08, 26.74it/s]
 59%|█████▉    | 327/554 [00:12<00:11, 20.39it/s]
 60%|█████▉    | 330/554 [00:12<00:10, 22.05it/s]
 60%|██████    | 334/554 [00:12<00:08, 25.08it/s]
 61%|██████    | 338/554 [00:12<00:08, 26.83it/s]
 62%|██████▏   | 342/554 [00:12<00:07, 28.65it/s]
 62%|██████▏   | 346/554 [00:13<00:06, 30.00it/s]
 63%|██████▎   | 350/554 [00:13<00:06, 30.88it/s]
 64%|██████▍   | 354/554 [00:13<00:06, 31.58it/s]
 65%|██████▍   | 358/554 [00:13<00:06, 31.56it/s]
 65%|██████▌   | 362/554 [00:13<00:06, 31.91it/s]
 66%|██████▌   | 366/554 [00:13<00:05, 32.35it/s]
 67%|██████▋   | 370/554 [00:13<00:05, 32.25it/s]
 68%|██████▊   | 374/554 [00:13<00:05, 31.43it/s]
 68%|██████▊   | 378/554 [00:14<00:05, 32.13it/s]
 69%|██████▉   | 382/554 [00:14<00:05, 32.48it/s]
 70%|██████▉   | 386/554 [00:14<00:05, 32.37it/s]
 70%|███████   | 390/554 [00:14<00:04, 32.96it/s]
 71%|███████   | 394/554 [00:14<00:04, 32.93it/s]
 72%|███████▏  | 398/554 [00:14<00:04, 33.14it/s]
 73%|███████▎  | 402/554 [00:14<00:04, 33.24it/s]
 73%|███████▎  | 406/554 [00:14<00:04, 32.92it/s]
 74%|███████▍  | 410/554 [00:15<00:04, 32.68it/s]
 75%|███████▍  | 414/554 [00:15<00:04, 32.34it/s]
 75%|███████▌  | 418/554 [00:15<00:04, 32.61it/s]
 76%|███████▌  | 422/554 [00:15<00:04, 32.52it/s]
 77%|███████▋  | 426/554 [00:15<00:03, 32.62it/s]
 78%|███████▊  | 430/554 [00:15<00:03, 32.40it/s]
 78%|███████▊  | 434/554 [00:15<00:03, 32.64it/s]
 79%|███████▉  | 438/554 [00:15<00:03, 32.64it/s]
 80%|███████▉  | 442/554 [00:16<00:03, 32.53it/s]
 81%|████████  | 446/554 [00:16<00:03, 32.63it/s]
 81%|████████  | 450/554 [00:16<00:03, 32.75it/s]
 82%|████████▏ | 454/554 [00:16<00:03, 33.13it/s]
 83%|████████▎ | 458/554 [00:16<00:02, 33.40it/s]
 83%|████████▎ | 462/554 [00:16<00:02, 33.58it/s]
 84%|████████▍ | 466/554 [00:16<00:02, 33.52it/s]
 85%|████████▍ | 470/554 [00:16<00:02, 33.33it/s]
 86%|████████▌ | 474/554 [00:16<00:02, 33.05it/s]
 86%|████████▋ | 478/554 [00:17<00:02, 33.43it/s]
 87%|████████▋ | 482/554 [00:17<00:02, 32.90it/s]
 88%|████████▊ | 486/554 [00:17<00:02, 33.21it/s]
 88%|████████▊ | 490/554 [00:17<00:01, 33.61it/s]
 89%|████████▉ | 494/554 [00:17<00:01, 33.63it/s]
 90%|████████▉ | 498/554 [00:17<00:01, 33.48it/s]
 91%|█████████ | 502/554 [00:17<00:01, 33.30it/s]
 91%|█████████▏| 506/554 [00:17<00:01, 33.61it/s]
 92%|█████████▏| 510/554 [00:18<00:01, 33.86it/s]
 93%|█████████▎| 514/554 [00:18<00:01, 33.19it/s]
 94%|█████████▎| 518/554 [00:18<00:01, 33.49it/s]
 94%|█████████▍| 522/554 [00:18<00:00, 34.29it/s]
 95%|█████████▍| 526/554 [00:18<00:00, 34.03it/s]
 96%|█████████▌| 530/554 [00:18<00:00, 33.88it/s]
 96%|█████████▋| 534/554 [00:18<00:00, 33.77it/s]
 97%|█████████▋| 538/554 [00:18<00:00, 34.01it/s]
 98%|█████████▊| 542/554 [00:19<00:00, 33.22it/s]
 99%|█████████▊| 546/554 [00:19<00:00, 33.29it/s]
 99%|█████████▉| 550/554 [00:19<00:00, 32.68it/s]
+FID: 102.64291574962897
+
+/mnt/petrelfs/leihaodong/anaconda3/envs/eval_xx/lib/python3.10/site-packages/torchmetrics/utilities/prints.py:43: UserWarning: Metric `InceptionScore` will save all extracted features in buffer. For large datasets this may lead to large memory footprint.
+  warnings.warn(*args, **kwargs)
+/mnt/petrelfs/leihaodong/anaconda3/envs/eval_xx/lib/python3.10/site-packages/open_clip/factory.py:450: UserWarning: QuickGELU mismatch between final model config (quick_gelu=False) and pretrained tag 'openai' (quick_gelu=True).
+  warnings.warn(
+Using a slow image processor as `use_fast` is unset and a slow processor was saved with this model. `use_fast=True` will be the default behavior in v4.52, even if the model was saved with a slow processor. This will result in minor differences in outputs. You'll still be able to use a slow processor with `use_fast=False`.
+DINO_I: 0.12038552016019821
+
+  warnings.warn(f"Importing from {__name__} is deprecated, please import via timm.layers", FutureWarning)
+/mnt/petrelfs/leihaodong/ICLR25/eval/HPSv2/hpsv2/img_score.py:107: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead.
+  with torch.cuda.amp.autocast():
+
+WARNING:huggingface_hub.utils._http:'(ReadTimeoutError("HTTPSConnectionPool(host='huggingface.co', port=443): Read timed out. (read timeout=10)"), '(Request ID: a67fcca2-7335-4bb8-89ad-5f427e21b640)')' thrown while requesting HEAD https://huggingface.co/xswu/HPSv2/resolve/main/HPS_v2.1_compressed.pt
+Retrying in 1s [Retry 1/5].
+WARNING:huggingface_hub.utils._http:Retrying in 1s [Retry 1/5].
+
+WARNING:huggingface_hub.utils._http:'(ReadTimeoutError("HTTPSConnectionPool(host='huggingface.co', port=443): Read timed out. (read timeout=10)"), '(Request ID: 241f7f09-1b72-4e2d-9975-cf1a1d5899d4)')' thrown while requesting HEAD https://huggingface.co/xswu/HPSv2/resolve/main/HPS_v2.1_compressed.pt
+Retrying in 1s [Retry 1/5].
+WARNING:huggingface_hub.utils._http:Retrying in 1s [Retry 1/5].
+
+WARNING:huggingface_hub.utils._http:'(ReadTimeoutError("HTTPSConnectionPool(host='huggingface.co', port=443): Read timed out. (read timeout=10)"), '(Request ID: ba68568b-efaf-45de-b261-b22e5ea15083)')' thrown while requesting HEAD https://huggingface.co/xswu/HPSv2/resolve/main/HPS_v2.1_compressed.pt
+Retrying in 1s [Retry 1/5].
+WARNING:huggingface_hub.utils._http:Retrying in 1s [Retry 1/5].
+'(ReadTimeoutError("HTTPSConnectionPool(host='huggingface.co', port=443): Read timed out. (read timeout=10)"), '(Request ID: c9f7997b-9a4d-4af8-93e9-3e2e85c296c6)')' thrown while requesting HEAD https://huggingface.co/xswu/HPSv2/resolve/main/HPS_v2.1_compressed.pt
+WARNING:huggingface_hub.utils._http:'(ReadTimeoutError("HTTPSConnectionPool(host='huggingface.co', port=443): Read timed out. (read timeout=10)"), '(Request ID: c9f7997b-9a4d-4af8-93e9-3e2e85c296c6)')' thrown while requesting HEAD https://huggingface.co/xswu/HPSv2/resolve/main/HPS_v2.1_compressed.pt
+Retrying in 2s [Retry 2/5].
+WARNING:huggingface_hub.utils._http:Retrying in 2s [Retry 2/5].
+
+WARNING:huggingface_hub.utils._http:'(ReadTimeoutError("HTTPSConnectionPool(host='huggingface.co', port=443): Read timed out. (read timeout=10)"), '(Request ID: e4fc11b2-5ccf-44af-af1f-6af66e4b39b5)')' thrown while requesting HEAD https://huggingface.co/xswu/HPSv2/resolve/main/HPS_v2.1_compressed.pt
+Retrying in 1s [Retry 1/5].
+WARNING:huggingface_hub.utils._http:Retrying in 1s [Retry 1/5].
+
+INFO:absl:FID: 102.64291574962897
+INFO:absl:DINO_I: 0.12038552016019821
+INFO:absl:IS: (tensor(17.6601, device='cuda:0'), tensor(1.5699, device='cuda:0'))
+INFO:absl:CLIP_T: 0.1882628591105826
+INFO:absl:HPSv2: 0.29271998216098827
+INFO:absl:Aesthetic: 6.625231840980613
+Scores:
+FID: 102.64291574962897
+DINO_I: 0.12038552016019821
+IS: (tensor(17.6601, device='cuda:0'), tensor(1.5699, device='cuda:0'))
+CLIP_T: 0.1882628591105826
+HPSv2: 0.29271998216098827
+Aesthetic: 6.625231840980613

scripts/eval/slurm-6226597.out

@@ -0,0 +1,51 @@
+loading annotations into memory...
+Done (t=0.03s)
+creating index...
+index created!
+一共497 张
+loading annotations into memory...
+Done (t=0.03s)
+creating index...
+index created!
+/mnt/petrelfs/leihaodong/ICLR25/exp/Emu3/MSCOCO2017Val/lantern/img/gt
+/mnt/petrelfs/leihaodong/ICLR25/exp/Emu3/MSCOCO2017Val/lantern/img
+
  0%|          | 0/497 [00:00<?, ?it/s]
  0%|          | 1/497 [00:00<04:51,  1.70it/s]
  1%|          | 3/497 [00:00<01:37,  5.07it/s]
  1%|          | 6/497 [00:00<00:51,  9.53it/s]
  2%|▏         | 9/497 [00:00<00:37, 13.09it/s]
  2%|▏         | 12/497 [00:01<00:29, 16.49it/s]
  3%|▎         | 16/497 [00:01<00:22, 21.05it/s]
  4%|▍         | 20/497 [00:01<00:19, 24.80it/s]
  5%|▍         | 23/497 [00:01<00:18, 25.14it/s]
  5%|▌         | 26/497 [00:01<00:19, 24.43it/s]
  6%|▌         | 29/497 [00:01<00:20, 22.46it/s]
  7%|▋         | 33/497 [00:01<00:18, 25.27it/s]
  7%|▋         | 37/497 [00:02<00:18, 25.31it/s]
  8%|▊         | 40/497 [00:02<00:17, 25.91it/s]
  9%|▉         | 44/497 [00:02<00:16, 27.88it/s]
  9%|▉         | 47/497 [00:02<00:17, 26.36it/s]
 10%|█         | 50/497 [00:02<00:20, 21.93it/s]
 11%|█         | 54/497 [00:02<00:18, 24.20it/s]
 12%|█▏        | 58/497 [00:02<00:17, 24.92it/s]
 12%|█▏        | 61/497 [00:03<00:18, 23.20it/s]
 13%|█▎        | 64/497 [00:03<00:20, 21.49it/s]
 13%|█▎        | 67/497 [00:03<00:18, 23.23it/s]
 14%|█▍        | 70/497 [00:03<00:19, 21.91it/s]
 15%|█▍        | 73/497 [00:03<00:21, 20.01it/s]
 15%|█▌        | 76/497 [00:03<00:19, 21.52it/s]
 16%|█▌        | 79/497 [00:03<00:20, 20.79it/s]
 16%|█▋        | 82/497 [00:04<00:21, 19.31it/s]
 17%|█▋        | 85/497 [00:04<00:20, 20.21it/s]
 18%|█▊        | 88/497 [00:04<00:18, 21.64it/s]
 18%|█▊        | 91/497 [00:04<00:17, 23.38it/s]
 19%|█▉        | 94/497 [00:04<00:16, 24.49it/s]
 20%|█▉        | 97/497 [00:04<00:16, 23.60it/s]
 20%|██        | 101/497 [00:04<00:14, 27.60it/s]
 21%|██        | 104/497 [00:04<00:14, 26.68it/s]
 22%|██▏       | 108/497 [00:05<00:14, 26.56it/s]
 22%|██▏       | 111/497 [00:05<00:16, 23.62it/s]
 23%|██▎       | 114/497 [00:05<00:16, 23.22it/s]
 24%|██▎       | 117/497 [00:05<00:16, 22.74it/s]
 24%|██▍       | 120/497 [00:05<00:15, 23.71it/s]
 25%|██▍       | 123/497 [00:05<00:16, 22.96it/s]
 26%|██▌       | 127/497 [00:05<00:15, 23.75it/s]
 26%|██▌       | 130/497 [00:06<00:15, 23.29it/s]
 27%|██▋       | 133/497 [00:06<00:15, 23.24it/s]
 27%|██▋       | 136/497 [00:06<00:15, 23.65it/s]
 28%|██▊       | 140/497 [00:06<00:14, 24.90it/s]
 29%|██▉       | 144/497 [00:06<00:14, 24.66it/s]
 30%|██▉       | 147/497 [00:06<00:16, 20.70it/s]
 30%|███       | 150/497 [00:06<00:16, 20.84it/s]
 31%|███       | 153/497 [00:07<00:16, 20.54it/s]
 32%|███▏      | 157/497 [00:07<00:14, 24.03it/s]
 32%|███▏      | 160/497 [00:07<00:13, 25.10it/s]
 33%|███▎      | 163/497 [00:07<00:14, 22.55it/s]
 33%|███▎      | 166/497 [00:07<00:14, 22.22it/s]
 34%|███▍      | 169/497 [00:07<00:14, 22.40it/s]
 35%|███▍      | 172/497 [00:07<00:13, 23.86it/s]
 35%|███▌      | 175/497 [00:07<00:12, 25.19it/s]
 36%|███▌      | 179/497 [00:08<00:11, 27.48it/s]
 37%|███▋      | 182/497 [00:08<00:13, 23.87it/s]
 37%|███▋      | 185/497 [00:08<00:12, 24.32it/s]
 38%|███▊      | 188/497 [00:08<00:12, 24.82it/s]
 38%|███▊      | 191/497 [00:08<00:12, 24.51it/s]
 39%|███▉      | 194/497 [00:08<00:12, 24.04it/s]
 40%|███▉      | 197/497 [00:08<00:12, 23.98it/s]
 40%|████      | 200/497 [00:08<00:11, 25.03it/s]
 41%|████      | 203/497 [00:09<00:12, 23.53it/s]
 42%|████▏     | 207/497 [00:09<00:11, 24.77it/s]
 42%|████▏     | 211/497 [00:09<00:10, 26.98it/s]
 43%|████▎     | 214/497 [00:09<00:11, 24.19it/s]
 44%|████▎     | 217/497 [00:09<00:12, 22.98it/s]
 44%|████▍     | 221/497 [00:09<00:10, 26.67it/s]
 45%|████▌     | 224/497 [00:09<00:11, 24.69it/s]
 46%|████▌     | 227/497 [00:10<00:11, 22.56it/s]
 46%|████▋     | 231/497 [00:10<00:10, 25.98it/s]
 47%|████▋     | 234/497 [00:10<00:10, 24.89it/s]
 48%|████▊     | 237/497 [00:10<00:10, 25.09it/s]
 48%|████▊     | 240/497 [00:10<00:11, 22.73it/s]
 49%|████▉     | 243/497 [00:10<00:11, 23.06it/s]
 49%|████▉     | 246/497 [00:10<00:11, 22.32it/s]
 50%|█████     | 249/497 [00:11<00:10, 23.20it/s]
 51%|█████     | 253/497 [00:11<00:08, 27.33it/s]
 52%|█████▏    | 256/497 [00:11<00:09, 24.72it/s]
 52%|█████▏    | 259/497 [00:11<00:12, 19.44it/s]
 53%|█████▎    | 265/497 [00:11<00:08, 27.08it/s]
 54%|█████▍    | 269/497 [00:11<00:08, 27.17it/s]
 55%|█████▌    | 275/497 [00:11<00:06, 33.31it/s]
 57%|█████▋    | 281/497 [00:11<00:05, 38.94it/s]
 58%|█████▊    | 288/497 [00:12<00:04, 44.89it/s]
 59%|█████▉    | 294/497 [00:12<00:04, 47.50it/s]
 60%|██████    | 300/497 [00:12<00:03, 49.76it/s]
 62%|██████▏   | 306/497 [00:12<00:03, 51.95it/s]
 63%|██████▎   | 312/497 [00:12<00:03, 53.12it/s]
 64%|██████▍   | 318/497 [00:12<00:03, 53.49it/s]
 65%|██████▌   | 325/497 [00:12<00:03, 54.90it/s]
 67%|██████▋   | 331/497 [00:12<00:03, 52.56it/s]
 68%|██████▊   | 337/497 [00:13<00:03, 48.73it/s]
 69%|██████▉   | 344/497 [00:13<00:02, 51.29it/s]
 71%|███████   | 351/497 [00:13<00:02, 54.75it/s]
 72%|███████▏  | 357/497 [00:13<00:02, 54.52it/s]
 73%|███████▎  | 363/497 [00:13<00:02, 54.75it/s]
 74%|███████▍  | 369/497 [00:13<00:02, 54.22it/s]
 75%|███████▌  | 375/497 [00:13<00:02, 53.64it/s]
 77%|███████▋  | 381/497 [00:13<00:02, 54.51it/s]
 78%|███████▊  | 387/497 [00:13<00:02, 47.61it/s]
 79%|███████▉  | 393/497 [00:14<00:02, 50.31it/s]
 80%|████████  | 399/497 [00:14<00:01, 52.54it/s]
 81%|████████▏ | 405/497 [00:14<00:01, 48.47it/s]
 83%|████████▎ | 412/497 [00:14<00:01, 50.30it/s]
 84%|████████▍ | 418/497 [00:14<00:01, 50.90it/s]
 85%|████████▌ | 424/497 [00:14<00:01, 47.22it/s]
 87%|████████▋ | 430/497 [00:14<00:01, 50.32it/s]
 88%|████████▊ | 436/497 [00:14<00:01, 51.60it/s]
 89%|████████▉ | 443/497 [00:15<00:00, 55.13it/s]
 90%|█████████ | 449/497 [00:15<00:01, 45.48it/s]
 91%|█████████▏| 454/497 [00:15<00:01, 40.51it/s]
 92%|█████████▏| 459/497 [00:15<00:00, 40.74it/s]
 93%|█████████▎| 464/497 [00:15<00:00, 42.86it/s]
 95%|█████████▍| 470/497 [00:15<00:00, 46.03it/s]
 96%|█████████▌| 475/497 [00:15<00:00, 41.95it/s]
 97%|█████████▋| 480/497 [00:15<00:00, 43.83it/s]
 98%|█████████▊| 486/497 [00:16<00:00, 44.91it/s]
 99%|█████████▉| 492/497 [00:16<00:00, 47.34it/s]
+
  0%|          | 0/497 [00:00<?, ?it/s]
  1%|          | 3/497 [00:00<00:18, 26.73it/s]
  1%|          | 6/497 [00:00<00:17, 28.20it/s]
  2%|▏         | 10/497 [00:00<00:16, 29.18it/s]
  3%|▎         | 13/497 [00:00<00:16, 29.45it/s]
  3%|▎         | 16/497 [00:00<00:16, 29.60it/s]
  4%|▍         | 20/497 [00:00<00:15, 30.36it/s]
  5%|▍         | 24/497 [00:00<00:15, 29.71it/s]
  5%|▌         | 27/497 [00:00<00:16, 29.05it/s]
  6%|▌         | 31/497 [00:01<00:16, 28.87it/s]
  7%|▋         | 34/497 [00:01<00:17, 26.95it/s]
  8%|▊         | 38/497 [00:01<00:16, 28.34it/s]
  8%|▊         | 41/497 [00:01<00:17, 25.42it/s]
  9%|▉         | 44/497 [00:01<00:17, 26.36it/s]
  9%|▉         | 47/497 [00:01<00:16, 27.09it/s]
 10%|█         | 50/497 [00:01<00:16, 27.59it/s]
 11%|█         | 53/497 [00:01<00:15, 28.08it/s]
 11%|█▏        | 56/497 [00:01<00:15, 27.74it/s]
 12%|█▏        | 60/497 [00:02<00:16, 27.06it/s]
 13%|█▎        | 63/497 [00:02<00:17, 25.17it/s]
 13%|█▎        | 66/497 [00:02<00:16, 25.74it/s]
 14%|█▍        | 69/497 [00:02<00:16, 26.60it/s]
 14%|█▍        | 72/497 [00:02<00:15, 27.44it/s]
 15%|█▌        | 76/497 [00:02<00:14, 28.62it/s]
 16%|█▌        | 79/497 [00:02<00:14, 28.78it/s]
 16%|█▋        | 82/497 [00:02<00:15, 26.93it/s]
 17%|█▋        | 85/497 [00:03<00:15, 27.35it/s]
 18%|█▊        | 89/497 [00:03<00:14, 28.46it/s]
 19%|█▊        | 93/497 [00:03<00:13, 29.63it/s]
 19%|█▉        | 96/497 [00:03<00:13, 29.53it/s]
 20%|██        | 100/497 [00:03<00:13, 29.68it/s]
 21%|██        | 103/497 [00:03<00:13, 29.67it/s]
 21%|██▏       | 106/497 [00:03<00:13, 28.23it/s]
 22%|██▏       | 109/497 [00:03<00:13, 28.45it/s]
 23%|██▎       | 112/497 [00:03<00:13, 27.97it/s]
 23%|██▎       | 115/497 [00:04<00:13, 27.91it/s]
 24%|██▎       | 118/497 [00:04<00:13, 28.21it/s]
 25%|██▍       | 122/497 [00:04<00:12, 29.31it/s]
 25%|██▌       | 126/497 [00:04<00:12, 29.50it/s]
 26%|██▌       | 130/497 [00:04<00:12, 29.86it/s]
 27%|██▋       | 133/497 [00:04<00:13, 27.55it/s]
 27%|██▋       | 136/497 [00:04<00:13, 27.69it/s]
 28%|██▊       | 139/497 [00:04<00:13, 27.47it/s]
 29%|██▉       | 143/497 [00:05<00:12, 28.23it/s]
 29%|██▉       | 146/497 [00:05<00:12, 28.35it/s]
 30%|███       | 150/497 [00:05<00:11, 29.18it/s]
 31%|███       | 154/497 [00:05<00:12, 28.50it/s]
 32%|███▏      | 157/497 [00:05<00:12, 28.15it/s]
 32%|███▏      | 160/497 [00:05<00:11, 28.59it/s]
 33%|███▎      | 163/497 [00:05<00:11, 27.96it/s]
 34%|███▎      | 167/497 [00:05<00:11, 29.18it/s]
 34%|███▍      | 171/497 [00:06<00:11, 29.41it/s]
 35%|███▌      | 174/497 [00:06<00:11, 28.88it/s]
 36%|███▌      | 178/497 [00:06<00:10, 29.48it/s]
 36%|███▋      | 181/497 [00:06<00:10, 29.58it/s]
 37%|███▋      | 185/497 [00:06<00:10, 29.97it/s]
 38%|███▊      | 188/497 [00:06<00:10, 29.92it/s]
 38%|███▊      | 191/497 [00:06<00:10, 29.56it/s]
 39%|███▉      | 194/497 [00:06<00:10, 28.65it/s]
 40%|███▉      | 197/497 [00:07<00:15, 19.15it/s]
 40%|████      | 200/497 [00:07<00:26, 11.39it/s]
 41%|████      | 202/497 [00:08<00:36,  8.02it/s]
 41%|████      | 204/497 [00:08<00:42,  6.83it/s]
 41%|████▏     | 206/497 [00:09<00:48,  5.94it/s]
 42%|████▏     | 207/497 [00:09<00:50,  5.77it/s]
 42%|████▏     | 208/497 [00:09<00:53,  5.43it/s]
 42%|████▏     | 209/497 [00:09<00:55,  5.16it/s]
 42%|████▏     | 210/497 [00:09<00:56,  5.08it/s]
 42%|████▏     | 211/497 [00:10<00:55,  5.12it/s]
 43%|████▎     | 212/497 [00:10<00:57,  4.93it/s]
 43%|████▎     | 213/497 [00:10<00:59,  4.80it/s]
 43%|████▎     | 214/497 [00:10<00:57,  4.89it/s]
 43%|████▎     | 215/497 [00:10<00:54,  5.14it/s]
 43%|████▎     | 216/497 [00:11<00:53,  5.24it/s]
 44%|████▎     | 217/497 [00:11<00:52,  5.35it/s]
 44%|████▍     | 218/497 [00:11<00:50,  5.53it/s]
 44%|████▍     | 219/497 [00:11<00:50,  5.50it/s]
 44%|████▍     | 220/497 [00:11<00:52,  5.27it/s]
 44%|████▍     | 221/497 [00:12<01:00,  4.57it/s]
 45%|████▍     | 222/497 [00:12<01:07,  4.09it/s]
 45%|████▍     | 223/497 [00:12<01:07,  4.08it/s]
 45%|████▌     | 224/497 [00:12<01:07,  4.06it/s]
 45%|████▌     | 225/497 [00:13<01:08,  3.96it/s]
 45%|████▌     | 226/497 [00:13<01:04,  4.21it/s]
 46%|████▌     | 227/497 [00:13<00:59,  4.52it/s]
 46%|████▌     | 228/497 [00:13<00:55,  4.83it/s]
 46%|████▌     | 229/497 [00:13<00:53,  4.98it/s]
 46%|████▋     | 230/497 [00:14<00:51,  5.20it/s]
 46%|████▋     | 231/497 [00:14<00:47,  5.65it/s]
 47%|████▋     | 232/497 [00:14<00:45,  5.79it/s]
 47%|████▋     | 233/497 [00:14<00:47,  5.60it/s]
 47%|████▋     | 234/497 [00:14<00:48,  5.45it/s]
 47%|████▋     | 235/497 [00:14<00:46,  5.67it/s]
 47%|████▋     | 236/497 [00:15<00:44,  5.92it/s]
 48%|████▊     | 237/497 [00:15<00:45,  5.72it/s]
 48%|████▊     | 238/497 [00:15<00:45,  5.65it/s]
 48%|████▊     | 239/497 [00:15<00:44,  5.86it/s]
 48%|████▊     | 240/497 [00:15<00:43,  5.90it/s]
 48%|████▊     | 241/497 [00:16<00:47,  5.40it/s]
 49%|████▊     | 242/497 [00:16<00:51,  5.00it/s]
 49%|████▉     | 243/497 [00:16<00:53,  4.77it/s]
 49%|████▉     | 244/497 [00:16<00:57,  4.42it/s]
 49%|████▉     | 245/497 [00:17<00:57,  4.35it/s]
 49%|████▉     | 246/497 [00:17<00:55,  4.55it/s]
 50%|████▉     | 247/497 [00:17<00:55,  4.51it/s]
 50%|████▉     | 248/497 [00:17<00:54,  4.58it/s]
 50%|█████     | 249/497 [00:17<00:49,  5.01it/s]
 50%|█████     | 250/497 [00:18<00:49,  4.97it/s]
 51%|█████     | 251/497 [00:18<00:51,  4.82it/s]
 51%|█████     | 252/497 [00:18<00:54,  4.51it/s]
 51%|█████     | 253/497 [00:18<00:49,  4.93it/s]
 51%|█████     | 254/497 [00:18<00:48,  5.01it/s]
 51%|█████▏    | 255/497 [00:19<00:47,  5.14it/s]
 52%|█████▏    | 256/497 [00:19<00:45,  5.25it/s]
 52%|█████▏    | 257/497 [00:19<00:45,  5.26it/s]
 52%|█████▏    | 258/497 [00:19<00:45,  5.20it/s]
 52%|█████▏    | 259/497 [00:19<00:47,  5.05it/s]
 52%|█████▏    | 260/497 [00:19<00:45,  5.23it/s]
 53%|█████▎    | 261/497 [00:20<00:44,  5.30it/s]
 53%|█████▎    | 262/497 [00:20<00:44,  5.24it/s]
 53%|█████▎    | 263/497 [00:20<00:44,  5.30it/s]
 53%|█████▎    | 264/497 [00:20<00:44,  5.21it/s]
 53%|█████▎    | 265/497 [00:20<00:43,  5.37it/s]
 54%|█████▎    | 266/497 [00:21<00:44,  5.16it/s]
 54%|█████▎    | 267/497 [00:21<00:45,  5.01it/s]
 54%|█████▍    | 268/497 [00:21<00:47,  4.81it/s]
 54%|█████▍    | 269/497 [00:21<00:43,  5.19it/s]
 54%|█████▍    | 270/497 [00:21<00:42,  5.32it/s]
 55%|█████▍    | 271/497 [00:22<00:40,  5.56it/s]
 55%|█████▍    | 272/497 [00:22<00:40,  5.60it/s]
 55%|█████▍    | 273/497 [00:22<00:39,  5.64it/s]
 55%|█████▌    | 274/497 [00:22<00:42,  5.30it/s]
 55%|█████▌    | 275/497 [00:22<00:42,  5.18it/s]
 56%|█████▌    | 276/497 [00:23<00:41,  5.31it/s]
 56%|█████▌    | 277/497 [00:23<00:43,  5.11it/s]
 56%|█████▌    | 278/497 [00:23<00:41,  5.21it/s]
 56%|█████▌    | 279/497 [00:23<00:40,  5.33it/s]
 56%|█████▋    | 280/497 [00:23<00:40,  5.42it/s]
 57%|█████▋    | 281/497 [00:23<00:41,  5.23it/s]
 57%|█████▋    | 282/497 [00:24<00:43,  4.97it/s]
 57%|█████▋    | 283/497 [00:24<00:43,  4.93it/s]
 57%|█████▋    | 284/497 [00:24<00:43,  4.87it/s]
 57%|█████▋    | 285/497 [00:24<00:46,  4.54it/s]
 58%|█████▊    | 286/497 [00:25<00:44,  4.74it/s]
 58%|█████▊    | 287/497 [00:25<00:42,  4.96it/s]
 58%|█████▊    | 288/497 [00:25<00:43,  4.77it/s]
 58%|█████▊    | 289/497 [00:25<00:44,  4.70it/s]
 58%|█████▊    | 290/497 [00:25<00:42,  4.87it/s]
 59%|█████▊    | 291/497 [00:26<00:41,  4.94it/s]
 59%|█████▉    | 292/497 [00:26<00:39,  5.13it/s]
 59%|█████▉    | 293/497 [00:26<00:39,  5.21it/s]
 59%|█████▉    | 294/497 [00:26<00:40,  5.00it/s]
 59%|█████▉    | 295/497 [00:26<00:39,  5.13it/s]
 60%|█████▉    | 296/497 [00:27<00:40,  5.01it/s]
 60%|█████▉    | 297/497 [00:27<00:40,  4.92it/s]
 60%|█████▉    | 298/497 [00:27<00:42,  4.64it/s]
 60%|██████    | 299/497 [00:27<00:44,  4.50it/s]
 60%|██████    | 300/497 [00:27<00:41,  4.76it/s]
 61%|██████    | 301/497 [00:28<00:39,  4.99it/s]
 61%|██████    | 302/497 [00:28<00:41,  4.73it/s]
 61%|██████    | 303/497 [00:28<00:39,  4.88it/s]
 61%|██████    | 304/497 [00:28<00:37,  5.20it/s]
 61%|██████▏   | 305/497 [00:28<00:34,  5.51it/s]
 62%|██████▏   | 306/497 [00:29<00:34,  5.55it/s]
 62%|██████▏   | 307/497 [00:29<00:35,  5.29it/s]
 62%|██████▏   | 308/497 [00:29<00:35,  5.32it/s]
 62%|██████▏   | 309/497 [00:29<00:35,  5.28it/s]
 62%|██████▏   | 310/497 [00:29<00:35,  5.31it/s]
 63%|██████▎   | 311/497 [00:29<00:33,  5.48it/s]
 63%|██████▎   | 312/497 [00:30<00:33,  5.59it/s]
 63%|██████▎   | 313/497 [00:30<00:33,  5.49it/s]
 63%|██████▎   | 314/497 [00:30<00:35,  5.16it/s]
 63%|██████▎   | 315/497 [00:30<00:34,  5.22it/s]
 64%|██████▎   | 316/497 [00:30<00:35,  5.03it/s]
 64%|██████▍   | 317/497 [00:31<00:36,  4.93it/s]
 64%|██████▍   | 318/497 [00:31<00:36,  4.89it/s]
 64%|██████▍   | 319/497 [00:31<00:35,  5.08it/s]
 64%|██████▍   | 320/497 [00:31<00:34,  5.08it/s]
 65%|██████▍   | 321/497 [00:31<00:34,  5.10it/s]
 65%|██████▍   | 322/497 [00:32<00:32,  5.34it/s]
 65%|██████▍   | 323/497 [00:32<00:35,  4.84it/s]
 65%|██████▌   | 324/497 [00:32<00:37,  4.64it/s]
 65%|██████▌   | 325/497 [00:32<00:36,  4.68it/s]
 66%|██████▌   | 326/497 [00:32<00:34,  5.02it/s]
 66%|██████▌   | 327/497 [00:33<00:32,  5.18it/s]
 66%|██████▌   | 328/497 [00:33<00:32,  5.27it/s]
 66%|██████▌   | 329/497 [00:33<00:31,  5.39it/s]
 66%|██████▋   | 330/497 [00:33<00:31,  5.36it/s]
 67%|██████▋   | 331/497 [00:33<00:31,  5.24it/s]
 67%|██████▋   | 332/497 [00:34<00:31,  5.31it/s]
 67%|██████▋   | 333/497 [00:34<00:29,  5.53it/s]
 67%|██████▋   | 334/497 [00:34<00:32,  5.09it/s]
 67%|██████▋   | 335/497 [00:34<00:32,  4.91it/s]
 68%|██████▊   | 336/497 [00:34<00:31,  5.05it/s]
 68%|██████▊   | 337/497 [00:35<00:31,  5.12it/s]
 68%|██████▊   | 338/497 [00:35<00:30,  5.14it/s]
 68%|██████▊   | 339/497 [00:35<00:30,  5.19it/s]
 68%|██████▊   | 340/497 [00:35<00:30,  5.16it/s]
 69%|██████▊   | 341/497 [00:35<00:31,  5.00it/s]
 69%|██████▉   | 342/497 [00:36<00:30,  5.10it/s]
 69%|██████▉   | 343/497 [00:36<00:29,  5.29it/s]
 69%|██████▉   | 344/497 [00:36<00:28,  5.28it/s]
 69%|██████▉   | 345/497 [00:36<00:29,  5.24it/s]
 70%|██████▉   | 346/497 [00:36<00:28,  5.26it/s]
 70%|██████▉   | 347/497 [00:36<00:26,  5.57it/s]
 70%|███████   | 348/497 [00:37<00:26,  5.62it/s]
 70%|███████   | 349/497 [00:37<00:26,  5.53it/s]
 70%|███████   | 350/497 [00:37<00:25,  5.83it/s]
 71%|███████   | 351/497 [00:37<00:24,  6.08it/s]
 71%|███████   | 352/497 [00:37<00:25,  5.77it/s]
 71%|███████   | 353/497 [00:37<00:24,  5.81it/s]
 71%|███████   | 354/497 [00:38<00:25,  5.60it/s]
 71%|███████▏  | 355/497 [00:38<00:24,  5.87it/s]
 72%|███████▏  | 356/497 [00:38<00:24,  5.86it/s]
 72%|███████▏  | 357/497 [00:38<00:23,  5.88it/s]
 72%|███████▏  | 358/497 [00:38<00:25,  5.44it/s]
 72%|██████��▏  | 359/497 [00:39<00:25,  5.37it/s]
 72%|███████▏  | 360/497 [00:39<00:26,  5.12it/s]
 73%|███████▎  | 361/497 [00:39<00:26,  5.15it/s]
 73%|███████▎  | 362/497 [00:39<00:25,  5.20it/s]
 73%|███████▎  | 363/497 [00:39<00:30,  4.43it/s]
 73%|███████▎  | 364/497 [00:40<00:32,  4.14it/s]
 73%|███████▎  | 365/497 [00:40<00:30,  4.35it/s]
 74%|███████▎  | 366/497 [00:40<00:28,  4.65it/s]
 74%|███████▍  | 367/497 [00:40<00:27,  4.76it/s]
 74%|███████▍  | 368/497 [00:40<00:25,  4.97it/s]
 74%|███████▍  | 369/497 [00:41<00:25,  5.00it/s]
 74%|███████▍  | 370/497 [00:41<00:24,  5.12it/s]
 75%|███████▍  | 371/497 [00:41<00:24,  5.10it/s]
 75%|███████▍  | 372/497 [00:41<00:23,  5.32it/s]
 75%|███████▌  | 373/497 [00:41<00:24,  5.10it/s]
 75%|███████▌  | 374/497 [00:42<00:24,  5.12it/s]
 75%|███████▌  | 375/497 [00:42<00:22,  5.33it/s]
 76%|███████▌  | 376/497 [00:42<00:22,  5.28it/s]
 76%|███████▌  | 377/497 [00:42<00:22,  5.38it/s]
 76%|███████▌  | 378/497 [00:42<00:20,  5.85it/s]
 76%|███████▋  | 379/497 [00:43<00:20,  5.68it/s]
 76%|███████▋  | 380/497 [00:43<00:21,  5.56it/s]
 77%|███████▋  | 381/497 [00:43<00:20,  5.55it/s]
 77%|███████▋  | 382/497 [00:43<00:20,  5.72it/s]
 77%|███████▋  | 383/497 [00:43<00:19,  5.95it/s]
 77%|███████▋  | 384/497 [00:43<00:18,  5.99it/s]
 77%|███████▋  | 385/497 [00:44<00:19,  5.68it/s]
 78%|███████▊  | 386/497 [00:44<00:19,  5.68it/s]
 78%|███████▊  | 387/497 [00:44<00:19,  5.66it/s]
 78%|███████▊  | 388/497 [00:44<00:20,  5.32it/s]
 78%|███████▊  | 389/497 [00:44<00:20,  5.18it/s]
 78%|███████▊  | 390/497 [00:45<00:20,  5.26it/s]
 79%|███████▊  | 391/497 [00:45<00:20,  5.20it/s]
 79%|███████▉  | 392/497 [00:45<00:19,  5.27it/s]
 79%|███████▉  | 393/497 [00:45<00:18,  5.62it/s]
 79%|███████▉  | 394/497 [00:45<00:19,  5.18it/s]
 79%|███████▉  | 395/497 [00:45<00:18,  5.39it/s]
 80%|███████▉  | 396/497 [00:46<00:18,  5.38it/s]
 80%|███████▉  | 397/497 [00:46<00:18,  5.36it/s]
 80%|████████  | 398/497 [00:46<00:18,  5.31it/s]
 80%|████████  | 399/497 [00:46<00:18,  5.31it/s]
 80%|████████  | 400/497 [00:46<00:17,  5.53it/s]
 81%|████████  | 401/497 [00:47<00:17,  5.53it/s]
 81%|████████  | 402/497 [00:47<00:17,  5.53it/s]
 81%|████████  | 403/497 [00:47<00:17,  5.45it/s]
 81%|████████▏ | 404/497 [00:47<00:16,  5.48it/s]
 81%|████████▏ | 405/497 [00:47<00:18,  5.09it/s]
 82%|████████▏ | 406/497 [00:48<00:18,  5.05it/s]
 82%|████████▏ | 407/497 [00:48<00:16,  5.31it/s]
 82%|████████▏ | 408/497 [00:48<00:17,  5.04it/s]
 82%|████████▏ | 409/497 [00:48<00:18,  4.64it/s]
 82%|████████▏ | 410/497 [00:48<00:19,  4.56it/s]
 83%|████████▎ | 411/497 [00:49<00:19,  4.48it/s]
 83%|████████▎ | 412/497 [00:49<00:18,  4.58it/s]
 83%|████████▎ | 413/497 [00:49<00:18,  4.60it/s]
 83%|████████▎ | 414/497 [00:49<00:17,  4.77it/s]
 84%|████████▎ | 415/497 [00:49<00:16,  4.90it/s]
 84%|████████▎ | 416/497 [00:50<00:16,  4.78it/s]
 84%|████████▍ | 417/497 [00:50<00:17,  4.70it/s]
 84%|████████▍ | 418/497 [00:50<00:16,  4.70it/s]
 84%|████████▍ | 419/497 [00:50<00:15,  4.94it/s]
 85%|████████▍ | 420/497 [00:50<00:15,  4.89it/s]
 85%|████████▍ | 421/497 [00:51<00:15,  4.88it/s]
 85%|████████▍ | 422/497 [00:51<00:16,  4.62it/s]
 85%|████████▌ | 423/497 [00:51<00:15,  4.84it/s]
 85%|████████▌ | 424/497 [00:51<00:14,  5.13it/s]
 86%|████████▌ | 425/497 [00:51<00:13,  5.51it/s]
 86%|████████▌ | 426/497 [00:52<00:14,  5.04it/s]
 86%|████████▌ | 427/497 [00:52<00:14,  4.82it/s]
 86%|████████▌ | 428/497 [00:52<00:14,  4.71it/s]
 86%|████████▋ | 429/497 [00:52<00:14,  4.66it/s]
 87%|████████▋ | 430/497 [00:53<00:15,  4.36it/s]
 87%|████████▋ | 431/497 [00:53<00:15,  4.21it/s]
 87%|████████▋ | 432/497 [00:53<00:15,  4.19it/s]
 87%|████████▋ | 433/497 [00:53<00:15,  4.24it/s]
 87%|████████▋ | 434/497 [00:54<00:14,  4.31it/s]
 88%|████████▊ | 435/497 [00:54<00:14,  4.41it/s]
 88%|████████▊ | 436/497 [00:54<00:13,  4.40it/s]
 88%|████████▊ | 437/497 [00:54<00:13,  4.57it/s]
 88%|████████▊ | 438/497 [00:54<00:13,  4.50it/s]
 88%|████████▊ | 439/497 [00:55<00:13,  4.30it/s]
 89%|████████▊ | 440/497 [00:55<00:14,  3.93it/s]
 89%|████████▊ | 441/497 [00:55<00:13,  4.16it/s]
 89%|████████▉ | 442/497 [00:55<00:13,  4.10it/s]
 89%|████████▉ | 443/497 [00:56<00:13,  4.10it/s]
 89%|████████▉ | 444/497 [00:56<00:12,  4.26it/s]
 90%|████████▉ | 445/497 [00:56<00:11,  4.38it/s]
 90%|████████▉ | 446/497 [00:56<00:11,  4.61it/s]
 90%|████████▉ | 447/497 [00:56<00:10,  4.91it/s]
 90%|█████████ | 448/497 [00:57<00:10,  4.57it/s]
 90%|█████████ | 449/497 [00:57<00:10,  4.64it/s]
 91%|█████████ | 450/497 [00:57<00:10,  4.63it/s]
 91%|█████████ | 451/497 [00:57<00:09,  4.82it/s]
 91%|█████████ | 452/497 [00:58<00:09,  4.63it/s]
 91%|█████████ | 453/497 [00:58<00:09,  4.62it/s]
 91%|█████████▏| 454/497 [00:58<00:09,  4.63it/s]
 92%|█████████▏| 455/497 [00:58<00:09,  4.60it/s]
 92%|█████████▏| 456/497 [00:58<00:08,  4.58it/s]
 92%|█████████▏| 457/497 [00:59<00:08,  4.68it/s]
 92%|█████████▏| 458/497 [00:59<00:08,  4.57it/s]
 92%|█████████▏| 459/497 [00:59<00:08,  4.49it/s]
 93%|█████████▎| 460/497 [00:59<00:07,  4.81it/s]
 93%|█████████▎| 461/497 [01:00<00:07,  4.81it/s]
 93%|█████████▎| 462/497 [01:00<00:07,  4.75it/s]
 93%|█████████▎| 463/497 [01:00<00:06,  5.02it/s]
 93%|█████████▎| 464/497 [01:00<00:06,  5.12it/s]
 94%|█████████▎| 465/497 [01:00<00:06,  5.21it/s]
 94%|█████████▍| 466/497 [01:00<00:05,  5.28it/s]
 94%|█████████▍| 467/497 [01:01<00:05,  5.24it/s]
 94%|█████████▍| 468/497 [01:01<00:05,  5.24it/s]
 94%|█████████▍| 469/497 [01:01<00:05,  5.31it/s]
 95%|█████████▍| 470/497 [01:01<00:04,  5.40it/s]
 95%|█████████▍| 471/497 [01:01<00:04,  5.55it/s]
 95%|█████████▍| 472/497 [01:02<00:04,  5.62it/s]
 95%|█████████▌| 473/497 [01:02<00:04,  5.55it/s]
 95%|█████████▌| 474/497 [01:02<00:04,  5.61it/s]
 96%|█████████▌| 475/497 [01:02<00:03,  5.59it/s]
 96%|█████████▌| 476/497 [01:02<00:03,  5.28it/s]
 96%|█████████▌| 477/497 [01:03<00:04,  4.96it/s]
 96%|█████████▌| 478/497 [01:03<00:03,  5.12it/s]
 96%|█████████▋| 479/497 [01:03<00:03,  5.34it/s]
 97%|█████████▋| 480/497 [01:03<00:03,  5.31it/s]
 97%|█████████▋| 481/497 [01:03<00:03,  5.32it/s]
 97%|█████████▋| 482/497 [01:03<00:02,  5.41it/s]
 97%|█████████▋| 483/497 [01:04<00:02,  4.91it/s]
 97%|█████████▋| 484/497 [01:04<00:02,  4.89it/s]
 98%|█████████▊| 485/497 [01:04<00:02,  4.84it/s]
 98%|█████████▊| 486/497 [01:04<00:02,  4.64it/s]
 98%|█████████▊| 487/497 [01:05<00:02,  4.31it/s]
 98%|█████████▊| 488/497 [01:05<00:01,  4.54it/s]
 98%|█████████▊| 489/497 [01:05<00:01,  4.60it/s]
 99%|█████████▊| 490/497 [01:05<00:01,  4.82it/s]
 99%|█████████▉| 491/497 [01:05<00:01,  5.17it/s]
 99%|█████████▉| 492/497 [01:06<00:00,  5.23it/s]
 99%|█████████▉| 493/497 [01:06<00:00,  5.20it/s]
 99%|█████████▉| 494/497 [01:06<00:00,  5.42it/s]
+FID: 106.39697359554248
+
+/mnt/petrelfs/leihaodong/anaconda3/envs/eval_xx/lib/python3.10/site-packages/torchmetrics/utilities/prints.py:43: UserWarning: Metric `InceptionScore` will save all extracted features in buffer. For large datasets this may lead to large memory footprint.
+  warnings.warn(*args, **kwargs)
+/mnt/petrelfs/leihaodong/anaconda3/envs/eval_xx/lib/python3.10/site-packages/open_clip/factory.py:450: UserWarning: QuickGELU mismatch between final model config (quick_gelu=False) and pretrained tag 'openai' (quick_gelu=True).
+  warnings.warn(
+Using a slow image processor as `use_fast` is unset and a slow processor was saved with this model. `use_fast=True` will be the default behavior in v4.52, even if the model was saved with a slow processor. This will result in minor differences in outputs. You'll still be able to use a slow processor with `use_fast=False`.
+DINO_I: 0.12376418709754944
+
+  warnings.warn(f"Importing from {__name__} is deprecated, please import via timm.layers", FutureWarning)
+/mnt/petrelfs/leihaodong/ICLR25/eval/HPSv2/hpsv2/img_score.py:107: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead.
+  with torch.cuda.amp.autocast():
+
+WARNING:huggingface_hub.utils._http:'(ReadTimeoutError("HTTPSConnectionPool(host='huggingface.co', port=443): Read timed out. (read timeout=10)"), '(Request ID: 49becba4-216e-4899-80ed-c5920f1749aa)')' thrown while requesting HEAD https://huggingface.co/xswu/HPSv2/resolve/main/HPS_v2.1_compressed.pt
+Retrying in 1s [Retry 1/5].
+WARNING:huggingface_hub.utils._http:Retrying in 1s [Retry 1/5].
+
+WARNING:huggingface_hub.utils._http:'(ReadTimeoutError("HTTPSConnectionPool(host='huggingface.co', port=443): Read timed out. (read timeout=10)"), '(Request ID: a6d58007-9463-45c9-83c7-9fda30ea5f9b)')' thrown while requesting HEAD https://huggingface.co/xswu/HPSv2/resolve/main/HPS_v2.1_compressed.pt
+Retrying in 1s [Retry 1/5].
+WARNING:huggingface_hub.utils._http:Retrying in 1s [Retry 1/5].
+
+WARNING:huggingface_hub.utils._http:'(ReadTimeoutError("HTTPSConnectionPool(host='huggingface.co', port=443): Read timed out. (read timeout=10)"), '(Request ID: 6a6c2332-d2e1-4d47-8be7-a76c2a556eb6)')' thrown while requesting HEAD https://huggingface.co/xswu/HPSv2/resolve/main/HPS_v2.1_compressed.pt
+Retrying in 1s [Retry 1/5].
+WARNING:huggingface_hub.utils._http:Retrying in 1s [Retry 1/5].
+
+INFO:absl:FID: 106.39697359554248
+INFO:absl:DINO_I: 0.12376418709754944
+INFO:absl:IS: (tensor(17.8868, device='cuda:0'), tensor(1.8582, device='cuda:0'))
+INFO:absl:CLIP_T: 0.1896880811728702
+INFO:absl:HPSv2: 0.291891976861167
+INFO:absl:Aesthetic: 6.570238386841127
+Scores:
+FID: 106.39697359554248
+DINO_I: 0.12376418709754944
+IS: (tensor(17.8868, device='cuda:0'), tensor(1.8582, device='cuda:0'))
+CLIP_T: 0.1896880811728702
+HPSv2: 0.291891976861167
+Aesthetic: 6.570238386841127

scripts/eval/slurm-6226598.out

@@ -0,0 +1,39 @@
+loading annotations into memory...
+Done (t=0.03s)
+creating index...
+index created!
+一共314 张
+loading annotations into memory...
+Done (t=0.03s)
+creating index...
+index created!
+/mnt/petrelfs/leihaodong/ICLR25/exp/Emu3/MSCOCO2017Val/ms_sjd/img/gt
+/mnt/petrelfs/leihaodong/ICLR25/exp/Emu3/MSCOCO2017Val/ms_sjd/img
+
  0%|          | 0/314 [00:00<?, ?it/s]
  0%|          | 1/314 [00:00<01:54,  2.74it/s]
  2%|▏         | 6/314 [00:00<00:19, 15.69it/s]
  3%|▎         | 10/314 [00:00<00:13, 21.87it/s]
  4%|▍         | 14/314 [00:00<00:15, 20.00it/s]
  5%|▌         | 17/314 [00:00<00:14, 20.88it/s]
  7%|▋         | 22/314 [00:01<00:10, 27.27it/s]
  9%|▉         | 28/314 [00:01<00:08, 34.62it/s]
 10%|█         | 32/314 [00:01<00:08, 31.85it/s]
 12%|█▏        | 37/314 [00:01<00:07, 35.73it/s]
 14%|█▎        | 43/314 [00:01<00:06, 40.36it/s]
 16%|█▌        | 49/314 [00:01<00:05, 44.50it/s]
 17%|█▋        | 54/314 [00:01<00:05, 44.99it/s]
 19%|█▉        | 59/314 [00:01<00:06, 36.77it/s]
 20%|██        | 64/314 [00:02<00:08, 29.42it/s]
 22%|██▏       | 69/314 [00:02<00:08, 30.34it/s]
 23%|██▎       | 73/314 [00:02<00:07, 32.03it/s]
 25%|██▍       | 77/314 [00:02<00:07, 31.14it/s]
 26%|██▌       | 81/314 [00:02<00:07, 30.10it/s]
 27%|██▋       | 85/314 [00:02<00:08, 27.02it/s]
 28%|██▊       | 88/314 [00:03<00:08, 27.65it/s]
 29%|██▉       | 91/314 [00:03<00:08, 27.19it/s]
 30%|██▉       | 94/314 [00:03<00:08, 24.79it/s]
 32%|███▏      | 99/314 [00:03<00:08, 26.55it/s]
 32%|███▏      | 102/314 [00:03<00:08, 24.32it/s]
 33%|███▎      | 105/314 [00:03<00:09, 22.66it/s]
 35%|███▍      | 109/314 [00:03<00:08, 25.46it/s]
 36%|███▌      | 113/314 [00:04<00:07, 26.66it/s]
 37%|███▋      | 116/314 [00:04<00:08, 23.13it/s]
 38%|███▊      | 119/314 [00:04<00:08, 21.83it/s]
 39%|███▉      | 122/314 [00:04<00:09, 21.10it/s]
 40%|███▉      | 125/314 [00:04<00:08, 22.94it/s]
 41%|████      | 128/314 [00:04<00:08, 21.73it/s]
 42%|████▏     | 131/314 [00:04<00:08, 22.59it/s]
 43%|████▎     | 134/314 [00:05<00:07, 22.64it/s]
 44%|████▎     | 137/314 [00:05<00:08, 21.99it/s]
 45%|████▌     | 142/314 [00:05<00:06, 27.80it/s]
 46%|████▌     | 145/314 [00:05<00:06, 25.20it/s]
 47%|████▋     | 149/314 [00:05<00:06, 26.40it/s]
 48%|████▊     | 152/314 [00:05<00:06, 26.35it/s]
 49%|████▉     | 155/314 [00:05<00:06, 24.52it/s]
 51%|█████     | 159/314 [00:05<00:05, 27.05it/s]
 52%|█████▏    | 162/314 [00:06<00:05, 25.76it/s]
 53%|█████▎    | 165/314 [00:06<00:06, 23.22it/s]
 54%|█████▎    | 168/314 [00:06<00:06, 22.25it/s]
 54%|█████▍    | 171/314 [00:06<00:06, 20.96it/s]
 55%|█████▌    | 174/314 [00:06<00:06, 22.82it/s]
 57%|█████▋    | 178/314 [00:06<00:05, 25.79it/s]
 58%|█████▊    | 181/314 [00:06<00:06, 20.68it/s]
 59%|█████▊    | 184/314 [00:07<00:06, 19.46it/s]
 60%|█████▉    | 188/314 [00:07<00:05, 22.82it/s]
 61%|██████    | 191/314 [00:07<00:05, 23.14it/s]
 62%|██████▏   | 195/314 [00:07<00:04, 26.42it/s]
 63%|██████▎   | 198/314 [00:07<00:04, 23.84it/s]
 64%|██████▍   | 201/314 [00:07<00:05, 21.02it/s]
 65%|██████▍   | 204/314 [00:08<00:05, 21.46it/s]
 66%|██████▌   | 207/314 [00:08<00:05, 20.78it/s]
 67%|██████▋   | 210/314 [00:08<00:05, 20.36it/s]
 68%|██████▊   | 215/314 [00:08<00:03, 26.67it/s]
 69%|██████▉   | 218/314 [00:08<00:03, 24.80it/s]
 70%|███████   | 221/314 [00:08<00:03, 23.72it/s]
 71%|███████▏  | 224/314 [00:08<00:03, 24.35it/s]
 72%|███████▏  | 227/314 [00:08<00:03, 24.09it/s]
 73%|███████▎  | 230/314 [00:09<00:03, 23.50it/s]
 74%|███████▍  | 233/314 [00:09<00:03, 24.25it/s]
 75%|███████▌  | 236/314 [00:09<00:03, 22.99it/s]
 76%|███████▌  | 239/314 [00:09<00:03, 21.49it/s]
 77%|███████▋  | 242/314 [00:09<00:03, 21.22it/s]
 78%|███████▊  | 245/314 [00:09<00:03, 21.93it/s]
 79%|███████▉  | 248/314 [00:09<00:02, 22.11it/s]
 80%|████████  | 252/314 [00:10<00:02, 24.16it/s]
 81%|████████  | 255/314 [00:10<00:02, 24.86it/s]
 82%|████████▏ | 258/314 [00:10<00:02, 23.82it/s]
 83%|████████▎ | 262/314 [00:10<00:02, 25.03it/s]
 84%|████████▍ | 265/314 [00:10<00:01, 25.41it/s]
 85%|████████▌ | 268/314 [00:10<00:01, 25.32it/s]
 86%|████████▋ | 271/314 [00:10<00:01, 23.93it/s]
 88%|████████▊ | 275/314 [00:10<00:01, 27.72it/s]
 89%|████████▊ | 278/314 [00:11<00:01, 23.66it/s]
 89%|████████▉ | 281/314 [00:11<00:01, 21.08it/s]
 90%|█████████ | 284/314 [00:11<00:01, 22.51it/s]
 91%|█████████▏| 287/314 [00:11<00:01, 22.68it/s]
 92%|█████████▏| 290/314 [00:11<00:01, 23.08it/s]
 93%|█████████▎| 293/314 [00:11<00:00, 22.01it/s]
 94%|█████████▍| 296/314 [00:11<00:00, 21.36it/s]
 95%|█████████▌| 299/314 [00:12<00:00, 18.64it/s]
 97%|█████████▋| 304/314 [00:12<00:00, 24.03it/s]
 98%|█████████▊| 307/314 [00:12<00:00, 23.04it/s]
 99%|█████████▊| 310/314 [00:12<00:00, 20.44it/s]
+
  0%|          | 0/314 [00:00<?, ?it/s]
  1%|▏         | 4/314 [00:00<00:10, 30.42it/s]
  3%|▎         | 8/314 [00:00<00:11, 27.61it/s]
  4%|▎         | 11/314 [00:00<00:10, 28.28it/s]
  4%|▍         | 14/314 [00:00<00:10, 27.89it/s]
  5%|▌         | 17/314 [00:00<00:10, 27.27it/s]
  6%|▋         | 20/314 [00:00<00:10, 27.40it/s]
  7%|▋         | 23/314 [00:00<00:10, 28.14it/s]
  8%|▊         | 26/314 [00:00<00:10, 27.41it/s]
  9%|▉         | 29/314 [00:01<00:10, 27.38it/s]
 10%|█         | 32/314 [00:01<00:10, 26.20it/s]
 11%|█         | 35/314 [00:01<00:11, 24.74it/s]
 12%|█▏        | 39/314 [00:01<00:10, 26.05it/s]
 13%|█▎        | 42/314 [00:01<00:10, 27.02it/s]
 14%|█▍        | 45/314 [00:01<00:10, 26.82it/s]
 15%|█▌        | 48/314 [00:01<00:10, 26.47it/s]
 16%|█▌        | 51/314 [00:01<00:10, 25.90it/s]
 17%|█▋        | 54/314 [00:02<00:10, 25.38it/s]
 18%|█▊        | 57/314 [00:02<00:11, 22.51it/s]
 19%|█▉        | 60/314 [00:02<00:11, 23.06it/s]
 20%|██        | 63/314 [00:02<00:10, 24.73it/s]
 21%|██        | 66/314 [00:02<00:09, 25.33it/s]
 22%|██▏       | 69/314 [00:02<00:10, 24.00it/s]
 23%|██▎       | 72/314 [00:02<00:09, 24.83it/s]
 24%|██▍       | 75/314 [00:02<00:09, 24.31it/s]
 25%|██▍       | 78/314 [00:03<00:10, 22.79it/s]
 26%|██▌       | 81/314 [00:03<00:10, 23.21it/s]
 27%|██▋       | 84/314 [00:03<00:09, 24.34it/s]
 28%|██▊       | 87/314 [00:03<00:10, 22.23it/s]
 29%|██▊       | 90/314 [00:03<00:09, 22.75it/s]
 30%|██▉       | 93/314 [00:03<00:09, 22.29it/s]
 31%|███       | 96/314 [00:03<00:09, 21.85it/s]
 32%|███▏      | 99/314 [00:03<00:09, 23.01it/s]
 32%|███▏      | 102/314 [00:04<00:09, 22.28it/s]
 33%|███▎      | 105/314 [00:04<00:09, 23.08it/s]
 34%|███▍      | 108/314 [00:04<00:08, 24.31it/s]
 35%|███▌      | 111/314 [00:04<00:08, 23.92it/s]
 36%|███▋      | 114/314 [00:04<00:07, 25.21it/s]
 37%|███▋      | 117/314 [00:04<00:07, 24.80it/s]
 38%|███▊      | 120/314 [00:04<00:07, 25.25it/s]
 39%|███▉      | 123/314 [00:04<00:07, 26.17it/s]
 40%|████      | 127/314 [00:05<00:06, 27.48it/s]
 42%|████▏     | 131/314 [00:05<00:06, 28.79it/s]
 43%|████▎     | 134/314 [00:05<00:06, 29.09it/s]
 44%|████▍     | 138/314 [00:05<00:05, 29.77it/s]
 45%|████▌     | 142/314 [00:05<00:05, 30.14it/s]
 46%|████▋     | 146/314 [00:05<00:05, 30.47it/s]
 48%|████▊     | 150/314 [00:05<00:05, 28.99it/s]
 49%|████▊     | 153/314 [00:05<00:05, 28.43it/s]
 50%|████▉     | 156/314 [00:06<00:05, 28.65it/s]
 51%|█████     | 159/314 [00:06<00:05, 26.23it/s]
 52%|█████▏    | 163/314 [00:06<00:05, 27.72it/s]
 53%|█████▎    | 166/314 [00:06<00:05, 27.92it/s]
 54%|█████▍    | 170/314 [00:06<00:05, 28.71it/s]
 55%|█████▌    | 174/314 [00:06<00:04, 29.37it/s]
 56%|█████▋    | 177/314 [00:06<00:04, 29.53it/s]
 57%|█████▋    | 180/314 [00:06<00:04, 27.73it/s]
 58%|█████▊    | 183/314 [00:07<00:05, 24.40it/s]
 59%|█████▉    | 186/314 [00:07<00:05, 25.58it/s]
 60%|██████    | 189/314 [00:07<00:04, 26.57it/s]
 61%|██████    | 192/314 [00:07<00:04, 26.96it/s]
 62%|██████▏   | 196/314 [00:07<00:04, 28.14it/s]
 63%|██████▎   | 199/314 [00:07<00:04, 28.11it/s]
 64%|██████▍   | 202/314 [00:07<00:03, 28.24it/s]
 66%|██████▌   | 206/314 [00:07<00:03, 29.08it/s]
 67%|██████▋   | 209/314 [00:07<00:03, 28.37it/s]
 68%|██████▊   | 213/314 [00:08<00:03, 29.12it/s]
 69%|██████▉   | 216/314 [00:08<00:03, 29.01it/s]
 70%|███████   | 220/314 [00:08<00:03, 28.53it/s]
 71%|███████   | 223/314 [00:08<00:03, 28.31it/s]
 72%|███████▏  | 226/314 [00:08<00:03, 28.69it/s]
 73%|███████▎  | 229/314 [00:08<00:02, 28.38it/s]
 74%|███████▍  | 232/314 [00:08<00:02, 27.82it/s]
 75%|███████▍  | 235/314 [00:08<00:02, 28.28it/s]
 76%|███████▌  | 239/314 [00:09<00:02, 29.14it/s]
 77%|███████▋  | 243/314 [00:09<00:02, 29.99it/s]
 78%|███████▊  | 246/314 [00:09<00:02, 29.72it/s]
 79%|███████▉  | 249/314 [00:09<00:02, 29.57it/s]
 80%|████████  | 252/314 [00:09<00:02, 27.25it/s]
 81%|████████  | 255/314 [00:09<00:02, 27.67it/s]
 82%|████████▏ | 258/314 [00:09<00:02, 27.08it/s]
 83%|████████▎ | 262/314 [00:09<00:01, 27.80it/s]
 85%|████████▍ | 266/314 [00:09<00:01, 28.50it/s]
 86%|████████▌ | 269/314 [00:10<00:01, 28.74it/s]
 87%|████████▋ | 272/314 [00:10<00:01, 28.30it/s]
 88%|████████▊ | 275/314 [00:10<00:01, 28.39it/s]
 89%|████████▉ | 279/314 [00:10<00:01, 27.88it/s]
 90%|█████████ | 283/314 [00:10<00:01, 28.76it/s]
 91%|█████████▏| 287/314 [00:10<00:00, 29.16it/s]
 92%|█████████▏| 290/314 [00:10<00:00, 28.28it/s]
 94%|█████████▎| 294/314 [00:10<00:00, 29.00it/s]
 95%|█████████▍| 298/314 [00:11<00:00, 29.45it/s]
 96%|█████████▌| 301/314 [00:11<00:00, 29.29it/s]
 97%|█████████▋| 304/314 [00:11<00:00, 29.38it/s]
 98%|█████████▊| 307/314 [00:11<00:00, 28.79it/s]
 99%|█████████▊| 310/314 [00:11<00:00, 28.91it/s]
+FID: 129.46357970318832
+
+/mnt/petrelfs/leihaodong/anaconda3/envs/eval_xx/lib/python3.10/site-packages/torchmetrics/utilities/prints.py:43: UserWarning: Metric `InceptionScore` will save all extracted features in buffer. For large datasets this may lead to large memory footprint.
+  warnings.warn(*args, **kwargs)
+/mnt/petrelfs/leihaodong/anaconda3/envs/eval_xx/lib/python3.10/site-packages/open_clip/factory.py:450: UserWarning: QuickGELU mismatch between final model config (quick_gelu=False) and pretrained tag 'openai' (quick_gelu=True).
+  warnings.warn(
+Using a slow image processor as `use_fast` is unset and a slow processor was saved with this model. `use_fast=True` will be the default behavior in v4.52, even if the model was saved with a slow processor. This will result in minor differences in outputs. You'll still be able to use a slow processor with `use_fast=False`.
+DINO_I: 0.13784950971603394
+
+  warnings.warn(f"Importing from {__name__} is deprecated, please import via timm.layers", FutureWarning)
+/mnt/petrelfs/leihaodong/ICLR25/eval/HPSv2/hpsv2/img_score.py:107: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead.
+  with torch.cuda.amp.autocast():
+
+INFO:absl:FID: 129.46357970318832
+INFO:absl:DINO_I: 0.13784950971603394
+INFO:absl:IS: (tensor(13.6355, device='cuda:0'), tensor(0.8688, device='cuda:0'))
+INFO:absl:CLIP_T: 0.1891890382690794
+INFO:absl:HPSv2: 0.29075515649880573
+INFO:absl:Aesthetic: 6.486864429370613
+Scores:
+FID: 129.46357970318832
+DINO_I: 0.13784950971603394
+IS: (tensor(13.6355, device='cuda:0'), tensor(0.8688, device='cuda:0'))
+CLIP_T: 0.1891890382690794
+HPSv2: 0.29075515649880573
+Aesthetic: 6.486864429370613

scripts/eval/slurm-6226599.out

@@ -0,0 +1,39 @@
+loading annotations into memory...
+Done (t=0.03s)
+creating index...
+index created!
+一共322 张
+loading annotations into memory...
+Done (t=0.03s)
+creating index...
+index created!
+/mnt/petrelfs/leihaodong/ICLR25/exp/Emu3/MSCOCO2017Val/mscr_gsd/img/gt
+/mnt/petrelfs/leihaodong/ICLR25/exp/Emu3/MSCOCO2017Val/mscr_gsd/img
+
  0%|          | 0/322 [00:00<?, ?it/s]
  0%|          | 1/322 [00:00<02:40,  2.01it/s]
  2%|▏         | 6/322 [00:00<00:25, 12.23it/s]
  3%|▎         | 9/322 [00:00<00:19, 15.94it/s]
  4%|▎         | 12/322 [00:00<00:20, 15.14it/s]
  5%|▍         | 15/322 [00:01<00:18, 16.88it/s]
  6%|▌         | 19/322 [00:01<00:14, 21.13it/s]
  8%|▊         | 25/322 [00:01<00:10, 29.47it/s]
  9%|▉         | 29/322 [00:01<00:10, 28.79it/s]
 11%|█         | 35/322 [00:01<00:08, 35.37it/s]
 12%|█▏        | 40/322 [00:01<00:07, 38.89it/s]
 14%|█▍        | 46/322 [00:01<00:06, 43.42it/s]
 16%|█▌        | 51/322 [00:01<00:06, 44.70it/s]
 17%|█▋        | 56/322 [00:02<00:07, 36.20it/s]
 19%|█▉        | 61/322 [00:02<00:08, 30.99it/s]
 20%|██        | 65/322 [00:02<00:08, 29.94it/s]
 21%|██▏       | 69/322 [00:02<00:08, 30.24it/s]
 23%|██▎       | 73/322 [00:02<00:08, 29.46it/s]
 24%|██▍       | 77/322 [00:02<00:08, 28.00it/s]
 25%|██▍       | 80/322 [00:02<00:08, 27.29it/s]
 26%|██▌       | 83/322 [00:03<00:09, 25.27it/s]
 27%|██▋       | 86/322 [00:03<00:09, 26.06it/s]
 28%|██▊       | 89/322 [00:03<00:09, 24.62it/s]
 29%|██▉       | 93/322 [00:03<00:08, 28.03it/s]
 30%|██▉       | 96/322 [00:03<00:08, 28.03it/s]
 31%|███       | 99/322 [00:03<00:08, 26.05it/s]
 32%|███▏      | 102/322 [00:03<00:08, 25.07it/s]
 33%|███▎      | 106/322 [00:03<00:07, 27.37it/s]
 34%|███▍      | 110/322 [00:04<00:07, 26.58it/s]
 35%|███▌      | 113/322 [00:04<00:08, 25.52it/s]
 36%|███▌      | 116/322 [00:04<00:08, 23.05it/s]
 37%|███▋      | 120/322 [00:04<00:08, 23.83it/s]
 38%|███▊      | 123/322 [00:04<00:09, 21.71it/s]
 39%|███▉      | 126/322 [00:04<00:09, 21.24it/s]
 40%|████      | 129/322 [00:05<00:08, 22.75it/s]
 41%|████      | 132/322 [00:05<00:08, 22.23it/s]
 42%|████▏     | 135/322 [00:05<00:08, 21.97it/s]
 43%|████▎     | 139/322 [00:05<00:07, 25.92it/s]
 44%|████▍     | 142/322 [00:05<00:07, 24.75it/s]
 46%|████▌     | 147/322 [00:05<00:06, 27.35it/s]
 47%|████▋     | 151/322 [00:05<00:06, 28.45it/s]
 48%|████▊     | 154/322 [00:05<00:06, 27.40it/s]
 49%|████▉     | 158/322 [00:06<00:05, 30.45it/s]
 50%|█████     | 162/322 [00:06<00:06, 26.26it/s]
 51%|█████     | 165/322 [00:06<00:06, 23.01it/s]
 52%|█████▏    | 168/322 [00:06<00:06, 23.69it/s]
 53%|█████▎    | 171/322 [00:06<00:06, 23.97it/s]
 54%|█████▍    | 174/322 [00:06<00:05, 24.99it/s]
 55%|█████▍    | 177/322 [00:06<00:05, 26.09it/s]
 56%|█████▌    | 180/322 [00:06<00:05, 26.34it/s]
 57%|█████▋    | 183/322 [00:07<00:05, 25.27it/s]
 58%|█████▊    | 186/322 [00:07<00:05, 22.67it/s]
 59%|█████▊    | 189/322 [00:07<00:05, 22.74it/s]
 60%|█████▉    | 192/322 [00:07<00:05, 23.98it/s]
 61%|██████    | 196/322 [00:07<00:04, 27.78it/s]
 62%|██████▏   | 199/322 [00:07<00:04, 27.70it/s]
 63%|██████▎   | 202/322 [00:07<00:04, 28.19it/s]
 64%|██████▎   | 205/322 [00:08<00:04, 23.73it/s]
 65%|██████▍   | 208/322 [00:08<00:04, 22.83it/s]
 66%|██████▌   | 211/322 [00:08<00:05, 21.93it/s]
 66%|██████▋   | 214/322 [00:08<00:04, 21.73it/s]
 68%|██████▊   | 219/322 [00:08<00:04, 24.53it/s]
 69%|██████▉   | 222/322 [00:08<00:04, 22.22it/s]
 70%|███████   | 226/322 [00:08<00:04, 23.41it/s]
 71%|███████   | 229/322 [00:09<00:04, 23.25it/s]
 72%|███████▏  | 232/322 [00:09<00:03, 23.49it/s]
 73%|███████▎  | 235/322 [00:09<00:03, 24.56it/s]
 74%|███████▍  | 239/322 [00:09<00:03, 25.18it/s]
 75%|███████▌  | 242/322 [00:09<00:03, 25.80it/s]
 76%|███████▌  | 245/322 [00:09<00:03, 22.93it/s]
 77%|███████▋  | 248/322 [00:09<00:03, 21.08it/s]
 78%|███████▊  | 252/322 [00:10<00:02, 23.73it/s]
 80%|███████▉  | 256/322 [00:10<00:02, 25.90it/s]
 80%|████████  | 259/322 [00:10<00:02, 26.67it/s]
 81%|████████▏ | 262/322 [00:10<00:02, 24.46it/s]
 82%|████████▏ | 265/322 [00:10<00:02, 25.43it/s]
 84%|████████▎ | 269/322 [00:10<00:02, 25.12it/s]
 85%|████████▍ | 273/322 [00:10<00:01, 25.94it/s]
 86%|████████▌ | 276/322 [00:10<00:01, 26.12it/s]
 87%|████████▋ | 280/322 [00:11<00:01, 27.76it/s]
 88%|███████���▊ | 284/322 [00:11<00:01, 27.10it/s]
 89%|████████▉ | 287/322 [00:11<00:01, 26.64it/s]
 90%|█████████ | 290/322 [00:11<00:01, 25.71it/s]
 92%|█████████▏| 295/322 [00:11<00:00, 28.17it/s]
 93%|█████████▎| 298/322 [00:11<00:00, 25.51it/s]
 94%|█████████▍| 303/322 [00:11<00:00, 26.73it/s]
 95%|█████████▌| 306/322 [00:12<00:00, 26.32it/s]
 96%|█████████▌| 309/322 [00:12<00:00, 24.88it/s]
 97%|█████████▋| 313/322 [00:12<00:00, 26.99it/s]
 99%|█████████▉| 318/322 [00:12<00:00, 27.70it/s]
+
  0%|          | 0/322 [00:00<?, ?it/s]
  1%|          | 2/322 [00:00<00:17, 17.90it/s]
  2%|▏         | 5/322 [00:00<00:19, 16.39it/s]
  2%|▏         | 8/322 [00:00<00:16, 19.61it/s]
  3%|▎         | 11/322 [00:00<00:14, 21.28it/s]
  5%|▍         | 15/322 [00:00<00:12, 24.28it/s]
  6%|▌         | 18/322 [00:00<00:12, 24.93it/s]
  7%|▋         | 21/322 [00:00<00:11, 25.80it/s]
  7%|▋         | 24/322 [00:01<00:11, 26.36it/s]
  8%|▊         | 27/322 [00:01<00:10, 27.11it/s]
  9%|▉         | 30/322 [00:01<00:10, 26.64it/s]
 10%|█         | 33/322 [00:01<00:10, 26.62it/s]
 11%|█         | 36/322 [00:01<00:11, 25.63it/s]
 12%|█▏        | 39/322 [00:01<00:11, 23.61it/s]
 13%|█▎        | 42/322 [00:01<00:11, 24.75it/s]
 14%|█▍        | 45/322 [00:01<00:10, 25.24it/s]
 15%|█▍        | 48/322 [00:01<00:10, 25.48it/s]
 16%|█▌        | 51/322 [00:02<00:10, 25.91it/s]
 17%|█▋        | 54/322 [00:02<00:10, 24.77it/s]
 18%|█▊        | 57/322 [00:02<00:10, 24.95it/s]
 19%|█▊        | 60/322 [00:02<00:11, 22.15it/s]
 20%|█▉        | 63/322 [00:02<00:10, 23.62it/s]
 20%|██        | 66/322 [00:02<00:10, 24.52it/s]
 21%|██▏       | 69/322 [00:02<00:11, 22.73it/s]
 22%|██▏       | 72/322 [00:02<00:10, 23.87it/s]
 23%|██▎       | 75/322 [00:03<00:10, 24.02it/s]
 24%|██▍       | 78/322 [00:03<00:10, 23.77it/s]
 25%|██▌       | 81/322 [00:03<00:10, 22.33it/s]
 26%|██▌       | 84/322 [00:03<00:10, 23.46it/s]
 27%|██▋       | 87/322 [00:03<00:09, 23.90it/s]
 28%|██▊       | 90/322 [00:03<00:11, 20.93it/s]
 29%|██▉       | 93/322 [00:03<00:11, 20.21it/s]
 30%|██▉       | 96/322 [00:04<00:11, 20.40it/s]
 31%|███       | 99/322 [00:04<00:10, 21.84it/s]
 32%|███▏      | 102/322 [00:04<00:09, 22.68it/s]
 33%|███▎      | 105/322 [00:04<00:10, 21.24it/s]
 34%|███▎      | 108/322 [00:04<00:09, 22.72it/s]
 34%|███▍      | 111/322 [00:04<00:08, 23.53it/s]
 35%|███▌      | 114/322 [00:04<00:08, 24.75it/s]
 36%|███▋      | 117/322 [00:04<00:08, 24.79it/s]
 37%|███▋      | 120/322 [00:05<00:08, 24.04it/s]
 38%|███▊      | 123/322 [00:05<00:07, 25.00it/s]
 39%|███▉      | 126/322 [00:05<00:07, 25.91it/s]
 40%|████      | 130/322 [00:05<00:07, 27.07it/s]
 41%|████▏     | 133/322 [00:05<00:06, 27.26it/s]
 42%|████▏     | 136/322 [00:05<00:06, 27.72it/s]
 43%|████▎     | 139/322 [00:05<00:06, 27.32it/s]
 44%|████▍     | 142/322 [00:05<00:06, 27.92it/s]
 45%|████▌     | 146/322 [00:05<00:06, 28.85it/s]
 46%|████▋     | 149/322 [00:06<00:06, 28.80it/s]
 47%|████▋     | 152/322 [00:06<00:06, 26.52it/s]
 48%|████▊     | 155/322 [00:06<00:06, 27.22it/s]
 49%|████▉     | 158/322 [00:06<00:06, 24.51it/s]
 50%|█████     | 161/322 [00:06<00:06, 25.85it/s]
 51%|█████     | 164/322 [00:06<00:05, 26.39it/s]
 52%|█████▏    | 168/322 [00:06<00:05, 27.88it/s]
 53%|█████▎    | 171/322 [00:06<00:05, 28.12it/s]
 54%|█████▍    | 174/322 [00:07<00:05, 28.38it/s]
 55%|█████▍    | 177/322 [00:07<00:05, 26.87it/s]
 56%|█████▌    | 180/322 [00:07<00:05, 26.56it/s]
 57%|█████▋    | 183/322 [00:07<00:05, 26.71it/s]
 58%|█████▊    | 186/322 [00:07<00:04, 27.53it/s]
 59%|█████▊    | 189/322 [00:07<00:04, 27.64it/s]
 60%|█████▉    | 192/322 [00:07<00:04, 27.64it/s]
 61%|██████    | 196/322 [00:07<00:04, 28.65it/s]
 62%|██████▏   | 199/322 [00:07<00:04, 28.25it/s]
 63%|██████▎   | 202/322 [00:08<00:04, 28.47it/s]
 64%|██████▍   | 206/322 [00:08<00:03, 29.20it/s]
 65%|██████▍   | 209/322 [00:08<00:04, 27.82it/s]
 66%|██████▌   | 212/322 [00:08<00:03, 27.98it/s]
 67%|██████▋   | 215/322 [00:08<00:03, 28.17it/s]
 68%|██████▊   | 218/322 [00:08<00:03, 28.46it/s]
 69%|██████▊   | 221/322 [00:08<00:03, 27.62it/s]
 70%|██████▉   | 224/322 [00:08<00:03, 28.12it/s]
 70%|███████   | 227/322 [00:08<00:03, 27.39it/s]
 71%|███████▏  | 230/322 [00:09<00:03, 27.58it/s]
 72%|███████▏  | 233/322 [00:09<00:03, 28.09it/s]
 74%|███████▎  | 237/322 [00:09<00:02, 28.84it/s]
 75%|███████▍  | 241/322 [00:09<00:02, 29.54it/s]
 76%|███████▌  | 244/322 [00:09<00:02, 29.53it/s]
 77%|███████▋  | 247/322 [00:09<00:02, 29.06it/s]
 78%|███████▊  | 250/322 [00:09<00:02, 27.58it/s]
 79%|███████▊  | 253/322 [00:09<00:02, 27.59it/s]
 80%|███████▉  | 256/322 [00:09<00:02, 25.93it/s]
 80%|████████  | 259/322 [00:10<00:02, 26.87it/s]
 81%|████████▏ | 262/322 [00:10<00:02, 27.34it/s]
 83%|████████▎ | 266/322 [00:10<00:01, 28.22it/s]
 84%|████████▎ | 269/322 [00:10<00:01, 28.47it/s]
 84%|████████▍ | 272/322 [00:10<00:01, 27.03it/s]
 86%|████████▌ | 276/322 [00:10<00:01, 28.35it/s]
 87%|████████▋ | 279/322 [00:10<00:01, 27.87it/s]
 88%|████████▊ | 282/322 [00:10<00:01, 28.43it/s]
 89%|████████▊ | 285/322 [00:11<00:01, 28.62it/s]
 89%|████████▉ | 288/322 [00:11<00:01, 28.10it/s]
 90%|█████████ | 291/322 [00:11<00:01, 28.00it/s]
 91%|█████████▏| 294/322 [00:11<00:00, 28.44it/s]
 92%|█████████▏| 297/322 [00:11<00:00, 27.55it/s]
 93%|█████████▎| 300/322 [00:11<00:00, 27.37it/s]
 94%|█████████▍| 304/322 [00:11<00:00, 28.21it/s]
 95%|█████████▌| 307/322 [00:11<00:00, 27.77it/s]
 96%|█████████▋| 310/322 [00:12<00:00, 21.79it/s]
 97%|█████████▋| 313/322 [00:12<00:00, 14.16it/s]
 98%|█████████▊| 315/322 [00:12<00:00,  8.74it/s]
 98%|█████████▊| 317/322 [00:13<00:00,  7.43it/s]
 99%|█████████▉| 319/322 [00:13<00:00,  5.96it/s]
 99%|█████████▉| 320/322 [00:14<00:00,  5.61it/s]
+FID: 153.26222254493456
+
+/mnt/petrelfs/leihaodong/anaconda3/envs/eval_xx/lib/python3.10/site-packages/torchmetrics/utilities/prints.py:43: UserWarning: Metric `InceptionScore` will save all extracted features in buffer. For large datasets this may lead to large memory footprint.
+  warnings.warn(*args, **kwargs)
+/mnt/petrelfs/leihaodong/anaconda3/envs/eval_xx/lib/python3.10/site-packages/open_clip/factory.py:450: UserWarning: QuickGELU mismatch between final model config (quick_gelu=False) and pretrained tag 'openai' (quick_gelu=True).
+  warnings.warn(
+Using a slow image processor as `use_fast` is unset and a slow processor was saved with this model. `use_fast=True` will be the default behavior in v4.52, even if the model was saved with a slow processor. This will result in minor differences in outputs. You'll still be able to use a slow processor with `use_fast=False`.
+DINO_I: 0.11627760529518127
+
+  warnings.warn(f"Importing from {__name__} is deprecated, please import via timm.layers", FutureWarning)
+/mnt/petrelfs/leihaodong/ICLR25/eval/HPSv2/hpsv2/img_score.py:107: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead.
+  with torch.cuda.amp.autocast():
+
+INFO:absl:FID: 153.26222254493456
+INFO:absl:DINO_I: 0.11627760529518127
+INFO:absl:IS: (tensor(11.0328, device='cuda:0'), tensor(1.5793, device='cuda:0'))
+INFO:absl:CLIP_T: 0.1856627132026305
+INFO:absl:HPSv2: 0.2726512458753882
+INFO:absl:Aesthetic: 6.4035010041657445
+Scores:
+FID: 153.26222254493456
+DINO_I: 0.11627760529518127
+IS: (tensor(11.0328, device='cuda:0'), tensor(1.5793, device='cuda:0'))
+CLIP_T: 0.1856627132026305
+HPSv2: 0.2726512458753882
+Aesthetic: 6.4035010041657445

scripts/eval/slurm-6226600.out

@@ -0,0 +1,43 @@
+loading annotations into memory...
+Done (t=0.03s)
+creating index...
+index created!
+一共305 张
+loading annotations into memory...
+Done (t=0.03s)
+creating index...
+index created!
+/mnt/petrelfs/leihaodong/ICLR25/exp/Emu3/MSCOCO2017Val/mscr_sjd_0.05/img/gt
+/mnt/petrelfs/leihaodong/ICLR25/exp/Emu3/MSCOCO2017Val/mscr_sjd_0.05/img
+
  0%|          | 0/305 [00:00<?, ?it/s]
  0%|          | 1/305 [00:00<01:56,  2.62it/s]
  2%|▏         | 5/305 [00:00<00:23, 12.69it/s]
  3%|▎         | 8/305 [00:00<00:17, 16.95it/s]
  4%|▎         | 11/305 [00:00<00:21, 13.97it/s]
  5%|▍         | 15/305 [00:00<00:15, 18.85it/s]
  6%|▌         | 19/305 [00:01<00:12, 23.29it/s]
  8%|▊         | 25/305 [00:01<00:08, 31.11it/s]
 10%|▉         | 29/305 [00:01<00:09, 28.85it/s]
 11%|█▏        | 35/305 [00:01<00:07, 35.58it/s]
 13%|█▎        | 40/305 [00:01<00:07, 37.82it/s]
 15%|█▌        | 46/305 [00:01<00:06, 40.76it/s]
 17%|█▋        | 52/305 [00:01<00:06, 37.86it/s]
 19%|█▊        | 57/305 [00:02<00:07, 31.81it/s]
 20%|██        | 61/305 [00:02<00:08, 28.66it/s]
 21%|██▏       | 65/305 [00:02<00:08, 29.01it/s]
 23%|██▎       | 70/305 [00:02<00:07, 31.70it/s]
 24%|██▍       | 74/305 [00:02<00:08, 28.75it/s]
 26%|██▌       | 78/305 [00:02<00:08, 26.84it/s]
 27%|██▋       | 81/305 [00:03<00:08, 26.16it/s]
 28%|██▊       | 84/305 [00:03<00:08, 25.98it/s]
 29%|██▊       | 87/305 [00:03<00:09, 24.17it/s]
 30%|███       | 92/305 [00:03<00:07, 29.53it/s]
 31%|███▏      | 96/305 [00:03<00:08, 25.70it/s]
 32%|███▏      | 99/305 [00:03<00:08, 24.40it/s]
 34%|███▍      | 104/305 [00:03<00:07, 28.41it/s]
 35%|███▌      | 108/305 [00:04<00:07, 28.12it/s]
 36%|███▋      | 111/305 [00:04<00:07, 25.84it/s]
 37%|███▋      | 114/305 [00:04<00:08, 22.54it/s]
 38%|███▊      | 117/305 [00:04<00:08, 22.76it/s]
 39%|███▉      | 120/305 [00:04<00:09, 19.60it/s]
 40%|████      | 123/305 [00:04<00:08, 21.73it/s]
 42%|████▏     | 127/305 [00:04<00:07, 25.14it/s]
 43%|████▎     | 130/305 [00:05<00:07, 23.90it/s]
 44%|████▎     | 133/305 [00:05<00:07, 22.13it/s]
 45%|████▍     | 137/305 [00:05<00:06, 25.39it/s]
 46%|████▌     | 140/305 [00:05<00:06, 24.07it/s]
 47%|████▋     | 144/305 [00:05<00:05, 27.10it/s]
 49%|████▊     | 148/305 [00:05<00:05, 28.33it/s]
 50%|████▉     | 151/305 [00:05<00:05, 26.33it/s]
 51%|█████     | 155/305 [00:05<00:05, 29.21it/s]
 52%|█████▏    | 159/305 [00:06<00:05, 29.07it/s]
 53%|█████▎    | 162/305 [00:06<00:05, 25.67it/s]
 54%|█████▍    | 165/305 [00:06<00:05, 24.13it/s]
 55%|█████▌    | 168/305 [00:06<00:05, 23.85it/s]
 56%|█████▋    | 172/305 [00:06<00:05, 25.91it/s]
 57%|█████▋    | 175/305 [00:06<00:04, 26.46it/s]
 58%|█████▊    | 178/305 [00:06<00:05, 23.38it/s]
 59%|█████▉    | 181/305 [00:07<00:04, 24.84it/s]
 60%|██████    | 184/305 [00:07<00:05, 21.48it/s]
 61%|██████▏   | 187/305 [00:07<00:05, 20.09it/s]
 62%|██████▏   | 190/305 [00:07<00:05, 20.92it/s]
 63%|██████▎   | 193/305 [00:07<00:05, 22.32it/s]
 65%|██████▍   | 197/305 [00:07<00:04, 22.74it/s]
 66%|██████▌   | 200/305 [00:07<00:05, 19.88it/s]
 67%|██████▋   | 203/305 [00:08<00:04, 20.89it/s]
 68%|██████▊   | 206/305 [00:08<00:04, 21.79it/s]
 69%|██████▊   | 209/305 [00:08<00:04, 22.40it/s]
 70%|██████▉   | 213/305 [00:08<00:03, 24.80it/s]
 71%|███████   | 216/305 [00:08<00:03, 22.64it/s]
 72%|███████▏  | 220/305 [00:08<00:03, 23.25it/s]
 73%|███████▎  | 223/305 [00:08<00:03, 23.34it/s]
 74%|███████▍  | 226/305 [00:09<00:03, 23.86it/s]
 75%|███████▌  | 229/305 [00:09<00:03, 22.08it/s]
 76%|███████▌  | 232/305 [00:09<00:03, 21.72it/s]
 77%|███████▋  | 235/305 [00:09<00:03, 20.27it/s]
 78%|███████▊  | 238/305 [00:09<00:03, 20.87it/s]
 79%|███████▉  | 241/305 [00:09<00:02, 21.48it/s]
 80%|████████  | 244/305 [00:09<00:02, 21.89it/s]
 81%|████████  | 247/305 [00:10<00:02, 21.82it/s]
 82%|████████▏ | 250/305 [00:10<00:02, 22.58it/s]
 83%|████████▎ | 253/305 [00:10<00:02, 21.30it/s]
 84%|████████▍ | 256/305 [00:10<00:02, 22.20it/s]
 85%|████████▍ | 259/305 [00:10<00:01, 24.06it/s]
 86%|████████▌ | 262/305 [00:10<00:01, 23.85it/s]
 87%|████████▋ | 265/305 [00:10<00:01, 21.69it/s]
 88%|████████▊ | 269/305 [00:10<00:01, 24.80it/s]
 89%|████████▉ | 272/305 [00:11<00:01, 25.50it/s]
 90%|█████████ | 275/305 [00:11<00:01, 25.30it/s]
 91%|████████��� | 278/305 [00:11<00:01, 24.05it/s]
 92%|█████████▏| 282/305 [00:11<00:00, 25.10it/s]
 93%|█████████▎| 285/305 [00:11<00:00, 24.12it/s]
 95%|█████████▍| 289/305 [00:11<00:00, 23.61it/s]
 96%|█████████▌| 292/305 [00:11<00:00, 23.82it/s]
 97%|█████████▋| 295/305 [00:12<00:00, 22.17it/s]
 98%|█████████▊| 298/305 [00:12<00:00, 23.20it/s]
 99%|█████████▉| 302/305 [00:12<00:00, 24.21it/s]
+
  0%|          | 0/305 [00:00<?, ?it/s]
  1%|          | 3/305 [00:00<00:12, 24.65it/s]
  2%|▏         | 6/305 [00:00<00:15, 18.93it/s]
  3%|▎         | 9/305 [00:00<00:13, 21.81it/s]
  4%|▍         | 12/305 [00:00<00:12, 23.29it/s]
  5%|▍         | 15/305 [00:00<00:11, 24.61it/s]
  6%|▌         | 18/305 [00:00<00:11, 25.65it/s]
  7%|▋         | 21/305 [00:00<00:10, 26.05it/s]
  8%|▊         | 24/305 [00:00<00:10, 26.45it/s]
  9%|▉         | 27/305 [00:01<00:10, 26.86it/s]
 10%|▉         | 30/305 [00:01<00:10, 27.23it/s]
 11%|█         | 33/305 [00:01<00:10, 26.91it/s]
 12%|█▏        | 36/305 [00:01<00:11, 23.20it/s]
 13%|█▎        | 39/305 [00:01<00:11, 23.29it/s]
 14%|█▍        | 42/305 [00:01<00:11, 23.14it/s]
 15%|█▍        | 45/305 [00:01<00:10, 24.46it/s]
 16%|█▌        | 48/305 [00:01<00:10, 24.97it/s]
 17%|█▋        | 51/305 [00:02<00:10, 24.71it/s]
 18%|█▊        | 54/305 [00:02<00:10, 24.81it/s]
 19%|█▊        | 57/305 [00:02<00:10, 24.08it/s]
 20%|█▉        | 60/305 [00:02<00:10, 22.47it/s]
 21%|██        | 63/305 [00:02<00:10, 23.67it/s]
 22%|██▏       | 66/305 [00:02<00:09, 25.10it/s]
 23%|██▎       | 69/305 [00:02<00:10, 23.01it/s]
 24%|██▎       | 72/305 [00:02<00:09, 24.01it/s]
 25%|██▍       | 75/305 [00:03<00:10, 22.76it/s]
 26%|██▌       | 78/305 [00:03<00:10, 20.66it/s]
 27%|██▋       | 81/305 [00:03<00:10, 21.89it/s]
 28%|██▊       | 84/305 [00:03<00:09, 23.28it/s]
 29%|██▊       | 87/305 [00:03<00:10, 20.65it/s]
 30%|██▉       | 90/305 [00:03<00:10, 20.84it/s]
 30%|███       | 93/305 [00:03<00:09, 21.42it/s]
 31%|███▏      | 96/305 [00:04<00:10, 20.75it/s]
 32%|███▏      | 99/305 [00:04<00:09, 22.55it/s]
 33%|███▎      | 102/305 [00:04<00:09, 21.50it/s]
 34%|███▍      | 105/305 [00:04<00:08, 22.86it/s]
 35%|███▌      | 108/305 [00:04<00:08, 24.14it/s]
 36%|███▋      | 111/305 [00:04<00:08, 23.88it/s]
 37%|███▋      | 114/305 [00:04<00:07, 24.87it/s]
 38%|███▊      | 117/305 [00:05<00:08, 22.52it/s]
 39%|███▉      | 120/305 [00:05<00:07, 23.77it/s]
 40%|████      | 123/305 [00:05<00:07, 25.06it/s]
 42%|████▏     | 127/305 [00:05<00:06, 26.69it/s]
 43%|████▎     | 131/305 [00:05<00:06, 27.92it/s]
 44%|████▍     | 134/305 [00:05<00:06, 27.94it/s]
 45%|████▍     | 137/305 [00:05<00:06, 27.85it/s]
 46%|████▌     | 140/305 [00:05<00:05, 28.28it/s]
 47%|████▋     | 144/305 [00:05<00:05, 28.87it/s]
 48%|████▊     | 147/305 [00:06<00:05, 26.40it/s]
 49%|████▉     | 150/305 [00:06<00:05, 25.95it/s]
 50%|█████     | 153/305 [00:06<00:05, 25.68it/s]
 51%|█████     | 156/305 [00:06<00:05, 25.84it/s]
 52%|█████▏    | 159/305 [00:06<00:05, 26.81it/s]
 53%|█████▎    | 162/305 [00:06<00:05, 27.15it/s]
 54%|█████▍    | 165/305 [00:06<00:05, 27.77it/s]
 55%|█████▌    | 168/305 [00:06<00:04, 28.12it/s]
 56%|█████▋    | 172/305 [00:06<00:04, 28.89it/s]
 57%|█████▋    | 175/305 [00:07<00:05, 25.12it/s]
 58%|█████▊    | 178/305 [00:07<00:05, 23.65it/s]
 59%|█████▉    | 181/305 [00:07<00:04, 24.93it/s]
 61%|██████    | 185/305 [00:07<00:04, 26.48it/s]
 62%|██████▏   | 189/305 [00:07<00:04, 27.38it/s]
 63%|██████▎   | 192/305 [00:07<00:04, 27.82it/s]
 64%|██████▍   | 195/305 [00:07<00:04, 27.12it/s]
 65%|██████▌   | 199/305 [00:08<00:03, 28.32it/s]
 66%|██████▌   | 202/305 [00:08<00:03, 28.32it/s]
 67%|██████▋   | 205/305 [00:08<00:03, 28.11it/s]
 68%|██████▊   | 208/305 [00:08<00:03, 28.39it/s]
 69%|██████▉   | 211/305 [00:08<00:03, 28.30it/s]
 70%|███████   | 214/305 [00:08<00:03, 28.37it/s]
 71%|███████   | 217/305 [00:08<00:03, 26.26it/s]
 72%|███████▏  | 220/305 [00:08<00:03, 27.15it/s]
 73%|███████▎  | 223/305 [00:08<00:02, 27.39it/s]
 74%|███████▍  | 226/305 [00:08<00:02, 27.81it/s]
 75%|███████▌  | 229/305 [00:09<00:02, 28.28it/s]
 76%|███████▋  | 233/305 [00:09<00:02, 29.12it/s]
 77%|███████▋  | 236/305 [00:09<00:02, 29.36it/s]
 79%|███████▊  | 240/305 [00:09<00:02, 29.58it/s]
 80%|███████▉  | 243/305 [00:09<00:02, 28.78it/s]
 81%|████████  | 246/305 [00:09<00:02, 27.70it/s]
 82%|████████▏ | 250/305 [00:09<00:01, 28.16it/s]
 83%|████████▎ | 253/305 [00:09<00:01, 27.51it/s]
 84%|████████▍ | 256/305 [00:10<00:01, 27.37it/s]
 85%|████████▍ | 259/305 [00:10<00:01, 27.85it/s]
 86%|████████▌ | 263/305 [00:10<00:01, 28.82it/s]
 87%|████████▋ | 266/305 [00:10<00:01, 28.39it/s]
 88%|████████▊ | 269/305 [00:10<00:01, 27.91it/s]
 89%|████████▉ | 272/305 [00:10<00:01, 28.32it/s]
 90%|█████████ | 275/305 [00:10<00:01, 26.85it/s]
 91%|█████████ | 278/305 [00:10<00:00, 27.22it/s]
 92%|█████████▏| 281/305 [00:10<00:00, 27.78it/s]
 93%|█████████▎| 284/305 [00:11<00:00, 28.40it/s]
 94%|█████████▍| 287/305 [00:11<00:00, 28.28it/s]
 95%|█████████▌| 290/305 [00:11<00:00, 28.51it/s]
 96%|█████████▌| 293/305 [00:11<00:00, 27.87it/s]
 97%|█████████▋| 296/305 [00:11<00:00, 28.30it/s]
 98%|█████████▊| 299/305 [00:11<00:00, 28.76it/s]
 99%|█████████▉| 302/305 [00:11<00:00, 27.18it/s]
+FID: 148.0970162151841
+
+/mnt/petrelfs/leihaodong/anaconda3/envs/eval_xx/lib/python3.10/site-packages/torchmetrics/utilities/prints.py:43: UserWarning: Metric `InceptionScore` will save all extracted features in buffer. For large datasets this may lead to large memory footprint.
+  warnings.warn(*args, **kwargs)
+/mnt/petrelfs/leihaodong/anaconda3/envs/eval_xx/lib/python3.10/site-packages/open_clip/factory.py:450: UserWarning: QuickGELU mismatch between final model config (quick_gelu=False) and pretrained tag 'openai' (quick_gelu=True).
+  warnings.warn(
+Using a slow image processor as `use_fast` is unset and a slow processor was saved with this model. `use_fast=True` will be the default behavior in v4.52, even if the model was saved with a slow processor. This will result in minor differences in outputs. You'll still be able to use a slow processor with `use_fast=False`.
+DINO_I: 0.12442810833454132
+
+  warnings.warn(f"Importing from {__name__} is deprecated, please import via timm.layers", FutureWarning)
+/mnt/petrelfs/leihaodong/ICLR25/eval/HPSv2/hpsv2/img_score.py:107: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead.
+  with torch.cuda.amp.autocast():
+
+WARNING:huggingface_hub.utils._http:'(ReadTimeoutError("HTTPSConnectionPool(host='huggingface.co', port=443): Read timed out. (read timeout=10)"), '(Request ID: fa702bc0-a189-422e-88f5-9c56344b0abf)')' thrown while requesting HEAD https://huggingface.co/xswu/HPSv2/resolve/main/HPS_v2.1_compressed.pt
+Retrying in 1s [Retry 1/5].
+WARNING:huggingface_hub.utils._http:Retrying in 1s [Retry 1/5].
+
+INFO:absl:FID: 148.0970162151841
+INFO:absl:DINO_I: 0.12442810833454132
+INFO:absl:IS: (tensor(12.1260, device='cuda:0'), tensor(1.7703, device='cuda:0'))
+INFO:absl:CLIP_T: 0.18574747233117214
+INFO:absl:HPSv2: 0.27285956711065573
+INFO:absl:Aesthetic: 6.442194496217321
+Scores:
+FID: 148.0970162151841
+DINO_I: 0.12442810833454132
+IS: (tensor(12.1260, device='cuda:0'), tensor(1.7703, device='cuda:0'))
+CLIP_T: 0.18574747233117214
+HPSv2: 0.27285956711065573
+Aesthetic: 6.442194496217321

scripts/eval/slurm-6226601.out

@@ -0,0 +1,43 @@
+loading annotations into memory...
+Done (t=0.04s)
+creating index...
+index created!
+一共497 张
+loading annotations into memory...
+Done (t=0.03s)
+creating index...
+index created!
+/mnt/petrelfs/leihaodong/ICLR25/exp/Emu3/MSCOCO2017Val/sjd/img/gt
+/mnt/petrelfs/leihaodong/ICLR25/exp/Emu3/MSCOCO2017Val/sjd/img
+
  0%|          | 0/497 [00:00<?, ?it/s]
  0%|          | 1/497 [00:00<03:53,  2.13it/s]
  2%|▏         | 8/497 [00:00<00:27, 17.78it/s]
  3%|▎         | 15/497 [00:00<00:15, 30.13it/s]
  5%|▍         | 23/497 [00:00<00:11, 41.98it/s]
  6%|▌         | 31/497 [00:00<00:09, 51.70it/s]
  8%|▊         | 39/497 [00:01<00:07, 57.54it/s]
  9%|▉         | 47/497 [00:01<00:07, 62.15it/s]
 11%|█         | 55/497 [00:01<00:06, 65.27it/s]
 13%|█▎        | 63/497 [00:01<00:06, 67.11it/s]
 14%|█▍        | 71/497 [00:01<00:06, 68.03it/s]
 16%|█▌        | 79/497 [00:01<00:05, 69.78it/s]
 18%|█▊        | 87/497 [00:01<00:05, 70.98it/s]
 19%|█▉        | 95/497 [00:01<00:05, 71.04it/s]
 21%|██        | 103/497 [00:01<00:05, 70.95it/s]
 22%|██▏       | 111/497 [00:02<00:05, 70.14it/s]
 24%|██▍       | 119/497 [00:02<00:05, 69.43it/s]
 25%|██▌       | 126/497 [00:02<00:05, 69.27it/s]
 27%|██▋       | 133/497 [00:02<00:05, 69.16it/s]
 28%|██▊       | 140/497 [00:02<00:05, 69.38it/s]
 30%|██▉       | 147/497 [00:02<00:05, 66.54it/s]
 31%|███       | 154/497 [00:02<00:05, 66.49it/s]
 32%|███▏      | 161/497 [00:02<00:05, 67.13it/s]
 34%|███▍      | 169/497 [00:02<00:04, 70.05it/s]
 36%|███▌      | 177/497 [00:02<00:04, 71.54it/s]
 37%|███▋      | 185/497 [00:03<00:04, 70.93it/s]
 39%|███▉      | 193/497 [00:03<00:04, 71.60it/s]
 40%|████      | 201/497 [00:03<00:04, 71.74it/s]
 42%|████▏     | 209/497 [00:03<00:04, 70.77it/s]
 44%|████▎     | 217/497 [00:03<00:03, 70.38it/s]
 45%|████▌     | 225/497 [00:03<00:03, 69.95it/s]
 47%|████▋     | 233/497 [00:03<00:03, 69.22it/s]
 48%|████▊     | 241/497 [00:03<00:03, 70.03it/s]
 50%|█████     | 249/497 [00:03<00:03, 69.94it/s]
 52%|█████▏    | 257/497 [00:04<00:03, 70.17it/s]
 53%|█████▎    | 265/497 [00:04<00:03, 69.02it/s]
 55%|█████▍    | 273/497 [00:04<00:03, 69.48it/s]
 56%|█████▋    | 280/497 [00:04<00:03, 69.12it/s]
 58%|█████▊    | 288/497 [00:04<00:03, 69.43it/s]
 59%|█████▉    | 295/497 [00:04<00:02, 69.50it/s]
 61%|██████    | 302/497 [00:04<00:02, 69.19it/s]
 62%|██████▏   | 309/497 [00:04<00:02, 68.99it/s]
 64%|██████▍   | 317/497 [00:04<00:02, 70.49it/s]
 65%|██████▌   | 325/497 [00:05<00:02, 72.72it/s]
 67%|██████▋   | 333/497 [00:05<00:02, 74.02it/s]
 69%|██████▊   | 341/497 [00:05<00:02, 74.61it/s]
 70%|███████   | 349/497 [00:05<00:01, 75.78it/s]
 72%|███████▏  | 357/497 [00:05<00:01, 74.47it/s]
 73%|███████▎  | 365/497 [00:05<00:01, 73.26it/s]
 75%|███████▌  | 373/497 [00:05<00:01, 74.02it/s]
 77%|███████▋  | 381/497 [00:05<00:01, 70.41it/s]
 78%|███████▊  | 389/497 [00:05<00:01, 71.25it/s]
 80%|███████▉  | 397/497 [00:06<00:01, 70.34it/s]
 81%|████████▏ | 405/497 [00:06<00:01, 71.53it/s]
 83%|████████▎ | 413/497 [00:06<00:01, 71.76it/s]
 85%|████████▍ | 421/497 [00:06<00:01, 70.54it/s]
 86%|████████▋ | 429/497 [00:06<00:00, 71.35it/s]
 88%|████████▊ | 437/497 [00:06<00:00, 71.41it/s]
 90%|████████▉ | 445/497 [00:06<00:00, 72.13it/s]
 91%|█████████ | 453/497 [00:06<00:00, 71.27it/s]
 93%|█████████▎| 461/497 [00:06<00:00, 71.72it/s]
 94%|█████████▍| 469/497 [00:07<00:00, 72.59it/s]
 96%|█████████▌| 477/497 [00:07<00:00, 72.54it/s]
 98%|█████████▊| 485/497 [00:07<00:00, 72.82it/s]
 99%|█████████▉| 493/497 [00:07<00:00, 70.22it/s]
+
  0%|          | 0/497 [00:00<?, ?it/s]
  1%|          | 4/497 [00:00<00:15, 30.90it/s]
  2%|▏         | 8/497 [00:00<00:15, 31.27it/s]
  2%|▏         | 12/497 [00:00<00:15, 30.68it/s]
  3%|▎         | 16/497 [00:00<00:15, 30.43it/s]
  4%|▍         | 20/497 [00:00<00:15, 31.00it/s]
  5%|▍         | 24/497 [00:00<00:15, 31.38it/s]
  6%|▌         | 28/497 [00:00<00:15, 30.92it/s]
  6%|▋         | 32/497 [00:01<00:15, 30.70it/s]
  7%|▋         | 36/497 [00:01<00:15, 30.71it/s]
  8%|▊         | 40/497 [00:01<00:14, 31.38it/s]
  9%|▉         | 44/497 [00:01<00:14, 31.40it/s]
 10%|▉         | 48/497 [00:01<00:14, 31.23it/s]
 10%|█         | 52/497 [00:01<00:14, 31.11it/s]
 11%|█▏        | 56/497 [00:01<00:14, 31.27it/s]
 12%|█▏        | 60/497 [00:01<00:13, 31.45it/s]
 13%|█▎        | 64/497 [00:02<00:13, 31.79it/s]
 14%|█▎        | 68/497 [00:02<00:13, 31.94it/s]
 14%|█▍        | 72/497 [00:02<00:13, 32.28it/s]
 15%|█▌        | 76/497 [00:02<00:13, 31.65it/s]
 16%|█▌        | 80/497 [00:02<00:13, 31.61it/s]
 17%|█▋        | 84/497 [00:02<00:13, 31.76it/s]
 18%|█▊        | 88/497 [00:02<00:12, 31.77it/s]
 19%|█▊        | 92/497 [00:02<00:12, 32.10it/s]
 19%|█▉        | 96/497 [00:03<00:12, 32.08it/s]
 20%|██        | 100/497 [00:03<00:12, 31.40it/s]
 21%|██        | 104/497 [00:03<00:12, 31.23it/s]
 22%|██▏       | 108/497 [00:03<00:12, 31.22it/s]
 23%|██▎       | 112/497 [00:03<00:12, 30.16it/s]
 23%|██▎       | 116/497 [00:03<00:12, 31.17it/s]
 24%|██▍       | 120/497 [00:03<00:12, 30.61it/s]
 25%|██▍       | 124/497 [00:03<00:11, 31.90it/s]
 26%|██▌       | 128/497 [00:04<00:11, 32.49it/s]
 27%|██▋       | 132/497 [00:04<00:11, 32.75it/s]
 27%|██▋       | 136/497 [00:04<00:11, 31.82it/s]
 28%|██▊       | 140/497 [00:04<00:11, 31.69it/s]
 29%|██▉       | 144/497 [00:04<00:10, 32.31it/s]
 30%|██▉       | 148/497 [00:04<00:10, 32.44it/s]
 31%|███       | 152/497 [00:04<00:10, 32.36it/s]
 31%|███▏      | 156/497 [00:04<00:10, 32.27it/s]
 32%|███▏      | 160/497 [00:05<00:10, 32.12it/s]
 33%|███▎      | 164/497 [00:05<00:10, 32.33it/s]
 34%|███▍      | 168/497 [00:05<00:10, 32.40it/s]
 35%|███▍      | 172/497 [00:05<00:10, 32.31it/s]
 35%|███▌      | 176/497 [00:05<00:09, 32.28it/s]
 36%|███▌      | 180/497 [00:05<00:09, 32.24it/s]
 37%|███▋      | 184/497 [00:05<00:09, 32.09it/s]
 38%|███▊      | 188/497 [00:05<00:09, 31.51it/s]
 39%|███▊      | 192/497 [00:06<00:09, 32.30it/s]
 39%|███▉      | 196/497 [00:06<00:09, 30.89it/s]
 40%|████      | 200/497 [00:06<00:09, 31.32it/s]
 41%|████      | 204/497 [00:06<00:09, 31.92it/s]
 42%|████▏     | 208/497 [00:06<00:09, 31.46it/s]
 43%|████▎     | 212/497 [00:06<00:09, 31.30it/s]
 43%|████▎     | 216/497 [00:06<00:08, 31.75it/s]
 44%|████▍     | 220/497 [00:06<00:08, 31.89it/s]
 45%|████▌     | 224/497 [00:07<00:08, 32.05it/s]
 46%|████▌     | 228/497 [00:07<00:08, 31.89it/s]
 47%|████▋     | 232/497 [00:07<00:08, 32.20it/s]
 47%|████▋     | 236/497 [00:07<00:08, 32.10it/s]
 48%|████▊     | 240/497 [00:07<00:07, 32.46it/s]
 49%|████▉     | 244/497 [00:07<00:07, 32.20it/s]
 50%|████▉     | 248/497 [00:07<00:07, 32.38it/s]
 51%|█████     | 252/497 [00:07<00:07, 32.81it/s]
 52%|█████▏    | 256/497 [00:08<00:07, 32.31it/s]
 52%|█████▏    | 260/497 [00:08<00:07, 32.34it/s]
 53%|█████▎    | 264/497 [00:08<00:07, 32.35it/s]
 54%|█████▍    | 268/497 [00:08<00:07, 31.95it/s]
 55%|█████▍    | 272/497 [00:08<00:07, 31.61it/s]
 56%|█████▌    | 276/497 [00:08<00:06, 31.61it/s]
 56%|█████▋    | 280/497 [00:08<00:06, 31.35it/s]
 57%|█████▋    | 284/497 [00:08<00:06, 31.14it/s]
 58%|█████▊    | 288/497 [00:09<00:06, 31.44it/s]
 59%|█████▉    | 292/497 [00:09<00:06, 31.06it/s]
 60%|█████▉    | 296/497 [00:09<00:06, 30.25it/s]
 60%|██████    | 300/497 [00:09<00:06, 30.96it/s]
 61%|██████    | 304/497 [00:09<00:06, 31.10it/s]
 62%|██████▏   | 308/497 [00:09<00:06, 31.15it/s]
 63%|██████▎   | 312/497 [00:09<00:05, 31.22it/s]
 64%|██████▎   | 316/497 [00:09<00:05, 31.33it/s]
 64%|██████▍   | 320/497 [00:10<00:05, 31.48it/s]
 65%|██████▌   | 324/497 [00:10<00:05, 31.27it/s]
 66%|██████▌   | 328/497 [00:10<00:05, 32.18it/s]
 67%|██████▋   | 332/497 [00:10<00:05, 32.93it/s]
 68%|██████▊   | 336/497 [00:10<00:04, 33.35it/s]
 68%|██████▊   | 340/497 [00:10<00:04, 32.87it/s]
 69%|██████▉   | 344/497 [00:10<00:04, 33.10it/s]
 70%|███████   | 348/497 [00:10<00:04, 32.74it/s]
 71%|███████   | 352/497 [00:11<00:04, 33.38it/s]
 72%|███████▏  | 356/497 [00:11<00:04, 32.85it/s]
 72%|███████▏  | 360/497 [00:11<00:04, 32.85it/s]
 73%|███████▎  | 364/497 [00:11<00:04, 32.65it/s]
 74%|███████▍  | 368/497 [00:11<00:03, 32.64it/s]
 75%|███████▍  | 372/497 [00:11<00:03, 32.32it/s]
 76%|███████▌  | 376/497 [00:11<00:03, 32.28it/s]
 76%|███████▋  | 380/497 [00:11<00:03, 32.71it/s]
 77%|███████▋  | 384/497 [00:12<00:03, 32.97it/s]
 78%|███████▊  | 388/497 [00:12<00:03, 32.67it/s]
 79%|███████▉  | 392/497 [00:12<00:03, 32.15it/s]
 80%|███████▉  | 396/497 [00:12<00:03, 32.28it/s]
 80%|████████  | 400/497 [00:12<00:02, 32.75it/s]
 81%|██████���█▏ | 404/497 [00:12<00:02, 33.44it/s]
 82%|████████▏ | 408/497 [00:12<00:02, 33.47it/s]
 83%|████████▎ | 412/497 [00:12<00:02, 33.29it/s]
 84%|████████▎ | 416/497 [00:13<00:02, 32.90it/s]
 85%|████████▍ | 420/497 [00:13<00:02, 32.85it/s]
 85%|████████▌ | 424/497 [00:13<00:02, 32.31it/s]
 86%|████████▌ | 428/497 [00:13<00:02, 32.33it/s]
 87%|████████▋ | 432/497 [00:13<00:02, 32.26it/s]
 88%|████████▊ | 436/497 [00:13<00:01, 32.52it/s]
 89%|████████▊ | 440/497 [00:13<00:01, 32.31it/s]
 89%|████████▉ | 444/497 [00:13<00:01, 32.13it/s]
 90%|█████████ | 448/497 [00:14<00:01, 32.34it/s]
 91%|█████████ | 452/497 [00:14<00:01, 32.32it/s]
 92%|█████████▏| 456/497 [00:14<00:01, 32.41it/s]
 93%|█████████▎| 460/497 [00:14<00:01, 31.79it/s]
 93%|█████████▎| 464/497 [00:14<00:01, 32.50it/s]
 94%|█████████▍| 468/497 [00:14<00:00, 33.36it/s]
 95%|█████████▍| 472/497 [00:14<00:00, 33.04it/s]
 96%|█████████▌| 476/497 [00:14<00:00, 33.52it/s]
 97%|█████████▋| 480/497 [00:14<00:00, 33.65it/s]
 97%|█████████▋| 484/497 [00:15<00:00, 33.30it/s]
 98%|█████████▊| 488/497 [00:15<00:00, 32.63it/s]
 99%|█████████▉| 492/497 [00:15<00:00, 32.61it/s]
+FID: 106.12328835883358
+
+/mnt/petrelfs/leihaodong/anaconda3/envs/eval_xx/lib/python3.10/site-packages/torchmetrics/utilities/prints.py:43: UserWarning: Metric `InceptionScore` will save all extracted features in buffer. For large datasets this may lead to large memory footprint.
+  warnings.warn(*args, **kwargs)
+/mnt/petrelfs/leihaodong/anaconda3/envs/eval_xx/lib/python3.10/site-packages/open_clip/factory.py:450: UserWarning: QuickGELU mismatch between final model config (quick_gelu=False) and pretrained tag 'openai' (quick_gelu=True).
+  warnings.warn(
+Using a slow image processor as `use_fast` is unset and a slow processor was saved with this model. `use_fast=True` will be the default behavior in v4.52, even if the model was saved with a slow processor. This will result in minor differences in outputs. You'll still be able to use a slow processor with `use_fast=False`.
+DINO_I: 0.12495101243257523
+
+  warnings.warn(f"Importing from {__name__} is deprecated, please import via timm.layers", FutureWarning)
+/mnt/petrelfs/leihaodong/ICLR25/eval/HPSv2/hpsv2/img_score.py:107: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead.
+  with torch.cuda.amp.autocast():
+
+WARNING:huggingface_hub.utils._http:'(ReadTimeoutError("HTTPSConnectionPool(host='huggingface.co', port=443): Read timed out. (read timeout=10)"), '(Request ID: 71d320ba-c4cd-4216-8a61-7118aca31918)')' thrown while requesting HEAD https://huggingface.co/xswu/HPSv2/resolve/main/HPS_v2.1_compressed.pt
+Retrying in 1s [Retry 1/5].
+WARNING:huggingface_hub.utils._http:Retrying in 1s [Retry 1/5].
+
+INFO:absl:FID: 106.12328835883358
+INFO:absl:DINO_I: 0.12495101243257523
+INFO:absl:IS: (tensor(17.3624, device='cuda:0'), tensor(1.3039, device='cuda:0'))
+INFO:absl:CLIP_T: 0.19042356838462818
+INFO:absl:HPSv2: 0.2899845361229879
+INFO:absl:Aesthetic: 6.541996148030763
+Scores:
+FID: 106.12328835883358
+DINO_I: 0.12495101243257523
+IS: (tensor(17.3624, device='cuda:0'), tensor(1.3039, device='cuda:0'))
+CLIP_T: 0.19042356838462818
+HPSv2: 0.2899845361229879
+Aesthetic: 6.541996148030763

scripts/eval/slurm-6234587.out

@@ -0,0 +1,22 @@
+/mnt/petrelfs/leihaodong/anaconda3/envs/eval_xx/lib/python3.10/site-packages/torchmetrics/utilities/prints.py:43: UserWarning: Metric `InceptionScore` will save all extracted features in buffer. For large datasets this may lead to large memory footprint.
+  warnings.warn(*args, **kwargs)
+/mnt/petrelfs/leihaodong/anaconda3/envs/eval_xx/lib/python3.10/site-packages/open_clip/factory.py:450: UserWarning: QuickGELU mismatch between final model config (quick_gelu=False) and pretrained tag 'openai' (quick_gelu=True).
+  warnings.warn(
+Using a slow image processor as `use_fast` is unset and a slow processor was saved with this model. `use_fast=True` will be the default behavior in v4.52, even if the model was saved with a slow processor. This will result in minor differences in outputs. You'll still be able to use a slow processor with `use_fast=False`.
+Loading PartiPrompts from /mnt/petrelfs/leihaodong/ICLR25/sjdtree/data/prompts/PartiPrompts.tsv for GPU 0, Node 0
+Number of unique prompts: 1632 | Max prompt length: 389
+一共932 张
+
+  warnings.warn(f"Importing from {__name__} is deprecated, please import via timm.layers", FutureWarning)
+/mnt/petrelfs/leihaodong/ICLR25/eval/HPSv2/hpsv2/img_score.py:107: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead.
+  with torch.cuda.amp.autocast():
+
+INFO:absl:IS: (tensor(17.2924, device='cuda:0'), tensor(1.6669, device='cuda:0'))
+INFO:absl:CLIP_T: 0.19844276723032217
+INFO:absl:HPSv2: 0.28268498105552575
+INFO:absl:Aesthetic: 6.514075135212599
+Scores:
+IS: (tensor(17.2924, device='cuda:0'), tensor(1.6669, device='cuda:0'))
+CLIP_T: 0.19844276723032217
+HPSv2: 0.28268498105552575
+Aesthetic: 6.514075135212599

scripts/eval/slurm-6234588.out

@@ -0,0 +1,22 @@
+/mnt/petrelfs/leihaodong/anaconda3/envs/eval_xx/lib/python3.10/site-packages/torchmetrics/utilities/prints.py:43: UserWarning: Metric `InceptionScore` will save all extracted features in buffer. For large datasets this may lead to large memory footprint.
+  warnings.warn(*args, **kwargs)
+/mnt/petrelfs/leihaodong/anaconda3/envs/eval_xx/lib/python3.10/site-packages/open_clip/factory.py:450: UserWarning: QuickGELU mismatch between final model config (quick_gelu=False) and pretrained tag 'openai' (quick_gelu=True).
+  warnings.warn(
+Using a slow image processor as `use_fast` is unset and a slow processor was saved with this model. `use_fast=True` will be the default behavior in v4.52, even if the model was saved with a slow processor. This will result in minor differences in outputs. You'll still be able to use a slow processor with `use_fast=False`.
+Loading PartiPrompts from /mnt/petrelfs/leihaodong/ICLR25/sjdtree/data/prompts/PartiPrompts.tsv for GPU 0, Node 0
+Number of unique prompts: 1632 | Max prompt length: 389
+一共147 张
+
+  warnings.warn(f"Importing from {__name__} is deprecated, please import via timm.layers", FutureWarning)
+/mnt/petrelfs/leihaodong/ICLR25/eval/HPSv2/hpsv2/img_score.py:107: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead.
+  with torch.cuda.amp.autocast():
+
+INFO:absl:IS: (tensor(6.2036, device='cuda:0'), tensor(0.9269, device='cuda:0'))
+INFO:absl:CLIP_T: 0.25545033842933423
+INFO:absl:HPSv2: 0.2595023849383503
+INFO:absl:Aesthetic: 6.3289138703119185
+Scores:
+IS: (tensor(6.2036, device='cuda:0'), tensor(0.9269, device='cuda:0'))
+CLIP_T: 0.25545033842933423
+HPSv2: 0.2595023849383503
+Aesthetic: 6.3289138703119185

scripts/eval/slurm-6234589.out

@@ -0,0 +1,22 @@
+/mnt/petrelfs/leihaodong/anaconda3/envs/eval_xx/lib/python3.10/site-packages/torchmetrics/utilities/prints.py:43: UserWarning: Metric `InceptionScore` will save all extracted features in buffer. For large datasets this may lead to large memory footprint.
+  warnings.warn(*args, **kwargs)
+/mnt/petrelfs/leihaodong/anaconda3/envs/eval_xx/lib/python3.10/site-packages/open_clip/factory.py:450: UserWarning: QuickGELU mismatch between final model config (quick_gelu=False) and pretrained tag 'openai' (quick_gelu=True).
+  warnings.warn(
+Using a slow image processor as `use_fast` is unset and a slow processor was saved with this model. `use_fast=True` will be the default behavior in v4.52, even if the model was saved with a slow processor. This will result in minor differences in outputs. You'll still be able to use a slow processor with `use_fast=False`.
+Loading PartiPrompts from /mnt/petrelfs/leihaodong/ICLR25/sjdtree/data/prompts/PartiPrompts.tsv for GPU 0, Node 0
+Number of unique prompts: 1632 | Max prompt length: 389
+一共696 张
+
+  warnings.warn(f"Importing from {__name__} is deprecated, please import via timm.layers", FutureWarning)
+/mnt/petrelfs/leihaodong/ICLR25/eval/HPSv2/hpsv2/img_score.py:107: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead.
+  with torch.cuda.amp.autocast():
+
+Traceback (most recent call last):
+  File "/mnt/petrelfs/leihaodong/ICLR25/eval/evaluation_metrics.py", line 436, in <module>
+    results = evaluate_quantitative_scores_text2img(
+  File "/mnt/petrelfs/leihaodong/ICLR25/eval/evaluation_metrics.py", line 309, in evaluate_quantitative_scores_text2img
+    image = Image.open(image_file)
+  File "/mnt/petrelfs/leihaodong/anaconda3/envs/eval_xx/lib/python3.10/site-packages/PIL/Image.py", line 3560, in open
+    raise UnidentifiedImageError(msg)
+PIL.UnidentifiedImageError: cannot identify image file '/mnt/petrelfs/leihaodong/ICLR25/exp/Lumina-mGPT-7B-768_PartiPrompts/mscr_sjd_0.05/img/695.png'
+srun: error: SH-IDC1-10-140-37-43: task 0: Exited with exit code 1

scripts/eval/slurm-6234590.out

@@ -0,0 +1,22 @@
+/mnt/petrelfs/leihaodong/anaconda3/envs/eval_xx/lib/python3.10/site-packages/torchmetrics/utilities/prints.py:43: UserWarning: Metric `InceptionScore` will save all extracted features in buffer. For large datasets this may lead to large memory footprint.
+  warnings.warn(*args, **kwargs)
+/mnt/petrelfs/leihaodong/anaconda3/envs/eval_xx/lib/python3.10/site-packages/open_clip/factory.py:450: UserWarning: QuickGELU mismatch between final model config (quick_gelu=False) and pretrained tag 'openai' (quick_gelu=True).
+  warnings.warn(
+Using a slow image processor as `use_fast` is unset and a slow processor was saved with this model. `use_fast=True` will be the default behavior in v4.52, even if the model was saved with a slow processor. This will result in minor differences in outputs. You'll still be able to use a slow processor with `use_fast=False`.
+Loading PartiPrompts from /mnt/petrelfs/leihaodong/ICLR25/sjdtree/data/prompts/PartiPrompts.tsv for GPU 0, Node 0
+Number of unique prompts: 1632 | Max prompt length: 389
+一共172 张
+
+  warnings.warn(f"Importing from {__name__} is deprecated, please import via timm.layers", FutureWarning)
+/mnt/petrelfs/leihaodong/ICLR25/eval/HPSv2/hpsv2/img_score.py:107: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead.
+  with torch.cuda.amp.autocast():
+
+INFO:absl:IS: (tensor(6.1117, device='cuda:0'), tensor(0.9643, device='cuda:0'))
+INFO:absl:CLIP_T: 0.25045652375664823
+INFO:absl:HPSv2: 0.23206205146257267
+INFO:absl:Aesthetic: 5.643340925837672
+Scores:
+IS: (tensor(6.1117, device='cuda:0'), tensor(0.9643, device='cuda:0'))
+CLIP_T: 0.25045652375664823
+HPSv2: 0.23206205146257267
+Aesthetic: 5.643340925837672

scripts/eval/slurm-6235595.out

@@ -0,0 +1,22 @@
+/mnt/petrelfs/leihaodong/anaconda3/envs/eval_xx/lib/python3.10/site-packages/torchmetrics/utilities/prints.py:43: UserWarning: Metric `InceptionScore` will save all extracted features in buffer. For large datasets this may lead to large memory footprint.
+  warnings.warn(*args, **kwargs)
+/mnt/petrelfs/leihaodong/anaconda3/envs/eval_xx/lib/python3.10/site-packages/open_clip/factory.py:450: UserWarning: QuickGELU mismatch between final model config (quick_gelu=False) and pretrained tag 'openai' (quick_gelu=True).
+  warnings.warn(
+Using a slow image processor as `use_fast` is unset and a slow processor was saved with this model. `use_fast=True` will be the default behavior in v4.52, even if the model was saved with a slow processor. This will result in minor differences in outputs. You'll still be able to use a slow processor with `use_fast=False`.
+Loading PartiPrompts from /mnt/petrelfs/leihaodong/ICLR25/sjdtree/data/prompts/PartiPrompts.tsv for GPU 0, Node 0
+Number of unique prompts: 1632 | Max prompt length: 389
+一共695 张
+
+  warnings.warn(f"Importing from {__name__} is deprecated, please import via timm.layers", FutureWarning)
+/mnt/petrelfs/leihaodong/ICLR25/eval/HPSv2/hpsv2/img_score.py:107: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead.
+  with torch.cuda.amp.autocast():
+
+INFO:absl:IS: (tensor(12.8385, device='cuda:0'), tensor(1.2827, device='cuda:0'))
+INFO:absl:CLIP_T: 0.1999332770946429
+INFO:absl:HPSv2: 0.25848292069469425
+INFO:absl:Aesthetic: 6.157701880297215
+Scores:
+IS: (tensor(12.8385, device='cuda:0'), tensor(1.2827, device='cuda:0'))
+CLIP_T: 0.1999332770946429
+HPSv2: 0.25848292069469425
+Aesthetic: 6.157701880297215

scripts/eval/slurm-6235610.out

@@ -0,0 +1,22 @@
+/mnt/petrelfs/leihaodong/anaconda3/envs/eval_xx/lib/python3.10/site-packages/torchmetrics/utilities/prints.py:43: UserWarning: Metric `InceptionScore` will save all extracted features in buffer. For large datasets this may lead to large memory footprint.
+  warnings.warn(*args, **kwargs)
+/mnt/petrelfs/leihaodong/anaconda3/envs/eval_xx/lib/python3.10/site-packages/open_clip/factory.py:450: UserWarning: QuickGELU mismatch between final model config (quick_gelu=False) and pretrained tag 'openai' (quick_gelu=True).
+  warnings.warn(
+Using a slow image processor as `use_fast` is unset and a slow processor was saved with this model. `use_fast=True` will be the default behavior in v4.52, even if the model was saved with a slow processor. This will result in minor differences in outputs. You'll still be able to use a slow processor with `use_fast=False`.
+Loading PartiPrompts from /mnt/petrelfs/leihaodong/ICLR25/sjdtree/data/prompts/PartiPrompts.tsv for GPU 0, Node 0
+Number of unique prompts: 1632 | Max prompt length: 389
+一共631 张
+
+  warnings.warn(f"Importing from {__name__} is deprecated, please import via timm.layers", FutureWarning)
+/mnt/petrelfs/leihaodong/ICLR25/eval/HPSv2/hpsv2/img_score.py:107: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead.
+  with torch.cuda.amp.autocast():
+
+INFO:absl:IS: (tensor(5.3711, device='cuda:0'), tensor(0.3064, device='cuda:0'))
+INFO:absl:CLIP_T: 0.1986828702494573
+INFO:absl:HPSv2: 0.1616893834810816
+INFO:absl:Aesthetic: 4.097829038103106
+Scores:
+IS: (tensor(5.3711, device='cuda:0'), tensor(0.3064, device='cuda:0'))
+CLIP_T: 0.1986828702494573
+HPSv2: 0.1616893834810816
+Aesthetic: 4.097829038103106

scripts/log/emu3/PP/slurm-6224918.out

@@ -0,0 +1,21 @@
+/mnt/petrelfs/leihaodong/anaconda3/envs/eval_xx/lib/python3.10/site-packages/torchmetrics/utilities/prints.py:43: UserWarning: Metric `InceptionScore` will save all extracted features in buffer. For large datasets this may lead to large memory footprint.
+  warnings.warn(*args, **kwargs)
+/mnt/petrelfs/leihaodong/anaconda3/envs/eval_xx/lib/python3.10/site-packages/open_clip/factory.py:450: UserWarning: QuickGELU mismatch between final model config (quick_gelu=False) and pretrained tag 'openai' (quick_gelu=True).
+  warnings.warn(
+Using a slow image processor as `use_fast` is unset and a slow processor was saved with this model. `use_fast=True` will be the default behavior in v4.52, even if the model was saved with a slow processor. This will result in minor differences in outputs. You'll still be able to use a slow processor with `use_fast=False`.
+Loading PartiPrompts from /mnt/petrelfs/leihaodong/ICLR25/sjdtree/data/prompts/PartiPrompts.tsv for GPU 0, Node 0
+Number of unique prompts: 1632 | Max prompt length: 389
+
+  warnings.warn(f"Importing from {__name__} is deprecated, please import via timm.layers", FutureWarning)
+/mnt/petrelfs/leihaodong/ICLR25/eval/HPSv2/hpsv2/img_score.py:107: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead.
+  with torch.cuda.amp.autocast():
+
+INFO:absl:IS: (tensor(12.2022, device='cuda:0'), tensor(1.0720, device='cuda:0'))
+INFO:absl:CLIP_T: 0.22810792344622313
+INFO:absl:HPSv2: 0.2903643798828125
+INFO:absl:Aesthetic: 6.61770681142807
+Scores:
+IS: (tensor(12.2022, device='cuda:0'), tensor(1.0720, device='cuda:0'))
+CLIP_T: 0.22810792344622313
+HPSv2: 0.2903643798828125
+Aesthetic: 6.61770681142807

scripts/log/emu3/PP/slurm-6224919.out

@@ -0,0 +1,21 @@
+/mnt/petrelfs/leihaodong/anaconda3/envs/eval_xx/lib/python3.10/site-packages/torchmetrics/utilities/prints.py:43: UserWarning: Metric `InceptionScore` will save all extracted features in buffer. For large datasets this may lead to large memory footprint.
+  warnings.warn(*args, **kwargs)
+/mnt/petrelfs/leihaodong/anaconda3/envs/eval_xx/lib/python3.10/site-packages/open_clip/factory.py:450: UserWarning: QuickGELU mismatch between final model config (quick_gelu=False) and pretrained tag 'openai' (quick_gelu=True).
+  warnings.warn(
+Using a slow image processor as `use_fast` is unset and a slow processor was saved with this model. `use_fast=True` will be the default behavior in v4.52, even if the model was saved with a slow processor. This will result in minor differences in outputs. You'll still be able to use a slow processor with `use_fast=False`.
+Loading PartiPrompts from /mnt/petrelfs/leihaodong/ICLR25/sjdtree/data/prompts/PartiPrompts.tsv for GPU 0, Node 0
+Number of unique prompts: 1632 | Max prompt length: 389
+
+  warnings.warn(f"Importing from {__name__} is deprecated, please import via timm.layers", FutureWarning)
+/mnt/petrelfs/leihaodong/ICLR25/eval/HPSv2/hpsv2/img_score.py:107: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead.
+  with torch.cuda.amp.autocast():
+
+INFO:absl:IS: (tensor(13.9641, device='cuda:0'), tensor(1.6822, device='cuda:0'))
+INFO:absl:CLIP_T: 0.2128829068272711
+INFO:absl:HPSv2: 0.2926989103618421
+INFO:absl:Aesthetic: 6.614537016233386
+Scores:
+IS: (tensor(13.9641, device='cuda:0'), tensor(1.6822, device='cuda:0'))
+CLIP_T: 0.2128829068272711
+HPSv2: 0.2926989103618421
+Aesthetic: 6.614537016233386

scripts/log/emu3/PP/slurm-6224920.out

@@ -0,0 +1,21 @@
+/mnt/petrelfs/leihaodong/anaconda3/envs/eval_xx/lib/python3.10/site-packages/torchmetrics/utilities/prints.py:43: UserWarning: Metric `InceptionScore` will save all extracted features in buffer. For large datasets this may lead to large memory footprint.
+  warnings.warn(*args, **kwargs)
+/mnt/petrelfs/leihaodong/anaconda3/envs/eval_xx/lib/python3.10/site-packages/open_clip/factory.py:450: UserWarning: QuickGELU mismatch between final model config (quick_gelu=False) and pretrained tag 'openai' (quick_gelu=True).
+  warnings.warn(
+Using a slow image processor as `use_fast` is unset and a slow processor was saved with this model. `use_fast=True` will be the default behavior in v4.52, even if the model was saved with a slow processor. This will result in minor differences in outputs. You'll still be able to use a slow processor with `use_fast=False`.
+Loading PartiPrompts from /mnt/petrelfs/leihaodong/ICLR25/sjdtree/data/prompts/PartiPrompts.tsv for GPU 0, Node 0
+Number of unique prompts: 1632 | Max prompt length: 389
+
+  warnings.warn(f"Importing from {__name__} is deprecated, please import via timm.layers", FutureWarning)
+/mnt/petrelfs/leihaodong/ICLR25/eval/HPSv2/hpsv2/img_score.py:107: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead.
+  with torch.cuda.amp.autocast():
+
+INFO:absl:IS: (tensor(14.5504, device='cuda:0'), tensor(1.4504, device='cuda:0'))
+INFO:absl:CLIP_T: 0.21511175778703154
+INFO:absl:HPSv2: 0.29233695056950937
+INFO:absl:Aesthetic: 6.610858074081278
+Scores:
+IS: (tensor(14.5504, device='cuda:0'), tensor(1.4504, device='cuda:0'))
+CLIP_T: 0.21511175778703154
+HPSv2: 0.29233695056950937
+Aesthetic: 6.610858074081278

scripts/log/emu3/PP/slurm-6224921.out

@@ -0,0 +1,21 @@
+/mnt/petrelfs/leihaodong/anaconda3/envs/eval_xx/lib/python3.10/site-packages/torchmetrics/utilities/prints.py:43: UserWarning: Metric `InceptionScore` will save all extracted features in buffer. For large datasets this may lead to large memory footprint.
+  warnings.warn(*args, **kwargs)
+/mnt/petrelfs/leihaodong/anaconda3/envs/eval_xx/lib/python3.10/site-packages/open_clip/factory.py:450: UserWarning: QuickGELU mismatch between final model config (quick_gelu=False) and pretrained tag 'openai' (quick_gelu=True).
+  warnings.warn(
+Using a slow image processor as `use_fast` is unset and a slow processor was saved with this model. `use_fast=True` will be the default behavior in v4.52, even if the model was saved with a slow processor. This will result in minor differences in outputs. You'll still be able to use a slow processor with `use_fast=False`.
+Loading PartiPrompts from /mnt/petrelfs/leihaodong/ICLR25/sjdtree/data/prompts/PartiPrompts.tsv for GPU 0, Node 0
+Number of unique prompts: 1632 | Max prompt length: 389
+
+  warnings.warn(f"Importing from {__name__} is deprecated, please import via timm.layers", FutureWarning)
+/mnt/petrelfs/leihaodong/ICLR25/eval/HPSv2/hpsv2/img_score.py:107: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead.
+  with torch.cuda.amp.autocast():
+
+INFO:absl:IS: (tensor(10.4139, device='cuda:0'), tensor(1.7621, device='cuda:0'))
+INFO:absl:CLIP_T: 0.25801803677504354
+INFO:absl:HPSv2: 0.27528296889414205
+INFO:absl:Aesthetic: 6.3358107563314405
+Scores:
+IS: (tensor(10.4139, device='cuda:0'), tensor(1.7621, device='cuda:0'))
+CLIP_T: 0.25801803677504354
+HPSv2: 0.27528296889414205
+Aesthetic: 6.3358107563314405

scripts/log/emu3/PP/slurm-6224922.out

@@ -0,0 +1,21 @@
+/mnt/petrelfs/leihaodong/anaconda3/envs/eval_xx/lib/python3.10/site-packages/torchmetrics/utilities/prints.py:43: UserWarning: Metric `InceptionScore` will save all extracted features in buffer. For large datasets this may lead to large memory footprint.
+  warnings.warn(*args, **kwargs)
+/mnt/petrelfs/leihaodong/anaconda3/envs/eval_xx/lib/python3.10/site-packages/open_clip/factory.py:450: UserWarning: QuickGELU mismatch between final model config (quick_gelu=False) and pretrained tag 'openai' (quick_gelu=True).
+  warnings.warn(
+Using a slow image processor as `use_fast` is unset and a slow processor was saved with this model. `use_fast=True` will be the default behavior in v4.52, even if the model was saved with a slow processor. This will result in minor differences in outputs. You'll still be able to use a slow processor with `use_fast=False`.
+Loading PartiPrompts from /mnt/petrelfs/leihaodong/ICLR25/sjdtree/data/prompts/PartiPrompts.tsv for GPU 0, Node 0
+Number of unique prompts: 1632 | Max prompt length: 389
+
+  warnings.warn(f"Importing from {__name__} is deprecated, please import via timm.layers", FutureWarning)
+/mnt/petrelfs/leihaodong/ICLR25/eval/HPSv2/hpsv2/img_score.py:107: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead.
+  with torch.cuda.amp.autocast():
+
+INFO:absl:IS: (tensor(9.1179, device='cuda:0'), tensor(1.1087, device='cuda:0'))
+INFO:absl:CLIP_T: 0.2371356459450908
+INFO:absl:HPSv2: 0.26797332763671877
+INFO:absl:Aesthetic: 6.121969189494848
+Scores:
+IS: (tensor(9.1179, device='cuda:0'), tensor(1.1087, device='cuda:0'))
+CLIP_T: 0.2371356459450908
+HPSv2: 0.26797332763671877
+Aesthetic: 6.121969189494848

scripts/log/emu3/PP/slurm-6224923.out

@@ -0,0 +1,21 @@
+/mnt/petrelfs/leihaodong/anaconda3/envs/eval_xx/lib/python3.10/site-packages/torchmetrics/utilities/prints.py:43: UserWarning: Metric `InceptionScore` will save all extracted features in buffer. For large datasets this may lead to large memory footprint.
+  warnings.warn(*args, **kwargs)
+/mnt/petrelfs/leihaodong/anaconda3/envs/eval_xx/lib/python3.10/site-packages/open_clip/factory.py:450: UserWarning: QuickGELU mismatch between final model config (quick_gelu=False) and pretrained tag 'openai' (quick_gelu=True).
+  warnings.warn(
+Using a slow image processor as `use_fast` is unset and a slow processor was saved with this model. `use_fast=True` will be the default behavior in v4.52, even if the model was saved with a slow processor. This will result in minor differences in outputs. You'll still be able to use a slow processor with `use_fast=False`.
+Loading PartiPrompts from /mnt/petrelfs/leihaodong/ICLR25/sjdtree/data/prompts/PartiPrompts.tsv for GPU 0, Node 0
+Number of unique prompts: 1632 | Max prompt length: 389
+
+  warnings.warn(f"Importing from {__name__} is deprecated, please import via timm.layers", FutureWarning)
+/mnt/petrelfs/leihaodong/ICLR25/eval/HPSv2/hpsv2/img_score.py:107: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead.
+  with torch.cuda.amp.autocast():
+
+INFO:absl:IS: (tensor(10.4369, device='cuda:0'), tensor(1.8989, device='cuda:0'))
+INFO:absl:CLIP_T: 0.2426145700125122
+INFO:absl:HPSv2: 0.268757468775699
+INFO:absl:Aesthetic: 6.344130627418819
+Scores:
+IS: (tensor(10.4369, device='cuda:0'), tensor(1.8989, device='cuda:0'))
+CLIP_T: 0.2426145700125122
+HPSv2: 0.268757468775699
+Aesthetic: 6.344130627418819

scripts/log/emu3/PP/slurm-6224924.out

@@ -0,0 +1,21 @@
+/mnt/petrelfs/leihaodong/anaconda3/envs/eval_xx/lib/python3.10/site-packages/torchmetrics/utilities/prints.py:43: UserWarning: Metric `InceptionScore` will save all extracted features in buffer. For large datasets this may lead to large memory footprint.
+  warnings.warn(*args, **kwargs)
+/mnt/petrelfs/leihaodong/anaconda3/envs/eval_xx/lib/python3.10/site-packages/open_clip/factory.py:450: UserWarning: QuickGELU mismatch between final model config (quick_gelu=False) and pretrained tag 'openai' (quick_gelu=True).
+  warnings.warn(
+Using a slow image processor as `use_fast` is unset and a slow processor was saved with this model. `use_fast=True` will be the default behavior in v4.52, even if the model was saved with a slow processor. This will result in minor differences in outputs. You'll still be able to use a slow processor with `use_fast=False`.
+Loading PartiPrompts from /mnt/petrelfs/leihaodong/ICLR25/sjdtree/data/prompts/PartiPrompts.tsv for GPU 0, Node 0
+Number of unique prompts: 1632 | Max prompt length: 389
+
+  warnings.warn(f"Importing from {__name__} is deprecated, please import via timm.layers", FutureWarning)
+/mnt/petrelfs/leihaodong/ICLR25/eval/HPSv2/hpsv2/img_score.py:107: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead.
+  with torch.cuda.amp.autocast():
+
+INFO:absl:IS: (tensor(14.7802, device='cuda:0'), tensor(1.8198, device='cuda:0'))
+INFO:absl:CLIP_T: 0.21445053784682183
+INFO:absl:HPSv2: 0.29132057388475835
+INFO:absl:Aesthetic: 6.598674116524622
+Scores:
+IS: (tensor(14.7802, device='cuda:0'), tensor(1.8198, device='cuda:0'))
+CLIP_T: 0.21445053784682183
+HPSv2: 0.29132057388475835
+Aesthetic: 6.598674116524622

scripts/log/emu3/T2I/slurm-6225299.out

@@ -0,0 +1,20 @@
+/mnt/petrelfs/leihaodong/anaconda3/envs/eval_xx/lib/python3.10/site-packages/torchmetrics/utilities/prints.py:43: UserWarning: Metric `InceptionScore` will save all extracted features in buffer. For large datasets this may lead to large memory footprint.
+  warnings.warn(*args, **kwargs)
+/mnt/petrelfs/leihaodong/anaconda3/envs/eval_xx/lib/python3.10/site-packages/open_clip/factory.py:450: UserWarning: QuickGELU mismatch between final model config (quick_gelu=False) and pretrained tag 'openai' (quick_gelu=True).
+  warnings.warn(
+Using a slow image processor as `use_fast` is unset and a slow processor was saved with this model. `use_fast=True` will be the default behavior in v4.52, even if the model was saved with a slow processor. This will result in minor differences in outputs. You'll still be able to use a slow processor with `use_fast=False`.
+Loading T2ICompBenchBench from /mnt/petrelfs/leihaodong/ICLR25/sjdtree/data/prompts/T2I-CompBench_val.json for GPU 0, Node 0
+
+  warnings.warn(f"Importing from {__name__} is deprecated, please import via timm.layers", FutureWarning)
+/mnt/petrelfs/leihaodong/ICLR25/eval/HPSv2/hpsv2/img_score.py:107: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead.
+  with torch.cuda.amp.autocast():
+
+INFO:absl:IS: (tensor(12.6427, device='cuda:0'), tensor(2.0444, device='cuda:0'))
+INFO:absl:CLIP_T: 0.22459603008375598
+INFO:absl:HPSv2: 0.28618738890955686
+INFO:absl:Aesthetic: 6.482009205237898
+Scores:
+IS: (tensor(12.6427, device='cuda:0'), tensor(2.0444, device='cuda:0'))
+CLIP_T: 0.22459603008375598
+HPSv2: 0.28618738890955686
+Aesthetic: 6.482009205237898

scripts/log/emu3/T2I/slurm-6225300.out

@@ -0,0 +1,20 @@
+/mnt/petrelfs/leihaodong/anaconda3/envs/eval_xx/lib/python3.10/site-packages/torchmetrics/utilities/prints.py:43: UserWarning: Metric `InceptionScore` will save all extracted features in buffer. For large datasets this may lead to large memory footprint.
+  warnings.warn(*args, **kwargs)
+/mnt/petrelfs/leihaodong/anaconda3/envs/eval_xx/lib/python3.10/site-packages/open_clip/factory.py:450: UserWarning: QuickGELU mismatch between final model config (quick_gelu=False) and pretrained tag 'openai' (quick_gelu=True).
+  warnings.warn(
+Using a slow image processor as `use_fast` is unset and a slow processor was saved with this model. `use_fast=True` will be the default behavior in v4.52, even if the model was saved with a slow processor. This will result in minor differences in outputs. You'll still be able to use a slow processor with `use_fast=False`.
+Loading T2ICompBenchBench from /mnt/petrelfs/leihaodong/ICLR25/sjdtree/data/prompts/T2I-CompBench_val.json for GPU 0, Node 0
+
+  warnings.warn(f"Importing from {__name__} is deprecated, please import via timm.layers", FutureWarning)
+/mnt/petrelfs/leihaodong/ICLR25/eval/HPSv2/hpsv2/img_score.py:107: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead.
+  with torch.cuda.amp.autocast():
+
+INFO:absl:IS: (tensor(15.4168, device='cuda:0'), tensor(2.0079, device='cuda:0'))
+INFO:absl:CLIP_T: 0.2170085100031744
+INFO:absl:HPSv2: 0.2852641428549459
+INFO:absl:Aesthetic: 6.5750199077636236
+Scores:
+IS: (tensor(15.4168, device='cuda:0'), tensor(2.0079, device='cuda:0'))
+CLIP_T: 0.2170085100031744
+HPSv2: 0.2852641428549459
+Aesthetic: 6.5750199077636236

scripts/log/emu3/T2I/slurm-6225301.out

@@ -0,0 +1,20 @@
+/mnt/petrelfs/leihaodong/anaconda3/envs/eval_xx/lib/python3.10/site-packages/torchmetrics/utilities/prints.py:43: UserWarning: Metric `InceptionScore` will save all extracted features in buffer. For large datasets this may lead to large memory footprint.
+  warnings.warn(*args, **kwargs)
+/mnt/petrelfs/leihaodong/anaconda3/envs/eval_xx/lib/python3.10/site-packages/open_clip/factory.py:450: UserWarning: QuickGELU mismatch between final model config (quick_gelu=False) and pretrained tag 'openai' (quick_gelu=True).
+  warnings.warn(
+Using a slow image processor as `use_fast` is unset and a slow processor was saved with this model. `use_fast=True` will be the default behavior in v4.52, even if the model was saved with a slow processor. This will result in minor differences in outputs. You'll still be able to use a slow processor with `use_fast=False`.
+Loading T2ICompBenchBench from /mnt/petrelfs/leihaodong/ICLR25/sjdtree/data/prompts/T2I-CompBench_val.json for GPU 0, Node 0
+
+  warnings.warn(f"Importing from {__name__} is deprecated, please import via timm.layers", FutureWarning)
+/mnt/petrelfs/leihaodong/ICLR25/eval/HPSv2/hpsv2/img_score.py:107: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead.
+  with torch.cuda.amp.autocast():
+
+INFO:absl:IS: (tensor(16.1048, device='cuda:0'), tensor(1.5941, device='cuda:0'))
+INFO:absl:CLIP_T: 0.22069272279019805
+INFO:absl:HPSv2: 0.2845417229942153
+INFO:absl:Aesthetic: 6.537994932120954
+Scores:
+IS: (tensor(16.1048, device='cuda:0'), tensor(1.5941, device='cuda:0'))
+CLIP_T: 0.22069272279019805
+HPSv2: 0.2845417229942153
+Aesthetic: 6.537994932120954

scripts/log/emu3/T2I/slurm-6225302.out

@@ -0,0 +1,20 @@
+/mnt/petrelfs/leihaodong/anaconda3/envs/eval_xx/lib/python3.10/site-packages/torchmetrics/utilities/prints.py:43: UserWarning: Metric `InceptionScore` will save all extracted features in buffer. For large datasets this may lead to large memory footprint.
+  warnings.warn(*args, **kwargs)
+/mnt/petrelfs/leihaodong/anaconda3/envs/eval_xx/lib/python3.10/site-packages/open_clip/factory.py:450: UserWarning: QuickGELU mismatch between final model config (quick_gelu=False) and pretrained tag 'openai' (quick_gelu=True).
+  warnings.warn(
+Using a slow image processor as `use_fast` is unset and a slow processor was saved with this model. `use_fast=True` will be the default behavior in v4.52, even if the model was saved with a slow processor. This will result in minor differences in outputs. You'll still be able to use a slow processor with `use_fast=False`.
+Loading T2ICompBenchBench from /mnt/petrelfs/leihaodong/ICLR25/sjdtree/data/prompts/T2I-CompBench_val.json for GPU 0, Node 0
+
+  warnings.warn(f"Importing from {__name__} is deprecated, please import via timm.layers", FutureWarning)
+/mnt/petrelfs/leihaodong/ICLR25/eval/HPSv2/hpsv2/img_score.py:107: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead.
+  with torch.cuda.amp.autocast():
+
+INFO:absl:IS: (tensor(9.2221, device='cuda:0'), tensor(1.8388, device='cuda:0'))
+INFO:absl:CLIP_T: 0.23495205142971282
+INFO:absl:HPSv2: 0.2865086244733146
+INFO:absl:Aesthetic: 6.499841072139668
+Scores:
+IS: (tensor(9.2221, device='cuda:0'), tensor(1.8388, device='cuda:0'))
+CLIP_T: 0.23495205142971282
+HPSv2: 0.2865086244733146
+Aesthetic: 6.499841072139668

scripts/log/emu3/T2I/slurm-6225303.out

@@ -0,0 +1,20 @@
+/mnt/petrelfs/leihaodong/anaconda3/envs/eval_xx/lib/python3.10/site-packages/torchmetrics/utilities/prints.py:43: UserWarning: Metric `InceptionScore` will save all extracted features in buffer. For large datasets this may lead to large memory footprint.
+  warnings.warn(*args, **kwargs)
+/mnt/petrelfs/leihaodong/anaconda3/envs/eval_xx/lib/python3.10/site-packages/open_clip/factory.py:450: UserWarning: QuickGELU mismatch between final model config (quick_gelu=False) and pretrained tag 'openai' (quick_gelu=True).
+  warnings.warn(
+Using a slow image processor as `use_fast` is unset and a slow processor was saved with this model. `use_fast=True` will be the default behavior in v4.52, even if the model was saved with a slow processor. This will result in minor differences in outputs. You'll still be able to use a slow processor with `use_fast=False`.
+Loading T2ICompBenchBench from /mnt/petrelfs/leihaodong/ICLR25/sjdtree/data/prompts/T2I-CompBench_val.json for GPU 0, Node 0
+
+  warnings.warn(f"Importing from {__name__} is deprecated, please import via timm.layers", FutureWarning)
+/mnt/petrelfs/leihaodong/ICLR25/eval/HPSv2/hpsv2/img_score.py:107: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead.
+  with torch.cuda.amp.autocast():
+
+INFO:absl:IS: (tensor(9.0747, device='cuda:0'), tensor(1.1286, device='cuda:0'))
+INFO:absl:CLIP_T: 0.2196346385104041
+INFO:absl:HPSv2: 0.26841844128024195
+INFO:absl:Aesthetic: 6.323781307281986
+Scores:
+IS: (tensor(9.0747, device='cuda:0'), tensor(1.1286, device='cuda:0'))
+CLIP_T: 0.2196346385104041
+HPSv2: 0.26841844128024195
+Aesthetic: 6.323781307281986

scripts/log/emu3/T2I/slurm-6225304.out

@@ -0,0 +1,20 @@
+/mnt/petrelfs/leihaodong/anaconda3/envs/eval_xx/lib/python3.10/site-packages/torchmetrics/utilities/prints.py:43: UserWarning: Metric `InceptionScore` will save all extracted features in buffer. For large datasets this may lead to large memory footprint.
+  warnings.warn(*args, **kwargs)
+/mnt/petrelfs/leihaodong/anaconda3/envs/eval_xx/lib/python3.10/site-packages/open_clip/factory.py:450: UserWarning: QuickGELU mismatch between final model config (quick_gelu=False) and pretrained tag 'openai' (quick_gelu=True).
+  warnings.warn(
+Using a slow image processor as `use_fast` is unset and a slow processor was saved with this model. `use_fast=True` will be the default behavior in v4.52, even if the model was saved with a slow processor. This will result in minor differences in outputs. You'll still be able to use a slow processor with `use_fast=False`.
+Loading T2ICompBenchBench from /mnt/petrelfs/leihaodong/ICLR25/sjdtree/data/prompts/T2I-CompBench_val.json for GPU 0, Node 0
+
+  warnings.warn(f"Importing from {__name__} is deprecated, please import via timm.layers", FutureWarning)
+/mnt/petrelfs/leihaodong/ICLR25/eval/HPSv2/hpsv2/img_score.py:107: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead.
+  with torch.cuda.amp.autocast():
+
+INFO:absl:IS: (tensor(9.9621, device='cuda:0'), tensor(0.9783, device='cuda:0'))
+INFO:absl:CLIP_T: 0.22185775332274985
+INFO:absl:HPSv2: 0.27371365906762296
+INFO:absl:Aesthetic: 6.499100695281732
+Scores:
+IS: (tensor(9.9621, device='cuda:0'), tensor(0.9783, device='cuda:0'))
+CLIP_T: 0.22185775332274985
+HPSv2: 0.27371365906762296
+Aesthetic: 6.499100695281732

scripts/log/emu3/T2I/slurm-6225305.out

@@ -0,0 +1,20 @@
+/mnt/petrelfs/leihaodong/anaconda3/envs/eval_xx/lib/python3.10/site-packages/torchmetrics/utilities/prints.py:43: UserWarning: Metric `InceptionScore` will save all extracted features in buffer. For large datasets this may lead to large memory footprint.
+  warnings.warn(*args, **kwargs)
+/mnt/petrelfs/leihaodong/anaconda3/envs/eval_xx/lib/python3.10/site-packages/open_clip/factory.py:450: UserWarning: QuickGELU mismatch between final model config (quick_gelu=False) and pretrained tag 'openai' (quick_gelu=True).
+  warnings.warn(
+Using a slow image processor as `use_fast` is unset and a slow processor was saved with this model. `use_fast=True` will be the default behavior in v4.52, even if the model was saved with a slow processor. This will result in minor differences in outputs. You'll still be able to use a slow processor with `use_fast=False`.
+Loading T2ICompBenchBench from /mnt/petrelfs/leihaodong/ICLR25/sjdtree/data/prompts/T2I-CompBench_val.json for GPU 0, Node 0
+
+  warnings.warn(f"Importing from {__name__} is deprecated, please import via timm.layers", FutureWarning)
+/mnt/petrelfs/leihaodong/ICLR25/eval/HPSv2/hpsv2/img_score.py:107: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead.
+  with torch.cuda.amp.autocast():
+
+INFO:absl:IS: (tensor(16.2280, device='cuda:0'), tensor(2.2855, device='cuda:0'))
+INFO:absl:CLIP_T: 0.21824728426104584
+INFO:absl:HPSv2: 0.2856275140131273
+INFO:absl:Aesthetic: 6.53763605670858
+Scores:
+IS: (tensor(16.2280, device='cuda:0'), tensor(2.2855, device='cuda:0'))
+CLIP_T: 0.21824728426104584
+HPSv2: 0.2856275140131273
+Aesthetic: 6.53763605670858

scripts/lumina_gpt/MSCOCO/base.sh

@@ -0,0 +1,24 @@
+#!/bin/bash
+cd /mnt/petrelfs/leihaodong/ICLR25/sjdtree
+tokenizer_path=/mnt/petrelfs/leihaodong/local_model/chameleon/tokenizer
+isp='random'
+method='speculative_jacobi'
+num_init_new_token=1
+benchmark_way='order'
+prompt="MSCOCO2017Val"
+num_images=4000
+slice='0-4000'
+output_path=/mnt/petrelfs/leihaodong/ICLR25/exp/Lumina-mGPT-7B-768_${prompt}/base
+mkdir -p ${output_path}
+
+# nohup 
+python /mnt/petrelfs/leihaodong/ICLR25/sjdtree/main.py \
+  --tokenizer_path=$tokenizer_path \
+  --output_path=$output_path \
+  --isp=$isp \
+  --method=$method \
+  --num_init_new_token=$num_init_new_token \
+  --benchmark_way=$benchmark_way \
+  --prompt=$prompt \
+  --num_images=$num_images \
+  --slice=$slice > ${output_path}.log 2>&1

scripts/lumina_gpt/MSCOCO/caccl+sjd++.sh

@@ -0,0 +1,58 @@
+#!/bin/bash
+#SBATCH -N 1                               # 节点数（可改成多节点）
+#SBATCH --job-name=MSCOCO2017Val              # 可选：任务名，便于查看
+
+# ==================== 环境变量 ====================
+export PARTITION=DataFrontier_Knowledge
+export CFSCTL=/mnt/petrelfs/leihaodong/app/cfs/bin/cfsctl
+export NNODES=$SLURM_NNODES                    # 实际分配到的节点数
+export MASTER_ADDR=$(scontrol show hostname $SLURM_NODELIST | head -n1)
+echo "Job started at $(date)"
+echo "Nodes: $SLURM_NODELIST"
+echo "Master node: $MASTER_ADDR"
+echo "Total nodes: $NNODES"
+# ==================== CFS 预加载（启动 + preload + 停止） ====================
+echo "Starting CFS preload..."
+# 取主节点（和你们原来脚本保持一致）
+MAIN_SERVER=$(scontrol show hostname $SLURM_NODELIST | head -n1)
+# 停止可能残留的旧 CFS 实例
+$CFSCTL -p $PARTITION -n $NNODES -X $MAIN_SERVER stop || true
+# 启动 CFS
+$CFSCTL -p $PARTITION -n $NNODES -X $MAIN_SERVER start
+# 开始预加载（这一步会卡住直到所有节点 preload 完成）
+$CFSCTL -p $PARTITION -n $NNODES -X $MAIN_SERVER -a preload
+# ==================== 真正要跑的程序 ====================
+cd /mnt/petrelfs/leihaodong/ICLR25/sjdtree
+isp='random'
+method='caccl'
+num_init_new_token=96
+benchmark_way='order'
+prompt="MSCOCO2017Val"
+num_images=4000
+slice='0-4000'
+output_path=/mnt/petrelfs/leihaodong/ICLR25/exp/Lumina-mGPT-7B-768/${prompt}/caccl
+tokenizer_path=/mnt/petrelfs/leihaodong/local_model/chameleon/tokenizer
+model_path=/nvme/leihaodong/mnt/local_model/models--Alpha-VLLM--Lumina-mGPT-7B-768/snapshots/755e7e40530a8edf4eea6078cbc484d2bfda091f
+mkdir -p ${output_path}
+
+srun \
+ -N 1 \
+ --ntasks-per-node 1 \
+ python main.py \
+  --tokenizer_path=$tokenizer_path \
+  --output_path=$output_path \
+  --model_path=$model_path \
+  --isp=$isp \
+  --method=$method \
+  --num_init_new_token=$num_init_new_token \
+  --benchmark_way=$benchmark_way \
+  --prompt=$prompt \
+  --num_images=$num_images \
+  --sjd_pp_threshold=0.5 \
+  --slice=$slice \
+  --target_size=768 > ${output_path}.log 2>&1
+
+# ==================== 清理 CFS ====================
+echo "Python script finished, stopping CFS..."
+$CFSCTL -p $PARTITION -n $NNODES -X $MAIN_SERVER stop
+echo "Job finished at $(date)"

scripts/lumina_gpt/MSCOCO/ms_gsd.sh

@@ -0,0 +1,31 @@
+#!/bin/sh
+cd /mnt/petrelfs/leihaodong/ICLR25/sjdtree
+isp='random'
+method='group_speculative_jacobi'
+num_init_new_token=16
+benchmark_way='order'
+prompt="MSCOCO2017Val"
+num_images=4000
+slice='0-4000'
+output_path=/mnt/petrelfs/leihaodong/ICLR25/exp/Lumina-mGPT-7B-768_${prompt}/ms_gsd
+tokenizer_path=/mnt/petrelfs/leihaodong/local_model/chameleon/tokenizer
+model_path=Alpha-VLLM/Lumina-mGPT-7B-768
+mkdir -p ${output_path}
+
+srun \
+ -N 1 \
+ --ntasks-per-node 1 \
+ python /mnt/petrelfs/leihaodong/ICLR25/sjdtree/main.py \
+  --tokenizer_path=$tokenizer_path \
+  --output_path=$output_path \
+  --model_path=$model_path \
+  --isp=$isp \
+  --method=$method \
+  --num_init_new_token=$num_init_new_token \
+  --benchmark_way=$benchmark_way \
+  --prompt=$prompt \
+  --num_images=$num_images \
+  --slice=$slice \
+  --static_tree \
+  --target_size=768 \
+  --tree_choices=Grouped_Tree_6 > ${output_path}.log 2>&1

scripts/lumina_gpt/MSCOCO/ms_lantern.sh

@@ -0,0 +1,31 @@
+#!/bin/sh
+cd /mnt/petrelfs/leihaodong/ICLR25/sjdtree
+isp='random'
+method='lantern'
+num_init_new_token=16
+benchmark_way='order'
+prompt="MSCOCO2017Val"
+num_images=4000
+slice='0-4000'
+output_path=/mnt/petrelfs/leihaodong/ICLR25/exp/Lumina-mGPT-7B-768_${prompt}/ms_lantern
+tokenizer_path=/mnt/petrelfs/leihaodong/local_model/chameleon/tokenizer
+model_path=Alpha-VLLM/Lumina-mGPT-7B-768
+mkdir -p ${output_path}
+
+srun \
+ -N 1 \
+ --ntasks-per-node 1 \
+ python /mnt/petrelfs/leihaodong/ICLR25/sjdtree/main.py \
+  --tokenizer_path=$tokenizer_path \
+  --output_path=$output_path \
+  --model_path=$model_path \
+  --isp=$isp \
+  --method=$method \
+  --num_init_new_token=$num_init_new_token \
+  --benchmark_way=$benchmark_way \
+  --prompt=$prompt \
+  --num_images=$num_images \
+  --slice=$slice \
+  --static_tree \
+  --target_size=768 \
+  --tree_choices=Grouped_Tree_6 > ${output_path}.log 2>&1

scripts/lumina_gpt/MSCOCO/ms_sjd.sh

@@ -0,0 +1,31 @@
+#!/bin/sh
+cd /mnt/petrelfs/leihaodong/ICLR25/sjdtree
+isp='random'
+method='speculative_jacobi'
+num_init_new_token=16
+benchmark_way='order'
+prompt="MSCOCO2017Val"
+num_images=4000
+slice='0-4000'
+output_path=/mnt/petrelfs/leihaodong/ICLR25/exp/Lumina-mGPT-7B-768_${prompt}/ms_sjd
+tokenizer_path=/mnt/petrelfs/leihaodong/local_model/chameleon/tokenizer
+model_path=Alpha-VLLM/Lumina-mGPT-7B-768
+mkdir -p ${output_path}
+
+srun \
+ -N 1 \
+ --ntasks-per-node 1 \
+ python /mnt/petrelfs/leihaodong/ICLR25/sjdtree/main.py \
+  --tokenizer_path=$tokenizer_path \
+  --output_path=$output_path \
+  --model_path=$model_path \
+  --isp=$isp \
+  --method=$method \
+  --num_init_new_token=$num_init_new_token \
+  --benchmark_way=$benchmark_way \
+  --prompt=$prompt \
+  --num_images=$num_images \
+  --slice=$slice \
+  --static_tree \
+  --target_size=768 \
+  --tree_choices=Grouped_Tree_6 > ${output_path}.log 2>&1

scripts/lumina_gpt/MSCOCO/mscr_gsd.sh

@@ -0,0 +1,33 @@
+#!/bin/sh
+cd /mnt/petrelfs/leihaodong/ICLR25/sjdtree
+isp='random'
+method='group_speculative_jacobi_groupsum'
+num_init_new_token=16
+benchmark_way='order'
+prompt="MSCOCO2017Val"
+num_images=4000
+slice='0-4000'
+groupsum_delta=0.008
+output_path=/mnt/petrelfs/leihaodong/ICLR25/exp/Lumina-mGPT-7B-768_${prompt}/mscr_gsd
+tokenizer_path=/mnt/petrelfs/leihaodong/local_model/chameleon/tokenizer
+model_path=Alpha-VLLM/Lumina-mGPT-7B-768
+mkdir -p ${output_path}
+
+srun \
+ -N 1 \
+ --ntasks-per-node 1 \
+ python /mnt/petrelfs/leihaodong/ICLR25/sjdtree/main.py \
+  --tokenizer_path=$tokenizer_path \
+  --output_path=$output_path \
+  --model_path=$model_path \
+  --isp=$isp \
+  --method=$method \
+  --num_init_new_token=$num_init_new_token \
+  --benchmark_way=$benchmark_way \
+  --prompt=$prompt \
+  --num_images=$num_images \
+  --slice=$slice \
+  --static_tree \
+  --groupsum_delta=$groupsum_delta \
+  --target_size=768 \
+  --tree_choices=Grouped_Tree_6 > ${output_path}.log 2>&1

scripts/lumina_gpt/MSCOCO/mscr_sjd.sh

@@ -0,0 +1,33 @@
+#!/bin/sh
+cd /mnt/petrelfs/leihaodong/ICLR25/sjdtree
+isp='random'
+method='speculative_jacobi_groupsum'
+num_init_new_token=16
+benchmark_way='order'
+prompt="MSCOCO2017Val"
+num_images=4000
+slice='0-4000'
+groupsum_delta=0.05
+output_path=/mnt/petrelfs/leihaodong/ICLR25/exp/Lumina-mGPT-7B-768_${prompt}/mscr_sjd_${groupsum_delta}
+tokenizer_path=/mnt/petrelfs/leihaodong/local_model/chameleon/tokenizer
+model_path=Alpha-VLLM/Lumina-mGPT-7B-768
+mkdir -p ${output_path}
+
+srun \
+ -N 1 \
+ --ntasks-per-node 1 \
+ python /mnt/petrelfs/leihaodong/ICLR25/sjdtree/main.py \
+  --tokenizer_path=$tokenizer_path \
+  --output_path=$output_path \
+  --model_path=$model_path \
+  --isp=$isp \
+  --method=$method \
+  --num_init_new_token=$num_init_new_token \
+  --benchmark_way=$benchmark_way \
+  --prompt=$prompt \
+  --num_images=$num_images \
+  --slice=$slice \
+  --static_tree \
+  --groupsum_delta=$groupsum_delta \
+  --target_size=768 \
+  --tree_choices=Grouped_Tree_6 > ${output_path}.log 2>&1

scripts/lumina_gpt/MSCOCO/sjd++.sh

@@ -0,0 +1,57 @@
+#!/bin/bash
+#SBATCH -N 1                               # 节点数（可改成多节点）
+#SBATCH --job-name=MSCOCO2017Val              # 可选：任务名，便于查看
+
+# ==================== 环境变量 ====================
+export PARTITION=DataFrontier_Knowledge
+export CFSCTL=/mnt/petrelfs/leihaodong/app/cfs/bin/cfsctl
+export NNODES=$SLURM_NNODES                    # 实际分配到的节点数
+export MASTER_ADDR=$(scontrol show hostname $SLURM_NODELIST | head -n1)
+echo "Job started at $(date)"
+echo "Nodes: $SLURM_NODELIST"
+echo "Master node: $MASTER_ADDR"
+echo "Total nodes: $NNODES"
+# ==================== CFS 预加载（启动 + preload + 停止） ====================
+echo "Starting CFS preload..."
+# 取主节点（和你们原来脚本保持一致）
+MAIN_SERVER=$(scontrol show hostname $SLURM_NODELIST | head -n1)
+# 停止可能残留的旧 CFS 实例
+$CFSCTL -p $PARTITION -n $NNODES -X $MAIN_SERVER stop || true
+# 启动 CFS
+$CFSCTL -p $PARTITION -n $NNODES -X $MAIN_SERVER start
+# 开始预加载（这一步会卡住直到所有节点 preload 完成）
+$CFSCTL -p $PARTITION -n $NNODES -X $MAIN_SERVER -a preload
+# ==================== 真正要跑的程序 ====================
+cd /mnt/petrelfs/leihaodong/ICLR25/sjdtree
+isp='random'
+method='sjd++'
+num_init_new_token=96
+benchmark_way='order'
+prompt="MSCOCO2017Val"
+num_images=4000
+slice='0-4000'
+output_path=/mnt/petrelfs/leihaodong/ICLR25/exp/Lumina-mGPT-7B-768/${prompt}/sjd++
+tokenizer_path=/mnt/petrelfs/leihaodong/local_model/chameleon/tokenizer
+model_path=/nvme/leihaodong/mnt/local_model/models--Alpha-VLLM--Lumina-mGPT-7B-768/snapshots/755e7e40530a8edf4eea6078cbc484d2bfda091f
+mkdir -p ${output_path}
+
+srun \
+ -N 1 \
+ --ntasks-per-node 1 \
+ python main.py \
+  --tokenizer_path=$tokenizer_path \
+  --output_path=$output_path \
+  --model_path=$model_path \
+  --isp=$isp \
+  --method=$method \
+  --num_init_new_token=$num_init_new_token \
+  --benchmark_way=$benchmark_way \
+  --prompt=$prompt \
+  --num_images=$num_images \
+  --slice=$slice \
+  --target_size=768 > ${output_path}.log 2>&1
+
+# ==================== 清理 CFS ====================
+echo "Python script finished, stopping CFS..."
+$CFSCTL -p $PARTITION -n $NNODES -X $MAIN_SERVER stop
+echo "Job finished at $(date)"