Add files using upload-large-folder tool

SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/clusters/image_D_labels_0.1_SFR-Embedding-Code-2B_R_dinov2-large.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9ac79d504529ef05454e08ec93f7dd95aba39581c22544f7560587f7796a2a4a
+size 12066688

SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/0ViT-B-16-cc3m-laclip-mix-inter-010-r19/benchmark_caltech101_epoch_40.pt_ViT-B-16_en_zeroshot_classification.json

@@ -0,0 +1 @@
+{"dataset": "caltech101", "model": "ViT-B-16", "pretrained": "/mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R_0.95_1.0_dinov2-large_0.0_0.05_rouge_0.7_top_4/logs/ViT-B-16-cc3m-laclip-mix-inter-010-r19/checkpoints/epoch_40.pt", "task": "zeroshot_classification", "metrics": {"acc1": 0.4881889763779528, "acc5": 0.7120516185476815, "mean_per_class_recall": 0.4222189537052795}, "language": "en"}

SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/0ViT-B-16-cc3m-laclip-mix-inter-010-r19/benchmark_cars_epoch_40.pt_ViT-B-16_en_zeroshot_classification.json

@@ -0,0 +1 @@
+{"dataset": "cars", "model": "ViT-B-16", "pretrained": "/mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R_0.95_1.0_dinov2-large_0.0_0.05_rouge_0.7_top_4/logs/ViT-B-16-cc3m-laclip-mix-inter-010-r19/checkpoints/epoch_40.pt", "task": "zeroshot_classification", "metrics": {"acc1": 0.008954110185300335, "acc5": 0.046884715831364256, "mean_per_class_recall": 0.00921639484283787}, "language": "en"}

SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/0ViT-B-16-cc3m-laclip-mix-inter-010-r19/benchmark_cifar100_epoch_40.pt_ViT-B-16_en_zeroshot_classification.json

@@ -0,0 +1 @@
+{"dataset": "cifar100", "model": "ViT-B-16", "pretrained": "/mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R_0.95_1.0_dinov2-large_0.0_0.05_rouge_0.7_top_4/logs/ViT-B-16-cc3m-laclip-mix-inter-010-r19/checkpoints/epoch_40.pt", "task": "zeroshot_classification", "metrics": {"acc1": 0.1927, "acc5": 0.4303, "mean_per_class_recall": 0.19220000000000004}, "language": "en"}

SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/0ViT-B-16-cc3m-laclip-mix-inter-010-r19/benchmark_cifar10_epoch_40.pt_ViT-B-16_en_zeroshot_classification.json

@@ -0,0 +1 @@
+{"dataset": "cifar10", "model": "ViT-B-16", "pretrained": "/mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R_0.95_1.0_dinov2-large_0.0_0.05_rouge_0.7_top_4/logs/ViT-B-16-cc3m-laclip-mix-inter-010-r19/checkpoints/epoch_40.pt", "task": "zeroshot_classification", "metrics": {"acc1": 0.4882, "acc5": 0.9177, "mean_per_class_recall": 0.4883}, "language": "en"}

SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/0ViT-B-16-cc3m-laclip-mix-inter-010-r19/benchmark_country211_epoch_40.pt_ViT-B-16_en_zeroshot_classification.json

@@ -0,0 +1 @@
+{"dataset": "country211", "model": "ViT-B-16", "pretrained": "/mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R_0.95_1.0_dinov2-large_0.0_0.05_rouge_0.7_top_4/logs/ViT-B-16-cc3m-laclip-mix-inter-010-r19/checkpoints/epoch_40.pt", "task": "zeroshot_classification", "metrics": {"acc1": 0.006066350710900474, "acc5": 0.02772511848341232, "mean_per_class_recall": 0.0060663507109004755}, "language": "en"}

SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/0ViT-B-16-cc3m-laclip-mix-inter-010-r19/benchmark_dtd_epoch_40.pt_ViT-B-16_en_zeroshot_classification.json

@@ -0,0 +1 @@
+{"dataset": "dtd", "model": "ViT-B-16", "pretrained": "/mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R_0.95_1.0_dinov2-large_0.0_0.05_rouge_0.7_top_4/logs/ViT-B-16-cc3m-laclip-mix-inter-010-r19/checkpoints/epoch_40.pt", "task": "zeroshot_classification", "metrics": {"acc1": 0.09148936170212765, "acc5": 0.2553191489361702, "mean_per_class_recall": 0.09042553191489362}, "language": "en"}

SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/0ViT-B-16-cc3m-laclip-mix-inter-010-r19/benchmark_eurosat_epoch_40.pt_ViT-B-16_en_zeroshot_classification.json

@@ -0,0 +1 @@
+{"dataset": "eurosat", "model": "ViT-B-16", "pretrained": "/mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R_0.95_1.0_dinov2-large_0.0_0.05_rouge_0.7_top_4/logs/ViT-B-16-cc3m-laclip-mix-inter-010-r19/checkpoints/epoch_40.pt", "task": "zeroshot_classification", "metrics": {"acc1": 0.12381481481481481, "acc5": 0.6881481481481482, "mean_per_class_recall": 0.1265}, "language": "en"}

SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/0ViT-B-16-cc3m-laclip-mix-inter-010-r19/benchmark_fgvc_aircraft_epoch_40.pt_ViT-B-16_en_zeroshot_classification.json

@@ -0,0 +1 @@
+{"dataset": "fgvc_aircraft", "model": "ViT-B-16", "pretrained": "/mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R_0.95_1.0_dinov2-large_0.0_0.05_rouge_0.7_top_4/logs/ViT-B-16-cc3m-laclip-mix-inter-010-r19/checkpoints/epoch_40.pt", "task": "zeroshot_classification", "metrics": {"acc1": 0.010801080108010801, "acc5": 0.05910591059105911, "mean_per_class_recall": 0.010739750445632799}, "language": "en"}

SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/0ViT-B-16-cc3m-laclip-mix-inter-010-r19/benchmark_flickr30k_epoch_40.pt_ViT-B-16_en_zeroshot_retrieval.json

@@ -0,0 +1 @@
+{"dataset": "flickr30k", "model": "ViT-B-16", "pretrained": "/mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R_0.95_1.0_dinov2-large_0.0_0.05_rouge_0.7_top_4/logs/ViT-B-16-cc3m-laclip-mix-inter-010-r19/checkpoints/epoch_40.pt", "task": "zeroshot_retrieval", "metrics": {"image_retrieval_recall@1": 0.13680000603199005, "text_retrieval_recall@1": 0.2370000034570694, "image_retrieval_recall@5": 0.3118000030517578, "text_retrieval_recall@5": 0.4779999852180481, "image_retrieval_recall@10": 0.4052000045776367, "text_retrieval_recall@10": 0.597000002861023}, "language": "en"}

SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/0ViT-B-16-cc3m-laclip-mix-inter-010-r19/benchmark_flowers_epoch_40.pt_ViT-B-16_en_zeroshot_classification.json

@@ -0,0 +1 @@
+{"dataset": "flowers", "model": "ViT-B-16", "pretrained": "/mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R_0.95_1.0_dinov2-large_0.0_0.05_rouge_0.7_top_4/logs/ViT-B-16-cc3m-laclip-mix-inter-010-r19/checkpoints/epoch_40.pt", "task": "zeroshot_classification", "metrics": {"acc1": 0.11790535046349, "acc5": 0.24735729386892177, "mean_per_class_recall": 0.11544495618699056}, "language": "en"}

SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/0ViT-B-16-cc3m-laclip-mix-inter-010-r19/benchmark_food101_epoch_40.pt_ViT-B-16_en_zeroshot_classification.json

@@ -0,0 +1 @@
+{"dataset": "food101", "model": "ViT-B-16", "pretrained": "/mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R_0.95_1.0_dinov2-large_0.0_0.05_rouge_0.7_top_4/logs/ViT-B-16-cc3m-laclip-mix-inter-010-r19/checkpoints/epoch_40.pt", "task": "zeroshot_classification", "metrics": {"acc1": 0.08982178217821782, "acc5": 0.23893069306930692, "mean_per_class_recall": 0.08998019801980198}, "language": "en"}

SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/0ViT-B-16-cc3m-laclip-mix-inter-010-r19/benchmark_gtsrb_epoch_40.pt_ViT-B-16_en_zeroshot_classification.json

@@ -0,0 +1 @@
+{"dataset": "gtsrb", "model": "ViT-B-16", "pretrained": "/mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R_0.95_1.0_dinov2-large_0.0_0.05_rouge_0.7_top_4/logs/ViT-B-16-cc3m-laclip-mix-inter-010-r19/checkpoints/epoch_40.pt", "task": "zeroshot_classification", "metrics": {"acc1": 0.04267616785431512, "acc5": 0.23024544734758512, "mean_per_class_recall": 0.04345437767303288}, "language": "en"}

SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/0ViT-B-16-cc3m-laclip-mix-inter-010-r19/benchmark_imagenet1k_epoch_40.pt_ViT-B-16_en_zeroshot_classification.json

@@ -0,0 +1 @@
+{"dataset": "imagenet1k", "model": "ViT-B-16", "pretrained": "/mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R_0.95_1.0_dinov2-large_0.0_0.05_rouge_0.7_top_4/logs/ViT-B-16-cc3m-laclip-mix-inter-010-r19/checkpoints/epoch_40.pt", "task": "zeroshot_classification", "metrics": {"acc1": 0.14988, "acc5": 0.30318, "mean_per_class_recall": 0.14981999999999998}, "language": "en"}

SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/0ViT-B-16-cc3m-laclip-mix-inter-010-r19/benchmark_mscoco_captions_epoch_40.pt_ViT-B-16_en_zeroshot_retrieval.json

@@ -0,0 +1 @@
+{"dataset": "mscoco_captions", "model": "ViT-B-16", "pretrained": "/mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R_0.95_1.0_dinov2-large_0.0_0.05_rouge_0.7_top_4/logs/ViT-B-16-cc3m-laclip-mix-inter-010-r19/checkpoints/epoch_40.pt", "task": "zeroshot_retrieval", "metrics": {"image_retrieval_recall@1": 0.06345462054014206, "text_retrieval_recall@1": 0.11240000277757645, "image_retrieval_recall@5": 0.17504997551441193, "text_retrieval_recall@5": 0.2754000127315521, "image_retrieval_recall@10": 0.25185924768447876, "text_retrieval_recall@10": 0.3758000135421753}, "language": "en"}

SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/0ViT-B-16-cc3m-laclip-mix-inter-010-r19/benchmark_pets_epoch_40.pt_ViT-B-16_en_zeroshot_classification.json

@@ -0,0 +1 @@
+{"dataset": "pets", "model": "ViT-B-16", "pretrained": "/mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R_0.95_1.0_dinov2-large_0.0_0.05_rouge_0.7_top_4/logs/ViT-B-16-cc3m-laclip-mix-inter-010-r19/checkpoints/epoch_40.pt", "task": "zeroshot_classification", "metrics": {"acc1": 0.11310983919324066, "acc5": 0.27773235213954756, "mean_per_class_recall": 0.11324831669133432}, "language": "en"}

SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/0ViT-B-16-cc3m-laclip-mix-inter-010-r19/benchmark_stl10_epoch_40.pt_ViT-B-16_en_zeroshot_classification.json

@@ -0,0 +1 @@
+{"dataset": "stl10", "model": "ViT-B-16", "pretrained": "/mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R_0.95_1.0_dinov2-large_0.0_0.05_rouge_0.7_top_4/logs/ViT-B-16-cc3m-laclip-mix-inter-010-r19/checkpoints/epoch_40.pt", "task": "zeroshot_classification", "metrics": {"acc1": 0.73375, "acc5": 0.973125, "mean_per_class_recall": 0.734125}, "language": "en"}

SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/0ViT-B-16-cc3m-laclip-mix-inter-010-r19/benchmark_sun397_epoch_40.pt_ViT-B-16_en_zeroshot_classification.json

@@ -0,0 +1 @@
+{"dataset": "sun397", "model": "ViT-B-16", "pretrained": "/mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R_0.95_1.0_dinov2-large_0.0_0.05_rouge_0.7_top_4/logs/ViT-B-16-cc3m-laclip-mix-inter-010-r19/checkpoints/epoch_40.pt", "task": "zeroshot_classification", "metrics": {"acc1": 0.2588962245066848, "acc5": 0.5254703275281829, "mean_per_class_recall": 0.22160657976490108}, "language": "en"}

SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/0ViT-B-16-cc3m-laclip-mix-inter-010-r19/benchmark_vtab_resisc45_epoch_40.pt_ViT-B-16_en_zeroshot_classification.json

@@ -0,0 +1 @@
+{"dataset": "vtab/resisc45", "model": "ViT-B-16", "pretrained": "/mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R_0.95_1.0_dinov2-large_0.0_0.05_rouge_0.7_top_4/logs/ViT-B-16-cc3m-laclip-mix-inter-010-r19/checkpoints/epoch_40.pt", "task": "zeroshot_classification", "metrics": {"acc1": 0.16253968253968254, "acc5": 0.42333333333333334, "mean_per_class_recall": 0.16592507845677368}, "language": "en"}

SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/0ViT-B-16-cc3m-laclip-mix-inter-010-r19/out.log

@@ -0,0 +1,281 @@
+2025-04-12,15:43:50 | INFO | Running in distributed mode with multiple processes. Device: cuda:0.Process (global: 0, local 0), total 4.
+2025-04-12,15:43:50 | INFO | Loaded ViT-B-16 model config.
+2025-04-12,15:43:51 | INFO | Model:
+2025-04-12,15:43:51 | INFO | CLIP(
+  (visual): VisionTransformer(
+    (conv1): Conv2d(3, 768, kernel_size=(16, 16), stride=(16, 16), bias=False)
+    (patch_dropout): Identity()
+    (ln_pre): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
+    (transformer): Transformer(
+      (resblocks): ModuleList(
+        (0-11): 12 x ResidualAttentionBlock(
+          (ln_1): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
+          (attn): MultiheadAttention(
+            (out_proj): NonDynamicallyQuantizableLinear(in_features=768, out_features=768, bias=True)
+          )
+          (ls_1): Identity()
+          (ln_2): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
+          (mlp): Sequential(
+            (c_fc): Linear(in_features=768, out_features=3072, bias=True)
+            (gelu): GELU(approximate='none')
+            (c_proj): Linear(in_features=3072, out_features=768, bias=True)
+          )
+          (ls_2): Identity()
+        )
+      )
+    )
+    (ln_post): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
+  )
+  (transformer): Transformer(
+    (resblocks): ModuleList(
+      (0-11): 12 x ResidualAttentionBlock(
+        (ln_1): LayerNorm((512,), eps=1e-05, elementwise_affine=True)
+        (attn): MultiheadAttention(
+          (out_proj): NonDynamicallyQuantizableLinear(in_features=512, out_features=512, bias=True)
+        )
+        (ls_1): Identity()
+        (ln_2): LayerNorm((512,), eps=1e-05, elementwise_affine=True)
+        (mlp): Sequential(
+          (c_fc): Linear(in_features=512, out_features=2048, bias=True)
+          (gelu): GELU(approximate='none')
+          (c_proj): Linear(in_features=2048, out_features=512, bias=True)
+        )
+        (ls_2): Identity()
+      )
+    )
+  )
+  (token_embedding): Embedding(49408, 512)
+  (ln_final): LayerNorm((512,), eps=1e-05, elementwise_affine=True)
+)
+2025-04-12,15:43:51 | INFO | Params:
+2025-04-12,15:43:51 | INFO |   accum_freq: 2
+2025-04-12,15:43:51 | INFO |   aug_cfg: {}
+2025-04-12,15:43:51 | INFO |   batch_size: 1024
+2025-04-12,15:43:51 | INFO |   beta1: 0.9
+2025-04-12,15:43:51 | INFO |   beta2: 0.98
+2025-04-12,15:43:51 | INFO |   cache_dir: None
+2025-04-12,15:43:51 | INFO |   caption_ratio: 0.1
+2025-04-12,15:43:51 | INFO |   checkpoint_path: /mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R_0.95_1.0_dinov2-large_0.0_0.05_rouge_0.7_top_4/logs/ViT-B-16-cc3m-laclip-mix-inter-010-r19/checkpoints
+2025-04-12,15:43:51 | INFO |   coca_caption_loss_weight: 2.0
+2025-04-12,15:43:51 | INFO |   coca_contrastive_loss_weight: 1.0
+2025-04-12,15:43:51 | INFO |   copy_codebase: False
+2025-04-12,15:43:51 | INFO |   csv_caption_key: title
+2025-04-12,15:43:51 | INFO |   csv_img_key: filepath
+2025-04-12,15:43:51 | INFO |   csv_separator: 	
+2025-04-12,15:43:51 | INFO |   dataset_resampled: False
+2025-04-12,15:43:51 | INFO |   dataset_type: synthetic
+2025-04-12,15:43:51 | INFO |   ddp_static_graph: False
+2025-04-12,15:43:51 | INFO |   debug: False
+2025-04-12,15:43:51 | INFO |   delete_previous_checkpoint: False
+2025-04-12,15:43:51 | INFO |   device: cuda:0
+2025-04-12,15:43:51 | INFO |   dist_backend: None
+2025-04-12,15:43:51 | INFO |   dist_url: None
+2025-04-12,15:43:51 | INFO |   distill: False
+2025-04-12,15:43:51 | INFO |   distill_model: None
+2025-04-12,15:43:51 | INFO |   distill_pretrained: None
+2025-04-12,15:43:51 | INFO |   distributed: True
+2025-04-12,15:43:51 | INFO |   epochs: 40
+2025-04-12,15:43:51 | INFO |   epochs_cooldown: None
+2025-04-12,15:43:51 | INFO |   eps: 1e-08
+2025-04-12,15:43:51 | INFO |   force_custom_text: False
+2025-04-12,15:43:51 | INFO |   force_image_size: None
+2025-04-12,15:43:51 | INFO |   force_patch_dropout: None
+2025-04-12,15:43:51 | INFO |   force_quick_gelu: False
+2025-04-12,15:43:51 | INFO |   gather_with_grad: True
+2025-04-12,15:43:51 | INFO |   grad_checkpointing: True
+2025-04-12,15:43:51 | INFO |   grad_clip_norm: None
+2025-04-12,15:43:51 | INFO |   horovod: False
+2025-04-12,15:43:51 | INFO |   image_interpolation: None
+2025-04-12,15:43:51 | INFO |   image_mean: None
+2025-04-12,15:43:51 | INFO |   image_resize_mode: None
+2025-04-12,15:43:51 | INFO |   image_std: None
+2025-04-12,15:43:51 | INFO |   imagenet_v2: None
+2025-04-12,15:43:51 | INFO |   imagenet_val: None
+2025-04-12,15:43:51 | INFO |   keep_func_name: 
+2025-04-12,15:43:51 | INFO |   local_loss: True
+2025-04-12,15:43:51 | INFO |   local_rank: 0
+2025-04-12,15:43:51 | INFO |   lock_image: False
+2025-04-12,15:43:51 | INFO |   lock_image_freeze_bn_stats: False
+2025-04-12,15:43:51 | INFO |   lock_image_unlocked_groups: 0
+2025-04-12,15:43:51 | INFO |   lock_text: False
+2025-04-12,15:43:51 | INFO |   lock_text_freeze_layer_norm: False
+2025-04-12,15:43:51 | INFO |   lock_text_unlocked_layers: 0
+2025-04-12,15:43:51 | INFO |   log_every_n_steps: 100
+2025-04-12,15:43:51 | INFO |   log_level: 20
+2025-04-12,15:43:51 | INFO |   log_local: False
+2025-04-12,15:43:51 | INFO |   log_path: /mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R_0.95_1.0_dinov2-large_0.0_0.05_rouge_0.7_top_4/logs/ViT-B-16-cc3m-laclip-mix-inter-010-r19/out.log
+2025-04-12,15:43:51 | INFO |   logs: /mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R_0.95_1.0_dinov2-large_0.0_0.05_rouge_0.7_top_4/logs
+2025-04-12,15:43:51 | INFO |   loss_dist_impl: None
+2025-04-12,15:43:51 | INFO |   lr: 0.001
+2025-04-12,15:43:51 | INFO |   lr_cooldown_end: 0.0
+2025-04-12,15:43:51 | INFO |   lr_cooldown_power: 1.0
+2025-04-12,15:43:51 | INFO |   lr_scheduler: cosine
+2025-04-12,15:43:51 | INFO |   map_func_name: use_low_inter_only
+2025-04-12,15:43:51 | INFO |   model: ViT-B-16
+2025-04-12,15:43:51 | INFO |   momentum: None
+2025-04-12,15:43:51 | INFO |   name: ViT-B-16-cc3m-laclip-mix-inter-010-r19
+2025-04-12,15:43:51 | INFO |   no_set_device_rank: False
+2025-04-12,15:43:51 | INFO |   opt: adamw
+2025-04-12,15:43:51 | INFO |   precision: amp
+2025-04-12,15:43:51 | INFO |   pretrained: 
+2025-04-12,15:43:51 | INFO |   pretrained_image: False
+2025-04-12,15:43:51 | INFO |   rank: 0
+2025-04-12,15:43:51 | INFO |   remote_sync: None
+2025-04-12,15:43:51 | INFO |   remote_sync_frequency: 300
+2025-04-12,15:43:51 | INFO |   remote_sync_protocol: s3
+2025-04-12,15:43:51 | INFO |   report_to: wandb
+2025-04-12,15:43:51 | INFO |   resume: /mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R_0.95_1.0_dinov2-large_0.0_0.05_rouge_0.7_top_4/logs/ViT-B-16-cc3m-laclip-mix-inter-010-r13/checkpoints/epoch_19.pt
+2025-04-12,15:43:51 | INFO |   save_frequency: 1
+2025-04-12,15:43:51 | INFO |   save_most_recent: False
+2025-04-12,15:43:51 | INFO |   seed: 0
+2025-04-12,15:43:51 | INFO |   siglip: False
+2025-04-12,15:43:51 | INFO |   skip_scheduler: False
+2025-04-12,15:43:51 | INFO |   tensorboard: False
+2025-04-12,15:43:51 | INFO |   tensorboard_path: 
+2025-04-12,15:43:51 | INFO |   torchcompile: False
+2025-04-12,15:43:51 | INFO |   torchscript: False
+2025-04-12,15:43:51 | INFO |   trace: False
+2025-04-12,15:43:51 | INFO |   train_data: /mnt/personal/zhudongy/cc3m-hgf-wds/{0000..0301}.tar
+2025-04-12,15:43:51 | INFO |   train_data_upsampling_factors: None
+2025-04-12,15:43:51 | INFO |   train_num_samples: 3016640
+2025-04-12,15:43:51 | INFO |   use_bn_sync: False
+2025-04-12,15:43:51 | INFO |   use_bnb_linear: None
+2025-04-12,15:43:51 | INFO |   val_data: None
+2025-04-12,15:43:51 | INFO |   val_frequency: 1
+2025-04-12,15:43:51 | INFO |   val_num_samples: None
+2025-04-12,15:43:51 | INFO |   wandb: True
+2025-04-12,15:43:51 | INFO |   wandb_notes: 
+2025-04-12,15:43:51 | INFO |   wandb_project_name: open-clip
+2025-04-12,15:43:51 | INFO |   warmup: 368
+2025-04-12,15:43:51 | INFO |   wd: 0.5
+2025-04-12,15:43:51 | INFO |   workers: 16
+2025-04-12,15:43:51 | INFO |   world_size: 4
+2025-04-12,15:43:51 | INFO |   zeroshot_frequency: 2
+2025-04-12,15:43:54 | INFO | Created AdamW (adamw) optimizer: lr: 0.001, betas: (0.9, 0.98), eps: 1e-08, weight_decay: 0.5, amsgrad: False, foreach: None, maximize: False, capturable: False, differentiable: False, fused: None
+2025-04-12,15:43:55 | INFO | => resuming checkpoint '/mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R_0.95_1.0_dinov2-large_0.0_0.05_rouge_0.7_top_4/logs/ViT-B-16-cc3m-laclip-mix-inter-010-r13/checkpoints/epoch_19.pt' (epoch 19)
+2025-04-12,15:46:31 | INFO | Start epoch 19
+2025-04-12,15:46:54 | INFO | Train Epoch: 19 [   8192/3080192 (0%)] Data (t): 18.084 Batch (t): 23.354, 350.774/s, 87.6934/s/gpu LR: 0.000560 Logit Scale: 54.396 Contrastive_loss: 1.1979 (1.1979) Loss: 1.1979 (1.1979)
+2025-04-12,15:54:33 | INFO | Train Epoch: 19 [ 827392/3080192 (27%)] Data (t): 0.460 Batch (t): 4.588, 1784.45/s, 446.112/s/gpu LR: 0.000549 Logit Scale: 55.313 Contrastive_loss: 1.4644 (1.3312) Loss: 1.4644 (1.3312)
+2025-04-12,16:02:12 | INFO | Train Epoch: 19 [1646592/3080192 (53%)] Data (t): 0.462 Batch (t): 4.595, 1786.67/s, 446.669/s/gpu LR: 0.000539 Logit Scale: 55.714 Contrastive_loss: 1.3151 (1.3258) Loss: 1.3151 (1.3258)
+2025-04-12,16:09:52 | INFO | Train Epoch: 19 [2465792/3080192 (80%)] Data (t): 0.467 Batch (t): 4.598, 1780.21/s, 445.053/s/gpu LR: 0.000528 Logit Scale: 55.816 Contrastive_loss: 1.2380 (1.3039) Loss: 1.2380 (1.3039)
+2025-04-12,16:15:37 | INFO | Train Epoch: 19 [3080192/3080192 (100%)] Data (t): 0.472 Batch (t): 4.606, 1783.63/s, 445.908/s/gpu LR: 0.000520 Logit Scale: 55.937 Contrastive_loss: 1.0830 (1.2597) Loss: 1.0830 (1.2597)
+2025-04-12,16:15:39 | INFO | Start epoch 20
+2025-04-12,16:16:00 | INFO | Train Epoch: 20 [   8192/3080192 (0%)] Data (t): 17.316 Batch (t): 21.512, 380.812/s, 95.2030/s/gpu LR: 0.000520 Logit Scale: 55.944 Contrastive_loss: 0.82071 (0.82071) Loss: 0.82071 (0.82071)
+2025-04-12,16:23:44 | INFO | Train Epoch: 20 [ 827392/3080192 (27%)] Data (t): 0.486 Batch (t): 4.632, 1772.86/s, 443.215/s/gpu LR: 0.000509 Logit Scale: 57.096 Contrastive_loss: 1.0051 (0.91291) Loss: 1.0051 (0.91291)
+2025-04-12,16:31:25 | INFO | Train Epoch: 20 [1646592/3080192 (53%)] Data (t): 0.476 Batch (t): 4.611, 1777.99/s, 444.498/s/gpu LR: 0.000498 Logit Scale: 57.470 Contrastive_loss: 1.2666 (1.0308) Loss: 1.2666 (1.0308)
+2025-04-12,16:39:06 | INFO | Train Epoch: 20 [2465792/3080192 (80%)] Data (t): 0.477 Batch (t): 4.613, 1777.20/s, 444.301/s/gpu LR: 0.000488 Logit Scale: 57.544 Contrastive_loss: 1.2081 (1.0751) Loss: 1.2081 (1.0751)
+2025-04-12,16:44:52 | INFO | Train Epoch: 20 [3080192/3080192 (100%)] Data (t): 0.481 Batch (t): 4.617, 1775.63/s, 443.906/s/gpu LR: 0.000480 Logit Scale: 57.738 Contrastive_loss: 0.96489 (1.0531) Loss: 0.96489 (1.0531)
+2025-04-12,16:44:54 | INFO | Start epoch 21
+2025-04-12,16:45:15 | INFO | Train Epoch: 21 [   8192/3080192 (0%)] Data (t): 17.309 Batch (t): 21.507, 380.897/s, 95.2242/s/gpu LR: 0.000479 Logit Scale: 57.744 Contrastive_loss: 0.66037 (0.66037) Loss: 0.66037 (0.66037)
+2025-04-12,16:52:59 | INFO | Train Epoch: 21 [ 827392/3080192 (27%)] Data (t): 0.493 Batch (t): 4.640, 1777.67/s, 444.418/s/gpu LR: 0.000469 Logit Scale: 58.678 Contrastive_loss: 1.0735 (0.86692) Loss: 1.0735 (0.86692)
+2025-04-12,17:00:41 | INFO | Train Epoch: 21 [1646592/3080192 (53%)] Data (t): 0.480 Batch (t): 4.616, 1774.96/s, 443.739/s/gpu LR: 0.000458 Logit Scale: 59.081 Contrastive_loss: 1.0338 (0.92254) Loss: 1.0338 (0.92254)
+2025-04-12,17:08:23 | INFO | Train Epoch: 21 [2465792/3080192 (80%)] Data (t): 0.481 Batch (t): 4.617, 1769.65/s, 442.414/s/gpu LR: 0.000447 Logit Scale: 59.303 Contrastive_loss: 1.0744 (0.96051) Loss: 1.0744 (0.96051)
+2025-04-12,17:14:09 | INFO | Train Epoch: 21 [3080192/3080192 (100%)] Data (t): 0.486 Batch (t): 4.624, 1772.05/s, 443.012/s/gpu LR: 0.000439 Logit Scale: 59.509 Contrastive_loss: 0.84948 (0.93830) Loss: 0.84948 (0.93830)
+2025-04-12,17:14:11 | INFO | Start epoch 22
+2025-04-12,17:14:32 | INFO | Train Epoch: 22 [   8192/3080192 (0%)] Data (t): 15.695 Batch (t): 20.957, 390.895/s, 97.7238/s/gpu LR: 0.000439 Logit Scale: 59.518 Contrastive_loss: 0.74400 (0.74400) Loss: 0.74400 (0.74400)
+2025-04-12,17:22:16 | INFO | Train Epoch: 22 [ 827392/3080192 (27%)] Data (t): 0.493 Batch (t): 4.641, 1769.69/s, 442.422/s/gpu LR: 0.000429 Logit Scale: 60.759 Contrastive_loss: 0.80650 (0.77525) Loss: 0.80650 (0.77525)
+2025-04-12,17:29:58 | INFO | Train Epoch: 22 [1646592/3080192 (53%)] Data (t): 0.482 Batch (t): 4.620, 1772.91/s, 443.227/s/gpu LR: 0.000418 Logit Scale: 61.141 Contrastive_loss: 0.98941 (0.84664) Loss: 0.98941 (0.84664)
+2025-04-12,17:37:40 | INFO | Train Epoch: 22 [2465792/3080192 (80%)] Data (t): 0.482 Batch (t): 4.621, 1776.46/s, 444.114/s/gpu LR: 0.000408 Logit Scale: 61.279 Contrastive_loss: 1.0364 (0.89408) Loss: 1.0364 (0.89408)
+2025-04-12,17:43:27 | INFO | Train Epoch: 22 [3080192/3080192 (100%)] Data (t): 0.488 Batch (t): 4.625, 1773.22/s, 443.305/s/gpu LR: 0.000400 Logit Scale: 61.486 Contrastive_loss: 0.85816 (0.88689) Loss: 0.85816 (0.88689)
+2025-04-12,17:43:28 | INFO | Start epoch 23
+2025-04-12,17:43:50 | INFO | Train Epoch: 23 [   8192/3080192 (0%)] Data (t): 17.715 Batch (t): 21.944, 373.307/s, 93.3267/s/gpu LR: 0.000400 Logit Scale: 61.490 Contrastive_loss: 0.81993 (0.81993) Loss: 0.81993 (0.81993)
+2025-04-12,17:51:34 | INFO | Train Epoch: 23 [ 827392/3080192 (27%)] Data (t): 0.496 Batch (t): 4.644, 1777.38/s, 444.346/s/gpu LR: 0.000389 Logit Scale: 62.490 Contrastive_loss: 0.85398 (0.83696) Loss: 0.85398 (0.83696)
+2025-04-12,17:59:17 | INFO | Train Epoch: 23 [1646592/3080192 (53%)] Data (t): 0.482 Batch (t): 4.620, 1771.82/s, 442.956/s/gpu LR: 0.000379 Logit Scale: 63.016 Contrastive_loss: 0.90926 (0.86106) Loss: 0.90926 (0.86106)
+2025-04-12,18:06:59 | INFO | Train Epoch: 23 [2465792/3080192 (80%)] Data (t): 0.485 Batch (t): 4.623, 1775.43/s, 443.859/s/gpu LR: 0.000368 Logit Scale: 63.236 Contrastive_loss: 0.81703 (0.85005) Loss: 0.81703 (0.85005)
+2025-04-12,18:12:46 | INFO | Train Epoch: 23 [3080192/3080192 (100%)] Data (t): 0.488 Batch (t): 4.625, 1774.43/s, 443.607/s/gpu LR: 0.000361 Logit Scale: 63.413 Contrastive_loss: 0.63382 (0.80680) Loss: 0.63382 (0.80680)
+2025-04-12,18:12:47 | INFO | Start epoch 24
+2025-04-12,18:13:09 | INFO | Train Epoch: 24 [   8192/3080192 (0%)] Data (t): 17.413 Batch (t): 21.628, 378.760/s, 94.6901/s/gpu LR: 0.000361 Logit Scale: 63.419 Contrastive_loss: 0.68919 (0.68919) Loss: 0.68919 (0.68919)
+2025-04-12,18:20:53 | INFO | Train Epoch: 24 [ 827392/3080192 (27%)] Data (t): 0.497 Batch (t): 4.645, 1771.21/s, 442.803/s/gpu LR: 0.000350 Logit Scale: 64.412 Contrastive_loss: 0.69677 (0.69298) Loss: 0.69677 (0.69298)
+2025-04-12,18:28:35 | INFO | Train Epoch: 24 [1646592/3080192 (53%)] Data (t): 0.482 Batch (t): 4.620, 1777.07/s, 444.267/s/gpu LR: 0.000340 Logit Scale: 64.917 Contrastive_loss: 0.83102 (0.73899) Loss: 0.83102 (0.73899)
+2025-04-12,18:36:18 | INFO | Train Epoch: 24 [2465792/3080192 (80%)] Data (t): 0.483 Batch (t): 4.621, 1775.57/s, 443.892/s/gpu LR: 0.000330 Logit Scale: 65.247 Contrastive_loss: 0.95238 (0.79234) Loss: 0.95238 (0.79234)
+2025-04-12,18:42:04 | INFO | Train Epoch: 24 [3080192/3080192 (100%)] Data (t): 0.489 Batch (t): 4.625, 1772.74/s, 443.185/s/gpu LR: 0.000322 Logit Scale: 65.503 Contrastive_loss: 0.62692 (0.75926) Loss: 0.62692 (0.75926)
+2025-04-12,18:42:06 | INFO | Start epoch 25
+2025-04-12,18:42:26 | INFO | Train Epoch: 25 [   8192/3080192 (0%)] Data (t): 15.808 Batch (t): 20.031, 408.965/s, 102.241/s/gpu LR: 0.000322 Logit Scale: 65.509 Contrastive_loss: 0.43343 (0.43343) Loss: 0.43343 (0.43343)
+2025-04-12,18:50:10 | INFO | Train Epoch: 25 [ 827392/3080192 (27%)] Data (t): 0.497 Batch (t): 4.644, 1771.92/s, 442.979/s/gpu LR: 0.000312 Logit Scale: 66.402 Contrastive_loss: 0.63256 (0.53299) Loss: 0.63256 (0.53299)
+2025-04-12,18:57:53 | INFO | Train Epoch: 25 [1646592/3080192 (53%)] Data (t): 0.485 Batch (t): 4.623, 1768.51/s, 442.126/s/gpu LR: 0.000303 Logit Scale: 66.900 Contrastive_loss: 0.67512 (0.58037) Loss: 0.67512 (0.58037)
+2025-04-12,19:05:35 | INFO | Train Epoch: 25 [2465792/3080192 (80%)] Data (t): 0.485 Batch (t): 4.623, 1771.08/s, 442.771/s/gpu LR: 0.000293 Logit Scale: 67.274 Contrastive_loss: 0.65914 (0.60006) Loss: 0.65914 (0.60006)
+2025-04-12,19:11:22 | INFO | Train Epoch: 25 [3080192/3080192 (100%)] Data (t): 0.488 Batch (t): 4.625, 1774.72/s, 443.679/s/gpu LR: 0.000285 Logit Scale: 67.623 Contrastive_loss: 0.61914 (0.60388) Loss: 0.61914 (0.60388)
+2025-04-12,19:11:23 | INFO | Start epoch 26
+2025-04-12,19:11:44 | INFO | Train Epoch: 26 [   8192/3080192 (0%)] Data (t): 16.774 Batch (t): 21.060, 388.985/s, 97.2463/s/gpu LR: 0.000285 Logit Scale: 67.627 Contrastive_loss: 0.37127 (0.37127) Loss: 0.37127 (0.37127)
+2025-04-12,19:19:29 | INFO | Train Epoch: 26 [ 827392/3080192 (27%)] Data (t): 0.501 Batch (t): 4.648, 1775.05/s, 443.763/s/gpu LR: 0.000276 Logit Scale: 68.558 Contrastive_loss: 0.56492 (0.46809) Loss: 0.56492 (0.46809)
+2025-04-12,19:27:11 | INFO | Train Epoch: 26 [1646592/3080192 (53%)] Data (t): 0.484 Batch (t): 4.622, 1773.63/s, 443.407/s/gpu LR: 0.000266 Logit Scale: 69.083 Contrastive_loss: 0.54506 (0.49375) Loss: 0.54506 (0.49375)
+2025-04-12,19:34:54 | INFO | Train Epoch: 26 [2465792/3080192 (80%)] Data (t): 0.486 Batch (t): 4.623, 1769.19/s, 442.298/s/gpu LR: 0.000257 Logit Scale: 69.391 Contrastive_loss: 0.63797 (0.52980) Loss: 0.63797 (0.52980)
+2025-04-12,19:40:41 | INFO | Train Epoch: 26 [3080192/3080192 (100%)] Data (t): 0.490 Batch (t): 4.626, 1773.01/s, 443.253/s/gpu LR: 0.000250 Logit Scale: 69.694 Contrastive_loss: 0.44939 (0.51372) Loss: 0.44939 (0.51372)
+2025-04-12,19:40:42 | INFO | Start epoch 27
+2025-04-12,19:41:03 | INFO | Train Epoch: 27 [   8192/3080192 (0%)] Data (t): 16.337 Batch (t): 20.520, 399.217/s, 99.8043/s/gpu LR: 0.000250 Logit Scale: 69.701 Contrastive_loss: 0.37365 (0.37365) Loss: 0.37365 (0.37365)
+2025-04-12,19:48:47 | INFO | Train Epoch: 27 [ 827392/3080192 (27%)] Data (t): 0.501 Batch (t): 4.647, 1769.65/s, 442.414/s/gpu LR: 0.000241 Logit Scale: 70.647 Contrastive_loss: 0.49643 (0.43504) Loss: 0.49643 (0.43504)
+2025-04-12,19:56:30 | INFO | Train Epoch: 27 [1646592/3080192 (53%)] Data (t): 0.487 Batch (t): 4.623, 1772.22/s, 443.055/s/gpu LR: 0.000231 Logit Scale: 71.111 Contrastive_loss: 0.47968 (0.44992) Loss: 0.47968 (0.44992)
+2025-04-12,20:04:12 | INFO | Train Epoch: 27 [2465792/3080192 (80%)] Data (t): 0.486 Batch (t): 4.623, 1777.43/s, 444.357/s/gpu LR: 0.000222 Logit Scale: 71.446 Contrastive_loss: 0.59746 (0.48680) Loss: 0.59746 (0.48680)
+2025-04-12,20:09:59 | INFO | Train Epoch: 27 [3080192/3080192 (100%)] Data (t): 0.487 Batch (t): 4.624, 1776.13/s, 444.033/s/gpu LR: 0.000216 Logit Scale: 71.788 Contrastive_loss: 0.49594 (0.48863) Loss: 0.49594 (0.48863)
+2025-04-12,20:10:00 | INFO | Start epoch 28
+2025-04-12,20:10:21 | INFO | Train Epoch: 28 [   8192/3080192 (0%)] Data (t): 16.807 Batch (t): 21.037, 389.416/s, 97.3540/s/gpu LR: 0.000216 Logit Scale: 71.793 Contrastive_loss: 0.36180 (0.36180) Loss: 0.36180 (0.36180)
+2025-04-12,20:18:06 | INFO | Train Epoch: 28 [ 827392/3080192 (27%)] Data (t): 0.495 Batch (t): 4.646, 1768.48/s, 442.119/s/gpu LR: 0.000207 Logit Scale: 72.591 Contrastive_loss: 0.44214 (0.40197) Loss: 0.44214 (0.40197)
+2025-04-12,20:25:48 | INFO | Train Epoch: 28 [1646592/3080192 (53%)] Data (t): 0.477 Batch (t): 4.621, 1775.15/s, 443.789/s/gpu LR: 0.000198 Logit Scale: 73.118 Contrastive_loss: 0.39160 (0.39851) Loss: 0.39160 (0.39851)
+2025-04-12,20:33:30 | INFO | Train Epoch: 28 [2465792/3080192 (80%)] Data (t): 0.482 Batch (t): 4.622, 1773.96/s, 443.491/s/gpu LR: 0.000190 Logit Scale: 73.503 Contrastive_loss: 0.38840 (0.39598) Loss: 0.38840 (0.39598)
+2025-04-12,20:39:17 | INFO | Train Epoch: 28 [3080192/3080192 (100%)] Data (t): 0.487 Batch (t): 4.624, 1772.09/s, 443.022/s/gpu LR: 0.000184 Logit Scale: 73.801 Contrastive_loss: 0.35650 (0.38809) Loss: 0.35650 (0.38809)
+2025-04-12,20:39:19 | INFO | Start epoch 29
+2025-04-12,20:39:40 | INFO | Train Epoch: 29 [   8192/3080192 (0%)] Data (t): 16.608 Batch (t): 20.831, 393.268/s, 98.3170/s/gpu LR: 0.000184 Logit Scale: 73.806 Contrastive_loss: 0.25394 (0.25394) Loss: 0.25394 (0.25394)
+2025-04-12,20:47:24 | INFO | Train Epoch: 29 [ 827392/3080192 (27%)] Data (t): 0.497 Batch (t): 4.645, 1770.72/s, 442.680/s/gpu LR: 0.000175 Logit Scale: 74.384 Contrastive_loss: 0.41378 (0.33386) Loss: 0.41378 (0.33386)
+2025-04-12,20:55:07 | INFO | Train Epoch: 29 [1646592/3080192 (53%)] Data (t): 0.487 Batch (t): 4.626, 1769.28/s, 442.321/s/gpu LR: 0.000167 Logit Scale: 74.811 Contrastive_loss: 0.48958 (0.38577) Loss: 0.48958 (0.38577)
+2025-04-12,21:02:50 | INFO | Train Epoch: 29 [2465792/3080192 (80%)] Data (t): 0.489 Batch (t): 4.626, 1756.99/s, 439.247/s/gpu LR: 0.000159 Logit Scale: 75.250 Contrastive_loss: 0.38187 (0.38479) Loss: 0.38187 (0.38479)
+2025-04-12,21:08:37 | INFO | Train Epoch: 29 [3080192/3080192 (100%)] Data (t): 0.492 Batch (t): 4.628, 1772.93/s, 443.234/s/gpu LR: 0.000154 Logit Scale: 75.680 Contrastive_loss: 0.38720 (0.38527) Loss: 0.38720 (0.38527)
+2025-04-12,21:08:39 | INFO | Start epoch 30
+2025-04-12,21:09:00 | INFO | Train Epoch: 30 [   8192/3080192 (0%)] Data (t): 16.737 Batch (t): 20.976, 390.540/s, 97.6351/s/gpu LR: 0.000153 Logit Scale: 75.685 Contrastive_loss: 0.33869 (0.33869) Loss: 0.33869 (0.33869)
+2025-04-12,21:16:45 | INFO | Train Epoch: 30 [ 827392/3080192 (27%)] Data (t): 0.502 Batch (t): 4.649, 1767.81/s, 441.952/s/gpu LR: 0.000146 Logit Scale: 76.288 Contrastive_loss: 0.31374 (0.32622) Loss: 0.31374 (0.32622)
+2025-04-12,21:24:27 | INFO | Train Epoch: 30 [1646592/3080192 (53%)] Data (t): 0.489 Batch (t): 4.627, 1770.43/s, 442.606/s/gpu LR: 0.000138 Logit Scale: 76.693 Contrastive_loss: 0.38628 (0.34624) Loss: 0.38628 (0.34624)
+2025-04-12,21:32:10 | INFO | Train Epoch: 30 [2465792/3080192 (80%)] Data (t): 0.488 Batch (t): 4.627, 1763.16/s, 440.790/s/gpu LR: 0.000131 Logit Scale: 77.063 Contrastive_loss: 0.47164 (0.37759) Loss: 0.47164 (0.37759)
+2025-04-12,21:37:57 | INFO | Train Epoch: 30 [3080192/3080192 (100%)] Data (t): 0.490 Batch (t): 4.627, 1775.68/s, 443.921/s/gpu LR: 0.000126 Logit Scale: 77.384 Contrastive_loss: 0.27648 (0.35737) Loss: 0.27648 (0.35737)
+2025-04-12,21:37:59 | INFO | Start epoch 31
+2025-04-12,21:38:20 | INFO | Train Epoch: 31 [   8192/3080192 (0%)] Data (t): 16.486 Batch (t): 20.697, 395.814/s, 98.9536/s/gpu LR: 0.000126 Logit Scale: 77.390 Contrastive_loss: 0.26433 (0.26433) Loss: 0.26433 (0.26433)
+2025-04-12,21:46:04 | INFO | Train Epoch: 31 [ 827392/3080192 (27%)] Data (t): 0.500 Batch (t): 4.645, 1770.66/s, 442.665/s/gpu LR: 0.000119 Logit Scale: 77.895 Contrastive_loss: 0.32327 (0.29380) Loss: 0.32327 (0.29380)
+2025-04-12,21:53:47 | INFO | Train Epoch: 31 [1646592/3080192 (53%)] Data (t): 0.488 Batch (t): 4.625, 1776.62/s, 444.155/s/gpu LR: 0.000112 Logit Scale: 78.244 Contrastive_loss: 0.26990 (0.28583) Loss: 0.26990 (0.28583)
+2025-04-12,22:01:29 | INFO | Train Epoch: 31 [2465792/3080192 (80%)] Data (t): 0.488 Batch (t): 4.625, 1764.68/s, 441.169/s/gpu LR: 0.000105 Logit Scale: 78.648 Contrastive_loss: 0.25305 (0.27764) Loss: 0.25305 (0.27764)
+2025-04-12,22:07:16 | INFO | Train Epoch: 31 [3080192/3080192 (100%)] Data (t): 0.491 Batch (t): 4.627, 1774.67/s, 443.669/s/gpu LR: 0.000100 Logit Scale: 78.891 Contrastive_loss: 0.27427 (0.27696) Loss: 0.27427 (0.27696)
+2025-04-12,22:07:19 | INFO | Start epoch 32
+2025-04-12,22:07:40 | INFO | Train Epoch: 32 [   8192/3080192 (0%)] Data (t): 16.630 Batch (t): 20.955, 390.942/s, 97.7355/s/gpu LR: 0.000100 Logit Scale: 78.896 Contrastive_loss: 0.28680 (0.28680) Loss: 0.28680 (0.28680)
+2025-04-12,22:15:24 | INFO | Train Epoch: 32 [ 827392/3080192 (27%)] Data (t): 0.507 Batch (t): 4.650, 1767.02/s, 441.754/s/gpu LR: 0.000094 Logit Scale: 79.300 Contrastive_loss: 0.26401 (0.27540) Loss: 0.26401 (0.27540)
+2025-04-12,22:23:07 | INFO | Train Epoch: 32 [1646592/3080192 (53%)] Data (t): 0.491 Batch (t): 4.628, 1770.42/s, 442.605/s/gpu LR: 0.000088 Logit Scale: 79.677 Contrastive_loss: 0.29907 (0.28329) Loss: 0.29907 (0.28329)
+2025-04-12,22:30:50 | INFO | Train Epoch: 32 [2465792/3080192 (80%)] Data (t): 0.491 Batch (t): 4.627, 1769.57/s, 442.393/s/gpu LR: 0.000082 Logit Scale: 79.980 Contrastive_loss: 0.27615 (0.28151) Loss: 0.27615 (0.28151)
+2025-04-12,22:36:37 | INFO | Train Epoch: 32 [3080192/3080192 (100%)] Data (t): 0.490 Batch (t): 4.625, 1776.28/s, 444.071/s/gpu LR: 0.000077 Logit Scale: 80.239 Contrastive_loss: 0.21444 (0.26809) Loss: 0.21444 (0.26809)
+2025-04-12,22:36:39 | INFO | Start epoch 33
+2025-04-12,22:37:00 | INFO | Train Epoch: 33 [   8192/3080192 (0%)] Data (t): 16.715 Batch (t): 20.936, 391.284/s, 97.8211/s/gpu LR: 0.000077 Logit Scale: 80.242 Contrastive_loss: 0.25637 (0.25637) Loss: 0.25637 (0.25637)
+2025-04-12,22:44:45 | INFO | Train Epoch: 33 [ 827392/3080192 (27%)] Data (t): 0.499 Batch (t): 4.647, 1766.21/s, 441.553/s/gpu LR: 0.000072 Logit Scale: 80.535 Contrastive_loss: 0.19438 (0.22538) Loss: 0.19438 (0.22538)
+2025-04-12,22:52:28 | INFO | Train Epoch: 33 [1646592/3080192 (53%)] Data (t): 0.491 Batch (t): 4.627, 1768.76/s, 442.189/s/gpu LR: 0.000066 Logit Scale: 80.776 Contrastive_loss: 0.18551 (0.21209) Loss: 0.18551 (0.21209)
+2025-04-12,23:00:10 | INFO | Train Epoch: 33 [2465792/3080192 (80%)] Data (t): 0.493 Batch (t): 4.628, 1766.99/s, 441.748/s/gpu LR: 0.000061 Logit Scale: 81.030 Contrastive_loss: 0.19242 (0.20717) Loss: 0.19242 (0.20717)
+2025-04-12,23:05:58 | INFO | Train Epoch: 33 [3080192/3080192 (100%)] Data (t): 0.494 Batch (t): 4.629, 1769.63/s, 442.407/s/gpu LR: 0.000057 Logit Scale: 81.198 Contrastive_loss: 0.20492 (0.20672) Loss: 0.20492 (0.20672)
+2025-04-12,23:06:00 | INFO | Start epoch 34
+2025-04-12,23:06:21 | INFO | Train Epoch: 34 [   8192/3080192 (0%)] Data (t): 16.615 Batch (t): 20.870, 392.516/s, 98.1291/s/gpu LR: 0.000057 Logit Scale: 81.200 Contrastive_loss: 0.16490 (0.16490) Loss: 0.16490 (0.16490)
+2025-04-12,23:14:05 | INFO | Train Epoch: 34 [ 827392/3080192 (27%)] Data (t): 0.502 Batch (t): 4.647, 1769.24/s, 442.309/s/gpu LR: 0.000052 Logit Scale: 81.441 Contrastive_loss: 0.21273 (0.18881) Loss: 0.21273 (0.18881)
+2025-04-12,23:21:48 | INFO | Train Epoch: 34 [1646592/3080192 (53%)] Data (t): 0.489 Batch (t): 4.625, 1765.33/s, 441.332/s/gpu LR: 0.000048 Logit Scale: 81.650 Contrastive_loss: 0.23317 (0.20360) Loss: 0.23317 (0.20360)
+2025-04-12,23:29:31 | INFO | Train Epoch: 34 [2465792/3080192 (80%)] Data (t): 0.490 Batch (t): 4.626, 1766.21/s, 441.554/s/gpu LR: 0.000043 Logit Scale: 81.851 Contrastive_loss: 0.19184 (0.20066) Loss: 0.19184 (0.20066)
+2025-04-12,23:35:18 | INFO | Train Epoch: 34 [3080192/3080192 (100%)] Data (t): 0.492 Batch (t): 4.627, 1774.03/s, 443.509/s/gpu LR: 0.000040 Logit Scale: 81.990 Contrastive_loss: 0.14899 (0.19033) Loss: 0.14899 (0.19033)
+2025-04-12,23:35:20 | INFO | Start epoch 35
+2025-04-12,23:35:41 | INFO | Train Epoch: 35 [   8192/3080192 (0%)] Data (t): 16.708 Batch (t): 20.902, 391.920/s, 97.9801/s/gpu LR: 0.000040 Logit Scale: 81.991 Contrastive_loss: 0.14559 (0.14559) Loss: 0.14559 (0.14559)
+2025-04-12,23:43:25 | INFO | Train Epoch: 35 [ 827392/3080192 (27%)] Data (t): 0.507 Batch (t): 4.647, 1771.83/s, 442.959/s/gpu LR: 0.000036 Logit Scale: 82.170 Contrastive_loss: 0.16298 (0.15429) Loss: 0.16298 (0.15429)
+2025-04-12,23:51:08 | INFO | Train Epoch: 35 [1646592/3080192 (53%)] Data (t): 0.493 Batch (t): 4.628, 1766.24/s, 441.559/s/gpu LR: 0.000032 Logit Scale: 82.309 Contrastive_loss: 0.15391 (0.15416) Loss: 0.15391 (0.15416)
+2025-04-12,23:58:51 | INFO | Train Epoch: 35 [2465792/3080192 (80%)] Data (t): 0.491 Batch (t): 4.625, 1767.08/s, 441.771/s/gpu LR: 0.000028 Logit Scale: 82.432 Contrastive_loss: 0.15303 (0.15388) Loss: 0.15303 (0.15388)
+2025-04-13,00:04:38 | INFO | Train Epoch: 35 [3080192/3080192 (100%)] Data (t): 0.495 Batch (t): 4.629, 1774.76/s, 443.690/s/gpu LR: 0.000026 Logit Scale: 82.527 Contrastive_loss: 0.16203 (0.15551) Loss: 0.16203 (0.15551)
+2025-04-13,00:04:41 | INFO | Start epoch 36
+2025-04-13,00:05:02 | INFO | Train Epoch: 36 [   8192/3080192 (0%)] Data (t): 16.804 Batch (t): 21.000, 390.097/s, 97.5243/s/gpu LR: 0.000026 Logit Scale: 82.528 Contrastive_loss: 0.42343 (0.42343) Loss: 0.42343 (0.42343)
+2025-04-13,00:12:46 | INFO | Train Epoch: 36 [ 827392/3080192 (27%)] Data (t): 0.503 Batch (t): 4.646, 1767.63/s, 441.908/s/gpu LR: 0.000022 Logit Scale: 82.645 Contrastive_loss: 0.14923 (0.28633) Loss: 0.14923 (0.28633)
+2025-04-13,00:20:29 | INFO | Train Epoch: 36 [1646592/3080192 (53%)] Data (t): 0.491 Batch (t): 4.626, 1775.05/s, 443.762/s/gpu LR: 0.000019 Logit Scale: 82.746 Contrastive_loss: 0.14070 (0.23779) Loss: 0.14070 (0.23779)
+2025-04-13,00:28:11 | INFO | Train Epoch: 36 [2465792/3080192 (80%)] Data (t): 0.490 Batch (t): 4.626, 1770.23/s, 442.559/s/gpu LR: 0.000017 Logit Scale: 82.828 Contrastive_loss: 0.16643 (0.21995) Loss: 0.16643 (0.21995)
+2025-04-13,00:33:58 | INFO | Train Epoch: 36 [3080192/3080192 (100%)] Data (t): 0.494 Batch (t): 4.629, 1773.34/s, 443.335/s/gpu LR: 0.000015 Logit Scale: 82.890 Contrastive_loss: 0.15245 (0.20645) Loss: 0.15245 (0.20645)
+2025-04-13,00:34:01 | INFO | Start epoch 37
+2025-04-13,00:34:22 | INFO | Train Epoch: 37 [   8192/3080192 (0%)] Data (t): 16.554 Batch (t): 20.750, 394.794/s, 98.6984/s/gpu LR: 0.000015 Logit Scale: 82.891 Contrastive_loss: 0.13314 (0.13314) Loss: 0.13314 (0.13314)
+2025-04-13,00:42:06 | INFO | Train Epoch: 37 [ 827392/3080192 (27%)] Data (t): 0.504 Batch (t): 4.646, 1775.16/s, 443.789/s/gpu LR: 0.000012 Logit Scale: 82.966 Contrastive_loss: 0.18545 (0.15929) Loss: 0.18545 (0.15929)
+2025-04-13,00:49:49 | INFO | Train Epoch: 37 [1646592/3080192 (53%)] Data (t): 0.491 Batch (t): 4.625, 1769.81/s, 442.452/s/gpu LR: 0.000010 Logit Scale: 83.020 Contrastive_loss: 0.15202 (0.15687) Loss: 0.15202 (0.15687)
+2025-04-13,00:57:31 | INFO | Train Epoch: 37 [2465792/3080192 (80%)] Data (t): 0.490 Batch (t): 4.624, 1765.39/s, 441.347/s/gpu LR: 0.000008 Logit Scale: 83.067 Contrastive_loss: 0.16217 (0.15819) Loss: 0.16217 (0.15819)
+2025-04-13,01:03:18 | INFO | Train Epoch: 37 [3080192/3080192 (100%)] Data (t): 0.493 Batch (t): 4.627, 1770.18/s, 442.545/s/gpu LR: 0.000006 Logit Scale: 83.094 Contrastive_loss: 0.12795 (0.15214) Loss: 0.12795 (0.15214)
+2025-04-13,01:03:20 | INFO | Start epoch 38
+2025-04-13,01:03:41 | INFO | Train Epoch: 38 [   8192/3080192 (0%)] Data (t): 16.456 Batch (t): 20.684, 396.051/s, 99.0126/s/gpu LR: 0.000006 Logit Scale: 83.094 Contrastive_loss: 0.15518 (0.15518) Loss: 0.15518 (0.15518)
+2025-04-13,01:11:26 | INFO | Train Epoch: 38 [ 827392/3080192 (27%)] Data (t): 0.504 Batch (t): 4.647, 1752.37/s, 438.093/s/gpu LR: 0.000005 Logit Scale: 83.130 Contrastive_loss: 0.12225 (0.13872) Loss: 0.12225 (0.13872)
+2025-04-13,01:19:08 | INFO | Train Epoch: 38 [1646592/3080192 (53%)] Data (t): 0.491 Batch (t): 4.626, 1771.39/s, 442.849/s/gpu LR: 0.000003 Logit Scale: 83.152 Contrastive_loss: 0.13913 (0.13885) Loss: 0.13913 (0.13885)
+2025-04-13,01:26:51 | INFO | Train Epoch: 38 [2465792/3080192 (80%)] Data (t): 0.491 Batch (t): 4.626, 1773.63/s, 443.407/s/gpu LR: 0.000002 Logit Scale: 83.167 Contrastive_loss: 0.17707 (0.14841) Loss: 0.17707 (0.14841)
+2025-04-13,01:32:38 | INFO | Train Epoch: 38 [3080192/3080192 (100%)] Data (t): 0.493 Batch (t): 4.626, 1772.34/s, 443.084/s/gpu LR: 0.000002 Logit Scale: 83.174 Contrastive_loss: 0.15010 (0.14875) Loss: 0.15010 (0.14875)
+2025-04-13,01:32:40 | INFO | Start epoch 39
+2025-04-13,01:33:01 | INFO | Train Epoch: 39 [   8192/3080192 (0%)] Data (t): 16.855 Batch (t): 21.055, 389.073/s, 97.2682/s/gpu LR: 0.000002 Logit Scale: 83.175 Contrastive_loss: 0.16583 (0.16583) Loss: 0.16583 (0.16583)
+2025-04-13,01:40:46 | INFO | Train Epoch: 39 [ 827392/3080192 (27%)] Data (t): 0.500 Batch (t): 4.648, 1772.85/s, 443.211/s/gpu LR: 0.000001 Logit Scale: 83.182 Contrastive_loss: 0.18496 (0.17540) Loss: 0.18496 (0.17540)
+2025-04-13,01:48:29 | INFO | Train Epoch: 39 [1646592/3080192 (53%)] Data (t): 0.491 Batch (t): 4.626, 1766.88/s, 441.719/s/gpu LR: 0.000000 Logit Scale: 83.186 Contrastive_loss: 0.13388 (0.16156) Loss: 0.13388 (0.16156)
+2025-04-13,01:56:11 | INFO | Train Epoch: 39 [2465792/3080192 (80%)] Data (t): 0.490 Batch (t): 4.625, 1766.41/s, 441.604/s/gpu LR: 0.000000 Logit Scale: 83.186 Contrastive_loss: 0.11790 (0.15064) Loss: 0.11790 (0.15064)
+2025-04-13,02:01:58 | INFO | Train Epoch: 39 [3080192/3080192 (100%)] Data (t): 0.494 Batch (t): 4.628, 1776.17/s, 444.043/s/gpu LR: 0.000000 Logit Scale: 83.186 Contrastive_loss: 0.16144 (0.15280) Loss: 0.16144 (0.15280)

SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/0ViT-B-16-cc3m-laclip-mix-inter-010-r19/params.txt

@@ -0,0 +1,103 @@
+accum_freq: 2
+aug_cfg: {}
+batch_size: 1024
+beta1: 0.9
+beta2: 0.98
+cache_dir: None
+caption_ratio: 0.1
+checkpoint_path: /mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R_0.95_1.0_dinov2-large_0.0_0.05_rouge_0.7_top_4/logs/ViT-B-16-cc3m-laclip-mix-inter-010-r19/checkpoints
+coca_caption_loss_weight: 2.0
+coca_contrastive_loss_weight: 1.0
+copy_codebase: False
+csv_caption_key: title
+csv_img_key: filepath
+csv_separator: 	
+dataset_resampled: False
+dataset_type: synthetic
+ddp_static_graph: False
+debug: False
+delete_previous_checkpoint: False
+device: cuda:0
+dist_backend: None
+dist_url: None
+distill: False
+distill_model: None
+distill_pretrained: None
+distributed: True
+epochs: 40
+epochs_cooldown: None
+eps: 1e-08
+force_custom_text: False
+force_image_size: None
+force_patch_dropout: None
+force_quick_gelu: False
+gather_with_grad: True
+grad_checkpointing: True
+grad_clip_norm: None
+horovod: False
+image_interpolation: None
+image_mean: None
+image_resize_mode: None
+image_std: None
+imagenet_v2: None
+imagenet_val: None
+keep_func_name: 
+local_loss: True
+local_rank: 0
+lock_image: False
+lock_image_freeze_bn_stats: False
+lock_image_unlocked_groups: 0
+lock_text: False
+lock_text_freeze_layer_norm: False
+lock_text_unlocked_layers: 0
+log_every_n_steps: 100
+log_level: 20
+log_local: False
+log_path: /mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R_0.95_1.0_dinov2-large_0.0_0.05_rouge_0.7_top_4/logs/ViT-B-16-cc3m-laclip-mix-inter-010-r19/out.log
+logs: /mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R_0.95_1.0_dinov2-large_0.0_0.05_rouge_0.7_top_4/logs
+loss_dist_impl: None
+lr: 0.001
+lr_cooldown_end: 0.0
+lr_cooldown_power: 1.0
+lr_scheduler: cosine
+map_func_name: use_low_inter_only
+model: ViT-B-16
+momentum: None
+name: ViT-B-16-cc3m-laclip-mix-inter-010-r19
+no_set_device_rank: False
+opt: adamw
+precision: amp
+pretrained: 
+pretrained_image: False
+rank: 0
+remote_sync: None
+remote_sync_frequency: 300
+remote_sync_protocol: s3
+report_to: wandb
+resume: /mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R_0.95_1.0_dinov2-large_0.0_0.05_rouge_0.7_top_4/logs/ViT-B-16-cc3m-laclip-mix-inter-010-r13/checkpoints/epoch_19.pt
+save_frequency: 1
+save_most_recent: False
+seed: 0
+siglip: False
+skip_scheduler: False
+tensorboard: False
+tensorboard_path: 
+torchcompile: False
+torchscript: False
+trace: False
+train_data: /mnt/personal/zhudongy/cc3m-hgf-wds/{0000..0301}.tar
+train_data_upsampling_factors: None
+train_num_samples: 3016640
+use_bn_sync: False
+use_bnb_linear: None
+val_data: None
+val_frequency: 1
+val_num_samples: None
+wandb: True
+wandb_notes: 
+wandb_project_name: open-clip
+warmup: 368
+wd: 0.5
+workers: 16
+world_size: 4
+zeroshot_frequency: 2

SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-010-filled-map_text_closest_image_closest/benchmark_cifar100_epoch_40.pt_ViT-B-16_en_zeroshot_classification.json

@@ -0,0 +1 @@
+{"dataset": "cifar100", "model": "ViT-B-16", "pretrained": "/mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-010-filled-map_text_closest_image_closest//checkpoints/epoch_40.pt", "task": "zeroshot_classification", "metrics": {"acc1": 0.1801, "acc5": 0.4249, "mean_per_class_recall": 0.18}, "language": "en"}

SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-010-filled-map_text_closest_image_closest/benchmark_country211_epoch_40.pt_ViT-B-16_en_zeroshot_classification.json

@@ -0,0 +1 @@
+{"dataset": "country211", "model": "ViT-B-16", "pretrained": "/mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-010-filled-map_text_closest_image_closest//checkpoints/epoch_40.pt", "task": "zeroshot_classification", "metrics": {"acc1": 0.005971563981042654, "acc5": 0.02919431279620853, "mean_per_class_recall": 0.005924170616113744}, "language": "en"}

SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-010-filled-map_text_closest_image_closest/benchmark_vtab_resisc45_epoch_40.pt_ViT-B-16_en_zeroshot_classification.json

@@ -0,0 +1 @@
+{"dataset": "vtab/resisc45", "model": "ViT-B-16", "pretrained": "/mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-010-filled-map_text_closest_image_closest//checkpoints/epoch_40.pt", "task": "zeroshot_classification", "metrics": {"acc1": 0.17333333333333334, "acc5": 0.44126984126984126, "mean_per_class_recall": 0.17410264610414472}, "language": "en"}

SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-010-filled-map_text_closest_image_closest/out.log

@@ -0,0 +1,394 @@
+2025-04-26,07:10:25 | INFO | Running in distributed mode with multiple processes. Device: cuda:0.Process (global: 0, local 0), total 4.
+2025-04-26,07:10:25 | INFO | Loaded ViT-B-16 model config.
+2025-04-26,07:10:26 | INFO | Model:
+2025-04-26,07:10:26 | INFO | CLIP(
+  (visual): VisionTransformer(
+    (conv1): Conv2d(3, 768, kernel_size=(16, 16), stride=(16, 16), bias=False)
+    (patch_dropout): Identity()
+    (ln_pre): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
+    (transformer): Transformer(
+      (resblocks): ModuleList(
+        (0-11): 12 x ResidualAttentionBlock(
+          (ln_1): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
+          (attn): MultiheadAttention(
+            (out_proj): NonDynamicallyQuantizableLinear(in_features=768, out_features=768, bias=True)
+          )
+          (ls_1): Identity()
+          (ln_2): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
+          (mlp): Sequential(
+            (c_fc): Linear(in_features=768, out_features=3072, bias=True)
+            (gelu): GELU(approximate='none')
+            (c_proj): Linear(in_features=3072, out_features=768, bias=True)
+          )
+          (ls_2): Identity()
+        )
+      )
+    )
+    (ln_post): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
+  )
+  (transformer): Transformer(
+    (resblocks): ModuleList(
+      (0-11): 12 x ResidualAttentionBlock(
+        (ln_1): LayerNorm((512,), eps=1e-05, elementwise_affine=True)
+        (attn): MultiheadAttention(
+          (out_proj): NonDynamicallyQuantizableLinear(in_features=512, out_features=512, bias=True)
+        )
+        (ls_1): Identity()
+        (ln_2): LayerNorm((512,), eps=1e-05, elementwise_affine=True)
+        (mlp): Sequential(
+          (c_fc): Linear(in_features=512, out_features=2048, bias=True)
+          (gelu): GELU(approximate='none')
+          (c_proj): Linear(in_features=2048, out_features=512, bias=True)
+        )
+        (ls_2): Identity()
+      )
+    )
+  )
+  (token_embedding): Embedding(49408, 512)
+  (ln_final): LayerNorm((512,), eps=1e-05, elementwise_affine=True)
+)
+2025-04-26,07:10:26 | INFO | Params:
+2025-04-26,07:10:26 | INFO |   accum_freq: 4
+2025-04-26,07:10:26 | INFO |   aug_cfg: {}
+2025-04-26,07:10:26 | INFO |   batch_size: 512
+2025-04-26,07:10:26 | INFO |   beta1: 0.9
+2025-04-26,07:10:26 | INFO |   beta2: 0.98
+2025-04-26,07:10:26 | INFO |   cache_dir: None
+2025-04-26,07:10:26 | INFO |   caption_ratio: 0.1
+2025-04-26,07:10:26 | INFO |   checkpoint_path: /mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-010-filled-map_text_closest_image_closest/checkpoints
+2025-04-26,07:10:26 | INFO |   coca_caption_loss_weight: 2.0
+2025-04-26,07:10:26 | INFO |   coca_contrastive_loss_weight: 1.0
+2025-04-26,07:10:26 | INFO |   copy_codebase: False
+2025-04-26,07:10:26 | INFO |   csv_caption_key: title
+2025-04-26,07:10:26 | INFO |   csv_img_key: filepath
+2025-04-26,07:10:26 | INFO |   csv_separator: 	
+2025-04-26,07:10:26 | INFO |   dataset_resampled: False
+2025-04-26,07:10:26 | INFO |   dataset_type: synthetic
+2025-04-26,07:10:26 | INFO |   ddp_static_graph: False
+2025-04-26,07:10:26 | INFO |   debug: False
+2025-04-26,07:10:26 | INFO |   delete_previous_checkpoint: False
+2025-04-26,07:10:26 | INFO |   device: cuda:0
+2025-04-26,07:10:26 | INFO |   dist_backend: None
+2025-04-26,07:10:26 | INFO |   dist_url: None
+2025-04-26,07:10:26 | INFO |   distill: False
+2025-04-26,07:10:26 | INFO |   distill_model: None
+2025-04-26,07:10:26 | INFO |   distill_pretrained: None
+2025-04-26,07:10:26 | INFO |   distributed: True
+2025-04-26,07:10:26 | INFO |   epochs: 40
+2025-04-26,07:10:26 | INFO |   epochs_cooldown: None
+2025-04-26,07:10:26 | INFO |   eps: 1e-08
+2025-04-26,07:10:26 | INFO |   force_custom_text: False
+2025-04-26,07:10:26 | INFO |   force_image_size: None
+2025-04-26,07:10:26 | INFO |   force_patch_dropout: None
+2025-04-26,07:10:26 | INFO |   force_quick_gelu: False
+2025-04-26,07:10:26 | INFO |   gather_with_grad: True
+2025-04-26,07:10:26 | INFO |   grad_checkpointing: True
+2025-04-26,07:10:26 | INFO |   grad_clip_norm: None
+2025-04-26,07:10:26 | INFO |   horovod: False
+2025-04-26,07:10:26 | INFO |   image_interpolation: None
+2025-04-26,07:10:26 | INFO |   image_mean: None
+2025-04-26,07:10:26 | INFO |   image_resize_mode: None
+2025-04-26,07:10:26 | INFO |   image_std: None
+2025-04-26,07:10:26 | INFO |   imagenet_v2: None
+2025-04-26,07:10:26 | INFO |   imagenet_val: None
+2025-04-26,07:10:26 | INFO |   keep_func_name: 
+2025-04-26,07:10:26 | INFO |   local_loss: False
+2025-04-26,07:10:26 | INFO |   local_rank: 0
+2025-04-26,07:10:26 | INFO |   lock_image: False
+2025-04-26,07:10:26 | INFO |   lock_image_freeze_bn_stats: False
+2025-04-26,07:10:26 | INFO |   lock_image_unlocked_groups: 0
+2025-04-26,07:10:26 | INFO |   lock_text: False
+2025-04-26,07:10:26 | INFO |   lock_text_freeze_layer_norm: False
+2025-04-26,07:10:26 | INFO |   lock_text_unlocked_layers: 0
+2025-04-26,07:10:26 | INFO |   log_every_n_steps: 100
+2025-04-26,07:10:26 | INFO |   log_level: 20
+2025-04-26,07:10:26 | INFO |   log_local: False
+2025-04-26,07:10:26 | INFO |   log_path: /mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-010-filled-map_text_closest_image_closest/out.log
+2025-04-26,07:10:26 | INFO |   logs: /mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs
+2025-04-26,07:10:26 | INFO |   loss_dist_impl: None
+2025-04-26,07:10:26 | INFO |   lr: 0.001
+2025-04-26,07:10:26 | INFO |   lr_cooldown_end: 0.0
+2025-04-26,07:10:26 | INFO |   lr_cooldown_power: 1.0
+2025-04-26,07:10:26 | INFO |   lr_scheduler: cosine
+2025-04-26,07:10:26 | INFO |   map_func_name: map_text_closest_image_closest
+2025-04-26,07:10:26 | INFO |   model: ViT-B-16
+2025-04-26,07:10:26 | INFO |   momentum: None
+2025-04-26,07:10:26 | INFO |   name: ViT-B-16-cc3m-laclip-mix-010-filled-map_text_closest_image_closest
+2025-04-26,07:10:26 | INFO |   no_set_device_rank: False
+2025-04-26,07:10:26 | INFO |   opt: adamw
+2025-04-26,07:10:26 | INFO |   precision: amp
+2025-04-26,07:10:26 | INFO |   pretrained: 
+2025-04-26,07:10:26 | INFO |   pretrained_image: False
+2025-04-26,07:10:26 | INFO |   rank: 0
+2025-04-26,07:10:26 | INFO |   remote_sync: None
+2025-04-26,07:10:26 | INFO |   remote_sync_frequency: 300
+2025-04-26,07:10:26 | INFO |   remote_sync_protocol: s3
+2025-04-26,07:10:26 | INFO |   report_to: tensorboard,wandb
+2025-04-26,07:10:26 | INFO |   resume: None
+2025-04-26,07:10:26 | INFO |   save_frequency: 1
+2025-04-26,07:10:26 | INFO |   save_most_recent: False
+2025-04-26,07:10:26 | INFO |   seed: 0
+2025-04-26,07:10:26 | INFO |   siglip: False
+2025-04-26,07:10:26 | INFO |   skip_scheduler: False
+2025-04-26,07:10:26 | INFO |   tensorboard: True
+2025-04-26,07:10:26 | INFO |   tensorboard_path: /mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-010-filled-map_text_closest_image_closest/tensorboard
+2025-04-26,07:10:26 | INFO |   torchcompile: False
+2025-04-26,07:10:26 | INFO |   torchscript: False
+2025-04-26,07:10:26 | INFO |   trace: False
+2025-04-26,07:10:26 | INFO |   train_data: /mnt/personal/zhudongy/cc3m-hgf-wds/{0000..0301}.tar
+2025-04-26,07:10:26 | INFO |   train_data_upsampling_factors: None
+2025-04-26,07:10:26 | INFO |   train_num_samples: 3016640
+2025-04-26,07:10:26 | INFO |   use_bn_sync: False
+2025-04-26,07:10:26 | INFO |   use_bnb_linear: None
+2025-04-26,07:10:26 | INFO |   val_data: None
+2025-04-26,07:10:26 | INFO |   val_frequency: 1
+2025-04-26,07:10:26 | INFO |   val_num_samples: None
+2025-04-26,07:10:26 | INFO |   wandb: True
+2025-04-26,07:10:26 | INFO |   wandb_notes: 
+2025-04-26,07:10:26 | INFO |   wandb_project_name: open-clip
+2025-04-26,07:10:26 | INFO |   warmup: 368
+2025-04-26,07:10:26 | INFO |   wd: 0.5
+2025-04-26,07:10:26 | INFO |   workers: 16
+2025-04-26,07:10:26 | INFO |   world_size: 4
+2025-04-26,07:10:26 | INFO |   zeroshot_frequency: 2
+2025-04-26,07:10:28 | INFO | Created AdamW (adamw) optimizer: lr: 0.001, betas: (0.9, 0.98), eps: 1e-08, weight_decay: 0.5, amsgrad: False, foreach: None, maximize: False, capturable: False, differentiable: False, fused: None
+2025-04-26,07:13:13 | INFO | Start epoch 0
+2025-04-26,07:13:34 | INFO | Train Epoch: 0 [   8192/3047424 (0%)] Data (t): 15.323 Batch (t): 20.238, 404.781/s, 101.195/s/gpu LR: 0.000003 Logit Scale: 14.286 Imm_image: -0.27226 (-0.27226) Imm_text: -0.27226 (-0.27226) Isd_image: -0.27166 (-0.27166) Isd_text: -0.27166 (-0.27166) Contrastive_loss: 9.1250 (9.1250) Loss: 9.1250 (9.1250)
+2025-04-26,07:21:45 | INFO | Train Epoch: 0 [ 827392/3047424 (27%)] Data (t): 0.684 Batch (t): 4.914, 1663.35/s, 415.837/s/gpu LR: 0.000274 Logit Scale: 14.245 Imm_image: 4.1377 (1.9327) Imm_text: 4.1377 (1.9327) Isd_image: 2.4432 (1.0858) Isd_text: 2.4432 (1.0858) Contrastive_loss: 8.1876 (8.6563) Loss: 8.1876 (8.6563)
+2025-04-26,07:29:58 | INFO | Train Epoch: 0 [1646592/3047424 (54%)] Data (t): 0.697 Batch (t): 4.928, 1664.28/s, 416.070/s/gpu LR: 0.000546 Logit Scale: 14.208 Imm_image: 6.1775 (3.3477) Imm_text: 6.1775 (3.3477) Isd_image: 2.4150 (1.5289) Isd_text: 2.4150 (1.5289) Contrastive_loss: 7.4246 (8.2457) Loss: 7.4246 (8.2457)
+2025-04-26,07:38:10 | INFO | Train Epoch: 0 [2465792/3047424 (81%)] Data (t): 0.691 Batch (t): 4.919, 1678.55/s, 419.637/s/gpu LR: 0.000818 Logit Scale: 14.137 Imm_image: 6.3221 (4.0913) Imm_text: 6.3221 (4.0913) Isd_image: 2.3277 (1.7286) Isd_text: 2.3277 (1.7286) Contrastive_loss: 7.3997 (8.0342) Loss: 7.3997 (8.0342)
+2025-04-26,07:43:57 | INFO | Train Epoch: 0 [3047424/3047424 (100%)] Data (t): 0.676 Batch (t): 4.897, 1711.41/s, 427.852/s/gpu LR: 0.001000 Logit Scale: 14.114 Imm_image: 6.6559 (4.6042) Imm_text: 6.6559 (4.6042) Isd_image: 2.3677 (1.8564) Isd_text: 2.3677 (1.8564) Contrastive_loss: 7.1586 (7.8591) Loss: 7.1586 (7.8591)
+2025-04-26,07:44:00 | INFO | Start epoch 1
+2025-04-26,07:44:17 | INFO | Train Epoch: 1 [   8192/3047424 (0%)] Data (t): 12.811 Batch (t): 17.417, 470.348/s, 117.587/s/gpu LR: 0.001000 Logit Scale: 14.116 Imm_image: 6.6618 (6.6618) Imm_text: 6.6618 (6.6618) Isd_image: 2.2505 (2.2505) Isd_text: 2.2505 (2.2505) Contrastive_loss: 7.1644 (7.1644) Loss: 7.1644 (7.1644)
+2025-04-26,07:52:31 | INFO | Train Epoch: 1 [ 827392/3047424 (27%)] Data (t): 0.702 Batch (t): 4.943, 1668.97/s, 417.244/s/gpu LR: 0.001000 Logit Scale: 14.569 Imm_image: 7.1003 (6.8810) Imm_text: 7.1003 (6.8810) Isd_image: 1.9572 (2.1038) Isd_text: 1.9572 (2.1038) Contrastive_loss: 6.7468 (6.9556) Loss: 6.7468 (6.9556)
+2025-04-26,08:00:44 | INFO | Train Epoch: 1 [1646592/3047424 (54%)] Data (t): 0.696 Batch (t): 4.924, 1661.59/s, 415.398/s/gpu LR: 0.001000 Logit Scale: 15.209 Imm_image: 7.4766 (7.0796) Imm_text: 7.4766 (7.0796) Isd_image: 1.6824 (1.9634) Isd_text: 1.6824 (1.9634) Contrastive_loss: 6.4242 (6.7785) Loss: 6.4242 (6.7785)
+2025-04-26,08:08:56 | INFO | Train Epoch: 1 [2465792/3047424 (81%)] Data (t): 0.691 Batch (t): 4.922, 1679.14/s, 419.785/s/gpu LR: 0.000999 Logit Scale: 16.186 Imm_image: 8.1138 (7.3381) Imm_text: 8.1138 (7.3381) Isd_image: 1.3676 (1.8144) Isd_text: 1.3676 (1.8144) Contrastive_loss: 6.0415 (6.5942) Loss: 6.0415 (6.5942)
+2025-04-26,08:14:44 | INFO | Train Epoch: 1 [3047424/3047424 (100%)] Data (t): 0.678 Batch (t): 4.902, 1714.24/s, 428.561/s/gpu LR: 0.000998 Logit Scale: 17.056 Imm_image: 8.3529 (7.5411) Imm_text: 8.3529 (7.5411) Isd_image: 1.2348 (1.6985) Isd_text: 1.2348 (1.6985) Contrastive_loss: 5.8203 (6.4394) Loss: 5.8203 (6.4394)
+2025-04-26,08:14:46 | INFO | Start epoch 2
+2025-04-26,08:15:04 | INFO | Train Epoch: 2 [   8192/3047424 (0%)] Data (t): 13.330 Batch (t): 18.078, 453.151/s, 113.288/s/gpu LR: 0.000998 Logit Scale: 17.070 Imm_image: 8.4309 (8.4309) Imm_text: 8.4309 (8.4309) Isd_image: 1.1545 (1.1545) Isd_text: 1.1545 (1.1545) Contrastive_loss: 5.7120 (5.7120) Loss: 5.7120 (5.7120)
+2025-04-26,08:23:19 | INFO | Train Epoch: 2 [ 827392/3047424 (27%)] Data (t): 0.702 Batch (t): 4.948, 1670.24/s, 417.560/s/gpu LR: 0.000997 Logit Scale: 18.528 Imm_image: 8.7349 (8.5829) Imm_text: 8.7349 (8.5829) Isd_image: 0.95538 (1.0549) Isd_text: 0.95538 (1.0549) Contrastive_loss: 5.6248 (5.6684) Loss: 5.6248 (5.6684)
+2025-04-26,08:31:32 | INFO | Train Epoch: 2 [1646592/3047424 (54%)] Data (t): 0.692 Batch (t): 4.929, 1662.44/s, 415.610/s/gpu LR: 0.000996 Logit Scale: 19.925 Imm_image: 9.5194 (8.8951) Imm_text: 9.5194 (8.8951) Isd_image: 0.95206 (1.0207) Isd_text: 0.95206 (1.0207) Contrastive_loss: 5.2383 (5.5250) Loss: 5.2383 (5.5250)
+2025-04-26,08:39:44 | INFO | Train Epoch: 2 [2465792/3047424 (81%)] Data (t): 0.687 Batch (t): 4.922, 1678.58/s, 419.644/s/gpu LR: 0.000995 Logit Scale: 21.434 Imm_image: 10.016 (9.1752) Imm_text: 10.016 (9.1752) Isd_image: 1.1104 (1.0431) Isd_text: 1.1104 (1.0431) Contrastive_loss: 5.0808 (5.4139) Loss: 5.0808 (5.4139)
+2025-04-26,08:45:32 | INFO | Train Epoch: 2 [3047424/3047424 (100%)] Data (t): 0.679 Batch (t): 4.905, 1713.87/s, 428.469/s/gpu LR: 0.000993 Logit Scale: 22.459 Imm_image: 10.597 (9.4596) Imm_text: 10.597 (9.4596) Isd_image: 1.2498 (1.0844) Isd_text: 1.2498 (1.0844) Contrastive_loss: 4.8430 (5.2997) Loss: 4.8430 (5.2997)
+2025-04-26,08:45:34 | INFO | Start epoch 3
+2025-04-26,08:45:53 | INFO | Train Epoch: 3 [   8192/3047424 (0%)] Data (t): 13.476 Batch (t): 18.133, 451.782/s, 112.946/s/gpu LR: 0.000993 Logit Scale: 22.473 Imm_image: 10.742 (10.742) Imm_text: 10.742 (10.742) Isd_image: 1.1603 (1.1603) Isd_text: 1.1603 (1.1603) Contrastive_loss: 4.7363 (4.7363) Loss: 4.7363 (4.7363)
+2025-04-26,08:54:08 | INFO | Train Epoch: 3 [ 827392/3047424 (27%)] Data (t): 0.706 Batch (t): 4.954, 1657.98/s, 414.495/s/gpu LR: 0.000992 Logit Scale: 24.092 Imm_image: 11.367 (11.054) Imm_text: 11.367 (11.054) Isd_image: 1.1550 (1.1576) Isd_text: 1.1550 (1.1576) Contrastive_loss: 4.6024 (4.6693) Loss: 4.6024 (4.6693)
+2025-04-26,09:02:21 | INFO | Train Epoch: 3 [1646592/3047424 (54%)] Data (t): 0.692 Batch (t): 4.927, 1674.18/s, 418.544/s/gpu LR: 0.000990 Logit Scale: 25.287 Imm_image: 11.905 (11.338) Imm_text: 11.905 (11.338) Isd_image: 1.4233 (1.2462) Isd_text: 1.4233 (1.2462) Contrastive_loss: 4.4603 (4.5996) Loss: 4.4603 (4.5996)
+2025-04-26,09:10:34 | INFO | Train Epoch: 3 [2465792/3047424 (81%)] Data (t): 0.696 Batch (t): 4.932, 1673.44/s, 418.359/s/gpu LR: 0.000987 Logit Scale: 26.572 Imm_image: 12.404 (11.604) Imm_text: 12.404 (11.604) Isd_image: 1.4679 (1.3016) Isd_text: 1.4679 (1.3016) Contrastive_loss: 4.3721 (4.5428) Loss: 4.3721 (4.5428)
+2025-04-26,09:16:20 | INFO | Train Epoch: 3 [3047424/3047424 (100%)] Data (t): 0.656 Batch (t): 4.878, 1711.61/s, 427.902/s/gpu LR: 0.000985 Logit Scale: 27.165 Imm_image: 12.715 (11.826) Imm_text: 12.715 (11.826) Isd_image: 1.5461 (1.3505) Isd_text: 1.5461 (1.3505) Contrastive_loss: 4.1237 (4.4590) Loss: 4.1237 (4.4590)
+2025-04-26,09:16:22 | INFO | Start epoch 4
+2025-04-26,09:16:40 | INFO | Train Epoch: 4 [   8192/3047424 (0%)] Data (t): 12.622 Batch (t): 17.191, 476.528/s, 119.132/s/gpu LR: 0.000985 Logit Scale: 27.172 Imm_image: 12.831 (12.831) Imm_text: 12.831 (12.831) Isd_image: 1.3067 (1.3067) Isd_text: 1.3067 (1.3067) Contrastive_loss: 3.9707 (3.9707) Loss: 3.9707 (3.9707)
+2025-04-26,09:24:53 | INFO | Train Epoch: 4 [ 827392/3047424 (27%)] Data (t): 0.686 Batch (t): 4.931, 1685.66/s, 421.416/s/gpu LR: 0.000983 Logit Scale: 28.422 Imm_image: 13.191 (13.011) Imm_text: 13.191 (13.011) Isd_image: 1.5714 (1.4391) Isd_text: 1.5714 (1.4391) Contrastive_loss: 4.0104 (3.9905) Loss: 4.0104 (3.9905)
+2025-04-26,09:33:05 | INFO | Train Epoch: 4 [1646592/3047424 (54%)] Data (t): 0.686 Batch (t): 4.920, 1671.61/s, 417.901/s/gpu LR: 0.000980 Logit Scale: 29.453 Imm_image: 13.635 (13.219) Imm_text: 13.635 (13.219) Isd_image: 1.8343 (1.5708) Isd_text: 1.8343 (1.5708) Contrastive_loss: 3.9989 (3.9933) Loss: 3.9989 (3.9933)
+2025-04-26,09:41:15 | INFO | Train Epoch: 4 [2465792/3047424 (81%)] Data (t): 0.670 Batch (t): 4.900, 1683.47/s, 420.868/s/gpu LR: 0.000977 Logit Scale: 30.264 Imm_image: 13.797 (13.364) Imm_text: 13.797 (13.364) Isd_image: 2.0846 (1.6992) Isd_text: 2.0846 (1.6992) Contrastive_loss: 3.9720 (3.9880) Loss: 3.9720 (3.9880)
+2025-04-26,09:47:02 | INFO | Train Epoch: 4 [3047424/3047424 (100%)] Data (t): 0.669 Batch (t): 4.894, 1708.23/s, 427.057/s/gpu LR: 0.000974 Logit Scale: 30.710 Imm_image: 14.276 (13.546) Imm_text: 14.276 (13.546) Isd_image: 1.7781 (1.7150) Isd_text: 1.7781 (1.7150) Contrastive_loss: 3.6126 (3.9129) Loss: 3.6126 (3.9129)
+2025-04-26,09:47:04 | INFO | Start epoch 5
+2025-04-26,09:47:22 | INFO | Train Epoch: 5 [   8192/3047424 (0%)] Data (t): 12.745 Batch (t): 17.245, 475.025/s, 118.756/s/gpu LR: 0.000974 Logit Scale: 30.718 Imm_image: 14.405 (14.405) Imm_text: 14.405 (14.405) Isd_image: 1.5662 (1.5662) Isd_text: 1.5662 (1.5662) Contrastive_loss: 3.5208 (3.5208) Loss: 3.5208 (3.5208)
+2025-04-26,09:55:36 | INFO | Train Epoch: 5 [ 827392/3047424 (27%)] Data (t): 0.703 Batch (t): 4.948, 1681.00/s, 420.249/s/gpu LR: 0.000971 Logit Scale: 31.926 Imm_image: 14.772 (14.588) Imm_text: 14.772 (14.588) Isd_image: 1.8027 (1.6845) Isd_text: 1.8027 (1.6845) Contrastive_loss: 3.5447 (3.5328) Loss: 3.5447 (3.5328)
+2025-04-26,10:03:46 | INFO | Train Epoch: 5 [1646592/3047424 (54%)] Data (t): 0.672 Batch (t): 4.899, 1660.04/s, 415.011/s/gpu LR: 0.000967 Logit Scale: 32.695 Imm_image: 15.030 (14.736) Imm_text: 15.030 (14.736) Isd_image: 2.1234 (1.8308) Isd_text: 2.1234 (1.8308) Contrastive_loss: 3.5942 (3.5532) Loss: 3.5942 (3.5532)
+2025-04-26,10:11:58 | INFO | Train Epoch: 5 [2465792/3047424 (81%)] Data (t): 0.686 Batch (t): 4.917, 1674.45/s, 418.613/s/gpu LR: 0.000963 Logit Scale: 33.383 Imm_image: 15.298 (14.876) Imm_text: 15.298 (14.876) Isd_image: 2.2905 (1.9457) Isd_text: 2.2905 (1.9457) Contrastive_loss: 3.5749 (3.5587) Loss: 3.5749 (3.5587)
+2025-04-26,10:17:46 | INFO | Train Epoch: 5 [3047424/3047424 (100%)] Data (t): 0.674 Batch (t): 4.899, 1710.74/s, 427.686/s/gpu LR: 0.000960 Logit Scale: 33.627 Imm_image: 15.616 (15.024) Imm_text: 15.616 (15.024) Isd_image: 2.4718 (2.0509) Isd_text: 2.4718 (2.0509) Contrastive_loss: 3.3490 (3.5167) Loss: 3.3490 (3.5167)
+2025-04-26,10:17:48 | INFO | Start epoch 6
+2025-04-26,10:18:06 | INFO | Train Epoch: 6 [   8192/3047424 (0%)] Data (t): 13.068 Batch (t): 17.821, 459.673/s, 114.918/s/gpu LR: 0.000960 Logit Scale: 33.632 Imm_image: 15.859 (15.859) Imm_text: 15.859 (15.859) Isd_image: 2.1425 (2.1425) Isd_text: 2.1425 (2.1425) Contrastive_loss: 3.1453 (3.1453) Loss: 3.1453 (3.1453)
+2025-04-26,10:26:21 | INFO | Train Epoch: 6 [ 827392/3047424 (27%)] Data (t): 0.700 Batch (t): 4.944, 1661.20/s, 415.299/s/gpu LR: 0.000955 Logit Scale: 34.652 Imm_image: 16.031 (15.945) Imm_text: 16.031 (15.945) Isd_image: 2.1247 (2.1336) Isd_text: 2.1247 (2.1336) Contrastive_loss: 3.2437 (3.1945) Loss: 3.2437 (3.1945)
+2025-04-26,10:34:33 | INFO | Train Epoch: 6 [1646592/3047424 (54%)] Data (t): 0.696 Batch (t): 4.928, 1658.25/s, 414.563/s/gpu LR: 0.000951 Logit Scale: 35.281 Imm_image: 16.110 (16.000) Imm_text: 16.110 (16.000) Isd_image: 2.1577 (2.1416) Isd_text: 2.1577 (2.1416) Contrastive_loss: 3.2704 (3.2198) Loss: 3.2704 (3.2198)
+2025-04-26,10:42:45 | INFO | Train Epoch: 6 [2465792/3047424 (81%)] Data (t): 0.686 Batch (t): 4.916, 1671.93/s, 417.983/s/gpu LR: 0.000946 Logit Scale: 35.760 Imm_image: 16.435 (16.109) Imm_text: 16.435 (16.109) Isd_image: 2.7697 (2.2986) Isd_text: 2.7697 (2.2986) Contrastive_loss: 3.3002 (3.2399) Loss: 3.3002 (3.2399)
+2025-04-26,10:48:33 | INFO | Train Epoch: 6 [3047424/3047424 (100%)] Data (t): 0.679 Batch (t): 4.907, 1715.68/s, 428.920/s/gpu LR: 0.000943 Logit Scale: 35.992 Imm_image: 16.595 (16.206) Imm_text: 16.595 (16.206) Isd_image: 2.5301 (2.3449) Isd_text: 2.5301 (2.3449) Contrastive_loss: 3.0446 (3.2008) Loss: 3.0446 (3.2008)
+2025-04-26,10:48:36 | INFO | Start epoch 7
+2025-04-26,10:48:53 | INFO | Train Epoch: 7 [   8192/3047424 (0%)] Data (t): 12.751 Batch (t): 17.393, 470.988/s, 117.747/s/gpu LR: 0.000943 Logit Scale: 35.998 Imm_image: 16.872 (16.872) Imm_text: 16.872 (16.872) Isd_image: 2.2662 (2.2662) Isd_text: 2.2662 (2.2662) Contrastive_loss: 2.8284 (2.8284) Loss: 2.8284 (2.8284)
+2025-04-26,10:57:08 | INFO | Train Epoch: 7 [ 827392/3047424 (27%)] Data (t): 0.707 Batch (t): 4.951, 1637.06/s, 409.265/s/gpu LR: 0.000937 Logit Scale: 36.905 Imm_image: 16.835 (16.853) Imm_text: 16.835 (16.853) Isd_image: 2.0720 (2.1691) Isd_text: 2.0720 (2.1691) Contrastive_loss: 2.9946 (2.9115) Loss: 2.9946 (2.9115)
+2025-04-26,11:05:19 | INFO | Train Epoch: 7 [1646592/3047424 (54%)] Data (t): 0.679 Batch (t): 4.911, 1666.05/s, 416.512/s/gpu LR: 0.000932 Logit Scale: 37.496 Imm_image: 17.195 (16.967) Imm_text: 17.195 (16.967) Isd_image: 2.6163 (2.3182) Isd_text: 2.6163 (2.3182) Contrastive_loss: 3.0063 (2.9431) Loss: 3.0063 (2.9431)
+2025-04-26,11:13:32 | INFO | Train Epoch: 7 [2465792/3047424 (81%)] Data (t): 0.690 Batch (t): 4.924, 1665.80/s, 416.449/s/gpu LR: 0.000927 Logit Scale: 37.826 Imm_image: 17.065 (16.992) Imm_text: 17.065 (16.992) Isd_image: 2.7598 (2.4286) Isd_text: 2.7598 (2.4286) Contrastive_loss: 3.1436 (2.9932) Loss: 3.1436 (2.9932)
+2025-04-26,11:19:20 | INFO | Train Epoch: 7 [3047424/3047424 (100%)] Data (t): 0.676 Batch (t): 4.904, 1714.48/s, 428.621/s/gpu LR: 0.000922 Logit Scale: 38.030 Imm_image: 17.659 (17.125) Imm_text: 17.659 (17.125) Isd_image: 2.7269 (2.4882) Isd_text: 2.7269 (2.4882) Contrastive_loss: 2.7796 (2.9505) Loss: 2.7796 (2.9505)
+2025-04-26,11:19:22 | INFO | Start epoch 8
+2025-04-26,11:19:40 | INFO | Train Epoch: 8 [   8192/3047424 (0%)] Data (t): 13.208 Batch (t): 17.719, 462.336/s, 115.584/s/gpu LR: 0.000922 Logit Scale: 38.036 Imm_image: 17.981 (17.981) Imm_text: 17.981 (17.981) Isd_image: 2.5675 (2.5675) Isd_text: 2.5675 (2.5675) Contrastive_loss: 2.5510 (2.5510) Loss: 2.5510 (2.5510)
+2025-04-26,11:27:55 | INFO | Train Epoch: 8 [ 827392/3047424 (27%)] Data (t): 0.701 Batch (t): 4.947, 1666.03/s, 416.508/s/gpu LR: 0.000917 Logit Scale: 38.964 Imm_image: 17.879 (17.930) Imm_text: 17.879 (17.930) Isd_image: 2.2639 (2.4157) Isd_text: 2.2639 (2.4157) Contrastive_loss: 2.7647 (2.6578) Loss: 2.7647 (2.6578)
+2025-04-26,11:36:08 | INFO | Train Epoch: 8 [1646592/3047424 (54%)] Data (t): 0.697 Batch (t): 4.932, 1658.67/s, 414.669/s/gpu LR: 0.000910 Logit Scale: 39.475 Imm_image: 18.072 (17.977) Imm_text: 18.072 (17.977) Isd_image: 2.7349 (2.5221) Isd_text: 2.7349 (2.5221) Contrastive_loss: 2.8409 (2.7188) Loss: 2.8409 (2.7188)
+2025-04-26,11:44:20 | INFO | Train Epoch: 8 [2465792/3047424 (81%)] Data (t): 0.693 Batch (t): 4.926, 1669.08/s, 417.269/s/gpu LR: 0.000904 Logit Scale: 39.705 Imm_image: 18.058 (17.997) Imm_text: 18.058 (17.997) Isd_image: 2.9109 (2.6193) Isd_text: 2.9109 (2.6193) Contrastive_loss: 2.8937 (2.7625) Loss: 2.8937 (2.7625)
+2025-04-26,11:50:09 | INFO | Train Epoch: 8 [3047424/3047424 (100%)] Data (t): 0.684 Batch (t): 4.911, 1716.61/s, 429.153/s/gpu LR: 0.000900 Logit Scale: 39.880 Imm_image: 18.549 (18.108) Imm_text: 18.549 (18.108) Isd_image: 3.1811 (2.7317) Isd_text: 3.1811 (2.7317) Contrastive_loss: 2.5660 (2.7232) Loss: 2.5660 (2.7232)
+2025-04-26,11:50:11 | INFO | Start epoch 9
+2025-04-26,11:50:29 | INFO | Train Epoch: 9 [   8192/3047424 (0%)] Data (t): 13.559 Batch (t): 18.083, 453.019/s, 113.255/s/gpu LR: 0.000900 Logit Scale: 39.885 Imm_image: 18.721 (18.721) Imm_text: 18.721 (18.721) Isd_image: 2.8998 (2.8998) Isd_text: 2.8998 (2.8998) Contrastive_loss: 2.4689 (2.4689) Loss: 2.4689 (2.4689)
+2025-04-26,11:58:44 | INFO | Train Epoch: 9 [ 827392/3047424 (27%)] Data (t): 0.703 Batch (t): 4.951, 1662.79/s, 415.697/s/gpu LR: 0.000893 Logit Scale: 40.810 Imm_image: 18.943 (18.832) Imm_text: 18.943 (18.832) Isd_image: 2.7050 (2.8024) Isd_text: 2.7050 (2.8024) Contrastive_loss: 2.5157 (2.4923) Loss: 2.5157 (2.4923)
+2025-04-26,12:06:58 | INFO | Train Epoch: 9 [1646592/3047424 (54%)] Data (t): 0.698 Batch (t): 4.934, 1663.60/s, 415.899/s/gpu LR: 0.000886 Logit Scale: 41.261 Imm_image: 18.963 (18.876) Imm_text: 18.963 (18.876) Isd_image: 2.9714 (2.8587) Isd_text: 2.9714 (2.8587) Contrastive_loss: 2.5983 (2.5276) Loss: 2.5983 (2.5276)
+2025-04-26,12:15:11 | INFO | Train Epoch: 9 [2465792/3047424 (81%)] Data (t): 0.694 Batch (t): 4.929, 1620.47/s, 405.117/s/gpu LR: 0.000879 Logit Scale: 41.381 Imm_image: 18.889 (18.879) Imm_text: 18.889 (18.879) Isd_image: 3.1413 (2.9294) Isd_text: 3.1413 (2.9294) Contrastive_loss: 2.6602 (2.5608) Loss: 2.6602 (2.5608)
+2025-04-26,12:20:58 | INFO | Train Epoch: 9 [3047424/3047424 (100%)] Data (t): 0.662 Batch (t): 4.886, 1714.91/s, 428.727/s/gpu LR: 0.000874 Logit Scale: 41.499 Imm_image: 19.170 (18.937) Imm_text: 19.170 (18.937) Isd_image: 3.2043 (2.9844) Isd_text: 3.2043 (2.9844) Contrastive_loss: 2.3764 (2.5239) Loss: 2.3764 (2.5239)
+2025-04-26,12:21:00 | INFO | Start epoch 10
+2025-04-26,12:21:17 | INFO | Train Epoch: 10 [   8192/3047424 (0%)] Data (t): 12.157 Batch (t): 16.656, 491.837/s, 122.959/s/gpu LR: 0.000874 Logit Scale: 41.505 Imm_image: 19.450 (19.450) Imm_text: 19.450 (19.450) Isd_image: 3.0018 (3.0018) Isd_text: 3.0018 (3.0018) Contrastive_loss: 2.2344 (2.2344) Loss: 2.2344 (2.2344)
+2025-04-26,12:29:30 | INFO | Train Epoch: 10 [ 827392/3047424 (27%)] Data (t): 0.697 Batch (t): 4.938, 1666.22/s, 416.556/s/gpu LR: 0.000867 Logit Scale: 42.476 Imm_image: 19.624 (19.537) Imm_text: 19.624 (19.537) Isd_image: 2.6374 (2.8196) Isd_text: 2.6374 (2.8196) Contrastive_loss: 2.3437 (2.2890) Loss: 2.3437 (2.2890)
+2025-04-26,12:37:43 | INFO | Train Epoch: 10 [1646592/3047424 (54%)] Data (t): 0.693 Batch (t): 4.928, 1666.23/s, 416.558/s/gpu LR: 0.000859 Logit Scale: 42.828 Imm_image: 19.572 (19.548) Imm_text: 19.572 (19.548) Isd_image: 3.1196 (2.9196) Isd_text: 3.1196 (2.9196) Contrastive_loss: 2.4127 (2.3302) Loss: 2.4127 (2.3302)
+2025-04-26,12:45:55 | INFO | Train Epoch: 10 [2465792/3047424 (81%)] Data (t): 0.692 Batch (t): 4.922, 1672.42/s, 418.106/s/gpu LR: 0.000852 Logit Scale: 42.987 Imm_image: 19.576 (19.555) Imm_text: 19.576 (19.555) Isd_image: 3.3384 (3.0243) Isd_text: 3.3384 (3.0243) Contrastive_loss: 2.5498 (2.3851) Loss: 2.5498 (2.3851)
+2025-04-26,12:51:44 | INFO | Train Epoch: 10 [3047424/3047424 (100%)] Data (t): 0.679 Batch (t): 4.910, 1707.36/s, 426.840/s/gpu LR: 0.000846 Logit Scale: 43.138 Imm_image: 20.022 (19.649) Imm_text: 20.022 (19.649) Isd_image: 3.2553 (3.0705) Isd_text: 3.2553 (3.0705) Contrastive_loss: 2.2203 (2.3522) Loss: 2.2203 (2.3522)
+2025-04-26,12:51:47 | INFO | Start epoch 11
+2025-04-26,12:52:04 | INFO | Train Epoch: 11 [   8192/3047424 (0%)] Data (t): 13.250 Batch (t): 17.785, 460.602/s, 115.150/s/gpu LR: 0.000846 Logit Scale: 43.141 Imm_image: 20.268 (20.268) Imm_text: 20.268 (20.268) Isd_image: 3.3257 (3.3257) Isd_text: 3.3257 (3.3257) Contrastive_loss: 2.0359 (2.0359) Loss: 2.0359 (2.0359)
+2025-04-26,13:00:19 | INFO | Train Epoch: 11 [ 827392/3047424 (27%)] Data (t): 0.705 Batch (t): 4.948, 1668.54/s, 417.136/s/gpu LR: 0.000838 Logit Scale: 44.100 Imm_image: 20.182 (20.225) Imm_text: 20.182 (20.225) Isd_image: 2.6004 (2.9630) Isd_text: 2.6004 (2.9630) Contrastive_loss: 2.2182 (2.1270) Loss: 2.2182 (2.1270)
+2025-04-26,13:08:33 | INFO | Train Epoch: 11 [1646592/3047424 (54%)] Data (t): 0.702 Batch (t): 4.942, 1651.65/s, 412.913/s/gpu LR: 0.000830 Logit Scale: 44.550 Imm_image: 20.331 (20.260) Imm_text: 20.331 (20.260) Isd_image: 2.8698 (2.9320) Isd_text: 2.8698 (2.9320) Contrastive_loss: 2.3181 (2.1907) Loss: 2.3181 (2.1907)
+2025-04-26,13:16:43 | INFO | Train Epoch: 11 [2465792/3047424 (81%)] Data (t): 0.668 Batch (t): 4.896, 1676.31/s, 419.078/s/gpu LR: 0.000822 Logit Scale: 44.520 Imm_image: 20.413 (20.299) Imm_text: 20.413 (20.299) Isd_image: 3.3902 (3.0465) Isd_text: 3.3902 (3.0465) Contrastive_loss: 2.3573 (2.2324) Loss: 2.3573 (2.2324)
+2025-04-26,13:22:31 | INFO | Train Epoch: 11 [3047424/3047424 (100%)] Data (t): 0.676 Batch (t): 4.903, 1715.16/s, 428.790/s/gpu LR: 0.000816 Logit Scale: 44.644 Imm_image: 20.852 (20.409) Imm_text: 20.852 (20.409) Isd_image: 3.2356 (3.0843) Isd_text: 3.2356 (3.0843) Contrastive_loss: 2.0466 (2.1952) Loss: 2.0466 (2.1952)
+2025-04-26,13:22:34 | INFO | Start epoch 12
+2025-04-26,13:22:51 | INFO | Train Epoch: 12 [   8192/3047424 (0%)] Data (t): 12.990 Batch (t): 17.512, 467.784/s, 116.946/s/gpu LR: 0.000816 Logit Scale: 44.648 Imm_image: 21.309 (21.309) Imm_text: 21.309 (21.309) Isd_image: 3.0604 (3.0604) Isd_text: 3.0604 (3.0604) Contrastive_loss: 1.7571 (1.7571) Loss: 1.7571 (1.7571)
+2025-04-26,13:31:06 | INFO | Train Epoch: 12 [ 827392/3047424 (27%)] Data (t): 0.704 Batch (t): 4.950, 1657.07/s, 414.269/s/gpu LR: 0.000808 Logit Scale: 45.764 Imm_image: 21.251 (21.280) Imm_text: 21.251 (21.280) Isd_image: 2.6780 (2.8692) Isd_text: 2.6780 (2.8692) Contrastive_loss: 2.0212 (1.8892) Loss: 2.0212 (1.8892)
+2025-04-26,13:39:19 | INFO | Train Epoch: 12 [1646592/3047424 (54%)] Data (t): 0.694 Batch (t): 4.926, 1667.01/s, 416.753/s/gpu LR: 0.000799 Logit Scale: 46.051 Imm_image: 21.089 (21.216) Imm_text: 21.089 (21.216) Isd_image: 3.0048 (2.9144) Isd_text: 3.0048 (2.9144) Contrastive_loss: 2.2064 (1.9949) Loss: 2.2064 (1.9949)
+2025-04-26,13:47:31 | INFO | Train Epoch: 12 [2465792/3047424 (81%)] Data (t): 0.694 Batch (t): 4.928, 1667.46/s, 416.865/s/gpu LR: 0.000790 Logit Scale: 45.927 Imm_image: 20.944 (21.148) Imm_text: 20.944 (21.148) Isd_image: 3.6934 (3.1091) Isd_text: 3.6934 (3.1091) Contrastive_loss: 2.2147 (2.0499) Loss: 2.2147 (2.0499)
+2025-04-26,13:53:20 | INFO | Train Epoch: 12 [3047424/3047424 (100%)] Data (t): 0.680 Batch (t): 4.909, 1709.31/s, 427.329/s/gpu LR: 0.000784 Logit Scale: 46.025 Imm_image: 21.522 (21.223) Imm_text: 21.522 (21.223) Isd_image: 3.3328 (3.1539) Isd_text: 3.3328 (3.1539) Contrastive_loss: 1.8286 (2.0056) Loss: 1.8286 (2.0056)
+2025-04-26,13:53:22 | INFO | Start epoch 13
+2025-04-26,13:53:40 | INFO | Train Epoch: 13 [   8192/3047424 (0%)] Data (t): 13.360 Batch (t): 18.005, 454.990/s, 113.748/s/gpu LR: 0.000784 Logit Scale: 46.031 Imm_image: 21.951 (21.951) Imm_text: 21.951 (21.951) Isd_image: 3.1097 (3.1097) Isd_text: 3.1097 (3.1097) Contrastive_loss: 1.6031 (1.6031) Loss: 1.6031 (1.6031)
+2025-04-26,14:01:56 | INFO | Train Epoch: 13 [ 827392/3047424 (27%)] Data (t): 0.705 Batch (t): 4.953, 1657.66/s, 414.415/s/gpu LR: 0.000775 Logit Scale: 47.201 Imm_image: 21.839 (21.895) Imm_text: 21.839 (21.895) Isd_image: 2.6220 (2.8658) Isd_text: 2.6220 (2.8658) Contrastive_loss: 1.9037 (1.7534) Loss: 1.9037 (1.7534)
+2025-04-26,14:10:08 | INFO | Train Epoch: 13 [1646592/3047424 (54%)] Data (t): 0.693 Batch (t): 4.927, 1659.33/s, 414.834/s/gpu LR: 0.000766 Logit Scale: 47.480 Imm_image: 21.781 (21.857) Imm_text: 21.781 (21.857) Isd_image: 3.5425 (3.0914) Isd_text: 3.5425 (3.0914) Contrastive_loss: 1.9959 (1.8343) Loss: 1.9959 (1.8343)
+2025-04-26,14:18:21 | INFO | Train Epoch: 13 [2465792/3047424 (81%)] Data (t): 0.691 Batch (t): 4.923, 1671.56/s, 417.889/s/gpu LR: 0.000756 Logit Scale: 47.338 Imm_image: 21.681 (21.813) Imm_text: 21.681 (21.813) Isd_image: 3.5550 (3.2073) Isd_text: 3.5550 (3.2073) Contrastive_loss: 2.0507 (1.8884) Loss: 2.0507 (1.8884)
+2025-04-26,14:24:09 | INFO | Train Epoch: 13 [3047424/3047424 (100%)] Data (t): 0.676 Batch (t): 4.906, 1710.02/s, 427.505/s/gpu LR: 0.000750 Logit Scale: 47.442 Imm_image: 22.179 (21.886) Imm_text: 22.179 (21.886) Isd_image: 3.4302 (3.2519) Isd_text: 3.4302 (3.2519) Contrastive_loss: 1.7243 (1.8556) Loss: 1.7243 (1.8556)
+2025-04-26,14:24:11 | INFO | Start epoch 14
+2025-04-26,14:24:29 | INFO | Train Epoch: 14 [   8192/3047424 (0%)] Data (t): 13.414 Batch (t): 18.126, 451.944/s, 112.986/s/gpu LR: 0.000750 Logit Scale: 47.446 Imm_image: 22.608 (22.608) Imm_text: 22.608 (22.608) Isd_image: 3.3672 (3.3672) Isd_text: 3.3672 (3.3672) Contrastive_loss: 1.4806 (1.4806) Loss: 1.4806 (1.4806)
+2025-04-26,14:32:44 | INFO | Train Epoch: 14 [ 827392/3047424 (27%)] Data (t): 0.707 Batch (t): 4.952, 1672.64/s, 418.160/s/gpu LR: 0.000740 Logit Scale: 48.607 Imm_image: 22.564 (22.586) Imm_text: 22.564 (22.586) Isd_image: 2.6599 (3.0135) Isd_text: 2.6599 (3.0135) Contrastive_loss: 1.7005 (1.5906) Loss: 1.7005 (1.5906)
+2025-04-26,14:40:58 | INFO | Train Epoch: 14 [1646592/3047424 (54%)] Data (t): 0.699 Batch (t): 4.935, 1669.22/s, 417.304/s/gpu LR: 0.000731 Logit Scale: 48.857 Imm_image: 22.620 (22.598) Imm_text: 22.620 (22.598) Isd_image: 3.3316 (3.1196) Isd_text: 3.3316 (3.1196) Contrastive_loss: 1.8648 (1.6820) Loss: 1.8648 (1.6820)
+2025-04-26,14:49:10 | INFO | Train Epoch: 14 [2465792/3047424 (81%)] Data (t): 0.690 Batch (t): 4.920, 1670.45/s, 417.612/s/gpu LR: 0.000721 Logit Scale: 48.682 Imm_image: 22.395 (22.547) Imm_text: 22.395 (22.547) Isd_image: 3.5906 (3.2373) Isd_text: 3.5906 (3.2373) Contrastive_loss: 1.9201 (1.7415) Loss: 1.9201 (1.7415)
+2025-04-26,14:54:58 | INFO | Train Epoch: 14 [3047424/3047424 (100%)] Data (t): 0.673 Batch (t): 4.897, 1716.16/s, 429.041/s/gpu LR: 0.000714 Logit Scale: 48.834 Imm_image: 22.965 (22.630) Imm_text: 22.965 (22.630) Isd_image: 3.4823 (3.2863) Isd_text: 3.4823 (3.2863) Contrastive_loss: 1.5692 (1.7071) Loss: 1.5692 (1.7071)
+2025-04-26,14:55:00 | INFO | Start epoch 15
+2025-04-26,14:55:18 | INFO | Train Epoch: 15 [   8192/3047424 (0%)] Data (t): 13.407 Batch (t): 17.985, 455.489/s, 113.872/s/gpu LR: 0.000714 Logit Scale: 48.839 Imm_image: 23.288 (23.288) Imm_text: 23.288 (23.288) Isd_image: 3.2415 (3.2415) Isd_text: 3.2415 (3.2415) Contrastive_loss: 1.4186 (1.4186) Loss: 1.4186 (1.4186)
+2025-04-26,15:03:33 | INFO | Train Epoch: 15 [ 827392/3047424 (27%)] Data (t): 0.703 Batch (t): 4.949, 1661.73/s, 415.432/s/gpu LR: 0.000704 Logit Scale: 50.058 Imm_image: 23.318 (23.303) Imm_text: 23.318 (23.303) Isd_image: 2.6949 (2.9682) Isd_text: 2.6949 (2.9682) Contrastive_loss: 1.5847 (1.5017) Loss: 1.5847 (1.5017)
+2025-04-26,15:11:46 | INFO | Train Epoch: 15 [1646592/3047424 (54%)] Data (t): 0.695 Batch (t): 4.932, 1661.28/s, 415.320/s/gpu LR: 0.000694 Logit Scale: 50.269 Imm_image: 23.286 (23.297) Imm_text: 23.286 (23.297) Isd_image: 3.1811 (3.0392) Isd_text: 3.1811 (3.0392) Contrastive_loss: 1.6936 (1.5657) Loss: 1.6936 (1.5657)
+2025-04-26,15:19:59 | INFO | Train Epoch: 15 [2465792/3047424 (81%)] Data (t): 0.692 Batch (t): 4.928, 1675.83/s, 418.958/s/gpu LR: 0.000684 Logit Scale: 50.091 Imm_image: 23.214 (23.277) Imm_text: 23.214 (23.277) Isd_image: 3.4567 (3.1436) Isd_text: 3.4567 (3.1436) Contrastive_loss: 1.7579 (1.6137) Loss: 1.7579 (1.6137)
+2025-04-26,15:25:46 | INFO | Train Epoch: 15 [3047424/3047424 (100%)] Data (t): 0.660 Batch (t): 4.882, 1715.82/s, 428.956/s/gpu LR: 0.000677 Logit Scale: 50.211 Imm_image: 23.706 (23.362) Imm_text: 23.706 (23.362) Isd_image: 3.5545 (3.2257) Isd_text: 3.5545 (3.2257) Contrastive_loss: 1.4196 (1.5749) Loss: 1.4196 (1.5749)
+2025-04-26,15:25:48 | INFO | Start epoch 16
+2025-04-26,15:26:06 | INFO | Train Epoch: 16 [   8192/3047424 (0%)] Data (t): 13.522 Batch (t): 18.104, 452.489/s, 113.122/s/gpu LR: 0.000677 Logit Scale: 50.215 Imm_image: 24.248 (24.248) Imm_text: 24.248 (24.248) Isd_image: 3.2084 (3.2084) Isd_text: 3.2084 (3.2084) Contrastive_loss: 1.1404 (1.1404) Loss: 1.1404 (1.1404)
+2025-04-26,15:34:20 | INFO | Train Epoch: 16 [ 827392/3047424 (27%)] Data (t): 0.704 Batch (t): 4.945, 1669.59/s, 417.399/s/gpu LR: 0.000667 Logit Scale: 51.514 Imm_image: 24.132 (24.190) Imm_text: 24.132 (24.190) Isd_image: 2.4709 (2.8396) Isd_text: 2.4709 (2.8396) Contrastive_loss: 1.4101 (1.2752) Loss: 1.4101 (1.2752)
+2025-04-26,15:42:34 | INFO | Train Epoch: 16 [1646592/3047424 (54%)] Data (t): 0.697 Batch (t): 4.933, 1668.92/s, 417.231/s/gpu LR: 0.000657 Logit Scale: 51.663 Imm_image: 23.885 (24.089) Imm_text: 23.885 (24.089) Isd_image: 2.9497 (2.8763) Isd_text: 2.9497 (2.8763) Contrastive_loss: 1.5870 (1.3792) Loss: 1.5870 (1.3792)
+2025-04-26,15:50:46 | INFO | Train Epoch: 16 [2465792/3047424 (81%)] Data (t): 0.693 Batch (t): 4.926, 1664.82/s, 416.205/s/gpu LR: 0.000646 Logit Scale: 51.417 Imm_image: 23.581 (23.962) Imm_text: 23.581 (23.962) Isd_image: 3.0984 (2.9318) Isd_text: 3.0984 (2.9318) Contrastive_loss: 1.6678 (1.4513) Loss: 1.6678 (1.4513)
+2025-04-26,15:56:35 | INFO | Train Epoch: 16 [3047424/3047424 (100%)] Data (t): 0.684 Batch (t): 4.913, 1711.28/s, 427.820/s/gpu LR: 0.000639 Logit Scale: 51.483 Imm_image: 24.422 (24.054) Imm_text: 24.422 (24.054) Isd_image: 3.4108 (3.0276) Isd_text: 3.4108 (3.0276) Contrastive_loss: 1.2393 (1.4089) Loss: 1.2393 (1.4089)
+2025-04-26,15:56:38 | INFO | Start epoch 17
+2025-04-26,15:56:56 | INFO | Train Epoch: 17 [   8192/3047424 (0%)] Data (t): 13.278 Batch (t): 17.872, 458.379/s, 114.595/s/gpu LR: 0.000639 Logit Scale: 51.488 Imm_image: 24.688 (24.688) Imm_text: 24.688 (24.688) Isd_image: 3.2552 (3.2552) Isd_text: 3.2552 (3.2552) Contrastive_loss: 1.1017 (1.1017) Loss: 1.1017 (1.1017)
+2025-04-26,16:05:11 | INFO | Train Epoch: 17 [ 827392/3047424 (27%)] Data (t): 0.709 Batch (t): 4.956, 1659.37/s, 414.841/s/gpu LR: 0.000628 Logit Scale: 52.929 Imm_image: 24.952 (24.820) Imm_text: 24.952 (24.820) Isd_image: 2.3677 (2.8114) Isd_text: 2.3677 (2.8114) Contrastive_loss: 1.2202 (1.1610) Loss: 1.2202 (1.1610)
+2025-04-26,16:13:24 | INFO | Train Epoch: 17 [1646592/3047424 (54%)] Data (t): 0.695 Batch (t): 4.929, 1662.90/s, 415.724/s/gpu LR: 0.000618 Logit Scale: 53.074 Imm_image: 24.670 (24.770) Imm_text: 24.670 (24.770) Isd_image: 2.6086 (2.7438) Isd_text: 2.6086 (2.7438) Contrastive_loss: 1.4462 (1.2560) Loss: 1.4462 (1.2560)
+2025-04-26,16:21:37 | INFO | Train Epoch: 17 [2465792/3047424 (81%)] Data (t): 0.691 Batch (t): 4.927, 1674.74/s, 418.686/s/gpu LR: 0.000607 Logit Scale: 52.849 Imm_image: 24.405 (24.679) Imm_text: 24.405 (24.679) Isd_image: 2.9512 (2.7956) Isd_text: 2.9512 (2.7956) Contrastive_loss: 1.4544 (1.3056) Loss: 1.4544 (1.3056)
+2025-04-26,16:27:24 | INFO | Train Epoch: 17 [3047424/3047424 (100%)] Data (t): 0.664 Batch (t): 4.889, 1718.16/s, 429.541/s/gpu LR: 0.000600 Logit Scale: 52.998 Imm_image: 25.070 (24.757) Imm_text: 25.070 (24.757) Isd_image: 3.2205 (2.8806) Isd_text: 3.2205 (2.8806) Contrastive_loss: 1.1236 (1.2692) Loss: 1.1236 (1.2692)
+2025-04-26,16:27:26 | INFO | Start epoch 18
+2025-04-26,16:27:44 | INFO | Train Epoch: 18 [   8192/3047424 (0%)] Data (t): 12.940 Batch (t): 17.521, 467.558/s, 116.890/s/gpu LR: 0.000600 Logit Scale: 53.002 Imm_image: 25.673 (25.673) Imm_text: 25.673 (25.673) Isd_image: 3.0190 (3.0190) Isd_text: 3.0190 (3.0190) Contrastive_loss: 0.85869 (0.85869) Loss: 0.85869 (0.85869)
+2025-04-26,16:35:58 | INFO | Train Epoch: 18 [ 827392/3047424 (27%)] Data (t): 0.699 Batch (t): 4.941, 1669.52/s, 417.380/s/gpu LR: 0.000589 Logit Scale: 54.468 Imm_image: 25.704 (25.689) Imm_text: 25.704 (25.689) Isd_image: 2.1135 (2.5663) Isd_text: 2.1135 (2.5663) Contrastive_loss: 1.0909 (0.97477) Loss: 1.0909 (0.97477)
+2025-04-26,16:44:10 | INFO | Train Epoch: 18 [1646592/3047424 (54%)] Data (t): 0.691 Batch (t): 4.924, 1665.67/s, 416.419/s/gpu LR: 0.000578 Logit Scale: 54.710 Imm_image: 25.605 (25.661) Imm_text: 25.605 (25.661) Isd_image: 2.9780 (2.7035) Isd_text: 2.9780 (2.7035) Contrastive_loss: 1.2648 (1.0714) Loss: 1.2648 (1.0714)
+2025-04-26,16:52:21 | INFO | Train Epoch: 18 [2465792/3047424 (81%)] Data (t): 0.676 Batch (t): 4.907, 1673.47/s, 418.369/s/gpu LR: 0.000568 Logit Scale: 54.239 Imm_image: 25.176 (25.540) Imm_text: 25.176 (25.540) Isd_image: 2.8536 (2.7410) Isd_text: 2.8536 (2.7410) Contrastive_loss: 1.3238 (1.1345) Loss: 1.3238 (1.1345)
+2025-04-26,16:58:09 | INFO | Train Epoch: 18 [3047424/3047424 (100%)] Data (t): 0.679 Batch (t): 4.902, 1708.41/s, 427.102/s/gpu LR: 0.000560 Logit Scale: 54.470 Imm_image: 26.030 (25.638) Imm_text: 26.030 (25.638) Isd_image: 2.7456 (2.7419) Isd_text: 2.7456 (2.7419) Contrastive_loss: 0.95385 (1.0984) Loss: 0.95385 (1.0984)
+2025-04-26,16:58:12 | INFO | Start epoch 19
+2025-04-26,16:58:30 | INFO | Train Epoch: 19 [   8192/3047424 (0%)] Data (t): 13.122 Batch (t): 17.721, 462.271/s, 115.568/s/gpu LR: 0.000560 Logit Scale: 54.472 Imm_image: 26.484 (26.484) Imm_text: 26.484 (26.484) Isd_image: 2.6400 (2.6400) Isd_text: 2.6400 (2.6400) Contrastive_loss: 0.79116 (0.79116) Loss: 0.79116 (0.79116)
+2025-04-26,17:06:44 | INFO | Train Epoch: 19 [ 827392/3047424 (27%)] Data (t): 0.705 Batch (t): 4.945, 1666.17/s, 416.543/s/gpu LR: 0.000549 Logit Scale: 56.006 Imm_image: 26.531 (26.507) Imm_text: 26.531 (26.507) Isd_image: 1.8994 (2.2697) Isd_text: 1.8994 (2.2697) Contrastive_loss: 0.97074 (0.88095) Loss: 0.97074 (0.88095)
+2025-04-26,17:14:57 | INFO | Train Epoch: 19 [1646592/3047424 (54%)] Data (t): 0.699 Batch (t): 4.931, 1660.66/s, 415.165/s/gpu LR: 0.000538 Logit Scale: 56.197 Imm_image: 26.343 (26.453) Imm_text: 26.343 (26.453) Isd_image: 2.4889 (2.3428) Isd_text: 2.4889 (2.3428) Contrastive_loss: 1.1094 (0.95709) Loss: 1.1094 (0.95709)
+2025-04-26,17:23:10 | INFO | Train Epoch: 19 [2465792/3047424 (81%)] Data (t): 0.692 Batch (t): 4.925, 1678.80/s, 419.701/s/gpu LR: 0.000528 Logit Scale: 56.028 Imm_image: 26.163 (26.380) Imm_text: 26.163 (26.380) Isd_image: 2.7418 (2.4425) Isd_text: 2.7418 (2.4425) Contrastive_loss: 1.1579 (1.0073) Loss: 1.1579 (1.0073)
+2025-04-26,17:28:58 | INFO | Train Epoch: 19 [3047424/3047424 (100%)] Data (t): 0.678 Batch (t): 4.906, 1708.30/s, 427.075/s/gpu LR: 0.000520 Logit Scale: 56.144 Imm_image: 26.881 (26.480) Imm_text: 26.881 (26.480) Isd_image: 2.6809 (2.4902) Isd_text: 2.6809 (2.4902) Contrastive_loss: 0.84861 (0.97555) Loss: 0.84861 (0.97555)
+2025-04-26,17:29:01 | INFO | Start epoch 20
+2025-04-26,17:29:19 | INFO | Train Epoch: 20 [   8192/3047424 (0%)] Data (t): 13.520 Batch (t): 18.149, 451.373/s, 112.843/s/gpu LR: 0.000520 Logit Scale: 56.147 Imm_image: 27.206 (27.206) Imm_text: 27.206 (27.206) Isd_image: 2.5496 (2.5496) Isd_text: 2.5496 (2.5496) Contrastive_loss: 0.72025 (0.72025) Loss: 0.72025 (0.72025)
+2025-04-26,17:37:34 | INFO | Train Epoch: 20 [ 827392/3047424 (27%)] Data (t): 0.705 Batch (t): 4.949, 1663.14/s, 415.784/s/gpu LR: 0.000509 Logit Scale: 57.831 Imm_image: 27.343 (27.275) Imm_text: 27.343 (27.275) Isd_image: 1.5551 (2.0524) Isd_text: 1.5551 (2.0524) Contrastive_loss: 0.84613 (0.78319) Loss: 0.84613 (0.78319)
+2025-04-26,17:45:47 | INFO | Train Epoch: 20 [1646592/3047424 (54%)] Data (t): 0.693 Batch (t): 4.930, 1660.35/s, 415.088/s/gpu LR: 0.000498 Logit Scale: 57.940 Imm_image: 27.171 (27.240) Imm_text: 27.171 (27.240) Isd_image: 2.4143 (2.1730) Isd_text: 2.4143 (2.1730) Contrastive_loss: 0.96746 (0.84461) Loss: 0.96746 (0.84461)
+2025-04-26,17:53:59 | INFO | Train Epoch: 20 [2465792/3047424 (81%)] Data (t): 0.694 Batch (t): 4.929, 1672.77/s, 418.193/s/gpu LR: 0.000487 Logit Scale: 57.662 Imm_image: 26.990 (27.178) Imm_text: 26.990 (27.178) Isd_image: 2.3142 (2.2083) Isd_text: 2.3142 (2.2083) Contrastive_loss: 1.0177 (0.88788) Loss: 1.0177 (0.88788)
+2025-04-26,17:59:48 | INFO | Train Epoch: 20 [3047424/3047424 (100%)] Data (t): 0.682 Batch (t): 4.909, 1709.94/s, 427.484/s/gpu LR: 0.000480 Logit Scale: 57.922 Imm_image: 27.752 (27.292) Imm_text: 27.752 (27.292) Isd_image: 2.2081 (2.2083) Isd_text: 2.2081 (2.2083) Contrastive_loss: 0.69428 (0.84916) Loss: 0.69428 (0.84916)
+2025-04-26,17:59:51 | INFO | Start epoch 21
+2025-04-26,18:00:09 | INFO | Train Epoch: 21 [   8192/3047424 (0%)] Data (t): 13.483 Batch (t): 18.118, 452.145/s, 113.036/s/gpu LR: 0.000480 Logit Scale: 57.928 Imm_image: 28.137 (28.137) Imm_text: 28.137 (28.137) Isd_image: 2.1418 (2.1418) Isd_text: 2.1418 (2.1418) Contrastive_loss: 0.57104 (0.57104) Loss: 0.57104 (0.57104)
+2025-04-26,18:08:24 | INFO | Train Epoch: 21 [ 827392/3047424 (27%)] Data (t): 0.709 Batch (t): 4.955, 1661.64/s, 415.410/s/gpu LR: 0.000469 Logit Scale: 59.657 Imm_image: 28.456 (28.296) Imm_text: 28.456 (28.296) Isd_image: 1.5023 (1.8220) Isd_text: 1.5023 (1.8220) Contrastive_loss: 0.72510 (0.64807) Loss: 0.72510 (0.64807)
+2025-04-26,18:16:37 | INFO | Train Epoch: 21 [1646592/3047424 (54%)] Data (t): 0.696 Batch (t): 4.930, 1671.51/s, 417.878/s/gpu LR: 0.000458 Logit Scale: 59.773 Imm_image: 28.309 (28.301) Imm_text: 28.309 (28.301) Isd_image: 2.1665 (1.9369) Isd_text: 2.1665 (1.9369) Contrastive_loss: 0.83912 (0.71175) Loss: 0.83912 (0.71175)
+2025-04-26,18:24:49 | INFO | Train Epoch: 21 [2465792/3047424 (81%)] Data (t): 0.690 Batch (t): 4.921, 1676.75/s, 419.187/s/gpu LR: 0.000447 Logit Scale: 59.544 Imm_image: 28.015 (28.229) Imm_text: 28.015 (28.229) Isd_image: 2.3414 (2.0380) Isd_text: 2.3414 (2.0380) Contrastive_loss: 0.88618 (0.75536) Loss: 0.88618 (0.75536)
+2025-04-26,18:30:38 | INFO | Train Epoch: 21 [3047424/3047424 (100%)] Data (t): 0.681 Batch (t): 4.907, 1709.13/s, 427.283/s/gpu LR: 0.000440 Logit Scale: 59.761 Imm_image: 28.689 (28.321) Imm_text: 28.689 (28.321) Isd_image: 2.2627 (2.0829) Isd_text: 2.2627 (2.0829) Contrastive_loss: 0.61415 (0.72712) Loss: 0.61415 (0.72712)
+2025-04-26,18:30:40 | INFO | Start epoch 22
+2025-04-26,18:30:58 | INFO | Train Epoch: 22 [   8192/3047424 (0%)] Data (t): 13.550 Batch (t): 18.163, 451.017/s, 112.754/s/gpu LR: 0.000440 Logit Scale: 59.765 Imm_image: 29.215 (29.215) Imm_text: 29.215 (29.215) Isd_image: 2.0924 (2.0924) Isd_text: 2.0924 (2.0924) Contrastive_loss: 0.49251 (0.49251) Loss: 0.49251 (0.49251)
+2025-04-26,18:39:13 | INFO | Train Epoch: 22 [ 827392/3047424 (27%)] Data (t): 0.705 Batch (t): 4.949, 1659.03/s, 414.758/s/gpu LR: 0.000429 Logit Scale: 61.553 Imm_image: 29.523 (29.369) Imm_text: 29.523 (29.369) Isd_image: 1.1648 (1.6286) Isd_text: 1.1648 (1.6286) Contrastive_loss: 0.59226 (0.54238) Loss: 0.59226 (0.54238)
+2025-04-26,18:47:26 | INFO | Train Epoch: 22 [1646592/3047424 (54%)] Data (t): 0.698 Batch (t): 4.931, 1662.28/s, 415.571/s/gpu LR: 0.000418 Logit Scale: 61.667 Imm_image: 29.247 (29.328) Imm_text: 29.247 (29.328) Isd_image: 1.9847 (1.7473) Isd_text: 1.9847 (1.7473) Contrastive_loss: 0.72890 (0.60456) Loss: 0.72890 (0.60456)
+2025-04-26,18:55:39 | INFO | Train Epoch: 22 [2465792/3047424 (81%)] Data (t): 0.694 Batch (t): 4.927, 1678.34/s, 419.586/s/gpu LR: 0.000407 Logit Scale: 61.313 Imm_image: 28.947 (29.233) Imm_text: 28.947 (29.233) Isd_image: 2.0178 (1.8149) Isd_text: 2.0178 (1.8149) Contrastive_loss: 0.80285 (0.65413) Loss: 0.80285 (0.65413)
+2025-04-26,19:01:28 | INFO | Train Epoch: 22 [3047424/3047424 (100%)] Data (t): 0.681 Batch (t): 4.910, 1708.12/s, 427.030/s/gpu LR: 0.000400 Logit Scale: 61.651 Imm_image: 29.803 (29.347) Imm_text: 29.803 (29.347) Isd_image: 1.9922 (1.8504) Isd_text: 1.9922 (1.8504) Contrastive_loss: 0.49208 (0.62172) Loss: 0.49208 (0.62172)
+2025-04-26,19:01:30 | INFO | Start epoch 23
+2025-04-26,19:01:48 | INFO | Train Epoch: 23 [   8192/3047424 (0%)] Data (t): 13.608 Batch (t): 18.178, 450.666/s, 112.666/s/gpu LR: 0.000400 Logit Scale: 61.655 Imm_image: 30.026 (30.026) Imm_text: 30.026 (30.026) Isd_image: 1.9201 (1.9201) Isd_text: 1.9201 (1.9201) Contrastive_loss: 0.44105 (0.44105) Loss: 0.44105 (0.44105)
+2025-04-26,19:10:03 | INFO | Train Epoch: 23 [ 827392/3047424 (27%)] Data (t): 0.705 Batch (t): 4.950, 1666.57/s, 416.643/s/gpu LR: 0.000389 Logit Scale: 63.412 Imm_image: 30.289 (30.158) Imm_text: 30.289 (30.158) Isd_image: 1.0218 (1.4709) Isd_text: 1.0218 (1.4709) Contrastive_loss: 0.54289 (0.49197) Loss: 0.54289 (0.49197)
+2025-04-26,19:18:16 | INFO | Train Epoch: 23 [1646592/3047424 (54%)] Data (t): 0.694 Batch (t): 4.930, 1670.78/s, 417.696/s/gpu LR: 0.000379 Logit Scale: 63.704 Imm_image: 30.358 (30.224) Imm_text: 30.358 (30.224) Isd_image: 1.8212 (1.5877) Isd_text: 1.8212 (1.5877) Contrastive_loss: 0.60601 (0.52998) Loss: 0.60601 (0.52998)
+2025-04-26,19:26:29 | INFO | Train Epoch: 23 [2465792/3047424 (81%)] Data (t): 0.691 Batch (t): 4.925, 1669.68/s, 417.419/s/gpu LR: 0.000368 Logit Scale: 63.518 Imm_image: 30.208 (30.220) Imm_text: 30.208 (30.220) Isd_image: 1.9523 (1.6789) Isd_text: 1.9523 (1.6789) Contrastive_loss: 0.65188 (0.56046) Loss: 0.65188 (0.56046)
+2025-04-26,19:32:17 | INFO | Train Epoch: 23 [3047424/3047424 (100%)] Data (t): 0.682 Batch (t): 4.908, 1714.14/s, 428.534/s/gpu LR: 0.000361 Logit Scale: 63.835 Imm_image: 30.962 (30.369) Imm_text: 30.962 (30.369) Isd_image: 1.8756 (1.7182) Isd_text: 1.8756 (1.7182) Contrastive_loss: 0.44037 (0.53644) Loss: 0.44037 (0.53644)
+2025-04-26,19:32:20 | INFO | Start epoch 24
+2025-04-26,19:32:38 | INFO | Train Epoch: 24 [   8192/3047424 (0%)] Data (t): 13.659 Batch (t): 18.352, 446.385/s, 111.596/s/gpu LR: 0.000361 Logit Scale: 63.838 Imm_image: 31.418 (31.418) Imm_text: 31.418 (31.418) Isd_image: 1.8595 (1.8595) Isd_text: 1.8595 (1.8595) Contrastive_loss: 0.33987 (0.33987) Loss: 0.33987 (0.33987)
+2025-04-26,19:40:53 | INFO | Train Epoch: 24 [ 827392/3047424 (27%)] Data (t): 0.704 Batch (t): 4.949, 1659.64/s, 414.911/s/gpu LR: 0.000350 Logit Scale: 65.536 Imm_image: 31.664 (31.541) Imm_text: 31.664 (31.541) Isd_image: 1.1723 (1.5159) Isd_text: 1.1723 (1.5159) Contrastive_loss: 0.43495 (0.38741) Loss: 0.43495 (0.38741)
+2025-04-26,19:49:06 | INFO | Train Epoch: 24 [1646592/3047424 (54%)] Data (t): 0.693 Batch (t): 4.928, 1661.47/s, 415.368/s/gpu LR: 0.000340 Logit Scale: 65.903 Imm_image: 31.519 (31.534) Imm_text: 31.519 (31.534) Isd_image: 1.6017 (1.5445) Isd_text: 1.6017 (1.5445) Contrastive_loss: 0.50552 (0.42678) Loss: 0.50552 (0.42678)
+2025-04-26,19:57:18 | INFO | Train Epoch: 24 [2465792/3047424 (81%)] Data (t): 0.692 Batch (t): 4.925, 1665.09/s, 416.274/s/gpu LR: 0.000330 Logit Scale: 65.893 Imm_image: 31.388 (31.497) Imm_text: 31.388 (31.497) Isd_image: 1.7976 (1.6078) Isd_text: 1.7976 (1.6078) Contrastive_loss: 0.54518 (0.45638) Loss: 0.54518 (0.45638)
+2025-04-26,20:03:07 | INFO | Train Epoch: 24 [3047424/3047424 (100%)] Data (t): 0.681 Batch (t): 4.906, 1716.17/s, 429.041/s/gpu LR: 0.000323 Logit Scale: 66.184 Imm_image: 32.252 (31.648) Imm_text: 32.252 (31.648) Isd_image: 1.6593 (1.6181) Isd_text: 1.6593 (1.6181) Contrastive_loss: 0.34162 (0.43343) Loss: 0.34162 (0.43343)
+2025-04-26,20:03:09 | INFO | Start epoch 25
+2025-04-26,20:03:26 | INFO | Train Epoch: 25 [   8192/3047424 (0%)] Data (t): 12.797 Batch (t): 17.366, 471.717/s, 117.929/s/gpu LR: 0.000323 Logit Scale: 66.188 Imm_image: 32.500 (32.500) Imm_text: 32.500 (32.500) Isd_image: 1.6436 (1.6436) Isd_text: 1.6436 (1.6436) Contrastive_loss: 0.31729 (0.31729) Loss: 0.31729 (0.31729)
+2025-04-26,20:11:41 | INFO | Train Epoch: 25 [ 827392/3047424 (27%)] Data (t): 0.705 Batch (t): 4.948, 1659.90/s, 414.975/s/gpu LR: 0.000312 Logit Scale: 67.872 Imm_image: 32.896 (32.698) Imm_text: 32.896 (32.698) Isd_image: 0.93667 (1.2902) Isd_text: 0.93667 (1.2902) Contrastive_loss: 0.36312 (0.34021) Loss: 0.36312 (0.34021)
+2025-04-26,20:19:54 | INFO | Train Epoch: 25 [1646592/3047424 (54%)] Data (t): 0.694 Batch (t): 4.929, 1663.48/s, 415.870/s/gpu LR: 0.000302 Logit Scale: 68.274 Imm_image: 32.830 (32.742) Imm_text: 32.830 (32.742) Isd_image: 1.4782 (1.3528) Isd_text: 1.4782 (1.3528) Contrastive_loss: 0.41686 (0.36576) Loss: 0.41686 (0.36576)
+2025-04-26,20:28:06 | INFO | Train Epoch: 25 [2465792/3047424 (81%)] Data (t): 0.692 Batch (t): 4.924, 1678.41/s, 419.602/s/gpu LR: 0.000293 Logit Scale: 68.243 Imm_image: 32.738 (32.741) Imm_text: 32.738 (32.741) Isd_image: 1.5896 (1.4120) Isd_text: 1.5896 (1.4120) Contrastive_loss: 0.45615 (0.38836) Loss: 0.45615 (0.38836)
+2025-04-26,20:33:55 | INFO | Train Epoch: 25 [3047424/3047424 (100%)] Data (t): 0.680 Batch (t): 4.907, 1709.89/s, 427.472/s/gpu LR: 0.000286 Logit Scale: 68.519 Imm_image: 33.529 (32.899) Imm_text: 33.529 (32.899) Isd_image: 1.6069 (1.4510) Isd_text: 1.6069 (1.4510) Contrastive_loss: 0.29083 (0.36885) Loss: 0.29083 (0.36885)
+2025-04-26,20:33:57 | INFO | Start epoch 26
+2025-04-26,20:34:15 | INFO | Train Epoch: 26 [   8192/3047424 (0%)] Data (t): 13.348 Batch (t): 17.913, 457.330/s, 114.332/s/gpu LR: 0.000286 Logit Scale: 68.526 Imm_image: 33.922 (33.922) Imm_text: 33.922 (33.922) Isd_image: 1.5964 (1.5964) Isd_text: 1.5964 (1.5964) Contrastive_loss: 0.24927 (0.24927) Loss: 0.24927 (0.24927)
+2025-04-26,20:42:30 | INFO | Train Epoch: 26 [ 827392/3047424 (27%)] Data (t): 0.703 Batch (t): 4.946, 1665.36/s, 416.341/s/gpu LR: 0.000276 Logit Scale: 70.137 Imm_image: 34.240 (34.081) Imm_text: 34.240 (34.081) Isd_image: 1.2326 (1.4145) Isd_text: 1.2326 (1.4145) Contrastive_loss: 0.29903 (0.27415) Loss: 0.29903 (0.27415)
+2025-04-26,20:50:43 | INFO | Train Epoch: 26 [1646592/3047424 (54%)] Data (t): 0.696 Batch (t): 4.929, 1662.18/s, 415.546/s/gpu LR: 0.000266 Logit Scale: 70.651 Imm_image: 34.228 (34.130) Imm_text: 34.228 (34.130) Isd_image: 1.3241 (1.3844) Isd_text: 1.3241 (1.3844) Contrastive_loss: 0.35615 (0.30148) Loss: 0.35615 (0.30148)
+2025-04-26,20:58:54 | INFO | Train Epoch: 26 [2465792/3047424 (81%)] Data (t): 0.689 Batch (t): 4.918, 1669.62/s, 417.405/s/gpu LR: 0.000257 Logit Scale: 70.734 Imm_image: 34.196 (34.147) Imm_text: 34.196 (34.147) Isd_image: 1.6361 (1.4473) Isd_text: 1.6361 (1.4473) Contrastive_loss: 0.35866 (0.31578) Loss: 0.35866 (0.31578)
+2025-04-26,21:04:43 | INFO | Train Epoch: 26 [3047424/3047424 (100%)] Data (t): 0.678 Batch (t): 4.904, 1719.75/s, 429.938/s/gpu LR: 0.000250 Logit Scale: 71.084 Imm_image: 35.105 (34.338) Imm_text: 35.105 (34.338) Isd_image: 1.6278 (1.4834) Isd_text: 1.6278 (1.4834) Contrastive_loss: 0.24082 (0.30079) Loss: 0.24082 (0.30079)
+2025-04-26,21:04:45 | INFO | Start epoch 27
+2025-04-26,21:05:03 | INFO | Train Epoch: 27 [   8192/3047424 (0%)] Data (t): 13.551 Batch (t): 18.137, 451.685/s, 112.921/s/gpu LR: 0.000250 Logit Scale: 71.087 Imm_image: 35.471 (35.471) Imm_text: 35.471 (35.471) Isd_image: 1.6427 (1.6427) Isd_text: 1.6427 (1.6427) Contrastive_loss: 0.19770 (0.19770) Loss: 0.19770 (0.19770)
+2025-04-26,21:13:18 | INFO | Train Epoch: 27 [ 827392/3047424 (27%)] Data (t): 0.705 Batch (t): 4.951, 1656.32/s, 414.080/s/gpu LR: 0.000241 Logit Scale: 72.521 Imm_image: 35.551 (35.511) Imm_text: 35.551 (35.511) Isd_image: 1.1859 (1.4143) Isd_text: 1.1859 (1.4143) Contrastive_loss: 0.26316 (0.23043) Loss: 0.26316 (0.23043)
+2025-04-26,21:21:31 | INFO | Train Epoch: 27 [1646592/3047424 (54%)] Data (t): 0.696 Batch (t): 4.931, 1660.20/s, 415.051/s/gpu LR: 0.000231 Logit Scale: 73.096 Imm_image: 35.445 (35.489) Imm_text: 35.445 (35.489) Isd_image: 1.2915 (1.3734) Isd_text: 1.2915 (1.3734) Contrastive_loss: 0.29878 (0.25321) Loss: 0.29878 (0.25321)
+2025-04-26,21:29:43 | INFO | Train Epoch: 27 [2465792/3047424 (81%)] Data (t): 0.692 Batch (t): 4.923, 1670.54/s, 417.635/s/gpu LR: 0.000222 Logit Scale: 73.344 Imm_image: 35.727 (35.548) Imm_text: 35.727 (35.548) Isd_image: 1.5990 (1.4298) Isd_text: 1.5990 (1.4298) Contrastive_loss: 0.29307 (0.26318) Loss: 0.29307 (0.26318)
+2025-04-26,21:35:32 | INFO | Train Epoch: 27 [3047424/3047424 (100%)] Data (t): 0.678 Batch (t): 4.903, 1710.91/s, 427.728/s/gpu LR: 0.000216 Logit Scale: 73.633 Imm_image: 36.496 (35.738) Imm_text: 36.496 (35.738) Isd_image: 1.5123 (1.4463) Isd_text: 1.5123 (1.4463) Contrastive_loss: 0.20083 (0.25071) Loss: 0.20083 (0.25071)
+2025-04-26,21:35:34 | INFO | Start epoch 28
+2025-04-26,21:35:52 | INFO | Train Epoch: 28 [   8192/3047424 (0%)] Data (t): 13.274 Batch (t): 17.846, 459.042/s, 114.761/s/gpu LR: 0.000216 Logit Scale: 73.638 Imm_image: 36.861 (36.861) Imm_text: 36.861 (36.861) Isd_image: 1.4300 (1.4300) Isd_text: 1.4300 (1.4300) Contrastive_loss: 0.16774 (0.16774) Loss: 0.16774 (0.16774)
+2025-04-26,21:44:08 | INFO | Train Epoch: 28 [ 827392/3047424 (27%)] Data (t): 0.715 Batch (t): 4.963, 1634.32/s, 408.581/s/gpu LR: 0.000207 Logit Scale: 75.002 Imm_image: 37.064 (36.962) Imm_text: 37.064 (36.962) Isd_image: 1.2946 (1.3623) Isd_text: 1.2946 (1.3623) Contrastive_loss: 0.20792 (0.18783) Loss: 0.20792 (0.18783)
+2025-04-26,21:52:25 | INFO | Train Epoch: 28 [1646592/3047424 (54%)] Data (t): 0.722 Batch (t): 4.974, 1637.07/s, 409.267/s/gpu LR: 0.000198 Logit Scale: 75.635 Imm_image: 36.999 (36.975) Imm_text: 36.999 (36.975) Isd_image: 1.3398 (1.3548) Isd_text: 1.3398 (1.3548) Contrastive_loss: 0.24134 (0.20567) Loss: 0.24134 (0.20567)
+2025-04-26,22:00:38 | INFO | Train Epoch: 28 [2465792/3047424 (81%)] Data (t): 0.689 Batch (t): 4.921, 1643.45/s, 410.863/s/gpu LR: 0.000190 Logit Scale: 76.018 Imm_image: 37.238 (37.040) Imm_text: 37.238 (37.040) Isd_image: 1.5507 (1.4038) Isd_text: 1.5507 (1.4038) Contrastive_loss: 0.23845 (0.21386) Loss: 0.23845 (0.21386)
+2025-04-26,22:06:24 | INFO | Train Epoch: 28 [3047424/3047424 (100%)] Data (t): 0.660 Batch (t): 4.882, 1719.76/s, 429.940/s/gpu LR: 0.000184 Logit Scale: 76.405 Imm_image: 38.002 (37.233) Imm_text: 38.002 (37.233) Isd_image: 1.5861 (1.4402) Isd_text: 1.5861 (1.4402) Contrastive_loss: 0.18176 (0.20744) Loss: 0.18176 (0.20744)
+2025-04-26,22:06:27 | INFO | Start epoch 29
+2025-04-26,22:06:45 | INFO | Train Epoch: 29 [   8192/3047424 (0%)] Data (t): 13.265 Batch (t): 17.903, 457.584/s, 114.396/s/gpu LR: 0.000184 Logit Scale: 76.410 Imm_image: 38.216 (38.216) Imm_text: 38.216 (38.216) Isd_image: 1.4878 (1.4878) Isd_text: 1.4878 (1.4878) Contrastive_loss: 0.14770 (0.14770) Loss: 0.14770 (0.14770)
+2025-04-26,22:15:00 | INFO | Train Epoch: 29 [ 827392/3047424 (27%)] Data (t): 0.703 Batch (t): 4.948, 1666.91/s, 416.727/s/gpu LR: 0.000175 Logit Scale: 77.564 Imm_image: 38.473 (38.344) Imm_text: 38.473 (38.344) Isd_image: 1.1820 (1.3349) Isd_text: 1.1820 (1.3349) Contrastive_loss: 0.16813 (0.15791) Loss: 0.16813 (0.15791)
+2025-04-26,22:23:13 | INFO | Train Epoch: 29 [1646592/3047424 (54%)] Data (t): 0.697 Batch (t): 4.930, 1663.12/s, 415.779/s/gpu LR: 0.000167 Logit Scale: 78.107 Imm_image: 38.727 (38.472) Imm_text: 38.727 (38.472) Isd_image: 1.4501 (1.3733) Isd_text: 1.4501 (1.3733) Contrastive_loss: 0.19228 (0.16937) Loss: 0.19228 (0.16937)
+2025-04-26,22:31:25 | INFO | Train Epoch: 29 [2465792/3047424 (81%)] Data (t): 0.694 Batch (t): 4.926, 1674.88/s, 418.719/s/gpu LR: 0.000159 Logit Scale: 78.509 Imm_image: 38.776 (38.548) Imm_text: 38.776 (38.548) Isd_image: 1.5894 (1.4273) Isd_text: 1.5894 (1.4273) Contrastive_loss: 0.18797 (0.17402) Loss: 0.18797 (0.17402)
+2025-04-26,22:37:14 | INFO | Train Epoch: 29 [3047424/3047424 (100%)] Data (t): 0.681 Batch (t): 4.908, 1713.61/s, 428.404/s/gpu LR: 0.000154 Logit Scale: 78.843 Imm_image: 39.552 (38.749) Imm_text: 39.552 (38.749) Isd_image: 1.6192 (1.4657) Isd_text: 1.6192 (1.4657) Contrastive_loss: 0.14366 (0.16795) Loss: 0.14366 (0.16795)
+2025-04-26,22:37:16 | INFO | Start epoch 30
+2025-04-26,22:37:34 | INFO | Train Epoch: 30 [   8192/3047424 (0%)] Data (t): 13.652 Batch (t): 18.134, 451.749/s, 112.937/s/gpu LR: 0.000154 Logit Scale: 78.848 Imm_image: 39.768 (39.768) Imm_text: 39.768 (39.768) Isd_image: 1.6412 (1.6412) Isd_text: 1.6412 (1.6412) Contrastive_loss: 0.13102 (0.13102) Loss: 0.13102 (0.13102)
+2025-04-26,22:45:49 | INFO | Train Epoch: 30 [ 827392/3047424 (27%)] Data (t): 0.709 Batch (t): 4.951, 1660.38/s, 415.096/s/gpu LR: 0.000146 Logit Scale: 79.899 Imm_image: 40.025 (39.896) Imm_text: 40.025 (39.896) Isd_image: 1.4185 (1.5298) Isd_text: 1.4185 (1.5298) Contrastive_loss: 0.14362 (0.13732) Loss: 0.14362 (0.13732)
+2025-04-26,22:54:03 | INFO | Train Epoch: 30 [1646592/3047424 (54%)] Data (t): 0.699 Batch (t): 4.936, 1663.43/s, 415.858/s/gpu LR: 0.000138 Logit Scale: 80.456 Imm_image: 40.066 (39.953) Imm_text: 40.066 (39.953) Isd_image: 1.5443 (1.5347) Isd_text: 1.5443 (1.5347) Contrastive_loss: 0.15916 (0.14460) Loss: 0.15916 (0.14460)
+2025-04-26,23:02:13 | INFO | Train Epoch: 30 [2465792/3047424 (81%)] Data (t): 0.679 Batch (t): 4.906, 1674.99/s, 418.748/s/gpu LR: 0.000131 Logit Scale: 80.908 Imm_image: 40.296 (40.039) Imm_text: 40.296 (40.039) Isd_image: 1.5791 (1.5458) Isd_text: 1.5791 (1.5458) Contrastive_loss: 0.15136 (0.14629) Loss: 0.15136 (0.14629)
+2025-04-26,23:08:02 | INFO | Train Epoch: 30 [3047424/3047424 (100%)] Data (t): 0.679 Batch (t): 4.903, 1714.81/s, 428.702/s/gpu LR: 0.000126 Logit Scale: 81.328 Imm_image: 41.039 (40.239) Imm_text: 41.039 (40.239) Isd_image: 1.5430 (1.5452) Isd_text: 1.5430 (1.5452) Contrastive_loss: 0.11773 (0.14058) Loss: 0.11773 (0.14058)
+2025-04-26,23:08:04 | INFO | Start epoch 31
+2025-04-26,23:08:22 | INFO | Train Epoch: 31 [   8192/3047424 (0%)] Data (t): 13.440 Batch (t): 18.115, 452.218/s, 113.054/s/gpu LR: 0.000126 Logit Scale: 81.333 Imm_image: 41.076 (41.076) Imm_text: 41.076 (41.076) Isd_image: 1.3890 (1.3890) Isd_text: 1.3890 (1.3890) Contrastive_loss: 0.10597 (0.10597) Loss: 0.10597 (0.10597)
+2025-04-26,23:16:38 | INFO | Train Epoch: 31 [ 827392/3047424 (27%)] Data (t): 0.712 Batch (t): 4.954, 1671.30/s, 417.826/s/gpu LR: 0.000119 Logit Scale: 82.111 Imm_image: 41.306 (41.191) Imm_text: 41.306 (41.191) Isd_image: 1.3598 (1.3744) Isd_text: 1.3598 (1.3744) Contrastive_loss: 0.11906 (0.11251) Loss: 0.11906 (0.11251)
+2025-04-26,23:24:50 | INFO | Train Epoch: 31 [1646592/3047424 (54%)] Data (t): 0.695 Batch (t): 4.928, 1656.04/s, 414.010/s/gpu LR: 0.000112 Logit Scale: 82.642 Imm_image: 41.451 (41.278) Imm_text: 41.451 (41.278) Isd_image: 1.3642 (1.3710) Isd_text: 1.3642 (1.3710) Contrastive_loss: 0.12228 (0.11577) Loss: 0.12228 (0.11577)
+2025-04-26,23:33:03 | INFO | Train Epoch: 31 [2465792/3047424 (81%)] Data (t): 0.692 Batch (t): 4.923, 1676.94/s, 419.234/s/gpu LR: 0.000105 Logit Scale: 83.172 Imm_image: 41.838 (41.418) Imm_text: 41.838 (41.418) Isd_image: 1.6475 (1.4401) Isd_text: 1.6475 (1.4401) Contrastive_loss: 0.12298 (0.11757) Loss: 0.12298 (0.11757)
+2025-04-26,23:38:51 | INFO | Train Epoch: 31 [3047424/3047424 (100%)] Data (t): 0.681 Batch (t): 4.904, 1716.89/s, 429.223/s/gpu LR: 0.000100 Logit Scale: 83.547 Imm_image: 42.496 (41.633) Imm_text: 42.496 (41.633) Isd_image: 1.5630 (1.4647) Isd_text: 1.5630 (1.4647) Contrastive_loss: 0.092837 (0.11263) Loss: 0.092837 (0.11263)
+2025-04-26,23:38:53 | INFO | Start epoch 32
+2025-04-26,23:39:12 | INFO | Train Epoch: 32 [   8192/3047424 (0%)] Data (t): 13.582 Batch (t): 18.352, 446.370/s, 111.593/s/gpu LR: 0.000100 Logit Scale: 83.553 Imm_image: 42.627 (42.627) Imm_text: 42.627 (42.627) Isd_image: 1.4939 (1.4939) Isd_text: 1.4939 (1.4939) Contrastive_loss: 0.091551 (0.091551) Loss: 0.091551 (0.091551)
+2025-04-26,23:47:27 | INFO | Train Epoch: 32 [ 827392/3047424 (27%)] Data (t): 0.706 Batch (t): 4.950, 1661.67/s, 415.419/s/gpu LR: 0.000094 Logit Scale: 84.160 Imm_image: 42.556 (42.592) Imm_text: 42.556 (42.592) Isd_image: 1.3681 (1.4310) Isd_text: 1.3681 (1.4310) Contrastive_loss: 0.091477 (0.091514) Loss: 0.091477 (0.091514)
+2025-04-26,23:55:40 | INFO | Train Epoch: 32 [1646592/3047424 (54%)] Data (t): 0.703 Batch (t): 4.934, 1653.70/s, 413.426/s/gpu LR: 0.000088 Logit Scale: 84.594 Imm_image: 42.840 (42.674) Imm_text: 42.840 (42.674) Isd_image: 1.5984 (1.4868) Isd_text: 1.5984 (1.4868) Contrastive_loss: 0.10604 (0.096357) Loss: 0.10604 (0.096357)
+2025-04-27,00:03:53 | INFO | Train Epoch: 32 [2465792/3047424 (81%)] Data (t): 0.693 Batch (t): 4.926, 1683.89/s, 420.973/s/gpu LR: 0.000082 Logit Scale: 85.030 Imm_image: 42.950 (42.743) Imm_text: 42.950 (42.743) Isd_image: 1.5832 (1.5109) Isd_text: 1.5832 (1.5109) Contrastive_loss: 0.099170 (0.097060) Loss: 0.099170 (0.097060)
+2025-04-27,00:09:41 | INFO | Train Epoch: 32 [3047424/3047424 (100%)] Data (t): 0.676 Batch (t): 4.902, 1723.93/s, 430.983/s/gpu LR: 0.000077 Logit Scale: 85.353 Imm_image: 43.558 (42.906) Imm_text: 43.558 (42.906) Isd_image: 1.4102 (1.4908) Isd_text: 1.4102 (1.4908) Contrastive_loss: 0.076849 (0.093018) Loss: 0.076849 (0.093018)
+2025-04-27,00:09:43 | INFO | Start epoch 33
+2025-04-27,00:10:02 | INFO | Train Epoch: 33 [   8192/3047424 (0%)] Data (t): 13.444 Batch (t): 18.107, 452.433/s, 113.108/s/gpu LR: 0.000077 Logit Scale: 85.358 Imm_image: 43.632 (43.632) Imm_text: 43.632 (43.632) Isd_image: 1.4252 (1.4252) Isd_text: 1.4252 (1.4252) Contrastive_loss: 0.076395 (0.076395) Loss: 0.076395 (0.076395)
+2025-04-27,00:18:17 | INFO | Train Epoch: 33 [ 827392/3047424 (27%)] Data (t): 0.711 Batch (t): 4.951, 1661.02/s, 415.255/s/gpu LR: 0.000072 Logit Scale: 85.860 Imm_image: 43.948 (43.790) Imm_text: 43.948 (43.790) Isd_image: 1.5108 (1.4680) Isd_text: 1.5108 (1.4680) Contrastive_loss: 0.078253 (0.077324) Loss: 0.078253 (0.077324)
+2025-04-27,00:26:30 | INFO | Train Epoch: 33 [1646592/3047424 (54%)] Data (t): 0.696 Batch (t): 4.930, 1669.57/s, 417.393/s/gpu LR: 0.000066 Logit Scale: 86.205 Imm_image: 44.008 (43.863) Imm_text: 44.008 (43.863) Isd_image: 1.6208 (1.5189) Isd_text: 1.6208 (1.5189) Contrastive_loss: 0.082241 (0.078963) Loss: 0.082241 (0.078963)
+2025-04-27,00:34:42 | INFO | Train Epoch: 33 [2465792/3047424 (81%)] Data (t): 0.695 Batch (t): 4.926, 1679.34/s, 419.834/s/gpu LR: 0.000061 Logit Scale: 86.583 Imm_image: 44.252 (43.960) Imm_text: 44.252 (43.960) Isd_image: 1.5330 (1.5225) Isd_text: 1.5330 (1.5225) Contrastive_loss: 0.075883 (0.078193) Loss: 0.075883 (0.078193)
+2025-04-27,00:40:31 | INFO | Train Epoch: 33 [3047424/3047424 (100%)] Data (t): 0.683 Batch (t): 4.908, 1709.82/s, 427.454/s/gpu LR: 0.000057 Logit Scale: 86.824 Imm_image: 44.804 (44.129) Imm_text: 44.804 (44.129) Isd_image: 1.6201 (1.5420) Isd_text: 1.6201 (1.5420) Contrastive_loss: 0.068852 (0.076325) Loss: 0.068852 (0.076325)
+2025-04-27,00:40:33 | INFO | Start epoch 34
+2025-04-27,00:40:51 | INFO | Train Epoch: 34 [   8192/3047424 (0%)] Data (t): 13.229 Batch (t): 17.831, 459.419/s, 114.855/s/gpu LR: 0.000057 Logit Scale: 86.825 Imm_image: 44.832 (44.832) Imm_text: 44.832 (44.832) Isd_image: 1.6502 (1.6502) Isd_text: 1.6502 (1.6502) Contrastive_loss: 0.068599 (0.068599) Loss: 0.068599 (0.068599)
+2025-04-27,00:49:06 | INFO | Train Epoch: 34 [ 827392/3047424 (27%)] Data (t): 0.710 Batch (t): 4.949, 1652.81/s, 413.202/s/gpu LR: 0.000052 Logit Scale: 87.176 Imm_image: 44.892 (44.862) Imm_text: 44.892 (44.862) Isd_image: 1.4637 (1.5569) Isd_text: 1.4637 (1.5569) Contrastive_loss: 0.072305 (0.070452) Loss: 0.072305 (0.070452)
+2025-04-27,00:57:19 | INFO | Train Epoch: 34 [1646592/3047424 (54%)] Data (t): 0.695 Batch (t): 4.925, 1663.77/s, 415.941/s/gpu LR: 0.000048 Logit Scale: 87.487 Imm_image: 45.108 (44.944) Imm_text: 45.108 (44.944) Isd_image: 1.4878 (1.5339) Isd_text: 1.4878 (1.5339) Contrastive_loss: 0.064124 (0.068342) Loss: 0.064124 (0.068342)
+2025-04-27,01:05:30 | INFO | Train Epoch: 34 [2465792/3047424 (81%)] Data (t): 0.684 Batch (t): 4.912, 1665.24/s, 416.311/s/gpu LR: 0.000043 Logit Scale: 87.770 Imm_image: 45.268 (45.025) Imm_text: 45.268 (45.025) Isd_image: 1.5024 (1.5260) Isd_text: 1.5024 (1.5260) Contrastive_loss: 0.065741 (0.067692) Loss: 0.065741 (0.067692)
+2025-04-27,01:11:18 | INFO | Train Epoch: 34 [3047424/3047424 (100%)] Data (t): 0.682 Batch (t): 4.901, 1713.24/s, 428.309/s/gpu LR: 0.000040 Logit Scale: 87.966 Imm_image: 45.567 (45.133) Imm_text: 45.567 (45.133) Isd_image: 1.4492 (1.5107) Isd_text: 1.4492 (1.5107) Contrastive_loss: 0.057280 (0.065610) Loss: 0.057280 (0.065610)
+2025-04-27,01:11:20 | INFO | Start epoch 35
+2025-04-27,01:11:38 | INFO | Train Epoch: 35 [   8192/3047424 (0%)] Data (t): 13.654 Batch (t): 18.234, 449.266/s, 112.317/s/gpu LR: 0.000040 Logit Scale: 87.968 Imm_image: 45.643 (45.643) Imm_text: 45.643 (45.643) Isd_image: 1.4399 (1.4399) Isd_text: 1.4399 (1.4399) Contrastive_loss: 0.057030 (0.057030) Loss: 0.057030 (0.057030)
+2025-04-27,01:19:53 | INFO | Train Epoch: 35 [ 827392/3047424 (27%)] Data (t): 0.709 Batch (t): 4.947, 1665.81/s, 416.453/s/gpu LR: 0.000036 Logit Scale: 88.263 Imm_image: 45.722 (45.682) Imm_text: 45.722 (45.682) Isd_image: 1.3973 (1.4186) Isd_text: 1.3973 (1.4186) Contrastive_loss: 0.064645 (0.060838) Loss: 0.064645 (0.060838)
+2025-04-27,01:28:06 | INFO | Train Epoch: 35 [1646592/3047424 (54%)] Data (t): 0.697 Batch (t): 4.929, 1665.17/s, 416.292/s/gpu LR: 0.000032 Logit Scale: 88.481 Imm_image: 45.843 (45.736) Imm_text: 45.843 (45.736) Isd_image: 1.4011 (1.4127) Isd_text: 1.4011 (1.4127) Contrastive_loss: 0.060792 (0.060822) Loss: 0.060792 (0.060822)
+2025-04-27,01:36:18 | INFO | Train Epoch: 35 [2465792/3047424 (81%)] Data (t): 0.695 Batch (t): 4.922, 1672.71/s, 418.179/s/gpu LR: 0.000028 Logit Scale: 88.686 Imm_image: 46.074 (45.820) Imm_text: 46.074 (45.820) Isd_image: 1.4744 (1.4281) Isd_text: 1.4744 (1.4281) Contrastive_loss: 0.069188 (0.062914) Loss: 0.069188 (0.062914)
+2025-04-27,01:42:06 | INFO | Train Epoch: 35 [3047424/3047424 (100%)] Data (t): 0.683 Batch (t): 4.904, 1713.75/s, 428.437/s/gpu LR: 0.000026 Logit Scale: 88.819 Imm_image: 46.400 (45.936) Imm_text: 46.400 (45.936) Isd_image: 1.4201 (1.4265) Isd_text: 1.4201 (1.4265) Contrastive_loss: 0.049764 (0.060284) Loss: 0.049764 (0.060284)
+2025-04-27,01:42:09 | INFO | Start epoch 36
+2025-04-27,01:42:27 | INFO | Train Epoch: 36 [   8192/3047424 (0%)] Data (t): 13.523 Batch (t): 18.137, 451.670/s, 112.918/s/gpu LR: 0.000026 Logit Scale: 88.820 Imm_image: 46.339 (46.339) Imm_text: 46.339 (46.339) Isd_image: 1.3509 (1.3509) Isd_text: 1.3509 (1.3509) Contrastive_loss: 0.052865 (0.052865) Loss: 0.052865 (0.052865)
+2025-04-27,01:50:42 | INFO | Train Epoch: 36 [ 827392/3047424 (27%)] Data (t): 0.707 Batch (t): 4.947, 1657.25/s, 414.312/s/gpu LR: 0.000022 Logit Scale: 89.006 Imm_image: 46.444 (46.391) Imm_text: 46.444 (46.391) Isd_image: 1.3992 (1.3751) Isd_text: 1.3992 (1.3751) Contrastive_loss: 0.049712 (0.051289) Loss: 0.049712 (0.051289)
+2025-04-27,01:58:55 | INFO | Train Epoch: 36 [1646592/3047424 (54%)] Data (t): 0.698 Batch (t): 4.927, 1659.79/s, 414.947/s/gpu LR: 0.000019 Logit Scale: 89.157 Imm_image: 46.548 (46.444) Imm_text: 46.548 (46.444) Isd_image: 1.4465 (1.3989) Isd_text: 1.4465 (1.3989) Contrastive_loss: 0.055454 (0.052677) Loss: 0.055454 (0.052677)
+2025-04-27,02:07:07 | INFO | Train Epoch: 36 [2465792/3047424 (81%)] Data (t): 0.694 Batch (t): 4.924, 1672.65/s, 418.162/s/gpu LR: 0.000016 Logit Scale: 89.271 Imm_image: 46.728 (46.515) Imm_text: 46.728 (46.515) Isd_image: 1.4507 (1.4118) Isd_text: 1.4507 (1.4118) Contrastive_loss: 0.058500 (0.054133) Loss: 0.058500 (0.054133)
+2025-04-27,02:12:55 | INFO | Train Epoch: 36 [3047424/3047424 (100%)] Data (t): 0.682 Batch (t): 4.902, 1721.97/s, 430.492/s/gpu LR: 0.000015 Logit Scale: 89.354 Imm_image: 46.920 (46.596) Imm_text: 46.920 (46.596) Isd_image: 1.4259 (1.4147) Isd_text: 1.4259 (1.4147) Contrastive_loss: 0.048069 (0.052920) Loss: 0.048069 (0.052920)
+2025-04-27,02:12:57 | INFO | Start epoch 37
+2025-04-27,02:13:16 | INFO | Train Epoch: 37 [   8192/3047424 (0%)] Data (t): 13.675 Batch (t): 18.328, 446.971/s, 111.743/s/gpu LR: 0.000015 Logit Scale: 89.355 Imm_image: 46.877 (46.877) Imm_text: 46.877 (46.877) Isd_image: 1.4588 (1.4588) Isd_text: 1.4588 (1.4588) Contrastive_loss: 0.049559 (0.049559) Loss: 0.049559 (0.049559)
+2025-04-27,02:21:30 | INFO | Train Epoch: 37 [ 827392/3047424 (27%)] Data (t): 0.706 Batch (t): 4.946, 1661.93/s, 415.482/s/gpu LR: 0.000012 Logit Scale: 89.458 Imm_image: 46.921 (46.899) Imm_text: 46.921 (46.899) Isd_image: 1.3657 (1.4123) Isd_text: 1.3657 (1.4123) Contrastive_loss: 0.047993 (0.048776) Loss: 0.047993 (0.048776)
+2025-04-27,02:29:43 | INFO | Train Epoch: 37 [1646592/3047424 (54%)] Data (t): 0.697 Batch (t): 4.926, 1661.51/s, 415.378/s/gpu LR: 0.000010 Logit Scale: 89.536 Imm_image: 47.068 (46.955) Imm_text: 47.068 (46.955) Isd_image: 1.3471 (1.3905) Isd_text: 1.3471 (1.3905) Contrastive_loss: 0.044614 (0.047389) Loss: 0.044614 (0.047389)
+2025-04-27,02:37:55 | INFO | Train Epoch: 37 [2465792/3047424 (81%)] Data (t): 0.694 Batch (t): 4.922, 1673.40/s, 418.350/s/gpu LR: 0.000008 Logit Scale: 89.600 Imm_image: 47.163 (47.007) Imm_text: 47.163 (47.007) Isd_image: 1.4368 (1.4021) Isd_text: 1.4368 (1.4021) Contrastive_loss: 0.041111 (0.045819) Loss: 0.041111 (0.045819)
+2025-04-27,02:43:43 | INFO | Train Epoch: 37 [3047424/3047424 (100%)] Data (t): 0.685 Batch (t): 4.905, 1719.68/s, 429.920/s/gpu LR: 0.000006 Logit Scale: 89.637 Imm_image: 47.201 (47.046) Imm_text: 47.201 (47.046) Isd_image: 1.3355 (1.3888) Isd_text: 1.3355 (1.3888) Contrastive_loss: 0.048584 (0.046372) Loss: 0.048584 (0.046372)
+2025-04-27,02:43:46 | INFO | Start epoch 38
+2025-04-27,02:44:04 | INFO | Train Epoch: 38 [   8192/3047424 (0%)] Data (t): 13.772 Batch (t): 18.423, 444.651/s, 111.163/s/gpu LR: 0.000006 Logit Scale: 89.638 Imm_image: 47.301 (47.301) Imm_text: 47.301 (47.301) Isd_image: 1.3752 (1.3752) Isd_text: 1.3752 (1.3752) Contrastive_loss: 0.041167 (0.041167) Loss: 0.041167 (0.041167)
+2025-04-27,02:52:19 | INFO | Train Epoch: 38 [ 827392/3047424 (27%)] Data (t): 0.708 Batch (t): 4.947, 1654.58/s, 413.645/s/gpu LR: 0.000005 Logit Scale: 89.679 Imm_image: 47.335 (47.318) Imm_text: 47.335 (47.318) Isd_image: 1.4212 (1.3982) Isd_text: 1.4212 (1.3982) Contrastive_loss: 0.045266 (0.043217) Loss: 0.045266 (0.043217)
+2025-04-27,03:00:32 | INFO | Train Epoch: 38 [1646592/3047424 (54%)] Data (t): 0.699 Batch (t): 4.929, 1667.75/s, 416.937/s/gpu LR: 0.000003 Logit Scale: 89.713 Imm_image: 47.334 (47.323) Imm_text: 47.334 (47.323) Isd_image: 1.2828 (1.3597) Isd_text: 1.2828 (1.3597) Contrastive_loss: 0.048032 (0.044822) Loss: 0.048032 (0.044822)
+2025-04-27,03:08:50 | INFO | Train Epoch: 38 [2465792/3047424 (81%)] Data (t): 0.733 Batch (t): 4.982, 1638.79/s, 409.696/s/gpu LR: 0.000002 Logit Scale: 89.736 Imm_image: 47.354 (47.331) Imm_text: 47.354 (47.331) Isd_image: 1.2347 (1.3285) Isd_text: 1.2347 (1.3285) Contrastive_loss: 0.047454 (0.045480) Loss: 0.047454 (0.045480)
+2025-04-27,03:14:48 | INFO | Train Epoch: 38 [3047424/3047424 (100%)] Data (t): 0.769 Batch (t): 5.038, 1673.33/s, 418.332/s/gpu LR: 0.000002 Logit Scale: 89.748 Imm_image: 47.391 (47.343) Imm_text: 47.391 (47.343) Isd_image: 1.3523 (1.3332) Isd_text: 1.3523 (1.3332) Contrastive_loss: 0.048341 (0.046052) Loss: 0.048341 (0.046052)
+2025-04-27,03:14:50 | INFO | Start epoch 39
+2025-04-27,03:15:09 | INFO | Train Epoch: 39 [   8192/3047424 (0%)] Data (t): 14.102 Batch (t): 18.899, 433.471/s, 108.368/s/gpu LR: 0.000002 Logit Scale: 89.748 Imm_image: 47.431 (47.431) Imm_text: 47.431 (47.431) Isd_image: 1.3615 (1.3615) Isd_text: 1.3615 (1.3615) Contrastive_loss: 0.044348 (0.044348) Loss: 0.044348 (0.044348)
+2025-04-27,03:23:36 | INFO | Train Epoch: 39 [ 827392/3047424 (27%)] Data (t): 0.793 Batch (t): 5.070, 1634.34/s, 408.585/s/gpu LR: 0.000001 Logit Scale: 89.758 Imm_image: 47.353 (47.392) Imm_text: 47.353 (47.392) Isd_image: 1.3247 (1.3431) Isd_text: 1.3247 (1.3431) Contrastive_loss: 0.046746 (0.045547) Loss: 0.046746 (0.045547)
+2025-04-27,03:31:57 | INFO | Train Epoch: 39 [1646592/3047424 (54%)] Data (t): 0.752 Batch (t): 5.006, 1607.04/s, 401.759/s/gpu LR: 0.000000 Logit Scale: 89.762 Imm_image: 47.399 (47.394) Imm_text: 47.399 (47.394) Isd_image: 1.3190 (1.3351) Isd_text: 1.3190 (1.3351) Contrastive_loss: 0.039496 (0.043530) Loss: 0.039496 (0.043530)
+2025-04-27,03:40:18 | INFO | Train Epoch: 39 [2465792/3047424 (81%)] Data (t): 0.760 Batch (t): 5.018, 1663.54/s, 415.885/s/gpu LR: 0.000000 Logit Scale: 89.763 Imm_image: 47.416 (47.400) Imm_text: 47.416 (47.400) Isd_image: 1.3053 (1.3276) Isd_text: 1.3053 (1.3276) Contrastive_loss: 0.045538 (0.044032) Loss: 0.045538 (0.044032)
+2025-04-27,03:46:12 | INFO | Train Epoch: 39 [3047424/3047424 (100%)] Data (t): 0.727 Batch (t): 4.973, 1695.31/s, 423.828/s/gpu LR: 0.000000 Logit Scale: 89.763 Imm_image: 47.447 (47.409) Imm_text: 47.447 (47.409) Isd_image: 1.3615 (1.3344) Isd_text: 1.3615 (1.3344) Contrastive_loss: 0.047040 (0.044633) Loss: 0.047040 (0.044633)

SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-010-filled-map_text_farest/benchmark_caltech101_epoch_40.pt_ViT-B-16_en_zeroshot_classification.json

@@ -0,0 +1 @@
+{"dataset": "caltech101", "model": "ViT-B-16", "pretrained": "/mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-010-filled-map_text_farest/checkpoints/epoch_40.pt", "task": "zeroshot_classification", "metrics": {"acc1": 0.502405949256343, "acc5": 0.7365485564304461, "mean_per_class_recall": 0.43103608193861165}, "language": "en"}

SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-010-filled-map_text_farest/benchmark_cars_epoch_40.pt_ViT-B-16_en_zeroshot_classification.json

@@ -0,0 +1 @@
+{"dataset": "cars", "model": "ViT-B-16", "pretrained": "/mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-010-filled-map_text_farest/checkpoints/epoch_40.pt", "task": "zeroshot_classification", "metrics": {"acc1": 0.012063176221862952, "acc5": 0.052481034697176965, "mean_per_class_recall": 0.012812957552955336}, "language": "en"}

SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-010-filled-map_text_farest/benchmark_cifar100_epoch_40.pt_ViT-B-16_en_zeroshot_classification.json

@@ -0,0 +1 @@
+{"dataset": "cifar100", "model": "ViT-B-16", "pretrained": "/mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-010-filled-map_text_farest/checkpoints/epoch_40.pt", "task": "zeroshot_classification", "metrics": {"acc1": 0.1979, "acc5": 0.4318, "mean_per_class_recall": 0.1975}, "language": "en"}

SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-010-filled-map_text_farest/benchmark_cifar10_epoch_40.pt_ViT-B-16_en_zeroshot_classification.json

@@ -0,0 +1 @@
+{"dataset": "cifar10", "model": "ViT-B-16", "pretrained": "/mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-010-filled-map_text_farest/checkpoints/epoch_40.pt", "task": "zeroshot_classification", "metrics": {"acc1": 0.4545, "acc5": 0.8871, "mean_per_class_recall": 0.4553}, "language": "en"}

SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-010-filled-map_text_farest/benchmark_country211_epoch_40.pt_ViT-B-16_en_zeroshot_classification.json

@@ -0,0 +1 @@
+{"dataset": "country211", "model": "ViT-B-16", "pretrained": "/mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-010-filled-map_text_farest/checkpoints/epoch_40.pt", "task": "zeroshot_classification", "metrics": {"acc1": 0.006682464454976303, "acc5": 0.036398104265402846, "mean_per_class_recall": 0.006729857819905213}, "language": "en"}

SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-010-filled-map_text_farest/benchmark_dtd_epoch_40.pt_ViT-B-16_en_zeroshot_classification.json

@@ -0,0 +1 @@
+{"dataset": "dtd", "model": "ViT-B-16", "pretrained": "/mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-010-filled-map_text_farest/checkpoints/epoch_40.pt", "task": "zeroshot_classification", "metrics": {"acc1": 0.08829787234042553, "acc5": 0.2712765957446808, "mean_per_class_recall": 0.08829787234042552}, "language": "en"}

SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-010-filled-map_text_farest/benchmark_eurosat_epoch_40.pt_ViT-B-16_en_zeroshot_classification.json

@@ -0,0 +1 @@
+{"dataset": "eurosat", "model": "ViT-B-16", "pretrained": "/mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-010-filled-map_text_farest/checkpoints/epoch_40.pt", "task": "zeroshot_classification", "metrics": {"acc1": 0.13437037037037036, "acc5": 0.5593333333333333, "mean_per_class_recall": 0.13030000000000003}, "language": "en"}

SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-010-filled-map_text_farest/benchmark_fgvc_aircraft_epoch_40.pt_ViT-B-16_en_zeroshot_classification.json

@@ -0,0 +1 @@
+{"dataset": "fgvc_aircraft", "model": "ViT-B-16", "pretrained": "/mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-010-filled-map_text_farest/checkpoints/epoch_40.pt", "task": "zeroshot_classification", "metrics": {"acc1": 0.0186018601860186, "acc5": 0.0702070207020702, "mean_per_class_recall": 0.019197860962566843}, "language": "en"}

SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-010-filled-map_text_farest/benchmark_flickr30k_epoch_40.pt_ViT-B-16_en_zeroshot_retrieval.json

@@ -0,0 +1 @@
+{"dataset": "flickr30k", "model": "ViT-B-16", "pretrained": "/mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-010-filled-map_text_farest/checkpoints/epoch_40.pt", "task": "zeroshot_retrieval", "metrics": {"image_retrieval_recall@1": 0.15539999306201935, "text_retrieval_recall@1": 0.24899999797344208, "image_retrieval_recall@5": 0.33820000290870667, "text_retrieval_recall@5": 0.5040000081062317, "image_retrieval_recall@10": 0.4323999881744385, "text_retrieval_recall@10": 0.6019999980926514}, "language": "en"}

SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-010-filled-map_text_farest/benchmark_flowers_epoch_40.pt_ViT-B-16_en_zeroshot_classification.json

@@ -0,0 +1 @@
+{"dataset": "flowers", "model": "ViT-B-16", "pretrained": "/mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-010-filled-map_text_farest/checkpoints/epoch_40.pt", "task": "zeroshot_classification", "metrics": {"acc1": 0.12554886973491625, "acc5": 0.2536997885835095, "mean_per_class_recall": 0.11819904573849162}, "language": "en"}

SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-010-filled-map_text_farest/benchmark_food101_epoch_40.pt_ViT-B-16_en_zeroshot_classification.json

@@ -0,0 +1 @@
+{"dataset": "food101", "model": "ViT-B-16", "pretrained": "/mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-010-filled-map_text_farest/checkpoints/epoch_40.pt", "task": "zeroshot_classification", "metrics": {"acc1": 0.1011881188118812, "acc5": 0.2512871287128713, "mean_per_class_recall": 0.10091089108910892}, "language": "en"}

SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-010-filled-map_text_farest/benchmark_gtsrb_epoch_40.pt_ViT-B-16_en_zeroshot_classification.json

@@ -0,0 +1 @@
+{"dataset": "gtsrb", "model": "ViT-B-16", "pretrained": "/mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-010-filled-map_text_farest/checkpoints/epoch_40.pt", "task": "zeroshot_classification", "metrics": {"acc1": 0.03776722090261283, "acc5": 0.14861441013460017, "mean_per_class_recall": 0.039929972052065076}, "language": "en"}

SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-010-filled-map_text_farest/benchmark_imagenet1k_epoch_40.pt_ViT-B-16_en_zeroshot_classification.json

@@ -0,0 +1 @@
+{"dataset": "imagenet1k", "model": "ViT-B-16", "pretrained": "/mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-010-filled-map_text_farest/checkpoints/epoch_40.pt", "task": "zeroshot_classification", "metrics": {"acc1": 0.15564, "acc5": 0.31012, "mean_per_class_recall": 0.15550000000000003}, "language": "en"}

SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-010-filled-map_text_farest/benchmark_mscoco_captions_epoch_40.pt_ViT-B-16_en_zeroshot_retrieval.json

@@ -0,0 +1 @@
+{"dataset": "mscoco_captions", "model": "ViT-B-16", "pretrained": "/mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-010-filled-map_text_farest/checkpoints/epoch_40.pt", "task": "zeroshot_retrieval", "metrics": {"image_retrieval_recall@1": 0.07149140536785126, "text_retrieval_recall@1": 0.12540000677108765, "image_retrieval_recall@5": 0.19992002844810486, "text_retrieval_recall@5": 0.2935999929904938, "image_retrieval_recall@10": 0.28032785654067993, "text_retrieval_recall@10": 0.39100000262260437}, "language": "en"}

SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-010-filled-map_text_farest/benchmark_pets_epoch_40.pt_ViT-B-16_en_zeroshot_classification.json

@@ -0,0 +1 @@
+{"dataset": "pets", "model": "ViT-B-16", "pretrained": "/mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-010-filled-map_text_farest/checkpoints/epoch_40.pt", "task": "zeroshot_classification", "metrics": {"acc1": 0.116107931316435, "acc5": 0.28672662850913055, "mean_per_class_recall": 0.11487310167490868}, "language": "en"}

SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-010-filled-map_text_farest/benchmark_stl10_epoch_40.pt_ViT-B-16_en_zeroshot_classification.json

@@ -0,0 +1 @@
+{"dataset": "stl10", "model": "ViT-B-16", "pretrained": "/mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-010-filled-map_text_farest/checkpoints/epoch_40.pt", "task": "zeroshot_classification", "metrics": {"acc1": 0.731125, "acc5": 0.9735, "mean_per_class_recall": 0.731375}, "language": "en"}

SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-010-filled-map_text_farest/benchmark_sun397_epoch_40.pt_ViT-B-16_en_zeroshot_classification.json

@@ -0,0 +1 @@
+{"dataset": "sun397", "model": "ViT-B-16", "pretrained": "/mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-010-filled-map_text_farest/checkpoints/epoch_40.pt", "task": "zeroshot_classification", "metrics": {"acc1": 0.2636224874487375, "acc5": 0.5219853982382258, "mean_per_class_recall": 0.22299183684324}, "language": "en"}

SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-010-filled-map_text_farest/benchmark_vtab_resisc45_epoch_40.pt_ViT-B-16_en_zeroshot_classification.json

@@ -0,0 +1 @@
+{"dataset": "vtab/resisc45", "model": "ViT-B-16", "pretrained": "/mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-010-filled-map_text_farest/checkpoints/epoch_40.pt", "task": "zeroshot_classification", "metrics": {"acc1": 0.15888888888888889, "acc5": 0.4058730158730159, "mean_per_class_recall": 0.16234808955215757}, "language": "en"}

SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-010-filled-map_text_farest/out.log

@@ -0,0 +1,394 @@
+2025-04-24,05:09:17 | INFO | Running in distributed mode with multiple processes. Device: cuda:0.Process (global: 0, local 0), total 4.
+2025-04-24,05:09:17 | INFO | Loaded ViT-B-16 model config.
+2025-04-24,05:09:18 | INFO | Model:
+2025-04-24,05:09:18 | INFO | CLIP(
+  (visual): VisionTransformer(
+    (conv1): Conv2d(3, 768, kernel_size=(16, 16), stride=(16, 16), bias=False)
+    (patch_dropout): Identity()
+    (ln_pre): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
+    (transformer): Transformer(
+      (resblocks): ModuleList(
+        (0-11): 12 x ResidualAttentionBlock(
+          (ln_1): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
+          (attn): MultiheadAttention(
+            (out_proj): NonDynamicallyQuantizableLinear(in_features=768, out_features=768, bias=True)
+          )
+          (ls_1): Identity()
+          (ln_2): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
+          (mlp): Sequential(
+            (c_fc): Linear(in_features=768, out_features=3072, bias=True)
+            (gelu): GELU(approximate='none')
+            (c_proj): Linear(in_features=3072, out_features=768, bias=True)
+          )
+          (ls_2): Identity()
+        )
+      )
+    )
+    (ln_post): LayerNorm((768,), eps=1e-05, elementwise_affine=True)
+  )
+  (transformer): Transformer(
+    (resblocks): ModuleList(
+      (0-11): 12 x ResidualAttentionBlock(
+        (ln_1): LayerNorm((512,), eps=1e-05, elementwise_affine=True)
+        (attn): MultiheadAttention(
+          (out_proj): NonDynamicallyQuantizableLinear(in_features=512, out_features=512, bias=True)
+        )
+        (ls_1): Identity()
+        (ln_2): LayerNorm((512,), eps=1e-05, elementwise_affine=True)
+        (mlp): Sequential(
+          (c_fc): Linear(in_features=512, out_features=2048, bias=True)
+          (gelu): GELU(approximate='none')
+          (c_proj): Linear(in_features=2048, out_features=512, bias=True)
+        )
+        (ls_2): Identity()
+      )
+    )
+  )
+  (token_embedding): Embedding(49408, 512)
+  (ln_final): LayerNorm((512,), eps=1e-05, elementwise_affine=True)
+)
+2025-04-24,05:09:18 | INFO | Params:
+2025-04-24,05:09:18 | INFO |   accum_freq: 4
+2025-04-24,05:09:18 | INFO |   aug_cfg: {}
+2025-04-24,05:09:18 | INFO |   batch_size: 512
+2025-04-24,05:09:18 | INFO |   beta1: 0.9
+2025-04-24,05:09:18 | INFO |   beta2: 0.98
+2025-04-24,05:09:18 | INFO |   cache_dir: None
+2025-04-24,05:09:18 | INFO |   caption_ratio: 0.1
+2025-04-24,05:09:18 | INFO |   checkpoint_path: /mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-010-filled-map_text_farest/checkpoints
+2025-04-24,05:09:18 | INFO |   coca_caption_loss_weight: 2.0
+2025-04-24,05:09:18 | INFO |   coca_contrastive_loss_weight: 1.0
+2025-04-24,05:09:18 | INFO |   copy_codebase: False
+2025-04-24,05:09:18 | INFO |   csv_caption_key: title
+2025-04-24,05:09:18 | INFO |   csv_img_key: filepath
+2025-04-24,05:09:18 | INFO |   csv_separator: 	
+2025-04-24,05:09:18 | INFO |   dataset_resampled: False
+2025-04-24,05:09:18 | INFO |   dataset_type: synthetic
+2025-04-24,05:09:18 | INFO |   ddp_static_graph: False
+2025-04-24,05:09:18 | INFO |   debug: False
+2025-04-24,05:09:18 | INFO |   delete_previous_checkpoint: False
+2025-04-24,05:09:18 | INFO |   device: cuda:0
+2025-04-24,05:09:18 | INFO |   dist_backend: None
+2025-04-24,05:09:18 | INFO |   dist_url: None
+2025-04-24,05:09:18 | INFO |   distill: False
+2025-04-24,05:09:18 | INFO |   distill_model: None
+2025-04-24,05:09:18 | INFO |   distill_pretrained: None
+2025-04-24,05:09:18 | INFO |   distributed: True
+2025-04-24,05:09:18 | INFO |   epochs: 40
+2025-04-24,05:09:18 | INFO |   epochs_cooldown: None
+2025-04-24,05:09:18 | INFO |   eps: 1e-08
+2025-04-24,05:09:18 | INFO |   force_custom_text: False
+2025-04-24,05:09:18 | INFO |   force_image_size: None
+2025-04-24,05:09:18 | INFO |   force_patch_dropout: None
+2025-04-24,05:09:18 | INFO |   force_quick_gelu: False
+2025-04-24,05:09:18 | INFO |   gather_with_grad: True
+2025-04-24,05:09:18 | INFO |   grad_checkpointing: True
+2025-04-24,05:09:18 | INFO |   grad_clip_norm: None
+2025-04-24,05:09:18 | INFO |   horovod: False
+2025-04-24,05:09:18 | INFO |   image_interpolation: None
+2025-04-24,05:09:18 | INFO |   image_mean: None
+2025-04-24,05:09:18 | INFO |   image_resize_mode: None
+2025-04-24,05:09:18 | INFO |   image_std: None
+2025-04-24,05:09:18 | INFO |   imagenet_v2: None
+2025-04-24,05:09:18 | INFO |   imagenet_val: None
+2025-04-24,05:09:18 | INFO |   keep_func_name: 
+2025-04-24,05:09:18 | INFO |   local_loss: False
+2025-04-24,05:09:18 | INFO |   local_rank: 0
+2025-04-24,05:09:18 | INFO |   lock_image: False
+2025-04-24,05:09:18 | INFO |   lock_image_freeze_bn_stats: False
+2025-04-24,05:09:18 | INFO |   lock_image_unlocked_groups: 0
+2025-04-24,05:09:18 | INFO |   lock_text: False
+2025-04-24,05:09:18 | INFO |   lock_text_freeze_layer_norm: False
+2025-04-24,05:09:18 | INFO |   lock_text_unlocked_layers: 0
+2025-04-24,05:09:18 | INFO |   log_every_n_steps: 100
+2025-04-24,05:09:18 | INFO |   log_level: 20
+2025-04-24,05:09:18 | INFO |   log_local: False
+2025-04-24,05:09:18 | INFO |   log_path: /mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-010-filled-map_text_farest/out.log
+2025-04-24,05:09:18 | INFO |   logs: /mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs
+2025-04-24,05:09:18 | INFO |   loss_dist_impl: None
+2025-04-24,05:09:18 | INFO |   lr: 0.001
+2025-04-24,05:09:18 | INFO |   lr_cooldown_end: 0.0
+2025-04-24,05:09:18 | INFO |   lr_cooldown_power: 1.0
+2025-04-24,05:09:18 | INFO |   lr_scheduler: cosine
+2025-04-24,05:09:18 | INFO |   map_func_name: map_text_farest
+2025-04-24,05:09:18 | INFO |   model: ViT-B-16
+2025-04-24,05:09:18 | INFO |   momentum: None
+2025-04-24,05:09:18 | INFO |   name: ViT-B-16-cc3m-laclip-mix-010-filled-map_text_farest
+2025-04-24,05:09:18 | INFO |   no_set_device_rank: False
+2025-04-24,05:09:18 | INFO |   opt: adamw
+2025-04-24,05:09:18 | INFO |   precision: amp
+2025-04-24,05:09:18 | INFO |   pretrained: 
+2025-04-24,05:09:18 | INFO |   pretrained_image: False
+2025-04-24,05:09:18 | INFO |   rank: 0
+2025-04-24,05:09:18 | INFO |   remote_sync: None
+2025-04-24,05:09:18 | INFO |   remote_sync_frequency: 300
+2025-04-24,05:09:18 | INFO |   remote_sync_protocol: s3
+2025-04-24,05:09:18 | INFO |   report_to: tensorboard,wandb
+2025-04-24,05:09:18 | INFO |   resume: None
+2025-04-24,05:09:18 | INFO |   save_frequency: 1
+2025-04-24,05:09:18 | INFO |   save_most_recent: False
+2025-04-24,05:09:18 | INFO |   seed: 0
+2025-04-24,05:09:18 | INFO |   siglip: False
+2025-04-24,05:09:18 | INFO |   skip_scheduler: False
+2025-04-24,05:09:18 | INFO |   tensorboard: True
+2025-04-24,05:09:18 | INFO |   tensorboard_path: /mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-010-filled-map_text_farest/tensorboard
+2025-04-24,05:09:18 | INFO |   torchcompile: False
+2025-04-24,05:09:18 | INFO |   torchscript: False
+2025-04-24,05:09:18 | INFO |   trace: False
+2025-04-24,05:09:18 | INFO |   train_data: /mnt/personal/zhudongy/cc3m-hgf-wds/{0000..0301}.tar
+2025-04-24,05:09:18 | INFO |   train_data_upsampling_factors: None
+2025-04-24,05:09:18 | INFO |   train_num_samples: 3016640
+2025-04-24,05:09:18 | INFO |   use_bn_sync: False
+2025-04-24,05:09:18 | INFO |   use_bnb_linear: None
+2025-04-24,05:09:18 | INFO |   val_data: None
+2025-04-24,05:09:18 | INFO |   val_frequency: 1
+2025-04-24,05:09:18 | INFO |   val_num_samples: None
+2025-04-24,05:09:18 | INFO |   wandb: True
+2025-04-24,05:09:18 | INFO |   wandb_notes: 
+2025-04-24,05:09:18 | INFO |   wandb_project_name: open-clip
+2025-04-24,05:09:18 | INFO |   warmup: 368
+2025-04-24,05:09:18 | INFO |   wd: 0.5
+2025-04-24,05:09:18 | INFO |   workers: 16
+2025-04-24,05:09:18 | INFO |   world_size: 4
+2025-04-24,05:09:18 | INFO |   zeroshot_frequency: 2
+2025-04-24,05:09:20 | INFO | Created AdamW (adamw) optimizer: lr: 0.001, betas: (0.9, 0.98), eps: 1e-08, weight_decay: 0.5, amsgrad: False, foreach: None, maximize: False, capturable: False, differentiable: False, fused: None
+2025-04-24,05:12:19 | INFO | Start epoch 0
+2025-04-24,05:12:37 | INFO | Train Epoch: 0 [   8192/3047424 (0%)] Data (t): 12.573 Batch (t): 18.123, 452.019/s, 113.005/s/gpu LR: 0.000003 Logit Scale: 14.286 Imm_image: -0.27248 (-0.27248) Imm_text: -0.27248 (-0.27248) Isd_image: -0.27242 (-0.27242) Isd_text: -0.27242 (-0.27242) Contrastive_loss: 9.1249 (9.1249) Loss: 9.1249 (9.1249)
+2025-04-24,05:20:36 | INFO | Train Epoch: 0 [ 827392/3047424 (27%)] Data (t): 0.585 Batch (t): 4.789, 1715.44/s, 428.859/s/gpu LR: 0.000274 Logit Scale: 14.256 Imm_image: 4.0758 (1.9017) Imm_text: 4.0758 (1.9017) Isd_image: 2.4775 (1.1025) Isd_text: 2.4775 (1.1025) Contrastive_loss: 8.2617 (8.6933) Loss: 8.2617 (8.6933)
+2025-04-24,05:28:38 | INFO | Train Epoch: 0 [1646592/3047424 (54%)] Data (t): 0.594 Batch (t): 4.825, 1691.58/s, 422.894/s/gpu LR: 0.000546 Logit Scale: 14.173 Imm_image: 5.6317 (3.1450) Imm_text: 5.6317 (3.1450) Isd_image: 2.9038 (1.7030) Isd_text: 2.9038 (1.7030) Contrastive_loss: 7.8079 (8.3981) Loss: 7.8079 (8.3981)
+2025-04-24,05:36:41 | INFO | Train Epoch: 0 [2465792/3047424 (81%)] Data (t): 0.617 Batch (t): 4.833, 1706.59/s, 426.648/s/gpu LR: 0.000818 Logit Scale: 14.103 Imm_image: 6.5782 (4.0033) Imm_text: 6.5782 (4.0033) Isd_image: 2.7935 (1.9756) Isd_text: 2.7935 (1.9756) Contrastive_loss: 7.4406 (8.1588) Loss: 7.4406 (8.1588)
+2025-04-24,05:42:22 | INFO | Train Epoch: 0 [3047424/3047424 (100%)] Data (t): 0.584 Batch (t): 4.800, 1745.27/s, 436.317/s/gpu LR: 0.001000 Logit Scale: 13.980 Imm_image: 6.1444 (4.4315) Imm_text: 6.1444 (4.4315) Isd_image: 2.7891 (2.1383) Isd_text: 2.7891 (2.1383) Contrastive_loss: 7.5790 (8.0428) Loss: 7.5790 (8.0428)
+2025-04-24,05:42:24 | INFO | Start epoch 1
+2025-04-24,05:42:38 | INFO | Train Epoch: 1 [   8192/3047424 (0%)] Data (t): 9.889 Batch (t): 14.276, 573.834/s, 143.458/s/gpu LR: 0.001000 Logit Scale: 13.978 Imm_image: 6.2627 (6.2627) Imm_text: 6.2627 (6.2627) Isd_image: 2.8676 (2.8676) Isd_text: 2.8676 (2.8676) Contrastive_loss: 7.5139 (7.5139) Loss: 7.5139 (7.5139)
+2025-04-24,05:50:43 | INFO | Train Epoch: 1 [ 827392/3047424 (27%)] Data (t): 0.609 Batch (t): 4.845, 1688.88/s, 422.221/s/gpu LR: 0.001000 Logit Scale: 14.156 Imm_image: 7.5471 (6.9049) Imm_text: 7.5471 (6.9049) Isd_image: 2.4874 (2.6775) Isd_text: 2.4874 (2.6775) Contrastive_loss: 6.8718 (7.1929) Loss: 6.8718 (7.1929)
+2025-04-24,05:58:48 | INFO | Train Epoch: 1 [1646592/3047424 (54%)] Data (t): 0.590 Batch (t): 4.847, 1707.76/s, 426.941/s/gpu LR: 0.001000 Logit Scale: 14.832 Imm_image: 7.5910 (7.1336) Imm_text: 7.5910 (7.1336) Isd_image: 2.1025 (2.4859) Isd_text: 2.1025 (2.4859) Contrastive_loss: 6.6634 (7.0164) Loss: 6.6634 (7.0164)
+2025-04-24,06:06:53 | INFO | Train Epoch: 1 [2465792/3047424 (81%)] Data (t): 0.613 Batch (t): 4.850, 1684.50/s, 421.125/s/gpu LR: 0.000999 Logit Scale: 15.812 Imm_image: 7.9976 (7.3496) Imm_text: 7.9976 (7.3496) Isd_image: 1.7529 (2.3026) Isd_text: 1.7529 (2.3026) Contrastive_loss: 6.2950 (6.8360) Loss: 6.2950 (6.8360)
+2025-04-24,06:12:36 | INFO | Train Epoch: 1 [3047424/3047424 (100%)] Data (t): 0.603 Batch (t): 4.833, 1742.21/s, 435.553/s/gpu LR: 0.000998 Logit Scale: 16.557 Imm_image: 8.3813 (7.5559) Imm_text: 8.3813 (7.5559) Isd_image: 1.5525 (2.1526) Isd_text: 1.5525 (2.1526) Contrastive_loss: 6.0689 (6.6826) Loss: 6.0689 (6.6826)
+2025-04-24,06:12:38 | INFO | Start epoch 2
+2025-04-24,06:12:52 | INFO | Train Epoch: 2 [   8192/3047424 (0%)] Data (t): 10.159 Batch (t): 14.470, 566.151/s, 141.538/s/gpu LR: 0.000998 Logit Scale: 16.570 Imm_image: 8.5871 (8.5871) Imm_text: 8.5871 (8.5871) Isd_image: 1.5260 (1.5260) Isd_text: 1.5260 (1.5260) Contrastive_loss: 5.8726 (5.8726) Loss: 5.8726 (5.8726)
+2025-04-24,06:20:55 | INFO | Train Epoch: 2 [ 827392/3047424 (27%)] Data (t): 0.605 Batch (t): 4.835, 1709.46/s, 427.365/s/gpu LR: 0.000997 Logit Scale: 17.754 Imm_image: 8.8804 (8.7338) Imm_text: 8.8804 (8.7338) Isd_image: 1.2524 (1.3892) Isd_text: 1.2524 (1.3892) Contrastive_loss: 5.7131 (5.7928) Loss: 5.7131 (5.7928)
+2025-04-24,06:28:57 | INFO | Train Epoch: 2 [1646592/3047424 (54%)] Data (t): 0.599 Batch (t): 4.818, 1698.38/s, 424.596/s/gpu LR: 0.000996 Logit Scale: 19.189 Imm_image: 9.4785 (8.9820) Imm_text: 9.4785 (8.9820) Isd_image: 1.2114 (1.3299) Isd_text: 1.2114 (1.3299) Contrastive_loss: 5.4268 (5.6708) Loss: 5.4268 (5.6708)
+2025-04-24,06:37:01 | INFO | Train Epoch: 2 [2465792/3047424 (81%)] Data (t): 0.607 Batch (t): 4.833, 1722.32/s, 430.579/s/gpu LR: 0.000995 Logit Scale: 20.503 Imm_image: 9.9785 (9.2311) Imm_text: 9.9785 (9.2311) Isd_image: 1.2081 (1.2995) Isd_text: 1.2081 (1.2995) Contrastive_loss: 5.1914 (5.5510) Loss: 5.1914 (5.5510)
+2025-04-24,06:42:39 | INFO | Train Epoch: 2 [3047424/3047424 (100%)] Data (t): 0.569 Batch (t): 4.772, 1735.67/s, 433.917/s/gpu LR: 0.000993 Logit Scale: 21.401 Imm_image: 10.517 (9.4883) Imm_text: 10.517 (9.4883) Isd_image: 1.5058 (1.3407) Isd_text: 1.5058 (1.3407) Contrastive_loss: 5.0237 (5.4455) Loss: 5.0237 (5.4455)
+2025-04-24,06:42:41 | INFO | Start epoch 3
+2025-04-24,06:42:56 | INFO | Train Epoch: 3 [   8192/3047424 (0%)] Data (t): 10.840 Batch (t): 15.184, 539.516/s, 134.879/s/gpu LR: 0.000993 Logit Scale: 21.417 Imm_image: 10.546 (10.546) Imm_text: 10.546 (10.546) Isd_image: 1.2462 (1.2462) Isd_text: 1.2462 (1.2462) Contrastive_loss: 4.9446 (4.9446) Loss: 4.9446 (4.9446)
+2025-04-24,06:50:57 | INFO | Train Epoch: 3 [ 827392/3047424 (27%)] Data (t): 0.601 Batch (t): 4.810, 1687.46/s, 421.865/s/gpu LR: 0.000992 Logit Scale: 22.945 Imm_image: 10.980 (10.763) Imm_text: 10.980 (10.763) Isd_image: 1.2356 (1.2409) Isd_text: 1.2356 (1.2409) Contrastive_loss: 4.8382 (4.8914) Loss: 4.8382 (4.8914)
+2025-04-24,06:59:01 | INFO | Train Epoch: 3 [1646592/3047424 (54%)] Data (t): 0.605 Batch (t): 4.839, 1587.84/s, 396.959/s/gpu LR: 0.000990 Logit Scale: 24.175 Imm_image: 11.207 (10.911) Imm_text: 11.207 (10.911) Isd_image: 1.4343 (1.3054) Isd_text: 1.4343 (1.3054) Contrastive_loss: 4.7593 (4.8474) Loss: 4.7593 (4.8474)
+2025-04-24,07:07:10 | INFO | Train Epoch: 3 [2465792/3047424 (81%)] Data (t): 0.639 Batch (t): 4.891, 1687.74/s, 421.934/s/gpu LR: 0.000987 Logit Scale: 25.375 Imm_image: 11.901 (11.159) Imm_text: 11.901 (11.159) Isd_image: 1.3960 (1.3280) Isd_text: 1.3960 (1.3280) Contrastive_loss: 4.5970 (4.7848) Loss: 4.5970 (4.7848)
+2025-04-24,07:12:55 | INFO | Train Epoch: 3 [3047424/3047424 (100%)] Data (t): 0.635 Batch (t): 4.854, 1732.30/s, 433.074/s/gpu LR: 0.000985 Logit Scale: 26.009 Imm_image: 12.362 (11.399) Imm_text: 12.362 (11.399) Isd_image: 1.9024 (1.4429) Isd_text: 1.9024 (1.4429) Contrastive_loss: 4.3290 (4.6936) Loss: 4.3290 (4.6936)
+2025-04-24,07:12:57 | INFO | Start epoch 4
+2025-04-24,07:13:11 | INFO | Train Epoch: 4 [   8192/3047424 (0%)] Data (t): 10.128 Batch (t): 14.552, 562.949/s, 140.737/s/gpu LR: 0.000985 Logit Scale: 26.021 Imm_image: 12.614 (12.614) Imm_text: 12.614 (12.614) Isd_image: 1.5877 (1.5877) Isd_text: 1.5877 (1.5877) Contrastive_loss: 4.1289 (4.1289) Loss: 4.1289 (4.1289)
+2025-04-24,07:21:17 | INFO | Train Epoch: 4 [ 827392/3047424 (27%)] Data (t): 0.623 Batch (t): 4.853, 1710.17/s, 427.544/s/gpu LR: 0.000983 Logit Scale: 27.294 Imm_image: 12.985 (12.800) Imm_text: 12.985 (12.800) Isd_image: 1.7843 (1.6860) Isd_text: 1.7843 (1.6860) Contrastive_loss: 4.1899 (4.1594) Loss: 4.1899 (4.1594)
+2025-04-24,07:29:18 | INFO | Train Epoch: 4 [1646592/3047424 (54%)] Data (t): 0.605 Batch (t): 4.819, 1708.23/s, 427.057/s/gpu LR: 0.000980 Logit Scale: 28.234 Imm_image: 12.913 (12.838) Imm_text: 12.913 (12.838) Isd_image: 1.8029 (1.7250) Isd_text: 1.8029 (1.7250) Contrastive_loss: 4.2980 (4.2056) Loss: 4.2980 (4.2056)
+2025-04-24,07:37:21 | INFO | Train Epoch: 4 [2465792/3047424 (81%)] Data (t): 0.604 Batch (t): 4.824, 1698.52/s, 424.631/s/gpu LR: 0.000977 Logit Scale: 29.032 Imm_image: 13.574 (13.022) Imm_text: 13.574 (13.022) Isd_image: 2.0237 (1.7996) Isd_text: 2.0237 (1.7996) Contrastive_loss: 4.0736 (4.1726) Loss: 4.0736 (4.1726)
+2025-04-24,07:43:03 | INFO | Train Epoch: 4 [3047424/3047424 (100%)] Data (t): 0.583 Batch (t): 4.823, 1740.81/s, 435.201/s/gpu LR: 0.000974 Logit Scale: 29.522 Imm_image: 13.967 (13.211) Imm_text: 13.967 (13.211) Isd_image: 2.0667 (1.8531) Isd_text: 2.0667 (1.8531) Contrastive_loss: 3.8135 (4.1008) Loss: 3.8135 (4.1008)
+2025-04-24,07:43:05 | INFO | Start epoch 5
+2025-04-24,07:43:20 | INFO | Train Epoch: 5 [   8192/3047424 (0%)] Data (t): 9.951 Batch (t): 14.357, 570.597/s, 142.649/s/gpu LR: 0.000974 Logit Scale: 29.532 Imm_image: 14.031 (14.031) Imm_text: 14.031 (14.031) Isd_image: 1.7906 (1.7906) Isd_text: 1.7906 (1.7906) Contrastive_loss: 3.7327 (3.7327) Loss: 3.7327 (3.7327)
+2025-04-24,07:51:23 | INFO | Train Epoch: 5 [ 827392/3047424 (27%)] Data (t): 0.614 Batch (t): 4.834, 1696.07/s, 424.017/s/gpu LR: 0.000971 Logit Scale: 30.600 Imm_image: 14.282 (14.157) Imm_text: 14.282 (14.157) Isd_image: 1.7615 (1.7760) Isd_text: 1.7615 (1.7760) Contrastive_loss: 3.7227 (3.7277) Loss: 3.7227 (3.7277)
+2025-04-24,07:59:29 | INFO | Train Epoch: 5 [1646592/3047424 (54%)] Data (t): 0.625 Batch (t): 4.857, 1682.89/s, 420.723/s/gpu LR: 0.000967 Logit Scale: 31.521 Imm_image: 14.629 (14.314) Imm_text: 14.629 (14.314) Isd_image: 2.0761 (1.8760) Isd_text: 2.0761 (1.8760) Contrastive_loss: 3.7711 (3.7421) Loss: 3.7711 (3.7421)
+2025-04-24,08:07:33 | INFO | Train Epoch: 5 [2465792/3047424 (81%)] Data (t): 0.613 Batch (t): 4.840, 1678.68/s, 419.670/s/gpu LR: 0.000963 Logit Scale: 32.162 Imm_image: 14.946 (14.472) Imm_text: 14.946 (14.472) Isd_image: 2.4353 (2.0158) Isd_text: 2.4353 (2.0158) Contrastive_loss: 3.7038 (3.7326) Loss: 3.7038 (3.7326)
+2025-04-24,08:13:15 | INFO | Train Epoch: 5 [3047424/3047424 (100%)] Data (t): 0.614 Batch (t): 4.820, 1736.84/s, 434.210/s/gpu LR: 0.000960 Logit Scale: 32.518 Imm_image: 15.194 (14.616) Imm_text: 15.194 (14.616) Isd_image: 2.5951 (2.1317) Isd_text: 2.5951 (2.1317) Contrastive_loss: 3.5552 (3.6971) Loss: 3.5552 (3.6971)
+2025-04-24,08:13:17 | INFO | Start epoch 6
+2025-04-24,08:13:32 | INFO | Train Epoch: 6 [   8192/3047424 (0%)] Data (t): 10.187 Batch (t): 14.584, 561.725/s, 140.431/s/gpu LR: 0.000960 Logit Scale: 32.526 Imm_image: 15.511 (15.511) Imm_text: 15.511 (15.511) Isd_image: 2.2997 (2.2997) Isd_text: 2.2997 (2.2997) Contrastive_loss: 3.3560 (3.3560) Loss: 3.3560 (3.3560)
+2025-04-24,08:21:36 | INFO | Train Epoch: 6 [ 827392/3047424 (27%)] Data (t): 0.616 Batch (t): 4.846, 1677.77/s, 419.441/s/gpu LR: 0.000955 Logit Scale: 33.495 Imm_image: 15.670 (15.591) Imm_text: 15.670 (15.591) Isd_image: 2.1325 (2.2161) Isd_text: 2.1325 (2.2161) Contrastive_loss: 3.3925 (3.3743) Loss: 3.3925 (3.3743)
+2025-04-24,08:29:40 | INFO | Train Epoch: 6 [1646592/3047424 (54%)] Data (t): 0.605 Batch (t): 4.839, 1671.29/s, 417.822/s/gpu LR: 0.000951 Logit Scale: 34.107 Imm_image: 15.774 (15.652) Imm_text: 15.774 (15.652) Isd_image: 2.5329 (2.3217) Isd_text: 2.5329 (2.3217) Contrastive_loss: 3.4465 (3.3983) Loss: 3.4465 (3.3983)
+2025-04-24,08:37:44 | INFO | Train Epoch: 6 [2465792/3047424 (81%)] Data (t): 0.610 Batch (t): 4.842, 1712.82/s, 428.204/s/gpu LR: 0.000946 Logit Scale: 34.679 Imm_image: 15.966 (15.730) Imm_text: 15.966 (15.730) Isd_image: 2.5756 (2.3852) Isd_text: 2.5756 (2.3852) Contrastive_loss: 3.4275 (3.4056) Loss: 3.4275 (3.4056)
+2025-04-24,08:43:27 | INFO | Train Epoch: 6 [3047424/3047424 (100%)] Data (t): 0.610 Batch (t): 4.831, 1732.92/s, 433.231/s/gpu LR: 0.000943 Logit Scale: 34.919 Imm_image: 16.362 (15.857) Imm_text: 16.362 (15.857) Isd_image: 2.8511 (2.4784) Isd_text: 2.8511 (2.4784) Contrastive_loss: 3.2097 (3.3664) Loss: 3.2097 (3.3664)
+2025-04-24,08:43:29 | INFO | Start epoch 7
+2025-04-24,08:43:44 | INFO | Train Epoch: 7 [   8192/3047424 (0%)] Data (t): 10.389 Batch (t): 14.748, 555.473/s, 138.868/s/gpu LR: 0.000943 Logit Scale: 34.924 Imm_image: 16.599 (16.599) Imm_text: 16.599 (16.599) Isd_image: 2.7080 (2.7080) Isd_text: 2.7080 (2.7080) Contrastive_loss: 3.0420 (3.0420) Loss: 3.0420 (3.0420)
+2025-04-24,08:51:50 | INFO | Train Epoch: 7 [ 827392/3047424 (27%)] Data (t): 0.628 Batch (t): 4.859, 1702.73/s, 425.683/s/gpu LR: 0.000937 Logit Scale: 35.780 Imm_image: 16.502 (16.551) Imm_text: 16.502 (16.551) Isd_image: 2.4893 (2.5987) Isd_text: 2.4893 (2.5987) Contrastive_loss: 3.2388 (3.1404) Loss: 3.2388 (3.1404)
+2025-04-24,08:59:50 | INFO | Train Epoch: 7 [1646592/3047424 (54%)] Data (t): 0.607 Batch (t): 4.805, 1707.89/s, 426.972/s/gpu LR: 0.000932 Logit Scale: 36.399 Imm_image: 16.843 (16.648) Imm_text: 16.843 (16.648) Isd_image: 2.7741 (2.6571) Isd_text: 2.7741 (2.6571) Contrastive_loss: 3.1931 (3.1580) Loss: 3.1931 (3.1580)
+2025-04-24,09:07:55 | INFO | Train Epoch: 7 [2465792/3047424 (81%)] Data (t): 0.619 Batch (t): 4.844, 1694.17/s, 423.542/s/gpu LR: 0.000927 Logit Scale: 36.773 Imm_image: 16.939 (16.721) Imm_text: 16.939 (16.721) Isd_image: 2.9737 (2.7363) Isd_text: 2.9737 (2.7363) Contrastive_loss: 3.2297 (3.1759) Loss: 3.2297 (3.1759)
+2025-04-24,09:13:37 | INFO | Train Epoch: 7 [3047424/3047424 (100%)] Data (t): 0.598 Batch (t): 4.816, 1735.94/s, 433.984/s/gpu LR: 0.000922 Logit Scale: 36.981 Imm_image: 17.322 (16.841) Imm_text: 17.322 (16.841) Isd_image: 2.9760 (2.7842) Isd_text: 2.9760 (2.7842) Contrastive_loss: 2.9649 (3.1337) Loss: 2.9649 (3.1337)
+2025-04-24,09:13:39 | INFO | Start epoch 8
+2025-04-24,09:13:53 | INFO | Train Epoch: 8 [   8192/3047424 (0%)] Data (t): 9.930 Batch (t): 14.306, 572.634/s, 143.158/s/gpu LR: 0.000922 Logit Scale: 36.988 Imm_image: 17.531 (17.531) Imm_text: 17.531 (17.531) Isd_image: 2.8298 (2.8298) Isd_text: 2.8298 (2.8298) Contrastive_loss: 2.8142 (2.8142) Loss: 2.8142 (2.8142)
+2025-04-24,09:21:58 | INFO | Train Epoch: 8 [ 827392/3047424 (27%)] Data (t): 0.621 Batch (t): 4.848, 1711.11/s, 427.776/s/gpu LR: 0.000917 Logit Scale: 37.840 Imm_image: 17.644 (17.588) Imm_text: 17.644 (17.588) Isd_image: 2.6992 (2.7645) Isd_text: 2.6992 (2.7645) Contrastive_loss: 2.9093 (2.8617) Loss: 2.9093 (2.8617)
+2025-04-24,09:30:02 | INFO | Train Epoch: 8 [1646592/3047424 (54%)] Data (t): 0.602 Batch (t): 4.839, 1677.93/s, 419.482/s/gpu LR: 0.000910 Logit Scale: 38.162 Imm_image: 17.679 (17.618) Imm_text: 17.679 (17.618) Isd_image: 2.8065 (2.7785) Isd_text: 2.8065 (2.7785) Contrastive_loss: 2.9823 (2.9019) Loss: 2.9823 (2.9019)
+2025-04-24,09:38:07 | INFO | Train Epoch: 8 [2465792/3047424 (81%)] Data (t): 0.619 Batch (t): 4.854, 1705.39/s, 426.348/s/gpu LR: 0.000904 Logit Scale: 38.597 Imm_image: 17.826 (17.670) Imm_text: 17.826 (17.670) Isd_image: 3.1703 (2.8764) Isd_text: 3.1703 (2.8764) Contrastive_loss: 3.0150 (2.9302) Loss: 3.0150 (2.9302)
+2025-04-24,09:43:51 | INFO | Train Epoch: 8 [3047424/3047424 (100%)] Data (t): 0.606 Batch (t): 4.839, 1732.50/s, 433.126/s/gpu LR: 0.000900 Logit Scale: 38.759 Imm_image: 18.172 (17.770) Imm_text: 18.172 (17.770) Isd_image: 3.4326 (2.9877) Isd_text: 3.4326 (2.9877) Contrastive_loss: 2.8022 (2.9046) Loss: 2.8022 (2.9046)
+2025-04-24,09:43:53 | INFO | Start epoch 9
+2025-04-24,09:44:08 | INFO | Train Epoch: 9 [   8192/3047424 (0%)] Data (t): 10.351 Batch (t): 14.785, 554.074/s, 138.519/s/gpu LR: 0.000900 Logit Scale: 38.761 Imm_image: 18.324 (18.324) Imm_text: 18.324 (18.324) Isd_image: 3.3085 (3.3085) Isd_text: 3.3085 (3.3085) Contrastive_loss: 2.6604 (2.6604) Loss: 2.6604 (2.6604)
+2025-04-24,09:52:12 | INFO | Train Epoch: 9 [ 827392/3047424 (27%)] Data (t): 0.624 Batch (t): 4.840, 1706.60/s, 426.650/s/gpu LR: 0.000893 Logit Scale: 39.597 Imm_image: 18.435 (18.379) Imm_text: 18.435 (18.379) Isd_image: 2.8770 (3.0927) Isd_text: 2.8770 (3.0927) Contrastive_loss: 2.7097 (2.6850) Loss: 2.7097 (2.6850)
+2025-04-24,10:00:17 | INFO | Train Epoch: 9 [1646592/3047424 (54%)] Data (t): 0.608 Batch (t): 4.850, 1687.08/s, 421.771/s/gpu LR: 0.000886 Logit Scale: 40.055 Imm_image: 18.559 (18.439) Imm_text: 18.559 (18.439) Isd_image: 3.3240 (3.1698) Isd_text: 3.3240 (3.1698) Contrastive_loss: 2.8195 (2.7298) Loss: 2.8195 (2.7298)
+2025-04-24,10:08:20 | INFO | Train Epoch: 9 [2465792/3047424 (81%)] Data (t): 0.616 Batch (t): 4.837, 1721.37/s, 430.343/s/gpu LR: 0.000879 Logit Scale: 40.321 Imm_image: 18.527 (18.461) Imm_text: 18.527 (18.461) Isd_image: 3.2977 (3.2018) Isd_text: 3.2977 (3.2018) Contrastive_loss: 2.8481 (2.7594) Loss: 2.8481 (2.7594)
+2025-04-24,10:13:59 | INFO | Train Epoch: 9 [3047424/3047424 (100%)] Data (t): 0.562 Batch (t): 4.768, 1745.04/s, 436.259/s/gpu LR: 0.000874 Logit Scale: 40.466 Imm_image: 18.969 (18.563) Imm_text: 18.969 (18.563) Isd_image: 3.5471 (3.2709) Isd_text: 3.5471 (3.2709) Contrastive_loss: 2.5811 (2.7238) Loss: 2.5811 (2.7238)
+2025-04-24,10:14:00 | INFO | Start epoch 10
+2025-04-24,10:14:14 | INFO | Train Epoch: 10 [   8192/3047424 (0%)] Data (t): 9.151 Batch (t): 13.488, 607.339/s, 151.835/s/gpu LR: 0.000874 Logit Scale: 40.469 Imm_image: 19.144 (19.144) Imm_text: 19.144 (19.144) Isd_image: 3.2687 (3.2687) Isd_text: 3.2687 (3.2687) Contrastive_loss: 2.4556 (2.4556) Loss: 2.4556 (2.4556)
+2025-04-24,10:22:18 | INFO | Train Epoch: 10 [ 827392/3047424 (27%)] Data (t): 0.615 Batch (t): 4.840, 1703.11/s, 425.778/s/gpu LR: 0.000867 Logit Scale: 41.302 Imm_image: 19.191 (19.168) Imm_text: 19.191 (19.168) Isd_image: 2.9065 (3.0876) Isd_text: 2.9065 (3.0876) Contrastive_loss: 2.5713 (2.5134) Loss: 2.5713 (2.5134)
+2025-04-24,10:30:24 | INFO | Train Epoch: 10 [1646592/3047424 (54%)] Data (t): 0.612 Batch (t): 4.857, 1649.66/s, 412.416/s/gpu LR: 0.000859 Logit Scale: 41.701 Imm_image: 19.253 (19.196) Imm_text: 19.253 (19.196) Isd_image: 3.3075 (3.1609) Isd_text: 3.3075 (3.1609) Contrastive_loss: 2.6768 (2.5679) Loss: 2.6768 (2.5679)
+2025-04-24,10:38:28 | INFO | Train Epoch: 10 [2465792/3047424 (81%)] Data (t): 0.605 Batch (t): 4.839, 1706.64/s, 426.660/s/gpu LR: 0.000852 Logit Scale: 41.953 Imm_image: 19.443 (19.258) Imm_text: 19.443 (19.258) Isd_image: 3.5797 (3.2656) Isd_text: 3.5797 (3.2656) Contrastive_loss: 2.6381 (2.5854) Loss: 2.6381 (2.5854)
+2025-04-24,10:44:09 | INFO | Train Epoch: 10 [3047424/3047424 (100%)] Data (t): 0.591 Batch (t): 4.808, 1744.83/s, 436.207/s/gpu LR: 0.000846 Logit Scale: 42.090 Imm_image: 19.719 (19.350) Imm_text: 19.719 (19.350) Isd_image: 3.5710 (3.3267) Isd_text: 3.5710 (3.3267) Contrastive_loss: 2.3999 (2.5483) Loss: 2.3999 (2.5483)
+2025-04-24,10:44:11 | INFO | Start epoch 11
+2025-04-24,10:44:24 | INFO | Train Epoch: 11 [   8192/3047424 (0%)] Data (t): 9.157 Batch (t): 13.523, 605.794/s, 151.448/s/gpu LR: 0.000846 Logit Scale: 42.093 Imm_image: 19.903 (19.903) Imm_text: 19.903 (19.903) Isd_image: 3.5206 (3.5206) Isd_text: 3.5206 (3.5206) Contrastive_loss: 2.2576 (2.2576) Loss: 2.2576 (2.2576)
+2025-04-24,10:52:29 | INFO | Train Epoch: 11 [ 827392/3047424 (27%)] Data (t): 0.625 Batch (t): 4.848, 1704.11/s, 426.028/s/gpu LR: 0.000838 Logit Scale: 42.893 Imm_image: 19.886 (19.894) Imm_text: 19.886 (19.894) Isd_image: 3.2376 (3.3791) Isd_text: 3.2376 (3.3791) Contrastive_loss: 2.4553 (2.3565) Loss: 2.4553 (2.3565)
+2025-04-24,11:00:32 | INFO | Train Epoch: 11 [1646592/3047424 (54%)] Data (t): 0.597 Batch (t): 4.827, 1672.73/s, 418.182/s/gpu LR: 0.000830 Logit Scale: 43.284 Imm_image: 20.012 (19.934) Imm_text: 20.012 (19.934) Isd_image: 3.5744 (3.4442) Isd_text: 3.5744 (3.4442) Contrastive_loss: 2.5284 (2.4138) Loss: 2.5284 (2.4138)
+2025-04-24,11:08:36 | INFO | Train Epoch: 11 [2465792/3047424 (81%)] Data (t): 0.603 Batch (t): 4.840, 1715.69/s, 428.922/s/gpu LR: 0.000822 Logit Scale: 43.374 Imm_image: 20.036 (19.959) Imm_text: 20.036 (19.959) Isd_image: 3.8859 (3.5546) Isd_text: 3.8859 (3.5546) Contrastive_loss: 2.5190 (2.4401) Loss: 2.5190 (2.4401)
+2025-04-24,11:14:16 | INFO | Train Epoch: 11 [3047424/3047424 (100%)] Data (t): 0.591 Batch (t): 4.788, 1736.18/s, 434.044/s/gpu LR: 0.000816 Logit Scale: 43.569 Imm_image: 20.538 (20.075) Imm_text: 20.538 (20.075) Isd_image: 3.7943 (3.6026) Isd_text: 3.7943 (3.6026) Contrastive_loss: 2.2289 (2.3979) Loss: 2.2289 (2.3979)
+2025-04-24,11:14:18 | INFO | Start epoch 12
+2025-04-24,11:14:32 | INFO | Train Epoch: 12 [   8192/3047424 (0%)] Data (t): 9.950 Batch (t): 14.333, 571.561/s, 142.890/s/gpu LR: 0.000816 Logit Scale: 43.576 Imm_image: 20.844 (20.844) Imm_text: 20.844 (20.844) Isd_image: 3.5646 (3.5646) Isd_text: 3.5646 (3.5646) Contrastive_loss: 2.0229 (2.0229) Loss: 2.0229 (2.0229)
+2025-04-24,11:22:37 | INFO | Train Epoch: 12 [ 827392/3047424 (27%)] Data (t): 0.603 Batch (t): 4.844, 1681.91/s, 420.478/s/gpu LR: 0.000808 Logit Scale: 44.516 Imm_image: 20.777 (20.811) Imm_text: 20.777 (20.811) Isd_image: 3.5357 (3.5501) Isd_text: 3.5357 (3.5501) Contrastive_loss: 2.2920 (2.1575) Loss: 2.2920 (2.1575)
+2025-04-24,11:30:43 | INFO | Train Epoch: 12 [1646592/3047424 (54%)] Data (t): 0.606 Batch (t): 4.861, 1683.17/s, 420.792/s/gpu LR: 0.000799 Logit Scale: 44.785 Imm_image: 20.706 (20.776) Imm_text: 20.706 (20.776) Isd_image: 3.5571 (3.5525) Isd_text: 3.5571 (3.5525) Contrastive_loss: 2.3800 (2.2316) Loss: 2.3800 (2.2316)
+2025-04-24,11:38:45 | INFO | Train Epoch: 12 [2465792/3047424 (81%)] Data (t): 0.604 Batch (t): 4.824, 1689.24/s, 422.309/s/gpu LR: 0.000790 Logit Scale: 44.848 Imm_image: 20.691 (20.755) Imm_text: 20.691 (20.755) Isd_image: 3.6952 (3.5881) Isd_text: 3.6952 (3.5881) Contrastive_loss: 2.4350 (2.2825) Loss: 2.4350 (2.2825)
+2025-04-24,11:44:30 | INFO | Train Epoch: 12 [3047424/3047424 (100%)] Data (t): 0.597 Batch (t): 4.859, 1734.36/s, 433.589/s/gpu LR: 0.000784 Logit Scale: 44.917 Imm_image: 21.170 (20.838) Imm_text: 21.170 (20.838) Isd_image: 4.0376 (3.6780) Isd_text: 4.0376 (3.6780) Contrastive_loss: 2.1143 (2.2488) Loss: 2.1143 (2.2488)
+2025-04-24,11:44:32 | INFO | Start epoch 13
+2025-04-24,11:44:47 | INFO | Train Epoch: 13 [   8192/3047424 (0%)] Data (t): 10.435 Batch (t): 14.817, 552.867/s, 138.217/s/gpu LR: 0.000784 Logit Scale: 44.924 Imm_image: 21.533 (21.533) Imm_text: 21.533 (21.533) Isd_image: 3.7475 (3.7475) Isd_text: 3.7475 (3.7475) Contrastive_loss: 1.8860 (1.8860) Loss: 1.8860 (1.8860)
+2025-04-24,11:52:54 | INFO | Train Epoch: 13 [ 827392/3047424 (27%)] Data (t): 0.620 Batch (t): 4.873, 1677.39/s, 419.348/s/gpu LR: 0.000775 Logit Scale: 45.843 Imm_image: 21.329 (21.431) Imm_text: 21.329 (21.431) Isd_image: 3.2676 (3.5075) Isd_text: 3.2676 (3.5075) Contrastive_loss: 2.1339 (2.0100) Loss: 2.1339 (2.0100)
+2025-04-24,12:01:00 | INFO | Train Epoch: 13 [1646592/3047424 (54%)] Data (t): 0.614 Batch (t): 4.861, 1680.33/s, 420.082/s/gpu LR: 0.000766 Logit Scale: 46.198 Imm_image: 21.390 (21.417) Imm_text: 21.390 (21.417) Isd_image: 3.7457 (3.5869) Isd_text: 3.7457 (3.5869) Contrastive_loss: 2.2104 (2.0768) Loss: 2.2104 (2.0768)
+2025-04-24,12:09:08 | INFO | Train Epoch: 13 [2465792/3047424 (81%)] Data (t): 0.637 Batch (t): 4.881, 1639.17/s, 409.791/s/gpu LR: 0.000756 Logit Scale: 46.175 Imm_image: 21.364 (21.404) Imm_text: 21.364 (21.404) Isd_image: 4.2811 (3.7605) Isd_text: 4.2811 (3.7605) Contrastive_loss: 2.2621 (2.1231) Loss: 2.2621 (2.1231)
+2025-04-24,12:14:51 | INFO | Train Epoch: 13 [3047424/3047424 (100%)] Data (t): 0.618 Batch (t): 4.830, 1732.40/s, 433.100/s/gpu LR: 0.000750 Logit Scale: 46.286 Imm_image: 21.843 (21.492) Imm_text: 21.843 (21.492) Isd_image: 4.3713 (3.8826) Isd_text: 4.3713 (3.8826) Contrastive_loss: 1.9808 (2.0946) Loss: 1.9808 (2.0946)
+2025-04-24,12:14:53 | INFO | Start epoch 14
+2025-04-24,12:15:08 | INFO | Train Epoch: 14 [   8192/3047424 (0%)] Data (t): 10.083 Batch (t): 14.487, 565.461/s, 141.365/s/gpu LR: 0.000750 Logit Scale: 46.292 Imm_image: 22.244 (22.244) Imm_text: 22.244 (22.244) Isd_image: 4.1004 (4.1004) Isd_text: 4.1004 (4.1004) Contrastive_loss: 1.7465 (1.7465) Loss: 1.7465 (1.7465)
+2025-04-24,12:23:15 | INFO | Train Epoch: 14 [ 827392/3047424 (27%)] Data (t): 0.632 Batch (t): 4.870, 1694.15/s, 423.536/s/gpu LR: 0.000740 Logit Scale: 47.167 Imm_image: 22.137 (22.190) Imm_text: 22.137 (22.190) Isd_image: 3.4623 (3.7813) Isd_text: 3.4623 (3.7813) Contrastive_loss: 1.9577 (1.8521) Loss: 1.9577 (1.8521)
+2025-04-24,12:31:14 | INFO | Train Epoch: 14 [1646592/3047424 (54%)] Data (t): 0.583 Batch (t): 4.789, 1719.38/s, 429.846/s/gpu LR: 0.000731 Logit Scale: 47.467 Imm_image: 22.096 (22.159) Imm_text: 22.096 (22.159) Isd_image: 3.6631 (3.7419) Isd_text: 3.6631 (3.7419) Contrastive_loss: 2.0578 (1.9207) Loss: 2.0578 (1.9207)
+2025-04-24,12:39:15 | INFO | Train Epoch: 14 [2465792/3047424 (81%)] Data (t): 0.604 Batch (t): 4.810, 1693.91/s, 423.477/s/gpu LR: 0.000721 Logit Scale: 47.458 Imm_image: 22.128 (22.151) Imm_text: 22.128 (22.151) Isd_image: 3.9999 (3.8064) Isd_text: 3.9999 (3.8064) Contrastive_loss: 2.0897 (1.9629) Loss: 2.0897 (1.9629)
+2025-04-24,12:44:58 | INFO | Train Epoch: 14 [3047424/3047424 (100%)] Data (t): 0.606 Batch (t): 4.837, 1743.30/s, 435.825/s/gpu LR: 0.000714 Logit Scale: 47.567 Imm_image: 22.639 (22.249) Imm_text: 22.639 (22.249) Isd_image: 4.0761 (3.8604) Isd_text: 4.0761 (3.8604) Contrastive_loss: 1.7311 (1.9166) Loss: 1.7311 (1.9166)
+2025-04-24,12:45:00 | INFO | Start epoch 15
+2025-04-24,12:45:14 | INFO | Train Epoch: 15 [   8192/3047424 (0%)] Data (t): 9.760 Batch (t): 14.087, 581.533/s, 145.383/s/gpu LR: 0.000714 Logit Scale: 47.574 Imm_image: 22.839 (22.839) Imm_text: 22.839 (22.839) Isd_image: 3.9018 (3.9018) Isd_text: 3.9018 (3.9018) Contrastive_loss: 1.6733 (1.6733) Loss: 1.6733 (1.6733)
+2025-04-24,12:53:19 | INFO | Train Epoch: 15 [ 827392/3047424 (27%)] Data (t): 0.628 Batch (t): 4.843, 1711.47/s, 427.867/s/gpu LR: 0.000704 Logit Scale: 48.482 Imm_image: 22.874 (22.857) Imm_text: 22.874 (22.857) Isd_image: 3.4022 (3.6520) Isd_text: 3.4022 (3.6520) Contrastive_loss: 1.8150 (1.7442) Loss: 1.8150 (1.7442)
+2025-04-24,13:01:21 | INFO | Train Epoch: 15 [1646592/3047424 (54%)] Data (t): 0.602 Batch (t): 4.826, 1703.33/s, 425.833/s/gpu LR: 0.000694 Logit Scale: 48.786 Imm_image: 22.871 (22.861) Imm_text: 22.871 (22.861) Isd_image: 3.9414 (3.7485) Isd_text: 3.9414 (3.7485) Contrastive_loss: 1.8890 (1.7924) Loss: 1.8890 (1.7924)
+2025-04-24,13:09:24 | INFO | Train Epoch: 15 [2465792/3047424 (81%)] Data (t): 0.606 Batch (t): 4.831, 1713.48/s, 428.370/s/gpu LR: 0.000684 Logit Scale: 48.805 Imm_image: 22.667 (22.813) Imm_text: 22.667 (22.813) Isd_image: 4.0410 (3.8216) Isd_text: 4.0410 (3.8216) Contrastive_loss: 1.9876 (1.8412) Loss: 1.9876 (1.8412)
+2025-04-24,13:15:05 | INFO | Train Epoch: 15 [3047424/3047424 (100%)] Data (t): 0.592 Batch (t): 4.792, 1732.51/s, 433.129/s/gpu LR: 0.000677 Logit Scale: 48.996 Imm_image: 23.216 (22.893) Imm_text: 23.216 (22.893) Isd_image: 4.2007 (3.8974) Isd_text: 4.2007 (3.8974) Contrastive_loss: 1.7208 (1.8172) Loss: 1.7208 (1.8172)
+2025-04-24,13:15:07 | INFO | Start epoch 16
+2025-04-24,13:15:21 | INFO | Train Epoch: 16 [   8192/3047424 (0%)] Data (t): 9.793 Batch (t): 14.156, 578.702/s, 144.675/s/gpu LR: 0.000677 Logit Scale: 48.999 Imm_image: 23.653 (23.653) Imm_text: 23.653 (23.653) Isd_image: 3.9960 (3.9960) Isd_text: 3.9960 (3.9960) Contrastive_loss: 1.4838 (1.4838) Loss: 1.4838 (1.4838)
+2025-04-24,13:23:26 | INFO | Train Epoch: 16 [ 827392/3047424 (27%)] Data (t): 0.624 Batch (t): 4.855, 1706.57/s, 426.643/s/gpu LR: 0.000667 Logit Scale: 49.989 Imm_image: 23.661 (23.657) Imm_text: 23.661 (23.657) Isd_image: 3.3644 (3.6802) Isd_text: 3.3644 (3.6802) Contrastive_loss: 1.6505 (1.5672) Loss: 1.6505 (1.5672)
+2025-04-24,13:31:31 | INFO | Train Epoch: 16 [1646592/3047424 (54%)] Data (t): 0.617 Batch (t): 4.846, 1708.44/s, 427.110/s/gpu LR: 0.000657 Logit Scale: 50.189 Imm_image: 23.596 (23.637) Imm_text: 23.596 (23.637) Isd_image: 4.2061 (3.8555) Isd_text: 4.2061 (3.8555) Contrastive_loss: 1.7709 (1.6351) Loss: 1.7709 (1.6351)
+2025-04-24,13:39:36 | INFO | Train Epoch: 16 [2465792/3047424 (81%)] Data (t): 0.625 Batch (t): 4.857, 1697.65/s, 424.413/s/gpu LR: 0.000646 Logit Scale: 50.212 Imm_image: 23.500 (23.603) Imm_text: 23.500 (23.603) Isd_image: 4.1922 (3.9397) Isd_text: 4.1922 (3.9397) Contrastive_loss: 1.8253 (1.6826) Loss: 1.8253 (1.6826)
+2025-04-24,13:45:21 | INFO | Train Epoch: 16 [3047424/3047424 (100%)] Data (t): 0.589 Batch (t): 4.858, 1737.04/s, 434.259/s/gpu LR: 0.000639 Logit Scale: 50.350 Imm_image: 24.069 (23.696) Imm_text: 24.069 (23.696) Isd_image: 4.3202 (4.0158) Isd_text: 4.3202 (4.0158) Contrastive_loss: 1.5298 (1.6521) Loss: 1.5298 (1.6521)
+2025-04-24,13:45:23 | INFO | Start epoch 17
+2025-04-24,13:45:38 | INFO | Train Epoch: 17 [   8192/3047424 (0%)] Data (t): 10.101 Batch (t): 14.447, 567.043/s, 141.761/s/gpu LR: 0.000639 Logit Scale: 50.353 Imm_image: 24.385 (24.385) Imm_text: 24.385 (24.385) Isd_image: 4.1519 (4.1519) Isd_text: 4.1519 (4.1519) Contrastive_loss: 1.3573 (1.3573) Loss: 1.3573 (1.3573)
+2025-04-24,13:53:43 | INFO | Train Epoch: 17 [ 827392/3047424 (27%)] Data (t): 0.622 Batch (t): 4.850, 1689.80/s, 422.451/s/gpu LR: 0.000628 Logit Scale: 51.392 Imm_image: 24.321 (24.353) Imm_text: 24.321 (24.353) Isd_image: 3.3838 (3.7679) Isd_text: 3.3838 (3.7679) Contrastive_loss: 1.4877 (1.4225) Loss: 1.4877 (1.4225)
+2025-04-24,14:01:48 | INFO | Train Epoch: 17 [1646592/3047424 (54%)] Data (t): 0.617 Batch (t): 4.852, 1678.44/s, 419.610/s/gpu LR: 0.000618 Logit Scale: 51.617 Imm_image: 24.236 (24.314) Imm_text: 24.236 (24.314) Isd_image: 3.6924 (3.7427) Isd_text: 3.6924 (3.7427) Contrastive_loss: 1.6406 (1.4952) Loss: 1.6406 (1.4952)
+2025-04-24,14:09:54 | INFO | Train Epoch: 17 [2465792/3047424 (81%)] Data (t): 0.639 Batch (t): 4.863, 1713.66/s, 428.416/s/gpu LR: 0.000607 Logit Scale: 51.594 Imm_image: 24.106 (24.262) Imm_text: 24.106 (24.262) Isd_image: 4.0283 (3.8141) Isd_text: 4.0283 (3.8141) Contrastive_loss: 1.6918 (1.5443) Loss: 1.6918 (1.5443)
+2025-04-24,14:15:38 | INFO | Train Epoch: 17 [3047424/3047424 (100%)] Data (t): 0.607 Batch (t): 4.839, 1733.70/s, 433.425/s/gpu LR: 0.000600 Logit Scale: 51.731 Imm_image: 24.823 (24.374) Imm_text: 24.823 (24.374) Isd_image: 4.2615 (3.9036) Isd_text: 4.2615 (3.9036) Contrastive_loss: 1.3592 (1.5073) Loss: 1.3592 (1.5073)
+2025-04-24,14:15:40 | INFO | Start epoch 18
+2025-04-24,14:15:55 | INFO | Train Epoch: 18 [   8192/3047424 (0%)] Data (t): 10.183 Batch (t): 14.446, 567.067/s, 141.767/s/gpu LR: 0.000600 Logit Scale: 51.734 Imm_image: 25.200 (25.200) Imm_text: 25.200 (25.200) Isd_image: 4.1812 (4.1812) Isd_text: 4.1812 (4.1812) Contrastive_loss: 1.1877 (1.1877) Loss: 1.1877 (1.1877)
+2025-04-24,14:24:01 | INFO | Train Epoch: 18 [ 827392/3047424 (27%)] Data (t): 0.617 Batch (t): 4.862, 1706.89/s, 426.722/s/gpu LR: 0.000589 Logit Scale: 52.754 Imm_image: 25.255 (25.228) Imm_text: 25.255 (25.228) Isd_image: 3.3625 (3.7718) Isd_text: 3.3625 (3.7718) Contrastive_loss: 1.3529 (1.2703) Loss: 1.3529 (1.2703)
+2025-04-24,14:32:07 | INFO | Train Epoch: 18 [1646592/3047424 (54%)] Data (t): 0.602 Batch (t): 4.856, 1687.03/s, 421.757/s/gpu LR: 0.000578 Logit Scale: 53.049 Imm_image: 25.020 (25.159) Imm_text: 25.020 (25.159) Isd_image: 3.9419 (3.8285) Isd_text: 3.9419 (3.8285) Contrastive_loss: 1.4816 (1.3407) Loss: 1.4816 (1.3407)
+2025-04-24,14:40:11 | INFO | Train Epoch: 18 [2465792/3047424 (81%)] Data (t): 0.593 Batch (t): 4.844, 1714.99/s, 428.748/s/gpu LR: 0.000568 Logit Scale: 53.007 Imm_image: 24.896 (25.093) Imm_text: 24.896 (25.093) Isd_image: 4.2207 (3.9266) Isd_text: 4.2207 (3.9266) Contrastive_loss: 1.5583 (1.3951) Loss: 1.5583 (1.3951)
+2025-04-24,14:45:55 | INFO | Train Epoch: 18 [3047424/3047424 (100%)] Data (t): 0.586 Batch (t): 4.844, 1733.78/s, 433.445/s/gpu LR: 0.000560 Logit Scale: 53.176 Imm_image: 25.472 (25.169) Imm_text: 25.472 (25.169) Isd_image: 4.2305 (3.9873) Isd_text: 4.2305 (3.9873) Contrastive_loss: 1.2939 (1.3749) Loss: 1.2939 (1.3749)
+2025-04-24,14:45:57 | INFO | Start epoch 19
+2025-04-24,14:46:12 | INFO | Train Epoch: 19 [   8192/3047424 (0%)] Data (t): 9.889 Batch (t): 15.274, 536.325/s, 134.081/s/gpu LR: 0.000560 Logit Scale: 53.180 Imm_image: 25.925 (25.925) Imm_text: 25.925 (25.925) Isd_image: 3.8176 (3.8176) Isd_text: 3.8176 (3.8176) Contrastive_loss: 1.0655 (1.0655) Loss: 1.0655 (1.0655)
+2025-04-24,14:54:18 | INFO | Train Epoch: 19 [ 827392/3047424 (27%)] Data (t): 0.625 Batch (t): 4.856, 1677.38/s, 419.346/s/gpu LR: 0.000549 Logit Scale: 54.292 Imm_image: 25.873 (25.899) Imm_text: 25.873 (25.899) Isd_image: 3.4075 (3.6126) Isd_text: 3.4075 (3.6126) Contrastive_loss: 1.2442 (1.1548) Loss: 1.2442 (1.1548)
+2025-04-24,15:02:23 | INFO | Train Epoch: 19 [1646592/3047424 (54%)] Data (t): 0.607 Batch (t): 4.848, 1706.70/s, 426.676/s/gpu LR: 0.000538 Logit Scale: 54.498 Imm_image: 25.949 (25.916) Imm_text: 25.949 (25.916) Isd_image: 4.1048 (3.7766) Isd_text: 4.1048 (3.7766) Contrastive_loss: 1.3645 (1.2247) Loss: 1.3645 (1.2247)
+2025-04-24,15:10:28 | INFO | Train Epoch: 19 [2465792/3047424 (81%)] Data (t): 0.628 Batch (t): 4.853, 1714.10/s, 428.525/s/gpu LR: 0.000528 Logit Scale: 54.666 Imm_image: 25.855 (25.900) Imm_text: 25.855 (25.900) Isd_image: 4.2441 (3.8935) Isd_text: 4.2441 (3.8935) Contrastive_loss: 1.3916 (1.2664) Loss: 1.3916 (1.2664)
+2025-04-24,15:16:11 | INFO | Train Epoch: 19 [3047424/3047424 (100%)] Data (t): 0.600 Batch (t): 4.827, 1734.36/s, 433.590/s/gpu LR: 0.000520 Logit Scale: 54.814 Imm_image: 26.478 (26.016) Imm_text: 26.478 (26.016) Isd_image: 4.2114 (3.9571) Isd_text: 4.2114 (3.9571) Contrastive_loss: 1.1059 (1.2343) Loss: 1.1059 (1.2343)
+2025-04-24,15:16:13 | INFO | Start epoch 20
+2025-04-24,15:16:27 | INFO | Train Epoch: 20 [   8192/3047424 (0%)] Data (t): 9.919 Batch (t): 14.324, 571.909/s, 142.977/s/gpu LR: 0.000520 Logit Scale: 54.819 Imm_image: 26.748 (26.748) Imm_text: 26.748 (26.748) Isd_image: 3.9890 (3.9890) Isd_text: 3.9890 (3.9890) Contrastive_loss: 1.0187 (1.0187) Loss: 1.0187 (1.0187)
+2025-04-24,15:24:35 | INFO | Train Epoch: 20 [ 827392/3047424 (27%)] Data (t): 0.632 Batch (t): 4.877, 1661.56/s, 415.391/s/gpu LR: 0.000509 Logit Scale: 55.963 Imm_image: 26.886 (26.817) Imm_text: 26.886 (26.817) Isd_image: 3.5723 (3.7807) Isd_text: 3.5723 (3.7807) Contrastive_loss: 1.1027 (1.0607) Loss: 1.1027 (1.0607)
+2025-04-24,15:32:40 | INFO | Train Epoch: 20 [1646592/3047424 (54%)] Data (t): 0.627 Batch (t): 4.859, 1697.00/s, 424.249/s/gpu LR: 0.000498 Logit Scale: 56.268 Imm_image: 26.772 (26.802) Imm_text: 26.772 (26.802) Isd_image: 3.8595 (3.8069) Isd_text: 3.8595 (3.8069) Contrastive_loss: 1.1891 (1.1035) Loss: 1.1891 (1.1035)
+2025-04-24,15:40:45 | INFO | Train Epoch: 20 [2465792/3047424 (81%)] Data (t): 0.590 Batch (t): 4.848, 1714.84/s, 428.710/s/gpu LR: 0.000487 Logit Scale: 56.273 Imm_image: 26.668 (26.768) Imm_text: 26.668 (26.768) Isd_image: 4.2049 (3.9064) Isd_text: 4.2049 (3.9064) Contrastive_loss: 1.2775 (1.1470) Loss: 1.2775 (1.1470)
+2025-04-24,15:46:28 | INFO | Train Epoch: 20 [3047424/3047424 (100%)] Data (t): 0.610 Batch (t): 4.823, 1736.29/s, 434.073/s/gpu LR: 0.000480 Logit Scale: 56.520 Imm_image: 27.326 (26.880) Imm_text: 27.326 (26.880) Isd_image: 4.4613 (4.0174) Isd_text: 4.4613 (4.0174) Contrastive_loss: 1.0086 (1.1193) Loss: 1.0086 (1.1193)
+2025-04-24,15:46:30 | INFO | Start epoch 21
+2025-04-24,15:46:44 | INFO | Train Epoch: 21 [   8192/3047424 (0%)] Data (t): 10.267 Batch (t): 14.714, 556.763/s, 139.191/s/gpu LR: 0.000480 Logit Scale: 56.524 Imm_image: 27.691 (27.691) Imm_text: 27.691 (27.691) Isd_image: 4.0270 (4.0270) Isd_text: 4.0270 (4.0270) Contrastive_loss: 0.83266 (0.83266) Loss: 0.83266 (0.83266)
+2025-04-24,15:54:51 | INFO | Train Epoch: 21 [ 827392/3047424 (27%)] Data (t): 0.623 Batch (t): 4.868, 1682.14/s, 420.535/s/gpu LR: 0.000469 Logit Scale: 57.717 Imm_image: 27.801 (27.746) Imm_text: 27.801 (27.746) Isd_image: 3.4194 (3.7232) Isd_text: 3.4194 (3.7232) Contrastive_loss: 1.0257 (0.92917) Loss: 1.0257 (0.92917)
+2025-04-24,16:02:57 | INFO | Train Epoch: 21 [1646592/3047424 (54%)] Data (t): 0.624 Batch (t): 4.855, 1703.23/s, 425.808/s/gpu LR: 0.000458 Logit Scale: 58.007 Imm_image: 27.684 (27.725) Imm_text: 27.684 (27.725) Isd_image: 4.2149 (3.8871) Isd_text: 4.2149 (3.8871) Contrastive_loss: 1.1036 (0.98732) Loss: 1.1036 (0.98732)
+2025-04-24,16:11:03 | INFO | Train Epoch: 21 [2465792/3047424 (81%)] Data (t): 0.609 Batch (t): 4.861, 1713.86/s, 428.464/s/gpu LR: 0.000447 Logit Scale: 58.004 Imm_image: 27.575 (27.688) Imm_text: 27.575 (27.688) Isd_image: 4.2724 (3.9834) Isd_text: 4.2724 (3.9834) Contrastive_loss: 1.2097 (1.0429) Loss: 1.2097 (1.0429)
+2025-04-24,16:16:47 | INFO | Train Epoch: 21 [3047424/3047424 (100%)] Data (t): 0.609 Batch (t): 4.846, 1735.70/s, 433.926/s/gpu LR: 0.000440 Logit Scale: 58.241 Imm_image: 28.328 (27.816) Imm_text: 28.328 (27.816) Isd_image: 4.2457 (4.0359) Isd_text: 4.2457 (4.0359) Contrastive_loss: 0.86375 (1.0071) Loss: 0.86375 (1.0071)
+2025-04-24,16:16:49 | INFO | Start epoch 22
+2025-04-24,16:17:04 | INFO | Train Epoch: 22 [   8192/3047424 (0%)] Data (t): 10.702 Batch (t): 15.128, 541.497/s, 135.374/s/gpu LR: 0.000440 Logit Scale: 58.247 Imm_image: 28.675 (28.675) Imm_text: 28.675 (28.675) Isd_image: 4.0029 (4.0029) Isd_text: 4.0029 (4.0029) Contrastive_loss: 0.72551 (0.72551) Loss: 0.72551 (0.72551)
+2025-04-24,16:25:09 | INFO | Train Epoch: 22 [ 827392/3047424 (27%)] Data (t): 0.614 Batch (t): 4.847, 1703.32/s, 425.831/s/gpu LR: 0.000429 Logit Scale: 59.525 Imm_image: 28.927 (28.801) Imm_text: 28.927 (28.801) Isd_image: 3.6898 (3.8463) Isd_text: 3.6898 (3.8463) Contrastive_loss: 0.87334 (0.79942) Loss: 0.87334 (0.79942)
+2025-04-24,16:33:14 | INFO | Train Epoch: 22 [1646592/3047424 (54%)] Data (t): 0.615 Batch (t): 4.858, 1696.91/s, 424.229/s/gpu LR: 0.000418 Logit Scale: 59.758 Imm_image: 28.669 (28.757) Imm_text: 28.669 (28.757) Isd_image: 4.0972 (3.9300) Isd_text: 4.0972 (3.9300) Contrastive_loss: 0.99460 (0.86448) Loss: 0.99460 (0.86448)
+2025-04-24,16:41:19 | INFO | Train Epoch: 22 [2465792/3047424 (81%)] Data (t): 0.613 Batch (t): 4.848, 1715.96/s, 428.991/s/gpu LR: 0.000407 Logit Scale: 59.837 Imm_image: 28.566 (28.709) Imm_text: 28.566 (28.709) Isd_image: 4.3656 (4.0389) Isd_text: 4.3656 (4.0389) Contrastive_loss: 1.0576 (0.91277) Loss: 1.0576 (0.91277)
+2025-04-24,16:47:05 | INFO | Train Epoch: 22 [3047424/3047424 (100%)] Data (t): 0.625 Batch (t): 4.876, 1721.65/s, 430.414/s/gpu LR: 0.000400 Logit Scale: 60.041 Imm_image: 29.388 (28.845) Imm_text: 29.388 (28.845) Isd_image: 4.0868 (4.0485) Isd_text: 4.0868 (4.0485) Contrastive_loss: 0.73261 (0.87674) Loss: 0.73261 (0.87674)
+2025-04-24,16:47:08 | INFO | Start epoch 23
+2025-04-24,16:47:23 | INFO | Train Epoch: 23 [   8192/3047424 (0%)] Data (t): 11.039 Batch (t): 15.433, 530.821/s, 132.705/s/gpu LR: 0.000400 Logit Scale: 60.047 Imm_image: 29.624 (29.624) Imm_text: 29.624 (29.624) Isd_image: 3.8282 (3.8282) Isd_text: 3.8282 (3.8282) Contrastive_loss: 0.67476 (0.67476) Loss: 0.67476 (0.67476)
+2025-04-24,16:55:37 | INFO | Train Epoch: 23 [ 827392/3047424 (27%)] Data (t): 0.670 Batch (t): 4.938, 1659.79/s, 414.948/s/gpu LR: 0.000389 Logit Scale: 61.257 Imm_image: 29.723 (29.673) Imm_text: 29.723 (29.673) Isd_image: 3.5416 (3.6849) Isd_text: 3.5416 (3.6849) Contrastive_loss: 0.78209 (0.72842) Loss: 0.78209 (0.72842)
+2025-04-24,17:03:47 | INFO | Train Epoch: 23 [1646592/3047424 (54%)] Data (t): 0.663 Batch (t): 4.907, 1671.15/s, 417.787/s/gpu LR: 0.000379 Logit Scale: 61.584 Imm_image: 29.741 (29.696) Imm_text: 29.741 (29.696) Isd_image: 4.3013 (3.8904) Isd_text: 4.3013 (3.8904) Contrastive_loss: 0.87943 (0.77876) Loss: 0.87943 (0.77876)
+2025-04-24,17:11:58 | INFO | Train Epoch: 23 [2465792/3047424 (81%)] Data (t): 0.661 Batch (t): 4.903, 1698.45/s, 424.611/s/gpu LR: 0.000368 Logit Scale: 61.747 Imm_image: 29.756 (29.711) Imm_text: 29.756 (29.711) Isd_image: 4.2756 (3.9867) Isd_text: 4.2756 (3.9867) Contrastive_loss: 0.91958 (0.81396) Loss: 0.91958 (0.81396)
+2025-04-24,17:17:45 | INFO | Train Epoch: 23 [3047424/3047424 (100%)] Data (t): 0.663 Batch (t): 4.890, 1721.41/s, 430.352/s/gpu LR: 0.000361 Logit Scale: 62.020 Imm_image: 30.337 (29.836) Imm_text: 30.337 (29.836) Isd_image: 4.3002 (4.0494) Isd_text: 4.3002 (4.0494) Contrastive_loss: 0.71312 (0.79379) Loss: 0.71312 (0.79379)
+2025-04-24,17:17:47 | INFO | Start epoch 24
+2025-04-24,17:18:02 | INFO | Train Epoch: 24 [   8192/3047424 (0%)] Data (t): 10.528 Batch (t): 14.953, 547.853/s, 136.963/s/gpu LR: 0.000361 Logit Scale: 62.021 Imm_image: 30.681 (30.681) Imm_text: 30.681 (30.681) Isd_image: 4.3031 (4.3031) Isd_text: 4.3031 (4.3031) Contrastive_loss: 0.59845 (0.59845) Loss: 0.59845 (0.59845)
+2025-04-24,17:26:16 | INFO | Train Epoch: 24 [ 827392/3047424 (27%)] Data (t): 0.686 Batch (t): 4.938, 1645.28/s, 411.321/s/gpu LR: 0.000350 Logit Scale: 63.187 Imm_image: 30.967 (30.824) Imm_text: 30.967 (30.824) Isd_image: 3.6465 (3.9748) Isd_text: 3.6465 (3.9748) Contrastive_loss: 0.69055 (0.64450) Loss: 0.69055 (0.64450)
+2025-04-24,17:34:26 | INFO | Train Epoch: 24 [1646592/3047424 (54%)] Data (t): 0.659 Batch (t): 4.902, 1678.38/s, 419.594/s/gpu LR: 0.000340 Logit Scale: 63.591 Imm_image: 30.924 (30.857) Imm_text: 30.924 (30.857) Isd_image: 4.3112 (4.0870) Isd_text: 4.3112 (4.0870) Contrastive_loss: 0.74675 (0.67858) Loss: 0.74675 (0.67858)
+2025-04-24,17:42:36 | INFO | Train Epoch: 24 [2465792/3047424 (81%)] Data (t): 0.661 Batch (t): 4.898, 1686.58/s, 421.644/s/gpu LR: 0.000330 Logit Scale: 63.837 Imm_image: 30.857 (30.857) Imm_text: 30.857 (30.857) Isd_image: 4.3047 (4.1414) Isd_text: 4.3047 (4.1414) Contrastive_loss: 0.78070 (0.70411) Loss: 0.78070 (0.70411)
+2025-04-24,17:48:22 | INFO | Train Epoch: 24 [3047424/3047424 (100%)] Data (t): 0.649 Batch (t): 4.869, 1721.05/s, 430.263/s/gpu LR: 0.000323 Logit Scale: 64.102 Imm_image: 31.553 (30.996) Imm_text: 31.553 (30.996) Isd_image: 4.4981 (4.2127) Isd_text: 4.4981 (4.2127) Contrastive_loss: 0.60120 (0.68353) Loss: 0.60120 (0.68353)
+2025-04-24,17:48:24 | INFO | Start epoch 25
+2025-04-24,17:48:39 | INFO | Train Epoch: 25 [   8192/3047424 (0%)] Data (t): 10.523 Batch (t): 14.979, 546.910/s, 136.727/s/gpu LR: 0.000323 Logit Scale: 64.105 Imm_image: 31.784 (31.784) Imm_text: 31.784 (31.784) Isd_image: 4.5983 (4.5983) Isd_text: 4.5983 (4.5983) Contrastive_loss: 0.52684 (0.52684) Loss: 0.52684 (0.52684)
+2025-04-24,17:56:48 | INFO | Train Epoch: 25 [ 827392/3047424 (27%)] Data (t): 0.660 Batch (t): 4.892, 1681.78/s, 420.445/s/gpu LR: 0.000312 Logit Scale: 65.212 Imm_image: 32.062 (31.923) Imm_text: 32.062 (31.923) Isd_image: 4.3580 (4.4782) Isd_text: 4.3580 (4.4782) Contrastive_loss: 0.59194 (0.55939) Loss: 0.59194 (0.55939)
+2025-04-24,18:04:57 | INFO | Train Epoch: 25 [1646592/3047424 (54%)] Data (t): 0.663 Batch (t): 4.888, 1673.50/s, 418.375/s/gpu LR: 0.000302 Logit Scale: 65.661 Imm_image: 32.079 (31.975) Imm_text: 32.079 (31.975) Isd_image: 4.5071 (4.4878) Isd_text: 4.5071 (4.4878) Contrastive_loss: 0.66737 (0.59538) Loss: 0.66737 (0.59538)
+2025-04-24,18:13:07 | INFO | Train Epoch: 25 [2465792/3047424 (81%)] Data (t): 0.665 Batch (t): 4.904, 1681.11/s, 420.278/s/gpu LR: 0.000293 Logit Scale: 65.872 Imm_image: 32.165 (32.023) Imm_text: 32.165 (32.023) Isd_image: 4.5679 (4.5078) Isd_text: 4.5679 (4.5078) Contrastive_loss: 0.67597 (0.61553) Loss: 0.67597 (0.61553)
+2025-04-24,18:18:55 | INFO | Train Epoch: 25 [3047424/3047424 (100%)] Data (t): 0.650 Batch (t): 4.897, 1713.27/s, 428.317/s/gpu LR: 0.000286 Logit Scale: 66.158 Imm_image: 32.943 (32.207) Imm_text: 32.943 (32.207) Isd_image: 4.5977 (4.5258) Isd_text: 4.5977 (4.5258) Contrastive_loss: 0.50315 (0.59305) Loss: 0.50315 (0.59305)
+2025-04-24,18:18:57 | INFO | Start epoch 26
+2025-04-24,18:19:13 | INFO | Train Epoch: 26 [   8192/3047424 (0%)] Data (t): 11.009 Batch (t): 15.437, 530.685/s, 132.671/s/gpu LR: 0.000286 Logit Scale: 66.162 Imm_image: 33.132 (33.132) Imm_text: 33.132 (33.132) Isd_image: 4.6732 (4.6732) Isd_text: 4.6732 (4.6732) Contrastive_loss: 0.49016 (0.49016) Loss: 0.49016 (0.49016)
+2025-04-24,18:27:27 | INFO | Train Epoch: 26 [ 827392/3047424 (27%)] Data (t): 0.678 Batch (t): 4.939, 1686.52/s, 421.630/s/gpu LR: 0.000276 Logit Scale: 67.246 Imm_image: 33.370 (33.251) Imm_text: 33.370 (33.251) Isd_image: 4.4638 (4.5685) Isd_text: 4.4638 (4.5685) Contrastive_loss: 0.51359 (0.50188) Loss: 0.51359 (0.50188)
+2025-04-24,18:35:35 | INFO | Train Epoch: 26 [1646592/3047424 (54%)] Data (t): 0.652 Batch (t): 4.889, 1669.70/s, 417.426/s/gpu LR: 0.000266 Logit Scale: 67.747 Imm_image: 33.339 (33.280) Imm_text: 33.339 (33.280) Isd_image: 4.8386 (4.6585) Isd_text: 4.8386 (4.6585) Contrastive_loss: 0.56741 (0.52372) Loss: 0.56741 (0.52372)
+2025-04-24,18:43:48 | INFO | Train Epoch: 26 [2465792/3047424 (81%)] Data (t): 0.671 Batch (t): 4.922, 1693.36/s, 423.341/s/gpu LR: 0.000257 Logit Scale: 68.040 Imm_image: 33.414 (33.314) Imm_text: 33.414 (33.314) Isd_image: 4.8976 (4.7183) Isd_text: 4.8976 (4.7183) Contrastive_loss: 0.56654 (0.53443) Loss: 0.56654 (0.53443)
+2025-04-24,18:49:35 | INFO | Train Epoch: 26 [3047424/3047424 (100%)] Data (t): 0.632 Batch (t): 4.892, 1722.62/s, 430.656/s/gpu LR: 0.000250 Logit Scale: 68.473 Imm_image: 34.133 (33.478) Imm_text: 34.133 (33.478) Isd_image: 4.8653 (4.7477) Isd_text: 4.8653 (4.7477) Contrastive_loss: 0.45428 (0.51840) Loss: 0.45428 (0.51840)
+2025-04-24,18:49:37 | INFO | Start epoch 27
+2025-04-24,18:49:53 | INFO | Train Epoch: 27 [   8192/3047424 (0%)] Data (t): 11.285 Batch (t): 15.638, 523.859/s, 130.965/s/gpu LR: 0.000250 Logit Scale: 68.477 Imm_image: 34.509 (34.509) Imm_text: 34.509 (34.509) Isd_image: 4.7748 (4.7748) Isd_text: 4.7748 (4.7748) Contrastive_loss: 0.39009 (0.39009) Loss: 0.39009 (0.39009)
+2025-04-24,18:58:06 | INFO | Train Epoch: 27 [ 827392/3047424 (27%)] Data (t): 0.691 Batch (t): 4.933, 1631.51/s, 407.878/s/gpu LR: 0.000241 Logit Scale: 69.411 Imm_image: 34.540 (34.524) Imm_text: 34.540 (34.524) Isd_image: 4.7571 (4.7660) Isd_text: 4.7571 (4.7660) Contrastive_loss: 0.45757 (0.42383) Loss: 0.45757 (0.42383)
+2025-04-24,19:06:17 | INFO | Train Epoch: 27 [1646592/3047424 (54%)] Data (t): 0.664 Batch (t): 4.909, 1677.44/s, 419.361/s/gpu LR: 0.000231 Logit Scale: 69.946 Imm_image: 34.659 (34.569) Imm_text: 34.659 (34.569) Isd_image: 5.0755 (4.8691) Isd_text: 5.0755 (4.8691) Contrastive_loss: 0.50388 (0.45051) Loss: 0.50388 (0.45051)
+2025-04-24,19:14:30 | INFO | Train Epoch: 27 [2465792/3047424 (81%)] Data (t): 0.666 Batch (t): 4.932, 1652.83/s, 413.207/s/gpu LR: 0.000222 Logit Scale: 70.327 Imm_image: 34.813 (34.630) Imm_text: 34.813 (34.630) Isd_image: 4.7184 (4.8314) Isd_text: 4.7184 (4.8314) Contrastive_loss: 0.50503 (0.46414) Loss: 0.50503 (0.46414)
+2025-04-24,19:20:18 | INFO | Train Epoch: 27 [3047424/3047424 (100%)] Data (t): 0.676 Batch (t): 4.905, 1722.88/s, 430.721/s/gpu LR: 0.000216 Logit Scale: 70.662 Imm_image: 35.478 (34.800) Imm_text: 35.478 (34.800) Isd_image: 5.1128 (4.8877) Isd_text: 5.1128 (4.8877) Contrastive_loss: 0.37966 (0.44725) Loss: 0.37966 (0.44725)
+2025-04-24,19:20:21 | INFO | Start epoch 28
+2025-04-24,19:20:35 | INFO | Train Epoch: 28 [   8192/3047424 (0%)] Data (t): 10.064 Batch (t): 14.499, 565.000/s, 141.250/s/gpu LR: 0.000216 Logit Scale: 70.668 Imm_image: 35.807 (35.807) Imm_text: 35.807 (35.807) Isd_image: 5.0841 (5.0841) Isd_text: 5.0841 (5.0841) Contrastive_loss: 0.33282 (0.33282) Loss: 0.33282 (0.33282)
+2025-04-24,19:28:48 | INFO | Train Epoch: 28 [ 827392/3047424 (27%)] Data (t): 0.671 Batch (t): 4.922, 1690.82/s, 422.704/s/gpu LR: 0.000207 Logit Scale: 71.493 Imm_image: 35.897 (35.852) Imm_text: 35.897 (35.852) Isd_image: 5.0708 (5.0774) Isd_text: 5.0708 (5.0774) Contrastive_loss: 0.42400 (0.37841) Loss: 0.42400 (0.37841)
+2025-04-24,19:36:59 | INFO | Train Epoch: 28 [1646592/3047424 (54%)] Data (t): 0.643 Batch (t): 4.914, 1663.66/s, 415.914/s/gpu LR: 0.000198 Logit Scale: 72.049 Imm_image: 35.777 (35.827) Imm_text: 35.777 (35.827) Isd_image: 5.3477 (5.1675) Isd_text: 5.3477 (5.1675) Contrastive_loss: 0.45098 (0.40260) Loss: 0.45098 (0.40260)
+2025-04-24,19:45:11 | INFO | Train Epoch: 28 [2465792/3047424 (81%)] Data (t): 0.669 Batch (t): 4.920, 1680.76/s, 420.190/s/gpu LR: 0.000190 Logit Scale: 72.450 Imm_image: 36.008 (35.872) Imm_text: 36.008 (35.872) Isd_image: 5.4943 (5.2492) Isd_text: 5.4943 (5.2492) Contrastive_loss: 0.46596 (0.41844) Loss: 0.46596 (0.41844)
+2025-04-24,19:50:57 | INFO | Train Epoch: 28 [3047424/3047424 (100%)] Data (t): 0.651 Batch (t): 4.874, 1721.76/s, 430.440/s/gpu LR: 0.000184 Logit Scale: 72.788 Imm_image: 36.747 (36.047) Imm_text: 36.747 (36.047) Isd_image: 5.5611 (5.3116) Isd_text: 5.5611 (5.3116) Contrastive_loss: 0.34426 (0.40360) Loss: 0.34426 (0.40360)
+2025-04-24,19:50:59 | INFO | Start epoch 29
+2025-04-24,19:51:15 | INFO | Train Epoch: 29 [   8192/3047424 (0%)] Data (t): 11.075 Batch (t): 15.499, 528.561/s, 132.140/s/gpu LR: 0.000184 Logit Scale: 72.795 Imm_image: 36.937 (36.937) Imm_text: 36.937 (36.937) Isd_image: 5.5310 (5.5310) Isd_text: 5.5310 (5.5310) Contrastive_loss: 0.30995 (0.30995) Loss: 0.30995 (0.30995)
+2025-04-24,19:59:25 | INFO | Train Epoch: 29 [ 827392/3047424 (27%)] Data (t): 0.659 Batch (t): 4.906, 1690.95/s, 422.737/s/gpu LR: 0.000175 Logit Scale: 73.616 Imm_image: 37.104 (37.020) Imm_text: 37.104 (37.020) Isd_image: 5.5140 (5.5225) Isd_text: 5.5140 (5.5225) Contrastive_loss: 0.33987 (0.32491) Loss: 0.33987 (0.32491)
+2025-04-24,20:07:36 | INFO | Train Epoch: 29 [1646592/3047424 (54%)] Data (t): 0.663 Batch (t): 4.909, 1595.18/s, 398.795/s/gpu LR: 0.000167 Logit Scale: 74.046 Imm_image: 37.125 (37.055) Imm_text: 37.125 (37.055) Isd_image: 5.6452 (5.5634) Isd_text: 5.6452 (5.5634) Contrastive_loss: 0.37285 (0.34089) Loss: 0.37285 (0.34089)
+2025-04-24,20:15:47 | INFO | Train Epoch: 29 [2465792/3047424 (81%)] Data (t): 0.664 Batch (t): 4.908, 1672.11/s, 418.027/s/gpu LR: 0.000159 Logit Scale: 74.464 Imm_image: 37.294 (37.115) Imm_text: 37.294 (37.115) Isd_image: 5.6046 (5.5737) Isd_text: 5.6046 (5.5737) Contrastive_loss: 0.40279 (0.35636) Loss: 0.40279 (0.35636)
+2025-04-24,20:21:36 | INFO | Train Epoch: 29 [3047424/3047424 (100%)] Data (t): 0.671 Batch (t): 4.916, 1718.00/s, 429.500/s/gpu LR: 0.000154 Logit Scale: 74.725 Imm_image: 37.907 (37.273) Imm_text: 37.907 (37.273) Isd_image: 5.6823 (5.5954) Isd_text: 5.6823 (5.5954) Contrastive_loss: 0.30236 (0.34556) Loss: 0.30236 (0.34556)
+2025-04-24,20:21:38 | INFO | Start epoch 30
+2025-04-24,20:21:53 | INFO | Train Epoch: 30 [   8192/3047424 (0%)] Data (t): 10.874 Batch (t): 15.294, 535.648/s, 133.912/s/gpu LR: 0.000154 Logit Scale: 74.730 Imm_image: 38.001 (38.001) Imm_text: 38.001 (38.001) Isd_image: 5.7280 (5.7280) Isd_text: 5.7280 (5.7280) Contrastive_loss: 0.29059 (0.29059) Loss: 0.29059 (0.29059)
+2025-04-24,20:30:04 | INFO | Train Epoch: 30 [ 827392/3047424 (27%)] Data (t): 0.674 Batch (t): 4.905, 1688.81/s, 422.203/s/gpu LR: 0.000146 Logit Scale: 75.346 Imm_image: 38.234 (38.117) Imm_text: 38.234 (38.117) Isd_image: 5.5996 (5.6638) Isd_text: 5.5996 (5.6638) Contrastive_loss: 0.32412 (0.30736) Loss: 0.32412 (0.30736)
+2025-04-24,20:38:18 | INFO | Train Epoch: 30 [1646592/3047424 (54%)] Data (t): 0.685 Batch (t): 4.943, 1678.86/s, 419.714/s/gpu LR: 0.000138 Logit Scale: 75.815 Imm_image: 38.327 (38.187) Imm_text: 38.327 (38.187) Isd_image: 5.5637 (5.6304) Isd_text: 5.5637 (5.6304) Contrastive_loss: 0.31641 (0.31037) Loss: 0.31641 (0.31037)
+2025-04-24,20:46:25 | INFO | Train Epoch: 30 [2465792/3047424 (81%)] Data (t): 0.650 Batch (t): 4.871, 1662.09/s, 415.522/s/gpu LR: 0.000131 Logit Scale: 76.202 Imm_image: 38.630 (38.298) Imm_text: 38.630 (38.298) Isd_image: 5.6330 (5.6311) Isd_text: 5.6330 (5.6311) Contrastive_loss: 0.30202 (0.30829) Loss: 0.30202 (0.30829)
+2025-04-24,20:52:13 | INFO | Train Epoch: 30 [3047424/3047424 (100%)] Data (t): 0.658 Batch (t): 4.889, 1715.96/s, 428.991/s/gpu LR: 0.000126 Logit Scale: 76.475 Imm_image: 39.034 (38.445) Imm_text: 39.034 (38.445) Isd_image: 5.6520 (5.6353) Isd_text: 5.6520 (5.6353) Contrastive_loss: 0.26744 (0.30012) Loss: 0.26744 (0.30012)
+2025-04-24,20:52:15 | INFO | Start epoch 31
+2025-04-24,20:52:30 | INFO | Train Epoch: 31 [   8192/3047424 (0%)] Data (t): 11.295 Batch (t): 15.700, 521.773/s, 130.443/s/gpu LR: 0.000126 Logit Scale: 76.477 Imm_image: 39.098 (39.098) Imm_text: 39.098 (39.098) Isd_image: 5.7717 (5.7717) Isd_text: 5.7717 (5.7717) Contrastive_loss: 0.25171 (0.25171) Loss: 0.25171 (0.25171)
+2025-04-24,21:00:43 | INFO | Train Epoch: 31 [ 827392/3047424 (27%)] Data (t): 0.679 Batch (t): 4.932, 1633.01/s, 408.252/s/gpu LR: 0.000119 Logit Scale: 77.014 Imm_image: 39.345 (39.222) Imm_text: 39.345 (39.222) Isd_image: 6.0412 (5.9064) Isd_text: 6.0412 (5.9064) Contrastive_loss: 0.26619 (0.25895) Loss: 0.26619 (0.25895)
+2025-04-24,21:08:53 | INFO | Train Epoch: 31 [1646592/3047424 (54%)] Data (t): 0.661 Batch (t): 4.895, 1682.47/s, 420.617/s/gpu LR: 0.000112 Logit Scale: 77.441 Imm_image: 39.491 (39.312) Imm_text: 39.491 (39.312) Isd_image: 5.7781 (5.8637) Isd_text: 5.7781 (5.8637) Contrastive_loss: 0.24729 (0.25506) Loss: 0.24729 (0.25506)
+2025-04-24,21:17:03 | INFO | Train Epoch: 31 [2465792/3047424 (81%)] Data (t): 0.668 Batch (t): 4.903, 1694.61/s, 423.652/s/gpu LR: 0.000105 Logit Scale: 77.763 Imm_image: 39.593 (39.382) Imm_text: 39.593 (39.382) Isd_image: 6.0166 (5.9019) Isd_text: 6.0166 (5.9019) Contrastive_loss: 0.27674 (0.26048) Loss: 0.27674 (0.26048)
+2025-04-24,21:22:49 | INFO | Train Epoch: 31 [3047424/3047424 (100%)] Data (t): 0.648 Batch (t): 4.866, 1713.21/s, 428.302/s/gpu LR: 0.000100 Logit Scale: 77.997 Imm_image: 40.085 (39.522) Imm_text: 40.085 (39.522) Isd_image: 5.9351 (5.9085) Isd_text: 5.9351 (5.9085) Contrastive_loss: 0.23148 (0.25468) Loss: 0.23148 (0.25468)
+2025-04-24,21:22:51 | INFO | Start epoch 32
+2025-04-24,21:23:07 | INFO | Train Epoch: 32 [   8192/3047424 (0%)] Data (t): 11.805 Batch (t): 16.200, 505.681/s, 126.420/s/gpu LR: 0.000100 Logit Scale: 78.000 Imm_image: 40.104 (40.104) Imm_text: 40.104 (40.104) Isd_image: 5.9143 (5.9143) Isd_text: 5.9143 (5.9143) Contrastive_loss: 0.21168 (0.21168) Loss: 0.21168 (0.21168)
+2025-04-24,21:31:19 | INFO | Train Epoch: 32 [ 827392/3047424 (27%)] Data (t): 0.668 Batch (t): 4.917, 1661.51/s, 415.377/s/gpu LR: 0.000094 Logit Scale: 78.446 Imm_image: 40.240 (40.172) Imm_text: 40.240 (40.172) Isd_image: 6.3322 (6.1232) Isd_text: 6.3322 (6.1232) Contrastive_loss: 0.23479 (0.22324) Loss: 0.23479 (0.22324)
+2025-04-24,21:39:30 | INFO | Train Epoch: 32 [1646592/3047424 (54%)] Data (t): 0.662 Batch (t): 4.913, 1658.97/s, 414.744/s/gpu LR: 0.000088 Logit Scale: 78.825 Imm_image: 40.452 (40.265) Imm_text: 40.452 (40.265) Isd_image: 6.0223 (6.0896) Isd_text: 6.0223 (6.0896) Contrastive_loss: 0.21849 (0.22166) Loss: 0.21849 (0.22166)
+2025-04-24,21:47:40 | INFO | Train Epoch: 32 [2465792/3047424 (81%)] Data (t): 0.661 Batch (t): 4.900, 1668.57/s, 417.142/s/gpu LR: 0.000082 Logit Scale: 79.163 Imm_image: 40.706 (40.376) Imm_text: 40.706 (40.376) Isd_image: 6.0437 (6.0781) Isd_text: 6.0437 (6.0781) Contrastive_loss: 0.23322 (0.22455) Loss: 0.23322 (0.22455)
+2025-04-24,21:53:28 | INFO | Train Epoch: 32 [3047424/3047424 (100%)] Data (t): 0.656 Batch (t): 4.900, 1715.26/s, 428.816/s/gpu LR: 0.000077 Logit Scale: 79.370 Imm_image: 41.104 (40.521) Imm_text: 41.104 (40.521) Isd_image: 6.3035 (6.1232) Isd_text: 6.3035 (6.1232) Contrastive_loss: 0.18004 (0.21565) Loss: 0.18004 (0.21565)
+2025-04-24,21:53:31 | INFO | Start epoch 33
+2025-04-24,21:53:46 | INFO | Train Epoch: 33 [   8192/3047424 (0%)] Data (t): 10.940 Batch (t): 15.363, 533.229/s, 133.307/s/gpu LR: 0.000077 Logit Scale: 79.373 Imm_image: 41.134 (41.134) Imm_text: 41.134 (41.134) Isd_image: 6.3588 (6.3588) Isd_text: 6.3588 (6.3588) Contrastive_loss: 0.18141 (0.18141) Loss: 0.18141 (0.18141)
+2025-04-24,22:02:01 | INFO | Train Epoch: 33 [ 827392/3047424 (27%)] Data (t): 0.664 Batch (t): 4.950, 1651.73/s, 412.934/s/gpu LR: 0.000072 Logit Scale: 79.769 Imm_image: 41.209 (41.172) Imm_text: 41.209 (41.172) Isd_image: 6.0914 (6.2251) Isd_text: 6.0914 (6.2251) Contrastive_loss: 0.18899 (0.18520) Loss: 0.18899 (0.18520)
+2025-04-24,22:10:11 | INFO | Train Epoch: 33 [1646592/3047424 (54%)] Data (t): 0.669 Batch (t): 4.896, 1675.98/s, 418.996/s/gpu LR: 0.000066 Logit Scale: 80.035 Imm_image: 41.430 (41.258) Imm_text: 41.430 (41.258) Isd_image: 6.1677 (6.2060) Isd_text: 6.1677 (6.2060) Contrastive_loss: 0.18902 (0.18647) Loss: 0.18902 (0.18647)
+2025-04-24,22:18:24 | INFO | Train Epoch: 33 [2465792/3047424 (81%)] Data (t): 0.659 Batch (t): 4.930, 1669.29/s, 417.322/s/gpu LR: 0.000061 Logit Scale: 80.312 Imm_image: 41.554 (41.332) Imm_text: 41.554 (41.332) Isd_image: 6.3227 (6.2352) Isd_text: 6.3227 (6.2352) Contrastive_loss: 0.20321 (0.19066) Loss: 0.20321 (0.19066)
+2025-04-24,22:24:10 | INFO | Train Epoch: 33 [3047424/3047424 (100%)] Data (t): 0.647 Batch (t): 4.875, 1722.03/s, 430.506/s/gpu LR: 0.000057 Logit Scale: 80.481 Imm_image: 41.958 (41.457) Imm_text: 41.958 (41.457) Isd_image: 6.0686 (6.2019) Isd_text: 6.0686 (6.2019) Contrastive_loss: 0.14970 (0.18246) Loss: 0.14970 (0.18246)
+2025-04-24,22:24:12 | INFO | Start epoch 34
+2025-04-24,22:24:28 | INFO | Train Epoch: 34 [   8192/3047424 (0%)] Data (t): 11.371 Batch (t): 15.794, 518.694/s, 129.674/s/gpu LR: 0.000057 Logit Scale: 80.485 Imm_image: 42.047 (42.047) Imm_text: 42.047 (42.047) Isd_image: 6.1005 (6.1005) Isd_text: 6.1005 (6.1005) Contrastive_loss: 0.14698 (0.14698) Loss: 0.14698 (0.14698)
+2025-04-24,22:32:38 | INFO | Train Epoch: 34 [ 827392/3047424 (27%)] Data (t): 0.662 Batch (t): 4.905, 1675.14/s, 418.786/s/gpu LR: 0.000052 Logit Scale: 80.753 Imm_image: 42.018 (42.033) Imm_text: 42.018 (42.033) Isd_image: 6.1029 (6.1017) Isd_text: 6.1029 (6.1017) Contrastive_loss: 0.17136 (0.15917) Loss: 0.17136 (0.15917)
+2025-04-24,22:40:50 | INFO | Train Epoch: 34 [1646592/3047424 (54%)] Data (t): 0.667 Batch (t): 4.915, 1641.72/s, 410.430/s/gpu LR: 0.000048 Logit Scale: 80.978 Imm_image: 42.162 (42.076) Imm_text: 42.162 (42.076) Isd_image: 6.0485 (6.0840) Isd_text: 6.0485 (6.0840) Contrastive_loss: 0.15972 (0.15935) Loss: 0.15972 (0.15935)
+2025-04-24,22:48:59 | INFO | Train Epoch: 34 [2465792/3047424 (81%)] Data (t): 0.661 Batch (t): 4.888, 1688.78/s, 422.194/s/gpu LR: 0.000043 Logit Scale: 81.174 Imm_image: 42.234 (42.115) Imm_text: 42.234 (42.115) Isd_image: 6.5003 (6.1881) Isd_text: 6.5003 (6.1881) Contrastive_loss: 0.17829 (0.16409) Loss: 0.17829 (0.16409)
+2025-04-24,22:54:44 | INFO | Train Epoch: 34 [3047424/3047424 (100%)] Data (t): 0.645 Batch (t): 4.859, 1722.19/s, 430.548/s/gpu LR: 0.000040 Logit Scale: 81.320 Imm_image: 42.667 (42.226) Imm_text: 42.667 (42.226) Isd_image: 6.2884 (6.2081) Isd_text: 6.2884 (6.2081) Contrastive_loss: 0.13980 (0.15923) Loss: 0.13980 (0.15923)
+2025-04-24,22:54:46 | INFO | Start epoch 35
+2025-04-24,22:55:02 | INFO | Train Epoch: 35 [   8192/3047424 (0%)] Data (t): 10.839 Batch (t): 15.390, 532.280/s, 133.070/s/gpu LR: 0.000040 Logit Scale: 81.322 Imm_image: 42.733 (42.733) Imm_text: 42.733 (42.733) Isd_image: 6.2185 (6.2185) Isd_text: 6.2185 (6.2185) Contrastive_loss: 0.14039 (0.14039) Loss: 0.14039 (0.14039)
+2025-04-24,23:03:14 | INFO | Train Epoch: 35 [ 827392/3047424 (27%)] Data (t): 0.667 Batch (t): 4.921, 1651.92/s, 412.980/s/gpu LR: 0.000036 Logit Scale: 81.532 Imm_image: 42.650 (42.691) Imm_text: 42.650 (42.691) Isd_image: 6.1043 (6.1614) Isd_text: 6.1043 (6.1614) Contrastive_loss: 0.15395 (0.14717) Loss: 0.15395 (0.14717)
+2025-04-24,23:11:24 | INFO | Train Epoch: 35 [1646592/3047424 (54%)] Data (t): 0.659 Batch (t): 4.904, 1649.26/s, 412.316/s/gpu LR: 0.000032 Logit Scale: 81.689 Imm_image: 42.860 (42.748) Imm_text: 42.860 (42.748) Isd_image: 6.2288 (6.1839) Isd_text: 6.2288 (6.1839) Contrastive_loss: 0.14504 (0.14646) Loss: 0.14504 (0.14646)
+2025-04-24,23:19:35 | INFO | Train Epoch: 35 [2465792/3047424 (81%)] Data (t): 0.669 Batch (t): 4.911, 1661.36/s, 415.341/s/gpu LR: 0.000028 Logit Scale: 81.828 Imm_image: 42.927 (42.792) Imm_text: 42.927 (42.792) Isd_image: 6.0876 (6.1598) Isd_text: 6.0876 (6.1598) Contrastive_loss: 0.12938 (0.14219) Loss: 0.12938 (0.14219)
+2025-04-24,23:25:23 | INFO | Train Epoch: 35 [3047424/3047424 (100%)] Data (t): 0.649 Batch (t): 4.894, 1688.89/s, 422.222/s/gpu LR: 0.000026 Logit Scale: 81.929 Imm_image: 43.110 (42.856) Imm_text: 43.110 (42.856) Isd_image: 6.0662 (6.1411) Isd_text: 6.0662 (6.1411) Contrastive_loss: 0.12332 (0.13842) Loss: 0.12332 (0.13842)
+2025-04-24,23:25:25 | INFO | Start epoch 36
+2025-04-24,23:25:40 | INFO | Train Epoch: 36 [   8192/3047424 (0%)] Data (t): 11.175 Batch (t): 15.687, 522.222/s, 130.556/s/gpu LR: 0.000026 Logit Scale: 81.930 Imm_image: 43.127 (43.127) Imm_text: 43.127 (43.127) Isd_image: 6.0050 (6.0050) Isd_text: 6.0050 (6.0050) Contrastive_loss: 0.12338 (0.12338) Loss: 0.12338 (0.12338)
+2025-04-24,23:33:51 | INFO | Train Epoch: 36 [ 827392/3047424 (27%)] Data (t): 0.653 Batch (t): 4.902, 1664.46/s, 416.115/s/gpu LR: 0.000022 Logit Scale: 82.077 Imm_image: 43.185 (43.156) Imm_text: 43.185 (43.156) Isd_image: 6.1256 (6.0653) Isd_text: 6.1256 (6.0653) Contrastive_loss: 0.14363 (0.13351) Loss: 0.14363 (0.13351)
+2025-04-24,23:42:02 | INFO | Train Epoch: 36 [1646592/3047424 (54%)] Data (t): 0.647 Batch (t): 4.914, 1665.22/s, 416.305/s/gpu LR: 0.000019 Logit Scale: 82.195 Imm_image: 43.395 (43.236) Imm_text: 43.395 (43.236) Isd_image: 6.0253 (6.0520) Isd_text: 6.0253 (6.0520) Contrastive_loss: 0.11268 (0.12656) Loss: 0.11268 (0.12656)
+2025-04-24,23:50:14 | INFO | Train Epoch: 36 [2465792/3047424 (81%)] Data (t): 0.671 Batch (t): 4.918, 1688.21/s, 422.052/s/gpu LR: 0.000016 Logit Scale: 82.295 Imm_image: 43.437 (43.286) Imm_text: 43.437 (43.286) Isd_image: 6.1446 (6.0751) Isd_text: 6.1446 (6.0751) Contrastive_loss: 0.12349 (0.12580) Loss: 0.12349 (0.12580)
+2025-04-24,23:55:59 | INFO | Train Epoch: 36 [3047424/3047424 (100%)] Data (t): 0.651 Batch (t): 4.868, 1721.02/s, 430.256/s/gpu LR: 0.000015 Logit Scale: 82.348 Imm_image: 43.462 (43.321) Imm_text: 43.462 (43.321) Isd_image: 6.0665 (6.0734) Isd_text: 6.0665 (6.0734) Contrastive_loss: 0.11807 (0.12425) Loss: 0.11807 (0.12425)
+2025-04-24,23:56:01 | INFO | Start epoch 37
+2025-04-24,23:56:17 | INFO | Train Epoch: 37 [   8192/3047424 (0%)] Data (t): 11.500 Batch (t): 15.851, 516.829/s, 129.207/s/gpu LR: 0.000015 Logit Scale: 82.349 Imm_image: 43.470 (43.470) Imm_text: 43.470 (43.470) Isd_image: 6.0364 (6.0364) Isd_text: 6.0364 (6.0364) Contrastive_loss: 0.13180 (0.13180) Loss: 0.13180 (0.13180)
+2025-04-25,00:04:29 | INFO | Train Epoch: 37 [ 827392/3047424 (27%)] Data (t): 0.678 Batch (t): 4.919, 1690.09/s, 422.523/s/gpu LR: 0.000012 Logit Scale: 82.436 Imm_image: 43.556 (43.513) Imm_text: 43.556 (43.513) Isd_image: 5.8971 (5.9667) Isd_text: 5.8971 (5.9667) Contrastive_loss: 0.11715 (0.12448) Loss: 0.11715 (0.12448)
+2025-04-25,00:12:39 | INFO | Train Epoch: 37 [1646592/3047424 (54%)] Data (t): 0.657 Batch (t): 4.902, 1665.45/s, 416.364/s/gpu LR: 0.000010 Logit Scale: 82.498 Imm_image: 43.617 (43.548) Imm_text: 43.617 (43.548) Isd_image: 5.7948 (5.9094) Isd_text: 5.7948 (5.9094) Contrastive_loss: 0.11226 (0.12040) Loss: 0.11226 (0.12040)
+2025-04-25,00:20:50 | INFO | Train Epoch: 37 [2465792/3047424 (81%)] Data (t): 0.663 Batch (t): 4.904, 1649.53/s, 412.381/s/gpu LR: 0.000008 Logit Scale: 82.553 Imm_image: 43.727 (43.593) Imm_text: 43.727 (43.593) Isd_image: 5.8659 (5.8985) Isd_text: 5.8659 (5.8985) Contrastive_loss: 0.10293 (0.11603) Loss: 0.10293 (0.11603)
+2025-04-25,00:26:38 | INFO | Train Epoch: 37 [3047424/3047424 (100%)] Data (t): 0.661 Batch (t): 4.909, 1715.83/s, 428.959/s/gpu LR: 0.000006 Logit Scale: 82.580 Imm_image: 43.770 (43.628) Imm_text: 43.770 (43.628) Isd_image: 5.9832 (5.9155) Isd_text: 5.9832 (5.9155) Contrastive_loss: 0.10681 (0.11419) Loss: 0.10681 (0.11419)
+2025-04-25,00:26:41 | INFO | Start epoch 38
+2025-04-25,00:26:57 | INFO | Train Epoch: 38 [   8192/3047424 (0%)] Data (t): 11.546 Batch (t): 16.004, 511.882/s, 127.970/s/gpu LR: 0.000006 Logit Scale: 82.580 Imm_image: 43.854 (43.854) Imm_text: 43.854 (43.854) Isd_image: 6.0387 (6.0387) Isd_text: 6.0387 (6.0387) Contrastive_loss: 0.10888 (0.10888) Loss: 0.10888 (0.10888)
+2025-04-25,00:35:08 | INFO | Train Epoch: 38 [ 827392/3047424 (27%)] Data (t): 0.670 Batch (t): 4.909, 1687.08/s, 421.769/s/gpu LR: 0.000005 Logit Scale: 82.617 Imm_image: 43.834 (43.844) Imm_text: 43.834 (43.844) Isd_image: 5.8917 (5.9652) Isd_text: 5.8917 (5.9652) Contrastive_loss: 0.11157 (0.11022) Loss: 0.11157 (0.11022)
+2025-04-25,00:43:21 | INFO | Train Epoch: 38 [1646592/3047424 (54%)] Data (t): 0.666 Batch (t): 4.929, 1675.09/s, 418.773/s/gpu LR: 0.000003 Logit Scale: 82.640 Imm_image: 43.827 (43.839) Imm_text: 43.827 (43.839) Isd_image: 5.9097 (5.9467) Isd_text: 5.9097 (5.9467) Contrastive_loss: 0.11789 (0.11278) Loss: 0.11789 (0.11278)
+2025-04-25,00:51:32 | INFO | Train Epoch: 38 [2465792/3047424 (81%)] Data (t): 0.661 Batch (t): 4.915, 1663.76/s, 415.941/s/gpu LR: 0.000002 Logit Scale: 82.660 Imm_image: 43.878 (43.848) Imm_text: 43.878 (43.848) Isd_image: 5.8302 (5.9175) Isd_text: 5.8302 (5.9175) Contrastive_loss: 0.10845 (0.11170) Loss: 0.10845 (0.11170)
+2025-04-25,00:57:21 | INFO | Train Epoch: 38 [3047424/3047424 (100%)] Data (t): 0.632 Batch (t): 4.909, 1728.57/s, 432.143/s/gpu LR: 0.000002 Logit Scale: 82.669 Imm_image: 43.890 (43.857) Imm_text: 43.890 (43.857) Isd_image: 5.8269 (5.8994) Isd_text: 5.8269 (5.8994) Contrastive_loss: 0.11058 (0.11147) Loss: 0.11058 (0.11147)
+2025-04-25,00:57:23 | INFO | Start epoch 39
+2025-04-25,00:57:39 | INFO | Train Epoch: 39 [   8192/3047424 (0%)] Data (t): 11.439 Batch (t): 15.799, 518.518/s, 129.629/s/gpu LR: 0.000002 Logit Scale: 82.669 Imm_image: 43.891 (43.891) Imm_text: 43.891 (43.891) Isd_image: 5.8348 (5.8348) Isd_text: 5.8348 (5.8348) Contrastive_loss: 0.10987 (0.10987) Loss: 0.10987 (0.10987)
+2025-04-25,01:05:49 | INFO | Train Epoch: 39 [ 827392/3047424 (27%)] Data (t): 0.676 Batch (t): 4.906, 1673.67/s, 418.418/s/gpu LR: 0.000001 Logit Scale: 82.677 Imm_image: 43.895 (43.893) Imm_text: 43.895 (43.893) Isd_image: 5.7788 (5.8068) Isd_text: 5.7788 (5.8068) Contrastive_loss: 0.10330 (0.10659) Loss: 0.10330 (0.10659)
+2025-04-25,01:13:58 | INFO | Train Epoch: 39 [1646592/3047424 (54%)] Data (t): 0.652 Batch (t): 4.889, 1676.94/s, 419.235/s/gpu LR: 0.000000 Logit Scale: 82.681 Imm_image: 43.834 (43.873) Imm_text: 43.834 (43.873) Isd_image: 5.9291 (5.8476) Isd_text: 5.9291 (5.8476) Contrastive_loss: 0.10881 (0.10733) Loss: 0.10881 (0.10733)
+2025-04-25,01:22:09 | INFO | Train Epoch: 39 [2465792/3047424 (81%)] Data (t): 0.670 Batch (t): 4.908, 1642.33/s, 410.582/s/gpu LR: 0.000000 Logit Scale: 82.681 Imm_image: 43.871 (43.873) Imm_text: 43.871 (43.873) Isd_image: 5.7890 (5.8329) Isd_text: 5.7890 (5.8329) Contrastive_loss: 0.10934 (0.10783) Loss: 0.10934 (0.10783)
+2025-04-25,01:27:57 | INFO | Train Epoch: 39 [3047424/3047424 (100%)] Data (t): 0.657 Batch (t): 4.898, 1714.23/s, 428.557/s/gpu LR: 0.000000 Logit Scale: 82.681 Imm_image: 43.864 (43.871) Imm_text: 43.864 (43.871) Isd_image: 5.7906 (5.8244) Isd_text: 5.7906 (5.8244) Contrastive_loss: 0.098484 (0.10596) Loss: 0.098484 (0.10596)

SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-010-filled-map_text_farest/params.txt

@@ -0,0 +1,103 @@
+accum_freq: 4
+aug_cfg: {}
+batch_size: 512
+beta1: 0.9
+beta2: 0.98
+cache_dir: None
+caption_ratio: 0.1
+checkpoint_path: /mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-010-filled-map_text_farest/checkpoints
+coca_caption_loss_weight: 2.0
+coca_contrastive_loss_weight: 1.0
+copy_codebase: False
+csv_caption_key: title
+csv_img_key: filepath
+csv_separator: 	
+dataset_resampled: False
+dataset_type: synthetic
+ddp_static_graph: False
+debug: False
+delete_previous_checkpoint: False
+device: cuda:0
+dist_backend: None
+dist_url: None
+distill: False
+distill_model: None
+distill_pretrained: None
+distributed: True
+epochs: 40
+epochs_cooldown: None
+eps: 1e-08
+force_custom_text: False
+force_image_size: None
+force_patch_dropout: None
+force_quick_gelu: False
+gather_with_grad: True
+grad_checkpointing: True
+grad_clip_norm: None
+horovod: False
+image_interpolation: None
+image_mean: None
+image_resize_mode: None
+image_std: None
+imagenet_v2: None
+imagenet_val: None
+keep_func_name: 
+local_loss: False
+local_rank: 0
+lock_image: False
+lock_image_freeze_bn_stats: False
+lock_image_unlocked_groups: 0
+lock_text: False
+lock_text_freeze_layer_norm: False
+lock_text_unlocked_layers: 0
+log_every_n_steps: 100
+log_level: 20
+log_local: False
+log_path: /mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-010-filled-map_text_farest/out.log
+logs: /mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs
+loss_dist_impl: None
+lr: 0.001
+lr_cooldown_end: 0.0
+lr_cooldown_power: 1.0
+lr_scheduler: cosine
+map_func_name: map_text_farest
+model: ViT-B-16
+momentum: None
+name: ViT-B-16-cc3m-laclip-mix-010-filled-map_text_farest
+no_set_device_rank: False
+opt: adamw
+precision: amp
+pretrained: 
+pretrained_image: False
+rank: 0
+remote_sync: None
+remote_sync_frequency: 300
+remote_sync_protocol: s3
+report_to: tensorboard,wandb
+resume: None
+save_frequency: 1
+save_most_recent: False
+seed: 0
+siglip: False
+skip_scheduler: False
+tensorboard: True
+tensorboard_path: /mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-010-filled-map_text_farest/tensorboard
+torchcompile: False
+torchscript: False
+trace: False
+train_data: /mnt/personal/zhudongy/cc3m-hgf-wds/{0000..0301}.tar
+train_data_upsampling_factors: None
+train_num_samples: 3016640
+use_bn_sync: False
+use_bnb_linear: None
+val_data: None
+val_frequency: 1
+val_num_samples: None
+wandb: True
+wandb_notes: 
+wandb_project_name: open-clip
+warmup: 368
+wd: 0.5
+workers: 16
+world_size: 4
+zeroshot_frequency: 2

SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-intra-010-filled-r32-b2/benchmark_caltech101_epoch_40.pt_ViT-B-16_en_zeroshot_classification.json

@@ -0,0 +1 @@
+{"dataset": "caltech101", "model": "ViT-B-16", "pretrained": "/mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R_0.95_1.0_dinov2-large_0.0_0.05_rouge_0.7_top_4/logs/ViT-B-16-cc3m-laclip-mix-intra-010-filled-r32-b2/checkpoints/epoch_40.pt", "task": "zeroshot_classification", "metrics": {"acc1": 0.4714566929133858, "acc5": 0.6971784776902887, "mean_per_class_recall": 0.4147448860593748}, "language": "en"}

SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-intra-010-filled-r32-b2/benchmark_cars_epoch_40.pt_ViT-B-16_en_zeroshot_classification.json

@@ -0,0 +1 @@
+{"dataset": "cars", "model": "ViT-B-16", "pretrained": "/mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R_0.95_1.0_dinov2-large_0.0_0.05_rouge_0.7_top_4/logs/ViT-B-16-cc3m-laclip-mix-intra-010-filled-r32-b2/checkpoints/epoch_40.pt", "task": "zeroshot_classification", "metrics": {"acc1": 0.011814450938937944, "acc5": 0.05098868299962691, "mean_per_class_recall": 0.012159162724500011}, "language": "en"}

SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-intra-010-filled-r32-b2/benchmark_cifar100_epoch_40.pt_ViT-B-16_en_zeroshot_classification.json

@@ -0,0 +1 @@
+{"dataset": "cifar100", "model": "ViT-B-16", "pretrained": "/mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R_0.95_1.0_dinov2-large_0.0_0.05_rouge_0.7_top_4/logs/ViT-B-16-cc3m-laclip-mix-intra-010-filled-r32-b2/checkpoints/epoch_40.pt", "task": "zeroshot_classification", "metrics": {"acc1": 0.1956, "acc5": 0.4428, "mean_per_class_recall": 0.1955}, "language": "en"}

SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-intra-010-filled-r32-b2/benchmark_cifar10_epoch_40.pt_ViT-B-16_en_zeroshot_classification.json

@@ -0,0 +1 @@
+{"dataset": "cifar10", "model": "ViT-B-16", "pretrained": "/mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R_0.95_1.0_dinov2-large_0.0_0.05_rouge_0.7_top_4/logs/ViT-B-16-cc3m-laclip-mix-intra-010-filled-r32-b2/checkpoints/epoch_40.pt", "task": "zeroshot_classification", "metrics": {"acc1": 0.4716, "acc5": 0.8997, "mean_per_class_recall": 0.4721}, "language": "en"}

SFR-Embedding-Code-2B_R#0.8#0.6#dinov2-large#0.0#0.2#rouge_0.2#top_8#inter_0.4/logs/ViT-B-16-cc3m-laclip-mix-intra-010-filled-r32-b2/benchmark_country211_epoch_40.pt_ViT-B-16_en_zeroshot_classification.json

@@ -0,0 +1 @@
+{"dataset": "country211", "model": "ViT-B-16", "pretrained": "/mnt/personal/zhudongy/cc3m_results/SFR-Embedding-Code-2B_R_0.95_1.0_dinov2-large_0.0_0.05_rouge_0.7_top_4/logs/ViT-B-16-cc3m-laclip-mix-intra-010-filled-r32-b2/checkpoints/epoch_40.pt", "task": "zeroshot_classification", "metrics": {"acc1": 0.005876777251184834, "acc5": 0.02928909952606635, "mean_per_class_recall": 0.005781990521327015}, "language": "en"}