End of training

all_results.json

@@ -1,13 +1,13 @@
 {
-    "epoch": 3.0,
-    "eval_accuracy": 0.9498432601880877,
-    "eval_loss": 0.1711730659008026,
-    "eval_runtime": 3.5127,
-    "eval_samples_per_second": 181.625,
-    "eval_steps_per_second": 5.694,
-    "total_flos": 4.282621662590853e+17,
-    "train_loss": 0.9712538012751827,
-    "train_runtime": 5653.2043,
-    "train_samples_per_second": 3.047,
-    "train_steps_per_second": 0.024
+    "epoch": 2.938775510204082,
+    "eval_accuracy": 0.8807692307692307,
+    "eval_loss": 0.3335144817829132,
+    "eval_runtime": 4.4218,
+    "eval_samples_per_second": 117.599,
+    "eval_steps_per_second": 3.845,
+    "total_flos": 3.424639439220572e+17,
+    "train_loss": 0.8488516895859329,
+    "train_runtime": 194.721,
+    "train_samples_per_second": 72.103,
+    "train_steps_per_second": 0.555
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 3.0,
-    "eval_accuracy": 0.9498432601880877,
-    "eval_loss": 0.1711730659008026,
-    "eval_runtime": 3.5127,
-    "eval_samples_per_second": 181.625,
-    "eval_steps_per_second": 5.694
+    "epoch": 2.938775510204082,
+    "eval_accuracy": 0.8807692307692307,
+    "eval_loss": 0.3335144817829132,
+    "eval_runtime": 4.4218,
+    "eval_samples_per_second": 117.599,
+    "eval_steps_per_second": 3.845
 }

runs/May13_09-43-49_c641b86eca15/events.out.tfevents.1715593637.c641b86eca15.935.3

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9205960e67c180cb3d7ff5fed93536d063194052280ae01f1b6dc11e1032d7f0
+size 405

train_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 3.0,
-    "total_flos": 4.282621662590853e+17,
-    "train_loss": 0.9712538012751827,
-    "train_runtime": 5653.2043,
-    "train_samples_per_second": 3.047,
-    "train_steps_per_second": 0.024
+    "epoch": 2.938775510204082,
+    "total_flos": 3.424639439220572e+17,
+    "train_loss": 0.8488516895859329,
+    "train_runtime": 194.721,
+    "train_samples_per_second": 72.103,
+    "train_steps_per_second": 0.555
 }

trainer_state.json

@@ -1,147 +1,126 @@
 {
-  "best_metric": 0.9498432601880877,
-  "best_model_checkpoint": "swin-tiny-patch4-window7-224-finetuned/checkpoint-135",
-  "epoch": 3.0,
+  "best_metric": 0.8807692307692307,
+  "best_model_checkpoint": "swin-tiny-patch4-window7-224-finetuned/checkpoint-108",
+  "epoch": 2.938775510204082,
   "eval_steps": 500,
-  "global_step": 135,
+  "global_step": 108,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.2222222222222222,
-      "grad_norm": 5.332180976867676,
-      "learning_rate": 3.571428571428572e-05,
-      "loss": 2.7613,
+      "epoch": 0.272108843537415,
+      "grad_norm": 8.761322975158691,
+      "learning_rate": 4.545454545454546e-05,
+      "loss": 1.8625,
       "step": 10
     },
     {
-      "epoch": 0.4444444444444444,
-      "grad_norm": 12.654841423034668,
-      "learning_rate": 4.75206611570248e-05,
-      "loss": 2.3428,
+      "epoch": 0.54421768707483,
+      "grad_norm": 13.928447723388672,
+      "learning_rate": 4.536082474226804e-05,
+      "loss": 1.5024,
       "step": 20
     },
     {
-      "epoch": 0.6666666666666666,
-      "grad_norm": 13.685147285461426,
-      "learning_rate": 4.338842975206612e-05,
-      "loss": 1.6575,
+      "epoch": 0.8163265306122449,
+      "grad_norm": 14.840970993041992,
+      "learning_rate": 4.020618556701031e-05,
+      "loss": 1.0835,
       "step": 30
     },
     {
-      "epoch": 0.8888888888888888,
-      "grad_norm": 11.329548835754395,
-      "learning_rate": 3.925619834710744e-05,
-      "loss": 1.1146,
-      "step": 40
+      "epoch": 0.9795918367346939,
+      "eval_accuracy": 0.6884615384615385,
+      "eval_loss": 0.7885341048240662,
+      "eval_runtime": 4.1617,
+      "eval_samples_per_second": 124.947,
+      "eval_steps_per_second": 4.085,
+      "step": 36
     },
     {
-      "epoch": 1.0,
-      "eval_accuracy": 0.7774294670846394,
-      "eval_loss": 0.6721618175506592,
-      "eval_runtime": 586.524,
-      "eval_samples_per_second": 1.088,
-      "eval_steps_per_second": 0.034,
-      "step": 45
+      "epoch": 1.08843537414966,
+      "grad_norm": 13.623029708862305,
+      "learning_rate": 3.5051546391752576e-05,
+      "loss": 0.857,
+      "step": 40
     },
     {
-      "epoch": 1.1111111111111112,
-      "grad_norm": 10.730902671813965,
-      "learning_rate": 3.512396694214876e-05,
-      "loss": 0.8703,
+      "epoch": 1.3605442176870748,
+      "grad_norm": 18.793283462524414,
+      "learning_rate": 2.9896907216494846e-05,
+      "loss": 0.7227,
       "step": 50
     },
     {
-      "epoch": 1.3333333333333333,
-      "grad_norm": 9.959417343139648,
-      "learning_rate": 3.099173553719008e-05,
-      "loss": 0.6665,
+      "epoch": 1.6326530612244898,
+      "grad_norm": 9.9773530960083,
+      "learning_rate": 2.4742268041237116e-05,
+      "loss": 0.6269,
       "step": 60
     },
     {
-      "epoch": 1.5555555555555556,
-      "grad_norm": 8.136271476745605,
-      "learning_rate": 2.6859504132231405e-05,
-      "loss": 0.6139,
+      "epoch": 1.9047619047619047,
+      "grad_norm": 10.044248580932617,
+      "learning_rate": 1.9587628865979382e-05,
+      "loss": 0.6076,
       "step": 70
     },
     {
-      "epoch": 1.7777777777777777,
-      "grad_norm": 7.427969455718994,
-      "learning_rate": 2.272727272727273e-05,
-      "loss": 0.5748,
-      "step": 80
+      "epoch": 1.9863945578231292,
+      "eval_accuracy": 0.8096153846153846,
+      "eval_loss": 0.5072689652442932,
+      "eval_runtime": 4.141,
+      "eval_samples_per_second": 125.573,
+      "eval_steps_per_second": 4.105,
+      "step": 73
     },
     {
-      "epoch": 2.0,
-      "grad_norm": 8.851130485534668,
-      "learning_rate": 1.859504132231405e-05,
-      "loss": 0.5419,
-      "step": 90
+      "epoch": 2.17687074829932,
+      "grad_norm": 11.315401077270508,
+      "learning_rate": 1.4432989690721649e-05,
+      "loss": 0.5397,
+      "step": 80
     },
     {
-      "epoch": 2.0,
-      "eval_accuracy": 0.9341692789968652,
-      "eval_loss": 0.2221757173538208,
-      "eval_runtime": 3.3094,
-      "eval_samples_per_second": 192.787,
-      "eval_steps_per_second": 6.043,
+      "epoch": 2.4489795918367347,
+      "grad_norm": 7.718348503112793,
+      "learning_rate": 9.278350515463918e-06,
+      "loss": 0.5192,
       "step": 90
     },
     {
-      "epoch": 2.2222222222222223,
-      "grad_norm": 9.288217544555664,
-      "learning_rate": 1.4462809917355372e-05,
-      "loss": 0.4968,
+      "epoch": 2.7210884353741496,
+      "grad_norm": 5.907692909240723,
+      "learning_rate": 4.123711340206186e-06,
+      "loss": 0.499,
       "step": 100
     },
     {
-      "epoch": 2.4444444444444446,
-      "grad_norm": 12.425986289978027,
-      "learning_rate": 1.0330578512396695e-05,
-      "loss": 0.4119,
-      "step": 110
-    },
-    {
-      "epoch": 2.6666666666666665,
-      "grad_norm": 7.838353157043457,
-      "learning_rate": 6.198347107438017e-06,
-      "loss": 0.4118,
-      "step": 120
-    },
-    {
-      "epoch": 2.888888888888889,
-      "grad_norm": 6.44113826751709,
-      "learning_rate": 2.066115702479339e-06,
-      "loss": 0.4282,
-      "step": 130
-    },
-    {
-      "epoch": 3.0,
-      "eval_accuracy": 0.9498432601880877,
-      "eval_loss": 0.1711730659008026,
-      "eval_runtime": 3.4263,
-      "eval_samples_per_second": 186.204,
-      "eval_steps_per_second": 5.837,
-      "step": 135
+      "epoch": 2.938775510204082,
+      "eval_accuracy": 0.8807692307692307,
+      "eval_loss": 0.3335144817829132,
+      "eval_runtime": 4.0967,
+      "eval_samples_per_second": 126.933,
+      "eval_steps_per_second": 4.15,
+      "step": 108
     },
     {
-      "epoch": 3.0,
-      "step": 135,
-      "total_flos": 4.282621662590853e+17,
-      "train_loss": 0.9712538012751827,
-      "train_runtime": 5653.2043,
-      "train_samples_per_second": 3.047,
-      "train_steps_per_second": 0.024
+      "epoch": 2.938775510204082,
+      "step": 108,
+      "total_flos": 3.424639439220572e+17,
+      "train_loss": 0.8488516895859329,
+      "train_runtime": 194.721,
+      "train_samples_per_second": 72.103,
+      "train_steps_per_second": 0.555
     }
   ],
   "logging_steps": 10,
-  "max_steps": 135,
+  "max_steps": 108,
   "num_input_tokens_seen": 0,
   "num_train_epochs": 3,
   "save_steps": 500,
-  "total_flos": 4.282621662590853e+17,
+  "total_flos": 3.424639439220572e+17,
   "train_batch_size": 32,
   "trial_name": null,
   "trial_params": null