End of training

all_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 19.73,
-    "total_flos": 5.187420432666501e+17,
-    "train_loss": 0.4364403486251831,
-    "train_runtime": 567.1784,
-    "train_samples_per_second": 11.954,
-    "train_steps_per_second": 0.071
+    "epoch": 29.73,
+    "total_flos": 7.813276079484396e+17,
+    "train_loss": 0.042075487226247786,
+    "train_runtime": 801.9686,
+    "train_samples_per_second": 12.681,
+    "train_steps_per_second": 0.075
 }

runs/Jul06_10-45-23_512daece0cbe/events.out.tfevents.1657104335.512daece0cbe.74.2

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:21663d0c8dff0bb6f06286c0f5565e84f94bfe353b6a5a5b084448ba27ff4b60
-size 4630
+oid sha256:93052b3fd5b54caf4537ef8d09df42e8c0927af4608878f6a4bf41bf184442ea
+size 15095

train_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 19.73,
-    "total_flos": 5.187420432666501e+17,
-    "train_loss": 0.4364403486251831,
-    "train_runtime": 567.1784,
-    "train_samples_per_second": 11.954,
-    "train_steps_per_second": 0.071
+    "epoch": 29.73,
+    "total_flos": 7.813276079484396e+17,
+    "train_loss": 0.042075487226247786,
+    "train_runtime": 801.9686,
+    "train_samples_per_second": 12.681,
+    "train_steps_per_second": 0.075
 }

trainer_state.json

@@ -1,229 +1,331 @@
 {
-  "best_metric": 0.910958904109589,
-  "best_model_checkpoint": "beit-base-patch16-224-pt22k-ft22k-rim_one-new/checkpoint-30",
-  "epoch": 19.727272727272727,
-  "global_step": 40,
+  "best_metric": 0.9178082191780822,
+  "best_model_checkpoint": "beit-base-patch16-224-pt22k-ft22k-rim_one-new/checkpoint-2",
+  "epoch": 29.727272727272727,
+  "global_step": 60,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
       "epoch": 0.73,
-      "eval_accuracy": 0.4794520547945205,
-      "eval_loss": 0.6921706199645996,
-      "eval_runtime": 25.0333,
-      "eval_samples_per_second": 5.832,
-      "eval_steps_per_second": 0.2,
+      "eval_accuracy": 0.9178082191780822,
+      "eval_loss": 0.24113720655441284,
+      "eval_runtime": 4.0491,
+      "eval_samples_per_second": 36.057,
+      "eval_steps_per_second": 1.235,
       "step": 2
     },
     {
       "epoch": 1.73,
-      "eval_accuracy": 0.6506849315068494,
-      "eval_loss": 0.7311218976974487,
-      "eval_runtime": 4.1824,
-      "eval_samples_per_second": 34.909,
-      "eval_steps_per_second": 1.195,
+      "eval_accuracy": 0.8972602739726028,
+      "eval_loss": 0.21820734441280365,
+      "eval_runtime": 4.3715,
+      "eval_samples_per_second": 33.398,
+      "eval_steps_per_second": 1.144,
       "step": 4
     },
     {
       "epoch": 2.73,
-      "eval_accuracy": 0.4178082191780822,
-      "eval_loss": 0.9364489316940308,
-      "eval_runtime": 4.1884,
-      "eval_samples_per_second": 34.858,
-      "eval_steps_per_second": 1.194,
+      "eval_accuracy": 0.8972602739726028,
+      "eval_loss": 0.3085276186466217,
+      "eval_runtime": 4.3762,
+      "eval_samples_per_second": 33.362,
+      "eval_steps_per_second": 1.143,
       "step": 6
     },
     {
       "epoch": 3.73,
-      "eval_accuracy": 0.7945205479452054,
-      "eval_loss": 0.4742239713668823,
-      "eval_runtime": 4.2915,
-      "eval_samples_per_second": 34.021,
-      "eval_steps_per_second": 1.165,
+      "eval_accuracy": 0.8972602739726028,
+      "eval_loss": 0.27944016456604004,
+      "eval_runtime": 4.3106,
+      "eval_samples_per_second": 33.87,
+      "eval_steps_per_second": 1.16,
       "step": 8
     },
     {
       "epoch": 4.73,
-      "learning_rate": 8.333333333333334e-05,
-      "loss": 0.8198,
+      "learning_rate": 4.62962962962963e-05,
+      "loss": 0.1392,
       "step": 10
     },
     {
       "epoch": 4.73,
-      "eval_accuracy": 0.8561643835616438,
-      "eval_loss": 0.36144283413887024,
-      "eval_runtime": 4.4462,
-      "eval_samples_per_second": 32.837,
-      "eval_steps_per_second": 1.125,
+      "eval_accuracy": 0.910958904109589,
+      "eval_loss": 0.23977190256118774,
+      "eval_runtime": 4.3753,
+      "eval_samples_per_second": 33.369,
+      "eval_steps_per_second": 1.143,
       "step": 10
     },
     {
       "epoch": 5.73,
-      "eval_accuracy": 0.8561643835616438,
-      "eval_loss": 0.3010813593864441,
-      "eval_runtime": 4.3245,
-      "eval_samples_per_second": 33.761,
+      "eval_accuracy": 0.8972602739726028,
+      "eval_loss": 0.29249313473701477,
+      "eval_runtime": 4.3239,
+      "eval_samples_per_second": 33.766,
       "eval_steps_per_second": 1.156,
       "step": 12
     },
     {
       "epoch": 6.73,
-      "eval_accuracy": 0.8561643835616438,
-      "eval_loss": 0.2709261178970337,
-      "eval_runtime": 4.3245,
-      "eval_samples_per_second": 33.761,
-      "eval_steps_per_second": 1.156,
+      "eval_accuracy": 0.910958904109589,
+      "eval_loss": 0.2797797918319702,
+      "eval_runtime": 4.3315,
+      "eval_samples_per_second": 33.707,
+      "eval_steps_per_second": 1.154,
       "step": 14
     },
     {
       "epoch": 7.73,
-      "eval_accuracy": 0.6027397260273972,
-      "eval_loss": 0.7967002987861633,
-      "eval_runtime": 4.3664,
-      "eval_samples_per_second": 33.437,
-      "eval_steps_per_second": 1.145,
+      "eval_accuracy": 0.9178082191780822,
+      "eval_loss": 0.21835216879844666,
+      "eval_runtime": 4.3475,
+      "eval_samples_per_second": 33.582,
+      "eval_steps_per_second": 1.15,
       "step": 16
     },
     {
       "epoch": 8.73,
-      "eval_accuracy": 0.8287671232876712,
-      "eval_loss": 0.39434078335762024,
-      "eval_runtime": 4.348,
-      "eval_samples_per_second": 33.578,
-      "eval_steps_per_second": 1.15,
+      "eval_accuracy": 0.910958904109589,
+      "eval_loss": 0.30065301060676575,
+      "eval_runtime": 4.3333,
+      "eval_samples_per_second": 33.693,
+      "eval_steps_per_second": 1.154,
       "step": 18
     },
     {
       "epoch": 9.73,
-      "learning_rate": 5.555555555555556e-05,
-      "loss": 0.5673,
+      "learning_rate": 3.7037037037037037e-05,
+      "loss": 0.0416,
       "step": 20
     },
     {
       "epoch": 9.73,
-      "eval_accuracy": 0.8835616438356164,
-      "eval_loss": 0.2661159336566925,
-      "eval_runtime": 4.3636,
-      "eval_samples_per_second": 33.458,
-      "eval_steps_per_second": 1.146,
+      "eval_accuracy": 0.9041095890410958,
+      "eval_loss": 0.3344126045703888,
+      "eval_runtime": 4.366,
+      "eval_samples_per_second": 33.44,
+      "eval_steps_per_second": 1.145,
       "step": 20
     },
     {
       "epoch": 10.73,
-      "eval_accuracy": 0.7671232876712328,
-      "eval_loss": 0.4626728296279907,
-      "eval_runtime": 4.4331,
-      "eval_samples_per_second": 32.934,
-      "eval_steps_per_second": 1.128,
+      "eval_accuracy": 0.910958904109589,
+      "eval_loss": 0.36264100670814514,
+      "eval_runtime": 4.3578,
+      "eval_samples_per_second": 33.503,
+      "eval_steps_per_second": 1.147,
       "step": 22
     },
     {
       "epoch": 11.73,
-      "eval_accuracy": 0.8835616438356164,
-      "eval_loss": 0.24509218335151672,
-      "eval_runtime": 4.4007,
-      "eval_samples_per_second": 33.177,
-      "eval_steps_per_second": 1.136,
+      "eval_accuracy": 0.8904109589041096,
+      "eval_loss": 0.48421263694763184,
+      "eval_runtime": 4.3873,
+      "eval_samples_per_second": 33.278,
+      "eval_steps_per_second": 1.14,
       "step": 24
     },
     {
       "epoch": 12.73,
-      "eval_accuracy": 0.9041095890410958,
-      "eval_loss": 0.23970794677734375,
-      "eval_runtime": 4.1384,
-      "eval_samples_per_second": 35.279,
-      "eval_steps_per_second": 1.208,
+      "eval_accuracy": 0.8972602739726028,
+      "eval_loss": 0.36641207337379456,
+      "eval_runtime": 4.1126,
+      "eval_samples_per_second": 35.5,
+      "eval_steps_per_second": 1.216,
       "step": 26
     },
     {
       "epoch": 13.73,
-      "eval_accuracy": 0.8904109589041096,
-      "eval_loss": 0.24433596432209015,
-      "eval_runtime": 4.1349,
-      "eval_samples_per_second": 35.309,
-      "eval_steps_per_second": 1.209,
+      "eval_accuracy": 0.910958904109589,
+      "eval_loss": 0.3457633852958679,
+      "eval_runtime": 4.131,
+      "eval_samples_per_second": 35.342,
+      "eval_steps_per_second": 1.21,
       "step": 28
     },
     {
       "epoch": 14.73,
       "learning_rate": 2.777777777777778e-05,
-      "loss": 0.2571,
+      "loss": 0.0263,
       "step": 30
     },
     {
       "epoch": 14.73,
       "eval_accuracy": 0.910958904109589,
-      "eval_loss": 0.24097245931625366,
-      "eval_runtime": 4.2188,
-      "eval_samples_per_second": 34.607,
-      "eval_steps_per_second": 1.185,
+      "eval_loss": 0.28100982308387756,
+      "eval_runtime": 4.0812,
+      "eval_samples_per_second": 35.773,
+      "eval_steps_per_second": 1.225,
       "step": 30
     },
     {
       "epoch": 15.73,
-      "eval_accuracy": 0.8904109589041096,
-      "eval_loss": 0.2714996933937073,
-      "eval_runtime": 4.1836,
-      "eval_samples_per_second": 34.898,
-      "eval_steps_per_second": 1.195,
+      "eval_accuracy": 0.8698630136986302,
+      "eval_loss": 0.4695476293563843,
+      "eval_runtime": 4.0905,
+      "eval_samples_per_second": 35.692,
+      "eval_steps_per_second": 1.222,
       "step": 32
     },
     {
       "epoch": 16.73,
       "eval_accuracy": 0.9041095890410958,
-      "eval_loss": 0.2597528100013733,
-      "eval_runtime": 4.1231,
-      "eval_samples_per_second": 35.41,
-      "eval_steps_per_second": 1.213,
+      "eval_loss": 0.3723364770412445,
+      "eval_runtime": 4.1081,
+      "eval_samples_per_second": 35.539,
+      "eval_steps_per_second": 1.217,
       "step": 34
     },
     {
       "epoch": 17.73,
-      "eval_accuracy": 0.8972602739726028,
-      "eval_loss": 0.25846657156944275,
-      "eval_runtime": 4.1153,
-      "eval_samples_per_second": 35.477,
-      "eval_steps_per_second": 1.215,
+      "eval_accuracy": 0.9041095890410958,
+      "eval_loss": 0.34467366337776184,
+      "eval_runtime": 4.1554,
+      "eval_samples_per_second": 35.135,
+      "eval_steps_per_second": 1.203,
       "step": 36
     },
     {
       "epoch": 18.73,
-      "eval_accuracy": 0.910958904109589,
-      "eval_loss": 0.2306668758392334,
-      "eval_runtime": 4.1812,
-      "eval_samples_per_second": 34.919,
-      "eval_steps_per_second": 1.196,
+      "eval_accuracy": 0.8904109589041096,
+      "eval_loss": 0.37077629566192627,
+      "eval_runtime": 4.0887,
+      "eval_samples_per_second": 35.708,
+      "eval_steps_per_second": 1.223,
       "step": 38
     },
     {
       "epoch": 19.73,
-      "learning_rate": 0.0,
-      "loss": 0.1015,
+      "learning_rate": 1.8518518518518518e-05,
+      "loss": 0.0264,
       "step": 40
     },
     {
       "epoch": 19.73,
+      "eval_accuracy": 0.910958904109589,
+      "eval_loss": 0.4051611125469208,
+      "eval_runtime": 4.1078,
+      "eval_samples_per_second": 35.542,
+      "eval_steps_per_second": 1.217,
+      "step": 40
+    },
+    {
+      "epoch": 20.73,
+      "eval_accuracy": 0.9041095890410958,
+      "eval_loss": 0.44917088747024536,
+      "eval_runtime": 4.0898,
+      "eval_samples_per_second": 35.699,
+      "eval_steps_per_second": 1.223,
+      "step": 42
+    },
+    {
+      "epoch": 21.73,
+      "eval_accuracy": 0.8904109589041096,
+      "eval_loss": 0.4648548662662506,
+      "eval_runtime": 4.151,
+      "eval_samples_per_second": 35.173,
+      "eval_steps_per_second": 1.205,
+      "step": 44
+    },
+    {
+      "epoch": 22.73,
+      "eval_accuracy": 0.9178082191780822,
+      "eval_loss": 0.4061222970485687,
+      "eval_runtime": 4.1096,
+      "eval_samples_per_second": 35.527,
+      "eval_steps_per_second": 1.217,
+      "step": 46
+    },
+    {
+      "epoch": 23.73,
+      "eval_accuracy": 0.910958904109589,
+      "eval_loss": 0.4136480391025543,
+      "eval_runtime": 4.0895,
+      "eval_samples_per_second": 35.701,
+      "eval_steps_per_second": 1.223,
+      "step": 48
+    },
+    {
+      "epoch": 24.73,
+      "learning_rate": 9.259259259259259e-06,
+      "loss": 0.0139,
+      "step": 50
+    },
+    {
+      "epoch": 24.73,
       "eval_accuracy": 0.8972602739726028,
-      "eval_loss": 0.2697519063949585,
-      "eval_runtime": 4.1223,
-      "eval_samples_per_second": 35.417,
+      "eval_loss": 0.41833144426345825,
+      "eval_runtime": 4.0666,
+      "eval_samples_per_second": 35.903,
+      "eval_steps_per_second": 1.23,
+      "step": 50
+    },
+    {
+      "epoch": 25.73,
+      "eval_accuracy": 0.8904109589041096,
+      "eval_loss": 0.4503590762615204,
+      "eval_runtime": 4.105,
+      "eval_samples_per_second": 35.567,
+      "eval_steps_per_second": 1.218,
+      "step": 52
+    },
+    {
+      "epoch": 26.73,
+      "eval_accuracy": 0.8972602739726028,
+      "eval_loss": 0.43678611516952515,
+      "eval_runtime": 4.1124,
+      "eval_samples_per_second": 35.503,
+      "eval_steps_per_second": 1.216,
+      "step": 54
+    },
+    {
+      "epoch": 27.73,
+      "eval_accuracy": 0.910958904109589,
+      "eval_loss": 0.4710823893547058,
+      "eval_runtime": 4.126,
+      "eval_samples_per_second": 35.385,
+      "eval_steps_per_second": 1.212,
+      "step": 56
+    },
+    {
+      "epoch": 28.73,
+      "eval_accuracy": 0.910958904109589,
+      "eval_loss": 0.39282917976379395,
+      "eval_runtime": 4.1214,
+      "eval_samples_per_second": 35.425,
       "eval_steps_per_second": 1.213,
-      "step": 40
+      "step": 58
     },
     {
-      "epoch": 19.73,
-      "step": 40,
-      "total_flos": 5.187420432666501e+17,
-      "train_loss": 0.4364403486251831,
-      "train_runtime": 567.1784,
-      "train_samples_per_second": 11.954,
-      "train_steps_per_second": 0.071
+      "epoch": 29.73,
+      "learning_rate": 0.0,
+      "loss": 0.005,
+      "step": 60
+    },
+    {
+      "epoch": 29.73,
+      "eval_accuracy": 0.8767123287671232,
+      "eval_loss": 0.45500674843788147,
+      "eval_runtime": 4.1417,
+      "eval_samples_per_second": 35.251,
+      "eval_steps_per_second": 1.207,
+      "step": 60
+    },
+    {
+      "epoch": 29.73,
+      "step": 60,
+      "total_flos": 7.813276079484396e+17,
+      "train_loss": 0.042075487226247786,
+      "train_runtime": 801.9686,
+      "train_samples_per_second": 12.681,
+      "train_steps_per_second": 0.075
     }
   ],
-  "max_steps": 40,
-  "num_train_epochs": 20,
-  "total_flos": 5.187420432666501e+17,
+  "max_steps": 60,
+  "num_train_epochs": 30,
+  "total_flos": 7.813276079484396e+17,
   "trial_name": null,
   "trial_params": null
 }