End of training

all_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 3.242021276595745,
+    "epoch": 4.125,
     "eval_accuracy": 0.7633136094674556,
-    "eval_loss": 0.875207781791687,
-    "eval_runtime": 11.2012,
-    "eval_samples_per_second": 15.088,
-    "eval_steps_per_second": 1.964
+    "eval_loss": 0.8172227144241333,
+    "eval_runtime": 11.0103,
+    "eval_samples_per_second": 15.349,
+    "eval_steps_per_second": 1.998
 }

test_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 3.242021276595745,
+    "epoch": 4.125,
     "eval_accuracy": 0.7633136094674556,
-    "eval_loss": 0.875207781791687,
-    "eval_runtime": 11.2012,
-    "eval_samples_per_second": 15.088,
-    "eval_steps_per_second": 1.964
+    "eval_loss": 0.8172227144241333,
+    "eval_runtime": 11.0103,
+    "eval_samples_per_second": 15.349,
+    "eval_steps_per_second": 1.998
 }

trainer_state.json

@@ -1,334 +1,245 @@
 {
   "best_metric": 0.7875,
   "best_model_checkpoint": "videomae-base-finetuned-gesturePhasev2/checkpoint-95",
-  "epoch": 3.242021276595745,
+  "epoch": 4.125,
   "eval_steps": 500,
-  "global_step": 376,
+  "global_step": 237,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
       "epoch": 0.026595744680851064,
-      "grad_norm": 13.211409568786621,
-      "learning_rate": 2.631578947368421e-06,
-      "loss": 1.594,
+      "grad_norm": 9.288494110107422,
+      "learning_rate": 6.578947368421053e-07,
+      "loss": 1.4443,
       "step": 10
     },
     {
       "epoch": 0.05319148936170213,
-      "grad_norm": 9.70760726928711,
-      "learning_rate": 5.263157894736842e-06,
-      "loss": 1.4073,
+      "grad_norm": 10.334871292114258,
+      "learning_rate": 1.3157894736842106e-06,
+      "loss": 1.418,
       "step": 20
     },
     {
       "epoch": 0.0797872340425532,
-      "grad_norm": 5.945734977722168,
-      "learning_rate": 7.894736842105265e-06,
-      "loss": 1.1347,
+      "grad_norm": 7.731433391571045,
+      "learning_rate": 1.973684210526316e-06,
+      "loss": 1.3362,
       "step": 30
     },
     {
       "epoch": 0.10638297872340426,
-      "grad_norm": 8.433605194091797,
-      "learning_rate": 9.940828402366864e-06,
-      "loss": 0.8755,
+      "grad_norm": 10.242834091186523,
+      "learning_rate": 2.631578947368421e-06,
+      "loss": 1.1665,
       "step": 40
     },
     {
-      "epoch": 0.13297872340425532,
-      "grad_norm": 5.273907661437988,
-      "learning_rate": 9.644970414201184e-06,
-      "loss": 0.6064,
+      "epoch": 0.125,
+      "eval_accuracy": 0.775,
+      "eval_loss": 1.0642327070236206,
+      "eval_runtime": 15.1556,
+      "eval_samples_per_second": 10.557,
+      "eval_steps_per_second": 1.32,
+      "step": 47
+    },
+    {
+      "epoch": 1.0066489361702127,
+      "grad_norm": 5.333619594573975,
+      "learning_rate": 3.289473684210527e-06,
+      "loss": 1.0034,
       "step": 50
     },
     {
-      "epoch": 0.1595744680851064,
-      "grad_norm": 6.961372375488281,
-      "learning_rate": 9.349112426035503e-06,
-      "loss": 0.9865,
+      "epoch": 1.0332446808510638,
+      "grad_norm": 4.666859149932861,
+      "learning_rate": 3.947368421052632e-06,
+      "loss": 0.8562,
       "step": 60
     },
     {
-      "epoch": 0.18617021276595744,
-      "grad_norm": 6.787417411804199,
-      "learning_rate": 9.053254437869822e-06,
-      "loss": 0.7214,
+      "epoch": 1.059840425531915,
+      "grad_norm": 4.381842136383057,
+      "learning_rate": 4.605263157894737e-06,
+      "loss": 0.7196,
       "step": 70
     },
     {
-      "epoch": 0.2127659574468085,
-      "grad_norm": 4.117732048034668,
-      "learning_rate": 8.757396449704143e-06,
-      "loss": 0.5468,
+      "epoch": 1.086436170212766,
+      "grad_norm": 4.5117340087890625,
+      "learning_rate": 4.997807075247147e-06,
+      "loss": 0.7522,
       "step": 80
     },
     {
-      "epoch": 0.2393617021276596,
-      "grad_norm": 5.041596412658691,
-      "learning_rate": 8.461538461538462e-06,
-      "loss": 0.806,
+      "epoch": 1.113031914893617,
+      "grad_norm": 3.7782816886901855,
+      "learning_rate": 4.973180832407471e-06,
+      "loss": 0.7316,
       "step": 90
     },
     {
-      "epoch": 0.2526595744680851,
+      "epoch": 1.1263297872340425,
       "eval_accuracy": 0.7875,
-      "eval_loss": 0.8249056935310364,
-      "eval_runtime": 22.7472,
-      "eval_samples_per_second": 7.034,
-      "eval_steps_per_second": 0.879,
+      "eval_loss": 0.7825992703437805,
+      "eval_runtime": 11.6767,
+      "eval_samples_per_second": 13.703,
+      "eval_steps_per_second": 1.713,
       "step": 95
     },
     {
-      "epoch": 1.0132978723404256,
-      "grad_norm": 3.2331671714782715,
-      "learning_rate": 8.165680473372781e-06,
-      "loss": 0.6714,
+      "epoch": 2.0132978723404253,
+      "grad_norm": 3.1058461666107178,
+      "learning_rate": 4.921457902821578e-06,
+      "loss": 0.785,
       "step": 100
     },
     {
-      "epoch": 1.0398936170212767,
-      "grad_norm": 5.518906116485596,
-      "learning_rate": 7.869822485207102e-06,
-      "loss": 0.7709,
+      "epoch": 2.0398936170212765,
+      "grad_norm": 3.0555734634399414,
+      "learning_rate": 4.84320497372973e-06,
+      "loss": 0.7795,
       "step": 110
     },
     {
-      "epoch": 1.0664893617021276,
-      "grad_norm": 4.540952205657959,
-      "learning_rate": 7.573964497041421e-06,
-      "loss": 0.7454,
+      "epoch": 2.0664893617021276,
+      "grad_norm": 3.8933866024017334,
+      "learning_rate": 4.7392794005985324e-06,
+      "loss": 0.774,
       "step": 120
     },
     {
-      "epoch": 1.0930851063829787,
-      "grad_norm": 5.326303482055664,
-      "learning_rate": 7.278106508875741e-06,
-      "loss": 0.6977,
+      "epoch": 2.0930851063829787,
+      "grad_norm": 4.3550591468811035,
+      "learning_rate": 4.610819813755038e-06,
+      "loss": 0.7218,
       "step": 130
     },
     {
-      "epoch": 1.1196808510638299,
-      "grad_norm": 9.28279972076416,
-      "learning_rate": 6.98224852071006e-06,
-      "loss": 0.6382,
+      "epoch": 2.11968085106383,
+      "grad_norm": 4.537750720977783,
+      "learning_rate": 4.4592336433146e-06,
+      "loss": 0.7259,
       "step": 140
     },
     {
-      "epoch": 1.1462765957446808,
-      "grad_norm": 3.8069820404052734,
-      "learning_rate": 6.686390532544379e-06,
-      "loss": 0.7963,
+      "epoch": 2.125,
+      "eval_accuracy": 0.7875,
+      "eval_loss": 0.8042387962341309,
+      "eval_runtime": 10.5863,
+      "eval_samples_per_second": 15.114,
+      "eval_steps_per_second": 1.889,
+      "step": 142
+    },
+    {
+      "epoch": 3.0199468085106385,
+      "grad_norm": 3.3319995403289795,
+      "learning_rate": 4.286181699082008e-06,
+      "loss": 0.6222,
       "step": 150
     },
     {
-      "epoch": 1.172872340425532,
-      "grad_norm": 3.917985677719116,
-      "learning_rate": 6.3905325443786995e-06,
-      "loss": 0.6782,
+      "epoch": 3.046542553191489,
+      "grad_norm": 4.209911823272705,
+      "learning_rate": 4.093559974371725e-06,
+      "loss": 0.7533,
       "step": 160
     },
     {
-      "epoch": 1.199468085106383,
-      "grad_norm": 5.1802802085876465,
-      "learning_rate": 6.0946745562130185e-06,
-      "loss": 0.6922,
+      "epoch": 3.0731382978723403,
+      "grad_norm": 2.9892210960388184,
+      "learning_rate": 3.88347887310836e-06,
+      "loss": 0.8408,
       "step": 170
     },
     {
-      "epoch": 1.226063829787234,
-      "grad_norm": 2.9730751514434814,
-      "learning_rate": 5.7988165680473375e-06,
-      "loss": 0.7682,
+      "epoch": 3.0997340425531914,
+      "grad_norm": 4.7717814445495605,
+      "learning_rate": 3.658240087799655e-06,
+      "loss": 0.7287,
       "step": 180
     },
     {
-      "epoch": 1.252659574468085,
-      "grad_norm": 6.959865570068359,
-      "learning_rate": 5.502958579881657e-06,
-      "loss": 0.8453,
+      "epoch": 3.1263297872340425,
+      "grad_norm": 5.496540546417236,
+      "learning_rate": 3.4203113817116955e-06,
+      "loss": 0.6643,
       "step": 190
     },
     {
-      "epoch": 1.252659574468085,
+      "epoch": 3.1263297872340425,
       "eval_accuracy": 0.7875,
-      "eval_loss": 0.8071285486221313,
-      "eval_runtime": 10.9141,
-      "eval_samples_per_second": 14.66,
-      "eval_steps_per_second": 1.832,
+      "eval_loss": 0.8022773861885071,
+      "eval_runtime": 10.4825,
+      "eval_samples_per_second": 15.264,
+      "eval_steps_per_second": 1.908,
       "step": 190
     },
     {
-      "epoch": 2.026595744680851,
-      "grad_norm": 3.4485950469970703,
-      "learning_rate": 5.207100591715976e-06,
-      "loss": 0.7081,
+      "epoch": 4.026595744680851,
+      "grad_norm": 3.0813848972320557,
+      "learning_rate": 3.1722995515381644e-06,
+      "loss": 0.6386,
       "step": 200
     },
     {
-      "epoch": 2.0531914893617023,
-      "grad_norm": 3.514798164367676,
-      "learning_rate": 4.911242603550296e-06,
-      "loss": 0.6794,
+      "epoch": 4.053191489361702,
+      "grad_norm": 3.0726428031921387,
+      "learning_rate": 2.9169218667902562e-06,
+      "loss": 0.6708,
       "step": 210
     },
     {
-      "epoch": 2.0797872340425534,
-      "grad_norm": 4.564563751220703,
-      "learning_rate": 4.615384615384616e-06,
-      "loss": 0.8428,
+      "epoch": 4.079787234042553,
+      "grad_norm": 5.47495174407959,
+      "learning_rate": 2.6569762988232838e-06,
+      "loss": 0.8216,
       "step": 220
     },
     {
-      "epoch": 2.106382978723404,
-      "grad_norm": 5.189121723175049,
-      "learning_rate": 4.319526627218935e-06,
-      "loss": 0.8315,
+      "epoch": 4.1063829787234045,
+      "grad_norm": 3.801624059677124,
+      "learning_rate": 2.3953108656770018e-06,
+      "loss": 0.761,
       "step": 230
     },
     {
-      "epoch": 2.132978723404255,
-      "grad_norm": 6.5788044929504395,
-      "learning_rate": 4.023668639053255e-06,
-      "loss": 0.6639,
-      "step": 240
-    },
-    {
-      "epoch": 2.1595744680851063,
-      "grad_norm": 6.100628852844238,
-      "learning_rate": 3.7278106508875745e-06,
-      "loss": 0.7014,
-      "step": 250
-    },
-    {
-      "epoch": 2.1861702127659575,
-      "grad_norm": 5.754176616668701,
-      "learning_rate": 3.4319526627218935e-06,
-      "loss": 0.7219,
-      "step": 260
-    },
-    {
-      "epoch": 2.2127659574468086,
-      "grad_norm": 6.1187424659729,
-      "learning_rate": 3.1360946745562134e-06,
-      "loss": 0.5346,
-      "step": 270
-    },
-    {
-      "epoch": 2.2393617021276597,
-      "grad_norm": 4.662169933319092,
-      "learning_rate": 2.840236686390533e-06,
-      "loss": 0.8814,
-      "step": 280
-    },
-    {
-      "epoch": 2.252659574468085,
+      "epoch": 4.125,
       "eval_accuracy": 0.7875,
-      "eval_loss": 0.8307636380195618,
-      "eval_runtime": 10.9219,
-      "eval_samples_per_second": 14.649,
-      "eval_steps_per_second": 1.831,
-      "step": 285
-    },
-    {
-      "epoch": 3.0132978723404253,
-      "grad_norm": 5.648519992828369,
-      "learning_rate": 2.5443786982248527e-06,
-      "loss": 0.601,
-      "step": 290
-    },
-    {
-      "epoch": 3.0398936170212765,
-      "grad_norm": 3.8109617233276367,
-      "learning_rate": 2.2485207100591717e-06,
-      "loss": 0.618,
-      "step": 300
+      "eval_loss": 0.8077355623245239,
+      "eval_runtime": 16.5222,
+      "eval_samples_per_second": 9.684,
+      "eval_steps_per_second": 1.21,
+      "step": 237
     },
     {
-      "epoch": 3.0664893617021276,
-      "grad_norm": 4.005825519561768,
-      "learning_rate": 1.952662721893491e-06,
-      "loss": 0.6362,
-      "step": 310
+      "epoch": 4.125,
+      "step": 237,
+      "total_flos": 4.691516857081528e+18,
+      "train_loss": 0.8600665486814604,
+      "train_runtime": 867.0351,
+      "train_samples_per_second": 6.939,
+      "train_steps_per_second": 0.434
     },
     {
-      "epoch": 3.0930851063829787,
-      "grad_norm": 6.642380237579346,
-      "learning_rate": 1.656804733727811e-06,
-      "loss": 0.8369,
-      "step": 320
-    },
-    {
-      "epoch": 3.11968085106383,
-      "grad_norm": 4.232272624969482,
-      "learning_rate": 1.3609467455621303e-06,
-      "loss": 0.7755,
-      "step": 330
-    },
-    {
-      "epoch": 3.146276595744681,
-      "grad_norm": 3.2232885360717773,
-      "learning_rate": 1.06508875739645e-06,
-      "loss": 0.8827,
-      "step": 340
-    },
-    {
-      "epoch": 3.172872340425532,
-      "grad_norm": 5.880322456359863,
-      "learning_rate": 7.692307692307694e-07,
-      "loss": 0.6081,
-      "step": 350
-    },
-    {
-      "epoch": 3.199468085106383,
-      "grad_norm": 6.9393310546875,
-      "learning_rate": 4.733727810650888e-07,
-      "loss": 0.8163,
-      "step": 360
-    },
-    {
-      "epoch": 3.226063829787234,
-      "grad_norm": 6.549450874328613,
-      "learning_rate": 1.775147928994083e-07,
-      "loss": 0.8272,
-      "step": 370
-    },
-    {
-      "epoch": 3.242021276595745,
-      "eval_accuracy": 0.7875,
-      "eval_loss": 0.8259578943252563,
-      "eval_runtime": 84.7179,
-      "eval_samples_per_second": 1.889,
-      "eval_steps_per_second": 0.236,
-      "step": 376
-    },
-    {
-      "epoch": 3.242021276595745,
-      "step": 376,
-      "total_flos": 3.7220613152994755e+18,
-      "train_loss": 0.7830839626332546,
-      "train_runtime": 814.9801,
-      "train_samples_per_second": 3.691,
-      "train_steps_per_second": 0.461
-    },
-    {
-      "epoch": 3.242021276595745,
+      "epoch": 4.125,
       "eval_accuracy": 0.7633136094674556,
-      "eval_loss": 0.8752076625823975,
-      "eval_runtime": 14.0179,
-      "eval_samples_per_second": 12.056,
-      "eval_steps_per_second": 1.569,
-      "step": 376
+      "eval_loss": 0.8172227144241333,
+      "eval_runtime": 60.5871,
+      "eval_samples_per_second": 2.789,
+      "eval_steps_per_second": 0.363,
+      "step": 237
     },
     {
-      "epoch": 3.242021276595745,
+      "epoch": 4.125,
       "eval_accuracy": 0.7633136094674556,
-      "eval_loss": 0.875207781791687,
-      "eval_runtime": 11.2012,
-      "eval_samples_per_second": 15.088,
-      "eval_steps_per_second": 1.964,
-      "step": 376
+      "eval_loss": 0.8172227144241333,
+      "eval_runtime": 11.0103,
+      "eval_samples_per_second": 15.349,
+      "eval_steps_per_second": 1.998,
+      "step": 237
     }
   ],
   "logging_steps": 10,
@@ -337,6 +248,15 @@
   "num_train_epochs": 9223372036854775807,
   "save_steps": 500,
   "stateful_callbacks": {
+    "EarlyStoppingCallback": {
+      "args": {
+        "early_stopping_patience": 3,
+        "early_stopping_threshold": 0.0
+      },
+      "attributes": {
+        "early_stopping_patience_counter": 0
+      }
+    },
     "TrainerControl": {
       "args": {
         "should_epoch_stop": false,
@@ -348,7 +268,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 3.7220613152994755e+18,
+  "total_flos": 4.691516857081528e+18,
   "train_batch_size": 8,
   "trial_name": null,
   "trial_params": null