End of training

all_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 7.078947368421052,
-    "eval_accuracy": 0.7972027972027972,
-    "eval_loss": 0.7915867567062378,
-    "eval_runtime": 1521.5044,
-    "eval_samples_per_second": 0.282,
-    "eval_steps_per_second": 0.018
+    "epoch": 3.2105263157894735,
+    "eval_accuracy": 0.5407925407925408,
+    "eval_loss": 0.9939900040626526,
+    "eval_runtime": 1531.0787,
+    "eval_samples_per_second": 0.28,
+    "eval_steps_per_second": 0.035
 }

test_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 7.078947368421052,
-    "eval_accuracy": 0.7972027972027972,
-    "eval_loss": 0.7915867567062378,
-    "eval_runtime": 1521.5044,
-    "eval_samples_per_second": 0.282,
-    "eval_steps_per_second": 0.018
+    "epoch": 3.2105263157894735,
+    "eval_accuracy": 0.5407925407925408,
+    "eval_loss": 0.9939900040626526,
+    "eval_runtime": 1531.0787,
+    "eval_samples_per_second": 0.28,
+    "eval_steps_per_second": 0.035
 }

trainer_state.json

@@ -1,7 +1,7 @@
 {
-  "best_metric": 0.5598526703499079,
+  "best_metric": 0.5524861878453039,
   "best_model_checkpoint": "videomae-base-finetuned-lift-data-resize/checkpoint-76",
-  "epoch": 7.078947368421052,
+  "epoch": 3.2105263157894735,
   "eval_steps": 500,
   "global_step": 76,
   "is_hyper_param_search": false,
@@ -10,150 +10,114 @@
   "log_history": [
     {
       "epoch": 0.13157894736842105,
-      "grad_norm": 5.784543514251709,
+      "grad_norm": 5.10792875289917,
       "learning_rate": 4.8529411764705885e-05,
-      "loss": 1.6087,
+      "loss": 1.6509,
       "step": 10
     },
     {
-      "epoch": 0.13157894736842105,
-      "eval_accuracy": 0.2596685082872928,
-      "eval_loss": 1.8482669591903687,
-      "eval_runtime": 1902.2155,
-      "eval_samples_per_second": 0.285,
-      "eval_steps_per_second": 0.018,
-      "step": 10
-    },
-    {
-      "epoch": 1.131578947368421,
-      "grad_norm": 5.533542633056641,
+      "epoch": 0.2631578947368421,
+      "grad_norm": 5.389455795288086,
       "learning_rate": 4.11764705882353e-05,
-      "loss": 1.3273,
+      "loss": 1.5445,
       "step": 20
     },
     {
-      "epoch": 1.131578947368421,
-      "eval_accuracy": 0.2983425414364641,
-      "eval_loss": 1.445150375366211,
-      "eval_runtime": 1919.0254,
-      "eval_samples_per_second": 0.283,
-      "eval_steps_per_second": 0.018,
+      "epoch": 0.2631578947368421,
+      "eval_accuracy": 0.11233885819521179,
+      "eval_loss": 1.6094709634780884,
+      "eval_runtime": 1927.838,
+      "eval_samples_per_second": 0.282,
+      "eval_steps_per_second": 0.035,
       "step": 20
     },
     {
-      "epoch": 2.1315789473684212,
-      "grad_norm": 5.178999900817871,
+      "epoch": 1.131578947368421,
+      "grad_norm": 5.348465442657471,
       "learning_rate": 3.382352941176471e-05,
-      "loss": 1.2351,
+      "loss": 1.3201,
       "step": 30
     },
     {
-      "epoch": 2.1315789473684212,
-      "eval_accuracy": 0.27992633517495397,
-      "eval_loss": 1.5890227556228638,
-      "eval_runtime": 1917.7799,
-      "eval_samples_per_second": 0.283,
-      "eval_steps_per_second": 0.018,
-      "step": 30
-    },
-    {
-      "epoch": 3.1315789473684212,
-      "grad_norm": 6.05620813369751,
+      "epoch": 1.263157894736842,
+      "grad_norm": 8.077356338500977,
       "learning_rate": 2.647058823529412e-05,
-      "loss": 1.1635,
+      "loss": 1.2942,
       "step": 40
     },
     {
-      "epoch": 3.1315789473684212,
-      "eval_accuracy": 0.29097605893186,
-      "eval_loss": 1.3829658031463623,
-      "eval_runtime": 1908.1474,
-      "eval_samples_per_second": 0.285,
-      "eval_steps_per_second": 0.018,
+      "epoch": 1.263157894736842,
+      "eval_accuracy": 0.5119705340699816,
+      "eval_loss": 1.2199088335037231,
+      "eval_runtime": 1935.0816,
+      "eval_samples_per_second": 0.281,
+      "eval_steps_per_second": 0.035,
       "step": 40
     },
     {
-      "epoch": 4.131578947368421,
-      "grad_norm": 6.08853006362915,
+      "epoch": 2.1315789473684212,
+      "grad_norm": 7.384807109832764,
       "learning_rate": 1.9117647058823528e-05,
-      "loss": 1.0374,
-      "step": 50
-    },
-    {
-      "epoch": 4.131578947368421,
-      "eval_accuracy": 0.3001841620626151,
-      "eval_loss": 1.3681721687316895,
-      "eval_runtime": 1920.2107,
-      "eval_samples_per_second": 0.283,
-      "eval_steps_per_second": 0.018,
+      "loss": 1.0568,
       "step": 50
     },
     {
-      "epoch": 5.131578947368421,
-      "grad_norm": 8.272621154785156,
+      "epoch": 2.263157894736842,
+      "grad_norm": 7.400702953338623,
       "learning_rate": 1.1764705882352942e-05,
-      "loss": 0.9699,
+      "loss": 1.2008,
       "step": 60
     },
     {
-      "epoch": 5.131578947368421,
-      "eval_accuracy": 0.5322283609576427,
-      "eval_loss": 1.2128217220306396,
-      "eval_runtime": 1939.0306,
-      "eval_samples_per_second": 0.28,
-      "eval_steps_per_second": 0.018,
+      "epoch": 2.263157894736842,
+      "eval_accuracy": 0.5267034990791897,
+      "eval_loss": 1.1846799850463867,
+      "eval_runtime": 1919.9448,
+      "eval_samples_per_second": 0.283,
+      "eval_steps_per_second": 0.035,
       "step": 60
     },
     {
-      "epoch": 6.131578947368421,
-      "grad_norm": 6.870037078857422,
+      "epoch": 3.1315789473684212,
+      "grad_norm": 8.256953239440918,
       "learning_rate": 4.411764705882353e-06,
-      "loss": 0.8748,
+      "loss": 1.0759,
       "step": 70
     },
     {
-      "epoch": 6.131578947368421,
-      "eval_accuracy": 0.5561694290976059,
-      "eval_loss": 1.0850194692611694,
-      "eval_runtime": 1915.7845,
-      "eval_samples_per_second": 0.283,
-      "eval_steps_per_second": 0.018,
-      "step": 70
-    },
-    {
-      "epoch": 7.078947368421052,
-      "eval_accuracy": 0.5598526703499079,
-      "eval_loss": 1.0720747709274292,
-      "eval_runtime": 1934.2154,
-      "eval_samples_per_second": 0.281,
-      "eval_steps_per_second": 0.018,
+      "epoch": 3.2105263157894735,
+      "eval_accuracy": 0.5524861878453039,
+      "eval_loss": 1.04552161693573,
+      "eval_runtime": 1926.4299,
+      "eval_samples_per_second": 0.282,
+      "eval_steps_per_second": 0.035,
       "step": 76
     },
     {
-      "epoch": 7.078947368421052,
+      "epoch": 3.2105263157894735,
       "step": 76,
-      "total_flos": 1.4890870230147072e+18,
-      "train_loss": 1.143978382411756,
-      "train_runtime": 20019.1589,
+      "total_flos": 7.46412658398167e+17,
+      "train_loss": 1.275582376279329,
+      "train_runtime": 10041.2113,
       "train_samples_per_second": 0.061,
-      "train_steps_per_second": 0.004
+      "train_steps_per_second": 0.008
     },
     {
-      "epoch": 7.078947368421052,
-      "eval_accuracy": 0.7972027972027972,
-      "eval_loss": 0.7926375269889832,
-      "eval_runtime": 1534.4439,
-      "eval_samples_per_second": 0.28,
-      "eval_steps_per_second": 0.018,
+      "epoch": 3.2105263157894735,
+      "eval_accuracy": 0.5407925407925408,
+      "eval_loss": 0.993989884853363,
+      "eval_runtime": 1508.347,
+      "eval_samples_per_second": 0.284,
+      "eval_steps_per_second": 0.036,
       "step": 76
     },
     {
-      "epoch": 7.078947368421052,
-      "eval_accuracy": 0.7972027972027972,
-      "eval_loss": 0.7915867567062378,
-      "eval_runtime": 1521.5044,
-      "eval_samples_per_second": 0.282,
-      "eval_steps_per_second": 0.018,
+      "epoch": 3.2105263157894735,
+      "eval_accuracy": 0.5407925407925408,
+      "eval_loss": 0.9939900040626526,
+      "eval_runtime": 1531.0787,
+      "eval_samples_per_second": 0.28,
+      "eval_steps_per_second": 0.035,
       "step": 76
     }
   ],
@@ -174,8 +138,8 @@
       "attributes": {}
     }
   },
-  "total_flos": 1.4890870230147072e+18,
-  "train_batch_size": 16,
+  "total_flos": 7.46412658398167e+17,
+  "train_batch_size": 8,
   "trial_name": null,
   "trial_params": null
 }