End of training

all_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 4.180952380952381,
-    "eval_accuracy": 0.8395061728395061,
-    "eval_loss": 0.38502195477485657,
-    "eval_runtime": 14.6842,
-    "eval_samples_per_second": 11.032,
-    "eval_steps_per_second": 1.43
+    "epoch": 4.184905660377359,
+    "eval_accuracy": 0.7978142076502732,
+    "eval_loss": 0.5372045040130615,
+    "eval_runtime": 17.466,
+    "eval_samples_per_second": 10.478,
+    "eval_steps_per_second": 1.317
 }

runs/Oct29_18-26-07_7c56bb07786b/events.out.tfevents.1730227090.7c56bb07786b.1505.1

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6690d3d3d8135b37dc4a7d03b3914f43d99d258eb1e258b70f5a5832c983c991
-size 411
+oid sha256:edabf379647a0aee3e379ea584cc632455901091857770579f0e324bc7853741
+size 734

test_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 4.180952380952381,
-    "eval_accuracy": 0.8395061728395061,
-    "eval_loss": 0.38502195477485657,
-    "eval_runtime": 14.6842,
-    "eval_samples_per_second": 11.032,
-    "eval_steps_per_second": 1.43
+    "epoch": 4.184905660377359,
+    "eval_accuracy": 0.7978142076502732,
+    "eval_loss": 0.5372045040130615,
+    "eval_runtime": 17.466,
+    "eval_samples_per_second": 10.478,
+    "eval_steps_per_second": 1.317
 }

trainer_state.json

@@ -1,235 +1,270 @@
 {
-  "best_metric": 0.9066666666666666,
-  "best_model_checkpoint": "videomae-base-finetuned-ucf101-subset/checkpoint-210",
-  "epoch": 4.180952380952381,
+  "best_metric": 0.8452380952380952,
+  "best_model_checkpoint": "videomae-base-finetuned-ucf101-subset/checkpoint-265",
+  "epoch": 4.184905660377359,
   "eval_steps": 500,
-  "global_step": 210,
+  "global_step": 265,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.047619047619047616,
-      "grad_norm": 7.849704265594482,
-      "learning_rate": 2.380952380952381e-05,
-      "loss": 2.3976,
+      "epoch": 0.03773584905660377,
+      "grad_norm": 8.033324241638184,
+      "learning_rate": 1.8518518518518518e-05,
+      "loss": 2.7609,
       "step": 10
     },
     {
-      "epoch": 0.09523809523809523,
-      "grad_norm": 7.76956033706665,
-      "learning_rate": 4.761904761904762e-05,
-      "loss": 2.3357,
+      "epoch": 0.07547169811320754,
+      "grad_norm": 7.882028102874756,
+      "learning_rate": 3.7037037037037037e-05,
+      "loss": 2.6513,
       "step": 20
     },
     {
-      "epoch": 0.14285714285714285,
-      "grad_norm": 15.013307571411133,
-      "learning_rate": 4.761904761904762e-05,
-      "loss": 2.282,
+      "epoch": 0.11320754716981132,
+      "grad_norm": 8.582988739013672,
+      "learning_rate": 4.936974789915967e-05,
+      "loss": 2.5533,
       "step": 30
     },
     {
-      "epoch": 0.19047619047619047,
-      "grad_norm": 7.262270450592041,
-      "learning_rate": 4.4973544973544974e-05,
-      "loss": 2.0368,
+      "epoch": 0.1509433962264151,
+      "grad_norm": 9.038335800170898,
+      "learning_rate": 4.726890756302521e-05,
+      "loss": 2.3765,
       "step": 40
     },
     {
-      "epoch": 0.20476190476190476,
-      "eval_accuracy": 0.29333333333333333,
-      "eval_loss": 2.0273361206054688,
-      "eval_runtime": 7.264,
-      "eval_samples_per_second": 10.325,
-      "eval_steps_per_second": 1.377,
-      "step": 43
+      "epoch": 0.18867924528301888,
+      "grad_norm": 10.162276268005371,
+      "learning_rate": 4.516806722689076e-05,
+      "loss": 2.2682,
+      "step": 50
     },
     {
-      "epoch": 1.0333333333333334,
-      "grad_norm": 8.3132963180542,
-      "learning_rate": 4.232804232804233e-05,
-      "loss": 1.947,
-      "step": 50
+      "epoch": 0.2037735849056604,
+      "eval_accuracy": 0.32142857142857145,
+      "eval_loss": 2.0374372005462646,
+      "eval_runtime": 21.9106,
+      "eval_samples_per_second": 3.834,
+      "eval_steps_per_second": 0.502,
+      "step": 54
     },
     {
-      "epoch": 1.0809523809523809,
-      "grad_norm": 14.42203426361084,
-      "learning_rate": 3.968253968253968e-05,
-      "loss": 1.594,
+      "epoch": 1.0226415094339623,
+      "grad_norm": 8.538351058959961,
+      "learning_rate": 4.3067226890756305e-05,
+      "loss": 1.9532,
       "step": 60
     },
     {
-      "epoch": 1.1285714285714286,
-      "grad_norm": 11.382922172546387,
-      "learning_rate": 3.7037037037037037e-05,
-      "loss": 1.2509,
+      "epoch": 1.060377358490566,
+      "grad_norm": 8.961767196655273,
+      "learning_rate": 4.096638655462185e-05,
+      "loss": 1.5809,
       "step": 70
     },
     {
-      "epoch": 1.1761904761904762,
-      "grad_norm": 9.775425910949707,
-      "learning_rate": 3.439153439153439e-05,
-      "loss": 1.1679,
+      "epoch": 1.0981132075471698,
+      "grad_norm": 8.71930980682373,
+      "learning_rate": 3.88655462184874e-05,
+      "loss": 1.4653,
       "step": 80
     },
     {
-      "epoch": 1.2047619047619047,
-      "eval_accuracy": 0.56,
-      "eval_loss": 1.0918534994125366,
-      "eval_runtime": 7.1505,
-      "eval_samples_per_second": 10.489,
-      "eval_steps_per_second": 1.398,
-      "step": 86
-    },
-    {
-      "epoch": 2.019047619047619,
-      "grad_norm": 16.352764129638672,
-      "learning_rate": 3.1746031746031745e-05,
-      "loss": 1.0361,
+      "epoch": 1.1358490566037736,
+      "grad_norm": 11.393306732177734,
+      "learning_rate": 3.6764705882352945e-05,
+      "loss": 1.3126,
       "step": 90
     },
     {
-      "epoch": 2.066666666666667,
-      "grad_norm": 5.250826358795166,
-      "learning_rate": 2.91005291005291e-05,
-      "loss": 0.7379,
+      "epoch": 1.1735849056603773,
+      "grad_norm": 8.893975257873535,
+      "learning_rate": 3.466386554621849e-05,
+      "loss": 1.0997,
       "step": 100
     },
     {
-      "epoch": 2.1142857142857143,
-      "grad_norm": 11.715250968933105,
-      "learning_rate": 2.6455026455026456e-05,
-      "loss": 0.7289,
-      "step": 110
+      "epoch": 1.2037735849056603,
+      "eval_accuracy": 0.6666666666666666,
+      "eval_loss": 1.0029432773590088,
+      "eval_runtime": 7.8357,
+      "eval_samples_per_second": 10.72,
+      "eval_steps_per_second": 1.404,
+      "step": 108
     },
     {
-      "epoch": 2.1619047619047618,
-      "grad_norm": 16.230215072631836,
-      "learning_rate": 2.380952380952381e-05,
-      "loss": 0.5097,
-      "step": 120
+      "epoch": 2.0075471698113208,
+      "grad_norm": 8.756331443786621,
+      "learning_rate": 3.2563025210084034e-05,
+      "loss": 1.0144,
+      "step": 110
     },
     {
-      "epoch": 2.204761904761905,
-      "eval_accuracy": 0.8266666666666667,
-      "eval_loss": 0.5804122090339661,
-      "eval_runtime": 6.6096,
-      "eval_samples_per_second": 11.347,
-      "eval_steps_per_second": 1.513,
-      "step": 129
+      "epoch": 2.0452830188679245,
+      "grad_norm": 15.8713960647583,
+      "learning_rate": 3.0462184873949578e-05,
+      "loss": 0.9754,
+      "step": 120
     },
     {
-      "epoch": 3.0047619047619047,
-      "grad_norm": 18.1605224609375,
-      "learning_rate": 2.1164021164021164e-05,
-      "loss": 0.4609,
+      "epoch": 2.0830188679245283,
+      "grad_norm": 25.470157623291016,
+      "learning_rate": 2.8361344537815126e-05,
+      "loss": 1.1946,
       "step": 130
     },
     {
-      "epoch": 3.052380952380952,
-      "grad_norm": 7.160711288452148,
-      "learning_rate": 1.8518518518518518e-05,
-      "loss": 0.3043,
+      "epoch": 2.120754716981132,
+      "grad_norm": 6.735860347747803,
+      "learning_rate": 2.6260504201680674e-05,
+      "loss": 0.782,
       "step": 140
     },
     {
-      "epoch": 3.1,
-      "grad_norm": 14.182634353637695,
-      "learning_rate": 1.5873015873015872e-05,
-      "loss": 0.336,
+      "epoch": 2.158490566037736,
+      "grad_norm": 8.800786018371582,
+      "learning_rate": 2.415966386554622e-05,
+      "loss": 0.8976,
       "step": 150
     },
     {
-      "epoch": 3.1476190476190475,
-      "grad_norm": 2.916877031326294,
-      "learning_rate": 1.3227513227513228e-05,
-      "loss": 0.2992,
+      "epoch": 2.1962264150943396,
+      "grad_norm": 8.588472366333008,
+      "learning_rate": 2.2058823529411766e-05,
+      "loss": 0.6996,
       "step": 160
     },
     {
-      "epoch": 3.1952380952380954,
-      "grad_norm": 8.268482208251953,
-      "learning_rate": 1.0582010582010582e-05,
-      "loss": 0.2293,
-      "step": 170
+      "epoch": 2.2037735849056603,
+      "eval_accuracy": 0.7857142857142857,
+      "eval_loss": 0.7632536292076111,
+      "eval_runtime": 8.3664,
+      "eval_samples_per_second": 10.04,
+      "eval_steps_per_second": 1.315,
+      "step": 162
     },
     {
-      "epoch": 3.204761904761905,
-      "eval_accuracy": 0.8266666666666667,
-      "eval_loss": 0.4854774475097656,
-      "eval_runtime": 6.5331,
-      "eval_samples_per_second": 11.48,
-      "eval_steps_per_second": 1.531,
-      "step": 172
+      "epoch": 3.030188679245283,
+      "grad_norm": 7.108087539672852,
+      "learning_rate": 1.9957983193277314e-05,
+      "loss": 0.8053,
+      "step": 170
     },
     {
-      "epoch": 4.038095238095238,
-      "grad_norm": 7.389148712158203,
-      "learning_rate": 7.936507936507936e-06,
-      "loss": 0.3101,
+      "epoch": 3.0679245283018868,
+      "grad_norm": 8.610198974609375,
+      "learning_rate": 1.785714285714286e-05,
+      "loss": 0.6667,
       "step": 180
     },
     {
-      "epoch": 4.085714285714285,
-      "grad_norm": 4.5858259201049805,
-      "learning_rate": 5.291005291005291e-06,
-      "loss": 0.1873,
+      "epoch": 3.1056603773584905,
+      "grad_norm": 11.434289932250977,
+      "learning_rate": 1.5756302521008403e-05,
+      "loss": 0.6006,
       "step": 190
     },
     {
-      "epoch": 4.133333333333334,
-      "grad_norm": 2.300278425216675,
-      "learning_rate": 2.6455026455026455e-06,
-      "loss": 0.2008,
+      "epoch": 3.1433962264150943,
+      "grad_norm": 4.931293487548828,
+      "learning_rate": 1.3655462184873949e-05,
+      "loss": 0.635,
       "step": 200
     },
     {
-      "epoch": 4.180952380952381,
-      "grad_norm": 2.982243537902832,
-      "learning_rate": 0.0,
-      "loss": 0.1882,
+      "epoch": 3.181132075471698,
+      "grad_norm": 6.236601829528809,
+      "learning_rate": 1.1554621848739497e-05,
+      "loss": 0.7031,
       "step": 210
     },
     {
-      "epoch": 4.180952380952381,
-      "eval_accuracy": 0.9066666666666666,
-      "eval_loss": 0.3435481786727905,
-      "eval_runtime": 6.9489,
-      "eval_samples_per_second": 10.793,
-      "eval_steps_per_second": 1.439,
-      "step": 210
+      "epoch": 3.2037735849056603,
+      "eval_accuracy": 0.7857142857142857,
+      "eval_loss": 0.5939908027648926,
+      "eval_runtime": 7.8516,
+      "eval_samples_per_second": 10.698,
+      "eval_steps_per_second": 1.401,
+      "step": 216
     },
     {
-      "epoch": 4.180952380952381,
-      "step": 210,
-      "total_flos": 2.0736178988651643e+18,
-      "train_loss": 0.9781298257055737,
-      "train_runtime": 354.3465,
-      "train_samples_per_second": 4.741,
-      "train_steps_per_second": 0.593
+      "epoch": 4.0150943396226415,
+      "grad_norm": 5.458017349243164,
+      "learning_rate": 9.453781512605041e-06,
+      "loss": 0.5998,
+      "step": 220
     },
     {
-      "epoch": 4.180952380952381,
-      "eval_accuracy": 0.8395061728395061,
-      "eval_loss": 0.3850219249725342,
-      "eval_runtime": 60.1785,
-      "eval_samples_per_second": 2.692,
-      "eval_steps_per_second": 0.349,
-      "step": 210
+      "epoch": 4.052830188679246,
+      "grad_norm": 4.542972087860107,
+      "learning_rate": 7.3529411764705884e-06,
+      "loss": 0.5022,
+      "step": 230
     },
     {
-      "epoch": 4.180952380952381,
-      "eval_accuracy": 0.8395061728395061,
-      "eval_loss": 0.38502195477485657,
-      "eval_runtime": 14.6842,
-      "eval_samples_per_second": 11.032,
-      "eval_steps_per_second": 1.43,
-      "step": 210
+      "epoch": 4.090566037735849,
+      "grad_norm": 17.187719345092773,
+      "learning_rate": 5.252100840336135e-06,
+      "loss": 0.4509,
+      "step": 240
+    },
+    {
+      "epoch": 4.128301886792453,
+      "grad_norm": 18.115018844604492,
+      "learning_rate": 3.1512605042016808e-06,
+      "loss": 0.5752,
+      "step": 250
+    },
+    {
+      "epoch": 4.166037735849057,
+      "grad_norm": 6.622856616973877,
+      "learning_rate": 1.0504201680672271e-06,
+      "loss": 0.4078,
+      "step": 260
+    },
+    {
+      "epoch": 4.184905660377359,
+      "eval_accuracy": 0.8452380952380952,
+      "eval_loss": 0.5297083854675293,
+      "eval_runtime": 8.7988,
+      "eval_samples_per_second": 9.547,
+      "eval_steps_per_second": 1.25,
+      "step": 265
+    },
+    {
+      "epoch": 4.184905660377359,
+      "step": 265,
+      "total_flos": 2.631970050168324e+18,
+      "train_loss": 1.195451885799192,
+      "train_runtime": 648.8571,
+      "train_samples_per_second": 3.267,
+      "train_steps_per_second": 0.408
+    },
+    {
+      "epoch": 4.184905660377359,
+      "eval_accuracy": 0.7978142076502732,
+      "eval_loss": 0.5372046828269958,
+      "eval_runtime": 49.961,
+      "eval_samples_per_second": 3.663,
+      "eval_steps_per_second": 0.46,
+      "step": 265
+    },
+    {
+      "epoch": 4.184905660377359,
+      "eval_accuracy": 0.7978142076502732,
+      "eval_loss": 0.5372045040130615,
+      "eval_runtime": 17.466,
+      "eval_samples_per_second": 10.478,
+      "eval_steps_per_second": 1.317,
+      "step": 265
     }
   ],
   "logging_steps": 10,
-  "max_steps": 210,
+  "max_steps": 265,
   "num_input_tokens_seen": 0,
   "num_train_epochs": 9223372036854775807,
   "save_steps": 500,
@@ -245,7 +280,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 2.0736178988651643e+18,
+  "total_flos": 2.631970050168324e+18,
   "train_batch_size": 8,
   "trial_name": null,
   "trial_params": null