🍻 cheers

all_results.json

@@ -1,12 +1,12 @@
 {
-    "epoch": 7.0,
-    "eval_accuracy": 0.7116727750501348,
-    "eval_loss": 0.9854417443275452,
-    "eval_runtime": 153.6785,
-    "eval_samples_per_second": 188.198,
-    "eval_steps_per_second": 23.53,
-    "train_loss": 0.6310260885072734,
-    "train_runtime": 24479.9902,
-    "train_samples_per_second": 66.162,
-    "train_steps_per_second": 2.068
+    "epoch": 3.0,
+    "eval_accuracy": 0.7164442293064104,
+    "eval_loss": 0.9865725040435791,
+    "eval_runtime": 151.7517,
+    "eval_samples_per_second": 190.588,
+    "eval_steps_per_second": 23.828,
+    "train_loss": 0.9066727288788254,
+    "train_runtime": 10400.7433,
+    "train_samples_per_second": 66.739,
+    "train_steps_per_second": 2.086
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 7.0,
-    "eval_accuracy": 0.7116727750501348,
-    "eval_loss": 0.9854417443275452,
-    "eval_runtime": 153.6785,
-    "eval_samples_per_second": 188.198,
-    "eval_steps_per_second": 23.53
+    "epoch": 3.0,
+    "eval_accuracy": 0.7164442293064104,
+    "eval_loss": 0.9865725040435791,
+    "eval_runtime": 151.7517,
+    "eval_samples_per_second": 190.588,
+    "eval_steps_per_second": 23.828
 }

runs/Jun04_16-32-48_adrian-development/events.out.tfevents.1685899738.adrian-development.13219.2

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2021eee5433a1ae5bd04475acb95834443fdd524cf9ea53c7ef3272e545953d6
+size 369

train_results.json

@@ -1,7 +1,7 @@
 {
-    "epoch": 7.0,
-    "train_loss": 0.6310260885072734,
-    "train_runtime": 24479.9902,
-    "train_samples_per_second": 66.162,
-    "train_steps_per_second": 2.068
+    "epoch": 3.0,
+    "train_loss": 0.9066727288788254,
+    "train_runtime": 10400.7433,
+    "train_samples_per_second": 66.739,
+    "train_steps_per_second": 2.086
 }

trainer_state.json

@@ -1,775 +1,340 @@
 {
-  "best_metric": 0.9854417443275452,
-  "best_model_checkpoint": "Transformers/vit-base-clothing-leafs-example-full-simple/checkpoint-14000",
-  "epoch": 7.0,
-  "global_step": 50617,
+  "best_metric": 0.9865725040435791,
+  "best_model_checkpoint": "Transformers/vit-base-clothing-leafs-example-full-simple/checkpoint-21000",
+  "epoch": 3.0,
+  "global_step": 21693,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
       "epoch": 0.14,
-      "learning_rate": 4.9012189580575696e-05,
-      "loss": 1.7292,
+      "learning_rate": 9.539482782464389e-05,
+      "loss": 1.558,
       "step": 1000
     },
     {
       "epoch": 0.14,
-      "eval_accuracy": 0.6539312633981053,
-      "eval_loss": 1.3212652206420898,
-      "eval_runtime": 154.4552,
-      "eval_samples_per_second": 187.252,
-      "eval_steps_per_second": 23.411,
+      "eval_accuracy": 0.6426249913560611,
+      "eval_loss": 1.2684125900268555,
+      "eval_runtime": 160.0475,
+      "eval_samples_per_second": 180.709,
+      "eval_steps_per_second": 22.593,
       "step": 1000
     },
     {
       "epoch": 0.28,
-      "learning_rate": 4.8025366971570815e-05,
-      "loss": 1.2308,
+      "learning_rate": 9.078504586733047e-05,
+      "loss": 1.2184,
       "step": 2000
     },
     {
       "epoch": 0.28,
-      "eval_accuracy": 0.6633012931332549,
-      "eval_loss": 1.198843002319336,
-      "eval_runtime": 151.3412,
-      "eval_samples_per_second": 191.105,
-      "eval_steps_per_second": 23.893,
+      "eval_accuracy": 0.6658944747942742,
+      "eval_loss": 1.1550828218460083,
+      "eval_runtime": 153.2478,
+      "eval_samples_per_second": 188.727,
+      "eval_steps_per_second": 23.596,
       "step": 2000
     },
     {
       "epoch": 0.41,
-      "learning_rate": 4.7037556552146516e-05,
-      "loss": 1.1582,
+      "learning_rate": 8.617526391001706e-05,
+      "loss": 1.1427,
       "step": 3000
     },
     {
       "epoch": 0.41,
-      "eval_accuracy": 0.6818684738261531,
-      "eval_loss": 1.1182668209075928,
-      "eval_runtime": 152.2609,
-      "eval_samples_per_second": 189.95,
-      "eval_steps_per_second": 23.749,
+      "eval_accuracy": 0.6677961413456884,
+      "eval_loss": 1.1250686645507812,
+      "eval_runtime": 152.762,
+      "eval_samples_per_second": 189.327,
+      "eval_steps_per_second": 23.671,
       "step": 3000
     },
     {
       "epoch": 0.55,
-      "learning_rate": 4.604974613272221e-05,
-      "loss": 1.109,
+      "learning_rate": 8.157009173466096e-05,
+      "loss": 1.1118,
       "step": 4000
     },
     {
       "epoch": 0.55,
-      "eval_accuracy": 0.6853952008851394,
-      "eval_loss": 1.0872125625610352,
-      "eval_runtime": 151.3068,
-      "eval_samples_per_second": 191.148,
-      "eval_steps_per_second": 23.898,
+      "eval_accuracy": 0.6727750501348454,
+      "eval_loss": 1.111645221710205,
+      "eval_runtime": 152.8701,
+      "eval_samples_per_second": 189.193,
+      "eval_steps_per_second": 23.654,
       "step": 4000
     },
     {
       "epoch": 0.69,
-      "learning_rate": 4.5062923523717335e-05,
-      "loss": 1.0772,
+      "learning_rate": 7.696030977734754e-05,
+      "loss": 1.0768,
       "step": 5000
     },
     {
       "epoch": 0.69,
-      "eval_accuracy": 0.6994329576101238,
-      "eval_loss": 1.0495002269744873,
-      "eval_runtime": 150.7718,
-      "eval_samples_per_second": 191.826,
-      "eval_steps_per_second": 23.983,
+      "eval_accuracy": 0.6808657769172256,
+      "eval_loss": 1.0770487785339355,
+      "eval_runtime": 153.4667,
+      "eval_samples_per_second": 188.458,
+      "eval_steps_per_second": 23.562,
       "step": 5000
     },
     {
       "epoch": 0.83,
-      "learning_rate": 4.407511310429303e-05,
-      "loss": 1.0371,
+      "learning_rate": 7.235513760199143e-05,
+      "loss": 1.0601,
       "step": 6000
     },
     {
       "epoch": 0.83,
-      "eval_accuracy": 0.6932093216236774,
-      "eval_loss": 1.05511474609375,
-      "eval_runtime": 151.6025,
-      "eval_samples_per_second": 190.775,
-      "eval_steps_per_second": 23.852,
+      "eval_accuracy": 0.6890256552105664,
+      "eval_loss": 1.054029107093811,
+      "eval_runtime": 153.5426,
+      "eval_samples_per_second": 188.365,
+      "eval_steps_per_second": 23.55,
       "step": 6000
     },
     {
       "epoch": 0.97,
-      "learning_rate": 4.308829049528815e-05,
-      "loss": 1.0287,
+      "learning_rate": 6.774535564467801e-05,
+      "loss": 1.0326,
       "step": 7000
     },
     {
       "epoch": 0.97,
-      "eval_accuracy": 0.7008159878293341,
-      "eval_loss": 1.0264155864715576,
-      "eval_runtime": 152.8647,
-      "eval_samples_per_second": 189.2,
-      "eval_steps_per_second": 23.655,
+      "eval_accuracy": 0.6928981398243551,
+      "eval_loss": 1.0408824682235718,
+      "eval_runtime": 153.2325,
+      "eval_samples_per_second": 188.746,
+      "eval_steps_per_second": 23.598,
       "step": 7000
     },
     {
       "epoch": 1.11,
-      "learning_rate": 4.210048007586384e-05,
-      "loss": 0.9387,
+      "learning_rate": 6.31401834693219e-05,
+      "loss": 0.9398,
       "step": 8000
     },
     {
       "epoch": 1.11,
-      "eval_accuracy": 0.7021644422930641,
-      "eval_loss": 1.025620937347412,
-      "eval_runtime": 152.5576,
-      "eval_samples_per_second": 189.581,
-      "eval_steps_per_second": 23.703,
+      "eval_accuracy": 0.6950764124196114,
+      "eval_loss": 1.0343307256698608,
+      "eval_runtime": 152.2,
+      "eval_samples_per_second": 190.026,
+      "eval_steps_per_second": 23.758,
       "step": 8000
     },
     {
       "epoch": 1.24,
-      "learning_rate": 4.111464527727839e-05,
-      "loss": 0.9005,
+      "learning_rate": 5.853040151200848e-05,
+      "loss": 0.8986,
       "step": 9000
     },
     {
       "epoch": 1.24,
-      "eval_accuracy": 0.7013346241615379,
-      "eval_loss": 1.0224932432174683,
-      "eval_runtime": 151.9509,
-      "eval_samples_per_second": 190.338,
-      "eval_steps_per_second": 23.797,
+      "eval_accuracy": 0.6950764124196114,
+      "eval_loss": 1.0352978706359863,
+      "eval_runtime": 152.4062,
+      "eval_samples_per_second": 189.769,
+      "eval_steps_per_second": 23.726,
       "step": 9000
     },
     {
       "epoch": 1.38,
-      "learning_rate": 4.012683485785408e-05,
-      "loss": 0.8957,
+      "learning_rate": 5.392061955469506e-05,
+      "loss": 0.8883,
       "step": 10000
     },
     {
       "epoch": 1.38,
-      "eval_accuracy": 0.6993638060991633,
-      "eval_loss": 1.0279977321624756,
-      "eval_runtime": 150.8357,
-      "eval_samples_per_second": 191.745,
-      "eval_steps_per_second": 23.973,
+      "eval_accuracy": 0.7006431090519328,
+      "eval_loss": 1.013888955116272,
+      "eval_runtime": 151.7157,
+      "eval_samples_per_second": 190.633,
+      "eval_steps_per_second": 23.834,
       "step": 10000
     },
     {
       "epoch": 1.52,
-      "learning_rate": 3.913902443842978e-05,
-      "loss": 0.8924,
+      "learning_rate": 4.9315447379338956e-05,
+      "loss": 0.8731,
       "step": 11000
     },
     {
       "epoch": 1.52,
-      "eval_accuracy": 0.7037203512896757,
-      "eval_loss": 1.016230821609497,
-      "eval_runtime": 151.3023,
-      "eval_samples_per_second": 191.154,
-      "eval_steps_per_second": 23.899,
+      "eval_accuracy": 0.7063826844616555,
+      "eval_loss": 0.9994178414344788,
+      "eval_runtime": 153.225,
+      "eval_samples_per_second": 188.755,
+      "eval_steps_per_second": 23.599,
       "step": 11000
     },
     {
       "epoch": 1.66,
-      "learning_rate": 3.8151214019005474e-05,
-      "loss": 0.8934,
+      "learning_rate": 4.470566542202554e-05,
+      "loss": 0.8752,
       "step": 12000
     },
     {
       "epoch": 1.66,
-      "eval_accuracy": 0.7084918055459511,
-      "eval_loss": 1.0030243396759033,
-      "eval_runtime": 151.4775,
-      "eval_samples_per_second": 190.933,
-      "eval_steps_per_second": 23.872,
+      "eval_accuracy": 0.7020261392711431,
+      "eval_loss": 1.004809856414795,
+      "eval_runtime": 154.2081,
+      "eval_samples_per_second": 187.552,
+      "eval_steps_per_second": 23.449,
       "step": 12000
     },
     {
       "epoch": 1.8,
-      "learning_rate": 3.7164391410000594e-05,
-      "loss": 0.8959,
+      "learning_rate": 4.010049324666943e-05,
+      "loss": 0.8579,
       "step": 13000
     },
     {
       "epoch": 1.8,
-      "eval_accuracy": 0.7021298665375838,
-      "eval_loss": 1.017107367515564,
-      "eval_runtime": 151.0427,
-      "eval_samples_per_second": 191.482,
-      "eval_steps_per_second": 23.94,
+      "eval_accuracy": 0.7090795933891155,
+      "eval_loss": 0.9911579489707947,
+      "eval_runtime": 153.7146,
+      "eval_samples_per_second": 188.154,
+      "eval_steps_per_second": 23.524,
       "step": 13000
     },
     {
       "epoch": 1.94,
-      "learning_rate": 3.6176580990576294e-05,
-      "loss": 0.8927,
+      "learning_rate": 3.5490711289356015e-05,
+      "loss": 0.864,
       "step": 14000
     },
     {
       "epoch": 1.94,
-      "eval_accuracy": 0.7116727750501348,
-      "eval_loss": 0.9854417443275452,
-      "eval_runtime": 151.5533,
-      "eval_samples_per_second": 190.837,
-      "eval_steps_per_second": 23.86,
+      "eval_accuracy": 0.7096328054767996,
+      "eval_loss": 0.9869005084037781,
+      "eval_runtime": 153.5275,
+      "eval_samples_per_second": 188.383,
+      "eval_steps_per_second": 23.553,
       "step": 14000
     },
     {
       "epoch": 2.07,
-      "learning_rate": 3.518877057115199e-05,
-      "loss": 0.8107,
+      "learning_rate": 3.08809293320426e-05,
+      "loss": 0.7798,
       "step": 15000
     },
     {
       "epoch": 2.07,
-      "eval_accuracy": 0.701887836249222,
-      "eval_loss": 1.0297770500183105,
-      "eval_runtime": 151.6336,
-      "eval_samples_per_second": 190.736,
-      "eval_steps_per_second": 23.847,
+      "eval_accuracy": 0.7094253509439181,
+      "eval_loss": 1.0022022724151611,
+      "eval_runtime": 153.635,
+      "eval_samples_per_second": 188.251,
+      "eval_steps_per_second": 23.536,
       "step": 15000
     },
     {
       "epoch": 2.21,
-      "learning_rate": 3.420194796214711e-05,
-      "loss": 0.7442,
+      "learning_rate": 2.6271147374729177e-05,
+      "loss": 0.6883,
       "step": 16000
     },
     {
       "epoch": 2.21,
-      "eval_accuracy": 0.707523684392504,
-      "eval_loss": 1.0102756023406982,
-      "eval_runtime": 154.4177,
-      "eval_samples_per_second": 187.297,
-      "eval_steps_per_second": 23.417,
+      "eval_accuracy": 0.7089067146117143,
+      "eval_loss": 1.00809645652771,
+      "eval_runtime": 154.6119,
+      "eval_samples_per_second": 187.062,
+      "eval_steps_per_second": 23.388,
       "step": 16000
     },
     {
       "epoch": 2.35,
-      "learning_rate": 3.32141375427228e-05,
-      "loss": 0.7421,
+      "learning_rate": 2.1661365417415756e-05,
+      "loss": 0.67,
       "step": 17000
     },
     {
       "epoch": 2.35,
-      "eval_accuracy": 0.708595532812392,
-      "eval_loss": 1.0141746997833252,
-      "eval_runtime": 154.77,
-      "eval_samples_per_second": 186.871,
-      "eval_steps_per_second": 23.364,
+      "eval_accuracy": 0.7111195629624507,
+      "eval_loss": 1.0066086053848267,
+      "eval_runtime": 152.9644,
+      "eval_samples_per_second": 189.077,
+      "eval_steps_per_second": 23.639,
       "step": 17000
     },
     {
       "epoch": 2.49,
-      "learning_rate": 3.222731493371792e-05,
-      "loss": 0.7311,
+      "learning_rate": 1.705619324205965e-05,
+      "loss": 0.678,
       "step": 18000
     },
     {
       "epoch": 2.49,
-      "eval_accuracy": 0.7110504114514902,
-      "eval_loss": 1.01434326171875,
-      "eval_runtime": 153.512,
-      "eval_samples_per_second": 188.402,
-      "eval_steps_per_second": 23.555,
+      "eval_accuracy": 0.7131595325357859,
+      "eval_loss": 0.9969001412391663,
+      "eval_runtime": 153.5427,
+      "eval_samples_per_second": 188.365,
+      "eval_steps_per_second": 23.55,
       "step": 18000
     },
     {
       "epoch": 2.63,
-      "learning_rate": 3.1239504514293614e-05,
-      "loss": 0.7362,
+      "learning_rate": 1.2446411284746232e-05,
+      "loss": 0.6701,
       "step": 19000
     },
     {
       "epoch": 2.63,
-      "eval_accuracy": 0.709943987276122,
-      "eval_loss": 1.0113046169281006,
-      "eval_runtime": 153.941,
-      "eval_samples_per_second": 187.877,
-      "eval_steps_per_second": 23.49,
+      "eval_accuracy": 0.7133324113131871,
+      "eval_loss": 0.9977221488952637,
+      "eval_runtime": 153.0562,
+      "eval_samples_per_second": 188.963,
+      "eval_steps_per_second": 23.625,
       "step": 19000
     },
     {
       "epoch": 2.77,
-      "learning_rate": 3.0252681905288737e-05,
-      "loss": 0.7392,
+      "learning_rate": 7.836629327432812e-06,
+      "loss": 0.6652,
       "step": 20000
     },
     {
       "epoch": 2.77,
-      "eval_accuracy": 0.7046193209321624,
-      "eval_loss": 1.03049898147583,
-      "eval_runtime": 154.7774,
-      "eval_samples_per_second": 186.862,
-      "eval_steps_per_second": 23.363,
+      "eval_accuracy": 0.7144388354885555,
+      "eval_loss": 0.9937713742256165,
+      "eval_runtime": 152.7221,
+      "eval_samples_per_second": 189.377,
+      "eval_steps_per_second": 23.677,
       "step": 20000
     },
     {
       "epoch": 2.9,
-      "learning_rate": 2.9264871485864437e-05,
-      "loss": 0.7515,
+      "learning_rate": 3.226847370119394e-06,
+      "loss": 0.6588,
       "step": 21000
     },
     {
       "epoch": 2.9,
-      "eval_accuracy": 0.7069013207938594,
-      "eval_loss": 1.0068926811218262,
-      "eval_runtime": 153.9311,
-      "eval_samples_per_second": 187.889,
-      "eval_steps_per_second": 23.491,
+      "eval_accuracy": 0.7164442293064104,
+      "eval_loss": 0.9865725040435791,
+      "eval_runtime": 151.9597,
+      "eval_samples_per_second": 190.327,
+      "eval_steps_per_second": 23.796,
       "step": 21000
     },
     {
-      "epoch": 3.04,
-      "learning_rate": 2.827706106644013e-05,
-      "loss": 0.6945,
-      "step": 22000
-    },
-    {
-      "epoch": 3.04,
-      "eval_accuracy": 0.7095636539658392,
-      "eval_loss": 1.0314772129058838,
-      "eval_runtime": 151.7505,
-      "eval_samples_per_second": 190.589,
-      "eval_steps_per_second": 23.829,
-      "step": 22000
-    },
-    {
-      "epoch": 3.18,
-      "learning_rate": 2.7289250647015823e-05,
-      "loss": 0.5912,
-      "step": 23000
-    },
-    {
-      "epoch": 3.18,
-      "eval_accuracy": 0.7004702302745315,
-      "eval_loss": 1.0634562969207764,
-      "eval_runtime": 152.7645,
-      "eval_samples_per_second": 189.324,
-      "eval_steps_per_second": 23.67,
-      "step": 23000
-    },
-    {
-      "epoch": 3.32,
-      "learning_rate": 2.6302428038010946e-05,
-      "loss": 0.58,
-      "step": 24000
-    },
-    {
-      "epoch": 3.32,
-      "eval_accuracy": 0.7049996542424452,
-      "eval_loss": 1.0592378377914429,
-      "eval_runtime": 153.0117,
-      "eval_samples_per_second": 189.018,
-      "eval_steps_per_second": 23.632,
-      "step": 24000
-    },
-    {
-      "epoch": 3.46,
-      "learning_rate": 2.5314617618586643e-05,
-      "loss": 0.576,
-      "step": 25000
-    },
-    {
-      "epoch": 3.46,
-      "eval_accuracy": 0.7018532604937417,
-      "eval_loss": 1.0623606443405151,
-      "eval_runtime": 154.2965,
-      "eval_samples_per_second": 187.444,
-      "eval_steps_per_second": 23.435,
-      "step": 25000
-    },
-    {
-      "epoch": 3.6,
-      "learning_rate": 2.4326807199162337e-05,
-      "loss": 0.5882,
-      "step": 26000
-    },
-    {
-      "epoch": 3.6,
-      "eval_accuracy": 0.7058986238849319,
-      "eval_loss": 1.0582630634307861,
-      "eval_runtime": 154.8374,
-      "eval_samples_per_second": 186.79,
-      "eval_steps_per_second": 23.354,
-      "step": 26000
-    },
-    {
-      "epoch": 3.73,
-      "learning_rate": 2.3338996779738033e-05,
-      "loss": 0.5867,
-      "step": 27000
-    },
-    {
-      "epoch": 3.73,
-      "eval_accuracy": 0.7076965631699053,
-      "eval_loss": 1.0526437759399414,
-      "eval_runtime": 153.5073,
-      "eval_samples_per_second": 188.408,
-      "eval_steps_per_second": 23.556,
-      "step": 27000
-    },
-    {
-      "epoch": 3.87,
-      "learning_rate": 2.2352174170733153e-05,
-      "loss": 0.593,
-      "step": 28000
-    },
-    {
-      "epoch": 3.87,
-      "eval_accuracy": 0.7047921997095636,
-      "eval_loss": 1.0647141933441162,
-      "eval_runtime": 153.5176,
-      "eval_samples_per_second": 188.395,
-      "eval_steps_per_second": 23.554,
-      "step": 28000
-    },
-    {
-      "epoch": 4.01,
-      "learning_rate": 2.136436375130885e-05,
-      "loss": 0.5745,
-      "step": 29000
-    },
-    {
-      "epoch": 4.01,
-      "eval_accuracy": 0.7044810179102413,
-      "eval_loss": 1.0814189910888672,
-      "eval_runtime": 147.6133,
-      "eval_samples_per_second": 195.931,
-      "eval_steps_per_second": 24.496,
-      "step": 29000
-    },
-    {
-      "epoch": 4.15,
-      "learning_rate": 2.0377541142303973e-05,
-      "loss": 0.4332,
-      "step": 30000
-    },
-    {
-      "epoch": 4.15,
-      "eval_accuracy": 0.7036166240232349,
-      "eval_loss": 1.1231409311294556,
-      "eval_runtime": 147.2084,
-      "eval_samples_per_second": 196.47,
-      "eval_steps_per_second": 24.564,
-      "step": 30000
-    },
-    {
-      "epoch": 4.29,
-      "learning_rate": 1.9389730722879666e-05,
-      "loss": 0.4427,
-      "step": 31000
-    },
-    {
-      "epoch": 4.29,
-      "eval_accuracy": 0.6981536546573542,
-      "eval_loss": 1.1329883337020874,
-      "eval_runtime": 156.5248,
-      "eval_samples_per_second": 184.776,
-      "eval_steps_per_second": 23.102,
-      "step": 31000
-    },
-    {
-      "epoch": 4.43,
-      "learning_rate": 1.8402908113874786e-05,
-      "loss": 0.4475,
-      "step": 32000
-    },
-    {
-      "epoch": 4.43,
-      "eval_accuracy": 0.7000207454532882,
-      "eval_loss": 1.126924991607666,
-      "eval_runtime": 155.4925,
-      "eval_samples_per_second": 186.003,
-      "eval_steps_per_second": 23.255,
-      "step": 32000
-    },
-    {
-      "epoch": 4.56,
-      "learning_rate": 1.7415097694450482e-05,
-      "loss": 0.4488,
-      "step": 33000
-    },
-    {
-      "epoch": 4.56,
-      "eval_accuracy": 0.7025101998478667,
-      "eval_loss": 1.1300145387649536,
-      "eval_runtime": 153.2715,
-      "eval_samples_per_second": 188.698,
-      "eval_steps_per_second": 23.592,
-      "step": 33000
-    },
-    {
-      "epoch": 4.7,
-      "learning_rate": 1.6428275085445602e-05,
-      "loss": 0.4431,
-      "step": 34000
-    },
-    {
-      "epoch": 4.7,
-      "eval_accuracy": 0.701887836249222,
-      "eval_loss": 1.1304574012756348,
-      "eval_runtime": 153.1234,
-      "eval_samples_per_second": 188.88,
-      "eval_steps_per_second": 23.615,
-      "step": 34000
-    },
-    {
-      "epoch": 4.84,
-      "learning_rate": 1.54404646660213e-05,
-      "loss": 0.4514,
-      "step": 35000
-    },
-    {
-      "epoch": 4.84,
-      "eval_accuracy": 0.6990872000553212,
-      "eval_loss": 1.1445332765579224,
-      "eval_runtime": 155.0001,
-      "eval_samples_per_second": 186.593,
-      "eval_steps_per_second": 23.329,
-      "step": 35000
-    },
-    {
-      "epoch": 4.98,
-      "learning_rate": 1.4452654246596994e-05,
-      "loss": 0.4427,
-      "step": 36000
-    },
-    {
-      "epoch": 4.98,
-      "eval_accuracy": 0.702821381647189,
-      "eval_loss": 1.1224807500839233,
-      "eval_runtime": 153.3684,
-      "eval_samples_per_second": 188.579,
-      "eval_steps_per_second": 23.577,
-      "step": 36000
-    },
-    {
-      "epoch": 5.12,
-      "learning_rate": 1.346681944801154e-05,
-      "loss": 0.3391,
-      "step": 37000
-    },
-    {
-      "epoch": 5.12,
-      "eval_accuracy": 0.6964248668833414,
-      "eval_loss": 1.1918517351150513,
-      "eval_runtime": 152.8672,
-      "eval_samples_per_second": 189.197,
-      "eval_steps_per_second": 23.655,
-      "step": 37000
-    },
-    {
-      "epoch": 5.26,
-      "learning_rate": 1.2479009028587233e-05,
-      "loss": 0.3222,
-      "step": 38000
-    },
-    {
-      "epoch": 5.26,
-      "eval_accuracy": 0.6971163819929466,
-      "eval_loss": 1.2108075618743896,
-      "eval_runtime": 155.4241,
-      "eval_samples_per_second": 186.084,
-      "eval_steps_per_second": 23.265,
-      "step": 38000
-    },
-    {
-      "epoch": 5.39,
-      "learning_rate": 1.149119860916293e-05,
-      "loss": 0.315,
-      "step": 39000
-    },
-    {
-      "epoch": 5.39,
-      "eval_accuracy": 0.6985685637231174,
-      "eval_loss": 1.2174683809280396,
-      "eval_runtime": 154.5842,
-      "eval_samples_per_second": 187.095,
-      "eval_steps_per_second": 23.392,
-      "step": 39000
-    },
-    {
-      "epoch": 5.53,
-      "learning_rate": 1.0503388189738627e-05,
-      "loss": 0.3307,
-      "step": 40000
-    },
-    {
-      "epoch": 5.53,
-      "eval_accuracy": 0.6943848973100062,
-      "eval_loss": 1.219739556312561,
-      "eval_runtime": 155.288,
-      "eval_samples_per_second": 186.247,
-      "eval_steps_per_second": 23.286,
-      "step": 40000
-    },
-    {
-      "epoch": 5.67,
-      "learning_rate": 9.516565580733746e-06,
-      "loss": 0.3187,
-      "step": 41000
-    },
-    {
-      "epoch": 5.67,
-      "eval_accuracy": 0.6988105940114792,
-      "eval_loss": 1.228104591369629,
-      "eval_runtime": 154.3425,
-      "eval_samples_per_second": 187.388,
-      "eval_steps_per_second": 23.428,
-      "step": 41000
-    },
-    {
-      "epoch": 5.81,
-      "learning_rate": 8.528755161309441e-06,
-      "loss": 0.3327,
-      "step": 42000
-    },
-    {
-      "epoch": 5.81,
-      "eval_accuracy": 0.6966668971717032,
-      "eval_loss": 1.2378689050674438,
-      "eval_runtime": 147.7901,
-      "eval_samples_per_second": 195.696,
-      "eval_steps_per_second": 24.467,
-      "step": 42000
-    },
-    {
-      "epoch": 5.95,
-      "learning_rate": 7.540944741885138e-06,
-      "loss": 0.3213,
-      "step": 43000
-    },
-    {
-      "epoch": 5.95,
-      "eval_accuracy": 0.6972201092593873,
-      "eval_loss": 1.2366853952407837,
-      "eval_runtime": 155.5771,
-      "eval_samples_per_second": 185.901,
-      "eval_steps_per_second": 23.242,
-      "step": 43000
-    },
-    {
-      "epoch": 6.08,
-      "learning_rate": 6.553134322460834e-06,
-      "loss": 0.2688,
-      "step": 44000
-    },
-    {
-      "epoch": 6.08,
-      "eval_accuracy": 0.6940737155106839,
-      "eval_loss": 1.2731064558029175,
-      "eval_runtime": 155.6374,
-      "eval_samples_per_second": 185.829,
-      "eval_steps_per_second": 23.233,
-      "step": 44000
-    },
-    {
-      "epoch": 6.22,
-      "learning_rate": 5.56532390303653e-06,
-      "loss": 0.2395,
-      "step": 45000
-    },
-    {
-      "epoch": 6.22,
-      "eval_accuracy": 0.696632321416223,
-      "eval_loss": 1.2903701066970825,
-      "eval_runtime": 153.3038,
-      "eval_samples_per_second": 188.658,
-      "eval_steps_per_second": 23.587,
-      "step": 45000
-    },
-    {
-      "epoch": 6.36,
-      "learning_rate": 4.5785012940316495e-06,
-      "loss": 0.2407,
-      "step": 46000
-    },
-    {
-      "epoch": 6.36,
-      "eval_accuracy": 0.6951455639305719,
-      "eval_loss": 1.2933955192565918,
-      "eval_runtime": 155.5305,
-      "eval_samples_per_second": 185.957,
-      "eval_steps_per_second": 23.249,
-      "step": 46000
-    },
-    {
-      "epoch": 6.5,
-      "learning_rate": 3.5906908746073454e-06,
-      "loss": 0.2312,
-      "step": 47000
-    },
-    {
-      "epoch": 6.5,
-      "eval_accuracy": 0.6955258972408547,
-      "eval_loss": 1.2975932359695435,
-      "eval_runtime": 155.813,
-      "eval_samples_per_second": 185.62,
-      "eval_steps_per_second": 23.207,
-      "step": 47000
-    },
-    {
-      "epoch": 6.64,
-      "learning_rate": 2.6028804551830413e-06,
-      "loss": 0.2369,
-      "step": 48000
-    },
-    {
-      "epoch": 6.64,
-      "eval_accuracy": 0.6934859276675195,
-      "eval_loss": 1.3018196821212769,
-      "eval_runtime": 150.7052,
-      "eval_samples_per_second": 191.911,
-      "eval_steps_per_second": 23.994,
-      "step": 48000
-    },
-    {
-      "epoch": 6.78,
-      "learning_rate": 1.6170456565975861e-06,
-      "loss": 0.2317,
-      "step": 49000
-    },
-    {
-      "epoch": 6.78,
-      "eval_accuracy": 0.6949381093976903,
-      "eval_loss": 1.3031286001205444,
-      "eval_runtime": 154.6116,
-      "eval_samples_per_second": 187.062,
-      "eval_steps_per_second": 23.388,
-      "step": 49000
-    },
-    {
-      "epoch": 6.91,
-      "learning_rate": 6.292352371732817e-07,
-      "loss": 0.2395,
-      "step": 50000
-    },
-    {
-      "epoch": 6.91,
-      "eval_accuracy": 0.6957679275292166,
-      "eval_loss": 1.3016529083251953,
-      "eval_runtime": 153.6886,
-      "eval_samples_per_second": 188.186,
-      "eval_steps_per_second": 23.528,
-      "step": 50000
-    },
-    {
-      "epoch": 7.0,
-      "step": 50617,
-      "total_flos": 1.2555854407514107e+20,
-      "train_loss": 0.6310260885072734,
-      "train_runtime": 24479.9902,
-      "train_samples_per_second": 66.162,
-      "train_steps_per_second": 2.068
+      "epoch": 3.0,
+      "step": 21693,
+      "total_flos": 5.381080460363188e+19,
+      "train_loss": 0.9066727288788254,
+      "train_runtime": 10400.7433,
+      "train_samples_per_second": 66.739,
+      "train_steps_per_second": 2.086
     }
   ],
-  "max_steps": 50617,
-  "num_train_epochs": 7,
-  "total_flos": 1.2555854407514107e+20,
+  "max_steps": 21693,
+  "num_train_epochs": 3,
+  "total_flos": 5.381080460363188e+19,
   "trial_name": null,
   "trial_params": null
 }