End of training

all_results.json

@@ -1,13 +1,13 @@
 {
     "epoch": 9.98,
-    "eval_accuracy": 0.8958625525946704,
-    "eval_loss": 0.31656068563461304,
-    "eval_runtime": 38.1413,
-    "eval_samples_per_second": 74.775,
-    "eval_steps_per_second": 1.18,
+    "eval_accuracy": 0.9119915848527349,
+    "eval_loss": 0.3377525210380554,
+    "eval_runtime": 37.4095,
+    "eval_samples_per_second": 76.237,
+    "eval_steps_per_second": 1.203,
     "total_flos": 6.364199987970048e+18,
-    "train_loss": 0.2624686861038208,
-    "train_runtime": 4808.7305,
-    "train_samples_per_second": 53.37,
-    "train_steps_per_second": 0.208
+    "train_loss": 0.11027517792582511,
+    "train_runtime": 4749.2843,
+    "train_samples_per_second": 54.038,
+    "train_steps_per_second": 0.211
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 9.98,
-    "eval_accuracy": 0.8958625525946704,
-    "eval_loss": 0.31656068563461304,
-    "eval_runtime": 38.1413,
-    "eval_samples_per_second": 74.775,
-    "eval_steps_per_second": 1.18
+    "eval_accuracy": 0.9119915848527349,
+    "eval_loss": 0.3377525210380554,
+    "eval_runtime": 37.4095,
+    "eval_samples_per_second": 76.237,
+    "eval_steps_per_second": 1.203
 }

runs/Apr23_22-49-44_f99a83d30fa0/events.out.tfevents.1713917390.f99a83d30fa0.34.5

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:223a0ce888c5e936e8e136c6a864c70618bcb19b4a1ddc5bea17105ef93fb3f8
+size 411

train_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 9.98,
     "total_flos": 6.364199987970048e+18,
-    "train_loss": 0.2624686861038208,
-    "train_runtime": 4808.7305,
-    "train_samples_per_second": 53.37,
-    "train_steps_per_second": 0.208
+    "train_loss": 0.11027517792582511,
+    "train_runtime": 4749.2843,
+    "train_samples_per_second": 54.038,
+    "train_steps_per_second": 0.211
 }

trainer_state.json

@@ -1,5 +1,5 @@
 {
-  "best_metric": 0.8958625525946704,
+  "best_metric": 0.9119915848527349,
   "best_model_checkpoint": "swin-tiny-patch4-window7-224-finetuned-eurosat/checkpoint-1000",
   "epoch": 9.975062344139651,
   "eval_steps": 500,
@@ -10,802 +10,802 @@
   "log_history": [
     {
       "epoch": 0.1,
-      "grad_norm": 4.312608242034912,
+      "grad_norm": 7.060407638549805,
       "learning_rate": 1e-05,
-      "loss": 0.3601,
+      "loss": 0.1774,
       "step": 10
     },
     {
       "epoch": 0.2,
-      "grad_norm": 4.751251697540283,
+      "grad_norm": 4.514194011688232,
       "learning_rate": 2e-05,
-      "loss": 0.3746,
+      "loss": 0.118,
       "step": 20
     },
     {
       "epoch": 0.3,
-      "grad_norm": 5.903082847595215,
+      "grad_norm": 5.400581359863281,
       "learning_rate": 3e-05,
-      "loss": 0.3844,
+      "loss": 0.1062,
       "step": 30
     },
     {
       "epoch": 0.4,
-      "grad_norm": 7.2845354080200195,
+      "grad_norm": 5.083698749542236,
       "learning_rate": 4e-05,
-      "loss": 0.3935,
+      "loss": 0.074,
       "step": 40
     },
     {
       "epoch": 0.5,
-      "grad_norm": 4.63520622253418,
+      "grad_norm": 2.7988944053649902,
       "learning_rate": 5e-05,
-      "loss": 0.3752,
+      "loss": 0.0639,
       "step": 50
     },
     {
       "epoch": 0.6,
-      "grad_norm": 5.532151222229004,
+      "grad_norm": 1.6099319458007812,
       "learning_rate": 6e-05,
-      "loss": 0.3769,
+      "loss": 0.0508,
       "step": 60
     },
     {
       "epoch": 0.7,
-      "grad_norm": 5.141340255737305,
+      "grad_norm": 2.6441915035247803,
       "learning_rate": 7e-05,
-      "loss": 0.3821,
+      "loss": 0.0524,
       "step": 70
     },
     {
       "epoch": 0.8,
-      "grad_norm": 6.834048748016357,
+      "grad_norm": 5.100390434265137,
       "learning_rate": 8e-05,
-      "loss": 0.3872,
+      "loss": 0.06,
       "step": 80
     },
     {
       "epoch": 0.9,
-      "grad_norm": 4.966386318206787,
+      "grad_norm": 4.311460971832275,
       "learning_rate": 9e-05,
-      "loss": 0.4001,
+      "loss": 0.0424,
       "step": 90
     },
     {
       "epoch": 1.0,
-      "grad_norm": 6.305124759674072,
+      "grad_norm": 2.94564151763916,
       "learning_rate": 0.0001,
-      "loss": 0.3657,
+      "loss": 0.0511,
       "step": 100
     },
     {
       "epoch": 1.0,
-      "eval_accuracy": 0.8369565217391305,
-      "eval_loss": 0.4743898808956146,
-      "eval_runtime": 37.4208,
-      "eval_samples_per_second": 76.214,
-      "eval_steps_per_second": 1.203,
+      "eval_accuracy": 0.8720196353436185,
+      "eval_loss": 0.5181019306182861,
+      "eval_runtime": 38.2631,
+      "eval_samples_per_second": 74.537,
+      "eval_steps_per_second": 1.176,
       "step": 100
     },
     {
       "epoch": 1.1,
-      "grad_norm": 4.9311909675598145,
+      "grad_norm": 4.313048839569092,
       "learning_rate": 9.888888888888889e-05,
-      "loss": 0.3983,
+      "loss": 0.0887,
       "step": 110
     },
     {
       "epoch": 1.2,
-      "grad_norm": 6.430706024169922,
+      "grad_norm": 5.640860557556152,
       "learning_rate": 9.777777777777778e-05,
-      "loss": 0.3911,
+      "loss": 0.084,
       "step": 120
     },
     {
       "epoch": 1.3,
-      "grad_norm": 5.201113224029541,
+      "grad_norm": 2.4771900177001953,
       "learning_rate": 9.666666666666667e-05,
-      "loss": 0.3525,
+      "loss": 0.0879,
       "step": 130
     },
     {
       "epoch": 1.4,
-      "grad_norm": 7.090891361236572,
+      "grad_norm": 7.712289810180664,
       "learning_rate": 9.555555555555557e-05,
-      "loss": 0.3367,
+      "loss": 0.1231,
       "step": 140
     },
     {
       "epoch": 1.5,
-      "grad_norm": 8.351272583007812,
+      "grad_norm": 7.235645771026611,
       "learning_rate": 9.444444444444444e-05,
-      "loss": 0.3358,
+      "loss": 0.143,
       "step": 150
     },
     {
       "epoch": 1.6,
-      "grad_norm": 5.695847511291504,
+      "grad_norm": 11.080587387084961,
       "learning_rate": 9.333333333333334e-05,
-      "loss": 0.342,
+      "loss": 0.1804,
       "step": 160
     },
     {
       "epoch": 1.7,
-      "grad_norm": 5.768601417541504,
+      "grad_norm": 4.359196662902832,
       "learning_rate": 9.222222222222223e-05,
-      "loss": 0.3539,
+      "loss": 0.1894,
       "step": 170
     },
     {
       "epoch": 1.8,
-      "grad_norm": 5.517068386077881,
+      "grad_norm": 4.214484214782715,
       "learning_rate": 9.111111111111112e-05,
-      "loss": 0.3498,
+      "loss": 0.1589,
       "step": 180
     },
     {
       "epoch": 1.9,
-      "grad_norm": 4.817891597747803,
+      "grad_norm": 6.100831508636475,
       "learning_rate": 9e-05,
-      "loss": 0.3135,
+      "loss": 0.1317,
       "step": 190
     },
     {
       "epoch": 2.0,
-      "grad_norm": 5.500814914703369,
+      "grad_norm": 4.592179298400879,
       "learning_rate": 8.888888888888889e-05,
-      "loss": 0.3018,
+      "loss": 0.1166,
       "step": 200
     },
     {
       "epoch": 2.0,
-      "eval_accuracy": 0.8492286115007013,
-      "eval_loss": 0.43779420852661133,
-      "eval_runtime": 40.6238,
-      "eval_samples_per_second": 70.205,
-      "eval_steps_per_second": 1.108,
+      "eval_accuracy": 0.8716690042075736,
+      "eval_loss": 0.4709436595439911,
+      "eval_runtime": 38.3291,
+      "eval_samples_per_second": 74.408,
+      "eval_steps_per_second": 1.174,
       "step": 200
     },
     {
       "epoch": 2.09,
-      "grad_norm": 7.610060691833496,
+      "grad_norm": 9.383784294128418,
       "learning_rate": 8.777777777777778e-05,
-      "loss": 0.3183,
+      "loss": 0.1538,
       "step": 210
     },
     {
       "epoch": 2.19,
-      "grad_norm": 6.4134135246276855,
+      "grad_norm": 5.486557960510254,
       "learning_rate": 8.666666666666667e-05,
-      "loss": 0.2902,
+      "loss": 0.1292,
       "step": 220
     },
     {
       "epoch": 2.29,
-      "grad_norm": 5.562830924987793,
+      "grad_norm": 5.093433856964111,
       "learning_rate": 8.555555555555556e-05,
-      "loss": 0.3122,
+      "loss": 0.1411,
       "step": 230
     },
     {
       "epoch": 2.39,
-      "grad_norm": 6.466961860656738,
+      "grad_norm": 6.046931266784668,
       "learning_rate": 8.444444444444444e-05,
-      "loss": 0.3168,
+      "loss": 0.1502,
       "step": 240
     },
     {
       "epoch": 2.49,
-      "grad_norm": 5.561737537384033,
+      "grad_norm": 6.943304061889648,
       "learning_rate": 8.333333333333334e-05,
-      "loss": 0.3021,
+      "loss": 0.1298,
       "step": 250
     },
     {
       "epoch": 2.59,
-      "grad_norm": 5.490133762359619,
+      "grad_norm": 4.685500621795654,
       "learning_rate": 8.222222222222222e-05,
-      "loss": 0.3186,
+      "loss": 0.1301,
       "step": 260
     },
     {
       "epoch": 2.69,
-      "grad_norm": 6.67871618270874,
+      "grad_norm": 4.650270462036133,
       "learning_rate": 8.111111111111112e-05,
-      "loss": 0.3407,
+      "loss": 0.1342,
       "step": 270
     },
     {
       "epoch": 2.79,
-      "grad_norm": 7.841400623321533,
+      "grad_norm": 8.815176010131836,
       "learning_rate": 8e-05,
-      "loss": 0.3052,
+      "loss": 0.1353,
       "step": 280
     },
     {
       "epoch": 2.89,
-      "grad_norm": 5.589080810546875,
+      "grad_norm": 6.133541584014893,
       "learning_rate": 7.88888888888889e-05,
-      "loss": 0.2989,
+      "loss": 0.1252,
       "step": 290
     },
     {
       "epoch": 2.99,
-      "grad_norm": 4.181407928466797,
+      "grad_norm": 5.344182014465332,
       "learning_rate": 7.777777777777778e-05,
-      "loss": 0.2923,
+      "loss": 0.1189,
       "step": 300
     },
     {
       "epoch": 2.99,
-      "eval_accuracy": 0.8523842917251052,
-      "eval_loss": 0.4396316409111023,
-      "eval_runtime": 38.1758,
-      "eval_samples_per_second": 74.707,
-      "eval_steps_per_second": 1.179,
+      "eval_accuracy": 0.8839410939691444,
+      "eval_loss": 0.41620370745658875,
+      "eval_runtime": 38.187,
+      "eval_samples_per_second": 74.685,
+      "eval_steps_per_second": 1.178,
       "step": 300
     },
     {
       "epoch": 3.09,
-      "grad_norm": 6.813866138458252,
+      "grad_norm": 5.3231706619262695,
       "learning_rate": 7.666666666666667e-05,
-      "loss": 0.2622,
+      "loss": 0.11,
       "step": 310
     },
     {
       "epoch": 3.19,
-      "grad_norm": 6.819497108459473,
+      "grad_norm": 6.189892292022705,
       "learning_rate": 7.555555555555556e-05,
-      "loss": 0.2816,
+      "loss": 0.1151,
       "step": 320
     },
     {
       "epoch": 3.29,
-      "grad_norm": 5.811374187469482,
+      "grad_norm": 4.102895259857178,
       "learning_rate": 7.444444444444444e-05,
-      "loss": 0.2859,
+      "loss": 0.1023,
       "step": 330
     },
     {
       "epoch": 3.39,
-      "grad_norm": 5.982662677764893,
+      "grad_norm": 4.935023784637451,
       "learning_rate": 7.333333333333333e-05,
-      "loss": 0.276,
+      "loss": 0.1298,
       "step": 340
     },
     {
       "epoch": 3.49,
-      "grad_norm": 6.501168727874756,
+      "grad_norm": 6.158393859863281,
       "learning_rate": 7.222222222222222e-05,
-      "loss": 0.2984,
+      "loss": 0.1382,
       "step": 350
     },
     {
       "epoch": 3.59,
-      "grad_norm": 5.3160529136657715,
+      "grad_norm": 4.13533878326416,
       "learning_rate": 7.111111111111112e-05,
-      "loss": 0.2907,
+      "loss": 0.1526,
       "step": 360
     },
     {
       "epoch": 3.69,
-      "grad_norm": 5.063723087310791,
+      "grad_norm": 2.808582067489624,
       "learning_rate": 7e-05,
-      "loss": 0.2864,
+      "loss": 0.1106,
       "step": 370
     },
     {
       "epoch": 3.79,
-      "grad_norm": 6.182292461395264,
+      "grad_norm": 4.059571266174316,
       "learning_rate": 6.88888888888889e-05,
-      "loss": 0.3045,
+      "loss": 0.1221,
       "step": 380
     },
     {
       "epoch": 3.89,
-      "grad_norm": 6.24503755569458,
+      "grad_norm": 4.626064777374268,
       "learning_rate": 6.777777777777778e-05,
-      "loss": 0.3198,
+      "loss": 0.1303,
       "step": 390
     },
     {
       "epoch": 3.99,
-      "grad_norm": 5.089386463165283,
+      "grad_norm": 6.2365031242370605,
       "learning_rate": 6.666666666666667e-05,
-      "loss": 0.3038,
+      "loss": 0.1306,
       "step": 400
     },
     {
       "epoch": 4.0,
-      "eval_accuracy": 0.861851332398317,
-      "eval_loss": 0.40673914551734924,
-      "eval_runtime": 38.2337,
-      "eval_samples_per_second": 74.594,
-      "eval_steps_per_second": 1.177,
+      "eval_accuracy": 0.8807854137447405,
+      "eval_loss": 0.4251408874988556,
+      "eval_runtime": 38.3771,
+      "eval_samples_per_second": 74.315,
+      "eval_steps_per_second": 1.173,
       "step": 401
     },
     {
       "epoch": 4.09,
-      "grad_norm": 5.16262149810791,
+      "grad_norm": 4.386157035827637,
       "learning_rate": 6.555555555555556e-05,
-      "loss": 0.2505,
+      "loss": 0.1149,
       "step": 410
     },
     {
       "epoch": 4.19,
-      "grad_norm": 6.915875434875488,
+      "grad_norm": 5.805418968200684,
       "learning_rate": 6.444444444444446e-05,
-      "loss": 0.2486,
+      "loss": 0.0997,
       "step": 420
     },
     {
       "epoch": 4.29,
-      "grad_norm": 5.945577621459961,
+      "grad_norm": 7.228994846343994,
       "learning_rate": 6.333333333333333e-05,
-      "loss": 0.2798,
+      "loss": 0.1262,
       "step": 430
     },
     {
       "epoch": 4.39,
-      "grad_norm": 6.409703731536865,
+      "grad_norm": 8.70606517791748,
       "learning_rate": 6.222222222222222e-05,
-      "loss": 0.2797,
+      "loss": 0.1422,
       "step": 440
     },
     {
       "epoch": 4.49,
-      "grad_norm": 9.431686401367188,
+      "grad_norm": 5.861635208129883,
       "learning_rate": 6.111111111111112e-05,
-      "loss": 0.2871,
+      "loss": 0.1128,
       "step": 450
     },
     {
       "epoch": 4.59,
-      "grad_norm": 6.450242042541504,
+      "grad_norm": 5.075900554656982,
       "learning_rate": 6e-05,
-      "loss": 0.2834,
+      "loss": 0.1142,
       "step": 460
     },
     {
       "epoch": 4.69,
-      "grad_norm": 6.083951473236084,
+      "grad_norm": 5.3067498207092285,
       "learning_rate": 5.8888888888888896e-05,
-      "loss": 0.3218,
+      "loss": 0.1359,
       "step": 470
     },
     {
       "epoch": 4.79,
-      "grad_norm": 4.859699249267578,
+      "grad_norm": 3.890895366668701,
       "learning_rate": 5.7777777777777776e-05,
-      "loss": 0.3078,
+      "loss": 0.1173,
       "step": 480
     },
     {
       "epoch": 4.89,
-      "grad_norm": 4.411416053771973,
+      "grad_norm": 5.006656646728516,
       "learning_rate": 5.666666666666667e-05,
-      "loss": 0.2685,
+      "loss": 0.1087,
       "step": 490
     },
     {
       "epoch": 4.99,
-      "grad_norm": 5.873050212860107,
+      "grad_norm": 4.467653274536133,
       "learning_rate": 5.555555555555556e-05,
-      "loss": 0.2755,
+      "loss": 0.1063,
       "step": 500
     },
     {
       "epoch": 5.0,
-      "eval_accuracy": 0.8772791023842917,
-      "eval_loss": 0.36721259355545044,
-      "eval_runtime": 38.0267,
-      "eval_samples_per_second": 75.0,
-      "eval_steps_per_second": 1.183,
+      "eval_accuracy": 0.8888499298737728,
+      "eval_loss": 0.41856977343559265,
+      "eval_runtime": 37.7302,
+      "eval_samples_per_second": 75.589,
+      "eval_steps_per_second": 1.193,
       "step": 501
     },
     {
       "epoch": 5.09,
-      "grad_norm": 5.115600109100342,
+      "grad_norm": 6.331502914428711,
       "learning_rate": 5.4444444444444446e-05,
-      "loss": 0.2579,
+      "loss": 0.1132,
       "step": 510
     },
     {
       "epoch": 5.19,
-      "grad_norm": 6.827105522155762,
+      "grad_norm": 6.853649139404297,
       "learning_rate": 5.333333333333333e-05,
-      "loss": 0.2852,
+      "loss": 0.1256,
       "step": 520
     },
     {
       "epoch": 5.29,
-      "grad_norm": 5.669814109802246,
+      "grad_norm": 6.621553421020508,
       "learning_rate": 5.222222222222223e-05,
-      "loss": 0.2786,
+      "loss": 0.121,
       "step": 530
     },
     {
       "epoch": 5.39,
-      "grad_norm": 5.156219959259033,
+      "grad_norm": 5.279974460601807,
       "learning_rate": 5.111111111111111e-05,
-      "loss": 0.2689,
+      "loss": 0.1119,
       "step": 540
     },
     {
       "epoch": 5.49,
-      "grad_norm": 5.50855827331543,
+      "grad_norm": 6.570211410522461,
       "learning_rate": 5e-05,
-      "loss": 0.258,
+      "loss": 0.1067,
       "step": 550
     },
     {
       "epoch": 5.59,
-      "grad_norm": 4.887132167816162,
+      "grad_norm": 4.3511199951171875,
       "learning_rate": 4.888888888888889e-05,
-      "loss": 0.2669,
+      "loss": 0.1202,
       "step": 560
     },
     {
       "epoch": 5.69,
-      "grad_norm": 4.258863925933838,
+      "grad_norm": 4.396121978759766,
       "learning_rate": 4.7777777777777784e-05,
-      "loss": 0.2416,
+      "loss": 0.1041,
       "step": 570
     },
     {
       "epoch": 5.79,
-      "grad_norm": 6.73113489151001,
+      "grad_norm": 5.755546569824219,
       "learning_rate": 4.666666666666667e-05,
-      "loss": 0.2575,
+      "loss": 0.1144,
       "step": 580
     },
     {
       "epoch": 5.89,
-      "grad_norm": 4.015745639801025,
+      "grad_norm": 4.096261501312256,
       "learning_rate": 4.555555555555556e-05,
-      "loss": 0.2308,
+      "loss": 0.109,
       "step": 590
     },
     {
       "epoch": 5.99,
-      "grad_norm": 7.089807510375977,
+      "grad_norm": 7.311570644378662,
       "learning_rate": 4.4444444444444447e-05,
-      "loss": 0.254,
+      "loss": 0.1188,
       "step": 600
     },
     {
       "epoch": 6.0,
-      "eval_accuracy": 0.8825385694249649,
-      "eval_loss": 0.34698864817619324,
-      "eval_runtime": 38.6025,
-      "eval_samples_per_second": 73.881,
-      "eval_steps_per_second": 1.166,
+      "eval_accuracy": 0.8842917251051894,
+      "eval_loss": 0.4293117821216583,
+      "eval_runtime": 36.9523,
+      "eval_samples_per_second": 77.181,
+      "eval_steps_per_second": 1.218,
       "step": 601
     },
     {
       "epoch": 6.08,
-      "grad_norm": 5.87188196182251,
+      "grad_norm": 3.6034762859344482,
       "learning_rate": 4.3333333333333334e-05,
-      "loss": 0.2432,
+      "loss": 0.1118,
       "step": 610
     },
     {
       "epoch": 6.18,
-      "grad_norm": 7.9145379066467285,
+      "grad_norm": 5.649061679840088,
       "learning_rate": 4.222222222222222e-05,
-      "loss": 0.2254,
+      "loss": 0.1005,
       "step": 620
     },
     {
       "epoch": 6.28,
-      "grad_norm": 6.642333030700684,
+      "grad_norm": 6.543485164642334,
       "learning_rate": 4.111111111111111e-05,
-      "loss": 0.2189,
+      "loss": 0.1012,
       "step": 630
     },
     {
       "epoch": 6.38,
-      "grad_norm": 5.588759899139404,
+      "grad_norm": 5.104783058166504,
       "learning_rate": 4e-05,
-      "loss": 0.2085,
+      "loss": 0.0862,
       "step": 640
     },
     {
       "epoch": 6.48,
-      "grad_norm": 4.943757057189941,
+      "grad_norm": 7.024726867675781,
       "learning_rate": 3.888888888888889e-05,
-      "loss": 0.2299,
+      "loss": 0.1063,
       "step": 650
     },
     {
       "epoch": 6.58,
-      "grad_norm": 4.3082594871521,
+      "grad_norm": 3.0800111293792725,
       "learning_rate": 3.777777777777778e-05,
-      "loss": 0.219,
+      "loss": 0.0974,
       "step": 660
     },
     {
       "epoch": 6.68,
-      "grad_norm": 5.564691066741943,
+      "grad_norm": 8.348942756652832,
       "learning_rate": 3.6666666666666666e-05,
-      "loss": 0.2346,
+      "loss": 0.1079,
       "step": 670
     },
     {
       "epoch": 6.78,
-      "grad_norm": 5.957446098327637,
+      "grad_norm": 5.644587516784668,
       "learning_rate": 3.555555555555556e-05,
-      "loss": 0.2294,
+      "loss": 0.0917,
       "step": 680
     },
     {
       "epoch": 6.88,
-      "grad_norm": 4.528319835662842,
+      "grad_norm": 6.3199872970581055,
       "learning_rate": 3.444444444444445e-05,
-      "loss": 0.208,
+      "loss": 0.1043,
       "step": 690
     },
     {
       "epoch": 6.98,
-      "grad_norm": 6.489378929138184,
+      "grad_norm": 7.193106174468994,
       "learning_rate": 3.3333333333333335e-05,
-      "loss": 0.2512,
+      "loss": 0.1293,
       "step": 700
     },
     {
       "epoch": 6.99,
-      "eval_accuracy": 0.8860448807854138,
-      "eval_loss": 0.3339594304561615,
-      "eval_runtime": 38.3448,
-      "eval_samples_per_second": 74.378,
-      "eval_steps_per_second": 1.174,
+      "eval_accuracy": 0.9032258064516129,
+      "eval_loss": 0.3595149517059326,
+      "eval_runtime": 37.0853,
+      "eval_samples_per_second": 76.904,
+      "eval_steps_per_second": 1.213,
       "step": 701
     },
     {
       "epoch": 7.08,
-      "grad_norm": 4.715758323669434,
+      "grad_norm": 5.787723541259766,
       "learning_rate": 3.222222222222223e-05,
-      "loss": 0.2247,
+      "loss": 0.1175,
       "step": 710
     },
     {
       "epoch": 7.18,
-      "grad_norm": 4.851717948913574,
+      "grad_norm": 3.8590798377990723,
       "learning_rate": 3.111111111111111e-05,
-      "loss": 0.2091,
+      "loss": 0.1003,
       "step": 720
     },
     {
       "epoch": 7.28,
-      "grad_norm": 5.768288612365723,
+      "grad_norm": 4.482466220855713,
       "learning_rate": 3e-05,
-      "loss": 0.2117,
+      "loss": 0.0991,
       "step": 730
     },
     {
       "epoch": 7.38,
-      "grad_norm": 4.091370582580566,
+      "grad_norm": 2.0709593296051025,
       "learning_rate": 2.8888888888888888e-05,
-      "loss": 0.1891,
+      "loss": 0.0827,
       "step": 740
     },
     {
       "epoch": 7.48,
-      "grad_norm": 3.7998056411743164,
+      "grad_norm": 5.126271724700928,
       "learning_rate": 2.777777777777778e-05,
-      "loss": 0.2005,
+      "loss": 0.1008,
       "step": 750
     },
     {
       "epoch": 7.58,
-      "grad_norm": 4.708491325378418,
+      "grad_norm": 4.332973003387451,
       "learning_rate": 2.6666666666666667e-05,
-      "loss": 0.1837,
+      "loss": 0.0805,
       "step": 760
     },
     {
       "epoch": 7.68,
-      "grad_norm": 4.464774131774902,
+      "grad_norm": 4.371252059936523,
       "learning_rate": 2.5555555555555554e-05,
-      "loss": 0.187,
+      "loss": 0.0951,
       "step": 770
     },
     {
       "epoch": 7.78,
-      "grad_norm": 4.849852561950684,
+      "grad_norm": 4.181329250335693,
       "learning_rate": 2.4444444444444445e-05,
-      "loss": 0.193,
+      "loss": 0.0937,
       "step": 780
     },
     {
       "epoch": 7.88,
-      "grad_norm": 4.364429950714111,
+      "grad_norm": 4.650885581970215,
       "learning_rate": 2.3333333333333336e-05,
-      "loss": 0.179,
+      "loss": 0.0857,
       "step": 790
     },
     {
       "epoch": 7.98,
-      "grad_norm": 6.141552925109863,
+      "grad_norm": 5.875858306884766,
       "learning_rate": 2.2222222222222223e-05,
-      "loss": 0.2092,
+      "loss": 0.1094,
       "step": 800
     },
     {
       "epoch": 8.0,
-      "eval_accuracy": 0.8860448807854138,
-      "eval_loss": 0.3440641760826111,
-      "eval_runtime": 38.2928,
-      "eval_samples_per_second": 74.479,
-      "eval_steps_per_second": 1.175,
+      "eval_accuracy": 0.8993688639551192,
+      "eval_loss": 0.36655670404434204,
+      "eval_runtime": 37.936,
+      "eval_samples_per_second": 75.179,
+      "eval_steps_per_second": 1.186,
       "step": 802
     },
     {
       "epoch": 8.08,
-      "grad_norm": 6.239747524261475,
+      "grad_norm": 6.7501349449157715,
       "learning_rate": 2.111111111111111e-05,
-      "loss": 0.1919,
+      "loss": 0.1025,
       "step": 810
     },
     {
       "epoch": 8.18,
-      "grad_norm": 4.517318248748779,
+      "grad_norm": 4.238629341125488,
       "learning_rate": 2e-05,
-      "loss": 0.1583,
+      "loss": 0.0783,
       "step": 820
     },
     {
       "epoch": 8.28,
-      "grad_norm": 5.20755672454834,
+      "grad_norm": 4.994195938110352,
       "learning_rate": 1.888888888888889e-05,
-      "loss": 0.1752,
+      "loss": 0.0989,
       "step": 830
     },
     {
       "epoch": 8.38,
-      "grad_norm": 4.3585991859436035,
+      "grad_norm": 4.091785430908203,
       "learning_rate": 1.777777777777778e-05,
-      "loss": 0.1748,
+      "loss": 0.093,
       "step": 840
     },
     {
       "epoch": 8.48,
-      "grad_norm": 4.46333646774292,
+      "grad_norm": 4.5268378257751465,
       "learning_rate": 1.6666666666666667e-05,
-      "loss": 0.1734,
+      "loss": 0.0993,
       "step": 850
     },
     {
       "epoch": 8.58,
-      "grad_norm": 4.966083526611328,
+      "grad_norm": 4.056771755218506,
       "learning_rate": 1.5555555555555555e-05,
-      "loss": 0.1783,
+      "loss": 0.0989,
       "step": 860
     },
     {
       "epoch": 8.68,
-      "grad_norm": 3.0647408962249756,
+      "grad_norm": 3.3963980674743652,
       "learning_rate": 1.4444444444444444e-05,
-      "loss": 0.1795,
+      "loss": 0.0975,
       "step": 870
     },
     {
       "epoch": 8.78,
-      "grad_norm": 4.360434532165527,
+      "grad_norm": 4.109832763671875,
       "learning_rate": 1.3333333333333333e-05,
-      "loss": 0.184,
+      "loss": 0.1079,
       "step": 880
     },
     {
       "epoch": 8.88,
-      "grad_norm": 4.1616973876953125,
+      "grad_norm": 4.538650035858154,
       "learning_rate": 1.2222222222222222e-05,
-      "loss": 0.1663,
+      "loss": 0.091,
       "step": 890
     },
     {
       "epoch": 8.98,
-      "grad_norm": 3.880200147628784,
+      "grad_norm": 3.982865571975708,
       "learning_rate": 1.1111111111111112e-05,
-      "loss": 0.1733,
+      "loss": 0.1047,
       "step": 900
     },
     {
       "epoch": 9.0,
-      "eval_accuracy": 0.894109396914446,
-      "eval_loss": 0.32196635007858276,
-      "eval_runtime": 38.3126,
-      "eval_samples_per_second": 74.44,
-      "eval_steps_per_second": 1.175,
+      "eval_accuracy": 0.9021739130434783,
+      "eval_loss": 0.3615379333496094,
+      "eval_runtime": 37.2047,
+      "eval_samples_per_second": 76.657,
+      "eval_steps_per_second": 1.21,
       "step": 902
     },
     {
       "epoch": 9.08,
-      "grad_norm": 5.077446460723877,
+      "grad_norm": 5.261457920074463,
       "learning_rate": 1e-05,
-      "loss": 0.1624,
+      "loss": 0.1016,
       "step": 910
     },
     {
       "epoch": 9.18,
-      "grad_norm": 4.26702880859375,
+      "grad_norm": 4.270400047302246,
       "learning_rate": 8.88888888888889e-06,
-      "loss": 0.17,
+      "loss": 0.109,
       "step": 920
     },
     {
       "epoch": 9.28,
-      "grad_norm": 3.957158327102661,
+      "grad_norm": 4.025336265563965,
       "learning_rate": 7.777777777777777e-06,
-      "loss": 0.1531,
+      "loss": 0.0955,
       "step": 930
     },
     {
       "epoch": 9.38,
-      "grad_norm": 3.6919987201690674,
+      "grad_norm": 4.099737167358398,
       "learning_rate": 6.666666666666667e-06,
-      "loss": 0.1559,
+      "loss": 0.1024,
       "step": 940
     },
     {
       "epoch": 9.48,
-      "grad_norm": 4.511406898498535,
+      "grad_norm": 4.68928861618042,
       "learning_rate": 5.555555555555556e-06,
-      "loss": 0.1542,
+      "loss": 0.1069,
       "step": 950
     },
     {
       "epoch": 9.58,
-      "grad_norm": 4.40770959854126,
+      "grad_norm": 4.062239170074463,
       "learning_rate": 4.444444444444445e-06,
-      "loss": 0.1502,
+      "loss": 0.0971,
       "step": 960
     },
     {
       "epoch": 9.68,
-      "grad_norm": 3.7929115295410156,
+      "grad_norm": 3.338601589202881,
       "learning_rate": 3.3333333333333333e-06,
-      "loss": 0.1565,
+      "loss": 0.1033,
       "step": 970
     },
     {
       "epoch": 9.78,
-      "grad_norm": 10.205160140991211,
+      "grad_norm": 7.083847522735596,
       "learning_rate": 2.2222222222222225e-06,
-      "loss": 0.1577,
+      "loss": 0.1139,
       "step": 980
     },
     {
       "epoch": 9.88,
-      "grad_norm": 3.9570119380950928,
+      "grad_norm": 5.833040714263916,
       "learning_rate": 1.1111111111111112e-06,
-      "loss": 0.1563,
+      "loss": 0.1198,
       "step": 990
     },
     {
       "epoch": 9.98,
-      "grad_norm": 6.175787448883057,
+      "grad_norm": 4.501120567321777,
       "learning_rate": 0.0,
-      "loss": 0.1386,
+      "loss": 0.0925,
       "step": 1000
     },
     {
       "epoch": 9.98,
-      "eval_accuracy": 0.8958625525946704,
-      "eval_loss": 0.31656068563461304,
-      "eval_runtime": 38.4379,
-      "eval_samples_per_second": 74.198,
-      "eval_steps_per_second": 1.171,
+      "eval_accuracy": 0.9119915848527349,
+      "eval_loss": 0.3377525210380554,
+      "eval_runtime": 37.2843,
+      "eval_samples_per_second": 76.493,
+      "eval_steps_per_second": 1.207,
       "step": 1000
     },
     {
       "epoch": 9.98,
       "step": 1000,
       "total_flos": 6.364199987970048e+18,
-      "train_loss": 0.2624686861038208,
-      "train_runtime": 4808.7305,
-      "train_samples_per_second": 53.37,
-      "train_steps_per_second": 0.208
+      "train_loss": 0.11027517792582511,
+      "train_runtime": 4749.2843,
+      "train_samples_per_second": 54.038,
+      "train_steps_per_second": 0.211
     }
   ],
   "logging_steps": 10,