End of training

all_results.json

@@ -1,13 +1,13 @@
 {
-    "epoch": 4.99,
-    "eval_accuracy": 0.9259259259259259,
-    "eval_loss": 0.20305776596069336,
-    "eval_runtime": 8.7135,
-    "eval_samples_per_second": 71.269,
-    "eval_steps_per_second": 8.952,
-    "total_flos": 6.917311166047027e+17,
-    "train_loss": 0.5214220342964961,
-    "train_runtime": 714.1939,
-    "train_samples_per_second": 39.072,
-    "train_steps_per_second": 1.218
+    "epoch": 9.97,
+    "eval_accuracy": 0.9726247987117552,
+    "eval_loss": 0.08394750207662582,
+    "eval_runtime": 8.4058,
+    "eval_samples_per_second": 73.878,
+    "eval_steps_per_second": 9.279,
+    "total_flos": 1.3833876610752307e+18,
+    "train_loss": 0.24886192696532983,
+    "train_runtime": 1438.0658,
+    "train_samples_per_second": 38.809,
+    "train_steps_per_second": 1.21
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 4.99,
-    "eval_accuracy": 0.9259259259259259,
-    "eval_loss": 0.20305776596069336,
-    "eval_runtime": 8.7135,
-    "eval_samples_per_second": 71.269,
-    "eval_steps_per_second": 8.952
+    "epoch": 9.97,
+    "eval_accuracy": 0.9726247987117552,
+    "eval_loss": 0.08394750207662582,
+    "eval_runtime": 8.4058,
+    "eval_samples_per_second": 73.878,
+    "eval_steps_per_second": 9.279
 }

runs/Apr17_12-51-18_5910a653ef0c/events.out.tfevents.1681737349.5910a653ef0c.229.10

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5f2d2e04027a81787cd460c45f4a5670b80215e417fa8f3114638ba29475e32a
+size 411

train_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 4.99,
-    "total_flos": 6.917311166047027e+17,
-    "train_loss": 0.5214220342964961,
-    "train_runtime": 714.1939,
-    "train_samples_per_second": 39.072,
-    "train_steps_per_second": 1.218
+    "epoch": 9.97,
+    "total_flos": 1.3833876610752307e+18,
+    "train_loss": 0.24886192696532983,
+    "train_runtime": 1438.0658,
+    "train_samples_per_second": 38.809,
+    "train_steps_per_second": 1.21
 }

trainer_state.json

@@ -1,592 +1,1159 @@
 {
-  "best_metric": 0.9259259259259259,
-  "best_model_checkpoint": "swin-tiny-patch4-window7-224-bottom_cleaned_data/checkpoint-870",
-  "epoch": 4.98567335243553,
-  "global_step": 870,
+  "best_metric": 0.9726247987117552,
+  "best_model_checkpoint": "swin-tiny-patch4-window7-224-bottom_cleaned_data/checkpoint-1047",
+  "epoch": 9.97134670487106,
+  "global_step": 1740,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
       "epoch": 0.06,
-      "learning_rate": 1.1494252873563219e-06,
-      "loss": 1.4361,
+      "learning_rate": 2.777777777777778e-05,
+      "loss": 1.3661,
       "step": 10
     },
     {
       "epoch": 0.11,
-      "learning_rate": 2.2988505747126437e-06,
-      "loss": 1.3935,
+      "learning_rate": 4.994192799070848e-05,
+      "loss": 1.1811,
       "step": 20
     },
     {
       "epoch": 0.17,
-      "learning_rate": 3.448275862068966e-06,
-      "loss": 1.3852,
+      "learning_rate": 4.965156794425087e-05,
+      "loss": 0.9235,
       "step": 30
     },
     {
       "epoch": 0.23,
-      "learning_rate": 4.5977011494252875e-06,
-      "loss": 1.3455,
+      "learning_rate": 4.9361207897793264e-05,
+      "loss": 0.84,
       "step": 40
     },
     {
       "epoch": 0.29,
-      "learning_rate": 5.747126436781609e-06,
-      "loss": 1.2788,
+      "learning_rate": 4.907084785133566e-05,
+      "loss": 0.6387,
       "step": 50
     },
     {
       "epoch": 0.34,
-      "learning_rate": 6.896551724137932e-06,
-      "loss": 1.2361,
+      "learning_rate": 4.878048780487805e-05,
+      "loss": 0.7581,
       "step": 60
     },
     {
       "epoch": 0.4,
-      "learning_rate": 8.045977011494253e-06,
-      "loss": 1.1974,
+      "learning_rate": 4.8490127758420445e-05,
+      "loss": 0.6915,
       "step": 70
     },
     {
       "epoch": 0.46,
-      "learning_rate": 9.195402298850575e-06,
-      "loss": 1.1278,
+      "learning_rate": 4.819976771196283e-05,
+      "loss": 0.6255,
       "step": 80
     },
     {
       "epoch": 0.52,
-      "learning_rate": 9.96168582375479e-06,
-      "loss": 1.0761,
+      "learning_rate": 4.7909407665505226e-05,
+      "loss": 0.527,
       "step": 90
     },
     {
       "epoch": 0.57,
-      "learning_rate": 9.833971902937422e-06,
-      "loss": 1.0022,
+      "learning_rate": 4.761904761904762e-05,
+      "loss": 0.5023,
       "step": 100
     },
     {
       "epoch": 0.63,
-      "learning_rate": 9.706257982120052e-06,
-      "loss": 0.9669,
+      "learning_rate": 4.7328687572590014e-05,
+      "loss": 0.5157,
       "step": 110
     },
     {
       "epoch": 0.69,
-      "learning_rate": 9.578544061302683e-06,
-      "loss": 0.8852,
+      "learning_rate": 4.703832752613241e-05,
+      "loss": 0.4598,
       "step": 120
     },
     {
       "epoch": 0.74,
-      "learning_rate": 9.450830140485315e-06,
-      "loss": 0.8496,
+      "learning_rate": 4.6747967479674795e-05,
+      "loss": 0.4344,
       "step": 130
     },
     {
       "epoch": 0.8,
-      "learning_rate": 9.323116219667945e-06,
-      "loss": 0.804,
+      "learning_rate": 4.6457607433217196e-05,
+      "loss": 0.4745,
       "step": 140
     },
     {
       "epoch": 0.86,
-      "learning_rate": 9.195402298850575e-06,
-      "loss": 0.7712,
+      "learning_rate": 4.616724738675958e-05,
+      "loss": 0.4485,
       "step": 150
     },
     {
       "epoch": 0.92,
-      "learning_rate": 9.067688378033207e-06,
-      "loss": 0.6497,
+      "learning_rate": 4.587688734030198e-05,
+      "loss": 0.3849,
       "step": 160
     },
     {
       "epoch": 0.97,
-      "learning_rate": 8.939974457215838e-06,
-      "loss": 0.6591,
+      "learning_rate": 4.5586527293844364e-05,
+      "loss": 0.4444,
       "step": 170
     },
     {
       "epoch": 1.0,
-      "eval_accuracy": 0.8067632850241546,
-      "eval_loss": 0.5231599807739258,
-      "eval_runtime": 9.148,
-      "eval_samples_per_second": 67.884,
-      "eval_steps_per_second": 8.526,
+      "eval_accuracy": 0.9162640901771336,
+      "eval_loss": 0.2271285504102707,
+      "eval_runtime": 8.1654,
+      "eval_samples_per_second": 76.052,
+      "eval_steps_per_second": 9.552,
       "step": 174
     },
     {
       "epoch": 1.03,
-      "learning_rate": 8.812260536398468e-06,
-      "loss": 0.6781,
+      "learning_rate": 4.529616724738676e-05,
+      "loss": 0.4314,
       "step": 180
     },
     {
       "epoch": 1.09,
-      "learning_rate": 8.684546615581098e-06,
-      "loss": 0.6175,
+      "learning_rate": 4.500580720092916e-05,
+      "loss": 0.325,
       "step": 190
     },
     {
       "epoch": 1.15,
-      "learning_rate": 8.55683269476373e-06,
-      "loss": 0.5585,
+      "learning_rate": 4.4715447154471546e-05,
+      "loss": 0.3656,
       "step": 200
     },
     {
       "epoch": 1.2,
-      "learning_rate": 8.429118773946362e-06,
-      "loss": 0.5456,
+      "learning_rate": 4.442508710801394e-05,
+      "loss": 0.3626,
       "step": 210
     },
     {
       "epoch": 1.26,
-      "learning_rate": 8.301404853128992e-06,
-      "loss": 0.5256,
+      "learning_rate": 4.413472706155633e-05,
+      "loss": 0.3937,
       "step": 220
     },
     {
       "epoch": 1.32,
-      "learning_rate": 8.173690932311623e-06,
-      "loss": 0.5278,
+      "learning_rate": 4.384436701509873e-05,
+      "loss": 0.3097,
       "step": 230
     },
     {
       "epoch": 1.38,
-      "learning_rate": 8.045977011494253e-06,
-      "loss": 0.4722,
+      "learning_rate": 4.3554006968641115e-05,
+      "loss": 0.3426,
       "step": 240
     },
     {
       "epoch": 1.43,
-      "learning_rate": 7.918263090676885e-06,
-      "loss": 0.509,
+      "learning_rate": 4.326364692218351e-05,
+      "loss": 0.3182,
       "step": 250
     },
     {
       "epoch": 1.49,
-      "learning_rate": 7.790549169859515e-06,
-      "loss": 0.4425,
+      "learning_rate": 4.29732868757259e-05,
+      "loss": 0.3147,
       "step": 260
     },
     {
       "epoch": 1.55,
-      "learning_rate": 7.662835249042147e-06,
-      "loss": 0.4515,
+      "learning_rate": 4.26829268292683e-05,
+      "loss": 0.3389,
       "step": 270
     },
     {
       "epoch": 1.6,
-      "learning_rate": 7.535121328224777e-06,
-      "loss": 0.477,
+      "learning_rate": 4.239256678281069e-05,
+      "loss": 0.3291,
       "step": 280
     },
     {
       "epoch": 1.66,
-      "learning_rate": 7.4074074074074075e-06,
-      "loss": 0.5036,
+      "learning_rate": 4.210220673635308e-05,
+      "loss": 0.3706,
       "step": 290
     },
     {
       "epoch": 1.72,
-      "learning_rate": 7.279693486590039e-06,
-      "loss": 0.5159,
+      "learning_rate": 4.181184668989547e-05,
+      "loss": 0.3576,
       "step": 300
     },
     {
       "epoch": 1.78,
-      "learning_rate": 7.15197956577267e-06,
-      "loss": 0.5527,
+      "learning_rate": 4.1521486643437866e-05,
+      "loss": 0.3543,
       "step": 310
     },
     {
       "epoch": 1.83,
-      "learning_rate": 7.0242656449553e-06,
-      "loss": 0.4243,
+      "learning_rate": 4.123112659698026e-05,
+      "loss": 0.305,
       "step": 320
     },
     {
       "epoch": 1.89,
-      "learning_rate": 6.896551724137932e-06,
-      "loss": 0.4485,
+      "learning_rate": 4.0940766550522653e-05,
+      "loss": 0.3487,
       "step": 330
     },
     {
       "epoch": 1.95,
-      "learning_rate": 6.7688378033205625e-06,
-      "loss": 0.4104,
+      "learning_rate": 4.065040650406504e-05,
+      "loss": 0.3518,
       "step": 340
     },
     {
       "epoch": 2.0,
-      "eval_accuracy": 0.8888888888888888,
-      "eval_loss": 0.31606417894363403,
-      "eval_runtime": 9.5164,
-      "eval_samples_per_second": 65.255,
-      "eval_steps_per_second": 8.196,
+      "eval_accuracy": 0.9033816425120773,
+      "eval_loss": 0.24492110311985016,
+      "eval_runtime": 9.5576,
+      "eval_samples_per_second": 64.975,
+      "eval_steps_per_second": 8.161,
       "step": 349
     },
     {
       "epoch": 2.01,
-      "learning_rate": 6.641123882503193e-06,
-      "loss": 0.3783,
+      "learning_rate": 4.0360046457607435e-05,
+      "loss": 0.2923,
       "step": 350
     },
     {
       "epoch": 2.06,
-      "learning_rate": 6.513409961685824e-06,
-      "loss": 0.372,
+      "learning_rate": 4.006968641114983e-05,
+      "loss": 0.3222,
       "step": 360
     },
     {
       "epoch": 2.12,
-      "learning_rate": 6.385696040868455e-06,
-      "loss": 0.4664,
+      "learning_rate": 3.977932636469222e-05,
+      "loss": 0.317,
       "step": 370
     },
     {
       "epoch": 2.18,
-      "learning_rate": 6.257982120051086e-06,
-      "loss": 0.4034,
+      "learning_rate": 3.948896631823461e-05,
+      "loss": 0.3217,
       "step": 380
     },
     {
       "epoch": 2.23,
-      "learning_rate": 6.130268199233717e-06,
-      "loss": 0.3513,
+      "learning_rate": 3.9198606271777003e-05,
+      "loss": 0.2979,
       "step": 390
     },
     {
       "epoch": 2.29,
-      "learning_rate": 6.002554278416348e-06,
-      "loss": 0.3816,
+      "learning_rate": 3.89082462253194e-05,
+      "loss": 0.259,
       "step": 400
     },
     {
       "epoch": 2.35,
-      "learning_rate": 5.874840357598979e-06,
-      "loss": 0.4518,
+      "learning_rate": 3.861788617886179e-05,
+      "loss": 0.3416,
       "step": 410
     },
     {
       "epoch": 2.41,
-      "learning_rate": 5.747126436781609e-06,
-      "loss": 0.4686,
+      "learning_rate": 3.8327526132404185e-05,
+      "loss": 0.3103,
       "step": 420
     },
     {
       "epoch": 2.46,
-      "learning_rate": 5.619412515964241e-06,
-      "loss": 0.3879,
+      "learning_rate": 3.803716608594657e-05,
+      "loss": 0.2513,
       "step": 430
     },
     {
       "epoch": 2.52,
-      "learning_rate": 5.491698595146872e-06,
-      "loss": 0.3775,
+      "learning_rate": 3.7746806039488966e-05,
+      "loss": 0.2599,
       "step": 440
     },
     {
       "epoch": 2.58,
-      "learning_rate": 5.3639846743295025e-06,
-      "loss": 0.457,
+      "learning_rate": 3.745644599303136e-05,
+      "loss": 0.3085,
       "step": 450
     },
     {
       "epoch": 2.64,
-      "learning_rate": 5.236270753512134e-06,
-      "loss": 0.3692,
+      "learning_rate": 3.7166085946573754e-05,
+      "loss": 0.2228,
       "step": 460
     },
     {
       "epoch": 2.69,
-      "learning_rate": 5.108556832694764e-06,
-      "loss": 0.4145,
+      "learning_rate": 3.687572590011615e-05,
+      "loss": 0.2387,
       "step": 470
     },
     {
       "epoch": 2.75,
-      "learning_rate": 4.980842911877395e-06,
-      "loss": 0.3576,
+      "learning_rate": 3.6585365853658535e-05,
+      "loss": 0.2562,
       "step": 480
     },
     {
       "epoch": 2.81,
-      "learning_rate": 4.853128991060026e-06,
-      "loss": 0.3297,
+      "learning_rate": 3.629500580720093e-05,
+      "loss": 0.2431,
       "step": 490
     },
     {
       "epoch": 2.87,
-      "learning_rate": 4.7254150702426575e-06,
-      "loss": 0.4084,
+      "learning_rate": 3.600464576074332e-05,
+      "loss": 0.2774,
       "step": 500
     },
     {
       "epoch": 2.92,
-      "learning_rate": 4.5977011494252875e-06,
-      "loss": 0.3237,
+      "learning_rate": 3.571428571428572e-05,
+      "loss": 0.2445,
       "step": 510
     },
     {
       "epoch": 2.98,
-      "learning_rate": 4.469987228607919e-06,
-      "loss": 0.3559,
+      "learning_rate": 3.5423925667828104e-05,
+      "loss": 0.225,
       "step": 520
     },
     {
       "epoch": 3.0,
-      "eval_accuracy": 0.9162640901771336,
-      "eval_loss": 0.2237011194229126,
-      "eval_runtime": 9.3284,
-      "eval_samples_per_second": 66.571,
-      "eval_steps_per_second": 8.362,
+      "eval_accuracy": 0.9500805152979066,
+      "eval_loss": 0.13247297704219818,
+      "eval_runtime": 8.7139,
+      "eval_samples_per_second": 71.265,
+      "eval_steps_per_second": 8.951,
       "step": 523
     },
     {
       "epoch": 3.04,
-      "learning_rate": 4.342273307790549e-06,
-      "loss": 0.3295,
+      "learning_rate": 3.51335656213705e-05,
+      "loss": 0.2198,
       "step": 530
     },
     {
       "epoch": 3.09,
-      "learning_rate": 4.214559386973181e-06,
-      "loss": 0.3902,
+      "learning_rate": 3.48432055749129e-05,
+      "loss": 0.2398,
       "step": 540
     },
     {
       "epoch": 3.15,
-      "learning_rate": 4.086845466155812e-06,
-      "loss": 0.2753,
+      "learning_rate": 3.4552845528455286e-05,
+      "loss": 0.1821,
       "step": 550
     },
     {
       "epoch": 3.21,
-      "learning_rate": 3.9591315453384425e-06,
-      "loss": 0.3705,
+      "learning_rate": 3.426248548199768e-05,
+      "loss": 0.2148,
       "step": 560
     },
     {
       "epoch": 3.27,
-      "learning_rate": 3.831417624521073e-06,
-      "loss": 0.3797,
+      "learning_rate": 3.397212543554007e-05,
+      "loss": 0.3183,
       "step": 570
     },
     {
       "epoch": 3.32,
-      "learning_rate": 3.7037037037037037e-06,
-      "loss": 0.3163,
+      "learning_rate": 3.368176538908247e-05,
+      "loss": 0.2292,
       "step": 580
     },
     {
       "epoch": 3.38,
-      "learning_rate": 3.575989782886335e-06,
-      "loss": 0.3488,
+      "learning_rate": 3.3391405342624855e-05,
+      "loss": 0.1987,
       "step": 590
     },
     {
       "epoch": 3.44,
-      "learning_rate": 3.448275862068966e-06,
-      "loss": 0.3248,
+      "learning_rate": 3.310104529616725e-05,
+      "loss": 0.2434,
       "step": 600
     },
     {
       "epoch": 3.5,
-      "learning_rate": 3.3205619412515967e-06,
-      "loss": 0.3806,
+      "learning_rate": 3.281068524970964e-05,
+      "loss": 0.2906,
       "step": 610
     },
     {
       "epoch": 3.55,
-      "learning_rate": 3.1928480204342275e-06,
-      "loss": 0.2763,
+      "learning_rate": 3.2520325203252037e-05,
+      "loss": 0.2117,
       "step": 620
     },
     {
       "epoch": 3.61,
-      "learning_rate": 3.0651340996168583e-06,
-      "loss": 0.3623,
+      "learning_rate": 3.222996515679443e-05,
+      "loss": 0.22,
       "step": 630
     },
     {
       "epoch": 3.67,
-      "learning_rate": 2.9374201787994896e-06,
-      "loss": 0.3353,
+      "learning_rate": 3.193960511033682e-05,
+      "loss": 0.1831,
       "step": 640
     },
     {
       "epoch": 3.72,
-      "learning_rate": 2.8097062579821204e-06,
-      "loss": 0.3229,
+      "learning_rate": 3.164924506387921e-05,
+      "loss": 0.1943,
       "step": 650
     },
     {
       "epoch": 3.78,
-      "learning_rate": 2.6819923371647512e-06,
-      "loss": 0.3127,
+      "learning_rate": 3.13588850174216e-05,
+      "loss": 0.2528,
       "step": 660
     },
     {
       "epoch": 3.84,
-      "learning_rate": 2.554278416347382e-06,
-      "loss": 0.28,
+      "learning_rate": 3.1068524970964e-05,
+      "loss": 0.2057,
       "step": 670
     },
     {
       "epoch": 3.9,
-      "learning_rate": 2.426564495530013e-06,
-      "loss": 0.4105,
+      "learning_rate": 3.077816492450639e-05,
+      "loss": 0.2584,
       "step": 680
     },
     {
       "epoch": 3.95,
-      "learning_rate": 2.2988505747126437e-06,
-      "loss": 0.3487,
+      "learning_rate": 3.048780487804878e-05,
+      "loss": 0.2195,
       "step": 690
     },
     {
       "epoch": 4.0,
-      "eval_accuracy": 0.9194847020933977,
-      "eval_loss": 0.19985385239124298,
-      "eval_runtime": 8.2263,
-      "eval_samples_per_second": 75.49,
-      "eval_steps_per_second": 9.482,
+      "eval_accuracy": 0.9549114331723028,
+      "eval_loss": 0.10237770527601242,
+      "eval_runtime": 9.1181,
+      "eval_samples_per_second": 68.106,
+      "eval_steps_per_second": 8.554,
       "step": 698
     },
     {
       "epoch": 4.01,
-      "learning_rate": 2.1711366538952746e-06,
-      "loss": 0.333,
+      "learning_rate": 3.0197444831591178e-05,
+      "loss": 0.2832,
       "step": 700
     },
     {
       "epoch": 4.07,
-      "learning_rate": 2.043422733077906e-06,
-      "loss": 0.3239,
+      "learning_rate": 2.9907084785133565e-05,
+      "loss": 0.2087,
       "step": 710
     },
     {
       "epoch": 4.13,
-      "learning_rate": 1.9157088122605367e-06,
-      "loss": 0.256,
+      "learning_rate": 2.9616724738675962e-05,
+      "loss": 0.1709,
       "step": 720
     },
     {
       "epoch": 4.18,
-      "learning_rate": 1.7879948914431675e-06,
-      "loss": 0.3252,
+      "learning_rate": 2.932636469221835e-05,
+      "loss": 0.1891,
       "step": 730
     },
     {
       "epoch": 4.24,
-      "learning_rate": 1.6602809706257983e-06,
-      "loss": 0.2693,
+      "learning_rate": 2.9036004645760743e-05,
+      "loss": 0.1923,
       "step": 740
     },
     {
       "epoch": 4.3,
-      "learning_rate": 1.5325670498084292e-06,
-      "loss": 0.3099,
+      "learning_rate": 2.874564459930314e-05,
+      "loss": 0.2178,
       "step": 750
     },
     {
       "epoch": 4.36,
-      "learning_rate": 1.4048531289910602e-06,
-      "loss": 0.3115,
+      "learning_rate": 2.8455284552845528e-05,
+      "loss": 0.2256,
       "step": 760
     },
     {
       "epoch": 4.41,
-      "learning_rate": 1.277139208173691e-06,
-      "loss": 0.3501,
+      "learning_rate": 2.8164924506387925e-05,
+      "loss": 0.2184,
       "step": 770
     },
     {
       "epoch": 4.47,
-      "learning_rate": 1.1494252873563219e-06,
-      "loss": 0.3733,
+      "learning_rate": 2.7874564459930312e-05,
+      "loss": 0.245,
       "step": 780
     },
     {
       "epoch": 4.53,
-      "learning_rate": 1.021711366538953e-06,
-      "loss": 0.3264,
+      "learning_rate": 2.758420441347271e-05,
+      "loss": 0.229,
       "step": 790
     },
     {
       "epoch": 4.58,
-      "learning_rate": 8.939974457215837e-07,
-      "loss": 0.3111,
+      "learning_rate": 2.7293844367015097e-05,
+      "loss": 0.2236,
       "step": 800
     },
     {
       "epoch": 4.64,
-      "learning_rate": 7.662835249042146e-07,
-      "loss": 0.4114,
+      "learning_rate": 2.7003484320557494e-05,
+      "loss": 0.2517,
       "step": 810
     },
     {
       "epoch": 4.7,
-      "learning_rate": 6.385696040868455e-07,
-      "loss": 0.2706,
+      "learning_rate": 2.6713124274099888e-05,
+      "loss": 0.1755,
       "step": 820
     },
     {
       "epoch": 4.76,
-      "learning_rate": 5.108556832694765e-07,
-      "loss": 0.2863,
+      "learning_rate": 2.642276422764228e-05,
+      "loss": 0.1663,
       "step": 830
     },
     {
       "epoch": 4.81,
-      "learning_rate": 3.831417624521073e-07,
-      "loss": 0.2813,
+      "learning_rate": 2.6132404181184672e-05,
+      "loss": 0.1958,
       "step": 840
     },
     {
       "epoch": 4.87,
-      "learning_rate": 2.5542784163473823e-07,
-      "loss": 0.3687,
+      "learning_rate": 2.5842044134727063e-05,
+      "loss": 0.2334,
       "step": 850
     },
     {
       "epoch": 4.93,
-      "learning_rate": 1.2771392081736911e-07,
-      "loss": 0.3689,
+      "learning_rate": 2.5551684088269457e-05,
+      "loss": 0.2576,
       "step": 860
     },
     {
       "epoch": 4.99,
-      "learning_rate": 0.0,
-      "loss": 0.3422,
+      "learning_rate": 2.5261324041811847e-05,
+      "loss": 0.2627,
       "step": 870
     },
     {
-      "epoch": 4.99,
-      "eval_accuracy": 0.9259259259259259,
-      "eval_loss": 0.20305776596069336,
-      "eval_runtime": 9.5579,
-      "eval_samples_per_second": 64.972,
-      "eval_steps_per_second": 8.161,
-      "step": 870
+      "epoch": 5.0,
+      "eval_accuracy": 0.9629629629629629,
+      "eval_loss": 0.1045805960893631,
+      "eval_runtime": 9.624,
+      "eval_samples_per_second": 64.526,
+      "eval_steps_per_second": 8.105,
+      "step": 872
     },
     {
-      "epoch": 4.99,
-      "step": 870,
-      "total_flos": 6.917311166047027e+17,
-      "train_loss": 0.5214220342964961,
-      "train_runtime": 714.1939,
-      "train_samples_per_second": 39.072,
-      "train_steps_per_second": 1.218
+      "epoch": 5.04,
+      "learning_rate": 2.497096399535424e-05,
+      "loss": 0.2139,
+      "step": 880
+    },
+    {
+      "epoch": 5.1,
+      "learning_rate": 2.4680603948896632e-05,
+      "loss": 0.2114,
+      "step": 890
+    },
+    {
+      "epoch": 5.16,
+      "learning_rate": 2.4390243902439026e-05,
+      "loss": 0.2097,
+      "step": 900
+    },
+    {
+      "epoch": 5.21,
+      "learning_rate": 2.4099883855981416e-05,
+      "loss": 0.2136,
+      "step": 910
+    },
+    {
+      "epoch": 5.27,
+      "learning_rate": 2.380952380952381e-05,
+      "loss": 0.2523,
+      "step": 920
+    },
+    {
+      "epoch": 5.33,
+      "learning_rate": 2.3519163763066204e-05,
+      "loss": 0.1787,
+      "step": 930
+    },
+    {
+      "epoch": 5.39,
+      "learning_rate": 2.3228803716608598e-05,
+      "loss": 0.1828,
+      "step": 940
+    },
+    {
+      "epoch": 5.44,
+      "learning_rate": 2.293844367015099e-05,
+      "loss": 0.1549,
+      "step": 950
+    },
+    {
+      "epoch": 5.5,
+      "learning_rate": 2.264808362369338e-05,
+      "loss": 0.1531,
+      "step": 960
+    },
+    {
+      "epoch": 5.56,
+      "learning_rate": 2.2357723577235773e-05,
+      "loss": 0.1957,
+      "step": 970
+    },
+    {
+      "epoch": 5.62,
+      "learning_rate": 2.2067363530778164e-05,
+      "loss": 0.222,
+      "step": 980
+    },
+    {
+      "epoch": 5.67,
+      "learning_rate": 2.1777003484320557e-05,
+      "loss": 0.2211,
+      "step": 990
+    },
+    {
+      "epoch": 5.73,
+      "learning_rate": 2.148664343786295e-05,
+      "loss": 0.1711,
+      "step": 1000
+    },
+    {
+      "epoch": 5.79,
+      "learning_rate": 2.1196283391405345e-05,
+      "loss": 0.1759,
+      "step": 1010
+    },
+    {
+      "epoch": 5.85,
+      "learning_rate": 2.0905923344947736e-05,
+      "loss": 0.2333,
+      "step": 1020
+    },
+    {
+      "epoch": 5.9,
+      "learning_rate": 2.061556329849013e-05,
+      "loss": 0.2269,
+      "step": 1030
+    },
+    {
+      "epoch": 5.96,
+      "learning_rate": 2.032520325203252e-05,
+      "loss": 0.142,
+      "step": 1040
+    },
+    {
+      "epoch": 6.0,
+      "eval_accuracy": 0.9726247987117552,
+      "eval_loss": 0.08394750207662582,
+      "eval_runtime": 8.0552,
+      "eval_samples_per_second": 77.093,
+      "eval_steps_per_second": 9.683,
+      "step": 1047
+    },
+    {
+      "epoch": 6.02,
+      "learning_rate": 2.0034843205574914e-05,
+      "loss": 0.1847,
+      "step": 1050
+    },
+    {
+      "epoch": 6.07,
+      "learning_rate": 1.9744483159117305e-05,
+      "loss": 0.1569,
+      "step": 1060
+    },
+    {
+      "epoch": 6.13,
+      "learning_rate": 1.94541231126597e-05,
+      "loss": 0.2001,
+      "step": 1070
+    },
+    {
+      "epoch": 6.19,
+      "learning_rate": 1.9163763066202093e-05,
+      "loss": 0.1721,
+      "step": 1080
+    },
+    {
+      "epoch": 6.25,
+      "learning_rate": 1.8873403019744483e-05,
+      "loss": 0.1406,
+      "step": 1090
+    },
+    {
+      "epoch": 6.3,
+      "learning_rate": 1.8583042973286877e-05,
+      "loss": 0.1943,
+      "step": 1100
+    },
+    {
+      "epoch": 6.36,
+      "learning_rate": 1.8292682926829268e-05,
+      "loss": 0.1812,
+      "step": 1110
+    },
+    {
+      "epoch": 6.42,
+      "learning_rate": 1.800232288037166e-05,
+      "loss": 0.105,
+      "step": 1120
+    },
+    {
+      "epoch": 6.48,
+      "learning_rate": 1.7711962833914052e-05,
+      "loss": 0.1356,
+      "step": 1130
+    },
+    {
+      "epoch": 6.53,
+      "learning_rate": 1.742160278745645e-05,
+      "loss": 0.1678,
+      "step": 1140
+    },
+    {
+      "epoch": 6.59,
+      "learning_rate": 1.713124274099884e-05,
+      "loss": 0.1989,
+      "step": 1150
+    },
+    {
+      "epoch": 6.65,
+      "learning_rate": 1.6840882694541234e-05,
+      "loss": 0.0919,
+      "step": 1160
+    },
+    {
+      "epoch": 6.7,
+      "learning_rate": 1.6550522648083624e-05,
+      "loss": 0.1554,
+      "step": 1170
+    },
+    {
+      "epoch": 6.76,
+      "learning_rate": 1.6260162601626018e-05,
+      "loss": 0.1257,
+      "step": 1180
+    },
+    {
+      "epoch": 6.82,
+      "learning_rate": 1.596980255516841e-05,
+      "loss": 0.2064,
+      "step": 1190
+    },
+    {
+      "epoch": 6.88,
+      "learning_rate": 1.56794425087108e-05,
+      "loss": 0.1877,
+      "step": 1200
+    },
+    {
+      "epoch": 6.93,
+      "learning_rate": 1.5389082462253197e-05,
+      "loss": 0.1878,
+      "step": 1210
+    },
+    {
+      "epoch": 6.99,
+      "learning_rate": 1.5098722415795589e-05,
+      "loss": 0.1516,
+      "step": 1220
+    },
+    {
+      "epoch": 7.0,
+      "eval_accuracy": 0.9629629629629629,
+      "eval_loss": 0.09183160960674286,
+      "eval_runtime": 9.7233,
+      "eval_samples_per_second": 63.867,
+      "eval_steps_per_second": 8.022,
+      "step": 1221
+    },
+    {
+      "epoch": 7.05,
+      "learning_rate": 1.4808362369337981e-05,
+      "loss": 0.1592,
+      "step": 1230
+    },
+    {
+      "epoch": 7.11,
+      "learning_rate": 1.4518002322880372e-05,
+      "loss": 0.198,
+      "step": 1240
+    },
+    {
+      "epoch": 7.16,
+      "learning_rate": 1.4227642276422764e-05,
+      "loss": 0.1892,
+      "step": 1250
+    },
+    {
+      "epoch": 7.22,
+      "learning_rate": 1.3937282229965156e-05,
+      "loss": 0.1777,
+      "step": 1260
+    },
+    {
+      "epoch": 7.28,
+      "learning_rate": 1.3646922183507548e-05,
+      "loss": 0.0918,
+      "step": 1270
+    },
+    {
+      "epoch": 7.34,
+      "learning_rate": 1.3356562137049944e-05,
+      "loss": 0.0896,
+      "step": 1280
+    },
+    {
+      "epoch": 7.39,
+      "learning_rate": 1.3066202090592336e-05,
+      "loss": 0.1273,
+      "step": 1290
+    },
+    {
+      "epoch": 7.45,
+      "learning_rate": 1.2775842044134728e-05,
+      "loss": 0.1521,
+      "step": 1300
+    },
+    {
+      "epoch": 7.51,
+      "learning_rate": 1.248548199767712e-05,
+      "loss": 0.1628,
+      "step": 1310
+    },
+    {
+      "epoch": 7.56,
+      "learning_rate": 1.2195121951219513e-05,
+      "loss": 0.1587,
+      "step": 1320
+    },
+    {
+      "epoch": 7.62,
+      "learning_rate": 1.1904761904761905e-05,
+      "loss": 0.205,
+      "step": 1330
+    },
+    {
+      "epoch": 7.68,
+      "learning_rate": 1.1614401858304299e-05,
+      "loss": 0.1415,
+      "step": 1340
+    },
+    {
+      "epoch": 7.74,
+      "learning_rate": 1.132404181184669e-05,
+      "loss": 0.1982,
+      "step": 1350
+    },
+    {
+      "epoch": 7.79,
+      "learning_rate": 1.1033681765389082e-05,
+      "loss": 0.1079,
+      "step": 1360
+    },
+    {
+      "epoch": 7.85,
+      "learning_rate": 1.0743321718931476e-05,
+      "loss": 0.1115,
+      "step": 1370
+    },
+    {
+      "epoch": 7.91,
+      "learning_rate": 1.0452961672473868e-05,
+      "loss": 0.2234,
+      "step": 1380
+    },
+    {
+      "epoch": 7.97,
+      "learning_rate": 1.016260162601626e-05,
+      "loss": 0.1498,
+      "step": 1390
+    },
+    {
+      "epoch": 8.0,
+      "eval_accuracy": 0.9726247987117552,
+      "eval_loss": 0.07796485722064972,
+      "eval_runtime": 9.3543,
+      "eval_samples_per_second": 66.386,
+      "eval_steps_per_second": 8.338,
+      "step": 1396
+    },
+    {
+      "epoch": 8.02,
+      "learning_rate": 9.872241579558652e-06,
+      "loss": 0.1299,
+      "step": 1400
+    },
+    {
+      "epoch": 8.08,
+      "learning_rate": 9.581881533101046e-06,
+      "loss": 0.1389,
+      "step": 1410
+    },
+    {
+      "epoch": 8.14,
+      "learning_rate": 9.291521486643439e-06,
+      "loss": 0.1499,
+      "step": 1420
+    },
+    {
+      "epoch": 8.19,
+      "learning_rate": 9.00116144018583e-06,
+      "loss": 0.2523,
+      "step": 1430
+    },
+    {
+      "epoch": 8.25,
+      "learning_rate": 8.710801393728225e-06,
+      "loss": 0.1067,
+      "step": 1440
+    },
+    {
+      "epoch": 8.31,
+      "learning_rate": 8.420441347270617e-06,
+      "loss": 0.1696,
+      "step": 1450
+    },
+    {
+      "epoch": 8.37,
+      "learning_rate": 8.130081300813009e-06,
+      "loss": 0.1279,
+      "step": 1460
+    },
+    {
+      "epoch": 8.42,
+      "learning_rate": 7.8397212543554e-06,
+      "loss": 0.1523,
+      "step": 1470
+    },
+    {
+      "epoch": 8.48,
+      "learning_rate": 7.5493612078977944e-06,
+      "loss": 0.1335,
+      "step": 1480
+    },
+    {
+      "epoch": 8.54,
+      "learning_rate": 7.259001161440186e-06,
+      "loss": 0.1122,
+      "step": 1490
+    },
+    {
+      "epoch": 8.6,
+      "learning_rate": 6.968641114982578e-06,
+      "loss": 0.1596,
+      "step": 1500
+    },
+    {
+      "epoch": 8.65,
+      "learning_rate": 6.678281068524972e-06,
+      "loss": 0.0975,
+      "step": 1510
+    },
+    {
+      "epoch": 8.71,
+      "learning_rate": 6.387921022067364e-06,
+      "loss": 0.1002,
+      "step": 1520
+    },
+    {
+      "epoch": 8.77,
+      "learning_rate": 6.0975609756097564e-06,
+      "loss": 0.0895,
+      "step": 1530
+    },
+    {
+      "epoch": 8.83,
+      "learning_rate": 5.8072009291521495e-06,
+      "loss": 0.0911,
+      "step": 1540
+    },
+    {
+      "epoch": 8.88,
+      "learning_rate": 5.516840882694541e-06,
+      "loss": 0.157,
+      "step": 1550
+    },
+    {
+      "epoch": 8.94,
+      "learning_rate": 5.226480836236934e-06,
+      "loss": 0.1513,
+      "step": 1560
+    },
+    {
+      "epoch": 9.0,
+      "learning_rate": 4.936120789779326e-06,
+      "loss": 0.1189,
+      "step": 1570
+    },
+    {
+      "epoch": 9.0,
+      "eval_accuracy": 0.966183574879227,
+      "eval_loss": 0.07211676239967346,
+      "eval_runtime": 8.3308,
+      "eval_samples_per_second": 74.542,
+      "eval_steps_per_second": 9.363,
+      "step": 1570
+    },
+    {
+      "epoch": 9.05,
+      "learning_rate": 4.645760743321719e-06,
+      "loss": 0.1239,
+      "step": 1580
+    },
+    {
+      "epoch": 9.11,
+      "learning_rate": 4.355400696864112e-06,
+      "loss": 0.1039,
+      "step": 1590
+    },
+    {
+      "epoch": 9.17,
+      "learning_rate": 4.0650406504065046e-06,
+      "loss": 0.1144,
+      "step": 1600
+    },
+    {
+      "epoch": 9.23,
+      "learning_rate": 3.7746806039488972e-06,
+      "loss": 0.1033,
+      "step": 1610
+    },
+    {
+      "epoch": 9.28,
+      "learning_rate": 3.484320557491289e-06,
+      "loss": 0.0829,
+      "step": 1620
+    },
+    {
+      "epoch": 9.34,
+      "learning_rate": 3.193960511033682e-06,
+      "loss": 0.1072,
+      "step": 1630
+    },
+    {
+      "epoch": 9.4,
+      "learning_rate": 2.9036004645760748e-06,
+      "loss": 0.1257,
+      "step": 1640
+    },
+    {
+      "epoch": 9.46,
+      "learning_rate": 2.613240418118467e-06,
+      "loss": 0.0804,
+      "step": 1650
+    },
+    {
+      "epoch": 9.51,
+      "learning_rate": 2.3228803716608596e-06,
+      "loss": 0.119,
+      "step": 1660
+    },
+    {
+      "epoch": 9.57,
+      "learning_rate": 2.0325203252032523e-06,
+      "loss": 0.0765,
+      "step": 1670
+    },
+    {
+      "epoch": 9.63,
+      "learning_rate": 1.7421602787456445e-06,
+      "loss": 0.1681,
+      "step": 1680
+    },
+    {
+      "epoch": 9.68,
+      "learning_rate": 1.4518002322880374e-06,
+      "loss": 0.1005,
+      "step": 1690
+    },
+    {
+      "epoch": 9.74,
+      "learning_rate": 1.1614401858304298e-06,
+      "loss": 0.1498,
+      "step": 1700
+    },
+    {
+      "epoch": 9.8,
+      "learning_rate": 8.710801393728223e-07,
+      "loss": 0.1365,
+      "step": 1710
+    },
+    {
+      "epoch": 9.86,
+      "learning_rate": 5.807200929152149e-07,
+      "loss": 0.122,
+      "step": 1720
+    },
+    {
+      "epoch": 9.91,
+      "learning_rate": 2.9036004645760745e-07,
+      "loss": 0.0906,
+      "step": 1730
+    },
+    {
+      "epoch": 9.97,
+      "learning_rate": 0.0,
+      "loss": 0.1594,
+      "step": 1740
+    },
+    {
+      "epoch": 9.97,
+      "eval_accuracy": 0.9726247987117552,
+      "eval_loss": 0.0668075904250145,
+      "eval_runtime": 9.2098,
+      "eval_samples_per_second": 67.428,
+      "eval_steps_per_second": 8.469,
+      "step": 1740
+    },
+    {
+      "epoch": 9.97,
+      "step": 1740,
+      "total_flos": 1.3833876610752307e+18,
+      "train_loss": 0.24886192696532983,
+      "train_runtime": 1438.0658,
+      "train_samples_per_second": 38.809,
+      "train_steps_per_second": 1.21
     }
   ],
-  "max_steps": 870,
-  "num_train_epochs": 5,
-  "total_flos": 6.917311166047027e+17,
+  "max_steps": 1740,
+  "num_train_epochs": 10,
+  "total_flos": 1.3833876610752307e+18,
   "trial_name": null,
   "trial_params": null
 }