End of training

all_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 5.0,
-    "train_loss": 1.4392181396484376,
-    "train_runtime": 183.8573,
+    "epoch": 3.0,
+    "train_loss": 1.999542236328125,
+    "train_runtime": 244.9371,
     "train_samples": 156,
-    "train_samples_per_second": 4.242,
-    "train_steps_per_second": 0.272
+    "train_samples_per_second": 1.911,
+    "train_steps_per_second": 0.122
 }

train_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 5.0,
-    "train_loss": 1.4392181396484376,
-    "train_runtime": 183.8573,
+    "epoch": 3.0,
+    "train_loss": 1.999542236328125,
+    "train_runtime": 244.9371,
     "train_samples": 156,
-    "train_samples_per_second": 4.242,
-    "train_steps_per_second": 0.272
+    "train_samples_per_second": 1.911,
+    "train_steps_per_second": 0.122
 }

trainer_state.json

@@ -1,8 +1,8 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 5.0,
-  "global_step": 50,
+  "epoch": 3.0,
+  "global_step": 30,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
@@ -17,9 +17,9 @@
       "epoch": 0.1,
       "eval_accuracy": 0.05291319857312723,
       "eval_loss": 2.6484375,
-      "eval_runtime": 1.5576,
-      "eval_samples_per_second": 18.618,
-      "eval_steps_per_second": 1.284,
+      "eval_runtime": 1.1592,
+      "eval_samples_per_second": 25.016,
+      "eval_steps_per_second": 1.725,
       "step": 1
     },
     {
@@ -32,14 +32,14 @@
       "epoch": 0.2,
       "eval_accuracy": 0.05291319857312723,
       "eval_loss": 2.6484375,
-      "eval_runtime": 1.275,
-      "eval_samples_per_second": 22.746,
-      "eval_steps_per_second": 1.569,
+      "eval_runtime": 1.6595,
+      "eval_samples_per_second": 17.475,
+      "eval_steps_per_second": 1.205,
       "step": 2
     },
     {
       "epoch": 0.3,
-      "learning_rate": 2.9970400926424075e-05,
+      "learning_rate": 2.99178284305241e-05,
       "loss": 2.6365,
       "step": 3
     },
@@ -47,729 +47,429 @@
       "epoch": 0.3,
       "eval_accuracy": 0.05601796802748051,
       "eval_loss": 2.55078125,
-      "eval_runtime": 1.1095,
-      "eval_samples_per_second": 26.137,
-      "eval_steps_per_second": 1.803,
+      "eval_runtime": 1.2012,
+      "eval_samples_per_second": 24.142,
+      "eval_steps_per_second": 1.665,
       "step": 3
     },
     {
       "epoch": 0.4,
-      "learning_rate": 2.988172051971717e-05,
+      "learning_rate": 2.9672214011007087e-05,
       "loss": 2.5088,
       "step": 4
     },
     {
       "epoch": 0.4,
-      "eval_accuracy": 0.056282203725723345,
+      "eval_accuracy": 0.05615008587660193,
       "eval_loss": 2.533203125,
-      "eval_runtime": 1.0069,
-      "eval_samples_per_second": 28.8,
-      "eval_steps_per_second": 1.986,
+      "eval_runtime": 0.9032,
+      "eval_samples_per_second": 32.109,
+      "eval_steps_per_second": 2.214,
       "step": 4
     },
     {
       "epoch": 0.5,
-      "learning_rate": 2.9734308760930333e-05,
-      "loss": 2.7297,
+      "learning_rate": 2.9265847744427305e-05,
+      "loss": 2.7307,
       "step": 5
     },
     {
       "epoch": 0.5,
-      "eval_accuracy": 0.05667855727308759,
+      "eval_accuracy": 0.05654643942396618,
       "eval_loss": 2.517578125,
-      "eval_runtime": 1.0127,
-      "eval_samples_per_second": 28.637,
-      "eval_steps_per_second": 1.975,
+      "eval_runtime": 1.6122,
+      "eval_samples_per_second": 17.988,
+      "eval_steps_per_second": 1.241,
       "step": 5
     },
     {
       "epoch": 0.6,
-      "learning_rate": 2.9528747416929467e-05,
-      "loss": 2.9702,
+      "learning_rate": 2.8703181864639013e-05,
+      "loss": 2.969,
       "step": 6
     },
     {
       "epoch": 0.6,
-      "eval_accuracy": 0.05720702866957326,
+      "eval_accuracy": 0.05714096974501255,
       "eval_loss": 2.494140625,
-      "eval_runtime": 1.2159,
-      "eval_samples_per_second": 23.851,
-      "eval_steps_per_second": 1.645,
+      "eval_runtime": 1.5031,
+      "eval_samples_per_second": 19.294,
+      "eval_steps_per_second": 1.331,
       "step": 6
     },
     {
       "epoch": 0.7,
-      "learning_rate": 2.9265847744427305e-05,
-      "loss": 2.729,
+      "learning_rate": 2.7990381056766583e-05,
+      "loss": 2.7283,
       "step": 7
     },
     {
       "epoch": 0.7,
-      "eval_accuracy": 0.05681067512220901,
+      "eval_accuracy": 0.056744616197648305,
       "eval_loss": 2.48828125,
-      "eval_runtime": 1.0262,
-      "eval_samples_per_second": 28.259,
-      "eval_steps_per_second": 1.949,
+      "eval_runtime": 0.999,
+      "eval_samples_per_second": 29.03,
+      "eval_steps_per_second": 2.002,
       "step": 7
     },
     {
       "epoch": 0.8,
-      "learning_rate": 2.894664728832377e-05,
-      "loss": 2.6172,
+      "learning_rate": 2.7135254915624213e-05,
+      "loss": 2.6157,
       "step": 8
     },
     {
       "epoch": 0.8,
       "eval_accuracy": 0.05780155899061963,
-      "eval_loss": 2.478515625,
-      "eval_runtime": 1.3127,
-      "eval_samples_per_second": 22.092,
-      "eval_steps_per_second": 1.524,
+      "eval_loss": 2.4765625,
+      "eval_runtime": 0.9999,
+      "eval_samples_per_second": 29.003,
+      "eval_steps_per_second": 2.0,
       "step": 8
     },
     {
       "epoch": 0.9,
-      "learning_rate": 2.8572405786990293e-05,
-      "loss": 2.6428,
+      "learning_rate": 2.6147172382160913e-05,
+      "loss": 2.6406,
       "step": 9
     },
     {
       "epoch": 0.9,
-      "eval_accuracy": 0.058065794688862464,
+      "eval_accuracy": 0.0583300303871053,
       "eval_loss": 2.458984375,
-      "eval_runtime": 1.0971,
-      "eval_samples_per_second": 26.432,
-      "eval_steps_per_second": 1.823,
+      "eval_runtime": 1.206,
+      "eval_samples_per_second": 24.046,
+      "eval_steps_per_second": 1.658,
       "step": 9
     },
     {
       "epoch": 1.0,
-      "learning_rate": 2.8144600200657953e-05,
-      "loss": 2.5681,
+      "learning_rate": 2.5036959095382875e-05,
+      "loss": 2.5701,
       "step": 10
     },
     {
       "epoch": 1.0,
-      "eval_accuracy": 0.05899061963271238,
-      "eval_loss": 2.435546875,
-      "eval_runtime": 1.3171,
-      "eval_samples_per_second": 22.018,
-      "eval_steps_per_second": 1.518,
+      "eval_accuracy": 0.05872638393446954,
+      "eval_loss": 2.4375,
+      "eval_runtime": 1.0035,
+      "eval_samples_per_second": 28.898,
+      "eval_steps_per_second": 1.993,
       "step": 10
     },
     {
       "epoch": 1.1,
-      "learning_rate": 2.7664918882530227e-05,
-      "loss": 2.1885,
+      "learning_rate": 2.3816778784387097e-05,
+      "loss": 2.2017,
       "step": 11
     },
     {
       "epoch": 1.1,
       "eval_accuracy": 0.05866032500990884,
       "eval_loss": 2.423828125,
-      "eval_runtime": 1.2181,
-      "eval_samples_per_second": 23.807,
-      "eval_steps_per_second": 1.642,
+      "eval_runtime": 1.1973,
+      "eval_samples_per_second": 24.222,
+      "eval_steps_per_second": 1.67,
       "step": 11
     },
     {
       "epoch": 1.2,
-      "learning_rate": 2.7135254915624213e-05,
-      "loss": 1.981,
+      "learning_rate": 2.25e-05,
+      "loss": 2.0039,
       "step": 12
     },
     {
       "epoch": 1.2,
-      "eval_accuracy": 0.05872638393446954,
+      "eval_accuracy": 0.05859426608534813,
       "eval_loss": 2.421875,
-      "eval_runtime": 1.6039,
-      "eval_samples_per_second": 18.081,
-      "eval_steps_per_second": 1.247,
+      "eval_runtime": 1.009,
+      "eval_samples_per_second": 28.742,
+      "eval_steps_per_second": 1.982,
       "step": 12
     },
     {
       "epoch": 1.3,
-      "learning_rate": 2.655769864163684e-05,
-      "loss": 1.8673,
+      "learning_rate": 2.1101049646137008e-05,
+      "loss": 1.8981,
       "step": 13
     },
     {
       "epoch": 1.3,
-      "eval_accuracy": 0.0591227374818338,
-      "eval_loss": 2.41796875,
-      "eval_runtime": 1.0123,
-      "eval_samples_per_second": 28.649,
-      "eval_steps_per_second": 1.976,
+      "eval_accuracy": 0.058858501783590964,
+      "eval_loss": 2.416015625,
+      "eval_runtime": 1.2002,
+      "eval_samples_per_second": 24.163,
+      "eval_steps_per_second": 1.666,
       "step": 13
     },
     {
       "epoch": 1.4,
-      "learning_rate": 2.5934529411321174e-05,
-      "loss": 1.7321,
+      "learning_rate": 1.963525491562421e-05,
+      "loss": 1.7683,
       "step": 14
     },
     {
       "epoch": 1.4,
-      "eval_accuracy": 0.05958514995375875,
-      "eval_loss": 2.41796875,
-      "eval_runtime": 1.0152,
-      "eval_samples_per_second": 28.567,
-      "eval_steps_per_second": 1.97,
+      "eval_accuracy": 0.059453032104637336,
+      "eval_loss": 2.416015625,
+      "eval_runtime": 1.0984,
+      "eval_samples_per_second": 26.402,
+      "eval_steps_per_second": 1.821,
       "step": 14
     },
     {
       "epoch": 1.5,
-      "learning_rate": 2.5268206588930332e-05,
-      "loss": 1.6355,
+      "learning_rate": 1.8118675362266388e-05,
+      "loss": 1.6746,
       "step": 15
     },
     {
       "epoch": 1.5,
-      "eval_accuracy": 0.060113621350244416,
-      "eval_loss": 2.41796875,
-      "eval_runtime": 1.4219,
-      "eval_samples_per_second": 20.396,
-      "eval_steps_per_second": 1.407,
+      "eval_accuracy": 0.059981503501123,
+      "eval_loss": 2.412109375,
+      "eval_runtime": 0.8904,
+      "eval_samples_per_second": 32.57,
+      "eval_steps_per_second": 2.246,
       "step": 15
     },
     {
       "epoch": 1.6,
-      "learning_rate": 2.4561359846230346e-05,
-      "loss": 1.7758,
+      "learning_rate": 1.6567926949014805e-05,
+      "loss": 1.8051,
       "step": 16
     },
     {
       "epoch": 1.6,
-      "eval_accuracy": 0.06017968027480513,
-      "eval_loss": 2.419921875,
-      "eval_runtime": 1.618,
-      "eval_samples_per_second": 17.923,
-      "eval_steps_per_second": 1.236,
+      "eval_accuracy": 0.06004756242568371,
+      "eval_loss": 2.41015625,
+      "eval_runtime": 1.1118,
+      "eval_samples_per_second": 26.085,
+      "eval_steps_per_second": 1.799,
       "step": 16
     },
     {
       "epoch": 1.7,
-      "learning_rate": 2.3816778784387097e-05,
-      "loss": 2.0162,
+      "learning_rate": 1.5e-05,
+      "loss": 2.0457,
       "step": 17
     },
     {
       "epoch": 1.7,
-      "eval_accuracy": 0.06050997489760867,
-      "eval_loss": 2.408203125,
-      "eval_runtime": 1.1225,
-      "eval_samples_per_second": 25.835,
-      "eval_steps_per_second": 1.782,
+      "eval_accuracy": 0.06017968027480513,
+      "eval_loss": 2.404296875,
+      "eval_runtime": 1.1017,
+      "eval_samples_per_second": 26.322,
+      "eval_steps_per_second": 1.815,
       "step": 17
     },
     {
       "epoch": 1.8,
-      "learning_rate": 2.303740192468495e-05,
-      "loss": 1.8037,
+      "learning_rate": 1.3432073050985201e-05,
+      "loss": 1.8257,
       "step": 18
     },
     {
       "epoch": 1.8,
-      "eval_accuracy": 0.06050997489760867,
-      "eval_loss": 2.396484375,
-      "eval_runtime": 0.912,
-      "eval_samples_per_second": 31.8,
-      "eval_steps_per_second": 2.193,
+      "eval_accuracy": 0.060576033822169376,
+      "eval_loss": 2.400390625,
+      "eval_runtime": 1.5124,
+      "eval_samples_per_second": 19.175,
+      "eval_steps_per_second": 1.322,
       "step": 18
     },
     {
       "epoch": 1.9,
-      "learning_rate": 2.222630511152573e-05,
-      "loss": 1.7204,
+      "learning_rate": 1.1881324637733613e-05,
+      "loss": 1.744,
       "step": 19
     },
     {
       "epoch": 1.9,
-      "eval_accuracy": 0.0607742105958515,
-      "eval_loss": 2.375,
-      "eval_runtime": 1.2202,
-      "eval_samples_per_second": 23.767,
-      "eval_steps_per_second": 1.639,
+      "eval_accuracy": 0.06070815167129079,
+      "eval_loss": 2.388671875,
+      "eval_runtime": 0.9026,
+      "eval_samples_per_second": 32.129,
+      "eval_steps_per_second": 2.216,
       "step": 19
     },
     {
       "epoch": 2.0,
-      "learning_rate": 2.138668937347609e-05,
-      "loss": 1.7831,
+      "learning_rate": 1.1881324637733613e-05,
+      "loss": 1.8232,
       "step": 20
     },
     {
       "epoch": 2.0,
-      "eval_accuracy": 0.060906328444972915,
-      "eval_loss": 2.357421875,
-      "eval_runtime": 1.0082,
-      "eval_samples_per_second": 28.765,
-      "eval_steps_per_second": 1.984,
+      "eval_accuracy": 0.06070815167129079,
+      "eval_loss": 2.388671875,
+      "eval_runtime": 1.3479,
+      "eval_samples_per_second": 21.515,
+      "eval_steps_per_second": 1.484,
       "step": 20
     },
     {
       "epoch": 2.1,
-      "learning_rate": 2.052186829027017e-05,
-      "loss": 1.299,
+      "learning_rate": 1.036474508437579e-05,
+      "loss": 1.4741,
       "step": 21
     },
     {
       "epoch": 2.1,
-      "eval_accuracy": 0.06163297661514071,
-      "eval_loss": 2.349609375,
-      "eval_runtime": 1.2114,
-      "eval_samples_per_second": 23.94,
-      "eval_steps_per_second": 1.651,
+      "eval_accuracy": 0.06097238736953362,
+      "eval_loss": 2.3828125,
+      "eval_runtime": 1.1016,
+      "eval_samples_per_second": 26.326,
+      "eval_steps_per_second": 1.816,
       "step": 21
     },
     {
       "epoch": 2.2,
-      "learning_rate": 1.963525491562421e-05,
-      "loss": 1.4463,
+      "learning_rate": 8.898950353863e-06,
+      "loss": 1.651,
       "step": 22
     },
     {
       "epoch": 2.2,
-      "eval_accuracy": 0.06196327123794425,
-      "eval_loss": 2.349609375,
-      "eval_runtime": 1.5179,
-      "eval_samples_per_second": 19.105,
-      "eval_steps_per_second": 1.318,
+      "eval_accuracy": 0.06084026952041221,
+      "eval_loss": 2.376953125,
+      "eval_runtime": 1.0964,
+      "eval_samples_per_second": 26.45,
+      "eval_steps_per_second": 1.824,
       "step": 22
     },
     {
       "epoch": 2.3,
-      "learning_rate": 1.8730348307472828e-05,
-      "loss": 1.1733,
+      "learning_rate": 7.500000000000004e-06,
+      "loss": 1.3732,
       "step": 23
     },
     {
       "epoch": 2.3,
-      "eval_accuracy": 0.061699035539701415,
-      "eval_loss": 2.365234375,
-      "eval_runtime": 1.3077,
-      "eval_samples_per_second": 22.176,
-      "eval_steps_per_second": 1.529,
+      "eval_accuracy": 0.06097238736953362,
+      "eval_loss": 2.373046875,
+      "eval_runtime": 1.1007,
+      "eval_samples_per_second": 26.348,
+      "eval_steps_per_second": 1.817,
       "step": 23
     },
     {
       "epoch": 2.4,
-      "learning_rate": 1.781071971878587e-05,
-      "loss": 1.1142,
+      "learning_rate": 6.1832212156129045e-06,
+      "loss": 1.3151,
       "step": 24
     },
     {
       "epoch": 2.4,
-      "eval_accuracy": 0.06255780155899061,
-      "eval_loss": 2.388671875,
-      "eval_runtime": 1.1042,
-      "eval_samples_per_second": 26.264,
-      "eval_steps_per_second": 1.811,
+      "eval_accuracy": 0.061038446294094335,
+      "eval_loss": 2.373046875,
+      "eval_runtime": 1.0081,
+      "eval_samples_per_second": 28.767,
+      "eval_steps_per_second": 1.984,
       "step": 24
     },
     {
       "epoch": 2.5,
-      "learning_rate": 1.6879998503464565e-05,
-      "loss": 1.3107,
+      "learning_rate": 4.963040904617131e-06,
+      "loss": 1.5302,
       "step": 25
     },
     {
       "epoch": 2.5,
-      "eval_accuracy": 0.06268991940811204,
-      "eval_loss": 2.421875,
-      "eval_runtime": 0.9167,
-      "eval_samples_per_second": 31.634,
-      "eval_steps_per_second": 2.182,
+      "eval_accuracy": 0.061038446294094335,
+      "eval_loss": 2.373046875,
+      "eval_runtime": 1.2116,
+      "eval_samples_per_second": 23.936,
+      "eval_steps_per_second": 1.651,
       "step": 25
     },
     {
       "epoch": 2.6,
-      "learning_rate": 1.5941857792939702e-05,
-      "loss": 1.011,
+      "learning_rate": 3.852827617839085e-06,
+      "loss": 1.2539,
       "step": 26
     },
     {
       "epoch": 2.6,
-      "eval_accuracy": 0.06216144801162637,
-      "eval_loss": 2.455078125,
-      "eval_runtime": 0.9207,
-      "eval_samples_per_second": 31.497,
-      "eval_steps_per_second": 2.172,
+      "eval_accuracy": 0.06117056414321575,
+      "eval_loss": 2.375,
+      "eval_runtime": 1.4027,
+      "eval_samples_per_second": 20.674,
+      "eval_steps_per_second": 1.426,
       "step": 26
     },
     {
       "epoch": 2.7,
-      "learning_rate": 1.5e-05,
-      "loss": 1.3403,
+      "learning_rate": 2.86474508437579e-06,
+      "loss": 1.6211,
       "step": 27
     },
     {
       "epoch": 2.7,
-      "eval_accuracy": 0.061566917690579995,
-      "eval_loss": 2.4765625,
-      "eval_runtime": 1.5266,
-      "eval_samples_per_second": 18.997,
-      "eval_steps_per_second": 1.31,
+      "eval_accuracy": 0.061236623067776455,
+      "eval_loss": 2.376953125,
+      "eval_runtime": 0.9053,
+      "eval_samples_per_second": 32.035,
+      "eval_steps_per_second": 2.209,
       "step": 27
     },
     {
       "epoch": 2.8,
-      "learning_rate": 1.5e-05,
-      "loss": 1.3108,
+      "learning_rate": 2.0096189432334194e-06,
+      "loss": 1.6047,
       "step": 28
     },
     {
       "epoch": 2.8,
-      "eval_accuracy": 0.061566917690579995,
-      "eval_loss": 2.4765625,
-      "eval_runtime": 1.252,
-      "eval_samples_per_second": 23.163,
-      "eval_steps_per_second": 1.597,
+      "eval_accuracy": 0.06130268199233716,
+      "eval_loss": 2.376953125,
+      "eval_runtime": 1.0962,
+      "eval_samples_per_second": 26.456,
+      "eval_steps_per_second": 1.825,
       "step": 28
     },
     {
       "epoch": 2.9,
-      "learning_rate": 1.40581422070603e-05,
-      "loss": 1.0076,
+      "learning_rate": 1.2968181353609854e-06,
+      "loss": 1.1953,
       "step": 29
     },
     {
       "epoch": 2.9,
-      "eval_accuracy": 0.06189721231338354,
-      "eval_loss": 2.4609375,
-      "eval_runtime": 0.9112,
-      "eval_samples_per_second": 31.825,
-      "eval_steps_per_second": 2.195,
+      "eval_accuracy": 0.06143479984145858,
+      "eval_loss": 2.37890625,
+      "eval_runtime": 1.52,
+      "eval_samples_per_second": 19.079,
+      "eval_steps_per_second": 1.316,
       "step": 29
     },
     {
       "epoch": 3.0,
-      "learning_rate": 1.3120001496535434e-05,
-      "loss": 0.8656,
+      "learning_rate": 7.341522555726971e-07,
+      "loss": 1.1621,
       "step": 30
     },
     {
       "epoch": 3.0,
-      "eval_accuracy": 0.062359624785308494,
-      "eval_loss": 2.451171875,
-      "eval_runtime": 1.5156,
-      "eval_samples_per_second": 19.134,
-      "eval_steps_per_second": 1.32,
+      "eval_accuracy": 0.06143479984145858,
+      "eval_loss": 2.37890625,
+      "eval_runtime": 1.3108,
+      "eval_samples_per_second": 22.124,
+      "eval_steps_per_second": 1.526,
       "step": 30
     },
     {
-      "epoch": 3.1,
-      "learning_rate": 1.2189280281214128e-05,
-      "loss": 0.6635,
-      "step": 31
-    },
-    {
-      "epoch": 3.1,
-      "eval_accuracy": 0.06282203725723345,
-      "eval_loss": 2.451171875,
-      "eval_runtime": 1.313,
-      "eval_samples_per_second": 22.087,
-      "eval_steps_per_second": 1.523,
-      "step": 31
-    },
-    {
-      "epoch": 3.2,
-      "learning_rate": 1.1269651692527181e-05,
-      "loss": 0.9996,
-      "step": 32
-    },
-    {
-      "epoch": 3.2,
-      "eval_accuracy": 0.06348262650284053,
-      "eval_loss": 2.443359375,
-      "eval_runtime": 1.01,
-      "eval_samples_per_second": 28.711,
-      "eval_steps_per_second": 1.98,
-      "step": 32
-    },
-    {
-      "epoch": 3.3,
-      "learning_rate": 1.036474508437579e-05,
-      "loss": 0.9029,
-      "step": 33
-    },
-    {
-      "epoch": 3.3,
-      "eval_accuracy": 0.06368080327652266,
-      "eval_loss": 2.447265625,
-      "eval_runtime": 1.4214,
-      "eval_samples_per_second": 20.402,
-      "eval_steps_per_second": 1.407,
-      "step": 33
-    },
-    {
-      "epoch": 3.4,
-      "learning_rate": 9.478131709729831e-06,
-      "loss": 0.8329,
-      "step": 34
-    },
-    {
-      "epoch": 3.4,
-      "eval_accuracy": 0.06374686220108336,
-      "eval_loss": 2.455078125,
-      "eval_runtime": 0.9136,
-      "eval_samples_per_second": 31.742,
-      "eval_steps_per_second": 2.189,
-      "step": 34
-    },
-    {
-      "epoch": 3.5,
-      "learning_rate": 8.61331062652391e-06,
-      "loss": 0.8012,
-      "step": 35
-    },
-    {
-      "epoch": 3.5,
-      "eval_accuracy": 0.06387898005020479,
-      "eval_loss": 2.46484375,
-      "eval_runtime": 1.6062,
-      "eval_samples_per_second": 18.055,
-      "eval_steps_per_second": 1.245,
-      "step": 35
-    },
-    {
-      "epoch": 3.6,
-      "learning_rate": 7.773694888474268e-06,
-      "loss": 0.5814,
-      "step": 36
-    },
-    {
-      "epoch": 3.6,
-      "eval_accuracy": 0.0640110978993262,
-      "eval_loss": 2.490234375,
-      "eval_runtime": 1.6209,
-      "eval_samples_per_second": 17.891,
-      "eval_steps_per_second": 1.234,
-      "step": 36
-    },
-    {
-      "epoch": 3.7,
-      "learning_rate": 6.962598075315047e-06,
-      "loss": 1.0688,
-      "step": 37
-    },
-    {
-      "epoch": 3.7,
-      "eval_accuracy": 0.06381292112564407,
-      "eval_loss": 2.509765625,
-      "eval_runtime": 1.2235,
-      "eval_samples_per_second": 23.703,
-      "eval_steps_per_second": 1.635,
-      "step": 37
-    },
-    {
-      "epoch": 3.8,
-      "learning_rate": 6.1832212156129045e-06,
-      "loss": 0.8688,
-      "step": 38
-    },
-    {
-      "epoch": 3.8,
-      "eval_accuracy": 0.06348262650284053,
-      "eval_loss": 2.517578125,
-      "eval_runtime": 1.5088,
-      "eval_samples_per_second": 19.221,
-      "eval_steps_per_second": 1.326,
-      "step": 38
-    },
-    {
-      "epoch": 3.9,
-      "learning_rate": 5.438640153769654e-06,
-      "loss": 0.7341,
-      "step": 39
-    },
-    {
-      "epoch": 3.9,
-      "eval_accuracy": 0.06381292112564407,
-      "eval_loss": 2.51953125,
-      "eval_runtime": 1.0209,
-      "eval_samples_per_second": 28.406,
-      "eval_steps_per_second": 1.959,
-      "step": 39
-    },
-    {
-      "epoch": 4.0,
-      "learning_rate": 4.731793411069669e-06,
-      "loss": 0.7102,
-      "step": 40
-    },
-    {
-      "epoch": 4.0,
-      "eval_accuracy": 0.0640110978993262,
-      "eval_loss": 2.51953125,
-      "eval_runtime": 1.3112,
-      "eval_samples_per_second": 22.116,
-      "eval_steps_per_second": 1.525,
-      "step": 40
-    },
-    {
-      "epoch": 4.1,
-      "learning_rate": 4.06547058867883e-06,
-      "loss": 0.7079,
-      "step": 41
-    },
-    {
-      "epoch": 4.1,
-      "eval_accuracy": 0.06414321574844761,
-      "eval_loss": 2.51953125,
-      "eval_runtime": 1.0046,
-      "eval_samples_per_second": 28.868,
-      "eval_steps_per_second": 1.991,
-      "step": 41
-    },
-    {
-      "epoch": 4.2,
-      "learning_rate": 3.442301358363163e-06,
-      "loss": 0.7656,
-      "step": 42
-    },
-    {
-      "epoch": 4.2,
-      "eval_accuracy": 0.06427533359756903,
-      "eval_loss": 2.51953125,
-      "eval_runtime": 1.3174,
-      "eval_samples_per_second": 22.012,
-      "eval_steps_per_second": 1.518,
-      "step": 42
-    },
-    {
-      "epoch": 4.3,
-      "learning_rate": 2.86474508437579e-06,
-      "loss": 0.6377,
-      "step": 43
-    },
-    {
-      "epoch": 4.3,
-      "eval_accuracy": 0.06447351037125115,
-      "eval_loss": 2.52734375,
-      "eval_runtime": 1.1043,
-      "eval_samples_per_second": 26.261,
-      "eval_steps_per_second": 1.811,
-      "step": 43
-    },
-    {
-      "epoch": 4.4,
-      "learning_rate": 2.335081117469777e-06,
-      "loss": 0.5898,
-      "step": 44
-    },
-    {
-      "epoch": 4.4,
-      "eval_accuracy": 0.06414321574844761,
-      "eval_loss": 2.53515625,
-      "eval_runtime": 1.0226,
-      "eval_samples_per_second": 28.359,
-      "eval_steps_per_second": 1.956,
-      "step": 44
-    },
-    {
-      "epoch": 4.5,
-      "learning_rate": 1.8553997993420495e-06,
-      "loss": 0.5958,
-      "step": 45
-    },
-    {
-      "epoch": 4.5,
-      "eval_accuracy": 0.06407715682388691,
-      "eval_loss": 2.54296875,
-      "eval_runtime": 0.91,
-      "eval_samples_per_second": 31.867,
-      "eval_steps_per_second": 2.198,
-      "step": 45
-    },
-    {
-      "epoch": 4.6,
-      "learning_rate": 1.4275942130097097e-06,
-      "loss": 0.7048,
-      "step": 46
-    },
-    {
-      "epoch": 4.6,
-      "eval_accuracy": 0.0640110978993262,
-      "eval_loss": 2.548828125,
-      "eval_runtime": 1.2089,
-      "eval_samples_per_second": 23.989,
-      "eval_steps_per_second": 1.654,
-      "step": 46
-    },
-    {
-      "epoch": 4.7,
-      "learning_rate": 1.0533527116762298e-06,
-      "loss": 0.5435,
-      "step": 47
-    },
-    {
-      "epoch": 4.7,
-      "eval_accuracy": 0.06414321574844761,
-      "eval_loss": 2.552734375,
-      "eval_runtime": 1.3311,
-      "eval_samples_per_second": 21.786,
-      "eval_steps_per_second": 1.502,
-      "step": 47
-    },
-    {
-      "epoch": 4.8,
-      "learning_rate": 7.341522555726971e-07,
-      "loss": 0.4769,
-      "step": 48
-    },
-    {
-      "epoch": 4.8,
-      "eval_accuracy": 0.0640110978993262,
-      "eval_loss": 2.552734375,
-      "eval_runtime": 1.0089,
-      "eval_samples_per_second": 28.743,
-      "eval_steps_per_second": 1.982,
-      "step": 48
-    },
-    {
-      "epoch": 4.9,
-      "learning_rate": 4.7125258307053385e-07,
-      "loss": 0.6583,
-      "step": 49
-    },
-    {
-      "epoch": 4.9,
-      "eval_accuracy": 0.06420927467300833,
-      "eval_loss": 2.5546875,
-      "eval_runtime": 1.0099,
-      "eval_samples_per_second": 28.715,
-      "eval_steps_per_second": 1.98,
-      "step": 49
-    },
-    {
-      "epoch": 5.0,
-      "learning_rate": 2.6569123906967083e-07,
-      "loss": 0.7168,
-      "step": 50
-    },
-    {
-      "epoch": 5.0,
-      "eval_accuracy": 0.06414321574844761,
-      "eval_loss": 2.5546875,
-      "eval_runtime": 1.5134,
-      "eval_samples_per_second": 19.163,
-      "eval_steps_per_second": 1.322,
-      "step": 50
-    },
-    {
-      "epoch": 5.0,
-      "step": 50,
-      "total_flos": 2477483753472.0,
-      "train_loss": 1.4392181396484376,
-      "train_runtime": 183.8573,
-      "train_samples_per_second": 4.242,
-      "train_steps_per_second": 0.272
+      "epoch": 3.0,
+      "step": 30,
+      "total_flos": 1466265894912.0,
+      "train_loss": 1.999542236328125,
+      "train_runtime": 244.9371,
+      "train_samples_per_second": 1.911,
+      "train_steps_per_second": 0.122
     }
   ],
-  "max_steps": 50,
-  "num_train_epochs": 5,
-  "total_flos": 2477483753472.0,
+  "max_steps": 30,
+  "num_train_epochs": 3,
+  "total_flos": 1466265894912.0,
   "trial_name": null,
   "trial_params": null
 }