diff --git "a/checkpoint-515/trainer_state.json" "b/checkpoint-515/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/checkpoint-515/trainer_state.json"
@@ -0,0 +1,5492 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 1.0078277886497065,
+  "eval_steps": 5,
+  "global_step": 515,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0019569471624266144,
+      "grad_norm": 3.932948112487793,
+      "learning_rate": 7.8125e-08,
+      "loss": 0.107,
+      "step": 1
+    },
+    {
+      "epoch": 0.003913894324853229,
+      "grad_norm": 4.482716083526611,
+      "learning_rate": 1.5625e-07,
+      "loss": 0.1529,
+      "step": 2
+    },
+    {
+      "epoch": 0.005870841487279843,
+      "grad_norm": 4.672689437866211,
+      "learning_rate": 2.3437500000000003e-07,
+      "loss": 0.1874,
+      "step": 3
+    },
+    {
+      "epoch": 0.007827788649706457,
+      "grad_norm": 4.226949214935303,
+      "learning_rate": 3.125e-07,
+      "loss": 0.1682,
+      "step": 4
+    },
+    {
+      "epoch": 0.009784735812133072,
+      "grad_norm": 4.327479362487793,
+      "learning_rate": 3.90625e-07,
+      "loss": 0.1438,
+      "step": 5
+    },
+    {
+      "epoch": 0.009784735812133072,
+      "eval_loss": 0.1470455378293991,
+      "eval_runtime": 107.3614,
+      "eval_samples_per_second": 28.427,
+      "eval_steps_per_second": 0.224,
+      "eval_sts-test_pearson_cosine": 0.8861388036460539,
+      "eval_sts-test_pearson_dot": 0.8769528313548112,
+      "eval_sts-test_pearson_euclidean": 0.9079831987750276,
+      "eval_sts-test_pearson_manhattan": 0.9086786527495163,
+      "eval_sts-test_pearson_max": 0.9086786527495163,
+      "eval_sts-test_spearman_cosine": 0.9077902566323186,
+      "eval_sts-test_spearman_dot": 0.8794770733264693,
+      "eval_sts-test_spearman_euclidean": 0.903967335376697,
+      "eval_sts-test_spearman_manhattan": 0.9043498244078092,
+      "eval_sts-test_spearman_max": 0.9077902566323186,
+      "step": 5
+    },
+    {
+      "epoch": 0.011741682974559686,
+      "grad_norm": 5.27250337600708,
+      "learning_rate": 4.6875000000000006e-07,
+      "loss": 0.2961,
+      "step": 6
+    },
+    {
+      "epoch": 0.0136986301369863,
+      "grad_norm": 5.903276443481445,
+      "learning_rate": 5.468750000000001e-07,
+      "loss": 0.3019,
+      "step": 7
+    },
+    {
+      "epoch": 0.015655577299412915,
+      "grad_norm": 4.000335693359375,
+      "learning_rate": 6.25e-07,
+      "loss": 0.1184,
+      "step": 8
+    },
+    {
+      "epoch": 0.01761252446183953,
+      "grad_norm": 5.876769065856934,
+      "learning_rate": 7.03125e-07,
+      "loss": 0.3176,
+      "step": 9
+    },
+    {
+      "epoch": 0.019569471624266144,
+      "grad_norm": 4.8437933921813965,
+      "learning_rate": 7.8125e-07,
+      "loss": 0.2234,
+      "step": 10
+    },
+    {
+      "epoch": 0.019569471624266144,
+      "eval_loss": 0.1467687040567398,
+      "eval_runtime": 107.2549,
+      "eval_samples_per_second": 28.456,
+      "eval_steps_per_second": 0.224,
+      "eval_sts-test_pearson_cosine": 0.8861409457129842,
+      "eval_sts-test_pearson_dot": 0.876972814890145,
+      "eval_sts-test_pearson_euclidean": 0.9080268416052204,
+      "eval_sts-test_pearson_manhattan": 0.9087444298597203,
+      "eval_sts-test_pearson_max": 0.9087444298597203,
+      "eval_sts-test_spearman_cosine": 0.9078342918735278,
+      "eval_sts-test_spearman_dot": 0.8794190309404447,
+      "eval_sts-test_spearman_euclidean": 0.9039501508923226,
+      "eval_sts-test_spearman_manhattan": 0.9044244247605487,
+      "eval_sts-test_spearman_max": 0.9078342918735278,
+      "step": 10
+    },
+    {
+      "epoch": 0.021526418786692758,
+      "grad_norm": 4.726498603820801,
+      "learning_rate": 8.59375e-07,
+      "loss": 0.1881,
+      "step": 11
+    },
+    {
+      "epoch": 0.023483365949119372,
+      "grad_norm": 4.818070411682129,
+      "learning_rate": 9.375000000000001e-07,
+      "loss": 0.1593,
+      "step": 12
+    },
+    {
+      "epoch": 0.025440313111545987,
+      "grad_norm": 4.98201322555542,
+      "learning_rate": 1.0156250000000001e-06,
+      "loss": 0.1833,
+      "step": 13
+    },
+    {
+      "epoch": 0.0273972602739726,
+      "grad_norm": 4.269514560699463,
+      "learning_rate": 1.0937500000000001e-06,
+      "loss": 0.1352,
+      "step": 14
+    },
+    {
+      "epoch": 0.029354207436399216,
+      "grad_norm": 6.1525492668151855,
+      "learning_rate": 1.1718750000000001e-06,
+      "loss": 0.3143,
+      "step": 15
+    },
+    {
+      "epoch": 0.029354207436399216,
+      "eval_loss": 0.1462097316980362,
+      "eval_runtime": 107.0721,
+      "eval_samples_per_second": 28.504,
+      "eval_steps_per_second": 0.224,
+      "eval_sts-test_pearson_cosine": 0.8860829119688085,
+      "eval_sts-test_pearson_dot": 0.8768990080043222,
+      "eval_sts-test_pearson_euclidean": 0.9080646402781543,
+      "eval_sts-test_pearson_manhattan": 0.9088063929836994,
+      "eval_sts-test_pearson_max": 0.9088063929836994,
+      "eval_sts-test_spearman_cosine": 0.907713597721555,
+      "eval_sts-test_spearman_dot": 0.8795110842851269,
+      "eval_sts-test_spearman_euclidean": 0.9040110126078148,
+      "eval_sts-test_spearman_manhattan": 0.9045081991218733,
+      "eval_sts-test_spearman_max": 0.907713597721555,
+      "step": 15
+    },
+    {
+      "epoch": 0.03131115459882583,
+      "grad_norm": 4.751354694366455,
+      "learning_rate": 1.25e-06,
+      "loss": 0.1583,
+      "step": 16
+    },
+    {
+      "epoch": 0.033268101761252444,
+      "grad_norm": 5.435980319976807,
+      "learning_rate": 1.328125e-06,
+      "loss": 0.2015,
+      "step": 17
+    },
+    {
+      "epoch": 0.03522504892367906,
+      "grad_norm": 4.1765851974487305,
+      "learning_rate": 1.40625e-06,
+      "loss": 0.1476,
+      "step": 18
+    },
+    {
+      "epoch": 0.03718199608610567,
+      "grad_norm": 4.689794540405273,
+      "learning_rate": 1.484375e-06,
+      "loss": 0.1676,
+      "step": 19
+    },
+    {
+      "epoch": 0.03913894324853229,
+      "grad_norm": 4.203744888305664,
+      "learning_rate": 1.5625e-06,
+      "loss": 0.1525,
+      "step": 20
+    },
+    {
+      "epoch": 0.03913894324853229,
+      "eval_loss": 0.14544810354709625,
+      "eval_runtime": 107.1845,
+      "eval_samples_per_second": 28.474,
+      "eval_steps_per_second": 0.224,
+      "eval_sts-test_pearson_cosine": 0.8861436293943533,
+      "eval_sts-test_pearson_dot": 0.8769239163708102,
+      "eval_sts-test_pearson_euclidean": 0.9082269545633608,
+      "eval_sts-test_pearson_manhattan": 0.9089828403051001,
+      "eval_sts-test_pearson_max": 0.9089828403051001,
+      "eval_sts-test_spearman_cosine": 0.907929343552723,
+      "eval_sts-test_spearman_dot": 0.8796122221358714,
+      "eval_sts-test_spearman_euclidean": 0.9043074002120102,
+      "eval_sts-test_spearman_manhattan": 0.9047217521412333,
+      "eval_sts-test_spearman_max": 0.907929343552723,
+      "step": 20
+    },
+    {
+      "epoch": 0.0410958904109589,
+      "grad_norm": 5.152130603790283,
+      "learning_rate": 1.640625e-06,
+      "loss": 0.1717,
+      "step": 21
+    },
+    {
+      "epoch": 0.043052837573385516,
+      "grad_norm": 5.343059062957764,
+      "learning_rate": 1.71875e-06,
+      "loss": 0.198,
+      "step": 22
+    },
+    {
+      "epoch": 0.04500978473581213,
+      "grad_norm": 5.224748134613037,
+      "learning_rate": 1.796875e-06,
+      "loss": 0.3062,
+      "step": 23
+    },
+    {
+      "epoch": 0.046966731898238745,
+      "grad_norm": 4.6179423332214355,
+      "learning_rate": 1.8750000000000003e-06,
+      "loss": 0.1241,
+      "step": 24
+    },
+    {
+      "epoch": 0.04892367906066536,
+      "grad_norm": 4.200148105621338,
+      "learning_rate": 1.953125e-06,
+      "loss": 0.1087,
+      "step": 25
+    },
+    {
+      "epoch": 0.04892367906066536,
+      "eval_loss": 0.14457188546657562,
+      "eval_runtime": 107.3809,
+      "eval_samples_per_second": 28.422,
+      "eval_steps_per_second": 0.224,
+      "eval_sts-test_pearson_cosine": 0.8862905994058754,
+      "eval_sts-test_pearson_dot": 0.877015249192232,
+      "eval_sts-test_pearson_euclidean": 0.9085054742522269,
+      "eval_sts-test_pearson_manhattan": 0.9092575877809899,
+      "eval_sts-test_pearson_max": 0.9092575877809899,
+      "eval_sts-test_spearman_cosine": 0.9082294902628751,
+      "eval_sts-test_spearman_dot": 0.8798810429630494,
+      "eval_sts-test_spearman_euclidean": 0.9047149499495015,
+      "eval_sts-test_spearman_manhattan": 0.9051023616193669,
+      "eval_sts-test_spearman_max": 0.9082294902628751,
+      "step": 25
+    },
+    {
+      "epoch": 0.050880626223091974,
+      "grad_norm": 4.890737533569336,
+      "learning_rate": 2.0312500000000002e-06,
+      "loss": 0.1767,
+      "step": 26
+    },
+    {
+      "epoch": 0.05283757338551859,
+      "grad_norm": 4.683767795562744,
+      "learning_rate": 2.109375e-06,
+      "loss": 0.1951,
+      "step": 27
+    },
+    {
+      "epoch": 0.0547945205479452,
+      "grad_norm": 4.656280040740967,
+      "learning_rate": 2.1875000000000002e-06,
+      "loss": 0.1621,
+      "step": 28
+    },
+    {
+      "epoch": 0.05675146771037182,
+      "grad_norm": 4.446409702301025,
+      "learning_rate": 2.265625e-06,
+      "loss": 0.221,
+      "step": 29
+    },
+    {
+      "epoch": 0.05870841487279843,
+      "grad_norm": 5.765133857727051,
+      "learning_rate": 2.3437500000000002e-06,
+      "loss": 0.2241,
+      "step": 30
+    },
+    {
+      "epoch": 0.05870841487279843,
+      "eval_loss": 0.14350731670856476,
+      "eval_runtime": 107.3747,
+      "eval_samples_per_second": 28.424,
+      "eval_steps_per_second": 0.224,
+      "eval_sts-test_pearson_cosine": 0.8863784941826807,
+      "eval_sts-test_pearson_dot": 0.8768948467465629,
+      "eval_sts-test_pearson_euclidean": 0.9088066170487232,
+      "eval_sts-test_pearson_manhattan": 0.9095658568102677,
+      "eval_sts-test_pearson_max": 0.9095658568102677,
+      "eval_sts-test_spearman_cosine": 0.9082580415676429,
+      "eval_sts-test_spearman_dot": 0.8801849487791585,
+      "eval_sts-test_spearman_euclidean": 0.9051721735871375,
+      "eval_sts-test_spearman_manhattan": 0.9054862826908437,
+      "eval_sts-test_spearman_max": 0.9082580415676429,
+      "step": 30
+    },
+    {
+      "epoch": 0.060665362035225046,
+      "grad_norm": 5.359245777130127,
+      "learning_rate": 2.421875e-06,
+      "loss": 0.2093,
+      "step": 31
+    },
+    {
+      "epoch": 0.06262230919765166,
+      "grad_norm": 4.439486503601074,
+      "learning_rate": 2.5e-06,
+      "loss": 0.1615,
+      "step": 32
+    },
+    {
+      "epoch": 0.06457925636007827,
+      "grad_norm": 3.689824342727661,
+      "learning_rate": 2.5781250000000004e-06,
+      "loss": 0.1615,
+      "step": 33
+    },
+    {
+      "epoch": 0.06653620352250489,
+      "grad_norm": 4.842885494232178,
+      "learning_rate": 2.65625e-06,
+      "loss": 0.1772,
+      "step": 34
+    },
+    {
+      "epoch": 0.0684931506849315,
+      "grad_norm": 5.209301948547363,
+      "learning_rate": 2.7343750000000004e-06,
+      "loss": 0.2324,
+      "step": 35
+    },
+    {
+      "epoch": 0.0684931506849315,
+      "eval_loss": 0.14226235449314117,
+      "eval_runtime": 107.3108,
+      "eval_samples_per_second": 28.441,
+      "eval_steps_per_second": 0.224,
+      "eval_sts-test_pearson_cosine": 0.8863574366132135,
+      "eval_sts-test_pearson_dot": 0.8765683077424664,
+      "eval_sts-test_pearson_euclidean": 0.9091012263251723,
+      "eval_sts-test_pearson_manhattan": 0.9098631032540263,
+      "eval_sts-test_pearson_max": 0.9098631032540263,
+      "eval_sts-test_spearman_cosine": 0.9083728733043733,
+      "eval_sts-test_spearman_dot": 0.8800282746130272,
+      "eval_sts-test_spearman_euclidean": 0.9052579170039636,
+      "eval_sts-test_spearman_manhattan": 0.9059997586640487,
+      "eval_sts-test_spearman_max": 0.9083728733043733,
+      "step": 35
+    },
+    {
+      "epoch": 0.07045009784735812,
+      "grad_norm": 4.740983009338379,
+      "learning_rate": 2.8125e-06,
+      "loss": 0.2611,
+      "step": 36
+    },
+    {
+      "epoch": 0.07240704500978473,
+      "grad_norm": 5.090059757232666,
+      "learning_rate": 2.8906250000000004e-06,
+      "loss": 0.214,
+      "step": 37
+    },
+    {
+      "epoch": 0.07436399217221135,
+      "grad_norm": 5.123153209686279,
+      "learning_rate": 2.96875e-06,
+      "loss": 0.1985,
+      "step": 38
+    },
+    {
+      "epoch": 0.07632093933463796,
+      "grad_norm": 5.401946067810059,
+      "learning_rate": 3.0468750000000004e-06,
+      "loss": 0.1855,
+      "step": 39
+    },
+    {
+      "epoch": 0.07827788649706457,
+      "grad_norm": 4.838700294494629,
+      "learning_rate": 3.125e-06,
+      "loss": 0.1234,
+      "step": 40
+    },
+    {
+      "epoch": 0.07827788649706457,
+      "eval_loss": 0.14100149273872375,
+      "eval_runtime": 107.3059,
+      "eval_samples_per_second": 28.442,
+      "eval_steps_per_second": 0.224,
+      "eval_sts-test_pearson_cosine": 0.8864265749012155,
+      "eval_sts-test_pearson_dot": 0.8764612424174422,
+      "eval_sts-test_pearson_euclidean": 0.9094092487009695,
+      "eval_sts-test_pearson_manhattan": 0.9101707626021143,
+      "eval_sts-test_pearson_max": 0.9101707626021143,
+      "eval_sts-test_spearman_cosine": 0.908505695048183,
+      "eval_sts-test_spearman_dot": 0.8802103674956289,
+      "eval_sts-test_spearman_euclidean": 0.9054564783507572,
+      "eval_sts-test_spearman_manhattan": 0.9063046490079084,
+      "eval_sts-test_spearman_max": 0.908505695048183,
+      "step": 40
+    },
+    {
+      "epoch": 0.08023483365949119,
+      "grad_norm": 3.8856801986694336,
+      "learning_rate": 3.2031250000000004e-06,
+      "loss": 0.1492,
+      "step": 41
+    },
+    {
+      "epoch": 0.0821917808219178,
+      "grad_norm": 5.678151607513428,
+      "learning_rate": 3.28125e-06,
+      "loss": 0.2022,
+      "step": 42
+    },
+    {
+      "epoch": 0.08414872798434442,
+      "grad_norm": 5.104148864746094,
+      "learning_rate": 3.3593750000000003e-06,
+      "loss": 0.2146,
+      "step": 43
+    },
+    {
+      "epoch": 0.08610567514677103,
+      "grad_norm": 4.76043701171875,
+      "learning_rate": 3.4375e-06,
+      "loss": 0.1688,
+      "step": 44
+    },
+    {
+      "epoch": 0.08806262230919765,
+      "grad_norm": 5.128803730010986,
+      "learning_rate": 3.5156250000000003e-06,
+      "loss": 0.175,
+      "step": 45
+    },
+    {
+      "epoch": 0.08806262230919765,
+      "eval_loss": 0.13962982594966888,
+      "eval_runtime": 107.4144,
+      "eval_samples_per_second": 28.413,
+      "eval_steps_per_second": 0.223,
+      "eval_sts-test_pearson_cosine": 0.886410908658177,
+      "eval_sts-test_pearson_dot": 0.8762836795862763,
+      "eval_sts-test_pearson_euclidean": 0.9096890242379734,
+      "eval_sts-test_pearson_manhattan": 0.9104590803642174,
+      "eval_sts-test_pearson_max": 0.9104590803642174,
+      "eval_sts-test_spearman_cosine": 0.9086694846648755,
+      "eval_sts-test_spearman_dot": 0.8801346931126159,
+      "eval_sts-test_spearman_euclidean": 0.9057376952773407,
+      "eval_sts-test_spearman_manhattan": 0.9064708999439774,
+      "eval_sts-test_spearman_max": 0.9086694846648755,
+      "step": 45
+    },
+    {
+      "epoch": 0.09001956947162426,
+      "grad_norm": 4.968522548675537,
+      "learning_rate": 3.59375e-06,
+      "loss": 0.2123,
+      "step": 46
+    },
+    {
+      "epoch": 0.09197651663405088,
+      "grad_norm": 4.343472957611084,
+      "learning_rate": 3.6718750000000003e-06,
+      "loss": 0.1118,
+      "step": 47
+    },
+    {
+      "epoch": 0.09393346379647749,
+      "grad_norm": 6.252938270568848,
+      "learning_rate": 3.7500000000000005e-06,
+      "loss": 0.3009,
+      "step": 48
+    },
+    {
+      "epoch": 0.0958904109589041,
+      "grad_norm": 3.411029815673828,
+      "learning_rate": 3.828125000000001e-06,
+      "loss": 0.1071,
+      "step": 49
+    },
+    {
+      "epoch": 0.09784735812133072,
+      "grad_norm": 5.379226207733154,
+      "learning_rate": 3.90625e-06,
+      "loss": 0.2608,
+      "step": 50
+    },
+    {
+      "epoch": 0.09784735812133072,
+      "eval_loss": 0.13823722302913666,
+      "eval_runtime": 107.3656,
+      "eval_samples_per_second": 28.426,
+      "eval_steps_per_second": 0.224,
+      "eval_sts-test_pearson_cosine": 0.8863074884351817,
+      "eval_sts-test_pearson_dot": 0.8763122134205692,
+      "eval_sts-test_pearson_euclidean": 0.9097700018848961,
+      "eval_sts-test_pearson_manhattan": 0.9105724410858811,
+      "eval_sts-test_pearson_max": 0.9105724410858811,
+      "eval_sts-test_spearman_cosine": 0.9085105281844131,
+      "eval_sts-test_spearman_dot": 0.8801239975611433,
+      "eval_sts-test_spearman_euclidean": 0.9059798443527296,
+      "eval_sts-test_spearman_manhattan": 0.9065691737139927,
+      "eval_sts-test_spearman_max": 0.9085105281844131,
+      "step": 50
+    },
+    {
+      "epoch": 0.09980430528375733,
+      "grad_norm": 4.599095821380615,
+      "learning_rate": 3.984375e-06,
+      "loss": 0.1368,
+      "step": 51
+    },
+    {
+      "epoch": 0.10176125244618395,
+      "grad_norm": 5.634761333465576,
+      "learning_rate": 4.0625000000000005e-06,
+      "loss": 0.2307,
+      "step": 52
+    },
+    {
+      "epoch": 0.10371819960861056,
+      "grad_norm": 4.678525924682617,
+      "learning_rate": 4.140625000000001e-06,
+      "loss": 0.1366,
+      "step": 53
+    },
+    {
+      "epoch": 0.10567514677103718,
+      "grad_norm": 4.931070327758789,
+      "learning_rate": 4.21875e-06,
+      "loss": 0.1857,
+      "step": 54
+    },
+    {
+      "epoch": 0.10763209393346379,
+      "grad_norm": 4.903087139129639,
+      "learning_rate": 4.296875e-06,
+      "loss": 0.2155,
+      "step": 55
+    },
+    {
+      "epoch": 0.10763209393346379,
+      "eval_loss": 0.1367325782775879,
+      "eval_runtime": 107.3012,
+      "eval_samples_per_second": 28.443,
+      "eval_steps_per_second": 0.224,
+      "eval_sts-test_pearson_cosine": 0.88603017002284,
+      "eval_sts-test_pearson_dot": 0.8761626193697236,
+      "eval_sts-test_pearson_euclidean": 0.9096799681812165,
+      "eval_sts-test_pearson_manhattan": 0.9104977957475867,
+      "eval_sts-test_pearson_max": 0.9104977957475867,
+      "eval_sts-test_spearman_cosine": 0.9084685067499666,
+      "eval_sts-test_spearman_dot": 0.8802836700617878,
+      "eval_sts-test_spearman_euclidean": 0.9058409364373706,
+      "eval_sts-test_spearman_manhattan": 0.9064240006220393,
+      "eval_sts-test_spearman_max": 0.9084685067499666,
+      "step": 55
+    },
+    {
+      "epoch": 0.1095890410958904,
+      "grad_norm": 5.408311367034912,
+      "learning_rate": 4.3750000000000005e-06,
+      "loss": 0.2022,
+      "step": 56
+    },
+    {
+      "epoch": 0.11154598825831702,
+      "grad_norm": 4.5926713943481445,
+      "learning_rate": 4.453125000000001e-06,
+      "loss": 0.2076,
+      "step": 57
+    },
+    {
+      "epoch": 0.11350293542074363,
+      "grad_norm": 6.475535869598389,
+      "learning_rate": 4.53125e-06,
+      "loss": 0.4133,
+      "step": 58
+    },
+    {
+      "epoch": 0.11545988258317025,
+      "grad_norm": 4.997581481933594,
+      "learning_rate": 4.609375e-06,
+      "loss": 0.1823,
+      "step": 59
+    },
+    {
+      "epoch": 0.11741682974559686,
+      "grad_norm": 3.899284601211548,
+      "learning_rate": 4.6875000000000004e-06,
+      "loss": 0.1136,
+      "step": 60
+    },
+    {
+      "epoch": 0.11741682974559686,
+      "eval_loss": 0.13528631627559662,
+      "eval_runtime": 107.3435,
+      "eval_samples_per_second": 28.432,
+      "eval_steps_per_second": 0.224,
+      "eval_sts-test_pearson_cosine": 0.8860224650016398,
+      "eval_sts-test_pearson_dot": 0.8762739756970772,
+      "eval_sts-test_pearson_euclidean": 0.9099016820022997,
+      "eval_sts-test_pearson_manhattan": 0.9107281338135995,
+      "eval_sts-test_pearson_max": 0.9107281338135995,
+      "eval_sts-test_spearman_cosine": 0.9087510214631306,
+      "eval_sts-test_spearman_dot": 0.8808623486228402,
+      "eval_sts-test_spearman_euclidean": 0.9060555634870038,
+      "eval_sts-test_spearman_manhattan": 0.9067256241238172,
+      "eval_sts-test_spearman_max": 0.9087510214631306,
+      "step": 60
+    },
+    {
+      "epoch": 0.11937377690802348,
+      "grad_norm": 4.476404190063477,
+      "learning_rate": 4.765625000000001e-06,
+      "loss": 0.1687,
+      "step": 61
+    },
+    {
+      "epoch": 0.12133072407045009,
+      "grad_norm": 4.893277168273926,
+      "learning_rate": 4.84375e-06,
+      "loss": 0.1591,
+      "step": 62
+    },
+    {
+      "epoch": 0.1232876712328767,
+      "grad_norm": 4.510354042053223,
+      "learning_rate": 4.921875e-06,
+      "loss": 0.1653,
+      "step": 63
+    },
+    {
+      "epoch": 0.12524461839530332,
+      "grad_norm": 4.400285243988037,
+      "learning_rate": 5e-06,
+      "loss": 0.1799,
+      "step": 64
+    },
+    {
+      "epoch": 0.12720156555772993,
+      "grad_norm": 4.631839752197266,
+      "learning_rate": 5.078125000000001e-06,
+      "loss": 0.1578,
+      "step": 65
+    },
+    {
+      "epoch": 0.12720156555772993,
+      "eval_loss": 0.1336735188961029,
+      "eval_runtime": 107.4984,
+      "eval_samples_per_second": 28.391,
+      "eval_steps_per_second": 0.223,
+      "eval_sts-test_pearson_cosine": 0.886014179849858,
+      "eval_sts-test_pearson_dot": 0.8762492282837839,
+      "eval_sts-test_pearson_euclidean": 0.9101155794045166,
+      "eval_sts-test_pearson_manhattan": 0.9109538919103571,
+      "eval_sts-test_pearson_max": 0.9109538919103571,
+      "eval_sts-test_spearman_cosine": 0.9089514176116413,
+      "eval_sts-test_spearman_dot": 0.8810853441583534,
+      "eval_sts-test_spearman_euclidean": 0.9061670836303911,
+      "eval_sts-test_spearman_manhattan": 0.9072153371772234,
+      "eval_sts-test_spearman_max": 0.9089514176116413,
+      "step": 65
+    },
+    {
+      "epoch": 0.12915851272015655,
+      "grad_norm": 4.043459415435791,
+      "learning_rate": 5.156250000000001e-06,
+      "loss": 0.1844,
+      "step": 66
+    },
+    {
+      "epoch": 0.13111545988258316,
+      "grad_norm": 4.447835922241211,
+      "learning_rate": 5.234375e-06,
+      "loss": 0.1489,
+      "step": 67
+    },
+    {
+      "epoch": 0.13307240704500978,
+      "grad_norm": 5.372109889984131,
+      "learning_rate": 5.3125e-06,
+      "loss": 0.1845,
+      "step": 68
+    },
+    {
+      "epoch": 0.1350293542074364,
+      "grad_norm": 3.5112483501434326,
+      "learning_rate": 5.390625000000001e-06,
+      "loss": 0.1364,
+      "step": 69
+    },
+    {
+      "epoch": 0.136986301369863,
+      "grad_norm": 4.305239200592041,
+      "learning_rate": 5.468750000000001e-06,
+      "loss": 0.1584,
+      "step": 70
+    },
+    {
+      "epoch": 0.136986301369863,
+      "eval_loss": 0.1320798397064209,
+      "eval_runtime": 107.505,
+      "eval_samples_per_second": 28.389,
+      "eval_steps_per_second": 0.223,
+      "eval_sts-test_pearson_cosine": 0.88578311613969,
+      "eval_sts-test_pearson_dot": 0.875928774505713,
+      "eval_sts-test_pearson_euclidean": 0.91024619729973,
+      "eval_sts-test_pearson_manhattan": 0.9110959495329505,
+      "eval_sts-test_pearson_max": 0.9110959495329505,
+      "eval_sts-test_spearman_cosine": 0.9086066538938818,
+      "eval_sts-test_spearman_dot": 0.8801235500485294,
+      "eval_sts-test_spearman_euclidean": 0.9060052183179386,
+      "eval_sts-test_spearman_manhattan": 0.907439182986703,
+      "eval_sts-test_spearman_max": 0.9086066538938818,
+      "step": 70
+    },
+    {
+      "epoch": 0.13894324853228962,
+      "grad_norm": 5.093306064605713,
+      "learning_rate": 5.546875e-06,
+      "loss": 0.2279,
+      "step": 71
+    },
+    {
+      "epoch": 0.14090019569471623,
+      "grad_norm": 4.953585147857666,
+      "learning_rate": 5.625e-06,
+      "loss": 0.2028,
+      "step": 72
+    },
+    {
+      "epoch": 0.14285714285714285,
+      "grad_norm": 4.1561102867126465,
+      "learning_rate": 5.7031250000000006e-06,
+      "loss": 0.2291,
+      "step": 73
+    },
+    {
+      "epoch": 0.14481409001956946,
+      "grad_norm": 5.00941801071167,
+      "learning_rate": 5.781250000000001e-06,
+      "loss": 0.2419,
+      "step": 74
+    },
+    {
+      "epoch": 0.14677103718199608,
+      "grad_norm": 3.6476099491119385,
+      "learning_rate": 5.859375e-06,
+      "loss": 0.1329,
+      "step": 75
+    },
+    {
+      "epoch": 0.14677103718199608,
+      "eval_loss": 0.13061992824077606,
+      "eval_runtime": 107.3395,
+      "eval_samples_per_second": 28.433,
+      "eval_steps_per_second": 0.224,
+      "eval_sts-test_pearson_cosine": 0.8854112983780439,
+      "eval_sts-test_pearson_dot": 0.8752625071185561,
+      "eval_sts-test_pearson_euclidean": 0.9103378320010516,
+      "eval_sts-test_pearson_manhattan": 0.9112261622276095,
+      "eval_sts-test_pearson_max": 0.9112261622276095,
+      "eval_sts-test_spearman_cosine": 0.9082604133844965,
+      "eval_sts-test_spearman_dot": 0.8794192099454903,
+      "eval_sts-test_spearman_euclidean": 0.9060063370994732,
+      "eval_sts-test_spearman_manhattan": 0.90766132824825,
+      "eval_sts-test_spearman_max": 0.9082604133844965,
+      "step": 75
+    },
+    {
+      "epoch": 0.1487279843444227,
+      "grad_norm": 4.10636568069458,
+      "learning_rate": 5.9375e-06,
+      "loss": 0.204,
+      "step": 76
+    },
+    {
+      "epoch": 0.1506849315068493,
+      "grad_norm": 4.767779350280762,
+      "learning_rate": 6.0156250000000005e-06,
+      "loss": 0.2239,
+      "step": 77
+    },
+    {
+      "epoch": 0.15264187866927592,
+      "grad_norm": 5.366302490234375,
+      "learning_rate": 6.093750000000001e-06,
+      "loss": 0.2181,
+      "step": 78
+    },
+    {
+      "epoch": 0.15459882583170254,
+      "grad_norm": 4.087960720062256,
+      "learning_rate": 6.171875e-06,
+      "loss": 0.1285,
+      "step": 79
+    },
+    {
+      "epoch": 0.15655577299412915,
+      "grad_norm": 3.7557668685913086,
+      "learning_rate": 6.25e-06,
+      "loss": 0.1067,
+      "step": 80
+    },
+    {
+      "epoch": 0.15655577299412915,
+      "eval_loss": 0.12924787402153015,
+      "eval_runtime": 107.2528,
+      "eval_samples_per_second": 28.456,
+      "eval_steps_per_second": 0.224,
+      "eval_sts-test_pearson_cosine": 0.8850894038300653,
+      "eval_sts-test_pearson_dot": 0.874941916465686,
+      "eval_sts-test_pearson_euclidean": 0.9101863990952803,
+      "eval_sts-test_pearson_manhattan": 0.9110826056950171,
+      "eval_sts-test_pearson_max": 0.9110826056950171,
+      "eval_sts-test_spearman_cosine": 0.9078700928826409,
+      "eval_sts-test_spearman_dot": 0.8792947566875607,
+      "eval_sts-test_spearman_euclidean": 0.9059290069197888,
+      "eval_sts-test_spearman_manhattan": 0.9075206750336968,
+      "eval_sts-test_spearman_max": 0.9078700928826409,
+      "step": 80
+    },
+    {
+      "epoch": 0.15851272015655576,
+      "grad_norm": 3.5708839893341064,
+      "learning_rate": 6.3281250000000005e-06,
+      "loss": 0.1189,
+      "step": 81
+    },
+    {
+      "epoch": 0.16046966731898238,
+      "grad_norm": 4.602839469909668,
+      "learning_rate": 6.406250000000001e-06,
+      "loss": 0.236,
+      "step": 82
+    },
+    {
+      "epoch": 0.162426614481409,
+      "grad_norm": 4.304513931274414,
+      "learning_rate": 6.484375000000001e-06,
+      "loss": 0.1584,
+      "step": 83
+    },
+    {
+      "epoch": 0.1643835616438356,
+      "grad_norm": 4.165163516998291,
+      "learning_rate": 6.5625e-06,
+      "loss": 0.1925,
+      "step": 84
+    },
+    {
+      "epoch": 0.16634050880626222,
+      "grad_norm": 3.9157192707061768,
+      "learning_rate": 6.6406250000000005e-06,
+      "loss": 0.129,
+      "step": 85
+    },
+    {
+      "epoch": 0.16634050880626222,
+      "eval_loss": 0.1278335303068161,
+      "eval_runtime": 107.1978,
+      "eval_samples_per_second": 28.471,
+      "eval_steps_per_second": 0.224,
+      "eval_sts-test_pearson_cosine": 0.8845993101894516,
+      "eval_sts-test_pearson_dot": 0.8740701762146532,
+      "eval_sts-test_pearson_euclidean": 0.9100055922999684,
+      "eval_sts-test_pearson_manhattan": 0.9108899080028133,
+      "eval_sts-test_pearson_max": 0.9108899080028133,
+      "eval_sts-test_spearman_cosine": 0.9078923342595523,
+      "eval_sts-test_spearman_dot": 0.8788126513485913,
+      "eval_sts-test_spearman_euclidean": 0.9057257466905491,
+      "eval_sts-test_spearman_manhattan": 0.9070083178420268,
+      "eval_sts-test_spearman_max": 0.9078923342595523,
+      "step": 85
+    },
+    {
+      "epoch": 0.16829745596868884,
+      "grad_norm": 4.233823776245117,
+      "learning_rate": 6.718750000000001e-06,
+      "loss": 0.1376,
+      "step": 86
+    },
+    {
+      "epoch": 0.17025440313111545,
+      "grad_norm": 4.670790195465088,
+      "learning_rate": 6.796875000000001e-06,
+      "loss": 0.1691,
+      "step": 87
+    },
+    {
+      "epoch": 0.17221135029354206,
+      "grad_norm": 3.742030382156372,
+      "learning_rate": 6.875e-06,
+      "loss": 0.1045,
+      "step": 88
+    },
+    {
+      "epoch": 0.17416829745596868,
+      "grad_norm": 4.242702960968018,
+      "learning_rate": 6.9531250000000004e-06,
+      "loss": 0.165,
+      "step": 89
+    },
+    {
+      "epoch": 0.1761252446183953,
+      "grad_norm": 5.499476909637451,
+      "learning_rate": 7.031250000000001e-06,
+      "loss": 0.2926,
+      "step": 90
+    },
+    {
+      "epoch": 0.1761252446183953,
+      "eval_loss": 0.12669824063777924,
+      "eval_runtime": 107.2778,
+      "eval_samples_per_second": 28.45,
+      "eval_steps_per_second": 0.224,
+      "eval_sts-test_pearson_cosine": 0.8844194771150324,
+      "eval_sts-test_pearson_dot": 0.873458365713796,
+      "eval_sts-test_pearson_euclidean": 0.9099396625521212,
+      "eval_sts-test_pearson_manhattan": 0.910745898918033,
+      "eval_sts-test_pearson_max": 0.910745898918033,
+      "eval_sts-test_spearman_cosine": 0.907622707909669,
+      "eval_sts-test_spearman_dot": 0.8783740442356941,
+      "eval_sts-test_spearman_euclidean": 0.9058808545625318,
+      "eval_sts-test_spearman_manhattan": 0.906889458491771,
+      "eval_sts-test_spearman_max": 0.907622707909669,
+      "step": 90
+    },
+    {
+      "epoch": 0.1780821917808219,
+      "grad_norm": 2.992021083831787,
+      "learning_rate": 7.109375000000001e-06,
+      "loss": 0.1048,
+      "step": 91
+    },
+    {
+      "epoch": 0.18003913894324852,
+      "grad_norm": 4.298286437988281,
+      "learning_rate": 7.1875e-06,
+      "loss": 0.1596,
+      "step": 92
+    },
+    {
+      "epoch": 0.18199608610567514,
+      "grad_norm": 5.210509300231934,
+      "learning_rate": 7.265625e-06,
+      "loss": 0.2474,
+      "step": 93
+    },
+    {
+      "epoch": 0.18395303326810175,
+      "grad_norm": 4.527407169342041,
+      "learning_rate": 7.343750000000001e-06,
+      "loss": 0.1652,
+      "step": 94
+    },
+    {
+      "epoch": 0.18590998043052837,
+      "grad_norm": 5.302050590515137,
+      "learning_rate": 7.421875000000001e-06,
+      "loss": 0.2483,
+      "step": 95
+    },
+    {
+      "epoch": 0.18590998043052837,
+      "eval_loss": 0.1252526491880417,
+      "eval_runtime": 107.5519,
+      "eval_samples_per_second": 28.377,
+      "eval_steps_per_second": 0.223,
+      "eval_sts-test_pearson_cosine": 0.884272350180128,
+      "eval_sts-test_pearson_dot": 0.8727334938335432,
+      "eval_sts-test_pearson_euclidean": 0.9099441972021025,
+      "eval_sts-test_pearson_manhattan": 0.9106991509833859,
+      "eval_sts-test_pearson_max": 0.9106991509833859,
+      "eval_sts-test_spearman_cosine": 0.9075948278738224,
+      "eval_sts-test_spearman_dot": 0.87780624023116,
+      "eval_sts-test_spearman_euclidean": 0.9060086194138042,
+      "eval_sts-test_spearman_manhattan": 0.9069788267607697,
+      "eval_sts-test_spearman_max": 0.9075948278738224,
+      "step": 95
+    },
+    {
+      "epoch": 0.18786692759295498,
+      "grad_norm": 3.690441608428955,
+      "learning_rate": 7.500000000000001e-06,
+      "loss": 0.1623,
+      "step": 96
+    },
+    {
+      "epoch": 0.1898238747553816,
+      "grad_norm": 4.585984706878662,
+      "learning_rate": 7.578125e-06,
+      "loss": 0.1955,
+      "step": 97
+    },
+    {
+      "epoch": 0.1917808219178082,
+      "grad_norm": 4.493942737579346,
+      "learning_rate": 7.656250000000001e-06,
+      "loss": 0.2023,
+      "step": 98
+    },
+    {
+      "epoch": 0.19373776908023482,
+      "grad_norm": 4.569936275482178,
+      "learning_rate": 7.734375e-06,
+      "loss": 0.1886,
+      "step": 99
+    },
+    {
+      "epoch": 0.19569471624266144,
+      "grad_norm": 3.7703664302825928,
+      "learning_rate": 7.8125e-06,
+      "loss": 0.1284,
+      "step": 100
+    },
+    {
+      "epoch": 0.19569471624266144,
+      "eval_loss": 0.12290485948324203,
+      "eval_runtime": 107.6958,
+      "eval_samples_per_second": 28.339,
+      "eval_steps_per_second": 0.223,
+      "eval_sts-test_pearson_cosine": 0.8836376979322419,
+      "eval_sts-test_pearson_dot": 0.8710695777275684,
+      "eval_sts-test_pearson_euclidean": 0.9098265834859519,
+      "eval_sts-test_pearson_manhattan": 0.9106248996071287,
+      "eval_sts-test_pearson_max": 0.9106248996071287,
+      "eval_sts-test_spearman_cosine": 0.9078868298544011,
+      "eval_sts-test_spearman_dot": 0.8773200625274038,
+      "eval_sts-test_spearman_euclidean": 0.9063156130669492,
+      "eval_sts-test_spearman_manhattan": 0.9071474495136926,
+      "eval_sts-test_spearman_max": 0.9078868298544011,
+      "step": 100
+    },
+    {
+      "epoch": 0.19765166340508805,
+      "grad_norm": 4.356619358062744,
+      "learning_rate": 7.890625e-06,
+      "loss": 0.2005,
+      "step": 101
+    },
+    {
+      "epoch": 0.19960861056751467,
+      "grad_norm": 4.293449878692627,
+      "learning_rate": 7.96875e-06,
+      "loss": 0.2301,
+      "step": 102
+    },
+    {
+      "epoch": 0.20156555772994128,
+      "grad_norm": 4.654509544372559,
+      "learning_rate": 8.046875e-06,
+      "loss": 0.2249,
+      "step": 103
+    },
+    {
+      "epoch": 0.2035225048923679,
+      "grad_norm": 4.510340213775635,
+      "learning_rate": 8.125000000000001e-06,
+      "loss": 0.214,
+      "step": 104
+    },
+    {
+      "epoch": 0.2054794520547945,
+      "grad_norm": 3.880908489227295,
+      "learning_rate": 8.203125000000001e-06,
+      "loss": 0.1429,
+      "step": 105
+    },
+    {
+      "epoch": 0.2054794520547945,
+      "eval_loss": 0.12076468020677567,
+      "eval_runtime": 107.7074,
+      "eval_samples_per_second": 28.336,
+      "eval_steps_per_second": 0.223,
+      "eval_sts-test_pearson_cosine": 0.8828542959864998,
+      "eval_sts-test_pearson_dot": 0.8689355363147886,
+      "eval_sts-test_pearson_euclidean": 0.9096459762354197,
+      "eval_sts-test_pearson_manhattan": 0.9104979967855148,
+      "eval_sts-test_pearson_max": 0.9104979967855148,
+      "eval_sts-test_spearman_cosine": 0.9076751563880199,
+      "eval_sts-test_spearman_dot": 0.8750991469270715,
+      "eval_sts-test_spearman_euclidean": 0.906379383614432,
+      "eval_sts-test_spearman_manhattan": 0.9071111562407043,
+      "eval_sts-test_spearman_max": 0.9076751563880199,
+      "step": 105
+    },
+    {
+      "epoch": 0.20743639921722112,
+      "grad_norm": 3.8524463176727295,
+      "learning_rate": 8.281250000000001e-06,
+      "loss": 0.17,
+      "step": 106
+    },
+    {
+      "epoch": 0.20939334637964774,
+      "grad_norm": 4.660905838012695,
+      "learning_rate": 8.359375e-06,
+      "loss": 0.1955,
+      "step": 107
+    },
+    {
+      "epoch": 0.21135029354207435,
+      "grad_norm": 4.391407012939453,
+      "learning_rate": 8.4375e-06,
+      "loss": 0.1964,
+      "step": 108
+    },
+    {
+      "epoch": 0.21330724070450097,
+      "grad_norm": 3.908740758895874,
+      "learning_rate": 8.515625e-06,
+      "loss": 0.1246,
+      "step": 109
+    },
+    {
+      "epoch": 0.21526418786692758,
+      "grad_norm": 3.295600414276123,
+      "learning_rate": 8.59375e-06,
+      "loss": 0.1295,
+      "step": 110
+    },
+    {
+      "epoch": 0.21526418786692758,
+      "eval_loss": 0.11901199817657471,
+      "eval_runtime": 107.5373,
+      "eval_samples_per_second": 28.381,
+      "eval_steps_per_second": 0.223,
+      "eval_sts-test_pearson_cosine": 0.8820675142963768,
+      "eval_sts-test_pearson_dot": 0.8664913359514981,
+      "eval_sts-test_pearson_euclidean": 0.9093761405951237,
+      "eval_sts-test_pearson_manhattan": 0.910248319457324,
+      "eval_sts-test_pearson_max": 0.910248319457324,
+      "eval_sts-test_spearman_cosine": 0.9071699146469111,
+      "eval_sts-test_spearman_dot": 0.8726812810253556,
+      "eval_sts-test_spearman_euclidean": 0.9064896954737618,
+      "eval_sts-test_spearman_manhattan": 0.9068174537121922,
+      "eval_sts-test_spearman_max": 0.9071699146469111,
+      "step": 110
+    },
+    {
+      "epoch": 0.2172211350293542,
+      "grad_norm": 5.0308518409729,
+      "learning_rate": 8.671875e-06,
+      "loss": 0.2203,
+      "step": 111
+    },
+    {
+      "epoch": 0.2191780821917808,
+      "grad_norm": 4.501624584197998,
+      "learning_rate": 8.750000000000001e-06,
+      "loss": 0.2195,
+      "step": 112
+    },
+    {
+      "epoch": 0.22113502935420742,
+      "grad_norm": 4.200097560882568,
+      "learning_rate": 8.828125000000001e-06,
+      "loss": 0.1823,
+      "step": 113
+    },
+    {
+      "epoch": 0.22309197651663404,
+      "grad_norm": 3.6750545501708984,
+      "learning_rate": 8.906250000000001e-06,
+      "loss": 0.174,
+      "step": 114
+    },
+    {
+      "epoch": 0.22504892367906065,
+      "grad_norm": 4.105295181274414,
+      "learning_rate": 8.984375000000002e-06,
+      "loss": 0.207,
+      "step": 115
+    },
+    {
+      "epoch": 0.22504892367906065,
+      "eval_loss": 0.11745984107255936,
+      "eval_runtime": 107.5979,
+      "eval_samples_per_second": 28.365,
+      "eval_steps_per_second": 0.223,
+      "eval_sts-test_pearson_cosine": 0.882042560326929,
+      "eval_sts-test_pearson_dot": 0.8653067979173212,
+      "eval_sts-test_pearson_euclidean": 0.9095832495385563,
+      "eval_sts-test_pearson_manhattan": 0.9103602950988618,
+      "eval_sts-test_pearson_max": 0.9103602950988618,
+      "eval_sts-test_spearman_cosine": 0.9068824772949942,
+      "eval_sts-test_spearman_dot": 0.8714208617482668,
+      "eval_sts-test_spearman_euclidean": 0.906395180809703,
+      "eval_sts-test_spearman_manhattan": 0.9068741088091138,
+      "eval_sts-test_spearman_max": 0.9068824772949942,
+      "step": 115
+    },
+    {
+      "epoch": 0.22700587084148727,
+      "grad_norm": 4.654273509979248,
+      "learning_rate": 9.0625e-06,
+      "loss": 0.2156,
+      "step": 116
+    },
+    {
+      "epoch": 0.22896281800391388,
+      "grad_norm": 4.661588191986084,
+      "learning_rate": 9.140625e-06,
+      "loss": 0.2202,
+      "step": 117
+    },
+    {
+      "epoch": 0.2309197651663405,
+      "grad_norm": 5.366416931152344,
+      "learning_rate": 9.21875e-06,
+      "loss": 0.2718,
+      "step": 118
+    },
+    {
+      "epoch": 0.2328767123287671,
+      "grad_norm": 3.672802448272705,
+      "learning_rate": 9.296875e-06,
+      "loss": 0.1387,
+      "step": 119
+    },
+    {
+      "epoch": 0.23483365949119372,
+      "grad_norm": 3.7878501415252686,
+      "learning_rate": 9.375000000000001e-06,
+      "loss": 0.1506,
+      "step": 120
+    },
+    {
+      "epoch": 0.23483365949119372,
+      "eval_loss": 0.11679373681545258,
+      "eval_runtime": 107.6687,
+      "eval_samples_per_second": 28.346,
+      "eval_steps_per_second": 0.223,
+      "eval_sts-test_pearson_cosine": 0.882107468031623,
+      "eval_sts-test_pearson_dot": 0.8647556765462645,
+      "eval_sts-test_pearson_euclidean": 0.9099443435071429,
+      "eval_sts-test_pearson_manhattan": 0.9105934104125866,
+      "eval_sts-test_pearson_max": 0.9105934104125866,
+      "eval_sts-test_spearman_cosine": 0.9068624287298908,
+      "eval_sts-test_spearman_dot": 0.8710628964083971,
+      "eval_sts-test_spearman_euclidean": 0.906624531024334,
+      "eval_sts-test_spearman_manhattan": 0.9069254385059298,
+      "eval_sts-test_spearman_max": 0.9069254385059298,
+      "step": 120
+    },
+    {
+      "epoch": 0.23679060665362034,
+      "grad_norm": 3.4761197566986084,
+      "learning_rate": 9.453125000000001e-06,
+      "loss": 0.1185,
+      "step": 121
+    },
+    {
+      "epoch": 0.23874755381604695,
+      "grad_norm": 3.9917871952056885,
+      "learning_rate": 9.531250000000001e-06,
+      "loss": 0.1681,
+      "step": 122
+    },
+    {
+      "epoch": 0.24070450097847357,
+      "grad_norm": 4.491674423217773,
+      "learning_rate": 9.609375000000001e-06,
+      "loss": 0.2321,
+      "step": 123
+    },
+    {
+      "epoch": 0.24266144814090018,
+      "grad_norm": 3.903496503829956,
+      "learning_rate": 9.6875e-06,
+      "loss": 0.1457,
+      "step": 124
+    },
+    {
+      "epoch": 0.2446183953033268,
+      "grad_norm": 5.046339988708496,
+      "learning_rate": 9.765625e-06,
+      "loss": 0.2027,
+      "step": 125
+    },
+    {
+      "epoch": 0.2446183953033268,
+      "eval_loss": 0.11647585779428482,
+      "eval_runtime": 107.5396,
+      "eval_samples_per_second": 28.38,
+      "eval_steps_per_second": 0.223,
+      "eval_sts-test_pearson_cosine": 0.8824938293263067,
+      "eval_sts-test_pearson_dot": 0.8653100788410637,
+      "eval_sts-test_pearson_euclidean": 0.9104636052712812,
+      "eval_sts-test_pearson_manhattan": 0.9109341151161342,
+      "eval_sts-test_pearson_max": 0.9109341151161342,
+      "eval_sts-test_spearman_cosine": 0.9070702535877924,
+      "eval_sts-test_spearman_dot": 0.8716920543922986,
+      "eval_sts-test_spearman_euclidean": 0.9070027239343528,
+      "eval_sts-test_spearman_manhattan": 0.9073061822378479,
+      "eval_sts-test_spearman_max": 0.9073061822378479,
+      "step": 125
+    },
+    {
+      "epoch": 0.2465753424657534,
+      "grad_norm": 4.304446697235107,
+      "learning_rate": 9.84375e-06,
+      "loss": 0.1821,
+      "step": 126
+    },
+    {
+      "epoch": 0.24853228962818003,
+      "grad_norm": 3.208357810974121,
+      "learning_rate": 9.921875e-06,
+      "loss": 0.1258,
+      "step": 127
+    },
+    {
+      "epoch": 0.25048923679060664,
+      "grad_norm": 4.275379657745361,
+      "learning_rate": 1e-05,
+      "loss": 0.184,
+      "step": 128
+    },
+    {
+      "epoch": 0.25244618395303325,
+      "grad_norm": 4.408608436584473,
+      "learning_rate": 1.0078125000000001e-05,
+      "loss": 0.2015,
+      "step": 129
+    },
+    {
+      "epoch": 0.25440313111545987,
+      "grad_norm": 3.565253973007202,
+      "learning_rate": 1.0156250000000001e-05,
+      "loss": 0.1323,
+      "step": 130
+    },
+    {
+      "epoch": 0.25440313111545987,
+      "eval_loss": 0.1154385656118393,
+      "eval_runtime": 107.5442,
+      "eval_samples_per_second": 28.379,
+      "eval_steps_per_second": 0.223,
+      "eval_sts-test_pearson_cosine": 0.8820850631122565,
+      "eval_sts-test_pearson_dot": 0.8648589750662984,
+      "eval_sts-test_pearson_euclidean": 0.9105884442785888,
+      "eval_sts-test_pearson_manhattan": 0.9109040210291837,
+      "eval_sts-test_pearson_max": 0.9109040210291837,
+      "eval_sts-test_spearman_cosine": 0.9074317095260507,
+      "eval_sts-test_spearman_dot": 0.8710452196601474,
+      "eval_sts-test_spearman_euclidean": 0.9070635408985837,
+      "eval_sts-test_spearman_manhattan": 0.9074422260724778,
+      "eval_sts-test_spearman_max": 0.9074422260724778,
+      "step": 130
+    },
+    {
+      "epoch": 0.2563600782778865,
+      "grad_norm": 4.261953353881836,
+      "learning_rate": 1.0234375000000001e-05,
+      "loss": 0.1939,
+      "step": 131
+    },
+    {
+      "epoch": 0.2583170254403131,
+      "grad_norm": 3.806480646133423,
+      "learning_rate": 1.0312500000000002e-05,
+      "loss": 0.1428,
+      "step": 132
+    },
+    {
+      "epoch": 0.2602739726027397,
+      "grad_norm": 2.824733257293701,
+      "learning_rate": 1.0390625e-05,
+      "loss": 0.1063,
+      "step": 133
+    },
+    {
+      "epoch": 0.2622309197651663,
+      "grad_norm": 4.076455116271973,
+      "learning_rate": 1.046875e-05,
+      "loss": 0.1602,
+      "step": 134
+    },
+    {
+      "epoch": 0.26418786692759294,
+      "grad_norm": 3.7571659088134766,
+      "learning_rate": 1.0546875e-05,
+      "loss": 0.1814,
+      "step": 135
+    },
+    {
+      "epoch": 0.26418786692759294,
+      "eval_loss": 0.11387230455875397,
+      "eval_runtime": 107.5968,
+      "eval_samples_per_second": 28.365,
+      "eval_steps_per_second": 0.223,
+      "eval_sts-test_pearson_cosine": 0.8812889193869892,
+      "eval_sts-test_pearson_dot": 0.8634898982579755,
+      "eval_sts-test_pearson_euclidean": 0.9104977472627025,
+      "eval_sts-test_pearson_manhattan": 0.9107178140804983,
+      "eval_sts-test_pearson_max": 0.9107178140804983,
+      "eval_sts-test_spearman_cosine": 0.9066986391131981,
+      "eval_sts-test_spearman_dot": 0.870129116588204,
+      "eval_sts-test_spearman_euclidean": 0.9070359293703052,
+      "eval_sts-test_spearman_manhattan": 0.9073414909830857,
+      "eval_sts-test_spearman_max": 0.9073414909830857,
+      "step": 135
+    },
+    {
+      "epoch": 0.26614481409001955,
+      "grad_norm": 3.864948034286499,
+      "learning_rate": 1.0625e-05,
+      "loss": 0.1518,
+      "step": 136
+    },
+    {
+      "epoch": 0.26810176125244617,
+      "grad_norm": 3.5900001525878906,
+      "learning_rate": 1.0703125000000001e-05,
+      "loss": 0.1379,
+      "step": 137
+    },
+    {
+      "epoch": 0.2700587084148728,
+      "grad_norm": 4.291954517364502,
+      "learning_rate": 1.0781250000000001e-05,
+      "loss": 0.1708,
+      "step": 138
+    },
+    {
+      "epoch": 0.2720156555772994,
+      "grad_norm": 3.8340342044830322,
+      "learning_rate": 1.0859375000000001e-05,
+      "loss": 0.2046,
+      "step": 139
+    },
+    {
+      "epoch": 0.273972602739726,
+      "grad_norm": 3.749396562576294,
+      "learning_rate": 1.0937500000000002e-05,
+      "loss": 0.1259,
+      "step": 140
+    },
+    {
+      "epoch": 0.273972602739726,
+      "eval_loss": 0.1124362125992775,
+      "eval_runtime": 107.5142,
+      "eval_samples_per_second": 28.387,
+      "eval_steps_per_second": 0.223,
+      "eval_sts-test_pearson_cosine": 0.8805714116282963,
+      "eval_sts-test_pearson_dot": 0.8618911680351633,
+      "eval_sts-test_pearson_euclidean": 0.9102979980912764,
+      "eval_sts-test_pearson_manhattan": 0.9105232760600299,
+      "eval_sts-test_pearson_max": 0.9105232760600299,
+      "eval_sts-test_spearman_cosine": 0.9063180743863257,
+      "eval_sts-test_spearman_dot": 0.8687826406354595,
+      "eval_sts-test_spearman_euclidean": 0.9070556199253175,
+      "eval_sts-test_spearman_manhattan": 0.9073570196707885,
+      "eval_sts-test_spearman_max": 0.9073570196707885,
+      "step": 140
+    },
+    {
+      "epoch": 0.2759295499021526,
+      "grad_norm": 2.8815276622772217,
+      "learning_rate": 1.1015625e-05,
+      "loss": 0.1181,
+      "step": 141
+    },
+    {
+      "epoch": 0.27788649706457924,
+      "grad_norm": 3.766554355621338,
+      "learning_rate": 1.109375e-05,
+      "loss": 0.2144,
+      "step": 142
+    },
+    {
+      "epoch": 0.27984344422700586,
+      "grad_norm": 4.289268493652344,
+      "learning_rate": 1.1171875e-05,
+      "loss": 0.1822,
+      "step": 143
+    },
+    {
+      "epoch": 0.28180039138943247,
+      "grad_norm": 3.9036617279052734,
+      "learning_rate": 1.125e-05,
+      "loss": 0.1667,
+      "step": 144
+    },
+    {
+      "epoch": 0.2837573385518591,
+      "grad_norm": 3.321366786956787,
+      "learning_rate": 1.1328125000000001e-05,
+      "loss": 0.0779,
+      "step": 145
+    },
+    {
+      "epoch": 0.2837573385518591,
+      "eval_loss": 0.1118142157793045,
+      "eval_runtime": 107.3173,
+      "eval_samples_per_second": 28.439,
+      "eval_steps_per_second": 0.224,
+      "eval_sts-test_pearson_cosine": 0.8796044904115364,
+      "eval_sts-test_pearson_dot": 0.8607678603166254,
+      "eval_sts-test_pearson_euclidean": 0.9097479995877322,
+      "eval_sts-test_pearson_manhattan": 0.9098650580518599,
+      "eval_sts-test_pearson_max": 0.9098650580518599,
+      "eval_sts-test_spearman_cosine": 0.9059690592987342,
+      "eval_sts-test_spearman_dot": 0.8685229490656053,
+      "eval_sts-test_spearman_euclidean": 0.90680836920613,
+      "eval_sts-test_spearman_manhattan": 0.9069437865231001,
+      "eval_sts-test_spearman_max": 0.9069437865231001,
+      "step": 145
+    },
+    {
+      "epoch": 0.2857142857142857,
+      "grad_norm": 3.460301160812378,
+      "learning_rate": 1.1406250000000001e-05,
+      "loss": 0.147,
+      "step": 146
+    },
+    {
+      "epoch": 0.2876712328767123,
+      "grad_norm": 3.8999266624450684,
+      "learning_rate": 1.1484375000000001e-05,
+      "loss": 0.1913,
+      "step": 147
+    },
+    {
+      "epoch": 0.2896281800391389,
+      "grad_norm": 3.539788007736206,
+      "learning_rate": 1.1562500000000002e-05,
+      "loss": 0.1357,
+      "step": 148
+    },
+    {
+      "epoch": 0.29158512720156554,
+      "grad_norm": 3.499439001083374,
+      "learning_rate": 1.1640625000000002e-05,
+      "loss": 0.1128,
+      "step": 149
+    },
+    {
+      "epoch": 0.29354207436399216,
+      "grad_norm": 3.2960240840911865,
+      "learning_rate": 1.171875e-05,
+      "loss": 0.0996,
+      "step": 150
+    },
+    {
+      "epoch": 0.29354207436399216,
+      "eval_loss": 0.11132737249135971,
+      "eval_runtime": 107.5867,
+      "eval_samples_per_second": 28.368,
+      "eval_steps_per_second": 0.223,
+      "eval_sts-test_pearson_cosine": 0.8787852416493207,
+      "eval_sts-test_pearson_dot": 0.8593025559452621,
+      "eval_sts-test_pearson_euclidean": 0.9091617970047303,
+      "eval_sts-test_pearson_manhattan": 0.9091664157178929,
+      "eval_sts-test_pearson_max": 0.9091664157178929,
+      "eval_sts-test_spearman_cosine": 0.9054375485671886,
+      "eval_sts-test_spearman_dot": 0.867029912731804,
+      "eval_sts-test_spearman_euclidean": 0.9062253050214613,
+      "eval_sts-test_spearman_manhattan": 0.9062610165280517,
+      "eval_sts-test_spearman_max": 0.9062610165280517,
+      "step": 150
+    },
+    {
+      "epoch": 0.29549902152641877,
+      "grad_norm": 4.271719932556152,
+      "learning_rate": 1.1796875e-05,
+      "loss": 0.1956,
+      "step": 151
+    },
+    {
+      "epoch": 0.2974559686888454,
+      "grad_norm": 3.168663501739502,
+      "learning_rate": 1.1875e-05,
+      "loss": 0.0942,
+      "step": 152
+    },
+    {
+      "epoch": 0.299412915851272,
+      "grad_norm": 3.816993236541748,
+      "learning_rate": 1.1953125000000001e-05,
+      "loss": 0.1406,
+      "step": 153
+    },
+    {
+      "epoch": 0.3013698630136986,
+      "grad_norm": 5.383023738861084,
+      "learning_rate": 1.2031250000000001e-05,
+      "loss": 0.2868,
+      "step": 154
+    },
+    {
+      "epoch": 0.30332681017612523,
+      "grad_norm": 3.123462677001953,
+      "learning_rate": 1.2109375000000001e-05,
+      "loss": 0.1102,
+      "step": 155
+    },
+    {
+      "epoch": 0.30332681017612523,
+      "eval_loss": 0.11142811924219131,
+      "eval_runtime": 107.3019,
+      "eval_samples_per_second": 28.443,
+      "eval_steps_per_second": 0.224,
+      "eval_sts-test_pearson_cosine": 0.8780761726881443,
+      "eval_sts-test_pearson_dot": 0.8581767032057357,
+      "eval_sts-test_pearson_euclidean": 0.9081534036571242,
+      "eval_sts-test_pearson_manhattan": 0.9081724370385316,
+      "eval_sts-test_pearson_max": 0.9081724370385316,
+      "eval_sts-test_spearman_cosine": 0.9048428490545583,
+      "eval_sts-test_spearman_dot": 0.8670075818523697,
+      "eval_sts-test_spearman_euclidean": 0.9052714766361651,
+      "eval_sts-test_spearman_manhattan": 0.9054467225757737,
+      "eval_sts-test_spearman_max": 0.9054467225757737,
+      "step": 155
+    },
+    {
+      "epoch": 0.30528375733855184,
+      "grad_norm": 4.1034979820251465,
+      "learning_rate": 1.2187500000000001e-05,
+      "loss": 0.1659,
+      "step": 156
+    },
+    {
+      "epoch": 0.30724070450097846,
+      "grad_norm": 3.60249400138855,
+      "learning_rate": 1.2265625000000002e-05,
+      "loss": 0.1645,
+      "step": 157
+    },
+    {
+      "epoch": 0.30919765166340507,
+      "grad_norm": 3.771853446960449,
+      "learning_rate": 1.234375e-05,
+      "loss": 0.151,
+      "step": 158
+    },
+    {
+      "epoch": 0.3111545988258317,
+      "grad_norm": 4.291686058044434,
+      "learning_rate": 1.2421875e-05,
+      "loss": 0.158,
+      "step": 159
+    },
+    {
+      "epoch": 0.3131115459882583,
+      "grad_norm": 5.1689453125,
+      "learning_rate": 1.25e-05,
+      "loss": 0.2323,
+      "step": 160
+    },
+    {
+      "epoch": 0.3131115459882583,
+      "eval_loss": 0.11126424372196198,
+      "eval_runtime": 107.301,
+      "eval_samples_per_second": 28.443,
+      "eval_steps_per_second": 0.224,
+      "eval_sts-test_pearson_cosine": 0.8777597983330929,
+      "eval_sts-test_pearson_dot": 0.8577739588604719,
+      "eval_sts-test_pearson_euclidean": 0.9075483317216817,
+      "eval_sts-test_pearson_manhattan": 0.9075908461381532,
+      "eval_sts-test_pearson_max": 0.9075908461381532,
+      "eval_sts-test_spearman_cosine": 0.9047649818597372,
+      "eval_sts-test_spearman_dot": 0.867389712873391,
+      "eval_sts-test_spearman_euclidean": 0.9048189966322366,
+      "eval_sts-test_spearman_manhattan": 0.9049692713679889,
+      "eval_sts-test_spearman_max": 0.9049692713679889,
+      "step": 160
+    },
+    {
+      "epoch": 0.3150684931506849,
+      "grad_norm": 3.304703712463379,
+      "learning_rate": 1.2578125e-05,
+      "loss": 0.1157,
+      "step": 161
+    },
+    {
+      "epoch": 0.31702544031311153,
+      "grad_norm": 4.064731121063232,
+      "learning_rate": 1.2656250000000001e-05,
+      "loss": 0.1507,
+      "step": 162
+    },
+    {
+      "epoch": 0.31898238747553814,
+      "grad_norm": 4.615545749664307,
+      "learning_rate": 1.2734375000000001e-05,
+      "loss": 0.1879,
+      "step": 163
+    },
+    {
+      "epoch": 0.32093933463796476,
+      "grad_norm": 3.767533540725708,
+      "learning_rate": 1.2812500000000001e-05,
+      "loss": 0.143,
+      "step": 164
+    },
+    {
+      "epoch": 0.32289628180039137,
+      "grad_norm": 4.727967262268066,
+      "learning_rate": 1.2890625000000002e-05,
+      "loss": 0.2227,
+      "step": 165
+    },
+    {
+      "epoch": 0.32289628180039137,
+      "eval_loss": 0.11155427247285843,
+      "eval_runtime": 107.2898,
+      "eval_samples_per_second": 28.446,
+      "eval_steps_per_second": 0.224,
+      "eval_sts-test_pearson_cosine": 0.8775899700998113,
+      "eval_sts-test_pearson_dot": 0.8571711542435376,
+      "eval_sts-test_pearson_euclidean": 0.907399950708088,
+      "eval_sts-test_pearson_manhattan": 0.9073879045697356,
+      "eval_sts-test_pearson_max": 0.907399950708088,
+      "eval_sts-test_spearman_cosine": 0.9049959431197784,
+      "eval_sts-test_spearman_dot": 0.8667648957618442,
+      "eval_sts-test_spearman_euclidean": 0.9048916279294749,
+      "eval_sts-test_spearman_manhattan": 0.9050786882020909,
+      "eval_sts-test_spearman_max": 0.9050786882020909,
+      "step": 165
+    },
+    {
+      "epoch": 0.324853228962818,
+      "grad_norm": 4.0150017738342285,
+      "learning_rate": 1.2968750000000002e-05,
+      "loss": 0.1624,
+      "step": 166
+    },
+    {
+      "epoch": 0.3268101761252446,
+      "grad_norm": 3.021153450012207,
+      "learning_rate": 1.3046875e-05,
+      "loss": 0.1345,
+      "step": 167
+    },
+    {
+      "epoch": 0.3287671232876712,
+      "grad_norm": 3.869710922241211,
+      "learning_rate": 1.3125e-05,
+      "loss": 0.1765,
+      "step": 168
+    },
+    {
+      "epoch": 0.33072407045009783,
+      "grad_norm": 3.538076162338257,
+      "learning_rate": 1.3203125e-05,
+      "loss": 0.1368,
+      "step": 169
+    },
+    {
+      "epoch": 0.33268101761252444,
+      "grad_norm": 3.378551483154297,
+      "learning_rate": 1.3281250000000001e-05,
+      "loss": 0.0962,
+      "step": 170
+    },
+    {
+      "epoch": 0.33268101761252444,
+      "eval_loss": 0.11131894588470459,
+      "eval_runtime": 107.3532,
+      "eval_samples_per_second": 28.43,
+      "eval_steps_per_second": 0.224,
+      "eval_sts-test_pearson_cosine": 0.8782576778514848,
+      "eval_sts-test_pearson_dot": 0.8576530243239538,
+      "eval_sts-test_pearson_euclidean": 0.9077401564122008,
+      "eval_sts-test_pearson_manhattan": 0.907609849534313,
+      "eval_sts-test_pearson_max": 0.9077401564122008,
+      "eval_sts-test_spearman_cosine": 0.9055560946586144,
+      "eval_sts-test_spearman_dot": 0.8666707838591381,
+      "eval_sts-test_spearman_euclidean": 0.9054064016892602,
+      "eval_sts-test_spearman_manhattan": 0.9054834186101147,
+      "eval_sts-test_spearman_max": 0.9055560946586144,
+      "step": 170
+    },
+    {
+      "epoch": 0.33463796477495106,
+      "grad_norm": 4.588249683380127,
+      "learning_rate": 1.3359375000000001e-05,
+      "loss": 0.1783,
+      "step": 171
+    },
+    {
+      "epoch": 0.33659491193737767,
+      "grad_norm": 4.370199680328369,
+      "learning_rate": 1.3437500000000001e-05,
+      "loss": 0.2019,
+      "step": 172
+    },
+    {
+      "epoch": 0.3385518590998043,
+      "grad_norm": 4.000157356262207,
+      "learning_rate": 1.3515625000000002e-05,
+      "loss": 0.1761,
+      "step": 173
+    },
+    {
+      "epoch": 0.3405088062622309,
+      "grad_norm": 4.3335862159729,
+      "learning_rate": 1.3593750000000002e-05,
+      "loss": 0.1855,
+      "step": 174
+    },
+    {
+      "epoch": 0.3424657534246575,
+      "grad_norm": 4.247244358062744,
+      "learning_rate": 1.3671875e-05,
+      "loss": 0.1922,
+      "step": 175
+    },
+    {
+      "epoch": 0.3424657534246575,
+      "eval_loss": 0.1105586364865303,
+      "eval_runtime": 107.3507,
+      "eval_samples_per_second": 28.43,
+      "eval_steps_per_second": 0.224,
+      "eval_sts-test_pearson_cosine": 0.8775475016000731,
+      "eval_sts-test_pearson_dot": 0.8543732981082479,
+      "eval_sts-test_pearson_euclidean": 0.9076643456809551,
+      "eval_sts-test_pearson_manhattan": 0.9075054089199206,
+      "eval_sts-test_pearson_max": 0.9076643456809551,
+      "eval_sts-test_spearman_cosine": 0.905357578063082,
+      "eval_sts-test_spearman_dot": 0.8628476388472094,
+      "eval_sts-test_spearman_euclidean": 0.9054710672619708,
+      "eval_sts-test_spearman_manhattan": 0.9055309444497123,
+      "eval_sts-test_spearman_max": 0.9055309444497123,
+      "step": 175
+    },
+    {
+      "epoch": 0.34442270058708413,
+      "grad_norm": 3.881108522415161,
+      "learning_rate": 1.375e-05,
+      "loss": 0.1538,
+      "step": 176
+    },
+    {
+      "epoch": 0.34637964774951074,
+      "grad_norm": 3.4271416664123535,
+      "learning_rate": 1.3828125e-05,
+      "loss": 0.1049,
+      "step": 177
+    },
+    {
+      "epoch": 0.34833659491193736,
+      "grad_norm": 3.7847940921783447,
+      "learning_rate": 1.3906250000000001e-05,
+      "loss": 0.1619,
+      "step": 178
+    },
+    {
+      "epoch": 0.350293542074364,
+      "grad_norm": 2.3725311756134033,
+      "learning_rate": 1.3984375000000001e-05,
+      "loss": 0.0731,
+      "step": 179
+    },
+    {
+      "epoch": 0.3522504892367906,
+      "grad_norm": 3.6820032596588135,
+      "learning_rate": 1.4062500000000001e-05,
+      "loss": 0.1205,
+      "step": 180
+    },
+    {
+      "epoch": 0.3522504892367906,
+      "eval_loss": 0.10974939167499542,
+      "eval_runtime": 107.353,
+      "eval_samples_per_second": 28.43,
+      "eval_steps_per_second": 0.224,
+      "eval_sts-test_pearson_cosine": 0.8782123578217031,
+      "eval_sts-test_pearson_dot": 0.852106566478191,
+      "eval_sts-test_pearson_euclidean": 0.9088860377565003,
+      "eval_sts-test_pearson_manhattan": 0.9087269620613702,
+      "eval_sts-test_pearson_max": 0.9088860377565003,
+      "eval_sts-test_spearman_cosine": 0.9058966517578029,
+      "eval_sts-test_spearman_dot": 0.8595467858069799,
+      "eval_sts-test_spearman_euclidean": 0.9064047128283795,
+      "eval_sts-test_spearman_manhattan": 0.9067846510375924,
+      "eval_sts-test_spearman_max": 0.9067846510375924,
+      "step": 180
+    },
+    {
+      "epoch": 0.3542074363992172,
+      "grad_norm": 3.7714688777923584,
+      "learning_rate": 1.4140625000000002e-05,
+      "loss": 0.169,
+      "step": 181
+    },
+    {
+      "epoch": 0.3561643835616438,
+      "grad_norm": 3.7113559246063232,
+      "learning_rate": 1.4218750000000002e-05,
+      "loss": 0.1688,
+      "step": 182
+    },
+    {
+      "epoch": 0.35812133072407043,
+      "grad_norm": 3.1639597415924072,
+      "learning_rate": 1.4296875000000002e-05,
+      "loss": 0.1274,
+      "step": 183
+    },
+    {
+      "epoch": 0.36007827788649704,
+      "grad_norm": 4.144288539886475,
+      "learning_rate": 1.4375e-05,
+      "loss": 0.1477,
+      "step": 184
+    },
+    {
+      "epoch": 0.36203522504892366,
+      "grad_norm": 3.4342098236083984,
+      "learning_rate": 1.4453125e-05,
+      "loss": 0.1418,
+      "step": 185
+    },
+    {
+      "epoch": 0.36203522504892366,
+      "eval_loss": 0.10942607372999191,
+      "eval_runtime": 107.2679,
+      "eval_samples_per_second": 28.452,
+      "eval_steps_per_second": 0.224,
+      "eval_sts-test_pearson_cosine": 0.8778855142398189,
+      "eval_sts-test_pearson_dot": 0.8501658695420333,
+      "eval_sts-test_pearson_euclidean": 0.9088432870055996,
+      "eval_sts-test_pearson_manhattan": 0.9086435133118579,
+      "eval_sts-test_pearson_max": 0.9088432870055996,
+      "eval_sts-test_spearman_cosine": 0.9055185931015683,
+      "eval_sts-test_spearman_dot": 0.8575025481866207,
+      "eval_sts-test_spearman_euclidean": 0.9063994321795352,
+      "eval_sts-test_spearman_manhattan": 0.9064969899293684,
+      "eval_sts-test_spearman_max": 0.9064969899293684,
+      "step": 185
+    },
+    {
+      "epoch": 0.3639921722113503,
+      "grad_norm": 4.744626045227051,
+      "learning_rate": 1.453125e-05,
+      "loss": 0.2477,
+      "step": 186
+    },
+    {
+      "epoch": 0.3659491193737769,
+      "grad_norm": 4.062248229980469,
+      "learning_rate": 1.4609375000000001e-05,
+      "loss": 0.1713,
+      "step": 187
+    },
+    {
+      "epoch": 0.3679060665362035,
+      "grad_norm": 3.989694833755493,
+      "learning_rate": 1.4687500000000001e-05,
+      "loss": 0.1703,
+      "step": 188
+    },
+    {
+      "epoch": 0.3698630136986301,
+      "grad_norm": 3.3543660640716553,
+      "learning_rate": 1.4765625000000001e-05,
+      "loss": 0.1176,
+      "step": 189
+    },
+    {
+      "epoch": 0.37181996086105673,
+      "grad_norm": 4.307045936584473,
+      "learning_rate": 1.4843750000000002e-05,
+      "loss": 0.1811,
+      "step": 190
+    },
+    {
+      "epoch": 0.37181996086105673,
+      "eval_loss": 0.10837770998477936,
+      "eval_runtime": 107.3429,
+      "eval_samples_per_second": 28.432,
+      "eval_steps_per_second": 0.224,
+      "eval_sts-test_pearson_cosine": 0.8774103555789884,
+      "eval_sts-test_pearson_dot": 0.84942827650618,
+      "eval_sts-test_pearson_euclidean": 0.9086430009253119,
+      "eval_sts-test_pearson_manhattan": 0.9084642534632353,
+      "eval_sts-test_pearson_max": 0.9086430009253119,
+      "eval_sts-test_spearman_cosine": 0.9048482639571866,
+      "eval_sts-test_spearman_dot": 0.8562155914115267,
+      "eval_sts-test_spearman_euclidean": 0.9060070531196555,
+      "eval_sts-test_spearman_manhattan": 0.9061608184537963,
+      "eval_sts-test_spearman_max": 0.9061608184537963,
+      "step": 190
+    },
+    {
+      "epoch": 0.37377690802348335,
+      "grad_norm": 4.140930652618408,
+      "learning_rate": 1.4921875000000002e-05,
+      "loss": 0.162,
+      "step": 191
+    },
+    {
+      "epoch": 0.37573385518590996,
+      "grad_norm": 2.7555642127990723,
+      "learning_rate": 1.5000000000000002e-05,
+      "loss": 0.1141,
+      "step": 192
+    },
+    {
+      "epoch": 0.3776908023483366,
+      "grad_norm": 4.070343017578125,
+      "learning_rate": 1.5078125e-05,
+      "loss": 0.154,
+      "step": 193
+    },
+    {
+      "epoch": 0.3796477495107632,
+      "grad_norm": 4.453440189361572,
+      "learning_rate": 1.515625e-05,
+      "loss": 0.2461,
+      "step": 194
+    },
+    {
+      "epoch": 0.3816046966731898,
+      "grad_norm": 3.7656772136688232,
+      "learning_rate": 1.5234375000000001e-05,
+      "loss": 0.1573,
+      "step": 195
+    },
+    {
+      "epoch": 0.3816046966731898,
+      "eval_loss": 0.10762027651071548,
+      "eval_runtime": 107.299,
+      "eval_samples_per_second": 28.444,
+      "eval_steps_per_second": 0.224,
+      "eval_sts-test_pearson_cosine": 0.8779461080888007,
+      "eval_sts-test_pearson_dot": 0.8521074278329072,
+      "eval_sts-test_pearson_euclidean": 0.9087045359990432,
+      "eval_sts-test_pearson_manhattan": 0.9086340705654771,
+      "eval_sts-test_pearson_max": 0.9087045359990432,
+      "eval_sts-test_spearman_cosine": 0.9045706718827756,
+      "eval_sts-test_spearman_dot": 0.8584340456924826,
+      "eval_sts-test_spearman_euclidean": 0.9055143864829975,
+      "eval_sts-test_spearman_manhattan": 0.9058283613329196,
+      "eval_sts-test_spearman_max": 0.9058283613329196,
+      "step": 195
+    },
+    {
+      "epoch": 0.3835616438356164,
+      "grad_norm": 3.063400983810425,
+      "learning_rate": 1.5312500000000003e-05,
+      "loss": 0.1197,
+      "step": 196
+    },
+    {
+      "epoch": 0.38551859099804303,
+      "grad_norm": 3.893153429031372,
+      "learning_rate": 1.5390625e-05,
+      "loss": 0.1395,
+      "step": 197
+    },
+    {
+      "epoch": 0.38747553816046965,
+      "grad_norm": 2.95540714263916,
+      "learning_rate": 1.546875e-05,
+      "loss": 0.0847,
+      "step": 198
+    },
+    {
+      "epoch": 0.38943248532289626,
+      "grad_norm": 3.4665300846099854,
+      "learning_rate": 1.5546875e-05,
+      "loss": 0.1848,
+      "step": 199
+    },
+    {
+      "epoch": 0.3913894324853229,
+      "grad_norm": 3.6926543712615967,
+      "learning_rate": 1.5625e-05,
+      "loss": 0.1377,
+      "step": 200
+    },
+    {
+      "epoch": 0.3913894324853229,
+      "eval_loss": 0.10723523795604706,
+      "eval_runtime": 107.245,
+      "eval_samples_per_second": 28.458,
+      "eval_steps_per_second": 0.224,
+      "eval_sts-test_pearson_cosine": 0.877994665901344,
+      "eval_sts-test_pearson_dot": 0.854134605280733,
+      "eval_sts-test_pearson_euclidean": 0.9085191117850383,
+      "eval_sts-test_pearson_manhattan": 0.9086424100414001,
+      "eval_sts-test_pearson_max": 0.9086424100414001,
+      "eval_sts-test_spearman_cosine": 0.904685279863199,
+      "eval_sts-test_spearman_dot": 0.8598855528557127,
+      "eval_sts-test_spearman_euclidean": 0.9052407772708506,
+      "eval_sts-test_spearman_manhattan": 0.9058868959828196,
+      "eval_sts-test_spearman_max": 0.9058868959828196,
+      "step": 200
+    },
+    {
+      "epoch": 0.3933463796477495,
+      "grad_norm": 3.303112268447876,
+      "learning_rate": 1.5703125e-05,
+      "loss": 0.1109,
+      "step": 201
+    },
+    {
+      "epoch": 0.3953033268101761,
+      "grad_norm": 3.4490058422088623,
+      "learning_rate": 1.578125e-05,
+      "loss": 0.1051,
+      "step": 202
+    },
+    {
+      "epoch": 0.3972602739726027,
+      "grad_norm": 2.6598286628723145,
+      "learning_rate": 1.5859375e-05,
+      "loss": 0.0975,
+      "step": 203
+    },
+    {
+      "epoch": 0.39921722113502933,
+      "grad_norm": 3.373512029647827,
+      "learning_rate": 1.59375e-05,
+      "loss": 0.127,
+      "step": 204
+    },
+    {
+      "epoch": 0.40117416829745595,
+      "grad_norm": 3.1471354961395264,
+      "learning_rate": 1.6015625e-05,
+      "loss": 0.1297,
+      "step": 205
+    },
+    {
+      "epoch": 0.40117416829745595,
+      "eval_loss": 0.10685314983129501,
+      "eval_runtime": 107.3321,
+      "eval_samples_per_second": 28.435,
+      "eval_steps_per_second": 0.224,
+      "eval_sts-test_pearson_cosine": 0.8785914848590666,
+      "eval_sts-test_pearson_dot": 0.8570818659891223,
+      "eval_sts-test_pearson_euclidean": 0.9086611488562145,
+      "eval_sts-test_pearson_manhattan": 0.9087606701935215,
+      "eval_sts-test_pearson_max": 0.9087606701935215,
+      "eval_sts-test_spearman_cosine": 0.9048987433800361,
+      "eval_sts-test_spearman_dot": 0.8616398023022556,
+      "eval_sts-test_spearman_euclidean": 0.9052247563192726,
+      "eval_sts-test_spearman_manhattan": 0.9056138237858093,
+      "eval_sts-test_spearman_max": 0.9056138237858093,
+      "step": 205
+    },
+    {
+      "epoch": 0.40313111545988256,
+      "grad_norm": 2.6924684047698975,
+      "learning_rate": 1.609375e-05,
+      "loss": 0.0783,
+      "step": 206
+    },
+    {
+      "epoch": 0.4050880626223092,
+      "grad_norm": 2.1100542545318604,
+      "learning_rate": 1.6171875000000002e-05,
+      "loss": 0.053,
+      "step": 207
+    },
+    {
+      "epoch": 0.4070450097847358,
+      "grad_norm": 3.7984156608581543,
+      "learning_rate": 1.6250000000000002e-05,
+      "loss": 0.1916,
+      "step": 208
+    },
+    {
+      "epoch": 0.4090019569471624,
+      "grad_norm": 4.329834461212158,
+      "learning_rate": 1.6328125000000002e-05,
+      "loss": 0.178,
+      "step": 209
+    },
+    {
+      "epoch": 0.410958904109589,
+      "grad_norm": 4.427723407745361,
+      "learning_rate": 1.6406250000000002e-05,
+      "loss": 0.2343,
+      "step": 210
+    },
+    {
+      "epoch": 0.410958904109589,
+      "eval_loss": 0.10670512914657593,
+      "eval_runtime": 107.2313,
+      "eval_samples_per_second": 28.462,
+      "eval_steps_per_second": 0.224,
+      "eval_sts-test_pearson_cosine": 0.8788965355860006,
+      "eval_sts-test_pearson_dot": 0.8580075676260999,
+      "eval_sts-test_pearson_euclidean": 0.908776492246521,
+      "eval_sts-test_pearson_manhattan": 0.9089340980301853,
+      "eval_sts-test_pearson_max": 0.9089340980301853,
+      "eval_sts-test_spearman_cosine": 0.90530862018312,
+      "eval_sts-test_spearman_dot": 0.8630207814775328,
+      "eval_sts-test_spearman_euclidean": 0.905449362900196,
+      "eval_sts-test_spearman_manhattan": 0.9056519071092534,
+      "eval_sts-test_spearman_max": 0.9056519071092534,
+      "step": 210
+    },
+    {
+      "epoch": 0.41291585127201563,
+      "grad_norm": 3.890899419784546,
+      "learning_rate": 1.6484375000000003e-05,
+      "loss": 0.1816,
+      "step": 211
+    },
+    {
+      "epoch": 0.41487279843444225,
+      "grad_norm": 4.071934700012207,
+      "learning_rate": 1.6562500000000003e-05,
+      "loss": 0.2522,
+      "step": 212
+    },
+    {
+      "epoch": 0.41682974559686886,
+      "grad_norm": 3.8046796321868896,
+      "learning_rate": 1.6640625000000003e-05,
+      "loss": 0.1787,
+      "step": 213
+    },
+    {
+      "epoch": 0.4187866927592955,
+      "grad_norm": 3.357276201248169,
+      "learning_rate": 1.671875e-05,
+      "loss": 0.1913,
+      "step": 214
+    },
+    {
+      "epoch": 0.4207436399217221,
+      "grad_norm": 3.8679873943328857,
+      "learning_rate": 1.6796875e-05,
+      "loss": 0.175,
+      "step": 215
+    },
+    {
+      "epoch": 0.4207436399217221,
+      "eval_loss": 0.10552908480167389,
+      "eval_runtime": 107.6412,
+      "eval_samples_per_second": 28.353,
+      "eval_steps_per_second": 0.223,
+      "eval_sts-test_pearson_cosine": 0.8791676817178924,
+      "eval_sts-test_pearson_dot": 0.8573342496118925,
+      "eval_sts-test_pearson_euclidean": 0.909475190469058,
+      "eval_sts-test_pearson_manhattan": 0.9097533727394405,
+      "eval_sts-test_pearson_max": 0.9097533727394405,
+      "eval_sts-test_spearman_cosine": 0.9056468502167161,
+      "eval_sts-test_spearman_dot": 0.8624976392318674,
+      "eval_sts-test_spearman_euclidean": 0.9066117769148375,
+      "eval_sts-test_spearman_manhattan": 0.9069566301351195,
+      "eval_sts-test_spearman_max": 0.9069566301351195,
+      "step": 215
+    },
+    {
+      "epoch": 0.4227005870841487,
+      "grad_norm": 3.436488389968872,
+      "learning_rate": 1.6875e-05,
+      "loss": 0.1533,
+      "step": 216
+    },
+    {
+      "epoch": 0.4246575342465753,
+      "grad_norm": 3.891040563583374,
+      "learning_rate": 1.6953125e-05,
+      "loss": 0.1819,
+      "step": 217
+    },
+    {
+      "epoch": 0.42661448140900193,
+      "grad_norm": 4.554884910583496,
+      "learning_rate": 1.703125e-05,
+      "loss": 0.2541,
+      "step": 218
+    },
+    {
+      "epoch": 0.42857142857142855,
+      "grad_norm": 3.4431850910186768,
+      "learning_rate": 1.7109375e-05,
+      "loss": 0.1103,
+      "step": 219
+    },
+    {
+      "epoch": 0.43052837573385516,
+      "grad_norm": 3.5396361351013184,
+      "learning_rate": 1.71875e-05,
+      "loss": 0.1693,
+      "step": 220
+    },
+    {
+      "epoch": 0.43052837573385516,
+      "eval_loss": 0.10396925359964371,
+      "eval_runtime": 107.33,
+      "eval_samples_per_second": 28.436,
+      "eval_steps_per_second": 0.224,
+      "eval_sts-test_pearson_cosine": 0.8794186180626897,
+      "eval_sts-test_pearson_dot": 0.8555325369075935,
+      "eval_sts-test_pearson_euclidean": 0.9099071011406157,
+      "eval_sts-test_pearson_manhattan": 0.9104095617945829,
+      "eval_sts-test_pearson_max": 0.9104095617945829,
+      "eval_sts-test_spearman_cosine": 0.9061536582519738,
+      "eval_sts-test_spearman_dot": 0.8609769018672648,
+      "eval_sts-test_spearman_euclidean": 0.9068523149448162,
+      "eval_sts-test_spearman_manhattan": 0.9075606826613808,
+      "eval_sts-test_spearman_max": 0.9075606826613808,
+      "step": 220
+    },
+    {
+      "epoch": 0.4324853228962818,
+      "grad_norm": 3.4416589736938477,
+      "learning_rate": 1.7265625e-05,
+      "loss": 0.1233,
+      "step": 221
+    },
+    {
+      "epoch": 0.4344422700587084,
+      "grad_norm": 2.9554316997528076,
+      "learning_rate": 1.734375e-05,
+      "loss": 0.0922,
+      "step": 222
+    },
+    {
+      "epoch": 0.436399217221135,
+      "grad_norm": 3.1570141315460205,
+      "learning_rate": 1.7421875e-05,
+      "loss": 0.1243,
+      "step": 223
+    },
+    {
+      "epoch": 0.4383561643835616,
+      "grad_norm": 3.8479344844818115,
+      "learning_rate": 1.7500000000000002e-05,
+      "loss": 0.1613,
+      "step": 224
+    },
+    {
+      "epoch": 0.44031311154598823,
+      "grad_norm": 3.004990339279175,
+      "learning_rate": 1.7578125000000002e-05,
+      "loss": 0.1188,
+      "step": 225
+    },
+    {
+      "epoch": 0.44031311154598823,
+      "eval_loss": 0.1029738187789917,
+      "eval_runtime": 107.2661,
+      "eval_samples_per_second": 28.453,
+      "eval_steps_per_second": 0.224,
+      "eval_sts-test_pearson_cosine": 0.8796765477862789,
+      "eval_sts-test_pearson_dot": 0.8576485629522204,
+      "eval_sts-test_pearson_euclidean": 0.9098263075403831,
+      "eval_sts-test_pearson_manhattan": 0.9104321398639006,
+      "eval_sts-test_pearson_max": 0.9104321398639006,
+      "eval_sts-test_spearman_cosine": 0.9064603386462892,
+      "eval_sts-test_spearman_dot": 0.8635142088856343,
+      "eval_sts-test_spearman_euclidean": 0.9066103896257344,
+      "eval_sts-test_spearman_manhattan": 0.9076328216947436,
+      "eval_sts-test_spearman_max": 0.9076328216947436,
+      "step": 225
+    },
+    {
+      "epoch": 0.44227005870841485,
+      "grad_norm": 3.595667839050293,
+      "learning_rate": 1.7656250000000002e-05,
+      "loss": 0.196,
+      "step": 226
+    },
+    {
+      "epoch": 0.44422700587084146,
+      "grad_norm": 3.9599428176879883,
+      "learning_rate": 1.7734375000000002e-05,
+      "loss": 0.2254,
+      "step": 227
+    },
+    {
+      "epoch": 0.4461839530332681,
+      "grad_norm": 3.2490875720977783,
+      "learning_rate": 1.7812500000000003e-05,
+      "loss": 0.1162,
+      "step": 228
+    },
+    {
+      "epoch": 0.4481409001956947,
+      "grad_norm": 4.811342239379883,
+      "learning_rate": 1.7890625000000003e-05,
+      "loss": 0.2579,
+      "step": 229
+    },
+    {
+      "epoch": 0.4500978473581213,
+      "grad_norm": 2.993255138397217,
+      "learning_rate": 1.7968750000000003e-05,
+      "loss": 0.1203,
+      "step": 230
+    },
+    {
+      "epoch": 0.4500978473581213,
+      "eval_loss": 0.102933868765831,
+      "eval_runtime": 107.2515,
+      "eval_samples_per_second": 28.456,
+      "eval_steps_per_second": 0.224,
+      "eval_sts-test_pearson_cosine": 0.8799758353085696,
+      "eval_sts-test_pearson_dot": 0.8592997081846103,
+      "eval_sts-test_pearson_euclidean": 0.9101945793558552,
+      "eval_sts-test_pearson_manhattan": 0.9106837055219174,
+      "eval_sts-test_pearson_max": 0.9106837055219174,
+      "eval_sts-test_spearman_cosine": 0.9071432428951217,
+      "eval_sts-test_spearman_dot": 0.865314059867535,
+      "eval_sts-test_spearman_euclidean": 0.9072587906520344,
+      "eval_sts-test_spearman_manhattan": 0.9077949555147645,
+      "eval_sts-test_spearman_max": 0.9077949555147645,
+      "step": 230
+    },
+    {
+      "epoch": 0.4520547945205479,
+      "grad_norm": 3.654191017150879,
+      "learning_rate": 1.8046875e-05,
+      "loss": 0.1654,
+      "step": 231
+    },
+    {
+      "epoch": 0.45401174168297453,
+      "grad_norm": 3.429565668106079,
+      "learning_rate": 1.8125e-05,
+      "loss": 0.1808,
+      "step": 232
+    },
+    {
+      "epoch": 0.45596868884540115,
+      "grad_norm": 3.5679566860198975,
+      "learning_rate": 1.8203125e-05,
+      "loss": 0.1397,
+      "step": 233
+    },
+    {
+      "epoch": 0.45792563600782776,
+      "grad_norm": 3.9862124919891357,
+      "learning_rate": 1.828125e-05,
+      "loss": 0.2177,
+      "step": 234
+    },
+    {
+      "epoch": 0.4598825831702544,
+      "grad_norm": 3.536984443664551,
+      "learning_rate": 1.8359375e-05,
+      "loss": 0.162,
+      "step": 235
+    },
+    {
+      "epoch": 0.4598825831702544,
+      "eval_loss": 0.10404225438833237,
+      "eval_runtime": 107.254,
+      "eval_samples_per_second": 28.456,
+      "eval_steps_per_second": 0.224,
+      "eval_sts-test_pearson_cosine": 0.8802088610554777,
+      "eval_sts-test_pearson_dot": 0.8618209119350905,
+      "eval_sts-test_pearson_euclidean": 0.9103461475031536,
+      "eval_sts-test_pearson_manhattan": 0.9106782364335553,
+      "eval_sts-test_pearson_max": 0.9106782364335553,
+      "eval_sts-test_spearman_cosine": 0.9077748174471387,
+      "eval_sts-test_spearman_dot": 0.8686349167216066,
+      "eval_sts-test_spearman_euclidean": 0.907571109705285,
+      "eval_sts-test_spearman_manhattan": 0.9080472631264893,
+      "eval_sts-test_spearman_max": 0.9080472631264893,
+      "step": 235
+    },
+    {
+      "epoch": 0.461839530332681,
+      "grad_norm": 3.2987570762634277,
+      "learning_rate": 1.84375e-05,
+      "loss": 0.177,
+      "step": 236
+    },
+    {
+      "epoch": 0.4637964774951076,
+      "grad_norm": 1.792919397354126,
+      "learning_rate": 1.8515625e-05,
+      "loss": 0.0556,
+      "step": 237
+    },
+    {
+      "epoch": 0.4657534246575342,
+      "grad_norm": 3.8270483016967773,
+      "learning_rate": 1.859375e-05,
+      "loss": 0.2285,
+      "step": 238
+    },
+    {
+      "epoch": 0.46771037181996084,
+      "grad_norm": 3.2458577156066895,
+      "learning_rate": 1.8671875e-05,
+      "loss": 0.1657,
+      "step": 239
+    },
+    {
+      "epoch": 0.46966731898238745,
+      "grad_norm": 4.352839469909668,
+      "learning_rate": 1.8750000000000002e-05,
+      "loss": 0.2555,
+      "step": 240
+    },
+    {
+      "epoch": 0.46966731898238745,
+      "eval_loss": 0.10528620332479477,
+      "eval_runtime": 107.3201,
+      "eval_samples_per_second": 28.438,
+      "eval_steps_per_second": 0.224,
+      "eval_sts-test_pearson_cosine": 0.8794205585889476,
+      "eval_sts-test_pearson_dot": 0.8616236846471828,
+      "eval_sts-test_pearson_euclidean": 0.9100171674371834,
+      "eval_sts-test_pearson_manhattan": 0.9102120642982687,
+      "eval_sts-test_pearson_max": 0.9102120642982687,
+      "eval_sts-test_spearman_cosine": 0.9076779309662261,
+      "eval_sts-test_spearman_dot": 0.8702396969551023,
+      "eval_sts-test_spearman_euclidean": 0.9078436896384199,
+      "eval_sts-test_spearman_manhattan": 0.9080407741935878,
+      "eval_sts-test_spearman_max": 0.9080407741935878,
+      "step": 240
+    },
+    {
+      "epoch": 0.47162426614481406,
+      "grad_norm": 3.644327163696289,
+      "learning_rate": 1.8828125000000002e-05,
+      "loss": 0.1606,
+      "step": 241
+    },
+    {
+      "epoch": 0.4735812133072407,
+      "grad_norm": 3.0316474437713623,
+      "learning_rate": 1.8906250000000002e-05,
+      "loss": 0.1257,
+      "step": 242
+    },
+    {
+      "epoch": 0.4755381604696673,
+      "grad_norm": 3.8527326583862305,
+      "learning_rate": 1.8984375000000002e-05,
+      "loss": 0.1898,
+      "step": 243
+    },
+    {
+      "epoch": 0.4774951076320939,
+      "grad_norm": 3.91603422164917,
+      "learning_rate": 1.9062500000000003e-05,
+      "loss": 0.1621,
+      "step": 244
+    },
+    {
+      "epoch": 0.4794520547945205,
+      "grad_norm": 3.6845171451568604,
+      "learning_rate": 1.9140625000000003e-05,
+      "loss": 0.1606,
+      "step": 245
+    },
+    {
+      "epoch": 0.4794520547945205,
+      "eval_loss": 0.10541080683469772,
+      "eval_runtime": 107.3443,
+      "eval_samples_per_second": 28.432,
+      "eval_steps_per_second": 0.224,
+      "eval_sts-test_pearson_cosine": 0.8782579310286232,
+      "eval_sts-test_pearson_dot": 0.8596847230641689,
+      "eval_sts-test_pearson_euclidean": 0.909741577402618,
+      "eval_sts-test_pearson_manhattan": 0.9098438643121189,
+      "eval_sts-test_pearson_max": 0.9098438643121189,
+      "eval_sts-test_spearman_cosine": 0.9078928712746891,
+      "eval_sts-test_spearman_dot": 0.8682800392187727,
+      "eval_sts-test_spearman_euclidean": 0.9083291960732551,
+      "eval_sts-test_spearman_manhattan": 0.908423397478484,
+      "eval_sts-test_spearman_max": 0.908423397478484,
+      "step": 245
+    },
+    {
+      "epoch": 0.48140900195694714,
+      "grad_norm": 3.31758451461792,
+      "learning_rate": 1.9218750000000003e-05,
+      "loss": 0.0983,
+      "step": 246
+    },
+    {
+      "epoch": 0.48336594911937375,
+      "grad_norm": 3.8613622188568115,
+      "learning_rate": 1.9296875000000003e-05,
+      "loss": 0.2028,
+      "step": 247
+    },
+    {
+      "epoch": 0.48532289628180036,
+      "grad_norm": 2.792924165725708,
+      "learning_rate": 1.9375e-05,
+      "loss": 0.0997,
+      "step": 248
+    },
+    {
+      "epoch": 0.487279843444227,
+      "grad_norm": 3.4162261486053467,
+      "learning_rate": 1.9453125e-05,
+      "loss": 0.1582,
+      "step": 249
+    },
+    {
+      "epoch": 0.4892367906066536,
+      "grad_norm": 4.499621391296387,
+      "learning_rate": 1.953125e-05,
+      "loss": 0.2394,
+      "step": 250
+    },
+    {
+      "epoch": 0.4892367906066536,
+      "eval_loss": 0.10517927259206772,
+      "eval_runtime": 107.2761,
+      "eval_samples_per_second": 28.45,
+      "eval_steps_per_second": 0.224,
+      "eval_sts-test_pearson_cosine": 0.8775291091187776,
+      "eval_sts-test_pearson_dot": 0.8591957018286404,
+      "eval_sts-test_pearson_euclidean": 0.9092406666480166,
+      "eval_sts-test_pearson_manhattan": 0.909395200356788,
+      "eval_sts-test_pearson_max": 0.909395200356788,
+      "eval_sts-test_spearman_cosine": 0.9073655224104529,
+      "eval_sts-test_spearman_dot": 0.866218124850164,
+      "eval_sts-test_spearman_euclidean": 0.9077081380676655,
+      "eval_sts-test_spearman_manhattan": 0.907968321901395,
+      "eval_sts-test_spearman_max": 0.907968321901395,
+      "step": 250
+    },
+    {
+      "epoch": 0.4911937377690802,
+      "grad_norm": 4.491675853729248,
+      "learning_rate": 1.9609375e-05,
+      "loss": 0.2186,
+      "step": 251
+    },
+    {
+      "epoch": 0.4931506849315068,
+      "grad_norm": 2.9051578044891357,
+      "learning_rate": 1.96875e-05,
+      "loss": 0.0993,
+      "step": 252
+    },
+    {
+      "epoch": 0.49510763209393344,
+      "grad_norm": 3.53365421295166,
+      "learning_rate": 1.9765625e-05,
+      "loss": 0.1805,
+      "step": 253
+    },
+    {
+      "epoch": 0.49706457925636005,
+      "grad_norm": 3.2181098461151123,
+      "learning_rate": 1.984375e-05,
+      "loss": 0.1178,
+      "step": 254
+    },
+    {
+      "epoch": 0.49902152641878667,
+      "grad_norm": 4.045453071594238,
+      "learning_rate": 1.9921875e-05,
+      "loss": 0.2198,
+      "step": 255
+    },
+    {
+      "epoch": 0.49902152641878667,
+      "eval_loss": 0.10428859293460846,
+      "eval_runtime": 107.2698,
+      "eval_samples_per_second": 28.452,
+      "eval_steps_per_second": 0.224,
+      "eval_sts-test_pearson_cosine": 0.8777129617944619,
+      "eval_sts-test_pearson_dot": 0.8588391652180615,
+      "eval_sts-test_pearson_euclidean": 0.9093230964292308,
+      "eval_sts-test_pearson_manhattan": 0.9095932968076137,
+      "eval_sts-test_pearson_max": 0.9095932968076137,
+      "eval_sts-test_spearman_cosine": 0.9069800350448274,
+      "eval_sts-test_spearman_dot": 0.8639776976998651,
+      "eval_sts-test_spearman_euclidean": 0.9072912800678044,
+      "eval_sts-test_spearman_manhattan": 0.9080281095866138,
+      "eval_sts-test_spearman_max": 0.9080281095866138,
+      "step": 255
+    },
+    {
+      "epoch": 0.5009784735812133,
+      "grad_norm": 2.8251521587371826,
+      "learning_rate": 2e-05,
+      "loss": 0.1064,
+      "step": 256
+    },
+    {
+      "epoch": 0.50293542074364,
+      "grad_norm": 3.3597464561462402,
+      "learning_rate": 1.999924308128909e-05,
+      "loss": 0.1436,
+      "step": 257
+    },
+    {
+      "epoch": 0.5048923679060665,
+      "grad_norm": 2.580488920211792,
+      "learning_rate": 1.9996972439741537e-05,
+      "loss": 0.0859,
+      "step": 258
+    },
+    {
+      "epoch": 0.5068493150684932,
+      "grad_norm": 3.937856674194336,
+      "learning_rate": 1.9993188419095562e-05,
+      "loss": 0.2157,
+      "step": 259
+    },
+    {
+      "epoch": 0.5088062622309197,
+      "grad_norm": 3.344531774520874,
+      "learning_rate": 1.9987891592190367e-05,
+      "loss": 0.1455,
+      "step": 260
+    },
+    {
+      "epoch": 0.5088062622309197,
+      "eval_loss": 0.10292962938547134,
+      "eval_runtime": 107.2285,
+      "eval_samples_per_second": 28.463,
+      "eval_steps_per_second": 0.224,
+      "eval_sts-test_pearson_cosine": 0.8767515459977318,
+      "eval_sts-test_pearson_dot": 0.8564862360521637,
+      "eval_sts-test_pearson_euclidean": 0.9083760527634203,
+      "eval_sts-test_pearson_manhattan": 0.9086626400377007,
+      "eval_sts-test_pearson_max": 0.9086626400377007,
+      "eval_sts-test_spearman_cosine": 0.9057508521481897,
+      "eval_sts-test_spearman_dot": 0.8601081456298736,
+      "eval_sts-test_spearman_euclidean": 0.9063700753520626,
+      "eval_sts-test_spearman_manhattan": 0.9068438122051519,
+      "eval_sts-test_spearman_max": 0.9068438122051519,
+      "step": 260
+    },
+    {
+      "epoch": 0.5107632093933464,
+      "grad_norm": 3.7637484073638916,
+      "learning_rate": 1.9981082760879432e-05,
+      "loss": 0.1974,
+      "step": 261
+    },
+    {
+      "epoch": 0.512720156555773,
+      "grad_norm": 3.182102918624878,
+      "learning_rate": 1.997276295590912e-05,
+      "loss": 0.1667,
+      "step": 262
+    },
+    {
+      "epoch": 0.5146771037181996,
+      "grad_norm": 3.7908170223236084,
+      "learning_rate": 1.9962933436762644e-05,
+      "loss": 0.1512,
+      "step": 263
+    },
+    {
+      "epoch": 0.5166340508806262,
+      "grad_norm": 3.4492650032043457,
+      "learning_rate": 1.9951595691469397e-05,
+      "loss": 0.1684,
+      "step": 264
+    },
+    {
+      "epoch": 0.5185909980430529,
+      "grad_norm": 3.816772222518921,
+      "learning_rate": 1.9938751436379684e-05,
+      "loss": 0.2132,
+      "step": 265
+    },
+    {
+      "epoch": 0.5185909980430529,
+      "eval_loss": 0.10117975622415543,
+      "eval_runtime": 107.3212,
+      "eval_samples_per_second": 28.438,
+      "eval_steps_per_second": 0.224,
+      "eval_sts-test_pearson_cosine": 0.8770393502752714,
+      "eval_sts-test_pearson_dot": 0.8567524208989885,
+      "eval_sts-test_pearson_euclidean": 0.9080912956763092,
+      "eval_sts-test_pearson_manhattan": 0.908247948105785,
+      "eval_sts-test_pearson_max": 0.908247948105785,
+      "eval_sts-test_spearman_cosine": 0.9053279079767796,
+      "eval_sts-test_spearman_dot": 0.8598375795035011,
+      "eval_sts-test_spearman_euclidean": 0.9057662913333698,
+      "eval_sts-test_spearman_manhattan": 0.9061448870047409,
+      "eval_sts-test_spearman_max": 0.9061448870047409,
+      "step": 265
+    },
+    {
+      "epoch": 0.5205479452054794,
+      "grad_norm": 3.5570499897003174,
+      "learning_rate": 1.992440261590491e-05,
+      "loss": 0.1645,
+      "step": 266
+    },
+    {
+      "epoch": 0.5225048923679061,
+      "grad_norm": 4.160579681396484,
+      "learning_rate": 1.9908551402223218e-05,
+      "loss": 0.203,
+      "step": 267
+    },
+    {
+      "epoch": 0.5244618395303327,
+      "grad_norm": 3.5718774795532227,
+      "learning_rate": 1.9891200194950644e-05,
+      "loss": 0.1539,
+      "step": 268
+    },
+    {
+      "epoch": 0.5264187866927593,
+      "grad_norm": 3.604438066482544,
+      "learning_rate": 1.9872351620777883e-05,
+      "loss": 0.1445,
+      "step": 269
+    },
+    {
+      "epoch": 0.5283757338551859,
+      "grad_norm": 3.4854915142059326,
+      "learning_rate": 1.9852008533072627e-05,
+      "loss": 0.1377,
+      "step": 270
+    },
+    {
+      "epoch": 0.5283757338551859,
+      "eval_loss": 0.09936786442995071,
+      "eval_runtime": 107.3119,
+      "eval_samples_per_second": 28.44,
+      "eval_steps_per_second": 0.224,
+      "eval_sts-test_pearson_cosine": 0.8772155384897071,
+      "eval_sts-test_pearson_dot": 0.8577040756637748,
+      "eval_sts-test_pearson_euclidean": 0.9081962404777727,
+      "eval_sts-test_pearson_manhattan": 0.9082660411148933,
+      "eval_sts-test_pearson_max": 0.9082660411148933,
+      "eval_sts-test_spearman_cosine": 0.9056296657323417,
+      "eval_sts-test_spearman_dot": 0.8627456954737598,
+      "eval_sts-test_spearman_euclidean": 0.9061553587999066,
+      "eval_sts-test_spearman_manhattan": 0.9063870360801298,
+      "eval_sts-test_spearman_max": 0.9063870360801298,
+      "step": 270
+    },
+    {
+      "epoch": 0.5303326810176126,
+      "grad_norm": 3.662992238998413,
+      "learning_rate": 1.9830174011447617e-05,
+      "loss": 0.1719,
+      "step": 271
+    },
+    {
+      "epoch": 0.5322896281800391,
+      "grad_norm": 3.5594613552093506,
+      "learning_rate": 1.980685136129445e-05,
+      "loss": 0.1896,
+      "step": 272
+    },
+    {
+      "epoch": 0.5342465753424658,
+      "grad_norm": 3.257335662841797,
+      "learning_rate": 1.978204411328318e-05,
+      "loss": 0.1452,
+      "step": 273
+    },
+    {
+      "epoch": 0.5362035225048923,
+      "grad_norm": 3.292863368988037,
+      "learning_rate": 1.9755756022827847e-05,
+      "loss": 0.1275,
+      "step": 274
+    },
+    {
+      "epoch": 0.538160469667319,
+      "grad_norm": 4.065443515777588,
+      "learning_rate": 1.972799106951796e-05,
+      "loss": 0.1883,
+      "step": 275
+    },
+    {
+      "epoch": 0.538160469667319,
+      "eval_loss": 0.09800439327955246,
+      "eval_runtime": 107.2596,
+      "eval_samples_per_second": 28.454,
+      "eval_steps_per_second": 0.224,
+      "eval_sts-test_pearson_cosine": 0.8773598628753827,
+      "eval_sts-test_pearson_dot": 0.8578251655808844,
+      "eval_sts-test_pearson_euclidean": 0.9082603623937704,
+      "eval_sts-test_pearson_manhattan": 0.9081101963076783,
+      "eval_sts-test_pearson_max": 0.9082603623937704,
+      "eval_sts-test_spearman_cosine": 0.9056689328319392,
+      "eval_sts-test_spearman_dot": 0.8647132741833555,
+      "eval_sts-test_spearman_euclidean": 0.9063065285608867,
+      "eval_sts-test_spearman_manhattan": 0.9067770433231558,
+      "eval_sts-test_spearman_max": 0.9067770433231558,
+      "step": 275
+    },
+    {
+      "epoch": 0.5401174168297456,
+      "grad_norm": 3.7186553478240967,
+      "learning_rate": 1.9698753456516047e-05,
+      "loss": 0.1462,
+      "step": 276
+    },
+    {
+      "epoch": 0.5420743639921722,
+      "grad_norm": 3.5399951934814453,
+      "learning_rate": 1.9668047609921382e-05,
+      "loss": 0.1595,
+      "step": 277
+    },
+    {
+      "epoch": 0.5440313111545988,
+      "grad_norm": 3.6143035888671875,
+      "learning_rate": 1.963587817809993e-05,
+      "loss": 0.1693,
+      "step": 278
+    },
+    {
+      "epoch": 0.5459882583170255,
+      "grad_norm": 4.133859634399414,
+      "learning_rate": 1.9602250030980657e-05,
+      "loss": 0.1929,
+      "step": 279
+    },
+    {
+      "epoch": 0.547945205479452,
+      "grad_norm": 3.6929726600646973,
+      "learning_rate": 1.9567168259318324e-05,
+      "loss": 0.154,
+      "step": 280
+    },
+    {
+      "epoch": 0.547945205479452,
+      "eval_loss": 0.0969705730676651,
+      "eval_runtime": 107.333,
+      "eval_samples_per_second": 28.435,
+      "eval_steps_per_second": 0.224,
+      "eval_sts-test_pearson_cosine": 0.8772326487842304,
+      "eval_sts-test_pearson_dot": 0.8584362564160372,
+      "eval_sts-test_pearson_euclidean": 0.9077579223693962,
+      "eval_sts-test_pearson_manhattan": 0.9072827835669532,
+      "eval_sts-test_pearson_max": 0.9077579223693962,
+      "eval_sts-test_spearman_cosine": 0.9052923754752349,
+      "eval_sts-test_spearman_dot": 0.866326959917868,
+      "eval_sts-test_spearman_euclidean": 0.9057464665245734,
+      "eval_sts-test_spearman_manhattan": 0.9059635996448444,
+      "eval_sts-test_spearman_max": 0.9059635996448444,
+      "step": 280
+    },
+    {
+      "epoch": 0.5499021526418787,
+      "grad_norm": 3.515667200088501,
+      "learning_rate": 1.953063817392281e-05,
+      "loss": 0.1468,
+      "step": 281
+    },
+    {
+      "epoch": 0.5518590998043053,
+      "grad_norm": 2.3627371788024902,
+      "learning_rate": 1.949266530485513e-05,
+      "loss": 0.0898,
+      "step": 282
+    },
+    {
+      "epoch": 0.5538160469667319,
+      "grad_norm": 3.26710844039917,
+      "learning_rate": 1.945325540059032e-05,
+      "loss": 0.1425,
+      "step": 283
+    },
+    {
+      "epoch": 0.5557729941291585,
+      "grad_norm": 3.6672258377075195,
+      "learning_rate": 1.941241442714716e-05,
+      "loss": 0.1362,
+      "step": 284
+    },
+    {
+      "epoch": 0.5577299412915852,
+      "grad_norm": 3.306119203567505,
+      "learning_rate": 1.9370148567185043e-05,
+      "loss": 0.1025,
+      "step": 285
+    },
+    {
+      "epoch": 0.5577299412915852,
+      "eval_loss": 0.09782103449106216,
+      "eval_runtime": 107.2147,
+      "eval_samples_per_second": 28.466,
+      "eval_steps_per_second": 0.224,
+      "eval_sts-test_pearson_cosine": 0.8780373092719852,
+      "eval_sts-test_pearson_dot": 0.861128415219923,
+      "eval_sts-test_pearson_euclidean": 0.9076094585437832,
+      "eval_sts-test_pearson_manhattan": 0.9068707688162918,
+      "eval_sts-test_pearson_max": 0.9076094585437832,
+      "eval_sts-test_spearman_cosine": 0.9052606468309083,
+      "eval_sts-test_spearman_dot": 0.868469739815811,
+      "eval_sts-test_spearman_euclidean": 0.9051151604801249,
+      "eval_sts-test_spearman_manhattan": 0.9048908224067698,
+      "eval_sts-test_spearman_max": 0.9052606468309083,
+      "step": 285
+    },
+    {
+      "epoch": 0.5596868884540117,
+      "grad_norm": 3.2606685161590576,
+      "learning_rate": 1.9326464219068023e-05,
+      "loss": 0.1578,
+      "step": 286
+    },
+    {
+      "epoch": 0.5616438356164384,
+      "grad_norm": 3.5152740478515625,
+      "learning_rate": 1.9281367995896187e-05,
+      "loss": 0.1235,
+      "step": 287
+    },
+    {
+      "epoch": 0.5636007827788649,
+      "grad_norm": 2.8671882152557373,
+      "learning_rate": 1.9234866724504554e-05,
+      "loss": 0.1109,
+      "step": 288
+    },
+    {
+      "epoch": 0.5655577299412916,
+      "grad_norm": 2.315185785293579,
+      "learning_rate": 1.9186967444429613e-05,
+      "loss": 0.0746,
+      "step": 289
+    },
+    {
+      "epoch": 0.5675146771037182,
+      "grad_norm": 3.4961392879486084,
+      "learning_rate": 1.913767740684362e-05,
+      "loss": 0.1471,
+      "step": 290
+    },
+    {
+      "epoch": 0.5675146771037182,
+      "eval_loss": 0.09924904257059097,
+      "eval_runtime": 107.3422,
+      "eval_samples_per_second": 28.432,
+      "eval_steps_per_second": 0.224,
+      "eval_sts-test_pearson_cosine": 0.8783681424807572,
+      "eval_sts-test_pearson_dot": 0.861120631953773,
+      "eval_sts-test_pearson_euclidean": 0.9077238606316402,
+      "eval_sts-test_pearson_manhattan": 0.9069786963498391,
+      "eval_sts-test_pearson_max": 0.9077238606316402,
+      "eval_sts-test_spearman_cosine": 0.9052591700392825,
+      "eval_sts-test_spearman_dot": 0.8684268233561366,
+      "eval_sts-test_spearman_euclidean": 0.9046835793152661,
+      "eval_sts-test_spearman_manhattan": 0.9045985071673613,
+      "eval_sts-test_spearman_max": 0.9052591700392825,
+      "step": 290
+    },
+    {
+      "epoch": 0.5694716242661448,
+      "grad_norm": 4.221432209014893,
+      "learning_rate": 1.9087004073456926e-05,
+      "loss": 0.2631,
+      "step": 291
+    },
+    {
+      "epoch": 0.5714285714285714,
+      "grad_norm": 3.4570438861846924,
+      "learning_rate": 1.9034955115388364e-05,
+      "loss": 0.11,
+      "step": 292
+    },
+    {
+      "epoch": 0.5733855185909981,
+      "grad_norm": 3.6059136390686035,
+      "learning_rate": 1.898153841200398e-05,
+      "loss": 0.1834,
+      "step": 293
+    },
+    {
+      "epoch": 0.5753424657534246,
+      "grad_norm": 3.3278088569641113,
+      "learning_rate": 1.892676204972423e-05,
+      "loss": 0.1277,
+      "step": 294
+    },
+    {
+      "epoch": 0.5772994129158513,
+      "grad_norm": 4.314577579498291,
+      "learning_rate": 1.8870634320799822e-05,
+      "loss": 0.2104,
+      "step": 295
+    },
+    {
+      "epoch": 0.5772994129158513,
+      "eval_loss": 0.09903673827648163,
+      "eval_runtime": 107.6434,
+      "eval_samples_per_second": 28.353,
+      "eval_steps_per_second": 0.223,
+      "eval_sts-test_pearson_cosine": 0.8781909348259072,
+      "eval_sts-test_pearson_dot": 0.8596231931866185,
+      "eval_sts-test_pearson_euclidean": 0.9076411156234586,
+      "eval_sts-test_pearson_manhattan": 0.9069147632233857,
+      "eval_sts-test_pearson_max": 0.9076411156234586,
+      "eval_sts-test_spearman_cosine": 0.9042011607174669,
+      "eval_sts-test_spearman_dot": 0.8660264551976247,
+      "eval_sts-test_spearman_euclidean": 0.9044265280698341,
+      "eval_sts-test_spearman_manhattan": 0.9041656729671835,
+      "eval_sts-test_spearman_max": 0.9044265280698341,
+      "step": 295
+    },
+    {
+      "epoch": 0.5792563600782779,
+      "grad_norm": 3.195991039276123,
+      "learning_rate": 1.8813163722056397e-05,
+      "loss": 0.1294,
+      "step": 296
+    },
+    {
+      "epoch": 0.5812133072407045,
+      "grad_norm": 3.6352145671844482,
+      "learning_rate": 1.875435895360826e-05,
+      "loss": 0.1672,
+      "step": 297
+    },
+    {
+      "epoch": 0.5831702544031311,
+      "grad_norm": 3.7248518466949463,
+      "learning_rate": 1.8694228917541313e-05,
+      "loss": 0.2171,
+      "step": 298
+    },
+    {
+      "epoch": 0.5851272015655578,
+      "grad_norm": 3.459801435470581,
+      "learning_rate": 1.8632782716565438e-05,
+      "loss": 0.1451,
+      "step": 299
+    },
+    {
+      "epoch": 0.5870841487279843,
+      "grad_norm": 2.6911542415618896,
+      "learning_rate": 1.857002965263648e-05,
+      "loss": 0.0871,
+      "step": 300
+    },
+    {
+      "epoch": 0.5870841487279843,
+      "eval_loss": 0.09800251573324203,
+      "eval_runtime": 107.2338,
+      "eval_samples_per_second": 28.461,
+      "eval_steps_per_second": 0.224,
+      "eval_sts-test_pearson_cosine": 0.8772925626670083,
+      "eval_sts-test_pearson_dot": 0.8566016359384749,
+      "eval_sts-test_pearson_euclidean": 0.9070931796775764,
+      "eval_sts-test_pearson_manhattan": 0.9064105714529896,
+      "eval_sts-test_pearson_max": 0.9070931796775764,
+      "eval_sts-test_spearman_cosine": 0.9032592361677008,
+      "eval_sts-test_spearman_dot": 0.8623085204012272,
+      "eval_sts-test_spearman_euclidean": 0.9038942565668446,
+      "eval_sts-test_spearman_manhattan": 0.9033954590073763,
+      "eval_sts-test_spearman_max": 0.9038942565668446,
+      "step": 300
+    },
+    {
+      "epoch": 0.589041095890411,
+      "grad_norm": 2.913508653640747,
+      "learning_rate": 1.850597922554809e-05,
+      "loss": 0.0897,
+      "step": 301
+    },
+    {
+      "epoch": 0.5909980430528375,
+      "grad_norm": 3.2928783893585205,
+      "learning_rate": 1.844064113149361e-05,
+      "loss": 0.1296,
+      "step": 302
+    },
+    {
+      "epoch": 0.5929549902152642,
+      "grad_norm": 3.2551913261413574,
+      "learning_rate": 1.8374025261598224e-05,
+      "loss": 0.1206,
+      "step": 303
+    },
+    {
+      "epoch": 0.5949119373776908,
+      "grad_norm": 3.246716022491455,
+      "learning_rate": 1.8306141700421606e-05,
+      "loss": 0.1665,
+      "step": 304
+    },
+    {
+      "epoch": 0.5968688845401174,
+      "grad_norm": 3.980085611343384,
+      "learning_rate": 1.8237000724431283e-05,
+      "loss": 0.1511,
+      "step": 305
+    },
+    {
+      "epoch": 0.5968688845401174,
+      "eval_loss": 0.09785618633031845,
+      "eval_runtime": 107.3879,
+      "eval_samples_per_second": 28.42,
+      "eval_steps_per_second": 0.223,
+      "eval_sts-test_pearson_cosine": 0.8767817266460346,
+      "eval_sts-test_pearson_dot": 0.8544828510438696,
+      "eval_sts-test_pearson_euclidean": 0.9070553577944469,
+      "eval_sts-test_pearson_manhattan": 0.9065146784679962,
+      "eval_sts-test_pearson_max": 0.9070553577944469,
+      "eval_sts-test_spearman_cosine": 0.9032290290662617,
+      "eval_sts-test_spearman_dot": 0.8599922398628699,
+      "eval_sts-test_spearman_euclidean": 0.9039456310149221,
+      "eval_sts-test_spearman_manhattan": 0.9035283702537087,
+      "eval_sts-test_spearman_max": 0.9039456310149221,
+      "step": 305
+    },
+    {
+      "epoch": 0.598825831702544,
+      "grad_norm": 3.576425790786743,
+      "learning_rate": 1.8166612800446927e-05,
+      "loss": 0.1566,
+      "step": 306
+    },
+    {
+      "epoch": 0.6007827788649707,
+      "grad_norm": 3.3370437622070312,
+      "learning_rate": 1.809498858405589e-05,
+      "loss": 0.1339,
+      "step": 307
+    },
+    {
+      "epoch": 0.6027397260273972,
+      "grad_norm": 3.3882863521575928,
+      "learning_rate": 1.802213891800007e-05,
+      "loss": 0.1474,
+      "step": 308
+    },
+    {
+      "epoch": 0.6046966731898239,
+      "grad_norm": 2.9576971530914307,
+      "learning_rate": 1.7948074830534535e-05,
+      "loss": 0.1022,
+      "step": 309
+    },
+    {
+      "epoch": 0.6066536203522505,
+      "grad_norm": 3.737396001815796,
+      "learning_rate": 1.7872807533758007e-05,
+      "loss": 0.1263,
+      "step": 310
+    },
+    {
+      "epoch": 0.6066536203522505,
+      "eval_loss": 0.09827280789613724,
+      "eval_runtime": 107.5911,
+      "eval_samples_per_second": 28.367,
+      "eval_steps_per_second": 0.223,
+      "eval_sts-test_pearson_cosine": 0.8774447336518928,
+      "eval_sts-test_pearson_dot": 0.8526883126161577,
+      "eval_sts-test_pearson_euclidean": 0.9083025051320742,
+      "eval_sts-test_pearson_manhattan": 0.9079128512948802,
+      "eval_sts-test_pearson_max": 0.9083025051320742,
+      "eval_sts-test_spearman_cosine": 0.9043404713941784,
+      "eval_sts-test_spearman_dot": 0.8595169367156317,
+      "eval_sts-test_spearman_euclidean": 0.9055969973115261,
+      "eval_sts-test_spearman_manhattan": 0.9051065234866762,
+      "eval_sts-test_spearman_max": 0.9055969973115261,
+      "step": 310
+    },
+    {
+      "epoch": 0.6086105675146771,
+      "grad_norm": 3.6634974479675293,
+      "learning_rate": 1.7796348421915536e-05,
+      "loss": 0.1713,
+      "step": 311
+    },
+    {
+      "epoch": 0.6105675146771037,
+      "grad_norm": 4.3175225257873535,
+      "learning_rate": 1.7718709069673595e-05,
+      "loss": 0.1628,
+      "step": 312
+    },
+    {
+      "epoch": 0.6125244618395304,
+      "grad_norm": 3.73574161529541,
+      "learning_rate": 1.763990123036787e-05,
+      "loss": 0.1585,
+      "step": 313
+    },
+    {
+      "epoch": 0.6144814090019569,
+      "grad_norm": 3.8439183235168457,
+      "learning_rate": 1.7559936834223982e-05,
+      "loss": 0.1419,
+      "step": 314
+    },
+    {
+      "epoch": 0.6164383561643836,
+      "grad_norm": 2.908531904220581,
+      "learning_rate": 1.747882798655147e-05,
+      "loss": 0.1136,
+      "step": 315
+    },
+    {
+      "epoch": 0.6164383561643836,
+      "eval_loss": 0.09831386059522629,
+      "eval_runtime": 107.369,
+      "eval_samples_per_second": 28.425,
+      "eval_steps_per_second": 0.224,
+      "eval_sts-test_pearson_cosine": 0.8766201899554078,
+      "eval_sts-test_pearson_dot": 0.8489007626542151,
+      "eval_sts-test_pearson_euclidean": 0.907814904603313,
+      "eval_sts-test_pearson_manhattan": 0.9075136258935672,
+      "eval_sts-test_pearson_max": 0.907814904603313,
+      "eval_sts-test_spearman_cosine": 0.9040450683177336,
+      "eval_sts-test_spearman_dot": 0.856758468963466,
+      "eval_sts-test_spearman_euclidean": 0.9053801326988233,
+      "eval_sts-test_spearman_manhattan": 0.9047017483273913,
+      "eval_sts-test_spearman_max": 0.9053801326988233,
+      "step": 315
+    },
+    {
+      "epoch": 0.6183953033268101,
+      "grad_norm": 4.284037113189697,
+      "learning_rate": 1.739658696591121e-05,
+      "loss": 0.255,
+      "step": 316
+    },
+    {
+      "epoch": 0.6203522504892368,
+      "grad_norm": 3.051182270050049,
+      "learning_rate": 1.7313226222256675e-05,
+      "loss": 0.1262,
+      "step": 317
+    },
+    {
+      "epoch": 0.6223091976516634,
+      "grad_norm": 3.270893096923828,
+      "learning_rate": 1.7228758375049186e-05,
+      "loss": 0.1393,
+      "step": 318
+    },
+    {
+      "epoch": 0.62426614481409,
+      "grad_norm": 3.4940428733825684,
+      "learning_rate": 1.714319621134755e-05,
+      "loss": 0.1134,
+      "step": 319
+    },
+    {
+      "epoch": 0.6262230919765166,
+      "grad_norm": 3.899348258972168,
+      "learning_rate": 1.705655268387229e-05,
+      "loss": 0.1441,
+      "step": 320
+    },
+    {
+      "epoch": 0.6262230919765166,
+      "eval_loss": 0.09844871610403061,
+      "eval_runtime": 107.3824,
+      "eval_samples_per_second": 28.422,
+      "eval_steps_per_second": 0.224,
+      "eval_sts-test_pearson_cosine": 0.8758902555910153,
+      "eval_sts-test_pearson_dot": 0.8487463478874753,
+      "eval_sts-test_pearson_euclidean": 0.9067076042499689,
+      "eval_sts-test_pearson_manhattan": 0.9065947885559749,
+      "eval_sts-test_pearson_max": 0.9067076042499689,
+      "eval_sts-test_spearman_cosine": 0.9032847891379552,
+      "eval_sts-test_spearman_dot": 0.8557776108162892,
+      "eval_sts-test_spearman_euclidean": 0.9042920057780914,
+      "eval_sts-test_spearman_manhattan": 0.9038587688165614,
+      "eval_sts-test_spearman_max": 0.9042920057780914,
+      "step": 320
+    },
+    {
+      "epoch": 0.6281800391389433,
+      "grad_norm": 4.422016143798828,
+      "learning_rate": 1.696884090904484e-05,
+      "loss": 0.1744,
+      "step": 321
+    },
+    {
+      "epoch": 0.6301369863013698,
+      "grad_norm": 3.950225353240967,
+      "learning_rate": 1.6880074165001906e-05,
+      "loss": 0.2124,
+      "step": 322
+    },
+    {
+      "epoch": 0.6320939334637965,
+      "grad_norm": 3.2186155319213867,
+      "learning_rate": 1.6790265889585377e-05,
+      "loss": 0.1267,
+      "step": 323
+    },
+    {
+      "epoch": 0.6340508806262231,
+      "grad_norm": 3.156022548675537,
+      "learning_rate": 1.669942967830807e-05,
+      "loss": 0.1435,
+      "step": 324
+    },
+    {
+      "epoch": 0.6360078277886497,
+      "grad_norm": 3.511422634124756,
+      "learning_rate": 1.6607579282295572e-05,
+      "loss": 0.1705,
+      "step": 325
+    },
+    {
+      "epoch": 0.6360078277886497,
+      "eval_loss": 0.09894353151321411,
+      "eval_runtime": 107.3913,
+      "eval_samples_per_second": 28.419,
+      "eval_steps_per_second": 0.223,
+      "eval_sts-test_pearson_cosine": 0.8748236722620941,
+      "eval_sts-test_pearson_dot": 0.8489660029264176,
+      "eval_sts-test_pearson_euclidean": 0.9056717167596496,
+      "eval_sts-test_pearson_manhattan": 0.9057306950198961,
+      "eval_sts-test_pearson_max": 0.9057306950198961,
+      "eval_sts-test_spearman_cosine": 0.9023375391880836,
+      "eval_sts-test_spearman_dot": 0.8556132394331987,
+      "eval_sts-test_spearman_euclidean": 0.9032693499527753,
+      "eval_sts-test_spearman_manhattan": 0.9032065639330431,
+      "eval_sts-test_spearman_max": 0.9032693499527753,
+      "step": 325
+    },
+    {
+      "epoch": 0.6379647749510763,
+      "grad_norm": 3.2097976207733154,
+      "learning_rate": 1.651472860620455e-05,
+      "loss": 0.1441,
+      "step": 326
+    },
+    {
+      "epoch": 0.639921722113503,
+      "grad_norm": 3.0201833248138428,
+      "learning_rate": 1.6420891706117818e-05,
+      "loss": 0.118,
+      "step": 327
+    },
+    {
+      "epoch": 0.6418786692759295,
+      "grad_norm": 3.370908737182617,
+      "learning_rate": 1.6326082787416465e-05,
+      "loss": 0.1956,
+      "step": 328
+    },
+    {
+      "epoch": 0.6438356164383562,
+      "grad_norm": 2.768566131591797,
+      "learning_rate": 1.6230316202629393e-05,
+      "loss": 0.0803,
+      "step": 329
+    },
+    {
+      "epoch": 0.6457925636007827,
+      "grad_norm": 3.2455928325653076,
+      "learning_rate": 1.613360644926059e-05,
+      "loss": 0.1651,
+      "step": 330
+    },
+    {
+      "epoch": 0.6457925636007827,
+      "eval_loss": 0.09914453327655792,
+      "eval_runtime": 107.2995,
+      "eval_samples_per_second": 28.444,
+      "eval_steps_per_second": 0.224,
+      "eval_sts-test_pearson_cosine": 0.8743455111936658,
+      "eval_sts-test_pearson_dot": 0.8479229638933452,
+      "eval_sts-test_pearson_euclidean": 0.9055198964101038,
+      "eval_sts-test_pearson_manhattan": 0.9055992524553022,
+      "eval_sts-test_pearson_max": 0.9055992524553022,
+      "eval_sts-test_spearman_cosine": 0.9022275405875834,
+      "eval_sts-test_spearman_dot": 0.8543106197166178,
+      "eval_sts-test_spearman_euclidean": 0.9029672341871219,
+      "eval_sts-test_spearman_manhattan": 0.9028108285285591,
+      "eval_sts-test_spearman_max": 0.9029672341871219,
+      "step": 330
+    },
+    {
+      "epoch": 0.6477495107632094,
+      "grad_norm": 3.465236186981201,
+      "learning_rate": 1.603596816759442e-05,
+      "loss": 0.1498,
+      "step": 331
+    },
+    {
+      "epoch": 0.649706457925636,
+      "grad_norm": 3.303255558013916,
+      "learning_rate": 1.5937416138479344e-05,
+      "loss": 0.1171,
+      "step": 332
+    },
+    {
+      "epoch": 0.6516634050880626,
+      "grad_norm": 3.893554449081421,
+      "learning_rate": 1.5837965281090334e-05,
+      "loss": 0.1976,
+      "step": 333
+    },
+    {
+      "epoch": 0.6536203522504892,
+      "grad_norm": 2.688338041305542,
+      "learning_rate": 1.5737630650670336e-05,
+      "loss": 0.0926,
+      "step": 334
+    },
+    {
+      "epoch": 0.6555772994129159,
+      "grad_norm": 3.4313673973083496,
+      "learning_rate": 1.5636427436251182e-05,
+      "loss": 0.1496,
+      "step": 335
+    },
+    {
+      "epoch": 0.6555772994129159,
+      "eval_loss": 0.09911184757947922,
+      "eval_runtime": 107.4795,
+      "eval_samples_per_second": 28.396,
+      "eval_steps_per_second": 0.223,
+      "eval_sts-test_pearson_cosine": 0.8746652537201496,
+      "eval_sts-test_pearson_dot": 0.8492828619583224,
+      "eval_sts-test_pearson_euclidean": 0.906047319803132,
+      "eval_sts-test_pearson_manhattan": 0.9060247174283395,
+      "eval_sts-test_pearson_max": 0.906047319803132,
+      "eval_sts-test_spearman_cosine": 0.9026310179602884,
+      "eval_sts-test_spearman_dot": 0.856069836553175,
+      "eval_sts-test_spearman_euclidean": 0.9034376594468683,
+      "eval_sts-test_spearman_manhattan": 0.9036356837785253,
+      "eval_sts-test_spearman_max": 0.9036356837785253,
+      "step": 335
+    },
+    {
+      "epoch": 0.6575342465753424,
+      "grad_norm": 3.2240829467773438,
+      "learning_rate": 1.5534370958354184e-05,
+      "loss": 0.1131,
+      "step": 336
+    },
+    {
+      "epoch": 0.6594911937377691,
+      "grad_norm": 3.2019200325012207,
+      "learning_rate": 1.5431476666670885e-05,
+      "loss": 0.1352,
+      "step": 337
+    },
+    {
+      "epoch": 0.6614481409001957,
+      "grad_norm": 3.5696215629577637,
+      "learning_rate": 1.5327760137724213e-05,
+      "loss": 0.1608,
+      "step": 338
+    },
+    {
+      "epoch": 0.6634050880626223,
+      "grad_norm": 3.2444350719451904,
+      "learning_rate": 1.5223237072510433e-05,
+      "loss": 0.1239,
+      "step": 339
+    },
+    {
+      "epoch": 0.6653620352250489,
+      "grad_norm": 3.1613712310791016,
+      "learning_rate": 1.5117923294122312e-05,
+      "loss": 0.1227,
+      "step": 340
+    },
+    {
+      "epoch": 0.6653620352250489,
+      "eval_loss": 0.09929565340280533,
+      "eval_runtime": 107.2604,
+      "eval_samples_per_second": 28.454,
+      "eval_steps_per_second": 0.224,
+      "eval_sts-test_pearson_cosine": 0.874603261003018,
+      "eval_sts-test_pearson_dot": 0.8513553588561518,
+      "eval_sts-test_pearson_euclidean": 0.9056124511024704,
+      "eval_sts-test_pearson_manhattan": 0.9053134930024975,
+      "eval_sts-test_pearson_max": 0.9056124511024704,
+      "eval_sts-test_spearman_cosine": 0.9019408192558488,
+      "eval_sts-test_spearman_dot": 0.8587178581922269,
+      "eval_sts-test_spearman_euclidean": 0.9026150865112329,
+      "eval_sts-test_spearman_manhattan": 0.9023037967369943,
+      "eval_sts-test_spearman_max": 0.9026150865112329,
+      "step": 340
+    },
+    {
+      "epoch": 0.6673189823874756,
+      "grad_norm": 3.3650641441345215,
+      "learning_rate": 1.5011834745353725e-05,
+      "loss": 0.1452,
+      "step": 341
+    },
+    {
+      "epoch": 0.6692759295499021,
+      "grad_norm": 3.7061643600463867,
+      "learning_rate": 1.4904987486286184e-05,
+      "loss": 0.1992,
+      "step": 342
+    },
+    {
+      "epoch": 0.6712328767123288,
+      "grad_norm": 3.262500286102295,
+      "learning_rate": 1.4797397691857614e-05,
+      "loss": 0.1349,
+      "step": 343
+    },
+    {
+      "epoch": 0.6731898238747553,
+      "grad_norm": 3.4780774116516113,
+      "learning_rate": 1.468908164941371e-05,
+      "loss": 0.1702,
+      "step": 344
+    },
+    {
+      "epoch": 0.675146771037182,
+      "grad_norm": 2.908043146133423,
+      "learning_rate": 1.4580055756242315e-05,
+      "loss": 0.1033,
+      "step": 345
+    },
+    {
+      "epoch": 0.675146771037182,
+      "eval_loss": 0.09903653711080551,
+      "eval_runtime": 107.5298,
+      "eval_samples_per_second": 28.383,
+      "eval_steps_per_second": 0.223,
+      "eval_sts-test_pearson_cosine": 0.8748590023241125,
+      "eval_sts-test_pearson_dot": 0.8538817887560792,
+      "eval_sts-test_pearson_euclidean": 0.905694726781384,
+      "eval_sts-test_pearson_manhattan": 0.9051916896005284,
+      "eval_sts-test_pearson_max": 0.905694726781384,
+      "eval_sts-test_spearman_cosine": 0.9022368936012142,
+      "eval_sts-test_spearman_dot": 0.86127843586652,
+      "eval_sts-test_spearman_euclidean": 0.9024703161806319,
+      "eval_sts-test_spearman_manhattan": 0.9023726241770144,
+      "eval_sts-test_spearman_max": 0.9024703161806319,
+      "step": 345
+    },
+    {
+      "epoch": 0.6771037181996086,
+      "grad_norm": 3.7118523120880127,
+      "learning_rate": 1.4470336517091139e-05,
+      "loss": 0.1788,
+      "step": 346
+    },
+    {
+      "epoch": 0.6790606653620352,
+      "grad_norm": 3.106895923614502,
+      "learning_rate": 1.435994054166919e-05,
+      "loss": 0.1084,
+      "step": 347
+    },
+    {
+      "epoch": 0.6810176125244618,
+      "grad_norm": 3.782027244567871,
+      "learning_rate": 1.4248884542132348e-05,
+      "loss": 0.1325,
+      "step": 348
+    },
+    {
+      "epoch": 0.6829745596868885,
+      "grad_norm": 3.8729352951049805,
+      "learning_rate": 1.4137185330553416e-05,
+      "loss": 0.1537,
+      "step": 349
+    },
+    {
+      "epoch": 0.684931506849315,
+      "grad_norm": 3.7617311477661133,
+      "learning_rate": 1.4024859816377046e-05,
+      "loss": 0.2099,
+      "step": 350
+    },
+    {
+      "epoch": 0.684931506849315,
+      "eval_loss": 0.09886621683835983,
+      "eval_runtime": 107.402,
+      "eval_samples_per_second": 28.417,
+      "eval_steps_per_second": 0.223,
+      "eval_sts-test_pearson_cosine": 0.8751440351237663,
+      "eval_sts-test_pearson_dot": 0.8544171062513735,
+      "eval_sts-test_pearson_euclidean": 0.906084032839116,
+      "eval_sts-test_pearson_manhattan": 0.9052674845128671,
+      "eval_sts-test_pearson_max": 0.906084032839116,
+      "eval_sts-test_spearman_cosine": 0.9021938876390173,
+      "eval_sts-test_spearman_dot": 0.861093657908235,
+      "eval_sts-test_spearman_euclidean": 0.902945753581654,
+      "eval_sts-test_spearman_manhattan": 0.9021002232489249,
+      "eval_sts-test_spearman_max": 0.902945753581654,
+      "step": 350
+    },
+    {
+      "epoch": 0.6868884540117417,
+      "grad_norm": 3.419968366622925,
+      "learning_rate": 1.3911925003859907e-05,
+      "loss": 0.1603,
+      "step": 351
+    },
+    {
+      "epoch": 0.6888454011741683,
+      "grad_norm": 3.050192356109619,
+      "learning_rate": 1.3798397989496549e-05,
+      "loss": 0.0982,
+      "step": 352
+    },
+    {
+      "epoch": 0.6908023483365949,
+      "grad_norm": 3.8518471717834473,
+      "learning_rate": 1.3684295959431241e-05,
+      "loss": 0.1537,
+      "step": 353
+    },
+    {
+      "epoch": 0.6927592954990215,
+      "grad_norm": 3.516019582748413,
+      "learning_rate": 1.3569636186856286e-05,
+      "loss": 0.1758,
+      "step": 354
+    },
+    {
+      "epoch": 0.6947162426614482,
+      "grad_norm": 3.678056240081787,
+      "learning_rate": 1.3454436029397135e-05,
+      "loss": 0.1521,
+      "step": 355
+    },
+    {
+      "epoch": 0.6947162426614482,
+      "eval_loss": 0.09901077300310135,
+      "eval_runtime": 107.3981,
+      "eval_samples_per_second": 28.418,
+      "eval_steps_per_second": 0.223,
+      "eval_sts-test_pearson_cosine": 0.8757904015245911,
+      "eval_sts-test_pearson_dot": 0.8540145186864471,
+      "eval_sts-test_pearson_euclidean": 0.9064035478905541,
+      "eval_sts-test_pearson_manhattan": 0.9052325995566524,
+      "eval_sts-test_pearson_max": 0.9064035478905541,
+      "eval_sts-test_spearman_cosine": 0.9019130734737861,
+      "eval_sts-test_spearman_dot": 0.859719212417121,
+      "eval_sts-test_spearman_euclidean": 0.9030878388365718,
+      "eval_sts-test_spearman_manhattan": 0.9015811981193075,
+      "eval_sts-test_spearman_max": 0.9030878388365718,
+      "step": 355
+    },
+    {
+      "epoch": 0.6966731898238747,
+      "grad_norm": 2.466977119445801,
+      "learning_rate": 1.3338712926484722e-05,
+      "loss": 0.089,
+      "step": 356
+    },
+    {
+      "epoch": 0.6986301369863014,
+      "grad_norm": 3.8731167316436768,
+      "learning_rate": 1.322248439671543e-05,
+      "loss": 0.1509,
+      "step": 357
+    },
+    {
+      "epoch": 0.700587084148728,
+      "grad_norm": 4.406742572784424,
+      "learning_rate": 1.3105768035199033e-05,
+      "loss": 0.1943,
+      "step": 358
+    },
+    {
+      "epoch": 0.7025440313111546,
+      "grad_norm": 3.6811671257019043,
+      "learning_rate": 1.2988581510895118e-05,
+      "loss": 0.1582,
+      "step": 359
+    },
+    {
+      "epoch": 0.7045009784735812,
+      "grad_norm": 3.6019861698150635,
+      "learning_rate": 1.2870942563938265e-05,
+      "loss": 0.1527,
+      "step": 360
+    },
+    {
+      "epoch": 0.7045009784735812,
+      "eval_loss": 0.09933393448591232,
+      "eval_runtime": 107.4727,
+      "eval_samples_per_second": 28.398,
+      "eval_steps_per_second": 0.223,
+      "eval_sts-test_pearson_cosine": 0.8759901613929976,
+      "eval_sts-test_pearson_dot": 0.8539317550786957,
+      "eval_sts-test_pearson_euclidean": 0.9058198972317745,
+      "eval_sts-test_pearson_manhattan": 0.9045671303429671,
+      "eval_sts-test_pearson_max": 0.9058198972317745,
+      "eval_sts-test_spearman_cosine": 0.9007536577936579,
+      "eval_sts-test_spearman_dot": 0.8583788673871872,
+      "eval_sts-test_spearman_euclidean": 0.9019342408204244,
+      "eval_sts-test_spearman_manhattan": 0.9006401238435077,
+      "eval_sts-test_spearman_max": 0.9019342408204244,
+      "step": 360
+    },
+    {
+      "epoch": 0.7064579256360078,
+      "grad_norm": 2.7014875411987305,
+      "learning_rate": 1.2752869002952492e-05,
+      "loss": 0.0754,
+      "step": 361
+    },
+    {
+      "epoch": 0.7084148727984344,
+      "grad_norm": 3.4292407035827637,
+      "learning_rate": 1.2634378702355317e-05,
+      "loss": 0.122,
+      "step": 362
+    },
+    {
+      "epoch": 0.7103718199608611,
+      "grad_norm": 3.9553112983703613,
+      "learning_rate": 1.2515489599651846e-05,
+      "loss": 0.1727,
+      "step": 363
+    },
+    {
+      "epoch": 0.7123287671232876,
+      "grad_norm": 2.3133935928344727,
+      "learning_rate": 1.2396219692719364e-05,
+      "loss": 0.074,
+      "step": 364
+    },
+    {
+      "epoch": 0.7142857142857143,
+      "grad_norm": 4.136401176452637,
+      "learning_rate": 1.2276587037082707e-05,
+      "loss": 0.1822,
+      "step": 365
+    },
+    {
+      "epoch": 0.7142857142857143,
+      "eval_loss": 0.09977750480175018,
+      "eval_runtime": 107.5146,
+      "eval_samples_per_second": 28.387,
+      "eval_steps_per_second": 0.223,
+      "eval_sts-test_pearson_cosine": 0.8763860642424627,
+      "eval_sts-test_pearson_dot": 0.8548955898076025,
+      "eval_sts-test_pearson_euclidean": 0.9053617182443816,
+      "eval_sts-test_pearson_manhattan": 0.9041710686717819,
+      "eval_sts-test_pearson_max": 0.9053617182443816,
+      "eval_sts-test_spearman_cosine": 0.9004566884230645,
+      "eval_sts-test_spearman_dot": 0.8588638368068857,
+      "eval_sts-test_spearman_euclidean": 0.90101585543415,
+      "eval_sts-test_spearman_manhattan": 0.9002594696141124,
+      "eval_sts-test_spearman_max": 0.90101585543415,
+      "step": 365
+    },
+    {
+      "epoch": 0.7162426614481409,
+      "grad_norm": 3.4787533283233643,
+      "learning_rate": 1.215660974318097e-05,
+      "loss": 0.1344,
+      "step": 366
+    },
+    {
+      "epoch": 0.7181996086105675,
+      "grad_norm": 3.3478775024414062,
+      "learning_rate": 1.2036305973625881e-05,
+      "loss": 0.1819,
+      "step": 367
+    },
+    {
+      "epoch": 0.7201565557729941,
+      "grad_norm": 3.435234546661377,
+      "learning_rate": 1.191569394045228e-05,
+      "loss": 0.1811,
+      "step": 368
+    },
+    {
+      "epoch": 0.7221135029354208,
+      "grad_norm": 3.827272653579712,
+      "learning_rate": 1.1794791902361095e-05,
+      "loss": 0.1564,
+      "step": 369
+    },
+    {
+      "epoch": 0.7240704500978473,
+      "grad_norm": 4.088834762573242,
+      "learning_rate": 1.1673618161955288e-05,
+      "loss": 0.1522,
+      "step": 370
+    },
+    {
+      "epoch": 0.7240704500978473,
+      "eval_loss": 0.09972013533115387,
+      "eval_runtime": 107.5001,
+      "eval_samples_per_second": 28.391,
+      "eval_steps_per_second": 0.223,
+      "eval_sts-test_pearson_cosine": 0.8760486874181325,
+      "eval_sts-test_pearson_dot": 0.8547648526750222,
+      "eval_sts-test_pearson_euclidean": 0.9048215352945358,
+      "eval_sts-test_pearson_manhattan": 0.9037328672460638,
+      "eval_sts-test_pearson_max": 0.9048215352945358,
+      "eval_sts-test_spearman_cosine": 0.8997958912973589,
+      "eval_sts-test_spearman_dot": 0.859370958100973,
+      "eval_sts-test_spearman_euclidean": 0.9005396125104228,
+      "eval_sts-test_spearman_manhattan": 0.8996819545858564,
+      "eval_sts-test_spearman_max": 0.9005396125104228,
+      "step": 370
+    },
+    {
+      "epoch": 0.726027397260274,
+      "grad_norm": 3.117750406265259,
+      "learning_rate": 1.1552191062969147e-05,
+      "loss": 0.1379,
+      "step": 371
+    },
+    {
+      "epoch": 0.7279843444227005,
+      "grad_norm": 2.870415449142456,
+      "learning_rate": 1.1430528987491303e-05,
+      "loss": 0.082,
+      "step": 372
+    },
+    {
+      "epoch": 0.7299412915851272,
+      "grad_norm": 3.0934267044067383,
+      "learning_rate": 1.1308650353182036e-05,
+      "loss": 0.1288,
+      "step": 373
+    },
+    {
+      "epoch": 0.7318982387475538,
+      "grad_norm": 4.175031661987305,
+      "learning_rate": 1.1186573610485099e-05,
+      "loss": 0.1809,
+      "step": 374
+    },
+    {
+      "epoch": 0.7338551859099804,
+      "grad_norm": 4.580765724182129,
+      "learning_rate": 1.1064317239834628e-05,
+      "loss": 0.2418,
+      "step": 375
+    },
+    {
+      "epoch": 0.7338551859099804,
+      "eval_loss": 0.0990942195057869,
+      "eval_runtime": 107.4134,
+      "eval_samples_per_second": 28.414,
+      "eval_steps_per_second": 0.223,
+      "eval_sts-test_pearson_cosine": 0.8759927956331243,
+      "eval_sts-test_pearson_dot": 0.8545842177056733,
+      "eval_sts-test_pearson_euclidean": 0.9050231511967292,
+      "eval_sts-test_pearson_manhattan": 0.9040000709018708,
+      "eval_sts-test_pearson_max": 0.9050231511967292,
+      "eval_sts-test_spearman_cosine": 0.9005535301527153,
+      "eval_sts-test_spearman_dot": 0.8591386543030902,
+      "eval_sts-test_spearman_euclidean": 0.9009244733583888,
+      "eval_sts-test_spearman_manhattan": 0.9002575453098726,
+      "eval_sts-test_spearman_max": 0.9009244733583888,
+      "step": 375
+    },
+    {
+      "epoch": 0.735812133072407,
+      "grad_norm": 2.8520517349243164,
+      "learning_rate": 1.094189974885752e-05,
+      "loss": 0.0789,
+      "step": 376
+    },
+    {
+      "epoch": 0.7377690802348337,
+      "grad_norm": 3.5163254737854004,
+      "learning_rate": 1.081933966957167e-05,
+      "loss": 0.132,
+      "step": 377
+    },
+    {
+      "epoch": 0.7397260273972602,
+      "grad_norm": 3.160409688949585,
+      "learning_rate": 1.0696655555580527e-05,
+      "loss": 0.1425,
+      "step": 378
+    },
+    {
+      "epoch": 0.7416829745596869,
+      "grad_norm": 3.013707160949707,
+      "learning_rate": 1.0573865979264362e-05,
+      "loss": 0.1514,
+      "step": 379
+    },
+    {
+      "epoch": 0.7436399217221135,
+      "grad_norm": 2.8930296897888184,
+      "learning_rate": 1.0450989528968747e-05,
+      "loss": 0.0997,
+      "step": 380
+    },
+    {
+      "epoch": 0.7436399217221135,
+      "eval_loss": 0.09842444956302643,
+      "eval_runtime": 107.304,
+      "eval_samples_per_second": 28.443,
+      "eval_steps_per_second": 0.224,
+      "eval_sts-test_pearson_cosine": 0.8764482210558782,
+      "eval_sts-test_pearson_dot": 0.8556595611954538,
+      "eval_sts-test_pearson_euclidean": 0.9055523774056253,
+      "eval_sts-test_pearson_manhattan": 0.9046224074077234,
+      "eval_sts-test_pearson_max": 0.9055523774056253,
+      "eval_sts-test_spearman_cosine": 0.9017774324005088,
+      "eval_sts-test_spearman_dot": 0.860274840931642,
+      "eval_sts-test_spearman_euclidean": 0.9016713719110112,
+      "eval_sts-test_spearman_manhattan": 0.9011951289872838,
+      "eval_sts-test_spearman_max": 0.9017774324005088,
+      "step": 380
+    },
+    {
+      "epoch": 0.7455968688845401,
+      "grad_norm": 4.028687477111816,
+      "learning_rate": 1.0328044806190547e-05,
+      "loss": 0.2002,
+      "step": 381
+    },
+    {
+      "epoch": 0.7475538160469667,
+      "grad_norm": 3.8006536960601807,
+      "learning_rate": 1.0205050422761989e-05,
+      "loss": 0.1943,
+      "step": 382
+    },
+    {
+      "epoch": 0.7495107632093934,
+      "grad_norm": 2.689953088760376,
+      "learning_rate": 1.0082024998033092e-05,
+      "loss": 0.1198,
+      "step": 383
+    },
+    {
+      "epoch": 0.7514677103718199,
+      "grad_norm": 3.1326684951782227,
+      "learning_rate": 9.95898715605304e-06,
+      "loss": 0.1171,
+      "step": 384
+    },
+    {
+      "epoch": 0.7534246575342466,
+      "grad_norm": 2.8200089931488037,
+      "learning_rate": 9.835955522750789e-06,
+      "loss": 0.0872,
+      "step": 385
+    },
+    {
+      "epoch": 0.7534246575342466,
+      "eval_loss": 0.09781364351511002,
+      "eval_runtime": 107.4026,
+      "eval_samples_per_second": 28.416,
+      "eval_steps_per_second": 0.223,
+      "eval_sts-test_pearson_cosine": 0.8764091455338202,
+      "eval_sts-test_pearson_dot": 0.8558042892931453,
+      "eval_sts-test_pearson_euclidean": 0.9058212793054227,
+      "eval_sts-test_pearson_manhattan": 0.9049414370234095,
+      "eval_sts-test_pearson_max": 0.9058212793054227,
+      "eval_sts-test_spearman_cosine": 0.902642877044557,
+      "eval_sts-test_spearman_dot": 0.861618590204356,
+      "eval_sts-test_spearman_euclidean": 0.9022327317339048,
+      "eval_sts-test_spearman_manhattan": 0.9017312490987527,
+      "eval_sts-test_spearman_max": 0.902642877044557,
+      "step": 385
+    },
+    {
+      "epoch": 0.7553816046966731,
+      "grad_norm": 3.0478882789611816,
+      "learning_rate": 9.712948723115384e-06,
+      "loss": 0.0937,
+      "step": 386
+    },
+    {
+      "epoch": 0.7573385518590998,
+      "grad_norm": 3.0198819637298584,
+      "learning_rate": 9.589985378376474e-06,
+      "loss": 0.0933,
+      "step": 387
+    },
+    {
+      "epoch": 0.7592954990215264,
+      "grad_norm": 3.319575786590576,
+      "learning_rate": 9.46708410318533e-06,
+      "loss": 0.1109,
+      "step": 388
+    },
+    {
+      "epoch": 0.761252446183953,
+      "grad_norm": 3.1134960651397705,
+      "learning_rate": 9.344263502796918e-06,
+      "loss": 0.0999,
+      "step": 389
+    },
+    {
+      "epoch": 0.7632093933463796,
+      "grad_norm": 3.596510887145996,
+      "learning_rate": 9.221542170253334e-06,
+      "loss": 0.1625,
+      "step": 390
+    },
+    {
+      "epoch": 0.7632093933463796,
+      "eval_loss": 0.09730728715658188,
+      "eval_runtime": 107.3398,
+      "eval_samples_per_second": 28.433,
+      "eval_steps_per_second": 0.224,
+      "eval_sts-test_pearson_cosine": 0.8764763861944456,
+      "eval_sts-test_pearson_dot": 0.8561556513692123,
+      "eval_sts-test_pearson_euclidean": 0.9062056976600675,
+      "eval_sts-test_pearson_manhattan": 0.9053202975786792,
+      "eval_sts-test_pearson_max": 0.9062056976600675,
+      "eval_sts-test_spearman_cosine": 0.9034457146739187,
+      "eval_sts-test_spearman_dot": 0.8626341305791112,
+      "eval_sts-test_spearman_euclidean": 0.9030662687285811,
+      "eval_sts-test_spearman_manhattan": 0.9025959777226189,
+      "eval_sts-test_spearman_max": 0.9034457146739187,
+      "step": 390
+    },
+    {
+      "epoch": 0.7651663405088063,
+      "grad_norm": 2.9087250232696533,
+      "learning_rate": 9.098938683569155e-06,
+      "loss": 0.1357,
+      "step": 391
+    },
+    {
+      "epoch": 0.7671232876712328,
+      "grad_norm": 3.40970516204834,
+      "learning_rate": 8.97647160291899e-06,
+      "loss": 0.1202,
+      "step": 392
+    },
+    {
+      "epoch": 0.7690802348336595,
+      "grad_norm": 2.962822437286377,
+      "learning_rate": 8.854159467827808e-06,
+      "loss": 0.116,
+      "step": 393
+    },
+    {
+      "epoch": 0.7710371819960861,
+      "grad_norm": 3.2842860221862793,
+      "learning_rate": 8.732020794364327e-06,
+      "loss": 0.1256,
+      "step": 394
+    },
+    {
+      "epoch": 0.7729941291585127,
+      "grad_norm": 3.8854291439056396,
+      "learning_rate": 8.610074072338006e-06,
+      "loss": 0.2402,
+      "step": 395
+    },
+    {
+      "epoch": 0.7729941291585127,
+      "eval_loss": 0.09690071642398834,
+      "eval_runtime": 107.3243,
+      "eval_samples_per_second": 28.437,
+      "eval_steps_per_second": 0.224,
+      "eval_sts-test_pearson_cosine": 0.8764916753522476,
+      "eval_sts-test_pearson_dot": 0.8559063099369891,
+      "eval_sts-test_pearson_euclidean": 0.906553160229111,
+      "eval_sts-test_pearson_manhattan": 0.9056446394888255,
+      "eval_sts-test_pearson_max": 0.906553160229111,
+      "eval_sts-test_spearman_cosine": 0.9036960084788808,
+      "eval_sts-test_spearman_dot": 0.8628471465833341,
+      "eval_sts-test_spearman_euclidean": 0.9032961559583487,
+      "eval_sts-test_spearman_manhattan": 0.9026765747443849,
+      "eval_sts-test_spearman_max": 0.9036960084788808,
+      "step": 395
+    },
+    {
+      "epoch": 0.7749510763209393,
+      "grad_norm": 4.216285705566406,
+      "learning_rate": 8.488337762499971e-06,
+      "loss": 0.2413,
+      "step": 396
+    },
+    {
+      "epoch": 0.776908023483366,
+      "grad_norm": 3.007631540298462,
+      "learning_rate": 8.366830293748364e-06,
+      "loss": 0.1144,
+      "step": 397
+    },
+    {
+      "epoch": 0.7788649706457925,
+      "grad_norm": 3.3882946968078613,
+      "learning_rate": 8.245570060338511e-06,
+      "loss": 0.1198,
+      "step": 398
+    },
+    {
+      "epoch": 0.7808219178082192,
+      "grad_norm": 3.6439969539642334,
+      "learning_rate": 8.124575419098321e-06,
+      "loss": 0.1361,
+      "step": 399
+    },
+    {
+      "epoch": 0.7827788649706457,
+      "grad_norm": 3.3682761192321777,
+      "learning_rate": 8.003864686649369e-06,
+      "loss": 0.1496,
+      "step": 400
+    },
+    {
+      "epoch": 0.7827788649706457,
+      "eval_loss": 0.09637484699487686,
+      "eval_runtime": 107.2784,
+      "eval_samples_per_second": 28.449,
+      "eval_steps_per_second": 0.224,
+      "eval_sts-test_pearson_cosine": 0.8766720014290486,
+      "eval_sts-test_pearson_dot": 0.8562700614904787,
+      "eval_sts-test_pearson_euclidean": 0.9067249528004879,
+      "eval_sts-test_pearson_manhattan": 0.905876462094101,
+      "eval_sts-test_pearson_max": 0.9067249528004879,
+      "eval_sts-test_spearman_cosine": 0.9038214015132995,
+      "eval_sts-test_spearman_dot": 0.8632292328530939,
+      "eval_sts-test_spearman_euclidean": 0.9036244959631774,
+      "eval_sts-test_spearman_manhattan": 0.9027354226531145,
+      "eval_sts-test_spearman_max": 0.9038214015132995,
+      "step": 400
+    },
+    {
+      "epoch": 0.7847358121330724,
+      "grad_norm": 3.5519731044769287,
+      "learning_rate": 7.883456136634053e-06,
+      "loss": 0.1606,
+      "step": 401
+    },
+    {
+      "epoch": 0.786692759295499,
+      "grad_norm": 4.0106329917907715,
+      "learning_rate": 7.763367996949262e-06,
+      "loss": 0.1739,
+      "step": 402
+    },
+    {
+      "epoch": 0.7886497064579256,
+      "grad_norm": 3.347114086151123,
+      "learning_rate": 7.64361844698699e-06,
+      "loss": 0.1121,
+      "step": 403
+    },
+    {
+      "epoch": 0.7906066536203522,
+      "grad_norm": 3.429165840148926,
+      "learning_rate": 7.524225614882216e-06,
+      "loss": 0.1176,
+      "step": 404
+    },
+    {
+      "epoch": 0.7925636007827789,
+      "grad_norm": 3.169438362121582,
+      "learning_rate": 7.4052075747686625e-06,
+      "loss": 0.1024,
+      "step": 405
+    },
+    {
+      "epoch": 0.7925636007827789,
+      "eval_loss": 0.09552557021379471,
+      "eval_runtime": 107.3797,
+      "eval_samples_per_second": 28.423,
+      "eval_steps_per_second": 0.224,
+      "eval_sts-test_pearson_cosine": 0.8765829897566032,
+      "eval_sts-test_pearson_dot": 0.856005218083783,
+      "eval_sts-test_pearson_euclidean": 0.9066659750196452,
+      "eval_sts-test_pearson_manhattan": 0.9058351541345429,
+      "eval_sts-test_pearson_max": 0.9066659750196452,
+      "eval_sts-test_spearman_cosine": 0.9033677132253135,
+      "eval_sts-test_spearman_dot": 0.8628391808588065,
+      "eval_sts-test_spearman_euclidean": 0.9033363425910781,
+      "eval_sts-test_spearman_manhattan": 0.9027725662000693,
+      "eval_sts-test_spearman_max": 0.9033677132253135,
+      "step": 405
+    },
+    {
+      "epoch": 0.7945205479452054,
+      "grad_norm": 3.571244239807129,
+      "learning_rate": 7.286582344042625e-06,
+      "loss": 0.1256,
+      "step": 406
+    },
+    {
+      "epoch": 0.7964774951076321,
+      "grad_norm": 3.144022226333618,
+      "learning_rate": 7.168367880635454e-06,
+      "loss": 0.1424,
+      "step": 407
+    },
+    {
+      "epoch": 0.7984344422700587,
+      "grad_norm": 3.899695634841919,
+      "learning_rate": 7.050582080294996e-06,
+      "loss": 0.181,
+      "step": 408
+    },
+    {
+      "epoch": 0.8003913894324853,
+      "grad_norm": 2.7152762413024902,
+      "learning_rate": 6.933242773876481e-06,
+      "loss": 0.0829,
+      "step": 409
+    },
+    {
+      "epoch": 0.8023483365949119,
+      "grad_norm": 4.248819351196289,
+      "learning_rate": 6.816367724643225e-06,
+      "loss": 0.2329,
+      "step": 410
+    },
+    {
+      "epoch": 0.8023483365949119,
+      "eval_loss": 0.09485543519258499,
+      "eval_runtime": 107.35,
+      "eval_samples_per_second": 28.43,
+      "eval_steps_per_second": 0.224,
+      "eval_sts-test_pearson_cosine": 0.876662263871305,
+      "eval_sts-test_pearson_dot": 0.8565606973853341,
+      "eval_sts-test_pearson_euclidean": 0.906503720630583,
+      "eval_sts-test_pearson_manhattan": 0.9056869432887309,
+      "eval_sts-test_pearson_max": 0.906503720630583,
+      "eval_sts-test_spearman_cosine": 0.9031667800616662,
+      "eval_sts-test_spearman_dot": 0.8635160884386128,
+      "eval_sts-test_spearman_euclidean": 0.9027914064811151,
+      "eval_sts-test_spearman_manhattan": 0.9026432798059095,
+      "eval_sts-test_spearman_max": 0.9031667800616662,
+      "step": 410
+    },
+    {
+      "epoch": 0.8043052837573386,
+      "grad_norm": 2.724637508392334,
+      "learning_rate": 6.699974625577545e-06,
+      "loss": 0.075,
+      "step": 411
+    },
+    {
+      "epoch": 0.8062622309197651,
+      "grad_norm": 2.825380325317383,
+      "learning_rate": 6.584081096702343e-06,
+      "loss": 0.1157,
+      "step": 412
+    },
+    {
+      "epoch": 0.8082191780821918,
+      "grad_norm": 3.5397439002990723,
+      "learning_rate": 6.4687046824137115e-06,
+      "loss": 0.1383,
+      "step": 413
+    },
+    {
+      "epoch": 0.8101761252446184,
+      "grad_norm": 2.7954063415527344,
+      "learning_rate": 6.353862848825011e-06,
+      "loss": 0.1042,
+      "step": 414
+    },
+    {
+      "epoch": 0.812133072407045,
+      "grad_norm": 3.3894357681274414,
+      "learning_rate": 6.2395729811227635e-06,
+      "loss": 0.1352,
+      "step": 415
+    },
+    {
+      "epoch": 0.812133072407045,
+      "eval_loss": 0.09433061629533768,
+      "eval_runtime": 107.4406,
+      "eval_samples_per_second": 28.406,
+      "eval_steps_per_second": 0.223,
+      "eval_sts-test_pearson_cosine": 0.8762374184553781,
+      "eval_sts-test_pearson_dot": 0.8565984619458994,
+      "eval_sts-test_pearson_euclidean": 0.9059975243362595,
+      "eval_sts-test_pearson_manhattan": 0.9052369107641345,
+      "eval_sts-test_pearson_max": 0.9059975243362595,
+      "eval_sts-test_spearman_cosine": 0.9025091804482506,
+      "eval_sts-test_spearman_dot": 0.8638915067704253,
+      "eval_sts-test_spearman_euclidean": 0.9023325717980691,
+      "eval_sts-test_spearman_manhattan": 0.9021176314896061,
+      "eval_sts-test_spearman_max": 0.9025091804482506,
+      "step": 415
+    },
+    {
+      "epoch": 0.8140900195694716,
+      "grad_norm": 2.5251049995422363,
+      "learning_rate": 6.125852380934841e-06,
+      "loss": 0.0778,
+      "step": 416
+    },
+    {
+      "epoch": 0.8160469667318982,
+      "grad_norm": 2.7683308124542236,
+      "learning_rate": 6.012718263711261e-06,
+      "loss": 0.1006,
+      "step": 417
+    },
+    {
+      "epoch": 0.8180039138943248,
+      "grad_norm": 4.330955982208252,
+      "learning_rate": 5.900187756118055e-06,
+      "loss": 0.2188,
+      "step": 418
+    },
+    {
+      "epoch": 0.8199608610567515,
+      "grad_norm": 4.001932621002197,
+      "learning_rate": 5.788277893444574e-06,
+      "loss": 0.1338,
+      "step": 419
+    },
+    {
+      "epoch": 0.821917808219178,
+      "grad_norm": 3.1434426307678223,
+      "learning_rate": 5.677005617024618e-06,
+      "loss": 0.1314,
+      "step": 420
+    },
+    {
+      "epoch": 0.821917808219178,
+      "eval_loss": 0.09401828795671463,
+      "eval_runtime": 107.3518,
+      "eval_samples_per_second": 28.43,
+      "eval_steps_per_second": 0.224,
+      "eval_sts-test_pearson_cosine": 0.875859101651175,
+      "eval_sts-test_pearson_dot": 0.85667243561015,
+      "eval_sts-test_pearson_euclidean": 0.9054683468020812,
+      "eval_sts-test_pearson_manhattan": 0.9047678088467972,
+      "eval_sts-test_pearson_max": 0.9054683468020812,
+      "eval_sts-test_spearman_cosine": 0.9022520195275646,
+      "eval_sts-test_spearman_dot": 0.8640136777140237,
+      "eval_sts-test_spearman_euclidean": 0.9018411134454687,
+      "eval_sts-test_spearman_manhattan": 0.9017817285216024,
+      "eval_sts-test_spearman_max": 0.9022520195275646,
+      "step": 420
+    },
+    {
+      "epoch": 0.8238747553816047,
+      "grad_norm": 3.4323763847351074,
+      "learning_rate": 5.566387771671788e-06,
+      "loss": 0.1298,
+      "step": 421
+    },
+    {
+      "epoch": 0.8258317025440313,
+      "grad_norm": 3.4278817176818848,
+      "learning_rate": 5.4564411031294695e-06,
+      "loss": 0.1573,
+      "step": 422
+    },
+    {
+      "epoch": 0.8277886497064579,
+      "grad_norm": 3.182588577270508,
+      "learning_rate": 5.34718225553579e-06,
+      "loss": 0.1283,
+      "step": 423
+    },
+    {
+      "epoch": 0.8297455968688845,
+      "grad_norm": 3.8493266105651855,
+      "learning_rate": 5.238627768903952e-06,
+      "loss": 0.1998,
+      "step": 424
+    },
+    {
+      "epoch": 0.8317025440313112,
+      "grad_norm": 2.763507127761841,
+      "learning_rate": 5.130794076618395e-06,
+      "loss": 0.0747,
+      "step": 425
+    },
+    {
+      "epoch": 0.8317025440313112,
+      "eval_loss": 0.09391138702630997,
+      "eval_runtime": 107.4165,
+      "eval_samples_per_second": 28.413,
+      "eval_steps_per_second": 0.223,
+      "eval_sts-test_pearson_cosine": 0.8759008831174315,
+      "eval_sts-test_pearson_dot": 0.8568367129286429,
+      "eval_sts-test_pearson_euclidean": 0.9053146972146215,
+      "eval_sts-test_pearson_manhattan": 0.9046055003987086,
+      "eval_sts-test_pearson_max": 0.9053146972146215,
+      "eval_sts-test_spearman_cosine": 0.9019464579147841,
+      "eval_sts-test_spearman_dot": 0.8642364494932302,
+      "eval_sts-test_spearman_euclidean": 0.9015885820774372,
+      "eval_sts-test_spearman_manhattan": 0.9014194670606389,
+      "eval_sts-test_spearman_max": 0.9019464579147841,
+      "step": 425
+    },
+    {
+      "epoch": 0.8336594911937377,
+      "grad_norm": 3.613516092300415,
+      "learning_rate": 5.02369750294697e-06,
+      "loss": 0.142,
+      "step": 426
+    },
+    {
+      "epoch": 0.8356164383561644,
+      "grad_norm": 3.1161856651306152,
+      "learning_rate": 4.917354260569775e-06,
+      "loss": 0.131,
+      "step": 427
+    },
+    {
+      "epoch": 0.837573385518591,
+      "grad_norm": 2.7324466705322266,
+      "learning_rate": 4.811780448124812e-06,
+      "loss": 0.1037,
+      "step": 428
+    },
+    {
+      "epoch": 0.8395303326810176,
+      "grad_norm": 3.0681588649749756,
+      "learning_rate": 4.706992047770877e-06,
+      "loss": 0.1145,
+      "step": 429
+    },
+    {
+      "epoch": 0.8414872798434442,
+      "grad_norm": 3.202045202255249,
+      "learning_rate": 4.6030049227681484e-06,
+      "loss": 0.1371,
+      "step": 430
+    },
+    {
+      "epoch": 0.8414872798434442,
+      "eval_loss": 0.09390870481729507,
+      "eval_runtime": 107.3691,
+      "eval_samples_per_second": 28.425,
+      "eval_steps_per_second": 0.224,
+      "eval_sts-test_pearson_cosine": 0.8761008129035103,
+      "eval_sts-test_pearson_dot": 0.8571168858418021,
+      "eval_sts-test_pearson_euclidean": 0.9053023363687696,
+      "eval_sts-test_pearson_manhattan": 0.9045322392753761,
+      "eval_sts-test_pearson_max": 0.9053023363687696,
+      "eval_sts-test_spearman_cosine": 0.9020062008487415,
+      "eval_sts-test_spearman_dot": 0.8644168418278989,
+      "eval_sts-test_spearman_euclidean": 0.9017023845351555,
+      "eval_sts-test_spearman_manhattan": 0.9012797088713137,
+      "eval_sts-test_spearman_max": 0.9020062008487415,
+      "step": 430
+    },
+    {
+      "epoch": 0.8434442270058709,
+      "grad_norm": 3.2915422916412354,
+      "learning_rate": 4.4998348150767525e-06,
+      "loss": 0.1457,
+      "step": 431
+    },
+    {
+      "epoch": 0.8454011741682974,
+      "grad_norm": 2.914283514022827,
+      "learning_rate": 4.397497342973677e-06,
+      "loss": 0.123,
+      "step": 432
+    },
+    {
+      "epoch": 0.8473581213307241,
+      "grad_norm": 2.803455114364624,
+      "learning_rate": 4.296007998688405e-06,
+      "loss": 0.1312,
+      "step": 433
+    },
+    {
+      "epoch": 0.8493150684931506,
+      "grad_norm": 3.1922266483306885,
+      "learning_rate": 4.195382146057672e-06,
+      "loss": 0.1301,
+      "step": 434
+    },
+    {
+      "epoch": 0.8512720156555773,
+      "grad_norm": 3.4597818851470947,
+      "learning_rate": 4.095635018199612e-06,
+      "loss": 0.1324,
+      "step": 435
+    },
+    {
+      "epoch": 0.8512720156555773,
+      "eval_loss": 0.09401452541351318,
+      "eval_runtime": 107.4416,
+      "eval_samples_per_second": 28.406,
+      "eval_steps_per_second": 0.223,
+      "eval_sts-test_pearson_cosine": 0.8763389911185249,
+      "eval_sts-test_pearson_dot": 0.8577101143502727,
+      "eval_sts-test_pearson_euclidean": 0.9054263092988353,
+      "eval_sts-test_pearson_manhattan": 0.904625946001658,
+      "eval_sts-test_pearson_max": 0.9054263092988353,
+      "eval_sts-test_spearman_cosine": 0.902128819304954,
+      "eval_sts-test_spearman_dot": 0.8650113175842223,
+      "eval_sts-test_spearman_euclidean": 0.9018843879152342,
+      "eval_sts-test_spearman_manhattan": 0.9012396564923683,
+      "eval_sts-test_spearman_max": 0.902128819304954,
+      "step": 435
+    },
+    {
+      "epoch": 0.8532289628180039,
+      "grad_norm": 4.107654571533203,
+      "learning_rate": 3.996781715207706e-06,
+      "loss": 0.1826,
+      "step": 436
+    },
+    {
+      "epoch": 0.8551859099804305,
+      "grad_norm": 3.786817789077759,
+      "learning_rate": 3.898837201864893e-06,
+      "loss": 0.2145,
+      "step": 437
+    },
+    {
+      "epoch": 0.8571428571428571,
+      "grad_norm": 3.1700594425201416,
+      "learning_rate": 3.8018163053781243e-06,
+      "loss": 0.132,
+      "step": 438
+    },
+    {
+      "epoch": 0.8590998043052838,
+      "grad_norm": 3.451833963394165,
+      "learning_rate": 3.7057337131337822e-06,
+      "loss": 0.1263,
+      "step": 439
+    },
+    {
+      "epoch": 0.8610567514677103,
+      "grad_norm": 3.5724422931671143,
+      "learning_rate": 3.610603970474239e-06,
+      "loss": 0.1461,
+      "step": 440
+    },
+    {
+      "epoch": 0.8610567514677103,
+      "eval_loss": 0.09419582784175873,
+      "eval_runtime": 107.4099,
+      "eval_samples_per_second": 28.415,
+      "eval_steps_per_second": 0.223,
+      "eval_sts-test_pearson_cosine": 0.8765822162435379,
+      "eval_sts-test_pearson_dot": 0.858518150878541,
+      "eval_sts-test_pearson_euclidean": 0.9054994480778518,
+      "eval_sts-test_pearson_manhattan": 0.9046991633393977,
+      "eval_sts-test_pearson_max": 0.9054994480778518,
+      "eval_sts-test_spearman_cosine": 0.9021761661395061,
+      "eval_sts-test_spearman_dot": 0.8659527051188517,
+      "eval_sts-test_spearman_euclidean": 0.9019459656509089,
+      "eval_sts-test_spearman_manhattan": 0.9013235203562157,
+      "eval_sts-test_spearman_max": 0.9021761661395061,
+      "step": 440
+    },
+    {
+      "epoch": 0.863013698630137,
+      "grad_norm": 3.693023204803467,
+      "learning_rate": 3.5164414784959368e-06,
+      "loss": 0.1349,
+      "step": 441
+    },
+    {
+      "epoch": 0.8649706457925636,
+      "grad_norm": 2.6913726329803467,
+      "learning_rate": 3.423260491869276e-06,
+      "loss": 0.0824,
+      "step": 442
+    },
+    {
+      "epoch": 0.8669275929549902,
+      "grad_norm": 3.2380688190460205,
+      "learning_rate": 3.3310751166807186e-06,
+      "loss": 0.1639,
+      "step": 443
+    },
+    {
+      "epoch": 0.8688845401174168,
+      "grad_norm": 3.5086116790771484,
+      "learning_rate": 3.2398993082973294e-06,
+      "loss": 0.1403,
+      "step": 444
+    },
+    {
+      "epoch": 0.8708414872798435,
+      "grad_norm": 3.3980696201324463,
+      "learning_rate": 3.1497468692541812e-06,
+      "loss": 0.1766,
+      "step": 445
+    },
+    {
+      "epoch": 0.8708414872798435,
+      "eval_loss": 0.09430497139692307,
+      "eval_runtime": 107.5105,
+      "eval_samples_per_second": 28.388,
+      "eval_steps_per_second": 0.223,
+      "eval_sts-test_pearson_cosine": 0.8769023883634329,
+      "eval_sts-test_pearson_dot": 0.8592785166761266,
+      "eval_sts-test_pearson_euclidean": 0.9057520441250233,
+      "eval_sts-test_pearson_manhattan": 0.9049586591092414,
+      "eval_sts-test_pearson_max": 0.9057520441250233,
+      "eval_sts-test_spearman_cosine": 0.9023938810261752,
+      "eval_sts-test_spearman_dot": 0.8662211679359387,
+      "eval_sts-test_spearman_euclidean": 0.9022854934710853,
+      "eval_sts-test_spearman_manhattan": 0.9016112262157012,
+      "eval_sts-test_spearman_max": 0.9023938810261752,
+      "step": 445
+    },
+    {
+      "epoch": 0.87279843444227,
+      "grad_norm": 3.6057686805725098,
+      "learning_rate": 3.0606314471648667e-06,
+      "loss": 0.1402,
+      "step": 446
+    },
+    {
+      "epoch": 0.8747553816046967,
+      "grad_norm": 3.3049042224884033,
+      "learning_rate": 2.972566532655462e-06,
+      "loss": 0.1203,
+      "step": 447
+    },
+    {
+      "epoch": 0.8767123287671232,
+      "grad_norm": 3.3699026107788086,
+      "learning_rate": 2.8855654573222824e-06,
+      "loss": 0.1398,
+      "step": 448
+    },
+    {
+      "epoch": 0.8786692759295499,
+      "grad_norm": 4.141995906829834,
+      "learning_rate": 2.79964139171369e-06,
+      "loss": 0.2226,
+      "step": 449
+    },
+    {
+      "epoch": 0.8806262230919765,
+      "grad_norm": 3.005411386489868,
+      "learning_rate": 2.7148073433362732e-06,
+      "loss": 0.0943,
+      "step": 450
+    },
+    {
+      "epoch": 0.8806262230919765,
+      "eval_loss": 0.09433107823133469,
+      "eval_runtime": 107.5997,
+      "eval_samples_per_second": 28.364,
+      "eval_steps_per_second": 0.223,
+      "eval_sts-test_pearson_cosine": 0.8770850204583311,
+      "eval_sts-test_pearson_dot": 0.859396639463007,
+      "eval_sts-test_pearson_euclidean": 0.9058757986724881,
+      "eval_sts-test_pearson_manhattan": 0.9050657797722572,
+      "eval_sts-test_pearson_max": 0.9058757986724881,
+      "eval_sts-test_spearman_cosine": 0.9023927622446406,
+      "eval_sts-test_spearman_dot": 0.8658192568573825,
+      "eval_sts-test_spearman_euclidean": 0.9024802062093993,
+      "eval_sts-test_spearman_manhattan": 0.9016976856527094,
+      "eval_sts-test_spearman_max": 0.9024802062093993,
+      "step": 450
+    },
+    {
+      "epoch": 0.8825831702544031,
+      "grad_norm": 3.1855523586273193,
+      "learning_rate": 2.6310761546857433e-06,
+      "loss": 0.1101,
+      "step": 451
+    },
+    {
+      "epoch": 0.8845401174168297,
+      "grad_norm": 3.945880651473999,
+      "learning_rate": 2.5484605013027753e-06,
+      "loss": 0.1536,
+      "step": 452
+    },
+    {
+      "epoch": 0.8864970645792564,
+      "grad_norm": 3.068859100341797,
+      "learning_rate": 2.4669728898541456e-06,
+      "loss": 0.1159,
+      "step": 453
+    },
+    {
+      "epoch": 0.8884540117416829,
+      "grad_norm": 2.9130120277404785,
+      "learning_rate": 2.3866256562394084e-06,
+      "loss": 0.1373,
+      "step": 454
+    },
+    {
+      "epoch": 0.8904109589041096,
+      "grad_norm": 3.2793118953704834,
+      "learning_rate": 2.3074309637234673e-06,
+      "loss": 0.1412,
+      "step": 455
+    },
+    {
+      "epoch": 0.8904109589041096,
+      "eval_loss": 0.0943305641412735,
+      "eval_runtime": 107.5498,
+      "eval_samples_per_second": 28.378,
+      "eval_steps_per_second": 0.223,
+      "eval_sts-test_pearson_cosine": 0.8773045053021702,
+      "eval_sts-test_pearson_dot": 0.8595737170903098,
+      "eval_sts-test_pearson_euclidean": 0.9060339759187682,
+      "eval_sts-test_pearson_manhattan": 0.9052200019449477,
+      "eval_sts-test_pearson_max": 0.9060339759187682,
+      "eval_sts-test_spearman_cosine": 0.902556909871424,
+      "eval_sts-test_spearman_dot": 0.8656299142704356,
+      "eval_sts-test_spearman_euclidean": 0.9027095564240303,
+      "eval_sts-test_spearman_manhattan": 0.9019418485348608,
+      "eval_sts-test_spearman_max": 0.9027095564240303,
+      "step": 455
+    },
+    {
+      "epoch": 0.8923679060665362,
+      "grad_norm": 2.4695982933044434,
+      "learning_rate": 2.229400801095235e-06,
+      "loss": 0.0626,
+      "step": 456
+    },
+    {
+      "epoch": 0.8943248532289628,
+      "grad_norm": 3.314206838607788,
+      "learning_rate": 2.15254698085274e-06,
+      "loss": 0.1447,
+      "step": 457
+    },
+    {
+      "epoch": 0.8962818003913894,
+      "grad_norm": 2.9935507774353027,
+      "learning_rate": 2.07688113741488e-06,
+      "loss": 0.1296,
+      "step": 458
+    },
+    {
+      "epoch": 0.898238747553816,
+      "grad_norm": 3.59999942779541,
+      "learning_rate": 2.0024147253601957e-06,
+      "loss": 0.1407,
+      "step": 459
+    },
+    {
+      "epoch": 0.9001956947162426,
+      "grad_norm": 3.781248092651367,
+      "learning_rate": 1.92915901769281e-06,
+      "loss": 0.1966,
+      "step": 460
+    },
+    {
+      "epoch": 0.9001956947162426,
+      "eval_loss": 0.0943402424454689,
+      "eval_runtime": 107.5535,
+      "eval_samples_per_second": 28.377,
+      "eval_steps_per_second": 0.223,
+      "eval_sts-test_pearson_cosine": 0.8773605385802614,
+      "eval_sts-test_pearson_dot": 0.8594795998035439,
+      "eval_sts-test_pearson_euclidean": 0.906101891634431,
+      "eval_sts-test_pearson_manhattan": 0.9052843548013115,
+      "eval_sts-test_pearson_max": 0.906101891634431,
+      "eval_sts-test_spearman_cosine": 0.9026339267922787,
+      "eval_sts-test_spearman_dot": 0.8655708873566604,
+      "eval_sts-test_spearman_euclidean": 0.9027678225663618,
+      "eval_sts-test_spearman_manhattan": 0.9020037842806264,
+      "eval_sts-test_spearman_max": 0.9027678225663618,
+      "step": 460
+    },
+    {
+      "epoch": 0.9021526418786693,
+      "grad_norm": 3.1703717708587646,
+      "learning_rate": 1.8571251041358895e-06,
+      "loss": 0.1019,
+      "step": 461
+    },
+    {
+      "epoch": 0.9041095890410958,
+      "grad_norm": 3.5208213329315186,
+      "learning_rate": 1.786323889452828e-06,
+      "loss": 0.1332,
+      "step": 462
+    },
+    {
+      "epoch": 0.9060665362035225,
+      "grad_norm": 3.7119576930999756,
+      "learning_rate": 1.7167660917964557e-06,
+      "loss": 0.1417,
+      "step": 463
+    },
+    {
+      "epoch": 0.9080234833659491,
+      "grad_norm": 3.853672981262207,
+      "learning_rate": 1.6484622410864815e-06,
+      "loss": 0.1782,
+      "step": 464
+    },
+    {
+      "epoch": 0.9099804305283757,
+      "grad_norm": 3.186000347137451,
+      "learning_rate": 1.5814226774154351e-06,
+      "loss": 0.1068,
+      "step": 465
+    },
+    {
+      "epoch": 0.9099804305283757,
+      "eval_loss": 0.09426674991846085,
+      "eval_runtime": 107.6056,
+      "eval_samples_per_second": 28.363,
+      "eval_steps_per_second": 0.223,
+      "eval_sts-test_pearson_cosine": 0.8773874434779254,
+      "eval_sts-test_pearson_dot": 0.8592176171000234,
+      "eval_sts-test_pearson_euclidean": 0.9061215425720242,
+      "eval_sts-test_pearson_manhattan": 0.9052975958981246,
+      "eval_sts-test_pearson_max": 0.9061215425720242,
+      "eval_sts-test_spearman_cosine": 0.9025369508088437,
+      "eval_sts-test_spearman_dot": 0.8649846905836943,
+      "eval_sts-test_spearman_euclidean": 0.9027074083634833,
+      "eval_sts-test_spearman_manhattan": 0.9020343941434181,
+      "eval_sts-test_spearman_max": 0.9027074083634833,
+      "step": 465
+    },
+    {
+      "epoch": 0.9119373776908023,
+      "grad_norm": 3.7655446529388428,
+      "learning_rate": 1.515657549483328e-06,
+      "loss": 0.1292,
+      "step": 466
+    },
+    {
+      "epoch": 0.913894324853229,
+      "grad_norm": 4.406928539276123,
+      "learning_rate": 1.4511768130613434e-06,
+      "loss": 0.1896,
+      "step": 467
+    },
+    {
+      "epoch": 0.9158512720156555,
+      "grad_norm": 3.7682666778564453,
+      "learning_rate": 1.3879902294846559e-06,
+      "loss": 0.1597,
+      "step": 468
+    },
+    {
+      "epoch": 0.9178082191780822,
+      "grad_norm": 3.596527338027954,
+      "learning_rate": 1.3261073641747358e-06,
+      "loss": 0.1588,
+      "step": 469
+    },
+    {
+      "epoch": 0.9197651663405088,
+      "grad_norm": 3.1445114612579346,
+      "learning_rate": 1.2655375851913209e-06,
+      "loss": 0.13,
+      "step": 470
+    },
+    {
+      "epoch": 0.9197651663405088,
+      "eval_loss": 0.09422677010297775,
+      "eval_runtime": 107.5943,
+      "eval_samples_per_second": 28.366,
+      "eval_steps_per_second": 0.223,
+      "eval_sts-test_pearson_cosine": 0.8774123494705477,
+      "eval_sts-test_pearson_dot": 0.8591397738989044,
+      "eval_sts-test_pearson_euclidean": 0.9061478507699354,
+      "eval_sts-test_pearson_manhattan": 0.905330010345305,
+      "eval_sts-test_pearson_max": 0.9061478507699354,
+      "eval_sts-test_spearman_cosine": 0.9025525689990692,
+      "eval_sts-test_spearman_dot": 0.8648119954659849,
+      "eval_sts-test_spearman_euclidean": 0.9027778468489135,
+      "eval_sts-test_spearman_manhattan": 0.9020008754486362,
+      "eval_sts-test_spearman_max": 0.9027778468489135,
+      "step": 470
+    },
+    {
+      "epoch": 0.9217221135029354,
+      "grad_norm": 3.6816020011901855,
+      "learning_rate": 1.2062900618142136e-06,
+      "loss": 0.1593,
+      "step": 471
+    },
+    {
+      "epoch": 0.923679060665362,
+      "grad_norm": 3.369281530380249,
+      "learning_rate": 1.1483737631552161e-06,
+      "loss": 0.1137,
+      "step": 472
+    },
+    {
+      "epoch": 0.9256360078277887,
+      "grad_norm": 3.330564260482788,
+      "learning_rate": 1.0917974568003531e-06,
+      "loss": 0.1295,
+      "step": 473
+    },
+    {
+      "epoch": 0.9275929549902152,
+      "grad_norm": 3.9396634101867676,
+      "learning_rate": 1.0365697074826043e-06,
+      "loss": 0.1367,
+      "step": 474
+    },
+    {
+      "epoch": 0.9295499021526419,
+      "grad_norm": 3.385772228240967,
+      "learning_rate": 9.82698875785325e-07,
+      "loss": 0.107,
+      "step": 475
+    },
+    {
+      "epoch": 0.9295499021526419,
+      "eval_loss": 0.09417176246643066,
+      "eval_runtime": 107.4989,
+      "eval_samples_per_second": 28.391,
+      "eval_steps_per_second": 0.223,
+      "eval_sts-test_pearson_cosine": 0.8774075619796129,
+      "eval_sts-test_pearson_dot": 0.8590208674593851,
+      "eval_sts-test_pearson_euclidean": 0.9061610232160783,
+      "eval_sts-test_pearson_manhattan": 0.9053388130087265,
+      "eval_sts-test_pearson_max": 0.9061610232160783,
+      "eval_sts-test_spearman_cosine": 0.9025310883936013,
+      "eval_sts-test_spearman_dot": 0.8647318907080943,
+      "eval_sts-test_spearman_euclidean": 0.9027404347943903,
+      "eval_sts-test_spearman_manhattan": 0.9020158671212021,
+      "eval_sts-test_spearman_max": 0.9027404347943903,
+      "step": 475
+    },
+    {
+      "epoch": 0.9315068493150684,
+      "grad_norm": 2.921172618865967,
+      "learning_rate": 9.301931168766165e-07,
+      "loss": 0.1442,
+      "step": 476
+    },
+    {
+      "epoch": 0.9334637964774951,
+      "grad_norm": 4.05568790435791,
+      "learning_rate": 8.790603792747499e-07,
+      "loss": 0.1841,
+      "step": 477
+    },
+    {
+      "epoch": 0.9354207436399217,
+      "grad_norm": 1.999931812286377,
+      "learning_rate": 8.293084036448895e-07,
+      "loss": 0.0436,
+      "step": 478
+    },
+    {
+      "epoch": 0.9373776908023483,
+      "grad_norm": 2.164313793182373,
+      "learning_rate": 7.809447216272892e-07,
+      "loss": 0.0908,
+      "step": 479
+    },
+    {
+      "epoch": 0.9393346379647749,
+      "grad_norm": 3.318042755126953,
+      "learning_rate": 7.33976654697115e-07,
+      "loss": 0.1233,
+      "step": 480
+    },
+    {
+      "epoch": 0.9393346379647749,
+      "eval_loss": 0.09411043673753738,
+      "eval_runtime": 107.5241,
+      "eval_samples_per_second": 28.384,
+      "eval_steps_per_second": 0.223,
+      "eval_sts-test_pearson_cosine": 0.8774878626138263,
+      "eval_sts-test_pearson_dot": 0.8591304690756056,
+      "eval_sts-test_pearson_euclidean": 0.9062269621782642,
+      "eval_sts-test_pearson_manhattan": 0.9054142728411033,
+      "eval_sts-test_pearson_max": 0.9062269621782642,
+      "eval_sts-test_spearman_cosine": 0.9025523452427622,
+      "eval_sts-test_spearman_dot": 0.8649809762289989,
+      "eval_sts-test_spearman_euclidean": 0.9027271436697569,
+      "eval_sts-test_spearman_manhattan": 0.9021333391823545,
+      "eval_sts-test_spearman_max": 0.9027271436697569,
+      "step": 480
+    },
+    {
+      "epoch": 0.9412915851272016,
+      "grad_norm": 3.7038323879241943,
+      "learning_rate": 6.884113130561043e-07,
+      "loss": 0.1676,
+      "step": 481
+    },
+    {
+      "epoch": 0.9432485322896281,
+      "grad_norm": 3.2897257804870605,
+      "learning_rate": 6.442555945561923e-07,
+      "loss": 0.1449,
+      "step": 482
+    },
+    {
+      "epoch": 0.9452054794520548,
+      "grad_norm": 3.172556161880493,
+      "learning_rate": 6.015161836552764e-07,
+      "loss": 0.1234,
+      "step": 483
+    },
+    {
+      "epoch": 0.9471624266144814,
+      "grad_norm": 2.7821803092956543,
+      "learning_rate": 5.601995504053193e-07,
+      "loss": 0.076,
+      "step": 484
+    },
+    {
+      "epoch": 0.949119373776908,
+      "grad_norm": 3.796891927719116,
+      "learning_rate": 5.203119494728826e-07,
+      "loss": 0.1369,
+      "step": 485
+    },
+    {
+      "epoch": 0.949119373776908,
+      "eval_loss": 0.09407136589288712,
+      "eval_runtime": 107.5103,
+      "eval_samples_per_second": 28.388,
+      "eval_steps_per_second": 0.223,
+      "eval_sts-test_pearson_cosine": 0.8775371989941387,
+      "eval_sts-test_pearson_dot": 0.8591526007541124,
+      "eval_sts-test_pearson_euclidean": 0.9062640446951766,
+      "eval_sts-test_pearson_manhattan": 0.905456867436393,
+      "eval_sts-test_pearson_max": 0.9062640446951766,
+      "eval_sts-test_spearman_cosine": 0.902595753966312,
+      "eval_sts-test_spearman_dot": 0.8649717127178908,
+      "eval_sts-test_spearman_euclidean": 0.9027548894518198,
+      "eval_sts-test_spearman_manhattan": 0.9021964384609165,
+      "eval_sts-test_spearman_max": 0.9027548894518198,
+      "step": 485
+    },
+    {
+      "epoch": 0.9510763209393346,
+      "grad_norm": 3.227461576461792,
+      "learning_rate": 4.818594191922577e-07,
+      "loss": 0.144,
+      "step": 486
+    },
+    {
+      "epoch": 0.9530332681017613,
+      "grad_norm": 2.9072837829589844,
+      "learning_rate": 4.448477806513729e-07,
+      "loss": 0.0874,
+      "step": 487
+    },
+    {
+      "epoch": 0.9549902152641878,
+      "grad_norm": 4.316675186157227,
+      "learning_rate": 4.0928263681057845e-07,
+      "loss": 0.195,
+      "step": 488
+    },
+    {
+      "epoch": 0.9569471624266145,
+      "grad_norm": 3.2540643215179443,
+      "learning_rate": 3.7516937165444025e-07,
+      "loss": 0.1585,
+      "step": 489
+    },
+    {
+      "epoch": 0.958904109589041,
+      "grad_norm": 3.0417003631591797,
+      "learning_rate": 3.4251314937669313e-07,
+      "loss": 0.1152,
+      "step": 490
+    },
+    {
+      "epoch": 0.958904109589041,
+      "eval_loss": 0.09405405074357986,
+      "eval_runtime": 107.5662,
+      "eval_samples_per_second": 28.373,
+      "eval_steps_per_second": 0.223,
+      "eval_sts-test_pearson_cosine": 0.8775483547748348,
+      "eval_sts-test_pearson_dot": 0.8591693308199357,
+      "eval_sts-test_pearson_euclidean": 0.9062789492300676,
+      "eval_sts-test_pearson_manhattan": 0.9054771724847893,
+      "eval_sts-test_pearson_max": 0.9062789492300676,
+      "eval_sts-test_spearman_cosine": 0.9026681615072432,
+      "eval_sts-test_spearman_dot": 0.8650296656013927,
+      "eval_sts-test_spearman_euclidean": 0.9028729432793703,
+      "eval_sts-test_spearman_manhattan": 0.9022379228802262,
+      "eval_sts-test_spearman_max": 0.9028729432793703,
+      "step": 490
+    },
+    {
+      "epoch": 0.9608610567514677,
+      "grad_norm": 2.9117610454559326,
+      "learning_rate": 3.1131891359847397e-07,
+      "loss": 0.0862,
+      "step": 491
+    },
+    {
+      "epoch": 0.9628180039138943,
+      "grad_norm": 4.31304407119751,
+      "learning_rate": 2.8159138661992824e-07,
+      "loss": 0.2244,
+      "step": 492
+    },
+    {
+      "epoch": 0.9647749510763209,
+      "grad_norm": 3.088024854660034,
+      "learning_rate": 2.5333506870533484e-07,
+      "loss": 0.0987,
+      "step": 493
+    },
+    {
+      "epoch": 0.9667318982387475,
+      "grad_norm": 3.0071840286254883,
+      "learning_rate": 2.2655423740183925e-07,
+      "loss": 0.105,
+      "step": 494
+    },
+    {
+      "epoch": 0.9686888454011742,
+      "grad_norm": 3.72670316696167,
+      "learning_rate": 2.0125294689190555e-07,
+      "loss": 0.1777,
+      "step": 495
+    },
+    {
+      "epoch": 0.9686888454011742,
+      "eval_loss": 0.09404183179140091,
+      "eval_runtime": 107.5451,
+      "eval_samples_per_second": 28.379,
+      "eval_steps_per_second": 0.223,
+      "eval_sts-test_pearson_cosine": 0.8775779600592797,
+      "eval_sts-test_pearson_dot": 0.8592464798748846,
+      "eval_sts-test_pearson_euclidean": 0.9062982054856592,
+      "eval_sts-test_pearson_manhattan": 0.905493676973923,
+      "eval_sts-test_pearson_max": 0.9062982054856592,
+      "eval_sts-test_spearman_cosine": 0.9026481576934011,
+      "eval_sts-test_spearman_dot": 0.8650861416932687,
+      "eval_sts-test_spearman_euclidean": 0.9028581753631112,
+      "eval_sts-test_spearman_manhattan": 0.9022116538897895,
+      "eval_sts-test_spearman_max": 0.9028581753631112,
+      "step": 495
+    },
+    {
+      "epoch": 0.9706457925636007,
+      "grad_norm": 3.807492256164551,
+      "learning_rate": 1.7743502737957107e-07,
+      "loss": 0.1838,
+      "step": 496
+    },
+    {
+      "epoch": 0.9726027397260274,
+      "grad_norm": 3.7314093112945557,
+      "learning_rate": 1.5510408451062552e-07,
+      "loss": 0.1813,
+      "step": 497
+    },
+    {
+      "epoch": 0.974559686888454,
+      "grad_norm": 3.0683677196502686,
+      "learning_rate": 1.3426349882676326e-07,
+      "loss": 0.0994,
+      "step": 498
+    },
+    {
+      "epoch": 0.9765166340508806,
+      "grad_norm": 3.496849298477173,
+      "learning_rate": 1.1491642525383595e-07,
+      "loss": 0.1576,
+      "step": 499
+    },
+    {
+      "epoch": 0.9784735812133072,
+      "grad_norm": 3.3572986125946045,
+      "learning_rate": 9.706579262424243e-08,
+      "loss": 0.1298,
+      "step": 500
+    },
+    {
+      "epoch": 0.9784735812133072,
+      "eval_loss": 0.09404946863651276,
+      "eval_runtime": 107.5442,
+      "eval_samples_per_second": 28.379,
+      "eval_steps_per_second": 0.223,
+      "eval_sts-test_pearson_cosine": 0.8775676935231487,
+      "eval_sts-test_pearson_dot": 0.8592566533355803,
+      "eval_sts-test_pearson_euclidean": 0.9062872900289646,
+      "eval_sts-test_pearson_manhattan": 0.9054889721563202,
+      "eval_sts-test_pearson_max": 0.9062872900289646,
+      "eval_sts-test_spearman_cosine": 0.9026514245354827,
+      "eval_sts-test_spearman_dot": 0.8649815132441356,
+      "eval_sts-test_spearman_euclidean": 0.9028325328903337,
+      "eval_sts-test_spearman_manhattan": 0.9022372963625668,
+      "eval_sts-test_spearman_max": 0.9028325328903337,
+      "step": 500
+    },
+    {
+      "epoch": 0.9804305283757339,
+      "grad_norm": 3.594297170639038,
+      "learning_rate": 8.071430323354778e-08,
+      "loss": 0.1884,
+      "step": 501
+    },
+    {
+      "epoch": 0.9823874755381604,
+      "grad_norm": 2.890631914138794,
+      "learning_rate": 6.586443243140839e-08,
+      "loss": 0.1032,
+      "step": 502
+    },
+    {
+      "epoch": 0.9843444227005871,
+      "grad_norm": 2.5980706214904785,
+      "learning_rate": 5.251842824683717e-08,
+      "loss": 0.1164,
+      "step": 503
+    },
+    {
+      "epoch": 0.9863013698630136,
+      "grad_norm": 3.6865243911743164,
+      "learning_rate": 4.067831104789033e-08,
+      "loss": 0.1466,
+      "step": 504
+    },
+    {
+      "epoch": 0.9882583170254403,
+      "grad_norm": 4.4932637214660645,
+      "learning_rate": 3.034587323581639e-08,
+      "loss": 0.2192,
+      "step": 505
+    },
+    {
+      "epoch": 0.9882583170254403,
+      "eval_loss": 0.09403973817825317,
+      "eval_runtime": 107.6061,
+      "eval_samples_per_second": 28.363,
+      "eval_steps_per_second": 0.223,
+      "eval_sts-test_pearson_cosine": 0.8775751164978143,
+      "eval_sts-test_pearson_dot": 0.859260290269255,
+      "eval_sts-test_pearson_euclidean": 0.906298388123929,
+      "eval_sts-test_pearson_manhattan": 0.9055001359483166,
+      "eval_sts-test_pearson_max": 0.906298388123929,
+      "eval_sts-test_spearman_cosine": 0.9026775592721354,
+      "eval_sts-test_spearman_dot": 0.8650394661276374,
+      "eval_sts-test_spearman_euclidean": 0.9028324881390722,
+      "eval_sts-test_spearman_manhattan": 0.9022903713585769,
+      "eval_sts-test_spearman_max": 0.9028324881390722,
+      "step": 505
+    },
+    {
+      "epoch": 0.9902152641878669,
+      "grad_norm": 2.892927408218384,
+      "learning_rate": 2.1522678973718848e-08,
+      "loss": 0.1302,
+      "step": 506
+    },
+    {
+      "epoch": 0.9921722113502935,
+      "grad_norm": 3.516493320465088,
+      "learning_rate": 1.421006394976221e-08,
+      "loss": 0.1371,
+      "step": 507
+    },
+    {
+      "epoch": 0.9941291585127201,
+      "grad_norm": 2.9676320552825928,
+      "learning_rate": 8.40913517497377e-09,
+      "loss": 0.1543,
+      "step": 508
+    },
+    {
+      "epoch": 0.9960861056751468,
+      "grad_norm": Infinity,
+      "learning_rate": 8.40913517497377e-09,
+      "loss": 0.1084,
+      "step": 509
+    },
+    {
+      "epoch": 0.9980430528375733,
+      "grad_norm": 0.0,
+      "learning_rate": 4.120770815659869e-09,
+      "loss": 0.0,
+      "step": 510
+    },
+    {
+      "epoch": 0.9980430528375733,
+      "eval_loss": 0.09404201060533524,
+      "eval_runtime": 107.6151,
+      "eval_samples_per_second": 28.36,
+      "eval_steps_per_second": 0.223,
+      "eval_sts-test_pearson_cosine": 0.8775732893584032,
+      "eval_sts-test_pearson_dot": 0.8592601719833955,
+      "eval_sts-test_pearson_euclidean": 0.9062985160748658,
+      "eval_sts-test_pearson_manhattan": 0.9054996976916896,
+      "eval_sts-test_pearson_max": 0.9062985160748658,
+      "eval_sts-test_spearman_cosine": 0.9026670874769698,
+      "eval_sts-test_spearman_dot": 0.8650449257815273,
+      "eval_sts-test_spearman_euclidean": 0.9028166014412784,
+      "eval_sts-test_spearman_manhattan": 0.9022697410270756,
+      "eval_sts-test_spearman_max": 0.9028166014412784,
+      "step": 510
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 0.0,
+      "learning_rate": 1.345620060465569e-09,
+      "loss": 0.0,
+      "step": 511
+    },
+    {
+      "epoch": 1.0019569471624266,
+      "grad_norm": 0.0,
+      "learning_rate": 8.410302209660437e-11,
+      "loss": 0.0,
+      "step": 512
+    },
+    {
+      "epoch": 1.0039138943248533,
+      "grad_norm": 0.0,
+      "learning_rate": 1.999966358932628e-05,
+      "loss": 0.0,
+      "step": 513
+    },
+    {
+      "epoch": 1.00587084148728,
+      "grad_norm": 0.0,
+      "learning_rate": 1.9997897495179932e-05,
+      "loss": 0.0,
+      "step": 514
+    },
+    {
+      "epoch": 1.0078277886497065,
+      "grad_norm": 0.0,
+      "learning_rate": 1.999461788189681e-05,
+      "loss": 0.0,
+      "step": 515
+    },
+    {
+      "epoch": 1.0078277886497065,
+      "eval_loss": 0.09377636015415192,
+      "eval_runtime": 107.6229,
+      "eval_samples_per_second": 28.358,
+      "eval_steps_per_second": 0.223,
+      "eval_sts-test_pearson_cosine": 0.8778303207829065,
+      "eval_sts-test_pearson_dot": 0.8593106541530915,
+      "eval_sts-test_pearson_euclidean": 0.9066522007596656,
+      "eval_sts-test_pearson_manhattan": 0.9058577790072662,
+      "eval_sts-test_pearson_max": 0.9066522007596656,
+      "eval_sts-test_spearman_cosine": 0.9030437588441012,
+      "eval_sts-test_spearman_dot": 0.865145302860828,
+      "eval_sts-test_spearman_euclidean": 0.9032262992393169,
+      "eval_sts-test_spearman_manhattan": 0.9024990017391836,
+      "eval_sts-test_spearman_max": 0.9032262992393169,
+      "step": 515
+    }
+  ],
+  "logging_steps": 1,
+  "max_steps": 1022,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 2,
+  "save_steps": 103,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 0.0,
+  "train_batch_size": 320,
+  "trial_name": null,
+  "trial_params": null
+}