diff --git "a/checkpoint-25830/trainer_state.json" "b/checkpoint-25830/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/checkpoint-25830/trainer_state.json"
@@ -0,0 +1,6935 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 5.0,
+  "eval_steps": 1292,
+  "global_step": 25830,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.012582268679829655,
+      "grad_norm": 4.108283996582031,
+      "learning_rate": 3.7166085946573755e-07,
+      "loss": 0.4577,
+      "step": 65
+    },
+    {
+      "epoch": 0.02516453735965931,
+      "grad_norm": 0.5773646831512451,
+      "learning_rate": 7.491289198606272e-07,
+      "loss": 0.4707,
+      "step": 130
+    },
+    {
+      "epoch": 0.03774680603948897,
+      "grad_norm": 0.6127182841300964,
+      "learning_rate": 1.1265969802555168e-06,
+      "loss": 0.5259,
+      "step": 195
+    },
+    {
+      "epoch": 0.05032907471931862,
+      "grad_norm": 1.513021469116211,
+      "learning_rate": 1.5040650406504067e-06,
+      "loss": 0.5501,
+      "step": 260
+    },
+    {
+      "epoch": 0.06291134339914828,
+      "grad_norm": 11.750137329101562,
+      "learning_rate": 1.8815331010452962e-06,
+      "loss": 0.5089,
+      "step": 325
+    },
+    {
+      "epoch": 0.07549361207897794,
+      "grad_norm": 4.16257905960083,
+      "learning_rate": 2.259001161440186e-06,
+      "loss": 0.4816,
+      "step": 390
+    },
+    {
+      "epoch": 0.08807588075880758,
+      "grad_norm": 16.527780532836914,
+      "learning_rate": 2.6306620209059233e-06,
+      "loss": 0.5822,
+      "step": 455
+    },
+    {
+      "epoch": 0.10065814943863724,
+      "grad_norm": 0.1845797300338745,
+      "learning_rate": 3.0081300813008134e-06,
+      "loss": 0.5686,
+      "step": 520
+    },
+    {
+      "epoch": 0.1132404181184669,
+      "grad_norm": 12.172422409057617,
+      "learning_rate": 3.3855981416957026e-06,
+      "loss": 0.5686,
+      "step": 585
+    },
+    {
+      "epoch": 0.12582268679829656,
+      "grad_norm": 0.6270273923873901,
+      "learning_rate": 3.7630662020905923e-06,
+      "loss": 0.517,
+      "step": 650
+    },
+    {
+      "epoch": 0.1384049554781262,
+      "grad_norm": 3.6368539333343506,
+      "learning_rate": 4.140534262485482e-06,
+      "loss": 0.3615,
+      "step": 715
+    },
+    {
+      "epoch": 0.15098722415795587,
+      "grad_norm": 9.541145324707031,
+      "learning_rate": 4.518002322880372e-06,
+      "loss": 0.5978,
+      "step": 780
+    },
+    {
+      "epoch": 0.16356949283778552,
+      "grad_norm": 9.86439323425293,
+      "learning_rate": 4.895470383275261e-06,
+      "loss": 0.5153,
+      "step": 845
+    },
+    {
+      "epoch": 0.17615176151761516,
+      "grad_norm": 4.669048309326172,
+      "learning_rate": 5.272938443670151e-06,
+      "loss": 0.5059,
+      "step": 910
+    },
+    {
+      "epoch": 0.18873403019744484,
+      "grad_norm": 9.666926383972168,
+      "learning_rate": 5.650406504065041e-06,
+      "loss": 0.5624,
+      "step": 975
+    },
+    {
+      "epoch": 0.20131629887727448,
+      "grad_norm": 6.078874588012695,
+      "learning_rate": 6.02787456445993e-06,
+      "loss": 0.5201,
+      "step": 1040
+    },
+    {
+      "epoch": 0.21389856755710415,
+      "grad_norm": 1.1067451238632202,
+      "learning_rate": 6.4053426248548205e-06,
+      "loss": 0.6127,
+      "step": 1105
+    },
+    {
+      "epoch": 0.2264808362369338,
+      "grad_norm": 1.1589373350143433,
+      "learning_rate": 6.78281068524971e-06,
+      "loss": 0.5333,
+      "step": 1170
+    },
+    {
+      "epoch": 0.23906310491676344,
+      "grad_norm": 1.977501630783081,
+      "learning_rate": 7.1602787456446e-06,
+      "loss": 0.494,
+      "step": 1235
+    },
+    {
+      "epoch": 0.2500967866821525,
+      "eval_StS-test_pearson_cosine": 0.8821101738384596,
+      "eval_StS-test_pearson_dot": 0.8032893366124795,
+      "eval_StS-test_pearson_euclidean": 0.8697205121607111,
+      "eval_StS-test_pearson_manhattan": 0.8704995590187196,
+      "eval_StS-test_pearson_max": 0.8821101738384596,
+      "eval_StS-test_spearman_cosine": 0.8943047751560564,
+      "eval_StS-test_spearman_dot": 0.8087424893555902,
+      "eval_StS-test_spearman_euclidean": 0.871583089708652,
+      "eval_StS-test_spearman_manhattan": 0.8737012027236009,
+      "eval_StS-test_spearman_max": 0.8943047751560564,
+      "eval_Vitaminc-test_cosine_accuracy": 0.5684210526315789,
+      "eval_Vitaminc-test_cosine_accuracy_threshold": 0.7028586268424988,
+      "eval_Vitaminc-test_cosine_ap": 0.5651043866206488,
+      "eval_Vitaminc-test_cosine_f1": 0.6755218216318786,
+      "eval_Vitaminc-test_cosine_f1_threshold": 0.5077509880065918,
+      "eval_Vitaminc-test_cosine_precision": 0.52046783625731,
+      "eval_Vitaminc-test_cosine_recall": 0.9621621621621622,
+      "eval_Vitaminc-test_dot_accuracy": 0.5684210526315789,
+      "eval_Vitaminc-test_dot_accuracy_threshold": 19.693286895751953,
+      "eval_Vitaminc-test_dot_ap": 0.5463931769790206,
+      "eval_Vitaminc-test_dot_f1": 0.6691449814126395,
+      "eval_Vitaminc-test_dot_f1_threshold": 13.839346885681152,
+      "eval_Vitaminc-test_dot_precision": 0.509915014164306,
+      "eval_Vitaminc-test_dot_recall": 0.972972972972973,
+      "eval_Vitaminc-test_euclidean_accuracy": 0.5894736842105263,
+      "eval_Vitaminc-test_euclidean_accuracy_threshold": 4.252468585968018,
+      "eval_Vitaminc-test_euclidean_ap": 0.5569049511912931,
+      "eval_Vitaminc-test_euclidean_f1": 0.6666666666666666,
+      "eval_Vitaminc-test_euclidean_f1_threshold": 6.922356128692627,
+      "eval_Vitaminc-test_euclidean_precision": 0.5041551246537396,
+      "eval_Vitaminc-test_euclidean_recall": 0.9837837837837838,
+      "eval_Vitaminc-test_manhattan_accuracy": 0.5815789473684211,
+      "eval_Vitaminc-test_manhattan_accuracy_threshold": 87.21337890625,
+      "eval_Vitaminc-test_manhattan_ap": 0.5572154085134091,
+      "eval_Vitaminc-test_manhattan_f1": 0.6666666666666667,
+      "eval_Vitaminc-test_manhattan_f1_threshold": 141.26380920410156,
+      "eval_Vitaminc-test_manhattan_precision": 0.505586592178771,
+      "eval_Vitaminc-test_manhattan_recall": 0.9783783783783784,
+      "eval_Vitaminc-test_max_accuracy": 0.5894736842105263,
+      "eval_Vitaminc-test_max_accuracy_threshold": 87.21337890625,
+      "eval_Vitaminc-test_max_ap": 0.5651043866206488,
+      "eval_Vitaminc-test_max_f1": 0.6755218216318786,
+      "eval_Vitaminc-test_max_f1_threshold": 141.26380920410156,
+      "eval_Vitaminc-test_max_precision": 0.52046783625731,
+      "eval_Vitaminc-test_max_recall": 0.9837837837837838,
+      "eval_mrpc-test_cosine_accuracy": 0.7473684210526316,
+      "eval_mrpc-test_cosine_accuracy_threshold": 0.7145693302154541,
+      "eval_mrpc-test_cosine_ap": 0.8563235829800693,
+      "eval_mrpc-test_cosine_f1": 0.8327645051194539,
+      "eval_mrpc-test_cosine_f1_threshold": 0.6522408723831177,
+      "eval_mrpc-test_cosine_precision": 0.7218934911242604,
+      "eval_mrpc-test_cosine_recall": 0.9838709677419355,
+      "eval_mrpc-test_dot_accuracy": 0.7026315789473684,
+      "eval_mrpc-test_dot_accuracy_threshold": 14.454626083374023,
+      "eval_mrpc-test_dot_ap": 0.796363256728503,
+      "eval_mrpc-test_dot_f1": 0.8054607508532423,
+      "eval_mrpc-test_dot_f1_threshold": 13.752894401550293,
+      "eval_mrpc-test_dot_precision": 0.6982248520710059,
+      "eval_mrpc-test_dot_recall": 0.9516129032258065,
+      "eval_mrpc-test_euclidean_accuracy": 0.7315789473684211,
+      "eval_mrpc-test_euclidean_accuracy_threshold": 3.890326499938965,
+      "eval_mrpc-test_euclidean_ap": 0.8252367395643119,
+      "eval_mrpc-test_euclidean_f1": 0.8165467625899281,
+      "eval_mrpc-test_euclidean_f1_threshold": 3.890326499938965,
+      "eval_mrpc-test_euclidean_precision": 0.737012987012987,
+      "eval_mrpc-test_euclidean_recall": 0.9153225806451613,
+      "eval_mrpc-test_manhattan_accuracy": 0.7289473684210527,
+      "eval_mrpc-test_manhattan_accuracy_threshold": 77.57926177978516,
+      "eval_mrpc-test_manhattan_ap": 0.8208816982117964,
+      "eval_mrpc-test_manhattan_f1": 0.815742397137746,
+      "eval_mrpc-test_manhattan_f1_threshold": 79.14703369140625,
+      "eval_mrpc-test_manhattan_precision": 0.7331189710610932,
+      "eval_mrpc-test_manhattan_recall": 0.9193548387096774,
+      "eval_mrpc-test_max_accuracy": 0.7473684210526316,
+      "eval_mrpc-test_max_accuracy_threshold": 77.57926177978516,
+      "eval_mrpc-test_max_ap": 0.8563235829800693,
+      "eval_mrpc-test_max_f1": 0.8327645051194539,
+      "eval_mrpc-test_max_f1_threshold": 79.14703369140625,
+      "eval_mrpc-test_max_precision": 0.737012987012987,
+      "eval_mrpc-test_max_recall": 0.9838709677419355,
+      "eval_nli-pairs_loss": 0.8093397617340088,
+      "eval_nli-pairs_runtime": 3.0363,
+      "eval_nli-pairs_samples_per_second": 52.696,
+      "eval_nli-pairs_steps_per_second": 1.647,
+      "eval_sequential_score": 0.5651043866206488,
+      "step": 1292
+    },
+    {
+      "epoch": 0.2500967866821525,
+      "eval_vitaminc-pairs_loss": 5.769770622253418,
+      "eval_vitaminc-pairs_runtime": 1.5488,
+      "eval_vitaminc-pairs_samples_per_second": 85.875,
+      "eval_vitaminc-pairs_steps_per_second": 3.228,
+      "step": 1292
+    },
+    {
+      "epoch": 0.2500967866821525,
+      "eval_qnli-contrastive_loss": 0.12109158933162689,
+      "eval_qnli-contrastive_runtime": 0.5097,
+      "eval_qnli-contrastive_samples_per_second": 313.889,
+      "eval_qnli-contrastive_steps_per_second": 9.809,
+      "step": 1292
+    },
+    {
+      "epoch": 0.2500967866821525,
+      "eval_scitail-pairs-qa_loss": 0.07553695887327194,
+      "eval_scitail-pairs-qa_runtime": 1.2071,
+      "eval_scitail-pairs-qa_samples_per_second": 132.548,
+      "eval_scitail-pairs-qa_steps_per_second": 4.142,
+      "step": 1292
+    },
+    {
+      "epoch": 0.2500967866821525,
+      "eval_scitail-pairs-pos_loss": 0.3979075253009796,
+      "eval_scitail-pairs-pos_runtime": 2.3649,
+      "eval_scitail-pairs-pos_samples_per_second": 67.656,
+      "eval_scitail-pairs-pos_steps_per_second": 2.114,
+      "step": 1292
+    },
+    {
+      "epoch": 0.2500967866821525,
+      "eval_xsum-pairs_loss": 0.313429057598114,
+      "eval_xsum-pairs_runtime": 1.4107,
+      "eval_xsum-pairs_samples_per_second": 113.419,
+      "eval_xsum-pairs_steps_per_second": 3.544,
+      "step": 1292
+    },
+    {
+      "epoch": 0.2500967866821525,
+      "eval_compression-pairs_loss": 0.08316509425640106,
+      "eval_compression-pairs_runtime": 0.3958,
+      "eval_compression-pairs_samples_per_second": 404.289,
+      "eval_compression-pairs_steps_per_second": 12.634,
+      "step": 1292
+    },
+    {
+      "epoch": 0.2500967866821525,
+      "eval_sciq_pairs_loss": 0.2692818343639374,
+      "eval_sciq_pairs_runtime": 7.8991,
+      "eval_sciq_pairs_samples_per_second": 20.255,
+      "eval_sciq_pairs_steps_per_second": 0.633,
+      "step": 1292
+    },
+    {
+      "epoch": 0.2500967866821525,
+      "eval_qasc_pairs_loss": 0.19870159029960632,
+      "eval_qasc_pairs_runtime": 1.4336,
+      "eval_qasc_pairs_samples_per_second": 111.608,
+      "eval_qasc_pairs_steps_per_second": 3.488,
+      "step": 1292
+    },
+    {
+      "epoch": 0.2500967866821525,
+      "eval_qasc_facts_sym_loss": 0.16445104777812958,
+      "eval_qasc_facts_sym_runtime": 0.3196,
+      "eval_qasc_facts_sym_samples_per_second": 500.598,
+      "eval_qasc_facts_sym_steps_per_second": 15.644,
+      "step": 1292
+    },
+    {
+      "epoch": 0.2500967866821525,
+      "eval_openbookqa_pairs_loss": 1.7182375192642212,
+      "eval_openbookqa_pairs_runtime": 1.2252,
+      "eval_openbookqa_pairs_samples_per_second": 130.592,
+      "eval_openbookqa_pairs_steps_per_second": 4.081,
+      "step": 1292
+    },
+    {
+      "epoch": 0.2500967866821525,
+      "eval_msmarco_pairs_loss": 0.4961338937282562,
+      "eval_msmarco_pairs_runtime": 3.1144,
+      "eval_msmarco_pairs_samples_per_second": 51.374,
+      "eval_msmarco_pairs_steps_per_second": 1.605,
+      "step": 1292
+    },
+    {
+      "epoch": 0.2500967866821525,
+      "eval_nq_pairs_loss": 0.4005078673362732,
+      "eval_nq_pairs_runtime": 7.7074,
+      "eval_nq_pairs_samples_per_second": 20.759,
+      "eval_nq_pairs_steps_per_second": 0.649,
+      "step": 1292
+    },
+    {
+      "epoch": 0.2500967866821525,
+      "eval_trivia_pairs_loss": 0.654505729675293,
+      "eval_trivia_pairs_runtime": 10.1383,
+      "eval_trivia_pairs_samples_per_second": 15.782,
+      "eval_trivia_pairs_steps_per_second": 0.493,
+      "step": 1292
+    },
+    {
+      "epoch": 0.2500967866821525,
+      "eval_quora_pairs_loss": 0.22833283245563507,
+      "eval_quora_pairs_runtime": 3.894,
+      "eval_quora_pairs_samples_per_second": 173.342,
+      "eval_quora_pairs_steps_per_second": 5.65,
+      "step": 1292
+    },
+    {
+      "epoch": 0.2500967866821525,
+      "eval_gooaq_pairs_loss": 0.4996432662010193,
+      "eval_gooaq_pairs_runtime": 2.2697,
+      "eval_gooaq_pairs_samples_per_second": 70.494,
+      "eval_gooaq_pairs_steps_per_second": 2.203,
+      "step": 1292
+    },
+    {
+      "epoch": 0.2500967866821525,
+      "eval_mrpc_pairs_loss": 0.04613902047276497,
+      "eval_mrpc_pairs_runtime": 0.3743,
+      "eval_mrpc_pairs_samples_per_second": 427.427,
+      "eval_mrpc_pairs_steps_per_second": 13.357,
+      "step": 1292
+    },
+    {
+      "epoch": 0.2516453735965931,
+      "grad_norm": 12.50217342376709,
+      "learning_rate": 7.53774680603949e-06,
+      "loss": 0.6236,
+      "step": 1300
+    },
+    {
+      "epoch": 0.26422764227642276,
+      "grad_norm": 1.8397300243377686,
+      "learning_rate": 7.91521486643438e-06,
+      "loss": 0.4947,
+      "step": 1365
+    },
+    {
+      "epoch": 0.2768099109562524,
+      "grad_norm": 5.885033130645752,
+      "learning_rate": 8.292682926829268e-06,
+      "loss": 0.5595,
+      "step": 1430
+    },
+    {
+      "epoch": 0.28939217963608205,
+      "grad_norm": 1.7783002853393555,
+      "learning_rate": 8.670150987224158e-06,
+      "loss": 0.641,
+      "step": 1495
+    },
+    {
+      "epoch": 0.30197444831591175,
+      "grad_norm": 5.305712699890137,
+      "learning_rate": 9.047619047619047e-06,
+      "loss": 0.5188,
+      "step": 1560
+    },
+    {
+      "epoch": 0.3145567169957414,
+      "grad_norm": 8.24319839477539,
+      "learning_rate": 9.425087108013936e-06,
+      "loss": 0.4927,
+      "step": 1625
+    },
+    {
+      "epoch": 0.32713898567557104,
+      "grad_norm": 11.07426929473877,
+      "learning_rate": 9.802555168408827e-06,
+      "loss": 0.657,
+      "step": 1690
+    },
+    {
+      "epoch": 0.3397212543554007,
+      "grad_norm": 9.04263687133789,
+      "learning_rate": 1.0180023228803716e-05,
+      "loss": 0.4665,
+      "step": 1755
+    },
+    {
+      "epoch": 0.3523035230352303,
+      "grad_norm": 1.4980370998382568,
+      "learning_rate": 1.0557491289198606e-05,
+      "loss": 0.4645,
+      "step": 1820
+    },
+    {
+      "epoch": 0.36488579171506,
+      "grad_norm": 16.511180877685547,
+      "learning_rate": 1.0934959349593495e-05,
+      "loss": 0.5887,
+      "step": 1885
+    },
+    {
+      "epoch": 0.37746806039488967,
+      "grad_norm": 5.706000804901123,
+      "learning_rate": 1.1312427409988386e-05,
+      "loss": 0.5308,
+      "step": 1950
+    },
+    {
+      "epoch": 0.3900503290747193,
+      "grad_norm": 1.0923340320587158,
+      "learning_rate": 1.1689895470383277e-05,
+      "loss": 0.536,
+      "step": 2015
+    },
+    {
+      "epoch": 0.40263259775454896,
+      "grad_norm": 6.521665573120117,
+      "learning_rate": 1.2067363530778166e-05,
+      "loss": 0.4841,
+      "step": 2080
+    },
+    {
+      "epoch": 0.4152148664343786,
+      "grad_norm": 7.254842758178711,
+      "learning_rate": 1.2444831591173055e-05,
+      "loss": 0.6499,
+      "step": 2145
+    },
+    {
+      "epoch": 0.4277971351142083,
+      "grad_norm": 14.938628196716309,
+      "learning_rate": 1.2822299651567945e-05,
+      "loss": 0.5982,
+      "step": 2210
+    },
+    {
+      "epoch": 0.44037940379403795,
+      "grad_norm": 10.192171096801758,
+      "learning_rate": 1.3199767711962834e-05,
+      "loss": 0.5281,
+      "step": 2275
+    },
+    {
+      "epoch": 0.4529616724738676,
+      "grad_norm": 8.02379035949707,
+      "learning_rate": 1.3577235772357725e-05,
+      "loss": 0.6657,
+      "step": 2340
+    },
+    {
+      "epoch": 0.46554394115369724,
+      "grad_norm": 2.842752695083618,
+      "learning_rate": 1.3954703832752614e-05,
+      "loss": 0.5746,
+      "step": 2405
+    },
+    {
+      "epoch": 0.4781262098335269,
+      "grad_norm": 0.4373825788497925,
+      "learning_rate": 1.432636469221835e-05,
+      "loss": 0.5853,
+      "step": 2470
+    },
+    {
+      "epoch": 0.4907084785133566,
+      "grad_norm": 16.475561141967773,
+      "learning_rate": 1.4703832752613242e-05,
+      "loss": 0.5828,
+      "step": 2535
+    },
+    {
+      "epoch": 0.500193573364305,
+      "eval_StS-test_pearson_cosine": 0.8850112014664042,
+      "eval_StS-test_pearson_dot": 0.8059948281019231,
+      "eval_StS-test_pearson_euclidean": 0.8736480607586303,
+      "eval_StS-test_pearson_manhattan": 0.8746434402751778,
+      "eval_StS-test_pearson_max": 0.8850112014664042,
+      "eval_StS-test_spearman_cosine": 0.8964908683344393,
+      "eval_StS-test_spearman_dot": 0.8112940990649002,
+      "eval_StS-test_spearman_euclidean": 0.8728524906348676,
+      "eval_StS-test_spearman_manhattan": 0.8747700566942432,
+      "eval_StS-test_spearman_max": 0.8964908683344393,
+      "eval_Vitaminc-test_cosine_accuracy": 0.5842105263157895,
+      "eval_Vitaminc-test_cosine_accuracy_threshold": 0.6876633763313293,
+      "eval_Vitaminc-test_cosine_ap": 0.5620310838989798,
+      "eval_Vitaminc-test_cosine_f1": 0.6755725190839694,
+      "eval_Vitaminc-test_cosine_f1_threshold": 0.5227451324462891,
+      "eval_Vitaminc-test_cosine_precision": 0.5221238938053098,
+      "eval_Vitaminc-test_cosine_recall": 0.9567567567567568,
+      "eval_Vitaminc-test_dot_accuracy": 0.5552631578947368,
+      "eval_Vitaminc-test_dot_accuracy_threshold": 19.32275390625,
+      "eval_Vitaminc-test_dot_ap": 0.5459852333889619,
+      "eval_Vitaminc-test_dot_f1": 0.6715596330275228,
+      "eval_Vitaminc-test_dot_f1_threshold": 12.946126937866211,
+      "eval_Vitaminc-test_dot_precision": 0.5083333333333333,
+      "eval_Vitaminc-test_dot_recall": 0.9891891891891892,
+      "eval_Vitaminc-test_euclidean_accuracy": 0.5736842105263158,
+      "eval_Vitaminc-test_euclidean_accuracy_threshold": 3.9492130279541016,
+      "eval_Vitaminc-test_euclidean_ap": 0.5569069466677845,
+      "eval_Vitaminc-test_euclidean_f1": 0.6654343807763402,
+      "eval_Vitaminc-test_euclidean_f1_threshold": 6.641269207000732,
+      "eval_Vitaminc-test_euclidean_precision": 0.5056179775280899,
+      "eval_Vitaminc-test_euclidean_recall": 0.972972972972973,
+      "eval_Vitaminc-test_manhattan_accuracy": 0.5789473684210527,
+      "eval_Vitaminc-test_manhattan_accuracy_threshold": 80.14714050292969,
+      "eval_Vitaminc-test_manhattan_ap": 0.556199495550533,
+      "eval_Vitaminc-test_manhattan_f1": 0.6691312384473197,
+      "eval_Vitaminc-test_manhattan_f1_threshold": 138.50221252441406,
+      "eval_Vitaminc-test_manhattan_precision": 0.5084269662921348,
+      "eval_Vitaminc-test_manhattan_recall": 0.9783783783783784,
+      "eval_Vitaminc-test_max_accuracy": 0.5842105263157895,
+      "eval_Vitaminc-test_max_accuracy_threshold": 80.14714050292969,
+      "eval_Vitaminc-test_max_ap": 0.5620310838989798,
+      "eval_Vitaminc-test_max_f1": 0.6755725190839694,
+      "eval_Vitaminc-test_max_f1_threshold": 138.50221252441406,
+      "eval_Vitaminc-test_max_precision": 0.5221238938053098,
+      "eval_Vitaminc-test_max_recall": 0.9891891891891892,
+      "eval_mrpc-test_cosine_accuracy": 0.7368421052631579,
+      "eval_mrpc-test_cosine_accuracy_threshold": 0.7595815658569336,
+      "eval_mrpc-test_cosine_ap": 0.8489522026016108,
+      "eval_mrpc-test_cosine_f1": 0.8235294117647058,
+      "eval_mrpc-test_cosine_f1_threshold": 0.6731016635894775,
+      "eval_mrpc-test_cosine_precision": 0.7212121212121212,
+      "eval_mrpc-test_cosine_recall": 0.9596774193548387,
+      "eval_mrpc-test_dot_accuracy": 0.6894736842105263,
+      "eval_mrpc-test_dot_accuracy_threshold": 15.645252227783203,
+      "eval_mrpc-test_dot_ap": 0.7828717059557433,
+      "eval_mrpc-test_dot_f1": 0.8071895424836601,
+      "eval_mrpc-test_dot_f1_threshold": 11.141586303710938,
+      "eval_mrpc-test_dot_precision": 0.6785714285714286,
+      "eval_mrpc-test_dot_recall": 0.9959677419354839,
+      "eval_mrpc-test_euclidean_accuracy": 0.7368421052631579,
+      "eval_mrpc-test_euclidean_accuracy_threshold": 3.7453079223632812,
+      "eval_mrpc-test_euclidean_ap": 0.8190549606804299,
+      "eval_mrpc-test_euclidean_f1": 0.8172043010752688,
+      "eval_mrpc-test_euclidean_f1_threshold": 3.905801773071289,
+      "eval_mrpc-test_euclidean_precision": 0.7354838709677419,
+      "eval_mrpc-test_euclidean_recall": 0.9193548387096774,
+      "eval_mrpc-test_manhattan_accuracy": 0.7315789473684211,
+      "eval_mrpc-test_manhattan_accuracy_threshold": 74.24961853027344,
+      "eval_mrpc-test_manhattan_ap": 0.815561648496177,
+      "eval_mrpc-test_manhattan_f1": 0.81508078994614,
+      "eval_mrpc-test_manhattan_f1_threshold": 78.01387023925781,
+      "eval_mrpc-test_manhattan_precision": 0.7346278317152104,
+      "eval_mrpc-test_manhattan_recall": 0.9153225806451613,
+      "eval_mrpc-test_max_accuracy": 0.7368421052631579,
+      "eval_mrpc-test_max_accuracy_threshold": 74.24961853027344,
+      "eval_mrpc-test_max_ap": 0.8489522026016108,
+      "eval_mrpc-test_max_f1": 0.8235294117647058,
+      "eval_mrpc-test_max_f1_threshold": 78.01387023925781,
+      "eval_mrpc-test_max_precision": 0.7354838709677419,
+      "eval_mrpc-test_max_recall": 0.9959677419354839,
+      "eval_nli-pairs_loss": 0.805642306804657,
+      "eval_nli-pairs_runtime": 3.1328,
+      "eval_nli-pairs_samples_per_second": 51.072,
+      "eval_nli-pairs_steps_per_second": 1.596,
+      "eval_sequential_score": 0.5620310838989798,
+      "step": 2584
+    },
+    {
+      "epoch": 0.500193573364305,
+      "eval_vitaminc-pairs_loss": 5.803024768829346,
+      "eval_vitaminc-pairs_runtime": 1.6236,
+      "eval_vitaminc-pairs_samples_per_second": 81.915,
+      "eval_vitaminc-pairs_steps_per_second": 3.08,
+      "step": 2584
+    },
+    {
+      "epoch": 0.500193573364305,
+      "eval_qnli-contrastive_loss": 0.1560516655445099,
+      "eval_qnli-contrastive_runtime": 0.5455,
+      "eval_qnli-contrastive_samples_per_second": 293.295,
+      "eval_qnli-contrastive_steps_per_second": 9.165,
+      "step": 2584
+    },
+    {
+      "epoch": 0.500193573364305,
+      "eval_scitail-pairs-qa_loss": 0.07059809565544128,
+      "eval_scitail-pairs-qa_runtime": 1.3304,
+      "eval_scitail-pairs-qa_samples_per_second": 120.266,
+      "eval_scitail-pairs-qa_steps_per_second": 3.758,
+      "step": 2584
+    },
+    {
+      "epoch": 0.500193573364305,
+      "eval_scitail-pairs-pos_loss": 0.40096405148506165,
+      "eval_scitail-pairs-pos_runtime": 2.6358,
+      "eval_scitail-pairs-pos_samples_per_second": 60.703,
+      "eval_scitail-pairs-pos_steps_per_second": 1.897,
+      "step": 2584
+    },
+    {
+      "epoch": 0.500193573364305,
+      "eval_xsum-pairs_loss": 0.3021017014980316,
+      "eval_xsum-pairs_runtime": 1.474,
+      "eval_xsum-pairs_samples_per_second": 108.547,
+      "eval_xsum-pairs_steps_per_second": 3.392,
+      "step": 2584
+    },
+    {
+      "epoch": 0.500193573364305,
+      "eval_compression-pairs_loss": 0.08080567419528961,
+      "eval_compression-pairs_runtime": 0.4395,
+      "eval_compression-pairs_samples_per_second": 364.011,
+      "eval_compression-pairs_steps_per_second": 11.375,
+      "step": 2584
+    },
+    {
+      "epoch": 0.500193573364305,
+      "eval_sciq_pairs_loss": 0.2838394045829773,
+      "eval_sciq_pairs_runtime": 8.1466,
+      "eval_sciq_pairs_samples_per_second": 19.64,
+      "eval_sciq_pairs_steps_per_second": 0.614,
+      "step": 2584
+    },
+    {
+      "epoch": 0.500193573364305,
+      "eval_qasc_pairs_loss": 0.19900637865066528,
+      "eval_qasc_pairs_runtime": 1.4833,
+      "eval_qasc_pairs_samples_per_second": 107.864,
+      "eval_qasc_pairs_steps_per_second": 3.371,
+      "step": 2584
+    },
+    {
+      "epoch": 0.500193573364305,
+      "eval_qasc_facts_sym_loss": 0.15245158970355988,
+      "eval_qasc_facts_sym_runtime": 0.3292,
+      "eval_qasc_facts_sym_samples_per_second": 485.978,
+      "eval_qasc_facts_sym_steps_per_second": 15.187,
+      "step": 2584
+    },
+    {
+      "epoch": 0.500193573364305,
+      "eval_openbookqa_pairs_loss": 1.7069475650787354,
+      "eval_openbookqa_pairs_runtime": 1.4335,
+      "eval_openbookqa_pairs_samples_per_second": 111.612,
+      "eval_openbookqa_pairs_steps_per_second": 3.488,
+      "step": 2584
+    },
+    {
+      "epoch": 0.500193573364305,
+      "eval_msmarco_pairs_loss": 0.5149500966072083,
+      "eval_msmarco_pairs_runtime": 3.2287,
+      "eval_msmarco_pairs_samples_per_second": 49.555,
+      "eval_msmarco_pairs_steps_per_second": 1.549,
+      "step": 2584
+    },
+    {
+      "epoch": 0.500193573364305,
+      "eval_nq_pairs_loss": 0.4162893295288086,
+      "eval_nq_pairs_runtime": 7.8277,
+      "eval_nq_pairs_samples_per_second": 20.44,
+      "eval_nq_pairs_steps_per_second": 0.639,
+      "step": 2584
+    },
+    {
+      "epoch": 0.500193573364305,
+      "eval_trivia_pairs_loss": 0.623663067817688,
+      "eval_trivia_pairs_runtime": 10.277,
+      "eval_trivia_pairs_samples_per_second": 15.569,
+      "eval_trivia_pairs_steps_per_second": 0.487,
+      "step": 2584
+    },
+    {
+      "epoch": 0.500193573364305,
+      "eval_quora_pairs_loss": 0.22309938073158264,
+      "eval_quora_pairs_runtime": 4.3318,
+      "eval_quora_pairs_samples_per_second": 155.824,
+      "eval_quora_pairs_steps_per_second": 5.079,
+      "step": 2584
+    },
+    {
+      "epoch": 0.500193573364305,
+      "eval_gooaq_pairs_loss": 0.49019113183021545,
+      "eval_gooaq_pairs_runtime": 2.2716,
+      "eval_gooaq_pairs_samples_per_second": 70.434,
+      "eval_gooaq_pairs_steps_per_second": 2.201,
+      "step": 2584
+    },
+    {
+      "epoch": 0.500193573364305,
+      "eval_mrpc_pairs_loss": 0.04494442045688629,
+      "eval_mrpc_pairs_runtime": 0.3756,
+      "eval_mrpc_pairs_samples_per_second": 425.958,
+      "eval_mrpc_pairs_steps_per_second": 13.311,
+      "step": 2584
+    },
+    {
+      "epoch": 0.5032907471931862,
+      "grad_norm": 0.5926259756088257,
+      "learning_rate": 1.5081300813008131e-05,
+      "loss": 0.4889,
+      "step": 2600
+    },
+    {
+      "epoch": 0.5158730158730159,
+      "grad_norm": 12.89693832397461,
+      "learning_rate": 1.545876887340302e-05,
+      "loss": 0.517,
+      "step": 2665
+    },
+    {
+      "epoch": 0.5284552845528455,
+      "grad_norm": 0.3751634359359741,
+      "learning_rate": 1.583623693379791e-05,
+      "loss": 0.5479,
+      "step": 2730
+    },
+    {
+      "epoch": 0.5410375532326752,
+      "grad_norm": 10.321626663208008,
+      "learning_rate": 1.62137049941928e-05,
+      "loss": 0.6409,
+      "step": 2795
+    },
+    {
+      "epoch": 0.5536198219125048,
+      "grad_norm": 0.40425410866737366,
+      "learning_rate": 1.659117305458769e-05,
+      "loss": 0.6137,
+      "step": 2860
+    },
+    {
+      "epoch": 0.5662020905923345,
+      "grad_norm": 0.26469847559928894,
+      "learning_rate": 1.696864111498258e-05,
+      "loss": 0.5947,
+      "step": 2925
+    },
+    {
+      "epoch": 0.5787843592721641,
+      "grad_norm": 0.3376736640930176,
+      "learning_rate": 1.7346109175377468e-05,
+      "loss": 0.514,
+      "step": 2990
+    },
+    {
+      "epoch": 0.5913666279519938,
+      "grad_norm": 11.042977333068848,
+      "learning_rate": 1.7723577235772357e-05,
+      "loss": 0.5256,
+      "step": 3055
+    },
+    {
+      "epoch": 0.6039488966318235,
+      "grad_norm": 28.415281295776367,
+      "learning_rate": 1.810104529616725e-05,
+      "loss": 0.5574,
+      "step": 3120
+    },
+    {
+      "epoch": 0.6165311653116531,
+      "grad_norm": 1.138208270072937,
+      "learning_rate": 1.8478513356562136e-05,
+      "loss": 0.5669,
+      "step": 3185
+    },
+    {
+      "epoch": 0.6291134339914828,
+      "grad_norm": 0.38507771492004395,
+      "learning_rate": 1.885598141695703e-05,
+      "loss": 0.5185,
+      "step": 3250
+    },
+    {
+      "epoch": 0.6416957026713124,
+      "grad_norm": 0.4718106687068939,
+      "learning_rate": 1.9233449477351914e-05,
+      "loss": 0.4875,
+      "step": 3315
+    },
+    {
+      "epoch": 0.6542779713511421,
+      "grad_norm": 0.3544551432132721,
+      "learning_rate": 1.9610917537746807e-05,
+      "loss": 0.5773,
+      "step": 3380
+    },
+    {
+      "epoch": 0.6668602400309718,
+      "grad_norm": 15.193903923034668,
+      "learning_rate": 1.9988385598141696e-05,
+      "loss": 0.6102,
+      "step": 3445
+    },
+    {
+      "epoch": 0.6794425087108014,
+      "grad_norm": 14.930959701538086,
+      "learning_rate": 2.0365853658536586e-05,
+      "loss": 0.5882,
+      "step": 3510
+    },
+    {
+      "epoch": 0.6920247773906311,
+      "grad_norm": 0.7249612212181091,
+      "learning_rate": 2.0743321718931475e-05,
+      "loss": 0.5888,
+      "step": 3575
+    },
+    {
+      "epoch": 0.7046070460704607,
+      "grad_norm": 0.21552956104278564,
+      "learning_rate": 2.1120789779326364e-05,
+      "loss": 0.6088,
+      "step": 3640
+    },
+    {
+      "epoch": 0.7171893147502904,
+      "grad_norm": 3.176356554031372,
+      "learning_rate": 2.1498257839721253e-05,
+      "loss": 0.688,
+      "step": 3705
+    },
+    {
+      "epoch": 0.72977158343012,
+      "grad_norm": 4.0542378425598145,
+      "learning_rate": 2.1875725900116146e-05,
+      "loss": 0.6541,
+      "step": 3770
+    },
+    {
+      "epoch": 0.7423538521099496,
+      "grad_norm": 0.5486993193626404,
+      "learning_rate": 2.2253193960511032e-05,
+      "loss": 0.4526,
+      "step": 3835
+    },
+    {
+      "epoch": 0.7502903600464577,
+      "eval_StS-test_pearson_cosine": 0.8754429584344908,
+      "eval_StS-test_pearson_dot": 0.7812438977199179,
+      "eval_StS-test_pearson_euclidean": 0.863532080020405,
+      "eval_StS-test_pearson_manhattan": 0.8642206125042041,
+      "eval_StS-test_pearson_max": 0.8754429584344908,
+      "eval_StS-test_spearman_cosine": 0.8899322603927042,
+      "eval_StS-test_spearman_dot": 0.7897822634826679,
+      "eval_StS-test_spearman_euclidean": 0.8661349339052001,
+      "eval_StS-test_spearman_manhattan": 0.8677664578276145,
+      "eval_StS-test_spearman_max": 0.8899322603927042,
+      "eval_Vitaminc-test_cosine_accuracy": 0.5815789473684211,
+      "eval_Vitaminc-test_cosine_accuracy_threshold": 0.7150823473930359,
+      "eval_Vitaminc-test_cosine_ap": 0.5695623413042502,
+      "eval_Vitaminc-test_cosine_f1": 0.6690909090909091,
+      "eval_Vitaminc-test_cosine_f1_threshold": 0.4144631028175354,
+      "eval_Vitaminc-test_cosine_precision": 0.5041095890410959,
+      "eval_Vitaminc-test_cosine_recall": 0.9945945945945946,
+      "eval_Vitaminc-test_dot_accuracy": 0.5631578947368421,
+      "eval_Vitaminc-test_dot_accuracy_threshold": 25.451148986816406,
+      "eval_Vitaminc-test_dot_ap": 0.5500277455332443,
+      "eval_Vitaminc-test_dot_f1": 0.6691449814126395,
+      "eval_Vitaminc-test_dot_f1_threshold": 14.098217964172363,
+      "eval_Vitaminc-test_dot_precision": 0.509915014164306,
+      "eval_Vitaminc-test_dot_recall": 0.972972972972973,
+      "eval_Vitaminc-test_euclidean_accuracy": 0.5736842105263158,
+      "eval_Vitaminc-test_euclidean_accuracy_threshold": 3.945739269256592,
+      "eval_Vitaminc-test_euclidean_ap": 0.5590599048028368,
+      "eval_Vitaminc-test_euclidean_f1": 0.6654478976234005,
+      "eval_Vitaminc-test_euclidean_f1_threshold": 6.931303024291992,
+      "eval_Vitaminc-test_euclidean_precision": 0.5027624309392266,
+      "eval_Vitaminc-test_euclidean_recall": 0.9837837837837838,
+      "eval_Vitaminc-test_manhattan_accuracy": 0.5736842105263158,
+      "eval_Vitaminc-test_manhattan_accuracy_threshold": 84.99573516845703,
+      "eval_Vitaminc-test_manhattan_ap": 0.559747734383099,
+      "eval_Vitaminc-test_manhattan_f1": 0.6679035250463822,
+      "eval_Vitaminc-test_manhattan_f1_threshold": 139.30535888671875,
+      "eval_Vitaminc-test_manhattan_precision": 0.5084745762711864,
+      "eval_Vitaminc-test_manhattan_recall": 0.972972972972973,
+      "eval_Vitaminc-test_max_accuracy": 0.5815789473684211,
+      "eval_Vitaminc-test_max_accuracy_threshold": 84.99573516845703,
+      "eval_Vitaminc-test_max_ap": 0.5695623413042502,
+      "eval_Vitaminc-test_max_f1": 0.6691449814126395,
+      "eval_Vitaminc-test_max_f1_threshold": 139.30535888671875,
+      "eval_Vitaminc-test_max_precision": 0.509915014164306,
+      "eval_Vitaminc-test_max_recall": 0.9945945945945946,
+      "eval_mrpc-test_cosine_accuracy": 0.7447368421052631,
+      "eval_mrpc-test_cosine_accuracy_threshold": 0.7351576089859009,
+      "eval_mrpc-test_cosine_ap": 0.8492225227540071,
+      "eval_mrpc-test_cosine_f1": 0.8274647887323944,
+      "eval_mrpc-test_cosine_f1_threshold": 0.7097899913787842,
+      "eval_mrpc-test_cosine_precision": 0.734375,
+      "eval_mrpc-test_cosine_recall": 0.9475806451612904,
+      "eval_mrpc-test_dot_accuracy": 0.6947368421052632,
+      "eval_mrpc-test_dot_accuracy_threshold": 13.770221710205078,
+      "eval_mrpc-test_dot_ap": 0.7664430695917603,
+      "eval_mrpc-test_dot_f1": 0.8060200668896322,
+      "eval_mrpc-test_dot_f1_threshold": 13.664058685302734,
+      "eval_mrpc-test_dot_precision": 0.6885714285714286,
+      "eval_mrpc-test_dot_recall": 0.9717741935483871,
+      "eval_mrpc-test_euclidean_accuracy": 0.7421052631578947,
+      "eval_mrpc-test_euclidean_accuracy_threshold": 3.573061466217041,
+      "eval_mrpc-test_euclidean_ap": 0.8229315066081502,
+      "eval_mrpc-test_euclidean_f1": 0.8216216216216217,
+      "eval_mrpc-test_euclidean_f1_threshold": 3.7641119956970215,
+      "eval_mrpc-test_euclidean_precision": 0.742671009771987,
+      "eval_mrpc-test_euclidean_recall": 0.9193548387096774,
+      "eval_mrpc-test_manhattan_accuracy": 0.7394736842105263,
+      "eval_mrpc-test_manhattan_accuracy_threshold": 75.05320739746094,
+      "eval_mrpc-test_manhattan_ap": 0.8195074272882318,
+      "eval_mrpc-test_manhattan_f1": 0.8228980322003578,
+      "eval_mrpc-test_manhattan_f1_threshold": 76.21920013427734,
+      "eval_mrpc-test_manhattan_precision": 0.7395498392282959,
+      "eval_mrpc-test_manhattan_recall": 0.9274193548387096,
+      "eval_mrpc-test_max_accuracy": 0.7447368421052631,
+      "eval_mrpc-test_max_accuracy_threshold": 75.05320739746094,
+      "eval_mrpc-test_max_ap": 0.8492225227540071,
+      "eval_mrpc-test_max_f1": 0.8274647887323944,
+      "eval_mrpc-test_max_f1_threshold": 76.21920013427734,
+      "eval_mrpc-test_max_precision": 0.742671009771987,
+      "eval_mrpc-test_max_recall": 0.9717741935483871,
+      "eval_nli-pairs_loss": 0.8814637064933777,
+      "eval_nli-pairs_runtime": 2.9865,
+      "eval_nli-pairs_samples_per_second": 53.575,
+      "eval_nli-pairs_steps_per_second": 1.674,
+      "eval_sequential_score": 0.5695623413042502,
+      "step": 3876
+    },
+    {
+      "epoch": 0.7502903600464577,
+      "eval_vitaminc-pairs_loss": 5.706007957458496,
+      "eval_vitaminc-pairs_runtime": 1.5554,
+      "eval_vitaminc-pairs_samples_per_second": 85.507,
+      "eval_vitaminc-pairs_steps_per_second": 3.215,
+      "step": 3876
+    },
+    {
+      "epoch": 0.7502903600464577,
+      "eval_qnli-contrastive_loss": 0.14242754876613617,
+      "eval_qnli-contrastive_runtime": 0.5128,
+      "eval_qnli-contrastive_samples_per_second": 311.993,
+      "eval_qnli-contrastive_steps_per_second": 9.75,
+      "step": 3876
+    },
+    {
+      "epoch": 0.7502903600464577,
+      "eval_scitail-pairs-qa_loss": 0.0674210786819458,
+      "eval_scitail-pairs-qa_runtime": 1.2622,
+      "eval_scitail-pairs-qa_samples_per_second": 126.759,
+      "eval_scitail-pairs-qa_steps_per_second": 3.961,
+      "step": 3876
+    },
+    {
+      "epoch": 0.7502903600464577,
+      "eval_scitail-pairs-pos_loss": 0.36303356289863586,
+      "eval_scitail-pairs-pos_runtime": 2.3597,
+      "eval_scitail-pairs-pos_samples_per_second": 67.806,
+      "eval_scitail-pairs-pos_steps_per_second": 2.119,
+      "step": 3876
+    },
+    {
+      "epoch": 0.7502903600464577,
+      "eval_xsum-pairs_loss": 0.32611021399497986,
+      "eval_xsum-pairs_runtime": 1.4024,
+      "eval_xsum-pairs_samples_per_second": 114.091,
+      "eval_xsum-pairs_steps_per_second": 3.565,
+      "step": 3876
+    },
+    {
+      "epoch": 0.7502903600464577,
+      "eval_compression-pairs_loss": 0.08624111860990524,
+      "eval_compression-pairs_runtime": 0.3948,
+      "eval_compression-pairs_samples_per_second": 405.263,
+      "eval_compression-pairs_steps_per_second": 12.664,
+      "step": 3876
+    },
+    {
+      "epoch": 0.7502903600464577,
+      "eval_sciq_pairs_loss": 0.2929747402667999,
+      "eval_sciq_pairs_runtime": 7.9352,
+      "eval_sciq_pairs_samples_per_second": 20.163,
+      "eval_sciq_pairs_steps_per_second": 0.63,
+      "step": 3876
+    },
+    {
+      "epoch": 0.7502903600464577,
+      "eval_qasc_pairs_loss": 0.20694346725940704,
+      "eval_qasc_pairs_runtime": 1.4441,
+      "eval_qasc_pairs_samples_per_second": 110.798,
+      "eval_qasc_pairs_steps_per_second": 3.462,
+      "step": 3876
+    },
+    {
+      "epoch": 0.7502903600464577,
+      "eval_qasc_facts_sym_loss": 0.14476129412651062,
+      "eval_qasc_facts_sym_runtime": 0.3295,
+      "eval_qasc_facts_sym_samples_per_second": 485.621,
+      "eval_qasc_facts_sym_steps_per_second": 15.176,
+      "step": 3876
+    },
+    {
+      "epoch": 0.7502903600464577,
+      "eval_openbookqa_pairs_loss": 1.7119003534317017,
+      "eval_openbookqa_pairs_runtime": 1.2433,
+      "eval_openbookqa_pairs_samples_per_second": 128.687,
+      "eval_openbookqa_pairs_steps_per_second": 4.021,
+      "step": 3876
+    },
+    {
+      "epoch": 0.7502903600464577,
+      "eval_msmarco_pairs_loss": 0.4974798262119293,
+      "eval_msmarco_pairs_runtime": 3.1059,
+      "eval_msmarco_pairs_samples_per_second": 51.516,
+      "eval_msmarco_pairs_steps_per_second": 1.61,
+      "step": 3876
+    },
+    {
+      "epoch": 0.7502903600464577,
+      "eval_nq_pairs_loss": 0.46937060356140137,
+      "eval_nq_pairs_runtime": 7.6944,
+      "eval_nq_pairs_samples_per_second": 20.794,
+      "eval_nq_pairs_steps_per_second": 0.65,
+      "step": 3876
+    },
+    {
+      "epoch": 0.7502903600464577,
+      "eval_trivia_pairs_loss": 0.6990238428115845,
+      "eval_trivia_pairs_runtime": 10.1876,
+      "eval_trivia_pairs_samples_per_second": 15.705,
+      "eval_trivia_pairs_steps_per_second": 0.491,
+      "step": 3876
+    },
+    {
+      "epoch": 0.7502903600464577,
+      "eval_quora_pairs_loss": 0.21933521330356598,
+      "eval_quora_pairs_runtime": 3.8953,
+      "eval_quora_pairs_samples_per_second": 173.287,
+      "eval_quora_pairs_steps_per_second": 5.648,
+      "step": 3876
+    },
+    {
+      "epoch": 0.7502903600464577,
+      "eval_gooaq_pairs_loss": 0.5000605583190918,
+      "eval_gooaq_pairs_runtime": 2.2758,
+      "eval_gooaq_pairs_samples_per_second": 70.306,
+      "eval_gooaq_pairs_steps_per_second": 2.197,
+      "step": 3876
+    },
+    {
+      "epoch": 0.7502903600464577,
+      "eval_mrpc_pairs_loss": 0.044196613132953644,
+      "eval_mrpc_pairs_runtime": 0.3742,
+      "eval_mrpc_pairs_samples_per_second": 427.588,
+      "eval_mrpc_pairs_steps_per_second": 13.362,
+      "step": 3876
+    },
+    {
+      "epoch": 0.7549361207897793,
+      "grad_norm": 1.587725281715393,
+      "learning_rate": 2.2630662020905924e-05,
+      "loss": 0.4466,
+      "step": 3900
+    },
+    {
+      "epoch": 0.7675183894696089,
+      "grad_norm": 9.404407501220703,
+      "learning_rate": 2.3008130081300814e-05,
+      "loss": 0.6293,
+      "step": 3965
+    },
+    {
+      "epoch": 0.7801006581494386,
+      "grad_norm": 3.672903537750244,
+      "learning_rate": 2.3385598141695703e-05,
+      "loss": 0.5591,
+      "step": 4030
+    },
+    {
+      "epoch": 0.7926829268292683,
+      "grad_norm": 1.3529752492904663,
+      "learning_rate": 2.3763066202090596e-05,
+      "loss": 0.5702,
+      "step": 4095
+    },
+    {
+      "epoch": 0.8052651955090979,
+      "grad_norm": 21.519527435302734,
+      "learning_rate": 2.414053426248548e-05,
+      "loss": 0.4645,
+      "step": 4160
+    },
+    {
+      "epoch": 0.8178474641889276,
+      "grad_norm": 5.5179362297058105,
+      "learning_rate": 2.4518002322880374e-05,
+      "loss": 0.5175,
+      "step": 4225
+    },
+    {
+      "epoch": 0.8304297328687572,
+      "grad_norm": 10.127123832702637,
+      "learning_rate": 2.489547038327526e-05,
+      "loss": 0.5188,
+      "step": 4290
+    },
+    {
+      "epoch": 0.8430120015485869,
+      "grad_norm": 5.518251895904541,
+      "learning_rate": 2.5272938443670153e-05,
+      "loss": 0.4908,
+      "step": 4355
+    },
+    {
+      "epoch": 0.8555942702284166,
+      "grad_norm": 0.7371014356613159,
+      "learning_rate": 2.5650406504065042e-05,
+      "loss": 0.5075,
+      "step": 4420
+    },
+    {
+      "epoch": 0.8681765389082462,
+      "grad_norm": 22.15941619873047,
+      "learning_rate": 2.602206736353078e-05,
+      "loss": 0.5625,
+      "step": 4485
+    },
+    {
+      "epoch": 0.8807588075880759,
+      "grad_norm": 1.0873440504074097,
+      "learning_rate": 2.6399535423925668e-05,
+      "loss": 0.4952,
+      "step": 4550
+    },
+    {
+      "epoch": 0.8933410762679055,
+      "grad_norm": 19.877098083496094,
+      "learning_rate": 2.6777003484320557e-05,
+      "loss": 0.5818,
+      "step": 4615
+    },
+    {
+      "epoch": 0.9059233449477352,
+      "grad_norm": 9.559219360351562,
+      "learning_rate": 2.715447154471545e-05,
+      "loss": 0.6346,
+      "step": 4680
+    },
+    {
+      "epoch": 0.9185056136275649,
+      "grad_norm": 2.896815299987793,
+      "learning_rate": 2.7531939605110336e-05,
+      "loss": 0.4971,
+      "step": 4745
+    },
+    {
+      "epoch": 0.9310878823073945,
+      "grad_norm": 4.5045318603515625,
+      "learning_rate": 2.7909407665505228e-05,
+      "loss": 0.4834,
+      "step": 4810
+    },
+    {
+      "epoch": 0.9436701509872242,
+      "grad_norm": 1.2259025573730469,
+      "learning_rate": 2.8286875725900114e-05,
+      "loss": 0.5992,
+      "step": 4875
+    },
+    {
+      "epoch": 0.9562524196670538,
+      "grad_norm": 9.3356294631958,
+      "learning_rate": 2.8664343786295007e-05,
+      "loss": 0.5244,
+      "step": 4940
+    },
+    {
+      "epoch": 0.9688346883468835,
+      "grad_norm": 2.221039056777954,
+      "learning_rate": 2.9041811846689896e-05,
+      "loss": 0.6051,
+      "step": 5005
+    },
+    {
+      "epoch": 0.9814169570267132,
+      "grad_norm": 1.3200041055679321,
+      "learning_rate": 2.9419279907084785e-05,
+      "loss": 0.4323,
+      "step": 5070
+    },
+    {
+      "epoch": 0.9939992257065428,
+      "grad_norm": 8.689266204833984,
+      "learning_rate": 2.9796747967479674e-05,
+      "loss": 0.5347,
+      "step": 5135
+    },
+    {
+      "epoch": 1.00038714672861,
+      "eval_StS-test_pearson_cosine": 0.8763915227055069,
+      "eval_StS-test_pearson_dot": 0.8008666276678159,
+      "eval_StS-test_pearson_euclidean": 0.862640534705167,
+      "eval_StS-test_pearson_manhattan": 0.8624528259321993,
+      "eval_StS-test_pearson_max": 0.8763915227055069,
+      "eval_StS-test_spearman_cosine": 0.8905227354354782,
+      "eval_StS-test_spearman_dot": 0.8046369691471742,
+      "eval_StS-test_spearman_euclidean": 0.8643543133924523,
+      "eval_StS-test_spearman_manhattan": 0.8653407373098487,
+      "eval_StS-test_spearman_max": 0.8905227354354782,
+      "eval_Vitaminc-test_cosine_accuracy": 0.5789473684210527,
+      "eval_Vitaminc-test_cosine_accuracy_threshold": 0.8158951997756958,
+      "eval_Vitaminc-test_cosine_ap": 0.5603755036121337,
+      "eval_Vitaminc-test_cosine_f1": 0.6778398510242085,
+      "eval_Vitaminc-test_cosine_f1_threshold": 0.5370069146156311,
+      "eval_Vitaminc-test_cosine_precision": 0.5170454545454546,
+      "eval_Vitaminc-test_cosine_recall": 0.9837837837837838,
+      "eval_Vitaminc-test_dot_accuracy": 0.5578947368421052,
+      "eval_Vitaminc-test_dot_accuracy_threshold": 25.046009063720703,
+      "eval_Vitaminc-test_dot_ap": 0.5384851861555453,
+      "eval_Vitaminc-test_dot_f1": 0.6703499079189686,
+      "eval_Vitaminc-test_dot_f1_threshold": 16.19637680053711,
+      "eval_Vitaminc-test_dot_precision": 0.5083798882681564,
+      "eval_Vitaminc-test_dot_recall": 0.9837837837837838,
+      "eval_Vitaminc-test_euclidean_accuracy": 0.5789473684210527,
+      "eval_Vitaminc-test_euclidean_accuracy_threshold": 4.248380661010742,
+      "eval_Vitaminc-test_euclidean_ap": 0.5580666089311308,
+      "eval_Vitaminc-test_euclidean_f1": 0.6703499079189686,
+      "eval_Vitaminc-test_euclidean_f1_threshold": 6.6477460861206055,
+      "eval_Vitaminc-test_euclidean_precision": 0.5083798882681564,
+      "eval_Vitaminc-test_euclidean_recall": 0.9837837837837838,
+      "eval_Vitaminc-test_manhattan_accuracy": 0.5736842105263158,
+      "eval_Vitaminc-test_manhattan_accuracy_threshold": 89.6431884765625,
+      "eval_Vitaminc-test_manhattan_ap": 0.5550489368095819,
+      "eval_Vitaminc-test_manhattan_f1": 0.6703499079189686,
+      "eval_Vitaminc-test_manhattan_f1_threshold": 138.67050170898438,
+      "eval_Vitaminc-test_manhattan_precision": 0.5083798882681564,
+      "eval_Vitaminc-test_manhattan_recall": 0.9837837837837838,
+      "eval_Vitaminc-test_max_accuracy": 0.5789473684210527,
+      "eval_Vitaminc-test_max_accuracy_threshold": 89.6431884765625,
+      "eval_Vitaminc-test_max_ap": 0.5603755036121337,
+      "eval_Vitaminc-test_max_f1": 0.6778398510242085,
+      "eval_Vitaminc-test_max_f1_threshold": 138.67050170898438,
+      "eval_Vitaminc-test_max_precision": 0.5170454545454546,
+      "eval_Vitaminc-test_max_recall": 0.9837837837837838,
+      "eval_mrpc-test_cosine_accuracy": 0.7526315789473684,
+      "eval_mrpc-test_cosine_accuracy_threshold": 0.7351824045181274,
+      "eval_mrpc-test_cosine_ap": 0.8519630000115989,
+      "eval_mrpc-test_cosine_f1": 0.8324514991181658,
+      "eval_mrpc-test_cosine_f1_threshold": 0.7194495797157288,
+      "eval_mrpc-test_cosine_precision": 0.7398119122257053,
+      "eval_mrpc-test_cosine_recall": 0.9516129032258065,
+      "eval_mrpc-test_dot_accuracy": 0.6868421052631579,
+      "eval_mrpc-test_dot_accuracy_threshold": 17.337886810302734,
+      "eval_mrpc-test_dot_ap": 0.7667271250415733,
+      "eval_mrpc-test_dot_f1": 0.8032786885245902,
+      "eval_mrpc-test_dot_f1_threshold": 12.53590202331543,
+      "eval_mrpc-test_dot_precision": 0.6767955801104972,
+      "eval_mrpc-test_dot_recall": 0.9879032258064516,
+      "eval_mrpc-test_euclidean_accuracy": 0.7473684210526316,
+      "eval_mrpc-test_euclidean_accuracy_threshold": 3.7708821296691895,
+      "eval_mrpc-test_euclidean_ap": 0.8286726870063765,
+      "eval_mrpc-test_euclidean_f1": 0.827338129496403,
+      "eval_mrpc-test_euclidean_f1_threshold": 3.813426971435547,
+      "eval_mrpc-test_euclidean_precision": 0.7467532467532467,
+      "eval_mrpc-test_euclidean_recall": 0.9274193548387096,
+      "eval_mrpc-test_manhattan_accuracy": 0.7421052631578947,
+      "eval_mrpc-test_manhattan_accuracy_threshold": 74.25254821777344,
+      "eval_mrpc-test_manhattan_ap": 0.824711624780047,
+      "eval_mrpc-test_manhattan_f1": 0.8243727598566308,
+      "eval_mrpc-test_manhattan_f1_threshold": 76.98123931884766,
+      "eval_mrpc-test_manhattan_precision": 0.7419354838709677,
+      "eval_mrpc-test_manhattan_recall": 0.9274193548387096,
+      "eval_mrpc-test_max_accuracy": 0.7526315789473684,
+      "eval_mrpc-test_max_accuracy_threshold": 74.25254821777344,
+      "eval_mrpc-test_max_ap": 0.8519630000115989,
+      "eval_mrpc-test_max_f1": 0.8324514991181658,
+      "eval_mrpc-test_max_f1_threshold": 76.98123931884766,
+      "eval_mrpc-test_max_precision": 0.7467532467532467,
+      "eval_mrpc-test_max_recall": 0.9879032258064516,
+      "eval_nli-pairs_loss": 0.8806689381599426,
+      "eval_nli-pairs_runtime": 3.1388,
+      "eval_nli-pairs_samples_per_second": 50.974,
+      "eval_nli-pairs_steps_per_second": 1.593,
+      "eval_sequential_score": 0.5603755036121337,
+      "step": 5168
+    },
+    {
+      "epoch": 1.00038714672861,
+      "eval_vitaminc-pairs_loss": 5.74862813949585,
+      "eval_vitaminc-pairs_runtime": 1.6555,
+      "eval_vitaminc-pairs_samples_per_second": 80.34,
+      "eval_vitaminc-pairs_steps_per_second": 3.02,
+      "step": 5168
+    },
+    {
+      "epoch": 1.00038714672861,
+      "eval_qnli-contrastive_loss": 0.15019993484020233,
+      "eval_qnli-contrastive_runtime": 0.5355,
+      "eval_qnli-contrastive_samples_per_second": 298.759,
+      "eval_qnli-contrastive_steps_per_second": 9.336,
+      "step": 5168
+    },
+    {
+      "epoch": 1.00038714672861,
+      "eval_scitail-pairs-qa_loss": 0.06865032762289047,
+      "eval_scitail-pairs-qa_runtime": 1.2902,
+      "eval_scitail-pairs-qa_samples_per_second": 124.008,
+      "eval_scitail-pairs-qa_steps_per_second": 3.875,
+      "step": 5168
+    },
+    {
+      "epoch": 1.00038714672861,
+      "eval_scitail-pairs-pos_loss": 0.35982105135917664,
+      "eval_scitail-pairs-pos_runtime": 2.4147,
+      "eval_scitail-pairs-pos_samples_per_second": 66.262,
+      "eval_scitail-pairs-pos_steps_per_second": 2.071,
+      "step": 5168
+    },
+    {
+      "epoch": 1.00038714672861,
+      "eval_xsum-pairs_loss": 0.2985193133354187,
+      "eval_xsum-pairs_runtime": 1.3999,
+      "eval_xsum-pairs_samples_per_second": 114.294,
+      "eval_xsum-pairs_steps_per_second": 3.572,
+      "step": 5168
+    },
+    {
+      "epoch": 1.00038714672861,
+      "eval_compression-pairs_loss": 0.0826263278722763,
+      "eval_compression-pairs_runtime": 0.4065,
+      "eval_compression-pairs_samples_per_second": 393.63,
+      "eval_compression-pairs_steps_per_second": 12.301,
+      "step": 5168
+    },
+    {
+      "epoch": 1.00038714672861,
+      "eval_sciq_pairs_loss": 0.23171819746494293,
+      "eval_sciq_pairs_runtime": 8.3356,
+      "eval_sciq_pairs_samples_per_second": 19.195,
+      "eval_sciq_pairs_steps_per_second": 0.6,
+      "step": 5168
+    },
+    {
+      "epoch": 1.00038714672861,
+      "eval_qasc_pairs_loss": 0.22728173434734344,
+      "eval_qasc_pairs_runtime": 1.4954,
+      "eval_qasc_pairs_samples_per_second": 106.994,
+      "eval_qasc_pairs_steps_per_second": 3.344,
+      "step": 5168
+    },
+    {
+      "epoch": 1.00038714672861,
+      "eval_qasc_facts_sym_loss": 0.14946210384368896,
+      "eval_qasc_facts_sym_runtime": 0.3454,
+      "eval_qasc_facts_sym_samples_per_second": 463.224,
+      "eval_qasc_facts_sym_steps_per_second": 14.476,
+      "step": 5168
+    },
+    {
+      "epoch": 1.00038714672861,
+      "eval_openbookqa_pairs_loss": 1.7004419565200806,
+      "eval_openbookqa_pairs_runtime": 1.4576,
+      "eval_openbookqa_pairs_samples_per_second": 109.771,
+      "eval_openbookqa_pairs_steps_per_second": 3.43,
+      "step": 5168
+    },
+    {
+      "epoch": 1.00038714672861,
+      "eval_msmarco_pairs_loss": 0.5557130575180054,
+      "eval_msmarco_pairs_runtime": 3.3251,
+      "eval_msmarco_pairs_samples_per_second": 48.119,
+      "eval_msmarco_pairs_steps_per_second": 1.504,
+      "step": 5168
+    },
+    {
+      "epoch": 1.00038714672861,
+      "eval_nq_pairs_loss": 0.44789552688598633,
+      "eval_nq_pairs_runtime": 7.8474,
+      "eval_nq_pairs_samples_per_second": 20.389,
+      "eval_nq_pairs_steps_per_second": 0.637,
+      "step": 5168
+    },
+    {
+      "epoch": 1.00038714672861,
+      "eval_trivia_pairs_loss": 0.7468730807304382,
+      "eval_trivia_pairs_runtime": 10.2776,
+      "eval_trivia_pairs_samples_per_second": 15.568,
+      "eval_trivia_pairs_steps_per_second": 0.486,
+      "step": 5168
+    },
+    {
+      "epoch": 1.00038714672861,
+      "eval_quora_pairs_loss": 0.2115297168493271,
+      "eval_quora_pairs_runtime": 4.0407,
+      "eval_quora_pairs_samples_per_second": 167.05,
+      "eval_quora_pairs_steps_per_second": 5.445,
+      "step": 5168
+    },
+    {
+      "epoch": 1.00038714672861,
+      "eval_gooaq_pairs_loss": 0.53202223777771,
+      "eval_gooaq_pairs_runtime": 2.2862,
+      "eval_gooaq_pairs_samples_per_second": 69.986,
+      "eval_gooaq_pairs_steps_per_second": 2.187,
+      "step": 5168
+    },
+    {
+      "epoch": 1.00038714672861,
+      "eval_mrpc_pairs_loss": 0.04773856699466705,
+      "eval_mrpc_pairs_runtime": 0.3785,
+      "eval_mrpc_pairs_samples_per_second": 422.744,
+      "eval_mrpc_pairs_steps_per_second": 13.211,
+      "step": 5168
+    },
+    {
+      "epoch": 1.0065814943863725,
+      "grad_norm": 1.1308379173278809,
+      "learning_rate": 2.99985958598248e-05,
+      "loss": 0.4946,
+      "step": 5200
+    },
+    {
+      "epoch": 1.019163763066202,
+      "grad_norm": 0.7707048058509827,
+      "learning_rate": 2.9986216360213095e-05,
+      "loss": 0.4134,
+      "step": 5265
+    },
+    {
+      "epoch": 1.0317460317460316,
+      "grad_norm": 19.599868774414062,
+      "learning_rate": 2.9960574370109496e-05,
+      "loss": 0.4702,
+      "step": 5330
+    },
+    {
+      "epoch": 1.0443283004258614,
+      "grad_norm": 0.9096149206161499,
+      "learning_rate": 2.992178447249302e-05,
+      "loss": 0.5392,
+      "step": 5395
+    },
+    {
+      "epoch": 1.056910569105691,
+      "grad_norm": 19.951955795288086,
+      "learning_rate": 2.986988075736407e-05,
+      "loss": 0.4845,
+      "step": 5460
+    },
+    {
+      "epoch": 1.0694928377855206,
+      "grad_norm": 13.416616439819336,
+      "learning_rate": 2.980490883963215e-05,
+      "loss": 0.4601,
+      "step": 5525
+    },
+    {
+      "epoch": 1.0820751064653504,
+      "grad_norm": 1.105792760848999,
+      "learning_rate": 2.9726925819027805e-05,
+      "loss": 0.5683,
+      "step": 5590
+    },
+    {
+      "epoch": 1.09465737514518,
+      "grad_norm": 6.264986991882324,
+      "learning_rate": 2.9636000229921248e-05,
+      "loss": 0.4742,
+      "step": 5655
+    },
+    {
+      "epoch": 1.1072396438250096,
+      "grad_norm": 0.4068054258823395,
+      "learning_rate": 2.9532211981091813e-05,
+      "loss": 0.5187,
+      "step": 5720
+    },
+    {
+      "epoch": 1.1198219125048394,
+      "grad_norm": 4.136714935302734,
+      "learning_rate": 2.941565228550117e-05,
+      "loss": 0.4888,
+      "step": 5785
+    },
+    {
+      "epoch": 1.132404181184669,
+      "grad_norm": 1.7270488739013672,
+      "learning_rate": 2.9286423580131986e-05,
+      "loss": 0.3486,
+      "step": 5850
+    },
+    {
+      "epoch": 1.1449864498644986,
+      "grad_norm": 3.33518123626709,
+      "learning_rate": 2.9144639435962487e-05,
+      "loss": 0.4714,
+      "step": 5915
+    },
+    {
+      "epoch": 1.1575687185443284,
+      "grad_norm": 3.2314696311950684,
+      "learning_rate": 2.8990424458156108e-05,
+      "loss": 0.5638,
+      "step": 5980
+    },
+    {
+      "epoch": 1.170150987224158,
+      "grad_norm": 28.55368423461914,
+      "learning_rate": 2.882391417655382e-05,
+      "loss": 0.457,
+      "step": 6045
+    },
+    {
+      "epoch": 1.1827332559039876,
+      "grad_norm": 17.361839294433594,
+      "learning_rate": 2.8645254926565508e-05,
+      "loss": 0.5419,
+      "step": 6110
+    },
+    {
+      "epoch": 1.1953155245838172,
+      "grad_norm": 14.851773262023926,
+      "learning_rate": 2.8454603720564953e-05,
+      "loss": 0.4478,
+      "step": 6175
+    },
+    {
+      "epoch": 1.207897793263647,
+      "grad_norm": 18.820308685302734,
+      "learning_rate": 2.825212810990158e-05,
+      "loss": 0.5012,
+      "step": 6240
+    },
+    {
+      "epoch": 1.2204800619434766,
+      "grad_norm": 0.15901756286621094,
+      "learning_rate": 2.803800603765008e-05,
+      "loss": 0.4505,
+      "step": 6305
+    },
+    {
+      "epoch": 1.2330623306233062,
+      "grad_norm": 8.319548606872559,
+      "learning_rate": 2.7812425682227454e-05,
+      "loss": 0.4677,
+      "step": 6370
+    },
+    {
+      "epoch": 1.245644599303136,
+      "grad_norm": 6.264440536499023,
+      "learning_rate": 2.757558529201485e-05,
+      "loss": 0.478,
+      "step": 6435
+    },
+    {
+      "epoch": 1.2504839334107627,
+      "eval_StS-test_pearson_cosine": 0.8749809466696259,
+      "eval_StS-test_pearson_dot": 0.7726333541295806,
+      "eval_StS-test_pearson_euclidean": 0.8635304000211516,
+      "eval_StS-test_pearson_manhattan": 0.8629148539097312,
+      "eval_StS-test_pearson_max": 0.8749809466696259,
+      "eval_StS-test_spearman_cosine": 0.8911637850925943,
+      "eval_StS-test_spearman_dot": 0.7791087210103667,
+      "eval_StS-test_spearman_euclidean": 0.8717319673612609,
+      "eval_StS-test_spearman_manhattan": 0.873430814633936,
+      "eval_StS-test_spearman_max": 0.8911637850925943,
+      "eval_Vitaminc-test_cosine_accuracy": 0.5789473684210527,
+      "eval_Vitaminc-test_cosine_accuracy_threshold": 0.7059692144393921,
+      "eval_Vitaminc-test_cosine_ap": 0.5681853466607133,
+      "eval_Vitaminc-test_cosine_f1": 0.6740740740740742,
+      "eval_Vitaminc-test_cosine_f1_threshold": 0.47524070739746094,
+      "eval_Vitaminc-test_cosine_precision": 0.5126760563380282,
+      "eval_Vitaminc-test_cosine_recall": 0.9837837837837838,
+      "eval_Vitaminc-test_dot_accuracy": 0.5605263157894737,
+      "eval_Vitaminc-test_dot_accuracy_threshold": 18.60018539428711,
+      "eval_Vitaminc-test_dot_ap": 0.542223191158599,
+      "eval_Vitaminc-test_dot_f1": 0.6678832116788321,
+      "eval_Vitaminc-test_dot_f1_threshold": 11.097679138183594,
+      "eval_Vitaminc-test_dot_precision": 0.5041322314049587,
+      "eval_Vitaminc-test_dot_recall": 0.9891891891891892,
+      "eval_Vitaminc-test_euclidean_accuracy": 0.5605263157894737,
+      "eval_Vitaminc-test_euclidean_accuracy_threshold": 3.57100772857666,
+      "eval_Vitaminc-test_euclidean_ap": 0.5599668277103405,
+      "eval_Vitaminc-test_euclidean_f1": 0.6691729323308271,
+      "eval_Vitaminc-test_euclidean_f1_threshold": 6.093693733215332,
+      "eval_Vitaminc-test_euclidean_precision": 0.5129682997118156,
+      "eval_Vitaminc-test_euclidean_recall": 0.9621621621621622,
+      "eval_Vitaminc-test_manhattan_accuracy": 0.5710526315789474,
+      "eval_Vitaminc-test_manhattan_accuracy_threshold": 77.09724426269531,
+      "eval_Vitaminc-test_manhattan_ap": 0.5609220991144057,
+      "eval_Vitaminc-test_manhattan_f1": 0.6666666666666666,
+      "eval_Vitaminc-test_manhattan_f1_threshold": 129.85740661621094,
+      "eval_Vitaminc-test_manhattan_precision": 0.5070422535211268,
+      "eval_Vitaminc-test_manhattan_recall": 0.972972972972973,
+      "eval_Vitaminc-test_max_accuracy": 0.5789473684210527,
+      "eval_Vitaminc-test_max_accuracy_threshold": 77.09724426269531,
+      "eval_Vitaminc-test_max_ap": 0.5681853466607133,
+      "eval_Vitaminc-test_max_f1": 0.6740740740740742,
+      "eval_Vitaminc-test_max_f1_threshold": 129.85740661621094,
+      "eval_Vitaminc-test_max_precision": 0.5129682997118156,
+      "eval_Vitaminc-test_max_recall": 0.9891891891891892,
+      "eval_mrpc-test_cosine_accuracy": 0.7421052631578947,
+      "eval_mrpc-test_cosine_accuracy_threshold": 0.6662319898605347,
+      "eval_mrpc-test_cosine_ap": 0.847874875651623,
+      "eval_mrpc-test_cosine_f1": 0.831615120274914,
+      "eval_mrpc-test_cosine_f1_threshold": 0.6662319898605347,
+      "eval_mrpc-test_cosine_precision": 0.7245508982035929,
+      "eval_mrpc-test_cosine_recall": 0.9758064516129032,
+      "eval_mrpc-test_dot_accuracy": 0.6973684210526315,
+      "eval_mrpc-test_dot_accuracy_threshold": 13.194486618041992,
+      "eval_mrpc-test_dot_ap": 0.7550727858117499,
+      "eval_mrpc-test_dot_f1": 0.803305785123967,
+      "eval_mrpc-test_dot_f1_threshold": 11.023738861083984,
+      "eval_mrpc-test_dot_precision": 0.680672268907563,
+      "eval_mrpc-test_dot_recall": 0.9798387096774194,
+      "eval_mrpc-test_euclidean_accuracy": 0.7526315789473684,
+      "eval_mrpc-test_euclidean_accuracy_threshold": 3.5304112434387207,
+      "eval_mrpc-test_euclidean_ap": 0.8291889523682024,
+      "eval_mrpc-test_euclidean_f1": 0.829090909090909,
+      "eval_mrpc-test_euclidean_f1_threshold": 3.5304112434387207,
+      "eval_mrpc-test_euclidean_precision": 0.7549668874172185,
+      "eval_mrpc-test_euclidean_recall": 0.9193548387096774,
+      "eval_mrpc-test_manhattan_accuracy": 0.7473684210526316,
+      "eval_mrpc-test_manhattan_accuracy_threshold": 70.23236846923828,
+      "eval_mrpc-test_manhattan_ap": 0.827542993951462,
+      "eval_mrpc-test_manhattan_f1": 0.8241758241758242,
+      "eval_mrpc-test_manhattan_f1_threshold": 70.23236846923828,
+      "eval_mrpc-test_manhattan_precision": 0.7550335570469798,
+      "eval_mrpc-test_manhattan_recall": 0.907258064516129,
+      "eval_mrpc-test_max_accuracy": 0.7526315789473684,
+      "eval_mrpc-test_max_accuracy_threshold": 70.23236846923828,
+      "eval_mrpc-test_max_ap": 0.847874875651623,
+      "eval_mrpc-test_max_f1": 0.831615120274914,
+      "eval_mrpc-test_max_f1_threshold": 70.23236846923828,
+      "eval_mrpc-test_max_precision": 0.7550335570469798,
+      "eval_mrpc-test_max_recall": 0.9798387096774194,
+      "eval_nli-pairs_loss": 0.8716467022895813,
+      "eval_nli-pairs_runtime": 3.0359,
+      "eval_nli-pairs_samples_per_second": 52.702,
+      "eval_nli-pairs_steps_per_second": 1.647,
+      "eval_sequential_score": 0.5681853466607133,
+      "step": 6460
+    },
+    {
+      "epoch": 1.2504839334107627,
+      "eval_vitaminc-pairs_loss": 5.813481330871582,
+      "eval_vitaminc-pairs_runtime": 1.6243,
+      "eval_vitaminc-pairs_samples_per_second": 81.88,
+      "eval_vitaminc-pairs_steps_per_second": 3.078,
+      "step": 6460
+    },
+    {
+      "epoch": 1.2504839334107627,
+      "eval_qnli-contrastive_loss": 0.09026918560266495,
+      "eval_qnli-contrastive_runtime": 0.5258,
+      "eval_qnli-contrastive_samples_per_second": 304.322,
+      "eval_qnli-contrastive_steps_per_second": 9.51,
+      "step": 6460
+    },
+    {
+      "epoch": 1.2504839334107627,
+      "eval_scitail-pairs-qa_loss": 0.06381334364414215,
+      "eval_scitail-pairs-qa_runtime": 1.2167,
+      "eval_scitail-pairs-qa_samples_per_second": 131.505,
+      "eval_scitail-pairs-qa_steps_per_second": 4.11,
+      "step": 6460
+    },
+    {
+      "epoch": 1.2504839334107627,
+      "eval_scitail-pairs-pos_loss": 0.3648618757724762,
+      "eval_scitail-pairs-pos_runtime": 2.3843,
+      "eval_scitail-pairs-pos_samples_per_second": 67.105,
+      "eval_scitail-pairs-pos_steps_per_second": 2.097,
+      "step": 6460
+    },
+    {
+      "epoch": 1.2504839334107627,
+      "eval_xsum-pairs_loss": 0.317168653011322,
+      "eval_xsum-pairs_runtime": 1.408,
+      "eval_xsum-pairs_samples_per_second": 113.634,
+      "eval_xsum-pairs_steps_per_second": 3.551,
+      "step": 6460
+    },
+    {
+      "epoch": 1.2504839334107627,
+      "eval_compression-pairs_loss": 0.08144789934158325,
+      "eval_compression-pairs_runtime": 0.406,
+      "eval_compression-pairs_samples_per_second": 394.049,
+      "eval_compression-pairs_steps_per_second": 12.314,
+      "step": 6460
+    },
+    {
+      "epoch": 1.2504839334107627,
+      "eval_sciq_pairs_loss": 0.25741466879844666,
+      "eval_sciq_pairs_runtime": 8.0381,
+      "eval_sciq_pairs_samples_per_second": 19.905,
+      "eval_sciq_pairs_steps_per_second": 0.622,
+      "step": 6460
+    },
+    {
+      "epoch": 1.2504839334107627,
+      "eval_qasc_pairs_loss": 0.19735541939735413,
+      "eval_qasc_pairs_runtime": 1.4554,
+      "eval_qasc_pairs_samples_per_second": 109.935,
+      "eval_qasc_pairs_steps_per_second": 3.435,
+      "step": 6460
+    },
+    {
+      "epoch": 1.2504839334107627,
+      "eval_qasc_facts_sym_loss": 0.14057289063930511,
+      "eval_qasc_facts_sym_runtime": 0.3302,
+      "eval_qasc_facts_sym_samples_per_second": 484.627,
+      "eval_qasc_facts_sym_steps_per_second": 15.145,
+      "step": 6460
+    },
+    {
+      "epoch": 1.2504839334107627,
+      "eval_openbookqa_pairs_loss": 1.6864427328109741,
+      "eval_openbookqa_pairs_runtime": 1.272,
+      "eval_openbookqa_pairs_samples_per_second": 125.791,
+      "eval_openbookqa_pairs_steps_per_second": 3.931,
+      "step": 6460
+    },
+    {
+      "epoch": 1.2504839334107627,
+      "eval_msmarco_pairs_loss": 0.5884890556335449,
+      "eval_msmarco_pairs_runtime": 3.1363,
+      "eval_msmarco_pairs_samples_per_second": 51.016,
+      "eval_msmarco_pairs_steps_per_second": 1.594,
+      "step": 6460
+    },
+    {
+      "epoch": 1.2504839334107627,
+      "eval_nq_pairs_loss": 0.4531589448451996,
+      "eval_nq_pairs_runtime": 7.7451,
+      "eval_nq_pairs_samples_per_second": 20.658,
+      "eval_nq_pairs_steps_per_second": 0.646,
+      "step": 6460
+    },
+    {
+      "epoch": 1.2504839334107627,
+      "eval_trivia_pairs_loss": 0.8866144418716431,
+      "eval_trivia_pairs_runtime": 10.2442,
+      "eval_trivia_pairs_samples_per_second": 15.619,
+      "eval_trivia_pairs_steps_per_second": 0.488,
+      "step": 6460
+    },
+    {
+      "epoch": 1.2504839334107627,
+      "eval_quora_pairs_loss": 0.2143346667289734,
+      "eval_quora_pairs_runtime": 3.9106,
+      "eval_quora_pairs_samples_per_second": 172.607,
+      "eval_quora_pairs_steps_per_second": 5.626,
+      "step": 6460
+    },
+    {
+      "epoch": 1.2504839334107627,
+      "eval_gooaq_pairs_loss": 0.535525381565094,
+      "eval_gooaq_pairs_runtime": 2.2827,
+      "eval_gooaq_pairs_samples_per_second": 70.094,
+      "eval_gooaq_pairs_steps_per_second": 2.19,
+      "step": 6460
+    },
+    {
+      "epoch": 1.2504839334107627,
+      "eval_mrpc_pairs_loss": 0.04031027480959892,
+      "eval_mrpc_pairs_runtime": 0.3763,
+      "eval_mrpc_pairs_samples_per_second": 425.243,
+      "eval_mrpc_pairs_steps_per_second": 13.289,
+      "step": 6460
+    },
+    {
+      "epoch": 1.2582268679829656,
+      "grad_norm": 3.637460947036743,
+      "learning_rate": 2.7327693011129508e-05,
+      "loss": 0.5004,
+      "step": 6500
+    },
+    {
+      "epoch": 1.2708091366627952,
+      "grad_norm": 11.11975383758545,
+      "learning_rate": 2.7068966696500025e-05,
+      "loss": 0.5453,
+      "step": 6565
+    },
+    {
+      "epoch": 1.2833914053426247,
+      "grad_norm": 0.5408668518066406,
+      "learning_rate": 2.6799633726405593e-05,
+      "loss": 0.529,
+      "step": 6630
+    },
+    {
+      "epoch": 1.2959736740224546,
+      "grad_norm": 1.3864606618881226,
+      "learning_rate": 2.6519930800647514e-05,
+      "loss": 0.4343,
+      "step": 6695
+    },
+    {
+      "epoch": 1.3085559427022841,
+      "grad_norm": 10.643692970275879,
+      "learning_rate": 2.623010373252868e-05,
+      "loss": 0.4235,
+      "step": 6760
+    },
+    {
+      "epoch": 1.321138211382114,
+      "grad_norm": 15.424538612365723,
+      "learning_rate": 2.59304072328237e-05,
+      "loss": 0.4493,
+      "step": 6825
+    },
+    {
+      "epoch": 1.3337204800619435,
+      "grad_norm": 6.388729095458984,
+      "learning_rate": 2.5621104685929617e-05,
+      "loss": 0.4349,
+      "step": 6890
+    },
+    {
+      "epoch": 1.3463027487417731,
+      "grad_norm": 17.623123168945312,
+      "learning_rate": 2.5302467918393965e-05,
+      "loss": 0.4016,
+      "step": 6955
+    },
+    {
+      "epoch": 1.3588850174216027,
+      "grad_norm": 0.2991964817047119,
+      "learning_rate": 2.49747769600235e-05,
+      "loss": 0.39,
+      "step": 7020
+    },
+    {
+      "epoch": 1.3714672861014323,
+      "grad_norm": 2.3669910430908203,
+      "learning_rate": 2.4638319797783612e-05,
+      "loss": 0.5162,
+      "step": 7085
+    },
+    {
+      "epoch": 1.3840495547812621,
+      "grad_norm": 0.24493175745010376,
+      "learning_rate": 2.4293392122704728e-05,
+      "loss": 0.4186,
+      "step": 7150
+    },
+    {
+      "epoch": 1.3966318234610917,
+      "grad_norm": 1.4514191150665283,
+      "learning_rate": 2.3940297070018048e-05,
+      "loss": 0.3867,
+      "step": 7215
+    },
+    {
+      "epoch": 1.4092140921409215,
+      "grad_norm": 12.809524536132812,
+      "learning_rate": 2.3579344952749066e-05,
+      "loss": 0.5076,
+      "step": 7280
+    },
+    {
+      "epoch": 1.4217963608207511,
+      "grad_norm": 6.879631996154785,
+      "learning_rate": 2.3210852989002988e-05,
+      "loss": 0.4872,
+      "step": 7345
+    },
+    {
+      "epoch": 1.4343786295005807,
+      "grad_norm": 2.3940792083740234,
+      "learning_rate": 2.283514502318168e-05,
+      "loss": 0.4508,
+      "step": 7410
+    },
+    {
+      "epoch": 1.4469608981804103,
+      "grad_norm": 3.3337996006011963,
+      "learning_rate": 2.2452551241377226e-05,
+      "loss": 0.4261,
+      "step": 7475
+    },
+    {
+      "epoch": 1.45954316686024,
+      "grad_norm": 18.75865364074707,
+      "learning_rate": 2.2063407881192124e-05,
+      "loss": 0.4444,
+      "step": 7540
+    },
+    {
+      "epoch": 1.4721254355400697,
+      "grad_norm": 3.5605080127716064,
+      "learning_rate": 2.1668056936241207e-05,
+      "loss": 0.4043,
+      "step": 7605
+    },
+    {
+      "epoch": 1.4847077042198993,
+      "grad_norm": 0.22081464529037476,
+      "learning_rate": 2.1266845855595004e-05,
+      "loss": 0.4451,
+      "step": 7670
+    },
+    {
+      "epoch": 1.497289972899729,
+      "grad_norm": 0.37092337012290955,
+      "learning_rate": 2.086012723842858e-05,
+      "loss": 0.4312,
+      "step": 7735
+    },
+    {
+      "epoch": 1.5005807200929153,
+      "eval_StS-test_pearson_cosine": 0.8818926010663196,
+      "eval_StS-test_pearson_dot": 0.7948939773009118,
+      "eval_StS-test_pearson_euclidean": 0.8683502643949064,
+      "eval_StS-test_pearson_manhattan": 0.8693690003312309,
+      "eval_StS-test_pearson_max": 0.8818926010663196,
+      "eval_StS-test_spearman_cosine": 0.8954799233919072,
+      "eval_StS-test_spearman_dot": 0.8002703657023814,
+      "eval_StS-test_spearman_euclidean": 0.8705533161218193,
+      "eval_StS-test_spearman_manhattan": 0.8724718674009548,
+      "eval_StS-test_spearman_max": 0.8954799233919072,
+      "eval_Vitaminc-test_cosine_accuracy": 0.5736842105263158,
+      "eval_Vitaminc-test_cosine_accuracy_threshold": 0.7514327764511108,
+      "eval_Vitaminc-test_cosine_ap": 0.5635780272169032,
+      "eval_Vitaminc-test_cosine_f1": 0.6728280961182994,
+      "eval_Vitaminc-test_cosine_f1_threshold": 0.4688807725906372,
+      "eval_Vitaminc-test_cosine_precision": 0.5112359550561798,
+      "eval_Vitaminc-test_cosine_recall": 0.9837837837837838,
+      "eval_Vitaminc-test_dot_accuracy": 0.5578947368421052,
+      "eval_Vitaminc-test_dot_accuracy_threshold": 18.673751831054688,
+      "eval_Vitaminc-test_dot_ap": 0.5505510423866934,
+      "eval_Vitaminc-test_dot_f1": 0.6679389312977099,
+      "eval_Vitaminc-test_dot_f1_threshold": 13.694689750671387,
+      "eval_Vitaminc-test_dot_precision": 0.5162241887905604,
+      "eval_Vitaminc-test_dot_recall": 0.9459459459459459,
+      "eval_Vitaminc-test_euclidean_accuracy": 0.5763157894736842,
+      "eval_Vitaminc-test_euclidean_accuracy_threshold": 3.70269513130188,
+      "eval_Vitaminc-test_euclidean_ap": 0.5537568551698975,
+      "eval_Vitaminc-test_euclidean_f1": 0.6654611211573237,
+      "eval_Vitaminc-test_euclidean_f1_threshold": 6.518295764923096,
+      "eval_Vitaminc-test_euclidean_precision": 0.5,
+      "eval_Vitaminc-test_euclidean_recall": 0.9945945945945946,
+      "eval_Vitaminc-test_manhattan_accuracy": 0.5789473684210527,
+      "eval_Vitaminc-test_manhattan_accuracy_threshold": 74.57982635498047,
+      "eval_Vitaminc-test_manhattan_ap": 0.5568365932864207,
+      "eval_Vitaminc-test_manhattan_f1": 0.6654676258992805,
+      "eval_Vitaminc-test_manhattan_f1_threshold": 139.71929931640625,
+      "eval_Vitaminc-test_manhattan_precision": 0.49865229110512127,
+      "eval_Vitaminc-test_manhattan_recall": 1.0,
+      "eval_Vitaminc-test_max_accuracy": 0.5789473684210527,
+      "eval_Vitaminc-test_max_accuracy_threshold": 74.57982635498047,
+      "eval_Vitaminc-test_max_ap": 0.5635780272169032,
+      "eval_Vitaminc-test_max_f1": 0.6728280961182994,
+      "eval_Vitaminc-test_max_f1_threshold": 139.71929931640625,
+      "eval_Vitaminc-test_max_precision": 0.5162241887905604,
+      "eval_Vitaminc-test_max_recall": 1.0,
+      "eval_mrpc-test_cosine_accuracy": 0.7315789473684211,
+      "eval_mrpc-test_cosine_accuracy_threshold": 0.7411860227584839,
+      "eval_mrpc-test_cosine_ap": 0.8441517199589379,
+      "eval_mrpc-test_cosine_f1": 0.8186528497409327,
+      "eval_mrpc-test_cosine_f1_threshold": 0.672976553440094,
+      "eval_mrpc-test_cosine_precision": 0.716012084592145,
+      "eval_mrpc-test_cosine_recall": 0.9556451612903226,
+      "eval_mrpc-test_dot_accuracy": 0.6921052631578948,
+      "eval_mrpc-test_dot_accuracy_threshold": 13.14952278137207,
+      "eval_mrpc-test_dot_ap": 0.761780840207299,
+      "eval_mrpc-test_dot_f1": 0.8045602605863192,
+      "eval_mrpc-test_dot_f1_threshold": 8.986488342285156,
+      "eval_mrpc-test_dot_precision": 0.674863387978142,
+      "eval_mrpc-test_dot_recall": 0.9959677419354839,
+      "eval_mrpc-test_euclidean_accuracy": 0.7421052631578947,
+      "eval_mrpc-test_euclidean_accuracy_threshold": 3.5243382453918457,
+      "eval_mrpc-test_euclidean_ap": 0.8217039781172437,
+      "eval_mrpc-test_euclidean_f1": 0.8231046931407942,
+      "eval_mrpc-test_euclidean_f1_threshold": 3.5243382453918457,
+      "eval_mrpc-test_euclidean_precision": 0.7450980392156863,
+      "eval_mrpc-test_euclidean_recall": 0.9193548387096774,
+      "eval_mrpc-test_manhattan_accuracy": 0.7368421052631579,
+      "eval_mrpc-test_manhattan_accuracy_threshold": 70.5282211303711,
+      "eval_mrpc-test_manhattan_ap": 0.8193721480476763,
+      "eval_mrpc-test_manhattan_f1": 0.8201438848920863,
+      "eval_mrpc-test_manhattan_f1_threshold": 70.67774963378906,
+      "eval_mrpc-test_manhattan_precision": 0.7402597402597403,
+      "eval_mrpc-test_manhattan_recall": 0.9193548387096774,
+      "eval_mrpc-test_max_accuracy": 0.7421052631578947,
+      "eval_mrpc-test_max_accuracy_threshold": 70.5282211303711,
+      "eval_mrpc-test_max_ap": 0.8441517199589379,
+      "eval_mrpc-test_max_f1": 0.8231046931407942,
+      "eval_mrpc-test_max_f1_threshold": 70.67774963378906,
+      "eval_mrpc-test_max_precision": 0.7450980392156863,
+      "eval_mrpc-test_max_recall": 0.9959677419354839,
+      "eval_nli-pairs_loss": 0.7820265293121338,
+      "eval_nli-pairs_runtime": 3.1498,
+      "eval_nli-pairs_samples_per_second": 50.797,
+      "eval_nli-pairs_steps_per_second": 1.587,
+      "eval_sequential_score": 0.5635780272169032,
+      "step": 7752
+    },
+    {
+      "epoch": 1.5005807200929153,
+      "eval_vitaminc-pairs_loss": 5.762389183044434,
+      "eval_vitaminc-pairs_runtime": 1.611,
+      "eval_vitaminc-pairs_samples_per_second": 82.56,
+      "eval_vitaminc-pairs_steps_per_second": 3.104,
+      "step": 7752
+    },
+    {
+      "epoch": 1.5005807200929153,
+      "eval_qnli-contrastive_loss": 0.16094566881656647,
+      "eval_qnli-contrastive_runtime": 0.5292,
+      "eval_qnli-contrastive_samples_per_second": 302.371,
+      "eval_qnli-contrastive_steps_per_second": 9.449,
+      "step": 7752
+    },
+    {
+      "epoch": 1.5005807200929153,
+      "eval_scitail-pairs-qa_loss": 0.05526223033666611,
+      "eval_scitail-pairs-qa_runtime": 1.3843,
+      "eval_scitail-pairs-qa_samples_per_second": 115.583,
+      "eval_scitail-pairs-qa_steps_per_second": 3.612,
+      "step": 7752
+    },
+    {
+      "epoch": 1.5005807200929153,
+      "eval_scitail-pairs-pos_loss": 0.3169253170490265,
+      "eval_scitail-pairs-pos_runtime": 2.5617,
+      "eval_scitail-pairs-pos_samples_per_second": 62.46,
+      "eval_scitail-pairs-pos_steps_per_second": 1.952,
+      "step": 7752
+    },
+    {
+      "epoch": 1.5005807200929153,
+      "eval_xsum-pairs_loss": 0.27240970730781555,
+      "eval_xsum-pairs_runtime": 1.421,
+      "eval_xsum-pairs_samples_per_second": 112.599,
+      "eval_xsum-pairs_steps_per_second": 3.519,
+      "step": 7752
+    },
+    {
+      "epoch": 1.5005807200929153,
+      "eval_compression-pairs_loss": 0.0707278698682785,
+      "eval_compression-pairs_runtime": 0.4155,
+      "eval_compression-pairs_samples_per_second": 385.033,
+      "eval_compression-pairs_steps_per_second": 12.032,
+      "step": 7752
+    },
+    {
+      "epoch": 1.5005807200929153,
+      "eval_sciq_pairs_loss": 0.25230515003204346,
+      "eval_sciq_pairs_runtime": 8.1954,
+      "eval_sciq_pairs_samples_per_second": 19.523,
+      "eval_sciq_pairs_steps_per_second": 0.61,
+      "step": 7752
+    },
+    {
+      "epoch": 1.5005807200929153,
+      "eval_qasc_pairs_loss": 0.17636016011238098,
+      "eval_qasc_pairs_runtime": 1.4827,
+      "eval_qasc_pairs_samples_per_second": 107.91,
+      "eval_qasc_pairs_steps_per_second": 3.372,
+      "step": 7752
+    },
+    {
+      "epoch": 1.5005807200929153,
+      "eval_qasc_facts_sym_loss": 0.12885577976703644,
+      "eval_qasc_facts_sym_runtime": 0.3403,
+      "eval_qasc_facts_sym_samples_per_second": 470.225,
+      "eval_qasc_facts_sym_steps_per_second": 14.695,
+      "step": 7752
+    },
+    {
+      "epoch": 1.5005807200929153,
+      "eval_openbookqa_pairs_loss": 1.6650470495224,
+      "eval_openbookqa_pairs_runtime": 1.344,
+      "eval_openbookqa_pairs_samples_per_second": 119.045,
+      "eval_openbookqa_pairs_steps_per_second": 3.72,
+      "step": 7752
+    },
+    {
+      "epoch": 1.5005807200929153,
+      "eval_msmarco_pairs_loss": 0.49048948287963867,
+      "eval_msmarco_pairs_runtime": 3.29,
+      "eval_msmarco_pairs_samples_per_second": 48.633,
+      "eval_msmarco_pairs_steps_per_second": 1.52,
+      "step": 7752
+    },
+    {
+      "epoch": 1.5005807200929153,
+      "eval_nq_pairs_loss": 0.4190645217895508,
+      "eval_nq_pairs_runtime": 7.8722,
+      "eval_nq_pairs_samples_per_second": 20.325,
+      "eval_nq_pairs_steps_per_second": 0.635,
+      "step": 7752
+    },
+    {
+      "epoch": 1.5005807200929153,
+      "eval_trivia_pairs_loss": 0.6449421644210815,
+      "eval_trivia_pairs_runtime": 10.3733,
+      "eval_trivia_pairs_samples_per_second": 15.424,
+      "eval_trivia_pairs_steps_per_second": 0.482,
+      "step": 7752
+    },
+    {
+      "epoch": 1.5005807200929153,
+      "eval_quora_pairs_loss": 0.18587273359298706,
+      "eval_quora_pairs_runtime": 4.2871,
+      "eval_quora_pairs_samples_per_second": 157.45,
+      "eval_quora_pairs_steps_per_second": 5.132,
+      "step": 7752
+    },
+    {
+      "epoch": 1.5005807200929153,
+      "eval_gooaq_pairs_loss": 0.4795718193054199,
+      "eval_gooaq_pairs_runtime": 2.2929,
+      "eval_gooaq_pairs_samples_per_second": 69.782,
+      "eval_gooaq_pairs_steps_per_second": 2.181,
+      "step": 7752
+    },
+    {
+      "epoch": 1.5005807200929153,
+      "eval_mrpc_pairs_loss": 0.03764195367693901,
+      "eval_mrpc_pairs_runtime": 0.3793,
+      "eval_mrpc_pairs_samples_per_second": 421.87,
+      "eval_mrpc_pairs_steps_per_second": 13.183,
+      "step": 7752
+    },
+    {
+      "epoch": 1.5098722415795587,
+      "grad_norm": 0.42758312821388245,
+      "learning_rate": 2.0448258524144327e-05,
+      "loss": 0.3881,
+      "step": 7800
+    },
+    {
+      "epoch": 1.5224545102593883,
+      "grad_norm": 0.5778383612632751,
+      "learning_rate": 2.0031601678240954e-05,
+      "loss": 0.3704,
+      "step": 7865
+    },
+    {
+      "epoch": 1.5350367789392179,
+      "grad_norm": 19.273122787475586,
+      "learning_rate": 1.9610522874204793e-05,
+      "loss": 0.451,
+      "step": 7930
+    },
+    {
+      "epoch": 1.5476190476190477,
+      "grad_norm": 5.756499290466309,
+      "learning_rate": 1.918539217170294e-05,
+      "loss": 0.4553,
+      "step": 7995
+    },
+    {
+      "epoch": 1.5602013162988773,
+      "grad_norm": 5.21534538269043,
+      "learning_rate": 1.8756583191361127e-05,
+      "loss": 0.411,
+      "step": 8060
+    },
+    {
+      "epoch": 1.572783584978707,
+      "grad_norm": 5.580538272857666,
+      "learning_rate": 1.8324472786412036e-05,
+      "loss": 0.446,
+      "step": 8125
+    },
+    {
+      "epoch": 1.5853658536585367,
+      "grad_norm": 16.46262550354004,
+      "learning_rate": 1.7889440711502724e-05,
+      "loss": 0.2831,
+      "step": 8190
+    },
+    {
+      "epoch": 1.5979481223383663,
+      "grad_norm": 0.19373361766338348,
+      "learning_rate": 1.7451869288952144e-05,
+      "loss": 0.3546,
+      "step": 8255
+    },
+    {
+      "epoch": 1.6105303910181958,
+      "grad_norm": 1.8671798706054688,
+      "learning_rate": 1.701214307275211e-05,
+      "loss": 0.3713,
+      "step": 8320
+    },
+    {
+      "epoch": 1.6231126596980254,
+      "grad_norm": 13.042320251464844,
+      "learning_rate": 1.657064851060702e-05,
+      "loss": 0.4221,
+      "step": 8385
+    },
+    {
+      "epoch": 1.6356949283778552,
+      "grad_norm": 0.6017350554466248,
+      "learning_rate": 1.612777360430923e-05,
+      "loss": 0.3154,
+      "step": 8450
+    },
+    {
+      "epoch": 1.6482771970576848,
+      "grad_norm": 4.0184407234191895,
+      "learning_rate": 1.5683907568748753e-05,
+      "loss": 0.3379,
+      "step": 8515
+    },
+    {
+      "epoch": 1.6608594657375146,
+      "grad_norm": 2.0760459899902344,
+      "learning_rate": 1.5239440489856724e-05,
+      "loss": 0.388,
+      "step": 8580
+    },
+    {
+      "epoch": 1.6734417344173442,
+      "grad_norm": 0.24721883237361908,
+      "learning_rate": 1.4794762981783464e-05,
+      "loss": 0.3323,
+      "step": 8645
+    },
+    {
+      "epoch": 1.6860240030971738,
+      "grad_norm": 0.26158830523490906,
+      "learning_rate": 1.435026584361225e-05,
+      "loss": 0.4388,
+      "step": 8710
+    },
+    {
+      "epoch": 1.6986062717770034,
+      "grad_norm": 8.231218338012695,
+      "learning_rate": 1.3906339715910606e-05,
+      "loss": 0.3223,
+      "step": 8775
+    },
+    {
+      "epoch": 1.711188540456833,
+      "grad_norm": 21.496234893798828,
+      "learning_rate": 1.346337473742089e-05,
+      "loss": 0.3775,
+      "step": 8840
+    },
+    {
+      "epoch": 1.7237708091366628,
+      "grad_norm": 1.0337002277374268,
+      "learning_rate": 1.3021760202191877e-05,
+      "loss": 0.4348,
+      "step": 8905
+    },
+    {
+      "epoch": 1.7363530778164924,
+      "grad_norm": 0.8793113231658936,
+      "learning_rate": 1.2581884217452726e-05,
+      "loss": 0.3758,
+      "step": 8970
+    },
+    {
+      "epoch": 1.7489353464963222,
+      "grad_norm": 9.16724967956543,
+      "learning_rate": 1.214413336252998e-05,
+      "loss": 0.2605,
+      "step": 9035
+    },
+    {
+      "epoch": 1.7506775067750677,
+      "eval_StS-test_pearson_cosine": 0.8777802459074545,
+      "eval_StS-test_pearson_dot": 0.7844466254188373,
+      "eval_StS-test_pearson_euclidean": 0.8614201032938403,
+      "eval_StS-test_pearson_manhattan": 0.8614174583863513,
+      "eval_StS-test_pearson_max": 0.8777802459074545,
+      "eval_StS-test_spearman_cosine": 0.892518681200198,
+      "eval_StS-test_spearman_dot": 0.7893342074296325,
+      "eval_StS-test_spearman_euclidean": 0.8654693632229482,
+      "eval_StS-test_spearman_manhattan": 0.8680543609352334,
+      "eval_StS-test_spearman_max": 0.892518681200198,
+      "eval_Vitaminc-test_cosine_accuracy": 0.5815789473684211,
+      "eval_Vitaminc-test_cosine_accuracy_threshold": 0.7277665138244629,
+      "eval_Vitaminc-test_cosine_ap": 0.5656681584201257,
+      "eval_Vitaminc-test_cosine_f1": 0.6703499079189686,
+      "eval_Vitaminc-test_cosine_f1_threshold": 0.4397270381450653,
+      "eval_Vitaminc-test_cosine_precision": 0.5083798882681564,
+      "eval_Vitaminc-test_cosine_recall": 0.9837837837837838,
+      "eval_Vitaminc-test_dot_accuracy": 0.5657894736842105,
+      "eval_Vitaminc-test_dot_accuracy_threshold": 20.005393981933594,
+      "eval_Vitaminc-test_dot_ap": 0.5480790470255866,
+      "eval_Vitaminc-test_dot_f1": 0.6666666666666667,
+      "eval_Vitaminc-test_dot_f1_threshold": 12.140420913696289,
+      "eval_Vitaminc-test_dot_precision": 0.5100286532951289,
+      "eval_Vitaminc-test_dot_recall": 0.9621621621621622,
+      "eval_Vitaminc-test_euclidean_accuracy": 0.5657894736842105,
+      "eval_Vitaminc-test_euclidean_accuracy_threshold": 3.4636647701263428,
+      "eval_Vitaminc-test_euclidean_ap": 0.5552496987243565,
+      "eval_Vitaminc-test_euclidean_f1": 0.663023679417122,
+      "eval_Vitaminc-test_euclidean_f1_threshold": 6.34542989730835,
+      "eval_Vitaminc-test_euclidean_precision": 0.5,
+      "eval_Vitaminc-test_euclidean_recall": 0.9837837837837838,
+      "eval_Vitaminc-test_manhattan_accuracy": 0.5684210526315789,
+      "eval_Vitaminc-test_manhattan_accuracy_threshold": 75.46658325195312,
+      "eval_Vitaminc-test_manhattan_ap": 0.5568456847059792,
+      "eval_Vitaminc-test_manhattan_f1": 0.6618962432915921,
+      "eval_Vitaminc-test_manhattan_f1_threshold": 144.27662658691406,
+      "eval_Vitaminc-test_manhattan_precision": 0.4946524064171123,
+      "eval_Vitaminc-test_manhattan_recall": 1.0,
+      "eval_Vitaminc-test_max_accuracy": 0.5815789473684211,
+      "eval_Vitaminc-test_max_accuracy_threshold": 75.46658325195312,
+      "eval_Vitaminc-test_max_ap": 0.5656681584201257,
+      "eval_Vitaminc-test_max_f1": 0.6703499079189686,
+      "eval_Vitaminc-test_max_f1_threshold": 144.27662658691406,
+      "eval_Vitaminc-test_max_precision": 0.5100286532951289,
+      "eval_Vitaminc-test_max_recall": 1.0,
+      "eval_mrpc-test_cosine_accuracy": 0.7447368421052631,
+      "eval_mrpc-test_cosine_accuracy_threshold": 0.719977617263794,
+      "eval_mrpc-test_cosine_ap": 0.850964159121461,
+      "eval_mrpc-test_cosine_f1": 0.8295254833040421,
+      "eval_mrpc-test_cosine_f1_threshold": 0.700160562992096,
+      "eval_mrpc-test_cosine_precision": 0.735202492211838,
+      "eval_mrpc-test_cosine_recall": 0.9516129032258065,
+      "eval_mrpc-test_dot_accuracy": 0.7026315789473684,
+      "eval_mrpc-test_dot_accuracy_threshold": 12.873937606811523,
+      "eval_mrpc-test_dot_ap": 0.7712669296028235,
+      "eval_mrpc-test_dot_f1": 0.8045977011494253,
+      "eval_mrpc-test_dot_f1_threshold": 9.337194442749023,
+      "eval_mrpc-test_dot_precision": 0.6786703601108033,
+      "eval_mrpc-test_dot_recall": 0.9879032258064516,
+      "eval_mrpc-test_euclidean_accuracy": 0.7368421052631579,
+      "eval_mrpc-test_euclidean_accuracy_threshold": 3.320279836654663,
+      "eval_mrpc-test_euclidean_ap": 0.8234369135440149,
+      "eval_mrpc-test_euclidean_f1": 0.8226950354609929,
+      "eval_mrpc-test_euclidean_f1_threshold": 3.4916229248046875,
+      "eval_mrpc-test_euclidean_precision": 0.7341772151898734,
+      "eval_mrpc-test_euclidean_recall": 0.9354838709677419,
+      "eval_mrpc-test_manhattan_accuracy": 0.7394736842105263,
+      "eval_mrpc-test_manhattan_accuracy_threshold": 67.87750244140625,
+      "eval_mrpc-test_manhattan_ap": 0.8215377503250268,
+      "eval_mrpc-test_manhattan_f1": 0.8235294117647058,
+      "eval_mrpc-test_manhattan_f1_threshold": 69.14122009277344,
+      "eval_mrpc-test_manhattan_precision": 0.7380191693290735,
+      "eval_mrpc-test_manhattan_recall": 0.9314516129032258,
+      "eval_mrpc-test_max_accuracy": 0.7447368421052631,
+      "eval_mrpc-test_max_accuracy_threshold": 67.87750244140625,
+      "eval_mrpc-test_max_ap": 0.850964159121461,
+      "eval_mrpc-test_max_f1": 0.8295254833040421,
+      "eval_mrpc-test_max_f1_threshold": 69.14122009277344,
+      "eval_mrpc-test_max_precision": 0.7380191693290735,
+      "eval_mrpc-test_max_recall": 0.9879032258064516,
+      "eval_nli-pairs_loss": 0.7929299473762512,
+      "eval_nli-pairs_runtime": 3.0029,
+      "eval_nli-pairs_samples_per_second": 53.282,
+      "eval_nli-pairs_steps_per_second": 1.665,
+      "eval_sequential_score": 0.5656681584201257,
+      "step": 9044
+    },
+    {
+      "epoch": 1.7506775067750677,
+      "eval_vitaminc-pairs_loss": 5.716288089752197,
+      "eval_vitaminc-pairs_runtime": 1.5553,
+      "eval_vitaminc-pairs_samples_per_second": 85.514,
+      "eval_vitaminc-pairs_steps_per_second": 3.215,
+      "step": 9044
+    },
+    {
+      "epoch": 1.7506775067750677,
+      "eval_qnli-contrastive_loss": 0.138463094830513,
+      "eval_qnli-contrastive_runtime": 0.5207,
+      "eval_qnli-contrastive_samples_per_second": 307.265,
+      "eval_qnli-contrastive_steps_per_second": 9.602,
+      "step": 9044
+    },
+    {
+      "epoch": 1.7506775067750677,
+      "eval_scitail-pairs-qa_loss": 0.0497589036822319,
+      "eval_scitail-pairs-qa_runtime": 1.2114,
+      "eval_scitail-pairs-qa_samples_per_second": 132.076,
+      "eval_scitail-pairs-qa_steps_per_second": 4.127,
+      "step": 9044
+    },
+    {
+      "epoch": 1.7506775067750677,
+      "eval_scitail-pairs-pos_loss": 0.28213563561439514,
+      "eval_scitail-pairs-pos_runtime": 2.5132,
+      "eval_scitail-pairs-pos_samples_per_second": 63.665,
+      "eval_scitail-pairs-pos_steps_per_second": 1.99,
+      "step": 9044
+    },
+    {
+      "epoch": 1.7506775067750677,
+      "eval_xsum-pairs_loss": 0.2819857597351074,
+      "eval_xsum-pairs_runtime": 1.404,
+      "eval_xsum-pairs_samples_per_second": 113.962,
+      "eval_xsum-pairs_steps_per_second": 3.561,
+      "step": 9044
+    },
+    {
+      "epoch": 1.7506775067750677,
+      "eval_compression-pairs_loss": 0.06524467468261719,
+      "eval_compression-pairs_runtime": 0.4016,
+      "eval_compression-pairs_samples_per_second": 398.384,
+      "eval_compression-pairs_steps_per_second": 12.449,
+      "step": 9044
+    },
+    {
+      "epoch": 1.7506775067750677,
+      "eval_sciq_pairs_loss": 0.25192952156066895,
+      "eval_sciq_pairs_runtime": 7.9683,
+      "eval_sciq_pairs_samples_per_second": 20.08,
+      "eval_sciq_pairs_steps_per_second": 0.627,
+      "step": 9044
+    },
+    {
+      "epoch": 1.7506775067750677,
+      "eval_qasc_pairs_loss": 0.15719862282276154,
+      "eval_qasc_pairs_runtime": 1.4347,
+      "eval_qasc_pairs_samples_per_second": 111.525,
+      "eval_qasc_pairs_steps_per_second": 3.485,
+      "step": 9044
+    },
+    {
+      "epoch": 1.7506775067750677,
+      "eval_qasc_facts_sym_loss": 0.11949674040079117,
+      "eval_qasc_facts_sym_runtime": 0.3223,
+      "eval_qasc_facts_sym_samples_per_second": 496.415,
+      "eval_qasc_facts_sym_steps_per_second": 15.513,
+      "step": 9044
+    },
+    {
+      "epoch": 1.7506775067750677,
+      "eval_openbookqa_pairs_loss": 1.6055030822753906,
+      "eval_openbookqa_pairs_runtime": 1.2319,
+      "eval_openbookqa_pairs_samples_per_second": 129.884,
+      "eval_openbookqa_pairs_steps_per_second": 4.059,
+      "step": 9044
+    },
+    {
+      "epoch": 1.7506775067750677,
+      "eval_msmarco_pairs_loss": 0.46120986342430115,
+      "eval_msmarco_pairs_runtime": 3.1151,
+      "eval_msmarco_pairs_samples_per_second": 51.364,
+      "eval_msmarco_pairs_steps_per_second": 1.605,
+      "step": 9044
+    },
+    {
+      "epoch": 1.7506775067750677,
+      "eval_nq_pairs_loss": 0.38706856966018677,
+      "eval_nq_pairs_runtime": 7.6899,
+      "eval_nq_pairs_samples_per_second": 20.807,
+      "eval_nq_pairs_steps_per_second": 0.65,
+      "step": 9044
+    },
+    {
+      "epoch": 1.7506775067750677,
+      "eval_trivia_pairs_loss": 0.6729233860969543,
+      "eval_trivia_pairs_runtime": 10.192,
+      "eval_trivia_pairs_samples_per_second": 15.699,
+      "eval_trivia_pairs_steps_per_second": 0.491,
+      "step": 9044
+    },
+    {
+      "epoch": 1.7506775067750677,
+      "eval_quora_pairs_loss": 0.18244178593158722,
+      "eval_quora_pairs_runtime": 3.9178,
+      "eval_quora_pairs_samples_per_second": 172.291,
+      "eval_quora_pairs_steps_per_second": 5.615,
+      "step": 9044
+    },
+    {
+      "epoch": 1.7506775067750677,
+      "eval_gooaq_pairs_loss": 0.4615425169467926,
+      "eval_gooaq_pairs_runtime": 2.2929,
+      "eval_gooaq_pairs_samples_per_second": 69.779,
+      "eval_gooaq_pairs_steps_per_second": 2.181,
+      "step": 9044
+    },
+    {
+      "epoch": 1.7506775067750677,
+      "eval_mrpc_pairs_loss": 0.03375188261270523,
+      "eval_mrpc_pairs_runtime": 0.3787,
+      "eval_mrpc_pairs_samples_per_second": 422.446,
+      "eval_mrpc_pairs_steps_per_second": 13.201,
+      "step": 9044
+    },
+    {
+      "epoch": 1.7615176151761518,
+      "grad_norm": 17.31819725036621,
+      "learning_rate": 1.1708892349107267e-05,
+      "loss": 0.325,
+      "step": 9100
+    },
+    {
+      "epoch": 1.7740998838559814,
+      "grad_norm": 1.4742263555526733,
+      "learning_rate": 1.1276543683126467e-05,
+      "loss": 0.2905,
+      "step": 9165
+    },
+    {
+      "epoch": 1.786682152535811,
+      "grad_norm": 17.93797492980957,
+      "learning_rate": 1.0847467328627261e-05,
+      "loss": 0.3873,
+      "step": 9230
+    },
+    {
+      "epoch": 1.7992644212156406,
+      "grad_norm": 10.806224822998047,
+      "learning_rate": 1.042204037382065e-05,
+      "loss": 0.2658,
+      "step": 9295
+    },
+    {
+      "epoch": 1.8118466898954704,
+      "grad_norm": 11.624892234802246,
+      "learning_rate": 1.0000636699689883e-05,
+      "loss": 0.3357,
+      "step": 9360
+    },
+    {
+      "epoch": 1.8244289585753002,
+      "grad_norm": 0.8121838569641113,
+      "learning_rate": 9.583626651409943e-06,
+      "loss": 0.2736,
+      "step": 9425
+    },
+    {
+      "epoch": 1.8370112272551298,
+      "grad_norm": 4.140751361846924,
+      "learning_rate": 9.171376712874502e-06,
+      "loss": 0.2758,
+      "step": 9490
+    },
+    {
+      "epoch": 1.8495934959349594,
+      "grad_norm": 0.2148134708404541,
+      "learning_rate": 8.764249184616278e-06,
+      "loss": 0.2423,
+      "step": 9555
+    },
+    {
+      "epoch": 1.862175764614789,
+      "grad_norm": 15.606979370117188,
+      "learning_rate": 8.362601865403897e-06,
+      "loss": 0.2992,
+      "step": 9620
+    },
+    {
+      "epoch": 1.8747580332946185,
+      "grad_norm": 0.496015340089798,
+      "learning_rate": 7.966787737795051e-06,
+      "loss": 0.2551,
+      "step": 9685
+    },
+    {
+      "epoch": 1.8873403019744484,
+      "grad_norm": 8.67402458190918,
+      "learning_rate": 7.577154657922359e-06,
+      "loss": 0.3565,
+      "step": 9750
+    },
+    {
+      "epoch": 1.899922570654278,
+      "grad_norm": 1.7114161252975464,
+      "learning_rate": 7.199887940715319e-06,
+      "loss": 0.2705,
+      "step": 9815
+    },
+    {
+      "epoch": 1.9125048393341078,
+      "grad_norm": 13.371696472167969,
+      "learning_rate": 6.823530434028392e-06,
+      "loss": 0.3363,
+      "step": 9880
+    },
+    {
+      "epoch": 1.9250871080139373,
+      "grad_norm": 1.002062201499939,
+      "learning_rate": 6.4543587119895626e-06,
+      "loss": 0.2787,
+      "step": 9945
+    },
+    {
+      "epoch": 1.937669376693767,
+      "grad_norm": 0.24811969697475433,
+      "learning_rate": 6.092697216397478e-06,
+      "loss": 0.273,
+      "step": 10010
+    },
+    {
+      "epoch": 1.9502516453735965,
+      "grad_norm": 6.3729352951049805,
+      "learning_rate": 5.7388637887853255e-06,
+      "loss": 0.264,
+      "step": 10075
+    },
+    {
+      "epoch": 1.962833914053426,
+      "grad_norm": 2.160889148712158,
+      "learning_rate": 5.393169391089864e-06,
+      "loss": 0.3532,
+      "step": 10140
+    },
+    {
+      "epoch": 1.975416182733256,
+      "grad_norm": 0.7218418121337891,
+      "learning_rate": 5.055917832366561e-06,
+      "loss": 0.221,
+      "step": 10205
+    },
+    {
+      "epoch": 1.9879984514130855,
+      "grad_norm": 0.2302940934896469,
+      "learning_rate": 4.727405501790925e-06,
+      "loss": 0.2625,
+      "step": 10270
+    },
+    {
+      "epoch": 2.0005807200929153,
+      "grad_norm": 8.200091361999512,
+      "learning_rate": 4.407921108180699e-06,
+      "loss": 0.2808,
+      "step": 10335
+    },
+    {
+      "epoch": 2.00077429345722,
+      "eval_StS-test_pearson_cosine": 0.8760335462570302,
+      "eval_StS-test_pearson_dot": 0.7858533850659265,
+      "eval_StS-test_pearson_euclidean": 0.8558007686663589,
+      "eval_StS-test_pearson_manhattan": 0.8557533448194315,
+      "eval_StS-test_pearson_max": 0.8760335462570302,
+      "eval_StS-test_spearman_cosine": 0.8899595181393952,
+      "eval_StS-test_spearman_dot": 0.7894333862187969,
+      "eval_StS-test_spearman_euclidean": 0.8594150783294261,
+      "eval_StS-test_spearman_manhattan": 0.8616158403353453,
+      "eval_StS-test_spearman_max": 0.8899595181393952,
+      "eval_Vitaminc-test_cosine_accuracy": 0.5789473684210527,
+      "eval_Vitaminc-test_cosine_accuracy_threshold": 0.7607381343841553,
+      "eval_Vitaminc-test_cosine_ap": 0.562686091779982,
+      "eval_Vitaminc-test_cosine_f1": 0.673076923076923,
+      "eval_Vitaminc-test_cosine_f1_threshold": 0.5096132159233093,
+      "eval_Vitaminc-test_cosine_precision": 0.5223880597014925,
+      "eval_Vitaminc-test_cosine_recall": 0.9459459459459459,
+      "eval_Vitaminc-test_dot_accuracy": 0.5710526315789474,
+      "eval_Vitaminc-test_dot_accuracy_threshold": 14.656951904296875,
+      "eval_Vitaminc-test_dot_ap": 0.5483043888699137,
+      "eval_Vitaminc-test_dot_f1": 0.6691176470588236,
+      "eval_Vitaminc-test_dot_f1_threshold": 9.500975608825684,
+      "eval_Vitaminc-test_dot_precision": 0.5069637883008357,
+      "eval_Vitaminc-test_dot_recall": 0.9837837837837838,
+      "eval_Vitaminc-test_euclidean_accuracy": 0.5710526315789474,
+      "eval_Vitaminc-test_euclidean_accuracy_threshold": 3.3688926696777344,
+      "eval_Vitaminc-test_euclidean_ap": 0.552131519622326,
+      "eval_Vitaminc-test_euclidean_f1": 0.6666666666666666,
+      "eval_Vitaminc-test_euclidean_f1_threshold": 5.8364458084106445,
+      "eval_Vitaminc-test_euclidean_precision": 0.5013623978201635,
+      "eval_Vitaminc-test_euclidean_recall": 0.9945945945945946,
+      "eval_Vitaminc-test_manhattan_accuracy": 0.5657894736842105,
+      "eval_Vitaminc-test_manhattan_accuracy_threshold": 67.55563354492188,
+      "eval_Vitaminc-test_manhattan_ap": 0.5549339934603554,
+      "eval_Vitaminc-test_manhattan_f1": 0.6642066420664207,
+      "eval_Vitaminc-test_manhattan_f1_threshold": 115.51136779785156,
+      "eval_Vitaminc-test_manhattan_precision": 0.5042016806722689,
+      "eval_Vitaminc-test_manhattan_recall": 0.972972972972973,
+      "eval_Vitaminc-test_max_accuracy": 0.5789473684210527,
+      "eval_Vitaminc-test_max_accuracy_threshold": 67.55563354492188,
+      "eval_Vitaminc-test_max_ap": 0.562686091779982,
+      "eval_Vitaminc-test_max_f1": 0.673076923076923,
+      "eval_Vitaminc-test_max_f1_threshold": 115.51136779785156,
+      "eval_Vitaminc-test_max_precision": 0.5223880597014925,
+      "eval_Vitaminc-test_max_recall": 0.9945945945945946,
+      "eval_mrpc-test_cosine_accuracy": 0.7394736842105263,
+      "eval_mrpc-test_cosine_accuracy_threshold": 0.6881712079048157,
+      "eval_mrpc-test_cosine_ap": 0.8489212567353354,
+      "eval_mrpc-test_cosine_f1": 0.8275862068965517,
+      "eval_mrpc-test_cosine_f1_threshold": 0.6637799143791199,
+      "eval_mrpc-test_cosine_precision": 0.7228915662650602,
+      "eval_mrpc-test_cosine_recall": 0.967741935483871,
+      "eval_mrpc-test_dot_accuracy": 0.6894736842105263,
+      "eval_mrpc-test_dot_accuracy_threshold": 11.309194564819336,
+      "eval_mrpc-test_dot_ap": 0.7692877575360217,
+      "eval_mrpc-test_dot_f1": 0.8065573770491804,
+      "eval_mrpc-test_dot_f1_threshold": 7.328646659851074,
+      "eval_mrpc-test_dot_precision": 0.6795580110497238,
+      "eval_mrpc-test_dot_recall": 0.9919354838709677,
+      "eval_mrpc-test_euclidean_accuracy": 0.7421052631578947,
+      "eval_mrpc-test_euclidean_accuracy_threshold": 3.1678833961486816,
+      "eval_mrpc-test_euclidean_ap": 0.8238273284368561,
+      "eval_mrpc-test_euclidean_f1": 0.8237410071942447,
+      "eval_mrpc-test_euclidean_f1_threshold": 3.2550790309906006,
+      "eval_mrpc-test_euclidean_precision": 0.7435064935064936,
+      "eval_mrpc-test_euclidean_recall": 0.9233870967741935,
+      "eval_mrpc-test_manhattan_accuracy": 0.7421052631578947,
+      "eval_mrpc-test_manhattan_accuracy_threshold": 64.18893432617188,
+      "eval_mrpc-test_manhattan_ap": 0.8221281490148798,
+      "eval_mrpc-test_manhattan_f1": 0.8243727598566308,
+      "eval_mrpc-test_manhattan_f1_threshold": 64.66268157958984,
+      "eval_mrpc-test_manhattan_precision": 0.7419354838709677,
+      "eval_mrpc-test_manhattan_recall": 0.9274193548387096,
+      "eval_mrpc-test_max_accuracy": 0.7421052631578947,
+      "eval_mrpc-test_max_accuracy_threshold": 64.18893432617188,
+      "eval_mrpc-test_max_ap": 0.8489212567353354,
+      "eval_mrpc-test_max_f1": 0.8275862068965517,
+      "eval_mrpc-test_max_f1_threshold": 64.66268157958984,
+      "eval_mrpc-test_max_precision": 0.7435064935064936,
+      "eval_mrpc-test_max_recall": 0.9919354838709677,
+      "eval_nli-pairs_loss": 0.74864262342453,
+      "eval_nli-pairs_runtime": 3.0791,
+      "eval_nli-pairs_samples_per_second": 51.964,
+      "eval_nli-pairs_steps_per_second": 1.624,
+      "eval_sequential_score": 0.562686091779982,
+      "step": 10336
+    },
+    {
+      "epoch": 2.00077429345722,
+      "eval_vitaminc-pairs_loss": 5.761841773986816,
+      "eval_vitaminc-pairs_runtime": 1.6287,
+      "eval_vitaminc-pairs_samples_per_second": 81.658,
+      "eval_vitaminc-pairs_steps_per_second": 3.07,
+      "step": 10336
+    },
+    {
+      "epoch": 2.00077429345722,
+      "eval_qnli-contrastive_loss": 0.11774346977472305,
+      "eval_qnli-contrastive_runtime": 0.5414,
+      "eval_qnli-contrastive_samples_per_second": 295.503,
+      "eval_qnli-contrastive_steps_per_second": 9.234,
+      "step": 10336
+    },
+    {
+      "epoch": 2.00077429345722,
+      "eval_scitail-pairs-qa_loss": 0.044219307601451874,
+      "eval_scitail-pairs-qa_runtime": 1.272,
+      "eval_scitail-pairs-qa_samples_per_second": 125.783,
+      "eval_scitail-pairs-qa_steps_per_second": 3.931,
+      "step": 10336
+    },
+    {
+      "epoch": 2.00077429345722,
+      "eval_scitail-pairs-pos_loss": 0.30131369829177856,
+      "eval_scitail-pairs-pos_runtime": 2.7269,
+      "eval_scitail-pairs-pos_samples_per_second": 58.675,
+      "eval_scitail-pairs-pos_steps_per_second": 1.834,
+      "step": 10336
+    },
+    {
+      "epoch": 2.00077429345722,
+      "eval_xsum-pairs_loss": 0.2668481469154358,
+      "eval_xsum-pairs_runtime": 1.4505,
+      "eval_xsum-pairs_samples_per_second": 110.31,
+      "eval_xsum-pairs_steps_per_second": 3.447,
+      "step": 10336
+    },
+    {
+      "epoch": 2.00077429345722,
+      "eval_compression-pairs_loss": 0.05984542891383171,
+      "eval_compression-pairs_runtime": 0.4143,
+      "eval_compression-pairs_samples_per_second": 386.167,
+      "eval_compression-pairs_steps_per_second": 12.068,
+      "step": 10336
+    },
+    {
+      "epoch": 2.00077429345722,
+      "eval_sciq_pairs_loss": 0.21106968820095062,
+      "eval_sciq_pairs_runtime": 8.1848,
+      "eval_sciq_pairs_samples_per_second": 19.548,
+      "eval_sciq_pairs_steps_per_second": 0.611,
+      "step": 10336
+    },
+    {
+      "epoch": 2.00077429345722,
+      "eval_qasc_pairs_loss": 0.14702823758125305,
+      "eval_qasc_pairs_runtime": 1.4839,
+      "eval_qasc_pairs_samples_per_second": 107.825,
+      "eval_qasc_pairs_steps_per_second": 3.37,
+      "step": 10336
+    },
+    {
+      "epoch": 2.00077429345722,
+      "eval_qasc_facts_sym_loss": 0.11709441989660263,
+      "eval_qasc_facts_sym_runtime": 0.3298,
+      "eval_qasc_facts_sym_samples_per_second": 485.098,
+      "eval_qasc_facts_sym_steps_per_second": 15.159,
+      "step": 10336
+    },
+    {
+      "epoch": 2.00077429345722,
+      "eval_openbookqa_pairs_loss": 1.611570119857788,
+      "eval_openbookqa_pairs_runtime": 1.3171,
+      "eval_openbookqa_pairs_samples_per_second": 121.475,
+      "eval_openbookqa_pairs_steps_per_second": 3.796,
+      "step": 10336
+    },
+    {
+      "epoch": 2.00077429345722,
+      "eval_msmarco_pairs_loss": 0.45828428864479065,
+      "eval_msmarco_pairs_runtime": 3.2279,
+      "eval_msmarco_pairs_samples_per_second": 49.567,
+      "eval_msmarco_pairs_steps_per_second": 1.549,
+      "step": 10336
+    },
+    {
+      "epoch": 2.00077429345722,
+      "eval_nq_pairs_loss": 0.361095130443573,
+      "eval_nq_pairs_runtime": 7.8042,
+      "eval_nq_pairs_samples_per_second": 20.502,
+      "eval_nq_pairs_steps_per_second": 0.641,
+      "step": 10336
+    },
+    {
+      "epoch": 2.00077429345722,
+      "eval_trivia_pairs_loss": 0.5992080569267273,
+      "eval_trivia_pairs_runtime": 10.3232,
+      "eval_trivia_pairs_samples_per_second": 15.499,
+      "eval_trivia_pairs_steps_per_second": 0.484,
+      "step": 10336
+    },
+    {
+      "epoch": 2.00077429345722,
+      "eval_quora_pairs_loss": 0.1803085058927536,
+      "eval_quora_pairs_runtime": 4.3744,
+      "eval_quora_pairs_samples_per_second": 154.308,
+      "eval_quora_pairs_steps_per_second": 5.029,
+      "step": 10336
+    },
+    {
+      "epoch": 2.00077429345722,
+      "eval_gooaq_pairs_loss": 0.44401103258132935,
+      "eval_gooaq_pairs_runtime": 2.2868,
+      "eval_gooaq_pairs_samples_per_second": 69.968,
+      "eval_gooaq_pairs_steps_per_second": 2.187,
+      "step": 10336
+    },
+    {
+      "epoch": 2.00077429345722,
+      "eval_mrpc_pairs_loss": 0.03005302883684635,
+      "eval_mrpc_pairs_runtime": 0.3764,
+      "eval_mrpc_pairs_samples_per_second": 425.052,
+      "eval_mrpc_pairs_steps_per_second": 13.283,
+      "step": 10336
+    },
+    {
+      "epoch": 2.013162988772745,
+      "grad_norm": 0.748429536819458,
+      "learning_rate": 4.097745426267871e-06,
+      "loss": 0.2003,
+      "step": 10400
+    },
+    {
+      "epoch": 2.0257452574525745,
+      "grad_norm": 0.1598404049873352,
+      "learning_rate": 3.797151049943393e-06,
+      "loss": 0.216,
+      "step": 10465
+    },
+    {
+      "epoch": 2.038327526132404,
+      "grad_norm": 1.4139065742492676,
+      "learning_rate": 3.5064021526915962e-06,
+      "loss": 0.2704,
+      "step": 10530
+    },
+    {
+      "epoch": 2.0509097948122337,
+      "grad_norm": 10.644986152648926,
+      "learning_rate": 3.225754255424692e-06,
+      "loss": 0.2788,
+      "step": 10595
+    },
+    {
+      "epoch": 2.0634920634920633,
+      "grad_norm": 0.1797364354133606,
+      "learning_rate": 2.955454001921588e-06,
+      "loss": 0.264,
+      "step": 10660
+    },
+    {
+      "epoch": 2.0760743321718933,
+      "grad_norm": 8.84312629699707,
+      "learning_rate": 2.6957389420681128e-06,
+      "loss": 0.2625,
+      "step": 10725
+    },
+    {
+      "epoch": 2.088656600851723,
+      "grad_norm": 2.1348655223846436,
+      "learning_rate": 2.446837323089423e-06,
+      "loss": 0.3213,
+      "step": 10790
+    },
+    {
+      "epoch": 2.1012388695315525,
+      "grad_norm": 0.9930692315101624,
+      "learning_rate": 2.208967888957853e-06,
+      "loss": 0.2799,
+      "step": 10855
+    },
+    {
+      "epoch": 2.113821138211382,
+      "grad_norm": 3.232008934020996,
+      "learning_rate": 1.982339688152608e-06,
+      "loss": 0.2719,
+      "step": 10920
+    },
+    {
+      "epoch": 2.1264034068912117,
+      "grad_norm": 1.367380142211914,
+      "learning_rate": 1.7671518899402124e-06,
+      "loss": 0.2086,
+      "step": 10985
+    },
+    {
+      "epoch": 2.1389856755710412,
+      "grad_norm": 2.2108681201934814,
+      "learning_rate": 1.5635936093371479e-06,
+      "loss": 0.208,
+      "step": 11050
+    },
+    {
+      "epoch": 2.1515679442508713,
+      "grad_norm": 0.841842770576477,
+      "learning_rate": 1.3718437409086144e-06,
+      "loss": 0.3249,
+      "step": 11115
+    },
+    {
+      "epoch": 2.164150212930701,
+      "grad_norm": 0.215606689453125,
+      "learning_rate": 1.1920708015492988e-06,
+      "loss": 0.2545,
+      "step": 11180
+    },
+    {
+      "epoch": 2.1767324816105305,
+      "grad_norm": 0.4588742256164551,
+      "learning_rate": 1.0244327823845117e-06,
+      "loss": 0.274,
+      "step": 11245
+    },
+    {
+      "epoch": 2.18931475029036,
+      "grad_norm": 7.016474723815918,
+      "learning_rate": 8.690770099217205e-07,
+      "loss": 0.216,
+      "step": 11310
+    },
+    {
+      "epoch": 2.2018970189701896,
+      "grad_norm": 2.9576404094696045,
+      "learning_rate": 7.261400165745497e-07,
+      "loss": 0.2467,
+      "step": 11375
+    },
+    {
+      "epoch": 2.2144792876500192,
+      "grad_norm": 5.329019546508789,
+      "learning_rate": 5.957474206730256e-07,
+      "loss": 0.2578,
+      "step": 11440
+    },
+    {
+      "epoch": 2.227061556329849,
+      "grad_norm": 0.16189192235469818,
+      "learning_rate": 4.780138160655507e-07,
+      "loss": 0.2608,
+      "step": 11505
+    },
+    {
+      "epoch": 2.239643825009679,
+      "grad_norm": 8.813294410705566,
+      "learning_rate": 3.7304267140955304e-07,
+      "loss": 0.2476,
+      "step": 11570
+    },
+    {
+      "epoch": 2.250871080139373,
+      "eval_StS-test_pearson_cosine": 0.8778872221291931,
+      "eval_StS-test_pearson_dot": 0.7893324466927049,
+      "eval_StS-test_pearson_euclidean": 0.8590444533278028,
+      "eval_StS-test_pearson_manhattan": 0.8591331390817218,
+      "eval_StS-test_pearson_max": 0.8778872221291931,
+      "eval_StS-test_spearman_cosine": 0.8919884140316408,
+      "eval_StS-test_spearman_dot": 0.7902921694428535,
+      "eval_StS-test_spearman_euclidean": 0.8625352693090679,
+      "eval_StS-test_spearman_manhattan": 0.8649523417867188,
+      "eval_StS-test_spearman_max": 0.8919884140316408,
+      "eval_Vitaminc-test_cosine_accuracy": 0.5789473684210527,
+      "eval_Vitaminc-test_cosine_accuracy_threshold": 0.6914482712745667,
+      "eval_Vitaminc-test_cosine_ap": 0.5637651212229153,
+      "eval_Vitaminc-test_cosine_f1": 0.6741573033707865,
+      "eval_Vitaminc-test_cosine_f1_threshold": 0.4676659405231476,
+      "eval_Vitaminc-test_cosine_precision": 0.5157593123209169,
+      "eval_Vitaminc-test_cosine_recall": 0.972972972972973,
+      "eval_Vitaminc-test_dot_accuracy": 0.5710526315789474,
+      "eval_Vitaminc-test_dot_accuracy_threshold": 14.829851150512695,
+      "eval_Vitaminc-test_dot_ap": 0.5497462011082384,
+      "eval_Vitaminc-test_dot_f1": 0.667953667953668,
+      "eval_Vitaminc-test_dot_f1_threshold": 11.509256362915039,
+      "eval_Vitaminc-test_dot_precision": 0.5195195195195195,
+      "eval_Vitaminc-test_dot_recall": 0.9351351351351351,
+      "eval_Vitaminc-test_euclidean_accuracy": 0.5710526315789474,
+      "eval_Vitaminc-test_euclidean_accuracy_threshold": 3.4082818031311035,
+      "eval_Vitaminc-test_euclidean_ap": 0.5529767634436812,
+      "eval_Vitaminc-test_euclidean_f1": 0.6654611211573237,
+      "eval_Vitaminc-test_euclidean_f1_threshold": 6.036683082580566,
+      "eval_Vitaminc-test_euclidean_precision": 0.5,
+      "eval_Vitaminc-test_euclidean_recall": 0.9945945945945946,
+      "eval_Vitaminc-test_manhattan_accuracy": 0.5710526315789474,
+      "eval_Vitaminc-test_manhattan_accuracy_threshold": 68.36351013183594,
+      "eval_Vitaminc-test_manhattan_ap": 0.5551197831302017,
+      "eval_Vitaminc-test_manhattan_f1": 0.6642728904847397,
+      "eval_Vitaminc-test_manhattan_f1_threshold": 127.92561340332031,
+      "eval_Vitaminc-test_manhattan_precision": 0.49731182795698925,
+      "eval_Vitaminc-test_manhattan_recall": 1.0,
+      "eval_Vitaminc-test_max_accuracy": 0.5789473684210527,
+      "eval_Vitaminc-test_max_accuracy_threshold": 68.36351013183594,
+      "eval_Vitaminc-test_max_ap": 0.5637651212229153,
+      "eval_Vitaminc-test_max_f1": 0.6741573033707865,
+      "eval_Vitaminc-test_max_f1_threshold": 127.92561340332031,
+      "eval_Vitaminc-test_max_precision": 0.5195195195195195,
+      "eval_Vitaminc-test_max_recall": 1.0,
+      "eval_mrpc-test_cosine_accuracy": 0.7394736842105263,
+      "eval_mrpc-test_cosine_accuracy_threshold": 0.694667637348175,
+      "eval_mrpc-test_cosine_ap": 0.8501212805448715,
+      "eval_mrpc-test_cosine_f1": 0.8253968253968255,
+      "eval_mrpc-test_cosine_f1_threshold": 0.694667637348175,
+      "eval_mrpc-test_cosine_precision": 0.7335423197492164,
+      "eval_mrpc-test_cosine_recall": 0.9435483870967742,
+      "eval_mrpc-test_dot_accuracy": 0.6894736842105263,
+      "eval_mrpc-test_dot_accuracy_threshold": 11.033209800720215,
+      "eval_mrpc-test_dot_ap": 0.7670949385599481,
+      "eval_mrpc-test_dot_f1": 0.8045977011494253,
+      "eval_mrpc-test_dot_f1_threshold": 8.026628494262695,
+      "eval_mrpc-test_dot_precision": 0.6786703601108033,
+      "eval_mrpc-test_dot_recall": 0.9879032258064516,
+      "eval_mrpc-test_euclidean_accuracy": 0.7421052631578947,
+      "eval_mrpc-test_euclidean_accuracy_threshold": 3.2092814445495605,
+      "eval_mrpc-test_euclidean_ap": 0.8242245566410124,
+      "eval_mrpc-test_euclidean_f1": 0.8224637681159419,
+      "eval_mrpc-test_euclidean_f1_threshold": 3.2241902351379395,
+      "eval_mrpc-test_euclidean_precision": 0.7467105263157895,
+      "eval_mrpc-test_euclidean_recall": 0.9153225806451613,
+      "eval_mrpc-test_manhattan_accuracy": 0.7447368421052631,
+      "eval_mrpc-test_manhattan_accuracy_threshold": 65.11294555664062,
+      "eval_mrpc-test_manhattan_ap": 0.8224581711224794,
+      "eval_mrpc-test_manhattan_f1": 0.8258527827648114,
+      "eval_mrpc-test_manhattan_f1_threshold": 65.11294555664062,
+      "eval_mrpc-test_manhattan_precision": 0.7443365695792881,
+      "eval_mrpc-test_manhattan_recall": 0.9274193548387096,
+      "eval_mrpc-test_max_accuracy": 0.7447368421052631,
+      "eval_mrpc-test_max_accuracy_threshold": 65.11294555664062,
+      "eval_mrpc-test_max_ap": 0.8501212805448715,
+      "eval_mrpc-test_max_f1": 0.8258527827648114,
+      "eval_mrpc-test_max_f1_threshold": 65.11294555664062,
+      "eval_mrpc-test_max_precision": 0.7467105263157895,
+      "eval_mrpc-test_max_recall": 0.9879032258064516,
+      "eval_nli-pairs_loss": 0.7365724444389343,
+      "eval_nli-pairs_runtime": 3.0211,
+      "eval_nli-pairs_samples_per_second": 52.96,
+      "eval_nli-pairs_steps_per_second": 1.655,
+      "eval_sequential_score": 0.5637651212229153,
+      "step": 11628
+    },
+    {
+      "epoch": 2.250871080139373,
+      "eval_vitaminc-pairs_loss": 5.7284016609191895,
+      "eval_vitaminc-pairs_runtime": 1.5892,
+      "eval_vitaminc-pairs_samples_per_second": 83.689,
+      "eval_vitaminc-pairs_steps_per_second": 3.146,
+      "step": 11628
+    },
+    {
+      "epoch": 2.250871080139373,
+      "eval_qnli-contrastive_loss": 0.1260131299495697,
+      "eval_qnli-contrastive_runtime": 0.5536,
+      "eval_qnli-contrastive_samples_per_second": 289.007,
+      "eval_qnli-contrastive_steps_per_second": 9.031,
+      "step": 11628
+    },
+    {
+      "epoch": 2.250871080139373,
+      "eval_scitail-pairs-qa_loss": 0.04436067119240761,
+      "eval_scitail-pairs-qa_runtime": 1.2776,
+      "eval_scitail-pairs-qa_samples_per_second": 125.233,
+      "eval_scitail-pairs-qa_steps_per_second": 3.914,
+      "step": 11628
+    },
+    {
+      "epoch": 2.250871080139373,
+      "eval_scitail-pairs-pos_loss": 0.28413012623786926,
+      "eval_scitail-pairs-pos_runtime": 2.4346,
+      "eval_scitail-pairs-pos_samples_per_second": 65.72,
+      "eval_scitail-pairs-pos_steps_per_second": 2.054,
+      "step": 11628
+    },
+    {
+      "epoch": 2.250871080139373,
+      "eval_xsum-pairs_loss": 0.2606610655784607,
+      "eval_xsum-pairs_runtime": 1.4129,
+      "eval_xsum-pairs_samples_per_second": 113.244,
+      "eval_xsum-pairs_steps_per_second": 3.539,
+      "step": 11628
+    },
+    {
+      "epoch": 2.250871080139373,
+      "eval_compression-pairs_loss": 0.0600699707865715,
+      "eval_compression-pairs_runtime": 0.4001,
+      "eval_compression-pairs_samples_per_second": 399.872,
+      "eval_compression-pairs_steps_per_second": 12.496,
+      "step": 11628
+    },
+    {
+      "epoch": 2.250871080139373,
+      "eval_sciq_pairs_loss": 0.2134503871202469,
+      "eval_sciq_pairs_runtime": 7.955,
+      "eval_sciq_pairs_samples_per_second": 20.113,
+      "eval_sciq_pairs_steps_per_second": 0.629,
+      "step": 11628
+    },
+    {
+      "epoch": 2.250871080139373,
+      "eval_qasc_pairs_loss": 0.1396595984697342,
+      "eval_qasc_pairs_runtime": 1.4536,
+      "eval_qasc_pairs_samples_per_second": 110.075,
+      "eval_qasc_pairs_steps_per_second": 3.44,
+      "step": 11628
+    },
+    {
+      "epoch": 2.250871080139373,
+      "eval_qasc_facts_sym_loss": 0.11638667434453964,
+      "eval_qasc_facts_sym_runtime": 0.3257,
+      "eval_qasc_facts_sym_samples_per_second": 491.178,
+      "eval_qasc_facts_sym_steps_per_second": 15.349,
+      "step": 11628
+    },
+    {
+      "epoch": 2.250871080139373,
+      "eval_openbookqa_pairs_loss": 1.5970699787139893,
+      "eval_openbookqa_pairs_runtime": 1.2308,
+      "eval_openbookqa_pairs_samples_per_second": 129.994,
+      "eval_openbookqa_pairs_steps_per_second": 4.062,
+      "step": 11628
+    },
+    {
+      "epoch": 2.250871080139373,
+      "eval_msmarco_pairs_loss": 0.45277172327041626,
+      "eval_msmarco_pairs_runtime": 3.1126,
+      "eval_msmarco_pairs_samples_per_second": 51.403,
+      "eval_msmarco_pairs_steps_per_second": 1.606,
+      "step": 11628
+    },
+    {
+      "epoch": 2.250871080139373,
+      "eval_nq_pairs_loss": 0.3480191230773926,
+      "eval_nq_pairs_runtime": 7.6989,
+      "eval_nq_pairs_samples_per_second": 20.782,
+      "eval_nq_pairs_steps_per_second": 0.649,
+      "step": 11628
+    },
+    {
+      "epoch": 2.250871080139373,
+      "eval_trivia_pairs_loss": 0.5698245167732239,
+      "eval_trivia_pairs_runtime": 10.2135,
+      "eval_trivia_pairs_samples_per_second": 15.666,
+      "eval_trivia_pairs_steps_per_second": 0.49,
+      "step": 11628
+    },
+    {
+      "epoch": 2.250871080139373,
+      "eval_quora_pairs_loss": 0.19794899225234985,
+      "eval_quora_pairs_runtime": 3.9215,
+      "eval_quora_pairs_samples_per_second": 172.129,
+      "eval_quora_pairs_steps_per_second": 5.61,
+      "step": 11628
+    },
+    {
+      "epoch": 2.250871080139373,
+      "eval_gooaq_pairs_loss": 0.429452508687973,
+      "eval_gooaq_pairs_runtime": 2.2884,
+      "eval_gooaq_pairs_samples_per_second": 69.918,
+      "eval_gooaq_pairs_steps_per_second": 2.185,
+      "step": 11628
+    },
+    {
+      "epoch": 2.250871080139373,
+      "eval_mrpc_pairs_loss": 0.029982861131429672,
+      "eval_mrpc_pairs_runtime": 0.3839,
+      "eval_mrpc_pairs_samples_per_second": 416.749,
+      "eval_mrpc_pairs_steps_per_second": 13.023,
+      "step": 11628
+    },
+    {
+      "epoch": 2.2522260936895084,
+      "grad_norm": 1.217717170715332,
+      "learning_rate": 2.809262392394196e-07,
+      "loss": 0.3296,
+      "step": 11635
+    },
+    {
+      "epoch": 2.264808362369338,
+      "grad_norm": 0.60965496301651,
+      "learning_rate": 2.0174547489152985e-07,
+      "loss": 0.2261,
+      "step": 11700
+    },
+    {
+      "epoch": 2.2773906310491676,
+      "grad_norm": 0.12040360271930695,
+      "learning_rate": 1.3556996535771416e-07,
+      "loss": 0.2999,
+      "step": 11765
+    },
+    {
+      "epoch": 2.289972899728997,
+      "grad_norm": 12.428990364074707,
+      "learning_rate": 8.245786812961842e-08,
+      "loss": 0.2888,
+      "step": 11830
+    },
+    {
+      "epoch": 2.302555168408827,
+      "grad_norm": 4.584768295288086,
+      "learning_rate": 4.245586008774671e-08,
+      "loss": 0.2253,
+      "step": 11895
+    },
+    {
+      "epoch": 2.315137437088657,
+      "grad_norm": 5.902719974517822,
+      "learning_rate": 1.5599096480093388e-08,
+      "loss": 0.2343,
+      "step": 11960
+    },
+    {
+      "epoch": 2.3277197057684864,
+      "grad_norm": 9.810800552368164,
+      "learning_rate": 1.9111800263971192e-09,
+      "loss": 0.331,
+      "step": 12025
+    },
+    {
+      "epoch": 2.340301974448316,
+      "grad_norm": 0.725075364112854,
+      "learning_rate": 2.99985958598248e-05,
+      "loss": 0.2101,
+      "step": 12090
+    },
+    {
+      "epoch": 2.3528842431281456,
+      "grad_norm": 3.3202953338623047,
+      "learning_rate": 2.9986216360213095e-05,
+      "loss": 0.2155,
+      "step": 12155
+    },
+    {
+      "epoch": 2.365466511807975,
+      "grad_norm": 0.6567480564117432,
+      "learning_rate": 2.9960574370109496e-05,
+      "loss": 0.2996,
+      "step": 12220
+    },
+    {
+      "epoch": 2.3780487804878048,
+      "grad_norm": 5.063121795654297,
+      "learning_rate": 2.992178447249302e-05,
+      "loss": 0.2964,
+      "step": 12285
+    },
+    {
+      "epoch": 2.3906310491676344,
+      "grad_norm": 12.657327651977539,
+      "learning_rate": 2.986988075736407e-05,
+      "loss": 0.2825,
+      "step": 12350
+    },
+    {
+      "epoch": 2.403213317847464,
+      "grad_norm": 4.071556568145752,
+      "learning_rate": 2.980490883963215e-05,
+      "loss": 0.2259,
+      "step": 12415
+    },
+    {
+      "epoch": 2.415795586527294,
+      "grad_norm": 21.96937370300293,
+      "learning_rate": 2.9728223774548642e-05,
+      "loss": 0.373,
+      "step": 12480
+    },
+    {
+      "epoch": 2.4283778552071236,
+      "grad_norm": 8.618860244750977,
+      "learning_rate": 2.963749672572836e-05,
+      "loss": 0.2906,
+      "step": 12545
+    },
+    {
+      "epoch": 2.440960123886953,
+      "grad_norm": 3.505587100982666,
+      "learning_rate": 2.9533905702009196e-05,
+      "loss": 0.295,
+      "step": 12610
+    },
+    {
+      "epoch": 2.4535423925667827,
+      "grad_norm": 2.3120741844177246,
+      "learning_rate": 2.9417541743024074e-05,
+      "loss": 0.3156,
+      "step": 12675
+    },
+    {
+      "epoch": 2.4661246612466123,
+      "grad_norm": 7.947314262390137,
+      "learning_rate": 2.9288507113735076e-05,
+      "loss": 0.3068,
+      "step": 12740
+    },
+    {
+      "epoch": 2.4787069299264424,
+      "grad_norm": 4.989195346832275,
+      "learning_rate": 2.9146915214559173e-05,
+      "loss": 0.3137,
+      "step": 12805
+    },
+    {
+      "epoch": 2.491289198606272,
+      "grad_norm": 26.050216674804688,
+      "learning_rate": 2.899289048170777e-05,
+      "loss": 0.2985,
+      "step": 12870
+    },
+    {
+      "epoch": 2.5009678668215254,
+      "eval_StS-test_pearson_cosine": 0.8795237562868311,
+      "eval_StS-test_pearson_dot": 0.7709910161255622,
+      "eval_StS-test_pearson_euclidean": 0.862412698225094,
+      "eval_StS-test_pearson_manhattan": 0.8626090591539304,
+      "eval_StS-test_pearson_max": 0.8795237562868311,
+      "eval_StS-test_spearman_cosine": 0.8945959623295415,
+      "eval_StS-test_spearman_dot": 0.7730415187911034,
+      "eval_StS-test_spearman_euclidean": 0.8705805738685102,
+      "eval_StS-test_spearman_manhattan": 0.8717799147229099,
+      "eval_StS-test_spearman_max": 0.8945959623295415,
+      "eval_Vitaminc-test_cosine_accuracy": 0.5763157894736842,
+      "eval_Vitaminc-test_cosine_accuracy_threshold": 0.7639559507369995,
+      "eval_Vitaminc-test_cosine_ap": 0.5674566161432741,
+      "eval_Vitaminc-test_cosine_f1": 0.6679316888045541,
+      "eval_Vitaminc-test_cosine_f1_threshold": 0.4714379906654358,
+      "eval_Vitaminc-test_cosine_precision": 0.5146198830409356,
+      "eval_Vitaminc-test_cosine_recall": 0.9513513513513514,
+      "eval_Vitaminc-test_dot_accuracy": 0.5578947368421052,
+      "eval_Vitaminc-test_dot_accuracy_threshold": 16.963916778564453,
+      "eval_Vitaminc-test_dot_ap": 0.5484232300840398,
+      "eval_Vitaminc-test_dot_f1": 0.6654545454545455,
+      "eval_Vitaminc-test_dot_f1_threshold": 7.844925880432129,
+      "eval_Vitaminc-test_dot_precision": 0.5013698630136987,
+      "eval_Vitaminc-test_dot_recall": 0.9891891891891892,
+      "eval_Vitaminc-test_euclidean_accuracy": 0.5631578947368421,
+      "eval_Vitaminc-test_euclidean_accuracy_threshold": 3.618887424468994,
+      "eval_Vitaminc-test_euclidean_ap": 0.5559673243165555,
+      "eval_Vitaminc-test_euclidean_f1": 0.6666666666666666,
+      "eval_Vitaminc-test_euclidean_f1_threshold": 6.2993974685668945,
+      "eval_Vitaminc-test_euclidean_precision": 0.5,
+      "eval_Vitaminc-test_euclidean_recall": 1.0,
+      "eval_Vitaminc-test_manhattan_accuracy": 0.5605263157894737,
+      "eval_Vitaminc-test_manhattan_accuracy_threshold": 72.00088500976562,
+      "eval_Vitaminc-test_manhattan_ap": 0.556701937854166,
+      "eval_Vitaminc-test_manhattan_f1": 0.6642728904847397,
+      "eval_Vitaminc-test_manhattan_f1_threshold": 135.06130981445312,
+      "eval_Vitaminc-test_manhattan_precision": 0.49731182795698925,
+      "eval_Vitaminc-test_manhattan_recall": 1.0,
+      "eval_Vitaminc-test_max_accuracy": 0.5763157894736842,
+      "eval_Vitaminc-test_max_accuracy_threshold": 72.00088500976562,
+      "eval_Vitaminc-test_max_ap": 0.5674566161432741,
+      "eval_Vitaminc-test_max_f1": 0.6679316888045541,
+      "eval_Vitaminc-test_max_f1_threshold": 135.06130981445312,
+      "eval_Vitaminc-test_max_precision": 0.5146198830409356,
+      "eval_Vitaminc-test_max_recall": 1.0,
+      "eval_mrpc-test_cosine_accuracy": 0.7289473684210527,
+      "eval_mrpc-test_cosine_accuracy_threshold": 0.7299143671989441,
+      "eval_mrpc-test_cosine_ap": 0.8489308274880782,
+      "eval_mrpc-test_cosine_f1": 0.8159722222222223,
+      "eval_mrpc-test_cosine_f1_threshold": 0.6859562397003174,
+      "eval_mrpc-test_cosine_precision": 0.7164634146341463,
+      "eval_mrpc-test_cosine_recall": 0.9475806451612904,
+      "eval_mrpc-test_dot_accuracy": 0.6894736842105263,
+      "eval_mrpc-test_dot_accuracy_threshold": 8.769765853881836,
+      "eval_mrpc-test_dot_ap": 0.7662559072996028,
+      "eval_mrpc-test_dot_f1": 0.8052373158756136,
+      "eval_mrpc-test_dot_f1_threshold": 8.27578353881836,
+      "eval_mrpc-test_dot_precision": 0.6776859504132231,
+      "eval_mrpc-test_dot_recall": 0.9919354838709677,
+      "eval_mrpc-test_euclidean_accuracy": 0.7394736842105263,
+      "eval_mrpc-test_euclidean_accuracy_threshold": 3.322539806365967,
+      "eval_mrpc-test_euclidean_ap": 0.8234854308664945,
+      "eval_mrpc-test_euclidean_f1": 0.8216216216216217,
+      "eval_mrpc-test_euclidean_f1_threshold": 3.322539806365967,
+      "eval_mrpc-test_euclidean_precision": 0.742671009771987,
+      "eval_mrpc-test_euclidean_recall": 0.9193548387096774,
+      "eval_mrpc-test_manhattan_accuracy": 0.7315789473684211,
+      "eval_mrpc-test_manhattan_accuracy_threshold": 65.09208679199219,
+      "eval_mrpc-test_manhattan_ap": 0.8216181552643164,
+      "eval_mrpc-test_manhattan_f1": 0.8180242634315426,
+      "eval_mrpc-test_manhattan_f1_threshold": 70.48411560058594,
+      "eval_mrpc-test_manhattan_precision": 0.7173252279635258,
+      "eval_mrpc-test_manhattan_recall": 0.9516129032258065,
+      "eval_mrpc-test_max_accuracy": 0.7394736842105263,
+      "eval_mrpc-test_max_accuracy_threshold": 65.09208679199219,
+      "eval_mrpc-test_max_ap": 0.8489308274880782,
+      "eval_mrpc-test_max_f1": 0.8216216216216217,
+      "eval_mrpc-test_max_f1_threshold": 70.48411560058594,
+      "eval_mrpc-test_max_precision": 0.742671009771987,
+      "eval_mrpc-test_max_recall": 0.9919354838709677,
+      "eval_nli-pairs_loss": 0.844605028629303,
+      "eval_nli-pairs_runtime": 3.1133,
+      "eval_nli-pairs_samples_per_second": 51.392,
+      "eval_nli-pairs_steps_per_second": 1.606,
+      "eval_sequential_score": 0.5674566161432741,
+      "step": 12920
+    },
+    {
+      "epoch": 2.5009678668215254,
+      "eval_vitaminc-pairs_loss": 6.087428092956543,
+      "eval_vitaminc-pairs_runtime": 1.7072,
+      "eval_vitaminc-pairs_samples_per_second": 77.905,
+      "eval_vitaminc-pairs_steps_per_second": 2.929,
+      "step": 12920
+    },
+    {
+      "epoch": 2.5009678668215254,
+      "eval_qnli-contrastive_loss": 0.17987525463104248,
+      "eval_qnli-contrastive_runtime": 0.5327,
+      "eval_qnli-contrastive_samples_per_second": 300.332,
+      "eval_qnli-contrastive_steps_per_second": 9.385,
+      "step": 12920
+    },
+    {
+      "epoch": 2.5009678668215254,
+      "eval_scitail-pairs-qa_loss": 0.04853365197777748,
+      "eval_scitail-pairs-qa_runtime": 1.3141,
+      "eval_scitail-pairs-qa_samples_per_second": 121.755,
+      "eval_scitail-pairs-qa_steps_per_second": 3.805,
+      "step": 12920
+    },
+    {
+      "epoch": 2.5009678668215254,
+      "eval_scitail-pairs-pos_loss": 0.311461478471756,
+      "eval_scitail-pairs-pos_runtime": 2.6161,
+      "eval_scitail-pairs-pos_samples_per_second": 61.161,
+      "eval_scitail-pairs-pos_steps_per_second": 1.911,
+      "step": 12920
+    },
+    {
+      "epoch": 2.5009678668215254,
+      "eval_xsum-pairs_loss": 0.27614933252334595,
+      "eval_xsum-pairs_runtime": 1.4525,
+      "eval_xsum-pairs_samples_per_second": 110.151,
+      "eval_xsum-pairs_steps_per_second": 3.442,
+      "step": 12920
+    },
+    {
+      "epoch": 2.5009678668215254,
+      "eval_compression-pairs_loss": 0.06246453523635864,
+      "eval_compression-pairs_runtime": 0.4412,
+      "eval_compression-pairs_samples_per_second": 362.657,
+      "eval_compression-pairs_steps_per_second": 11.333,
+      "step": 12920
+    },
+    {
+      "epoch": 2.5009678668215254,
+      "eval_sciq_pairs_loss": 0.25065740942955017,
+      "eval_sciq_pairs_runtime": 8.0752,
+      "eval_sciq_pairs_samples_per_second": 19.814,
+      "eval_sciq_pairs_steps_per_second": 0.619,
+      "step": 12920
+    },
+    {
+      "epoch": 2.5009678668215254,
+      "eval_qasc_pairs_loss": 0.147993803024292,
+      "eval_qasc_pairs_runtime": 1.5149,
+      "eval_qasc_pairs_samples_per_second": 105.617,
+      "eval_qasc_pairs_steps_per_second": 3.301,
+      "step": 12920
+    },
+    {
+      "epoch": 2.5009678668215254,
+      "eval_qasc_facts_sym_loss": 0.12534400820732117,
+      "eval_qasc_facts_sym_runtime": 0.3356,
+      "eval_qasc_facts_sym_samples_per_second": 476.803,
+      "eval_qasc_facts_sym_steps_per_second": 14.9,
+      "step": 12920
+    },
+    {
+      "epoch": 2.5009678668215254,
+      "eval_openbookqa_pairs_loss": 1.685699224472046,
+      "eval_openbookqa_pairs_runtime": 1.3649,
+      "eval_openbookqa_pairs_samples_per_second": 117.226,
+      "eval_openbookqa_pairs_steps_per_second": 3.663,
+      "step": 12920
+    },
+    {
+      "epoch": 2.5009678668215254,
+      "eval_msmarco_pairs_loss": 0.5117918848991394,
+      "eval_msmarco_pairs_runtime": 3.2325,
+      "eval_msmarco_pairs_samples_per_second": 49.497,
+      "eval_msmarco_pairs_steps_per_second": 1.547,
+      "step": 12920
+    },
+    {
+      "epoch": 2.5009678668215254,
+      "eval_nq_pairs_loss": 0.40486058592796326,
+      "eval_nq_pairs_runtime": 7.8919,
+      "eval_nq_pairs_samples_per_second": 20.274,
+      "eval_nq_pairs_steps_per_second": 0.634,
+      "step": 12920
+    },
+    {
+      "epoch": 2.5009678668215254,
+      "eval_trivia_pairs_loss": 0.6596993803977966,
+      "eval_trivia_pairs_runtime": 10.3415,
+      "eval_trivia_pairs_samples_per_second": 15.472,
+      "eval_trivia_pairs_steps_per_second": 0.483,
+      "step": 12920
+    },
+    {
+      "epoch": 2.5009678668215254,
+      "eval_quora_pairs_loss": 0.16216738522052765,
+      "eval_quora_pairs_runtime": 4.0361,
+      "eval_quora_pairs_samples_per_second": 167.241,
+      "eval_quora_pairs_steps_per_second": 5.451,
+      "step": 12920
+    },
+    {
+      "epoch": 2.5009678668215254,
+      "eval_gooaq_pairs_loss": 0.46012258529663086,
+      "eval_gooaq_pairs_runtime": 2.2861,
+      "eval_gooaq_pairs_samples_per_second": 69.989,
+      "eval_gooaq_pairs_steps_per_second": 2.187,
+      "step": 12920
+    },
+    {
+      "epoch": 2.5009678668215254,
+      "eval_mrpc_pairs_loss": 0.03315501660108566,
+      "eval_mrpc_pairs_runtime": 0.3835,
+      "eval_mrpc_pairs_samples_per_second": 417.178,
+      "eval_mrpc_pairs_steps_per_second": 13.037,
+      "step": 12920
+    },
+    {
+      "epoch": 2.5038714672861015,
+      "grad_norm": 0.28498581051826477,
+      "learning_rate": 2.8826568277827527e-05,
+      "loss": 0.2487,
+      "step": 12935
+    },
+    {
+      "epoch": 2.516453735965931,
+      "grad_norm": 1.0256837606430054,
+      "learning_rate": 2.8648094773038627e-05,
+      "loss": 0.2624,
+      "step": 13000
+    },
+    {
+      "epoch": 2.5290360046457607,
+      "grad_norm": 4.241197109222412,
+      "learning_rate": 2.845762681647511e-05,
+      "loss": 0.2677,
+      "step": 13065
+    },
+    {
+      "epoch": 2.5416182733255903,
+      "grad_norm": 0.36626845598220825,
+      "learning_rate": 2.8255331798439983e-05,
+      "loss": 0.3675,
+      "step": 13130
+    },
+    {
+      "epoch": 2.55420054200542,
+      "grad_norm": 8.30843734741211,
+      "learning_rate": 2.8041387503296447e-05,
+      "loss": 0.2948,
+      "step": 13195
+    },
+    {
+      "epoch": 2.5667828106852495,
+      "grad_norm": 2.066960573196411,
+      "learning_rate": 2.7815981953224384e-05,
+      "loss": 0.2964,
+      "step": 13260
+    },
+    {
+      "epoch": 2.5793650793650795,
+      "grad_norm": 8.503743171691895,
+      "learning_rate": 2.757931324297952e-05,
+      "loss": 0.2434,
+      "step": 13325
+    },
+    {
+      "epoch": 2.591947348044909,
+      "grad_norm": 0.9377574920654297,
+      "learning_rate": 2.7331589365800378e-05,
+      "loss": 0.2722,
+      "step": 13390
+    },
+    {
+      "epoch": 2.6045296167247387,
+      "grad_norm": 0.3242267370223999,
+      "learning_rate": 2.707302803061613e-05,
+      "loss": 0.3233,
+      "step": 13455
+    },
+    {
+      "epoch": 2.6171118854045683,
+      "grad_norm": 9.376032829284668,
+      "learning_rate": 2.68038564707159e-05,
+      "loss": 0.2968,
+      "step": 13520
+    },
+    {
+      "epoch": 2.629694154084398,
+      "grad_norm": 17.351438522338867,
+      "learning_rate": 2.6524311244047752e-05,
+      "loss": 0.3172,
+      "step": 13585
+    },
+    {
+      "epoch": 2.642276422764228,
+      "grad_norm": 9.639175415039062,
+      "learning_rate": 2.6234638025322753e-05,
+      "loss": 0.2398,
+      "step": 13650
+    },
+    {
+      "epoch": 2.6548586914440575,
+      "grad_norm": 2.589085578918457,
+      "learning_rate": 2.593509139010695e-05,
+      "loss": 0.3085,
+      "step": 13715
+    },
+    {
+      "epoch": 2.667440960123887,
+      "grad_norm": 1.8706110715866089,
+      "learning_rate": 2.5625934591090952e-05,
+      "loss": 0.2809,
+      "step": 13780
+    },
+    {
+      "epoch": 2.6800232288037167,
+      "grad_norm": 3.431440591812134,
+      "learning_rate": 2.530743932673362e-05,
+      "loss": 0.3496,
+      "step": 13845
+    },
+    {
+      "epoch": 2.6926054974835463,
+      "grad_norm": 2.126661777496338,
+      "learning_rate": 2.497988550248348e-05,
+      "loss": 0.2281,
+      "step": 13910
+    },
+    {
+      "epoch": 2.705187766163376,
+      "grad_norm": 0.4347963333129883,
+      "learning_rate": 2.464356098478738e-05,
+      "loss": 0.3075,
+      "step": 13975
+    },
+    {
+      "epoch": 2.7177700348432055,
+      "grad_norm": 4.2169623374938965,
+      "learning_rate": 2.4298761348102788e-05,
+      "loss": 0.3335,
+      "step": 14040
+    },
+    {
+      "epoch": 2.730352303523035,
+      "grad_norm": 0.31253141164779663,
+      "learning_rate": 2.394578961513602e-05,
+      "loss": 0.3423,
+      "step": 14105
+    },
+    {
+      "epoch": 2.7429345722028646,
+      "grad_norm": 9.057765007019043,
+      "learning_rate": 2.3584955990534625e-05,
+      "loss": 0.2579,
+      "step": 14170
+    },
+    {
+      "epoch": 2.751064653503678,
+      "eval_StS-test_pearson_cosine": 0.8786355725766506,
+      "eval_StS-test_pearson_dot": 0.7543362401661523,
+      "eval_StS-test_pearson_euclidean": 0.8562858498968491,
+      "eval_StS-test_pearson_manhattan": 0.8534764973498815,
+      "eval_StS-test_pearson_max": 0.8786355725766506,
+      "eval_StS-test_spearman_cosine": 0.8947025850013365,
+      "eval_StS-test_spearman_dot": 0.7682338653004566,
+      "eval_StS-test_spearman_euclidean": 0.8617967923645825,
+      "eval_StS-test_spearman_manhattan": 0.8596798835071182,
+      "eval_StS-test_spearman_max": 0.8947025850013365,
+      "eval_Vitaminc-test_cosine_accuracy": 0.5736842105263158,
+      "eval_Vitaminc-test_cosine_accuracy_threshold": 0.6747879385948181,
+      "eval_Vitaminc-test_cosine_ap": 0.563396078554423,
+      "eval_Vitaminc-test_cosine_f1": 0.6703096539162112,
+      "eval_Vitaminc-test_cosine_f1_threshold": 0.38167619705200195,
+      "eval_Vitaminc-test_cosine_precision": 0.5054945054945055,
+      "eval_Vitaminc-test_cosine_recall": 0.9945945945945946,
+      "eval_Vitaminc-test_dot_accuracy": 0.5657894736842105,
+      "eval_Vitaminc-test_dot_accuracy_threshold": 17.101760864257812,
+      "eval_Vitaminc-test_dot_ap": 0.543179161382298,
+      "eval_Vitaminc-test_dot_f1": 0.6691176470588236,
+      "eval_Vitaminc-test_dot_f1_threshold": 9.36573600769043,
+      "eval_Vitaminc-test_dot_precision": 0.5069637883008357,
+      "eval_Vitaminc-test_dot_recall": 0.9837837837837838,
+      "eval_Vitaminc-test_euclidean_accuracy": 0.5684210526315789,
+      "eval_Vitaminc-test_euclidean_accuracy_threshold": 3.356025457382202,
+      "eval_Vitaminc-test_euclidean_ap": 0.549990518332784,
+      "eval_Vitaminc-test_euclidean_f1": 0.6618962432915921,
+      "eval_Vitaminc-test_euclidean_f1_threshold": 6.548671722412109,
+      "eval_Vitaminc-test_euclidean_precision": 0.4946524064171123,
+      "eval_Vitaminc-test_euclidean_recall": 1.0,
+      "eval_Vitaminc-test_manhattan_accuracy": 0.5684210526315789,
+      "eval_Vitaminc-test_manhattan_accuracy_threshold": 65.38542175292969,
+      "eval_Vitaminc-test_manhattan_ap": 0.549871250700025,
+      "eval_Vitaminc-test_manhattan_f1": 0.6630824372759856,
+      "eval_Vitaminc-test_manhattan_f1_threshold": 133.0389404296875,
+      "eval_Vitaminc-test_manhattan_precision": 0.4959785522788204,
+      "eval_Vitaminc-test_manhattan_recall": 1.0,
+      "eval_Vitaminc-test_max_accuracy": 0.5736842105263158,
+      "eval_Vitaminc-test_max_accuracy_threshold": 65.38542175292969,
+      "eval_Vitaminc-test_max_ap": 0.563396078554423,
+      "eval_Vitaminc-test_max_f1": 0.6703096539162112,
+      "eval_Vitaminc-test_max_f1_threshold": 133.0389404296875,
+      "eval_Vitaminc-test_max_precision": 0.5069637883008357,
+      "eval_Vitaminc-test_max_recall": 1.0,
+      "eval_mrpc-test_cosine_accuracy": 0.7368421052631579,
+      "eval_mrpc-test_cosine_accuracy_threshold": 0.6817317605018616,
+      "eval_mrpc-test_cosine_ap": 0.8447794771234458,
+      "eval_mrpc-test_cosine_f1": 0.8257839721254355,
+      "eval_mrpc-test_cosine_f1_threshold": 0.6817317605018616,
+      "eval_mrpc-test_cosine_precision": 0.7269938650306749,
+      "eval_mrpc-test_cosine_recall": 0.9556451612903226,
+      "eval_mrpc-test_dot_accuracy": 0.6868421052631579,
+      "eval_mrpc-test_dot_accuracy_threshold": 7.941440582275391,
+      "eval_mrpc-test_dot_ap": 0.7665690166165218,
+      "eval_mrpc-test_dot_f1": 0.8052373158756136,
+      "eval_mrpc-test_dot_f1_threshold": 7.941440582275391,
+      "eval_mrpc-test_dot_precision": 0.6776859504132231,
+      "eval_mrpc-test_dot_recall": 0.9919354838709677,
+      "eval_mrpc-test_euclidean_accuracy": 0.7394736842105263,
+      "eval_mrpc-test_euclidean_accuracy_threshold": 3.0255215167999268,
+      "eval_mrpc-test_euclidean_ap": 0.8199592936110021,
+      "eval_mrpc-test_euclidean_f1": 0.8170055452865065,
+      "eval_mrpc-test_euclidean_f1_threshold": 3.059394359588623,
+      "eval_mrpc-test_euclidean_precision": 0.7542662116040956,
+      "eval_mrpc-test_euclidean_recall": 0.8911290322580645,
+      "eval_mrpc-test_manhattan_accuracy": 0.7394736842105263,
+      "eval_mrpc-test_manhattan_accuracy_threshold": 62.93813705444336,
+      "eval_mrpc-test_manhattan_ap": 0.8196183938787889,
+      "eval_mrpc-test_manhattan_f1": 0.8218694885361553,
+      "eval_mrpc-test_manhattan_f1_threshold": 64.45368957519531,
+      "eval_mrpc-test_manhattan_precision": 0.7304075235109718,
+      "eval_mrpc-test_manhattan_recall": 0.9395161290322581,
+      "eval_mrpc-test_max_accuracy": 0.7394736842105263,
+      "eval_mrpc-test_max_accuracy_threshold": 62.93813705444336,
+      "eval_mrpc-test_max_ap": 0.8447794771234458,
+      "eval_mrpc-test_max_f1": 0.8257839721254355,
+      "eval_mrpc-test_max_f1_threshold": 64.45368957519531,
+      "eval_mrpc-test_max_precision": 0.7542662116040956,
+      "eval_mrpc-test_max_recall": 0.9919354838709677,
+      "eval_nli-pairs_loss": 0.8535138964653015,
+      "eval_nli-pairs_runtime": 3.0393,
+      "eval_nli-pairs_samples_per_second": 52.643,
+      "eval_nli-pairs_steps_per_second": 1.645,
+      "eval_sequential_score": 0.563396078554423,
+      "step": 14212
+    },
+    {
+      "epoch": 2.751064653503678,
+      "eval_vitaminc-pairs_loss": 5.795119285583496,
+      "eval_vitaminc-pairs_runtime": 1.5715,
+      "eval_vitaminc-pairs_samples_per_second": 84.631,
+      "eval_vitaminc-pairs_steps_per_second": 3.182,
+      "step": 14212
+    },
+    {
+      "epoch": 2.751064653503678,
+      "eval_qnli-contrastive_loss": 0.15990221500396729,
+      "eval_qnli-contrastive_runtime": 0.5245,
+      "eval_qnli-contrastive_samples_per_second": 305.076,
+      "eval_qnli-contrastive_steps_per_second": 9.534,
+      "step": 14212
+    },
+    {
+      "epoch": 2.751064653503678,
+      "eval_scitail-pairs-qa_loss": 0.044501952826976776,
+      "eval_scitail-pairs-qa_runtime": 1.214,
+      "eval_scitail-pairs-qa_samples_per_second": 131.796,
+      "eval_scitail-pairs-qa_steps_per_second": 4.119,
+      "step": 14212
+    },
+    {
+      "epoch": 2.751064653503678,
+      "eval_scitail-pairs-pos_loss": 0.2757861316204071,
+      "eval_scitail-pairs-pos_runtime": 2.4956,
+      "eval_scitail-pairs-pos_samples_per_second": 64.112,
+      "eval_scitail-pairs-pos_steps_per_second": 2.003,
+      "step": 14212
+    },
+    {
+      "epoch": 2.751064653503678,
+      "eval_xsum-pairs_loss": 0.27693209052085876,
+      "eval_xsum-pairs_runtime": 1.424,
+      "eval_xsum-pairs_samples_per_second": 112.363,
+      "eval_xsum-pairs_steps_per_second": 3.511,
+      "step": 14212
+    },
+    {
+      "epoch": 2.751064653503678,
+      "eval_compression-pairs_loss": 0.060699742287397385,
+      "eval_compression-pairs_runtime": 0.3966,
+      "eval_compression-pairs_samples_per_second": 403.385,
+      "eval_compression-pairs_steps_per_second": 12.606,
+      "step": 14212
+    },
+    {
+      "epoch": 2.751064653503678,
+      "eval_sciq_pairs_loss": 0.24653755128383636,
+      "eval_sciq_pairs_runtime": 7.9551,
+      "eval_sciq_pairs_samples_per_second": 20.113,
+      "eval_sciq_pairs_steps_per_second": 0.629,
+      "step": 14212
+    },
+    {
+      "epoch": 2.751064653503678,
+      "eval_qasc_pairs_loss": 0.1548977792263031,
+      "eval_qasc_pairs_runtime": 1.4497,
+      "eval_qasc_pairs_samples_per_second": 110.365,
+      "eval_qasc_pairs_steps_per_second": 3.449,
+      "step": 14212
+    },
+    {
+      "epoch": 2.751064653503678,
+      "eval_qasc_facts_sym_loss": 0.11749088764190674,
+      "eval_qasc_facts_sym_runtime": 0.3316,
+      "eval_qasc_facts_sym_samples_per_second": 482.562,
+      "eval_qasc_facts_sym_steps_per_second": 15.08,
+      "step": 14212
+    },
+    {
+      "epoch": 2.751064653503678,
+      "eval_openbookqa_pairs_loss": 1.6478569507598877,
+      "eval_openbookqa_pairs_runtime": 1.2421,
+      "eval_openbookqa_pairs_samples_per_second": 128.816,
+      "eval_openbookqa_pairs_steps_per_second": 4.025,
+      "step": 14212
+    },
+    {
+      "epoch": 2.751064653503678,
+      "eval_msmarco_pairs_loss": 0.45152410864830017,
+      "eval_msmarco_pairs_runtime": 3.128,
+      "eval_msmarco_pairs_samples_per_second": 51.151,
+      "eval_msmarco_pairs_steps_per_second": 1.598,
+      "step": 14212
+    },
+    {
+      "epoch": 2.751064653503678,
+      "eval_nq_pairs_loss": 0.41965824365615845,
+      "eval_nq_pairs_runtime": 7.6797,
+      "eval_nq_pairs_samples_per_second": 20.834,
+      "eval_nq_pairs_steps_per_second": 0.651,
+      "step": 14212
+    },
+    {
+      "epoch": 2.751064653503678,
+      "eval_trivia_pairs_loss": 0.7053402066230774,
+      "eval_trivia_pairs_runtime": 10.1801,
+      "eval_trivia_pairs_samples_per_second": 15.717,
+      "eval_trivia_pairs_steps_per_second": 0.491,
+      "step": 14212
+    },
+    {
+      "epoch": 2.751064653503678,
+      "eval_quora_pairs_loss": 0.19425351917743683,
+      "eval_quora_pairs_runtime": 3.9249,
+      "eval_quora_pairs_samples_per_second": 171.98,
+      "eval_quora_pairs_steps_per_second": 5.605,
+      "step": 14212
+    },
+    {
+      "epoch": 2.751064653503678,
+      "eval_gooaq_pairs_loss": 0.4620817303657532,
+      "eval_gooaq_pairs_runtime": 2.2762,
+      "eval_gooaq_pairs_samples_per_second": 70.292,
+      "eval_gooaq_pairs_steps_per_second": 2.197,
+      "step": 14212
+    },
+    {
+      "epoch": 2.751064653503678,
+      "eval_mrpc_pairs_loss": 0.02996860072016716,
+      "eval_mrpc_pairs_runtime": 0.3789,
+      "eval_mrpc_pairs_samples_per_second": 422.261,
+      "eval_mrpc_pairs_steps_per_second": 13.196,
+      "step": 14212
+    },
+    {
+      "epoch": 2.7555168408826947,
+      "grad_norm": 4.241976737976074,
+      "learning_rate": 2.3216577588268073e-05,
+      "loss": 0.2323,
+      "step": 14235
+    },
+    {
+      "epoch": 2.7680991095625243,
+      "grad_norm": 8.405671119689941,
+      "learning_rate": 2.2840978152936186e-05,
+      "loss": 0.2952,
+      "step": 14300
+    },
+    {
+      "epoch": 2.780681378242354,
+      "grad_norm": 7.208241939544678,
+      "learning_rate": 2.2458487775250408e-05,
+      "loss": 0.2626,
+      "step": 14365
+    },
+    {
+      "epoch": 2.7932636469221834,
+      "grad_norm": 0.3114963173866272,
+      "learning_rate": 2.2069442601937893e-05,
+      "loss": 0.2741,
+      "step": 14430
+    },
+    {
+      "epoch": 2.805845915602013,
+      "grad_norm": 5.421462535858154,
+      "learning_rate": 2.16741845403232e-05,
+      "loss": 0.2424,
+      "step": 14495
+    },
+    {
+      "epoch": 2.818428184281843,
+      "grad_norm": 2.307614803314209,
+      "learning_rate": 2.127306095784752e-05,
+      "loss": 0.2544,
+      "step": 14560
+    },
+    {
+      "epoch": 2.8310104529616726,
+      "grad_norm": 11.653911590576172,
+      "learning_rate": 2.0866424376789318e-05,
+      "loss": 0.2889,
+      "step": 14625
+    },
+    {
+      "epoch": 2.8435927216415022,
+      "grad_norm": 10.62498950958252,
+      "learning_rate": 2.0454632164454574e-05,
+      "loss": 0.2222,
+      "step": 14690
+    },
+    {
+      "epoch": 2.856174990321332,
+      "grad_norm": 9.023911476135254,
+      "learning_rate": 2.003804621910928e-05,
+      "loss": 0.2335,
+      "step": 14755
+    },
+    {
+      "epoch": 2.8687572590011614,
+      "grad_norm": 15.53839111328125,
+      "learning_rate": 1.9617032651929685e-05,
+      "loss": 0.2741,
+      "step": 14820
+    },
+    {
+      "epoch": 2.881339527680991,
+      "grad_norm": 12.925213813781738,
+      "learning_rate": 1.919196146525036e-05,
+      "loss": 0.2814,
+      "step": 14885
+    },
+    {
+      "epoch": 2.8939217963608206,
+      "grad_norm": 10.400435447692871,
+      "learning_rate": 1.8763206227392307e-05,
+      "loss": 0.2493,
+      "step": 14950
+    },
+    {
+      "epoch": 2.90650406504065,
+      "grad_norm": 0.2103821039199829,
+      "learning_rate": 1.833114374435731e-05,
+      "loss": 0.3201,
+      "step": 15015
+    },
+    {
+      "epoch": 2.91908633372048,
+      "grad_norm": 5.022848606109619,
+      "learning_rate": 1.7896153728676896e-05,
+      "loss": 0.3087,
+      "step": 15080
+    },
+    {
+      "epoch": 2.93166860240031,
+      "grad_norm": 0.6551499366760254,
+      "learning_rate": 1.7458618465706787e-05,
+      "loss": 0.2326,
+      "step": 15145
+    },
+    {
+      "epoch": 2.9442508710801394,
+      "grad_norm": 0.16861538589000702,
+      "learning_rate": 1.7018922477660426e-05,
+      "loss": 0.335,
+      "step": 15210
+    },
+    {
+      "epoch": 2.956833139759969,
+      "grad_norm": 1.4138991832733154,
+      "learning_rate": 1.6577452185676585e-05,
+      "loss": 0.2784,
+      "step": 15275
+    },
+    {
+      "epoch": 2.9694154084397986,
+      "grad_norm": 5.632164001464844,
+      "learning_rate": 1.613459557021811e-05,
+      "loss": 0.2806,
+      "step": 15340
+    },
+    {
+      "epoch": 2.9819976771196286,
+      "grad_norm": 3.5955312252044678,
+      "learning_rate": 1.5690741830100336e-05,
+      "loss": 0.185,
+      "step": 15405
+    },
+    {
+      "epoch": 2.994579945799458,
+      "grad_norm": 0.3451797068119049,
+      "learning_rate": 1.5246281040448716e-05,
+      "loss": 0.2096,
+      "step": 15470
+    },
+    {
+      "epoch": 3.0011614401858306,
+      "eval_StS-test_pearson_cosine": 0.87492343918585,
+      "eval_StS-test_pearson_dot": 0.7692943484657655,
+      "eval_StS-test_pearson_euclidean": 0.8576504398982409,
+      "eval_StS-test_pearson_manhattan": 0.8566044794239935,
+      "eval_StS-test_pearson_max": 0.87492343918585,
+      "eval_StS-test_spearman_cosine": 0.8910954765225738,
+      "eval_StS-test_spearman_dot": 0.7676996572528597,
+      "eval_StS-test_spearman_euclidean": 0.8619204921788826,
+      "eval_StS-test_spearman_manhattan": 0.8625681099677316,
+      "eval_StS-test_spearman_max": 0.8910954765225738,
+      "eval_Vitaminc-test_cosine_accuracy": 0.5710526315789474,
+      "eval_Vitaminc-test_cosine_accuracy_threshold": 0.7102299928665161,
+      "eval_Vitaminc-test_cosine_ap": 0.5511272681050597,
+      "eval_Vitaminc-test_cosine_f1": 0.6741154562383612,
+      "eval_Vitaminc-test_cosine_f1_threshold": 0.4787493944168091,
+      "eval_Vitaminc-test_cosine_precision": 0.5142045454545454,
+      "eval_Vitaminc-test_cosine_recall": 0.9783783783783784,
+      "eval_Vitaminc-test_dot_accuracy": 0.5552631578947368,
+      "eval_Vitaminc-test_dot_accuracy_threshold": 15.257586479187012,
+      "eval_Vitaminc-test_dot_ap": 0.5382903253286273,
+      "eval_Vitaminc-test_dot_f1": 0.6718446601941749,
+      "eval_Vitaminc-test_dot_f1_threshold": 11.204628944396973,
+      "eval_Vitaminc-test_dot_precision": 0.5242424242424243,
+      "eval_Vitaminc-test_dot_recall": 0.9351351351351351,
+      "eval_Vitaminc-test_euclidean_accuracy": 0.5657894736842105,
+      "eval_Vitaminc-test_euclidean_accuracy_threshold": 3.175092935562134,
+      "eval_Vitaminc-test_euclidean_ap": 0.5490562334663566,
+      "eval_Vitaminc-test_euclidean_f1": 0.6642468239564427,
+      "eval_Vitaminc-test_euclidean_f1_threshold": 5.7642717361450195,
+      "eval_Vitaminc-test_euclidean_precision": 0.5,
+      "eval_Vitaminc-test_euclidean_recall": 0.9891891891891892,
+      "eval_Vitaminc-test_manhattan_accuracy": 0.5684210526315789,
+      "eval_Vitaminc-test_manhattan_accuracy_threshold": 66.32759094238281,
+      "eval_Vitaminc-test_manhattan_ap": 0.5526446299587435,
+      "eval_Vitaminc-test_manhattan_f1": 0.6678966789667896,
+      "eval_Vitaminc-test_manhattan_f1_threshold": 113.66679382324219,
+      "eval_Vitaminc-test_manhattan_precision": 0.5070028011204482,
+      "eval_Vitaminc-test_manhattan_recall": 0.9783783783783784,
+      "eval_Vitaminc-test_max_accuracy": 0.5710526315789474,
+      "eval_Vitaminc-test_max_accuracy_threshold": 66.32759094238281,
+      "eval_Vitaminc-test_max_ap": 0.5526446299587435,
+      "eval_Vitaminc-test_max_f1": 0.6741154562383612,
+      "eval_Vitaminc-test_max_f1_threshold": 113.66679382324219,
+      "eval_Vitaminc-test_max_precision": 0.5242424242424243,
+      "eval_Vitaminc-test_max_recall": 0.9891891891891892,
+      "eval_mrpc-test_cosine_accuracy": 0.7394736842105263,
+      "eval_mrpc-test_cosine_accuracy_threshold": 0.7056349515914917,
+      "eval_mrpc-test_cosine_ap": 0.848334505377635,
+      "eval_mrpc-test_cosine_f1": 0.8237347294938918,
+      "eval_mrpc-test_cosine_f1_threshold": 0.6817628145217896,
+      "eval_mrpc-test_cosine_precision": 0.7261538461538461,
+      "eval_mrpc-test_cosine_recall": 0.9516129032258065,
+      "eval_mrpc-test_dot_accuracy": 0.6868421052631579,
+      "eval_mrpc-test_dot_accuracy_threshold": 8.654559135437012,
+      "eval_mrpc-test_dot_ap": 0.7545877870990263,
+      "eval_mrpc-test_dot_f1": 0.8045602605863192,
+      "eval_mrpc-test_dot_f1_threshold": 6.8338165283203125,
+      "eval_mrpc-test_dot_precision": 0.674863387978142,
+      "eval_mrpc-test_dot_recall": 0.9959677419354839,
+      "eval_mrpc-test_euclidean_accuracy": 0.7421052631578947,
+      "eval_mrpc-test_euclidean_accuracy_threshold": 3.064603805541992,
+      "eval_mrpc-test_euclidean_ap": 0.8262105822554444,
+      "eval_mrpc-test_euclidean_f1": 0.8231046931407942,
+      "eval_mrpc-test_euclidean_f1_threshold": 3.064603805541992,
+      "eval_mrpc-test_euclidean_precision": 0.7450980392156863,
+      "eval_mrpc-test_euclidean_recall": 0.9193548387096774,
+      "eval_mrpc-test_manhattan_accuracy": 0.7421052631578947,
+      "eval_mrpc-test_manhattan_accuracy_threshold": 61.945098876953125,
+      "eval_mrpc-test_manhattan_ap": 0.8276315163519761,
+      "eval_mrpc-test_manhattan_f1": 0.825,
+      "eval_mrpc-test_manhattan_f1_threshold": 62.05877685546875,
+      "eval_mrpc-test_manhattan_precision": 0.7403846153846154,
+      "eval_mrpc-test_manhattan_recall": 0.9314516129032258,
+      "eval_mrpc-test_max_accuracy": 0.7421052631578947,
+      "eval_mrpc-test_max_accuracy_threshold": 61.945098876953125,
+      "eval_mrpc-test_max_ap": 0.848334505377635,
+      "eval_mrpc-test_max_f1": 0.825,
+      "eval_mrpc-test_max_f1_threshold": 62.05877685546875,
+      "eval_mrpc-test_max_precision": 0.7450980392156863,
+      "eval_mrpc-test_max_recall": 0.9959677419354839,
+      "eval_nli-pairs_loss": 0.7936088442802429,
+      "eval_nli-pairs_runtime": 3.1903,
+      "eval_nli-pairs_samples_per_second": 50.152,
+      "eval_nli-pairs_steps_per_second": 1.567,
+      "eval_sequential_score": 0.5526446299587435,
+      "step": 15504
+    },
+    {
+      "epoch": 3.0011614401858306,
+      "eval_vitaminc-pairs_loss": 5.943141937255859,
+      "eval_vitaminc-pairs_runtime": 1.6452,
+      "eval_vitaminc-pairs_samples_per_second": 80.84,
+      "eval_vitaminc-pairs_steps_per_second": 3.039,
+      "step": 15504
+    },
+    {
+      "epoch": 3.0011614401858306,
+      "eval_qnli-contrastive_loss": 0.13478858768939972,
+      "eval_qnli-contrastive_runtime": 0.5401,
+      "eval_qnli-contrastive_samples_per_second": 296.259,
+      "eval_qnli-contrastive_steps_per_second": 9.258,
+      "step": 15504
+    },
+    {
+      "epoch": 3.0011614401858306,
+      "eval_scitail-pairs-qa_loss": 0.03971510007977486,
+      "eval_scitail-pairs-qa_runtime": 1.3721,
+      "eval_scitail-pairs-qa_samples_per_second": 116.61,
+      "eval_scitail-pairs-qa_steps_per_second": 3.644,
+      "step": 15504
+    },
+    {
+      "epoch": 3.0011614401858306,
+      "eval_scitail-pairs-pos_loss": 0.2736285924911499,
+      "eval_scitail-pairs-pos_runtime": 2.8426,
+      "eval_scitail-pairs-pos_samples_per_second": 56.286,
+      "eval_scitail-pairs-pos_steps_per_second": 1.759,
+      "step": 15504
+    },
+    {
+      "epoch": 3.0011614401858306,
+      "eval_xsum-pairs_loss": 0.27379804849624634,
+      "eval_xsum-pairs_runtime": 1.4208,
+      "eval_xsum-pairs_samples_per_second": 112.611,
+      "eval_xsum-pairs_steps_per_second": 3.519,
+      "step": 15504
+    },
+    {
+      "epoch": 3.0011614401858306,
+      "eval_compression-pairs_loss": 0.055795930325984955,
+      "eval_compression-pairs_runtime": 0.4254,
+      "eval_compression-pairs_samples_per_second": 376.141,
+      "eval_compression-pairs_steps_per_second": 11.754,
+      "step": 15504
+    },
+    {
+      "epoch": 3.0011614401858306,
+      "eval_sciq_pairs_loss": 0.20566284656524658,
+      "eval_sciq_pairs_runtime": 8.2359,
+      "eval_sciq_pairs_samples_per_second": 19.427,
+      "eval_sciq_pairs_steps_per_second": 0.607,
+      "step": 15504
+    },
+    {
+      "epoch": 3.0011614401858306,
+      "eval_qasc_pairs_loss": 0.13599181175231934,
+      "eval_qasc_pairs_runtime": 1.4778,
+      "eval_qasc_pairs_samples_per_second": 108.271,
+      "eval_qasc_pairs_steps_per_second": 3.383,
+      "step": 15504
+    },
+    {
+      "epoch": 3.0011614401858306,
+      "eval_qasc_facts_sym_loss": 0.11813600361347198,
+      "eval_qasc_facts_sym_runtime": 0.3695,
+      "eval_qasc_facts_sym_samples_per_second": 432.998,
+      "eval_qasc_facts_sym_steps_per_second": 13.531,
+      "step": 15504
+    },
+    {
+      "epoch": 3.0011614401858306,
+      "eval_openbookqa_pairs_loss": 1.5948731899261475,
+      "eval_openbookqa_pairs_runtime": 1.3515,
+      "eval_openbookqa_pairs_samples_per_second": 118.389,
+      "eval_openbookqa_pairs_steps_per_second": 3.7,
+      "step": 15504
+    },
+    {
+      "epoch": 3.0011614401858306,
+      "eval_msmarco_pairs_loss": 0.48051756620407104,
+      "eval_msmarco_pairs_runtime": 3.3017,
+      "eval_msmarco_pairs_samples_per_second": 48.46,
+      "eval_msmarco_pairs_steps_per_second": 1.514,
+      "step": 15504
+    },
+    {
+      "epoch": 3.0011614401858306,
+      "eval_nq_pairs_loss": 0.38785767555236816,
+      "eval_nq_pairs_runtime": 7.8378,
+      "eval_nq_pairs_samples_per_second": 20.414,
+      "eval_nq_pairs_steps_per_second": 0.638,
+      "step": 15504
+    },
+    {
+      "epoch": 3.0011614401858306,
+      "eval_trivia_pairs_loss": 0.6309649348258972,
+      "eval_trivia_pairs_runtime": 10.3017,
+      "eval_trivia_pairs_samples_per_second": 15.531,
+      "eval_trivia_pairs_steps_per_second": 0.485,
+      "step": 15504
+    },
+    {
+      "epoch": 3.0011614401858306,
+      "eval_quora_pairs_loss": 0.175103560090065,
+      "eval_quora_pairs_runtime": 4.0555,
+      "eval_quora_pairs_samples_per_second": 166.441,
+      "eval_quora_pairs_steps_per_second": 5.425,
+      "step": 15504
+    },
+    {
+      "epoch": 3.0011614401858306,
+      "eval_gooaq_pairs_loss": 0.42644429206848145,
+      "eval_gooaq_pairs_runtime": 2.2853,
+      "eval_gooaq_pairs_samples_per_second": 70.011,
+      "eval_gooaq_pairs_steps_per_second": 2.188,
+      "step": 15504
+    },
+    {
+      "epoch": 3.0011614401858306,
+      "eval_mrpc_pairs_loss": 0.02719317004084587,
+      "eval_mrpc_pairs_runtime": 0.3819,
+      "eval_mrpc_pairs_samples_per_second": 418.95,
+      "eval_mrpc_pairs_steps_per_second": 13.092,
+      "step": 15504
+    },
+    {
+      "epoch": 3.0071622144792878,
+      "grad_norm": 0.18203336000442505,
+      "learning_rate": 1.480160380988634e-05,
+      "loss": 0.2311,
+      "step": 15535
+    },
+    {
+      "epoch": 3.0197444831591174,
+      "grad_norm": 4.591137409210205,
+      "learning_rate": 1.4357100937252624e-05,
+      "loss": 0.1838,
+      "step": 15600
+    },
+    {
+      "epoch": 3.032326751838947,
+      "grad_norm": 0.6497881412506104,
+      "learning_rate": 1.391316306815472e-05,
+      "loss": 0.2076,
+      "step": 15665
+    },
+    {
+      "epoch": 3.0449090205187765,
+      "grad_norm": 10.108341217041016,
+      "learning_rate": 1.3470180351653784e-05,
+      "loss": 0.2866,
+      "step": 15730
+    },
+    {
+      "epoch": 3.057491289198606,
+      "grad_norm": 0.8622055649757385,
+      "learning_rate": 1.302854209738737e-05,
+      "loss": 0.2769,
+      "step": 15795
+    },
+    {
+      "epoch": 3.0700735578784357,
+      "grad_norm": 15.875243186950684,
+      "learning_rate": 1.258863643342987e-05,
+      "loss": 0.2398,
+      "step": 15860
+    },
+    {
+      "epoch": 3.0826558265582658,
+      "grad_norm": 7.3159708976745605,
+      "learning_rate": 1.2150849965190985e-05,
+      "loss": 0.3044,
+      "step": 15925
+    },
+    {
+      "epoch": 3.0952380952380953,
+      "grad_norm": 9.284073829650879,
+      "learning_rate": 1.1715567435652663e-05,
+      "loss": 0.261,
+      "step": 15990
+    },
+    {
+      "epoch": 3.107820363917925,
+      "grad_norm": 10.888318061828613,
+      "learning_rate": 1.1283171387242675e-05,
+      "loss": 0.2806,
+      "step": 16055
+    },
+    {
+      "epoch": 3.1204026325977545,
+      "grad_norm": 2.058668851852417,
+      "learning_rate": 1.0854041825642137e-05,
+      "loss": 0.2093,
+      "step": 16120
+    },
+    {
+      "epoch": 3.132984901277584,
+      "grad_norm": 7.410093307495117,
+      "learning_rate": 1.042855588582242e-05,
+      "loss": 0.1835,
+      "step": 16185
+    },
+    {
+      "epoch": 3.1455671699574137,
+      "grad_norm": 7.427304744720459,
+      "learning_rate": 1.0007087500604994e-05,
+      "loss": 0.2615,
+      "step": 16250
+    },
+    {
+      "epoch": 3.1581494386372437,
+      "grad_norm": 7.91264533996582,
+      "learning_rate": 9.590007072035367e-06,
+      "loss": 0.3112,
+      "step": 16315
+    },
+    {
+      "epoch": 3.1707317073170733,
+      "grad_norm": 0.32049843668937683,
+      "learning_rate": 9.177681145859981e-06,
+      "loss": 0.2247,
+      "step": 16380
+    },
+    {
+      "epoch": 3.183313975996903,
+      "grad_norm": 0.22129254043102264,
+      "learning_rate": 8.77047208939235e-06,
+      "loss": 0.2808,
+      "step": 16445
+    },
+    {
+      "epoch": 3.1958962446767325,
+      "grad_norm": 17.339345932006836,
+      "learning_rate": 8.368737773051066e-06,
+      "loss": 0.2256,
+      "step": 16510
+    },
+    {
+      "epoch": 3.208478513356562,
+      "grad_norm": 0.13545936346054077,
+      "learning_rate": 7.972831255850175e-06,
+      "loss": 0.266,
+      "step": 16575
+    },
+    {
+      "epoch": 3.2210607820363917,
+      "grad_norm": 7.021492004394531,
+      "learning_rate": 7.583100475117644e-06,
+      "loss": 0.2396,
+      "step": 16640
+    },
+    {
+      "epoch": 3.2336430507162213,
+      "grad_norm": 2.81817889213562,
+      "learning_rate": 7.1998879407153136e-06,
+      "loss": 0.2493,
+      "step": 16705
+    },
+    {
+      "epoch": 3.2462253193960513,
+      "grad_norm": 1.4829257726669312,
+      "learning_rate": 6.823530434028392e-06,
+      "loss": 0.2642,
+      "step": 16770
+    },
+    {
+      "epoch": 3.251258226867983,
+      "eval_StS-test_pearson_cosine": 0.880777931752262,
+      "eval_StS-test_pearson_dot": 0.7745050681792695,
+      "eval_StS-test_pearson_euclidean": 0.8649445292928557,
+      "eval_StS-test_pearson_manhattan": 0.8633416689400311,
+      "eval_StS-test_pearson_max": 0.880777931752262,
+      "eval_StS-test_spearman_cosine": 0.8965067413194601,
+      "eval_StS-test_spearman_dot": 0.7722399877819839,
+      "eval_StS-test_spearman_euclidean": 0.8683118316991572,
+      "eval_StS-test_spearman_manhattan": 0.8680132006430414,
+      "eval_StS-test_spearman_max": 0.8965067413194601,
+      "eval_Vitaminc-test_cosine_accuracy": 0.5789473684210527,
+      "eval_Vitaminc-test_cosine_accuracy_threshold": 0.7058641910552979,
+      "eval_Vitaminc-test_cosine_ap": 0.5596672617305367,
+      "eval_Vitaminc-test_cosine_f1": 0.6766355140186916,
+      "eval_Vitaminc-test_cosine_f1_threshold": 0.43707022070884705,
+      "eval_Vitaminc-test_cosine_precision": 0.5171428571428571,
+      "eval_Vitaminc-test_cosine_recall": 0.9783783783783784,
+      "eval_Vitaminc-test_dot_accuracy": 0.5631578947368421,
+      "eval_Vitaminc-test_dot_accuracy_threshold": 12.52048110961914,
+      "eval_Vitaminc-test_dot_ap": 0.5494024707334075,
+      "eval_Vitaminc-test_dot_f1": 0.6703096539162112,
+      "eval_Vitaminc-test_dot_f1_threshold": 7.3003621101379395,
+      "eval_Vitaminc-test_dot_precision": 0.5054945054945055,
+      "eval_Vitaminc-test_dot_recall": 0.9945945945945946,
+      "eval_Vitaminc-test_euclidean_accuracy": 0.5605263157894737,
+      "eval_Vitaminc-test_euclidean_accuracy_threshold": 3.11747145652771,
+      "eval_Vitaminc-test_euclidean_ap": 0.547341808113373,
+      "eval_Vitaminc-test_euclidean_f1": 0.6678832116788321,
+      "eval_Vitaminc-test_euclidean_f1_threshold": 5.6768622398376465,
+      "eval_Vitaminc-test_euclidean_precision": 0.5041322314049587,
+      "eval_Vitaminc-test_euclidean_recall": 0.9891891891891892,
+      "eval_Vitaminc-test_manhattan_accuracy": 0.5684210526315789,
+      "eval_Vitaminc-test_manhattan_accuracy_threshold": 63.42762756347656,
+      "eval_Vitaminc-test_manhattan_ap": 0.5490956668956571,
+      "eval_Vitaminc-test_manhattan_f1": 0.6654545454545455,
+      "eval_Vitaminc-test_manhattan_f1_threshold": 118.83499908447266,
+      "eval_Vitaminc-test_manhattan_precision": 0.5013698630136987,
+      "eval_Vitaminc-test_manhattan_recall": 0.9891891891891892,
+      "eval_Vitaminc-test_max_accuracy": 0.5789473684210527,
+      "eval_Vitaminc-test_max_accuracy_threshold": 63.42762756347656,
+      "eval_Vitaminc-test_max_ap": 0.5596672617305367,
+      "eval_Vitaminc-test_max_f1": 0.6766355140186916,
+      "eval_Vitaminc-test_max_f1_threshold": 118.83499908447266,
+      "eval_Vitaminc-test_max_precision": 0.5171428571428571,
+      "eval_Vitaminc-test_max_recall": 0.9945945945945946,
+      "eval_mrpc-test_cosine_accuracy": 0.7315789473684211,
+      "eval_mrpc-test_cosine_accuracy_threshold": 0.7081175446510315,
+      "eval_mrpc-test_cosine_ap": 0.8471091006168546,
+      "eval_mrpc-test_cosine_f1": 0.8204225352112677,
+      "eval_mrpc-test_cosine_f1_threshold": 0.7004733085632324,
+      "eval_mrpc-test_cosine_precision": 0.728125,
+      "eval_mrpc-test_cosine_recall": 0.9395161290322581,
+      "eval_mrpc-test_dot_accuracy": 0.6921052631578948,
+      "eval_mrpc-test_dot_accuracy_threshold": 7.834476470947266,
+      "eval_mrpc-test_dot_ap": 0.7544989265631823,
+      "eval_mrpc-test_dot_f1": 0.8072487644151566,
+      "eval_mrpc-test_dot_f1_threshold": 7.739718437194824,
+      "eval_mrpc-test_dot_precision": 0.6824512534818942,
+      "eval_mrpc-test_dot_recall": 0.9879032258064516,
+      "eval_mrpc-test_euclidean_accuracy": 0.7421052631578947,
+      "eval_mrpc-test_euclidean_accuracy_threshold": 3.0166122913360596,
+      "eval_mrpc-test_euclidean_ap": 0.8252951646149229,
+      "eval_mrpc-test_euclidean_f1": 0.8231046931407942,
+      "eval_mrpc-test_euclidean_f1_threshold": 3.0470428466796875,
+      "eval_mrpc-test_euclidean_precision": 0.7450980392156863,
+      "eval_mrpc-test_euclidean_recall": 0.9193548387096774,
+      "eval_mrpc-test_manhattan_accuracy": 0.7394736842105263,
+      "eval_mrpc-test_manhattan_accuracy_threshold": 60.53413391113281,
+      "eval_mrpc-test_manhattan_ap": 0.8247798907091635,
+      "eval_mrpc-test_manhattan_f1": 0.8220640569395017,
+      "eval_mrpc-test_manhattan_f1_threshold": 62.061195373535156,
+      "eval_mrpc-test_manhattan_precision": 0.7356687898089171,
+      "eval_mrpc-test_manhattan_recall": 0.9314516129032258,
+      "eval_mrpc-test_max_accuracy": 0.7421052631578947,
+      "eval_mrpc-test_max_accuracy_threshold": 60.53413391113281,
+      "eval_mrpc-test_max_ap": 0.8471091006168546,
+      "eval_mrpc-test_max_f1": 0.8231046931407942,
+      "eval_mrpc-test_max_f1_threshold": 62.061195373535156,
+      "eval_mrpc-test_max_precision": 0.7450980392156863,
+      "eval_mrpc-test_max_recall": 0.9879032258064516,
+      "eval_nli-pairs_loss": 0.7501987814903259,
+      "eval_nli-pairs_runtime": 3.0677,
+      "eval_nli-pairs_samples_per_second": 52.156,
+      "eval_nli-pairs_steps_per_second": 1.63,
+      "eval_sequential_score": 0.5596672617305367,
+      "step": 16796
+    },
+    {
+      "epoch": 3.251258226867983,
+      "eval_vitaminc-pairs_loss": 6.001572132110596,
+      "eval_vitaminc-pairs_runtime": 1.6281,
+      "eval_vitaminc-pairs_samples_per_second": 81.691,
+      "eval_vitaminc-pairs_steps_per_second": 3.071,
+      "step": 16796
+    },
+    {
+      "epoch": 3.251258226867983,
+      "eval_qnli-contrastive_loss": 0.1142602190375328,
+      "eval_qnli-contrastive_runtime": 0.5259,
+      "eval_qnli-contrastive_samples_per_second": 304.23,
+      "eval_qnli-contrastive_steps_per_second": 9.507,
+      "step": 16796
+    },
+    {
+      "epoch": 3.251258226867983,
+      "eval_scitail-pairs-qa_loss": 0.03872662037611008,
+      "eval_scitail-pairs-qa_runtime": 1.2213,
+      "eval_scitail-pairs-qa_samples_per_second": 131.01,
+      "eval_scitail-pairs-qa_steps_per_second": 4.094,
+      "step": 16796
+    },
+    {
+      "epoch": 3.251258226867983,
+      "eval_scitail-pairs-pos_loss": 0.2690686285495758,
+      "eval_scitail-pairs-pos_runtime": 2.3945,
+      "eval_scitail-pairs-pos_samples_per_second": 66.82,
+      "eval_scitail-pairs-pos_steps_per_second": 2.088,
+      "step": 16796
+    },
+    {
+      "epoch": 3.251258226867983,
+      "eval_xsum-pairs_loss": 0.2524264454841614,
+      "eval_xsum-pairs_runtime": 1.4129,
+      "eval_xsum-pairs_samples_per_second": 113.239,
+      "eval_xsum-pairs_steps_per_second": 3.539,
+      "step": 16796
+    },
+    {
+      "epoch": 3.251258226867983,
+      "eval_compression-pairs_loss": 0.053744763135910034,
+      "eval_compression-pairs_runtime": 0.4045,
+      "eval_compression-pairs_samples_per_second": 395.564,
+      "eval_compression-pairs_steps_per_second": 12.361,
+      "step": 16796
+    },
+    {
+      "epoch": 3.251258226867983,
+      "eval_sciq_pairs_loss": 0.2147907316684723,
+      "eval_sciq_pairs_runtime": 8.0242,
+      "eval_sciq_pairs_samples_per_second": 19.94,
+      "eval_sciq_pairs_steps_per_second": 0.623,
+      "step": 16796
+    },
+    {
+      "epoch": 3.251258226867983,
+      "eval_qasc_pairs_loss": 0.1279141902923584,
+      "eval_qasc_pairs_runtime": 1.4536,
+      "eval_qasc_pairs_samples_per_second": 110.074,
+      "eval_qasc_pairs_steps_per_second": 3.44,
+      "step": 16796
+    },
+    {
+      "epoch": 3.251258226867983,
+      "eval_qasc_facts_sym_loss": 0.11657889187335968,
+      "eval_qasc_facts_sym_runtime": 0.33,
+      "eval_qasc_facts_sym_samples_per_second": 484.857,
+      "eval_qasc_facts_sym_steps_per_second": 15.152,
+      "step": 16796
+    },
+    {
+      "epoch": 3.251258226867983,
+      "eval_openbookqa_pairs_loss": 1.5521962642669678,
+      "eval_openbookqa_pairs_runtime": 1.2491,
+      "eval_openbookqa_pairs_samples_per_second": 128.096,
+      "eval_openbookqa_pairs_steps_per_second": 4.003,
+      "step": 16796
+    },
+    {
+      "epoch": 3.251258226867983,
+      "eval_msmarco_pairs_loss": 0.45534777641296387,
+      "eval_msmarco_pairs_runtime": 3.1507,
+      "eval_msmarco_pairs_samples_per_second": 50.783,
+      "eval_msmarco_pairs_steps_per_second": 1.587,
+      "step": 16796
+    },
+    {
+      "epoch": 3.251258226867983,
+      "eval_nq_pairs_loss": 0.3408084511756897,
+      "eval_nq_pairs_runtime": 7.7393,
+      "eval_nq_pairs_samples_per_second": 20.674,
+      "eval_nq_pairs_steps_per_second": 0.646,
+      "step": 16796
+    },
+    {
+      "epoch": 3.251258226867983,
+      "eval_trivia_pairs_loss": 0.6149572134017944,
+      "eval_trivia_pairs_runtime": 10.2599,
+      "eval_trivia_pairs_samples_per_second": 15.595,
+      "eval_trivia_pairs_steps_per_second": 0.487,
+      "step": 16796
+    },
+    {
+      "epoch": 3.251258226867983,
+      "eval_quora_pairs_loss": 0.16969382762908936,
+      "eval_quora_pairs_runtime": 3.9818,
+      "eval_quora_pairs_samples_per_second": 169.521,
+      "eval_quora_pairs_steps_per_second": 5.525,
+      "step": 16796
+    },
+    {
+      "epoch": 3.251258226867983,
+      "eval_gooaq_pairs_loss": 0.3949827253818512,
+      "eval_gooaq_pairs_runtime": 2.286,
+      "eval_gooaq_pairs_samples_per_second": 69.99,
+      "eval_gooaq_pairs_steps_per_second": 2.187,
+      "step": 16796
+    },
+    {
+      "epoch": 3.251258226867983,
+      "eval_mrpc_pairs_loss": 0.025785308331251144,
+      "eval_mrpc_pairs_runtime": 0.3822,
+      "eval_mrpc_pairs_samples_per_second": 418.675,
+      "eval_mrpc_pairs_steps_per_second": 13.084,
+      "step": 16796
+    },
+    {
+      "epoch": 3.258807588075881,
+      "grad_norm": 0.9428536295890808,
+      "learning_rate": 6.4543587119895626e-06,
+      "loss": 0.2925,
+      "step": 16835
+    },
+    {
+      "epoch": 3.2713898567557105,
+      "grad_norm": 0.49789130687713623,
+      "learning_rate": 6.092697216397478e-06,
+      "loss": 0.2983,
+      "step": 16900
+    },
+    {
+      "epoch": 3.28397212543554,
+      "grad_norm": 0.5712814927101135,
+      "learning_rate": 5.7388637887853255e-06,
+      "loss": 0.2797,
+      "step": 16965
+    },
+    {
+      "epoch": 3.2965543941153697,
+      "grad_norm": 0.5957891345024109,
+      "learning_rate": 5.393169391089869e-06,
+      "loss": 0.2351,
+      "step": 17030
+    },
+    {
+      "epoch": 3.3091366627951992,
+      "grad_norm": 0.17854127287864685,
+      "learning_rate": 5.055917832366551e-06,
+      "loss": 0.2431,
+      "step": 17095
+    },
+    {
+      "epoch": 3.3217189314750293,
+      "grad_norm": 0.43127721548080444,
+      "learning_rate": 4.727405501790925e-06,
+      "loss": 0.2406,
+      "step": 17160
+    },
+    {
+      "epoch": 3.334301200154859,
+      "grad_norm": 0.40976136922836304,
+      "learning_rate": 4.407921108180699e-06,
+      "loss": 0.2354,
+      "step": 17225
+    },
+    {
+      "epoch": 3.3468834688346885,
+      "grad_norm": 4.054069519042969,
+      "learning_rate": 4.097745426267878e-06,
+      "loss": 0.1614,
+      "step": 17290
+    },
+    {
+      "epoch": 3.359465737514518,
+      "grad_norm": 0.5118033289909363,
+      "learning_rate": 3.797151049943393e-06,
+      "loss": 0.1372,
+      "step": 17355
+    },
+    {
+      "epoch": 3.3720480061943476,
+      "grad_norm": 0.12063035368919373,
+      "learning_rate": 3.5064021526915996e-06,
+      "loss": 0.2192,
+      "step": 17420
+    },
+    {
+      "epoch": 3.3846302748741772,
+      "grad_norm": 1.3429920673370361,
+      "learning_rate": 3.225754255424692e-06,
+      "loss": 0.1628,
+      "step": 17485
+    },
+    {
+      "epoch": 3.397212543554007,
+      "grad_norm": 6.818206310272217,
+      "learning_rate": 2.9554540019215848e-06,
+      "loss": 0.1496,
+      "step": 17550
+    },
+    {
+      "epoch": 3.4097948122338364,
+      "grad_norm": 0.1729680895805359,
+      "learning_rate": 2.6957389420681094e-06,
+      "loss": 0.1813,
+      "step": 17615
+    },
+    {
+      "epoch": 3.4223770809136664,
+      "grad_norm": 2.8938841819763184,
+      "learning_rate": 2.446837323089423e-06,
+      "loss": 0.1914,
+      "step": 17680
+    },
+    {
+      "epoch": 3.434959349593496,
+      "grad_norm": 0.3303506076335907,
+      "learning_rate": 2.208967888957853e-06,
+      "loss": 0.196,
+      "step": 17745
+    },
+    {
+      "epoch": 3.4475416182733256,
+      "grad_norm": 3.809260606765747,
+      "learning_rate": 1.982339688152608e-06,
+      "loss": 0.1858,
+      "step": 17810
+    },
+    {
+      "epoch": 3.460123886953155,
+      "grad_norm": 3.7293314933776855,
+      "learning_rate": 1.7671518899402124e-06,
+      "loss": 0.19,
+      "step": 17875
+    },
+    {
+      "epoch": 3.472706155632985,
+      "grad_norm": 0.11209724098443985,
+      "learning_rate": 1.5635936093371428e-06,
+      "loss": 0.166,
+      "step": 17940
+    },
+    {
+      "epoch": 3.4852884243128144,
+      "grad_norm": 1.0682430267333984,
+      "learning_rate": 1.3718437409086144e-06,
+      "loss": 0.1681,
+      "step": 18005
+    },
+    {
+      "epoch": 3.4978706929926444,
+      "grad_norm": 0.11844446510076523,
+      "learning_rate": 1.1920708015492954e-06,
+      "loss": 0.1615,
+      "step": 18070
+    },
+    {
+      "epoch": 3.5013550135501355,
+      "eval_StS-test_pearson_cosine": 0.8813824345718434,
+      "eval_StS-test_pearson_dot": 0.7717544169066424,
+      "eval_StS-test_pearson_euclidean": 0.8642445322185591,
+      "eval_StS-test_pearson_manhattan": 0.8626796723170516,
+      "eval_StS-test_pearson_max": 0.8813824345718434,
+      "eval_StS-test_spearman_cosine": 0.8968829857988841,
+      "eval_StS-test_spearman_dot": 0.7725145356884126,
+      "eval_StS-test_spearman_euclidean": 0.868454031751171,
+      "eval_StS-test_spearman_manhattan": 0.8683110654171217,
+      "eval_StS-test_spearman_max": 0.8968829857988841,
+      "eval_Vitaminc-test_cosine_accuracy": 0.5789473684210527,
+      "eval_Vitaminc-test_cosine_accuracy_threshold": 0.7279139161109924,
+      "eval_Vitaminc-test_cosine_ap": 0.5599139775011265,
+      "eval_Vitaminc-test_cosine_f1": 0.6741996233521658,
+      "eval_Vitaminc-test_cosine_f1_threshold": 0.4386724531650543,
+      "eval_Vitaminc-test_cosine_precision": 0.5173410404624278,
+      "eval_Vitaminc-test_cosine_recall": 0.9675675675675676,
+      "eval_Vitaminc-test_dot_accuracy": 0.5657894736842105,
+      "eval_Vitaminc-test_dot_accuracy_threshold": 12.463031768798828,
+      "eval_Vitaminc-test_dot_ap": 0.5548755673509098,
+      "eval_Vitaminc-test_dot_f1": 0.6691042047531992,
+      "eval_Vitaminc-test_dot_f1_threshold": 6.735769271850586,
+      "eval_Vitaminc-test_dot_precision": 0.505524861878453,
+      "eval_Vitaminc-test_dot_recall": 0.9891891891891892,
+      "eval_Vitaminc-test_euclidean_accuracy": 0.5657894736842105,
+      "eval_Vitaminc-test_euclidean_accuracy_threshold": 3.1054494380950928,
+      "eval_Vitaminc-test_euclidean_ap": 0.5460311037339075,
+      "eval_Vitaminc-test_euclidean_f1": 0.667870036101083,
+      "eval_Vitaminc-test_euclidean_f1_threshold": 5.527202606201172,
+      "eval_Vitaminc-test_euclidean_precision": 0.5013550135501355,
+      "eval_Vitaminc-test_euclidean_recall": 1.0,
+      "eval_Vitaminc-test_manhattan_accuracy": 0.5657894736842105,
+      "eval_Vitaminc-test_manhattan_accuracy_threshold": 61.94684600830078,
+      "eval_Vitaminc-test_manhattan_ap": 0.5511678056221634,
+      "eval_Vitaminc-test_manhattan_f1": 0.667870036101083,
+      "eval_Vitaminc-test_manhattan_f1_threshold": 113.38865661621094,
+      "eval_Vitaminc-test_manhattan_precision": 0.5013550135501355,
+      "eval_Vitaminc-test_manhattan_recall": 1.0,
+      "eval_Vitaminc-test_max_accuracy": 0.5789473684210527,
+      "eval_Vitaminc-test_max_accuracy_threshold": 61.94684600830078,
+      "eval_Vitaminc-test_max_ap": 0.5599139775011265,
+      "eval_Vitaminc-test_max_f1": 0.6741996233521658,
+      "eval_Vitaminc-test_max_f1_threshold": 113.38865661621094,
+      "eval_Vitaminc-test_max_precision": 0.5173410404624278,
+      "eval_Vitaminc-test_max_recall": 1.0,
+      "eval_mrpc-test_cosine_accuracy": 0.7289473684210527,
+      "eval_mrpc-test_cosine_accuracy_threshold": 0.7165935635566711,
+      "eval_mrpc-test_cosine_ap": 0.8464822906177363,
+      "eval_mrpc-test_cosine_f1": 0.8193979933110368,
+      "eval_mrpc-test_cosine_f1_threshold": 0.6241698265075684,
+      "eval_mrpc-test_cosine_precision": 0.7,
+      "eval_mrpc-test_cosine_recall": 0.9879032258064516,
+      "eval_mrpc-test_dot_accuracy": 0.6894736842105263,
+      "eval_mrpc-test_dot_accuracy_threshold": 6.886929512023926,
+      "eval_mrpc-test_dot_ap": 0.7573697765974963,
+      "eval_mrpc-test_dot_f1": 0.805921052631579,
+      "eval_mrpc-test_dot_f1_threshold": 6.723896026611328,
+      "eval_mrpc-test_dot_precision": 0.6805555555555556,
+      "eval_mrpc-test_dot_recall": 0.9879032258064516,
+      "eval_mrpc-test_euclidean_accuracy": 0.7368421052631579,
+      "eval_mrpc-test_euclidean_accuracy_threshold": 2.9247827529907227,
+      "eval_mrpc-test_euclidean_ap": 0.8222108381881498,
+      "eval_mrpc-test_euclidean_f1": 0.8214285714285715,
+      "eval_mrpc-test_euclidean_f1_threshold": 3.0410943031311035,
+      "eval_mrpc-test_euclidean_precision": 0.7371794871794872,
+      "eval_mrpc-test_euclidean_recall": 0.9274193548387096,
+      "eval_mrpc-test_manhattan_accuracy": 0.7421052631578947,
+      "eval_mrpc-test_manhattan_accuracy_threshold": 59.87488555908203,
+      "eval_mrpc-test_manhattan_ap": 0.8232871898556722,
+      "eval_mrpc-test_manhattan_f1": 0.8262411347517731,
+      "eval_mrpc-test_manhattan_f1_threshold": 59.87488555908203,
+      "eval_mrpc-test_manhattan_precision": 0.7373417721518988,
+      "eval_mrpc-test_manhattan_recall": 0.9395161290322581,
+      "eval_mrpc-test_max_accuracy": 0.7421052631578947,
+      "eval_mrpc-test_max_accuracy_threshold": 59.87488555908203,
+      "eval_mrpc-test_max_ap": 0.8464822906177363,
+      "eval_mrpc-test_max_f1": 0.8262411347517731,
+      "eval_mrpc-test_max_f1_threshold": 59.87488555908203,
+      "eval_mrpc-test_max_precision": 0.7373417721518988,
+      "eval_mrpc-test_max_recall": 0.9879032258064516,
+      "eval_nli-pairs_loss": 0.7530286908149719,
+      "eval_nli-pairs_runtime": 3.1538,
+      "eval_nli-pairs_samples_per_second": 50.732,
+      "eval_nli-pairs_steps_per_second": 1.585,
+      "eval_sequential_score": 0.5599139775011265,
+      "step": 18088
+    },
+    {
+      "epoch": 3.5013550135501355,
+      "eval_vitaminc-pairs_loss": 6.0385422706604,
+      "eval_vitaminc-pairs_runtime": 1.7005,
+      "eval_vitaminc-pairs_samples_per_second": 78.211,
+      "eval_vitaminc-pairs_steps_per_second": 2.94,
+      "step": 18088
+    },
+    {
+      "epoch": 3.5013550135501355,
+      "eval_qnli-contrastive_loss": 0.13678142428398132,
+      "eval_qnli-contrastive_runtime": 0.554,
+      "eval_qnli-contrastive_samples_per_second": 288.79,
+      "eval_qnli-contrastive_steps_per_second": 9.025,
+      "step": 18088
+    },
+    {
+      "epoch": 3.5013550135501355,
+      "eval_scitail-pairs-qa_loss": 0.03590356931090355,
+      "eval_scitail-pairs-qa_runtime": 1.3971,
+      "eval_scitail-pairs-qa_samples_per_second": 114.52,
+      "eval_scitail-pairs-qa_steps_per_second": 3.579,
+      "step": 18088
+    },
+    {
+      "epoch": 3.5013550135501355,
+      "eval_scitail-pairs-pos_loss": 0.27764028310775757,
+      "eval_scitail-pairs-pos_runtime": 2.6797,
+      "eval_scitail-pairs-pos_samples_per_second": 59.709,
+      "eval_scitail-pairs-pos_steps_per_second": 1.866,
+      "step": 18088
+    },
+    {
+      "epoch": 3.5013550135501355,
+      "eval_xsum-pairs_loss": 0.24325545132160187,
+      "eval_xsum-pairs_runtime": 1.4199,
+      "eval_xsum-pairs_samples_per_second": 112.683,
+      "eval_xsum-pairs_steps_per_second": 3.521,
+      "step": 18088
+    },
+    {
+      "epoch": 3.5013550135501355,
+      "eval_compression-pairs_loss": 0.051899444311857224,
+      "eval_compression-pairs_runtime": 0.4875,
+      "eval_compression-pairs_samples_per_second": 328.228,
+      "eval_compression-pairs_steps_per_second": 10.257,
+      "step": 18088
+    },
+    {
+      "epoch": 3.5013550135501355,
+      "eval_sciq_pairs_loss": 0.21445715427398682,
+      "eval_sciq_pairs_runtime": 8.2376,
+      "eval_sciq_pairs_samples_per_second": 19.423,
+      "eval_sciq_pairs_steps_per_second": 0.607,
+      "step": 18088
+    },
+    {
+      "epoch": 3.5013550135501355,
+      "eval_qasc_pairs_loss": 0.1192399114370346,
+      "eval_qasc_pairs_runtime": 1.6296,
+      "eval_qasc_pairs_samples_per_second": 98.182,
+      "eval_qasc_pairs_steps_per_second": 3.068,
+      "step": 18088
+    },
+    {
+      "epoch": 3.5013550135501355,
+      "eval_qasc_facts_sym_loss": 0.11172313988208771,
+      "eval_qasc_facts_sym_runtime": 0.3505,
+      "eval_qasc_facts_sym_samples_per_second": 456.501,
+      "eval_qasc_facts_sym_steps_per_second": 14.266,
+      "step": 18088
+    },
+    {
+      "epoch": 3.5013550135501355,
+      "eval_openbookqa_pairs_loss": 1.5495846271514893,
+      "eval_openbookqa_pairs_runtime": 1.4896,
+      "eval_openbookqa_pairs_samples_per_second": 107.414,
+      "eval_openbookqa_pairs_steps_per_second": 3.357,
+      "step": 18088
+    },
+    {
+      "epoch": 3.5013550135501355,
+      "eval_msmarco_pairs_loss": 0.4486960768699646,
+      "eval_msmarco_pairs_runtime": 3.2657,
+      "eval_msmarco_pairs_samples_per_second": 48.995,
+      "eval_msmarco_pairs_steps_per_second": 1.531,
+      "step": 18088
+    },
+    {
+      "epoch": 3.5013550135501355,
+      "eval_nq_pairs_loss": 0.3250276446342468,
+      "eval_nq_pairs_runtime": 7.7729,
+      "eval_nq_pairs_samples_per_second": 20.584,
+      "eval_nq_pairs_steps_per_second": 0.643,
+      "step": 18088
+    },
+    {
+      "epoch": 3.5013550135501355,
+      "eval_trivia_pairs_loss": 0.56200110912323,
+      "eval_trivia_pairs_runtime": 10.4885,
+      "eval_trivia_pairs_samples_per_second": 15.255,
+      "eval_trivia_pairs_steps_per_second": 0.477,
+      "step": 18088
+    },
+    {
+      "epoch": 3.5013550135501355,
+      "eval_quora_pairs_loss": 0.1906571239233017,
+      "eval_quora_pairs_runtime": 4.1399,
+      "eval_quora_pairs_samples_per_second": 163.049,
+      "eval_quora_pairs_steps_per_second": 5.314,
+      "step": 18088
+    },
+    {
+      "epoch": 3.5013550135501355,
+      "eval_gooaq_pairs_loss": 0.40848416090011597,
+      "eval_gooaq_pairs_runtime": 2.2935,
+      "eval_gooaq_pairs_samples_per_second": 69.763,
+      "eval_gooaq_pairs_steps_per_second": 2.18,
+      "step": 18088
+    },
+    {
+      "epoch": 3.5013550135501355,
+      "eval_mrpc_pairs_loss": 0.024090532213449478,
+      "eval_mrpc_pairs_runtime": 0.3833,
+      "eval_mrpc_pairs_samples_per_second": 417.394,
+      "eval_mrpc_pairs_steps_per_second": 13.044,
+      "step": 18088
+    },
+    {
+      "epoch": 3.510452961672474,
+      "grad_norm": 0.44903191924095154,
+      "learning_rate": 1.024432782384515e-06,
+      "loss": 0.136,
+      "step": 18135
+    },
+    {
+      "epoch": 3.5230352303523036,
+      "grad_norm": 0.4198465645313263,
+      "learning_rate": 8.690770099217205e-07,
+      "loss": 0.1507,
+      "step": 18200
+    },
+    {
+      "epoch": 3.535617499032133,
+      "grad_norm": 9.77443790435791,
+      "learning_rate": 7.26140016574553e-07,
+      "loss": 0.207,
+      "step": 18265
+    },
+    {
+      "epoch": 3.5481997677119628,
+      "grad_norm": 4.438516616821289,
+      "learning_rate": 5.957474206730273e-07,
+      "loss": 0.2192,
+      "step": 18330
+    },
+    {
+      "epoch": 3.5607820363917924,
+      "grad_norm": 1.4462083578109741,
+      "learning_rate": 4.78013816065549e-07,
+      "loss": 0.1574,
+      "step": 18395
+    },
+    {
+      "epoch": 3.573364305071622,
+      "grad_norm": 0.18752352893352509,
+      "learning_rate": 3.730426714095514e-07,
+      "loss": 0.1859,
+      "step": 18460
+    },
+    {
+      "epoch": 3.5859465737514515,
+      "grad_norm": 0.1693798154592514,
+      "learning_rate": 2.809262392394196e-07,
+      "loss": 0.13,
+      "step": 18525
+    },
+    {
+      "epoch": 3.5985288424312816,
+      "grad_norm": 1.442649006843567,
+      "learning_rate": 2.0174547489152985e-07,
+      "loss": 0.1555,
+      "step": 18590
+    },
+    {
+      "epoch": 3.611111111111111,
+      "grad_norm": 2.7645862102508545,
+      "learning_rate": 1.3556996535771416e-07,
+      "loss": 0.2212,
+      "step": 18655
+    },
+    {
+      "epoch": 3.6236933797909407,
+      "grad_norm": 0.19195497035980225,
+      "learning_rate": 8.317578698789685e-08,
+      "loss": 0.1647,
+      "step": 18720
+    },
+    {
+      "epoch": 3.6362756484707703,
+      "grad_norm": 0.15366147458553314,
+      "learning_rate": 4.2971789421488715e-08,
+      "loss": 0.1529,
+      "step": 18785
+    },
+    {
+      "epoch": 3.6488579171506,
+      "grad_norm": 2.829648733139038,
+      "learning_rate": 1.591258287146735e-08,
+      "loss": 0.1512,
+      "step": 18850
+    },
+    {
+      "epoch": 3.66144018583043,
+      "grad_norm": 0.14906300604343414,
+      "learning_rate": 2.0219479719363333e-09,
+      "loss": 0.1788,
+      "step": 18915
+    },
+    {
+      "epoch": 3.6740224545102595,
+      "grad_norm": 0.3509635031223297,
+      "learning_rate": 2.9998687907672072e-05,
+      "loss": 0.184,
+      "step": 18980
+    },
+    {
+      "epoch": 3.686604723190089,
+      "grad_norm": 1.5732251405715942,
+      "learning_rate": 2.9986216360213095e-05,
+      "loss": 0.2365,
+      "step": 19045
+    },
+    {
+      "epoch": 3.6991869918699187,
+      "grad_norm": 4.517373561859131,
+      "learning_rate": 2.9960574370109496e-05,
+      "loss": 0.1695,
+      "step": 19110
+    },
+    {
+      "epoch": 3.7117692605497483,
+      "grad_norm": 2.6720194816589355,
+      "learning_rate": 2.992178447249302e-05,
+      "loss": 0.2396,
+      "step": 19175
+    },
+    {
+      "epoch": 3.724351529229578,
+      "grad_norm": 1.2681653499603271,
+      "learning_rate": 2.986988075736407e-05,
+      "loss": 0.232,
+      "step": 19240
+    },
+    {
+      "epoch": 3.7369337979094075,
+      "grad_norm": 7.320454120635986,
+      "learning_rate": 2.980490883963215e-05,
+      "loss": 0.2263,
+      "step": 19305
+    },
+    {
+      "epoch": 3.749516066589237,
+      "grad_norm": 0.3230324387550354,
+      "learning_rate": 2.9726925819027805e-05,
+      "loss": 0.1583,
+      "step": 19370
+    },
+    {
+      "epoch": 3.751451800232288,
+      "eval_StS-test_pearson_cosine": 0.8727518027563739,
+      "eval_StS-test_pearson_dot": 0.7544441009977498,
+      "eval_StS-test_pearson_euclidean": 0.8528252057552665,
+      "eval_StS-test_pearson_manhattan": 0.8503574183592293,
+      "eval_StS-test_pearson_max": 0.8727518027563739,
+      "eval_StS-test_spearman_cosine": 0.888096795979988,
+      "eval_StS-test_spearman_dot": 0.7570130879859552,
+      "eval_StS-test_spearman_euclidean": 0.858368008662364,
+      "eval_StS-test_spearman_manhattan": 0.8576093894472316,
+      "eval_StS-test_spearman_max": 0.888096795979988,
+      "eval_Vitaminc-test_cosine_accuracy": 0.5815789473684211,
+      "eval_Vitaminc-test_cosine_accuracy_threshold": 0.7381868362426758,
+      "eval_Vitaminc-test_cosine_ap": 0.5598426418690512,
+      "eval_Vitaminc-test_cosine_f1": 0.6716417910447761,
+      "eval_Vitaminc-test_cosine_f1_threshold": 0.4170137643814087,
+      "eval_Vitaminc-test_cosine_precision": 0.5128205128205128,
+      "eval_Vitaminc-test_cosine_recall": 0.972972972972973,
+      "eval_Vitaminc-test_dot_accuracy": 0.5605263157894737,
+      "eval_Vitaminc-test_dot_accuracy_threshold": 13.731411933898926,
+      "eval_Vitaminc-test_dot_ap": 0.5487102325187027,
+      "eval_Vitaminc-test_dot_f1": 0.6691588785046729,
+      "eval_Vitaminc-test_dot_f1_threshold": 7.848942756652832,
+      "eval_Vitaminc-test_dot_precision": 0.5114285714285715,
+      "eval_Vitaminc-test_dot_recall": 0.9675675675675676,
+      "eval_Vitaminc-test_euclidean_accuracy": 0.5789473684210527,
+      "eval_Vitaminc-test_euclidean_accuracy_threshold": 3.16796612739563,
+      "eval_Vitaminc-test_euclidean_ap": 0.5499801975077707,
+      "eval_Vitaminc-test_euclidean_f1": 0.6654676258992805,
+      "eval_Vitaminc-test_euclidean_f1_threshold": 5.5990071296691895,
+      "eval_Vitaminc-test_euclidean_precision": 0.49865229110512127,
+      "eval_Vitaminc-test_euclidean_recall": 1.0,
+      "eval_Vitaminc-test_manhattan_accuracy": 0.5842105263157895,
+      "eval_Vitaminc-test_manhattan_accuracy_threshold": 62.038818359375,
+      "eval_Vitaminc-test_manhattan_ap": 0.5517991149943702,
+      "eval_Vitaminc-test_manhattan_f1": 0.6642599277978339,
+      "eval_Vitaminc-test_manhattan_f1_threshold": 111.26097106933594,
+      "eval_Vitaminc-test_manhattan_precision": 0.4986449864498645,
+      "eval_Vitaminc-test_manhattan_recall": 0.9945945945945946,
+      "eval_Vitaminc-test_max_accuracy": 0.5842105263157895,
+      "eval_Vitaminc-test_max_accuracy_threshold": 62.038818359375,
+      "eval_Vitaminc-test_max_ap": 0.5598426418690512,
+      "eval_Vitaminc-test_max_f1": 0.6716417910447761,
+      "eval_Vitaminc-test_max_f1_threshold": 111.26097106933594,
+      "eval_Vitaminc-test_max_precision": 0.5128205128205128,
+      "eval_Vitaminc-test_max_recall": 1.0,
+      "eval_mrpc-test_cosine_accuracy": 0.7394736842105263,
+      "eval_mrpc-test_cosine_accuracy_threshold": 0.7163187265396118,
+      "eval_mrpc-test_cosine_ap": 0.8453611137786435,
+      "eval_mrpc-test_cosine_f1": 0.8266199649737302,
+      "eval_mrpc-test_cosine_f1_threshold": 0.6975057721138,
+      "eval_mrpc-test_cosine_precision": 0.7306501547987616,
+      "eval_mrpc-test_cosine_recall": 0.9516129032258065,
+      "eval_mrpc-test_dot_accuracy": 0.6894736842105263,
+      "eval_mrpc-test_dot_accuracy_threshold": 7.592863082885742,
+      "eval_mrpc-test_dot_ap": 0.7556545113140425,
+      "eval_mrpc-test_dot_f1": 0.8039538714991763,
+      "eval_mrpc-test_dot_f1_threshold": 7.086015224456787,
+      "eval_mrpc-test_dot_precision": 0.6796657381615598,
+      "eval_mrpc-test_dot_recall": 0.9838709677419355,
+      "eval_mrpc-test_euclidean_accuracy": 0.7394736842105263,
+      "eval_mrpc-test_euclidean_accuracy_threshold": 2.8540396690368652,
+      "eval_mrpc-test_euclidean_ap": 0.8243914030378248,
+      "eval_mrpc-test_euclidean_f1": 0.8186714542190305,
+      "eval_mrpc-test_euclidean_f1_threshold": 2.9808104038238525,
+      "eval_mrpc-test_euclidean_precision": 0.7378640776699029,
+      "eval_mrpc-test_euclidean_recall": 0.9193548387096774,
+      "eval_mrpc-test_manhattan_accuracy": 0.7421052631578947,
+      "eval_mrpc-test_manhattan_accuracy_threshold": 57.382469177246094,
+      "eval_mrpc-test_manhattan_ap": 0.8242227392776778,
+      "eval_mrpc-test_manhattan_f1": 0.8231046931407942,
+      "eval_mrpc-test_manhattan_f1_threshold": 58.107505798339844,
+      "eval_mrpc-test_manhattan_precision": 0.7450980392156863,
+      "eval_mrpc-test_manhattan_recall": 0.9193548387096774,
+      "eval_mrpc-test_max_accuracy": 0.7421052631578947,
+      "eval_mrpc-test_max_accuracy_threshold": 57.382469177246094,
+      "eval_mrpc-test_max_ap": 0.8453611137786435,
+      "eval_mrpc-test_max_f1": 0.8266199649737302,
+      "eval_mrpc-test_max_f1_threshold": 58.107505798339844,
+      "eval_mrpc-test_max_precision": 0.7450980392156863,
+      "eval_mrpc-test_max_recall": 0.9838709677419355,
+      "eval_nli-pairs_loss": 0.8451490998268127,
+      "eval_nli-pairs_runtime": 3.108,
+      "eval_nli-pairs_samples_per_second": 51.479,
+      "eval_nli-pairs_steps_per_second": 1.609,
+      "eval_sequential_score": 0.5598426418690512,
+      "step": 19380
+    },
+    {
+      "epoch": 3.751451800232288,
+      "eval_vitaminc-pairs_loss": 6.042627334594727,
+      "eval_vitaminc-pairs_runtime": 1.5811,
+      "eval_vitaminc-pairs_samples_per_second": 84.119,
+      "eval_vitaminc-pairs_steps_per_second": 3.162,
+      "step": 19380
+    },
+    {
+      "epoch": 3.751451800232288,
+      "eval_qnli-contrastive_loss": 0.14806851744651794,
+      "eval_qnli-contrastive_runtime": 0.5234,
+      "eval_qnli-contrastive_samples_per_second": 305.715,
+      "eval_qnli-contrastive_steps_per_second": 9.554,
+      "step": 19380
+    },
+    {
+      "epoch": 3.751451800232288,
+      "eval_scitail-pairs-qa_loss": 0.03740249201655388,
+      "eval_scitail-pairs-qa_runtime": 1.22,
+      "eval_scitail-pairs-qa_samples_per_second": 131.142,
+      "eval_scitail-pairs-qa_steps_per_second": 4.098,
+      "step": 19380
+    },
+    {
+      "epoch": 3.751451800232288,
+      "eval_scitail-pairs-pos_loss": 0.27747786045074463,
+      "eval_scitail-pairs-pos_runtime": 2.4093,
+      "eval_scitail-pairs-pos_samples_per_second": 66.41,
+      "eval_scitail-pairs-pos_steps_per_second": 2.075,
+      "step": 19380
+    },
+    {
+      "epoch": 3.751451800232288,
+      "eval_xsum-pairs_loss": 0.2729683518409729,
+      "eval_xsum-pairs_runtime": 1.4228,
+      "eval_xsum-pairs_samples_per_second": 112.455,
+      "eval_xsum-pairs_steps_per_second": 3.514,
+      "step": 19380
+    },
+    {
+      "epoch": 3.751451800232288,
+      "eval_compression-pairs_loss": 0.055649857968091965,
+      "eval_compression-pairs_runtime": 0.4125,
+      "eval_compression-pairs_samples_per_second": 387.884,
+      "eval_compression-pairs_steps_per_second": 12.121,
+      "step": 19380
+    },
+    {
+      "epoch": 3.751451800232288,
+      "eval_sciq_pairs_loss": 0.22532808780670166,
+      "eval_sciq_pairs_runtime": 7.9946,
+      "eval_sciq_pairs_samples_per_second": 20.013,
+      "eval_sciq_pairs_steps_per_second": 0.625,
+      "step": 19380
+    },
+    {
+      "epoch": 3.751451800232288,
+      "eval_qasc_pairs_loss": 0.13675269484519958,
+      "eval_qasc_pairs_runtime": 1.4473,
+      "eval_qasc_pairs_samples_per_second": 110.553,
+      "eval_qasc_pairs_steps_per_second": 3.455,
+      "step": 19380
+    },
+    {
+      "epoch": 3.751451800232288,
+      "eval_qasc_facts_sym_loss": 0.10979139804840088,
+      "eval_qasc_facts_sym_runtime": 0.3399,
+      "eval_qasc_facts_sym_samples_per_second": 470.673,
+      "eval_qasc_facts_sym_steps_per_second": 14.709,
+      "step": 19380
+    },
+    {
+      "epoch": 3.751451800232288,
+      "eval_openbookqa_pairs_loss": 1.6146221160888672,
+      "eval_openbookqa_pairs_runtime": 1.2438,
+      "eval_openbookqa_pairs_samples_per_second": 128.633,
+      "eval_openbookqa_pairs_steps_per_second": 4.02,
+      "step": 19380
+    },
+    {
+      "epoch": 3.751451800232288,
+      "eval_msmarco_pairs_loss": 0.5042445063591003,
+      "eval_msmarco_pairs_runtime": 3.1207,
+      "eval_msmarco_pairs_samples_per_second": 51.271,
+      "eval_msmarco_pairs_steps_per_second": 1.602,
+      "step": 19380
+    },
+    {
+      "epoch": 3.751451800232288,
+      "eval_nq_pairs_loss": 0.34877270460128784,
+      "eval_nq_pairs_runtime": 7.7338,
+      "eval_nq_pairs_samples_per_second": 20.688,
+      "eval_nq_pairs_steps_per_second": 0.647,
+      "step": 19380
+    },
+    {
+      "epoch": 3.751451800232288,
+      "eval_trivia_pairs_loss": 0.6838275790214539,
+      "eval_trivia_pairs_runtime": 10.2133,
+      "eval_trivia_pairs_samples_per_second": 15.666,
+      "eval_trivia_pairs_steps_per_second": 0.49,
+      "step": 19380
+    },
+    {
+      "epoch": 3.751451800232288,
+      "eval_quora_pairs_loss": 0.16937024891376495,
+      "eval_quora_pairs_runtime": 3.9408,
+      "eval_quora_pairs_samples_per_second": 171.286,
+      "eval_quora_pairs_steps_per_second": 5.583,
+      "step": 19380
+    },
+    {
+      "epoch": 3.751451800232288,
+      "eval_gooaq_pairs_loss": 0.44292718172073364,
+      "eval_gooaq_pairs_runtime": 2.2878,
+      "eval_gooaq_pairs_samples_per_second": 69.935,
+      "eval_gooaq_pairs_steps_per_second": 2.185,
+      "step": 19380
+    },
+    {
+      "epoch": 3.751451800232288,
+      "eval_mrpc_pairs_loss": 0.025487110018730164,
+      "eval_mrpc_pairs_runtime": 0.3866,
+      "eval_mrpc_pairs_samples_per_second": 413.842,
+      "eval_mrpc_pairs_steps_per_second": 12.933,
+      "step": 19380
+    },
+    {
+      "epoch": 3.762098335269067,
+      "grad_norm": 1.5713428258895874,
+      "learning_rate": 2.9636000229921248e-05,
+      "loss": 0.1775,
+      "step": 19435
+    },
+    {
+      "epoch": 3.7746806039488967,
+      "grad_norm": 1.9037290811538696,
+      "learning_rate": 2.9532211981091813e-05,
+      "loss": 0.179,
+      "step": 19500
+    },
+    {
+      "epoch": 3.7872628726287263,
+      "grad_norm": 0.3006834387779236,
+      "learning_rate": 2.9415652285501175e-05,
+      "loss": 0.225,
+      "step": 19565
+    },
+    {
+      "epoch": 3.799845141308556,
+      "grad_norm": 4.1254754066467285,
+      "learning_rate": 2.9288507113735076e-05,
+      "loss": 0.168,
+      "step": 19630
+    },
+    {
+      "epoch": 3.8124274099883855,
+      "grad_norm": 1.7347090244293213,
+      "learning_rate": 2.9146915214559173e-05,
+      "loss": 0.2078,
+      "step": 19695
+    },
+    {
+      "epoch": 3.8250096786682155,
+      "grad_norm": 0.538968026638031,
+      "learning_rate": 2.899289048170777e-05,
+      "loss": 0.1772,
+      "step": 19760
+    },
+    {
+      "epoch": 3.837591947348045,
+      "grad_norm": 2.847313404083252,
+      "learning_rate": 2.8826568277827517e-05,
+      "loss": 0.1729,
+      "step": 19825
+    },
+    {
+      "epoch": 3.8501742160278747,
+      "grad_norm": 3.284494638442993,
+      "learning_rate": 2.8648094773038627e-05,
+      "loss": 0.1674,
+      "step": 19890
+    },
+    {
+      "epoch": 3.8627564847077043,
+      "grad_norm": 5.920264720916748,
+      "learning_rate": 2.8457626816475117e-05,
+      "loss": 0.1964,
+      "step": 19955
+    },
+    {
+      "epoch": 3.875338753387534,
+      "grad_norm": 0.21167056262493134,
+      "learning_rate": 2.8255331798439983e-05,
+      "loss": 0.1475,
+      "step": 20020
+    },
+    {
+      "epoch": 3.8879210220673635,
+      "grad_norm": 1.4526911973953247,
+      "learning_rate": 2.804138750329644e-05,
+      "loss": 0.2376,
+      "step": 20085
+    },
+    {
+      "epoch": 3.900503290747193,
+      "grad_norm": 4.268107891082764,
+      "learning_rate": 2.7815981953224388e-05,
+      "loss": 0.1967,
+      "step": 20150
+    },
+    {
+      "epoch": 3.9130855594270226,
+      "grad_norm": 1.4916267395019531,
+      "learning_rate": 2.757931324297952e-05,
+      "loss": 0.238,
+      "step": 20215
+    },
+    {
+      "epoch": 3.925667828106852,
+      "grad_norm": 2.0566771030426025,
+      "learning_rate": 2.733158936580037e-05,
+      "loss": 0.2088,
+      "step": 20280
+    },
+    {
+      "epoch": 3.9382500967866823,
+      "grad_norm": 0.14723509550094604,
+      "learning_rate": 2.707302803061613e-05,
+      "loss": 0.2234,
+      "step": 20345
+    },
+    {
+      "epoch": 3.950832365466512,
+      "grad_norm": 4.088009834289551,
+      "learning_rate": 2.68038564707159e-05,
+      "loss": 0.2084,
+      "step": 20410
+    },
+    {
+      "epoch": 3.9634146341463414,
+      "grad_norm": 5.46230936050415,
+      "learning_rate": 2.652431124404776e-05,
+      "loss": 0.276,
+      "step": 20475
+    },
+    {
+      "epoch": 3.975996902826171,
+      "grad_norm": 0.2442116141319275,
+      "learning_rate": 2.6234638025322753e-05,
+      "loss": 0.1641,
+      "step": 20540
+    },
+    {
+      "epoch": 3.9885791715060006,
+      "grad_norm": 3.910892963409424,
+      "learning_rate": 2.593509139010695e-05,
+      "loss": 0.1624,
+      "step": 20605
+    },
+    {
+      "epoch": 4.001161440185831,
+      "grad_norm": 0.3063616156578064,
+      "learning_rate": 2.5625934591090952e-05,
+      "loss": 0.2049,
+      "step": 20670
+    },
+    {
+      "epoch": 4.00154858691444,
+      "eval_StS-test_pearson_cosine": 0.8750931016826595,
+      "eval_StS-test_pearson_dot": 0.7528798867496775,
+      "eval_StS-test_pearson_euclidean": 0.8596040324559132,
+      "eval_StS-test_pearson_manhattan": 0.8576391829193253,
+      "eval_StS-test_pearson_max": 0.8750931016826595,
+      "eval_StS-test_spearman_cosine": 0.8938139167778957,
+      "eval_StS-test_spearman_dot": 0.7453007954801198,
+      "eval_StS-test_spearman_euclidean": 0.8649762059986812,
+      "eval_StS-test_spearman_manhattan": 0.8633117319487409,
+      "eval_StS-test_spearman_max": 0.8938139167778957,
+      "eval_Vitaminc-test_cosine_accuracy": 0.5710526315789474,
+      "eval_Vitaminc-test_cosine_accuracy_threshold": 0.785514235496521,
+      "eval_Vitaminc-test_cosine_ap": 0.5538123804323005,
+      "eval_Vitaminc-test_cosine_f1": 0.6778398510242085,
+      "eval_Vitaminc-test_cosine_f1_threshold": 0.45756226778030396,
+      "eval_Vitaminc-test_cosine_precision": 0.5170454545454546,
+      "eval_Vitaminc-test_cosine_recall": 0.9837837837837838,
+      "eval_Vitaminc-test_dot_accuracy": 0.5552631578947368,
+      "eval_Vitaminc-test_dot_accuracy_threshold": 11.234939575195312,
+      "eval_Vitaminc-test_dot_ap": 0.5407038172991465,
+      "eval_Vitaminc-test_dot_f1": 0.6715328467153285,
+      "eval_Vitaminc-test_dot_f1_threshold": 7.919400215148926,
+      "eval_Vitaminc-test_dot_precision": 0.5068870523415978,
+      "eval_Vitaminc-test_dot_recall": 0.9945945945945946,
+      "eval_Vitaminc-test_euclidean_accuracy": 0.5763157894736842,
+      "eval_Vitaminc-test_euclidean_accuracy_threshold": 3.191226005554199,
+      "eval_Vitaminc-test_euclidean_ap": 0.5517470320076553,
+      "eval_Vitaminc-test_euclidean_f1": 0.6703703703703704,
+      "eval_Vitaminc-test_euclidean_f1_threshold": 5.36379337310791,
+      "eval_Vitaminc-test_euclidean_precision": 0.5098591549295775,
+      "eval_Vitaminc-test_euclidean_recall": 0.9783783783783784,
+      "eval_Vitaminc-test_manhattan_accuracy": 0.5815789473684211,
+      "eval_Vitaminc-test_manhattan_accuracy_threshold": 64.27323913574219,
+      "eval_Vitaminc-test_manhattan_ap": 0.5526594530059742,
+      "eval_Vitaminc-test_manhattan_f1": 0.6654411764705882,
+      "eval_Vitaminc-test_manhattan_f1_threshold": 114.15983581542969,
+      "eval_Vitaminc-test_manhattan_precision": 0.5041782729805014,
+      "eval_Vitaminc-test_manhattan_recall": 0.9783783783783784,
+      "eval_Vitaminc-test_max_accuracy": 0.5815789473684211,
+      "eval_Vitaminc-test_max_accuracy_threshold": 64.27323913574219,
+      "eval_Vitaminc-test_max_ap": 0.5538123804323005,
+      "eval_Vitaminc-test_max_f1": 0.6778398510242085,
+      "eval_Vitaminc-test_max_f1_threshold": 114.15983581542969,
+      "eval_Vitaminc-test_max_precision": 0.5170454545454546,
+      "eval_Vitaminc-test_max_recall": 0.9945945945945946,
+      "eval_mrpc-test_cosine_accuracy": 0.7368421052631579,
+      "eval_mrpc-test_cosine_accuracy_threshold": 0.7361759543418884,
+      "eval_mrpc-test_cosine_ap": 0.8453284249119957,
+      "eval_mrpc-test_cosine_f1": 0.823321554770318,
+      "eval_mrpc-test_cosine_f1_threshold": 0.7002996206283569,
+      "eval_mrpc-test_cosine_precision": 0.7327044025157232,
+      "eval_mrpc-test_cosine_recall": 0.9395161290322581,
+      "eval_mrpc-test_dot_accuracy": 0.6973684210526315,
+      "eval_mrpc-test_dot_accuracy_threshold": 8.213135719299316,
+      "eval_mrpc-test_dot_ap": 0.7568205630034259,
+      "eval_mrpc-test_dot_f1": 0.8099173553719008,
+      "eval_mrpc-test_dot_f1_threshold": 7.905356407165527,
+      "eval_mrpc-test_dot_precision": 0.6862745098039216,
+      "eval_mrpc-test_dot_recall": 0.9879032258064516,
+      "eval_mrpc-test_euclidean_accuracy": 0.7315789473684211,
+      "eval_mrpc-test_euclidean_accuracy_threshold": 2.9826178550720215,
+      "eval_mrpc-test_euclidean_ap": 0.8276910261659405,
+      "eval_mrpc-test_euclidean_f1": 0.8165467625899281,
+      "eval_mrpc-test_euclidean_f1_threshold": 3.032474994659424,
+      "eval_mrpc-test_euclidean_precision": 0.737012987012987,
+      "eval_mrpc-test_euclidean_recall": 0.9153225806451613,
+      "eval_mrpc-test_manhattan_accuracy": 0.7342105263157894,
+      "eval_mrpc-test_manhattan_accuracy_threshold": 58.31708526611328,
+      "eval_mrpc-test_manhattan_ap": 0.8277820763569756,
+      "eval_mrpc-test_manhattan_f1": 0.818018018018018,
+      "eval_mrpc-test_manhattan_f1_threshold": 59.713478088378906,
+      "eval_mrpc-test_manhattan_precision": 0.739413680781759,
+      "eval_mrpc-test_manhattan_recall": 0.9153225806451613,
+      "eval_mrpc-test_max_accuracy": 0.7368421052631579,
+      "eval_mrpc-test_max_accuracy_threshold": 58.31708526611328,
+      "eval_mrpc-test_max_ap": 0.8453284249119957,
+      "eval_mrpc-test_max_f1": 0.823321554770318,
+      "eval_mrpc-test_max_f1_threshold": 59.713478088378906,
+      "eval_mrpc-test_max_precision": 0.739413680781759,
+      "eval_mrpc-test_max_recall": 0.9879032258064516,
+      "eval_nli-pairs_loss": 0.8549426198005676,
+      "eval_nli-pairs_runtime": 3.2134,
+      "eval_nli-pairs_samples_per_second": 49.792,
+      "eval_nli-pairs_steps_per_second": 1.556,
+      "eval_sequential_score": 0.5538123804323005,
+      "step": 20672
+    },
+    {
+      "epoch": 4.00154858691444,
+      "eval_vitaminc-pairs_loss": 5.757671356201172,
+      "eval_vitaminc-pairs_runtime": 1.6437,
+      "eval_vitaminc-pairs_samples_per_second": 80.913,
+      "eval_vitaminc-pairs_steps_per_second": 3.042,
+      "step": 20672
+    },
+    {
+      "epoch": 4.00154858691444,
+      "eval_qnli-contrastive_loss": 0.14272728562355042,
+      "eval_qnli-contrastive_runtime": 0.5575,
+      "eval_qnli-contrastive_samples_per_second": 287.005,
+      "eval_qnli-contrastive_steps_per_second": 8.969,
+      "step": 20672
+    },
+    {
+      "epoch": 4.00154858691444,
+      "eval_scitail-pairs-qa_loss": 0.037104763090610504,
+      "eval_scitail-pairs-qa_runtime": 1.2412,
+      "eval_scitail-pairs-qa_samples_per_second": 128.906,
+      "eval_scitail-pairs-qa_steps_per_second": 4.028,
+      "step": 20672
+    },
+    {
+      "epoch": 4.00154858691444,
+      "eval_scitail-pairs-pos_loss": 0.2630062401294708,
+      "eval_scitail-pairs-pos_runtime": 2.7981,
+      "eval_scitail-pairs-pos_samples_per_second": 57.182,
+      "eval_scitail-pairs-pos_steps_per_second": 1.787,
+      "step": 20672
+    },
+    {
+      "epoch": 4.00154858691444,
+      "eval_xsum-pairs_loss": 0.2699679732322693,
+      "eval_xsum-pairs_runtime": 1.4262,
+      "eval_xsum-pairs_samples_per_second": 112.188,
+      "eval_xsum-pairs_steps_per_second": 3.506,
+      "step": 20672
+    },
+    {
+      "epoch": 4.00154858691444,
+      "eval_compression-pairs_loss": 0.05051081255078316,
+      "eval_compression-pairs_runtime": 0.4202,
+      "eval_compression-pairs_samples_per_second": 380.805,
+      "eval_compression-pairs_steps_per_second": 11.9,
+      "step": 20672
+    },
+    {
+      "epoch": 4.00154858691444,
+      "eval_sciq_pairs_loss": 0.22690464556217194,
+      "eval_sciq_pairs_runtime": 8.2116,
+      "eval_sciq_pairs_samples_per_second": 19.485,
+      "eval_sciq_pairs_steps_per_second": 0.609,
+      "step": 20672
+    },
+    {
+      "epoch": 4.00154858691444,
+      "eval_qasc_pairs_loss": 0.14700038731098175,
+      "eval_qasc_pairs_runtime": 1.5727,
+      "eval_qasc_pairs_samples_per_second": 101.735,
+      "eval_qasc_pairs_steps_per_second": 3.179,
+      "step": 20672
+    },
+    {
+      "epoch": 4.00154858691444,
+      "eval_qasc_facts_sym_loss": 0.1227593868970871,
+      "eval_qasc_facts_sym_runtime": 0.3467,
+      "eval_qasc_facts_sym_samples_per_second": 461.554,
+      "eval_qasc_facts_sym_steps_per_second": 14.424,
+      "step": 20672
+    },
+    {
+      "epoch": 4.00154858691444,
+      "eval_openbookqa_pairs_loss": 1.6408637762069702,
+      "eval_openbookqa_pairs_runtime": 1.382,
+      "eval_openbookqa_pairs_samples_per_second": 115.772,
+      "eval_openbookqa_pairs_steps_per_second": 3.618,
+      "step": 20672
+    },
+    {
+      "epoch": 4.00154858691444,
+      "eval_msmarco_pairs_loss": 0.508281409740448,
+      "eval_msmarco_pairs_runtime": 3.2197,
+      "eval_msmarco_pairs_samples_per_second": 49.694,
+      "eval_msmarco_pairs_steps_per_second": 1.553,
+      "step": 20672
+    },
+    {
+      "epoch": 4.00154858691444,
+      "eval_nq_pairs_loss": 0.4071750044822693,
+      "eval_nq_pairs_runtime": 7.8411,
+      "eval_nq_pairs_samples_per_second": 20.405,
+      "eval_nq_pairs_steps_per_second": 0.638,
+      "step": 20672
+    },
+    {
+      "epoch": 4.00154858691444,
+      "eval_trivia_pairs_loss": 0.6768335103988647,
+      "eval_trivia_pairs_runtime": 10.4692,
+      "eval_trivia_pairs_samples_per_second": 15.283,
+      "eval_trivia_pairs_steps_per_second": 0.478,
+      "step": 20672
+    },
+    {
+      "epoch": 4.00154858691444,
+      "eval_quora_pairs_loss": 0.1442810744047165,
+      "eval_quora_pairs_runtime": 4.0313,
+      "eval_quora_pairs_samples_per_second": 167.439,
+      "eval_quora_pairs_steps_per_second": 5.457,
+      "step": 20672
+    },
+    {
+      "epoch": 4.00154858691444,
+      "eval_gooaq_pairs_loss": 0.3749992251396179,
+      "eval_gooaq_pairs_runtime": 2.3247,
+      "eval_gooaq_pairs_samples_per_second": 68.827,
+      "eval_gooaq_pairs_steps_per_second": 2.151,
+      "step": 20672
+    },
+    {
+      "epoch": 4.00154858691444,
+      "eval_mrpc_pairs_loss": 0.02576417662203312,
+      "eval_mrpc_pairs_runtime": 0.388,
+      "eval_mrpc_pairs_samples_per_second": 412.418,
+      "eval_mrpc_pairs_steps_per_second": 12.888,
+      "step": 20672
+    },
+    {
+      "epoch": 4.01374370886566,
+      "grad_norm": 2.267441987991333,
+      "learning_rate": 2.530743932673362e-05,
+      "loss": 0.1442,
+      "step": 20735
+    },
+    {
+      "epoch": 4.02632597754549,
+      "grad_norm": 5.09822940826416,
+      "learning_rate": 2.497988550248347e-05,
+      "loss": 0.1301,
+      "step": 20800
+    },
+    {
+      "epoch": 4.038908246225319,
+      "grad_norm": 2.3969829082489014,
+      "learning_rate": 2.464356098478738e-05,
+      "loss": 0.2229,
+      "step": 20865
+    },
+    {
+      "epoch": 4.051490514905149,
+      "grad_norm": 0.22940613329410553,
+      "learning_rate": 2.4298761348102794e-05,
+      "loss": 0.2103,
+      "step": 20930
+    },
+    {
+      "epoch": 4.064072783584979,
+      "grad_norm": 1.4287828207015991,
+      "learning_rate": 2.394578961513602e-05,
+      "loss": 0.2235,
+      "step": 20995
+    },
+    {
+      "epoch": 4.076655052264808,
+      "grad_norm": 4.12492036819458,
+      "learning_rate": 2.3584955990534625e-05,
+      "loss": 0.2242,
+      "step": 21060
+    },
+    {
+      "epoch": 4.089237320944638,
+      "grad_norm": 1.9469503164291382,
+      "learning_rate": 2.3216577588268062e-05,
+      "loss": 0.2803,
+      "step": 21125
+    },
+    {
+      "epoch": 4.101819589624467,
+      "grad_norm": 0.6210708618164062,
+      "learning_rate": 2.2840978152936186e-05,
+      "loss": 0.2094,
+      "step": 21190
+    },
+    {
+      "epoch": 4.114401858304297,
+      "grad_norm": 8.642997741699219,
+      "learning_rate": 2.2458487775250408e-05,
+      "loss": 0.2196,
+      "step": 21255
+    },
+    {
+      "epoch": 4.1269841269841265,
+      "grad_norm": 0.306015282869339,
+      "learning_rate": 2.20694426019379e-05,
+      "loss": 0.1926,
+      "step": 21320
+    },
+    {
+      "epoch": 4.139566395663957,
+      "grad_norm": 15.104653358459473,
+      "learning_rate": 2.16741845403232e-05,
+      "loss": 0.1866,
+      "step": 21385
+    },
+    {
+      "epoch": 4.152148664343787,
+      "grad_norm": 16.314868927001953,
+      "learning_rate": 2.127306095784754e-05,
+      "loss": 0.2873,
+      "step": 21450
+    },
+    {
+      "epoch": 4.164730933023616,
+      "grad_norm": 2.250847578048706,
+      "learning_rate": 2.08664243767893e-05,
+      "loss": 0.2281,
+      "step": 21515
+    },
+    {
+      "epoch": 4.177313201703446,
+      "grad_norm": 0.29784318804740906,
+      "learning_rate": 2.0454632164454574e-05,
+      "loss": 0.2539,
+      "step": 21580
+    },
+    {
+      "epoch": 4.189895470383275,
+      "grad_norm": 3.4132087230682373,
+      "learning_rate": 2.003804621910928e-05,
+      "loss": 0.2232,
+      "step": 21645
+    },
+    {
+      "epoch": 4.202477739063105,
+      "grad_norm": 0.33203577995300293,
+      "learning_rate": 1.9617032651929685e-05,
+      "loss": 0.2111,
+      "step": 21710
+    },
+    {
+      "epoch": 4.2150600077429345,
+      "grad_norm": 3.6074330806732178,
+      "learning_rate": 1.919196146525035e-05,
+      "loss": 0.2502,
+      "step": 21775
+    },
+    {
+      "epoch": 4.227642276422764,
+      "grad_norm": 4.078161716461182,
+      "learning_rate": 1.8763206227392307e-05,
+      "loss": 0.2432,
+      "step": 21840
+    },
+    {
+      "epoch": 4.240224545102594,
+      "grad_norm": 6.575666904449463,
+      "learning_rate": 1.8337814009344705e-05,
+      "loss": 0.2261,
+      "step": 21905
+    },
+    {
+      "epoch": 4.251645373596594,
+      "eval_StS-test_pearson_cosine": 0.8738325263775029,
+      "eval_StS-test_pearson_dot": 0.7708615429651562,
+      "eval_StS-test_pearson_euclidean": 0.8601389926129998,
+      "eval_StS-test_pearson_manhattan": 0.8578889010939907,
+      "eval_StS-test_pearson_max": 0.8738325263775029,
+      "eval_StS-test_spearman_cosine": 0.8918479654814222,
+      "eval_StS-test_spearman_dot": 0.7754460023496,
+      "eval_StS-test_spearman_euclidean": 0.864699468715008,
+      "eval_StS-test_spearman_manhattan": 0.8639355949944897,
+      "eval_StS-test_spearman_max": 0.8918479654814222,
+      "eval_Vitaminc-test_cosine_accuracy": 0.5842105263157895,
+      "eval_Vitaminc-test_cosine_accuracy_threshold": 0.6854072213172913,
+      "eval_Vitaminc-test_cosine_ap": 0.5568954471097083,
+      "eval_Vitaminc-test_cosine_f1": 0.6779661016949152,
+      "eval_Vitaminc-test_cosine_f1_threshold": 0.4551093280315399,
+      "eval_Vitaminc-test_cosine_precision": 0.5202312138728323,
+      "eval_Vitaminc-test_cosine_recall": 0.972972972972973,
+      "eval_Vitaminc-test_dot_accuracy": 0.5605263157894737,
+      "eval_Vitaminc-test_dot_accuracy_threshold": 12.5089750289917,
+      "eval_Vitaminc-test_dot_ap": 0.538198865243154,
+      "eval_Vitaminc-test_dot_f1": 0.6715867158671587,
+      "eval_Vitaminc-test_dot_f1_threshold": 7.747126579284668,
+      "eval_Vitaminc-test_dot_precision": 0.5098039215686274,
+      "eval_Vitaminc-test_dot_recall": 0.9837837837837838,
+      "eval_Vitaminc-test_euclidean_accuracy": 0.5657894736842105,
+      "eval_Vitaminc-test_euclidean_accuracy_threshold": 3.2167530059814453,
+      "eval_Vitaminc-test_euclidean_ap": 0.550658853361431,
+      "eval_Vitaminc-test_euclidean_f1": 0.6666666666666666,
+      "eval_Vitaminc-test_euclidean_f1_threshold": 5.500119686126709,
+      "eval_Vitaminc-test_euclidean_precision": 0.5027472527472527,
+      "eval_Vitaminc-test_euclidean_recall": 0.9891891891891892,
+      "eval_Vitaminc-test_manhattan_accuracy": 0.5684210526315789,
+      "eval_Vitaminc-test_manhattan_accuracy_threshold": 63.02822494506836,
+      "eval_Vitaminc-test_manhattan_ap": 0.5526335700218852,
+      "eval_Vitaminc-test_manhattan_f1": 0.6666666666666666,
+      "eval_Vitaminc-test_manhattan_f1_threshold": 113.22319030761719,
+      "eval_Vitaminc-test_manhattan_precision": 0.5027472527472527,
+      "eval_Vitaminc-test_manhattan_recall": 0.9891891891891892,
+      "eval_Vitaminc-test_max_accuracy": 0.5842105263157895,
+      "eval_Vitaminc-test_max_accuracy_threshold": 63.02822494506836,
+      "eval_Vitaminc-test_max_ap": 0.5568954471097083,
+      "eval_Vitaminc-test_max_f1": 0.6779661016949152,
+      "eval_Vitaminc-test_max_f1_threshold": 113.22319030761719,
+      "eval_Vitaminc-test_max_precision": 0.5202312138728323,
+      "eval_Vitaminc-test_max_recall": 0.9891891891891892,
+      "eval_mrpc-test_cosine_accuracy": 0.7368421052631579,
+      "eval_mrpc-test_cosine_accuracy_threshold": 0.7062541842460632,
+      "eval_mrpc-test_cosine_ap": 0.8471734940379884,
+      "eval_mrpc-test_cosine_f1": 0.8226950354609929,
+      "eval_mrpc-test_cosine_f1_threshold": 0.6996989250183105,
+      "eval_mrpc-test_cosine_precision": 0.7341772151898734,
+      "eval_mrpc-test_cosine_recall": 0.9354838709677419,
+      "eval_mrpc-test_dot_accuracy": 0.6921052631578948,
+      "eval_mrpc-test_dot_accuracy_threshold": 7.777188301086426,
+      "eval_mrpc-test_dot_ap": 0.7490093903788431,
+      "eval_mrpc-test_dot_f1": 0.8072487644151566,
+      "eval_mrpc-test_dot_f1_threshold": 6.582016944885254,
+      "eval_mrpc-test_dot_precision": 0.6824512534818942,
+      "eval_mrpc-test_dot_recall": 0.9879032258064516,
+      "eval_mrpc-test_euclidean_accuracy": 0.7342105263157894,
+      "eval_mrpc-test_euclidean_accuracy_threshold": 2.782139539718628,
+      "eval_mrpc-test_euclidean_ap": 0.827030801717203,
+      "eval_mrpc-test_euclidean_f1": 0.8193202146690519,
+      "eval_mrpc-test_euclidean_f1_threshold": 2.9623780250549316,
+      "eval_mrpc-test_euclidean_precision": 0.7363344051446945,
+      "eval_mrpc-test_euclidean_recall": 0.9233870967741935,
+      "eval_mrpc-test_manhattan_accuracy": 0.7421052631578947,
+      "eval_mrpc-test_manhattan_accuracy_threshold": 55.787437438964844,
+      "eval_mrpc-test_manhattan_ap": 0.8262345455513962,
+      "eval_mrpc-test_manhattan_f1": 0.8209764918625678,
+      "eval_mrpc-test_manhattan_f1_threshold": 56.79448318481445,
+      "eval_mrpc-test_manhattan_precision": 0.7442622950819672,
+      "eval_mrpc-test_manhattan_recall": 0.9153225806451613,
+      "eval_mrpc-test_max_accuracy": 0.7421052631578947,
+      "eval_mrpc-test_max_accuracy_threshold": 55.787437438964844,
+      "eval_mrpc-test_max_ap": 0.8471734940379884,
+      "eval_mrpc-test_max_f1": 0.8226950354609929,
+      "eval_mrpc-test_max_f1_threshold": 56.79448318481445,
+      "eval_mrpc-test_max_precision": 0.7442622950819672,
+      "eval_mrpc-test_max_recall": 0.9879032258064516,
+      "eval_nli-pairs_loss": 0.7817401885986328,
+      "eval_nli-pairs_runtime": 3.0345,
+      "eval_nli-pairs_samples_per_second": 52.727,
+      "eval_nli-pairs_steps_per_second": 1.648,
+      "eval_sequential_score": 0.5568954471097083,
+      "step": 21964
+    },
+    {
+      "epoch": 4.251645373596594,
+      "eval_vitaminc-pairs_loss": 5.9666008949279785,
+      "eval_vitaminc-pairs_runtime": 1.5793,
+      "eval_vitaminc-pairs_samples_per_second": 84.215,
+      "eval_vitaminc-pairs_steps_per_second": 3.166,
+      "step": 21964
+    },
+    {
+      "epoch": 4.251645373596594,
+      "eval_qnli-contrastive_loss": 0.10523457825183868,
+      "eval_qnli-contrastive_runtime": 0.5394,
+      "eval_qnli-contrastive_samples_per_second": 296.625,
+      "eval_qnli-contrastive_steps_per_second": 9.27,
+      "step": 21964
+    },
+    {
+      "epoch": 4.251645373596594,
+      "eval_scitail-pairs-qa_loss": 0.03577850013971329,
+      "eval_scitail-pairs-qa_runtime": 1.2305,
+      "eval_scitail-pairs-qa_samples_per_second": 130.026,
+      "eval_scitail-pairs-qa_steps_per_second": 4.063,
+      "step": 21964
+    },
+    {
+      "epoch": 4.251645373596594,
+      "eval_scitail-pairs-pos_loss": 0.2494029849767685,
+      "eval_scitail-pairs-pos_runtime": 2.4173,
+      "eval_scitail-pairs-pos_samples_per_second": 66.188,
+      "eval_scitail-pairs-pos_steps_per_second": 2.068,
+      "step": 21964
+    },
+    {
+      "epoch": 4.251645373596594,
+      "eval_xsum-pairs_loss": 0.26588425040245056,
+      "eval_xsum-pairs_runtime": 1.4156,
+      "eval_xsum-pairs_samples_per_second": 113.027,
+      "eval_xsum-pairs_steps_per_second": 3.532,
+      "step": 21964
+    },
+    {
+      "epoch": 4.251645373596594,
+      "eval_compression-pairs_loss": 0.05197296291589737,
+      "eval_compression-pairs_runtime": 0.4041,
+      "eval_compression-pairs_samples_per_second": 395.981,
+      "eval_compression-pairs_steps_per_second": 12.374,
+      "step": 21964
+    },
+    {
+      "epoch": 4.251645373596594,
+      "eval_sciq_pairs_loss": 0.22115518152713776,
+      "eval_sciq_pairs_runtime": 7.9696,
+      "eval_sciq_pairs_samples_per_second": 20.076,
+      "eval_sciq_pairs_steps_per_second": 0.627,
+      "step": 21964
+    },
+    {
+      "epoch": 4.251645373596594,
+      "eval_qasc_pairs_loss": 0.12823519110679626,
+      "eval_qasc_pairs_runtime": 1.4565,
+      "eval_qasc_pairs_samples_per_second": 109.856,
+      "eval_qasc_pairs_steps_per_second": 3.433,
+      "step": 21964
+    },
+    {
+      "epoch": 4.251645373596594,
+      "eval_qasc_facts_sym_loss": 0.1139301210641861,
+      "eval_qasc_facts_sym_runtime": 0.3301,
+      "eval_qasc_facts_sym_samples_per_second": 484.765,
+      "eval_qasc_facts_sym_steps_per_second": 15.149,
+      "step": 21964
+    },
+    {
+      "epoch": 4.251645373596594,
+      "eval_openbookqa_pairs_loss": 1.6211611032485962,
+      "eval_openbookqa_pairs_runtime": 1.2439,
+      "eval_openbookqa_pairs_samples_per_second": 128.625,
+      "eval_openbookqa_pairs_steps_per_second": 4.02,
+      "step": 21964
+    },
+    {
+      "epoch": 4.251645373596594,
+      "eval_msmarco_pairs_loss": 0.5529596209526062,
+      "eval_msmarco_pairs_runtime": 3.1232,
+      "eval_msmarco_pairs_samples_per_second": 51.229,
+      "eval_msmarco_pairs_steps_per_second": 1.601,
+      "step": 21964
+    },
+    {
+      "epoch": 4.251645373596594,
+      "eval_nq_pairs_loss": 0.36157506704330444,
+      "eval_nq_pairs_runtime": 7.7301,
+      "eval_nq_pairs_samples_per_second": 20.698,
+      "eval_nq_pairs_steps_per_second": 0.647,
+      "step": 21964
+    },
+    {
+      "epoch": 4.251645373596594,
+      "eval_trivia_pairs_loss": 0.6664417386054993,
+      "eval_trivia_pairs_runtime": 10.1737,
+      "eval_trivia_pairs_samples_per_second": 15.727,
+      "eval_trivia_pairs_steps_per_second": 0.491,
+      "step": 21964
+    },
+    {
+      "epoch": 4.251645373596594,
+      "eval_quora_pairs_loss": 0.16971223056316376,
+      "eval_quora_pairs_runtime": 3.9513,
+      "eval_quora_pairs_samples_per_second": 170.83,
+      "eval_quora_pairs_steps_per_second": 5.568,
+      "step": 21964
+    },
+    {
+      "epoch": 4.251645373596594,
+      "eval_gooaq_pairs_loss": 0.4112943112850189,
+      "eval_gooaq_pairs_runtime": 2.2892,
+      "eval_gooaq_pairs_samples_per_second": 69.895,
+      "eval_gooaq_pairs_steps_per_second": 2.184,
+      "step": 21964
+    },
+    {
+      "epoch": 4.251645373596594,
+      "eval_mrpc_pairs_loss": 0.023606743663549423,
+      "eval_mrpc_pairs_runtime": 0.3846,
+      "eval_mrpc_pairs_samples_per_second": 415.966,
+      "eval_mrpc_pairs_steps_per_second": 12.999,
+      "step": 21964
+    },
+    {
+      "epoch": 4.252806813782423,
+      "grad_norm": 0.3098312020301819,
+      "learning_rate": 1.7902866143381558e-05,
+      "loss": 0.2843,
+      "step": 21970
+    },
+    {
+      "epoch": 4.265389082462253,
+      "grad_norm": 1.2247912883758545,
+      "learning_rate": 1.7465367131009776e-05,
+      "loss": 0.2221,
+      "step": 22035
+    },
+    {
+      "epoch": 4.2779713511420825,
+      "grad_norm": 4.876120567321777,
+      "learning_rate": 1.702570146258441e-05,
+      "loss": 0.3073,
+      "step": 22100
+    },
+    {
+      "epoch": 4.290553619821912,
+      "grad_norm": 16.738489151000977,
+      "learning_rate": 1.6584255532598227e-05,
+      "loss": 0.2695,
+      "step": 22165
+    },
+    {
+      "epoch": 4.303135888501743,
+      "grad_norm": 0.21949075162410736,
+      "learning_rate": 1.6141417300103866e-05,
+      "loss": 0.2286,
+      "step": 22230
+    },
+    {
+      "epoch": 4.315718157181572,
+      "grad_norm": 16.98850440979004,
+      "learning_rate": 1.569757594776104e-05,
+      "loss": 0.2236,
+      "step": 22295
+    },
+    {
+      "epoch": 4.328300425861402,
+      "grad_norm": 0.08736617863178253,
+      "learning_rate": 1.525312153980833e-05,
+      "loss": 0.311,
+      "step": 22360
+    },
+    {
+      "epoch": 4.340882694541231,
+      "grad_norm": 0.7681021690368652,
+      "learning_rate": 1.48084446792604e-05,
+      "loss": 0.1523,
+      "step": 22425
+    },
+    {
+      "epoch": 4.353464963221061,
+      "grad_norm": 0.5627409815788269,
+      "learning_rate": 1.4363936164631424e-05,
+      "loss": 0.137,
+      "step": 22490
+    },
+    {
+      "epoch": 4.3660472319008905,
+      "grad_norm": 12.019976615905762,
+      "learning_rate": 1.3919986646487015e-05,
+      "loss": 0.1888,
+      "step": 22555
+    },
+    {
+      "epoch": 4.37862950058072,
+      "grad_norm": 0.4080559015274048,
+      "learning_rate": 1.347698628412585e-05,
+      "loss": 0.1886,
+      "step": 22620
+    },
+    {
+      "epoch": 4.39121176926055,
+      "grad_norm": 1.7257754802703857,
+      "learning_rate": 1.3035324402693452e-05,
+      "loss": 0.1539,
+      "step": 22685
+    },
+    {
+      "epoch": 4.403794037940379,
+      "grad_norm": 1.4404957294464111,
+      "learning_rate": 1.2595389151028544e-05,
+      "loss": 0.1141,
+      "step": 22750
+    },
+    {
+      "epoch": 4.416376306620209,
+      "grad_norm": 0.5989357829093933,
+      "learning_rate": 1.2157567160543606e-05,
+      "loss": 0.2414,
+      "step": 22815
+    },
+    {
+      "epoch": 4.4289585753000384,
+      "grad_norm": 1.9353371858596802,
+      "learning_rate": 1.1722243205438878e-05,
+      "loss": 0.1844,
+      "step": 22880
+    },
+    {
+      "epoch": 4.441540843979868,
+      "grad_norm": 0.2850472033023834,
+      "learning_rate": 1.1289799864548468e-05,
+      "loss": 0.1882,
+      "step": 22945
+    },
+    {
+      "epoch": 4.454123112659698,
+      "grad_norm": 4.457968235015869,
+      "learning_rate": 1.0860617185115858e-05,
+      "loss": 0.2267,
+      "step": 23010
+    },
+    {
+      "epoch": 4.466705381339528,
+      "grad_norm": 2.144094944000244,
+      "learning_rate": 1.0435072348794313e-05,
+      "loss": 0.1586,
+      "step": 23075
+    },
+    {
+      "epoch": 4.479287650019358,
+      "grad_norm": 3.16145658493042,
+      "learning_rate": 1.0013539340165799e-05,
+      "loss": 0.1833,
+      "step": 23140
+    },
+    {
+      "epoch": 4.491869918699187,
+      "grad_norm": 5.41245174407959,
+      "learning_rate": 9.596388618069195e-06,
+      "loss": 0.1505,
+      "step": 23205
+    },
+    {
+      "epoch": 4.501742160278746,
+      "eval_StS-test_pearson_cosine": 0.8816702865221286,
+      "eval_StS-test_pearson_dot": 0.7894071456966113,
+      "eval_StS-test_pearson_euclidean": 0.8662344171069792,
+      "eval_StS-test_pearson_manhattan": 0.8645345748649402,
+      "eval_StS-test_pearson_max": 0.8816702865221286,
+      "eval_StS-test_spearman_cosine": 0.8968976546264207,
+      "eval_StS-test_spearman_dot": 0.7860133600255552,
+      "eval_StS-test_spearman_euclidean": 0.8683799213314533,
+      "eval_StS-test_spearman_manhattan": 0.8675563870810289,
+      "eval_StS-test_spearman_max": 0.8968976546264207,
+      "eval_Vitaminc-test_cosine_accuracy": 0.5842105263157895,
+      "eval_Vitaminc-test_cosine_accuracy_threshold": 0.6739277243614197,
+      "eval_Vitaminc-test_cosine_ap": 0.5589703949384612,
+      "eval_Vitaminc-test_cosine_f1": 0.6716981132075472,
+      "eval_Vitaminc-test_cosine_f1_threshold": 0.432912677526474,
+      "eval_Vitaminc-test_cosine_precision": 0.5159420289855072,
+      "eval_Vitaminc-test_cosine_recall": 0.9621621621621622,
+      "eval_Vitaminc-test_dot_accuracy": 0.5578947368421052,
+      "eval_Vitaminc-test_dot_accuracy_threshold": 10.692544937133789,
+      "eval_Vitaminc-test_dot_ap": 0.5593745513398436,
+      "eval_Vitaminc-test_dot_f1": 0.6691312384473197,
+      "eval_Vitaminc-test_dot_f1_threshold": 6.927554607391357,
+      "eval_Vitaminc-test_dot_precision": 0.5084269662921348,
+      "eval_Vitaminc-test_dot_recall": 0.9783783783783784,
+      "eval_Vitaminc-test_euclidean_accuracy": 0.5736842105263158,
+      "eval_Vitaminc-test_euclidean_accuracy_threshold": 2.90975284576416,
+      "eval_Vitaminc-test_euclidean_ap": 0.5526828083289721,
+      "eval_Vitaminc-test_euclidean_f1": 0.6666666666666666,
+      "eval_Vitaminc-test_euclidean_f1_threshold": 5.231945037841797,
+      "eval_Vitaminc-test_euclidean_precision": 0.5013623978201635,
+      "eval_Vitaminc-test_euclidean_recall": 0.9945945945945946,
+      "eval_Vitaminc-test_manhattan_accuracy": 0.5736842105263158,
+      "eval_Vitaminc-test_manhattan_accuracy_threshold": 59.42338180541992,
+      "eval_Vitaminc-test_manhattan_ap": 0.5548476157371564,
+      "eval_Vitaminc-test_manhattan_f1": 0.6678765880217786,
+      "eval_Vitaminc-test_manhattan_f1_threshold": 106.84965515136719,
+      "eval_Vitaminc-test_manhattan_precision": 0.5027322404371585,
+      "eval_Vitaminc-test_manhattan_recall": 0.9945945945945946,
+      "eval_Vitaminc-test_max_accuracy": 0.5842105263157895,
+      "eval_Vitaminc-test_max_accuracy_threshold": 59.42338180541992,
+      "eval_Vitaminc-test_max_ap": 0.5593745513398436,
+      "eval_Vitaminc-test_max_f1": 0.6716981132075472,
+      "eval_Vitaminc-test_max_f1_threshold": 106.84965515136719,
+      "eval_Vitaminc-test_max_precision": 0.5159420289855072,
+      "eval_Vitaminc-test_max_recall": 0.9945945945945946,
+      "eval_mrpc-test_cosine_accuracy": 0.7289473684210527,
+      "eval_mrpc-test_cosine_accuracy_threshold": 0.7656036615371704,
+      "eval_mrpc-test_cosine_ap": 0.8388786109810218,
+      "eval_mrpc-test_cosine_f1": 0.8180300500834724,
+      "eval_mrpc-test_cosine_f1_threshold": 0.6135265827178955,
+      "eval_mrpc-test_cosine_precision": 0.698005698005698,
+      "eval_mrpc-test_cosine_recall": 0.9879032258064516,
+      "eval_mrpc-test_dot_accuracy": 0.6947368421052632,
+      "eval_mrpc-test_dot_accuracy_threshold": 7.60512638092041,
+      "eval_mrpc-test_dot_ap": 0.7526445714043273,
+      "eval_mrpc-test_dot_f1": 0.804635761589404,
+      "eval_mrpc-test_dot_f1_threshold": 6.214287757873535,
+      "eval_mrpc-test_dot_precision": 0.6825842696629213,
+      "eval_mrpc-test_dot_recall": 0.9798387096774194,
+      "eval_mrpc-test_euclidean_accuracy": 0.7289473684210527,
+      "eval_mrpc-test_euclidean_accuracy_threshold": 2.930023670196533,
+      "eval_mrpc-test_euclidean_ap": 0.8236665039155787,
+      "eval_mrpc-test_euclidean_f1": 0.8189806678383128,
+      "eval_mrpc-test_euclidean_f1_threshold": 2.930023670196533,
+      "eval_mrpc-test_euclidean_precision": 0.7258566978193146,
+      "eval_mrpc-test_euclidean_recall": 0.9395161290322581,
+      "eval_mrpc-test_manhattan_accuracy": 0.7342105263157894,
+      "eval_mrpc-test_manhattan_accuracy_threshold": 55.630340576171875,
+      "eval_mrpc-test_manhattan_ap": 0.8215020127627809,
+      "eval_mrpc-test_manhattan_f1": 0.8188153310104529,
+      "eval_mrpc-test_manhattan_f1_threshold": 59.091949462890625,
+      "eval_mrpc-test_manhattan_precision": 0.7208588957055214,
+      "eval_mrpc-test_manhattan_recall": 0.9475806451612904,
+      "eval_mrpc-test_max_accuracy": 0.7342105263157894,
+      "eval_mrpc-test_max_accuracy_threshold": 55.630340576171875,
+      "eval_mrpc-test_max_ap": 0.8388786109810218,
+      "eval_mrpc-test_max_f1": 0.8189806678383128,
+      "eval_mrpc-test_max_f1_threshold": 59.091949462890625,
+      "eval_mrpc-test_max_precision": 0.7258566978193146,
+      "eval_mrpc-test_max_recall": 0.9879032258064516,
+      "eval_nli-pairs_loss": 0.7658803462982178,
+      "eval_nli-pairs_runtime": 3.343,
+      "eval_nli-pairs_samples_per_second": 47.862,
+      "eval_nli-pairs_steps_per_second": 1.496,
+      "eval_sequential_score": 0.5593745513398436,
+      "step": 23256
+    },
+    {
+      "epoch": 4.501742160278746,
+      "eval_vitaminc-pairs_loss": 6.087721824645996,
+      "eval_vitaminc-pairs_runtime": 1.6495,
+      "eval_vitaminc-pairs_samples_per_second": 80.629,
+      "eval_vitaminc-pairs_steps_per_second": 3.031,
+      "step": 23256
+    },
+    {
+      "epoch": 4.501742160278746,
+      "eval_qnli-contrastive_loss": 0.16503135859966278,
+      "eval_qnli-contrastive_runtime": 0.5683,
+      "eval_qnli-contrastive_samples_per_second": 281.564,
+      "eval_qnli-contrastive_steps_per_second": 8.799,
+      "step": 23256
+    },
+    {
+      "epoch": 4.501742160278746,
+      "eval_scitail-pairs-qa_loss": 0.0311884768307209,
+      "eval_scitail-pairs-qa_runtime": 1.3796,
+      "eval_scitail-pairs-qa_samples_per_second": 115.973,
+      "eval_scitail-pairs-qa_steps_per_second": 3.624,
+      "step": 23256
+    },
+    {
+      "epoch": 4.501742160278746,
+      "eval_scitail-pairs-pos_loss": 0.2662835419178009,
+      "eval_scitail-pairs-pos_runtime": 2.8353,
+      "eval_scitail-pairs-pos_samples_per_second": 56.432,
+      "eval_scitail-pairs-pos_steps_per_second": 1.763,
+      "step": 23256
+    },
+    {
+      "epoch": 4.501742160278746,
+      "eval_xsum-pairs_loss": 0.2549589276313782,
+      "eval_xsum-pairs_runtime": 1.48,
+      "eval_xsum-pairs_samples_per_second": 108.107,
+      "eval_xsum-pairs_steps_per_second": 3.378,
+      "step": 23256
+    },
+    {
+      "epoch": 4.501742160278746,
+      "eval_compression-pairs_loss": 0.0466877818107605,
+      "eval_compression-pairs_runtime": 0.4094,
+      "eval_compression-pairs_samples_per_second": 390.835,
+      "eval_compression-pairs_steps_per_second": 12.214,
+      "step": 23256
+    },
+    {
+      "epoch": 4.501742160278746,
+      "eval_sciq_pairs_loss": 0.21744883060455322,
+      "eval_sciq_pairs_runtime": 8.389,
+      "eval_sciq_pairs_samples_per_second": 19.073,
+      "eval_sciq_pairs_steps_per_second": 0.596,
+      "step": 23256
+    },
+    {
+      "epoch": 4.501742160278746,
+      "eval_qasc_pairs_loss": 0.11268162727355957,
+      "eval_qasc_pairs_runtime": 1.5933,
+      "eval_qasc_pairs_samples_per_second": 100.418,
+      "eval_qasc_pairs_steps_per_second": 3.138,
+      "step": 23256
+    },
+    {
+      "epoch": 4.501742160278746,
+      "eval_qasc_facts_sym_loss": 0.10673767328262329,
+      "eval_qasc_facts_sym_runtime": 0.3455,
+      "eval_qasc_facts_sym_samples_per_second": 463.139,
+      "eval_qasc_facts_sym_steps_per_second": 14.473,
+      "step": 23256
+    },
+    {
+      "epoch": 4.501742160278746,
+      "eval_openbookqa_pairs_loss": 1.5697686672210693,
+      "eval_openbookqa_pairs_runtime": 1.3837,
+      "eval_openbookqa_pairs_samples_per_second": 115.633,
+      "eval_openbookqa_pairs_steps_per_second": 3.614,
+      "step": 23256
+    },
+    {
+      "epoch": 4.501742160278746,
+      "eval_msmarco_pairs_loss": 0.4660285413265228,
+      "eval_msmarco_pairs_runtime": 3.212,
+      "eval_msmarco_pairs_samples_per_second": 49.813,
+      "eval_msmarco_pairs_steps_per_second": 1.557,
+      "step": 23256
+    },
+    {
+      "epoch": 4.501742160278746,
+      "eval_nq_pairs_loss": 0.3551401197910309,
+      "eval_nq_pairs_runtime": 7.8628,
+      "eval_nq_pairs_samples_per_second": 20.349,
+      "eval_nq_pairs_steps_per_second": 0.636,
+      "step": 23256
+    },
+    {
+      "epoch": 4.501742160278746,
+      "eval_trivia_pairs_loss": 0.557299017906189,
+      "eval_trivia_pairs_runtime": 10.3596,
+      "eval_trivia_pairs_samples_per_second": 15.445,
+      "eval_trivia_pairs_steps_per_second": 0.483,
+      "step": 23256
+    },
+    {
+      "epoch": 4.501742160278746,
+      "eval_quora_pairs_loss": 0.17031456530094147,
+      "eval_quora_pairs_runtime": 4.0048,
+      "eval_quora_pairs_samples_per_second": 168.546,
+      "eval_quora_pairs_steps_per_second": 5.493,
+      "step": 23256
+    },
+    {
+      "epoch": 4.501742160278746,
+      "eval_gooaq_pairs_loss": 0.36391741037368774,
+      "eval_gooaq_pairs_runtime": 2.3095,
+      "eval_gooaq_pairs_samples_per_second": 69.278,
+      "eval_gooaq_pairs_steps_per_second": 2.165,
+      "step": 23256
+    },
+    {
+      "epoch": 4.501742160278746,
+      "eval_mrpc_pairs_loss": 0.021536488085985184,
+      "eval_mrpc_pairs_runtime": 0.3852,
+      "eval_mrpc_pairs_samples_per_second": 415.417,
+      "eval_mrpc_pairs_steps_per_second": 12.982,
+      "step": 23256
+    },
+    {
+      "epoch": 4.504452187379017,
+      "grad_norm": 0.9454036951065063,
+      "learning_rate": 9.183986790027585e-06,
+      "loss": 0.1539,
+      "step": 23270
+    },
+    {
+      "epoch": 4.5170344560588465,
+      "grad_norm": 0.7937942147254944,
+      "learning_rate": 8.776696290059776e-06,
+      "loss": 0.1661,
+      "step": 23335
+    },
+    {
+      "epoch": 4.529616724738676,
+      "grad_norm": 2.6932709217071533,
+      "learning_rate": 8.374875060160005e-06,
+      "loss": 0.1719,
+      "step": 23400
+    },
+    {
+      "epoch": 4.542198993418506,
+      "grad_norm": 1.9167306423187256,
+      "learning_rate": 7.978876235725097e-06,
+      "loss": 0.2357,
+      "step": 23465
+    },
+    {
+      "epoch": 4.554781262098335,
+      "grad_norm": 0.4528341591358185,
+      "learning_rate": 7.5890478352061e-06,
+      "loss": 0.1764,
+      "step": 23530
+    },
+    {
+      "epoch": 4.567363530778165,
+      "grad_norm": 15.758140563964844,
+      "learning_rate": 7.205732454256714e-06,
+      "loss": 0.1741,
+      "step": 23595
+    },
+    {
+      "epoch": 4.579945799457994,
+      "grad_norm": 0.13203494250774384,
+      "learning_rate": 6.8292669646475355e-06,
+      "loss": 0.1251,
+      "step": 23660
+    },
+    {
+      "epoch": 4.592528068137824,
+      "grad_norm": 0.5477350354194641,
+      "learning_rate": 6.459982218210615e-06,
+      "loss": 0.1734,
+      "step": 23725
+    },
+    {
+      "epoch": 4.605110336817654,
+      "grad_norm": 0.11947894096374512,
+      "learning_rate": 6.098202756074636e-06,
+      "loss": 0.1539,
+      "step": 23790
+    },
+    {
+      "epoch": 4.617692605497483,
+      "grad_norm": 2.542245864868164,
+      "learning_rate": 5.744246523446167e-06,
+      "loss": 0.1804,
+      "step": 23855
+    },
+    {
+      "epoch": 4.630274874177314,
+      "grad_norm": 1.1431196928024292,
+      "learning_rate": 5.398424590187569e-06,
+      "loss": 0.1678,
+      "step": 23920
+    },
+    {
+      "epoch": 4.642857142857143,
+      "grad_norm": 3.039541721343994,
+      "learning_rate": 5.061040877437311e-06,
+      "loss": 0.1526,
+      "step": 23985
+    },
+    {
+      "epoch": 4.655439411536973,
+      "grad_norm": 6.675982475280762,
+      "learning_rate": 4.7323918905127275e-06,
+      "loss": 0.1954,
+      "step": 24050
+    },
+    {
+      "epoch": 4.668021680216802,
+      "grad_norm": 11.611279487609863,
+      "learning_rate": 4.412766458330265e-06,
+      "loss": 0.1876,
+      "step": 24115
+    },
+    {
+      "epoch": 4.680603948896632,
+      "grad_norm": 1.0078673362731934,
+      "learning_rate": 4.1024454795717974e-06,
+      "loss": 0.1567,
+      "step": 24180
+    },
+    {
+      "epoch": 4.693186217576462,
+      "grad_norm": 0.6833789944648743,
+      "learning_rate": 3.801701675820568e-06,
+      "loss": 0.0897,
+      "step": 24245
+    },
+    {
+      "epoch": 4.705768486256291,
+      "grad_norm": 0.08990409970283508,
+      "learning_rate": 3.5107993518832366e-06,
+      "loss": 0.1338,
+      "step": 24310
+    },
+    {
+      "epoch": 4.718350754936121,
+      "grad_norm": 2.577784299850464,
+      "learning_rate": 3.2299941635091016e-06,
+      "loss": 0.1611,
+      "step": 24375
+    },
+    {
+      "epoch": 4.73093302361595,
+      "grad_norm": 2.1565229892730713,
+      "learning_rate": 2.959532892710328e-06,
+      "loss": 0.1611,
+      "step": 24440
+    },
+    {
+      "epoch": 4.74351529229578,
+      "grad_norm": 1.29633367061615,
+      "learning_rate": 2.699653230880698e-06,
+      "loss": 0.1218,
+      "step": 24505
+    },
+    {
+      "epoch": 4.751838946960898,
+      "eval_StS-test_pearson_cosine": 0.8792609870436672,
+      "eval_StS-test_pearson_dot": 0.7762148715297421,
+      "eval_StS-test_pearson_euclidean": 0.8598391820879342,
+      "eval_StS-test_pearson_manhattan": 0.8576575270233212,
+      "eval_StS-test_pearson_max": 0.8792609870436672,
+      "eval_StS-test_spearman_cosine": 0.895012053474811,
+      "eval_StS-test_spearman_dot": 0.7765599574914741,
+      "eval_StS-test_spearman_euclidean": 0.8626746231706642,
+      "eval_StS-test_spearman_manhattan": 0.8613839852851796,
+      "eval_StS-test_spearman_max": 0.895012053474811,
+      "eval_Vitaminc-test_cosine_accuracy": 0.5763157894736842,
+      "eval_Vitaminc-test_cosine_accuracy_threshold": 0.7384297847747803,
+      "eval_Vitaminc-test_cosine_ap": 0.5582610977968616,
+      "eval_Vitaminc-test_cosine_f1": 0.6716981132075472,
+      "eval_Vitaminc-test_cosine_f1_threshold": 0.4387393295764923,
+      "eval_Vitaminc-test_cosine_precision": 0.5159420289855072,
+      "eval_Vitaminc-test_cosine_recall": 0.9621621621621622,
+      "eval_Vitaminc-test_dot_accuracy": 0.5526315789473685,
+      "eval_Vitaminc-test_dot_accuracy_threshold": 10.325508117675781,
+      "eval_Vitaminc-test_dot_ap": 0.558739606374971,
+      "eval_Vitaminc-test_dot_f1": 0.6704545454545454,
+      "eval_Vitaminc-test_dot_f1_threshold": 6.89784049987793,
+      "eval_Vitaminc-test_dot_precision": 0.5160349854227405,
+      "eval_Vitaminc-test_dot_recall": 0.9567567567567568,
+      "eval_Vitaminc-test_euclidean_accuracy": 0.5842105263157895,
+      "eval_Vitaminc-test_euclidean_accuracy_threshold": 2.8687446117401123,
+      "eval_Vitaminc-test_euclidean_ap": 0.5521974352836823,
+      "eval_Vitaminc-test_euclidean_f1": 0.6654478976234005,
+      "eval_Vitaminc-test_euclidean_f1_threshold": 4.835182189941406,
+      "eval_Vitaminc-test_euclidean_precision": 0.5027624309392266,
+      "eval_Vitaminc-test_euclidean_recall": 0.9837837837837838,
+      "eval_Vitaminc-test_manhattan_accuracy": 0.5789473684210527,
+      "eval_Vitaminc-test_manhattan_accuracy_threshold": 56.989898681640625,
+      "eval_Vitaminc-test_manhattan_ap": 0.5519449989523413,
+      "eval_Vitaminc-test_manhattan_f1": 0.6654611211573237,
+      "eval_Vitaminc-test_manhattan_f1_threshold": 102.37645721435547,
+      "eval_Vitaminc-test_manhattan_precision": 0.5,
+      "eval_Vitaminc-test_manhattan_recall": 0.9945945945945946,
+      "eval_Vitaminc-test_max_accuracy": 0.5842105263157895,
+      "eval_Vitaminc-test_max_accuracy_threshold": 56.989898681640625,
+      "eval_Vitaminc-test_max_ap": 0.558739606374971,
+      "eval_Vitaminc-test_max_f1": 0.6716981132075472,
+      "eval_Vitaminc-test_max_f1_threshold": 102.37645721435547,
+      "eval_Vitaminc-test_max_precision": 0.5160349854227405,
+      "eval_Vitaminc-test_max_recall": 0.9945945945945946,
+      "eval_mrpc-test_cosine_accuracy": 0.7315789473684211,
+      "eval_mrpc-test_cosine_accuracy_threshold": 0.7699094414710999,
+      "eval_mrpc-test_cosine_ap": 0.8427559709422051,
+      "eval_mrpc-test_cosine_f1": 0.8216783216783218,
+      "eval_mrpc-test_cosine_f1_threshold": 0.6925866007804871,
+      "eval_mrpc-test_cosine_precision": 0.7253086419753086,
+      "eval_mrpc-test_cosine_recall": 0.9475806451612904,
+      "eval_mrpc-test_dot_accuracy": 0.6973684210526315,
+      "eval_mrpc-test_dot_accuracy_threshold": 6.628315448760986,
+      "eval_mrpc-test_dot_ap": 0.7544902048971998,
+      "eval_mrpc-test_dot_f1": 0.8040885860306644,
+      "eval_mrpc-test_dot_f1_threshold": 6.628315448760986,
+      "eval_mrpc-test_dot_precision": 0.696165191740413,
+      "eval_mrpc-test_dot_recall": 0.9516129032258065,
+      "eval_mrpc-test_euclidean_accuracy": 0.7289473684210527,
+      "eval_mrpc-test_euclidean_accuracy_threshold": 2.609426975250244,
+      "eval_mrpc-test_euclidean_ap": 0.825937819365436,
+      "eval_mrpc-test_euclidean_f1": 0.8148148148148148,
+      "eval_mrpc-test_euclidean_f1_threshold": 2.775911569595337,
+      "eval_mrpc-test_euclidean_precision": 0.7241379310344828,
+      "eval_mrpc-test_euclidean_recall": 0.9314516129032258,
+      "eval_mrpc-test_manhattan_accuracy": 0.7342105263157894,
+      "eval_mrpc-test_manhattan_accuracy_threshold": 52.07958221435547,
+      "eval_mrpc-test_manhattan_ap": 0.8243923902541919,
+      "eval_mrpc-test_manhattan_f1": 0.8185053380782917,
+      "eval_mrpc-test_manhattan_f1_threshold": 53.8984489440918,
+      "eval_mrpc-test_manhattan_precision": 0.732484076433121,
+      "eval_mrpc-test_manhattan_recall": 0.9274193548387096,
+      "eval_mrpc-test_max_accuracy": 0.7342105263157894,
+      "eval_mrpc-test_max_accuracy_threshold": 52.07958221435547,
+      "eval_mrpc-test_max_ap": 0.8427559709422051,
+      "eval_mrpc-test_max_f1": 0.8216783216783218,
+      "eval_mrpc-test_max_f1_threshold": 53.8984489440918,
+      "eval_mrpc-test_max_precision": 0.732484076433121,
+      "eval_mrpc-test_max_recall": 0.9516129032258065,
+      "eval_nli-pairs_loss": 0.7502982020378113,
+      "eval_nli-pairs_runtime": 3.0243,
+      "eval_nli-pairs_samples_per_second": 52.905,
+      "eval_nli-pairs_steps_per_second": 1.653,
+      "eval_sequential_score": 0.558739606374971,
+      "step": 24548
+    },
+    {
+      "epoch": 4.751838946960898,
+      "eval_vitaminc-pairs_loss": 6.014657497406006,
+      "eval_vitaminc-pairs_runtime": 1.6124,
+      "eval_vitaminc-pairs_samples_per_second": 82.484,
+      "eval_vitaminc-pairs_steps_per_second": 3.101,
+      "step": 24548
+    },
+    {
+      "epoch": 4.751838946960898,
+      "eval_qnli-contrastive_loss": 0.1470354050397873,
+      "eval_qnli-contrastive_runtime": 0.5367,
+      "eval_qnli-contrastive_samples_per_second": 298.11,
+      "eval_qnli-contrastive_steps_per_second": 9.316,
+      "step": 24548
+    },
+    {
+      "epoch": 4.751838946960898,
+      "eval_scitail-pairs-qa_loss": 0.028492147102952003,
+      "eval_scitail-pairs-qa_runtime": 1.2207,
+      "eval_scitail-pairs-qa_samples_per_second": 131.069,
+      "eval_scitail-pairs-qa_steps_per_second": 4.096,
+      "step": 24548
+    },
+    {
+      "epoch": 4.751838946960898,
+      "eval_scitail-pairs-pos_loss": 0.2434484362602234,
+      "eval_scitail-pairs-pos_runtime": 2.394,
+      "eval_scitail-pairs-pos_samples_per_second": 66.835,
+      "eval_scitail-pairs-pos_steps_per_second": 2.089,
+      "step": 24548
+    },
+    {
+      "epoch": 4.751838946960898,
+      "eval_xsum-pairs_loss": 0.2599203288555145,
+      "eval_xsum-pairs_runtime": 1.4307,
+      "eval_xsum-pairs_samples_per_second": 111.832,
+      "eval_xsum-pairs_steps_per_second": 3.495,
+      "step": 24548
+    },
+    {
+      "epoch": 4.751838946960898,
+      "eval_compression-pairs_loss": 0.04551355913281441,
+      "eval_compression-pairs_runtime": 0.4296,
+      "eval_compression-pairs_samples_per_second": 372.43,
+      "eval_compression-pairs_steps_per_second": 11.638,
+      "step": 24548
+    },
+    {
+      "epoch": 4.751838946960898,
+      "eval_sciq_pairs_loss": 0.2139088362455368,
+      "eval_sciq_pairs_runtime": 8.0693,
+      "eval_sciq_pairs_samples_per_second": 19.828,
+      "eval_sciq_pairs_steps_per_second": 0.62,
+      "step": 24548
+    },
+    {
+      "epoch": 4.751838946960898,
+      "eval_qasc_pairs_loss": 0.11057784408330917,
+      "eval_qasc_pairs_runtime": 1.4817,
+      "eval_qasc_pairs_samples_per_second": 107.982,
+      "eval_qasc_pairs_steps_per_second": 3.374,
+      "step": 24548
+    },
+    {
+      "epoch": 4.751838946960898,
+      "eval_qasc_facts_sym_loss": 0.10339337587356567,
+      "eval_qasc_facts_sym_runtime": 0.3423,
+      "eval_qasc_facts_sym_samples_per_second": 467.489,
+      "eval_qasc_facts_sym_steps_per_second": 14.609,
+      "step": 24548
+    },
+    {
+      "epoch": 4.751838946960898,
+      "eval_openbookqa_pairs_loss": 1.5395641326904297,
+      "eval_openbookqa_pairs_runtime": 1.2771,
+      "eval_openbookqa_pairs_samples_per_second": 125.281,
+      "eval_openbookqa_pairs_steps_per_second": 3.915,
+      "step": 24548
+    },
+    {
+      "epoch": 4.751838946960898,
+      "eval_msmarco_pairs_loss": 0.4512082636356354,
+      "eval_msmarco_pairs_runtime": 3.1793,
+      "eval_msmarco_pairs_samples_per_second": 50.326,
+      "eval_msmarco_pairs_steps_per_second": 1.573,
+      "step": 24548
+    },
+    {
+      "epoch": 4.751838946960898,
+      "eval_nq_pairs_loss": 0.3511877954006195,
+      "eval_nq_pairs_runtime": 7.7842,
+      "eval_nq_pairs_samples_per_second": 20.554,
+      "eval_nq_pairs_steps_per_second": 0.642,
+      "step": 24548
+    },
+    {
+      "epoch": 4.751838946960898,
+      "eval_trivia_pairs_loss": 0.5725733041763306,
+      "eval_trivia_pairs_runtime": 10.2494,
+      "eval_trivia_pairs_samples_per_second": 15.611,
+      "eval_trivia_pairs_steps_per_second": 0.488,
+      "step": 24548
+    },
+    {
+      "epoch": 4.751838946960898,
+      "eval_quora_pairs_loss": 0.17618750035762787,
+      "eval_quora_pairs_runtime": 4.0433,
+      "eval_quora_pairs_samples_per_second": 166.942,
+      "eval_quora_pairs_steps_per_second": 5.441,
+      "step": 24548
+    },
+    {
+      "epoch": 4.751838946960898,
+      "eval_gooaq_pairs_loss": 0.35624396800994873,
+      "eval_gooaq_pairs_runtime": 2.3577,
+      "eval_gooaq_pairs_samples_per_second": 67.864,
+      "eval_gooaq_pairs_steps_per_second": 2.121,
+      "step": 24548
+    },
+    {
+      "epoch": 4.751838946960898,
+      "eval_mrpc_pairs_loss": 0.01982388086616993,
+      "eval_mrpc_pairs_runtime": 0.3996,
+      "eval_mrpc_pairs_samples_per_second": 400.421,
+      "eval_mrpc_pairs_steps_per_second": 12.513,
+      "step": 24548
+    },
+    {
+      "epoch": 4.7560975609756095,
+      "grad_norm": 0.15101125836372375,
+      "learning_rate": 2.4505835699037006e-06,
+      "loss": 0.0839,
+      "step": 24570
+    },
+    {
+      "epoch": 4.768679829655439,
+      "grad_norm": 0.5723939538002014,
+      "learning_rate": 2.2125428014332173e-06,
+      "loss": 0.1318,
+      "step": 24635
+    },
+    {
+      "epoch": 4.781262098335269,
+      "grad_norm": 0.13490521907806396,
+      "learning_rate": 1.985740124523413e-06,
+      "loss": 0.129,
+      "step": 24700
+    },
+    {
+      "epoch": 4.793844367015099,
+      "grad_norm": 0.1680539846420288,
+      "learning_rate": 1.7703748617768645e-06,
+      "loss": 0.1326,
+      "step": 24765
+    },
+    {
+      "epoch": 4.806426635694928,
+      "grad_norm": 0.649811863899231,
+      "learning_rate": 1.5666362841724996e-06,
+      "loss": 0.103,
+      "step": 24830
+    },
+    {
+      "epoch": 4.819008904374758,
+      "grad_norm": 0.36529257893562317,
+      "learning_rate": 1.3747034447271883e-06,
+      "loss": 0.1251,
+      "step": 24895
+    },
+    {
+      "epoch": 4.831591173054588,
+      "grad_norm": 2.842799186706543,
+      "learning_rate": 1.1947450211373478e-06,
+      "loss": 0.115,
+      "step": 24960
+    },
+    {
+      "epoch": 4.8441734417344176,
+      "grad_norm": 0.2723024785518646,
+      "learning_rate": 1.0269191675386908e-06,
+      "loss": 0.1023,
+      "step": 25025
+    },
+    {
+      "epoch": 4.856755710414247,
+      "grad_norm": 1.8682302236557007,
+      "learning_rate": 8.713733755145653e-07,
+      "loss": 0.1003,
+      "step": 25090
+    },
+    {
+      "epoch": 4.869337979094077,
+      "grad_norm": 2.646700620651245,
+      "learning_rate": 7.282443444748149e-07,
+      "loss": 0.1094,
+      "step": 25155
+    },
+    {
+      "epoch": 4.881920247773906,
+      "grad_norm": 0.5898552536964417,
+      "learning_rate": 5.97657861519324e-07,
+      "loss": 0.133,
+      "step": 25220
+    },
+    {
+      "epoch": 4.894502516453736,
+      "grad_norm": 7.81072473526001,
+      "learning_rate": 4.797286908916226e-07,
+      "loss": 0.1164,
+      "step": 25285
+    },
+    {
+      "epoch": 4.9070847851335655,
+      "grad_norm": 3.0226893424987793,
+      "learning_rate": 3.7456047311982044e-07,
+      "loss": 0.1303,
+      "step": 25350
+    },
+    {
+      "epoch": 4.919667053813395,
+      "grad_norm": 0.5438377261161804,
+      "learning_rate": 2.822456339334556e-07,
+      "loss": 0.1446,
+      "step": 25415
+    },
+    {
+      "epoch": 4.932249322493225,
+      "grad_norm": 2.313892126083374,
+      "learning_rate": 2.028653030363481e-07,
+      "loss": 0.1189,
+      "step": 25480
+    },
+    {
+      "epoch": 4.944831591173054,
+      "grad_norm": 0.06484068930149078,
+      "learning_rate": 1.3648924280681707e-07,
+      "loss": 0.1272,
+      "step": 25545
+    },
+    {
+      "epoch": 4.957413859852885,
+      "grad_norm": 0.3085618019104004,
+      "learning_rate": 8.317578698789685e-08,
+      "loss": 0.1283,
+      "step": 25610
+    },
+    {
+      "epoch": 4.969996128532713,
+      "grad_norm": 0.0721147358417511,
+      "learning_rate": 4.2971789421488715e-08,
+      "loss": 0.1473,
+      "step": 25675
+    },
+    {
+      "epoch": 4.982578397212544,
+      "grad_norm": 2.4143240451812744,
+      "learning_rate": 1.5912582871465686e-08,
+      "loss": 0.0909,
+      "step": 25740
+    },
+    {
+      "epoch": 4.9951606658923735,
+      "grad_norm": 0.8749054670333862,
+      "learning_rate": 2.0219479719363333e-09,
+      "loss": 0.1042,
+      "step": 25805
+    }
+  ],
+  "logging_steps": 65,
+  "max_steps": 25830,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 5,
+  "save_steps": 2583,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 0.0,
+  "train_batch_size": 32,
+  "trial_name": null,
+  "trial_params": null
+}