Training in progress, step 2134, checkpoint

last-checkpoint/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2d13135869408c362410b6d65a9ecf14fcbec80f6a15aee0cf299eff51f090af
+oid sha256:0d1940ecafb946ec472d4f0bbf359034ac82aafe67fc708bd8bdef5a90ecb5bd
 size 735217848

last-checkpoint/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d9a0b7a39de0aa8d9888d2f2710d48cfcf1e88b35f597dd289a68fd843ed1b1b
+oid sha256:2788a180080964955e9955aa3cc88e16f1bfabe1adbc739556ede5713c083501
 size 1470521978

last-checkpoint/trainer_state.json

@@ -10,1829 +10,1829 @@
   "log_history": [
     {
       "epoch": 0.004686035613870665,
-      "grad_norm": 30.10508155822754,
+      "grad_norm": 12.236315727233887,
       "learning_rate": 9.994142455482662e-05,
-      "loss": 3.8567,
+      "loss": 1.5156,
       "step": 10
     },
     {
       "epoch": 0.00937207122774133,
-      "grad_norm": 8.754369735717773,
+      "grad_norm": 7.22237491607666,
       "learning_rate": 9.988284910965324e-05,
-      "loss": 1.8518,
+      "loss": 0.9095,
       "step": 20
     },
     {
       "epoch": 0.014058106841611996,
-      "grad_norm": 6.94720458984375,
+      "grad_norm": 6.729518890380859,
       "learning_rate": 9.982427366447986e-05,
-      "loss": 1.0621,
+      "loss": 0.8282,
       "step": 30
     },
     {
       "epoch": 0.01874414245548266,
-      "grad_norm": 7.401930809020996,
+      "grad_norm": 5.495608329772949,
       "learning_rate": 9.976569821930647e-05,
-      "loss": 0.8079,
+      "loss": 0.6827,
       "step": 40
     },
     {
       "epoch": 0.023430178069353328,
-      "grad_norm": 4.625001430511475,
+      "grad_norm": 5.464115142822266,
       "learning_rate": 9.970712277413309e-05,
-      "loss": 0.7732,
+      "loss": 0.699,
       "step": 50
     },
     {
       "epoch": 0.028116213683223992,
-      "grad_norm": 5.161655426025391,
+      "grad_norm": 6.352676868438721,
       "learning_rate": 9.964854732895971e-05,
-      "loss": 0.6538,
+      "loss": 0.6191,
       "step": 60
     },
     {
       "epoch": 0.03280224929709466,
-      "grad_norm": 4.755752086639404,
+      "grad_norm": 6.076754570007324,
       "learning_rate": 9.958997188378632e-05,
-      "loss": 0.5888,
+      "loss": 0.574,
       "step": 70
     },
     {
       "epoch": 0.03748828491096532,
-      "grad_norm": 5.150806903839111,
+      "grad_norm": 4.699342250823975,
       "learning_rate": 9.953139643861293e-05,
-      "loss": 0.6152,
+      "loss": 0.6077,
       "step": 80
     },
     {
       "epoch": 0.04217432052483599,
-      "grad_norm": 4.075490951538086,
+      "grad_norm": 4.326422214508057,
       "learning_rate": 9.947282099343956e-05,
-      "loss": 0.5545,
+      "loss": 0.5481,
       "step": 90
     },
     {
       "epoch": 0.046860356138706656,
-      "grad_norm": 4.118444442749023,
+      "grad_norm": 4.6638078689575195,
       "learning_rate": 9.941424554826618e-05,
-      "loss": 0.6139,
+      "loss": 0.6155,
       "step": 100
     },
     {
       "epoch": 0.046860356138706656,
-      "eval_loss": 0.08172167092561722,
-      "eval_pearson_cosine": 0.7774916057538377,
-      "eval_pearson_dot": 0.7673975678531928,
-      "eval_pearson_euclidean": 0.7739600680365157,
-      "eval_pearson_manhattan": 0.7732699847297653,
-      "eval_runtime": 4.6785,
-      "eval_samples_per_second": 320.614,
-      "eval_spearman_cosine": 0.7788712392049503,
-      "eval_spearman_dot": 0.7662758145209008,
-      "eval_spearman_euclidean": 0.779851450790825,
-      "eval_spearman_manhattan": 0.7790465578192005,
-      "eval_steps_per_second": 20.092,
+      "eval_loss": 0.08715534955263138,
+      "eval_pearson_cosine": 0.7868586661747727,
+      "eval_pearson_dot": 0.7575333730744234,
+      "eval_pearson_euclidean": 0.7895270519850293,
+      "eval_pearson_manhattan": 0.7883817170623928,
+      "eval_runtime": 4.8232,
+      "eval_samples_per_second": 310.995,
+      "eval_spearman_cosine": 0.7900534456476194,
+      "eval_spearman_dot": 0.7544776204268622,
+      "eval_spearman_euclidean": 0.7914128821733017,
+      "eval_spearman_manhattan": 0.7903422305815181,
+      "eval_steps_per_second": 19.489,
       "step": 100
     },
     {
       "epoch": 0.05154639175257732,
-      "grad_norm": 4.516867160797119,
+      "grad_norm": 3.8049023151397705,
       "learning_rate": 9.935567010309279e-05,
-      "loss": 0.5911,
+      "loss": 0.5402,
       "step": 110
     },
     {
       "epoch": 0.056232427366447985,
-      "grad_norm": 3.0860955715179443,
+      "grad_norm": 3.812105655670166,
       "learning_rate": 9.929709465791941e-05,
-      "loss": 0.5118,
+      "loss": 0.4821,
       "step": 120
     },
     {
       "epoch": 0.06091846298031865,
-      "grad_norm": 4.2761616706848145,
+      "grad_norm": 4.27888822555542,
       "learning_rate": 9.923851921274601e-05,
-      "loss": 0.5347,
+      "loss": 0.5086,
       "step": 130
     },
     {
       "epoch": 0.06560449859418932,
-      "grad_norm": 3.3774375915527344,
+      "grad_norm": 3.5396862030029297,
       "learning_rate": 9.917994376757263e-05,
-      "loss": 0.516,
+      "loss": 0.4709,
       "step": 140
     },
     {
       "epoch": 0.07029053420805999,
-      "grad_norm": 3.043034315109253,
+      "grad_norm": 2.9794766902923584,
       "learning_rate": 9.912136832239926e-05,
-      "loss": 0.5583,
+      "loss": 0.5376,
       "step": 150
     },
     {
       "epoch": 0.07497656982193064,
-      "grad_norm": 3.3026859760284424,
+      "grad_norm": 4.1056413650512695,
       "learning_rate": 9.906279287722588e-05,
-      "loss": 0.5691,
+      "loss": 0.5403,
       "step": 160
     },
     {
       "epoch": 0.07966260543580131,
-      "grad_norm": 3.694498062133789,
+      "grad_norm": 3.5660178661346436,
       "learning_rate": 9.90042174320525e-05,
-      "loss": 0.5126,
+      "loss": 0.5095,
       "step": 170
     },
     {
       "epoch": 0.08434864104967198,
-      "grad_norm": 2.7696762084960938,
+      "grad_norm": 3.226407527923584,
       "learning_rate": 9.89456419868791e-05,
-      "loss": 0.4945,
+      "loss": 0.4684,
       "step": 180
     },
     {
       "epoch": 0.08903467666354264,
-      "grad_norm": 2.7977497577667236,
+      "grad_norm": 3.224576950073242,
       "learning_rate": 9.888706654170572e-05,
-      "loss": 0.5133,
+      "loss": 0.5005,
       "step": 190
     },
     {
       "epoch": 0.09372071227741331,
-      "grad_norm": 4.13506555557251,
+      "grad_norm": 2.986266851425171,
       "learning_rate": 9.882849109653233e-05,
-      "loss": 0.4496,
+      "loss": 0.4415,
       "step": 200
     },
     {
       "epoch": 0.09372071227741331,
-      "eval_loss": 0.06149223446846008,
-      "eval_pearson_cosine": 0.7878806911951841,
-      "eval_pearson_dot": 0.7892236854270749,
-      "eval_pearson_euclidean": 0.7815123592368103,
-      "eval_pearson_manhattan": 0.7802018326188076,
-      "eval_runtime": 4.629,
-      "eval_samples_per_second": 324.041,
-      "eval_spearman_cosine": 0.7881602244285756,
-      "eval_spearman_dot": 0.7890157817002775,
-      "eval_spearman_euclidean": 0.7853020907543472,
-      "eval_spearman_manhattan": 0.7841560791576906,
-      "eval_steps_per_second": 20.307,
+      "eval_loss": 0.05990051105618477,
+      "eval_pearson_cosine": 0.8037077163940438,
+      "eval_pearson_dot": 0.7711149700928424,
+      "eval_pearson_euclidean": 0.8065472278987844,
+      "eval_pearson_manhattan": 0.8064149560284655,
+      "eval_runtime": 4.6772,
+      "eval_samples_per_second": 320.707,
+      "eval_spearman_cosine": 0.8062807649835092,
+      "eval_spearman_dot": 0.7710757773956901,
+      "eval_spearman_euclidean": 0.8127702338196209,
+      "eval_spearman_manhattan": 0.8128386190479225,
+      "eval_steps_per_second": 20.098,
       "step": 200
     },
     {
       "epoch": 0.09840674789128398,
-      "grad_norm": 3.0135467052459717,
+      "grad_norm": 3.6775708198547363,
       "learning_rate": 9.876991565135896e-05,
-      "loss": 0.5474,
+      "loss": 0.5227,
       "step": 210
     },
     {
       "epoch": 0.10309278350515463,
-      "grad_norm": 2.703244924545288,
+      "grad_norm": 2.6270956993103027,
       "learning_rate": 9.871134020618558e-05,
-      "loss": 0.4499,
+      "loss": 0.4477,
       "step": 220
     },
     {
       "epoch": 0.1077788191190253,
-      "grad_norm": 3.320786714553833,
+      "grad_norm": 3.687191963195801,
       "learning_rate": 9.865276476101218e-05,
-      "loss": 0.5042,
+      "loss": 0.4674,
       "step": 230
     },
     {
       "epoch": 0.11246485473289597,
-      "grad_norm": 3.642885684967041,
+      "grad_norm": 2.5306522846221924,
       "learning_rate": 9.85941893158388e-05,
-      "loss": 0.4496,
+      "loss": 0.4539,
       "step": 240
     },
     {
       "epoch": 0.11715089034676664,
-      "grad_norm": 2.59848952293396,
+      "grad_norm": 2.426499366760254,
       "learning_rate": 9.853561387066542e-05,
-      "loss": 0.467,
+      "loss": 0.4485,
       "step": 250
     },
     {
       "epoch": 0.1218369259606373,
-      "grad_norm": 2.8087403774261475,
+      "grad_norm": 2.7819628715515137,
       "learning_rate": 9.847703842549204e-05,
-      "loss": 0.4818,
+      "loss": 0.4653,
       "step": 260
     },
     {
       "epoch": 0.12652296157450796,
-      "grad_norm": 3.655815362930298,
+      "grad_norm": 4.526126384735107,
       "learning_rate": 9.841846298031867e-05,
-      "loss": 0.4195,
+      "loss": 0.3939,
       "step": 270
     },
     {
       "epoch": 0.13120899718837864,
-      "grad_norm": 2.4474446773529053,
+      "grad_norm": 2.6432578563690186,
       "learning_rate": 9.835988753514527e-05,
-      "loss": 0.4295,
+      "loss": 0.4241,
       "step": 280
     },
     {
       "epoch": 0.1358950328022493,
-      "grad_norm": 4.437081336975098,
+      "grad_norm": 3.9447426795959473,
       "learning_rate": 9.830131208997189e-05,
-      "loss": 0.4783,
+      "loss": 0.459,
       "step": 290
     },
     {
       "epoch": 0.14058106841611998,
-      "grad_norm": 2.853497266769409,
+      "grad_norm": 3.0471863746643066,
       "learning_rate": 9.82427366447985e-05,
-      "loss": 0.445,
+      "loss": 0.4497,
       "step": 300
     },
     {
       "epoch": 0.14058106841611998,
-      "eval_loss": 0.06087055429816246,
-      "eval_pearson_cosine": 0.8085847462809284,
-      "eval_pearson_dot": 0.8099837052074292,
-      "eval_pearson_euclidean": 0.797235616826157,
-      "eval_pearson_manhattan": 0.796084644837677,
-      "eval_runtime": 4.8442,
-      "eval_samples_per_second": 309.65,
-      "eval_spearman_cosine": 0.8087098487869666,
-      "eval_spearman_dot": 0.8088350013000529,
-      "eval_spearman_euclidean": 0.8038077082663364,
-      "eval_spearman_manhattan": 0.8027157437095755,
-      "eval_steps_per_second": 19.405,
+      "eval_loss": 0.06257154792547226,
+      "eval_pearson_cosine": 0.8157783815016391,
+      "eval_pearson_dot": 0.7943799683165018,
+      "eval_pearson_euclidean": 0.8109480529101205,
+      "eval_pearson_manhattan": 0.8107189557075429,
+      "eval_runtime": 4.6882,
+      "eval_samples_per_second": 319.95,
+      "eval_spearman_cosine": 0.8164514602392956,
+      "eval_spearman_dot": 0.793312162219592,
+      "eval_spearman_euclidean": 0.818050705442233,
+      "eval_spearman_manhattan": 0.8178521574840566,
+      "eval_steps_per_second": 20.05,
       "step": 300
     },
     {
       "epoch": 0.14526710402999063,
-      "grad_norm": 3.850320339202881,
+      "grad_norm": 3.6450355052948,
       "learning_rate": 9.818416119962512e-05,
-      "loss": 0.4192,
+      "loss": 0.4274,
       "step": 310
     },
     {
       "epoch": 0.14995313964386128,
-      "grad_norm": 3.8749234676361084,
+      "grad_norm": 2.9165449142456055,
       "learning_rate": 9.812558575445174e-05,
-      "loss": 0.4282,
+      "loss": 0.4263,
       "step": 320
     },
     {
       "epoch": 0.15463917525773196,
-      "grad_norm": 2.7373576164245605,
+      "grad_norm": 2.6363539695739746,
       "learning_rate": 9.806701030927836e-05,
-      "loss": 0.5132,
+      "loss": 0.5083,
       "step": 330
     },
     {
       "epoch": 0.15932521087160262,
-      "grad_norm": 2.587904691696167,
+      "grad_norm": 2.2315175533294678,
       "learning_rate": 9.800843486410497e-05,
-      "loss": 0.4452,
+      "loss": 0.4517,
       "step": 340
     },
     {
       "epoch": 0.1640112464854733,
-      "grad_norm": 2.770646333694458,
+      "grad_norm": 2.2491111755371094,
       "learning_rate": 9.794985941893159e-05,
-      "loss": 0.5072,
+      "loss": 0.4797,
       "step": 350
     },
     {
       "epoch": 0.16869728209934395,
-      "grad_norm": 3.2103078365325928,
+      "grad_norm": 3.327615976333618,
       "learning_rate": 9.78912839737582e-05,
-      "loss": 0.4488,
+      "loss": 0.4437,
       "step": 360
     },
     {
       "epoch": 0.1733833177132146,
-      "grad_norm": 2.8420042991638184,
+      "grad_norm": 2.88334321975708,
       "learning_rate": 9.783270852858482e-05,
-      "loss": 0.4132,
+      "loss": 0.4006,
       "step": 370
     },
     {
       "epoch": 0.1780693533270853,
-      "grad_norm": 2.8594541549682617,
+      "grad_norm": 2.7329928874969482,
       "learning_rate": 9.777413308341144e-05,
-      "loss": 0.4217,
+      "loss": 0.4184,
       "step": 380
     },
     {
       "epoch": 0.18275538894095594,
-      "grad_norm": 2.6843366622924805,
+      "grad_norm": 3.280235767364502,
       "learning_rate": 9.771555763823806e-05,
-      "loss": 0.423,
+      "loss": 0.4022,
       "step": 390
     },
     {
       "epoch": 0.18744142455482662,
-      "grad_norm": 2.808396339416504,
+      "grad_norm": 2.5065557956695557,
       "learning_rate": 9.765698219306467e-05,
-      "loss": 0.4398,
+      "loss": 0.4479,
       "step": 400
     },
     {
       "epoch": 0.18744142455482662,
-      "eval_loss": 0.06297692656517029,
-      "eval_pearson_cosine": 0.7991067883059557,
-      "eval_pearson_dot": 0.8028010593184547,
-      "eval_pearson_euclidean": 0.7880801281434806,
-      "eval_pearson_manhattan": 0.7862328250693538,
-      "eval_runtime": 5.1278,
-      "eval_samples_per_second": 292.522,
-      "eval_spearman_cosine": 0.8008051119507676,
-      "eval_spearman_dot": 0.8028962103300248,
-      "eval_spearman_euclidean": 0.7941679469832046,
-      "eval_spearman_manhattan": 0.7922002625333209,
-      "eval_steps_per_second": 18.331,
+      "eval_loss": 0.0645737424492836,
+      "eval_pearson_cosine": 0.8129588336284996,
+      "eval_pearson_dot": 0.7848457662394352,
+      "eval_pearson_euclidean": 0.813348529089879,
+      "eval_pearson_manhattan": 0.8134040940780514,
+      "eval_runtime": 4.6989,
+      "eval_samples_per_second": 319.226,
+      "eval_spearman_cosine": 0.8152062220540486,
+      "eval_spearman_dot": 0.7852216128574532,
+      "eval_spearman_euclidean": 0.8209486088575746,
+      "eval_spearman_manhattan": 0.8208551347463181,
+      "eval_steps_per_second": 20.005,
       "step": 400
     },
     {
       "epoch": 0.19212746016869728,
-      "grad_norm": 3.4171650409698486,
+      "grad_norm": 3.4038398265838623,
       "learning_rate": 9.759840674789129e-05,
-      "loss": 0.472,
+      "loss": 0.4605,
       "step": 410
     },
     {
       "epoch": 0.19681349578256796,
-      "grad_norm": 2.304884433746338,
+      "grad_norm": 2.5624849796295166,
       "learning_rate": 9.753983130271791e-05,
-      "loss": 0.4269,
+      "loss": 0.4283,
       "step": 420
     },
     {
       "epoch": 0.2014995313964386,
-      "grad_norm": 3.582045793533325,
+      "grad_norm": 4.289618015289307,
       "learning_rate": 9.748125585754451e-05,
-      "loss": 0.4709,
+      "loss": 0.451,
       "step": 430
     },
     {
       "epoch": 0.20618556701030927,
-      "grad_norm": 2.555410623550415,
+      "grad_norm": 2.115318536758423,
       "learning_rate": 9.742268041237114e-05,
-      "loss": 0.428,
+      "loss": 0.4354,
       "step": 440
     },
     {
       "epoch": 0.21087160262417995,
-      "grad_norm": 3.291259288787842,
+      "grad_norm": 3.031923770904541,
       "learning_rate": 9.736410496719776e-05,
-      "loss": 0.3884,
+      "loss": 0.3964,
       "step": 450
     },
     {
       "epoch": 0.2155576382380506,
-      "grad_norm": 2.4803812503814697,
+      "grad_norm": 3.1766481399536133,
       "learning_rate": 9.730552952202438e-05,
-      "loss": 0.4434,
+      "loss": 0.4381,
       "step": 460
     },
     {
       "epoch": 0.22024367385192128,
-      "grad_norm": 2.9194233417510986,
+      "grad_norm": 2.5575857162475586,
       "learning_rate": 9.7246954076851e-05,
-      "loss": 0.3969,
+      "loss": 0.3923,
       "step": 470
     },
     {
       "epoch": 0.22492970946579194,
-      "grad_norm": 2.5176608562469482,
+      "grad_norm": 2.7524373531341553,
       "learning_rate": 9.71883786316776e-05,
-      "loss": 0.3668,
+      "loss": 0.3636,
       "step": 480
     },
     {
       "epoch": 0.2296157450796626,
-      "grad_norm": 3.111013412475586,
+      "grad_norm": 2.397308111190796,
       "learning_rate": 9.712980318650421e-05,
-      "loss": 0.4047,
+      "loss": 0.4024,
       "step": 490
     },
     {
       "epoch": 0.23430178069353327,
-      "grad_norm": 2.3099541664123535,
+      "grad_norm": 2.7153968811035156,
       "learning_rate": 9.707122774133085e-05,
-      "loss": 0.4509,
+      "loss": 0.4284,
       "step": 500
     },
     {
       "epoch": 0.23430178069353327,
-      "eval_loss": 0.06479576230049133,
-      "eval_pearson_cosine": 0.7937691374860236,
-      "eval_pearson_dot": 0.8037986869849263,
-      "eval_pearson_euclidean": 0.7796252574393066,
-      "eval_pearson_manhattan": 0.7788957424132832,
-      "eval_runtime": 5.1806,
-      "eval_samples_per_second": 289.541,
-      "eval_spearman_cosine": 0.7968672368227576,
-      "eval_spearman_dot": 0.8062882580483537,
-      "eval_spearman_euclidean": 0.785478492768414,
-      "eval_spearman_manhattan": 0.7846800638460941,
-      "eval_steps_per_second": 18.145,
+      "eval_loss": 0.06505414098501205,
+      "eval_pearson_cosine": 0.8050664803014342,
+      "eval_pearson_dot": 0.7819900690459463,
+      "eval_pearson_euclidean": 0.8067253667157352,
+      "eval_pearson_manhattan": 0.8066066749239482,
+      "eval_runtime": 4.7695,
+      "eval_samples_per_second": 314.5,
+      "eval_spearman_cosine": 0.8089450551178121,
+      "eval_spearman_dot": 0.7819509905282449,
+      "eval_spearman_euclidean": 0.813757984849266,
+      "eval_spearman_manhattan": 0.813378975698625,
+      "eval_steps_per_second": 19.709,
       "step": 500
     },
     {
       "epoch": 0.23898781630740393,
-      "grad_norm": 2.587941884994507,
+      "grad_norm": 2.5172924995422363,
       "learning_rate": 9.701265229615746e-05,
-      "loss": 0.4402,
+      "loss": 0.4338,
       "step": 510
     },
     {
       "epoch": 0.2436738519212746,
-      "grad_norm": 2.8946237564086914,
+      "grad_norm": 3.0745673179626465,
       "learning_rate": 9.695407685098408e-05,
-      "loss": 0.4303,
+      "loss": 0.4411,
       "step": 520
     },
     {
       "epoch": 0.24835988753514526,
-      "grad_norm": 2.144695997238159,
+      "grad_norm": 2.51690411567688,
       "learning_rate": 9.689550140581068e-05,
-      "loss": 0.3955,
+      "loss": 0.3888,
       "step": 530
     },
     {
       "epoch": 0.2530459231490159,
-      "grad_norm": 2.310426712036133,
+      "grad_norm": 2.5043115615844727,
       "learning_rate": 9.68369259606373e-05,
-      "loss": 0.4027,
+      "loss": 0.3946,
       "step": 540
     },
     {
       "epoch": 0.25773195876288657,
-      "grad_norm": 1.9948300123214722,
+      "grad_norm": 1.931721806526184,
       "learning_rate": 9.677835051546392e-05,
-      "loss": 0.3806,
+      "loss": 0.3734,
       "step": 550
     },
     {
       "epoch": 0.2624179943767573,
-      "grad_norm": 2.4448132514953613,
+      "grad_norm": 2.396235704421997,
       "learning_rate": 9.671977507029055e-05,
-      "loss": 0.3917,
+      "loss": 0.3781,
       "step": 560
     },
     {
       "epoch": 0.26710402999062793,
-      "grad_norm": 2.1171720027923584,
+      "grad_norm": 1.8347618579864502,
       "learning_rate": 9.666119962511717e-05,
-      "loss": 0.4086,
+      "loss": 0.4067,
       "step": 570
     },
     {
       "epoch": 0.2717900656044986,
-      "grad_norm": 3.1124324798583984,
+      "grad_norm": 2.65472412109375,
       "learning_rate": 9.660262417994377e-05,
-      "loss": 0.3754,
+      "loss": 0.3646,
       "step": 580
     },
     {
       "epoch": 0.27647610121836924,
-      "grad_norm": 2.3518879413604736,
+      "grad_norm": 2.131566047668457,
       "learning_rate": 9.654404873477039e-05,
-      "loss": 0.4155,
+      "loss": 0.4212,
       "step": 590
     },
     {
       "epoch": 0.28116213683223995,
-      "grad_norm": 2.3229143619537354,
+      "grad_norm": 2.4860639572143555,
       "learning_rate": 9.6485473289597e-05,
-      "loss": 0.3899,
+      "loss": 0.3853,
       "step": 600
     },
     {
       "epoch": 0.28116213683223995,
-      "eval_loss": 0.05898291990160942,
-      "eval_pearson_cosine": 0.8002203582671399,
-      "eval_pearson_dot": 0.8059872147596572,
-      "eval_pearson_euclidean": 0.7924462958823995,
-      "eval_pearson_manhattan": 0.7914827034937701,
-      "eval_runtime": 4.8482,
-      "eval_samples_per_second": 309.39,
-      "eval_spearman_cosine": 0.8056451862347204,
-      "eval_spearman_dot": 0.8090541661182976,
-      "eval_spearman_euclidean": 0.797456379085547,
-      "eval_spearman_manhattan": 0.796591059583954,
-      "eval_steps_per_second": 19.388,
+      "eval_loss": 0.05714314430952072,
+      "eval_pearson_cosine": 0.8037633908878945,
+      "eval_pearson_dot": 0.7922442881383063,
+      "eval_pearson_euclidean": 0.8033561562960063,
+      "eval_pearson_manhattan": 0.8035567720397765,
+      "eval_runtime": 4.8432,
+      "eval_samples_per_second": 309.711,
+      "eval_spearman_cosine": 0.8084109915495988,
+      "eval_spearman_dot": 0.7948649982391062,
+      "eval_spearman_euclidean": 0.8103898218799952,
+      "eval_spearman_manhattan": 0.8106767456523252,
+      "eval_steps_per_second": 19.409,
       "step": 600
     },
     {
       "epoch": 0.2858481724461106,
-      "grad_norm": 2.2351458072662354,
+      "grad_norm": 2.178133249282837,
       "learning_rate": 9.642689784442362e-05,
-      "loss": 0.4096,
+      "loss": 0.4,
       "step": 610
     },
     {
       "epoch": 0.29053420805998126,
-      "grad_norm": 2.529383420944214,
+      "grad_norm": 2.625169038772583,
       "learning_rate": 9.636832239925024e-05,
-      "loss": 0.4094,
+      "loss": 0.3901,
       "step": 620
     },
     {
       "epoch": 0.2952202436738519,
-      "grad_norm": 2.727684736251831,
+      "grad_norm": 2.9042716026306152,
       "learning_rate": 9.630974695407685e-05,
-      "loss": 0.4017,
+      "loss": 0.3999,
       "step": 630
     },
     {
       "epoch": 0.29990627928772257,
-      "grad_norm": 2.1631858348846436,
+      "grad_norm": 1.7897489070892334,
       "learning_rate": 9.625117150890347e-05,
-      "loss": 0.3829,
+      "loss": 0.3519,
       "step": 640
     },
     {
       "epoch": 0.3045923149015933,
-      "grad_norm": 2.08465838432312,
+      "grad_norm": 2.162492513656616,
       "learning_rate": 9.619259606373009e-05,
-      "loss": 0.3593,
+      "loss": 0.3609,
       "step": 650
     },
     {
       "epoch": 0.30927835051546393,
-      "grad_norm": 2.7151734828948975,
+      "grad_norm": 2.502537727355957,
       "learning_rate": 9.61340206185567e-05,
-      "loss": 0.3575,
+      "loss": 0.3664,
       "step": 660
     },
     {
       "epoch": 0.3139643861293346,
-      "grad_norm": 2.9916470050811768,
+      "grad_norm": 2.679335594177246,
       "learning_rate": 9.607544517338332e-05,
-      "loss": 0.4043,
+      "loss": 0.4041,
       "step": 670
     },
     {
       "epoch": 0.31865042174320524,
-      "grad_norm": 2.2090351581573486,
+      "grad_norm": 2.2340409755706787,
       "learning_rate": 9.601686972820994e-05,
-      "loss": 0.3969,
+      "loss": 0.3847,
       "step": 680
     },
     {
       "epoch": 0.3233364573570759,
-      "grad_norm": 2.5505614280700684,
+      "grad_norm": 3.3209805488586426,
       "learning_rate": 9.595829428303656e-05,
-      "loss": 0.4162,
+      "loss": 0.4274,
       "step": 690
     },
     {
       "epoch": 0.3280224929709466,
-      "grad_norm": 1.9354357719421387,
+      "grad_norm": 1.8717986345291138,
       "learning_rate": 9.589971883786317e-05,
-      "loss": 0.3944,
+      "loss": 0.3798,
       "step": 700
     },
     {
       "epoch": 0.3280224929709466,
-      "eval_loss": 0.06544993817806244,
-      "eval_pearson_cosine": 0.7849690888771903,
-      "eval_pearson_dot": 0.7906563077972777,
-      "eval_pearson_euclidean": 0.7744169020961635,
-      "eval_pearson_manhattan": 0.773555246133725,
-      "eval_runtime": 4.7217,
-      "eval_samples_per_second": 317.682,
-      "eval_spearman_cosine": 0.7912145091954246,
-      "eval_spearman_dot": 0.795880260533099,
-      "eval_spearman_euclidean": 0.7788547557160983,
-      "eval_spearman_manhattan": 0.7778138343707646,
-      "eval_steps_per_second": 19.908,
+      "eval_loss": 0.06265808641910553,
+      "eval_pearson_cosine": 0.7997962383626067,
+      "eval_pearson_dot": 0.782403525826652,
+      "eval_pearson_euclidean": 0.7963005835260759,
+      "eval_pearson_manhattan": 0.7967635425851256,
+      "eval_runtime": 4.6788,
+      "eval_samples_per_second": 320.593,
+      "eval_spearman_cosine": 0.8040677308135645,
+      "eval_spearman_dot": 0.7846759313363996,
+      "eval_spearman_euclidean": 0.8039856887485184,
+      "eval_spearman_manhattan": 0.8042998695210408,
+      "eval_steps_per_second": 20.09,
       "step": 700
     },
     {
       "epoch": 0.33270852858481725,
-      "grad_norm": 2.754939317703247,
+      "grad_norm": 2.667065382003784,
       "learning_rate": 9.584114339268979e-05,
-      "loss": 0.3421,
+      "loss": 0.3517,
       "step": 710
     },
     {
       "epoch": 0.3373945641986879,
-      "grad_norm": 2.395989418029785,
+      "grad_norm": 2.4856221675872803,
       "learning_rate": 9.578256794751641e-05,
-      "loss": 0.4181,
+      "loss": 0.4113,
       "step": 720
     },
     {
       "epoch": 0.34208059981255856,
-      "grad_norm": 1.9646469354629517,
+      "grad_norm": 2.3127479553222656,
       "learning_rate": 9.572399250234303e-05,
-      "loss": 0.3604,
+      "loss": 0.3333,
       "step": 730
     },
     {
       "epoch": 0.3467666354264292,
-      "grad_norm": 1.77524995803833,
+      "grad_norm": 1.7665061950683594,
       "learning_rate": 9.566541705716964e-05,
-      "loss": 0.4152,
+      "loss": 0.4113,
       "step": 740
     },
     {
       "epoch": 0.3514526710402999,
-      "grad_norm": 2.225848913192749,
+      "grad_norm": 1.8380166292190552,
       "learning_rate": 9.560684161199626e-05,
-      "loss": 0.3582,
+      "loss": 0.3634,
       "step": 750
     },
     {
       "epoch": 0.3561387066541706,
-      "grad_norm": 2.9699714183807373,
+      "grad_norm": 2.980945348739624,
       "learning_rate": 9.554826616682288e-05,
-      "loss": 0.3498,
+      "loss": 0.3486,
       "step": 760
     },
     {
       "epoch": 0.36082474226804123,
-      "grad_norm": 2.2419025897979736,
+      "grad_norm": 2.6140224933624268,
       "learning_rate": 9.54896907216495e-05,
-      "loss": 0.4014,
+      "loss": 0.389,
       "step": 770
     },
     {
       "epoch": 0.3655107778819119,
-      "grad_norm": 2.090801954269409,
+      "grad_norm": 2.7822117805480957,
       "learning_rate": 9.54311152764761e-05,
-      "loss": 0.3788,
+      "loss": 0.3747,
       "step": 780
     },
     {
       "epoch": 0.3701968134957826,
-      "grad_norm": 2.0521080493927,
+      "grad_norm": 1.884682059288025,
       "learning_rate": 9.537253983130271e-05,
-      "loss": 0.3381,
+      "loss": 0.3311,
       "step": 790
     },
     {
       "epoch": 0.37488284910965325,
-      "grad_norm": 2.189819574356079,
+      "grad_norm": 1.9988900423049927,
       "learning_rate": 9.531396438612934e-05,
-      "loss": 0.3866,
+      "loss": 0.3615,
       "step": 800
     },
     {
       "epoch": 0.37488284910965325,
-      "eval_loss": 0.057871297001838684,
-      "eval_pearson_cosine": 0.8000694370275774,
-      "eval_pearson_dot": 0.808069225934787,
-      "eval_pearson_euclidean": 0.7859619382036129,
-      "eval_pearson_manhattan": 0.7850475265092101,
-      "eval_runtime": 4.9109,
-      "eval_samples_per_second": 305.445,
-      "eval_spearman_cosine": 0.802358950496682,
-      "eval_spearman_dot": 0.8096070418813107,
-      "eval_spearman_euclidean": 0.7916939807330186,
-      "eval_spearman_manhattan": 0.7908012039061529,
-      "eval_steps_per_second": 19.141,
+      "eval_loss": 0.05634882301092148,
+      "eval_pearson_cosine": 0.8038418277696451,
+      "eval_pearson_dot": 0.7799485472675656,
+      "eval_pearson_euclidean": 0.8069547387629541,
+      "eval_pearson_manhattan": 0.8073835773742246,
+      "eval_runtime": 4.7355,
+      "eval_samples_per_second": 316.756,
+      "eval_spearman_cosine": 0.8046853584108842,
+      "eval_spearman_dot": 0.7772894170177264,
+      "eval_spearman_euclidean": 0.8106627528697075,
+      "eval_spearman_manhattan": 0.8113494849262057,
+      "eval_steps_per_second": 19.85,
       "step": 800
     },
     {
       "epoch": 0.3795688847235239,
-      "grad_norm": 2.161396026611328,
+      "grad_norm": 2.149801015853882,
       "learning_rate": 9.525538894095596e-05,
-      "loss": 0.411,
+      "loss": 0.4051,
       "step": 810
     },
     {
       "epoch": 0.38425492033739456,
-      "grad_norm": 2.709839344024658,
+      "grad_norm": 2.5438592433929443,
       "learning_rate": 9.519681349578258e-05,
-      "loss": 0.379,
+      "loss": 0.3536,
       "step": 820
     },
     {
       "epoch": 0.3889409559512652,
-      "grad_norm": 2.1237142086029053,
+      "grad_norm": 2.192567825317383,
       "learning_rate": 9.513823805060918e-05,
-      "loss": 0.3333,
+      "loss": 0.3437,
       "step": 830
     },
     {
       "epoch": 0.3936269915651359,
-      "grad_norm": 2.088275194168091,
+      "grad_norm": 2.1283817291259766,
       "learning_rate": 9.50796626054358e-05,
-      "loss": 0.4006,
+      "loss": 0.406,
       "step": 840
     },
     {
       "epoch": 0.3983130271790066,
-      "grad_norm": 2.1435964107513428,
+      "grad_norm": 2.416581630706787,
       "learning_rate": 9.502108716026242e-05,
-      "loss": 0.3328,
+      "loss": 0.3383,
       "step": 850
     },
     {
       "epoch": 0.4029990627928772,
-      "grad_norm": 2.671003580093384,
+      "grad_norm": 2.987426996231079,
       "learning_rate": 9.496251171508905e-05,
-      "loss": 0.3751,
+      "loss": 0.4024,
       "step": 860
     },
     {
       "epoch": 0.4076850984067479,
-      "grad_norm": 2.3052737712860107,
+      "grad_norm": 3.472496747970581,
       "learning_rate": 9.490393626991566e-05,
-      "loss": 0.3988,
+      "loss": 0.4088,
       "step": 870
     },
     {
       "epoch": 0.41237113402061853,
-      "grad_norm": 1.9277461767196655,
+      "grad_norm": 1.7949929237365723,
       "learning_rate": 9.484536082474227e-05,
-      "loss": 0.3717,
+      "loss": 0.3556,
       "step": 880
     },
     {
       "epoch": 0.41705716963448924,
-      "grad_norm": 2.213792324066162,
+      "grad_norm": 2.5681650638580322,
       "learning_rate": 9.478678537956888e-05,
-      "loss": 0.3521,
+      "loss": 0.3488,
       "step": 890
     },
     {
       "epoch": 0.4217432052483599,
-      "grad_norm": 2.0909924507141113,
+      "grad_norm": 1.9793102741241455,
       "learning_rate": 9.47282099343955e-05,
-      "loss": 0.3292,
+      "loss": 0.3333,
       "step": 900
     },
     {
       "epoch": 0.4217432052483599,
-      "eval_loss": 0.05752936750650406,
-      "eval_pearson_cosine": 0.7969533250152949,
-      "eval_pearson_dot": 0.8037643743757386,
-      "eval_pearson_euclidean": 0.7823638645148634,
-      "eval_pearson_manhattan": 0.7813551521507094,
-      "eval_runtime": 4.8299,
-      "eval_samples_per_second": 310.563,
-      "eval_spearman_cosine": 0.7992619351747234,
-      "eval_spearman_dot": 0.8055866654381859,
-      "eval_spearman_euclidean": 0.7873785407587978,
-      "eval_spearman_manhattan": 0.7866152504590422,
-      "eval_steps_per_second": 19.462,
+      "eval_loss": 0.0584479495882988,
+      "eval_pearson_cosine": 0.8053522096959256,
+      "eval_pearson_dot": 0.7824338806172406,
+      "eval_pearson_euclidean": 0.8050956474463948,
+      "eval_pearson_manhattan": 0.8050994959105964,
+      "eval_runtime": 4.7733,
+      "eval_samples_per_second": 314.248,
+      "eval_spearman_cosine": 0.8066755205383609,
+      "eval_spearman_dot": 0.7821157253633244,
+      "eval_spearman_euclidean": 0.8097694547473674,
+      "eval_spearman_manhattan": 0.8100692397448628,
+      "eval_steps_per_second": 19.693,
       "step": 900
     },
     {
       "epoch": 0.42642924086223055,
-      "grad_norm": 2.122093915939331,
+      "grad_norm": 1.8367053270339966,
       "learning_rate": 9.466963448922212e-05,
-      "loss": 0.3008,
+      "loss": 0.2852,
       "step": 910
     },
     {
       "epoch": 0.4311152764761012,
-      "grad_norm": 2.1320900917053223,
+      "grad_norm": 2.7685294151306152,
       "learning_rate": 9.461105904404875e-05,
-      "loss": 0.3691,
+      "loss": 0.3903,
       "step": 920
     },
     {
       "epoch": 0.43580131208997186,
-      "grad_norm": 2.6174728870391846,
+      "grad_norm": 2.694925546646118,
       "learning_rate": 9.455248359887535e-05,
-      "loss": 0.3433,
+      "loss": 0.3556,
       "step": 930
     },
     {
       "epoch": 0.44048734770384257,
-      "grad_norm": 1.940846562385559,
+      "grad_norm": 2.1870169639587402,
       "learning_rate": 9.449390815370197e-05,
-      "loss": 0.3517,
+      "loss": 0.3371,
       "step": 940
     },
     {
       "epoch": 0.4451733833177132,
-      "grad_norm": 2.081587791442871,
+      "grad_norm": 2.213475465774536,
       "learning_rate": 9.443533270852859e-05,
-      "loss": 0.3455,
+      "loss": 0.3506,
       "step": 950
     },
     {
       "epoch": 0.4498594189315839,
-      "grad_norm": 2.1150238513946533,
+      "grad_norm": 2.083057165145874,
       "learning_rate": 9.43767572633552e-05,
-      "loss": 0.3791,
+      "loss": 0.3821,
       "step": 960
     },
     {
       "epoch": 0.45454545454545453,
-      "grad_norm": 2.286648750305176,
+      "grad_norm": 2.3329579830169678,
       "learning_rate": 9.431818181818182e-05,
-      "loss": 0.363,
+      "loss": 0.3715,
       "step": 970
     },
     {
       "epoch": 0.4592314901593252,
-      "grad_norm": 2.0841028690338135,
+      "grad_norm": 2.198944330215454,
       "learning_rate": 9.425960637300844e-05,
-      "loss": 0.3541,
+      "loss": 0.3549,
       "step": 980
     },
     {
       "epoch": 0.4639175257731959,
-      "grad_norm": 2.329029083251953,
+      "grad_norm": 11.298641204833984,
       "learning_rate": 9.420103092783506e-05,
-      "loss": 0.3635,
+      "loss": 0.4068,
       "step": 990
     },
     {
       "epoch": 0.46860356138706655,
-      "grad_norm": 1.905613660812378,
+      "grad_norm": 2.49027681350708,
       "learning_rate": 9.414245548266167e-05,
-      "loss": 0.3714,
+      "loss": 0.3693,
       "step": 1000
     },
     {
       "epoch": 0.46860356138706655,
-      "eval_loss": 0.05644960328936577,
-      "eval_pearson_cosine": 0.803875799789175,
-      "eval_pearson_dot": 0.8110916308252172,
-      "eval_pearson_euclidean": 0.7894191369326222,
-      "eval_pearson_manhattan": 0.7890558407124857,
-      "eval_runtime": 4.8648,
-      "eval_samples_per_second": 308.335,
-      "eval_spearman_cosine": 0.8060476875796366,
-      "eval_spearman_dot": 0.8116367057743546,
-      "eval_spearman_euclidean": 0.7952310888165739,
-      "eval_spearman_manhattan": 0.7948288092961023,
-      "eval_steps_per_second": 19.322,
+      "eval_loss": 0.05829383432865143,
+      "eval_pearson_cosine": 0.7877790961258597,
+      "eval_pearson_dot": 0.7795745583242422,
+      "eval_pearson_euclidean": 0.7679074299205411,
+      "eval_pearson_manhattan": 0.7685797828363299,
+      "eval_runtime": 4.7202,
+      "eval_samples_per_second": 317.781,
+      "eval_spearman_cosine": 0.7864141944499116,
+      "eval_spearman_dot": 0.7772556312389166,
+      "eval_spearman_euclidean": 0.7804453284579596,
+      "eval_spearman_manhattan": 0.7808739773322116,
+      "eval_steps_per_second": 19.914,
       "step": 1000
     },
     {
       "epoch": 0.4732895970009372,
-      "grad_norm": 2.0706255435943604,
+      "grad_norm": 2.1024911403656006,
       "learning_rate": 9.408388003748829e-05,
-      "loss": 0.3365,
+      "loss": 0.3561,
       "step": 1010
     },
     {
       "epoch": 0.47797563261480785,
-      "grad_norm": 1.9742926359176636,
+      "grad_norm": 1.8707431554794312,
       "learning_rate": 9.402530459231491e-05,
-      "loss": 0.3723,
+      "loss": 0.3774,
       "step": 1020
     },
     {
       "epoch": 0.48266166822867856,
-      "grad_norm": 2.0600244998931885,
+      "grad_norm": 1.9497177600860596,
       "learning_rate": 9.396672914714152e-05,
-      "loss": 0.3605,
+      "loss": 0.3473,
       "step": 1030
     },
     {
       "epoch": 0.4873477038425492,
-      "grad_norm": 1.6684983968734741,
+      "grad_norm": 2.7375972270965576,
       "learning_rate": 9.390815370196814e-05,
-      "loss": 0.346,
+      "loss": 0.3697,
       "step": 1040
     },
     {
       "epoch": 0.49203373945641987,
-      "grad_norm": 2.3354814052581787,
+      "grad_norm": 2.511812210083008,
       "learning_rate": 9.384957825679476e-05,
-      "loss": 0.3657,
+      "loss": 0.3686,
       "step": 1050
     },
     {
       "epoch": 0.4967197750702905,
-      "grad_norm": 1.809728980064392,
+      "grad_norm": 2.1349568367004395,
       "learning_rate": 9.379100281162138e-05,
-      "loss": 0.3579,
+      "loss": 0.3496,
       "step": 1060
     },
     {
       "epoch": 0.5014058106841612,
-      "grad_norm": 2.339686155319214,
+      "grad_norm": 2.2301406860351562,
       "learning_rate": 9.373242736644799e-05,
-      "loss": 0.382,
+      "loss": 0.3654,
       "step": 1070
     },
     {
       "epoch": 0.5060918462980318,
-      "grad_norm": 2.170480966567993,
+      "grad_norm": 2.1164608001708984,
       "learning_rate": 9.36738519212746e-05,
-      "loss": 0.3636,
+      "loss": 0.3638,
       "step": 1080
     },
     {
       "epoch": 0.5107778819119025,
-      "grad_norm": 2.4314892292022705,
+      "grad_norm": 2.4411444664001465,
       "learning_rate": 9.361527647610123e-05,
-      "loss": 0.3638,
+      "loss": 0.3587,
       "step": 1090
     },
     {
       "epoch": 0.5154639175257731,
-      "grad_norm": 2.1810007095336914,
+      "grad_norm": 2.258974313735962,
       "learning_rate": 9.355670103092784e-05,
-      "loss": 0.3583,
+      "loss": 0.3623,
       "step": 1100
     },
     {
       "epoch": 0.5154639175257731,
-      "eval_loss": 0.057510774582624435,
-      "eval_pearson_cosine": 0.7969929903339477,
-      "eval_pearson_dot": 0.8054373741957335,
-      "eval_pearson_euclidean": 0.7808607374061509,
-      "eval_pearson_manhattan": 0.7798589297759975,
-      "eval_runtime": 4.7384,
-      "eval_samples_per_second": 316.564,
-      "eval_spearman_cosine": 0.7973557927542151,
-      "eval_spearman_dot": 0.8047121394369887,
-      "eval_spearman_euclidean": 0.7843951879137491,
-      "eval_spearman_manhattan": 0.783835500667981,
-      "eval_steps_per_second": 19.838,
+      "eval_loss": 0.057412039488554,
+      "eval_pearson_cosine": 0.8054045239651941,
+      "eval_pearson_dot": 0.7882266367922739,
+      "eval_pearson_euclidean": 0.8054283430349685,
+      "eval_pearson_manhattan": 0.8054310326990333,
+      "eval_runtime": 4.884,
+      "eval_samples_per_second": 307.123,
+      "eval_spearman_cosine": 0.8089804612877154,
+      "eval_spearman_dot": 0.7907022600197556,
+      "eval_spearman_euclidean": 0.8116072764746199,
+      "eval_spearman_manhattan": 0.8115702025237405,
+      "eval_steps_per_second": 19.246,
       "step": 1100
     },
     {
       "epoch": 0.5201499531396439,
-      "grad_norm": 2.1650052070617676,
+      "grad_norm": 2.1118857860565186,
       "learning_rate": 9.349812558575446e-05,
-      "loss": 0.3565,
+      "loss": 0.3515,
       "step": 1110
     },
     {
       "epoch": 0.5248359887535146,
-      "grad_norm": 1.9347268342971802,
+      "grad_norm": 1.9648866653442383,
       "learning_rate": 9.343955014058108e-05,
-      "loss": 0.3732,
+      "loss": 0.3609,
       "step": 1120
     },
     {
       "epoch": 0.5295220243673852,
-      "grad_norm": 2.7452573776245117,
+      "grad_norm": 2.2861597537994385,
       "learning_rate": 9.338097469540768e-05,
-      "loss": 0.4022,
+      "loss": 0.4036,
       "step": 1130
     },
     {
       "epoch": 0.5342080599812559,
-      "grad_norm": 2.7319650650024414,
+      "grad_norm": 2.585618734359741,
       "learning_rate": 9.33223992502343e-05,
-      "loss": 0.4039,
+      "loss": 0.4082,
       "step": 1140
     },
     {
       "epoch": 0.5388940955951266,
-      "grad_norm": 1.8111310005187988,
+      "grad_norm": 1.9095607995986938,
       "learning_rate": 9.326382380506093e-05,
-      "loss": 0.3445,
+      "loss": 0.3412,
       "step": 1150
     },
     {
       "epoch": 0.5435801312089972,
-      "grad_norm": 2.470259189605713,
+      "grad_norm": 2.4912240505218506,
       "learning_rate": 9.320524835988755e-05,
-      "loss": 0.329,
+      "loss": 0.3241,
       "step": 1160
     },
     {
       "epoch": 0.5482661668228679,
-      "grad_norm": 2.465682029724121,
+      "grad_norm": 2.18245530128479,
       "learning_rate": 9.314667291471416e-05,
-      "loss": 0.3621,
+      "loss": 0.3488,
       "step": 1170
     },
     {
       "epoch": 0.5529522024367385,
-      "grad_norm": 1.7908481359481812,
+      "grad_norm": 1.8920319080352783,
       "learning_rate": 9.308809746954077e-05,
-      "loss": 0.3657,
+      "loss": 0.3681,
       "step": 1180
     },
     {
       "epoch": 0.5576382380506092,
-      "grad_norm": 2.1389901638031006,
+      "grad_norm": 2.3703997135162354,
       "learning_rate": 9.302952202436738e-05,
-      "loss": 0.3388,
+      "loss": 0.3351,
       "step": 1190
     },
     {
       "epoch": 0.5623242736644799,
-      "grad_norm": 2.2106475830078125,
+      "grad_norm": 2.7007317543029785,
       "learning_rate": 9.2970946579194e-05,
-      "loss": 0.3913,
+      "loss": 0.3795,
       "step": 1200
     },
     {
       "epoch": 0.5623242736644799,
-      "eval_loss": 0.06687228381633759,
-      "eval_pearson_cosine": 0.7966150240375356,
-      "eval_pearson_dot": 0.8042618343582149,
-      "eval_pearson_euclidean": 0.7828267973995509,
-      "eval_pearson_manhattan": 0.7833213095644886,
-      "eval_runtime": 4.9951,
-      "eval_samples_per_second": 300.291,
-      "eval_spearman_cosine": 0.8000396168871208,
-      "eval_spearman_dot": 0.8048980184574105,
-      "eval_spearman_euclidean": 0.7859538030557693,
-      "eval_spearman_manhattan": 0.7864758987006061,
-      "eval_steps_per_second": 18.818,
+      "eval_loss": 0.05915520712733269,
+      "eval_pearson_cosine": 0.8027526360325794,
+      "eval_pearson_dot": 0.7769488034596463,
+      "eval_pearson_euclidean": 0.8068785294837539,
+      "eval_pearson_manhattan": 0.8075156260399368,
+      "eval_runtime": 4.746,
+      "eval_samples_per_second": 316.058,
+      "eval_spearman_cosine": 0.8065060357142443,
+      "eval_spearman_dot": 0.7771498848133976,
+      "eval_spearman_euclidean": 0.8122246926374114,
+      "eval_spearman_manhattan": 0.8132591524706569,
+      "eval_steps_per_second": 19.806,
       "step": 1200
     },
     {
       "epoch": 0.5670103092783505,
-      "grad_norm": 1.9803818464279175,
+      "grad_norm": 1.776810646057129,
       "learning_rate": 9.291237113402063e-05,
-      "loss": 0.3435,
+      "loss": 0.3432,
       "step": 1210
     },
     {
       "epoch": 0.5716963448922212,
-      "grad_norm": 1.7664510011672974,
+      "grad_norm": 1.6577345132827759,
       "learning_rate": 9.285379568884725e-05,
-      "loss": 0.3798,
+      "loss": 0.36,
       "step": 1220
     },
     {
       "epoch": 0.5763823805060918,
-      "grad_norm": 1.8455970287322998,
+      "grad_norm": 2.0708200931549072,
       "learning_rate": 9.279522024367385e-05,
-      "loss": 0.3612,
+      "loss": 0.3591,
       "step": 1230
     },
     {
       "epoch": 0.5810684161199625,
-      "grad_norm": 1.92662513256073,
+      "grad_norm": 2.041905403137207,
       "learning_rate": 9.273664479850047e-05,
-      "loss": 0.3214,
+      "loss": 0.3246,
       "step": 1240
     },
     {
       "epoch": 0.5857544517338332,
-      "grad_norm": 1.968152403831482,
+      "grad_norm": 2.0132291316986084,
       "learning_rate": 9.267806935332709e-05,
-      "loss": 0.3362,
+      "loss": 0.3307,
       "step": 1250
     },
     {
       "epoch": 0.5904404873477038,
-      "grad_norm": 2.589308261871338,
+      "grad_norm": 2.6334540843963623,
       "learning_rate": 9.26194939081537e-05,
-      "loss": 0.3742,
+      "loss": 0.3801,
       "step": 1260
     },
     {
       "epoch": 0.5951265229615745,
-      "grad_norm": 1.9938191175460815,
+      "grad_norm": 2.337106227874756,
       "learning_rate": 9.256091846298033e-05,
-      "loss": 0.3507,
+      "loss": 0.3392,
       "step": 1270
     },
     {
       "epoch": 0.5998125585754451,
-      "grad_norm": 2.208752155303955,
+      "grad_norm": 2.2226593494415283,
       "learning_rate": 9.250234301780694e-05,
-      "loss": 0.3396,
+      "loss": 0.3418,
       "step": 1280
     },
     {
       "epoch": 0.6044985941893158,
-      "grad_norm": 1.8320244550704956,
+      "grad_norm": 1.7794549465179443,
       "learning_rate": 9.244376757263355e-05,
-      "loss": 0.3242,
+      "loss": 0.314,
       "step": 1290
     },
     {
       "epoch": 0.6091846298031866,
-      "grad_norm": 2.33010196685791,
+      "grad_norm": 1.9184062480926514,
       "learning_rate": 9.238519212746017e-05,
-      "loss": 0.3208,
+      "loss": 0.3053,
       "step": 1300
     },
     {
       "epoch": 0.6091846298031866,
-      "eval_loss": 0.045637644827365875,
-      "eval_pearson_cosine": 0.8113760130294838,
-      "eval_pearson_dot": 0.8168668534470953,
-      "eval_pearson_euclidean": 0.7946685783976477,
-      "eval_pearson_manhattan": 0.7948391219509574,
-      "eval_runtime": 4.6512,
-      "eval_samples_per_second": 322.5,
-      "eval_spearman_cosine": 0.8117070589902026,
-      "eval_spearman_dot": 0.8166875703437103,
-      "eval_spearman_euclidean": 0.8003786649119463,
-      "eval_spearman_manhattan": 0.8005194055765069,
-      "eval_steps_per_second": 20.21,
+      "eval_loss": 0.045989640057086945,
+      "eval_pearson_cosine": 0.820789253991002,
+      "eval_pearson_dot": 0.8058718276562473,
+      "eval_pearson_euclidean": 0.8166045784832683,
+      "eval_pearson_manhattan": 0.8162631376424372,
+      "eval_runtime": 5.0529,
+      "eval_samples_per_second": 296.861,
+      "eval_spearman_cosine": 0.8220198254899076,
+      "eval_spearman_dot": 0.8055773386238695,
+      "eval_spearman_euclidean": 0.8240178580854226,
+      "eval_spearman_manhattan": 0.8238001131690327,
+      "eval_steps_per_second": 18.603,
       "step": 1300
     },
     {
       "epoch": 0.6138706654170571,
-      "grad_norm": 1.4652146100997925,
+      "grad_norm": 1.4235422611236572,
       "learning_rate": 9.232661668228679e-05,
-      "loss": 0.3707,
+      "loss": 0.3753,
       "step": 1310
     },
     {
       "epoch": 0.6185567010309279,
-      "grad_norm": 2.13409161567688,
+      "grad_norm": 2.533778190612793,
       "learning_rate": 9.22680412371134e-05,
-      "loss": 0.3617,
+      "loss": 0.3428,
       "step": 1320
     },
     {
       "epoch": 0.6232427366447985,
-      "grad_norm": 1.738176941871643,
+      "grad_norm": 1.4714725017547607,
       "learning_rate": 9.220946579194002e-05,
-      "loss": 0.3376,
+      "loss": 0.3234,
       "step": 1330
     },
     {
       "epoch": 0.6279287722586692,
-      "grad_norm": 1.750163197517395,
+      "grad_norm": 2.222369432449341,
       "learning_rate": 9.215089034676664e-05,
-      "loss": 0.3404,
+      "loss": 0.3245,
       "step": 1340
     },
     {
       "epoch": 0.6326148078725399,
-      "grad_norm": 2.1063754558563232,
+      "grad_norm": 2.2116310596466064,
       "learning_rate": 9.209231490159326e-05,
-      "loss": 0.3752,
+      "loss": 0.3559,
       "step": 1350
     },
     {
       "epoch": 0.6373008434864105,
-      "grad_norm": 1.4449872970581055,
+      "grad_norm": 1.7387092113494873,
       "learning_rate": 9.203373945641987e-05,
-      "loss": 0.2979,
+      "loss": 0.3136,
       "step": 1360
     },
     {
       "epoch": 0.6419868791002812,
-      "grad_norm": 2.006011724472046,
+      "grad_norm": 2.0696067810058594,
       "learning_rate": 9.197516401124649e-05,
-      "loss": 0.3501,
+      "loss": 0.3445,
       "step": 1370
     },
     {
       "epoch": 0.6466729147141518,
-      "grad_norm": 2.036403179168701,
+      "grad_norm": 1.9684154987335205,
       "learning_rate": 9.191658856607311e-05,
-      "loss": 0.3557,
+      "loss": 0.3642,
       "step": 1380
     },
     {
       "epoch": 0.6513589503280225,
-      "grad_norm": 2.405442237854004,
+      "grad_norm": 2.100874662399292,
       "learning_rate": 9.185801312089973e-05,
-      "loss": 0.3584,
+      "loss": 0.3503,
       "step": 1390
     },
     {
       "epoch": 0.6560449859418932,
-      "grad_norm": 1.8007841110229492,
+      "grad_norm": 2.0006463527679443,
       "learning_rate": 9.179943767572634e-05,
-      "loss": 0.3247,
+      "loss": 0.3254,
       "step": 1400
     },
     {
       "epoch": 0.6560449859418932,
-      "eval_loss": 0.053980741649866104,
-      "eval_pearson_cosine": 0.8086150300067487,
-      "eval_pearson_dot": 0.8174400112152469,
-      "eval_pearson_euclidean": 0.7907082515140829,
-      "eval_pearson_manhattan": 0.7906201840168094,
-      "eval_runtime": 4.7766,
-      "eval_samples_per_second": 314.029,
-      "eval_spearman_cosine": 0.8111256668964958,
-      "eval_spearman_dot": 0.8195298336175879,
-      "eval_spearman_euclidean": 0.7970481482249502,
-      "eval_spearman_manhattan": 0.7967056753253113,
-      "eval_steps_per_second": 19.679,
+      "eval_loss": 0.05193132907152176,
+      "eval_pearson_cosine": 0.8103823201459353,
+      "eval_pearson_dot": 0.7882454399014449,
+      "eval_pearson_euclidean": 0.811911388094785,
+      "eval_pearson_manhattan": 0.8115250997508099,
+      "eval_runtime": 5.1172,
+      "eval_samples_per_second": 293.129,
+      "eval_spearman_cosine": 0.8131502946747682,
+      "eval_spearman_dot": 0.7891548472904836,
+      "eval_spearman_euclidean": 0.8194175695170715,
+      "eval_spearman_manhattan": 0.8189330381324939,
+      "eval_steps_per_second": 18.369,
       "step": 1400
     },
     {
       "epoch": 0.6607310215557638,
-      "grad_norm": 1.750858187675476,
+      "grad_norm": 1.919021487236023,
       "learning_rate": 9.174086223055296e-05,
-      "loss": 0.3304,
+      "loss": 0.3202,
       "step": 1410
     },
     {
       "epoch": 0.6654170571696345,
-      "grad_norm": 1.7640630006790161,
+      "grad_norm": 1.8564157485961914,
       "learning_rate": 9.168228678537958e-05,
-      "loss": 0.3282,
+      "loss": 0.3096,
       "step": 1420
     },
     {
       "epoch": 0.6701030927835051,
-      "grad_norm": 1.7033814191818237,
+      "grad_norm": 1.6826527118682861,
       "learning_rate": 9.162371134020618e-05,
-      "loss": 0.3544,
+      "loss": 0.3406,
       "step": 1430
     },
     {
       "epoch": 0.6747891283973758,
-      "grad_norm": 2.1169116497039795,
+      "grad_norm": 2.024946689605713,
       "learning_rate": 9.156513589503281e-05,
-      "loss": 0.3599,
+      "loss": 0.3516,
       "step": 1440
     },
     {
       "epoch": 0.6794751640112465,
-      "grad_norm": 1.8890944719314575,
+      "grad_norm": 1.8498772382736206,
       "learning_rate": 9.150656044985943e-05,
-      "loss": 0.3097,
+      "loss": 0.3091,
       "step": 1450
     },
     {
       "epoch": 0.6841611996251171,
-      "grad_norm": 1.9908314943313599,
+      "grad_norm": 1.799222469329834,
       "learning_rate": 9.144798500468605e-05,
-      "loss": 0.3187,
+      "loss": 0.3191,
       "step": 1460
     },
     {
       "epoch": 0.6888472352389878,
-      "grad_norm": 2.442484140396118,
+      "grad_norm": 2.3396248817443848,
       "learning_rate": 9.138940955951266e-05,
-      "loss": 0.3184,
+      "loss": 0.3233,
       "step": 1470
     },
     {
       "epoch": 0.6935332708528584,
-      "grad_norm": 2.19154953956604,
+      "grad_norm": 2.101897716522217,
       "learning_rate": 9.133083411433927e-05,
-      "loss": 0.3585,
+      "loss": 0.3432,
       "step": 1480
     },
     {
       "epoch": 0.6982193064667291,
-      "grad_norm": 1.869988203048706,
+      "grad_norm": 1.8764758110046387,
       "learning_rate": 9.127225866916588e-05,
-      "loss": 0.3149,
+      "loss": 0.2991,
       "step": 1490
     },
     {
       "epoch": 0.7029053420805998,
-      "grad_norm": 2.1679623126983643,
+      "grad_norm": 2.519561767578125,
       "learning_rate": 9.121368322399251e-05,
-      "loss": 0.3269,
+      "loss": 0.3399,
       "step": 1500
     },
     {
       "epoch": 0.7029053420805998,
-      "eval_loss": 0.04739663749933243,
-      "eval_pearson_cosine": 0.8141597203524853,
-      "eval_pearson_dot": 0.8207163609193167,
-      "eval_pearson_euclidean": 0.796674853711913,
-      "eval_pearson_manhattan": 0.7960896801887358,
-      "eval_runtime": 4.6642,
-      "eval_samples_per_second": 321.601,
-      "eval_spearman_cosine": 0.8154525852016754,
-      "eval_spearman_dot": 0.8205176794102382,
-      "eval_spearman_euclidean": 0.8031080794349076,
-      "eval_spearman_manhattan": 0.8025975518083548,
-      "eval_steps_per_second": 20.154,
+      "eval_loss": 0.0482293963432312,
+      "eval_pearson_cosine": 0.8191894627160892,
+      "eval_pearson_dot": 0.8010005297958998,
+      "eval_pearson_euclidean": 0.8162180404574002,
+      "eval_pearson_manhattan": 0.8162308024373885,
+      "eval_runtime": 8.1047,
+      "eval_samples_per_second": 185.079,
+      "eval_spearman_cosine": 0.8200045044762955,
+      "eval_spearman_dot": 0.8002723589197207,
+      "eval_spearman_euclidean": 0.8238380749416172,
+      "eval_spearman_manhattan": 0.8242054788884601,
+      "eval_steps_per_second": 11.598,
       "step": 1500
     },
     {
       "epoch": 0.7075913776944704,
-      "grad_norm": 1.590140700340271,
+      "grad_norm": 1.77321457862854,
       "learning_rate": 9.115510777881913e-05,
-      "loss": 0.2743,
+      "loss": 0.2938,
       "step": 1510
     },
     {
       "epoch": 0.7122774133083412,
-      "grad_norm": 1.7297847270965576,
+      "grad_norm": 1.7782044410705566,
       "learning_rate": 9.109653233364575e-05,
-      "loss": 0.3274,
+      "loss": 0.3371,
       "step": 1520
     },
     {
       "epoch": 0.7169634489222118,
-      "grad_norm": 1.7958931922912598,
+      "grad_norm": 2.416205883026123,
       "learning_rate": 9.103795688847235e-05,
-      "loss": 0.3184,
+      "loss": 0.3222,
       "step": 1530
     },
     {
       "epoch": 0.7216494845360825,
-      "grad_norm": 1.4942468404769897,
+      "grad_norm": 1.8090990781784058,
       "learning_rate": 9.097938144329897e-05,
-      "loss": 0.3055,
+      "loss": 0.3139,
       "step": 1540
     },
     {
       "epoch": 0.7263355201499532,
-      "grad_norm": 1.6255275011062622,
+      "grad_norm": 1.3428120613098145,
       "learning_rate": 9.092080599812559e-05,
-      "loss": 0.3008,
+      "loss": 0.2989,
       "step": 1550
     },
     {
       "epoch": 0.7310215557638238,
-      "grad_norm": 1.8744940757751465,
+      "grad_norm": 1.667705774307251,
       "learning_rate": 9.086223055295222e-05,
-      "loss": 0.3223,
+      "loss": 0.3242,
       "step": 1560
     },
     {
       "epoch": 0.7357075913776945,
-      "grad_norm": 2.46989369392395,
+      "grad_norm": 2.157198905944824,
       "learning_rate": 9.080365510777883e-05,
-      "loss": 0.3255,
+      "loss": 0.3402,
       "step": 1570
     },
     {
       "epoch": 0.7403936269915652,
-      "grad_norm": 1.6579105854034424,
+      "grad_norm": 1.7279380559921265,
       "learning_rate": 9.074507966260544e-05,
-      "loss": 0.3162,
+      "loss": 0.3095,
       "step": 1580
     },
     {
       "epoch": 0.7450796626054358,
-      "grad_norm": 1.74732506275177,
+      "grad_norm": 1.914881706237793,
       "learning_rate": 9.068650421743205e-05,
-      "loss": 0.3602,
+      "loss": 0.3367,
       "step": 1590
     },
     {
       "epoch": 0.7497656982193065,
-      "grad_norm": 2.1454904079437256,
+      "grad_norm": 2.0836734771728516,
       "learning_rate": 9.062792877225867e-05,
-      "loss": 0.3411,
+      "loss": 0.3418,
       "step": 1600
     },
     {
       "epoch": 0.7497656982193065,
-      "eval_loss": 0.04926175996661186,
-      "eval_pearson_cosine": 0.8161287135448276,
-      "eval_pearson_dot": 0.8223874860494362,
-      "eval_pearson_euclidean": 0.8018699359881083,
-      "eval_pearson_manhattan": 0.8013834703697782,
-      "eval_runtime": 4.7094,
-      "eval_samples_per_second": 318.509,
-      "eval_spearman_cosine": 0.8189562831716073,
-      "eval_spearman_dot": 0.8236193691999756,
-      "eval_spearman_euclidean": 0.8078062231815545,
-      "eval_spearman_manhattan": 0.8072727629308311,
-      "eval_steps_per_second": 19.96,
+      "eval_loss": 0.05068698152899742,
+      "eval_pearson_cosine": 0.8195360561590519,
+      "eval_pearson_dot": 0.7972572526105068,
+      "eval_pearson_euclidean": 0.8191033648513582,
+      "eval_pearson_manhattan": 0.8185097560985408,
+      "eval_runtime": 7.465,
+      "eval_samples_per_second": 200.937,
+      "eval_spearman_cosine": 0.8226781085038538,
+      "eval_spearman_dot": 0.7978348943334088,
+      "eval_spearman_euclidean": 0.8252181755470007,
+      "eval_spearman_manhattan": 0.824758318207313,
+      "eval_steps_per_second": 12.592,
       "step": 1600
     },
     {
       "epoch": 0.7544517338331771,
-      "grad_norm": 2.28283953666687,
+      "grad_norm": 2.3218607902526855,
       "learning_rate": 9.056935332708529e-05,
-      "loss": 0.3234,
+      "loss": 0.3163,
       "step": 1610
     },
     {
       "epoch": 0.7591377694470478,
-      "grad_norm": 1.6917240619659424,
+      "grad_norm": 1.7433195114135742,
       "learning_rate": 9.051077788191192e-05,
-      "loss": 0.352,
+      "loss": 0.3435,
       "step": 1620
     },
     {
       "epoch": 0.7638238050609185,
-      "grad_norm": 1.712475061416626,
+      "grad_norm": 2.307793140411377,
       "learning_rate": 9.045220243673852e-05,
-      "loss": 0.3107,
+      "loss": 0.3101,
       "step": 1630
     },
     {
       "epoch": 0.7685098406747891,
-      "grad_norm": 1.6429861783981323,
+      "grad_norm": 2.072714328765869,
       "learning_rate": 9.039362699156514e-05,
-      "loss": 0.3175,
+      "loss": 0.3343,
       "step": 1640
     },
     {
       "epoch": 0.7731958762886598,
-      "grad_norm": 1.5003321170806885,
+      "grad_norm": 1.672747015953064,
       "learning_rate": 9.033505154639176e-05,
-      "loss": 0.3502,
+      "loss": 0.3475,
       "step": 1650
     },
     {
       "epoch": 0.7778819119025304,
-      "grad_norm": 1.9713871479034424,
+      "grad_norm": 1.9799168109893799,
       "learning_rate": 9.027647610121837e-05,
-      "loss": 0.3691,
+      "loss": 0.3658,
       "step": 1660
     },
     {
       "epoch": 0.7825679475164011,
-      "grad_norm": 2.0830156803131104,
+      "grad_norm": 2.181130886077881,
       "learning_rate": 9.021790065604499e-05,
-      "loss": 0.3226,
+      "loss": 0.2987,
       "step": 1670
     },
     {
       "epoch": 0.7872539831302718,
-      "grad_norm": 2.0364925861358643,
+      "grad_norm": 1.9070302248001099,
       "learning_rate": 9.015932521087161e-05,
-      "loss": 0.3344,
+      "loss": 0.3301,
       "step": 1680
     },
     {
       "epoch": 0.7919400187441424,
-      "grad_norm": 1.7688632011413574,
+      "grad_norm": 1.7219617366790771,
       "learning_rate": 9.010074976569822e-05,
-      "loss": 0.328,
+      "loss": 0.3414,
       "step": 1690
     },
     {
       "epoch": 0.7966260543580131,
-      "grad_norm": 2.0706257820129395,
+      "grad_norm": 1.836569905281067,
       "learning_rate": 9.004217432052484e-05,
-      "loss": 0.3284,
+      "loss": 0.329,
       "step": 1700
     },
     {
       "epoch": 0.7966260543580131,
-      "eval_loss": 0.042656708508729935,
-      "eval_pearson_cosine": 0.8112926503111737,
-      "eval_pearson_dot": 0.818754256660867,
-      "eval_pearson_euclidean": 0.7946703633011838,
-      "eval_pearson_manhattan": 0.7944877991875288,
-      "eval_runtime": 4.8215,
-      "eval_samples_per_second": 311.103,
-      "eval_spearman_cosine": 0.812852551114096,
-      "eval_spearman_dot": 0.8198186422414758,
-      "eval_spearman_euclidean": 0.8004155438327007,
-      "eval_spearman_manhattan": 0.8002132702523219,
-      "eval_steps_per_second": 19.496,
+      "eval_loss": 0.048954952508211136,
+      "eval_pearson_cosine": 0.806191191357641,
+      "eval_pearson_dot": 0.7809211131919511,
+      "eval_pearson_euclidean": 0.8070190601338894,
+      "eval_pearson_manhattan": 0.8065529338804547,
+      "eval_runtime": 8.988,
+      "eval_samples_per_second": 166.888,
+      "eval_spearman_cosine": 0.807531064838451,
+      "eval_spearman_dot": 0.7785249437738705,
+      "eval_spearman_euclidean": 0.8131489823225247,
+      "eval_spearman_manhattan": 0.8125461186879448,
+      "eval_steps_per_second": 10.458,
       "step": 1700
     },
     {
       "epoch": 0.8013120899718837,
-      "grad_norm": 1.6608582735061646,
+      "grad_norm": 1.7277847528457642,
       "learning_rate": 8.998359887535146e-05,
-      "loss": 0.3321,
+      "loss": 0.3387,
       "step": 1710
     },
     {
       "epoch": 0.8059981255857545,
-      "grad_norm": 2.178337574005127,
+      "grad_norm": 2.1043872833251953,
       "learning_rate": 8.992502343017808e-05,
-      "loss": 0.3289,
+      "loss": 0.3209,
       "step": 1720
     },
     {
       "epoch": 0.8106841611996252,
-      "grad_norm": 2.123746633529663,
+      "grad_norm": 1.989742398262024,
       "learning_rate": 8.986644798500469e-05,
-      "loss": 0.3393,
+      "loss": 0.3244,
       "step": 1730
     },
     {
       "epoch": 0.8153701968134958,
-      "grad_norm": 1.5920015573501587,
+      "grad_norm": 1.8288764953613281,
       "learning_rate": 8.980787253983131e-05,
-      "loss": 0.3137,
+      "loss": 0.3177,
       "step": 1740
     },
     {
       "epoch": 0.8200562324273665,
-      "grad_norm": 1.6964048147201538,
+      "grad_norm": 1.7958714962005615,
       "learning_rate": 8.974929709465793e-05,
-      "loss": 0.3242,
+      "loss": 0.3055,
       "step": 1750
     },
     {
       "epoch": 0.8247422680412371,
-      "grad_norm": 2.2545530796051025,
+      "grad_norm": 2.055933952331543,
       "learning_rate": 8.969072164948454e-05,
-      "loss": 0.3473,
+      "loss": 0.3535,
       "step": 1760
     },
     {
       "epoch": 0.8294283036551078,
-      "grad_norm": 1.829145073890686,
+      "grad_norm": 2.179781436920166,
       "learning_rate": 8.963214620431116e-05,
-      "loss": 0.2871,
+      "loss": 0.2989,
       "step": 1770
     },
     {
       "epoch": 0.8341143392689785,
-      "grad_norm": 1.7790557146072388,
+      "grad_norm": 1.6793406009674072,
       "learning_rate": 8.957357075913777e-05,
-      "loss": 0.2979,
+      "loss": 0.3113,
       "step": 1780
     },
     {
       "epoch": 0.8388003748828491,
-      "grad_norm": 1.6724668741226196,
+      "grad_norm": 1.4585118293762207,
       "learning_rate": 8.95149953139644e-05,
-      "loss": 0.3094,
+      "loss": 0.3185,
       "step": 1790
     },
     {
       "epoch": 0.8434864104967198,
-      "grad_norm": 1.939370036125183,
+      "grad_norm": 2.087599277496338,
       "learning_rate": 8.945641986879101e-05,
-      "loss": 0.2845,
+      "loss": 0.2774,
       "step": 1800
     },
     {
       "epoch": 0.8434864104967198,
-      "eval_loss": 0.04392844811081886,
-      "eval_pearson_cosine": 0.8097501259779989,
-      "eval_pearson_dot": 0.8164764225020082,
-      "eval_pearson_euclidean": 0.7921053664943116,
-      "eval_pearson_manhattan": 0.7914097303155287,
-      "eval_runtime": 4.6495,
-      "eval_samples_per_second": 322.617,
-      "eval_spearman_cosine": 0.8123240085388628,
-      "eval_spearman_dot": 0.8183737971290131,
-      "eval_spearman_euclidean": 0.7989029211686818,
-      "eval_spearman_manhattan": 0.7978239561899777,
-      "eval_steps_per_second": 20.217,
+      "eval_loss": 0.04364720731973648,
+      "eval_pearson_cosine": 0.8131129034275588,
+      "eval_pearson_dot": 0.781156666273711,
+      "eval_pearson_euclidean": 0.812444188325955,
+      "eval_pearson_manhattan": 0.8124379952641316,
+      "eval_runtime": 8.1518,
+      "eval_samples_per_second": 184.008,
+      "eval_spearman_cosine": 0.8136297625669392,
+      "eval_spearman_dot": 0.7793439007721238,
+      "eval_spearman_euclidean": 0.8212822580110872,
+      "eval_spearman_manhattan": 0.8211590955757442,
+      "eval_steps_per_second": 11.531,
       "step": 1800
     },
     {
       "epoch": 0.8481724461105904,
-      "grad_norm": 1.730468511581421,
+      "grad_norm": 2.398632049560547,
       "learning_rate": 8.939784442361763e-05,
-      "loss": 0.3206,
+      "loss": 0.3166,
       "step": 1810
     },
     {
       "epoch": 0.8528584817244611,
-      "grad_norm": 1.5852857828140259,
+      "grad_norm": 2.182029962539673,
       "learning_rate": 8.933926897844423e-05,
-      "loss": 0.3159,
+      "loss": 0.3131,
       "step": 1820
     },
     {
       "epoch": 0.8575445173383318,
-      "grad_norm": 1.507568597793579,
+      "grad_norm": 1.647484540939331,
       "learning_rate": 8.928069353327085e-05,
-      "loss": 0.3019,
+      "loss": 0.2974,
       "step": 1830
     },
     {
       "epoch": 0.8622305529522024,
-      "grad_norm": 1.661995768547058,
+      "grad_norm": 1.591589331626892,
       "learning_rate": 8.922211808809747e-05,
-      "loss": 0.3407,
+      "loss": 0.3261,
       "step": 1840
     },
     {
       "epoch": 0.8669165885660731,
-      "grad_norm": 1.877208948135376,
+      "grad_norm": 1.9286956787109375,
       "learning_rate": 8.91635426429241e-05,
-      "loss": 0.322,
+      "loss": 0.3333,
       "step": 1850
     },
     {
       "epoch": 0.8716026241799437,
-      "grad_norm": 1.6956888437271118,
+      "grad_norm": 1.7909513711929321,
       "learning_rate": 8.910496719775072e-05,
-      "loss": 0.3434,
+      "loss": 0.3419,
       "step": 1860
     },
     {
       "epoch": 0.8762886597938144,
-      "grad_norm": 1.8181349039077759,
+      "grad_norm": 1.7410364151000977,
       "learning_rate": 8.904639175257732e-05,
-      "loss": 0.316,
+      "loss": 0.3146,
       "step": 1870
     },
     {
       "epoch": 0.8809746954076851,
-      "grad_norm": 1.7894842624664307,
+      "grad_norm": 1.7161133289337158,
       "learning_rate": 8.898781630740394e-05,
-      "loss": 0.2881,
+      "loss": 0.3005,
       "step": 1880
     },
     {
       "epoch": 0.8856607310215557,
-      "grad_norm": 2.1654884815216064,
+      "grad_norm": 2.411195755004883,
       "learning_rate": 8.892924086223055e-05,
-      "loss": 0.3423,
+      "loss": 0.3334,
       "step": 1890
     },
     {
       "epoch": 0.8903467666354264,
-      "grad_norm": 1.6889327764511108,
+      "grad_norm": 2.475139617919922,
       "learning_rate": 8.887066541705717e-05,
-      "loss": 0.3199,
+      "loss": 0.3113,
       "step": 1900
     },
     {
       "epoch": 0.8903467666354264,
-      "eval_loss": 0.05001167580485344,
-      "eval_pearson_cosine": 0.8149866383088806,
-      "eval_pearson_dot": 0.8233746991643329,
-      "eval_pearson_euclidean": 0.7989509074968737,
-      "eval_pearson_manhattan": 0.7988165127794176,
-      "eval_runtime": 4.7223,
-      "eval_samples_per_second": 317.643,
-      "eval_spearman_cosine": 0.819837111960085,
-      "eval_spearman_dot": 0.8273224990756979,
-      "eval_spearman_euclidean": 0.805295090022205,
-      "eval_spearman_manhattan": 0.8050664910203635,
-      "eval_steps_per_second": 19.906,
+      "eval_loss": 0.05027003958821297,
+      "eval_pearson_cosine": 0.8191568367121675,
+      "eval_pearson_dot": 0.7944326448061823,
+      "eval_pearson_euclidean": 0.8190699446663956,
+      "eval_pearson_manhattan": 0.8186706917312714,
+      "eval_runtime": 8.5204,
+      "eval_samples_per_second": 176.048,
+      "eval_spearman_cosine": 0.8218975251970799,
+      "eval_spearman_dot": 0.7945150952501927,
+      "eval_spearman_euclidean": 0.8259739792538716,
+      "eval_spearman_manhattan": 0.8255935921333749,
+      "eval_steps_per_second": 11.032,
       "step": 1900
     },
     {
       "epoch": 0.895032802249297,
-      "grad_norm": 2.3411951065063477,
+      "grad_norm": 2.144102096557617,
       "learning_rate": 8.88120899718838e-05,
-      "loss": 0.3137,
+      "loss": 0.3513,
       "step": 1910
     },
     {
       "epoch": 0.8997188378631678,
-      "grad_norm": 2.0821354389190674,
+      "grad_norm": 2.026171922683716,
       "learning_rate": 8.87535145267104e-05,
-      "loss": 0.2934,
+      "loss": 0.3079,
       "step": 1920
     },
     {
       "epoch": 0.9044048734770385,
-      "grad_norm": 1.8365401029586792,
+      "grad_norm": 1.6299704313278198,
       "learning_rate": 8.869493908153702e-05,
-      "loss": 0.3322,
+      "loss": 0.3211,
       "step": 1930
     },
     {
       "epoch": 0.9090909090909091,
-      "grad_norm": 2.154188632965088,
+      "grad_norm": 2.224339485168457,
       "learning_rate": 8.863636363636364e-05,
       "loss": 0.2862,
       "step": 1940
     },
     {
       "epoch": 0.9137769447047798,
-      "grad_norm": 2.136406421661377,
+      "grad_norm": 2.2343602180480957,
       "learning_rate": 8.857778819119026e-05,
-      "loss": 0.3336,
+      "loss": 0.3254,
       "step": 1950
     },
     {
       "epoch": 0.9184629803186504,
-      "grad_norm": 2.426302671432495,
+      "grad_norm": 2.2561750411987305,
       "learning_rate": 8.851921274601687e-05,
-      "loss": 0.3486,
+      "loss": 0.3543,
       "step": 1960
     },
     {
       "epoch": 0.9231490159325211,
-      "grad_norm": 2.407034397125244,
+      "grad_norm": 2.290877342224121,
       "learning_rate": 8.846063730084349e-05,
-      "loss": 0.3377,
+      "loss": 0.3441,
       "step": 1970
     },
     {
       "epoch": 0.9278350515463918,
-      "grad_norm": 2.170762062072754,
+      "grad_norm": 2.0177667140960693,
       "learning_rate": 8.840206185567011e-05,
-      "loss": 0.2904,
+      "loss": 0.294,
       "step": 1980
     },
     {
       "epoch": 0.9325210871602624,
-      "grad_norm": 1.6726795434951782,
+      "grad_norm": 1.5551223754882812,
       "learning_rate": 8.834348641049672e-05,
-      "loss": 0.3535,
+      "loss": 0.3575,
       "step": 1990
     },
     {
       "epoch": 0.9372071227741331,
-      "grad_norm": 1.835772156715393,
+      "grad_norm": 2.153193235397339,
       "learning_rate": 8.828491096532334e-05,
-      "loss": 0.3544,
+      "loss": 0.3495,
       "step": 2000
     },
     {
       "epoch": 0.9372071227741331,
-      "eval_loss": 0.06010111793875694,
-      "eval_pearson_cosine": 0.7969487570615001,
-      "eval_pearson_dot": 0.8113433756657784,
-      "eval_pearson_euclidean": 0.7775976461343248,
-      "eval_pearson_manhattan": 0.7779361917572194,
-      "eval_runtime": 4.7319,
-      "eval_samples_per_second": 316.998,
-      "eval_spearman_cosine": 0.8015033854161633,
-      "eval_spearman_dot": 0.8149867800941926,
-      "eval_spearman_euclidean": 0.7812746949762143,
-      "eval_spearman_manhattan": 0.7818782293686074,
-      "eval_steps_per_second": 19.865,
+      "eval_loss": 0.0519048236310482,
+      "eval_pearson_cosine": 0.8078889701459957,
+      "eval_pearson_dot": 0.7809871272349067,
+      "eval_pearson_euclidean": 0.8121397338443259,
+      "eval_pearson_manhattan": 0.8123168684391544,
+      "eval_runtime": 7.7437,
+      "eval_samples_per_second": 193.706,
+      "eval_spearman_cosine": 0.8110291834886367,
+      "eval_spearman_dot": 0.7809741520898368,
+      "eval_spearman_euclidean": 0.8183453813795359,
+      "eval_spearman_manhattan": 0.8184355582057805,
+      "eval_steps_per_second": 12.139,
       "step": 2000
     },
     {
       "epoch": 0.9418931583880038,
-      "grad_norm": 1.3811417818069458,
+      "grad_norm": 1.2223283052444458,
       "learning_rate": 8.822633552014996e-05,
-      "loss": 0.2759,
+      "loss": 0.2682,
       "step": 2010
     },
     {
       "epoch": 0.9465791940018744,
-      "grad_norm": 2.8094141483306885,
+      "grad_norm": 2.9590303897857666,
       "learning_rate": 8.816776007497658e-05,
-      "loss": 0.3046,
+      "loss": 0.3167,
       "step": 2020
     },
     {
       "epoch": 0.9512652296157451,
-      "grad_norm": 1.8067703247070312,
+      "grad_norm": 1.9395133256912231,
       "learning_rate": 8.810918462980319e-05,
-      "loss": 0.314,
+      "loss": 0.2964,
       "step": 2030
     },
     {
       "epoch": 0.9559512652296157,
-      "grad_norm": 1.9540753364562988,
+      "grad_norm": 1.6027612686157227,
       "learning_rate": 8.805060918462981e-05,
-      "loss": 0.2965,
+      "loss": 0.2914,
       "step": 2040
     },
     {
       "epoch": 0.9606373008434864,
-      "grad_norm": 2.20998477935791,
+      "grad_norm": 2.282153606414795,
       "learning_rate": 8.799203373945643e-05,
-      "loss": 0.3278,
+      "loss": 0.3169,
       "step": 2050
     },
     {
       "epoch": 0.9653233364573571,
-      "grad_norm": 2.156224250793457,
+      "grad_norm": 2.311307668685913,
       "learning_rate": 8.793345829428304e-05,
-      "loss": 0.3202,
+      "loss": 0.3217,
       "step": 2060
     },
     {
       "epoch": 0.9700093720712277,
-      "grad_norm": 1.6349718570709229,
+      "grad_norm": 1.5998715162277222,
       "learning_rate": 8.787488284910965e-05,
-      "loss": 0.3192,
+      "loss": 0.3168,
       "step": 2070
     },
     {
       "epoch": 0.9746954076850984,
-      "grad_norm": 1.7507764101028442,
+      "grad_norm": 2.0307867527008057,
       "learning_rate": 8.781630740393628e-05,
-      "loss": 0.3148,
+      "loss": 0.3251,
       "step": 2080
     },
     {
       "epoch": 0.979381443298969,
-      "grad_norm": 1.6444741487503052,
+      "grad_norm": 1.5513020753860474,
       "learning_rate": 8.77577319587629e-05,
-      "loss": 0.3009,
+      "loss": 0.322,
       "step": 2090
     },
     {
       "epoch": 0.9840674789128397,
-      "grad_norm": 1.5341808795928955,
+      "grad_norm": 1.8426481485366821,
       "learning_rate": 8.769915651358951e-05,
-      "loss": 0.3119,
+      "loss": 0.3015,
       "step": 2100
     },
     {
       "epoch": 0.9840674789128397,
-      "eval_loss": 0.04616248607635498,
-      "eval_pearson_cosine": 0.8028400286804569,
-      "eval_pearson_dot": 0.8142363636471543,
-      "eval_pearson_euclidean": 0.7825157658447595,
-      "eval_pearson_manhattan": 0.7827871893341012,
-      "eval_runtime": 4.9143,
-      "eval_samples_per_second": 305.23,
-      "eval_spearman_cosine": 0.805895283206979,
-      "eval_spearman_dot": 0.8167650665649905,
-      "eval_spearman_euclidean": 0.7880766458177777,
-      "eval_spearman_manhattan": 0.7880451603002018,
-      "eval_steps_per_second": 19.128,
+      "eval_loss": 0.04664711281657219,
+      "eval_pearson_cosine": 0.8138371307223906,
+      "eval_pearson_dot": 0.7830982581720178,
+      "eval_pearson_euclidean": 0.815901838580178,
+      "eval_pearson_manhattan": 0.8155626928902295,
+      "eval_runtime": 5.703,
+      "eval_samples_per_second": 263.019,
+      "eval_spearman_cosine": 0.8154952876855104,
+      "eval_spearman_dot": 0.7815036030773422,
+      "eval_spearman_euclidean": 0.8247759821039686,
+      "eval_spearman_manhattan": 0.824388025102129,
+      "eval_steps_per_second": 16.483,
       "step": 2100
     },
     {
       "epoch": 0.9887535145267105,
-      "grad_norm": 2.1978578567504883,
+      "grad_norm": 2.0757949352264404,
       "learning_rate": 8.764058106841613e-05,
-      "loss": 0.3473,
+      "loss": 0.3508,
       "step": 2110
     },
     {
       "epoch": 0.993439550140581,
-      "grad_norm": 1.6111299991607666,
+      "grad_norm": 1.5524593591690063,
       "learning_rate": 8.758200562324273e-05,
-      "loss": 0.2743,
+      "loss": 0.2877,
       "step": 2120
     },
     {
       "epoch": 0.9981255857544518,
-      "grad_norm": 2.0017919540405273,
+      "grad_norm": 2.00026273727417,
       "learning_rate": 8.752343017806935e-05,
-      "loss": 0.3201,
+      "loss": 0.3261,
       "step": 2130
     }
   ],