ReToken-InternVL3.5-8B / trainer_state.json
AvaXiao's picture
Upload folder using huggingface_hub
2f2860c verified
Raw
History Blame Contribute Delete
246 kB
{
"best_global_step": null,
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 1.0,
"eval_steps": 500,
"global_step": 1105,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.0009053870529651426,
"grad_norm": 28.25,
"learning_rate": 0.0,
"loss": 1.6612,
"retrieval/pos_neg_gap_layer35": 0.07887314260005951,
"step": 1
},
{
"epoch": 0.0018107741059302852,
"grad_norm": 25.875,
"learning_rate": 2.9999999999999997e-06,
"loss": 1.6402,
"retrieval/pos_neg_gap_layer35": 0.0631406232714653,
"step": 2
},
{
"epoch": 0.002716161158895428,
"grad_norm": 24.5,
"learning_rate": 5.999999999999999e-06,
"loss": 1.6315,
"retrieval/pos_neg_gap_layer35": 0.07465092837810516,
"step": 3
},
{
"epoch": 0.0036215482118605704,
"grad_norm": 20.5,
"learning_rate": 8.999999999999999e-06,
"loss": 1.6444,
"retrieval/pos_neg_gap_layer35": 0.0639418214559555,
"step": 4
},
{
"epoch": 0.004526935264825713,
"grad_norm": 27.875,
"learning_rate": 1.1999999999999999e-05,
"loss": 1.6035,
"retrieval/pos_neg_gap_layer35": 0.06895973533391953,
"step": 5
},
{
"epoch": 0.005432322317790856,
"grad_norm": 19.375,
"learning_rate": 1.4999999999999999e-05,
"loss": 1.5717,
"retrieval/pos_neg_gap_layer35": 0.07463866472244263,
"step": 6
},
{
"epoch": 0.006337709370755998,
"grad_norm": 17.625,
"learning_rate": 1.7999999999999997e-05,
"loss": 1.5505,
"retrieval/pos_neg_gap_layer35": 0.06858367472887039,
"step": 7
},
{
"epoch": 0.007243096423721141,
"grad_norm": 18.625,
"learning_rate": 2.1e-05,
"loss": 1.5117,
"retrieval/pos_neg_gap_layer35": 0.07422861456871033,
"step": 8
},
{
"epoch": 0.008148483476686284,
"grad_norm": 15.75,
"learning_rate": 2.3999999999999997e-05,
"loss": 1.5048,
"retrieval/pos_neg_gap_layer35": 0.06635777652263641,
"step": 9
},
{
"epoch": 0.009053870529651425,
"grad_norm": 11.625,
"learning_rate": 2.6999999999999996e-05,
"loss": 1.4444,
"retrieval/pos_neg_gap_layer35": 0.06809765100479126,
"step": 10
},
{
"epoch": 0.009959257582616569,
"grad_norm": 8.875,
"learning_rate": 2.9999999999999997e-05,
"loss": 1.3648,
"retrieval/pos_neg_gap_layer35": 0.0853898823261261,
"step": 11
},
{
"epoch": 0.010864644635581712,
"grad_norm": 7.15625,
"learning_rate": 3.2999999999999996e-05,
"loss": 1.3223,
"retrieval/pos_neg_gap_layer35": 0.08562920987606049,
"step": 12
},
{
"epoch": 0.011770031688546853,
"grad_norm": 5.625,
"learning_rate": 3.5999999999999994e-05,
"loss": 1.2621,
"retrieval/pos_neg_gap_layer35": 0.08723468333482742,
"step": 13
},
{
"epoch": 0.012675418741511997,
"grad_norm": 6.53125,
"learning_rate": 3.9e-05,
"loss": 1.2681,
"retrieval/pos_neg_gap_layer35": 0.07135563343763351,
"step": 14
},
{
"epoch": 0.013580805794477138,
"grad_norm": 8.5625,
"learning_rate": 4.2e-05,
"loss": 1.2629,
"retrieval/pos_neg_gap_layer35": 0.07767720893025398,
"step": 15
},
{
"epoch": 0.014486192847442281,
"grad_norm": 9.125,
"learning_rate": 4.4999999999999996e-05,
"loss": 1.2877,
"retrieval/pos_neg_gap_layer35": 0.06703323870897293,
"step": 16
},
{
"epoch": 0.015391579900407425,
"grad_norm": 12.5,
"learning_rate": 4.7999999999999994e-05,
"loss": 1.2641,
"retrieval/pos_neg_gap_layer35": 0.08898844942450523,
"step": 17
},
{
"epoch": 0.016296966953372568,
"grad_norm": 12.125,
"learning_rate": 5.1e-05,
"loss": 1.2586,
"retrieval/pos_neg_gap_layer35": 0.0924573764204979,
"step": 18
},
{
"epoch": 0.017202354006337708,
"grad_norm": 12.375,
"learning_rate": 5.399999999999999e-05,
"loss": 1.2697,
"retrieval/pos_neg_gap_layer35": 0.09430341422557831,
"step": 19
},
{
"epoch": 0.01810774105930285,
"grad_norm": 12.6875,
"learning_rate": 5.6999999999999996e-05,
"loss": 1.2698,
"retrieval/pos_neg_gap_layer35": 0.09796033054590225,
"step": 20
},
{
"epoch": 0.019013128112267994,
"grad_norm": 11.1875,
"learning_rate": 5.9999999999999995e-05,
"loss": 1.2117,
"retrieval/pos_neg_gap_layer35": 0.124989815056324,
"step": 21
},
{
"epoch": 0.019918515165233137,
"grad_norm": 12.875,
"learning_rate": 6.299999999999999e-05,
"loss": 1.1927,
"retrieval/pos_neg_gap_layer35": 0.13117096200585365,
"step": 22
},
{
"epoch": 0.02082390221819828,
"grad_norm": 9.875,
"learning_rate": 6.599999999999999e-05,
"loss": 1.1805,
"retrieval/pos_neg_gap_layer35": 0.13677415251731873,
"step": 23
},
{
"epoch": 0.021729289271163424,
"grad_norm": 9.0,
"learning_rate": 6.9e-05,
"loss": 1.1417,
"retrieval/pos_neg_gap_layer35": 0.14854301512241364,
"step": 24
},
{
"epoch": 0.022634676324128564,
"grad_norm": 6.4375,
"learning_rate": 7.199999999999999e-05,
"loss": 1.1067,
"retrieval/pos_neg_gap_layer35": 0.16532203555107117,
"step": 25
},
{
"epoch": 0.023540063377093707,
"grad_norm": 6.5625,
"learning_rate": 7.5e-05,
"loss": 1.0978,
"retrieval/pos_neg_gap_layer35": 0.1693706437945366,
"step": 26
},
{
"epoch": 0.02444545043005885,
"grad_norm": 6.5625,
"learning_rate": 7.8e-05,
"loss": 1.065,
"retrieval/pos_neg_gap_layer35": 0.19212086498737335,
"step": 27
},
{
"epoch": 0.025350837483023993,
"grad_norm": 5.59375,
"learning_rate": 8.1e-05,
"loss": 1.075,
"retrieval/pos_neg_gap_layer35": 0.18652977421879768,
"step": 28
},
{
"epoch": 0.026256224535989137,
"grad_norm": 6.375,
"learning_rate": 8.4e-05,
"loss": 1.0384,
"retrieval/pos_neg_gap_layer35": 0.20841557160019875,
"step": 29
},
{
"epoch": 0.027161611588954276,
"grad_norm": 6.0,
"learning_rate": 8.699999999999999e-05,
"loss": 1.017,
"retrieval/pos_neg_gap_layer35": 0.22711872309446335,
"step": 30
},
{
"epoch": 0.02806699864191942,
"grad_norm": 5.46875,
"learning_rate": 8.999999999999999e-05,
"loss": 0.989,
"retrieval/pos_neg_gap_layer35": 0.2427726574242115,
"step": 31
},
{
"epoch": 0.028972385694884563,
"grad_norm": 6.5,
"learning_rate": 9.3e-05,
"loss": 1.0278,
"retrieval/pos_neg_gap_layer35": 0.2173640839755535,
"step": 32
},
{
"epoch": 0.029877772747849706,
"grad_norm": 5.1875,
"learning_rate": 9.599999999999999e-05,
"loss": 0.9617,
"retrieval/pos_neg_gap_layer35": 0.2581077106297016,
"step": 33
},
{
"epoch": 0.03078315980081485,
"grad_norm": 5.46875,
"learning_rate": 9.9e-05,
"loss": 0.9065,
"retrieval/pos_neg_gap_layer35": 0.2899169661104679,
"step": 34
},
{
"epoch": 0.03168854685377999,
"grad_norm": 4.875,
"learning_rate": 0.000102,
"loss": 0.8726,
"retrieval/pos_neg_gap_layer35": 0.3188772052526474,
"step": 35
},
{
"epoch": 0.032593933906745136,
"grad_norm": 6.21875,
"learning_rate": 0.00010499999999999999,
"loss": 0.8397,
"retrieval/pos_neg_gap_layer35": 0.33143163472414017,
"step": 36
},
{
"epoch": 0.033499320959710276,
"grad_norm": 6.03125,
"learning_rate": 0.00010799999999999998,
"loss": 0.792,
"retrieval/pos_neg_gap_layer35": 0.36378253251314163,
"step": 37
},
{
"epoch": 0.034404708012675415,
"grad_norm": 6.09375,
"learning_rate": 0.00011099999999999999,
"loss": 0.8065,
"retrieval/pos_neg_gap_layer35": 0.3615504018962383,
"step": 38
},
{
"epoch": 0.03531009506564056,
"grad_norm": 6.15625,
"learning_rate": 0.00011399999999999999,
"loss": 0.7609,
"retrieval/pos_neg_gap_layer35": 0.39002688229084015,
"step": 39
},
{
"epoch": 0.0362154821186057,
"grad_norm": 6.15625,
"learning_rate": 0.000117,
"loss": 0.719,
"retrieval/pos_neg_gap_layer35": 0.4134774021804333,
"step": 40
},
{
"epoch": 0.03712086917157085,
"grad_norm": 4.625,
"learning_rate": 0.00011999999999999999,
"loss": 0.7105,
"retrieval/pos_neg_gap_layer35": 0.430039681494236,
"step": 41
},
{
"epoch": 0.03802625622453599,
"grad_norm": 4.53125,
"learning_rate": 0.00012299999999999998,
"loss": 0.7237,
"retrieval/pos_neg_gap_layer35": 0.42825623974204063,
"step": 42
},
{
"epoch": 0.038931643277501135,
"grad_norm": 3.84375,
"learning_rate": 0.00012599999999999997,
"loss": 0.6861,
"retrieval/pos_neg_gap_layer35": 0.4501771777868271,
"step": 43
},
{
"epoch": 0.039837030330466275,
"grad_norm": 5.21875,
"learning_rate": 0.000129,
"loss": 0.6557,
"retrieval/pos_neg_gap_layer35": 0.4686204716563225,
"step": 44
},
{
"epoch": 0.040742417383431415,
"grad_norm": 3.328125,
"learning_rate": 0.00013199999999999998,
"loss": 0.6488,
"retrieval/pos_neg_gap_layer35": 0.47880157828330994,
"step": 45
},
{
"epoch": 0.04164780443639656,
"grad_norm": 3.0625,
"learning_rate": 0.000135,
"loss": 0.6012,
"retrieval/pos_neg_gap_layer35": 0.508323710411787,
"step": 46
},
{
"epoch": 0.0425531914893617,
"grad_norm": 3.65625,
"learning_rate": 0.000138,
"loss": 0.5614,
"retrieval/pos_neg_gap_layer35": 0.530730850994587,
"step": 47
},
{
"epoch": 0.04345857854232685,
"grad_norm": 4.40625,
"learning_rate": 0.00014099999999999998,
"loss": 0.5424,
"retrieval/pos_neg_gap_layer35": 0.5443335846066475,
"step": 48
},
{
"epoch": 0.04436396559529199,
"grad_norm": 5.53125,
"learning_rate": 0.00014399999999999998,
"loss": 0.5855,
"retrieval/pos_neg_gap_layer35": 0.5243937522172928,
"step": 49
},
{
"epoch": 0.04526935264825713,
"grad_norm": 3.65625,
"learning_rate": 0.000147,
"loss": 0.5274,
"retrieval/pos_neg_gap_layer35": 0.55564384162426,
"step": 50
},
{
"epoch": 0.046174739701222274,
"grad_norm": 3.25,
"learning_rate": 0.00015,
"loss": 0.5822,
"retrieval/pos_neg_gap_layer35": 0.5306160524487495,
"step": 51
},
{
"epoch": 0.047080126754187414,
"grad_norm": 2.5,
"learning_rate": 0.00015299999999999998,
"loss": 0.5218,
"retrieval/pos_neg_gap_layer35": 0.5698486715555191,
"step": 52
},
{
"epoch": 0.04798551380715256,
"grad_norm": 2.25,
"learning_rate": 0.000156,
"loss": 0.5278,
"retrieval/pos_neg_gap_layer35": 0.5760174617171288,
"step": 53
},
{
"epoch": 0.0488909008601177,
"grad_norm": 2.84375,
"learning_rate": 0.000159,
"loss": 0.5298,
"retrieval/pos_neg_gap_layer35": 0.5720288977026939,
"step": 54
},
{
"epoch": 0.04979628791308284,
"grad_norm": 3.328125,
"learning_rate": 0.000162,
"loss": 0.4759,
"retrieval/pos_neg_gap_layer35": 0.6073333248496056,
"step": 55
},
{
"epoch": 0.05070167496604799,
"grad_norm": 2.09375,
"learning_rate": 0.000165,
"loss": 0.4881,
"retrieval/pos_neg_gap_layer35": 0.6045046746730804,
"step": 56
},
{
"epoch": 0.05160706201901313,
"grad_norm": 3.0,
"learning_rate": 0.000168,
"loss": 0.4922,
"retrieval/pos_neg_gap_layer35": 0.5916033461689949,
"step": 57
},
{
"epoch": 0.05251244907197827,
"grad_norm": 2.46875,
"learning_rate": 0.00017099999999999998,
"loss": 0.4203,
"retrieval/pos_neg_gap_layer35": 0.6445182487368584,
"step": 58
},
{
"epoch": 0.05341783612494341,
"grad_norm": 3.1875,
"learning_rate": 0.00017399999999999997,
"loss": 0.4356,
"retrieval/pos_neg_gap_layer35": 0.6368029490113258,
"step": 59
},
{
"epoch": 0.05432322317790855,
"grad_norm": 1.8828125,
"learning_rate": 0.00017699999999999997,
"loss": 0.418,
"retrieval/pos_neg_gap_layer35": 0.6428860053420067,
"step": 60
},
{
"epoch": 0.0552286102308737,
"grad_norm": 3.375,
"learning_rate": 0.00017999999999999998,
"loss": 0.4644,
"retrieval/pos_neg_gap_layer35": 0.6171751543879509,
"step": 61
},
{
"epoch": 0.05613399728383884,
"grad_norm": 2.359375,
"learning_rate": 0.00018299999999999998,
"loss": 0.4479,
"retrieval/pos_neg_gap_layer35": 0.6299770139157772,
"step": 62
},
{
"epoch": 0.057039384336803986,
"grad_norm": 1.953125,
"learning_rate": 0.000186,
"loss": 0.4361,
"retrieval/pos_neg_gap_layer35": 0.645499549806118,
"step": 63
},
{
"epoch": 0.057944771389769126,
"grad_norm": 1.9296875,
"learning_rate": 0.00018899999999999999,
"loss": 0.3707,
"retrieval/pos_neg_gap_layer35": 0.6785326227545738,
"step": 64
},
{
"epoch": 0.058850158442734266,
"grad_norm": 1.7421875,
"learning_rate": 0.00019199999999999998,
"loss": 0.441,
"retrieval/pos_neg_gap_layer35": 0.6410124003887177,
"step": 65
},
{
"epoch": 0.05975554549569941,
"grad_norm": 2.109375,
"learning_rate": 0.000195,
"loss": 0.3501,
"retrieval/pos_neg_gap_layer35": 0.6975989118218422,
"step": 66
},
{
"epoch": 0.06066093254866455,
"grad_norm": 2.671875,
"learning_rate": 0.000198,
"loss": 0.3914,
"retrieval/pos_neg_gap_layer35": 0.669355496764183,
"step": 67
},
{
"epoch": 0.0615663196016297,
"grad_norm": 2.609375,
"learning_rate": 0.000201,
"loss": 0.3747,
"retrieval/pos_neg_gap_layer35": 0.6800299286842346,
"step": 68
},
{
"epoch": 0.06247170665459484,
"grad_norm": 2.734375,
"learning_rate": 0.000204,
"loss": 0.4268,
"retrieval/pos_neg_gap_layer35": 0.652707114815712,
"step": 69
},
{
"epoch": 0.06337709370755998,
"grad_norm": 2.515625,
"learning_rate": 0.00020699999999999996,
"loss": 0.3494,
"retrieval/pos_neg_gap_layer35": 0.7074983268976212,
"step": 70
},
{
"epoch": 0.06428248076052512,
"grad_norm": 2.34375,
"learning_rate": 0.00020999999999999998,
"loss": 0.3832,
"retrieval/pos_neg_gap_layer35": 0.6818146035075188,
"step": 71
},
{
"epoch": 0.06518786781349027,
"grad_norm": 1.8359375,
"learning_rate": 0.00021299999999999997,
"loss": 0.3184,
"retrieval/pos_neg_gap_layer35": 0.7209471315145493,
"step": 72
},
{
"epoch": 0.06609325486645541,
"grad_norm": 3.515625,
"learning_rate": 0.00021599999999999996,
"loss": 0.4201,
"retrieval/pos_neg_gap_layer35": 0.6767289042472839,
"step": 73
},
{
"epoch": 0.06699864191942055,
"grad_norm": 2.3125,
"learning_rate": 0.00021899999999999998,
"loss": 0.3128,
"retrieval/pos_neg_gap_layer35": 0.7245900556445122,
"step": 74
},
{
"epoch": 0.06790402897238569,
"grad_norm": 2.140625,
"learning_rate": 0.00022199999999999998,
"loss": 0.3145,
"retrieval/pos_neg_gap_layer35": 0.7261751815676689,
"step": 75
},
{
"epoch": 0.06880941602535083,
"grad_norm": 1.734375,
"learning_rate": 0.000225,
"loss": 0.3366,
"retrieval/pos_neg_gap_layer35": 0.7103222459554672,
"step": 76
},
{
"epoch": 0.06971480307831598,
"grad_norm": 2.265625,
"learning_rate": 0.00022799999999999999,
"loss": 0.341,
"retrieval/pos_neg_gap_layer35": 0.70716293156147,
"step": 77
},
{
"epoch": 0.07062019013128112,
"grad_norm": 1.921875,
"learning_rate": 0.00023099999999999998,
"loss": 0.348,
"retrieval/pos_neg_gap_layer35": 0.7140689194202423,
"step": 78
},
{
"epoch": 0.07152557718424626,
"grad_norm": 2.0625,
"learning_rate": 0.000234,
"loss": 0.3664,
"retrieval/pos_neg_gap_layer35": 0.6952437534928322,
"step": 79
},
{
"epoch": 0.0724309642372114,
"grad_norm": 2.890625,
"learning_rate": 0.000237,
"loss": 0.3564,
"retrieval/pos_neg_gap_layer35": 0.7034005150198936,
"step": 80
},
{
"epoch": 0.07333635129017656,
"grad_norm": 4.40625,
"learning_rate": 0.00023999999999999998,
"loss": 0.4206,
"retrieval/pos_neg_gap_layer35": 0.6754265502095222,
"step": 81
},
{
"epoch": 0.0742417383431417,
"grad_norm": 2.25,
"learning_rate": 0.000243,
"loss": 0.3774,
"retrieval/pos_neg_gap_layer35": 0.6931728795170784,
"step": 82
},
{
"epoch": 0.07514712539610684,
"grad_norm": 2.453125,
"learning_rate": 0.00024599999999999996,
"loss": 0.362,
"retrieval/pos_neg_gap_layer35": 0.6995893642306328,
"step": 83
},
{
"epoch": 0.07605251244907198,
"grad_norm": 1.5703125,
"learning_rate": 0.000249,
"loss": 0.3262,
"retrieval/pos_neg_gap_layer35": 0.7296346575021744,
"step": 84
},
{
"epoch": 0.07695789950203712,
"grad_norm": 2.359375,
"learning_rate": 0.00025199999999999995,
"loss": 0.4007,
"retrieval/pos_neg_gap_layer35": 0.6771382838487625,
"step": 85
},
{
"epoch": 0.07786328655500227,
"grad_norm": 1.9609375,
"learning_rate": 0.00025499999999999996,
"loss": 0.3468,
"retrieval/pos_neg_gap_layer35": 0.7118361592292786,
"step": 86
},
{
"epoch": 0.07876867360796741,
"grad_norm": 1.7890625,
"learning_rate": 0.000258,
"loss": 0.3315,
"retrieval/pos_neg_gap_layer35": 0.7185137122869492,
"step": 87
},
{
"epoch": 0.07967406066093255,
"grad_norm": 1.5390625,
"learning_rate": 0.000261,
"loss": 0.3134,
"retrieval/pos_neg_gap_layer35": 0.7288002595305443,
"step": 88
},
{
"epoch": 0.08057944771389769,
"grad_norm": 3.140625,
"learning_rate": 0.00026399999999999997,
"loss": 0.3357,
"retrieval/pos_neg_gap_layer35": 0.7221949473023415,
"step": 89
},
{
"epoch": 0.08148483476686283,
"grad_norm": 2.3125,
"learning_rate": 0.000267,
"loss": 0.3382,
"retrieval/pos_neg_gap_layer35": 0.713361568748951,
"step": 90
},
{
"epoch": 0.08239022181982798,
"grad_norm": 2.703125,
"learning_rate": 0.00027,
"loss": 0.3731,
"retrieval/pos_neg_gap_layer35": 0.7051924467086792,
"step": 91
},
{
"epoch": 0.08329560887279312,
"grad_norm": 2.09375,
"learning_rate": 0.00027299999999999997,
"loss": 0.3871,
"retrieval/pos_neg_gap_layer35": 0.6929271668195724,
"step": 92
},
{
"epoch": 0.08420099592575826,
"grad_norm": 2.671875,
"learning_rate": 0.000276,
"loss": 0.2989,
"retrieval/pos_neg_gap_layer35": 0.7462679743766785,
"step": 93
},
{
"epoch": 0.0851063829787234,
"grad_norm": 1.671875,
"learning_rate": 0.000279,
"loss": 0.2864,
"retrieval/pos_neg_gap_layer35": 0.7504495307803154,
"step": 94
},
{
"epoch": 0.08601177003168854,
"grad_norm": 3.171875,
"learning_rate": 0.00028199999999999997,
"loss": 0.3076,
"retrieval/pos_neg_gap_layer35": 0.7340373545885086,
"step": 95
},
{
"epoch": 0.0869171570846537,
"grad_norm": 3.390625,
"learning_rate": 0.000285,
"loss": 0.3382,
"retrieval/pos_neg_gap_layer35": 0.7167180255055428,
"step": 96
},
{
"epoch": 0.08782254413761884,
"grad_norm": 2.8125,
"learning_rate": 0.00028799999999999995,
"loss": 0.3195,
"retrieval/pos_neg_gap_layer35": 0.731058344244957,
"step": 97
},
{
"epoch": 0.08872793119058398,
"grad_norm": 2.90625,
"learning_rate": 0.00029099999999999997,
"loss": 0.3641,
"retrieval/pos_neg_gap_layer35": 0.7082844749093056,
"step": 98
},
{
"epoch": 0.08963331824354911,
"grad_norm": 2.265625,
"learning_rate": 0.000294,
"loss": 0.318,
"retrieval/pos_neg_gap_layer35": 0.731484092772007,
"step": 99
},
{
"epoch": 0.09053870529651425,
"grad_norm": 2.25,
"learning_rate": 0.00029699999999999996,
"loss": 0.3589,
"retrieval/pos_neg_gap_layer35": 0.7069174870848656,
"step": 100
},
{
"epoch": 0.09144409234947941,
"grad_norm": 2.109375,
"learning_rate": 0.0003,
"loss": 0.3064,
"retrieval/pos_neg_gap_layer35": 0.741381324827671,
"step": 101
},
{
"epoch": 0.09234947940244455,
"grad_norm": 2.8125,
"learning_rate": 0.0002999999283858206,
"loss": 0.3495,
"retrieval/pos_neg_gap_layer35": 0.7186024561524391,
"step": 102
},
{
"epoch": 0.09325486645540969,
"grad_norm": 3.203125,
"learning_rate": 0.0002999997135433509,
"loss": 0.3728,
"retrieval/pos_neg_gap_layer35": 0.7052524164319038,
"step": 103
},
{
"epoch": 0.09416025350837483,
"grad_norm": 3.890625,
"learning_rate": 0.0002999993554727959,
"loss": 0.3731,
"retrieval/pos_neg_gap_layer35": 0.703025110065937,
"step": 104
},
{
"epoch": 0.09506564056133997,
"grad_norm": 3.078125,
"learning_rate": 0.0002999988541744976,
"loss": 0.3037,
"retrieval/pos_neg_gap_layer35": 0.7558679953217506,
"step": 105
},
{
"epoch": 0.09597102761430512,
"grad_norm": 1.96875,
"learning_rate": 0.0002999982096489347,
"loss": 0.3396,
"retrieval/pos_neg_gap_layer35": 0.7202946022152901,
"step": 106
},
{
"epoch": 0.09687641466727026,
"grad_norm": 2.15625,
"learning_rate": 0.00029999742189672256,
"loss": 0.3056,
"retrieval/pos_neg_gap_layer35": 0.7423298731446266,
"step": 107
},
{
"epoch": 0.0977818017202354,
"grad_norm": 2.0,
"learning_rate": 0.00029999649091861345,
"loss": 0.3485,
"retrieval/pos_neg_gap_layer35": 0.7212969064712524,
"step": 108
},
{
"epoch": 0.09868718877320054,
"grad_norm": 2.203125,
"learning_rate": 0.0002999954167154962,
"loss": 0.3144,
"retrieval/pos_neg_gap_layer35": 0.7376366779208183,
"step": 109
},
{
"epoch": 0.09959257582616568,
"grad_norm": 2.0,
"learning_rate": 0.0002999941992883966,
"loss": 0.3211,
"retrieval/pos_neg_gap_layer35": 0.7397533133625984,
"step": 110
},
{
"epoch": 0.10049796287913083,
"grad_norm": 1.75,
"learning_rate": 0.0002999928386384771,
"loss": 0.2829,
"retrieval/pos_neg_gap_layer35": 0.7592035382986069,
"step": 111
},
{
"epoch": 0.10140334993209597,
"grad_norm": 1.40625,
"learning_rate": 0.0002999913347670369,
"loss": 0.3325,
"retrieval/pos_neg_gap_layer35": 0.730476588010788,
"step": 112
},
{
"epoch": 0.10230873698506111,
"grad_norm": 1.6796875,
"learning_rate": 0.000299989687675512,
"loss": 0.3156,
"retrieval/pos_neg_gap_layer35": 0.7384232357144356,
"step": 113
},
{
"epoch": 0.10321412403802625,
"grad_norm": 1.6875,
"learning_rate": 0.00029998789736547515,
"loss": 0.3574,
"retrieval/pos_neg_gap_layer35": 0.7202979549765587,
"step": 114
},
{
"epoch": 0.10411951109099139,
"grad_norm": 1.453125,
"learning_rate": 0.00029998596383863587,
"loss": 0.2612,
"retrieval/pos_neg_gap_layer35": 0.7724613100290298,
"step": 115
},
{
"epoch": 0.10502489814395655,
"grad_norm": 1.7265625,
"learning_rate": 0.0002999838870968403,
"loss": 0.268,
"retrieval/pos_neg_gap_layer35": 0.7666561678051949,
"step": 116
},
{
"epoch": 0.10593028519692169,
"grad_norm": 1.84375,
"learning_rate": 0.0002999816671420715,
"loss": 0.3519,
"retrieval/pos_neg_gap_layer35": 0.7121048867702484,
"step": 117
},
{
"epoch": 0.10683567224988683,
"grad_norm": 2.015625,
"learning_rate": 0.0002999793039764492,
"loss": 0.3565,
"retrieval/pos_neg_gap_layer35": 0.7223715335130692,
"step": 118
},
{
"epoch": 0.10774105930285197,
"grad_norm": 1.65625,
"learning_rate": 0.0002999767976022298,
"loss": 0.3344,
"retrieval/pos_neg_gap_layer35": 0.7269682660698891,
"step": 119
},
{
"epoch": 0.1086464463558171,
"grad_norm": 1.609375,
"learning_rate": 0.00029997414802180667,
"loss": 0.342,
"retrieval/pos_neg_gap_layer35": 0.7178287133574486,
"step": 120
},
{
"epoch": 0.10955183340878226,
"grad_norm": 2.046875,
"learning_rate": 0.0002999713552377097,
"loss": 0.3301,
"retrieval/pos_neg_gap_layer35": 0.7330742478370667,
"step": 121
},
{
"epoch": 0.1104572204617474,
"grad_norm": 1.59375,
"learning_rate": 0.00029996841925260557,
"loss": 0.2606,
"retrieval/pos_neg_gap_layer35": 0.7765707522630692,
"step": 122
},
{
"epoch": 0.11136260751471254,
"grad_norm": 2.6875,
"learning_rate": 0.0002999653400692978,
"loss": 0.3701,
"retrieval/pos_neg_gap_layer35": 0.7033884972333908,
"step": 123
},
{
"epoch": 0.11226799456767768,
"grad_norm": 1.53125,
"learning_rate": 0.0002999621176907265,
"loss": 0.2848,
"retrieval/pos_neg_gap_layer35": 0.7586625516414642,
"step": 124
},
{
"epoch": 0.11317338162064282,
"grad_norm": 2.15625,
"learning_rate": 0.0002999587521199686,
"loss": 0.3447,
"retrieval/pos_neg_gap_layer35": 0.7313715294003487,
"step": 125
},
{
"epoch": 0.11407876867360797,
"grad_norm": 3.84375,
"learning_rate": 0.0002999552433602377,
"loss": 0.3343,
"retrieval/pos_neg_gap_layer35": 0.7272748127579689,
"step": 126
},
{
"epoch": 0.11498415572657311,
"grad_norm": 1.921875,
"learning_rate": 0.00029995159141488426,
"loss": 0.3711,
"retrieval/pos_neg_gap_layer35": 0.7181159630417824,
"step": 127
},
{
"epoch": 0.11588954277953825,
"grad_norm": 1.8203125,
"learning_rate": 0.00029994779628739524,
"loss": 0.4032,
"retrieval/pos_neg_gap_layer35": 0.6918591111898422,
"step": 128
},
{
"epoch": 0.11679492983250339,
"grad_norm": 1.6171875,
"learning_rate": 0.0002999438579813945,
"loss": 0.2771,
"retrieval/pos_neg_gap_layer35": 0.7596332877874374,
"step": 129
},
{
"epoch": 0.11770031688546853,
"grad_norm": 1.734375,
"learning_rate": 0.0002999397765006425,
"loss": 0.2919,
"retrieval/pos_neg_gap_layer35": 0.7467661499977112,
"step": 130
},
{
"epoch": 0.11860570393843368,
"grad_norm": 2.484375,
"learning_rate": 0.00029993555184903655,
"loss": 0.377,
"retrieval/pos_neg_gap_layer35": 0.7095595598220825,
"step": 131
},
{
"epoch": 0.11951109099139882,
"grad_norm": 1.8125,
"learning_rate": 0.00029993118403061055,
"loss": 0.293,
"retrieval/pos_neg_gap_layer35": 0.752009741961956,
"step": 132
},
{
"epoch": 0.12041647804436396,
"grad_norm": 2.46875,
"learning_rate": 0.00029992667304953505,
"loss": 0.335,
"retrieval/pos_neg_gap_layer35": 0.7303048372268677,
"step": 133
},
{
"epoch": 0.1213218650973291,
"grad_norm": 1.34375,
"learning_rate": 0.00029992201891011755,
"loss": 0.3348,
"retrieval/pos_neg_gap_layer35": 0.7341319844126701,
"step": 134
},
{
"epoch": 0.12222725215029426,
"grad_norm": 3.515625,
"learning_rate": 0.000299917221616802,
"loss": 0.3263,
"retrieval/pos_neg_gap_layer35": 0.7310595661401749,
"step": 135
},
{
"epoch": 0.1231326392032594,
"grad_norm": 1.4921875,
"learning_rate": 0.00029991228117416907,
"loss": 0.2915,
"retrieval/pos_neg_gap_layer35": 0.7516295239329338,
"step": 136
},
{
"epoch": 0.12403802625622454,
"grad_norm": 1.9453125,
"learning_rate": 0.00029990719758693626,
"loss": 0.3253,
"retrieval/pos_neg_gap_layer35": 0.7412228733301163,
"step": 137
},
{
"epoch": 0.12494341330918968,
"grad_norm": 1.5,
"learning_rate": 0.0002999019708599576,
"loss": 0.3009,
"retrieval/pos_neg_gap_layer35": 0.7491727843880653,
"step": 138
},
{
"epoch": 0.12584880036215482,
"grad_norm": 1.2734375,
"learning_rate": 0.00029989660099822386,
"loss": 0.2899,
"retrieval/pos_neg_gap_layer35": 0.7603149563074112,
"step": 139
},
{
"epoch": 0.12675418741511996,
"grad_norm": 2.921875,
"learning_rate": 0.00029989108800686253,
"loss": 0.3438,
"retrieval/pos_neg_gap_layer35": 0.7235552668571472,
"step": 140
},
{
"epoch": 0.1276595744680851,
"grad_norm": 1.5859375,
"learning_rate": 0.00029988543189113766,
"loss": 0.3198,
"retrieval/pos_neg_gap_layer35": 0.7430799305438995,
"step": 141
},
{
"epoch": 0.12856496152105024,
"grad_norm": 1.5390625,
"learning_rate": 0.00029987963265645014,
"loss": 0.3203,
"retrieval/pos_neg_gap_layer35": 0.7481404691934586,
"step": 142
},
{
"epoch": 0.1294703485740154,
"grad_norm": 2.359375,
"learning_rate": 0.0002998736903083372,
"loss": 0.3345,
"retrieval/pos_neg_gap_layer35": 0.734069436788559,
"step": 143
},
{
"epoch": 0.13037573562698054,
"grad_norm": 1.5078125,
"learning_rate": 0.00029986760485247317,
"loss": 0.3528,
"retrieval/pos_neg_gap_layer35": 0.7340828254818916,
"step": 144
},
{
"epoch": 0.13128112267994568,
"grad_norm": 1.9765625,
"learning_rate": 0.0002998613762946686,
"loss": 0.3316,
"retrieval/pos_neg_gap_layer35": 0.7414320856332779,
"step": 145
},
{
"epoch": 0.13218650973291082,
"grad_norm": 3.046875,
"learning_rate": 0.000299855004640871,
"loss": 0.3087,
"retrieval/pos_neg_gap_layer35": 0.7459285333752632,
"step": 146
},
{
"epoch": 0.13309189678587596,
"grad_norm": 1.359375,
"learning_rate": 0.0002998484898971642,
"loss": 0.2536,
"retrieval/pos_neg_gap_layer35": 0.7747001349925995,
"step": 147
},
{
"epoch": 0.1339972838388411,
"grad_norm": 2.0625,
"learning_rate": 0.000299841832069769,
"loss": 0.2895,
"retrieval/pos_neg_gap_layer35": 0.7498420625925064,
"step": 148
},
{
"epoch": 0.13490267089180624,
"grad_norm": 1.890625,
"learning_rate": 0.0002998350311650426,
"loss": 0.2719,
"retrieval/pos_neg_gap_layer35": 0.7645679637789726,
"step": 149
},
{
"epoch": 0.13580805794477138,
"grad_norm": 1.578125,
"learning_rate": 0.00029982808718947887,
"loss": 0.3503,
"retrieval/pos_neg_gap_layer35": 0.7216777876019478,
"step": 150
},
{
"epoch": 0.13671344499773652,
"grad_norm": 1.3515625,
"learning_rate": 0.0002998210001497083,
"loss": 0.3402,
"retrieval/pos_neg_gap_layer35": 0.7371743246912956,
"step": 151
},
{
"epoch": 0.13761883205070166,
"grad_norm": 1.203125,
"learning_rate": 0.0002998137700524981,
"loss": 0.3628,
"retrieval/pos_neg_gap_layer35": 0.7205822914838791,
"step": 152
},
{
"epoch": 0.13852421910366683,
"grad_norm": 2.53125,
"learning_rate": 0.00029980639690475186,
"loss": 0.3025,
"retrieval/pos_neg_gap_layer35": 0.7401210814714432,
"step": 153
},
{
"epoch": 0.13942960615663197,
"grad_norm": 1.453125,
"learning_rate": 0.00029979888071350987,
"loss": 0.2689,
"retrieval/pos_neg_gap_layer35": 0.7667993903160095,
"step": 154
},
{
"epoch": 0.1403349932095971,
"grad_norm": 1.171875,
"learning_rate": 0.00029979122148594904,
"loss": 0.2659,
"retrieval/pos_neg_gap_layer35": 0.7736154273152351,
"step": 155
},
{
"epoch": 0.14124038026256225,
"grad_norm": 1.1640625,
"learning_rate": 0.00029978341922938284,
"loss": 0.2522,
"retrieval/pos_neg_gap_layer35": 0.7741991057991982,
"step": 156
},
{
"epoch": 0.1421457673155274,
"grad_norm": 2.21875,
"learning_rate": 0.0002997754739512613,
"loss": 0.3138,
"retrieval/pos_neg_gap_layer35": 0.7516636252403259,
"step": 157
},
{
"epoch": 0.14305115436849253,
"grad_norm": 1.671875,
"learning_rate": 0.000299767385659171,
"loss": 0.2939,
"retrieval/pos_neg_gap_layer35": 0.7644387111067772,
"step": 158
},
{
"epoch": 0.14395654142145767,
"grad_norm": 2.609375,
"learning_rate": 0.000299759154360835,
"loss": 0.3501,
"retrieval/pos_neg_gap_layer35": 0.7274163588881493,
"step": 159
},
{
"epoch": 0.1448619284744228,
"grad_norm": 1.2890625,
"learning_rate": 0.00029975078006411323,
"loss": 0.318,
"retrieval/pos_neg_gap_layer35": 0.7465367317199707,
"step": 160
},
{
"epoch": 0.14576731552738795,
"grad_norm": 2.375,
"learning_rate": 0.00029974226277700174,
"loss": 0.3428,
"retrieval/pos_neg_gap_layer35": 0.7237431034445763,
"step": 161
},
{
"epoch": 0.14667270258035311,
"grad_norm": 2.234375,
"learning_rate": 0.00029973360250763336,
"loss": 0.2729,
"retrieval/pos_neg_gap_layer35": 0.7726488709449768,
"step": 162
},
{
"epoch": 0.14757808963331825,
"grad_norm": 1.1953125,
"learning_rate": 0.0002997247992642774,
"loss": 0.2619,
"retrieval/pos_neg_gap_layer35": 0.7700003609061241,
"step": 163
},
{
"epoch": 0.1484834766862834,
"grad_norm": 2.34375,
"learning_rate": 0.00029971585305533973,
"loss": 0.3559,
"retrieval/pos_neg_gap_layer35": 0.721222847700119,
"step": 164
},
{
"epoch": 0.14938886373924853,
"grad_norm": 1.4375,
"learning_rate": 0.00029970676388936263,
"loss": 0.2896,
"retrieval/pos_neg_gap_layer35": 0.7561898753046989,
"step": 165
},
{
"epoch": 0.15029425079221367,
"grad_norm": 1.7578125,
"learning_rate": 0.0002996975317750249,
"loss": 0.2904,
"retrieval/pos_neg_gap_layer35": 0.7605533972382545,
"step": 166
},
{
"epoch": 0.1511996378451788,
"grad_norm": 1.1328125,
"learning_rate": 0.0002996881567211421,
"loss": 0.2954,
"retrieval/pos_neg_gap_layer35": 0.7566156536340714,
"step": 167
},
{
"epoch": 0.15210502489814395,
"grad_norm": 1.4375,
"learning_rate": 0.00029967863873666577,
"loss": 0.2569,
"retrieval/pos_neg_gap_layer35": 0.7757222056388855,
"step": 168
},
{
"epoch": 0.1530104119511091,
"grad_norm": 1.6328125,
"learning_rate": 0.0002996689778306844,
"loss": 0.2713,
"retrieval/pos_neg_gap_layer35": 0.7690919637680054,
"step": 169
},
{
"epoch": 0.15391579900407423,
"grad_norm": 1.4453125,
"learning_rate": 0.0002996591740124226,
"loss": 0.2689,
"retrieval/pos_neg_gap_layer35": 0.765770711004734,
"step": 170
},
{
"epoch": 0.15482118605703937,
"grad_norm": 2.109375,
"learning_rate": 0.0002996492272912418,
"loss": 0.309,
"retrieval/pos_neg_gap_layer35": 0.7466119304299355,
"step": 171
},
{
"epoch": 0.15572657311000454,
"grad_norm": 1.2109375,
"learning_rate": 0.0002996391376766395,
"loss": 0.2442,
"retrieval/pos_neg_gap_layer35": 0.7834091782569885,
"step": 172
},
{
"epoch": 0.15663196016296968,
"grad_norm": 2.09375,
"learning_rate": 0.00029962890517825,
"loss": 0.3142,
"retrieval/pos_neg_gap_layer35": 0.754226379096508,
"step": 173
},
{
"epoch": 0.15753734721593482,
"grad_norm": 1.875,
"learning_rate": 0.00029961852980584376,
"loss": 0.2492,
"retrieval/pos_neg_gap_layer35": 0.7825815603137016,
"step": 174
},
{
"epoch": 0.15844273426889996,
"grad_norm": 1.5546875,
"learning_rate": 0.0002996080115693277,
"loss": 0.2439,
"retrieval/pos_neg_gap_layer35": 0.7873627096414566,
"step": 175
},
{
"epoch": 0.1593481213218651,
"grad_norm": 1.5234375,
"learning_rate": 0.00029959735047874533,
"loss": 0.3097,
"retrieval/pos_neg_gap_layer35": 0.7510440424084663,
"step": 176
},
{
"epoch": 0.16025350837483024,
"grad_norm": 1.6015625,
"learning_rate": 0.00029958654654427643,
"loss": 0.2605,
"retrieval/pos_neg_gap_layer35": 0.7733475565910339,
"step": 177
},
{
"epoch": 0.16115889542779538,
"grad_norm": 1.4140625,
"learning_rate": 0.0002995755997762372,
"loss": 0.2748,
"retrieval/pos_neg_gap_layer35": 0.7716283276677132,
"step": 178
},
{
"epoch": 0.16206428248076052,
"grad_norm": 1.2265625,
"learning_rate": 0.00029956451018508016,
"loss": 0.2592,
"retrieval/pos_neg_gap_layer35": 0.7734121307730675,
"step": 179
},
{
"epoch": 0.16296966953372566,
"grad_norm": 1.1015625,
"learning_rate": 0.0002995532777813944,
"loss": 0.2663,
"retrieval/pos_neg_gap_layer35": 0.7744711861014366,
"step": 180
},
{
"epoch": 0.1638750565866908,
"grad_norm": 1.640625,
"learning_rate": 0.0002995419025759051,
"loss": 0.2905,
"retrieval/pos_neg_gap_layer35": 0.7610870823264122,
"step": 181
},
{
"epoch": 0.16478044363965597,
"grad_norm": 2.96875,
"learning_rate": 0.000299530384579474,
"loss": 0.3651,
"retrieval/pos_neg_gap_layer35": 0.7282858788967133,
"step": 182
},
{
"epoch": 0.1656858306926211,
"grad_norm": 1.578125,
"learning_rate": 0.00029951872380309914,
"loss": 0.3109,
"retrieval/pos_neg_gap_layer35": 0.7556559443473816,
"step": 183
},
{
"epoch": 0.16659121774558625,
"grad_norm": 1.921875,
"learning_rate": 0.00029950692025791485,
"loss": 0.2876,
"retrieval/pos_neg_gap_layer35": 0.7645461037755013,
"step": 184
},
{
"epoch": 0.16749660479855139,
"grad_norm": 1.6328125,
"learning_rate": 0.00029949497395519186,
"loss": 0.279,
"retrieval/pos_neg_gap_layer35": 0.7647104114294052,
"step": 185
},
{
"epoch": 0.16840199185151652,
"grad_norm": 2.734375,
"learning_rate": 0.0002994828849063371,
"loss": 0.4012,
"retrieval/pos_neg_gap_layer35": 0.6940660253167152,
"step": 186
},
{
"epoch": 0.16930737890448166,
"grad_norm": 1.25,
"learning_rate": 0.00029947065312289396,
"loss": 0.2632,
"retrieval/pos_neg_gap_layer35": 0.7750058546662331,
"step": 187
},
{
"epoch": 0.1702127659574468,
"grad_norm": 1.3046875,
"learning_rate": 0.0002994582786165419,
"loss": 0.2553,
"retrieval/pos_neg_gap_layer35": 0.7848618552088737,
"step": 188
},
{
"epoch": 0.17111815301041194,
"grad_norm": 1.875,
"learning_rate": 0.00029944576139909695,
"loss": 0.3292,
"retrieval/pos_neg_gap_layer35": 0.7478398308157921,
"step": 189
},
{
"epoch": 0.17202354006337708,
"grad_norm": 2.3125,
"learning_rate": 0.00029943310148251113,
"loss": 0.3205,
"retrieval/pos_neg_gap_layer35": 0.7450714185833931,
"step": 190
},
{
"epoch": 0.17292892711634222,
"grad_norm": 1.78125,
"learning_rate": 0.0002994202988788728,
"loss": 0.3072,
"retrieval/pos_neg_gap_layer35": 0.7579996809363365,
"step": 191
},
{
"epoch": 0.1738343141693074,
"grad_norm": 1.8984375,
"learning_rate": 0.00029940735360040673,
"loss": 0.2775,
"retrieval/pos_neg_gap_layer35": 0.7690539658069611,
"step": 192
},
{
"epoch": 0.17473970122227253,
"grad_norm": 1.7421875,
"learning_rate": 0.00029939426565947367,
"loss": 0.2475,
"retrieval/pos_neg_gap_layer35": 0.7885555922985077,
"step": 193
},
{
"epoch": 0.17564508827523767,
"grad_norm": 1.5,
"learning_rate": 0.0002993810350685708,
"loss": 0.2589,
"retrieval/pos_neg_gap_layer35": 0.7795660942792892,
"step": 194
},
{
"epoch": 0.1765504753282028,
"grad_norm": 1.46875,
"learning_rate": 0.0002993676618403314,
"loss": 0.2491,
"retrieval/pos_neg_gap_layer35": 0.7819913625717163,
"step": 195
},
{
"epoch": 0.17745586238116795,
"grad_norm": 3.34375,
"learning_rate": 0.00029935414598752496,
"loss": 0.3505,
"retrieval/pos_neg_gap_layer35": 0.7359964326024055,
"step": 196
},
{
"epoch": 0.1783612494341331,
"grad_norm": 1.96875,
"learning_rate": 0.0002993404875230571,
"loss": 0.2568,
"retrieval/pos_neg_gap_layer35": 0.7808651998639107,
"step": 197
},
{
"epoch": 0.17926663648709823,
"grad_norm": 1.7421875,
"learning_rate": 0.0002993266864599699,
"loss": 0.2654,
"retrieval/pos_neg_gap_layer35": 0.7818052247166634,
"step": 198
},
{
"epoch": 0.18017202354006337,
"grad_norm": 1.484375,
"learning_rate": 0.00029931274281144114,
"loss": 0.2794,
"retrieval/pos_neg_gap_layer35": 0.7739764004945755,
"step": 199
},
{
"epoch": 0.1810774105930285,
"grad_norm": 1.5078125,
"learning_rate": 0.00029929865659078514,
"loss": 0.2814,
"retrieval/pos_neg_gap_layer35": 0.7717407643795013,
"step": 200
},
{
"epoch": 0.18198279764599365,
"grad_norm": 2.09375,
"learning_rate": 0.00029928442781145215,
"loss": 0.3094,
"retrieval/pos_neg_gap_layer35": 0.757408894598484,
"step": 201
},
{
"epoch": 0.18288818469895882,
"grad_norm": 1.4921875,
"learning_rate": 0.0002992700564870286,
"loss": 0.267,
"retrieval/pos_neg_gap_layer35": 0.7801089286804199,
"step": 202
},
{
"epoch": 0.18379357175192396,
"grad_norm": 2.359375,
"learning_rate": 0.0002992555426312371,
"loss": 0.3313,
"retrieval/pos_neg_gap_layer35": 0.7356933578848839,
"step": 203
},
{
"epoch": 0.1846989588048891,
"grad_norm": 1.4296875,
"learning_rate": 0.0002992408862579362,
"loss": 0.2941,
"retrieval/pos_neg_gap_layer35": 0.7602184936404228,
"step": 204
},
{
"epoch": 0.18560434585785424,
"grad_norm": 1.5,
"learning_rate": 0.00029922608738112066,
"loss": 0.2845,
"retrieval/pos_neg_gap_layer35": 0.763976477086544,
"step": 205
},
{
"epoch": 0.18650973291081938,
"grad_norm": 1.3984375,
"learning_rate": 0.0002992111460149213,
"loss": 0.2999,
"retrieval/pos_neg_gap_layer35": 0.7604339420795441,
"step": 206
},
{
"epoch": 0.18741511996378452,
"grad_norm": 0.94921875,
"learning_rate": 0.0002991960621736049,
"loss": 0.2297,
"retrieval/pos_neg_gap_layer35": 0.7965266406536102,
"step": 207
},
{
"epoch": 0.18832050701674966,
"grad_norm": 1.71875,
"learning_rate": 0.0002991808358715744,
"loss": 0.3421,
"retrieval/pos_neg_gap_layer35": 0.7317541837692261,
"step": 208
},
{
"epoch": 0.1892258940697148,
"grad_norm": 1.6015625,
"learning_rate": 0.0002991654671233687,
"loss": 0.2396,
"retrieval/pos_neg_gap_layer35": 0.7898482084274292,
"step": 209
},
{
"epoch": 0.19013128112267993,
"grad_norm": 1.2109375,
"learning_rate": 0.00029914995594366273,
"loss": 0.3041,
"retrieval/pos_neg_gap_layer35": 0.7563532441854477,
"step": 210
},
{
"epoch": 0.19103666817564507,
"grad_norm": 1.296875,
"learning_rate": 0.0002991343023472675,
"loss": 0.2419,
"retrieval/pos_neg_gap_layer35": 0.7919302806258202,
"step": 211
},
{
"epoch": 0.19194205522861024,
"grad_norm": 1.625,
"learning_rate": 0.0002991185063491299,
"loss": 0.2714,
"retrieval/pos_neg_gap_layer35": 0.7735611945390701,
"step": 212
},
{
"epoch": 0.19284744228157538,
"grad_norm": 1.9140625,
"learning_rate": 0.0002991025679643328,
"loss": 0.2752,
"retrieval/pos_neg_gap_layer35": 0.7702867761254311,
"step": 213
},
{
"epoch": 0.19375282933454052,
"grad_norm": 1.6796875,
"learning_rate": 0.00029908648720809505,
"loss": 0.2232,
"retrieval/pos_neg_gap_layer35": 0.8008521497249603,
"step": 214
},
{
"epoch": 0.19465821638750566,
"grad_norm": 1.7578125,
"learning_rate": 0.0002990702640957715,
"loss": 0.2765,
"retrieval/pos_neg_gap_layer35": 0.7676773741841316,
"step": 215
},
{
"epoch": 0.1955636034404708,
"grad_norm": 1.5,
"learning_rate": 0.0002990538986428528,
"loss": 0.3107,
"retrieval/pos_neg_gap_layer35": 0.7472456768155098,
"step": 216
},
{
"epoch": 0.19646899049343594,
"grad_norm": 1.3046875,
"learning_rate": 0.0002990373908649657,
"loss": 0.2513,
"retrieval/pos_neg_gap_layer35": 0.7812482789158821,
"step": 217
},
{
"epoch": 0.19737437754640108,
"grad_norm": 1.3125,
"learning_rate": 0.00029902074077787274,
"loss": 0.2572,
"retrieval/pos_neg_gap_layer35": 0.7806036919355392,
"step": 218
},
{
"epoch": 0.19827976459936622,
"grad_norm": 1.1953125,
"learning_rate": 0.00029900394839747226,
"loss": 0.2519,
"retrieval/pos_neg_gap_layer35": 0.7906462252140045,
"step": 219
},
{
"epoch": 0.19918515165233136,
"grad_norm": 1.6640625,
"learning_rate": 0.0002989870137397986,
"loss": 0.2997,
"retrieval/pos_neg_gap_layer35": 0.7555447071790695,
"step": 220
},
{
"epoch": 0.20009053870529653,
"grad_norm": 1.171875,
"learning_rate": 0.00029896993682102197,
"loss": 0.2489,
"retrieval/pos_neg_gap_layer35": 0.7838186696171761,
"step": 221
},
{
"epoch": 0.20099592575826167,
"grad_norm": 1.7578125,
"learning_rate": 0.00029895271765744823,
"loss": 0.2314,
"retrieval/pos_neg_gap_layer35": 0.7931787371635437,
"step": 222
},
{
"epoch": 0.2019013128112268,
"grad_norm": 2.484375,
"learning_rate": 0.0002989353562655193,
"loss": 0.3127,
"retrieval/pos_neg_gap_layer35": 0.7563988417387009,
"step": 223
},
{
"epoch": 0.20280669986419195,
"grad_norm": 1.578125,
"learning_rate": 0.0002989178526618129,
"loss": 0.3159,
"retrieval/pos_neg_gap_layer35": 0.7468860968947411,
"step": 224
},
{
"epoch": 0.2037120869171571,
"grad_norm": 2.34375,
"learning_rate": 0.00029890020686304227,
"loss": 0.2914,
"retrieval/pos_neg_gap_layer35": 0.7625986859202385,
"step": 225
},
{
"epoch": 0.20461747397012223,
"grad_norm": 1.390625,
"learning_rate": 0.0002988824188860567,
"loss": 0.2305,
"retrieval/pos_neg_gap_layer35": 0.798121802508831,
"step": 226
},
{
"epoch": 0.20552286102308737,
"grad_norm": 1.8203125,
"learning_rate": 0.00029886448874784106,
"loss": 0.3246,
"retrieval/pos_neg_gap_layer35": 0.7463467866182327,
"step": 227
},
{
"epoch": 0.2064282480760525,
"grad_norm": 1.625,
"learning_rate": 0.0002988464164655161,
"loss": 0.3211,
"retrieval/pos_neg_gap_layer35": 0.7552456259727478,
"step": 228
},
{
"epoch": 0.20733363512901765,
"grad_norm": 1.5546875,
"learning_rate": 0.0002988282020563383,
"loss": 0.3171,
"retrieval/pos_neg_gap_layer35": 0.7504011541604996,
"step": 229
},
{
"epoch": 0.20823902218198279,
"grad_norm": 1.1171875,
"learning_rate": 0.0002988098455376997,
"loss": 0.277,
"retrieval/pos_neg_gap_layer35": 0.7734598815441132,
"step": 230
},
{
"epoch": 0.20914440923494795,
"grad_norm": 1.6328125,
"learning_rate": 0.00029879134692712805,
"loss": 0.3382,
"retrieval/pos_neg_gap_layer35": 0.7357949614524841,
"step": 231
},
{
"epoch": 0.2100497962879131,
"grad_norm": 0.9921875,
"learning_rate": 0.0002987727062422871,
"loss": 0.2301,
"retrieval/pos_neg_gap_layer35": 0.7988517954945564,
"step": 232
},
{
"epoch": 0.21095518334087823,
"grad_norm": 1.4453125,
"learning_rate": 0.0002987539235009758,
"loss": 0.2901,
"retrieval/pos_neg_gap_layer35": 0.7612474113702774,
"step": 233
},
{
"epoch": 0.21186057039384337,
"grad_norm": 1.21875,
"learning_rate": 0.00029873499872112906,
"loss": 0.2738,
"retrieval/pos_neg_gap_layer35": 0.7799316123127937,
"step": 234
},
{
"epoch": 0.2127659574468085,
"grad_norm": 2.078125,
"learning_rate": 0.00029871593192081733,
"loss": 0.307,
"retrieval/pos_neg_gap_layer35": 0.7531334534287453,
"step": 235
},
{
"epoch": 0.21367134449977365,
"grad_norm": 1.578125,
"learning_rate": 0.0002986967231182466,
"loss": 0.2926,
"retrieval/pos_neg_gap_layer35": 0.7630181834101677,
"step": 236
},
{
"epoch": 0.2145767315527388,
"grad_norm": 1.4140625,
"learning_rate": 0.0002986773723317585,
"loss": 0.299,
"retrieval/pos_neg_gap_layer35": 0.7590935006737709,
"step": 237
},
{
"epoch": 0.21548211860570393,
"grad_norm": 2.171875,
"learning_rate": 0.0002986578795798302,
"loss": 0.307,
"retrieval/pos_neg_gap_layer35": 0.7521366402506828,
"step": 238
},
{
"epoch": 0.21638750565866907,
"grad_norm": 1.4375,
"learning_rate": 0.0002986382448810746,
"loss": 0.306,
"retrieval/pos_neg_gap_layer35": 0.757026419043541,
"step": 239
},
{
"epoch": 0.2172928927116342,
"grad_norm": 1.1171875,
"learning_rate": 0.0002986184682542399,
"loss": 0.2654,
"retrieval/pos_neg_gap_layer35": 0.7712252661585808,
"step": 240
},
{
"epoch": 0.21819827976459938,
"grad_norm": 1.6796875,
"learning_rate": 0.00029859854971820993,
"loss": 0.3425,
"retrieval/pos_neg_gap_layer35": 0.732631616294384,
"step": 241
},
{
"epoch": 0.21910366681756452,
"grad_norm": 1.4609375,
"learning_rate": 0.0002985784892920041,
"loss": 0.2983,
"retrieval/pos_neg_gap_layer35": 0.7568358108401299,
"step": 242
},
{
"epoch": 0.22000905387052966,
"grad_norm": 1.0625,
"learning_rate": 0.0002985582869947771,
"loss": 0.2625,
"retrieval/pos_neg_gap_layer35": 0.7773241177201271,
"step": 243
},
{
"epoch": 0.2209144409234948,
"grad_norm": 1.671875,
"learning_rate": 0.0002985379428458193,
"loss": 0.2911,
"retrieval/pos_neg_gap_layer35": 0.7591343745589256,
"step": 244
},
{
"epoch": 0.22181982797645994,
"grad_norm": 1.6484375,
"learning_rate": 0.0002985174568645564,
"loss": 0.2892,
"retrieval/pos_neg_gap_layer35": 0.7668918743729591,
"step": 245
},
{
"epoch": 0.22272521502942508,
"grad_norm": 1.1484375,
"learning_rate": 0.0002984968290705495,
"loss": 0.2648,
"retrieval/pos_neg_gap_layer35": 0.7803701534867287,
"step": 246
},
{
"epoch": 0.22363060208239022,
"grad_norm": 1.9375,
"learning_rate": 0.0002984760594834953,
"loss": 0.2676,
"retrieval/pos_neg_gap_layer35": 0.7729732096195221,
"step": 247
},
{
"epoch": 0.22453598913535536,
"grad_norm": 1.3671875,
"learning_rate": 0.00029845514812322564,
"loss": 0.2929,
"retrieval/pos_neg_gap_layer35": 0.7643093690276146,
"step": 248
},
{
"epoch": 0.2254413761883205,
"grad_norm": 0.84765625,
"learning_rate": 0.00029843409500970793,
"loss": 0.2234,
"retrieval/pos_neg_gap_layer35": 0.7997985556721687,
"step": 249
},
{
"epoch": 0.22634676324128564,
"grad_norm": 1.296875,
"learning_rate": 0.0002984129001630448,
"loss": 0.2601,
"retrieval/pos_neg_gap_layer35": 0.7787963002920151,
"step": 250
},
{
"epoch": 0.2272521502942508,
"grad_norm": 1.359375,
"learning_rate": 0.00029839156360347434,
"loss": 0.2659,
"retrieval/pos_neg_gap_layer35": 0.7742995023727417,
"step": 251
},
{
"epoch": 0.22815753734721594,
"grad_norm": 1.1875,
"learning_rate": 0.0002983700853513698,
"loss": 0.2485,
"retrieval/pos_neg_gap_layer35": 0.7829974815249443,
"step": 252
},
{
"epoch": 0.22906292440018108,
"grad_norm": 2.421875,
"learning_rate": 0.0002983484654272398,
"loss": 0.3415,
"retrieval/pos_neg_gap_layer35": 0.7404558211565018,
"step": 253
},
{
"epoch": 0.22996831145314622,
"grad_norm": 1.3984375,
"learning_rate": 0.0002983267038517284,
"loss": 0.2943,
"retrieval/pos_neg_gap_layer35": 0.763766422867775,
"step": 254
},
{
"epoch": 0.23087369850611136,
"grad_norm": 1.5703125,
"learning_rate": 0.0002983048006456145,
"loss": 0.2388,
"retrieval/pos_neg_gap_layer35": 0.7977386638522148,
"step": 255
},
{
"epoch": 0.2317790855590765,
"grad_norm": 1.09375,
"learning_rate": 0.00029828275582981283,
"loss": 0.2355,
"retrieval/pos_neg_gap_layer35": 0.7938683852553368,
"step": 256
},
{
"epoch": 0.23268447261204164,
"grad_norm": 2.046875,
"learning_rate": 0.00029826056942537275,
"loss": 0.2796,
"retrieval/pos_neg_gap_layer35": 0.7639481723308563,
"step": 257
},
{
"epoch": 0.23358985966500678,
"grad_norm": 3.34375,
"learning_rate": 0.0002982382414534792,
"loss": 0.3693,
"retrieval/pos_neg_gap_layer35": 0.7198377102613449,
"step": 258
},
{
"epoch": 0.23449524671797192,
"grad_norm": 1.0390625,
"learning_rate": 0.00029821577193545214,
"loss": 0.2555,
"retrieval/pos_neg_gap_layer35": 0.7830753549933434,
"step": 259
},
{
"epoch": 0.23540063377093706,
"grad_norm": 1.828125,
"learning_rate": 0.0002981931608927467,
"loss": 0.2957,
"retrieval/pos_neg_gap_layer35": 0.7566682025790215,
"step": 260
},
{
"epoch": 0.23630602082390223,
"grad_norm": 1.546875,
"learning_rate": 0.0002981704083469532,
"loss": 0.3236,
"retrieval/pos_neg_gap_layer35": 0.7584602981805801,
"step": 261
},
{
"epoch": 0.23721140787686737,
"grad_norm": 1.578125,
"learning_rate": 0.00029814751431979705,
"loss": 0.268,
"retrieval/pos_neg_gap_layer35": 0.776491217315197,
"step": 262
},
{
"epoch": 0.2381167949298325,
"grad_norm": 2.34375,
"learning_rate": 0.00029812447883313865,
"loss": 0.2986,
"retrieval/pos_neg_gap_layer35": 0.7511299550533295,
"step": 263
},
{
"epoch": 0.23902218198279765,
"grad_norm": 1.625,
"learning_rate": 0.0002981013019089737,
"loss": 0.3625,
"retrieval/pos_neg_gap_layer35": 0.7233055010437965,
"step": 264
},
{
"epoch": 0.2399275690357628,
"grad_norm": 1.40625,
"learning_rate": 0.00029807798356943266,
"loss": 0.2806,
"retrieval/pos_neg_gap_layer35": 0.7668991237878799,
"step": 265
},
{
"epoch": 0.24083295608872793,
"grad_norm": 1.7109375,
"learning_rate": 0.0002980545238367813,
"loss": 0.2668,
"retrieval/pos_neg_gap_layer35": 0.7790426909923553,
"step": 266
},
{
"epoch": 0.24173834314169307,
"grad_norm": 1.7734375,
"learning_rate": 0.0002980309227334203,
"loss": 0.2734,
"retrieval/pos_neg_gap_layer35": 0.7722537368535995,
"step": 267
},
{
"epoch": 0.2426437301946582,
"grad_norm": 1.6953125,
"learning_rate": 0.0002980071802818852,
"loss": 0.2418,
"retrieval/pos_neg_gap_layer35": 0.7910887822508812,
"step": 268
},
{
"epoch": 0.24354911724762335,
"grad_norm": 2.0625,
"learning_rate": 0.0002979832965048467,
"loss": 0.2593,
"retrieval/pos_neg_gap_layer35": 0.7870487570762634,
"step": 269
},
{
"epoch": 0.24445450430058852,
"grad_norm": 2.609375,
"learning_rate": 0.0002979592714251103,
"loss": 0.325,
"retrieval/pos_neg_gap_layer35": 0.7660333216190338,
"step": 270
},
{
"epoch": 0.24535989135355366,
"grad_norm": 1.0625,
"learning_rate": 0.0002979351050656165,
"loss": 0.247,
"retrieval/pos_neg_gap_layer35": 0.7936612889170647,
"step": 271
},
{
"epoch": 0.2462652784065188,
"grad_norm": 1.9453125,
"learning_rate": 0.00029791079744944076,
"loss": 0.2691,
"retrieval/pos_neg_gap_layer35": 0.7733396291732788,
"step": 272
},
{
"epoch": 0.24717066545948393,
"grad_norm": 0.95703125,
"learning_rate": 0.00029788634859979326,
"loss": 0.2561,
"retrieval/pos_neg_gap_layer35": 0.7847837656736374,
"step": 273
},
{
"epoch": 0.24807605251244907,
"grad_norm": 1.453125,
"learning_rate": 0.0002978617585400192,
"loss": 0.3273,
"retrieval/pos_neg_gap_layer35": 0.7622664049267769,
"step": 274
},
{
"epoch": 0.24898143956541421,
"grad_norm": 0.90625,
"learning_rate": 0.0002978370272935985,
"loss": 0.2791,
"retrieval/pos_neg_gap_layer35": 0.7823470085859299,
"step": 275
},
{
"epoch": 0.24988682661837935,
"grad_norm": 2.046875,
"learning_rate": 0.0002978121548841459,
"loss": 0.28,
"retrieval/pos_neg_gap_layer35": 0.770948201417923,
"step": 276
},
{
"epoch": 0.2507922136713445,
"grad_norm": 1.609375,
"learning_rate": 0.0002977871413354109,
"loss": 0.3371,
"retrieval/pos_neg_gap_layer35": 0.7531165331602097,
"step": 277
},
{
"epoch": 0.25169760072430963,
"grad_norm": 1.21875,
"learning_rate": 0.00029776198667127806,
"loss": 0.2419,
"retrieval/pos_neg_gap_layer35": 0.791239432990551,
"step": 278
},
{
"epoch": 0.2526029877772748,
"grad_norm": 1.421875,
"learning_rate": 0.0002977366909157663,
"loss": 0.278,
"retrieval/pos_neg_gap_layer35": 0.7707710117101669,
"step": 279
},
{
"epoch": 0.2535083748302399,
"grad_norm": 1.296875,
"learning_rate": 0.0002977112540930293,
"loss": 0.2766,
"retrieval/pos_neg_gap_layer35": 0.7738260924816132,
"step": 280
},
{
"epoch": 0.2544137618832051,
"grad_norm": 1.421875,
"learning_rate": 0.00029768567622735586,
"loss": 0.243,
"retrieval/pos_neg_gap_layer35": 0.7910718247294426,
"step": 281
},
{
"epoch": 0.2553191489361702,
"grad_norm": 2.21875,
"learning_rate": 0.00029765995734316885,
"loss": 0.3213,
"retrieval/pos_neg_gap_layer35": 0.7470906898379326,
"step": 282
},
{
"epoch": 0.25622453598913536,
"grad_norm": 1.0625,
"learning_rate": 0.0002976340974650263,
"loss": 0.2698,
"retrieval/pos_neg_gap_layer35": 0.7754645571112633,
"step": 283
},
{
"epoch": 0.25712992304210047,
"grad_norm": 1.8515625,
"learning_rate": 0.0002976080966176205,
"loss": 0.308,
"retrieval/pos_neg_gap_layer35": 0.7581334412097931,
"step": 284
},
{
"epoch": 0.25803531009506564,
"grad_norm": 0.76953125,
"learning_rate": 0.0002975819548257787,
"loss": 0.2156,
"retrieval/pos_neg_gap_layer35": 0.8084847703576088,
"step": 285
},
{
"epoch": 0.2589406971480308,
"grad_norm": 1.03125,
"learning_rate": 0.00029755567211446236,
"loss": 0.2417,
"retrieval/pos_neg_gap_layer35": 0.7908913791179657,
"step": 286
},
{
"epoch": 0.2598460842009959,
"grad_norm": 0.97265625,
"learning_rate": 0.0002975292485087678,
"loss": 0.2794,
"retrieval/pos_neg_gap_layer35": 0.7675927877426147,
"step": 287
},
{
"epoch": 0.2607514712539611,
"grad_norm": 1.90625,
"learning_rate": 0.00029750268403392564,
"loss": 0.3432,
"retrieval/pos_neg_gap_layer35": 0.7499352619051933,
"step": 288
},
{
"epoch": 0.2616568583069262,
"grad_norm": 1.2734375,
"learning_rate": 0.00029747597871530125,
"loss": 0.2911,
"retrieval/pos_neg_gap_layer35": 0.7630149275064468,
"step": 289
},
{
"epoch": 0.26256224535989137,
"grad_norm": 1.4453125,
"learning_rate": 0.00029744913257839426,
"loss": 0.2746,
"retrieval/pos_neg_gap_layer35": 0.7714230418205261,
"step": 290
},
{
"epoch": 0.2634676324128565,
"grad_norm": 2.25,
"learning_rate": 0.00029742214564883884,
"loss": 0.2839,
"retrieval/pos_neg_gap_layer35": 0.7616220414638519,
"step": 291
},
{
"epoch": 0.26437301946582165,
"grad_norm": 0.89453125,
"learning_rate": 0.0002973950179524037,
"loss": 0.2716,
"retrieval/pos_neg_gap_layer35": 0.7741082385182381,
"step": 292
},
{
"epoch": 0.26527840651878676,
"grad_norm": 1.671875,
"learning_rate": 0.00029736774951499194,
"loss": 0.3218,
"retrieval/pos_neg_gap_layer35": 0.7503506392240524,
"step": 293
},
{
"epoch": 0.2661837935717519,
"grad_norm": 1.5703125,
"learning_rate": 0.00029734034036264074,
"loss": 0.3035,
"retrieval/pos_neg_gap_layer35": 0.754734069108963,
"step": 294
},
{
"epoch": 0.2670891806247171,
"grad_norm": 1.3828125,
"learning_rate": 0.00029731279052152204,
"loss": 0.373,
"retrieval/pos_neg_gap_layer35": 0.7435363903641701,
"step": 295
},
{
"epoch": 0.2679945676776822,
"grad_norm": 1.328125,
"learning_rate": 0.000297285100017942,
"loss": 0.2771,
"retrieval/pos_neg_gap_layer35": 0.7681750729680061,
"step": 296
},
{
"epoch": 0.2688999547306474,
"grad_norm": 1.2578125,
"learning_rate": 0.00029725726887834097,
"loss": 0.2394,
"retrieval/pos_neg_gap_layer35": 0.790833093225956,
"step": 297
},
{
"epoch": 0.2698053417836125,
"grad_norm": 3.453125,
"learning_rate": 0.00029722929712929375,
"loss": 0.2768,
"retrieval/pos_neg_gap_layer35": 0.7736025080084801,
"step": 298
},
{
"epoch": 0.27071072883657765,
"grad_norm": 0.828125,
"learning_rate": 0.00029720118479750925,
"loss": 0.2566,
"retrieval/pos_neg_gap_layer35": 0.7836650535464287,
"step": 299
},
{
"epoch": 0.27161611588954276,
"grad_norm": 0.98046875,
"learning_rate": 0.0002971729319098307,
"loss": 0.2449,
"retrieval/pos_neg_gap_layer35": 0.7929629236459732,
"step": 300
},
{
"epoch": 0.27252150294250793,
"grad_norm": 1.140625,
"learning_rate": 0.0002971445384932356,
"loss": 0.2541,
"retrieval/pos_neg_gap_layer35": 0.7878819778561592,
"step": 301
},
{
"epoch": 0.27342688999547304,
"grad_norm": 1.0078125,
"learning_rate": 0.00029711600457483557,
"loss": 0.248,
"retrieval/pos_neg_gap_layer35": 0.7902123779058456,
"step": 302
},
{
"epoch": 0.2743322770484382,
"grad_norm": 1.0390625,
"learning_rate": 0.0002970873301818762,
"loss": 0.259,
"retrieval/pos_neg_gap_layer35": 0.7832691743969917,
"step": 303
},
{
"epoch": 0.2752376641014033,
"grad_norm": 0.91015625,
"learning_rate": 0.0002970585153417377,
"loss": 0.2523,
"retrieval/pos_neg_gap_layer35": 0.7938778847455978,
"step": 304
},
{
"epoch": 0.2761430511543685,
"grad_norm": 1.7109375,
"learning_rate": 0.0002970295600819338,
"loss": 0.2961,
"retrieval/pos_neg_gap_layer35": 0.76498993486166,
"step": 305
},
{
"epoch": 0.27704843820733366,
"grad_norm": 2.03125,
"learning_rate": 0.0002970004644301128,
"loss": 0.304,
"retrieval/pos_neg_gap_layer35": 0.7620171904563904,
"step": 306
},
{
"epoch": 0.27795382526029877,
"grad_norm": 1.3515625,
"learning_rate": 0.0002969712284140567,
"loss": 0.2494,
"retrieval/pos_neg_gap_layer35": 0.7837408632040024,
"step": 307
},
{
"epoch": 0.27885921231326394,
"grad_norm": 1.140625,
"learning_rate": 0.0002969418520616818,
"loss": 0.2673,
"retrieval/pos_neg_gap_layer35": 0.7841784879565239,
"step": 308
},
{
"epoch": 0.27976459936622905,
"grad_norm": 2.28125,
"learning_rate": 0.0002969123354010382,
"loss": 0.3388,
"retrieval/pos_neg_gap_layer35": 0.7401595488190651,
"step": 309
},
{
"epoch": 0.2806699864191942,
"grad_norm": 1.28125,
"learning_rate": 0.00029688267846031006,
"loss": 0.2796,
"retrieval/pos_neg_gap_layer35": 0.7716855257749557,
"step": 310
},
{
"epoch": 0.28157537347215933,
"grad_norm": 1.6484375,
"learning_rate": 0.00029685288126781546,
"loss": 0.348,
"retrieval/pos_neg_gap_layer35": 0.7355174273252487,
"step": 311
},
{
"epoch": 0.2824807605251245,
"grad_norm": 1.25,
"learning_rate": 0.0002968229438520065,
"loss": 0.3104,
"retrieval/pos_neg_gap_layer35": 0.7479116767644882,
"step": 312
},
{
"epoch": 0.2833861475780896,
"grad_norm": 1.2734375,
"learning_rate": 0.00029679286624146906,
"loss": 0.2566,
"retrieval/pos_neg_gap_layer35": 0.7934438735246658,
"step": 313
},
{
"epoch": 0.2842915346310548,
"grad_norm": 1.0703125,
"learning_rate": 0.00029676264846492293,
"loss": 0.2392,
"retrieval/pos_neg_gap_layer35": 0.7941480353474617,
"step": 314
},
{
"epoch": 0.28519692168401994,
"grad_norm": 2.390625,
"learning_rate": 0.0002967322905512216,
"loss": 0.2669,
"retrieval/pos_neg_gap_layer35": 0.7706975638866425,
"step": 315
},
{
"epoch": 0.28610230873698506,
"grad_norm": 2.59375,
"learning_rate": 0.0002967017925293527,
"loss": 0.3325,
"retrieval/pos_neg_gap_layer35": 0.7526589781045914,
"step": 316
},
{
"epoch": 0.2870076957899502,
"grad_norm": 2.078125,
"learning_rate": 0.00029667115442843725,
"loss": 0.294,
"retrieval/pos_neg_gap_layer35": 0.7613546326756477,
"step": 317
},
{
"epoch": 0.28791308284291534,
"grad_norm": 2.5625,
"learning_rate": 0.00029664037627773036,
"loss": 0.2855,
"retrieval/pos_neg_gap_layer35": 0.7656847089529037,
"step": 318
},
{
"epoch": 0.2888184698958805,
"grad_norm": 1.0625,
"learning_rate": 0.0002966094581066207,
"loss": 0.2689,
"retrieval/pos_neg_gap_layer35": 0.7779642567038536,
"step": 319
},
{
"epoch": 0.2897238569488456,
"grad_norm": 1.4375,
"learning_rate": 0.00029657839994463056,
"loss": 0.3193,
"retrieval/pos_neg_gap_layer35": 0.755174309015274,
"step": 320
},
{
"epoch": 0.2906292440018108,
"grad_norm": 1.4453125,
"learning_rate": 0.00029654720182141617,
"loss": 0.2426,
"retrieval/pos_neg_gap_layer35": 0.7940244078636169,
"step": 321
},
{
"epoch": 0.2915346310547759,
"grad_norm": 1.375,
"learning_rate": 0.00029651586376676706,
"loss": 0.2492,
"retrieval/pos_neg_gap_layer35": 0.787509560585022,
"step": 322
},
{
"epoch": 0.29244001810774106,
"grad_norm": 1.4140625,
"learning_rate": 0.0002964843858106067,
"loss": 0.2806,
"retrieval/pos_neg_gap_layer35": 0.7716903761029243,
"step": 323
},
{
"epoch": 0.29334540516070623,
"grad_norm": 1.0390625,
"learning_rate": 0.00029645276798299187,
"loss": 0.2255,
"retrieval/pos_neg_gap_layer35": 0.7995646595954895,
"step": 324
},
{
"epoch": 0.29425079221367134,
"grad_norm": 1.3046875,
"learning_rate": 0.0002964210103141131,
"loss": 0.2665,
"retrieval/pos_neg_gap_layer35": 0.7748811691999435,
"step": 325
},
{
"epoch": 0.2951561792666365,
"grad_norm": 1.453125,
"learning_rate": 0.0002963891128342944,
"loss": 0.2417,
"retrieval/pos_neg_gap_layer35": 0.7860556393861771,
"step": 326
},
{
"epoch": 0.2960615663196016,
"grad_norm": 1.2265625,
"learning_rate": 0.00029635707557399327,
"loss": 0.2572,
"retrieval/pos_neg_gap_layer35": 0.7816877588629723,
"step": 327
},
{
"epoch": 0.2969669533725668,
"grad_norm": 1.3203125,
"learning_rate": 0.0002963248985638006,
"loss": 0.2845,
"retrieval/pos_neg_gap_layer35": 0.7769095003604889,
"step": 328
},
{
"epoch": 0.2978723404255319,
"grad_norm": 1.6015625,
"learning_rate": 0.00029629258183444084,
"loss": 0.2731,
"retrieval/pos_neg_gap_layer35": 0.7785501778125763,
"step": 329
},
{
"epoch": 0.29877772747849707,
"grad_norm": 1.7265625,
"learning_rate": 0.0002962601254167718,
"loss": 0.3245,
"retrieval/pos_neg_gap_layer35": 0.7486740723252296,
"step": 330
},
{
"epoch": 0.2996831145314622,
"grad_norm": 1.796875,
"learning_rate": 0.00029622752934178463,
"loss": 0.29,
"retrieval/pos_neg_gap_layer35": 0.7675061076879501,
"step": 331
},
{
"epoch": 0.30058850158442735,
"grad_norm": 1.46875,
"learning_rate": 0.000296194793640604,
"loss": 0.2874,
"retrieval/pos_neg_gap_layer35": 0.7624929323792458,
"step": 332
},
{
"epoch": 0.30149388863739246,
"grad_norm": 1.1875,
"learning_rate": 0.00029616191834448767,
"loss": 0.3067,
"retrieval/pos_neg_gap_layer35": 0.756493404507637,
"step": 333
},
{
"epoch": 0.3023992756903576,
"grad_norm": 1.4921875,
"learning_rate": 0.0002961289034848268,
"loss": 0.3077,
"retrieval/pos_neg_gap_layer35": 0.7559342756867409,
"step": 334
},
{
"epoch": 0.3033046627433228,
"grad_norm": 1.1171875,
"learning_rate": 0.0002960957490931459,
"loss": 0.2625,
"retrieval/pos_neg_gap_layer35": 0.7778562232851982,
"step": 335
},
{
"epoch": 0.3042100497962879,
"grad_norm": 0.92578125,
"learning_rate": 0.00029606245520110255,
"loss": 0.2381,
"retrieval/pos_neg_gap_layer35": 0.7934528291225433,
"step": 336
},
{
"epoch": 0.3051154368492531,
"grad_norm": 1.6484375,
"learning_rate": 0.0002960290218404877,
"loss": 0.3123,
"retrieval/pos_neg_gap_layer35": 0.7531040459871292,
"step": 337
},
{
"epoch": 0.3060208239022182,
"grad_norm": 0.890625,
"learning_rate": 0.0002959954490432253,
"loss": 0.2346,
"retrieval/pos_neg_gap_layer35": 0.7940370664000511,
"step": 338
},
{
"epoch": 0.30692621095518335,
"grad_norm": 1.5625,
"learning_rate": 0.00029596173684137255,
"loss": 0.3081,
"retrieval/pos_neg_gap_layer35": 0.7578252330422401,
"step": 339
},
{
"epoch": 0.30783159800814847,
"grad_norm": 1.6640625,
"learning_rate": 0.00029592788526711973,
"loss": 0.3149,
"retrieval/pos_neg_gap_layer35": 0.7444384396076202,
"step": 340
},
{
"epoch": 0.30873698506111363,
"grad_norm": 1.6328125,
"learning_rate": 0.00029589389435279027,
"loss": 0.2309,
"retrieval/pos_neg_gap_layer35": 0.8017411008477211,
"step": 341
},
{
"epoch": 0.30964237211407875,
"grad_norm": 1.578125,
"learning_rate": 0.00029585976413084057,
"loss": 0.2768,
"retrieval/pos_neg_gap_layer35": 0.779523566365242,
"step": 342
},
{
"epoch": 0.3105477591670439,
"grad_norm": 1.40625,
"learning_rate": 0.00029582549463386,
"loss": 0.256,
"retrieval/pos_neg_gap_layer35": 0.7844132408499718,
"step": 343
},
{
"epoch": 0.3114531462200091,
"grad_norm": 1.8203125,
"learning_rate": 0.0002957910858945711,
"loss": 0.2694,
"retrieval/pos_neg_gap_layer35": 0.7762969955801964,
"step": 344
},
{
"epoch": 0.3123585332729742,
"grad_norm": 1.234375,
"learning_rate": 0.00029575653794582915,
"loss": 0.2371,
"retrieval/pos_neg_gap_layer35": 0.80193892121315,
"step": 345
},
{
"epoch": 0.31326392032593936,
"grad_norm": 0.96875,
"learning_rate": 0.0002957218508206225,
"loss": 0.2698,
"retrieval/pos_neg_gap_layer35": 0.7818038314580917,
"step": 346
},
{
"epoch": 0.31416930737890447,
"grad_norm": 0.89453125,
"learning_rate": 0.0002956870245520723,
"loss": 0.2716,
"retrieval/pos_neg_gap_layer35": 0.7843402400612831,
"step": 347
},
{
"epoch": 0.31507469443186964,
"grad_norm": 1.515625,
"learning_rate": 0.0002956520591734327,
"loss": 0.3372,
"retrieval/pos_neg_gap_layer35": 0.7558187395334244,
"step": 348
},
{
"epoch": 0.31598008148483475,
"grad_norm": 1.40625,
"learning_rate": 0.00029561695471809046,
"loss": 0.2882,
"retrieval/pos_neg_gap_layer35": 0.761152058839798,
"step": 349
},
{
"epoch": 0.3168854685377999,
"grad_norm": 1.1953125,
"learning_rate": 0.00029558171121956545,
"loss": 0.284,
"retrieval/pos_neg_gap_layer35": 0.7713920548558235,
"step": 350
},
{
"epoch": 0.31779085559076503,
"grad_norm": 1.1484375,
"learning_rate": 0.00029554632871151,
"loss": 0.2801,
"retrieval/pos_neg_gap_layer35": 0.7796169146895409,
"step": 351
},
{
"epoch": 0.3186962426437302,
"grad_norm": 1.4609375,
"learning_rate": 0.00029551080722770934,
"loss": 0.3007,
"retrieval/pos_neg_gap_layer35": 0.7574788630008698,
"step": 352
},
{
"epoch": 0.3196016296966953,
"grad_norm": 1.2421875,
"learning_rate": 0.00029547514680208135,
"loss": 0.2527,
"retrieval/pos_neg_gap_layer35": 0.7874152734875679,
"step": 353
},
{
"epoch": 0.3205070167496605,
"grad_norm": 1.1171875,
"learning_rate": 0.00029543934746867665,
"loss": 0.2431,
"retrieval/pos_neg_gap_layer35": 0.7894258424639702,
"step": 354
},
{
"epoch": 0.32141240380262565,
"grad_norm": 1.015625,
"learning_rate": 0.00029540340926167833,
"loss": 0.2783,
"retrieval/pos_neg_gap_layer35": 0.76955696195364,
"step": 355
},
{
"epoch": 0.32231779085559076,
"grad_norm": 1.2578125,
"learning_rate": 0.0002953673322154022,
"loss": 0.2971,
"retrieval/pos_neg_gap_layer35": 0.7695530131459236,
"step": 356
},
{
"epoch": 0.3232231779085559,
"grad_norm": 0.6796875,
"learning_rate": 0.00029533111636429675,
"loss": 0.2245,
"retrieval/pos_neg_gap_layer35": 0.8112246319651604,
"step": 357
},
{
"epoch": 0.32412856496152104,
"grad_norm": 1.265625,
"learning_rate": 0.0002952947617429428,
"loss": 0.2538,
"retrieval/pos_neg_gap_layer35": 0.792332299053669,
"step": 358
},
{
"epoch": 0.3250339520144862,
"grad_norm": 0.8984375,
"learning_rate": 0.0002952582683860538,
"loss": 0.2293,
"retrieval/pos_neg_gap_layer35": 0.8019958436489105,
"step": 359
},
{
"epoch": 0.3259393390674513,
"grad_norm": 1.609375,
"learning_rate": 0.0002952216363284756,
"loss": 0.2476,
"retrieval/pos_neg_gap_layer35": 0.7923199757933617,
"step": 360
},
{
"epoch": 0.3268447261204165,
"grad_norm": 1.6640625,
"learning_rate": 0.00029518486560518663,
"loss": 0.2674,
"retrieval/pos_neg_gap_layer35": 0.7762177586555481,
"step": 361
},
{
"epoch": 0.3277501131733816,
"grad_norm": 1.2578125,
"learning_rate": 0.0002951479562512975,
"loss": 0.2777,
"retrieval/pos_neg_gap_layer35": 0.780401736497879,
"step": 362
},
{
"epoch": 0.32865550022634676,
"grad_norm": 1.421875,
"learning_rate": 0.0002951109083020514,
"loss": 0.2857,
"retrieval/pos_neg_gap_layer35": 0.7718285620212555,
"step": 363
},
{
"epoch": 0.32956088727931193,
"grad_norm": 1.171875,
"learning_rate": 0.0002950737217928238,
"loss": 0.2554,
"retrieval/pos_neg_gap_layer35": 0.7930465415120125,
"step": 364
},
{
"epoch": 0.33046627433227704,
"grad_norm": 1.53125,
"learning_rate": 0.00029503639675912234,
"loss": 0.2679,
"retrieval/pos_neg_gap_layer35": 0.7791837900876999,
"step": 365
},
{
"epoch": 0.3313716613852422,
"grad_norm": 1.203125,
"learning_rate": 0.0002949989332365871,
"loss": 0.3068,
"retrieval/pos_neg_gap_layer35": 0.7515399605035782,
"step": 366
},
{
"epoch": 0.3322770484382073,
"grad_norm": 1.75,
"learning_rate": 0.00029496133126099037,
"loss": 0.3122,
"retrieval/pos_neg_gap_layer35": 0.751657597720623,
"step": 367
},
{
"epoch": 0.3331824354911725,
"grad_norm": 1.671875,
"learning_rate": 0.0002949235908682366,
"loss": 0.2502,
"retrieval/pos_neg_gap_layer35": 0.7890942916274071,
"step": 368
},
{
"epoch": 0.3340878225441376,
"grad_norm": 1.765625,
"learning_rate": 0.00029488571209436236,
"loss": 0.283,
"retrieval/pos_neg_gap_layer35": 0.7788895815610886,
"step": 369
},
{
"epoch": 0.33499320959710277,
"grad_norm": 1.703125,
"learning_rate": 0.00029484769497553653,
"loss": 0.2911,
"retrieval/pos_neg_gap_layer35": 0.7591556534171104,
"step": 370
},
{
"epoch": 0.3358985966500679,
"grad_norm": 1.5859375,
"learning_rate": 0.0002948095395480599,
"loss": 0.325,
"retrieval/pos_neg_gap_layer35": 0.7432632967829704,
"step": 371
},
{
"epoch": 0.33680398370303305,
"grad_norm": 1.140625,
"learning_rate": 0.00029477124584836536,
"loss": 0.2565,
"retrieval/pos_neg_gap_layer35": 0.7880156114697456,
"step": 372
},
{
"epoch": 0.33770937075599816,
"grad_norm": 1.46875,
"learning_rate": 0.0002947328139130179,
"loss": 0.2281,
"retrieval/pos_neg_gap_layer35": 0.8014998584985733,
"step": 373
},
{
"epoch": 0.33861475780896333,
"grad_norm": 1.1328125,
"learning_rate": 0.0002946942437787145,
"loss": 0.2665,
"retrieval/pos_neg_gap_layer35": 0.7816552892327309,
"step": 374
},
{
"epoch": 0.3395201448619285,
"grad_norm": 1.046875,
"learning_rate": 0.0002946555354822841,
"loss": 0.271,
"retrieval/pos_neg_gap_layer35": 0.7805389016866684,
"step": 375
},
{
"epoch": 0.3404255319148936,
"grad_norm": 1.296875,
"learning_rate": 0.00029461668906068746,
"loss": 0.3452,
"retrieval/pos_neg_gap_layer35": 0.7369760423898697,
"step": 376
},
{
"epoch": 0.3413309189678588,
"grad_norm": 1.0234375,
"learning_rate": 0.0002945777045510173,
"loss": 0.2546,
"retrieval/pos_neg_gap_layer35": 0.7848569974303246,
"step": 377
},
{
"epoch": 0.3422363060208239,
"grad_norm": 1.125,
"learning_rate": 0.00029453858199049833,
"loss": 0.2277,
"retrieval/pos_neg_gap_layer35": 0.8045710921287537,
"step": 378
},
{
"epoch": 0.34314169307378906,
"grad_norm": 0.796875,
"learning_rate": 0.0002944993214164869,
"loss": 0.2185,
"retrieval/pos_neg_gap_layer35": 0.8064287751913071,
"step": 379
},
{
"epoch": 0.34404708012675417,
"grad_norm": 1.53125,
"learning_rate": 0.0002944599228664711,
"loss": 0.2619,
"retrieval/pos_neg_gap_layer35": 0.7849665805697441,
"step": 380
},
{
"epoch": 0.34495246717971934,
"grad_norm": 1.6484375,
"learning_rate": 0.0002944203863780709,
"loss": 0.2724,
"retrieval/pos_neg_gap_layer35": 0.7698182314634323,
"step": 381
},
{
"epoch": 0.34585785423268445,
"grad_norm": 1.859375,
"learning_rate": 0.000294380711989038,
"loss": 0.3325,
"retrieval/pos_neg_gap_layer35": 0.736426867544651,
"step": 382
},
{
"epoch": 0.3467632412856496,
"grad_norm": 2.765625,
"learning_rate": 0.00029434089973725565,
"loss": 0.2317,
"retrieval/pos_neg_gap_layer35": 0.7975165322422981,
"step": 383
},
{
"epoch": 0.3476686283386148,
"grad_norm": 1.8671875,
"learning_rate": 0.0002943009496607389,
"loss": 0.2225,
"retrieval/pos_neg_gap_layer35": 0.8110938891768456,
"step": 384
},
{
"epoch": 0.3485740153915799,
"grad_norm": 1.3515625,
"learning_rate": 0.0002942608617976342,
"loss": 0.3072,
"retrieval/pos_neg_gap_layer35": 0.7532743141055107,
"step": 385
},
{
"epoch": 0.34947940244454506,
"grad_norm": 0.8828125,
"learning_rate": 0.00029422063618621974,
"loss": 0.2244,
"retrieval/pos_neg_gap_layer35": 0.8045077249407768,
"step": 386
},
{
"epoch": 0.3503847894975102,
"grad_norm": 1.9453125,
"learning_rate": 0.00029418027286490516,
"loss": 0.2291,
"retrieval/pos_neg_gap_layer35": 0.8010301887989044,
"step": 387
},
{
"epoch": 0.35129017655047534,
"grad_norm": 1.625,
"learning_rate": 0.0002941397718722316,
"loss": 0.3063,
"retrieval/pos_neg_gap_layer35": 0.7690878808498383,
"step": 388
},
{
"epoch": 0.35219556360344045,
"grad_norm": 1.1953125,
"learning_rate": 0.0002940991332468717,
"loss": 0.2803,
"retrieval/pos_neg_gap_layer35": 0.7694816514849663,
"step": 389
},
{
"epoch": 0.3531009506564056,
"grad_norm": 0.75,
"learning_rate": 0.0002940583570276294,
"loss": 0.2071,
"retrieval/pos_neg_gap_layer35": 0.8144046440720558,
"step": 390
},
{
"epoch": 0.35400633770937073,
"grad_norm": 1.7734375,
"learning_rate": 0.0002940174432534402,
"loss": 0.319,
"retrieval/pos_neg_gap_layer35": 0.7469370365142822,
"step": 391
},
{
"epoch": 0.3549117247623359,
"grad_norm": 0.734375,
"learning_rate": 0.0002939763919633708,
"loss": 0.2386,
"retrieval/pos_neg_gap_layer35": 0.8075713068246841,
"step": 392
},
{
"epoch": 0.35581711181530107,
"grad_norm": 1.15625,
"learning_rate": 0.00029393520319661925,
"loss": 0.2563,
"retrieval/pos_neg_gap_layer35": 0.7867434173822403,
"step": 393
},
{
"epoch": 0.3567224988682662,
"grad_norm": 0.86328125,
"learning_rate": 0.000293893876992515,
"loss": 0.2395,
"retrieval/pos_neg_gap_layer35": 0.7949795052409172,
"step": 394
},
{
"epoch": 0.35762788592123135,
"grad_norm": 1.296875,
"learning_rate": 0.0002938524133905184,
"loss": 0.247,
"retrieval/pos_neg_gap_layer35": 0.7915525734424591,
"step": 395
},
{
"epoch": 0.35853327297419646,
"grad_norm": 4.125,
"learning_rate": 0.00029381081243022144,
"loss": 0.2858,
"retrieval/pos_neg_gap_layer35": 0.765034444630146,
"step": 396
},
{
"epoch": 0.3594386600271616,
"grad_norm": 1.34375,
"learning_rate": 0.00029376907415134683,
"loss": 0.2347,
"retrieval/pos_neg_gap_layer35": 0.7994118854403496,
"step": 397
},
{
"epoch": 0.36034404708012674,
"grad_norm": 2.046875,
"learning_rate": 0.00029372719859374873,
"loss": 0.3178,
"retrieval/pos_neg_gap_layer35": 0.7565224170684814,
"step": 398
},
{
"epoch": 0.3612494341330919,
"grad_norm": 1.28125,
"learning_rate": 0.00029368518579741225,
"loss": 0.2571,
"retrieval/pos_neg_gap_layer35": 0.7909009158611298,
"step": 399
},
{
"epoch": 0.362154821186057,
"grad_norm": 1.1875,
"learning_rate": 0.00029364303580245346,
"loss": 0.2622,
"retrieval/pos_neg_gap_layer35": 0.7819407060742378,
"step": 400
},
{
"epoch": 0.3630602082390222,
"grad_norm": 1.203125,
"learning_rate": 0.0002936007486491196,
"loss": 0.2522,
"retrieval/pos_neg_gap_layer35": 0.7852613776922226,
"step": 401
},
{
"epoch": 0.3639655952919873,
"grad_norm": 1.46875,
"learning_rate": 0.0002935583243777888,
"loss": 0.2827,
"retrieval/pos_neg_gap_layer35": 0.770590141415596,
"step": 402
},
{
"epoch": 0.36487098234495247,
"grad_norm": 1.3671875,
"learning_rate": 0.0002935157630289701,
"loss": 0.2455,
"retrieval/pos_neg_gap_layer35": 0.7887911945581436,
"step": 403
},
{
"epoch": 0.36577636939791763,
"grad_norm": 0.828125,
"learning_rate": 0.00029347306464330345,
"loss": 0.2416,
"retrieval/pos_neg_gap_layer35": 0.8057910799980164,
"step": 404
},
{
"epoch": 0.36668175645088275,
"grad_norm": 0.94921875,
"learning_rate": 0.0002934302292615596,
"loss": 0.2224,
"retrieval/pos_neg_gap_layer35": 0.8071197345852852,
"step": 405
},
{
"epoch": 0.3675871435038479,
"grad_norm": 1.2734375,
"learning_rate": 0.0002933872569246403,
"loss": 0.2954,
"retrieval/pos_neg_gap_layer35": 0.7712145671248436,
"step": 406
},
{
"epoch": 0.368492530556813,
"grad_norm": 2.015625,
"learning_rate": 0.00029334414767357775,
"loss": 0.3268,
"retrieval/pos_neg_gap_layer35": 0.7557507753372192,
"step": 407
},
{
"epoch": 0.3693979176097782,
"grad_norm": 1.34375,
"learning_rate": 0.0002933009015495352,
"loss": 0.2319,
"retrieval/pos_neg_gap_layer35": 0.7998365238308907,
"step": 408
},
{
"epoch": 0.3703033046627433,
"grad_norm": 1.25,
"learning_rate": 0.0002932575185938064,
"loss": 0.2603,
"retrieval/pos_neg_gap_layer35": 0.7861762195825577,
"step": 409
},
{
"epoch": 0.37120869171570847,
"grad_norm": 2.1875,
"learning_rate": 0.00029321399884781583,
"loss": 0.3556,
"retrieval/pos_neg_gap_layer35": 0.732573077082634,
"step": 410
},
{
"epoch": 0.3721140787686736,
"grad_norm": 1.0859375,
"learning_rate": 0.0002931703423531186,
"loss": 0.2559,
"retrieval/pos_neg_gap_layer35": 0.7922574654221535,
"step": 411
},
{
"epoch": 0.37301946582163875,
"grad_norm": 1.3125,
"learning_rate": 0.00029312654915140024,
"loss": 0.3615,
"retrieval/pos_neg_gap_layer35": 0.7290648967027664,
"step": 412
},
{
"epoch": 0.3739248528746039,
"grad_norm": 1.0703125,
"learning_rate": 0.0002930826192844771,
"loss": 0.2928,
"retrieval/pos_neg_gap_layer35": 0.7701489627361298,
"step": 413
},
{
"epoch": 0.37483023992756903,
"grad_norm": 0.9921875,
"learning_rate": 0.0002930385527942958,
"loss": 0.2869,
"retrieval/pos_neg_gap_layer35": 0.7785881981253624,
"step": 414
},
{
"epoch": 0.3757356269805342,
"grad_norm": 0.8515625,
"learning_rate": 0.0002929943497229334,
"loss": 0.2222,
"retrieval/pos_neg_gap_layer35": 0.8094303458929062,
"step": 415
},
{
"epoch": 0.3766410140334993,
"grad_norm": 1.9765625,
"learning_rate": 0.0002929500101125976,
"loss": 0.3153,
"retrieval/pos_neg_gap_layer35": 0.7456752061843872,
"step": 416
},
{
"epoch": 0.3775464010864645,
"grad_norm": 1.6171875,
"learning_rate": 0.0002929055340056262,
"loss": 0.324,
"retrieval/pos_neg_gap_layer35": 0.75583166629076,
"step": 417
},
{
"epoch": 0.3784517881394296,
"grad_norm": 1.2734375,
"learning_rate": 0.0002928609214444876,
"loss": 0.2516,
"retrieval/pos_neg_gap_layer35": 0.7837498188018799,
"step": 418
},
{
"epoch": 0.37935717519239476,
"grad_norm": 0.68359375,
"learning_rate": 0.0002928161724717803,
"loss": 0.213,
"retrieval/pos_neg_gap_layer35": 0.8139834925532341,
"step": 419
},
{
"epoch": 0.38026256224535987,
"grad_norm": 1.4765625,
"learning_rate": 0.00029277128713023305,
"loss": 0.2999,
"retrieval/pos_neg_gap_layer35": 0.7576423212885857,
"step": 420
},
{
"epoch": 0.38116794929832504,
"grad_norm": 1.2265625,
"learning_rate": 0.0002927262654627049,
"loss": 0.2831,
"retrieval/pos_neg_gap_layer35": 0.7662830278277397,
"step": 421
},
{
"epoch": 0.38207333635129015,
"grad_norm": 1.4140625,
"learning_rate": 0.00029268110751218507,
"loss": 0.2719,
"retrieval/pos_neg_gap_layer35": 0.780327171087265,
"step": 422
},
{
"epoch": 0.3829787234042553,
"grad_norm": 1.2734375,
"learning_rate": 0.0002926358133217929,
"loss": 0.3331,
"retrieval/pos_neg_gap_layer35": 0.7449531629681587,
"step": 423
},
{
"epoch": 0.3838841104572205,
"grad_norm": 1.1328125,
"learning_rate": 0.0002925903829347778,
"loss": 0.2409,
"retrieval/pos_neg_gap_layer35": 0.800158716738224,
"step": 424
},
{
"epoch": 0.3847894975101856,
"grad_norm": 0.81640625,
"learning_rate": 0.00029254481639451926,
"loss": 0.2294,
"retrieval/pos_neg_gap_layer35": 0.8019485473632812,
"step": 425
},
{
"epoch": 0.38569488456315076,
"grad_norm": 0.890625,
"learning_rate": 0.0002924991137445267,
"loss": 0.2407,
"retrieval/pos_neg_gap_layer35": 0.7939884588122368,
"step": 426
},
{
"epoch": 0.3866002716161159,
"grad_norm": 1.828125,
"learning_rate": 0.00029245327502843956,
"loss": 0.267,
"retrieval/pos_neg_gap_layer35": 0.7782160714268684,
"step": 427
},
{
"epoch": 0.38750565866908104,
"grad_norm": 1.3125,
"learning_rate": 0.0002924073002900273,
"loss": 0.2351,
"retrieval/pos_neg_gap_layer35": 0.7982724979519844,
"step": 428
},
{
"epoch": 0.38841104572204616,
"grad_norm": 1.4921875,
"learning_rate": 0.00029236118957318895,
"loss": 0.2468,
"retrieval/pos_neg_gap_layer35": 0.7967758774757385,
"step": 429
},
{
"epoch": 0.3893164327750113,
"grad_norm": 1.0390625,
"learning_rate": 0.00029231494292195385,
"loss": 0.2484,
"retrieval/pos_neg_gap_layer35": 0.7943631261587143,
"step": 430
},
{
"epoch": 0.39022181982797643,
"grad_norm": 1.5703125,
"learning_rate": 0.00029226856038048067,
"loss": 0.3277,
"retrieval/pos_neg_gap_layer35": 0.7554822564125061,
"step": 431
},
{
"epoch": 0.3911272068809416,
"grad_norm": 1.421875,
"learning_rate": 0.00029222204199305814,
"loss": 0.2792,
"retrieval/pos_neg_gap_layer35": 0.772708535194397,
"step": 432
},
{
"epoch": 0.39203259393390677,
"grad_norm": 0.98046875,
"learning_rate": 0.0002921753878041046,
"loss": 0.2382,
"retrieval/pos_neg_gap_layer35": 0.797708049416542,
"step": 433
},
{
"epoch": 0.3929379809868719,
"grad_norm": 1.5390625,
"learning_rate": 0.0002921285978581681,
"loss": 0.2689,
"retrieval/pos_neg_gap_layer35": 0.779910109937191,
"step": 434
},
{
"epoch": 0.39384336803983705,
"grad_norm": 1.765625,
"learning_rate": 0.0002920816721999262,
"loss": 0.2883,
"retrieval/pos_neg_gap_layer35": 0.7671637684106827,
"step": 435
},
{
"epoch": 0.39474875509280216,
"grad_norm": 1.75,
"learning_rate": 0.00029203461087418624,
"loss": 0.2853,
"retrieval/pos_neg_gap_layer35": 0.7753506228327751,
"step": 436
},
{
"epoch": 0.39565414214576733,
"grad_norm": 1.6875,
"learning_rate": 0.00029198741392588495,
"loss": 0.2552,
"retrieval/pos_neg_gap_layer35": 0.785978227853775,
"step": 437
},
{
"epoch": 0.39655952919873244,
"grad_norm": 1.15625,
"learning_rate": 0.0002919400814000886,
"loss": 0.2499,
"retrieval/pos_neg_gap_layer35": 0.7916012182831764,
"step": 438
},
{
"epoch": 0.3974649162516976,
"grad_norm": 1.015625,
"learning_rate": 0.0002918926133419929,
"loss": 0.2264,
"retrieval/pos_neg_gap_layer35": 0.8069272264838219,
"step": 439
},
{
"epoch": 0.3983703033046627,
"grad_norm": 1.25,
"learning_rate": 0.000291845009796923,
"loss": 0.3025,
"retrieval/pos_neg_gap_layer35": 0.769988551735878,
"step": 440
},
{
"epoch": 0.3992756903576279,
"grad_norm": 1.609375,
"learning_rate": 0.0002917972708103334,
"loss": 0.3102,
"retrieval/pos_neg_gap_layer35": 0.7566511854529381,
"step": 441
},
{
"epoch": 0.40018107741059306,
"grad_norm": 1.0859375,
"learning_rate": 0.00029174939642780806,
"loss": 0.2676,
"retrieval/pos_neg_gap_layer35": 0.7832580506801605,
"step": 442
},
{
"epoch": 0.40108646446355817,
"grad_norm": 1.3359375,
"learning_rate": 0.00029170138669506,
"loss": 0.3198,
"retrieval/pos_neg_gap_layer35": 0.7615250423550606,
"step": 443
},
{
"epoch": 0.40199185151652334,
"grad_norm": 1.1875,
"learning_rate": 0.00029165324165793163,
"loss": 0.2657,
"retrieval/pos_neg_gap_layer35": 0.7852713689208031,
"step": 444
},
{
"epoch": 0.40289723856948845,
"grad_norm": 1.3828125,
"learning_rate": 0.0002916049613623945,
"loss": 0.3489,
"retrieval/pos_neg_gap_layer35": 0.7336863875389099,
"step": 445
},
{
"epoch": 0.4038026256224536,
"grad_norm": 0.90234375,
"learning_rate": 0.00029155654585454933,
"loss": 0.2639,
"retrieval/pos_neg_gap_layer35": 0.7801369875669479,
"step": 446
},
{
"epoch": 0.4047080126754187,
"grad_norm": 0.88671875,
"learning_rate": 0.000291507995180626,
"loss": 0.2289,
"retrieval/pos_neg_gap_layer35": 0.8003855273127556,
"step": 447
},
{
"epoch": 0.4056133997283839,
"grad_norm": 0.9453125,
"learning_rate": 0.00029145930938698325,
"loss": 0.2584,
"retrieval/pos_neg_gap_layer35": 0.7870207652449608,
"step": 448
},
{
"epoch": 0.406518786781349,
"grad_norm": 0.9765625,
"learning_rate": 0.0002914104885201091,
"loss": 0.2392,
"retrieval/pos_neg_gap_layer35": 0.795634850859642,
"step": 449
},
{
"epoch": 0.4074241738343142,
"grad_norm": 1.3359375,
"learning_rate": 0.00029136153262662045,
"loss": 0.2587,
"retrieval/pos_neg_gap_layer35": 0.7824196368455887,
"step": 450
},
{
"epoch": 0.4083295608872793,
"grad_norm": 1.2890625,
"learning_rate": 0.0002913124417532631,
"loss": 0.2386,
"retrieval/pos_neg_gap_layer35": 0.7909270823001862,
"step": 451
},
{
"epoch": 0.40923494794024445,
"grad_norm": 1.484375,
"learning_rate": 0.0002912632159469117,
"loss": 0.3079,
"retrieval/pos_neg_gap_layer35": 0.7537304386496544,
"step": 452
},
{
"epoch": 0.4101403349932096,
"grad_norm": 1.875,
"learning_rate": 0.0002912138552545698,
"loss": 0.3499,
"retrieval/pos_neg_gap_layer35": 0.7476329058408737,
"step": 453
},
{
"epoch": 0.41104572204617473,
"grad_norm": 1.8046875,
"learning_rate": 0.0002911643597233698,
"loss": 0.3524,
"retrieval/pos_neg_gap_layer35": 0.731356717646122,
"step": 454
},
{
"epoch": 0.4119511090991399,
"grad_norm": 1.1953125,
"learning_rate": 0.0002911147294005728,
"loss": 0.3276,
"retrieval/pos_neg_gap_layer35": 0.7598177939653397,
"step": 455
},
{
"epoch": 0.412856496152105,
"grad_norm": 1.3359375,
"learning_rate": 0.00029106496433356846,
"loss": 0.2244,
"retrieval/pos_neg_gap_layer35": 0.8059214875102043,
"step": 456
},
{
"epoch": 0.4137618832050702,
"grad_norm": 1.3671875,
"learning_rate": 0.0002910150645698754,
"loss": 0.2265,
"retrieval/pos_neg_gap_layer35": 0.8031329587101936,
"step": 457
},
{
"epoch": 0.4146672702580353,
"grad_norm": 1.359375,
"learning_rate": 0.00029096503015714064,
"loss": 0.3215,
"retrieval/pos_neg_gap_layer35": 0.7656263709068298,
"step": 458
},
{
"epoch": 0.41557265731100046,
"grad_norm": 1.4765625,
"learning_rate": 0.0002909148611431398,
"loss": 0.3223,
"retrieval/pos_neg_gap_layer35": 0.7442685440182686,
"step": 459
},
{
"epoch": 0.41647804436396557,
"grad_norm": 2.1875,
"learning_rate": 0.000290864557575777,
"loss": 0.3749,
"retrieval/pos_neg_gap_layer35": 0.7375131472945213,
"step": 460
},
{
"epoch": 0.41738343141693074,
"grad_norm": 1.40625,
"learning_rate": 0.000290814119503085,
"loss": 0.2118,
"retrieval/pos_neg_gap_layer35": 0.8139390721917152,
"step": 461
},
{
"epoch": 0.4182888184698959,
"grad_norm": 1.296875,
"learning_rate": 0.0002907635469732249,
"loss": 0.3003,
"retrieval/pos_neg_gap_layer35": 0.7680317088961601,
"step": 462
},
{
"epoch": 0.419194205522861,
"grad_norm": 1.3828125,
"learning_rate": 0.000290712840034486,
"loss": 0.2474,
"retrieval/pos_neg_gap_layer35": 0.795130118727684,
"step": 463
},
{
"epoch": 0.4200995925758262,
"grad_norm": 2.15625,
"learning_rate": 0.00029066199873528626,
"loss": 0.3366,
"retrieval/pos_neg_gap_layer35": 0.7404636740684509,
"step": 464
},
{
"epoch": 0.4210049796287913,
"grad_norm": 1.0859375,
"learning_rate": 0.00029061102312417177,
"loss": 0.2638,
"retrieval/pos_neg_gap_layer35": 0.7796059623360634,
"step": 465
},
{
"epoch": 0.42191036668175647,
"grad_norm": 1.671875,
"learning_rate": 0.00029055991324981683,
"loss": 0.2969,
"retrieval/pos_neg_gap_layer35": 0.7672151476144791,
"step": 466
},
{
"epoch": 0.4228157537347216,
"grad_norm": 1.15625,
"learning_rate": 0.00029050866916102404,
"loss": 0.3147,
"retrieval/pos_neg_gap_layer35": 0.7457406744360924,
"step": 467
},
{
"epoch": 0.42372114078768675,
"grad_norm": 1.1015625,
"learning_rate": 0.0002904572909067241,
"loss": 0.2406,
"retrieval/pos_neg_gap_layer35": 0.7960643768310547,
"step": 468
},
{
"epoch": 0.42462652784065186,
"grad_norm": 0.74609375,
"learning_rate": 0.0002904057785359758,
"loss": 0.2029,
"retrieval/pos_neg_gap_layer35": 0.8208494409918785,
"step": 469
},
{
"epoch": 0.425531914893617,
"grad_norm": 1.9921875,
"learning_rate": 0.0002903541320979661,
"loss": 0.3072,
"retrieval/pos_neg_gap_layer35": 0.7495509386062622,
"step": 470
},
{
"epoch": 0.42643730194658214,
"grad_norm": 1.109375,
"learning_rate": 0.00029030235164200977,
"loss": 0.2367,
"retrieval/pos_neg_gap_layer35": 0.7917436361312866,
"step": 471
},
{
"epoch": 0.4273426889995473,
"grad_norm": 2.015625,
"learning_rate": 0.0002902504372175498,
"loss": 0.3126,
"retrieval/pos_neg_gap_layer35": 0.7567478120326996,
"step": 472
},
{
"epoch": 0.42824807605251247,
"grad_norm": 1.15625,
"learning_rate": 0.0002901983888741569,
"loss": 0.2607,
"retrieval/pos_neg_gap_layer35": 0.787869893014431,
"step": 473
},
{
"epoch": 0.4291534631054776,
"grad_norm": 1.0078125,
"learning_rate": 0.00029014620666152976,
"loss": 0.282,
"retrieval/pos_neg_gap_layer35": 0.771853007376194,
"step": 474
},
{
"epoch": 0.43005885015844275,
"grad_norm": 1.609375,
"learning_rate": 0.00029009389062949485,
"loss": 0.3319,
"retrieval/pos_neg_gap_layer35": 0.7407456561923027,
"step": 475
},
{
"epoch": 0.43096423721140786,
"grad_norm": 1.1171875,
"learning_rate": 0.0002900414408280064,
"loss": 0.2627,
"retrieval/pos_neg_gap_layer35": 0.781107522547245,
"step": 476
},
{
"epoch": 0.43186962426437303,
"grad_norm": 0.96875,
"learning_rate": 0.0002899888573071465,
"loss": 0.2382,
"retrieval/pos_neg_gap_layer35": 0.7990758046507835,
"step": 477
},
{
"epoch": 0.43277501131733814,
"grad_norm": 1.0703125,
"learning_rate": 0.00028993614011712485,
"loss": 0.3148,
"retrieval/pos_neg_gap_layer35": 0.7507627010345459,
"step": 478
},
{
"epoch": 0.4336803983703033,
"grad_norm": 0.84765625,
"learning_rate": 0.00028988328930827856,
"loss": 0.2508,
"retrieval/pos_neg_gap_layer35": 0.7851749658584595,
"step": 479
},
{
"epoch": 0.4345857854232684,
"grad_norm": 1.28125,
"learning_rate": 0.00028983030493107267,
"loss": 0.2659,
"retrieval/pos_neg_gap_layer35": 0.7831493988633156,
"step": 480
},
{
"epoch": 0.4354911724762336,
"grad_norm": 0.78515625,
"learning_rate": 0.0002897771870360996,
"loss": 0.2557,
"retrieval/pos_neg_gap_layer35": 0.7896689549088478,
"step": 481
},
{
"epoch": 0.43639655952919876,
"grad_norm": 1.53125,
"learning_rate": 0.00028972393567407926,
"loss": 0.2889,
"retrieval/pos_neg_gap_layer35": 0.767035223543644,
"step": 482
},
{
"epoch": 0.43730194658216387,
"grad_norm": 2.046875,
"learning_rate": 0.00028967055089585905,
"loss": 0.357,
"retrieval/pos_neg_gap_layer35": 0.7265632897615433,
"step": 483
},
{
"epoch": 0.43820733363512904,
"grad_norm": 1.1875,
"learning_rate": 0.00028961703275241367,
"loss": 0.2645,
"retrieval/pos_neg_gap_layer35": 0.7778621688485146,
"step": 484
},
{
"epoch": 0.43911272068809415,
"grad_norm": 1.65625,
"learning_rate": 0.0002895633812948453,
"loss": 0.3168,
"retrieval/pos_neg_gap_layer35": 0.756355457007885,
"step": 485
},
{
"epoch": 0.4400181077410593,
"grad_norm": 1.46875,
"learning_rate": 0.0002895095965743833,
"loss": 0.3435,
"retrieval/pos_neg_gap_layer35": 0.7490367516875267,
"step": 486
},
{
"epoch": 0.44092349479402443,
"grad_norm": 1.15625,
"learning_rate": 0.0002894556786423843,
"loss": 0.2838,
"retrieval/pos_neg_gap_layer35": 0.7642712742090225,
"step": 487
},
{
"epoch": 0.4418288818469896,
"grad_norm": 1.84375,
"learning_rate": 0.00028940162755033223,
"loss": 0.2445,
"retrieval/pos_neg_gap_layer35": 0.7890613973140717,
"step": 488
},
{
"epoch": 0.4427342688999547,
"grad_norm": 1.078125,
"learning_rate": 0.000289347443349838,
"loss": 0.2814,
"retrieval/pos_neg_gap_layer35": 0.7762579545378685,
"step": 489
},
{
"epoch": 0.4436396559529199,
"grad_norm": 1.0078125,
"learning_rate": 0.0002892931260926397,
"loss": 0.2353,
"retrieval/pos_neg_gap_layer35": 0.80211241543293,
"step": 490
},
{
"epoch": 0.44454504300588504,
"grad_norm": 1.7265625,
"learning_rate": 0.0002892386758306025,
"loss": 0.2215,
"retrieval/pos_neg_gap_layer35": 0.8068276047706604,
"step": 491
},
{
"epoch": 0.44545043005885016,
"grad_norm": 1.3203125,
"learning_rate": 0.00028918409261571854,
"loss": 0.2283,
"retrieval/pos_neg_gap_layer35": 0.8011421412229538,
"step": 492
},
{
"epoch": 0.4463558171118153,
"grad_norm": 1.828125,
"learning_rate": 0.00028912937650010697,
"loss": 0.2619,
"retrieval/pos_neg_gap_layer35": 0.7819690853357315,
"step": 493
},
{
"epoch": 0.44726120416478043,
"grad_norm": 1.1015625,
"learning_rate": 0.0002890745275360137,
"loss": 0.2376,
"retrieval/pos_neg_gap_layer35": 0.7940984517335892,
"step": 494
},
{
"epoch": 0.4481665912177456,
"grad_norm": 1.6796875,
"learning_rate": 0.00028901954577581165,
"loss": 0.3004,
"retrieval/pos_neg_gap_layer35": 0.7598761767148972,
"step": 495
},
{
"epoch": 0.4490719782707107,
"grad_norm": 1.7265625,
"learning_rate": 0.0002889644312720004,
"loss": 0.2958,
"retrieval/pos_neg_gap_layer35": 0.7623439803719521,
"step": 496
},
{
"epoch": 0.4499773653236759,
"grad_norm": 0.96875,
"learning_rate": 0.0002889091840772064,
"loss": 0.2003,
"retrieval/pos_neg_gap_layer35": 0.8202961310744286,
"step": 497
},
{
"epoch": 0.450882752376641,
"grad_norm": 1.109375,
"learning_rate": 0.0002888538042441827,
"loss": 0.2868,
"retrieval/pos_neg_gap_layer35": 0.7704854235053062,
"step": 498
},
{
"epoch": 0.45178813942960616,
"grad_norm": 1.734375,
"learning_rate": 0.0002887982918258091,
"loss": 0.311,
"retrieval/pos_neg_gap_layer35": 0.7590202987194061,
"step": 499
},
{
"epoch": 0.4526935264825713,
"grad_norm": 1.1953125,
"learning_rate": 0.000288742646875092,
"loss": 0.2627,
"retrieval/pos_neg_gap_layer35": 0.7818822711706161,
"step": 500
},
{
"epoch": 0.45359891353553644,
"grad_norm": 1.3046875,
"learning_rate": 0.00028868686944516416,
"loss": 0.3382,
"retrieval/pos_neg_gap_layer35": 0.7484894320368767,
"step": 501
},
{
"epoch": 0.4545043005885016,
"grad_norm": 1.4140625,
"learning_rate": 0.00028863095958928505,
"loss": 0.2988,
"retrieval/pos_neg_gap_layer35": 0.7617760673165321,
"step": 502
},
{
"epoch": 0.4554096876414667,
"grad_norm": 1.09375,
"learning_rate": 0.00028857491736084055,
"loss": 0.2337,
"retrieval/pos_neg_gap_layer35": 0.8000829368829727,
"step": 503
},
{
"epoch": 0.4563150746944319,
"grad_norm": 1.0859375,
"learning_rate": 0.0002885187428133429,
"loss": 0.2779,
"retrieval/pos_neg_gap_layer35": 0.7754890769720078,
"step": 504
},
{
"epoch": 0.457220461747397,
"grad_norm": 1.0546875,
"learning_rate": 0.0002884624360004306,
"loss": 0.2412,
"retrieval/pos_neg_gap_layer35": 0.7956714108586311,
"step": 505
},
{
"epoch": 0.45812584880036217,
"grad_norm": 0.98046875,
"learning_rate": 0.0002884059969758686,
"loss": 0.296,
"retrieval/pos_neg_gap_layer35": 0.7719065546989441,
"step": 506
},
{
"epoch": 0.4590312358533273,
"grad_norm": 0.91796875,
"learning_rate": 0.00028834942579354807,
"loss": 0.2579,
"retrieval/pos_neg_gap_layer35": 0.7840260714292526,
"step": 507
},
{
"epoch": 0.45993662290629245,
"grad_norm": 1.109375,
"learning_rate": 0.00028829272250748624,
"loss": 0.2844,
"retrieval/pos_neg_gap_layer35": 0.7701221331954002,
"step": 508
},
{
"epoch": 0.46084200995925756,
"grad_norm": 0.875,
"learning_rate": 0.00028823588717182664,
"loss": 0.285,
"retrieval/pos_neg_gap_layer35": 0.7729863002896309,
"step": 509
},
{
"epoch": 0.4617473970122227,
"grad_norm": 0.87890625,
"learning_rate": 0.00028817891984083874,
"loss": 0.233,
"retrieval/pos_neg_gap_layer35": 0.8023329526185989,
"step": 510
},
{
"epoch": 0.4626527840651879,
"grad_norm": 1.09375,
"learning_rate": 0.00028812182056891817,
"loss": 0.3042,
"retrieval/pos_neg_gap_layer35": 0.7860079184174538,
"step": 511
},
{
"epoch": 0.463558171118153,
"grad_norm": 0.9375,
"learning_rate": 0.00028806458941058646,
"loss": 0.2073,
"retrieval/pos_neg_gap_layer35": 0.8159942030906677,
"step": 512
},
{
"epoch": 0.4644635581711182,
"grad_norm": 1.71875,
"learning_rate": 0.00028800722642049114,
"loss": 0.3803,
"retrieval/pos_neg_gap_layer35": 0.7316062971949577,
"step": 513
},
{
"epoch": 0.4653689452240833,
"grad_norm": 1.40625,
"learning_rate": 0.00028794973165340563,
"loss": 0.2411,
"retrieval/pos_neg_gap_layer35": 0.7968330532312393,
"step": 514
},
{
"epoch": 0.46627433227704845,
"grad_norm": 1.1015625,
"learning_rate": 0.0002878921051642291,
"loss": 0.2603,
"retrieval/pos_neg_gap_layer35": 0.7833465114235878,
"step": 515
},
{
"epoch": 0.46717971933001357,
"grad_norm": 1.34375,
"learning_rate": 0.00028783434700798647,
"loss": 0.2466,
"retrieval/pos_neg_gap_layer35": 0.7969139218330383,
"step": 516
},
{
"epoch": 0.46808510638297873,
"grad_norm": 1.234375,
"learning_rate": 0.0002877764572398286,
"loss": 0.3268,
"retrieval/pos_neg_gap_layer35": 0.7666233777999878,
"step": 517
},
{
"epoch": 0.46899049343594384,
"grad_norm": 1.265625,
"learning_rate": 0.0002877184359150317,
"loss": 0.252,
"retrieval/pos_neg_gap_layer35": 0.7953454852104187,
"step": 518
},
{
"epoch": 0.469895880488909,
"grad_norm": 1.453125,
"learning_rate": 0.0002876602830889978,
"loss": 0.2725,
"retrieval/pos_neg_gap_layer35": 0.7735436633229256,
"step": 519
},
{
"epoch": 0.4708012675418741,
"grad_norm": 1.7265625,
"learning_rate": 0.0002876019988172545,
"loss": 0.2827,
"retrieval/pos_neg_gap_layer35": 0.7730310633778572,
"step": 520
},
{
"epoch": 0.4717066545948393,
"grad_norm": 1.2265625,
"learning_rate": 0.0002875435831554549,
"loss": 0.309,
"retrieval/pos_neg_gap_layer35": 0.7545478045940399,
"step": 521
},
{
"epoch": 0.47261204164780446,
"grad_norm": 1.359375,
"learning_rate": 0.0002874850361593775,
"loss": 0.2702,
"retrieval/pos_neg_gap_layer35": 0.7812429666519165,
"step": 522
},
{
"epoch": 0.47351742870076957,
"grad_norm": 1.3515625,
"learning_rate": 0.00028742635788492614,
"loss": 0.3307,
"retrieval/pos_neg_gap_layer35": 0.7380700260400772,
"step": 523
},
{
"epoch": 0.47442281575373474,
"grad_norm": 1.1015625,
"learning_rate": 0.00028736754838813033,
"loss": 0.2654,
"retrieval/pos_neg_gap_layer35": 0.7768687754869461,
"step": 524
},
{
"epoch": 0.47532820280669985,
"grad_norm": 0.921875,
"learning_rate": 0.0002873086077251444,
"loss": 0.2411,
"retrieval/pos_neg_gap_layer35": 0.793905571103096,
"step": 525
},
{
"epoch": 0.476233589859665,
"grad_norm": 0.96875,
"learning_rate": 0.00028724953595224835,
"loss": 0.2537,
"retrieval/pos_neg_gap_layer35": 0.7902905493974686,
"step": 526
},
{
"epoch": 0.47713897691263013,
"grad_norm": 1.9140625,
"learning_rate": 0.00028719033312584716,
"loss": 0.2778,
"retrieval/pos_neg_gap_layer35": 0.7717692255973816,
"step": 527
},
{
"epoch": 0.4780443639655953,
"grad_norm": 0.96484375,
"learning_rate": 0.000287130999302471,
"loss": 0.2775,
"retrieval/pos_neg_gap_layer35": 0.7794568166136742,
"step": 528
},
{
"epoch": 0.4789497510185604,
"grad_norm": 0.88671875,
"learning_rate": 0.00028707153453877497,
"loss": 0.264,
"retrieval/pos_neg_gap_layer35": 0.7819542810320854,
"step": 529
},
{
"epoch": 0.4798551380715256,
"grad_norm": 1.5078125,
"learning_rate": 0.00028701193889153954,
"loss": 0.3302,
"retrieval/pos_neg_gap_layer35": 0.7466500848531723,
"step": 530
},
{
"epoch": 0.48076052512449075,
"grad_norm": 1.0859375,
"learning_rate": 0.00028695221241766987,
"loss": 0.2511,
"retrieval/pos_neg_gap_layer35": 0.7928766086697578,
"step": 531
},
{
"epoch": 0.48166591217745586,
"grad_norm": 0.9765625,
"learning_rate": 0.00028689235517419617,
"loss": 0.2791,
"retrieval/pos_neg_gap_layer35": 0.7742719799280167,
"step": 532
},
{
"epoch": 0.482571299230421,
"grad_norm": 1.0078125,
"learning_rate": 0.00028683236721827336,
"loss": 0.2367,
"retrieval/pos_neg_gap_layer35": 0.7971755042672157,
"step": 533
},
{
"epoch": 0.48347668628338614,
"grad_norm": 1.1328125,
"learning_rate": 0.00028677224860718136,
"loss": 0.2361,
"retrieval/pos_neg_gap_layer35": 0.7928821891546249,
"step": 534
},
{
"epoch": 0.4843820733363513,
"grad_norm": 1.9296875,
"learning_rate": 0.0002867119993983247,
"loss": 0.4113,
"retrieval/pos_neg_gap_layer35": 0.7236201986670494,
"step": 535
},
{
"epoch": 0.4852874603893164,
"grad_norm": 1.7890625,
"learning_rate": 0.0002866516196492329,
"loss": 0.2469,
"retrieval/pos_neg_gap_layer35": 0.7850403636693954,
"step": 536
},
{
"epoch": 0.4861928474422816,
"grad_norm": 1.2421875,
"learning_rate": 0.00028659110941755964,
"loss": 0.2956,
"retrieval/pos_neg_gap_layer35": 0.7769482210278511,
"step": 537
},
{
"epoch": 0.4870982344952467,
"grad_norm": 1.375,
"learning_rate": 0.0002865304687610836,
"loss": 0.2781,
"retrieval/pos_neg_gap_layer35": 0.7851505950093269,
"step": 538
},
{
"epoch": 0.48800362154821186,
"grad_norm": 1.1171875,
"learning_rate": 0.00028646969773770786,
"loss": 0.2781,
"retrieval/pos_neg_gap_layer35": 0.7753990516066551,
"step": 539
},
{
"epoch": 0.48890900860117703,
"grad_norm": 1.71875,
"learning_rate": 0.0002864087964054599,
"loss": 0.3523,
"retrieval/pos_neg_gap_layer35": 0.7379433214664459,
"step": 540
},
{
"epoch": 0.48981439565414214,
"grad_norm": 0.8984375,
"learning_rate": 0.0002863477648224918,
"loss": 0.2196,
"retrieval/pos_neg_gap_layer35": 0.8133002519607544,
"step": 541
},
{
"epoch": 0.4907197827071073,
"grad_norm": 0.796875,
"learning_rate": 0.00028628660304707985,
"loss": 0.204,
"retrieval/pos_neg_gap_layer35": 0.8220289722084999,
"step": 542
},
{
"epoch": 0.4916251697600724,
"grad_norm": 0.84765625,
"learning_rate": 0.0002862253111376247,
"loss": 0.2066,
"retrieval/pos_neg_gap_layer35": 0.8168248310685158,
"step": 543
},
{
"epoch": 0.4925305568130376,
"grad_norm": 1.4765625,
"learning_rate": 0.0002861638891526514,
"loss": 0.2586,
"retrieval/pos_neg_gap_layer35": 0.7817580029368401,
"step": 544
},
{
"epoch": 0.4934359438660027,
"grad_norm": 1.7421875,
"learning_rate": 0.000286102337150809,
"loss": 0.2922,
"retrieval/pos_neg_gap_layer35": 0.7679023966193199,
"step": 545
},
{
"epoch": 0.49434133091896787,
"grad_norm": 2.484375,
"learning_rate": 0.00028604065519087073,
"loss": 0.3238,
"retrieval/pos_neg_gap_layer35": 0.7503202557563782,
"step": 546
},
{
"epoch": 0.495246717971933,
"grad_norm": 1.1640625,
"learning_rate": 0.0002859788433317341,
"loss": 0.2504,
"retrieval/pos_neg_gap_layer35": 0.7931062504649162,
"step": 547
},
{
"epoch": 0.49615210502489815,
"grad_norm": 1.390625,
"learning_rate": 0.0002859169016324204,
"loss": 0.2638,
"retrieval/pos_neg_gap_layer35": 0.7866487726569176,
"step": 548
},
{
"epoch": 0.49705749207786326,
"grad_norm": 0.98046875,
"learning_rate": 0.00028585483015207506,
"loss": 0.2483,
"retrieval/pos_neg_gap_layer35": 0.7919879779219627,
"step": 549
},
{
"epoch": 0.49796287913082843,
"grad_norm": 1.46875,
"learning_rate": 0.0002857926289499674,
"loss": 0.2934,
"retrieval/pos_neg_gap_layer35": 0.7709784060716629,
"step": 550
},
{
"epoch": 0.4988682661837936,
"grad_norm": 1.6484375,
"learning_rate": 0.0002857302980854906,
"loss": 0.3493,
"retrieval/pos_neg_gap_layer35": 0.7413239553570747,
"step": 551
},
{
"epoch": 0.4997736532367587,
"grad_norm": 1.09375,
"learning_rate": 0.00028566783761816155,
"loss": 0.2198,
"retrieval/pos_neg_gap_layer35": 0.8130913749337196,
"step": 552
},
{
"epoch": 0.5006790402897239,
"grad_norm": 0.94921875,
"learning_rate": 0.0002856052476076211,
"loss": 0.2653,
"retrieval/pos_neg_gap_layer35": 0.781051442027092,
"step": 553
},
{
"epoch": 0.501584427342689,
"grad_norm": 1.015625,
"learning_rate": 0.0002855425281136336,
"loss": 0.248,
"retrieval/pos_neg_gap_layer35": 0.791755348443985,
"step": 554
},
{
"epoch": 0.5024898143956541,
"grad_norm": 1.4375,
"learning_rate": 0.0002854796791960872,
"loss": 0.204,
"retrieval/pos_neg_gap_layer35": 0.8181781023740768,
"step": 555
},
{
"epoch": 0.5033952014486193,
"grad_norm": 1.21875,
"learning_rate": 0.00028541670091499356,
"loss": 0.2138,
"retrieval/pos_neg_gap_layer35": 0.8068717867136002,
"step": 556
},
{
"epoch": 0.5043005885015844,
"grad_norm": 1.5703125,
"learning_rate": 0.0002853535933304877,
"loss": 0.3136,
"retrieval/pos_neg_gap_layer35": 0.7625513002276421,
"step": 557
},
{
"epoch": 0.5052059755545496,
"grad_norm": 0.87109375,
"learning_rate": 0.0002852903565028284,
"loss": 0.2395,
"retrieval/pos_neg_gap_layer35": 0.8047147840261459,
"step": 558
},
{
"epoch": 0.5061113626075147,
"grad_norm": 0.984375,
"learning_rate": 0.0002852269904923976,
"loss": 0.2512,
"retrieval/pos_neg_gap_layer35": 0.7887162417173386,
"step": 559
},
{
"epoch": 0.5070167496604798,
"grad_norm": 0.9765625,
"learning_rate": 0.0002851634953597008,
"loss": 0.2254,
"retrieval/pos_neg_gap_layer35": 0.8046447411179543,
"step": 560
},
{
"epoch": 0.507922136713445,
"grad_norm": 1.0234375,
"learning_rate": 0.00028509987116536665,
"loss": 0.2535,
"retrieval/pos_neg_gap_layer35": 0.7920905947685242,
"step": 561
},
{
"epoch": 0.5088275237664102,
"grad_norm": 1.3671875,
"learning_rate": 0.0002850361179701471,
"loss": 0.2326,
"retrieval/pos_neg_gap_layer35": 0.7981530502438545,
"step": 562
},
{
"epoch": 0.5097329108193753,
"grad_norm": 1.4140625,
"learning_rate": 0.0002849722358349172,
"loss": 0.2246,
"retrieval/pos_neg_gap_layer35": 0.8085191622376442,
"step": 563
},
{
"epoch": 0.5106382978723404,
"grad_norm": 1.6015625,
"learning_rate": 0.0002849082248206752,
"loss": 0.1975,
"retrieval/pos_neg_gap_layer35": 0.8257662206888199,
"step": 564
},
{
"epoch": 0.5115436849253056,
"grad_norm": 0.7890625,
"learning_rate": 0.00028484408498854243,
"loss": 0.2245,
"retrieval/pos_neg_gap_layer35": 0.8027700185775757,
"step": 565
},
{
"epoch": 0.5124490719782707,
"grad_norm": 1.21875,
"learning_rate": 0.0002847798163997631,
"loss": 0.2453,
"retrieval/pos_neg_gap_layer35": 0.790862962603569,
"step": 566
},
{
"epoch": 0.5133544590312359,
"grad_norm": 1.25,
"learning_rate": 0.00028471541911570444,
"loss": 0.2986,
"retrieval/pos_neg_gap_layer35": 0.7710316553711891,
"step": 567
},
{
"epoch": 0.5142598460842009,
"grad_norm": 0.67578125,
"learning_rate": 0.00028465089319785667,
"loss": 0.2117,
"retrieval/pos_neg_gap_layer35": 0.8109322339296341,
"step": 568
},
{
"epoch": 0.5151652331371661,
"grad_norm": 0.98046875,
"learning_rate": 0.0002845862387078326,
"loss": 0.2266,
"retrieval/pos_neg_gap_layer35": 0.805977076292038,
"step": 569
},
{
"epoch": 0.5160706201901313,
"grad_norm": 0.75390625,
"learning_rate": 0.000284521455707368,
"loss": 0.2191,
"retrieval/pos_neg_gap_layer35": 0.8096857890486717,
"step": 570
},
{
"epoch": 0.5169760072430964,
"grad_norm": 0.90625,
"learning_rate": 0.0002844565442583213,
"loss": 0.2361,
"retrieval/pos_neg_gap_layer35": 0.8015604168176651,
"step": 571
},
{
"epoch": 0.5178813942960616,
"grad_norm": 2.515625,
"learning_rate": 0.00028439150442267355,
"loss": 0.2641,
"retrieval/pos_neg_gap_layer35": 0.7798870578408241,
"step": 572
},
{
"epoch": 0.5187867813490267,
"grad_norm": 1.609375,
"learning_rate": 0.00028432633626252846,
"loss": 0.22,
"retrieval/pos_neg_gap_layer35": 0.8114399835467339,
"step": 573
},
{
"epoch": 0.5196921684019918,
"grad_norm": 0.94921875,
"learning_rate": 0.0002842610398401121,
"loss": 0.2316,
"retrieval/pos_neg_gap_layer35": 0.8069865331053734,
"step": 574
},
{
"epoch": 0.520597555454957,
"grad_norm": 1.25,
"learning_rate": 0.00028419561521777326,
"loss": 0.2783,
"retrieval/pos_neg_gap_layer35": 0.7835790663957596,
"step": 575
},
{
"epoch": 0.5215029425079222,
"grad_norm": 1.6015625,
"learning_rate": 0.00028413006245798296,
"loss": 0.2941,
"retrieval/pos_neg_gap_layer35": 0.7716887816786766,
"step": 576
},
{
"epoch": 0.5224083295608872,
"grad_norm": 0.77734375,
"learning_rate": 0.0002840643816233346,
"loss": 0.2495,
"retrieval/pos_neg_gap_layer35": 0.7945374175906181,
"step": 577
},
{
"epoch": 0.5233137166138524,
"grad_norm": 1.375,
"learning_rate": 0.000283998572776544,
"loss": 0.2432,
"retrieval/pos_neg_gap_layer35": 0.801528237760067,
"step": 578
},
{
"epoch": 0.5242191036668176,
"grad_norm": 1.1328125,
"learning_rate": 0.000283932635980449,
"loss": 0.2325,
"retrieval/pos_neg_gap_layer35": 0.8080378547310829,
"step": 579
},
{
"epoch": 0.5251244907197827,
"grad_norm": 0.96875,
"learning_rate": 0.00028386657129800974,
"loss": 0.2314,
"retrieval/pos_neg_gap_layer35": 0.8107351958751678,
"step": 580
},
{
"epoch": 0.5260298777727479,
"grad_norm": 1.0546875,
"learning_rate": 0.00028380037879230857,
"loss": 0.2501,
"retrieval/pos_neg_gap_layer35": 0.794516958296299,
"step": 581
},
{
"epoch": 0.526935264825713,
"grad_norm": 1.625,
"learning_rate": 0.00028373405852654966,
"loss": 0.247,
"retrieval/pos_neg_gap_layer35": 0.8054124191403389,
"step": 582
},
{
"epoch": 0.5278406518786781,
"grad_norm": 1.140625,
"learning_rate": 0.00028366761056405936,
"loss": 0.2725,
"retrieval/pos_neg_gap_layer35": 0.7909405082464218,
"step": 583
},
{
"epoch": 0.5287460389316433,
"grad_norm": 1.2890625,
"learning_rate": 0.00028360103496828583,
"loss": 0.2329,
"retrieval/pos_neg_gap_layer35": 0.8033862411975861,
"step": 584
},
{
"epoch": 0.5296514259846085,
"grad_norm": 1.3125,
"learning_rate": 0.00028353433180279927,
"loss": 0.2672,
"retrieval/pos_neg_gap_layer35": 0.7800335958600044,
"step": 585
},
{
"epoch": 0.5305568130375735,
"grad_norm": 1.484375,
"learning_rate": 0.0002834675011312915,
"loss": 0.32,
"retrieval/pos_neg_gap_layer35": 0.752836748957634,
"step": 586
},
{
"epoch": 0.5314622000905387,
"grad_norm": 1.1484375,
"learning_rate": 0.0002834005430175761,
"loss": 0.2713,
"retrieval/pos_neg_gap_layer35": 0.7938762605190277,
"step": 587
},
{
"epoch": 0.5323675871435039,
"grad_norm": 1.546875,
"learning_rate": 0.0002833334575255886,
"loss": 0.285,
"retrieval/pos_neg_gap_layer35": 0.7780323550105095,
"step": 588
},
{
"epoch": 0.533272974196469,
"grad_norm": 0.9609375,
"learning_rate": 0.00028326624471938577,
"loss": 0.2915,
"retrieval/pos_neg_gap_layer35": 0.7747867852449417,
"step": 589
},
{
"epoch": 0.5341783612494342,
"grad_norm": 0.92578125,
"learning_rate": 0.00028319890466314623,
"loss": 0.2421,
"retrieval/pos_neg_gap_layer35": 0.7945190295577049,
"step": 590
},
{
"epoch": 0.5350837483023992,
"grad_norm": 1.1953125,
"learning_rate": 0.0002831314374211701,
"loss": 0.3317,
"retrieval/pos_neg_gap_layer35": 0.7551342695951462,
"step": 591
},
{
"epoch": 0.5359891353553644,
"grad_norm": 0.94140625,
"learning_rate": 0.0002830638430578786,
"loss": 0.271,
"retrieval/pos_neg_gap_layer35": 0.7866533771157265,
"step": 592
},
{
"epoch": 0.5368945224083296,
"grad_norm": 1.3203125,
"learning_rate": 0.00028299612163781493,
"loss": 0.25,
"retrieval/pos_neg_gap_layer35": 0.7868204638361931,
"step": 593
},
{
"epoch": 0.5377999094612947,
"grad_norm": 1.0546875,
"learning_rate": 0.00028292827322564306,
"loss": 0.2475,
"retrieval/pos_neg_gap_layer35": 0.7897237464785576,
"step": 594
},
{
"epoch": 0.5387052965142598,
"grad_norm": 1.53125,
"learning_rate": 0.0002828602978861485,
"loss": 0.3064,
"retrieval/pos_neg_gap_layer35": 0.7602527812123299,
"step": 595
},
{
"epoch": 0.539610683567225,
"grad_norm": 1.0078125,
"learning_rate": 0.0002827921956842379,
"loss": 0.2642,
"retrieval/pos_neg_gap_layer35": 0.793736033141613,
"step": 596
},
{
"epoch": 0.5405160706201901,
"grad_norm": 1.65625,
"learning_rate": 0.000282723966684939,
"loss": 0.2982,
"retrieval/pos_neg_gap_layer35": 0.7552542015910149,
"step": 597
},
{
"epoch": 0.5414214576731553,
"grad_norm": 0.7421875,
"learning_rate": 0.00028265561095340065,
"loss": 0.2442,
"retrieval/pos_neg_gap_layer35": 0.7980299592018127,
"step": 598
},
{
"epoch": 0.5423268447261205,
"grad_norm": 0.70703125,
"learning_rate": 0.00028258712855489275,
"loss": 0.27,
"retrieval/pos_neg_gap_layer35": 0.7755333557724953,
"step": 599
},
{
"epoch": 0.5432322317790855,
"grad_norm": 0.83203125,
"learning_rate": 0.00028251851955480613,
"loss": 0.3533,
"retrieval/pos_neg_gap_layer35": 0.7247518301010132,
"step": 600
},
{
"epoch": 0.5441376188320507,
"grad_norm": 1.375,
"learning_rate": 0.0002824497840186524,
"loss": 0.2632,
"retrieval/pos_neg_gap_layer35": 0.7812588214874268,
"step": 601
},
{
"epoch": 0.5450430058850159,
"grad_norm": 0.90625,
"learning_rate": 0.0002823809220120641,
"loss": 0.269,
"retrieval/pos_neg_gap_layer35": 0.7786519750952721,
"step": 602
},
{
"epoch": 0.545948392937981,
"grad_norm": 1.671875,
"learning_rate": 0.00028231193360079464,
"loss": 0.2856,
"retrieval/pos_neg_gap_layer35": 0.773596316576004,
"step": 603
},
{
"epoch": 0.5468537799909461,
"grad_norm": 0.92578125,
"learning_rate": 0.00028224281885071785,
"loss": 0.2251,
"retrieval/pos_neg_gap_layer35": 0.8041056841611862,
"step": 604
},
{
"epoch": 0.5477591670439113,
"grad_norm": 1.046875,
"learning_rate": 0.0002821735778278284,
"loss": 0.2793,
"retrieval/pos_neg_gap_layer35": 0.782682366669178,
"step": 605
},
{
"epoch": 0.5486645540968764,
"grad_norm": 0.95703125,
"learning_rate": 0.00028210421059824146,
"loss": 0.2231,
"retrieval/pos_neg_gap_layer35": 0.8023364916443825,
"step": 606
},
{
"epoch": 0.5495699411498416,
"grad_norm": 1.0078125,
"learning_rate": 0.00028203471722819276,
"loss": 0.2528,
"retrieval/pos_neg_gap_layer35": 0.7923801466822624,
"step": 607
},
{
"epoch": 0.5504753282028066,
"grad_norm": 1.2578125,
"learning_rate": 0.0002819650977840384,
"loss": 0.3098,
"retrieval/pos_neg_gap_layer35": 0.7563069015741348,
"step": 608
},
{
"epoch": 0.5513807152557718,
"grad_norm": 1.21875,
"learning_rate": 0.0002818953523322549,
"loss": 0.27,
"retrieval/pos_neg_gap_layer35": 0.7792688980698586,
"step": 609
},
{
"epoch": 0.552286102308737,
"grad_norm": 1.53125,
"learning_rate": 0.00028182548093943913,
"loss": 0.278,
"retrieval/pos_neg_gap_layer35": 0.77851802110672,
"step": 610
},
{
"epoch": 0.5531914893617021,
"grad_norm": 1.109375,
"learning_rate": 0.0002817554836723083,
"loss": 0.2605,
"retrieval/pos_neg_gap_layer35": 0.7843850031495094,
"step": 611
},
{
"epoch": 0.5540968764146673,
"grad_norm": 1.015625,
"learning_rate": 0.0002816853605976995,
"loss": 0.2254,
"retrieval/pos_neg_gap_layer35": 0.8049393445253372,
"step": 612
},
{
"epoch": 0.5550022634676324,
"grad_norm": 1.3515625,
"learning_rate": 0.0002816151117825702,
"loss": 0.2404,
"retrieval/pos_neg_gap_layer35": 0.7932192757725716,
"step": 613
},
{
"epoch": 0.5559076505205975,
"grad_norm": 1.546875,
"learning_rate": 0.000281544737293998,
"loss": 0.2484,
"retrieval/pos_neg_gap_layer35": 0.799022451043129,
"step": 614
},
{
"epoch": 0.5568130375735627,
"grad_norm": 1.3515625,
"learning_rate": 0.00028147423719918026,
"loss": 0.2338,
"retrieval/pos_neg_gap_layer35": 0.7996276021003723,
"step": 615
},
{
"epoch": 0.5577184246265279,
"grad_norm": 2.109375,
"learning_rate": 0.00028140361156543447,
"loss": 0.3471,
"retrieval/pos_neg_gap_layer35": 0.7567324712872505,
"step": 616
},
{
"epoch": 0.5586238116794929,
"grad_norm": 1.6640625,
"learning_rate": 0.00028133286046019785,
"loss": 0.2535,
"retrieval/pos_neg_gap_layer35": 0.7902892753481865,
"step": 617
},
{
"epoch": 0.5595291987324581,
"grad_norm": 1.703125,
"learning_rate": 0.00028126198395102763,
"loss": 0.2459,
"retrieval/pos_neg_gap_layer35": 0.798321969807148,
"step": 618
},
{
"epoch": 0.5604345857854233,
"grad_norm": 1.203125,
"learning_rate": 0.00028119098210560047,
"loss": 0.253,
"retrieval/pos_neg_gap_layer35": 0.7879588603973389,
"step": 619
},
{
"epoch": 0.5613399728383884,
"grad_norm": 1.25,
"learning_rate": 0.0002811198549917131,
"loss": 0.3195,
"retrieval/pos_neg_gap_layer35": 0.7678583860397339,
"step": 620
},
{
"epoch": 0.5622453598913536,
"grad_norm": 1.1953125,
"learning_rate": 0.00028104860267728145,
"loss": 0.2708,
"retrieval/pos_neg_gap_layer35": 0.7841521054506302,
"step": 621
},
{
"epoch": 0.5631507469443187,
"grad_norm": 0.82421875,
"learning_rate": 0.0002809772252303413,
"loss": 0.2327,
"retrieval/pos_neg_gap_layer35": 0.804856464266777,
"step": 622
},
{
"epoch": 0.5640561339972838,
"grad_norm": 0.83203125,
"learning_rate": 0.00028090572271904777,
"loss": 0.2392,
"retrieval/pos_neg_gap_layer35": 0.795910581946373,
"step": 623
},
{
"epoch": 0.564961521050249,
"grad_norm": 0.76953125,
"learning_rate": 0.0002808340952116755,
"loss": 0.2163,
"retrieval/pos_neg_gap_layer35": 0.8119159936904907,
"step": 624
},
{
"epoch": 0.5658669081032142,
"grad_norm": 0.875,
"learning_rate": 0.0002807623427766184,
"loss": 0.2218,
"retrieval/pos_neg_gap_layer35": 0.8046466782689095,
"step": 625
},
{
"epoch": 0.5667722951561792,
"grad_norm": 0.76171875,
"learning_rate": 0.0002806904654823897,
"loss": 0.2498,
"retrieval/pos_neg_gap_layer35": 0.7915834188461304,
"step": 626
},
{
"epoch": 0.5676776822091444,
"grad_norm": 0.7578125,
"learning_rate": 0.0002806184633976218,
"loss": 0.2394,
"retrieval/pos_neg_gap_layer35": 0.8020771741867065,
"step": 627
},
{
"epoch": 0.5685830692621096,
"grad_norm": 0.96484375,
"learning_rate": 0.0002805463365910664,
"loss": 0.2379,
"retrieval/pos_neg_gap_layer35": 0.7995203211903572,
"step": 628
},
{
"epoch": 0.5694884563150747,
"grad_norm": 1.015625,
"learning_rate": 0.0002804740851315941,
"loss": 0.2601,
"retrieval/pos_neg_gap_layer35": 0.7985266149044037,
"step": 629
},
{
"epoch": 0.5703938433680399,
"grad_norm": 1.140625,
"learning_rate": 0.00028040170908819463,
"loss": 0.2982,
"retrieval/pos_neg_gap_layer35": 0.7619927078485489,
"step": 630
},
{
"epoch": 0.5712992304210049,
"grad_norm": 1.4609375,
"learning_rate": 0.0002803292085299767,
"loss": 0.2691,
"retrieval/pos_neg_gap_layer35": 0.7806361317634583,
"step": 631
},
{
"epoch": 0.5722046174739701,
"grad_norm": 0.99609375,
"learning_rate": 0.0002802565835261679,
"loss": 0.2285,
"retrieval/pos_neg_gap_layer35": 0.8028188347816467,
"step": 632
},
{
"epoch": 0.5731100045269353,
"grad_norm": 0.91015625,
"learning_rate": 0.0002801838341461146,
"loss": 0.3042,
"retrieval/pos_neg_gap_layer35": 0.7637373581528664,
"step": 633
},
{
"epoch": 0.5740153915799004,
"grad_norm": 1.25,
"learning_rate": 0.00028011096045928196,
"loss": 0.2739,
"retrieval/pos_neg_gap_layer35": 0.7869913280010223,
"step": 634
},
{
"epoch": 0.5749207786328655,
"grad_norm": 0.78515625,
"learning_rate": 0.0002800379625352539,
"loss": 0.2289,
"retrieval/pos_neg_gap_layer35": 0.8007155880331993,
"step": 635
},
{
"epoch": 0.5758261656858307,
"grad_norm": 0.73828125,
"learning_rate": 0.00027996484044373273,
"loss": 0.2386,
"retrieval/pos_neg_gap_layer35": 0.8090836703777313,
"step": 636
},
{
"epoch": 0.5767315527387958,
"grad_norm": 2.453125,
"learning_rate": 0.0002798915942545397,
"loss": 0.461,
"retrieval/pos_neg_gap_layer35": 0.6802124008536339,
"step": 637
},
{
"epoch": 0.577636939791761,
"grad_norm": 1.3046875,
"learning_rate": 0.00027981822403761424,
"loss": 0.2418,
"retrieval/pos_neg_gap_layer35": 0.7938051894307137,
"step": 638
},
{
"epoch": 0.5785423268447262,
"grad_norm": 0.96484375,
"learning_rate": 0.0002797447298630144,
"loss": 0.1973,
"retrieval/pos_neg_gap_layer35": 0.828748919069767,
"step": 639
},
{
"epoch": 0.5794477138976912,
"grad_norm": 0.76953125,
"learning_rate": 0.0002796711118009164,
"loss": 0.1981,
"retrieval/pos_neg_gap_layer35": 0.8263879418373108,
"step": 640
},
{
"epoch": 0.5803531009506564,
"grad_norm": 2.03125,
"learning_rate": 0.000279597369921615,
"loss": 0.3304,
"retrieval/pos_neg_gap_layer35": 0.7615088671445847,
"step": 641
},
{
"epoch": 0.5812584880036216,
"grad_norm": 1.03125,
"learning_rate": 0.00027952350429552293,
"loss": 0.269,
"retrieval/pos_neg_gap_layer35": 0.7872718051075935,
"step": 642
},
{
"epoch": 0.5821638750565867,
"grad_norm": 1.1953125,
"learning_rate": 0.00027944951499317135,
"loss": 0.287,
"retrieval/pos_neg_gap_layer35": 0.7705860212445259,
"step": 643
},
{
"epoch": 0.5830692621095518,
"grad_norm": 1.0546875,
"learning_rate": 0.0002793754020852092,
"loss": 0.2887,
"retrieval/pos_neg_gap_layer35": 0.7756828814744949,
"step": 644
},
{
"epoch": 0.583974649162517,
"grad_norm": 0.66796875,
"learning_rate": 0.00027930116564240374,
"loss": 0.2013,
"retrieval/pos_neg_gap_layer35": 0.8224857077002525,
"step": 645
},
{
"epoch": 0.5848800362154821,
"grad_norm": 1.0546875,
"learning_rate": 0.00027922680573564,
"loss": 0.2718,
"retrieval/pos_neg_gap_layer35": 0.7815973311662674,
"step": 646
},
{
"epoch": 0.5857854232684473,
"grad_norm": 2.1875,
"learning_rate": 0.00027915232243592097,
"loss": 0.322,
"retrieval/pos_neg_gap_layer35": 0.7526913285255432,
"step": 647
},
{
"epoch": 0.5866908103214125,
"grad_norm": 1.0625,
"learning_rate": 0.0002790777158143675,
"loss": 0.2537,
"retrieval/pos_neg_gap_layer35": 0.7880778089165688,
"step": 648
},
{
"epoch": 0.5875961973743775,
"grad_norm": 0.84375,
"learning_rate": 0.00027900298594221815,
"loss": 0.2274,
"retrieval/pos_neg_gap_layer35": 0.8052579313516617,
"step": 649
},
{
"epoch": 0.5885015844273427,
"grad_norm": 1.234375,
"learning_rate": 0.0002789281328908291,
"loss": 0.2643,
"retrieval/pos_neg_gap_layer35": 0.7882343903183937,
"step": 650
},
{
"epoch": 0.5894069714803079,
"grad_norm": 0.81640625,
"learning_rate": 0.00027885315673167426,
"loss": 0.2116,
"retrieval/pos_neg_gap_layer35": 0.8235296681523323,
"step": 651
},
{
"epoch": 0.590312358533273,
"grad_norm": 0.87109375,
"learning_rate": 0.000278778057536345,
"loss": 0.225,
"retrieval/pos_neg_gap_layer35": 0.8073238134384155,
"step": 652
},
{
"epoch": 0.5912177455862381,
"grad_norm": 0.6796875,
"learning_rate": 0.0002787028353765503,
"loss": 0.2221,
"retrieval/pos_neg_gap_layer35": 0.807576984167099,
"step": 653
},
{
"epoch": 0.5921231326392032,
"grad_norm": 0.94921875,
"learning_rate": 0.0002786274903241164,
"loss": 0.2534,
"retrieval/pos_neg_gap_layer35": 0.7969003319740295,
"step": 654
},
{
"epoch": 0.5930285196921684,
"grad_norm": 0.66015625,
"learning_rate": 0.000278552022450987,
"loss": 0.2455,
"retrieval/pos_neg_gap_layer35": 0.8003811836242676,
"step": 655
},
{
"epoch": 0.5939339067451336,
"grad_norm": 0.76171875,
"learning_rate": 0.000278476431829223,
"loss": 0.2561,
"retrieval/pos_neg_gap_layer35": 0.7966005727648735,
"step": 656
},
{
"epoch": 0.5948392937980986,
"grad_norm": 0.890625,
"learning_rate": 0.00027840071853100254,
"loss": 0.2388,
"retrieval/pos_neg_gap_layer35": 0.797157496213913,
"step": 657
},
{
"epoch": 0.5957446808510638,
"grad_norm": 0.890625,
"learning_rate": 0.00027832488262862084,
"loss": 0.2301,
"retrieval/pos_neg_gap_layer35": 0.8001188486814499,
"step": 658
},
{
"epoch": 0.596650067904029,
"grad_norm": 1.1015625,
"learning_rate": 0.00027824892419449034,
"loss": 0.2808,
"retrieval/pos_neg_gap_layer35": 0.7757779508829117,
"step": 659
},
{
"epoch": 0.5975554549569941,
"grad_norm": 0.828125,
"learning_rate": 0.00027817284330114043,
"loss": 0.2489,
"retrieval/pos_neg_gap_layer35": 0.7913022562861443,
"step": 660
},
{
"epoch": 0.5984608420099593,
"grad_norm": 1.0078125,
"learning_rate": 0.00027809664002121726,
"loss": 0.2898,
"retrieval/pos_neg_gap_layer35": 0.7776209190487862,
"step": 661
},
{
"epoch": 0.5993662290629244,
"grad_norm": 0.828125,
"learning_rate": 0.00027802031442748397,
"loss": 0.2382,
"retrieval/pos_neg_gap_layer35": 0.799986720085144,
"step": 662
},
{
"epoch": 0.6002716161158895,
"grad_norm": 1.0234375,
"learning_rate": 0.0002779438665928206,
"loss": 0.3194,
"retrieval/pos_neg_gap_layer35": 0.7608011737465858,
"step": 663
},
{
"epoch": 0.6011770031688547,
"grad_norm": 1.0546875,
"learning_rate": 0.00027786729659022374,
"loss": 0.2769,
"retrieval/pos_neg_gap_layer35": 0.7709820121526718,
"step": 664
},
{
"epoch": 0.6020823902218199,
"grad_norm": 0.77734375,
"learning_rate": 0.00027779060449280665,
"loss": 0.2508,
"retrieval/pos_neg_gap_layer35": 0.8012598231434822,
"step": 665
},
{
"epoch": 0.6029877772747849,
"grad_norm": 1.359375,
"learning_rate": 0.00027771379037379925,
"loss": 0.2566,
"retrieval/pos_neg_gap_layer35": 0.7825501784682274,
"step": 666
},
{
"epoch": 0.6038931643277501,
"grad_norm": 1.3984375,
"learning_rate": 0.000277636854306548,
"loss": 0.3002,
"retrieval/pos_neg_gap_layer35": 0.7643223255872726,
"step": 667
},
{
"epoch": 0.6047985513807153,
"grad_norm": 0.953125,
"learning_rate": 0.0002775597963645157,
"loss": 0.2734,
"retrieval/pos_neg_gap_layer35": 0.7792612761259079,
"step": 668
},
{
"epoch": 0.6057039384336804,
"grad_norm": 0.80859375,
"learning_rate": 0.00027748261662128155,
"loss": 0.2549,
"retrieval/pos_neg_gap_layer35": 0.7965288087725639,
"step": 669
},
{
"epoch": 0.6066093254866456,
"grad_norm": 1.0390625,
"learning_rate": 0.00027740531515054107,
"loss": 0.2375,
"retrieval/pos_neg_gap_layer35": 0.7989114820957184,
"step": 670
},
{
"epoch": 0.6075147125396106,
"grad_norm": 0.7109375,
"learning_rate": 0.0002773278920261061,
"loss": 0.2199,
"retrieval/pos_neg_gap_layer35": 0.8126061409711838,
"step": 671
},
{
"epoch": 0.6084200995925758,
"grad_norm": 1.0546875,
"learning_rate": 0.0002772503473219044,
"loss": 0.2635,
"retrieval/pos_neg_gap_layer35": 0.7824723571538925,
"step": 672
},
{
"epoch": 0.609325486645541,
"grad_norm": 1.6875,
"learning_rate": 0.00027717268111198014,
"loss": 0.2746,
"retrieval/pos_neg_gap_layer35": 0.7778177633881569,
"step": 673
},
{
"epoch": 0.6102308736985061,
"grad_norm": 1.0,
"learning_rate": 0.0002770948934704932,
"loss": 0.2626,
"retrieval/pos_neg_gap_layer35": 0.7910703495144844,
"step": 674
},
{
"epoch": 0.6111362607514712,
"grad_norm": 1.5,
"learning_rate": 0.0002770169844717196,
"loss": 0.3083,
"retrieval/pos_neg_gap_layer35": 0.7652135938405991,
"step": 675
},
{
"epoch": 0.6120416478044364,
"grad_norm": 0.7421875,
"learning_rate": 0.0002769389541900512,
"loss": 0.2598,
"retrieval/pos_neg_gap_layer35": 0.7849985063076019,
"step": 676
},
{
"epoch": 0.6129470348574015,
"grad_norm": 0.87890625,
"learning_rate": 0.00027686080269999575,
"loss": 0.2252,
"retrieval/pos_neg_gap_layer35": 0.8066645115613937,
"step": 677
},
{
"epoch": 0.6138524219103667,
"grad_norm": 0.93359375,
"learning_rate": 0.0002767825300761765,
"loss": 0.2563,
"retrieval/pos_neg_gap_layer35": 0.7834104672074318,
"step": 678
},
{
"epoch": 0.6147578089633319,
"grad_norm": 2.21875,
"learning_rate": 0.0002767041363933326,
"loss": 0.29,
"retrieval/pos_neg_gap_layer35": 0.7713447213172913,
"step": 679
},
{
"epoch": 0.6156631960162969,
"grad_norm": 1.1875,
"learning_rate": 0.0002766256217263187,
"loss": 0.2978,
"retrieval/pos_neg_gap_layer35": 0.7650850564241409,
"step": 680
},
{
"epoch": 0.6165685830692621,
"grad_norm": 0.96875,
"learning_rate": 0.00027654698615010495,
"loss": 0.2387,
"retrieval/pos_neg_gap_layer35": 0.796810045838356,
"step": 681
},
{
"epoch": 0.6174739701222273,
"grad_norm": 0.66015625,
"learning_rate": 0.00027646822973977704,
"loss": 0.2637,
"retrieval/pos_neg_gap_layer35": 0.7967437654733658,
"step": 682
},
{
"epoch": 0.6183793571751924,
"grad_norm": 1.078125,
"learning_rate": 0.0002763893525705359,
"loss": 0.2412,
"retrieval/pos_neg_gap_layer35": 0.8047746047377586,
"step": 683
},
{
"epoch": 0.6192847442281575,
"grad_norm": 1.9765625,
"learning_rate": 0.00027631035471769786,
"loss": 0.2574,
"retrieval/pos_neg_gap_layer35": 0.7872433438897133,
"step": 684
},
{
"epoch": 0.6201901312811227,
"grad_norm": 0.58203125,
"learning_rate": 0.00027623123625669454,
"loss": 0.1882,
"retrieval/pos_neg_gap_layer35": 0.8339928165078163,
"step": 685
},
{
"epoch": 0.6210955183340878,
"grad_norm": 1.03125,
"learning_rate": 0.0002761519972630726,
"loss": 0.2792,
"retrieval/pos_neg_gap_layer35": 0.7754400223493576,
"step": 686
},
{
"epoch": 0.622000905387053,
"grad_norm": 1.0,
"learning_rate": 0.00027607263781249383,
"loss": 0.2597,
"retrieval/pos_neg_gap_layer35": 0.7908517196774483,
"step": 687
},
{
"epoch": 0.6229062924400182,
"grad_norm": 0.6875,
"learning_rate": 0.00027599315798073514,
"loss": 0.2054,
"retrieval/pos_neg_gap_layer35": 0.8196249306201935,
"step": 688
},
{
"epoch": 0.6238116794929832,
"grad_norm": 1.1875,
"learning_rate": 0.0002759135578436882,
"loss": 0.2107,
"retrieval/pos_neg_gap_layer35": 0.8125434145331383,
"step": 689
},
{
"epoch": 0.6247170665459484,
"grad_norm": 1.0625,
"learning_rate": 0.0002758338374773597,
"loss": 0.2306,
"retrieval/pos_neg_gap_layer35": 0.8070158213376999,
"step": 690
},
{
"epoch": 0.6256224535989136,
"grad_norm": 1.59375,
"learning_rate": 0.00027575399695787113,
"loss": 0.3218,
"retrieval/pos_neg_gap_layer35": 0.7601017504930496,
"step": 691
},
{
"epoch": 0.6265278406518787,
"grad_norm": 1.4140625,
"learning_rate": 0.0002756740363614586,
"loss": 0.2385,
"retrieval/pos_neg_gap_layer35": 0.7986306101083755,
"step": 692
},
{
"epoch": 0.6274332277048438,
"grad_norm": 1.7578125,
"learning_rate": 0.00027559395576447297,
"loss": 0.326,
"retrieval/pos_neg_gap_layer35": 0.7464891001582146,
"step": 693
},
{
"epoch": 0.6283386147578089,
"grad_norm": 0.984375,
"learning_rate": 0.0002755137552433797,
"loss": 0.2308,
"retrieval/pos_neg_gap_layer35": 0.806078590452671,
"step": 694
},
{
"epoch": 0.6292440018107741,
"grad_norm": 0.6875,
"learning_rate": 0.00027543343487475866,
"loss": 0.2358,
"retrieval/pos_neg_gap_layer35": 0.8070366531610489,
"step": 695
},
{
"epoch": 0.6301493888637393,
"grad_norm": 1.2734375,
"learning_rate": 0.0002753529947353042,
"loss": 0.2801,
"retrieval/pos_neg_gap_layer35": 0.7792836129665375,
"step": 696
},
{
"epoch": 0.6310547759167043,
"grad_norm": 1.203125,
"learning_rate": 0.00027527243490182507,
"loss": 0.2783,
"retrieval/pos_neg_gap_layer35": 0.7827678993344307,
"step": 697
},
{
"epoch": 0.6319601629696695,
"grad_norm": 0.62890625,
"learning_rate": 0.0002751917554512443,
"loss": 0.1863,
"retrieval/pos_neg_gap_layer35": 0.8301127701997757,
"step": 698
},
{
"epoch": 0.6328655500226347,
"grad_norm": 0.61328125,
"learning_rate": 0.0002751109564605991,
"loss": 0.2043,
"retrieval/pos_neg_gap_layer35": 0.8225157856941223,
"step": 699
},
{
"epoch": 0.6337709370755998,
"grad_norm": 1.171875,
"learning_rate": 0.00027503003800704093,
"loss": 0.2118,
"retrieval/pos_neg_gap_layer35": 0.8165381997823715,
"step": 700
},
{
"epoch": 0.634676324128565,
"grad_norm": 1.375,
"learning_rate": 0.0002749490001678351,
"loss": 0.2515,
"retrieval/pos_neg_gap_layer35": 0.7921962887048721,
"step": 701
},
{
"epoch": 0.6355817111815301,
"grad_norm": 0.92578125,
"learning_rate": 0.0002748678430203612,
"loss": 0.2261,
"retrieval/pos_neg_gap_layer35": 0.8052646666765213,
"step": 702
},
{
"epoch": 0.6364870982344952,
"grad_norm": 1.4375,
"learning_rate": 0.00027478656664211247,
"loss": 0.2677,
"retrieval/pos_neg_gap_layer35": 0.7784762158989906,
"step": 703
},
{
"epoch": 0.6373924852874604,
"grad_norm": 1.109375,
"learning_rate": 0.00027470517111069623,
"loss": 0.2714,
"retrieval/pos_neg_gap_layer35": 0.7768310010433197,
"step": 704
},
{
"epoch": 0.6382978723404256,
"grad_norm": 1.484375,
"learning_rate": 0.00027462365650383344,
"loss": 0.2942,
"retrieval/pos_neg_gap_layer35": 0.7640905231237411,
"step": 705
},
{
"epoch": 0.6392032593933906,
"grad_norm": 0.75390625,
"learning_rate": 0.0002745420228993587,
"loss": 0.2692,
"retrieval/pos_neg_gap_layer35": 0.7996787205338478,
"step": 706
},
{
"epoch": 0.6401086464463558,
"grad_norm": 1.1953125,
"learning_rate": 0.00027446027037522045,
"loss": 0.284,
"retrieval/pos_neg_gap_layer35": 0.7731221541762352,
"step": 707
},
{
"epoch": 0.641014033499321,
"grad_norm": 1.234375,
"learning_rate": 0.0002743783990094805,
"loss": 0.2686,
"retrieval/pos_neg_gap_layer35": 0.7799250334501266,
"step": 708
},
{
"epoch": 0.6419194205522861,
"grad_norm": 0.96484375,
"learning_rate": 0.00027429640888031415,
"loss": 0.2848,
"retrieval/pos_neg_gap_layer35": 0.7671584561467171,
"step": 709
},
{
"epoch": 0.6428248076052513,
"grad_norm": 1.3046875,
"learning_rate": 0.00027421430006601023,
"loss": 0.3217,
"retrieval/pos_neg_gap_layer35": 0.7622716128826141,
"step": 710
},
{
"epoch": 0.6437301946582163,
"grad_norm": 0.72265625,
"learning_rate": 0.0002741320726449707,
"loss": 0.2457,
"retrieval/pos_neg_gap_layer35": 0.795741505920887,
"step": 711
},
{
"epoch": 0.6446355817111815,
"grad_norm": 0.81640625,
"learning_rate": 0.00027404972669571104,
"loss": 0.2102,
"retrieval/pos_neg_gap_layer35": 0.815107561647892,
"step": 712
},
{
"epoch": 0.6455409687641467,
"grad_norm": 1.0859375,
"learning_rate": 0.00027396726229685965,
"loss": 0.26,
"retrieval/pos_neg_gap_layer35": 0.7908284887671471,
"step": 713
},
{
"epoch": 0.6464463558171119,
"grad_norm": 1.1875,
"learning_rate": 0.00027388467952715813,
"loss": 0.3093,
"retrieval/pos_neg_gap_layer35": 0.7736351862549782,
"step": 714
},
{
"epoch": 0.6473517428700769,
"grad_norm": 0.67578125,
"learning_rate": 0.0002738019784654611,
"loss": 0.2413,
"retrieval/pos_neg_gap_layer35": 0.7960422784090042,
"step": 715
},
{
"epoch": 0.6482571299230421,
"grad_norm": 1.4140625,
"learning_rate": 0.00027371915919073624,
"loss": 0.2607,
"retrieval/pos_neg_gap_layer35": 0.7843823358416557,
"step": 716
},
{
"epoch": 0.6491625169760072,
"grad_norm": 0.78125,
"learning_rate": 0.0002736362217820639,
"loss": 0.2375,
"retrieval/pos_neg_gap_layer35": 0.7991039603948593,
"step": 717
},
{
"epoch": 0.6500679040289724,
"grad_norm": 1.21875,
"learning_rate": 0.00027355316631863736,
"loss": 0.3422,
"retrieval/pos_neg_gap_layer35": 0.7430773824453354,
"step": 718
},
{
"epoch": 0.6509732910819376,
"grad_norm": 0.66015625,
"learning_rate": 0.00027346999287976265,
"loss": 0.243,
"retrieval/pos_neg_gap_layer35": 0.8046166300773621,
"step": 719
},
{
"epoch": 0.6518786781349026,
"grad_norm": 0.89453125,
"learning_rate": 0.00027338670154485837,
"loss": 0.2791,
"retrieval/pos_neg_gap_layer35": 0.7783932238817215,
"step": 720
},
{
"epoch": 0.6527840651878678,
"grad_norm": 2.078125,
"learning_rate": 0.00027330329239345577,
"loss": 0.2414,
"retrieval/pos_neg_gap_layer35": 0.7974641993641853,
"step": 721
},
{
"epoch": 0.653689452240833,
"grad_norm": 0.8984375,
"learning_rate": 0.0002732197655051985,
"loss": 0.2576,
"retrieval/pos_neg_gap_layer35": 0.7874908596277237,
"step": 722
},
{
"epoch": 0.6545948392937981,
"grad_norm": 1.3359375,
"learning_rate": 0.00027313612095984267,
"loss": 0.2634,
"retrieval/pos_neg_gap_layer35": 0.7804579138755798,
"step": 723
},
{
"epoch": 0.6555002263467632,
"grad_norm": 1.6015625,
"learning_rate": 0.0002730523588372568,
"loss": 0.2665,
"retrieval/pos_neg_gap_layer35": 0.7944665104150772,
"step": 724
},
{
"epoch": 0.6564056133997284,
"grad_norm": 2.875,
"learning_rate": 0.00027296847921742166,
"loss": 0.3371,
"retrieval/pos_neg_gap_layer35": 0.7493275552988052,
"step": 725
},
{
"epoch": 0.6573110004526935,
"grad_norm": 0.93359375,
"learning_rate": 0.0002728844821804301,
"loss": 0.283,
"retrieval/pos_neg_gap_layer35": 0.7749064043164253,
"step": 726
},
{
"epoch": 0.6582163875056587,
"grad_norm": 0.84765625,
"learning_rate": 0.0002728003678064873,
"loss": 0.2166,
"retrieval/pos_neg_gap_layer35": 0.8127419948577881,
"step": 727
},
{
"epoch": 0.6591217745586239,
"grad_norm": 0.93359375,
"learning_rate": 0.0002727161361759102,
"loss": 0.2708,
"retrieval/pos_neg_gap_layer35": 0.7807655408978462,
"step": 728
},
{
"epoch": 0.6600271616115889,
"grad_norm": 0.78125,
"learning_rate": 0.00027263178736912797,
"loss": 0.2396,
"retrieval/pos_neg_gap_layer35": 0.8016004860401154,
"step": 729
},
{
"epoch": 0.6609325486645541,
"grad_norm": 0.734375,
"learning_rate": 0.0002725473214666815,
"loss": 0.2538,
"retrieval/pos_neg_gap_layer35": 0.7917163372039795,
"step": 730
},
{
"epoch": 0.6618379357175193,
"grad_norm": 0.9921875,
"learning_rate": 0.0002724627385492235,
"loss": 0.2612,
"retrieval/pos_neg_gap_layer35": 0.7879665046930313,
"step": 731
},
{
"epoch": 0.6627433227704844,
"grad_norm": 0.828125,
"learning_rate": 0.00027237803869751854,
"loss": 0.2625,
"retrieval/pos_neg_gap_layer35": 0.7773905023932457,
"step": 732
},
{
"epoch": 0.6636487098234495,
"grad_norm": 0.91015625,
"learning_rate": 0.00027229322199244277,
"loss": 0.2539,
"retrieval/pos_neg_gap_layer35": 0.7822834774851799,
"step": 733
},
{
"epoch": 0.6645540968764146,
"grad_norm": 0.9609375,
"learning_rate": 0.0002722082885149838,
"loss": 0.247,
"retrieval/pos_neg_gap_layer35": 0.7895412370562553,
"step": 734
},
{
"epoch": 0.6654594839293798,
"grad_norm": 1.640625,
"learning_rate": 0.0002721232383462409,
"loss": 0.2723,
"retrieval/pos_neg_gap_layer35": 0.7848855927586555,
"step": 735
},
{
"epoch": 0.666364870982345,
"grad_norm": 1.1484375,
"learning_rate": 0.00027203807156742473,
"loss": 0.2722,
"retrieval/pos_neg_gap_layer35": 0.7759896814823151,
"step": 736
},
{
"epoch": 0.6672702580353101,
"grad_norm": 1.390625,
"learning_rate": 0.0002719527882598573,
"loss": 0.2334,
"retrieval/pos_neg_gap_layer35": 0.802346833050251,
"step": 737
},
{
"epoch": 0.6681756450882752,
"grad_norm": 1.1953125,
"learning_rate": 0.0002718673885049717,
"loss": 0.2767,
"retrieval/pos_neg_gap_layer35": 0.7833170518279076,
"step": 738
},
{
"epoch": 0.6690810321412404,
"grad_norm": 0.90234375,
"learning_rate": 0.0002717818723843126,
"loss": 0.2467,
"retrieval/pos_neg_gap_layer35": 0.7995397970080376,
"step": 739
},
{
"epoch": 0.6699864191942055,
"grad_norm": 0.86328125,
"learning_rate": 0.00027169623997953547,
"loss": 0.2556,
"retrieval/pos_neg_gap_layer35": 0.7944035604596138,
"step": 740
},
{
"epoch": 0.6708918062471707,
"grad_norm": 1.46875,
"learning_rate": 0.0002716104913724069,
"loss": 0.2455,
"retrieval/pos_neg_gap_layer35": 0.7977171540260315,
"step": 741
},
{
"epoch": 0.6717971933001358,
"grad_norm": 1.6015625,
"learning_rate": 0.00027152462664480434,
"loss": 0.37,
"retrieval/pos_neg_gap_layer35": 0.7310309484601021,
"step": 742
},
{
"epoch": 0.6727025803531009,
"grad_norm": 0.81640625,
"learning_rate": 0.0002714386458787164,
"loss": 0.2499,
"retrieval/pos_neg_gap_layer35": 0.7907652631402016,
"step": 743
},
{
"epoch": 0.6736079674060661,
"grad_norm": 0.72265625,
"learning_rate": 0.0002713525491562421,
"loss": 0.2156,
"retrieval/pos_neg_gap_layer35": 0.8126255422830582,
"step": 744
},
{
"epoch": 0.6745133544590313,
"grad_norm": 1.1875,
"learning_rate": 0.0002712663365595916,
"loss": 0.3037,
"retrieval/pos_neg_gap_layer35": 0.7629745155572891,
"step": 745
},
{
"epoch": 0.6754187415119963,
"grad_norm": 1.21875,
"learning_rate": 0.00027118000817108527,
"loss": 0.2631,
"retrieval/pos_neg_gap_layer35": 0.7838939130306244,
"step": 746
},
{
"epoch": 0.6763241285649615,
"grad_norm": 0.87109375,
"learning_rate": 0.0002710935640731545,
"loss": 0.2489,
"retrieval/pos_neg_gap_layer35": 0.7939401865005493,
"step": 747
},
{
"epoch": 0.6772295156179267,
"grad_norm": 0.84765625,
"learning_rate": 0.0002710070043483408,
"loss": 0.21,
"retrieval/pos_neg_gap_layer35": 0.817916713654995,
"step": 748
},
{
"epoch": 0.6781349026708918,
"grad_norm": 0.90625,
"learning_rate": 0.0002709203290792962,
"loss": 0.2485,
"retrieval/pos_neg_gap_layer35": 0.7902071177959442,
"step": 749
},
{
"epoch": 0.679040289723857,
"grad_norm": 0.9140625,
"learning_rate": 0.00027083353834878314,
"loss": 0.2125,
"retrieval/pos_neg_gap_layer35": 0.8137314766645432,
"step": 750
},
{
"epoch": 0.679945676776822,
"grad_norm": 1.015625,
"learning_rate": 0.00027074663223967423,
"loss": 0.2645,
"retrieval/pos_neg_gap_layer35": 0.7793468460440636,
"step": 751
},
{
"epoch": 0.6808510638297872,
"grad_norm": 1.8828125,
"learning_rate": 0.0002706596108349522,
"loss": 0.3139,
"retrieval/pos_neg_gap_layer35": 0.7691495642066002,
"step": 752
},
{
"epoch": 0.6817564508827524,
"grad_norm": 0.8828125,
"learning_rate": 0.0002705724742177101,
"loss": 0.2743,
"retrieval/pos_neg_gap_layer35": 0.7933579981327057,
"step": 753
},
{
"epoch": 0.6826618379357176,
"grad_norm": 1.3203125,
"learning_rate": 0.00027048522247115066,
"loss": 0.3246,
"retrieval/pos_neg_gap_layer35": 0.7475208938121796,
"step": 754
},
{
"epoch": 0.6835672249886826,
"grad_norm": 1.3984375,
"learning_rate": 0.00027039785567858676,
"loss": 0.2892,
"retrieval/pos_neg_gap_layer35": 0.777754932641983,
"step": 755
},
{
"epoch": 0.6844726120416478,
"grad_norm": 1.1015625,
"learning_rate": 0.00027031037392344104,
"loss": 0.2118,
"retrieval/pos_neg_gap_layer35": 0.8151483461260796,
"step": 756
},
{
"epoch": 0.6853779990946129,
"grad_norm": 1.953125,
"learning_rate": 0.0002702227772892461,
"loss": 0.2315,
"retrieval/pos_neg_gap_layer35": 0.8042129352688789,
"step": 757
},
{
"epoch": 0.6862833861475781,
"grad_norm": 0.9140625,
"learning_rate": 0.000270135065859644,
"loss": 0.229,
"retrieval/pos_neg_gap_layer35": 0.8079781457781792,
"step": 758
},
{
"epoch": 0.6871887732005433,
"grad_norm": 0.67578125,
"learning_rate": 0.0002700472397183864,
"loss": 0.19,
"retrieval/pos_neg_gap_layer35": 0.8301509842276573,
"step": 759
},
{
"epoch": 0.6880941602535083,
"grad_norm": 1.1015625,
"learning_rate": 0.00026995929894933475,
"loss": 0.2242,
"retrieval/pos_neg_gap_layer35": 0.8051323816180229,
"step": 760
},
{
"epoch": 0.6889995473064735,
"grad_norm": 1.2890625,
"learning_rate": 0.0002698712436364597,
"loss": 0.261,
"retrieval/pos_neg_gap_layer35": 0.7857527360320091,
"step": 761
},
{
"epoch": 0.6899049343594387,
"grad_norm": 1.1796875,
"learning_rate": 0.0002697830738638415,
"loss": 0.265,
"retrieval/pos_neg_gap_layer35": 0.7946971207857132,
"step": 762
},
{
"epoch": 0.6908103214124038,
"grad_norm": 0.9140625,
"learning_rate": 0.0002696947897156694,
"loss": 0.2586,
"retrieval/pos_neg_gap_layer35": 0.7946853935718536,
"step": 763
},
{
"epoch": 0.6917157084653689,
"grad_norm": 0.8359375,
"learning_rate": 0.000269606391276242,
"loss": 0.2464,
"retrieval/pos_neg_gap_layer35": 0.7964846715331078,
"step": 764
},
{
"epoch": 0.6926210955183341,
"grad_norm": 0.88671875,
"learning_rate": 0.0002695178786299673,
"loss": 0.187,
"retrieval/pos_neg_gap_layer35": 0.8331781327724457,
"step": 765
},
{
"epoch": 0.6935264825712992,
"grad_norm": 1.1015625,
"learning_rate": 0.0002694292518613619,
"loss": 0.2465,
"retrieval/pos_neg_gap_layer35": 0.7964694648981094,
"step": 766
},
{
"epoch": 0.6944318696242644,
"grad_norm": 0.79296875,
"learning_rate": 0.00026934051105505165,
"loss": 0.2692,
"retrieval/pos_neg_gap_layer35": 0.7875032424926758,
"step": 767
},
{
"epoch": 0.6953372566772296,
"grad_norm": 0.8359375,
"learning_rate": 0.0002692516562957712,
"loss": 0.206,
"retrieval/pos_neg_gap_layer35": 0.8187586590647697,
"step": 768
},
{
"epoch": 0.6962426437301946,
"grad_norm": 1.2109375,
"learning_rate": 0.00026916268766836397,
"loss": 0.2868,
"retrieval/pos_neg_gap_layer35": 0.7804899960756302,
"step": 769
},
{
"epoch": 0.6971480307831598,
"grad_norm": 0.515625,
"learning_rate": 0.00026907360525778225,
"loss": 0.1932,
"retrieval/pos_neg_gap_layer35": 0.8266654089093208,
"step": 770
},
{
"epoch": 0.698053417836125,
"grad_norm": 1.2578125,
"learning_rate": 0.00026898440914908684,
"loss": 0.2839,
"retrieval/pos_neg_gap_layer35": 0.776340164244175,
"step": 771
},
{
"epoch": 0.6989588048890901,
"grad_norm": 1.53125,
"learning_rate": 0.0002688950994274472,
"loss": 0.2864,
"retrieval/pos_neg_gap_layer35": 0.769797295331955,
"step": 772
},
{
"epoch": 0.6998641919420552,
"grad_norm": 2.140625,
"learning_rate": 0.0002688056761781412,
"loss": 0.3698,
"retrieval/pos_neg_gap_layer35": 0.7393320947885513,
"step": 773
},
{
"epoch": 0.7007695789950203,
"grad_norm": 1.1640625,
"learning_rate": 0.0002687161394865551,
"loss": 0.2859,
"retrieval/pos_neg_gap_layer35": 0.7718316987156868,
"step": 774
},
{
"epoch": 0.7016749660479855,
"grad_norm": 1.078125,
"learning_rate": 0.0002686264894381836,
"loss": 0.2109,
"retrieval/pos_neg_gap_layer35": 0.8176923841238022,
"step": 775
},
{
"epoch": 0.7025803531009507,
"grad_norm": 0.90625,
"learning_rate": 0.0002685367261186295,
"loss": 0.2574,
"retrieval/pos_neg_gap_layer35": 0.7869426533579826,
"step": 776
},
{
"epoch": 0.7034857401539159,
"grad_norm": 0.7421875,
"learning_rate": 0.00026844684961360396,
"loss": 0.2792,
"retrieval/pos_neg_gap_layer35": 0.7818774953484535,
"step": 777
},
{
"epoch": 0.7043911272068809,
"grad_norm": 0.640625,
"learning_rate": 0.00026835686000892587,
"loss": 0.2529,
"retrieval/pos_neg_gap_layer35": 0.8022956848144531,
"step": 778
},
{
"epoch": 0.7052965142598461,
"grad_norm": 1.8828125,
"learning_rate": 0.0002682667573905224,
"loss": 0.2882,
"retrieval/pos_neg_gap_layer35": 0.7685422599315643,
"step": 779
},
{
"epoch": 0.7062019013128112,
"grad_norm": 0.859375,
"learning_rate": 0.0002681765418444286,
"loss": 0.2275,
"retrieval/pos_neg_gap_layer35": 0.8139750063419342,
"step": 780
},
{
"epoch": 0.7071072883657764,
"grad_norm": 1.15625,
"learning_rate": 0.00026808621345678725,
"loss": 0.2326,
"retrieval/pos_neg_gap_layer35": 0.8015415892004967,
"step": 781
},
{
"epoch": 0.7080126754187415,
"grad_norm": 0.59375,
"learning_rate": 0.0002679957723138489,
"loss": 0.2249,
"retrieval/pos_neg_gap_layer35": 0.8081740587949753,
"step": 782
},
{
"epoch": 0.7089180624717066,
"grad_norm": 1.0625,
"learning_rate": 0.0002679052185019719,
"loss": 0.2668,
"retrieval/pos_neg_gap_layer35": 0.7833994030952454,
"step": 783
},
{
"epoch": 0.7098234495246718,
"grad_norm": 1.3984375,
"learning_rate": 0.000267814552107622,
"loss": 0.2579,
"retrieval/pos_neg_gap_layer35": 0.7939979285001755,
"step": 784
},
{
"epoch": 0.710728836577637,
"grad_norm": 1.0234375,
"learning_rate": 0.0002677237732173725,
"loss": 0.2455,
"retrieval/pos_neg_gap_layer35": 0.8013731613755226,
"step": 785
},
{
"epoch": 0.7116342236306021,
"grad_norm": 0.8125,
"learning_rate": 0.00026763288191790417,
"loss": 0.2362,
"retrieval/pos_neg_gap_layer35": 0.7975328117609024,
"step": 786
},
{
"epoch": 0.7125396106835672,
"grad_norm": 0.76953125,
"learning_rate": 0.00026754187829600514,
"loss": 0.2541,
"retrieval/pos_neg_gap_layer35": 0.7948419600725174,
"step": 787
},
{
"epoch": 0.7134449977365324,
"grad_norm": 0.8984375,
"learning_rate": 0.0002674507624385707,
"loss": 0.2558,
"retrieval/pos_neg_gap_layer35": 0.7940822616219521,
"step": 788
},
{
"epoch": 0.7143503847894975,
"grad_norm": 1.53125,
"learning_rate": 0.0002673595344326033,
"loss": 0.1973,
"retrieval/pos_neg_gap_layer35": 0.8256388828158379,
"step": 789
},
{
"epoch": 0.7152557718424627,
"grad_norm": 0.69921875,
"learning_rate": 0.00026726819436521265,
"loss": 0.2316,
"retrieval/pos_neg_gap_layer35": 0.8052102625370026,
"step": 790
},
{
"epoch": 0.7161611588954278,
"grad_norm": 1.171875,
"learning_rate": 0.00026717674232361517,
"loss": 0.2602,
"retrieval/pos_neg_gap_layer35": 0.7890394553542137,
"step": 791
},
{
"epoch": 0.7170665459483929,
"grad_norm": 1.3046875,
"learning_rate": 0.00026708517839513446,
"loss": 0.2487,
"retrieval/pos_neg_gap_layer35": 0.7922657579183578,
"step": 792
},
{
"epoch": 0.7179719330013581,
"grad_norm": 1.0859375,
"learning_rate": 0.0002669935026672009,
"loss": 0.2768,
"retrieval/pos_neg_gap_layer35": 0.7824727445840836,
"step": 793
},
{
"epoch": 0.7188773200543233,
"grad_norm": 1.296875,
"learning_rate": 0.0002669017152273515,
"loss": 0.2927,
"retrieval/pos_neg_gap_layer35": 0.7892505452036858,
"step": 794
},
{
"epoch": 0.7197827071072883,
"grad_norm": 1.296875,
"learning_rate": 0.00026680981616323,
"loss": 0.2172,
"retrieval/pos_neg_gap_layer35": 0.8107941299676895,
"step": 795
},
{
"epoch": 0.7206880941602535,
"grad_norm": 1.2578125,
"learning_rate": 0.00026671780556258693,
"loss": 0.2902,
"retrieval/pos_neg_gap_layer35": 0.7728852853178978,
"step": 796
},
{
"epoch": 0.7215934812132186,
"grad_norm": 1.484375,
"learning_rate": 0.00026662568351327895,
"loss": 0.2961,
"retrieval/pos_neg_gap_layer35": 0.7766953334212303,
"step": 797
},
{
"epoch": 0.7224988682661838,
"grad_norm": 0.94140625,
"learning_rate": 0.0002665334501032694,
"loss": 0.21,
"retrieval/pos_neg_gap_layer35": 0.8152530640363693,
"step": 798
},
{
"epoch": 0.723404255319149,
"grad_norm": 1.3515625,
"learning_rate": 0.00026644110542062787,
"loss": 0.2824,
"retrieval/pos_neg_gap_layer35": 0.7741586863994598,
"step": 799
},
{
"epoch": 0.724309642372114,
"grad_norm": 0.7109375,
"learning_rate": 0.0002663486495535302,
"loss": 0.2428,
"retrieval/pos_neg_gap_layer35": 0.8054796233773232,
"step": 800
},
{
"epoch": 0.7252150294250792,
"grad_norm": 0.796875,
"learning_rate": 0.00026625608259025834,
"loss": 0.225,
"retrieval/pos_neg_gap_layer35": 0.8150393888354301,
"step": 801
},
{
"epoch": 0.7261204164780444,
"grad_norm": 1.1171875,
"learning_rate": 0.00026616340461920054,
"loss": 0.2679,
"retrieval/pos_neg_gap_layer35": 0.7768324166536331,
"step": 802
},
{
"epoch": 0.7270258035310095,
"grad_norm": 1.6640625,
"learning_rate": 0.0002660706157288508,
"loss": 0.2856,
"retrieval/pos_neg_gap_layer35": 0.7810106575489044,
"step": 803
},
{
"epoch": 0.7279311905839746,
"grad_norm": 1.859375,
"learning_rate": 0.00026597771600780917,
"loss": 0.2387,
"retrieval/pos_neg_gap_layer35": 0.7984187006950378,
"step": 804
},
{
"epoch": 0.7288365776369398,
"grad_norm": 1.0625,
"learning_rate": 0.0002658847055447814,
"loss": 0.279,
"retrieval/pos_neg_gap_layer35": 0.7705641388893127,
"step": 805
},
{
"epoch": 0.7297419646899049,
"grad_norm": 1.0703125,
"learning_rate": 0.0002657915844285791,
"loss": 0.3293,
"retrieval/pos_neg_gap_layer35": 0.7598812952637672,
"step": 806
},
{
"epoch": 0.7306473517428701,
"grad_norm": 0.94140625,
"learning_rate": 0.0002656983527481195,
"loss": 0.219,
"retrieval/pos_neg_gap_layer35": 0.8103760927915573,
"step": 807
},
{
"epoch": 0.7315527387958353,
"grad_norm": 1.25,
"learning_rate": 0.0002656050105924254,
"loss": 0.3557,
"retrieval/pos_neg_gap_layer35": 0.7475313097238541,
"step": 808
},
{
"epoch": 0.7324581258488003,
"grad_norm": 1.296875,
"learning_rate": 0.0002655115580506251,
"loss": 0.254,
"retrieval/pos_neg_gap_layer35": 0.7876361235976219,
"step": 809
},
{
"epoch": 0.7333635129017655,
"grad_norm": 1.328125,
"learning_rate": 0.00026541799521195234,
"loss": 0.261,
"retrieval/pos_neg_gap_layer35": 0.7863417491316795,
"step": 810
},
{
"epoch": 0.7342688999547307,
"grad_norm": 1.203125,
"learning_rate": 0.00026532432216574606,
"loss": 0.2901,
"retrieval/pos_neg_gap_layer35": 0.7657178938388824,
"step": 811
},
{
"epoch": 0.7351742870076958,
"grad_norm": 0.84765625,
"learning_rate": 0.0002652305390014505,
"loss": 0.2428,
"retrieval/pos_neg_gap_layer35": 0.7951228842139244,
"step": 812
},
{
"epoch": 0.7360796740606609,
"grad_norm": 0.86328125,
"learning_rate": 0.00026513664580861515,
"loss": 0.2448,
"retrieval/pos_neg_gap_layer35": 0.7958799079060555,
"step": 813
},
{
"epoch": 0.736985061113626,
"grad_norm": 1.046875,
"learning_rate": 0.00026504264267689434,
"loss": 0.272,
"retrieval/pos_neg_gap_layer35": 0.781773753464222,
"step": 814
},
{
"epoch": 0.7378904481665912,
"grad_norm": 1.15625,
"learning_rate": 0.0002649485296960476,
"loss": 0.2579,
"retrieval/pos_neg_gap_layer35": 0.7895345464348793,
"step": 815
},
{
"epoch": 0.7387958352195564,
"grad_norm": 0.79296875,
"learning_rate": 0.0002648543069559392,
"loss": 0.2233,
"retrieval/pos_neg_gap_layer35": 0.8058580607175827,
"step": 816
},
{
"epoch": 0.7397012222725216,
"grad_norm": 1.25,
"learning_rate": 0.00026475997454653824,
"loss": 0.2476,
"retrieval/pos_neg_gap_layer35": 0.7939665913581848,
"step": 817
},
{
"epoch": 0.7406066093254866,
"grad_norm": 0.796875,
"learning_rate": 0.0002646655325579186,
"loss": 0.2263,
"retrieval/pos_neg_gap_layer35": 0.8070612028241158,
"step": 818
},
{
"epoch": 0.7415119963784518,
"grad_norm": 1.2734375,
"learning_rate": 0.00026457098108025875,
"loss": 0.2519,
"retrieval/pos_neg_gap_layer35": 0.7897947207093239,
"step": 819
},
{
"epoch": 0.7424173834314169,
"grad_norm": 1.09375,
"learning_rate": 0.0002644763202038417,
"loss": 0.2488,
"retrieval/pos_neg_gap_layer35": 0.7886952385306358,
"step": 820
},
{
"epoch": 0.7433227704843821,
"grad_norm": 0.90625,
"learning_rate": 0.00026438155001905493,
"loss": 0.2587,
"retrieval/pos_neg_gap_layer35": 0.7879218980669975,
"step": 821
},
{
"epoch": 0.7442281575373472,
"grad_norm": 1.1484375,
"learning_rate": 0.0002642866706163903,
"loss": 0.2496,
"retrieval/pos_neg_gap_layer35": 0.7932540699839592,
"step": 822
},
{
"epoch": 0.7451335445903123,
"grad_norm": 1.1640625,
"learning_rate": 0.0002641916820864439,
"loss": 0.2762,
"retrieval/pos_neg_gap_layer35": 0.7802223041653633,
"step": 823
},
{
"epoch": 0.7460389316432775,
"grad_norm": 1.515625,
"learning_rate": 0.00026409658451991615,
"loss": 0.2793,
"retrieval/pos_neg_gap_layer35": 0.7764236778020859,
"step": 824
},
{
"epoch": 0.7469443186962427,
"grad_norm": 0.92578125,
"learning_rate": 0.0002640013780076115,
"loss": 0.2171,
"retrieval/pos_neg_gap_layer35": 0.8078960403800011,
"step": 825
},
{
"epoch": 0.7478497057492078,
"grad_norm": 1.5859375,
"learning_rate": 0.00026390606264043836,
"loss": 0.2791,
"retrieval/pos_neg_gap_layer35": 0.7736713513731956,
"step": 826
},
{
"epoch": 0.7487550928021729,
"grad_norm": 0.5859375,
"learning_rate": 0.0002638106385094092,
"loss": 0.2034,
"retrieval/pos_neg_gap_layer35": 0.8198371529579163,
"step": 827
},
{
"epoch": 0.7496604798551381,
"grad_norm": 1.3828125,
"learning_rate": 0.0002637151057056403,
"loss": 0.2744,
"retrieval/pos_neg_gap_layer35": 0.778425395488739,
"step": 828
},
{
"epoch": 0.7505658669081032,
"grad_norm": 0.97265625,
"learning_rate": 0.00026361946432035174,
"loss": 0.2677,
"retrieval/pos_neg_gap_layer35": 0.7889441475272179,
"step": 829
},
{
"epoch": 0.7514712539610684,
"grad_norm": 0.71484375,
"learning_rate": 0.0002635237144448672,
"loss": 0.2165,
"retrieval/pos_neg_gap_layer35": 0.8167901411652565,
"step": 830
},
{
"epoch": 0.7523766410140335,
"grad_norm": 1.125,
"learning_rate": 0.00026342785617061395,
"loss": 0.2452,
"retrieval/pos_neg_gap_layer35": 0.7951151728630066,
"step": 831
},
{
"epoch": 0.7532820280669986,
"grad_norm": 0.9453125,
"learning_rate": 0.0002633318895891229,
"loss": 0.2874,
"retrieval/pos_neg_gap_layer35": 0.7802886888384819,
"step": 832
},
{
"epoch": 0.7541874151199638,
"grad_norm": 0.88671875,
"learning_rate": 0.0002632358147920283,
"loss": 0.2106,
"retrieval/pos_neg_gap_layer35": 0.8160812854766846,
"step": 833
},
{
"epoch": 0.755092802172929,
"grad_norm": 0.90234375,
"learning_rate": 0.0002631396318710676,
"loss": 0.2136,
"retrieval/pos_neg_gap_layer35": 0.8146700114011765,
"step": 834
},
{
"epoch": 0.7559981892258941,
"grad_norm": 1.015625,
"learning_rate": 0.0002630433409180817,
"loss": 0.249,
"retrieval/pos_neg_gap_layer35": 0.7956576943397522,
"step": 835
},
{
"epoch": 0.7569035762788592,
"grad_norm": 1.1484375,
"learning_rate": 0.0002629469420250146,
"loss": 0.2578,
"retrieval/pos_neg_gap_layer35": 0.7848058491945267,
"step": 836
},
{
"epoch": 0.7578089633318243,
"grad_norm": 1.1875,
"learning_rate": 0.00026285043528391326,
"loss": 0.2256,
"retrieval/pos_neg_gap_layer35": 0.8082139268517494,
"step": 837
},
{
"epoch": 0.7587143503847895,
"grad_norm": 1.359375,
"learning_rate": 0.00026275382078692775,
"loss": 0.2713,
"retrieval/pos_neg_gap_layer35": 0.7795856818556786,
"step": 838
},
{
"epoch": 0.7596197374377547,
"grad_norm": 0.8046875,
"learning_rate": 0.0002626570986263109,
"loss": 0.2067,
"retrieval/pos_neg_gap_layer35": 0.815989650785923,
"step": 839
},
{
"epoch": 0.7605251244907197,
"grad_norm": 1.3671875,
"learning_rate": 0.0002625602688944185,
"loss": 0.2261,
"retrieval/pos_neg_gap_layer35": 0.8062889724969864,
"step": 840
},
{
"epoch": 0.7614305115436849,
"grad_norm": 1.0,
"learning_rate": 0.000262463331683709,
"loss": 0.2589,
"retrieval/pos_neg_gap_layer35": 0.7887637838721275,
"step": 841
},
{
"epoch": 0.7623358985966501,
"grad_norm": 0.9296875,
"learning_rate": 0.00026236628708674326,
"loss": 0.2603,
"retrieval/pos_neg_gap_layer35": 0.8031280636787415,
"step": 842
},
{
"epoch": 0.7632412856496152,
"grad_norm": 1.515625,
"learning_rate": 0.0002622691351961851,
"loss": 0.3097,
"retrieval/pos_neg_gap_layer35": 0.7644767537713051,
"step": 843
},
{
"epoch": 0.7641466727025803,
"grad_norm": 0.74609375,
"learning_rate": 0.0002621718761048005,
"loss": 0.2041,
"retrieval/pos_neg_gap_layer35": 0.8252085968852043,
"step": 844
},
{
"epoch": 0.7650520597555455,
"grad_norm": 1.078125,
"learning_rate": 0.0002620745099054577,
"loss": 0.28,
"retrieval/pos_neg_gap_layer35": 0.7784862369298935,
"step": 845
},
{
"epoch": 0.7659574468085106,
"grad_norm": 1.90625,
"learning_rate": 0.00026197703669112756,
"loss": 0.2623,
"retrieval/pos_neg_gap_layer35": 0.7795697599649429,
"step": 846
},
{
"epoch": 0.7668628338614758,
"grad_norm": 0.8125,
"learning_rate": 0.0002618794565548829,
"loss": 0.2364,
"retrieval/pos_neg_gap_layer35": 0.8108007684350014,
"step": 847
},
{
"epoch": 0.767768220914441,
"grad_norm": 1.859375,
"learning_rate": 0.0002617817695898986,
"loss": 0.3034,
"retrieval/pos_neg_gap_layer35": 0.7762005999684334,
"step": 848
},
{
"epoch": 0.768673607967406,
"grad_norm": 1.328125,
"learning_rate": 0.0002616839758894517,
"loss": 0.2853,
"retrieval/pos_neg_gap_layer35": 0.7714657038450241,
"step": 849
},
{
"epoch": 0.7695789950203712,
"grad_norm": 2.078125,
"learning_rate": 0.00026158607554692095,
"loss": 0.3253,
"retrieval/pos_neg_gap_layer35": 0.7641889005899429,
"step": 850
},
{
"epoch": 0.7704843820733364,
"grad_norm": 1.3515625,
"learning_rate": 0.0002614880686557872,
"loss": 0.3744,
"retrieval/pos_neg_gap_layer35": 0.7332576885819435,
"step": 851
},
{
"epoch": 0.7713897691263015,
"grad_norm": 1.03125,
"learning_rate": 0.00026138995530963283,
"loss": 0.298,
"retrieval/pos_neg_gap_layer35": 0.7759854197502136,
"step": 852
},
{
"epoch": 0.7722951561792666,
"grad_norm": 0.76171875,
"learning_rate": 0.0002612917356021419,
"loss": 0.2495,
"retrieval/pos_neg_gap_layer35": 0.7907742783427238,
"step": 853
},
{
"epoch": 0.7732005432322318,
"grad_norm": 0.96484375,
"learning_rate": 0.0002611934096271001,
"loss": 0.2808,
"retrieval/pos_neg_gap_layer35": 0.7738053500652313,
"step": 854
},
{
"epoch": 0.7741059302851969,
"grad_norm": 1.34375,
"learning_rate": 0.0002610949774783946,
"loss": 0.2479,
"retrieval/pos_neg_gap_layer35": 0.7904229015111923,
"step": 855
},
{
"epoch": 0.7750113173381621,
"grad_norm": 1.0703125,
"learning_rate": 0.00026099643925001374,
"loss": 0.2253,
"retrieval/pos_neg_gap_layer35": 0.8009040504693985,
"step": 856
},
{
"epoch": 0.7759167043911273,
"grad_norm": 1.7890625,
"learning_rate": 0.0002608977950360475,
"loss": 0.3042,
"retrieval/pos_neg_gap_layer35": 0.764371819794178,
"step": 857
},
{
"epoch": 0.7768220914440923,
"grad_norm": 1.125,
"learning_rate": 0.00026079904493068676,
"loss": 0.2502,
"retrieval/pos_neg_gap_layer35": 0.7939741462469101,
"step": 858
},
{
"epoch": 0.7777274784970575,
"grad_norm": 1.53125,
"learning_rate": 0.0002607001890282236,
"loss": 0.29,
"retrieval/pos_neg_gap_layer35": 0.772497832775116,
"step": 859
},
{
"epoch": 0.7786328655500226,
"grad_norm": 1.6328125,
"learning_rate": 0.00026060122742305126,
"loss": 0.2907,
"retrieval/pos_neg_gap_layer35": 0.7709087654948235,
"step": 860
},
{
"epoch": 0.7795382526029878,
"grad_norm": 0.91796875,
"learning_rate": 0.0002605021602096637,
"loss": 0.2198,
"retrieval/pos_neg_gap_layer35": 0.808967225253582,
"step": 861
},
{
"epoch": 0.7804436396559529,
"grad_norm": 0.828125,
"learning_rate": 0.0002604029874826558,
"loss": 0.2286,
"retrieval/pos_neg_gap_layer35": 0.8064554184675217,
"step": 862
},
{
"epoch": 0.781349026708918,
"grad_norm": 0.8515625,
"learning_rate": 0.0002603037093367233,
"loss": 0.2288,
"retrieval/pos_neg_gap_layer35": 0.8130983859300613,
"step": 863
},
{
"epoch": 0.7822544137618832,
"grad_norm": 1.421875,
"learning_rate": 0.0002602043258666624,
"loss": 0.3374,
"retrieval/pos_neg_gap_layer35": 0.7558306455612183,
"step": 864
},
{
"epoch": 0.7831598008148484,
"grad_norm": 1.0078125,
"learning_rate": 0.0002601048371673701,
"loss": 0.2104,
"retrieval/pos_neg_gap_layer35": 0.8190824091434479,
"step": 865
},
{
"epoch": 0.7840651878678135,
"grad_norm": 1.4375,
"learning_rate": 0.00026000524333384364,
"loss": 0.2582,
"retrieval/pos_neg_gap_layer35": 0.7884343862533569,
"step": 866
},
{
"epoch": 0.7849705749207786,
"grad_norm": 1.53125,
"learning_rate": 0.00025990554446118084,
"loss": 0.2515,
"retrieval/pos_neg_gap_layer35": 0.7916184440255165,
"step": 867
},
{
"epoch": 0.7858759619737438,
"grad_norm": 2.25,
"learning_rate": 0.0002598057406445797,
"loss": 0.2675,
"retrieval/pos_neg_gap_layer35": 0.7847841680049896,
"step": 868
},
{
"epoch": 0.7867813490267089,
"grad_norm": 1.1015625,
"learning_rate": 0.00025970583197933844,
"loss": 0.2045,
"retrieval/pos_neg_gap_layer35": 0.8189889416098595,
"step": 869
},
{
"epoch": 0.7876867360796741,
"grad_norm": 0.8515625,
"learning_rate": 0.0002596058185608555,
"loss": 0.2301,
"retrieval/pos_neg_gap_layer35": 0.8102581202983856,
"step": 870
},
{
"epoch": 0.7885921231326392,
"grad_norm": 0.8515625,
"learning_rate": 0.00025950570048462917,
"loss": 0.2495,
"retrieval/pos_neg_gap_layer35": 0.7887965142726898,
"step": 871
},
{
"epoch": 0.7894975101856043,
"grad_norm": 1.2265625,
"learning_rate": 0.00025940547784625783,
"loss": 0.2492,
"retrieval/pos_neg_gap_layer35": 0.8017110899090767,
"step": 872
},
{
"epoch": 0.7904028972385695,
"grad_norm": 0.7265625,
"learning_rate": 0.0002593051507414396,
"loss": 0.2044,
"retrieval/pos_neg_gap_layer35": 0.8205394297838211,
"step": 873
},
{
"epoch": 0.7913082842915347,
"grad_norm": 0.890625,
"learning_rate": 0.0002592047192659725,
"loss": 0.2097,
"retrieval/pos_neg_gap_layer35": 0.8150735050439835,
"step": 874
},
{
"epoch": 0.7922136713444998,
"grad_norm": 1.4375,
"learning_rate": 0.000259104183515754,
"loss": 0.2385,
"retrieval/pos_neg_gap_layer35": 0.8078443184494972,
"step": 875
},
{
"epoch": 0.7931190583974649,
"grad_norm": 1.3828125,
"learning_rate": 0.00025900354358678117,
"loss": 0.2836,
"retrieval/pos_neg_gap_layer35": 0.7848035022616386,
"step": 876
},
{
"epoch": 0.79402444545043,
"grad_norm": 0.796875,
"learning_rate": 0.00025890279957515075,
"loss": 0.2428,
"retrieval/pos_neg_gap_layer35": 0.7967529892921448,
"step": 877
},
{
"epoch": 0.7949298325033952,
"grad_norm": 1.078125,
"learning_rate": 0.00025880195157705865,
"loss": 0.2708,
"retrieval/pos_neg_gap_layer35": 0.780790701508522,
"step": 878
},
{
"epoch": 0.7958352195563604,
"grad_norm": 1.234375,
"learning_rate": 0.00025870099968880033,
"loss": 0.2452,
"retrieval/pos_neg_gap_layer35": 0.7903577834367752,
"step": 879
},
{
"epoch": 0.7967406066093254,
"grad_norm": 0.86328125,
"learning_rate": 0.00025859994400677003,
"loss": 0.2506,
"retrieval/pos_neg_gap_layer35": 0.7957386821508408,
"step": 880
},
{
"epoch": 0.7976459936622906,
"grad_norm": 0.87890625,
"learning_rate": 0.0002584987846274615,
"loss": 0.2246,
"retrieval/pos_neg_gap_layer35": 0.8059256225824356,
"step": 881
},
{
"epoch": 0.7985513807152558,
"grad_norm": 1.3359375,
"learning_rate": 0.0002583975216474673,
"loss": 0.229,
"retrieval/pos_neg_gap_layer35": 0.8019848540425301,
"step": 882
},
{
"epoch": 0.7994567677682209,
"grad_norm": 1.0703125,
"learning_rate": 0.000258296155163479,
"loss": 0.2385,
"retrieval/pos_neg_gap_layer35": 0.7972767874598503,
"step": 883
},
{
"epoch": 0.8003621548211861,
"grad_norm": 1.515625,
"learning_rate": 0.00025819468527228694,
"loss": 0.2829,
"retrieval/pos_neg_gap_layer35": 0.7711047306656837,
"step": 884
},
{
"epoch": 0.8012675418741512,
"grad_norm": 1.1875,
"learning_rate": 0.00025809311207078027,
"loss": 0.2827,
"retrieval/pos_neg_gap_layer35": 0.7782229855656624,
"step": 885
},
{
"epoch": 0.8021729289271163,
"grad_norm": 0.7578125,
"learning_rate": 0.00025799143565594665,
"loss": 0.2014,
"retrieval/pos_neg_gap_layer35": 0.8231864497065544,
"step": 886
},
{
"epoch": 0.8030783159800815,
"grad_norm": 0.97265625,
"learning_rate": 0.0002578896561248725,
"loss": 0.2191,
"retrieval/pos_neg_gap_layer35": 0.8173988163471222,
"step": 887
},
{
"epoch": 0.8039837030330467,
"grad_norm": 0.94140625,
"learning_rate": 0.00025778777357474256,
"loss": 0.2653,
"retrieval/pos_neg_gap_layer35": 0.7789407521486282,
"step": 888
},
{
"epoch": 0.8048890900860117,
"grad_norm": 0.83984375,
"learning_rate": 0.0002576857881028399,
"loss": 0.2552,
"retrieval/pos_neg_gap_layer35": 0.7950991243124008,
"step": 889
},
{
"epoch": 0.8057944771389769,
"grad_norm": 0.6796875,
"learning_rate": 0.0002575836998065459,
"loss": 0.2031,
"retrieval/pos_neg_gap_layer35": 0.8203428909182549,
"step": 890
},
{
"epoch": 0.8066998641919421,
"grad_norm": 1.3046875,
"learning_rate": 0.0002574815087833403,
"loss": 0.2322,
"retrieval/pos_neg_gap_layer35": 0.8104332163929939,
"step": 891
},
{
"epoch": 0.8076052512449072,
"grad_norm": 1.6796875,
"learning_rate": 0.0002573792151308007,
"loss": 0.4525,
"retrieval/pos_neg_gap_layer35": 0.7075380086898804,
"step": 892
},
{
"epoch": 0.8085106382978723,
"grad_norm": 1.7578125,
"learning_rate": 0.0002572768189466029,
"loss": 0.3194,
"retrieval/pos_neg_gap_layer35": 0.7556198164820671,
"step": 893
},
{
"epoch": 0.8094160253508375,
"grad_norm": 1.53125,
"learning_rate": 0.0002571743203285202,
"loss": 0.2833,
"retrieval/pos_neg_gap_layer35": 0.7823956087231636,
"step": 894
},
{
"epoch": 0.8103214124038026,
"grad_norm": 0.490234375,
"learning_rate": 0.0002570717193744242,
"loss": 0.1903,
"retrieval/pos_neg_gap_layer35": 0.8307401612401009,
"step": 895
},
{
"epoch": 0.8112267994567678,
"grad_norm": 0.6953125,
"learning_rate": 0.00025696901618228404,
"loss": 0.3126,
"retrieval/pos_neg_gap_layer35": 0.7726689204573631,
"step": 896
},
{
"epoch": 0.812132186509733,
"grad_norm": 1.0078125,
"learning_rate": 0.0002568662108501664,
"loss": 0.324,
"retrieval/pos_neg_gap_layer35": 0.7571419104933739,
"step": 897
},
{
"epoch": 0.813037573562698,
"grad_norm": 0.99609375,
"learning_rate": 0.00025676330347623543,
"loss": 0.211,
"retrieval/pos_neg_gap_layer35": 0.817098468542099,
"step": 898
},
{
"epoch": 0.8139429606156632,
"grad_norm": 1.0859375,
"learning_rate": 0.000256660294158753,
"loss": 0.2407,
"retrieval/pos_neg_gap_layer35": 0.7962794974446297,
"step": 899
},
{
"epoch": 0.8148483476686283,
"grad_norm": 1.0703125,
"learning_rate": 0.000256557182996078,
"loss": 0.2428,
"retrieval/pos_neg_gap_layer35": 0.7971868067979813,
"step": 900
},
{
"epoch": 0.8157537347215935,
"grad_norm": 1.28125,
"learning_rate": 0.0002564539700866669,
"loss": 0.2679,
"retrieval/pos_neg_gap_layer35": 0.7822175770998001,
"step": 901
},
{
"epoch": 0.8166591217745586,
"grad_norm": 2.203125,
"learning_rate": 0.0002563506555290729,
"loss": 0.2511,
"retrieval/pos_neg_gap_layer35": 0.7967018857598305,
"step": 902
},
{
"epoch": 0.8175645088275237,
"grad_norm": 0.625,
"learning_rate": 0.00025624723942194665,
"loss": 0.1978,
"retrieval/pos_neg_gap_layer35": 0.8219863623380661,
"step": 903
},
{
"epoch": 0.8184698958804889,
"grad_norm": 0.90625,
"learning_rate": 0.0002561437218640356,
"loss": 0.214,
"retrieval/pos_neg_gap_layer35": 0.8201027736067772,
"step": 904
},
{
"epoch": 0.8193752829334541,
"grad_norm": 0.765625,
"learning_rate": 0.000256040102954184,
"loss": 0.198,
"retrieval/pos_neg_gap_layer35": 0.8235468491911888,
"step": 905
},
{
"epoch": 0.8202806699864192,
"grad_norm": 1.0234375,
"learning_rate": 0.00025593638279133304,
"loss": 0.2527,
"retrieval/pos_neg_gap_layer35": 0.7874069735407829,
"step": 906
},
{
"epoch": 0.8211860570393843,
"grad_norm": 0.73828125,
"learning_rate": 0.0002558325614745205,
"loss": 0.2106,
"retrieval/pos_neg_gap_layer35": 0.8157206922769547,
"step": 907
},
{
"epoch": 0.8220914440923495,
"grad_norm": 0.70703125,
"learning_rate": 0.00025572863910288074,
"loss": 0.2587,
"retrieval/pos_neg_gap_layer35": 0.7890689224004745,
"step": 908
},
{
"epoch": 0.8229968311453146,
"grad_norm": 0.87109375,
"learning_rate": 0.00025562461577564464,
"loss": 0.2673,
"retrieval/pos_neg_gap_layer35": 0.7909923046827316,
"step": 909
},
{
"epoch": 0.8239022181982798,
"grad_norm": 0.6796875,
"learning_rate": 0.0002555204915921395,
"loss": 0.2622,
"retrieval/pos_neg_gap_layer35": 0.7884917035698891,
"step": 910
},
{
"epoch": 0.8248076052512449,
"grad_norm": 2.578125,
"learning_rate": 0.0002554162666517889,
"loss": 0.2786,
"retrieval/pos_neg_gap_layer35": 0.7719915434718132,
"step": 911
},
{
"epoch": 0.82571299230421,
"grad_norm": 2.390625,
"learning_rate": 0.0002553119410541125,
"loss": 0.2772,
"retrieval/pos_neg_gap_layer35": 0.7800751775503159,
"step": 912
},
{
"epoch": 0.8266183793571752,
"grad_norm": 1.109375,
"learning_rate": 0.0002552075148987263,
"loss": 0.2903,
"retrieval/pos_neg_gap_layer35": 0.774979867041111,
"step": 913
},
{
"epoch": 0.8275237664101404,
"grad_norm": 1.234375,
"learning_rate": 0.0002551029882853422,
"loss": 0.2703,
"retrieval/pos_neg_gap_layer35": 0.7849760055541992,
"step": 914
},
{
"epoch": 0.8284291534631055,
"grad_norm": 0.8046875,
"learning_rate": 0.000254998361313768,
"loss": 0.2745,
"retrieval/pos_neg_gap_layer35": 0.7865116968750954,
"step": 915
},
{
"epoch": 0.8293345405160706,
"grad_norm": 1.078125,
"learning_rate": 0.0002548936340839074,
"loss": 0.2798,
"retrieval/pos_neg_gap_layer35": 0.7695112898945808,
"step": 916
},
{
"epoch": 0.8302399275690358,
"grad_norm": 0.87890625,
"learning_rate": 0.0002547888066957598,
"loss": 0.2109,
"retrieval/pos_neg_gap_layer35": 0.8211440369486809,
"step": 917
},
{
"epoch": 0.8311453146220009,
"grad_norm": 1.203125,
"learning_rate": 0.0002546838792494202,
"loss": 0.2482,
"retrieval/pos_neg_gap_layer35": 0.78993871062994,
"step": 918
},
{
"epoch": 0.8320507016749661,
"grad_norm": 1.1640625,
"learning_rate": 0.00025457885184507916,
"loss": 0.2496,
"retrieval/pos_neg_gap_layer35": 0.7901500687003136,
"step": 919
},
{
"epoch": 0.8329560887279311,
"grad_norm": 1.453125,
"learning_rate": 0.0002544737245830228,
"loss": 0.3179,
"retrieval/pos_neg_gap_layer35": 0.770226463675499,
"step": 920
},
{
"epoch": 0.8338614757808963,
"grad_norm": 1.1171875,
"learning_rate": 0.0002543684975636324,
"loss": 0.231,
"retrieval/pos_neg_gap_layer35": 0.8057003617286682,
"step": 921
},
{
"epoch": 0.8347668628338615,
"grad_norm": 1.2109375,
"learning_rate": 0.00025426317088738455,
"loss": 0.28,
"retrieval/pos_neg_gap_layer35": 0.7737095281481743,
"step": 922
},
{
"epoch": 0.8356722498868266,
"grad_norm": 0.9375,
"learning_rate": 0.00025415774465485105,
"loss": 0.2694,
"retrieval/pos_neg_gap_layer35": 0.7754654660820961,
"step": 923
},
{
"epoch": 0.8365776369397918,
"grad_norm": 0.890625,
"learning_rate": 0.00025405221896669883,
"loss": 0.2748,
"retrieval/pos_neg_gap_layer35": 0.7828161790966988,
"step": 924
},
{
"epoch": 0.8374830239927569,
"grad_norm": 1.78125,
"learning_rate": 0.0002539465939236896,
"loss": 0.2375,
"retrieval/pos_neg_gap_layer35": 0.8012754917144775,
"step": 925
},
{
"epoch": 0.838388411045722,
"grad_norm": 1.2109375,
"learning_rate": 0.0002538408696266801,
"loss": 0.2565,
"retrieval/pos_neg_gap_layer35": 0.7867497950792313,
"step": 926
},
{
"epoch": 0.8392937980986872,
"grad_norm": 0.60546875,
"learning_rate": 0.00025373504617662177,
"loss": 0.1983,
"retrieval/pos_neg_gap_layer35": 0.8302271440625191,
"step": 927
},
{
"epoch": 0.8401991851516524,
"grad_norm": 0.98046875,
"learning_rate": 0.00025362912367456067,
"loss": 0.2455,
"retrieval/pos_neg_gap_layer35": 0.8058625012636185,
"step": 928
},
{
"epoch": 0.8411045722046174,
"grad_norm": 1.0390625,
"learning_rate": 0.00025352310222163753,
"loss": 0.2417,
"retrieval/pos_neg_gap_layer35": 0.7997393757104874,
"step": 929
},
{
"epoch": 0.8420099592575826,
"grad_norm": 1.484375,
"learning_rate": 0.00025341698191908767,
"loss": 0.2704,
"retrieval/pos_neg_gap_layer35": 0.7864329069852829,
"step": 930
},
{
"epoch": 0.8429153463105478,
"grad_norm": 1.1796875,
"learning_rate": 0.0002533107628682405,
"loss": 0.2907,
"retrieval/pos_neg_gap_layer35": 0.7752860710024834,
"step": 931
},
{
"epoch": 0.8438207333635129,
"grad_norm": 0.984375,
"learning_rate": 0.00025320444517051995,
"loss": 0.2442,
"retrieval/pos_neg_gap_layer35": 0.8008893802762032,
"step": 932
},
{
"epoch": 0.8447261204164781,
"grad_norm": 0.7265625,
"learning_rate": 0.0002530980289274441,
"loss": 0.2299,
"retrieval/pos_neg_gap_layer35": 0.8153559789061546,
"step": 933
},
{
"epoch": 0.8456315074694432,
"grad_norm": 0.75390625,
"learning_rate": 0.00025299151424062506,
"loss": 0.262,
"retrieval/pos_neg_gap_layer35": 0.7927069589495659,
"step": 934
},
{
"epoch": 0.8465368945224083,
"grad_norm": 1.4453125,
"learning_rate": 0.0002528849012117691,
"loss": 0.3038,
"retrieval/pos_neg_gap_layer35": 0.7629602923989296,
"step": 935
},
{
"epoch": 0.8474422815753735,
"grad_norm": 1.0078125,
"learning_rate": 0.00025277818994267617,
"loss": 0.2511,
"retrieval/pos_neg_gap_layer35": 0.7986046001315117,
"step": 936
},
{
"epoch": 0.8483476686283387,
"grad_norm": 0.80078125,
"learning_rate": 0.00025267138053524014,
"loss": 0.2423,
"retrieval/pos_neg_gap_layer35": 0.8008102253079414,
"step": 937
},
{
"epoch": 0.8492530556813037,
"grad_norm": 1.0078125,
"learning_rate": 0.00025256447309144866,
"loss": 0.2343,
"retrieval/pos_neg_gap_layer35": 0.8064358606934547,
"step": 938
},
{
"epoch": 0.8501584427342689,
"grad_norm": 0.8359375,
"learning_rate": 0.0002524574677133829,
"loss": 0.2284,
"retrieval/pos_neg_gap_layer35": 0.8028170242905617,
"step": 939
},
{
"epoch": 0.851063829787234,
"grad_norm": 0.7265625,
"learning_rate": 0.0002523503645032174,
"loss": 0.2198,
"retrieval/pos_neg_gap_layer35": 0.8098101541399956,
"step": 940
},
{
"epoch": 0.8519692168401992,
"grad_norm": 1.140625,
"learning_rate": 0.0002522431635632205,
"loss": 0.2271,
"retrieval/pos_neg_gap_layer35": 0.805466927587986,
"step": 941
},
{
"epoch": 0.8528746038931643,
"grad_norm": 0.9453125,
"learning_rate": 0.00025213586499575353,
"loss": 0.2273,
"retrieval/pos_neg_gap_layer35": 0.8046845644712448,
"step": 942
},
{
"epoch": 0.8537799909461294,
"grad_norm": 0.48828125,
"learning_rate": 0.00025202846890327114,
"loss": 0.1722,
"retrieval/pos_neg_gap_layer35": 0.8433089032769203,
"step": 943
},
{
"epoch": 0.8546853779990946,
"grad_norm": 1.78125,
"learning_rate": 0.0002519209753883211,
"loss": 0.256,
"retrieval/pos_neg_gap_layer35": 0.7957945615053177,
"step": 944
},
{
"epoch": 0.8555907650520598,
"grad_norm": 1.3125,
"learning_rate": 0.00025181338455354426,
"loss": 0.2704,
"retrieval/pos_neg_gap_layer35": 0.7908289209008217,
"step": 945
},
{
"epoch": 0.8564961521050249,
"grad_norm": 1.0234375,
"learning_rate": 0.0002517056965016742,
"loss": 0.3482,
"retrieval/pos_neg_gap_layer35": 0.7530882582068443,
"step": 946
},
{
"epoch": 0.85740153915799,
"grad_norm": 0.64453125,
"learning_rate": 0.0002515979113355377,
"loss": 0.2599,
"retrieval/pos_neg_gap_layer35": 0.7947349473834038,
"step": 947
},
{
"epoch": 0.8583069262109552,
"grad_norm": 1.328125,
"learning_rate": 0.00025149002915805384,
"loss": 0.2575,
"retrieval/pos_neg_gap_layer35": 0.7914715334773064,
"step": 948
},
{
"epoch": 0.8592123132639203,
"grad_norm": 0.84375,
"learning_rate": 0.00025138205007223463,
"loss": 0.224,
"retrieval/pos_neg_gap_layer35": 0.8096958175301552,
"step": 949
},
{
"epoch": 0.8601177003168855,
"grad_norm": 0.77734375,
"learning_rate": 0.0002512739741811845,
"loss": 0.2097,
"retrieval/pos_neg_gap_layer35": 0.8225469812750816,
"step": 950
},
{
"epoch": 0.8610230873698506,
"grad_norm": 0.80078125,
"learning_rate": 0.0002511658015881003,
"loss": 0.2423,
"retrieval/pos_neg_gap_layer35": 0.8017240315675735,
"step": 951
},
{
"epoch": 0.8619284744228157,
"grad_norm": 0.9296875,
"learning_rate": 0.0002510575323962713,
"loss": 0.2868,
"retrieval/pos_neg_gap_layer35": 0.7798621132969856,
"step": 952
},
{
"epoch": 0.8628338614757809,
"grad_norm": 0.703125,
"learning_rate": 0.0002509491667090789,
"loss": 0.2209,
"retrieval/pos_neg_gap_layer35": 0.8111787438392639,
"step": 953
},
{
"epoch": 0.8637392485287461,
"grad_norm": 1.28125,
"learning_rate": 0.0002508407046299968,
"loss": 0.2639,
"retrieval/pos_neg_gap_layer35": 0.7841088175773621,
"step": 954
},
{
"epoch": 0.8646446355817112,
"grad_norm": 1.1171875,
"learning_rate": 0.0002507321462625906,
"loss": 0.2139,
"retrieval/pos_neg_gap_layer35": 0.8107197359204292,
"step": 955
},
{
"epoch": 0.8655500226346763,
"grad_norm": 0.71484375,
"learning_rate": 0.0002506234917105178,
"loss": 0.1991,
"retrieval/pos_neg_gap_layer35": 0.8198460191488266,
"step": 956
},
{
"epoch": 0.8664554096876415,
"grad_norm": 1.4609375,
"learning_rate": 0.0002505147410775279,
"loss": 0.2564,
"retrieval/pos_neg_gap_layer35": 0.7866636589169502,
"step": 957
},
{
"epoch": 0.8673607967406066,
"grad_norm": 1.8671875,
"learning_rate": 0.00025040589446746205,
"loss": 0.2915,
"retrieval/pos_neg_gap_layer35": 0.7707198858261108,
"step": 958
},
{
"epoch": 0.8682661837935718,
"grad_norm": 1.6328125,
"learning_rate": 0.000250296951984253,
"loss": 0.342,
"retrieval/pos_neg_gap_layer35": 0.7583250477910042,
"step": 959
},
{
"epoch": 0.8691715708465368,
"grad_norm": 1.0625,
"learning_rate": 0.0002501879137319252,
"loss": 0.2161,
"retrieval/pos_neg_gap_layer35": 0.8148963078856468,
"step": 960
},
{
"epoch": 0.870076957899502,
"grad_norm": 1.828125,
"learning_rate": 0.00025007877981459434,
"loss": 0.2339,
"retrieval/pos_neg_gap_layer35": 0.8054594323039055,
"step": 961
},
{
"epoch": 0.8709823449524672,
"grad_norm": 0.9140625,
"learning_rate": 0.0002499695503364677,
"loss": 0.2557,
"retrieval/pos_neg_gap_layer35": 0.786057636141777,
"step": 962
},
{
"epoch": 0.8718877320054323,
"grad_norm": 1.25,
"learning_rate": 0.00024986022540184354,
"loss": 0.2697,
"retrieval/pos_neg_gap_layer35": 0.793047197163105,
"step": 963
},
{
"epoch": 0.8727931190583975,
"grad_norm": 0.859375,
"learning_rate": 0.00024975080511511154,
"loss": 0.2268,
"retrieval/pos_neg_gap_layer35": 0.8046604320406914,
"step": 964
},
{
"epoch": 0.8736985061113626,
"grad_norm": 0.77734375,
"learning_rate": 0.00024964128958075216,
"loss": 0.2394,
"retrieval/pos_neg_gap_layer35": 0.808429129421711,
"step": 965
},
{
"epoch": 0.8746038931643277,
"grad_norm": 0.6171875,
"learning_rate": 0.00024953167890333697,
"loss": 0.2098,
"retrieval/pos_neg_gap_layer35": 0.8192198127508163,
"step": 966
},
{
"epoch": 0.8755092802172929,
"grad_norm": 1.6796875,
"learning_rate": 0.00024942197318752834,
"loss": 0.2592,
"retrieval/pos_neg_gap_layer35": 0.78922089189291,
"step": 967
},
{
"epoch": 0.8764146672702581,
"grad_norm": 1.34375,
"learning_rate": 0.0002493121725380795,
"loss": 0.2992,
"retrieval/pos_neg_gap_layer35": 0.7702390253543854,
"step": 968
},
{
"epoch": 0.8773200543232231,
"grad_norm": 2.0625,
"learning_rate": 0.0002492022770598341,
"loss": 0.3289,
"retrieval/pos_neg_gap_layer35": 0.7490772530436516,
"step": 969
},
{
"epoch": 0.8782254413761883,
"grad_norm": 1.0546875,
"learning_rate": 0.0002490922868577265,
"loss": 0.2331,
"retrieval/pos_neg_gap_layer35": 0.8054832220077515,
"step": 970
},
{
"epoch": 0.8791308284291535,
"grad_norm": 0.76953125,
"learning_rate": 0.00024898220203678153,
"loss": 0.2149,
"retrieval/pos_neg_gap_layer35": 0.8208700343966484,
"step": 971
},
{
"epoch": 0.8800362154821186,
"grad_norm": 1.109375,
"learning_rate": 0.00024887202270211425,
"loss": 0.2631,
"retrieval/pos_neg_gap_layer35": 0.7888341471552849,
"step": 972
},
{
"epoch": 0.8809416025350838,
"grad_norm": 0.90234375,
"learning_rate": 0.0002487617489589301,
"loss": 0.2245,
"retrieval/pos_neg_gap_layer35": 0.8110646679997444,
"step": 973
},
{
"epoch": 0.8818469895880489,
"grad_norm": 1.3203125,
"learning_rate": 0.00024865138091252447,
"loss": 0.2889,
"retrieval/pos_neg_gap_layer35": 0.7819212302565575,
"step": 974
},
{
"epoch": 0.882752376641014,
"grad_norm": 0.796875,
"learning_rate": 0.00024854091866828305,
"loss": 0.2204,
"retrieval/pos_neg_gap_layer35": 0.8170958906412125,
"step": 975
},
{
"epoch": 0.8836577636939792,
"grad_norm": 1.9765625,
"learning_rate": 0.0002484303623316813,
"loss": 0.3003,
"retrieval/pos_neg_gap_layer35": 0.7705280780792236,
"step": 976
},
{
"epoch": 0.8845631507469444,
"grad_norm": 0.7734375,
"learning_rate": 0.0002483197120082846,
"loss": 0.2312,
"retrieval/pos_neg_gap_layer35": 0.8062387406826019,
"step": 977
},
{
"epoch": 0.8854685377999094,
"grad_norm": 1.8671875,
"learning_rate": 0.00024820896780374793,
"loss": 0.2474,
"retrieval/pos_neg_gap_layer35": 0.7935343310236931,
"step": 978
},
{
"epoch": 0.8863739248528746,
"grad_norm": 1.140625,
"learning_rate": 0.0002480981298238162,
"loss": 0.283,
"retrieval/pos_neg_gap_layer35": 0.7839161157608032,
"step": 979
},
{
"epoch": 0.8872793119058398,
"grad_norm": 0.66015625,
"learning_rate": 0.0002479871981743235,
"loss": 0.1971,
"retrieval/pos_neg_gap_layer35": 0.826920635998249,
"step": 980
},
{
"epoch": 0.8881846989588049,
"grad_norm": 1.5859375,
"learning_rate": 0.0002478761729611937,
"loss": 0.2963,
"retrieval/pos_neg_gap_layer35": 0.7682082876563072,
"step": 981
},
{
"epoch": 0.8890900860117701,
"grad_norm": 0.578125,
"learning_rate": 0.00024776505429043987,
"loss": 0.2101,
"retrieval/pos_neg_gap_layer35": 0.8158812075853348,
"step": 982
},
{
"epoch": 0.8899954730647351,
"grad_norm": 0.76953125,
"learning_rate": 0.0002476538422681642,
"loss": 0.249,
"retrieval/pos_neg_gap_layer35": 0.8046075403690338,
"step": 983
},
{
"epoch": 0.8909008601177003,
"grad_norm": 0.984375,
"learning_rate": 0.0002475425370005581,
"loss": 0.2753,
"retrieval/pos_neg_gap_layer35": 0.7882667854428291,
"step": 984
},
{
"epoch": 0.8918062471706655,
"grad_norm": 1.5390625,
"learning_rate": 0.00024743113859390225,
"loss": 0.3586,
"retrieval/pos_neg_gap_layer35": 0.75099878013134,
"step": 985
},
{
"epoch": 0.8927116342236306,
"grad_norm": 1.140625,
"learning_rate": 0.00024731964715456585,
"loss": 0.3276,
"retrieval/pos_neg_gap_layer35": 0.7522145956754684,
"step": 986
},
{
"epoch": 0.8936170212765957,
"grad_norm": 1.328125,
"learning_rate": 0.00024720806278900724,
"loss": 0.2891,
"retrieval/pos_neg_gap_layer35": 0.7677125707268715,
"step": 987
},
{
"epoch": 0.8945224083295609,
"grad_norm": 0.7265625,
"learning_rate": 0.00024709638560377334,
"loss": 0.2336,
"retrieval/pos_neg_gap_layer35": 0.8017161190509796,
"step": 988
},
{
"epoch": 0.895427795382526,
"grad_norm": 1.1640625,
"learning_rate": 0.0002469846157054998,
"loss": 0.2511,
"retrieval/pos_neg_gap_layer35": 0.8000732958316803,
"step": 989
},
{
"epoch": 0.8963331824354912,
"grad_norm": 2.09375,
"learning_rate": 0.0002468727532009107,
"loss": 0.2916,
"retrieval/pos_neg_gap_layer35": 0.7691463455557823,
"step": 990
},
{
"epoch": 0.8972385694884563,
"grad_norm": 0.796875,
"learning_rate": 0.0002467607981968186,
"loss": 0.1947,
"retrieval/pos_neg_gap_layer35": 0.828714907169342,
"step": 991
},
{
"epoch": 0.8981439565414214,
"grad_norm": 0.8671875,
"learning_rate": 0.0002466487508001244,
"loss": 0.2301,
"retrieval/pos_neg_gap_layer35": 0.8040446117520332,
"step": 992
},
{
"epoch": 0.8990493435943866,
"grad_norm": 0.6953125,
"learning_rate": 0.0002465366111178172,
"loss": 0.1849,
"retrieval/pos_neg_gap_layer35": 0.8361316546797752,
"step": 993
},
{
"epoch": 0.8999547306473518,
"grad_norm": 1.265625,
"learning_rate": 0.0002464243792569742,
"loss": 0.2457,
"retrieval/pos_neg_gap_layer35": 0.7937345951795578,
"step": 994
},
{
"epoch": 0.9008601177003169,
"grad_norm": 0.92578125,
"learning_rate": 0.0002463120553247606,
"loss": 0.2198,
"retrieval/pos_neg_gap_layer35": 0.809515506029129,
"step": 995
},
{
"epoch": 0.901765504753282,
"grad_norm": 1.5,
"learning_rate": 0.0002461996394284297,
"loss": 0.3211,
"retrieval/pos_neg_gap_layer35": 0.7533524036407471,
"step": 996
},
{
"epoch": 0.9026708918062472,
"grad_norm": 1.09375,
"learning_rate": 0.00024608713167532226,
"loss": 0.2841,
"retrieval/pos_neg_gap_layer35": 0.7767128497362137,
"step": 997
},
{
"epoch": 0.9035762788592123,
"grad_norm": 0.625,
"learning_rate": 0.00024597453217286707,
"loss": 0.2438,
"retrieval/pos_neg_gap_layer35": 0.8083752021193504,
"step": 998
},
{
"epoch": 0.9044816659121775,
"grad_norm": 0.734375,
"learning_rate": 0.0002458618410285804,
"loss": 0.2301,
"retrieval/pos_neg_gap_layer35": 0.8134948164224625,
"step": 999
},
{
"epoch": 0.9053870529651425,
"grad_norm": 1.1875,
"learning_rate": 0.00024574905835006596,
"loss": 0.2928,
"retrieval/pos_neg_gap_layer35": 0.7761303260922432,
"step": 1000
},
{
"epoch": 0.9062924400181077,
"grad_norm": 1.0234375,
"learning_rate": 0.00024563618424501506,
"loss": 0.3168,
"retrieval/pos_neg_gap_layer35": 0.7696171626448631,
"step": 1001
},
{
"epoch": 0.9071978270710729,
"grad_norm": 1.1328125,
"learning_rate": 0.00024552321882120604,
"loss": 0.2416,
"retrieval/pos_neg_gap_layer35": 0.7994090393185616,
"step": 1002
},
{
"epoch": 0.908103214124038,
"grad_norm": 0.95703125,
"learning_rate": 0.00024541016218650473,
"loss": 0.2638,
"retrieval/pos_neg_gap_layer35": 0.781087338924408,
"step": 1003
},
{
"epoch": 0.9090086011770032,
"grad_norm": 0.79296875,
"learning_rate": 0.0002452970144488638,
"loss": 0.2374,
"retrieval/pos_neg_gap_layer35": 0.8026145547628403,
"step": 1004
},
{
"epoch": 0.9099139882299683,
"grad_norm": 1.0703125,
"learning_rate": 0.0002451837757163231,
"loss": 0.3587,
"retrieval/pos_neg_gap_layer35": 0.7399043366312981,
"step": 1005
},
{
"epoch": 0.9108193752829334,
"grad_norm": 1.2421875,
"learning_rate": 0.0002450704460970092,
"loss": 0.222,
"retrieval/pos_neg_gap_layer35": 0.8084927052259445,
"step": 1006
},
{
"epoch": 0.9117247623358986,
"grad_norm": 1.109375,
"learning_rate": 0.0002449570256991356,
"loss": 0.3526,
"retrieval/pos_neg_gap_layer35": 0.757655419409275,
"step": 1007
},
{
"epoch": 0.9126301493888638,
"grad_norm": 1.34375,
"learning_rate": 0.0002448435146310024,
"loss": 0.2589,
"retrieval/pos_neg_gap_layer35": 0.7890852391719818,
"step": 1008
},
{
"epoch": 0.9135355364418288,
"grad_norm": 0.671875,
"learning_rate": 0.0002447299130009962,
"loss": 0.2436,
"retrieval/pos_neg_gap_layer35": 0.7984486147761345,
"step": 1009
},
{
"epoch": 0.914440923494794,
"grad_norm": 1.0625,
"learning_rate": 0.0002446162209175903,
"loss": 0.2328,
"retrieval/pos_neg_gap_layer35": 0.8044926673173904,
"step": 1010
},
{
"epoch": 0.9153463105477592,
"grad_norm": 0.90625,
"learning_rate": 0.00024450243848934413,
"loss": 0.2656,
"retrieval/pos_neg_gap_layer35": 0.7824936211109161,
"step": 1011
},
{
"epoch": 0.9162516976007243,
"grad_norm": 0.83984375,
"learning_rate": 0.00024438856582490365,
"loss": 0.2192,
"retrieval/pos_neg_gap_layer35": 0.8096484243869781,
"step": 1012
},
{
"epoch": 0.9171570846536895,
"grad_norm": 0.78125,
"learning_rate": 0.0002442746030330006,
"loss": 0.252,
"retrieval/pos_neg_gap_layer35": 0.7951190397143364,
"step": 1013
},
{
"epoch": 0.9180624717066546,
"grad_norm": 0.59375,
"learning_rate": 0.0002441605502224532,
"loss": 0.2397,
"retrieval/pos_neg_gap_layer35": 0.8010373562574387,
"step": 1014
},
{
"epoch": 0.9189678587596197,
"grad_norm": 0.875,
"learning_rate": 0.00024404640750216533,
"loss": 0.2754,
"retrieval/pos_neg_gap_layer35": 0.7778530344367027,
"step": 1015
},
{
"epoch": 0.9198732458125849,
"grad_norm": 0.7734375,
"learning_rate": 0.00024393217498112687,
"loss": 0.2131,
"retrieval/pos_neg_gap_layer35": 0.8125851005315781,
"step": 1016
},
{
"epoch": 0.9207786328655501,
"grad_norm": 1.484375,
"learning_rate": 0.00024381785276841338,
"loss": 0.328,
"retrieval/pos_neg_gap_layer35": 0.7676214352250099,
"step": 1017
},
{
"epoch": 0.9216840199185151,
"grad_norm": 1.25,
"learning_rate": 0.000243703440973186,
"loss": 0.2416,
"retrieval/pos_neg_gap_layer35": 0.7968010827898979,
"step": 1018
},
{
"epoch": 0.9225894069714803,
"grad_norm": 0.8203125,
"learning_rate": 0.00024358893970469162,
"loss": 0.2852,
"retrieval/pos_neg_gap_layer35": 0.7854959294199944,
"step": 1019
},
{
"epoch": 0.9234947940244455,
"grad_norm": 0.87890625,
"learning_rate": 0.00024347434907226234,
"loss": 0.2463,
"retrieval/pos_neg_gap_layer35": 0.7948428839445114,
"step": 1020
},
{
"epoch": 0.9244001810774106,
"grad_norm": 1.046875,
"learning_rate": 0.0002433596691853157,
"loss": 0.2692,
"retrieval/pos_neg_gap_layer35": 0.7844233512878418,
"step": 1021
},
{
"epoch": 0.9253055681303758,
"grad_norm": 1.015625,
"learning_rate": 0.00024324490015335451,
"loss": 0.3028,
"retrieval/pos_neg_gap_layer35": 0.7612828239798546,
"step": 1022
},
{
"epoch": 0.9262109551833408,
"grad_norm": 1.078125,
"learning_rate": 0.00024313004208596656,
"loss": 0.249,
"retrieval/pos_neg_gap_layer35": 0.8040665984153748,
"step": 1023
},
{
"epoch": 0.927116342236306,
"grad_norm": 0.8984375,
"learning_rate": 0.00024301509509282474,
"loss": 0.2471,
"retrieval/pos_neg_gap_layer35": 0.7934447228908539,
"step": 1024
},
{
"epoch": 0.9280217292892712,
"grad_norm": 0.91796875,
"learning_rate": 0.0002429000592836869,
"loss": 0.2192,
"retrieval/pos_neg_gap_layer35": 0.8096957430243492,
"step": 1025
},
{
"epoch": 0.9289271163422363,
"grad_norm": 1.6796875,
"learning_rate": 0.00024278493476839557,
"loss": 0.2532,
"retrieval/pos_neg_gap_layer35": 0.7925464063882828,
"step": 1026
},
{
"epoch": 0.9298325033952014,
"grad_norm": 1.2578125,
"learning_rate": 0.00024266972165687807,
"loss": 0.2689,
"retrieval/pos_neg_gap_layer35": 0.7839167192578316,
"step": 1027
},
{
"epoch": 0.9307378904481666,
"grad_norm": 1.6171875,
"learning_rate": 0.00024255442005914634,
"loss": 0.316,
"retrieval/pos_neg_gap_layer35": 0.7535669580101967,
"step": 1028
},
{
"epoch": 0.9316432775011317,
"grad_norm": 2.015625,
"learning_rate": 0.00024243903008529678,
"loss": 0.3462,
"retrieval/pos_neg_gap_layer35": 0.7483504042029381,
"step": 1029
},
{
"epoch": 0.9325486645540969,
"grad_norm": 1.8359375,
"learning_rate": 0.0002423235518455101,
"loss": 0.3492,
"retrieval/pos_neg_gap_layer35": 0.7414084002375603,
"step": 1030
},
{
"epoch": 0.9334540516070621,
"grad_norm": 1.109375,
"learning_rate": 0.00024220798545005135,
"loss": 0.2419,
"retrieval/pos_neg_gap_layer35": 0.7989005818963051,
"step": 1031
},
{
"epoch": 0.9343594386600271,
"grad_norm": 0.90234375,
"learning_rate": 0.00024209233100926987,
"loss": 0.2675,
"retrieval/pos_neg_gap_layer35": 0.7933980748057365,
"step": 1032
},
{
"epoch": 0.9352648257129923,
"grad_norm": 1.5078125,
"learning_rate": 0.00024197658863359887,
"loss": 0.2507,
"retrieval/pos_neg_gap_layer35": 0.7955424338579178,
"step": 1033
},
{
"epoch": 0.9361702127659575,
"grad_norm": 0.8671875,
"learning_rate": 0.00024186075843355567,
"loss": 0.1915,
"retrieval/pos_neg_gap_layer35": 0.829775907099247,
"step": 1034
},
{
"epoch": 0.9370755998189226,
"grad_norm": 0.9296875,
"learning_rate": 0.00024174484051974138,
"loss": 0.2219,
"retrieval/pos_neg_gap_layer35": 0.8063685148954391,
"step": 1035
},
{
"epoch": 0.9379809868718877,
"grad_norm": 1.0546875,
"learning_rate": 0.00024162883500284088,
"loss": 0.2842,
"retrieval/pos_neg_gap_layer35": 0.7720093727111816,
"step": 1036
},
{
"epoch": 0.9388863739248529,
"grad_norm": 1.34375,
"learning_rate": 0.00024151274199362265,
"loss": 0.2258,
"retrieval/pos_neg_gap_layer35": 0.804417110979557,
"step": 1037
},
{
"epoch": 0.939791760977818,
"grad_norm": 1.3203125,
"learning_rate": 0.0002413965616029389,
"loss": 0.2555,
"retrieval/pos_neg_gap_layer35": 0.7887859493494034,
"step": 1038
},
{
"epoch": 0.9406971480307832,
"grad_norm": 0.72265625,
"learning_rate": 0.000241280293941725,
"loss": 0.2173,
"retrieval/pos_neg_gap_layer35": 0.8091920614242554,
"step": 1039
},
{
"epoch": 0.9416025350837482,
"grad_norm": 0.75390625,
"learning_rate": 0.00024116393912099993,
"loss": 0.2053,
"retrieval/pos_neg_gap_layer35": 0.8142609223723412,
"step": 1040
},
{
"epoch": 0.9425079221367134,
"grad_norm": 1.53125,
"learning_rate": 0.00024104749725186562,
"loss": 0.2079,
"retrieval/pos_neg_gap_layer35": 0.8196741342544556,
"step": 1041
},
{
"epoch": 0.9434133091896786,
"grad_norm": 1.1953125,
"learning_rate": 0.00024093096844550734,
"loss": 0.2766,
"retrieval/pos_neg_gap_layer35": 0.791561096906662,
"step": 1042
},
{
"epoch": 0.9443186962426438,
"grad_norm": 0.7421875,
"learning_rate": 0.0002408143528131932,
"loss": 0.2317,
"retrieval/pos_neg_gap_layer35": 0.8067691922187805,
"step": 1043
},
{
"epoch": 0.9452240832956089,
"grad_norm": 0.8984375,
"learning_rate": 0.00024069765046627443,
"loss": 0.274,
"retrieval/pos_neg_gap_layer35": 0.7875049188733101,
"step": 1044
},
{
"epoch": 0.946129470348574,
"grad_norm": 0.98828125,
"learning_rate": 0.0002405808615161849,
"loss": 0.2337,
"retrieval/pos_neg_gap_layer35": 0.8071285411715508,
"step": 1045
},
{
"epoch": 0.9470348574015391,
"grad_norm": 1.1640625,
"learning_rate": 0.0002404639860744411,
"loss": 0.2552,
"retrieval/pos_neg_gap_layer35": 0.7892273962497711,
"step": 1046
},
{
"epoch": 0.9479402444545043,
"grad_norm": 1.0390625,
"learning_rate": 0.00024034702425264237,
"loss": 0.2504,
"retrieval/pos_neg_gap_layer35": 0.7992671504616737,
"step": 1047
},
{
"epoch": 0.9488456315074695,
"grad_norm": 1.109375,
"learning_rate": 0.00024022997616247025,
"loss": 0.225,
"retrieval/pos_neg_gap_layer35": 0.8117284551262856,
"step": 1048
},
{
"epoch": 0.9497510185604345,
"grad_norm": 0.81640625,
"learning_rate": 0.00024011284191568884,
"loss": 0.2323,
"retrieval/pos_neg_gap_layer35": 0.8015414252877235,
"step": 1049
},
{
"epoch": 0.9506564056133997,
"grad_norm": 0.984375,
"learning_rate": 0.00023999562162414442,
"loss": 0.2665,
"retrieval/pos_neg_gap_layer35": 0.7895231992006302,
"step": 1050
},
{
"epoch": 0.9515617926663649,
"grad_norm": 1.40625,
"learning_rate": 0.0002398783153997655,
"loss": 0.25,
"retrieval/pos_neg_gap_layer35": 0.7909038290381432,
"step": 1051
},
{
"epoch": 0.95246717971933,
"grad_norm": 0.96875,
"learning_rate": 0.00023976092335456256,
"loss": 0.2148,
"retrieval/pos_neg_gap_layer35": 0.8117726370692253,
"step": 1052
},
{
"epoch": 0.9533725667722952,
"grad_norm": 0.75390625,
"learning_rate": 0.0002396434456006281,
"loss": 0.2217,
"retrieval/pos_neg_gap_layer35": 0.8146641999483109,
"step": 1053
},
{
"epoch": 0.9542779538252603,
"grad_norm": 1.3203125,
"learning_rate": 0.00023952588225013634,
"loss": 0.3371,
"retrieval/pos_neg_gap_layer35": 0.7596519812941551,
"step": 1054
},
{
"epoch": 0.9551833408782254,
"grad_norm": 1.0546875,
"learning_rate": 0.00023940823341534343,
"loss": 0.2694,
"retrieval/pos_neg_gap_layer35": 0.7805606946349144,
"step": 1055
},
{
"epoch": 0.9560887279311906,
"grad_norm": 1.21875,
"learning_rate": 0.00023929049920858698,
"loss": 0.2271,
"retrieval/pos_neg_gap_layer35": 0.8084366023540497,
"step": 1056
},
{
"epoch": 0.9569941149841558,
"grad_norm": 0.9765625,
"learning_rate": 0.00023917267974228615,
"loss": 0.2223,
"retrieval/pos_neg_gap_layer35": 0.814299687743187,
"step": 1057
},
{
"epoch": 0.9578995020371208,
"grad_norm": 0.83984375,
"learning_rate": 0.00023905477512894155,
"loss": 0.2979,
"retrieval/pos_neg_gap_layer35": 0.7661475986242294,
"step": 1058
},
{
"epoch": 0.958804889090086,
"grad_norm": 0.80078125,
"learning_rate": 0.00023893678548113516,
"loss": 0.2775,
"retrieval/pos_neg_gap_layer35": 0.7816267684102058,
"step": 1059
},
{
"epoch": 0.9597102761430512,
"grad_norm": 1.09375,
"learning_rate": 0.0002388187109115299,
"loss": 0.2731,
"retrieval/pos_neg_gap_layer35": 0.7876512184739113,
"step": 1060
},
{
"epoch": 0.9606156631960163,
"grad_norm": 0.80859375,
"learning_rate": 0.0002387005515328701,
"loss": 0.2669,
"retrieval/pos_neg_gap_layer35": 0.7844613716006279,
"step": 1061
},
{
"epoch": 0.9615210502489815,
"grad_norm": 1.1796875,
"learning_rate": 0.00023858230745798083,
"loss": 0.2428,
"retrieval/pos_neg_gap_layer35": 0.7970415353775024,
"step": 1062
},
{
"epoch": 0.9624264373019465,
"grad_norm": 0.92578125,
"learning_rate": 0.0002384639787997682,
"loss": 0.2168,
"retrieval/pos_neg_gap_layer35": 0.8126402944326401,
"step": 1063
},
{
"epoch": 0.9633318243549117,
"grad_norm": 0.890625,
"learning_rate": 0.00023834556567121893,
"loss": 0.2334,
"retrieval/pos_neg_gap_layer35": 0.8052059933543205,
"step": 1064
},
{
"epoch": 0.9642372114078769,
"grad_norm": 1.203125,
"learning_rate": 0.0002382270681854005,
"loss": 0.3192,
"retrieval/pos_neg_gap_layer35": 0.7662748098373413,
"step": 1065
},
{
"epoch": 0.965142598460842,
"grad_norm": 0.921875,
"learning_rate": 0.0002381084864554609,
"loss": 0.2504,
"retrieval/pos_neg_gap_layer35": 0.7988310679793358,
"step": 1066
},
{
"epoch": 0.9660479855138071,
"grad_norm": 0.84375,
"learning_rate": 0.00023798982059462863,
"loss": 0.2353,
"retrieval/pos_neg_gap_layer35": 0.7999881058931351,
"step": 1067
},
{
"epoch": 0.9669533725667723,
"grad_norm": 1.1640625,
"learning_rate": 0.00023787107071621239,
"loss": 0.2298,
"retrieval/pos_neg_gap_layer35": 0.8047187626361847,
"step": 1068
},
{
"epoch": 0.9678587596197374,
"grad_norm": 1.3203125,
"learning_rate": 0.00023775223693360125,
"loss": 0.2904,
"retrieval/pos_neg_gap_layer35": 0.7723774686455727,
"step": 1069
},
{
"epoch": 0.9687641466727026,
"grad_norm": 1.09375,
"learning_rate": 0.00023763331936026426,
"loss": 0.2336,
"retrieval/pos_neg_gap_layer35": 0.8043026030063629,
"step": 1070
},
{
"epoch": 0.9696695337256678,
"grad_norm": 1.265625,
"learning_rate": 0.00023751431810975062,
"loss": 0.2343,
"retrieval/pos_neg_gap_layer35": 0.8013731390237808,
"step": 1071
},
{
"epoch": 0.9705749207786328,
"grad_norm": 1.515625,
"learning_rate": 0.00023739523329568933,
"loss": 0.2379,
"retrieval/pos_neg_gap_layer35": 0.8011744990944862,
"step": 1072
},
{
"epoch": 0.971480307831598,
"grad_norm": 1.3046875,
"learning_rate": 0.0002372760650317891,
"loss": 0.2661,
"retrieval/pos_neg_gap_layer35": 0.7876545116305351,
"step": 1073
},
{
"epoch": 0.9723856948845632,
"grad_norm": 0.734375,
"learning_rate": 0.00023715681343183863,
"loss": 0.2241,
"retrieval/pos_neg_gap_layer35": 0.8045293688774109,
"step": 1074
},
{
"epoch": 0.9732910819375283,
"grad_norm": 0.77734375,
"learning_rate": 0.00023703747860970582,
"loss": 0.2774,
"retrieval/pos_neg_gap_layer35": 0.7876648306846619,
"step": 1075
},
{
"epoch": 0.9741964689904934,
"grad_norm": 0.83203125,
"learning_rate": 0.00023691806067933831,
"loss": 0.2376,
"retrieval/pos_neg_gap_layer35": 0.8056944906711578,
"step": 1076
},
{
"epoch": 0.9751018560434586,
"grad_norm": 1.46875,
"learning_rate": 0.00023679855975476295,
"loss": 0.2728,
"retrieval/pos_neg_gap_layer35": 0.7801089212298393,
"step": 1077
},
{
"epoch": 0.9760072430964237,
"grad_norm": 0.953125,
"learning_rate": 0.00023667897595008587,
"loss": 0.2512,
"retrieval/pos_neg_gap_layer35": 0.7894560471177101,
"step": 1078
},
{
"epoch": 0.9769126301493889,
"grad_norm": 0.6953125,
"learning_rate": 0.00023655930937949238,
"loss": 0.2439,
"retrieval/pos_neg_gap_layer35": 0.8021877631545067,
"step": 1079
},
{
"epoch": 0.9778180172023541,
"grad_norm": 0.96875,
"learning_rate": 0.00023643956015724678,
"loss": 0.2979,
"retrieval/pos_neg_gap_layer35": 0.7732494473457336,
"step": 1080
},
{
"epoch": 0.9787234042553191,
"grad_norm": 0.96875,
"learning_rate": 0.00023631972839769229,
"loss": 0.2586,
"retrieval/pos_neg_gap_layer35": 0.7902258113026619,
"step": 1081
},
{
"epoch": 0.9796287913082843,
"grad_norm": 1.03125,
"learning_rate": 0.00023619981421525095,
"loss": 0.3189,
"retrieval/pos_neg_gap_layer35": 0.7587589547038078,
"step": 1082
},
{
"epoch": 0.9805341783612495,
"grad_norm": 0.98828125,
"learning_rate": 0.0002360798177244236,
"loss": 0.3212,
"retrieval/pos_neg_gap_layer35": 0.7602540403604507,
"step": 1083
},
{
"epoch": 0.9814395654142146,
"grad_norm": 0.75,
"learning_rate": 0.0002359597390397894,
"loss": 0.2138,
"retrieval/pos_neg_gap_layer35": 0.8161752745509148,
"step": 1084
},
{
"epoch": 0.9823449524671797,
"grad_norm": 1.0234375,
"learning_rate": 0.00023583957827600627,
"loss": 0.2316,
"retrieval/pos_neg_gap_layer35": 0.8030286207795143,
"step": 1085
},
{
"epoch": 0.9832503395201448,
"grad_norm": 0.796875,
"learning_rate": 0.00023571933554781044,
"loss": 0.2602,
"retrieval/pos_neg_gap_layer35": 0.7874865904450417,
"step": 1086
},
{
"epoch": 0.98415572657311,
"grad_norm": 1.109375,
"learning_rate": 0.0002355990109700163,
"loss": 0.2712,
"retrieval/pos_neg_gap_layer35": 0.7793728858232498,
"step": 1087
},
{
"epoch": 0.9850611136260752,
"grad_norm": 1.5546875,
"learning_rate": 0.00023547860465751648,
"loss": 0.2615,
"retrieval/pos_neg_gap_layer35": 0.7819217294454575,
"step": 1088
},
{
"epoch": 0.9859665006790402,
"grad_norm": 1.2578125,
"learning_rate": 0.00023535811672528164,
"loss": 0.2923,
"retrieval/pos_neg_gap_layer35": 0.7694869115948677,
"step": 1089
},
{
"epoch": 0.9868718877320054,
"grad_norm": 0.66796875,
"learning_rate": 0.0002352375472883604,
"loss": 0.2146,
"retrieval/pos_neg_gap_layer35": 0.8132204115390778,
"step": 1090
},
{
"epoch": 0.9877772747849706,
"grad_norm": 1.2734375,
"learning_rate": 0.0002351168964618791,
"loss": 0.2711,
"retrieval/pos_neg_gap_layer35": 0.7849212735891342,
"step": 1091
},
{
"epoch": 0.9886826618379357,
"grad_norm": 0.9765625,
"learning_rate": 0.00023499616436104196,
"loss": 0.2621,
"retrieval/pos_neg_gap_layer35": 0.793477900326252,
"step": 1092
},
{
"epoch": 0.9895880488909009,
"grad_norm": 1.3046875,
"learning_rate": 0.00023487535110113065,
"loss": 0.2342,
"retrieval/pos_neg_gap_layer35": 0.800565280020237,
"step": 1093
},
{
"epoch": 0.990493435943866,
"grad_norm": 1.1484375,
"learning_rate": 0.0002347544567975045,
"loss": 0.2594,
"retrieval/pos_neg_gap_layer35": 0.7821277529001236,
"step": 1094
},
{
"epoch": 0.9913988229968311,
"grad_norm": 0.8515625,
"learning_rate": 0.0002346334815656,
"loss": 0.2619,
"retrieval/pos_neg_gap_layer35": 0.7906242087483406,
"step": 1095
},
{
"epoch": 0.9923042100497963,
"grad_norm": 0.65625,
"learning_rate": 0.0002345124255209311,
"loss": 0.2341,
"retrieval/pos_neg_gap_layer35": 0.8005865514278412,
"step": 1096
},
{
"epoch": 0.9932095971027615,
"grad_norm": 1.375,
"learning_rate": 0.00023439128877908886,
"loss": 0.283,
"retrieval/pos_neg_gap_layer35": 0.7740660235285759,
"step": 1097
},
{
"epoch": 0.9941149841557265,
"grad_norm": 1.6484375,
"learning_rate": 0.0002342700714557414,
"loss": 0.3048,
"retrieval/pos_neg_gap_layer35": 0.7623828798532486,
"step": 1098
},
{
"epoch": 0.9950203712086917,
"grad_norm": 1.265625,
"learning_rate": 0.0002341487736666338,
"loss": 0.3056,
"retrieval/pos_neg_gap_layer35": 0.7716804295778275,
"step": 1099
},
{
"epoch": 0.9959257582616569,
"grad_norm": 0.84375,
"learning_rate": 0.00023402739552758788,
"loss": 0.2424,
"retrieval/pos_neg_gap_layer35": 0.8089796975255013,
"step": 1100
},
{
"epoch": 0.996831145314622,
"grad_norm": 1.359375,
"learning_rate": 0.0002339059371545023,
"loss": 0.3238,
"retrieval/pos_neg_gap_layer35": 0.7645340412855148,
"step": 1101
},
{
"epoch": 0.9977365323675872,
"grad_norm": 0.91015625,
"learning_rate": 0.0002337843986633523,
"loss": 0.289,
"retrieval/pos_neg_gap_layer35": 0.770873673260212,
"step": 1102
},
{
"epoch": 0.9986419194205522,
"grad_norm": 1.03125,
"learning_rate": 0.00023366278017018953,
"loss": 0.3406,
"retrieval/pos_neg_gap_layer35": 0.754501074552536,
"step": 1103
},
{
"epoch": 0.9995473064735174,
"grad_norm": 1.359375,
"learning_rate": 0.00023354108179114217,
"loss": 0.2723,
"retrieval/pos_neg_gap_layer35": 0.7776569947600365,
"step": 1104
},
{
"epoch": 1.0,
"grad_norm": 1.2421875,
"learning_rate": 0.00023341930364241458,
"loss": 0.287,
"retrieval/pos_neg_gap_layer35": 0.772953063249588,
"step": 1105
}
],
"logging_steps": 1.0,
"max_steps": 3315,
"num_input_tokens_seen": 0,
"num_train_epochs": 3,
"save_steps": 500,
"stateful_callbacks": {
"TrainerControl": {
"args": {
"should_epoch_stop": false,
"should_evaluate": false,
"should_log": false,
"should_save": true,
"should_training_stop": false
},
"attributes": {}
}
},
"total_flos": 3.305242693376929e+22,
"train_batch_size": 8,
"trial_name": null,
"trial_params": null
}