mikhail-panzo's picture
Training in progress, step 10500, checkpoint
3a4c441
{
"best_metric": 0.35496434569358826,
"best_model_checkpoint": "mikhail_panzo/zlm_b128_le5_s12000/checkpoint-10500",
"epoch": 17.59455497382199,
"eval_steps": 500,
"global_step": 10500,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.08376963350785341,
"grad_norm": 6.650441646575928,
"learning_rate": 2.4500000000000004e-07,
"loss": 1.1088,
"step": 50
},
{
"epoch": 0.16753926701570682,
"grad_norm": 13.608625411987305,
"learning_rate": 4.95e-07,
"loss": 1.1077,
"step": 100
},
{
"epoch": 0.2513089005235602,
"grad_norm": 2.7265217304229736,
"learning_rate": 7.450000000000001e-07,
"loss": 0.9759,
"step": 150
},
{
"epoch": 0.33507853403141363,
"grad_norm": 5.083759307861328,
"learning_rate": 9.950000000000002e-07,
"loss": 0.8705,
"step": 200
},
{
"epoch": 0.418848167539267,
"grad_norm": 1.8121678829193115,
"learning_rate": 1.2450000000000002e-06,
"loss": 0.8185,
"step": 250
},
{
"epoch": 0.5026178010471204,
"grad_norm": 2.275899648666382,
"learning_rate": 1.495e-06,
"loss": 0.784,
"step": 300
},
{
"epoch": 0.5863874345549738,
"grad_norm": 2.2861170768737793,
"learning_rate": 1.745e-06,
"loss": 0.7545,
"step": 350
},
{
"epoch": 0.6701570680628273,
"grad_norm": 3.1991238594055176,
"learning_rate": 1.9950000000000004e-06,
"loss": 0.7227,
"step": 400
},
{
"epoch": 0.7539267015706806,
"grad_norm": 1.843092679977417,
"learning_rate": 2.245e-06,
"loss": 0.6829,
"step": 450
},
{
"epoch": 0.837696335078534,
"grad_norm": 1.4700403213500977,
"learning_rate": 2.4950000000000003e-06,
"loss": 0.6644,
"step": 500
},
{
"epoch": 0.837696335078534,
"eval_loss": 0.5727962851524353,
"eval_runtime": 293.5061,
"eval_samples_per_second": 28.923,
"eval_steps_per_second": 3.618,
"step": 500
},
{
"epoch": 0.9214659685863874,
"grad_norm": 1.4362367391586304,
"learning_rate": 2.7450000000000004e-06,
"loss": 0.6487,
"step": 550
},
{
"epoch": 1.0052356020942408,
"grad_norm": 1.8555901050567627,
"learning_rate": 2.995e-06,
"loss": 0.6301,
"step": 600
},
{
"epoch": 1.0890052356020943,
"grad_norm": 1.9496757984161377,
"learning_rate": 3.2450000000000003e-06,
"loss": 0.6237,
"step": 650
},
{
"epoch": 1.1727748691099475,
"grad_norm": 3.0705084800720215,
"learning_rate": 3.495e-06,
"loss": 0.6156,
"step": 700
},
{
"epoch": 1.256544502617801,
"grad_norm": 2.2215688228607178,
"learning_rate": 3.745e-06,
"loss": 0.5945,
"step": 750
},
{
"epoch": 1.3403141361256545,
"grad_norm": 2.021375894546509,
"learning_rate": 3.995000000000001e-06,
"loss": 0.5891,
"step": 800
},
{
"epoch": 1.4240837696335078,
"grad_norm": 2.0354769229888916,
"learning_rate": 4.245e-06,
"loss": 0.5781,
"step": 850
},
{
"epoch": 1.5078534031413613,
"grad_norm": 1.8674426078796387,
"learning_rate": 4.495e-06,
"loss": 0.5677,
"step": 900
},
{
"epoch": 1.5916230366492146,
"grad_norm": 2.9264817237854004,
"learning_rate": 4.745e-06,
"loss": 0.5576,
"step": 950
},
{
"epoch": 1.675392670157068,
"grad_norm": 1.9513416290283203,
"learning_rate": 4.9950000000000005e-06,
"loss": 0.5594,
"step": 1000
},
{
"epoch": 1.675392670157068,
"eval_loss": 0.48248979449272156,
"eval_runtime": 294.2627,
"eval_samples_per_second": 28.848,
"eval_steps_per_second": 3.609,
"step": 1000
},
{
"epoch": 1.7591623036649215,
"grad_norm": 1.934941291809082,
"learning_rate": 5.245e-06,
"loss": 0.5571,
"step": 1050
},
{
"epoch": 1.8429319371727748,
"grad_norm": 1.7669196128845215,
"learning_rate": 5.495000000000001e-06,
"loss": 0.5477,
"step": 1100
},
{
"epoch": 1.9267015706806283,
"grad_norm": 3.954806089401245,
"learning_rate": 5.745000000000001e-06,
"loss": 0.5294,
"step": 1150
},
{
"epoch": 2.0104712041884816,
"grad_norm": 2.8323569297790527,
"learning_rate": 5.995000000000001e-06,
"loss": 0.5261,
"step": 1200
},
{
"epoch": 2.094240837696335,
"grad_norm": 2.193530797958374,
"learning_rate": 6.245000000000001e-06,
"loss": 0.5262,
"step": 1250
},
{
"epoch": 2.1780104712041886,
"grad_norm": 1.5814998149871826,
"learning_rate": 6.4950000000000005e-06,
"loss": 0.5204,
"step": 1300
},
{
"epoch": 2.261780104712042,
"grad_norm": 5.6610918045043945,
"learning_rate": 6.745000000000001e-06,
"loss": 0.5157,
"step": 1350
},
{
"epoch": 2.345549738219895,
"grad_norm": 9.156899452209473,
"learning_rate": 6.995000000000001e-06,
"loss": 0.5178,
"step": 1400
},
{
"epoch": 2.4293193717277486,
"grad_norm": 2.3170149326324463,
"learning_rate": 7.245000000000001e-06,
"loss": 0.5163,
"step": 1450
},
{
"epoch": 2.513089005235602,
"grad_norm": 2.1565325260162354,
"learning_rate": 7.495000000000001e-06,
"loss": 0.5042,
"step": 1500
},
{
"epoch": 2.513089005235602,
"eval_loss": 0.4464746415615082,
"eval_runtime": 293.8295,
"eval_samples_per_second": 28.891,
"eval_steps_per_second": 3.614,
"step": 1500
},
{
"epoch": 2.5997905759162303,
"grad_norm": 2.4758172035217285,
"learning_rate": 7.745e-06,
"loss": 0.5031,
"step": 1550
},
{
"epoch": 2.683560209424084,
"grad_norm": 2.1877381801605225,
"learning_rate": 7.990000000000001e-06,
"loss": 0.4994,
"step": 1600
},
{
"epoch": 2.7673298429319373,
"grad_norm": 2.4486210346221924,
"learning_rate": 8.24e-06,
"loss": 0.4965,
"step": 1650
},
{
"epoch": 2.8510994764397903,
"grad_norm": 2.596200704574585,
"learning_rate": 8.49e-06,
"loss": 0.4988,
"step": 1700
},
{
"epoch": 2.934869109947644,
"grad_norm": 1.7787096500396729,
"learning_rate": 8.740000000000001e-06,
"loss": 0.4947,
"step": 1750
},
{
"epoch": 3.0186387434554973,
"grad_norm": 2.09403133392334,
"learning_rate": 8.99e-06,
"loss": 0.4845,
"step": 1800
},
{
"epoch": 3.102408376963351,
"grad_norm": 2.0056636333465576,
"learning_rate": 9.240000000000001e-06,
"loss": 0.4833,
"step": 1850
},
{
"epoch": 3.1861780104712043,
"grad_norm": 3.200199842453003,
"learning_rate": 9.49e-06,
"loss": 0.484,
"step": 1900
},
{
"epoch": 3.269947643979058,
"grad_norm": 2.5462379455566406,
"learning_rate": 9.74e-06,
"loss": 0.4726,
"step": 1950
},
{
"epoch": 3.353717277486911,
"grad_norm": 1.3994622230529785,
"learning_rate": 9.990000000000001e-06,
"loss": 0.4795,
"step": 2000
},
{
"epoch": 3.353717277486911,
"eval_loss": 0.42617055773735046,
"eval_runtime": 269.3627,
"eval_samples_per_second": 31.515,
"eval_steps_per_second": 3.943,
"step": 2000
},
{
"epoch": 3.4374869109947643,
"grad_norm": 1.8979076147079468,
"learning_rate": 9.952e-06,
"loss": 0.4792,
"step": 2050
},
{
"epoch": 3.521256544502618,
"grad_norm": 1.4493324756622314,
"learning_rate": 9.902000000000001e-06,
"loss": 0.4751,
"step": 2100
},
{
"epoch": 3.6050261780104713,
"grad_norm": 3.3374929428100586,
"learning_rate": 9.852e-06,
"loss": 0.4736,
"step": 2150
},
{
"epoch": 3.6887958115183244,
"grad_norm": 3.5062992572784424,
"learning_rate": 9.802e-06,
"loss": 0.4675,
"step": 2200
},
{
"epoch": 3.772565445026178,
"grad_norm": 2.250505208969116,
"learning_rate": 9.752e-06,
"loss": 0.4611,
"step": 2250
},
{
"epoch": 3.8563350785340313,
"grad_norm": 1.793270468711853,
"learning_rate": 9.702e-06,
"loss": 0.4605,
"step": 2300
},
{
"epoch": 3.940104712041885,
"grad_norm": 1.663677453994751,
"learning_rate": 9.652e-06,
"loss": 0.4693,
"step": 2350
},
{
"epoch": 4.023874345549738,
"grad_norm": 2.1321282386779785,
"learning_rate": 9.602e-06,
"loss": 0.4587,
"step": 2400
},
{
"epoch": 4.107643979057592,
"grad_norm": 1.7361410856246948,
"learning_rate": 9.552000000000001e-06,
"loss": 0.4611,
"step": 2450
},
{
"epoch": 4.191413612565445,
"grad_norm": 2.167386770248413,
"learning_rate": 9.502000000000002e-06,
"loss": 0.455,
"step": 2500
},
{
"epoch": 4.191413612565445,
"eval_loss": 0.40905508399009705,
"eval_runtime": 269.8231,
"eval_samples_per_second": 31.461,
"eval_steps_per_second": 3.936,
"step": 2500
},
{
"epoch": 4.275183246073299,
"grad_norm": 1.430746078491211,
"learning_rate": 9.452000000000002e-06,
"loss": 0.4581,
"step": 2550
},
{
"epoch": 4.358952879581151,
"grad_norm": 2.1168527603149414,
"learning_rate": 9.402e-06,
"loss": 0.4516,
"step": 2600
},
{
"epoch": 4.442722513089005,
"grad_norm": 2.1330721378326416,
"learning_rate": 9.353000000000002e-06,
"loss": 0.4505,
"step": 2650
},
{
"epoch": 4.526492146596858,
"grad_norm": 1.274557113647461,
"learning_rate": 9.303e-06,
"loss": 0.4575,
"step": 2700
},
{
"epoch": 4.610261780104712,
"grad_norm": 1.835204839706421,
"learning_rate": 9.253000000000001e-06,
"loss": 0.4491,
"step": 2750
},
{
"epoch": 4.694031413612565,
"grad_norm": 2.0255746841430664,
"learning_rate": 9.203000000000002e-06,
"loss": 0.4484,
"step": 2800
},
{
"epoch": 4.777801047120419,
"grad_norm": 2.4793522357940674,
"learning_rate": 9.153e-06,
"loss": 0.4489,
"step": 2850
},
{
"epoch": 4.861570680628272,
"grad_norm": 1.7192201614379883,
"learning_rate": 9.103e-06,
"loss": 0.4502,
"step": 2900
},
{
"epoch": 4.945340314136126,
"grad_norm": 1.9378846883773804,
"learning_rate": 9.053000000000001e-06,
"loss": 0.4483,
"step": 2950
},
{
"epoch": 5.029109947643979,
"grad_norm": 2.4576103687286377,
"learning_rate": 9.003e-06,
"loss": 0.4474,
"step": 3000
},
{
"epoch": 5.029109947643979,
"eval_loss": 0.39903518557548523,
"eval_runtime": 269.2955,
"eval_samples_per_second": 31.523,
"eval_steps_per_second": 3.944,
"step": 3000
},
{
"epoch": 5.112879581151833,
"grad_norm": 1.4718371629714966,
"learning_rate": 8.953e-06,
"loss": 0.4428,
"step": 3050
},
{
"epoch": 5.196649214659685,
"grad_norm": 2.4222004413604736,
"learning_rate": 8.903000000000001e-06,
"loss": 0.4417,
"step": 3100
},
{
"epoch": 5.280418848167539,
"grad_norm": 1.5284855365753174,
"learning_rate": 8.853e-06,
"loss": 0.4395,
"step": 3150
},
{
"epoch": 5.364188481675392,
"grad_norm": 1.7470890283584595,
"learning_rate": 8.803e-06,
"loss": 0.4359,
"step": 3200
},
{
"epoch": 5.447958115183246,
"grad_norm": 2.1919684410095215,
"learning_rate": 8.753e-06,
"loss": 0.4389,
"step": 3250
},
{
"epoch": 5.531727748691099,
"grad_norm": 1.6179628372192383,
"learning_rate": 8.703e-06,
"loss": 0.4423,
"step": 3300
},
{
"epoch": 5.615497382198953,
"grad_norm": 1.420087218284607,
"learning_rate": 8.653e-06,
"loss": 0.4324,
"step": 3350
},
{
"epoch": 5.699267015706806,
"grad_norm": 1.6972101926803589,
"learning_rate": 8.603e-06,
"loss": 0.4327,
"step": 3400
},
{
"epoch": 5.78303664921466,
"grad_norm": 8.797286033630371,
"learning_rate": 8.553000000000001e-06,
"loss": 0.4347,
"step": 3450
},
{
"epoch": 5.866806282722513,
"grad_norm": 1.5441597700119019,
"learning_rate": 8.503e-06,
"loss": 0.4342,
"step": 3500
},
{
"epoch": 5.866806282722513,
"eval_loss": 0.39290204644203186,
"eval_runtime": 272.1215,
"eval_samples_per_second": 31.196,
"eval_steps_per_second": 3.903,
"step": 3500
},
{
"epoch": 5.950575916230367,
"grad_norm": 1.957065224647522,
"learning_rate": 8.453000000000002e-06,
"loss": 0.4288,
"step": 3550
},
{
"epoch": 6.0343455497382195,
"grad_norm": 1.3891559839248657,
"learning_rate": 8.403e-06,
"loss": 0.429,
"step": 3600
},
{
"epoch": 6.118115183246073,
"grad_norm": 1.6092897653579712,
"learning_rate": 8.353000000000001e-06,
"loss": 0.431,
"step": 3650
},
{
"epoch": 6.201884816753926,
"grad_norm": 1.5566798448562622,
"learning_rate": 8.303000000000002e-06,
"loss": 0.428,
"step": 3700
},
{
"epoch": 6.28565445026178,
"grad_norm": 2.70001220703125,
"learning_rate": 8.253e-06,
"loss": 0.433,
"step": 3750
},
{
"epoch": 6.369424083769633,
"grad_norm": 2.6352884769439697,
"learning_rate": 8.203000000000001e-06,
"loss": 0.431,
"step": 3800
},
{
"epoch": 6.453193717277487,
"grad_norm": 2.049609422683716,
"learning_rate": 8.153000000000001e-06,
"loss": 0.4279,
"step": 3850
},
{
"epoch": 6.53696335078534,
"grad_norm": 1.422606110572815,
"learning_rate": 8.103e-06,
"loss": 0.4266,
"step": 3900
},
{
"epoch": 6.620732984293194,
"grad_norm": 1.659340739250183,
"learning_rate": 8.053e-06,
"loss": 0.4293,
"step": 3950
},
{
"epoch": 6.704502617801047,
"grad_norm": 1.6214194297790527,
"learning_rate": 8.003000000000001e-06,
"loss": 0.4258,
"step": 4000
},
{
"epoch": 6.704502617801047,
"eval_loss": 0.385530561208725,
"eval_runtime": 270.5177,
"eval_samples_per_second": 31.381,
"eval_steps_per_second": 3.926,
"step": 4000
},
{
"epoch": 6.788272251308901,
"grad_norm": 2.0542335510253906,
"learning_rate": 7.953e-06,
"loss": 0.4232,
"step": 4050
},
{
"epoch": 6.872041884816754,
"grad_norm": 2.0001602172851562,
"learning_rate": 7.903e-06,
"loss": 0.4201,
"step": 4100
},
{
"epoch": 6.955811518324607,
"grad_norm": 1.5019524097442627,
"learning_rate": 7.853000000000001e-06,
"loss": 0.4202,
"step": 4150
},
{
"epoch": 7.0395811518324605,
"grad_norm": 1.5901544094085693,
"learning_rate": 7.803000000000001e-06,
"loss": 0.4214,
"step": 4200
},
{
"epoch": 7.123350785340314,
"grad_norm": 1.3859797716140747,
"learning_rate": 7.753e-06,
"loss": 0.4268,
"step": 4250
},
{
"epoch": 7.2071204188481675,
"grad_norm": 1.6223664283752441,
"learning_rate": 7.703e-06,
"loss": 0.4163,
"step": 4300
},
{
"epoch": 7.290890052356021,
"grad_norm": 1.9755197763442993,
"learning_rate": 7.653000000000001e-06,
"loss": 0.421,
"step": 4350
},
{
"epoch": 7.374659685863874,
"grad_norm": 1.9882097244262695,
"learning_rate": 7.603000000000001e-06,
"loss": 0.4215,
"step": 4400
},
{
"epoch": 7.458429319371728,
"grad_norm": 1.4274579286575317,
"learning_rate": 7.553e-06,
"loss": 0.424,
"step": 4450
},
{
"epoch": 7.542198952879581,
"grad_norm": 1.5370174646377563,
"learning_rate": 7.503e-06,
"loss": 0.4175,
"step": 4500
},
{
"epoch": 7.542198952879581,
"eval_loss": 0.3783535957336426,
"eval_runtime": 264.9436,
"eval_samples_per_second": 32.041,
"eval_steps_per_second": 4.008,
"step": 4500
},
{
"epoch": 7.625968586387435,
"grad_norm": 1.5206329822540283,
"learning_rate": 7.4530000000000005e-06,
"loss": 0.418,
"step": 4550
},
{
"epoch": 7.7097382198952875,
"grad_norm": 1.5614436864852905,
"learning_rate": 7.403e-06,
"loss": 0.4206,
"step": 4600
},
{
"epoch": 7.793507853403141,
"grad_norm": 1.9834551811218262,
"learning_rate": 7.353e-06,
"loss": 0.421,
"step": 4650
},
{
"epoch": 7.8772774869109945,
"grad_norm": 2.0368449687957764,
"learning_rate": 7.303e-06,
"loss": 0.4206,
"step": 4700
},
{
"epoch": 7.961047120418848,
"grad_norm": 1.4374061822891235,
"learning_rate": 7.253e-06,
"loss": 0.4127,
"step": 4750
},
{
"epoch": 8.044816753926701,
"grad_norm": 1.9674043655395508,
"learning_rate": 7.203000000000001e-06,
"loss": 0.4111,
"step": 4800
},
{
"epoch": 8.128586387434556,
"grad_norm": 1.4639081954956055,
"learning_rate": 7.153000000000001e-06,
"loss": 0.4153,
"step": 4850
},
{
"epoch": 8.212356020942408,
"grad_norm": 1.189695954322815,
"learning_rate": 7.103000000000001e-06,
"loss": 0.4257,
"step": 4900
},
{
"epoch": 8.296125654450261,
"grad_norm": 1.5426299571990967,
"learning_rate": 7.053000000000001e-06,
"loss": 0.4119,
"step": 4950
},
{
"epoch": 8.379895287958115,
"grad_norm": 1.526410460472107,
"learning_rate": 7.0030000000000005e-06,
"loss": 0.4185,
"step": 5000
},
{
"epoch": 8.379895287958115,
"eval_loss": 0.3761734962463379,
"eval_runtime": 274.1558,
"eval_samples_per_second": 30.964,
"eval_steps_per_second": 3.874,
"step": 5000
},
{
"epoch": 8.463664921465968,
"grad_norm": 1.3419328927993774,
"learning_rate": 6.953000000000001e-06,
"loss": 0.412,
"step": 5050
},
{
"epoch": 8.547434554973822,
"grad_norm": 2.1306657791137695,
"learning_rate": 6.903000000000001e-06,
"loss": 0.4108,
"step": 5100
},
{
"epoch": 8.631204188481675,
"grad_norm": 2.0914053916931152,
"learning_rate": 6.853000000000001e-06,
"loss": 0.4177,
"step": 5150
},
{
"epoch": 8.71497382198953,
"grad_norm": 1.6214958429336548,
"learning_rate": 6.803000000000001e-06,
"loss": 0.4083,
"step": 5200
},
{
"epoch": 8.798743455497382,
"grad_norm": 1.5650039911270142,
"learning_rate": 6.753e-06,
"loss": 0.4118,
"step": 5250
},
{
"epoch": 8.882513089005236,
"grad_norm": 1.564424991607666,
"learning_rate": 6.703000000000001e-06,
"loss": 0.4093,
"step": 5300
},
{
"epoch": 8.966282722513089,
"grad_norm": 1.623558759689331,
"learning_rate": 6.6530000000000005e-06,
"loss": 0.412,
"step": 5350
},
{
"epoch": 9.050052356020942,
"grad_norm": 1.5212219953536987,
"learning_rate": 6.603e-06,
"loss": 0.4125,
"step": 5400
},
{
"epoch": 9.133821989528796,
"grad_norm": 2.202746629714966,
"learning_rate": 6.553000000000001e-06,
"loss": 0.4049,
"step": 5450
},
{
"epoch": 9.217591623036649,
"grad_norm": 1.6582872867584229,
"learning_rate": 6.503e-06,
"loss": 0.4093,
"step": 5500
},
{
"epoch": 9.217591623036649,
"eval_loss": 0.373293936252594,
"eval_runtime": 279.0862,
"eval_samples_per_second": 30.417,
"eval_steps_per_second": 3.805,
"step": 5500
},
{
"epoch": 9.301361256544503,
"grad_norm": 2.3431801795959473,
"learning_rate": 6.453000000000001e-06,
"loss": 0.4099,
"step": 5550
},
{
"epoch": 9.385130890052356,
"grad_norm": 1.7502645254135132,
"learning_rate": 6.403e-06,
"loss": 0.4097,
"step": 5600
},
{
"epoch": 9.46890052356021,
"grad_norm": 1.638600468635559,
"learning_rate": 6.353e-06,
"loss": 0.4068,
"step": 5650
},
{
"epoch": 9.552670157068063,
"grad_norm": 1.3677990436553955,
"learning_rate": 6.3030000000000005e-06,
"loss": 0.4086,
"step": 5700
},
{
"epoch": 9.636439790575917,
"grad_norm": 1.4438238143920898,
"learning_rate": 6.253e-06,
"loss": 0.4067,
"step": 5750
},
{
"epoch": 9.72020942408377,
"grad_norm": 1.7439764738082886,
"learning_rate": 6.204e-06,
"loss": 0.4085,
"step": 5800
},
{
"epoch": 9.803979057591622,
"grad_norm": 1.257203459739685,
"learning_rate": 6.154e-06,
"loss": 0.4069,
"step": 5850
},
{
"epoch": 9.887748691099477,
"grad_norm": 1.2076969146728516,
"learning_rate": 6.104000000000001e-06,
"loss": 0.4,
"step": 5900
},
{
"epoch": 9.97151832460733,
"grad_norm": 1.1376831531524658,
"learning_rate": 6.054000000000001e-06,
"loss": 0.4088,
"step": 5950
},
{
"epoch": 10.055287958115183,
"grad_norm": 1.4293948411941528,
"learning_rate": 6.004000000000001e-06,
"loss": 0.4012,
"step": 6000
},
{
"epoch": 10.055287958115183,
"eval_loss": 0.3698088824748993,
"eval_runtime": 276.0941,
"eval_samples_per_second": 30.747,
"eval_steps_per_second": 3.847,
"step": 6000
},
{
"epoch": 10.139057591623036,
"grad_norm": 1.1719565391540527,
"learning_rate": 5.954000000000001e-06,
"loss": 0.4036,
"step": 6050
},
{
"epoch": 10.22282722513089,
"grad_norm": 1.0520621538162231,
"learning_rate": 5.9040000000000006e-06,
"loss": 0.4074,
"step": 6100
},
{
"epoch": 10.306596858638743,
"grad_norm": 1.9477139711380005,
"learning_rate": 5.854000000000001e-06,
"loss": 0.4049,
"step": 6150
},
{
"epoch": 10.390366492146597,
"grad_norm": 1.147050142288208,
"learning_rate": 5.804000000000001e-06,
"loss": 0.4009,
"step": 6200
},
{
"epoch": 10.47413612565445,
"grad_norm": 1.32716703414917,
"learning_rate": 5.754e-06,
"loss": 0.4037,
"step": 6250
},
{
"epoch": 10.557905759162304,
"grad_norm": 1.6168224811553955,
"learning_rate": 5.704000000000001e-06,
"loss": 0.401,
"step": 6300
},
{
"epoch": 10.641675392670157,
"grad_norm": 1.3314645290374756,
"learning_rate": 5.654e-06,
"loss": 0.4039,
"step": 6350
},
{
"epoch": 10.72544502617801,
"grad_norm": 1.0647807121276855,
"learning_rate": 5.604000000000001e-06,
"loss": 0.4039,
"step": 6400
},
{
"epoch": 10.809214659685864,
"grad_norm": 1.3919047117233276,
"learning_rate": 5.5540000000000005e-06,
"loss": 0.4014,
"step": 6450
},
{
"epoch": 10.892984293193717,
"grad_norm": 1.3037397861480713,
"learning_rate": 5.504e-06,
"loss": 0.403,
"step": 6500
},
{
"epoch": 10.892984293193717,
"eval_loss": 0.3670424520969391,
"eval_runtime": 281.4305,
"eval_samples_per_second": 30.164,
"eval_steps_per_second": 3.774,
"step": 6500
},
{
"epoch": 10.976753926701571,
"grad_norm": 0.9927046298980713,
"learning_rate": 5.454000000000001e-06,
"loss": 0.4002,
"step": 6550
},
{
"epoch": 11.060523560209424,
"grad_norm": 2.82547926902771,
"learning_rate": 5.404e-06,
"loss": 0.4035,
"step": 6600
},
{
"epoch": 11.144293193717278,
"grad_norm": 1.4662278890609741,
"learning_rate": 5.354e-06,
"loss": 0.4031,
"step": 6650
},
{
"epoch": 11.22806282722513,
"grad_norm": 0.9803363084793091,
"learning_rate": 5.304e-06,
"loss": 0.3984,
"step": 6700
},
{
"epoch": 11.311832460732985,
"grad_norm": 1.580372929573059,
"learning_rate": 5.254e-06,
"loss": 0.3987,
"step": 6750
},
{
"epoch": 11.395602094240838,
"grad_norm": 1.5269379615783691,
"learning_rate": 5.2040000000000005e-06,
"loss": 0.3999,
"step": 6800
},
{
"epoch": 11.479371727748692,
"grad_norm": 1.2878773212432861,
"learning_rate": 5.154e-06,
"loss": 0.3991,
"step": 6850
},
{
"epoch": 11.563141361256545,
"grad_norm": 1.2725164890289307,
"learning_rate": 5.104e-06,
"loss": 0.3979,
"step": 6900
},
{
"epoch": 11.646910994764397,
"grad_norm": 2.5339043140411377,
"learning_rate": 5.054e-06,
"loss": 0.3966,
"step": 6950
},
{
"epoch": 11.730680628272252,
"grad_norm": 1.1876580715179443,
"learning_rate": 5.004e-06,
"loss": 0.4017,
"step": 7000
},
{
"epoch": 11.730680628272252,
"eval_loss": 0.3627295196056366,
"eval_runtime": 270.756,
"eval_samples_per_second": 31.353,
"eval_steps_per_second": 3.922,
"step": 7000
},
{
"epoch": 11.814450261780104,
"grad_norm": 1.1595991849899292,
"learning_rate": 4.954e-06,
"loss": 0.4003,
"step": 7050
},
{
"epoch": 11.898219895287959,
"grad_norm": 1.2361228466033936,
"learning_rate": 4.904000000000001e-06,
"loss": 0.4019,
"step": 7100
},
{
"epoch": 11.981989528795811,
"grad_norm": 1.5332549810409546,
"learning_rate": 4.8540000000000005e-06,
"loss": 0.395,
"step": 7150
},
{
"epoch": 12.065759162303666,
"grad_norm": 1.4963312149047852,
"learning_rate": 4.804e-06,
"loss": 0.406,
"step": 7200
},
{
"epoch": 12.149528795811518,
"grad_norm": 1.5511187314987183,
"learning_rate": 4.7540000000000006e-06,
"loss": 0.3991,
"step": 7250
},
{
"epoch": 12.233298429319373,
"grad_norm": 1.8427493572235107,
"learning_rate": 4.704e-06,
"loss": 0.3982,
"step": 7300
},
{
"epoch": 12.317068062827225,
"grad_norm": 1.3739436864852905,
"learning_rate": 4.654e-06,
"loss": 0.3966,
"step": 7350
},
{
"epoch": 12.400837696335078,
"grad_norm": 1.659621238708496,
"learning_rate": 4.604e-06,
"loss": 0.4019,
"step": 7400
},
{
"epoch": 12.484607329842932,
"grad_norm": 1.8510388135910034,
"learning_rate": 4.554000000000001e-06,
"loss": 0.4013,
"step": 7450
},
{
"epoch": 12.568376963350785,
"grad_norm": 1.100199580192566,
"learning_rate": 4.504e-06,
"loss": 0.3961,
"step": 7500
},
{
"epoch": 12.568376963350785,
"eval_loss": 0.36272022128105164,
"eval_runtime": 264.6514,
"eval_samples_per_second": 32.076,
"eval_steps_per_second": 4.013,
"step": 7500
},
{
"epoch": 12.652146596858639,
"grad_norm": 1.439460039138794,
"learning_rate": 4.454000000000001e-06,
"loss": 0.3937,
"step": 7550
},
{
"epoch": 12.735916230366492,
"grad_norm": 1.253373622894287,
"learning_rate": 4.4040000000000005e-06,
"loss": 0.4011,
"step": 7600
},
{
"epoch": 12.819685863874346,
"grad_norm": 1.6404370069503784,
"learning_rate": 4.354e-06,
"loss": 0.4016,
"step": 7650
},
{
"epoch": 12.903455497382199,
"grad_norm": 1.3606590032577515,
"learning_rate": 4.304000000000001e-06,
"loss": 0.3964,
"step": 7700
},
{
"epoch": 12.987225130890053,
"grad_norm": 1.249391794204712,
"learning_rate": 4.254e-06,
"loss": 0.4013,
"step": 7750
},
{
"epoch": 13.070994764397906,
"grad_norm": 1.1404308080673218,
"learning_rate": 4.205e-06,
"loss": 0.3939,
"step": 7800
},
{
"epoch": 13.15476439790576,
"grad_norm": 0.9418945908546448,
"learning_rate": 4.155e-06,
"loss": 0.3968,
"step": 7850
},
{
"epoch": 13.238534031413613,
"grad_norm": 1.3190369606018066,
"learning_rate": 4.1050000000000005e-06,
"loss": 0.3977,
"step": 7900
},
{
"epoch": 13.322303664921465,
"grad_norm": 1.2260388135910034,
"learning_rate": 4.055000000000001e-06,
"loss": 0.3973,
"step": 7950
},
{
"epoch": 13.40607329842932,
"grad_norm": 1.035477638244629,
"learning_rate": 4.005000000000001e-06,
"loss": 0.3957,
"step": 8000
},
{
"epoch": 13.40607329842932,
"eval_loss": 0.3603293001651764,
"eval_runtime": 270.7314,
"eval_samples_per_second": 31.356,
"eval_steps_per_second": 3.923,
"step": 8000
},
{
"epoch": 13.489842931937172,
"grad_norm": 1.4767895936965942,
"learning_rate": 3.955e-06,
"loss": 0.3921,
"step": 8050
},
{
"epoch": 13.573612565445027,
"grad_norm": 15.019492149353027,
"learning_rate": 3.905000000000001e-06,
"loss": 0.3947,
"step": 8100
},
{
"epoch": 13.65738219895288,
"grad_norm": 1.3216609954833984,
"learning_rate": 3.855e-06,
"loss": 0.3969,
"step": 8150
},
{
"epoch": 13.741151832460734,
"grad_norm": 1.0909982919692993,
"learning_rate": 3.8050000000000004e-06,
"loss": 0.3916,
"step": 8200
},
{
"epoch": 13.824921465968586,
"grad_norm": 1.6264184713363647,
"learning_rate": 3.7550000000000005e-06,
"loss": 0.3951,
"step": 8250
},
{
"epoch": 13.90869109947644,
"grad_norm": 1.1289631128311157,
"learning_rate": 3.705e-06,
"loss": 0.3926,
"step": 8300
},
{
"epoch": 13.992460732984293,
"grad_norm": 1.1617991924285889,
"learning_rate": 3.655e-06,
"loss": 0.393,
"step": 8350
},
{
"epoch": 14.076230366492146,
"grad_norm": 1.1278605461120605,
"learning_rate": 3.6050000000000002e-06,
"loss": 0.3969,
"step": 8400
},
{
"epoch": 14.16,
"grad_norm": 0.9024301767349243,
"learning_rate": 3.5550000000000003e-06,
"loss": 0.3888,
"step": 8450
},
{
"epoch": 14.243769633507853,
"grad_norm": 0.9253320097923279,
"learning_rate": 3.505e-06,
"loss": 0.3919,
"step": 8500
},
{
"epoch": 14.243769633507853,
"eval_loss": 0.35886332392692566,
"eval_runtime": 262.8233,
"eval_samples_per_second": 32.299,
"eval_steps_per_second": 4.041,
"step": 8500
},
{
"epoch": 14.327539267015707,
"grad_norm": 1.3649754524230957,
"learning_rate": 3.455e-06,
"loss": 0.3961,
"step": 8550
},
{
"epoch": 14.41130890052356,
"grad_norm": 1.2764108180999756,
"learning_rate": 3.4050000000000004e-06,
"loss": 0.3946,
"step": 8600
},
{
"epoch": 14.495078534031414,
"grad_norm": 1.174688458442688,
"learning_rate": 3.3550000000000005e-06,
"loss": 0.3955,
"step": 8650
},
{
"epoch": 14.578848167539267,
"grad_norm": 1.0795891284942627,
"learning_rate": 3.3050000000000005e-06,
"loss": 0.3933,
"step": 8700
},
{
"epoch": 14.662617801047121,
"grad_norm": 0.9259040355682373,
"learning_rate": 3.255e-06,
"loss": 0.3913,
"step": 8750
},
{
"epoch": 14.746387434554974,
"grad_norm": 1.1176104545593262,
"learning_rate": 3.2050000000000002e-06,
"loss": 0.3942,
"step": 8800
},
{
"epoch": 14.830157068062828,
"grad_norm": 1.0351334810256958,
"learning_rate": 3.1550000000000003e-06,
"loss": 0.3981,
"step": 8850
},
{
"epoch": 14.91392670157068,
"grad_norm": 2.326068878173828,
"learning_rate": 3.1050000000000003e-06,
"loss": 0.3948,
"step": 8900
},
{
"epoch": 14.997696335078533,
"grad_norm": 1.4447543621063232,
"learning_rate": 3.0550000000000004e-06,
"loss": 0.3908,
"step": 8950
},
{
"epoch": 15.081465968586388,
"grad_norm": 1.365200161933899,
"learning_rate": 3.005e-06,
"loss": 0.3933,
"step": 9000
},
{
"epoch": 15.081465968586388,
"eval_loss": 0.35749351978302,
"eval_runtime": 283.7549,
"eval_samples_per_second": 29.917,
"eval_steps_per_second": 3.743,
"step": 9000
},
{
"epoch": 15.16523560209424,
"grad_norm": 1.0825339555740356,
"learning_rate": 2.955e-06,
"loss": 0.3923,
"step": 9050
},
{
"epoch": 15.249005235602095,
"grad_norm": 1.0156179666519165,
"learning_rate": 2.905e-06,
"loss": 0.3913,
"step": 9100
},
{
"epoch": 15.332774869109947,
"grad_norm": 1.0104918479919434,
"learning_rate": 2.855e-06,
"loss": 0.3893,
"step": 9150
},
{
"epoch": 15.416544502617802,
"grad_norm": 0.9494548439979553,
"learning_rate": 2.8050000000000007e-06,
"loss": 0.391,
"step": 9200
},
{
"epoch": 15.500314136125654,
"grad_norm": 1.2736467123031616,
"learning_rate": 2.7550000000000003e-06,
"loss": 0.3946,
"step": 9250
},
{
"epoch": 15.584083769633509,
"grad_norm": 0.81021648645401,
"learning_rate": 2.7050000000000004e-06,
"loss": 0.3961,
"step": 9300
},
{
"epoch": 15.667853403141361,
"grad_norm": 1.0645405054092407,
"learning_rate": 2.6550000000000004e-06,
"loss": 0.3912,
"step": 9350
},
{
"epoch": 15.751623036649214,
"grad_norm": 4.980652809143066,
"learning_rate": 2.6050000000000005e-06,
"loss": 0.3902,
"step": 9400
},
{
"epoch": 15.835392670157068,
"grad_norm": 1.4880826473236084,
"learning_rate": 2.555e-06,
"loss": 0.3939,
"step": 9450
},
{
"epoch": 15.91916230366492,
"grad_norm": 0.8299314379692078,
"learning_rate": 2.505e-06,
"loss": 0.3928,
"step": 9500
},
{
"epoch": 15.91916230366492,
"eval_loss": 0.3572653830051422,
"eval_runtime": 261.7949,
"eval_samples_per_second": 32.426,
"eval_steps_per_second": 4.057,
"step": 9500
},
{
"epoch": 16.002931937172775,
"grad_norm": 1.1118539571762085,
"learning_rate": 2.4550000000000002e-06,
"loss": 0.3938,
"step": 9550
},
{
"epoch": 16.086701570680628,
"grad_norm": 1.1094309091567993,
"learning_rate": 2.4050000000000003e-06,
"loss": 0.3903,
"step": 9600
},
{
"epoch": 16.17047120418848,
"grad_norm": 0.9645286202430725,
"learning_rate": 2.355e-06,
"loss": 0.3928,
"step": 9650
},
{
"epoch": 16.254240837696337,
"grad_norm": 1.376489520072937,
"learning_rate": 2.3050000000000004e-06,
"loss": 0.39,
"step": 9700
},
{
"epoch": 16.33801047120419,
"grad_norm": 1.0610970258712769,
"learning_rate": 2.2550000000000004e-06,
"loss": 0.3861,
"step": 9750
},
{
"epoch": 16.42178010471204,
"grad_norm": 1.481760859489441,
"learning_rate": 2.205e-06,
"loss": 0.3964,
"step": 9800
},
{
"epoch": 16.505549738219894,
"grad_norm": 1.0688999891281128,
"learning_rate": 2.155e-06,
"loss": 0.3965,
"step": 9850
},
{
"epoch": 16.589319371727747,
"grad_norm": 1.2465980052947998,
"learning_rate": 2.105e-06,
"loss": 0.3904,
"step": 9900
},
{
"epoch": 16.673089005235603,
"grad_norm": 2.9295108318328857,
"learning_rate": 2.0550000000000002e-06,
"loss": 0.3944,
"step": 9950
},
{
"epoch": 16.756858638743456,
"grad_norm": 1.1967060565948486,
"learning_rate": 2.0050000000000003e-06,
"loss": 0.3951,
"step": 10000
},
{
"epoch": 16.756858638743456,
"eval_loss": 0.35580456256866455,
"eval_runtime": 268.9079,
"eval_samples_per_second": 31.568,
"eval_steps_per_second": 3.949,
"step": 10000
},
{
"epoch": 16.84062827225131,
"grad_norm": 1.9004417657852173,
"learning_rate": 1.9550000000000003e-06,
"loss": 0.3919,
"step": 10050
},
{
"epoch": 16.92439790575916,
"grad_norm": 1.0050182342529297,
"learning_rate": 1.9050000000000002e-06,
"loss": 0.3887,
"step": 10100
},
{
"epoch": 17.008167539267017,
"grad_norm": 1.1736068725585938,
"learning_rate": 1.8550000000000002e-06,
"loss": 0.3875,
"step": 10150
},
{
"epoch": 17.09193717277487,
"grad_norm": 1.0031588077545166,
"learning_rate": 1.805e-06,
"loss": 0.3901,
"step": 10200
},
{
"epoch": 17.175706806282722,
"grad_norm": 0.9444441199302673,
"learning_rate": 1.7550000000000001e-06,
"loss": 0.3911,
"step": 10250
},
{
"epoch": 17.259476439790575,
"grad_norm": 2.094243049621582,
"learning_rate": 1.7050000000000002e-06,
"loss": 0.3934,
"step": 10300
},
{
"epoch": 17.34324607329843,
"grad_norm": 1.1646901369094849,
"learning_rate": 1.6550000000000002e-06,
"loss": 0.3905,
"step": 10350
},
{
"epoch": 17.427015706806284,
"grad_norm": 0.9865755438804626,
"learning_rate": 1.605e-06,
"loss": 0.3936,
"step": 10400
},
{
"epoch": 17.510785340314136,
"grad_norm": 0.8808874487876892,
"learning_rate": 1.5550000000000001e-06,
"loss": 0.3908,
"step": 10450
},
{
"epoch": 17.59455497382199,
"grad_norm": 0.9992907047271729,
"learning_rate": 1.505e-06,
"loss": 0.3911,
"step": 10500
},
{
"epoch": 17.59455497382199,
"eval_loss": 0.35496434569358826,
"eval_runtime": 270.622,
"eval_samples_per_second": 31.368,
"eval_steps_per_second": 3.924,
"step": 10500
}
],
"logging_steps": 50,
"max_steps": 12000,
"num_input_tokens_seen": 0,
"num_train_epochs": 21,
"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": 1.8839142441042106e+17,
"train_batch_size": 16,
"trial_name": null,
"trial_params": null
}