kanishka's picture
End of training
f1354e2 verified
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 20.0,
"eval_steps": 500,
"global_step": 371900,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.05,
"grad_norm": 1.4397599697113037,
"learning_rate": 3.125e-06,
"loss": 7.5907,
"step": 1000
},
{
"epoch": 0.11,
"grad_norm": 0.7548799514770508,
"learning_rate": 6.25e-06,
"loss": 5.82,
"step": 2000
},
{
"epoch": 0.16,
"grad_norm": 0.921891987323761,
"learning_rate": 9.375000000000001e-06,
"loss": 5.3784,
"step": 3000
},
{
"epoch": 0.22,
"grad_norm": 0.9153477549552917,
"learning_rate": 1.25e-05,
"loss": 5.1593,
"step": 4000
},
{
"epoch": 0.27,
"grad_norm": 1.0300954580307007,
"learning_rate": 1.5625e-05,
"loss": 5.015,
"step": 5000
},
{
"epoch": 0.32,
"grad_norm": 1.0274837017059326,
"learning_rate": 1.8750000000000002e-05,
"loss": 4.8818,
"step": 6000
},
{
"epoch": 0.38,
"grad_norm": 1.128319263458252,
"learning_rate": 2.1875e-05,
"loss": 4.757,
"step": 7000
},
{
"epoch": 0.43,
"grad_norm": 1.1055282354354858,
"learning_rate": 2.5e-05,
"loss": 4.6511,
"step": 8000
},
{
"epoch": 0.48,
"grad_norm": 1.071848750114441,
"learning_rate": 2.8125000000000003e-05,
"loss": 4.5641,
"step": 9000
},
{
"epoch": 0.54,
"grad_norm": 0.9956591725349426,
"learning_rate": 3.125e-05,
"loss": 4.4822,
"step": 10000
},
{
"epoch": 0.59,
"grad_norm": 1.0482920408248901,
"learning_rate": 3.4371875e-05,
"loss": 4.4079,
"step": 11000
},
{
"epoch": 0.65,
"grad_norm": 1.0127135515213013,
"learning_rate": 3.7496875e-05,
"loss": 4.3425,
"step": 12000
},
{
"epoch": 0.7,
"grad_norm": 1.0505046844482422,
"learning_rate": 4.0621875e-05,
"loss": 4.2861,
"step": 13000
},
{
"epoch": 0.75,
"grad_norm": 1.0681025981903076,
"learning_rate": 4.3746875000000005e-05,
"loss": 4.2289,
"step": 14000
},
{
"epoch": 0.81,
"grad_norm": 1.1013708114624023,
"learning_rate": 4.686875e-05,
"loss": 4.1873,
"step": 15000
},
{
"epoch": 0.86,
"grad_norm": 1.0665634870529175,
"learning_rate": 4.9990625000000004e-05,
"loss": 4.1308,
"step": 16000
},
{
"epoch": 0.91,
"grad_norm": 1.1269334554672241,
"learning_rate": 5.3115625000000005e-05,
"loss": 4.0923,
"step": 17000
},
{
"epoch": 0.97,
"grad_norm": 0.9998357892036438,
"learning_rate": 5.62375e-05,
"loss": 4.0514,
"step": 18000
},
{
"epoch": 1.0,
"eval_accuracy": 0.3101297245280225,
"eval_loss": 4.2435173988342285,
"eval_runtime": 151.7646,
"eval_samples_per_second": 381.631,
"eval_steps_per_second": 5.963,
"step": 18595
},
{
"epoch": 1.02,
"grad_norm": 0.9669459462165833,
"learning_rate": 5.93625e-05,
"loss": 4.0035,
"step": 19000
},
{
"epoch": 1.08,
"grad_norm": 1.1156255006790161,
"learning_rate": 6.24875e-05,
"loss": 3.9592,
"step": 20000
},
{
"epoch": 1.13,
"grad_norm": 0.9864040613174438,
"learning_rate": 6.56125e-05,
"loss": 3.9219,
"step": 21000
},
{
"epoch": 1.18,
"grad_norm": 1.012749433517456,
"learning_rate": 6.8734375e-05,
"loss": 3.8856,
"step": 22000
},
{
"epoch": 1.24,
"grad_norm": 1.0206193923950195,
"learning_rate": 7.185937500000001e-05,
"loss": 3.8566,
"step": 23000
},
{
"epoch": 1.29,
"grad_norm": 0.9851728081703186,
"learning_rate": 7.498125e-05,
"loss": 3.8218,
"step": 24000
},
{
"epoch": 1.34,
"grad_norm": 0.9994772672653198,
"learning_rate": 7.810625e-05,
"loss": 3.7989,
"step": 25000
},
{
"epoch": 1.4,
"grad_norm": 0.9388182163238525,
"learning_rate": 8.1228125e-05,
"loss": 3.7688,
"step": 26000
},
{
"epoch": 1.45,
"grad_norm": 0.8769707083702087,
"learning_rate": 8.435e-05,
"loss": 3.7461,
"step": 27000
},
{
"epoch": 1.51,
"grad_norm": 0.9455641508102417,
"learning_rate": 8.747500000000001e-05,
"loss": 3.7314,
"step": 28000
},
{
"epoch": 1.56,
"grad_norm": 0.932089626789093,
"learning_rate": 9.0596875e-05,
"loss": 3.7058,
"step": 29000
},
{
"epoch": 1.61,
"grad_norm": 0.8848442435264587,
"learning_rate": 9.3721875e-05,
"loss": 3.6882,
"step": 30000
},
{
"epoch": 1.67,
"grad_norm": 0.8709959387779236,
"learning_rate": 9.684375e-05,
"loss": 3.6697,
"step": 31000
},
{
"epoch": 1.72,
"grad_norm": 0.8558521866798401,
"learning_rate": 9.9965625e-05,
"loss": 3.6505,
"step": 32000
},
{
"epoch": 1.77,
"grad_norm": 0.8082968592643738,
"learning_rate": 9.970903206825538e-05,
"loss": 3.6347,
"step": 33000
},
{
"epoch": 1.83,
"grad_norm": 0.849818766117096,
"learning_rate": 9.941512209473375e-05,
"loss": 3.6204,
"step": 34000
},
{
"epoch": 1.88,
"grad_norm": 0.8615933060646057,
"learning_rate": 9.912091791703443e-05,
"loss": 3.5962,
"step": 35000
},
{
"epoch": 1.94,
"grad_norm": 0.8331784009933472,
"learning_rate": 9.882671373933511e-05,
"loss": 3.5822,
"step": 36000
},
{
"epoch": 1.99,
"grad_norm": 0.8385667204856873,
"learning_rate": 9.853280376581348e-05,
"loss": 3.5672,
"step": 37000
},
{
"epoch": 2.0,
"eval_accuracy": 0.36221752126950446,
"eval_loss": 3.765214681625366,
"eval_runtime": 153.4159,
"eval_samples_per_second": 377.523,
"eval_steps_per_second": 5.899,
"step": 37190
},
{
"epoch": 2.04,
"grad_norm": 0.8255208730697632,
"learning_rate": 9.823859958811416e-05,
"loss": 3.5267,
"step": 38000
},
{
"epoch": 2.1,
"grad_norm": 0.8173959851264954,
"learning_rate": 9.794468961459253e-05,
"loss": 3.509,
"step": 39000
},
{
"epoch": 2.15,
"grad_norm": 0.826338529586792,
"learning_rate": 9.765048543689321e-05,
"loss": 3.5054,
"step": 40000
},
{
"epoch": 2.2,
"grad_norm": 0.8249008655548096,
"learning_rate": 9.735628125919388e-05,
"loss": 3.4969,
"step": 41000
},
{
"epoch": 2.26,
"grad_norm": 0.7982118725776672,
"learning_rate": 9.706207708149456e-05,
"loss": 3.4873,
"step": 42000
},
{
"epoch": 2.31,
"grad_norm": 0.796518862247467,
"learning_rate": 9.676816710797295e-05,
"loss": 3.4752,
"step": 43000
},
{
"epoch": 2.37,
"grad_norm": 0.7944287061691284,
"learning_rate": 9.647396293027361e-05,
"loss": 3.4747,
"step": 44000
},
{
"epoch": 2.42,
"grad_norm": 0.801404595375061,
"learning_rate": 9.61797587525743e-05,
"loss": 3.464,
"step": 45000
},
{
"epoch": 2.47,
"grad_norm": 0.782745361328125,
"learning_rate": 9.588584877905267e-05,
"loss": 3.4534,
"step": 46000
},
{
"epoch": 2.53,
"grad_norm": 0.9451091289520264,
"learning_rate": 9.559164460135335e-05,
"loss": 3.4443,
"step": 47000
},
{
"epoch": 2.58,
"grad_norm": 0.7674155235290527,
"learning_rate": 9.529773462783173e-05,
"loss": 3.4388,
"step": 48000
},
{
"epoch": 2.64,
"grad_norm": 0.7728825211524963,
"learning_rate": 9.50035304501324e-05,
"loss": 3.4275,
"step": 49000
},
{
"epoch": 2.69,
"grad_norm": 0.7956791520118713,
"learning_rate": 9.470932627243308e-05,
"loss": 3.4232,
"step": 50000
},
{
"epoch": 2.74,
"grad_norm": 0.799748420715332,
"learning_rate": 9.441541629891145e-05,
"loss": 3.4107,
"step": 51000
},
{
"epoch": 2.8,
"grad_norm": 0.7880944013595581,
"learning_rate": 9.412121212121212e-05,
"loss": 3.4125,
"step": 52000
},
{
"epoch": 2.85,
"grad_norm": 0.7877218127250671,
"learning_rate": 9.38273021476905e-05,
"loss": 3.4058,
"step": 53000
},
{
"epoch": 2.9,
"grad_norm": 0.784431517124176,
"learning_rate": 9.353309796999118e-05,
"loss": 3.3989,
"step": 54000
},
{
"epoch": 2.96,
"grad_norm": 0.7350291609764099,
"learning_rate": 9.323889379229186e-05,
"loss": 3.3933,
"step": 55000
},
{
"epoch": 3.0,
"eval_accuracy": 0.37918697076492014,
"eval_loss": 3.5858614444732666,
"eval_runtime": 153.7959,
"eval_samples_per_second": 376.59,
"eval_steps_per_second": 5.884,
"step": 55785
},
{
"epoch": 3.01,
"grad_norm": 0.7721139788627625,
"learning_rate": 9.294468961459253e-05,
"loss": 3.3787,
"step": 56000
},
{
"epoch": 3.07,
"grad_norm": 0.7534685730934143,
"learning_rate": 9.26507796410709e-05,
"loss": 3.3358,
"step": 57000
},
{
"epoch": 3.12,
"grad_norm": 0.7684205770492554,
"learning_rate": 9.235657546337158e-05,
"loss": 3.3365,
"step": 58000
},
{
"epoch": 3.17,
"grad_norm": 0.7345510125160217,
"learning_rate": 9.206237128567227e-05,
"loss": 3.3292,
"step": 59000
},
{
"epoch": 3.23,
"grad_norm": 0.772286593914032,
"learning_rate": 9.176846131215064e-05,
"loss": 3.3285,
"step": 60000
},
{
"epoch": 3.28,
"grad_norm": 0.7555109262466431,
"learning_rate": 9.147425713445132e-05,
"loss": 3.3297,
"step": 61000
},
{
"epoch": 3.33,
"grad_norm": 0.7533125281333923,
"learning_rate": 9.118034716092969e-05,
"loss": 3.3248,
"step": 62000
},
{
"epoch": 3.39,
"grad_norm": 0.7516513466835022,
"learning_rate": 9.088614298323037e-05,
"loss": 3.3264,
"step": 63000
},
{
"epoch": 3.44,
"grad_norm": 0.7214173078536987,
"learning_rate": 9.059193880553104e-05,
"loss": 3.3206,
"step": 64000
},
{
"epoch": 3.5,
"grad_norm": 0.7576092481613159,
"learning_rate": 9.029802883200942e-05,
"loss": 3.316,
"step": 65000
},
{
"epoch": 3.55,
"grad_norm": 0.7526379823684692,
"learning_rate": 9.000411885848779e-05,
"loss": 3.313,
"step": 66000
},
{
"epoch": 3.6,
"grad_norm": 0.7402895092964172,
"learning_rate": 8.970991468078847e-05,
"loss": 3.3134,
"step": 67000
},
{
"epoch": 3.66,
"grad_norm": 0.7132677435874939,
"learning_rate": 8.941571050308914e-05,
"loss": 3.3069,
"step": 68000
},
{
"epoch": 3.71,
"grad_norm": 0.8066958785057068,
"learning_rate": 8.912150632538982e-05,
"loss": 3.307,
"step": 69000
},
{
"epoch": 3.76,
"grad_norm": 0.7379071712493896,
"learning_rate": 8.88273021476905e-05,
"loss": 3.302,
"step": 70000
},
{
"epoch": 3.82,
"grad_norm": 0.7558432817459106,
"learning_rate": 8.853339217416887e-05,
"loss": 3.3019,
"step": 71000
},
{
"epoch": 3.87,
"grad_norm": 0.7654557228088379,
"learning_rate": 8.823918799646955e-05,
"loss": 3.2993,
"step": 72000
},
{
"epoch": 3.93,
"grad_norm": 0.7307350635528564,
"learning_rate": 8.794527802294792e-05,
"loss": 3.2961,
"step": 73000
},
{
"epoch": 3.98,
"grad_norm": 0.7173799872398376,
"learning_rate": 8.76510738452486e-05,
"loss": 3.2939,
"step": 74000
},
{
"epoch": 4.0,
"eval_accuracy": 0.3863495861517887,
"eval_loss": 3.5397000312805176,
"eval_runtime": 153.3856,
"eval_samples_per_second": 377.597,
"eval_steps_per_second": 5.9,
"step": 74380
},
{
"epoch": 4.03,
"grad_norm": 0.7380111217498779,
"learning_rate": 8.735686966754927e-05,
"loss": 3.2551,
"step": 75000
},
{
"epoch": 4.09,
"grad_norm": 0.7536616921424866,
"learning_rate": 8.706266548984995e-05,
"loss": 3.2352,
"step": 76000
},
{
"epoch": 4.14,
"grad_norm": 0.7641943097114563,
"learning_rate": 8.676875551632834e-05,
"loss": 3.2392,
"step": 77000
},
{
"epoch": 4.19,
"grad_norm": 0.7541372776031494,
"learning_rate": 8.647455133862902e-05,
"loss": 3.2448,
"step": 78000
},
{
"epoch": 4.25,
"grad_norm": 0.7148016691207886,
"learning_rate": 8.618064136510739e-05,
"loss": 3.2369,
"step": 79000
},
{
"epoch": 4.3,
"grad_norm": 0.7816110253334045,
"learning_rate": 8.588643718740806e-05,
"loss": 3.238,
"step": 80000
},
{
"epoch": 4.36,
"grad_norm": 0.7324749231338501,
"learning_rate": 8.559223300970874e-05,
"loss": 3.2387,
"step": 81000
},
{
"epoch": 4.41,
"grad_norm": 0.7476484179496765,
"learning_rate": 8.529802883200941e-05,
"loss": 3.2402,
"step": 82000
},
{
"epoch": 4.46,
"grad_norm": 0.7432982325553894,
"learning_rate": 8.500382465431009e-05,
"loss": 3.2346,
"step": 83000
},
{
"epoch": 4.52,
"grad_norm": 0.7481192350387573,
"learning_rate": 8.470991468078847e-05,
"loss": 3.2354,
"step": 84000
},
{
"epoch": 4.57,
"grad_norm": 0.7329268455505371,
"learning_rate": 8.441571050308915e-05,
"loss": 3.2349,
"step": 85000
},
{
"epoch": 4.62,
"grad_norm": 0.7217251062393188,
"learning_rate": 8.412150632538982e-05,
"loss": 3.2331,
"step": 86000
},
{
"epoch": 4.68,
"grad_norm": 0.702445387840271,
"learning_rate": 8.382759635186819e-05,
"loss": 3.2324,
"step": 87000
},
{
"epoch": 4.73,
"grad_norm": 0.7485525012016296,
"learning_rate": 8.353339217416887e-05,
"loss": 3.2318,
"step": 88000
},
{
"epoch": 4.79,
"grad_norm": 0.7345052361488342,
"learning_rate": 8.323948220064726e-05,
"loss": 3.2277,
"step": 89000
},
{
"epoch": 4.84,
"grad_norm": 0.7187086343765259,
"learning_rate": 8.294527802294793e-05,
"loss": 3.2262,
"step": 90000
},
{
"epoch": 4.89,
"grad_norm": 0.7242982983589172,
"learning_rate": 8.26510738452486e-05,
"loss": 3.2304,
"step": 91000
},
{
"epoch": 4.95,
"grad_norm": 0.7377817630767822,
"learning_rate": 8.235716387172698e-05,
"loss": 3.2248,
"step": 92000
},
{
"epoch": 5.0,
"eval_accuracy": 0.3919453815222192,
"eval_loss": 3.4727883338928223,
"eval_runtime": 153.4803,
"eval_samples_per_second": 377.364,
"eval_steps_per_second": 5.897,
"step": 92975
},
{
"epoch": 5.0,
"grad_norm": 0.7362983822822571,
"learning_rate": 8.206295969402766e-05,
"loss": 3.2199,
"step": 93000
},
{
"epoch": 5.06,
"grad_norm": 0.7538195848464966,
"learning_rate": 8.176904972050603e-05,
"loss": 3.1732,
"step": 94000
},
{
"epoch": 5.11,
"grad_norm": 0.7742282748222351,
"learning_rate": 8.147484554280671e-05,
"loss": 3.1708,
"step": 95000
},
{
"epoch": 5.16,
"grad_norm": 0.7649984955787659,
"learning_rate": 8.118064136510739e-05,
"loss": 3.17,
"step": 96000
},
{
"epoch": 5.22,
"grad_norm": 0.7696561217308044,
"learning_rate": 8.088643718740806e-05,
"loss": 3.1709,
"step": 97000
},
{
"epoch": 5.27,
"grad_norm": 0.7525352835655212,
"learning_rate": 8.059252721388644e-05,
"loss": 3.1775,
"step": 98000
},
{
"epoch": 5.32,
"grad_norm": 0.7478598952293396,
"learning_rate": 8.029832303618712e-05,
"loss": 3.1765,
"step": 99000
},
{
"epoch": 5.38,
"grad_norm": 0.717166543006897,
"learning_rate": 8.000411885848779e-05,
"loss": 3.1735,
"step": 100000
},
{
"epoch": 5.43,
"grad_norm": 0.743232786655426,
"learning_rate": 7.971020888496618e-05,
"loss": 3.1801,
"step": 101000
},
{
"epoch": 5.49,
"grad_norm": 0.7384679913520813,
"learning_rate": 7.941600470726684e-05,
"loss": 3.1796,
"step": 102000
},
{
"epoch": 5.54,
"grad_norm": 0.7780394554138184,
"learning_rate": 7.912209473374523e-05,
"loss": 3.1745,
"step": 103000
},
{
"epoch": 5.59,
"grad_norm": 0.7568414211273193,
"learning_rate": 7.88281847602236e-05,
"loss": 3.1776,
"step": 104000
},
{
"epoch": 5.65,
"grad_norm": 0.7363548278808594,
"learning_rate": 7.853398058252428e-05,
"loss": 3.176,
"step": 105000
},
{
"epoch": 5.7,
"grad_norm": 0.7693814635276794,
"learning_rate": 7.824007060900265e-05,
"loss": 3.1777,
"step": 106000
},
{
"epoch": 5.75,
"grad_norm": 0.7400150299072266,
"learning_rate": 7.794586643130333e-05,
"loss": 3.1755,
"step": 107000
},
{
"epoch": 5.81,
"grad_norm": 0.7456902861595154,
"learning_rate": 7.76519564577817e-05,
"loss": 3.1769,
"step": 108000
},
{
"epoch": 5.86,
"grad_norm": 0.7355201244354248,
"learning_rate": 7.735775228008238e-05,
"loss": 3.1732,
"step": 109000
},
{
"epoch": 5.92,
"grad_norm": 0.6928702592849731,
"learning_rate": 7.706354810238305e-05,
"loss": 3.1744,
"step": 110000
},
{
"epoch": 5.97,
"grad_norm": 0.7330304384231567,
"learning_rate": 7.676934392468373e-05,
"loss": 3.173,
"step": 111000
},
{
"epoch": 6.0,
"eval_accuracy": 0.3949880570450938,
"eval_loss": 3.467240571975708,
"eval_runtime": 153.7673,
"eval_samples_per_second": 376.66,
"eval_steps_per_second": 5.886,
"step": 111570
},
{
"epoch": 6.02,
"grad_norm": 0.7390985488891602,
"learning_rate": 7.647543395116212e-05,
"loss": 3.1468,
"step": 112000
},
{
"epoch": 6.08,
"grad_norm": 0.775302529335022,
"learning_rate": 7.618152397764049e-05,
"loss": 3.1168,
"step": 113000
},
{
"epoch": 6.13,
"grad_norm": 0.7603143453598022,
"learning_rate": 7.588731979994115e-05,
"loss": 3.1219,
"step": 114000
},
{
"epoch": 6.18,
"grad_norm": 0.7535637617111206,
"learning_rate": 7.559311562224184e-05,
"loss": 3.123,
"step": 115000
},
{
"epoch": 6.24,
"grad_norm": 0.7628033757209778,
"learning_rate": 7.529891144454252e-05,
"loss": 3.1289,
"step": 116000
},
{
"epoch": 6.29,
"grad_norm": 0.7593590617179871,
"learning_rate": 7.50050014710209e-05,
"loss": 3.124,
"step": 117000
},
{
"epoch": 6.35,
"grad_norm": 0.7466164231300354,
"learning_rate": 7.471109149749926e-05,
"loss": 3.126,
"step": 118000
},
{
"epoch": 6.4,
"grad_norm": 0.7527124881744385,
"learning_rate": 7.441688731979994e-05,
"loss": 3.1335,
"step": 119000
},
{
"epoch": 6.45,
"grad_norm": 0.7380790114402771,
"learning_rate": 7.412268314210062e-05,
"loss": 3.1333,
"step": 120000
},
{
"epoch": 6.51,
"grad_norm": 0.7652768492698669,
"learning_rate": 7.382847896440129e-05,
"loss": 3.1307,
"step": 121000
},
{
"epoch": 6.56,
"grad_norm": 0.7151010036468506,
"learning_rate": 7.353456899087967e-05,
"loss": 3.1341,
"step": 122000
},
{
"epoch": 6.61,
"grad_norm": 0.75682532787323,
"learning_rate": 7.324065901735804e-05,
"loss": 3.1289,
"step": 123000
},
{
"epoch": 6.67,
"grad_norm": 0.762443482875824,
"learning_rate": 7.294645483965872e-05,
"loss": 3.13,
"step": 124000
},
{
"epoch": 6.72,
"grad_norm": 0.7441132068634033,
"learning_rate": 7.265225066195939e-05,
"loss": 3.1321,
"step": 125000
},
{
"epoch": 6.78,
"grad_norm": 0.7210931777954102,
"learning_rate": 7.235834068843778e-05,
"loss": 3.1297,
"step": 126000
},
{
"epoch": 6.83,
"grad_norm": 0.757232129573822,
"learning_rate": 7.206413651073846e-05,
"loss": 3.1349,
"step": 127000
},
{
"epoch": 6.88,
"grad_norm": 0.7122140526771545,
"learning_rate": 7.176993233303914e-05,
"loss": 3.1319,
"step": 128000
},
{
"epoch": 6.94,
"grad_norm": 0.7202630639076233,
"learning_rate": 7.14757281553398e-05,
"loss": 3.1335,
"step": 129000
},
{
"epoch": 6.99,
"grad_norm": 0.7444923520088196,
"learning_rate": 7.118181818181818e-05,
"loss": 3.1332,
"step": 130000
},
{
"epoch": 7.0,
"eval_accuracy": 0.39866983551009927,
"eval_loss": 3.424886465072632,
"eval_runtime": 153.9395,
"eval_samples_per_second": 376.239,
"eval_steps_per_second": 5.879,
"step": 130165
},
{
"epoch": 7.04,
"grad_norm": 0.7459632754325867,
"learning_rate": 7.088790820829656e-05,
"loss": 3.0788,
"step": 131000
},
{
"epoch": 7.1,
"grad_norm": 0.781525194644928,
"learning_rate": 7.059370403059724e-05,
"loss": 3.0751,
"step": 132000
},
{
"epoch": 7.15,
"grad_norm": 0.7643530368804932,
"learning_rate": 7.029979405707561e-05,
"loss": 3.0821,
"step": 133000
},
{
"epoch": 7.21,
"grad_norm": 0.7509598731994629,
"learning_rate": 7.000558987937629e-05,
"loss": 3.0821,
"step": 134000
},
{
"epoch": 7.26,
"grad_norm": 0.7690565586090088,
"learning_rate": 6.971138570167697e-05,
"loss": 3.0853,
"step": 135000
},
{
"epoch": 7.31,
"grad_norm": 0.76784348487854,
"learning_rate": 6.941718152397764e-05,
"loss": 3.0896,
"step": 136000
},
{
"epoch": 7.37,
"grad_norm": 0.756600558757782,
"learning_rate": 6.912356575463373e-05,
"loss": 3.0854,
"step": 137000
},
{
"epoch": 7.42,
"grad_norm": 0.7614617347717285,
"learning_rate": 6.88293615769344e-05,
"loss": 3.0915,
"step": 138000
},
{
"epoch": 7.48,
"grad_norm": 0.763688862323761,
"learning_rate": 6.853515739923508e-05,
"loss": 3.0949,
"step": 139000
},
{
"epoch": 7.53,
"grad_norm": 0.7517523169517517,
"learning_rate": 6.824124742571345e-05,
"loss": 3.0958,
"step": 140000
},
{
"epoch": 7.58,
"grad_norm": 0.7513327598571777,
"learning_rate": 6.794704324801413e-05,
"loss": 3.0964,
"step": 141000
},
{
"epoch": 7.64,
"grad_norm": 0.7641260623931885,
"learning_rate": 6.76528390703148e-05,
"loss": 3.0958,
"step": 142000
},
{
"epoch": 7.69,
"grad_norm": 0.7590881586074829,
"learning_rate": 6.735892909679318e-05,
"loss": 3.094,
"step": 143000
},
{
"epoch": 7.74,
"grad_norm": 0.7330321073532104,
"learning_rate": 6.706472491909386e-05,
"loss": 3.0967,
"step": 144000
},
{
"epoch": 7.8,
"grad_norm": 0.749355673789978,
"learning_rate": 6.677052074139453e-05,
"loss": 3.0984,
"step": 145000
},
{
"epoch": 7.85,
"grad_norm": 0.7481942176818848,
"learning_rate": 6.647631656369521e-05,
"loss": 3.0988,
"step": 146000
},
{
"epoch": 7.91,
"grad_norm": 0.7246766090393066,
"learning_rate": 6.618240659017358e-05,
"loss": 3.1007,
"step": 147000
},
{
"epoch": 7.96,
"grad_norm": 0.7423948049545288,
"learning_rate": 6.588849661665197e-05,
"loss": 3.0958,
"step": 148000
},
{
"epoch": 8.0,
"eval_accuracy": 0.3998490150370919,
"eval_loss": 3.4232170581817627,
"eval_runtime": 153.6442,
"eval_samples_per_second": 376.962,
"eval_steps_per_second": 5.89,
"step": 148760
},
{
"epoch": 8.01,
"grad_norm": 0.754693865776062,
"learning_rate": 6.559458664313034e-05,
"loss": 3.0855,
"step": 149000
},
{
"epoch": 8.07,
"grad_norm": 0.7467632293701172,
"learning_rate": 6.5300382465431e-05,
"loss": 3.0395,
"step": 150000
},
{
"epoch": 8.12,
"grad_norm": 0.7670211791992188,
"learning_rate": 6.500617828773169e-05,
"loss": 3.0438,
"step": 151000
},
{
"epoch": 8.17,
"grad_norm": 0.7683054804801941,
"learning_rate": 6.471197411003237e-05,
"loss": 3.046,
"step": 152000
},
{
"epoch": 8.23,
"grad_norm": 0.7673893570899963,
"learning_rate": 6.441776993233303e-05,
"loss": 3.0528,
"step": 153000
},
{
"epoch": 8.28,
"grad_norm": 0.7494843602180481,
"learning_rate": 6.412356575463372e-05,
"loss": 3.0557,
"step": 154000
},
{
"epoch": 8.34,
"grad_norm": 0.7590730786323547,
"learning_rate": 6.38296557811121e-05,
"loss": 3.0553,
"step": 155000
},
{
"epoch": 8.39,
"grad_norm": 0.7913221120834351,
"learning_rate": 6.353545160341278e-05,
"loss": 3.0609,
"step": 156000
},
{
"epoch": 8.44,
"grad_norm": 0.7681291699409485,
"learning_rate": 6.324154162989114e-05,
"loss": 3.0592,
"step": 157000
},
{
"epoch": 8.5,
"grad_norm": 0.7939081788063049,
"learning_rate": 6.294733745219182e-05,
"loss": 3.0597,
"step": 158000
},
{
"epoch": 8.55,
"grad_norm": 0.7624540328979492,
"learning_rate": 6.26534274786702e-05,
"loss": 3.0635,
"step": 159000
},
{
"epoch": 8.6,
"grad_norm": 0.7824980020523071,
"learning_rate": 6.235922330097088e-05,
"loss": 3.0587,
"step": 160000
},
{
"epoch": 8.66,
"grad_norm": 0.777001142501831,
"learning_rate": 6.206531332744924e-05,
"loss": 3.0606,
"step": 161000
},
{
"epoch": 8.71,
"grad_norm": 0.7782078385353088,
"learning_rate": 6.177110914974992e-05,
"loss": 3.0672,
"step": 162000
},
{
"epoch": 8.77,
"grad_norm": 0.7476241588592529,
"learning_rate": 6.14769049720506e-05,
"loss": 3.058,
"step": 163000
},
{
"epoch": 8.82,
"grad_norm": 0.7654275298118591,
"learning_rate": 6.118299499852899e-05,
"loss": 3.0678,
"step": 164000
},
{
"epoch": 8.87,
"grad_norm": 0.7455354928970337,
"learning_rate": 6.0888790820829656e-05,
"loss": 3.0664,
"step": 165000
},
{
"epoch": 8.93,
"grad_norm": 0.7513304948806763,
"learning_rate": 6.059458664313034e-05,
"loss": 3.0674,
"step": 166000
},
{
"epoch": 8.98,
"grad_norm": 0.730146586894989,
"learning_rate": 6.030038246543102e-05,
"loss": 3.0709,
"step": 167000
},
{
"epoch": 9.0,
"eval_accuracy": 0.40124062423958307,
"eval_loss": 3.413806915283203,
"eval_runtime": 153.6351,
"eval_samples_per_second": 376.984,
"eval_steps_per_second": 5.891,
"step": 167355
},
{
"epoch": 9.03,
"grad_norm": 0.7589735388755798,
"learning_rate": 6.0006472491909396e-05,
"loss": 3.03,
"step": 168000
},
{
"epoch": 9.09,
"grad_norm": 0.8093643188476562,
"learning_rate": 5.9712268314210063e-05,
"loss": 3.0117,
"step": 169000
},
{
"epoch": 9.14,
"grad_norm": 0.7802730798721313,
"learning_rate": 5.941835834068844e-05,
"loss": 3.0147,
"step": 170000
},
{
"epoch": 9.2,
"grad_norm": 0.8009030222892761,
"learning_rate": 5.912415416298912e-05,
"loss": 3.0232,
"step": 171000
},
{
"epoch": 9.25,
"grad_norm": 0.7963329553604126,
"learning_rate": 5.88302441894675e-05,
"loss": 3.024,
"step": 172000
},
{
"epoch": 9.3,
"grad_norm": 0.7515285611152649,
"learning_rate": 5.853604001176817e-05,
"loss": 3.025,
"step": 173000
},
{
"epoch": 9.36,
"grad_norm": 0.7860144972801208,
"learning_rate": 5.824183583406885e-05,
"loss": 3.029,
"step": 174000
},
{
"epoch": 9.41,
"grad_norm": 0.766737163066864,
"learning_rate": 5.794763165636953e-05,
"loss": 3.0294,
"step": 175000
},
{
"epoch": 9.46,
"grad_norm": 0.7924851775169373,
"learning_rate": 5.765372168284789e-05,
"loss": 3.0303,
"step": 176000
},
{
"epoch": 9.52,
"grad_norm": 0.805383563041687,
"learning_rate": 5.735981170932627e-05,
"loss": 3.0318,
"step": 177000
},
{
"epoch": 9.57,
"grad_norm": 0.781491756439209,
"learning_rate": 5.706560753162695e-05,
"loss": 3.0328,
"step": 178000
},
{
"epoch": 9.63,
"grad_norm": 0.7737943530082703,
"learning_rate": 5.677169755810533e-05,
"loss": 3.0322,
"step": 179000
},
{
"epoch": 9.68,
"grad_norm": 0.7793260812759399,
"learning_rate": 5.6477493380406e-05,
"loss": 3.0339,
"step": 180000
},
{
"epoch": 9.73,
"grad_norm": 0.7405052185058594,
"learning_rate": 5.6183583406884374e-05,
"loss": 3.0366,
"step": 181000
},
{
"epoch": 9.79,
"grad_norm": 0.798646092414856,
"learning_rate": 5.5889379229185055e-05,
"loss": 3.04,
"step": 182000
},
{
"epoch": 9.84,
"grad_norm": 0.7703911662101746,
"learning_rate": 5.559546925566343e-05,
"loss": 3.0363,
"step": 183000
},
{
"epoch": 9.9,
"grad_norm": 0.7801164984703064,
"learning_rate": 5.530155928214181e-05,
"loss": 3.0383,
"step": 184000
},
{
"epoch": 9.95,
"grad_norm": 0.7496475577354431,
"learning_rate": 5.5007355104442484e-05,
"loss": 3.0426,
"step": 185000
},
{
"epoch": 10.0,
"eval_accuracy": 0.40137438372216294,
"eval_loss": 3.426945924758911,
"eval_runtime": 153.8591,
"eval_samples_per_second": 376.435,
"eval_steps_per_second": 5.882,
"step": 185950
},
{
"epoch": 10.0,
"grad_norm": 0.7879915237426758,
"learning_rate": 5.471315092674316e-05,
"loss": 3.0403,
"step": 186000
},
{
"epoch": 10.06,
"grad_norm": 0.7904853820800781,
"learning_rate": 5.441924095322154e-05,
"loss": 2.9838,
"step": 187000
},
{
"epoch": 10.11,
"grad_norm": 0.7917992472648621,
"learning_rate": 5.412503677552222e-05,
"loss": 2.9899,
"step": 188000
},
{
"epoch": 10.16,
"grad_norm": 0.8023027181625366,
"learning_rate": 5.383083259782289e-05,
"loss": 2.9929,
"step": 189000
},
{
"epoch": 10.22,
"grad_norm": 0.7836917638778687,
"learning_rate": 5.3536628420123566e-05,
"loss": 2.9918,
"step": 190000
},
{
"epoch": 10.27,
"grad_norm": 0.792332112789154,
"learning_rate": 5.3242718446601943e-05,
"loss": 2.9955,
"step": 191000
},
{
"epoch": 10.33,
"grad_norm": 0.8201410174369812,
"learning_rate": 5.294851426890262e-05,
"loss": 3.0028,
"step": 192000
},
{
"epoch": 10.38,
"grad_norm": 0.807855486869812,
"learning_rate": 5.2654604295380995e-05,
"loss": 3.0062,
"step": 193000
},
{
"epoch": 10.43,
"grad_norm": 0.7855746150016785,
"learning_rate": 5.2360400117681677e-05,
"loss": 3.0054,
"step": 194000
},
{
"epoch": 10.49,
"grad_norm": 0.781585156917572,
"learning_rate": 5.206619593998235e-05,
"loss": 3.0082,
"step": 195000
},
{
"epoch": 10.54,
"grad_norm": 0.8157842755317688,
"learning_rate": 5.1771991762283025e-05,
"loss": 3.0068,
"step": 196000
},
{
"epoch": 10.59,
"grad_norm": 0.7841827273368835,
"learning_rate": 5.14780817887614e-05,
"loss": 3.0081,
"step": 197000
},
{
"epoch": 10.65,
"grad_norm": 0.8037407398223877,
"learning_rate": 5.1183877611062084e-05,
"loss": 3.0082,
"step": 198000
},
{
"epoch": 10.7,
"grad_norm": 0.8202306628227234,
"learning_rate": 5.088967343336276e-05,
"loss": 3.0149,
"step": 199000
},
{
"epoch": 10.76,
"grad_norm": 0.7935760021209717,
"learning_rate": 5.059576345984113e-05,
"loss": 3.0136,
"step": 200000
},
{
"epoch": 10.81,
"grad_norm": 0.7972075343132019,
"learning_rate": 5.030155928214181e-05,
"loss": 3.0142,
"step": 201000
},
{
"epoch": 10.86,
"grad_norm": 0.7736737728118896,
"learning_rate": 5.000764930862019e-05,
"loss": 3.0111,
"step": 202000
},
{
"epoch": 10.92,
"grad_norm": 0.7707228660583496,
"learning_rate": 4.971344513092086e-05,
"loss": 3.0161,
"step": 203000
},
{
"epoch": 10.97,
"grad_norm": 0.8111630082130432,
"learning_rate": 4.941924095322154e-05,
"loss": 3.0138,
"step": 204000
},
{
"epoch": 11.0,
"eval_accuracy": 0.4036656950797068,
"eval_loss": 3.402326822280884,
"eval_runtime": 154.1124,
"eval_samples_per_second": 375.817,
"eval_steps_per_second": 5.872,
"step": 204545
},
{
"epoch": 11.02,
"grad_norm": 0.7924003005027771,
"learning_rate": 4.9125330979699914e-05,
"loss": 2.9853,
"step": 205000
},
{
"epoch": 11.08,
"grad_norm": 0.8233866691589355,
"learning_rate": 4.8831126802000595e-05,
"loss": 2.9635,
"step": 206000
},
{
"epoch": 11.13,
"grad_norm": 0.8010469675064087,
"learning_rate": 4.853692262430127e-05,
"loss": 2.9704,
"step": 207000
},
{
"epoch": 11.19,
"grad_norm": 0.8223724365234375,
"learning_rate": 4.824301265077965e-05,
"loss": 2.9697,
"step": 208000
},
{
"epoch": 11.24,
"grad_norm": 0.814317524433136,
"learning_rate": 4.794880847308032e-05,
"loss": 2.9726,
"step": 209000
},
{
"epoch": 11.29,
"grad_norm": 0.8176049590110779,
"learning_rate": 4.76548984995587e-05,
"loss": 2.9755,
"step": 210000
},
{
"epoch": 11.35,
"grad_norm": 0.8245882987976074,
"learning_rate": 4.736069432185937e-05,
"loss": 2.9801,
"step": 211000
},
{
"epoch": 11.4,
"grad_norm": 0.8146276473999023,
"learning_rate": 4.706678434833775e-05,
"loss": 2.9788,
"step": 212000
},
{
"epoch": 11.45,
"grad_norm": 0.8188950419425964,
"learning_rate": 4.6772580170638425e-05,
"loss": 2.9817,
"step": 213000
},
{
"epoch": 11.51,
"grad_norm": 0.7914262413978577,
"learning_rate": 4.64783759929391e-05,
"loss": 2.9856,
"step": 214000
},
{
"epoch": 11.56,
"grad_norm": 0.81853187084198,
"learning_rate": 4.618417181523978e-05,
"loss": 2.9851,
"step": 215000
},
{
"epoch": 11.62,
"grad_norm": 0.8388851881027222,
"learning_rate": 4.5889967637540455e-05,
"loss": 2.9852,
"step": 216000
},
{
"epoch": 11.67,
"grad_norm": 0.8250225782394409,
"learning_rate": 4.559605766401883e-05,
"loss": 2.9838,
"step": 217000
},
{
"epoch": 11.72,
"grad_norm": 0.7764880061149597,
"learning_rate": 4.530185348631951e-05,
"loss": 2.9913,
"step": 218000
},
{
"epoch": 11.78,
"grad_norm": 0.8230167627334595,
"learning_rate": 4.5007943512797884e-05,
"loss": 2.9888,
"step": 219000
},
{
"epoch": 11.83,
"grad_norm": 0.7996993064880371,
"learning_rate": 4.471373933509856e-05,
"loss": 2.9869,
"step": 220000
},
{
"epoch": 11.88,
"grad_norm": 0.8135858178138733,
"learning_rate": 4.441953515739924e-05,
"loss": 2.9924,
"step": 221000
},
{
"epoch": 11.94,
"grad_norm": 0.8289496898651123,
"learning_rate": 4.412562518387761e-05,
"loss": 2.9901,
"step": 222000
},
{
"epoch": 11.99,
"grad_norm": 0.7562257647514343,
"learning_rate": 4.383142100617829e-05,
"loss": 2.995,
"step": 223000
},
{
"epoch": 12.0,
"eval_accuracy": 0.4034856471573994,
"eval_loss": 3.4036989212036133,
"eval_runtime": 153.8908,
"eval_samples_per_second": 376.358,
"eval_steps_per_second": 5.881,
"step": 223140
},
{
"epoch": 12.05,
"grad_norm": 0.8186893463134766,
"learning_rate": 4.3537216828478966e-05,
"loss": 2.9427,
"step": 224000
},
{
"epoch": 12.1,
"grad_norm": 0.861786961555481,
"learning_rate": 4.324330685495734e-05,
"loss": 2.9419,
"step": 225000
},
{
"epoch": 12.15,
"grad_norm": 0.8424795866012573,
"learning_rate": 4.294910267725802e-05,
"loss": 2.9471,
"step": 226000
},
{
"epoch": 12.21,
"grad_norm": 0.822487473487854,
"learning_rate": 4.265489849955869e-05,
"loss": 2.953,
"step": 227000
},
{
"epoch": 12.26,
"grad_norm": 0.8305549025535583,
"learning_rate": 4.2360694321859373e-05,
"loss": 2.951,
"step": 228000
},
{
"epoch": 12.32,
"grad_norm": 0.8566609621047974,
"learning_rate": 4.206678434833775e-05,
"loss": 2.9529,
"step": 229000
},
{
"epoch": 12.37,
"grad_norm": 0.8459467887878418,
"learning_rate": 4.1772580170638425e-05,
"loss": 2.9578,
"step": 230000
},
{
"epoch": 12.42,
"grad_norm": 0.8066169023513794,
"learning_rate": 4.14786701971168e-05,
"loss": 2.965,
"step": 231000
},
{
"epoch": 12.48,
"grad_norm": 0.8078910112380981,
"learning_rate": 4.118446601941748e-05,
"loss": 2.9581,
"step": 232000
},
{
"epoch": 12.53,
"grad_norm": 0.8605197668075562,
"learning_rate": 4.0890556045895854e-05,
"loss": 2.9648,
"step": 233000
},
{
"epoch": 12.58,
"grad_norm": 0.8278983235359192,
"learning_rate": 4.059635186819653e-05,
"loss": 2.9654,
"step": 234000
},
{
"epoch": 12.64,
"grad_norm": 0.7852072715759277,
"learning_rate": 4.03021476904972e-05,
"loss": 2.9638,
"step": 235000
},
{
"epoch": 12.69,
"grad_norm": 0.8360777497291565,
"learning_rate": 4.0007943512797884e-05,
"loss": 2.9643,
"step": 236000
},
{
"epoch": 12.75,
"grad_norm": 0.8392547965049744,
"learning_rate": 3.9714033539276255e-05,
"loss": 2.9719,
"step": 237000
},
{
"epoch": 12.8,
"grad_norm": 0.828360915184021,
"learning_rate": 3.9419829361576936e-05,
"loss": 2.9705,
"step": 238000
},
{
"epoch": 12.85,
"grad_norm": 0.8145458698272705,
"learning_rate": 3.912591938805531e-05,
"loss": 2.9723,
"step": 239000
},
{
"epoch": 12.91,
"grad_norm": 0.837388813495636,
"learning_rate": 3.883200941453369e-05,
"loss": 2.9677,
"step": 240000
},
{
"epoch": 12.96,
"grad_norm": 0.8371777534484863,
"learning_rate": 3.8537805236834365e-05,
"loss": 2.9702,
"step": 241000
},
{
"epoch": 13.0,
"eval_accuracy": 0.4042875322322731,
"eval_loss": 3.3991425037384033,
"eval_runtime": 153.5581,
"eval_samples_per_second": 377.173,
"eval_steps_per_second": 5.894,
"step": 241735
},
{
"epoch": 13.01,
"grad_norm": 0.8215118646621704,
"learning_rate": 3.824360105913504e-05,
"loss": 2.9556,
"step": 242000
},
{
"epoch": 13.07,
"grad_norm": 0.8516760468482971,
"learning_rate": 3.7949396881435714e-05,
"loss": 2.9217,
"step": 243000
},
{
"epoch": 13.12,
"grad_norm": 0.847839891910553,
"learning_rate": 3.765548690791409e-05,
"loss": 2.9274,
"step": 244000
},
{
"epoch": 13.18,
"grad_norm": 0.8394870758056641,
"learning_rate": 3.736128273021477e-05,
"loss": 2.9288,
"step": 245000
},
{
"epoch": 13.23,
"grad_norm": 0.8435803055763245,
"learning_rate": 3.706707855251545e-05,
"loss": 2.9338,
"step": 246000
},
{
"epoch": 13.28,
"grad_norm": 0.8720064163208008,
"learning_rate": 3.677287437481612e-05,
"loss": 2.9338,
"step": 247000
},
{
"epoch": 13.34,
"grad_norm": 0.8536742329597473,
"learning_rate": 3.64789644012945e-05,
"loss": 2.9371,
"step": 248000
},
{
"epoch": 13.39,
"grad_norm": 0.845740020275116,
"learning_rate": 3.6184760223595173e-05,
"loss": 2.9405,
"step": 249000
},
{
"epoch": 13.44,
"grad_norm": 0.8853806257247925,
"learning_rate": 3.589055604589585e-05,
"loss": 2.9408,
"step": 250000
},
{
"epoch": 13.5,
"grad_norm": 0.8576515316963196,
"learning_rate": 3.559635186819653e-05,
"loss": 2.9412,
"step": 251000
},
{
"epoch": 13.55,
"grad_norm": 0.8389966487884521,
"learning_rate": 3.5302441894674906e-05,
"loss": 2.9444,
"step": 252000
},
{
"epoch": 13.61,
"grad_norm": 0.8276740908622742,
"learning_rate": 3.500882612533098e-05,
"loss": 2.9425,
"step": 253000
},
{
"epoch": 13.66,
"grad_norm": 0.8876621127128601,
"learning_rate": 3.471462194763166e-05,
"loss": 2.9441,
"step": 254000
},
{
"epoch": 13.71,
"grad_norm": 0.8480891585350037,
"learning_rate": 3.4420417769932335e-05,
"loss": 2.9461,
"step": 255000
},
{
"epoch": 13.77,
"grad_norm": 0.8812869191169739,
"learning_rate": 3.412621359223302e-05,
"loss": 2.946,
"step": 256000
},
{
"epoch": 13.82,
"grad_norm": 0.8724462985992432,
"learning_rate": 3.383200941453369e-05,
"loss": 2.9463,
"step": 257000
},
{
"epoch": 13.87,
"grad_norm": 0.837191641330719,
"learning_rate": 3.3537805236834366e-05,
"loss": 2.9517,
"step": 258000
},
{
"epoch": 13.93,
"grad_norm": 0.8628780841827393,
"learning_rate": 3.324360105913504e-05,
"loss": 2.9476,
"step": 259000
},
{
"epoch": 13.98,
"grad_norm": 0.8205897808074951,
"learning_rate": 3.294969108561342e-05,
"loss": 2.954,
"step": 260000
},
{
"epoch": 14.0,
"eval_accuracy": 0.4042310321193401,
"eval_loss": 3.4180350303649902,
"eval_runtime": 153.7406,
"eval_samples_per_second": 376.725,
"eval_steps_per_second": 5.887,
"step": 260330
},
{
"epoch": 14.04,
"grad_norm": 0.8445749282836914,
"learning_rate": 3.26554869079141e-05,
"loss": 2.9212,
"step": 261000
},
{
"epoch": 14.09,
"grad_norm": 0.8567079901695251,
"learning_rate": 3.236157693439247e-05,
"loss": 2.9086,
"step": 262000
},
{
"epoch": 14.14,
"grad_norm": 0.8678041100502014,
"learning_rate": 3.2067666960870846e-05,
"loss": 2.9132,
"step": 263000
},
{
"epoch": 14.2,
"grad_norm": 0.8548084497451782,
"learning_rate": 3.177346278317152e-05,
"loss": 2.9131,
"step": 264000
},
{
"epoch": 14.25,
"grad_norm": 0.9028843641281128,
"learning_rate": 3.14792586054722e-05,
"loss": 2.9163,
"step": 265000
},
{
"epoch": 14.3,
"grad_norm": 0.912269651889801,
"learning_rate": 3.118505442777288e-05,
"loss": 2.9202,
"step": 266000
},
{
"epoch": 14.36,
"grad_norm": 0.8899259567260742,
"learning_rate": 3.0891144454251254e-05,
"loss": 2.9205,
"step": 267000
},
{
"epoch": 14.41,
"grad_norm": 0.8662390112876892,
"learning_rate": 3.059694027655193e-05,
"loss": 2.9197,
"step": 268000
},
{
"epoch": 14.47,
"grad_norm": 0.8414704203605652,
"learning_rate": 3.0303030303030306e-05,
"loss": 2.9215,
"step": 269000
},
{
"epoch": 14.52,
"grad_norm": 0.8585665225982666,
"learning_rate": 3.000882612533098e-05,
"loss": 2.9224,
"step": 270000
},
{
"epoch": 14.57,
"grad_norm": 0.8473167419433594,
"learning_rate": 2.971462194763166e-05,
"loss": 2.9261,
"step": 271000
},
{
"epoch": 14.63,
"grad_norm": 0.8617039918899536,
"learning_rate": 2.9420417769932336e-05,
"loss": 2.9276,
"step": 272000
},
{
"epoch": 14.68,
"grad_norm": 0.8747395873069763,
"learning_rate": 2.9126507796410713e-05,
"loss": 2.9274,
"step": 273000
},
{
"epoch": 14.74,
"grad_norm": 0.8597936034202576,
"learning_rate": 2.8832303618711388e-05,
"loss": 2.9252,
"step": 274000
},
{
"epoch": 14.79,
"grad_norm": 0.8555459976196289,
"learning_rate": 2.8538393645189765e-05,
"loss": 2.926,
"step": 275000
},
{
"epoch": 14.84,
"grad_norm": 0.876461386680603,
"learning_rate": 2.824418946749044e-05,
"loss": 2.9302,
"step": 276000
},
{
"epoch": 14.9,
"grad_norm": 0.8898986577987671,
"learning_rate": 2.7950279493968817e-05,
"loss": 2.9278,
"step": 277000
},
{
"epoch": 14.95,
"grad_norm": 0.8596899509429932,
"learning_rate": 2.765607531626949e-05,
"loss": 2.9299,
"step": 278000
},
{
"epoch": 15.0,
"eval_accuracy": 0.4048775922701934,
"eval_loss": 3.4059524536132812,
"eval_runtime": 153.5689,
"eval_samples_per_second": 377.147,
"eval_steps_per_second": 5.893,
"step": 278925
},
{
"epoch": 15.0,
"grad_norm": 0.8894351124763489,
"learning_rate": 2.7361871138570166e-05,
"loss": 2.9266,
"step": 279000
},
{
"epoch": 15.06,
"grad_norm": 0.8618838787078857,
"learning_rate": 2.7067666960870847e-05,
"loss": 2.8883,
"step": 280000
},
{
"epoch": 15.11,
"grad_norm": 0.9119635224342346,
"learning_rate": 2.6773756987349217e-05,
"loss": 2.8918,
"step": 281000
},
{
"epoch": 15.17,
"grad_norm": 0.9044109582901001,
"learning_rate": 2.64795528096499e-05,
"loss": 2.8922,
"step": 282000
},
{
"epoch": 15.22,
"grad_norm": 0.8991778492927551,
"learning_rate": 2.6185348631950573e-05,
"loss": 2.8929,
"step": 283000
},
{
"epoch": 15.27,
"grad_norm": 0.8891802430152893,
"learning_rate": 2.589143865842895e-05,
"loss": 2.8997,
"step": 284000
},
{
"epoch": 15.33,
"grad_norm": 0.9004053473472595,
"learning_rate": 2.5597528684907328e-05,
"loss": 2.9055,
"step": 285000
},
{
"epoch": 15.38,
"grad_norm": 0.895466148853302,
"learning_rate": 2.5303324507208002e-05,
"loss": 2.906,
"step": 286000
},
{
"epoch": 15.43,
"grad_norm": 0.9080808162689209,
"learning_rate": 2.5009414533686383e-05,
"loss": 2.9072,
"step": 287000
},
{
"epoch": 15.49,
"grad_norm": 0.8760538101196289,
"learning_rate": 2.4715210355987057e-05,
"loss": 2.9057,
"step": 288000
},
{
"epoch": 15.54,
"grad_norm": 0.8684205412864685,
"learning_rate": 2.4421006178287735e-05,
"loss": 2.9073,
"step": 289000
},
{
"epoch": 15.6,
"grad_norm": 0.8719987273216248,
"learning_rate": 2.412680200058841e-05,
"loss": 2.9046,
"step": 290000
},
{
"epoch": 15.65,
"grad_norm": 0.8654382824897766,
"learning_rate": 2.3832597822889084e-05,
"loss": 2.9077,
"step": 291000
},
{
"epoch": 15.7,
"grad_norm": 0.8941999673843384,
"learning_rate": 2.3538393645189762e-05,
"loss": 2.9111,
"step": 292000
},
{
"epoch": 15.76,
"grad_norm": 0.8872021436691284,
"learning_rate": 2.324418946749044e-05,
"loss": 2.914,
"step": 293000
},
{
"epoch": 15.81,
"grad_norm": 0.8763778209686279,
"learning_rate": 2.2950573698146517e-05,
"loss": 2.9142,
"step": 294000
},
{
"epoch": 15.86,
"grad_norm": 0.8603065609931946,
"learning_rate": 2.265636952044719e-05,
"loss": 2.9092,
"step": 295000
},
{
"epoch": 15.92,
"grad_norm": 0.8278149366378784,
"learning_rate": 2.236245954692557e-05,
"loss": 2.911,
"step": 296000
},
{
"epoch": 15.97,
"grad_norm": 0.8719545006752014,
"learning_rate": 2.2068255369226243e-05,
"loss": 2.9106,
"step": 297000
},
{
"epoch": 16.0,
"eval_accuracy": 0.4049153485882295,
"eval_loss": 3.4084088802337646,
"eval_runtime": 153.4722,
"eval_samples_per_second": 377.384,
"eval_steps_per_second": 5.897,
"step": 297520
},
{
"epoch": 16.03,
"grad_norm": 0.8873878121376038,
"learning_rate": 2.177405119152692e-05,
"loss": 2.8945,
"step": 298000
},
{
"epoch": 16.08,
"grad_norm": 0.9068606495857239,
"learning_rate": 2.1480141218005298e-05,
"loss": 2.8766,
"step": 299000
},
{
"epoch": 16.13,
"grad_norm": 0.9187499284744263,
"learning_rate": 2.1185937040305972e-05,
"loss": 2.879,
"step": 300000
},
{
"epoch": 16.19,
"grad_norm": 0.8892787098884583,
"learning_rate": 2.089173286260665e-05,
"loss": 2.8821,
"step": 301000
},
{
"epoch": 16.24,
"grad_norm": 0.8787720203399658,
"learning_rate": 2.0597822889085024e-05,
"loss": 2.8863,
"step": 302000
},
{
"epoch": 16.29,
"grad_norm": 0.8958452343940735,
"learning_rate": 2.0303618711385702e-05,
"loss": 2.8852,
"step": 303000
},
{
"epoch": 16.35,
"grad_norm": 0.9086764454841614,
"learning_rate": 2.000941453368638e-05,
"loss": 2.8864,
"step": 304000
},
{
"epoch": 16.4,
"grad_norm": 0.9159351587295532,
"learning_rate": 1.9715210355987058e-05,
"loss": 2.8904,
"step": 305000
},
{
"epoch": 16.46,
"grad_norm": 0.8615668416023254,
"learning_rate": 1.942130038246543e-05,
"loss": 2.886,
"step": 306000
},
{
"epoch": 16.51,
"grad_norm": 0.8897385597229004,
"learning_rate": 1.912709620476611e-05,
"loss": 2.8863,
"step": 307000
},
{
"epoch": 16.56,
"grad_norm": 0.8810350298881531,
"learning_rate": 1.8833186231244483e-05,
"loss": 2.8905,
"step": 308000
},
{
"epoch": 16.62,
"grad_norm": 0.9394462704658508,
"learning_rate": 1.853898205354516e-05,
"loss": 2.8913,
"step": 309000
},
{
"epoch": 16.67,
"grad_norm": 0.8938777446746826,
"learning_rate": 1.824507208002354e-05,
"loss": 2.8938,
"step": 310000
},
{
"epoch": 16.72,
"grad_norm": 0.9033794403076172,
"learning_rate": 1.7950867902324213e-05,
"loss": 2.8934,
"step": 311000
},
{
"epoch": 16.78,
"grad_norm": 0.9354248642921448,
"learning_rate": 1.765695792880259e-05,
"loss": 2.8917,
"step": 312000
},
{
"epoch": 16.83,
"grad_norm": 0.8875927329063416,
"learning_rate": 1.7362753751103268e-05,
"loss": 2.8934,
"step": 313000
},
{
"epoch": 16.89,
"grad_norm": 0.9174014925956726,
"learning_rate": 1.7068549573403943e-05,
"loss": 2.8956,
"step": 314000
},
{
"epoch": 16.94,
"grad_norm": 0.8926222920417786,
"learning_rate": 1.6774345395704617e-05,
"loss": 2.8933,
"step": 315000
},
{
"epoch": 16.99,
"grad_norm": 0.9039079546928406,
"learning_rate": 1.6480435422182998e-05,
"loss": 2.8923,
"step": 316000
},
{
"epoch": 17.0,
"eval_accuracy": 0.4055247570595917,
"eval_loss": 3.4154162406921387,
"eval_runtime": 153.7223,
"eval_samples_per_second": 376.77,
"eval_steps_per_second": 5.887,
"step": 316115
},
{
"epoch": 17.05,
"grad_norm": 0.93231600522995,
"learning_rate": 1.6186231244483672e-05,
"loss": 2.8687,
"step": 317000
},
{
"epoch": 17.1,
"grad_norm": 0.8828312158584595,
"learning_rate": 1.589202706678435e-05,
"loss": 2.865,
"step": 318000
},
{
"epoch": 17.16,
"grad_norm": 0.9261276721954346,
"learning_rate": 1.5598117093262727e-05,
"loss": 2.8687,
"step": 319000
},
{
"epoch": 17.21,
"grad_norm": 0.9356419444084167,
"learning_rate": 1.5303912915563402e-05,
"loss": 2.8681,
"step": 320000
},
{
"epoch": 17.26,
"grad_norm": 0.9402567744255066,
"learning_rate": 1.5009708737864078e-05,
"loss": 2.8676,
"step": 321000
},
{
"epoch": 17.32,
"grad_norm": 0.9250918626785278,
"learning_rate": 1.4715504560164756e-05,
"loss": 2.8729,
"step": 322000
},
{
"epoch": 17.37,
"grad_norm": 0.9387281537055969,
"learning_rate": 1.442159458664313e-05,
"loss": 2.8744,
"step": 323000
},
{
"epoch": 17.42,
"grad_norm": 0.9035524725914001,
"learning_rate": 1.4127390408943808e-05,
"loss": 2.8718,
"step": 324000
},
{
"epoch": 17.48,
"grad_norm": 0.8882308602333069,
"learning_rate": 1.3833480435422183e-05,
"loss": 2.8704,
"step": 325000
},
{
"epoch": 17.53,
"grad_norm": 0.9479586482048035,
"learning_rate": 1.3539276257722861e-05,
"loss": 2.8766,
"step": 326000
},
{
"epoch": 17.59,
"grad_norm": 0.8966055512428284,
"learning_rate": 1.3245072080023537e-05,
"loss": 2.874,
"step": 327000
},
{
"epoch": 17.64,
"grad_norm": 0.9272529482841492,
"learning_rate": 1.2951162106501913e-05,
"loss": 2.8737,
"step": 328000
},
{
"epoch": 17.69,
"grad_norm": 0.9414493441581726,
"learning_rate": 1.2657252132980289e-05,
"loss": 2.8767,
"step": 329000
},
{
"epoch": 17.75,
"grad_norm": 0.9303755760192871,
"learning_rate": 1.2363047955280965e-05,
"loss": 2.8767,
"step": 330000
},
{
"epoch": 17.8,
"grad_norm": 0.9229914546012878,
"learning_rate": 1.2068843777581643e-05,
"loss": 2.8778,
"step": 331000
},
{
"epoch": 17.85,
"grad_norm": 0.9260939955711365,
"learning_rate": 1.1774639599882319e-05,
"loss": 2.8765,
"step": 332000
},
{
"epoch": 17.91,
"grad_norm": 0.9214343428611755,
"learning_rate": 1.1480729626360694e-05,
"loss": 2.8752,
"step": 333000
},
{
"epoch": 17.96,
"grad_norm": 0.9373890161514282,
"learning_rate": 1.118652544866137e-05,
"loss": 2.8795,
"step": 334000
},
{
"epoch": 18.0,
"eval_accuracy": 0.40571911476079897,
"eval_loss": 3.419478178024292,
"eval_runtime": 153.6818,
"eval_samples_per_second": 376.87,
"eval_steps_per_second": 5.889,
"step": 334710
},
{
"epoch": 18.02,
"grad_norm": 0.9543386101722717,
"learning_rate": 1.0892321270962048e-05,
"loss": 2.8708,
"step": 335000
},
{
"epoch": 18.07,
"grad_norm": 0.893997311592102,
"learning_rate": 1.0598411297440426e-05,
"loss": 2.8512,
"step": 336000
},
{
"epoch": 18.12,
"grad_norm": 0.9292035102844238,
"learning_rate": 1.03045013239188e-05,
"loss": 2.8533,
"step": 337000
},
{
"epoch": 18.18,
"grad_norm": 0.9333125948905945,
"learning_rate": 1.0010297146219477e-05,
"loss": 2.855,
"step": 338000
},
{
"epoch": 18.23,
"grad_norm": 0.9297951459884644,
"learning_rate": 9.716092968520154e-06,
"loss": 2.8587,
"step": 339000
},
{
"epoch": 18.28,
"grad_norm": 0.9800245761871338,
"learning_rate": 9.421888790820831e-06,
"loss": 2.8595,
"step": 340000
},
{
"epoch": 18.34,
"grad_norm": 0.9462539553642273,
"learning_rate": 9.127684613121506e-06,
"loss": 2.8594,
"step": 341000
},
{
"epoch": 18.39,
"grad_norm": 0.9382343888282776,
"learning_rate": 8.833774639599883e-06,
"loss": 2.8603,
"step": 342000
},
{
"epoch": 18.45,
"grad_norm": 0.9466389417648315,
"learning_rate": 8.539864666078259e-06,
"loss": 2.8575,
"step": 343000
},
{
"epoch": 18.5,
"grad_norm": 0.9292024970054626,
"learning_rate": 8.245660488378935e-06,
"loss": 2.8561,
"step": 344000
},
{
"epoch": 18.55,
"grad_norm": 0.9281138181686401,
"learning_rate": 7.951456310679611e-06,
"loss": 2.859,
"step": 345000
},
{
"epoch": 18.61,
"grad_norm": 0.9447932839393616,
"learning_rate": 7.657546337157988e-06,
"loss": 2.8596,
"step": 346000
},
{
"epoch": 18.66,
"grad_norm": 0.9081346988677979,
"learning_rate": 7.363342159458665e-06,
"loss": 2.8632,
"step": 347000
},
{
"epoch": 18.71,
"grad_norm": 0.924645185470581,
"learning_rate": 7.0691379817593416e-06,
"loss": 2.8609,
"step": 348000
},
{
"epoch": 18.77,
"grad_norm": 0.9200485944747925,
"learning_rate": 6.774933804060019e-06,
"loss": 2.8629,
"step": 349000
},
{
"epoch": 18.82,
"grad_norm": 0.9209837317466736,
"learning_rate": 6.481023830538393e-06,
"loss": 2.8599,
"step": 350000
},
{
"epoch": 18.88,
"grad_norm": 0.9589852094650269,
"learning_rate": 6.186819652839071e-06,
"loss": 2.8617,
"step": 351000
},
{
"epoch": 18.93,
"grad_norm": 0.896015465259552,
"learning_rate": 5.892615475139747e-06,
"loss": 2.8612,
"step": 352000
},
{
"epoch": 18.98,
"grad_norm": 0.971016526222229,
"learning_rate": 5.598705501618123e-06,
"loss": 2.8628,
"step": 353000
},
{
"epoch": 19.0,
"eval_accuracy": 0.40566355519671377,
"eval_loss": 3.422478199005127,
"eval_runtime": 153.9588,
"eval_samples_per_second": 376.192,
"eval_steps_per_second": 5.878,
"step": 353305
},
{
"epoch": 19.04,
"grad_norm": 0.9984861612319946,
"learning_rate": 5.304795528096499e-06,
"loss": 2.8502,
"step": 354000
},
{
"epoch": 19.09,
"grad_norm": 0.9807627201080322,
"learning_rate": 5.0105913503971756e-06,
"loss": 2.8423,
"step": 355000
},
{
"epoch": 19.14,
"grad_norm": 0.9239267110824585,
"learning_rate": 4.7163871726978526e-06,
"loss": 2.8449,
"step": 356000
},
{
"epoch": 19.2,
"grad_norm": 0.9425461292266846,
"learning_rate": 4.422182994998529e-06,
"loss": 2.8478,
"step": 357000
},
{
"epoch": 19.25,
"grad_norm": 0.9452289938926697,
"learning_rate": 4.128273021476905e-06,
"loss": 2.8426,
"step": 358000
},
{
"epoch": 19.31,
"grad_norm": 0.8920537233352661,
"learning_rate": 3.834068843777581e-06,
"loss": 2.8425,
"step": 359000
},
{
"epoch": 19.36,
"grad_norm": 0.968367338180542,
"learning_rate": 3.5401588702559578e-06,
"loss": 2.8503,
"step": 360000
},
{
"epoch": 19.41,
"grad_norm": 0.9560465216636658,
"learning_rate": 3.245954692556635e-06,
"loss": 2.8474,
"step": 361000
},
{
"epoch": 19.47,
"grad_norm": 0.9240676760673523,
"learning_rate": 2.951750514857311e-06,
"loss": 2.8452,
"step": 362000
},
{
"epoch": 19.52,
"grad_norm": 0.9011372327804565,
"learning_rate": 2.657840541335687e-06,
"loss": 2.8438,
"step": 363000
},
{
"epoch": 19.58,
"grad_norm": 0.9103080034255981,
"learning_rate": 2.3636363636363636e-06,
"loss": 2.85,
"step": 364000
},
{
"epoch": 19.63,
"grad_norm": 0.9446433782577515,
"learning_rate": 2.070020594292439e-06,
"loss": 2.848,
"step": 365000
},
{
"epoch": 19.68,
"grad_norm": 0.927521824836731,
"learning_rate": 1.7758164165931158e-06,
"loss": 2.8497,
"step": 366000
},
{
"epoch": 19.74,
"grad_norm": 0.9345549941062927,
"learning_rate": 1.4816122388937924e-06,
"loss": 2.8488,
"step": 367000
},
{
"epoch": 19.79,
"grad_norm": 0.9598424434661865,
"learning_rate": 1.187408061194469e-06,
"loss": 2.8465,
"step": 368000
},
{
"epoch": 19.84,
"grad_norm": 0.9485284090042114,
"learning_rate": 8.932038834951457e-07,
"loss": 2.8497,
"step": 369000
},
{
"epoch": 19.9,
"grad_norm": 0.9556813836097717,
"learning_rate": 5.992939099735216e-07,
"loss": 2.8439,
"step": 370000
},
{
"epoch": 19.95,
"grad_norm": 0.9353678226470947,
"learning_rate": 3.050897322741983e-07,
"loss": 2.8497,
"step": 371000
},
{
"epoch": 20.0,
"eval_accuracy": 0.4055823320854937,
"eval_loss": 3.4263744354248047,
"eval_runtime": 153.7409,
"eval_samples_per_second": 376.725,
"eval_steps_per_second": 5.887,
"step": 371900
},
{
"epoch": 20.0,
"step": 371900,
"total_flos": 1.56693365388288e+18,
"train_loss": 3.1576745831950372,
"train_runtime": 81271.0608,
"train_samples_per_second": 146.433,
"train_steps_per_second": 4.576
}
],
"logging_steps": 1000,
"max_steps": 371900,
"num_input_tokens_seen": 0,
"num_train_epochs": 20,
"save_steps": 5000,
"total_flos": 1.56693365388288e+18,
"train_batch_size": 32,
"trial_name": null,
"trial_params": null
}