kanishka's picture
End of training
1c2dcfa verified
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 20.0,
"eval_steps": 500,
"global_step": 372040,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.05,
"grad_norm": 1.1385914087295532,
"learning_rate": 3.125e-06,
"loss": 7.5337,
"step": 1000
},
{
"epoch": 0.11,
"grad_norm": 0.8339954614639282,
"learning_rate": 6.25e-06,
"loss": 5.7974,
"step": 2000
},
{
"epoch": 0.16,
"grad_norm": 0.7770102024078369,
"learning_rate": 9.375000000000001e-06,
"loss": 5.3786,
"step": 3000
},
{
"epoch": 0.22,
"grad_norm": 0.8822804689407349,
"learning_rate": 1.25e-05,
"loss": 5.1546,
"step": 4000
},
{
"epoch": 0.27,
"grad_norm": 0.9704174995422363,
"learning_rate": 1.5625e-05,
"loss": 5.0064,
"step": 5000
},
{
"epoch": 0.32,
"grad_norm": 1.011856198310852,
"learning_rate": 1.8750000000000002e-05,
"loss": 4.8697,
"step": 6000
},
{
"epoch": 0.38,
"grad_norm": 1.127143383026123,
"learning_rate": 2.1875e-05,
"loss": 4.7448,
"step": 7000
},
{
"epoch": 0.43,
"grad_norm": 1.1060676574707031,
"learning_rate": 2.5e-05,
"loss": 4.6483,
"step": 8000
},
{
"epoch": 0.48,
"grad_norm": 1.0898035764694214,
"learning_rate": 2.8125000000000003e-05,
"loss": 4.554,
"step": 9000
},
{
"epoch": 0.54,
"grad_norm": 1.0667864084243774,
"learning_rate": 3.125e-05,
"loss": 4.4709,
"step": 10000
},
{
"epoch": 0.59,
"grad_norm": 1.0186491012573242,
"learning_rate": 3.4375e-05,
"loss": 4.4155,
"step": 11000
},
{
"epoch": 0.65,
"grad_norm": 1.0487744808197021,
"learning_rate": 3.7496875e-05,
"loss": 4.3425,
"step": 12000
},
{
"epoch": 0.7,
"grad_norm": 1.0958940982818604,
"learning_rate": 4.0621875e-05,
"loss": 4.2827,
"step": 13000
},
{
"epoch": 0.75,
"grad_norm": 1.0781230926513672,
"learning_rate": 4.374375e-05,
"loss": 4.2321,
"step": 14000
},
{
"epoch": 0.81,
"grad_norm": 1.0080424547195435,
"learning_rate": 4.686875e-05,
"loss": 4.1839,
"step": 15000
},
{
"epoch": 0.86,
"grad_norm": 1.0729751586914062,
"learning_rate": 4.999375e-05,
"loss": 4.1402,
"step": 16000
},
{
"epoch": 0.91,
"grad_norm": 0.9728330373764038,
"learning_rate": 5.311875000000001e-05,
"loss": 4.0886,
"step": 17000
},
{
"epoch": 0.97,
"grad_norm": 0.982369065284729,
"learning_rate": 5.6240625e-05,
"loss": 4.0517,
"step": 18000
},
{
"epoch": 1.0,
"eval_accuracy": 0.30857265198935596,
"eval_loss": 4.261654376983643,
"eval_runtime": 152.7298,
"eval_samples_per_second": 379.238,
"eval_steps_per_second": 5.932,
"step": 18602
},
{
"epoch": 1.02,
"grad_norm": 0.9907498955726624,
"learning_rate": 5.93625e-05,
"loss": 4.0017,
"step": 19000
},
{
"epoch": 1.08,
"grad_norm": 1.011263370513916,
"learning_rate": 6.24875e-05,
"loss": 3.9591,
"step": 20000
},
{
"epoch": 1.13,
"grad_norm": 0.9945884346961975,
"learning_rate": 6.56125e-05,
"loss": 3.919,
"step": 21000
},
{
"epoch": 1.18,
"grad_norm": 0.9863882064819336,
"learning_rate": 6.8734375e-05,
"loss": 3.8924,
"step": 22000
},
{
"epoch": 1.24,
"grad_norm": 1.0095977783203125,
"learning_rate": 7.185625e-05,
"loss": 3.8609,
"step": 23000
},
{
"epoch": 1.29,
"grad_norm": 0.9618180394172668,
"learning_rate": 7.498125e-05,
"loss": 3.8246,
"step": 24000
},
{
"epoch": 1.34,
"grad_norm": 1.0228387117385864,
"learning_rate": 7.8103125e-05,
"loss": 3.7991,
"step": 25000
},
{
"epoch": 1.4,
"grad_norm": 0.953283965587616,
"learning_rate": 8.122500000000001e-05,
"loss": 3.7693,
"step": 26000
},
{
"epoch": 1.45,
"grad_norm": 1.013651967048645,
"learning_rate": 8.435e-05,
"loss": 3.7465,
"step": 27000
},
{
"epoch": 1.51,
"grad_norm": 0.9061474800109863,
"learning_rate": 8.7471875e-05,
"loss": 3.7264,
"step": 28000
},
{
"epoch": 1.56,
"grad_norm": 0.892005205154419,
"learning_rate": 9.0596875e-05,
"loss": 3.7033,
"step": 29000
},
{
"epoch": 1.61,
"grad_norm": 0.8759484887123108,
"learning_rate": 9.3721875e-05,
"loss": 3.6875,
"step": 30000
},
{
"epoch": 1.67,
"grad_norm": 0.8649481534957886,
"learning_rate": 9.684375e-05,
"loss": 3.6669,
"step": 31000
},
{
"epoch": 1.72,
"grad_norm": 0.9080563187599182,
"learning_rate": 9.996875e-05,
"loss": 3.655,
"step": 32000
},
{
"epoch": 1.77,
"grad_norm": 0.8319312930107117,
"learning_rate": 9.970885778143747e-05,
"loss": 3.6316,
"step": 33000
},
{
"epoch": 1.83,
"grad_norm": 0.8593823909759521,
"learning_rate": 9.941506881543349e-05,
"loss": 3.618,
"step": 34000
},
{
"epoch": 1.88,
"grad_norm": 0.8492996692657471,
"learning_rate": 9.912098576638043e-05,
"loss": 3.5981,
"step": 35000
},
{
"epoch": 1.94,
"grad_norm": 0.8345481753349304,
"learning_rate": 9.882690271732737e-05,
"loss": 3.5809,
"step": 36000
},
{
"epoch": 1.99,
"grad_norm": 0.92176753282547,
"learning_rate": 9.853281966827432e-05,
"loss": 3.5614,
"step": 37000
},
{
"epoch": 2.0,
"eval_accuracy": 0.36166690750187913,
"eval_loss": 3.7325448989868164,
"eval_runtime": 154.634,
"eval_samples_per_second": 374.568,
"eval_steps_per_second": 5.859,
"step": 37204
},
{
"epoch": 2.04,
"grad_norm": 0.8371630907058716,
"learning_rate": 9.823903070227034e-05,
"loss": 3.5263,
"step": 38000
},
{
"epoch": 2.1,
"grad_norm": 0.8102852702140808,
"learning_rate": 9.794494765321727e-05,
"loss": 3.5146,
"step": 39000
},
{
"epoch": 2.15,
"grad_norm": 0.8165059089660645,
"learning_rate": 9.765086460416421e-05,
"loss": 3.504,
"step": 40000
},
{
"epoch": 2.2,
"grad_norm": 0.8198731541633606,
"learning_rate": 9.735707563816023e-05,
"loss": 3.4944,
"step": 41000
},
{
"epoch": 2.26,
"grad_norm": 0.8228951096534729,
"learning_rate": 9.706299258910717e-05,
"loss": 3.4795,
"step": 42000
},
{
"epoch": 2.31,
"grad_norm": 0.8009123802185059,
"learning_rate": 9.676890954005412e-05,
"loss": 3.4731,
"step": 43000
},
{
"epoch": 2.37,
"grad_norm": 0.7662414312362671,
"learning_rate": 9.647512057405012e-05,
"loss": 3.4634,
"step": 44000
},
{
"epoch": 2.42,
"grad_norm": 0.7814858555793762,
"learning_rate": 9.618103752499706e-05,
"loss": 3.4628,
"step": 45000
},
{
"epoch": 2.47,
"grad_norm": 0.7715092301368713,
"learning_rate": 9.588695447594401e-05,
"loss": 3.451,
"step": 46000
},
{
"epoch": 2.53,
"grad_norm": 0.7964122891426086,
"learning_rate": 9.559287142689096e-05,
"loss": 3.4396,
"step": 47000
},
{
"epoch": 2.58,
"grad_norm": 0.7903115153312683,
"learning_rate": 9.529908246088695e-05,
"loss": 3.4363,
"step": 48000
},
{
"epoch": 2.63,
"grad_norm": 0.7684825658798218,
"learning_rate": 9.50049994118339e-05,
"loss": 3.4255,
"step": 49000
},
{
"epoch": 2.69,
"grad_norm": 0.774431049823761,
"learning_rate": 9.47112104458299e-05,
"loss": 3.4233,
"step": 50000
},
{
"epoch": 2.74,
"grad_norm": 0.7562047839164734,
"learning_rate": 9.441712739677684e-05,
"loss": 3.407,
"step": 51000
},
{
"epoch": 2.8,
"grad_norm": 0.8174190521240234,
"learning_rate": 9.412333843077286e-05,
"loss": 3.4048,
"step": 52000
},
{
"epoch": 2.85,
"grad_norm": 0.7771363258361816,
"learning_rate": 9.382954946476885e-05,
"loss": 3.3993,
"step": 53000
},
{
"epoch": 2.9,
"grad_norm": 0.8073471188545227,
"learning_rate": 9.35354664157158e-05,
"loss": 3.3877,
"step": 54000
},
{
"epoch": 2.96,
"grad_norm": 0.7597507834434509,
"learning_rate": 9.324138336666275e-05,
"loss": 3.3871,
"step": 55000
},
{
"epoch": 3.0,
"eval_accuracy": 0.3793839146396705,
"eval_loss": 3.5926387310028076,
"eval_runtime": 154.4428,
"eval_samples_per_second": 375.032,
"eval_steps_per_second": 5.866,
"step": 55806
},
{
"epoch": 3.01,
"grad_norm": 0.7489961385726929,
"learning_rate": 9.294730031760969e-05,
"loss": 3.3712,
"step": 56000
},
{
"epoch": 3.06,
"grad_norm": 0.7571219205856323,
"learning_rate": 9.26535113516057e-05,
"loss": 3.328,
"step": 57000
},
{
"epoch": 3.12,
"grad_norm": 0.739839494228363,
"learning_rate": 9.23597223856017e-05,
"loss": 3.3289,
"step": 58000
},
{
"epoch": 3.17,
"grad_norm": 0.7352610230445862,
"learning_rate": 9.206563933654865e-05,
"loss": 3.328,
"step": 59000
},
{
"epoch": 3.23,
"grad_norm": 0.7837926745414734,
"learning_rate": 9.17715562874956e-05,
"loss": 3.3302,
"step": 60000
},
{
"epoch": 3.28,
"grad_norm": 0.7600297927856445,
"learning_rate": 9.147747323844255e-05,
"loss": 3.321,
"step": 61000
},
{
"epoch": 3.33,
"grad_norm": 0.7775222063064575,
"learning_rate": 9.118368427243854e-05,
"loss": 3.3186,
"step": 62000
},
{
"epoch": 3.39,
"grad_norm": 0.7667526006698608,
"learning_rate": 9.088960122338549e-05,
"loss": 3.321,
"step": 63000
},
{
"epoch": 3.44,
"grad_norm": 0.75556880235672,
"learning_rate": 9.059551817433244e-05,
"loss": 3.3154,
"step": 64000
},
{
"epoch": 3.49,
"grad_norm": 0.7374812364578247,
"learning_rate": 9.030172920832843e-05,
"loss": 3.3126,
"step": 65000
},
{
"epoch": 3.55,
"grad_norm": 0.7322445511817932,
"learning_rate": 9.000764615927539e-05,
"loss": 3.3114,
"step": 66000
},
{
"epoch": 3.6,
"grad_norm": 0.7468699812889099,
"learning_rate": 8.971385719327139e-05,
"loss": 3.3054,
"step": 67000
},
{
"epoch": 3.66,
"grad_norm": 0.7796970009803772,
"learning_rate": 8.941977414421833e-05,
"loss": 3.3025,
"step": 68000
},
{
"epoch": 3.71,
"grad_norm": 0.7635445594787598,
"learning_rate": 8.912598517821433e-05,
"loss": 3.2961,
"step": 69000
},
{
"epoch": 3.76,
"grad_norm": 0.7578097581863403,
"learning_rate": 8.883190212916128e-05,
"loss": 3.2999,
"step": 70000
},
{
"epoch": 3.82,
"grad_norm": 0.7045617699623108,
"learning_rate": 8.853781908010823e-05,
"loss": 3.2919,
"step": 71000
},
{
"epoch": 3.87,
"grad_norm": 0.7356869578361511,
"learning_rate": 8.824373603105518e-05,
"loss": 3.2901,
"step": 72000
},
{
"epoch": 3.92,
"grad_norm": 0.7233588099479675,
"learning_rate": 8.794994706505117e-05,
"loss": 3.2872,
"step": 73000
},
{
"epoch": 3.98,
"grad_norm": 0.7042175531387329,
"learning_rate": 8.765615809904718e-05,
"loss": 3.2873,
"step": 74000
},
{
"epoch": 4.0,
"eval_accuracy": 0.38890029805120985,
"eval_loss": 3.490251064300537,
"eval_runtime": 153.9343,
"eval_samples_per_second": 376.271,
"eval_steps_per_second": 5.886,
"step": 74408
},
{
"epoch": 4.03,
"grad_norm": 0.7667938470840454,
"learning_rate": 8.736207504999411e-05,
"loss": 3.2467,
"step": 75000
},
{
"epoch": 4.09,
"grad_norm": 0.7835637927055359,
"learning_rate": 8.706799200094107e-05,
"loss": 3.2311,
"step": 76000
},
{
"epoch": 4.14,
"grad_norm": 0.773997962474823,
"learning_rate": 8.677390895188802e-05,
"loss": 3.2309,
"step": 77000
},
{
"epoch": 4.19,
"grad_norm": 0.7710655331611633,
"learning_rate": 8.648011998588402e-05,
"loss": 3.2363,
"step": 78000
},
{
"epoch": 4.25,
"grad_norm": 0.7279992699623108,
"learning_rate": 8.618633101988002e-05,
"loss": 3.232,
"step": 79000
},
{
"epoch": 4.3,
"grad_norm": 0.7493019700050354,
"learning_rate": 8.589224797082696e-05,
"loss": 3.2336,
"step": 80000
},
{
"epoch": 4.35,
"grad_norm": 0.7595520615577698,
"learning_rate": 8.559816492177391e-05,
"loss": 3.2314,
"step": 81000
},
{
"epoch": 4.41,
"grad_norm": 0.7674165368080139,
"learning_rate": 8.530408187272086e-05,
"loss": 3.2277,
"step": 82000
},
{
"epoch": 4.46,
"grad_norm": 0.7552059292793274,
"learning_rate": 8.501029290671685e-05,
"loss": 3.2294,
"step": 83000
},
{
"epoch": 4.52,
"grad_norm": 0.7497634291648865,
"learning_rate": 8.471650394071287e-05,
"loss": 3.2297,
"step": 84000
},
{
"epoch": 4.57,
"grad_norm": 0.7449929714202881,
"learning_rate": 8.442242089165981e-05,
"loss": 3.2281,
"step": 85000
},
{
"epoch": 4.62,
"grad_norm": 0.7487680315971375,
"learning_rate": 8.412833784260675e-05,
"loss": 3.2247,
"step": 86000
},
{
"epoch": 4.68,
"grad_norm": 0.7540663480758667,
"learning_rate": 8.38342547935537e-05,
"loss": 3.2283,
"step": 87000
},
{
"epoch": 4.73,
"grad_norm": 0.7613685727119446,
"learning_rate": 8.354046582754971e-05,
"loss": 3.2239,
"step": 88000
},
{
"epoch": 4.78,
"grad_norm": 0.7112619876861572,
"learning_rate": 8.324638277849665e-05,
"loss": 3.2241,
"step": 89000
},
{
"epoch": 4.84,
"grad_norm": 0.7479850649833679,
"learning_rate": 8.295229972944359e-05,
"loss": 3.2193,
"step": 90000
},
{
"epoch": 4.89,
"grad_norm": 0.7664503455162048,
"learning_rate": 8.265821668039054e-05,
"loss": 3.2191,
"step": 91000
},
{
"epoch": 4.95,
"grad_norm": 0.7236024737358093,
"learning_rate": 8.23641336313375e-05,
"loss": 3.2178,
"step": 92000
},
{
"epoch": 5.0,
"grad_norm": 0.7357888221740723,
"learning_rate": 8.20703446653335e-05,
"loss": 3.2166,
"step": 93000
},
{
"epoch": 5.0,
"eval_accuracy": 0.39298697266241545,
"eval_loss": 3.4704902172088623,
"eval_runtime": 154.103,
"eval_samples_per_second": 375.859,
"eval_steps_per_second": 5.879,
"step": 93010
},
{
"epoch": 5.05,
"grad_norm": 0.7487916946411133,
"learning_rate": 8.177626161628044e-05,
"loss": 3.1601,
"step": 94000
},
{
"epoch": 5.11,
"grad_norm": 0.7451751232147217,
"learning_rate": 8.148247265027644e-05,
"loss": 3.1606,
"step": 95000
},
{
"epoch": 5.16,
"grad_norm": 0.7442896962165833,
"learning_rate": 8.118838960122339e-05,
"loss": 3.1632,
"step": 96000
},
{
"epoch": 5.21,
"grad_norm": 0.9075034260749817,
"learning_rate": 8.089430655217034e-05,
"loss": 3.1711,
"step": 97000
},
{
"epoch": 5.27,
"grad_norm": 0.7450466752052307,
"learning_rate": 8.06002235031173e-05,
"loss": 3.1694,
"step": 98000
},
{
"epoch": 5.32,
"grad_norm": 0.752282977104187,
"learning_rate": 8.030643453711328e-05,
"loss": 3.1673,
"step": 99000
},
{
"epoch": 5.38,
"grad_norm": 0.742364764213562,
"learning_rate": 8.001235148806023e-05,
"loss": 3.1734,
"step": 100000
},
{
"epoch": 5.43,
"grad_norm": 0.7386059165000916,
"learning_rate": 7.971826843900719e-05,
"loss": 3.1743,
"step": 101000
},
{
"epoch": 5.48,
"grad_norm": 0.7800270915031433,
"learning_rate": 7.942447947300318e-05,
"loss": 3.1678,
"step": 102000
},
{
"epoch": 5.54,
"grad_norm": 0.7261184453964233,
"learning_rate": 7.913039642395013e-05,
"loss": 3.1706,
"step": 103000
},
{
"epoch": 5.59,
"grad_norm": 0.7511762976646423,
"learning_rate": 7.883631337489708e-05,
"loss": 3.1692,
"step": 104000
},
{
"epoch": 5.64,
"grad_norm": 0.7737910151481628,
"learning_rate": 7.854223032584402e-05,
"loss": 3.1706,
"step": 105000
},
{
"epoch": 5.7,
"grad_norm": 0.7110986113548279,
"learning_rate": 7.824844135984002e-05,
"loss": 3.1682,
"step": 106000
},
{
"epoch": 5.75,
"grad_norm": 0.7741941213607788,
"learning_rate": 7.795435831078697e-05,
"loss": 3.1632,
"step": 107000
},
{
"epoch": 5.81,
"grad_norm": 0.7159838676452637,
"learning_rate": 7.766056934478297e-05,
"loss": 3.1684,
"step": 108000
},
{
"epoch": 5.86,
"grad_norm": 0.7344425916671753,
"learning_rate": 7.736648629572993e-05,
"loss": 3.1628,
"step": 109000
},
{
"epoch": 5.91,
"grad_norm": 0.7268814444541931,
"learning_rate": 7.707240324667686e-05,
"loss": 3.1657,
"step": 110000
},
{
"epoch": 5.97,
"grad_norm": 0.7119531631469727,
"learning_rate": 7.677861428067287e-05,
"loss": 3.1683,
"step": 111000
},
{
"epoch": 6.0,
"eval_accuracy": 0.3964901341200214,
"eval_loss": 3.4386308193206787,
"eval_runtime": 154.2891,
"eval_samples_per_second": 375.406,
"eval_steps_per_second": 5.872,
"step": 111612
},
{
"epoch": 6.02,
"grad_norm": 0.7270282506942749,
"learning_rate": 7.648482531466886e-05,
"loss": 3.1402,
"step": 112000
},
{
"epoch": 6.07,
"grad_norm": 0.759936511516571,
"learning_rate": 7.619074226561581e-05,
"loss": 3.1064,
"step": 113000
},
{
"epoch": 6.13,
"grad_norm": 0.7575330138206482,
"learning_rate": 7.589665921656276e-05,
"loss": 3.1131,
"step": 114000
},
{
"epoch": 6.18,
"grad_norm": 0.7685191631317139,
"learning_rate": 7.560257616750971e-05,
"loss": 3.1157,
"step": 115000
},
{
"epoch": 6.24,
"grad_norm": 0.7537042498588562,
"learning_rate": 7.53087872015057e-05,
"loss": 3.1189,
"step": 116000
},
{
"epoch": 6.29,
"grad_norm": 0.7388986945152283,
"learning_rate": 7.501470415245265e-05,
"loss": 3.119,
"step": 117000
},
{
"epoch": 6.34,
"grad_norm": 0.7611464262008667,
"learning_rate": 7.47206211033996e-05,
"loss": 3.1211,
"step": 118000
},
{
"epoch": 6.4,
"grad_norm": 0.7273160815238953,
"learning_rate": 7.442653805434656e-05,
"loss": 3.1224,
"step": 119000
},
{
"epoch": 6.45,
"grad_norm": 0.7437482476234436,
"learning_rate": 7.413274908834256e-05,
"loss": 3.1248,
"step": 120000
},
{
"epoch": 6.5,
"grad_norm": 0.7418404221534729,
"learning_rate": 7.38386660392895e-05,
"loss": 3.1241,
"step": 121000
},
{
"epoch": 6.56,
"grad_norm": 0.7766395807266235,
"learning_rate": 7.354458299023645e-05,
"loss": 3.1262,
"step": 122000
},
{
"epoch": 6.61,
"grad_norm": 0.7430386543273926,
"learning_rate": 7.325049994118339e-05,
"loss": 3.1262,
"step": 123000
},
{
"epoch": 6.67,
"grad_norm": 0.7568381428718567,
"learning_rate": 7.29567109751794e-05,
"loss": 3.1218,
"step": 124000
},
{
"epoch": 6.72,
"grad_norm": 0.742994487285614,
"learning_rate": 7.266292200917539e-05,
"loss": 3.1226,
"step": 125000
},
{
"epoch": 6.77,
"grad_norm": 0.7728854417800903,
"learning_rate": 7.236883896012234e-05,
"loss": 3.1304,
"step": 126000
},
{
"epoch": 6.83,
"grad_norm": 0.7517836093902588,
"learning_rate": 7.207504999411833e-05,
"loss": 3.1249,
"step": 127000
},
{
"epoch": 6.88,
"grad_norm": 0.7383127808570862,
"learning_rate": 7.178096694506528e-05,
"loss": 3.1212,
"step": 128000
},
{
"epoch": 6.93,
"grad_norm": 0.7259215712547302,
"learning_rate": 7.148717797906129e-05,
"loss": 3.1247,
"step": 129000
},
{
"epoch": 6.99,
"grad_norm": 0.7265549898147583,
"learning_rate": 7.119309493000824e-05,
"loss": 3.122,
"step": 130000
},
{
"epoch": 7.0,
"eval_accuracy": 0.3987158276834467,
"eval_loss": 3.423013687133789,
"eval_runtime": 154.1006,
"eval_samples_per_second": 375.865,
"eval_steps_per_second": 5.879,
"step": 130214
},
{
"epoch": 7.04,
"grad_norm": 0.757236897945404,
"learning_rate": 7.089901188095519e-05,
"loss": 3.0764,
"step": 131000
},
{
"epoch": 7.1,
"grad_norm": 0.7935017347335815,
"learning_rate": 7.060492883190213e-05,
"loss": 3.0706,
"step": 132000
},
{
"epoch": 7.15,
"grad_norm": 0.7874728441238403,
"learning_rate": 7.031084578284908e-05,
"loss": 3.076,
"step": 133000
},
{
"epoch": 7.2,
"grad_norm": 0.7797748446464539,
"learning_rate": 7.001705681684508e-05,
"loss": 3.073,
"step": 134000
},
{
"epoch": 7.26,
"grad_norm": 0.7578076124191284,
"learning_rate": 6.972326785084107e-05,
"loss": 3.077,
"step": 135000
},
{
"epoch": 7.31,
"grad_norm": 0.756395161151886,
"learning_rate": 6.942918480178802e-05,
"loss": 3.0835,
"step": 136000
},
{
"epoch": 7.36,
"grad_norm": 0.720391035079956,
"learning_rate": 6.913539583578404e-05,
"loss": 3.0818,
"step": 137000
},
{
"epoch": 7.42,
"grad_norm": 0.7686936259269714,
"learning_rate": 6.884131278673098e-05,
"loss": 3.0802,
"step": 138000
},
{
"epoch": 7.47,
"grad_norm": 0.7382326722145081,
"learning_rate": 6.854722973767792e-05,
"loss": 3.0847,
"step": 139000
},
{
"epoch": 7.53,
"grad_norm": 0.7340770959854126,
"learning_rate": 6.825314668862487e-05,
"loss": 3.0861,
"step": 140000
},
{
"epoch": 7.58,
"grad_norm": 0.7532581090927124,
"learning_rate": 6.795935772262087e-05,
"loss": 3.0868,
"step": 141000
},
{
"epoch": 7.63,
"grad_norm": 0.7558372020721436,
"learning_rate": 6.766527467356781e-05,
"loss": 3.0869,
"step": 142000
},
{
"epoch": 7.69,
"grad_norm": 0.7301390767097473,
"learning_rate": 6.737119162451476e-05,
"loss": 3.0858,
"step": 143000
},
{
"epoch": 7.74,
"grad_norm": 0.740253210067749,
"learning_rate": 6.707740265851076e-05,
"loss": 3.0902,
"step": 144000
},
{
"epoch": 7.79,
"grad_norm": 0.765359103679657,
"learning_rate": 6.678331960945772e-05,
"loss": 3.0886,
"step": 145000
},
{
"epoch": 7.85,
"grad_norm": 0.7424470782279968,
"learning_rate": 6.648953064345372e-05,
"loss": 3.0862,
"step": 146000
},
{
"epoch": 7.9,
"grad_norm": 0.7344245314598083,
"learning_rate": 6.619544759440066e-05,
"loss": 3.0876,
"step": 147000
},
{
"epoch": 7.96,
"grad_norm": 0.7477200627326965,
"learning_rate": 6.590165862839667e-05,
"loss": 3.0883,
"step": 148000
},
{
"epoch": 8.0,
"eval_accuracy": 0.4019737201422278,
"eval_loss": 3.410255193710327,
"eval_runtime": 153.6826,
"eval_samples_per_second": 376.887,
"eval_steps_per_second": 5.895,
"step": 148816
},
{
"epoch": 8.01,
"grad_norm": 0.7679169774055481,
"learning_rate": 6.560757557934361e-05,
"loss": 3.0781,
"step": 149000
},
{
"epoch": 8.06,
"grad_norm": 0.7777008414268494,
"learning_rate": 6.531349253029055e-05,
"loss": 3.0328,
"step": 150000
},
{
"epoch": 8.12,
"grad_norm": 0.7723888754844666,
"learning_rate": 6.50194094812375e-05,
"loss": 3.0361,
"step": 151000
},
{
"epoch": 8.17,
"grad_norm": 0.7410120964050293,
"learning_rate": 6.47256205152335e-05,
"loss": 3.0396,
"step": 152000
},
{
"epoch": 8.22,
"grad_norm": 0.775367021560669,
"learning_rate": 6.443153746618044e-05,
"loss": 3.0435,
"step": 153000
},
{
"epoch": 8.28,
"grad_norm": 0.7430744767189026,
"learning_rate": 6.413774850017646e-05,
"loss": 3.0487,
"step": 154000
},
{
"epoch": 8.33,
"grad_norm": 0.7859558463096619,
"learning_rate": 6.38436654511234e-05,
"loss": 3.0461,
"step": 155000
},
{
"epoch": 8.39,
"grad_norm": 0.7532193064689636,
"learning_rate": 6.354958240207035e-05,
"loss": 3.0479,
"step": 156000
},
{
"epoch": 8.44,
"grad_norm": 0.7526784539222717,
"learning_rate": 6.325579343606635e-05,
"loss": 3.0521,
"step": 157000
},
{
"epoch": 8.49,
"grad_norm": 0.7687983512878418,
"learning_rate": 6.296171038701329e-05,
"loss": 3.0505,
"step": 158000
},
{
"epoch": 8.55,
"grad_norm": 0.7615057229995728,
"learning_rate": 6.26679214210093e-05,
"loss": 3.0543,
"step": 159000
},
{
"epoch": 8.6,
"grad_norm": 0.7644217014312744,
"learning_rate": 6.237383837195624e-05,
"loss": 3.0556,
"step": 160000
},
{
"epoch": 8.65,
"grad_norm": 0.7489671111106873,
"learning_rate": 6.20797553229032e-05,
"loss": 3.0523,
"step": 161000
},
{
"epoch": 8.71,
"grad_norm": 0.7821494936943054,
"learning_rate": 6.17859663568992e-05,
"loss": 3.0606,
"step": 162000
},
{
"epoch": 8.76,
"grad_norm": 0.7804410457611084,
"learning_rate": 6.149188330784615e-05,
"loss": 3.0582,
"step": 163000
},
{
"epoch": 8.82,
"grad_norm": 0.7675402164459229,
"learning_rate": 6.119809434184214e-05,
"loss": 3.0566,
"step": 164000
},
{
"epoch": 8.87,
"grad_norm": 0.7614428400993347,
"learning_rate": 6.090401129278909e-05,
"loss": 3.0547,
"step": 165000
},
{
"epoch": 8.92,
"grad_norm": 0.7919172644615173,
"learning_rate": 6.0609928243736035e-05,
"loss": 3.0551,
"step": 166000
},
{
"epoch": 8.98,
"grad_norm": 0.7789074182510376,
"learning_rate": 6.031613927773203e-05,
"loss": 3.059,
"step": 167000
},
{
"epoch": 9.0,
"eval_accuracy": 0.4021935956374767,
"eval_loss": 3.4160611629486084,
"eval_runtime": 154.2434,
"eval_samples_per_second": 375.517,
"eval_steps_per_second": 5.874,
"step": 167418
},
{
"epoch": 9.03,
"grad_norm": 0.796655535697937,
"learning_rate": 6.002205622867898e-05,
"loss": 3.0247,
"step": 168000
},
{
"epoch": 9.09,
"grad_norm": 0.7840985059738159,
"learning_rate": 5.972797317962593e-05,
"loss": 3.0027,
"step": 169000
},
{
"epoch": 9.14,
"grad_norm": 0.8092240691184998,
"learning_rate": 5.943418421362194e-05,
"loss": 3.0089,
"step": 170000
},
{
"epoch": 9.19,
"grad_norm": 0.7674739360809326,
"learning_rate": 5.9140101164568875e-05,
"loss": 3.0123,
"step": 171000
},
{
"epoch": 9.25,
"grad_norm": 0.7991486191749573,
"learning_rate": 5.884631219856488e-05,
"loss": 3.0115,
"step": 172000
},
{
"epoch": 9.3,
"grad_norm": 0.803733766078949,
"learning_rate": 5.855222914951183e-05,
"loss": 3.0176,
"step": 173000
},
{
"epoch": 9.35,
"grad_norm": 0.7872689962387085,
"learning_rate": 5.8258440183507825e-05,
"loss": 3.0212,
"step": 174000
},
{
"epoch": 9.41,
"grad_norm": 0.7754504680633545,
"learning_rate": 5.796435713445477e-05,
"loss": 3.0189,
"step": 175000
},
{
"epoch": 9.46,
"grad_norm": 0.7785583138465881,
"learning_rate": 5.767027408540172e-05,
"loss": 3.0226,
"step": 176000
},
{
"epoch": 9.52,
"grad_norm": 0.7682950496673584,
"learning_rate": 5.737619103634867e-05,
"loss": 3.0225,
"step": 177000
},
{
"epoch": 9.57,
"grad_norm": 0.8099873065948486,
"learning_rate": 5.708240207034466e-05,
"loss": 3.024,
"step": 178000
},
{
"epoch": 9.62,
"grad_norm": 0.7960034012794495,
"learning_rate": 5.6788319021291615e-05,
"loss": 3.024,
"step": 179000
},
{
"epoch": 9.68,
"grad_norm": 0.8109017014503479,
"learning_rate": 5.649453005528762e-05,
"loss": 3.0264,
"step": 180000
},
{
"epoch": 9.73,
"grad_norm": 0.7674356698989868,
"learning_rate": 5.6200447006234556e-05,
"loss": 3.0301,
"step": 181000
},
{
"epoch": 9.78,
"grad_norm": 0.789279580116272,
"learning_rate": 5.5906658040230565e-05,
"loss": 3.0311,
"step": 182000
},
{
"epoch": 9.84,
"grad_norm": 0.7678402066230774,
"learning_rate": 5.561257499117751e-05,
"loss": 3.0311,
"step": 183000
},
{
"epoch": 9.89,
"grad_norm": 0.7561872601509094,
"learning_rate": 5.5318786025173505e-05,
"loss": 3.0295,
"step": 184000
},
{
"epoch": 9.95,
"grad_norm": 0.8036589026451111,
"learning_rate": 5.502470297612046e-05,
"loss": 3.0335,
"step": 185000
},
{
"epoch": 10.0,
"grad_norm": 0.7822810411453247,
"learning_rate": 5.47306199270674e-05,
"loss": 3.0294,
"step": 186000
},
{
"epoch": 10.0,
"eval_accuracy": 0.40386782023038625,
"eval_loss": 3.4003984928131104,
"eval_runtime": 154.1461,
"eval_samples_per_second": 375.754,
"eval_steps_per_second": 5.878,
"step": 186020
},
{
"epoch": 10.05,
"grad_norm": 0.7965357303619385,
"learning_rate": 5.443683096106341e-05,
"loss": 2.9764,
"step": 187000
},
{
"epoch": 10.11,
"grad_norm": 0.7956034541130066,
"learning_rate": 5.414274791201035e-05,
"loss": 2.9774,
"step": 188000
},
{
"epoch": 10.16,
"grad_norm": 0.7924751043319702,
"learning_rate": 5.38486648629573e-05,
"loss": 2.9838,
"step": 189000
},
{
"epoch": 10.21,
"grad_norm": 0.8389952182769775,
"learning_rate": 5.355458181390425e-05,
"loss": 2.9866,
"step": 190000
},
{
"epoch": 10.27,
"grad_norm": 0.8086053133010864,
"learning_rate": 5.3260792847900256e-05,
"loss": 2.9918,
"step": 191000
},
{
"epoch": 10.32,
"grad_norm": 0.7812045216560364,
"learning_rate": 5.2966709798847195e-05,
"loss": 2.9899,
"step": 192000
},
{
"epoch": 10.38,
"grad_norm": 0.8031854629516602,
"learning_rate": 5.26729208328432e-05,
"loss": 2.9933,
"step": 193000
},
{
"epoch": 10.43,
"grad_norm": 0.8311526775360107,
"learning_rate": 5.237913186683919e-05,
"loss": 2.9941,
"step": 194000
},
{
"epoch": 10.48,
"grad_norm": 0.8191061019897461,
"learning_rate": 5.2085048817786145e-05,
"loss": 2.9984,
"step": 195000
},
{
"epoch": 10.54,
"grad_norm": 0.8086720705032349,
"learning_rate": 5.179096576873309e-05,
"loss": 2.9989,
"step": 196000
},
{
"epoch": 10.59,
"grad_norm": 0.8062707781791687,
"learning_rate": 5.14971768027291e-05,
"loss": 2.9972,
"step": 197000
},
{
"epoch": 10.64,
"grad_norm": 0.7883285284042358,
"learning_rate": 5.1203093753676044e-05,
"loss": 3.0017,
"step": 198000
},
{
"epoch": 10.7,
"grad_norm": 0.8313503265380859,
"learning_rate": 5.090901070462298e-05,
"loss": 2.999,
"step": 199000
},
{
"epoch": 10.75,
"grad_norm": 0.8046980500221252,
"learning_rate": 5.061522173861899e-05,
"loss": 3.0024,
"step": 200000
},
{
"epoch": 10.81,
"grad_norm": 0.804289698600769,
"learning_rate": 5.032113868956594e-05,
"loss": 3.0023,
"step": 201000
},
{
"epoch": 10.86,
"grad_norm": 0.7925947308540344,
"learning_rate": 5.002705564051289e-05,
"loss": 3.005,
"step": 202000
},
{
"epoch": 10.91,
"grad_norm": 0.7959842681884766,
"learning_rate": 4.9733266674508884e-05,
"loss": 3.0082,
"step": 203000
},
{
"epoch": 10.97,
"grad_norm": 0.7851746082305908,
"learning_rate": 4.9439477708504887e-05,
"loss": 3.0081,
"step": 204000
},
{
"epoch": 11.0,
"eval_accuracy": 0.4041393560543386,
"eval_loss": 3.404827833175659,
"eval_runtime": 153.9153,
"eval_samples_per_second": 376.317,
"eval_steps_per_second": 5.886,
"step": 204622
},
{
"epoch": 11.02,
"grad_norm": 0.7984915375709534,
"learning_rate": 4.914539465945183e-05,
"loss": 2.9866,
"step": 205000
},
{
"epoch": 11.07,
"grad_norm": 0.832140326499939,
"learning_rate": 4.8851311610398784e-05,
"loss": 2.9503,
"step": 206000
},
{
"epoch": 11.13,
"grad_norm": 0.8821794390678406,
"learning_rate": 4.855722856134572e-05,
"loss": 2.9585,
"step": 207000
},
{
"epoch": 11.18,
"grad_norm": 0.861076295375824,
"learning_rate": 4.8263439595341724e-05,
"loss": 2.963,
"step": 208000
},
{
"epoch": 11.24,
"grad_norm": 0.8196221590042114,
"learning_rate": 4.7969356546288676e-05,
"loss": 2.9638,
"step": 209000
},
{
"epoch": 11.29,
"grad_norm": 0.8027163147926331,
"learning_rate": 4.767556758028467e-05,
"loss": 2.9643,
"step": 210000
},
{
"epoch": 11.34,
"grad_norm": 0.8131231069564819,
"learning_rate": 4.7381484531231624e-05,
"loss": 2.9705,
"step": 211000
},
{
"epoch": 11.4,
"grad_norm": 0.8047467470169067,
"learning_rate": 4.708740148217857e-05,
"loss": 2.9758,
"step": 212000
},
{
"epoch": 11.45,
"grad_norm": 0.8449523448944092,
"learning_rate": 4.6793612516174565e-05,
"loss": 2.9711,
"step": 213000
},
{
"epoch": 11.5,
"grad_norm": 0.8473635315895081,
"learning_rate": 4.6499529467121517e-05,
"loss": 2.9766,
"step": 214000
},
{
"epoch": 11.56,
"grad_norm": 0.8421896696090698,
"learning_rate": 4.620574050111752e-05,
"loss": 2.9753,
"step": 215000
},
{
"epoch": 11.61,
"grad_norm": 0.8112220764160156,
"learning_rate": 4.5911657452064464e-05,
"loss": 2.9797,
"step": 216000
},
{
"epoch": 11.67,
"grad_norm": 0.8139015436172485,
"learning_rate": 4.5617868486060466e-05,
"loss": 2.9775,
"step": 217000
},
{
"epoch": 11.72,
"grad_norm": 0.8164324164390564,
"learning_rate": 4.532378543700742e-05,
"loss": 2.9746,
"step": 218000
},
{
"epoch": 11.77,
"grad_norm": 0.8094378113746643,
"learning_rate": 4.502970238795436e-05,
"loss": 2.9792,
"step": 219000
},
{
"epoch": 11.83,
"grad_norm": 0.8222824931144714,
"learning_rate": 4.473561933890131e-05,
"loss": 2.981,
"step": 220000
},
{
"epoch": 11.88,
"grad_norm": 0.8311740756034851,
"learning_rate": 4.4441536289848254e-05,
"loss": 2.9817,
"step": 221000
},
{
"epoch": 11.93,
"grad_norm": 0.809837818145752,
"learning_rate": 4.41474532407952e-05,
"loss": 2.9842,
"step": 222000
},
{
"epoch": 11.99,
"grad_norm": 0.80256187915802,
"learning_rate": 4.385395835784026e-05,
"loss": 2.9849,
"step": 223000
},
{
"epoch": 12.0,
"eval_accuracy": 0.40463795548169496,
"eval_loss": 3.406789779663086,
"eval_runtime": 154.0508,
"eval_samples_per_second": 375.986,
"eval_steps_per_second": 5.881,
"step": 223224
},
{
"epoch": 12.04,
"grad_norm": 0.8648255467414856,
"learning_rate": 4.3559875308787204e-05,
"loss": 2.9414,
"step": 224000
},
{
"epoch": 12.1,
"grad_norm": 0.8371706008911133,
"learning_rate": 4.326579225973415e-05,
"loss": 2.9348,
"step": 225000
},
{
"epoch": 12.15,
"grad_norm": 0.8452836871147156,
"learning_rate": 4.297200329373015e-05,
"loss": 2.9398,
"step": 226000
},
{
"epoch": 12.2,
"grad_norm": 0.8612444996833801,
"learning_rate": 4.26779202446771e-05,
"loss": 2.9431,
"step": 227000
},
{
"epoch": 12.26,
"grad_norm": 0.8397490978240967,
"learning_rate": 4.238383719562405e-05,
"loss": 2.9411,
"step": 228000
},
{
"epoch": 12.31,
"grad_norm": 0.8516615629196167,
"learning_rate": 4.209004822962005e-05,
"loss": 2.9496,
"step": 229000
},
{
"epoch": 12.36,
"grad_norm": 0.866407573223114,
"learning_rate": 4.179625926361605e-05,
"loss": 2.9463,
"step": 230000
},
{
"epoch": 12.42,
"grad_norm": 0.8084841370582581,
"learning_rate": 4.150217621456299e-05,
"loss": 2.949,
"step": 231000
},
{
"epoch": 12.47,
"grad_norm": 0.8507031202316284,
"learning_rate": 4.1208387248558994e-05,
"loss": 2.9496,
"step": 232000
},
{
"epoch": 12.53,
"grad_norm": 0.82049560546875,
"learning_rate": 4.0914304199505946e-05,
"loss": 2.9493,
"step": 233000
},
{
"epoch": 12.58,
"grad_norm": 0.843075156211853,
"learning_rate": 4.062022115045289e-05,
"loss": 2.9587,
"step": 234000
},
{
"epoch": 12.63,
"grad_norm": 0.8294638991355896,
"learning_rate": 4.0326138101399836e-05,
"loss": 2.9545,
"step": 235000
},
{
"epoch": 12.69,
"grad_norm": 0.8557348251342773,
"learning_rate": 4.003205505234678e-05,
"loss": 2.9546,
"step": 236000
},
{
"epoch": 12.74,
"grad_norm": 0.8104209303855896,
"learning_rate": 3.9738266086342784e-05,
"loss": 2.9598,
"step": 237000
},
{
"epoch": 12.79,
"grad_norm": 0.8661037683486938,
"learning_rate": 3.9444183037289735e-05,
"loss": 2.9611,
"step": 238000
},
{
"epoch": 12.85,
"grad_norm": 0.8609908223152161,
"learning_rate": 3.915009998823668e-05,
"loss": 2.9567,
"step": 239000
},
{
"epoch": 12.9,
"grad_norm": 0.837121307849884,
"learning_rate": 3.8856016939183626e-05,
"loss": 2.9652,
"step": 240000
},
{
"epoch": 12.96,
"grad_norm": 0.8167340755462646,
"learning_rate": 3.856222797317963e-05,
"loss": 2.9618,
"step": 241000
},
{
"epoch": 13.0,
"eval_accuracy": 0.4047847406809369,
"eval_loss": 3.412700653076172,
"eval_runtime": 153.7925,
"eval_samples_per_second": 376.618,
"eval_steps_per_second": 5.891,
"step": 241826
},
{
"epoch": 13.01,
"grad_norm": 0.8533641695976257,
"learning_rate": 3.826814492412657e-05,
"loss": 2.9526,
"step": 242000
},
{
"epoch": 13.06,
"grad_norm": 0.8695389032363892,
"learning_rate": 3.7974061875073525e-05,
"loss": 2.9135,
"step": 243000
},
{
"epoch": 13.12,
"grad_norm": 0.8457789421081543,
"learning_rate": 3.767997882602047e-05,
"loss": 2.9143,
"step": 244000
},
{
"epoch": 13.17,
"grad_norm": 0.8333380818367004,
"learning_rate": 3.7386189860016466e-05,
"loss": 2.9226,
"step": 245000
},
{
"epoch": 13.22,
"grad_norm": 0.8502881526947021,
"learning_rate": 3.7092694977061526e-05,
"loss": 2.9231,
"step": 246000
},
{
"epoch": 13.28,
"grad_norm": 0.8744940757751465,
"learning_rate": 3.679861192800847e-05,
"loss": 2.926,
"step": 247000
},
{
"epoch": 13.33,
"grad_norm": 0.867820143699646,
"learning_rate": 3.6504528878955416e-05,
"loss": 2.9293,
"step": 248000
},
{
"epoch": 13.39,
"grad_norm": 0.8131315112113953,
"learning_rate": 3.621044582990237e-05,
"loss": 2.9304,
"step": 249000
},
{
"epoch": 13.44,
"grad_norm": 0.8412694931030273,
"learning_rate": 3.591636278084931e-05,
"loss": 2.9291,
"step": 250000
},
{
"epoch": 13.49,
"grad_norm": 0.8476005792617798,
"learning_rate": 3.562257381484531e-05,
"loss": 2.9297,
"step": 251000
},
{
"epoch": 13.55,
"grad_norm": 0.8458747863769531,
"learning_rate": 3.532878484884131e-05,
"loss": 2.9349,
"step": 252000
},
{
"epoch": 13.6,
"grad_norm": 0.8815217614173889,
"learning_rate": 3.503470179978826e-05,
"loss": 2.9321,
"step": 253000
},
{
"epoch": 13.65,
"grad_norm": 0.8400765061378479,
"learning_rate": 3.474061875073521e-05,
"loss": 2.9336,
"step": 254000
},
{
"epoch": 13.71,
"grad_norm": 0.8505930304527283,
"learning_rate": 3.444653570168216e-05,
"loss": 2.9395,
"step": 255000
},
{
"epoch": 13.76,
"grad_norm": 0.8682374358177185,
"learning_rate": 3.415274673567816e-05,
"loss": 2.9371,
"step": 256000
},
{
"epoch": 13.82,
"grad_norm": 0.8451393246650696,
"learning_rate": 3.38586636866251e-05,
"loss": 2.9445,
"step": 257000
},
{
"epoch": 13.87,
"grad_norm": 0.8744685053825378,
"learning_rate": 3.35648747206211e-05,
"loss": 2.938,
"step": 258000
},
{
"epoch": 13.92,
"grad_norm": 0.8651999831199646,
"learning_rate": 3.3270791671568055e-05,
"loss": 2.9409,
"step": 259000
},
{
"epoch": 13.98,
"grad_norm": 0.8568054437637329,
"learning_rate": 3.2976708622515e-05,
"loss": 2.9398,
"step": 260000
},
{
"epoch": 14.0,
"eval_accuracy": 0.40543314834367516,
"eval_loss": 3.4078917503356934,
"eval_runtime": 153.8367,
"eval_samples_per_second": 376.51,
"eval_steps_per_second": 5.889,
"step": 260428
},
{
"epoch": 14.03,
"grad_norm": 0.8709078431129456,
"learning_rate": 3.2682919656511e-05,
"loss": 2.914,
"step": 261000
},
{
"epoch": 14.08,
"grad_norm": 0.874599039554596,
"learning_rate": 3.238883660745795e-05,
"loss": 2.8959,
"step": 262000
},
{
"epoch": 14.14,
"grad_norm": 0.8708846569061279,
"learning_rate": 3.209475355840489e-05,
"loss": 2.8994,
"step": 263000
},
{
"epoch": 14.19,
"grad_norm": 0.9044806361198425,
"learning_rate": 3.1800964592400895e-05,
"loss": 2.9021,
"step": 264000
},
{
"epoch": 14.25,
"grad_norm": 0.8368488550186157,
"learning_rate": 3.150688154334784e-05,
"loss": 2.9034,
"step": 265000
},
{
"epoch": 14.3,
"grad_norm": 0.8695858120918274,
"learning_rate": 3.121279849429479e-05,
"loss": 2.9071,
"step": 266000
},
{
"epoch": 14.35,
"grad_norm": 0.8245272636413574,
"learning_rate": 3.0919009528290795e-05,
"loss": 2.9094,
"step": 267000
},
{
"epoch": 14.41,
"grad_norm": 0.8766151666641235,
"learning_rate": 3.062492647923774e-05,
"loss": 2.9142,
"step": 268000
},
{
"epoch": 14.46,
"grad_norm": 0.8801271319389343,
"learning_rate": 3.033113751323374e-05,
"loss": 2.9104,
"step": 269000
},
{
"epoch": 14.51,
"grad_norm": 0.8736371994018555,
"learning_rate": 3.0037054464180687e-05,
"loss": 2.9144,
"step": 270000
},
{
"epoch": 14.57,
"grad_norm": 0.8868771195411682,
"learning_rate": 2.9743265498176686e-05,
"loss": 2.9199,
"step": 271000
},
{
"epoch": 14.62,
"grad_norm": 0.8758333325386047,
"learning_rate": 2.9449182449123635e-05,
"loss": 2.9148,
"step": 272000
},
{
"epoch": 14.68,
"grad_norm": 0.8853691220283508,
"learning_rate": 2.9155393483119637e-05,
"loss": 2.9151,
"step": 273000
},
{
"epoch": 14.73,
"grad_norm": 0.8897860646247864,
"learning_rate": 2.886131043406658e-05,
"loss": 2.9161,
"step": 274000
},
{
"epoch": 14.78,
"grad_norm": 0.8651039600372314,
"learning_rate": 2.856752146806258e-05,
"loss": 2.9228,
"step": 275000
},
{
"epoch": 14.84,
"grad_norm": 0.8585176467895508,
"learning_rate": 2.827343841900953e-05,
"loss": 2.9196,
"step": 276000
},
{
"epoch": 14.89,
"grad_norm": 0.8638288974761963,
"learning_rate": 2.797935536995648e-05,
"loss": 2.9218,
"step": 277000
},
{
"epoch": 14.94,
"grad_norm": 0.8799272775650024,
"learning_rate": 2.7685566403952477e-05,
"loss": 2.9239,
"step": 278000
},
{
"epoch": 15.0,
"grad_norm": 0.8739799857139587,
"learning_rate": 2.7391483354899426e-05,
"loss": 2.9226,
"step": 279000
},
{
"epoch": 15.0,
"eval_accuracy": 0.406453792523118,
"eval_loss": 3.3963141441345215,
"eval_runtime": 153.9816,
"eval_samples_per_second": 376.155,
"eval_steps_per_second": 5.884,
"step": 279030
},
{
"epoch": 15.05,
"grad_norm": 0.8708630204200745,
"learning_rate": 2.709740030584637e-05,
"loss": 2.8824,
"step": 280000
},
{
"epoch": 15.11,
"grad_norm": 0.9101833701133728,
"learning_rate": 2.680331725679332e-05,
"loss": 2.8831,
"step": 281000
},
{
"epoch": 15.16,
"grad_norm": 0.9030382633209229,
"learning_rate": 2.6509528290789322e-05,
"loss": 2.8836,
"step": 282000
},
{
"epoch": 15.21,
"grad_norm": 0.8790870308876038,
"learning_rate": 2.621544524173627e-05,
"loss": 2.8874,
"step": 283000
},
{
"epoch": 15.27,
"grad_norm": 0.8721250295639038,
"learning_rate": 2.592165627573227e-05,
"loss": 2.8902,
"step": 284000
},
{
"epoch": 15.32,
"grad_norm": 0.9108452200889587,
"learning_rate": 2.5627573226679215e-05,
"loss": 2.892,
"step": 285000
},
{
"epoch": 15.37,
"grad_norm": 0.9115946888923645,
"learning_rate": 2.5333490177626163e-05,
"loss": 2.8947,
"step": 286000
},
{
"epoch": 15.43,
"grad_norm": 0.9250348806381226,
"learning_rate": 2.5039701211622162e-05,
"loss": 2.8961,
"step": 287000
},
{
"epoch": 15.48,
"grad_norm": 0.8640110492706299,
"learning_rate": 2.474561816256911e-05,
"loss": 2.894,
"step": 288000
},
{
"epoch": 15.54,
"grad_norm": 0.8953117728233337,
"learning_rate": 2.445153511351606e-05,
"loss": 2.8987,
"step": 289000
},
{
"epoch": 15.59,
"grad_norm": 0.9380891919136047,
"learning_rate": 2.4157452064463004e-05,
"loss": 2.8939,
"step": 290000
},
{
"epoch": 15.64,
"grad_norm": 0.8643067479133606,
"learning_rate": 2.386395718150806e-05,
"loss": 2.8975,
"step": 291000
},
{
"epoch": 15.7,
"grad_norm": 0.8749162554740906,
"learning_rate": 2.3569874132455006e-05,
"loss": 2.9003,
"step": 292000
},
{
"epoch": 15.75,
"grad_norm": 0.8664742112159729,
"learning_rate": 2.3275791083401954e-05,
"loss": 2.9002,
"step": 293000
},
{
"epoch": 15.8,
"grad_norm": 0.9035205245018005,
"learning_rate": 2.29817080343489e-05,
"loss": 2.9042,
"step": 294000
},
{
"epoch": 15.86,
"grad_norm": 0.8949082493782043,
"learning_rate": 2.2687624985295848e-05,
"loss": 2.9004,
"step": 295000
},
{
"epoch": 15.91,
"grad_norm": 0.8694750070571899,
"learning_rate": 2.23941301023409e-05,
"loss": 2.9055,
"step": 296000
},
{
"epoch": 15.97,
"grad_norm": 0.898671567440033,
"learning_rate": 2.210004705328785e-05,
"loss": 2.9009,
"step": 297000
},
{
"epoch": 16.0,
"eval_accuracy": 0.40682327471800617,
"eval_loss": 3.4036316871643066,
"eval_runtime": 153.8703,
"eval_samples_per_second": 376.428,
"eval_steps_per_second": 5.888,
"step": 297632
},
{
"epoch": 16.02,
"grad_norm": 0.8959853053092957,
"learning_rate": 2.1805964004234798e-05,
"loss": 2.8874,
"step": 298000
},
{
"epoch": 16.07,
"grad_norm": 0.9309787154197693,
"learning_rate": 2.1511880955181746e-05,
"loss": 2.8624,
"step": 299000
},
{
"epoch": 16.13,
"grad_norm": 0.8814511299133301,
"learning_rate": 2.1218091989177742e-05,
"loss": 2.8708,
"step": 300000
},
{
"epoch": 16.18,
"grad_norm": 0.9105135202407837,
"learning_rate": 2.0924008940124694e-05,
"loss": 2.871,
"step": 301000
},
{
"epoch": 16.23,
"grad_norm": 0.902424156665802,
"learning_rate": 2.062992589107164e-05,
"loss": 2.8719,
"step": 302000
},
{
"epoch": 16.29,
"grad_norm": 0.9054093956947327,
"learning_rate": 2.0335842842018588e-05,
"loss": 2.8707,
"step": 303000
},
{
"epoch": 16.34,
"grad_norm": 0.9166373610496521,
"learning_rate": 2.0042053876014587e-05,
"loss": 2.8774,
"step": 304000
},
{
"epoch": 16.4,
"grad_norm": 0.9339081645011902,
"learning_rate": 1.9747970826961535e-05,
"loss": 2.879,
"step": 305000
},
{
"epoch": 16.45,
"grad_norm": 0.9023671746253967,
"learning_rate": 1.9454181860957534e-05,
"loss": 2.8791,
"step": 306000
},
{
"epoch": 16.5,
"grad_norm": 0.9114070534706116,
"learning_rate": 1.9160098811904483e-05,
"loss": 2.8824,
"step": 307000
},
{
"epoch": 16.56,
"grad_norm": 0.8930648565292358,
"learning_rate": 1.886601576285143e-05,
"loss": 2.8833,
"step": 308000
},
{
"epoch": 16.61,
"grad_norm": 0.9150155186653137,
"learning_rate": 1.857222679684743e-05,
"loss": 2.8808,
"step": 309000
},
{
"epoch": 16.66,
"grad_norm": 0.9177570343017578,
"learning_rate": 1.8278143747794375e-05,
"loss": 2.885,
"step": 310000
},
{
"epoch": 16.72,
"grad_norm": 0.9326618313789368,
"learning_rate": 1.7984060698741327e-05,
"loss": 2.8825,
"step": 311000
},
{
"epoch": 16.77,
"grad_norm": 0.9235589504241943,
"learning_rate": 1.7690271732737326e-05,
"loss": 2.8832,
"step": 312000
},
{
"epoch": 16.83,
"grad_norm": 0.9126114249229431,
"learning_rate": 1.739618868368427e-05,
"loss": 2.8817,
"step": 313000
},
{
"epoch": 16.88,
"grad_norm": 0.8965770602226257,
"learning_rate": 1.7102399717680274e-05,
"loss": 2.8865,
"step": 314000
},
{
"epoch": 16.93,
"grad_norm": 0.9391665458679199,
"learning_rate": 1.6808316668627222e-05,
"loss": 2.8835,
"step": 315000
},
{
"epoch": 16.99,
"grad_norm": 0.9226199984550476,
"learning_rate": 1.651452770262322e-05,
"loss": 2.8845,
"step": 316000
},
{
"epoch": 17.0,
"eval_accuracy": 0.4067043014512522,
"eval_loss": 3.409031867980957,
"eval_runtime": 153.8139,
"eval_samples_per_second": 376.565,
"eval_steps_per_second": 5.89,
"step": 316234
},
{
"epoch": 17.04,
"grad_norm": 0.9422922730445862,
"learning_rate": 1.622044465357017e-05,
"loss": 2.8622,
"step": 317000
},
{
"epoch": 17.09,
"grad_norm": 0.8947985172271729,
"learning_rate": 1.592665568756617e-05,
"loss": 2.8557,
"step": 318000
},
{
"epoch": 17.15,
"grad_norm": 0.9117223620414734,
"learning_rate": 1.5632572638513114e-05,
"loss": 2.8571,
"step": 319000
},
{
"epoch": 17.2,
"grad_norm": 0.9183168411254883,
"learning_rate": 1.5338783672509116e-05,
"loss": 2.8595,
"step": 320000
},
{
"epoch": 17.26,
"grad_norm": 0.9229071736335754,
"learning_rate": 1.5044700623456065e-05,
"loss": 2.863,
"step": 321000
},
{
"epoch": 17.31,
"grad_norm": 0.9431742429733276,
"learning_rate": 1.4750617574403012e-05,
"loss": 2.8624,
"step": 322000
},
{
"epoch": 17.36,
"grad_norm": 0.9331133365631104,
"learning_rate": 1.4456828608399012e-05,
"loss": 2.861,
"step": 323000
},
{
"epoch": 17.42,
"grad_norm": 0.9034436941146851,
"learning_rate": 1.4162745559345961e-05,
"loss": 2.8647,
"step": 324000
},
{
"epoch": 17.47,
"grad_norm": 0.9203888773918152,
"learning_rate": 1.3868662510292906e-05,
"loss": 2.8648,
"step": 325000
},
{
"epoch": 17.52,
"grad_norm": 0.9167985320091248,
"learning_rate": 1.3574873544288907e-05,
"loss": 2.8638,
"step": 326000
},
{
"epoch": 17.58,
"grad_norm": 0.9315596222877502,
"learning_rate": 1.3280790495235857e-05,
"loss": 2.8658,
"step": 327000
},
{
"epoch": 17.63,
"grad_norm": 0.883280336856842,
"learning_rate": 1.2987001529231858e-05,
"loss": 2.8622,
"step": 328000
},
{
"epoch": 17.69,
"grad_norm": 0.9321721792221069,
"learning_rate": 1.2692918480178803e-05,
"loss": 2.86,
"step": 329000
},
{
"epoch": 17.74,
"grad_norm": 0.9126923680305481,
"learning_rate": 1.2398835431125751e-05,
"loss": 2.8686,
"step": 330000
},
{
"epoch": 17.79,
"grad_norm": 0.8995952010154724,
"learning_rate": 1.2105046465121752e-05,
"loss": 2.8676,
"step": 331000
},
{
"epoch": 17.85,
"grad_norm": 0.920877993106842,
"learning_rate": 1.1810963416068699e-05,
"loss": 2.8663,
"step": 332000
},
{
"epoch": 17.9,
"grad_norm": 0.9144251942634583,
"learning_rate": 1.1516880367015646e-05,
"loss": 2.8701,
"step": 333000
},
{
"epoch": 17.96,
"grad_norm": 0.9452138543128967,
"learning_rate": 1.1222797317962593e-05,
"loss": 2.8685,
"step": 334000
},
{
"epoch": 18.0,
"eval_accuracy": 0.4071263172963953,
"eval_loss": 3.405372381210327,
"eval_runtime": 155.4934,
"eval_samples_per_second": 372.498,
"eval_steps_per_second": 5.827,
"step": 334836
},
{
"epoch": 18.01,
"grad_norm": 0.9502813220024109,
"learning_rate": 1.0929008351958593e-05,
"loss": 2.8643,
"step": 335000
},
{
"epoch": 18.06,
"grad_norm": 0.9388921856880188,
"learning_rate": 1.063492530290554e-05,
"loss": 2.8412,
"step": 336000
},
{
"epoch": 18.12,
"grad_norm": 0.9569390416145325,
"learning_rate": 1.0341136336901541e-05,
"loss": 2.8438,
"step": 337000
},
{
"epoch": 18.17,
"grad_norm": 0.9436145424842834,
"learning_rate": 1.004705328784849e-05,
"loss": 2.8426,
"step": 338000
},
{
"epoch": 18.22,
"grad_norm": 0.9303746223449707,
"learning_rate": 9.752970238795436e-06,
"loss": 2.8428,
"step": 339000
},
{
"epoch": 18.28,
"grad_norm": 0.914296567440033,
"learning_rate": 9.459181272791437e-06,
"loss": 2.8489,
"step": 340000
},
{
"epoch": 18.33,
"grad_norm": 0.9756318926811218,
"learning_rate": 9.165098223738384e-06,
"loss": 2.847,
"step": 341000
},
{
"epoch": 18.39,
"grad_norm": 0.9441861510276794,
"learning_rate": 8.871015174685332e-06,
"loss": 2.8477,
"step": 342000
},
{
"epoch": 18.44,
"grad_norm": 0.9245781302452087,
"learning_rate": 8.577226208681333e-06,
"loss": 2.8496,
"step": 343000
},
{
"epoch": 18.49,
"grad_norm": 0.9517726302146912,
"learning_rate": 8.283143159628278e-06,
"loss": 2.8483,
"step": 344000
},
{
"epoch": 18.55,
"grad_norm": 0.9460880160331726,
"learning_rate": 7.989060110575227e-06,
"loss": 2.8541,
"step": 345000
},
{
"epoch": 18.6,
"grad_norm": 0.9118568301200867,
"learning_rate": 7.694977061522173e-06,
"loss": 2.8503,
"step": 346000
},
{
"epoch": 18.65,
"grad_norm": 0.918589174747467,
"learning_rate": 7.401188095518174e-06,
"loss": 2.8526,
"step": 347000
},
{
"epoch": 18.71,
"grad_norm": 0.9284429550170898,
"learning_rate": 7.107399129514175e-06,
"loss": 2.8541,
"step": 348000
},
{
"epoch": 18.76,
"grad_norm": 0.954234778881073,
"learning_rate": 6.8133160804611226e-06,
"loss": 2.8519,
"step": 349000
},
{
"epoch": 18.82,
"grad_norm": 0.9434486627578735,
"learning_rate": 6.519527114457123e-06,
"loss": 2.8508,
"step": 350000
},
{
"epoch": 18.87,
"grad_norm": 0.9406701922416687,
"learning_rate": 6.22544406540407e-06,
"loss": 2.8559,
"step": 351000
},
{
"epoch": 18.92,
"grad_norm": 0.9509878158569336,
"learning_rate": 5.931361016351018e-06,
"loss": 2.8522,
"step": 352000
},
{
"epoch": 18.98,
"grad_norm": 0.9960227012634277,
"learning_rate": 5.6372779672979655e-06,
"loss": 2.8513,
"step": 353000
},
{
"epoch": 19.0,
"eval_accuracy": 0.4068732555821874,
"eval_loss": 3.4187262058258057,
"eval_runtime": 154.0999,
"eval_samples_per_second": 375.867,
"eval_steps_per_second": 5.879,
"step": 353438
},
{
"epoch": 19.03,
"grad_norm": 0.9322279691696167,
"learning_rate": 5.343194918244912e-06,
"loss": 2.8414,
"step": 354000
},
{
"epoch": 19.08,
"grad_norm": 0.9578444361686707,
"learning_rate": 5.049405952240913e-06,
"loss": 2.8361,
"step": 355000
},
{
"epoch": 19.14,
"grad_norm": 0.9322285652160645,
"learning_rate": 4.75532290318786e-06,
"loss": 2.8365,
"step": 356000
},
{
"epoch": 19.19,
"grad_norm": 0.9732059836387634,
"learning_rate": 4.4612398541348075e-06,
"loss": 2.8372,
"step": 357000
},
{
"epoch": 19.25,
"grad_norm": 0.9745797514915466,
"learning_rate": 4.167156805081755e-06,
"loss": 2.8352,
"step": 358000
},
{
"epoch": 19.3,
"grad_norm": 0.9456592202186584,
"learning_rate": 3.873367839077756e-06,
"loss": 2.8367,
"step": 359000
},
{
"epoch": 19.35,
"grad_norm": 0.9315751791000366,
"learning_rate": 3.5792847900247036e-06,
"loss": 2.8368,
"step": 360000
},
{
"epoch": 19.41,
"grad_norm": 0.8824998140335083,
"learning_rate": 3.2857899070697564e-06,
"loss": 2.8355,
"step": 361000
},
{
"epoch": 19.46,
"grad_norm": 0.985062837600708,
"learning_rate": 2.9917068580167037e-06,
"loss": 2.8372,
"step": 362000
},
{
"epoch": 19.51,
"grad_norm": 0.9548268914222717,
"learning_rate": 2.6976238089636514e-06,
"loss": 2.839,
"step": 363000
},
{
"epoch": 19.57,
"grad_norm": 0.9308351278305054,
"learning_rate": 2.403540759910599e-06,
"loss": 2.8368,
"step": 364000
},
{
"epoch": 19.62,
"grad_norm": 0.9744262099266052,
"learning_rate": 2.1094577108575463e-06,
"loss": 2.8369,
"step": 365000
},
{
"epoch": 19.68,
"grad_norm": 0.962074875831604,
"learning_rate": 1.8153746618044938e-06,
"loss": 2.8354,
"step": 366000
},
{
"epoch": 19.73,
"grad_norm": 0.9616215229034424,
"learning_rate": 1.521585695800494e-06,
"loss": 2.8404,
"step": 367000
},
{
"epoch": 19.78,
"grad_norm": 0.9153415560722351,
"learning_rate": 1.2277967297964945e-06,
"loss": 2.8384,
"step": 368000
},
{
"epoch": 19.84,
"grad_norm": 0.9574062824249268,
"learning_rate": 9.337136807434419e-07,
"loss": 2.8356,
"step": 369000
},
{
"epoch": 19.89,
"grad_norm": 0.930395781993866,
"learning_rate": 6.396306316903894e-07,
"loss": 2.8368,
"step": 370000
},
{
"epoch": 19.94,
"grad_norm": 0.9612599015235901,
"learning_rate": 3.4584166568638985e-07,
"loss": 2.8348,
"step": 371000
},
{
"epoch": 20.0,
"grad_norm": 0.9463599920272827,
"learning_rate": 5.175861663333726e-08,
"loss": 2.8368,
"step": 372000
},
{
"epoch": 20.0,
"eval_accuracy": 0.40681131693060796,
"eval_loss": 3.4240403175354004,
"eval_runtime": 155.2617,
"eval_samples_per_second": 373.054,
"eval_steps_per_second": 5.835,
"step": 372040
},
{
"epoch": 20.0,
"step": 372040,
"total_flos": 1.56748665397248e+18,
"train_loss": 3.1496440902626293,
"train_runtime": 81299.3395,
"train_samples_per_second": 146.434,
"train_steps_per_second": 4.576
}
],
"logging_steps": 1000,
"max_steps": 372040,
"num_input_tokens_seen": 0,
"num_train_epochs": 20,
"save_steps": 5000,
"total_flos": 1.56748665397248e+18,
"train_batch_size": 32,
"trial_name": null,
"trial_params": null
}