joelniklaus's picture
Training in progress, step 200000
70055f0
raw
history blame
No virus
25.2 kB
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 0.2,
"global_step": 200000,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.0,
"learning_rate": 2.0000000000000003e-06,
"loss": 4.7524,
"step": 1000
},
{
"epoch": 0.0,
"learning_rate": 4.000000000000001e-06,
"loss": 3.5548,
"step": 2000
},
{
"epoch": 0.0,
"learning_rate": 6e-06,
"loss": 2.9636,
"step": 3000
},
{
"epoch": 0.0,
"learning_rate": 8.000000000000001e-06,
"loss": 2.7618,
"step": 4000
},
{
"epoch": 0.01,
"learning_rate": 1e-05,
"loss": 2.6257,
"step": 5000
},
{
"epoch": 0.01,
"learning_rate": 1.2e-05,
"loss": 2.5517,
"step": 6000
},
{
"epoch": 0.01,
"learning_rate": 1.4000000000000001e-05,
"loss": 2.4628,
"step": 7000
},
{
"epoch": 0.01,
"learning_rate": 1.6000000000000003e-05,
"loss": 2.4338,
"step": 8000
},
{
"epoch": 0.01,
"learning_rate": 1.8e-05,
"loss": 2.3362,
"step": 9000
},
{
"epoch": 0.01,
"learning_rate": 2e-05,
"loss": 2.2746,
"step": 10000
},
{
"epoch": 0.01,
"learning_rate": 2.2000000000000003e-05,
"loss": 2.1795,
"step": 11000
},
{
"epoch": 0.01,
"learning_rate": 2.4e-05,
"loss": 2.2309,
"step": 12000
},
{
"epoch": 0.01,
"learning_rate": 2.6000000000000002e-05,
"loss": 1.8605,
"step": 13000
},
{
"epoch": 0.01,
"learning_rate": 2.8000000000000003e-05,
"loss": 1.7232,
"step": 14000
},
{
"epoch": 0.01,
"learning_rate": 3e-05,
"loss": 1.6872,
"step": 15000
},
{
"epoch": 0.02,
"learning_rate": 3.2000000000000005e-05,
"loss": 2.1581,
"step": 16000
},
{
"epoch": 0.02,
"learning_rate": 3.4000000000000007e-05,
"loss": 2.0653,
"step": 17000
},
{
"epoch": 0.02,
"learning_rate": 3.6e-05,
"loss": 2.049,
"step": 18000
},
{
"epoch": 0.02,
"learning_rate": 3.8e-05,
"loss": 2.0773,
"step": 19000
},
{
"epoch": 0.02,
"learning_rate": 4e-05,
"loss": 1.9397,
"step": 20000
},
{
"epoch": 0.02,
"learning_rate": 4.2e-05,
"loss": 1.8899,
"step": 21000
},
{
"epoch": 0.02,
"learning_rate": 4.4000000000000006e-05,
"loss": 1.8474,
"step": 22000
},
{
"epoch": 0.02,
"learning_rate": 4.600000000000001e-05,
"loss": 1.8034,
"step": 23000
},
{
"epoch": 0.02,
"learning_rate": 4.8e-05,
"loss": 1.7666,
"step": 24000
},
{
"epoch": 0.03,
"learning_rate": 5e-05,
"loss": 1.7899,
"step": 25000
},
{
"epoch": 0.03,
"learning_rate": 5.2000000000000004e-05,
"loss": 1.4265,
"step": 26000
},
{
"epoch": 0.03,
"learning_rate": 5.4000000000000005e-05,
"loss": 1.3535,
"step": 27000
},
{
"epoch": 0.03,
"learning_rate": 5.6000000000000006e-05,
"loss": 1.1703,
"step": 28000
},
{
"epoch": 0.03,
"learning_rate": 5.8e-05,
"loss": 1.129,
"step": 29000
},
{
"epoch": 0.03,
"learning_rate": 6e-05,
"loss": 1.2811,
"step": 30000
},
{
"epoch": 0.03,
"learning_rate": 6.2e-05,
"loss": 1.2189,
"step": 31000
},
{
"epoch": 0.03,
"learning_rate": 6.400000000000001e-05,
"loss": 1.4371,
"step": 32000
},
{
"epoch": 0.03,
"learning_rate": 6.6e-05,
"loss": 1.3551,
"step": 33000
},
{
"epoch": 0.03,
"learning_rate": 6.800000000000001e-05,
"loss": 1.3616,
"step": 34000
},
{
"epoch": 0.04,
"learning_rate": 7e-05,
"loss": 1.2483,
"step": 35000
},
{
"epoch": 0.04,
"learning_rate": 7.2e-05,
"loss": 1.2174,
"step": 36000
},
{
"epoch": 0.04,
"learning_rate": 7.4e-05,
"loss": 1.1861,
"step": 37000
},
{
"epoch": 0.04,
"learning_rate": 7.6e-05,
"loss": 1.1465,
"step": 38000
},
{
"epoch": 0.04,
"learning_rate": 7.800000000000001e-05,
"loss": 1.1957,
"step": 39000
},
{
"epoch": 0.04,
"learning_rate": 8e-05,
"loss": 1.136,
"step": 40000
},
{
"epoch": 0.04,
"learning_rate": 8.2e-05,
"loss": 1.1122,
"step": 41000
},
{
"epoch": 0.04,
"learning_rate": 8.4e-05,
"loss": 1.0905,
"step": 42000
},
{
"epoch": 0.04,
"learning_rate": 8.6e-05,
"loss": 1.0996,
"step": 43000
},
{
"epoch": 0.04,
"learning_rate": 8.800000000000001e-05,
"loss": 1.0632,
"step": 44000
},
{
"epoch": 0.04,
"learning_rate": 9e-05,
"loss": 1.1387,
"step": 45000
},
{
"epoch": 0.05,
"learning_rate": 9.200000000000001e-05,
"loss": 1.2049,
"step": 46000
},
{
"epoch": 0.05,
"learning_rate": 9.4e-05,
"loss": 1.1783,
"step": 47000
},
{
"epoch": 0.05,
"learning_rate": 9.6e-05,
"loss": 1.1043,
"step": 48000
},
{
"epoch": 0.05,
"learning_rate": 9.8e-05,
"loss": 1.0633,
"step": 49000
},
{
"epoch": 0.05,
"learning_rate": 0.0001,
"loss": 1.1331,
"step": 50000
},
{
"epoch": 0.05,
"eval_loss": 0.753776490688324,
"eval_runtime": 33.8225,
"eval_samples_per_second": 147.83,
"eval_steps_per_second": 2.336,
"step": 50000
},
{
"epoch": 0.05,
"learning_rate": 9.999972660400536e-05,
"loss": 0.9559,
"step": 51000
},
{
"epoch": 0.05,
"learning_rate": 9.999890641901125e-05,
"loss": 0.9156,
"step": 52000
},
{
"epoch": 0.05,
"learning_rate": 9.999753945398704e-05,
"loss": 0.9237,
"step": 53000
},
{
"epoch": 0.05,
"learning_rate": 9.99956257238817e-05,
"loss": 0.9444,
"step": 54000
},
{
"epoch": 0.06,
"learning_rate": 9.999316524962345e-05,
"loss": 0.9576,
"step": 55000
},
{
"epoch": 0.06,
"learning_rate": 9.999015805811965e-05,
"loss": 0.994,
"step": 56000
},
{
"epoch": 0.06,
"learning_rate": 9.998660418225645e-05,
"loss": 1.1439,
"step": 57000
},
{
"epoch": 0.06,
"learning_rate": 9.998250366089848e-05,
"loss": 1.0131,
"step": 58000
},
{
"epoch": 0.06,
"learning_rate": 9.997785653888835e-05,
"loss": 1.0879,
"step": 59000
},
{
"epoch": 0.06,
"learning_rate": 9.997266286704631e-05,
"loss": 0.925,
"step": 60000
},
{
"epoch": 0.06,
"learning_rate": 9.996692270216947e-05,
"loss": 0.9413,
"step": 61000
},
{
"epoch": 0.06,
"learning_rate": 9.996063610703137e-05,
"loss": 0.8956,
"step": 62000
},
{
"epoch": 0.06,
"learning_rate": 9.995380315038119e-05,
"loss": 0.8672,
"step": 63000
},
{
"epoch": 0.06,
"learning_rate": 9.994642390694308e-05,
"loss": 0.8875,
"step": 64000
},
{
"epoch": 0.07,
"learning_rate": 9.993849845741524e-05,
"loss": 0.9124,
"step": 65000
},
{
"epoch": 0.07,
"learning_rate": 9.993002688846913e-05,
"loss": 0.9428,
"step": 66000
},
{
"epoch": 0.07,
"learning_rate": 9.992100929274846e-05,
"loss": 0.9931,
"step": 67000
},
{
"epoch": 0.07,
"learning_rate": 9.991144576886823e-05,
"loss": 1.1243,
"step": 68000
},
{
"epoch": 0.07,
"learning_rate": 9.990133642141359e-05,
"loss": 0.9759,
"step": 69000
},
{
"epoch": 0.07,
"learning_rate": 9.989068136093873e-05,
"loss": 1.0443,
"step": 70000
},
{
"epoch": 0.07,
"learning_rate": 9.987948070396571e-05,
"loss": 0.9112,
"step": 71000
},
{
"epoch": 0.07,
"learning_rate": 9.986773457298311e-05,
"loss": 0.9212,
"step": 72000
},
{
"epoch": 0.07,
"learning_rate": 9.985544309644475e-05,
"loss": 0.8929,
"step": 73000
},
{
"epoch": 0.07,
"learning_rate": 9.984260640876821e-05,
"loss": 0.8609,
"step": 74000
},
{
"epoch": 0.07,
"learning_rate": 9.98292246503335e-05,
"loss": 0.8797,
"step": 75000
},
{
"epoch": 0.08,
"learning_rate": 9.981529796748134e-05,
"loss": 0.9014,
"step": 76000
},
{
"epoch": 0.08,
"learning_rate": 9.980082651251175e-05,
"loss": 0.9251,
"step": 77000
},
{
"epoch": 0.08,
"learning_rate": 9.97858104436822e-05,
"loss": 0.988,
"step": 78000
},
{
"epoch": 0.08,
"learning_rate": 9.977024992520602e-05,
"loss": 1.0805,
"step": 79000
},
{
"epoch": 0.08,
"learning_rate": 9.975414512725057e-05,
"loss": 0.9457,
"step": 80000
},
{
"epoch": 0.08,
"learning_rate": 9.973749622593534e-05,
"loss": 0.9887,
"step": 81000
},
{
"epoch": 0.08,
"learning_rate": 9.972030340333001e-05,
"loss": 0.8866,
"step": 82000
},
{
"epoch": 0.08,
"learning_rate": 9.970256684745258e-05,
"loss": 0.8843,
"step": 83000
},
{
"epoch": 0.08,
"learning_rate": 9.968428675226714e-05,
"loss": 0.8685,
"step": 84000
},
{
"epoch": 0.09,
"learning_rate": 9.966546331768191e-05,
"loss": 0.8369,
"step": 85000
},
{
"epoch": 0.09,
"learning_rate": 9.964609674954696e-05,
"loss": 0.8639,
"step": 86000
},
{
"epoch": 0.09,
"learning_rate": 9.962618725965196e-05,
"loss": 0.881,
"step": 87000
},
{
"epoch": 0.09,
"learning_rate": 9.96057350657239e-05,
"loss": 0.8954,
"step": 88000
},
{
"epoch": 0.09,
"learning_rate": 9.95847403914247e-05,
"loss": 0.9711,
"step": 89000
},
{
"epoch": 0.09,
"learning_rate": 9.956320346634876e-05,
"loss": 1.0358,
"step": 90000
},
{
"epoch": 0.09,
"learning_rate": 9.954112452602045e-05,
"loss": 0.927,
"step": 91000
},
{
"epoch": 0.09,
"learning_rate": 9.95185038118915e-05,
"loss": 0.9126,
"step": 92000
},
{
"epoch": 0.09,
"learning_rate": 9.949534157133844e-05,
"loss": 0.8692,
"step": 93000
},
{
"epoch": 0.09,
"learning_rate": 9.94716380576598e-05,
"loss": 0.8355,
"step": 94000
},
{
"epoch": 0.1,
"learning_rate": 9.944739353007344e-05,
"loss": 0.8358,
"step": 95000
},
{
"epoch": 0.1,
"learning_rate": 9.942260825371358e-05,
"loss": 0.8086,
"step": 96000
},
{
"epoch": 0.1,
"learning_rate": 9.939728249962807e-05,
"loss": 0.8348,
"step": 97000
},
{
"epoch": 0.1,
"learning_rate": 9.937141654477528e-05,
"loss": 0.8471,
"step": 98000
},
{
"epoch": 0.1,
"learning_rate": 9.934501067202117e-05,
"loss": 0.8733,
"step": 99000
},
{
"epoch": 0.1,
"learning_rate": 9.931806517013612e-05,
"loss": 0.9549,
"step": 100000
},
{
"epoch": 0.1,
"eval_loss": 0.777112603187561,
"eval_runtime": 32.3706,
"eval_samples_per_second": 154.461,
"eval_steps_per_second": 2.44,
"step": 100000
},
{
"epoch": 0.1,
"learning_rate": 9.929058033379181e-05,
"loss": 1.0084,
"step": 101000
},
{
"epoch": 0.1,
"learning_rate": 9.926255646355804e-05,
"loss": 0.9332,
"step": 102000
},
{
"epoch": 0.1,
"learning_rate": 9.923399386589933e-05,
"loss": 0.8486,
"step": 103000
},
{
"epoch": 0.1,
"learning_rate": 9.92048928531717e-05,
"loss": 0.8585,
"step": 104000
},
{
"epoch": 0.1,
"learning_rate": 9.917525374361912e-05,
"loss": 0.8168,
"step": 105000
},
{
"epoch": 0.11,
"learning_rate": 9.914507686137019e-05,
"loss": 0.8111,
"step": 106000
},
{
"epoch": 0.11,
"learning_rate": 9.911436253643445e-05,
"loss": 0.7897,
"step": 107000
},
{
"epoch": 0.11,
"learning_rate": 9.90831111046988e-05,
"loss": 0.8249,
"step": 108000
},
{
"epoch": 0.11,
"learning_rate": 9.905132290792394e-05,
"loss": 0.8336,
"step": 109000
},
{
"epoch": 0.11,
"learning_rate": 9.901899829374047e-05,
"loss": 0.8623,
"step": 110000
},
{
"epoch": 0.11,
"learning_rate": 9.89861376156452e-05,
"loss": 0.9608,
"step": 111000
},
{
"epoch": 0.11,
"learning_rate": 9.895274123299723e-05,
"loss": 0.9681,
"step": 112000
},
{
"epoch": 0.11,
"learning_rate": 9.891880951101407e-05,
"loss": 0.9459,
"step": 113000
},
{
"epoch": 0.11,
"learning_rate": 9.888434282076758e-05,
"loss": 0.7867,
"step": 114000
},
{
"epoch": 0.12,
"learning_rate": 9.884934153917997e-05,
"loss": 0.8544,
"step": 115000
},
{
"epoch": 0.12,
"learning_rate": 9.881380604901964e-05,
"loss": 0.8081,
"step": 116000
},
{
"epoch": 0.12,
"learning_rate": 9.877773673889701e-05,
"loss": 0.7888,
"step": 117000
},
{
"epoch": 0.12,
"learning_rate": 9.87411340032603e-05,
"loss": 0.7873,
"step": 118000
},
{
"epoch": 0.12,
"learning_rate": 9.870399824239117e-05,
"loss": 0.8214,
"step": 119000
},
{
"epoch": 0.12,
"learning_rate": 9.86663298624003e-05,
"loss": 0.8268,
"step": 120000
},
{
"epoch": 0.12,
"learning_rate": 9.862812927522309e-05,
"loss": 0.8562,
"step": 121000
},
{
"epoch": 0.12,
"learning_rate": 9.858939689861506e-05,
"loss": 0.977,
"step": 122000
},
{
"epoch": 0.12,
"learning_rate": 9.855013315614725e-05,
"loss": 0.9267,
"step": 123000
},
{
"epoch": 0.12,
"learning_rate": 9.851033847720166e-05,
"loss": 0.9484,
"step": 124000
},
{
"epoch": 0.12,
"learning_rate": 9.847001329696653e-05,
"loss": 0.7635,
"step": 125000
},
{
"epoch": 0.13,
"learning_rate": 9.842915805643155e-05,
"loss": 0.8435,
"step": 126000
},
{
"epoch": 0.13,
"learning_rate": 9.838777320238312e-05,
"loss": 0.7942,
"step": 127000
},
{
"epoch": 0.13,
"learning_rate": 9.834585918739936e-05,
"loss": 0.7748,
"step": 128000
},
{
"epoch": 0.13,
"learning_rate": 9.830341646984521e-05,
"loss": 0.7881,
"step": 129000
},
{
"epoch": 0.13,
"learning_rate": 9.826044551386744e-05,
"loss": 0.8121,
"step": 130000
},
{
"epoch": 0.13,
"learning_rate": 9.821694678938953e-05,
"loss": 0.8211,
"step": 131000
},
{
"epoch": 0.13,
"learning_rate": 9.817292077210659e-05,
"loss": 0.8477,
"step": 132000
},
{
"epoch": 0.13,
"learning_rate": 9.812836794348004e-05,
"loss": 0.9879,
"step": 133000
},
{
"epoch": 0.13,
"learning_rate": 9.808328879073251e-05,
"loss": 0.8801,
"step": 134000
},
{
"epoch": 0.14,
"learning_rate": 9.803768380684242e-05,
"loss": 0.9278,
"step": 135000
},
{
"epoch": 0.14,
"learning_rate": 9.799155349053851e-05,
"loss": 0.772,
"step": 136000
},
{
"epoch": 0.14,
"learning_rate": 9.794489834629455e-05,
"loss": 0.8189,
"step": 137000
},
{
"epoch": 0.14,
"learning_rate": 9.789771888432375e-05,
"loss": 0.7837,
"step": 138000
},
{
"epoch": 0.14,
"learning_rate": 9.785001562057309e-05,
"loss": 0.7684,
"step": 139000
},
{
"epoch": 0.14,
"learning_rate": 9.780178907671789e-05,
"loss": 0.7807,
"step": 140000
},
{
"epoch": 0.14,
"learning_rate": 9.775303978015585e-05,
"loss": 0.8037,
"step": 141000
},
{
"epoch": 0.14,
"learning_rate": 9.77037682640015e-05,
"loss": 0.8205,
"step": 142000
},
{
"epoch": 0.14,
"learning_rate": 9.765397506708023e-05,
"loss": 0.8609,
"step": 143000
},
{
"epoch": 0.14,
"learning_rate": 9.760366073392246e-05,
"loss": 0.9872,
"step": 144000
},
{
"epoch": 0.14,
"learning_rate": 9.755282581475769e-05,
"loss": 0.8561,
"step": 145000
},
{
"epoch": 0.15,
"learning_rate": 9.750147086550844e-05,
"loss": 0.9079,
"step": 146000
},
{
"epoch": 0.15,
"learning_rate": 9.744959644778422e-05,
"loss": 0.7751,
"step": 147000
},
{
"epoch": 0.15,
"learning_rate": 9.739720312887535e-05,
"loss": 0.7998,
"step": 148000
},
{
"epoch": 0.15,
"learning_rate": 9.734429148174675e-05,
"loss": 0.7823,
"step": 149000
},
{
"epoch": 0.15,
"learning_rate": 9.729086208503174e-05,
"loss": 0.7551,
"step": 150000
},
{
"epoch": 0.15,
"eval_loss": 0.7697679400444031,
"eval_runtime": 22.8281,
"eval_samples_per_second": 219.028,
"eval_steps_per_second": 3.461,
"step": 150000
},
{
"epoch": 0.15,
"learning_rate": 9.723691552302562e-05,
"loss": 0.7738,
"step": 151000
},
{
"epoch": 0.15,
"learning_rate": 9.718245238567939e-05,
"loss": 0.7963,
"step": 152000
},
{
"epoch": 0.15,
"learning_rate": 9.712747326859315e-05,
"loss": 0.8171,
"step": 153000
},
{
"epoch": 0.15,
"learning_rate": 9.707197877300974e-05,
"loss": 0.8773,
"step": 154000
},
{
"epoch": 0.15,
"learning_rate": 9.701596950580806e-05,
"loss": 0.9735,
"step": 155000
},
{
"epoch": 0.16,
"learning_rate": 9.695944607949649e-05,
"loss": 0.8433,
"step": 156000
},
{
"epoch": 0.16,
"learning_rate": 9.690240911220618e-05,
"loss": 0.8706,
"step": 157000
},
{
"epoch": 0.16,
"learning_rate": 9.684485922768422e-05,
"loss": 0.7855,
"step": 158000
},
{
"epoch": 0.16,
"learning_rate": 9.6786797055287e-05,
"loss": 0.785,
"step": 159000
},
{
"epoch": 0.16,
"learning_rate": 9.672822322997305e-05,
"loss": 0.765,
"step": 160000
},
{
"epoch": 0.16,
"learning_rate": 9.66691383922964e-05,
"loss": 0.7384,
"step": 161000
},
{
"epoch": 0.16,
"learning_rate": 9.660954318839933e-05,
"loss": 0.7583,
"step": 162000
},
{
"epoch": 0.16,
"learning_rate": 9.654943827000548e-05,
"loss": 0.7785,
"step": 163000
},
{
"epoch": 0.16,
"learning_rate": 9.648882429441257e-05,
"loss": 0.8002,
"step": 164000
},
{
"epoch": 0.17,
"learning_rate": 9.642770192448536e-05,
"loss": 0.8721,
"step": 165000
},
{
"epoch": 0.17,
"learning_rate": 9.636607182864827e-05,
"loss": 0.9439,
"step": 166000
},
{
"epoch": 0.17,
"learning_rate": 9.630393468087818e-05,
"loss": 0.8458,
"step": 167000
},
{
"epoch": 0.17,
"learning_rate": 9.624129116069694e-05,
"loss": 0.8241,
"step": 168000
},
{
"epoch": 0.17,
"learning_rate": 9.617814195316411e-05,
"loss": 0.7928,
"step": 169000
},
{
"epoch": 0.17,
"learning_rate": 9.611448774886924e-05,
"loss": 0.7656,
"step": 170000
},
{
"epoch": 0.17,
"learning_rate": 9.605032924392457e-05,
"loss": 0.7647,
"step": 171000
},
{
"epoch": 0.17,
"learning_rate": 9.598566713995718e-05,
"loss": 0.7523,
"step": 172000
},
{
"epoch": 0.17,
"learning_rate": 9.59205021441015e-05,
"loss": 0.7739,
"step": 173000
},
{
"epoch": 0.17,
"learning_rate": 9.58548349689915e-05,
"loss": 0.7857,
"step": 174000
},
{
"epoch": 0.17,
"learning_rate": 9.578866633275288e-05,
"loss": 0.8132,
"step": 175000
},
{
"epoch": 0.18,
"learning_rate": 9.572199695899522e-05,
"loss": 0.8928,
"step": 176000
},
{
"epoch": 0.18,
"learning_rate": 9.565482757680415e-05,
"loss": 0.9479,
"step": 177000
},
{
"epoch": 0.18,
"learning_rate": 9.558715892073323e-05,
"loss": 0.8774,
"step": 178000
},
{
"epoch": 0.18,
"learning_rate": 9.551899173079607e-05,
"loss": 0.7822,
"step": 179000
},
{
"epoch": 0.18,
"learning_rate": 9.545032675245813e-05,
"loss": 0.7981,
"step": 180000
},
{
"epoch": 0.18,
"learning_rate": 9.538116473662861e-05,
"loss": 0.7582,
"step": 181000
},
{
"epoch": 0.18,
"learning_rate": 9.531150643965223e-05,
"loss": 0.7543,
"step": 182000
},
{
"epoch": 0.18,
"learning_rate": 9.524135262330098e-05,
"loss": 0.741,
"step": 183000
},
{
"epoch": 0.18,
"learning_rate": 9.517070405476575e-05,
"loss": 0.7712,
"step": 184000
},
{
"epoch": 0.18,
"learning_rate": 9.509956150664796e-05,
"loss": 0.7824,
"step": 185000
},
{
"epoch": 0.19,
"learning_rate": 9.502792575695112e-05,
"loss": 0.8096,
"step": 186000
},
{
"epoch": 0.19,
"learning_rate": 9.49557975890723e-05,
"loss": 0.9033,
"step": 187000
},
{
"epoch": 0.19,
"learning_rate": 9.488317779179361e-05,
"loss": 0.9156,
"step": 188000
},
{
"epoch": 0.19,
"learning_rate": 9.481006715927351e-05,
"loss": 0.896,
"step": 189000
},
{
"epoch": 0.19,
"learning_rate": 9.473646649103818e-05,
"loss": 0.7307,
"step": 190000
},
{
"epoch": 0.19,
"learning_rate": 9.46623765919727e-05,
"loss": 0.8032,
"step": 191000
},
{
"epoch": 0.19,
"learning_rate": 9.458779827231237e-05,
"loss": 0.7586,
"step": 192000
},
{
"epoch": 0.19,
"learning_rate": 9.451273234763371e-05,
"loss": 0.7422,
"step": 193000
},
{
"epoch": 0.19,
"learning_rate": 9.443717963884569e-05,
"loss": 0.7412,
"step": 194000
},
{
"epoch": 0.2,
"learning_rate": 9.43611409721806e-05,
"loss": 0.7738,
"step": 195000
},
{
"epoch": 0.2,
"learning_rate": 9.428461717918511e-05,
"loss": 0.7806,
"step": 196000
},
{
"epoch": 0.2,
"learning_rate": 9.420760909671118e-05,
"loss": 0.807,
"step": 197000
},
{
"epoch": 0.2,
"learning_rate": 9.413011756690685e-05,
"loss": 0.9269,
"step": 198000
},
{
"epoch": 0.2,
"learning_rate": 9.405214343720707e-05,
"loss": 0.8776,
"step": 199000
},
{
"epoch": 0.2,
"learning_rate": 9.397368756032445e-05,
"loss": 0.8979,
"step": 200000
},
{
"epoch": 0.2,
"eval_loss": 0.7604823112487793,
"eval_runtime": 22.7684,
"eval_samples_per_second": 219.603,
"eval_steps_per_second": 3.47,
"step": 200000
}
],
"max_steps": 1000000,
"num_train_epochs": 9223372036854775807,
"total_flos": 1.19299796631552e+19,
"trial_name": null,
"trial_params": null
}