{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 6315,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.02,
      "learning_rate": 1.9841646872525734e-05,
      "loss": 0.6389,
      "step": 50
    },
    {
      "epoch": 0.05,
      "learning_rate": 1.9683293745051467e-05,
      "loss": 0.4715,
      "step": 100
    },
    {
      "epoch": 0.07,
      "learning_rate": 1.95249406175772e-05,
      "loss": 0.4132,
      "step": 150
    },
    {
      "epoch": 0.1,
      "learning_rate": 1.936658749010293e-05,
      "loss": 0.3877,
      "step": 200
    },
    {
      "epoch": 0.12,
      "learning_rate": 1.9208234362628665e-05,
      "loss": 0.3773,
      "step": 250
    },
    {
      "epoch": 0.14,
      "learning_rate": 1.9049881235154395e-05,
      "loss": 0.375,
      "step": 300
    },
    {
      "epoch": 0.17,
      "learning_rate": 1.8891528107680128e-05,
      "loss": 0.3556,
      "step": 350
    },
    {
      "epoch": 0.19,
      "learning_rate": 1.873317498020586e-05,
      "loss": 0.3407,
      "step": 400
    },
    {
      "epoch": 0.21,
      "learning_rate": 1.8574821852731593e-05,
      "loss": 0.3301,
      "step": 450
    },
    {
      "epoch": 0.24,
      "learning_rate": 1.8416468725257326e-05,
      "loss": 0.3449,
      "step": 500
    },
    {
      "epoch": 0.24,
      "eval_accuracy": 0.856651376146789,
      "eval_loss": 0.3367556929588318,
      "eval_runtime": 1.965,
      "eval_samples_per_second": 443.767,
      "eval_steps_per_second": 27.99,
      "step": 500
    },
    {
      "epoch": 0.26,
      "learning_rate": 1.825811559778306e-05,
      "loss": 0.357,
      "step": 550
    },
    {
      "epoch": 0.29,
      "learning_rate": 1.809976247030879e-05,
      "loss": 0.3244,
      "step": 600
    },
    {
      "epoch": 0.31,
      "learning_rate": 1.794140934283452e-05,
      "loss": 0.3156,
      "step": 650
    },
    {
      "epoch": 0.33,
      "learning_rate": 1.7783056215360253e-05,
      "loss": 0.2856,
      "step": 700
    },
    {
      "epoch": 0.36,
      "learning_rate": 1.762470308788599e-05,
      "loss": 0.3261,
      "step": 750
    },
    {
      "epoch": 0.38,
      "learning_rate": 1.746634996041172e-05,
      "loss": 0.2993,
      "step": 800
    },
    {
      "epoch": 0.4,
      "learning_rate": 1.730799683293745e-05,
      "loss": 0.3101,
      "step": 850
    },
    {
      "epoch": 0.43,
      "learning_rate": 1.7149643705463184e-05,
      "loss": 0.2983,
      "step": 900
    },
    {
      "epoch": 0.45,
      "learning_rate": 1.6991290577988917e-05,
      "loss": 0.2862,
      "step": 950
    },
    {
      "epoch": 0.48,
      "learning_rate": 1.6832937450514647e-05,
      "loss": 0.2987,
      "step": 1000
    },
    {
      "epoch": 0.48,
      "eval_accuracy": 0.8715596330275229,
      "eval_loss": 0.3037159740924835,
      "eval_runtime": 1.9639,
      "eval_samples_per_second": 444.01,
      "eval_steps_per_second": 28.005,
      "step": 1000
    },
    {
      "epoch": 0.5,
      "learning_rate": 1.6674584323040383e-05,
      "loss": 0.275,
      "step": 1050
    },
    {
      "epoch": 0.52,
      "learning_rate": 1.6516231195566115e-05,
      "loss": 0.2834,
      "step": 1100
    },
    {
      "epoch": 0.55,
      "learning_rate": 1.6357878068091845e-05,
      "loss": 0.2945,
      "step": 1150
    },
    {
      "epoch": 0.57,
      "learning_rate": 1.6199524940617578e-05,
      "loss": 0.2662,
      "step": 1200
    },
    {
      "epoch": 0.59,
      "learning_rate": 1.604117181314331e-05,
      "loss": 0.2675,
      "step": 1250
    },
    {
      "epoch": 0.62,
      "learning_rate": 1.5882818685669043e-05,
      "loss": 0.2683,
      "step": 1300
    },
    {
      "epoch": 0.64,
      "learning_rate": 1.5724465558194776e-05,
      "loss": 0.2769,
      "step": 1350
    },
    {
      "epoch": 0.67,
      "learning_rate": 1.556611243072051e-05,
      "loss": 0.288,
      "step": 1400
    },
    {
      "epoch": 0.69,
      "learning_rate": 1.540775930324624e-05,
      "loss": 0.2738,
      "step": 1450
    },
    {
      "epoch": 0.71,
      "learning_rate": 1.5249406175771972e-05,
      "loss": 0.2492,
      "step": 1500
    },
    {
      "epoch": 0.71,
      "eval_accuracy": 0.8841743119266054,
      "eval_loss": 0.33466824889183044,
      "eval_runtime": 1.9704,
      "eval_samples_per_second": 442.547,
      "eval_steps_per_second": 27.913,
      "step": 1500
    },
    {
      "epoch": 0.74,
      "learning_rate": 1.5091053048297705e-05,
      "loss": 0.2659,
      "step": 1550
    },
    {
      "epoch": 0.76,
      "learning_rate": 1.4932699920823436e-05,
      "loss": 0.2639,
      "step": 1600
    },
    {
      "epoch": 0.78,
      "learning_rate": 1.477434679334917e-05,
      "loss": 0.2343,
      "step": 1650
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.4615993665874902e-05,
      "loss": 0.2715,
      "step": 1700
    },
    {
      "epoch": 0.83,
      "learning_rate": 1.4457640538400634e-05,
      "loss": 0.2496,
      "step": 1750
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.4299287410926367e-05,
      "loss": 0.2757,
      "step": 1800
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.4140934283452098e-05,
      "loss": 0.2486,
      "step": 1850
    },
    {
      "epoch": 0.9,
      "learning_rate": 1.3982581155977833e-05,
      "loss": 0.2602,
      "step": 1900
    },
    {
      "epoch": 0.93,
      "learning_rate": 1.3824228028503564e-05,
      "loss": 0.2215,
      "step": 1950
    },
    {
      "epoch": 0.95,
      "learning_rate": 1.3665874901029297e-05,
      "loss": 0.24,
      "step": 2000
    },
    {
      "epoch": 0.95,
      "eval_accuracy": 0.8830275229357798,
      "eval_loss": 0.2952757775783539,
      "eval_runtime": 1.9706,
      "eval_samples_per_second": 442.496,
      "eval_steps_per_second": 27.91,
      "step": 2000
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.3507521773555028e-05,
      "loss": 0.2555,
      "step": 2050
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.3349168646080762e-05,
      "loss": 0.2616,
      "step": 2100
    },
    {
      "epoch": 1.02,
      "learning_rate": 1.3190815518606495e-05,
      "loss": 0.202,
      "step": 2150
    },
    {
      "epoch": 1.05,
      "learning_rate": 1.3032462391132226e-05,
      "loss": 0.2014,
      "step": 2200
    },
    {
      "epoch": 1.07,
      "learning_rate": 1.2874109263657959e-05,
      "loss": 0.1972,
      "step": 2250
    },
    {
      "epoch": 1.09,
      "learning_rate": 1.271575613618369e-05,
      "loss": 0.1989,
      "step": 2300
    },
    {
      "epoch": 1.12,
      "learning_rate": 1.2557403008709424e-05,
      "loss": 0.177,
      "step": 2350
    },
    {
      "epoch": 1.14,
      "learning_rate": 1.2399049881235155e-05,
      "loss": 0.1908,
      "step": 2400
    },
    {
      "epoch": 1.16,
      "learning_rate": 1.2240696753760888e-05,
      "loss": 0.2098,
      "step": 2450
    },
    {
      "epoch": 1.19,
      "learning_rate": 1.208234362628662e-05,
      "loss": 0.195,
      "step": 2500
    },
    {
      "epoch": 1.19,
      "eval_accuracy": 0.8841743119266054,
      "eval_loss": 0.34449705481529236,
      "eval_runtime": 1.966,
      "eval_samples_per_second": 443.536,
      "eval_steps_per_second": 27.975,
      "step": 2500
    },
    {
      "epoch": 1.21,
      "learning_rate": 1.1923990498812352e-05,
      "loss": 0.1947,
      "step": 2550
    },
    {
      "epoch": 1.24,
      "learning_rate": 1.1765637371338086e-05,
      "loss": 0.209,
      "step": 2600
    },
    {
      "epoch": 1.26,
      "learning_rate": 1.1607284243863817e-05,
      "loss": 0.2054,
      "step": 2650
    },
    {
      "epoch": 1.28,
      "learning_rate": 1.144893111638955e-05,
      "loss": 0.1899,
      "step": 2700
    },
    {
      "epoch": 1.31,
      "learning_rate": 1.1290577988915281e-05,
      "loss": 0.2151,
      "step": 2750
    },
    {
      "epoch": 1.33,
      "learning_rate": 1.1132224861441014e-05,
      "loss": 0.1973,
      "step": 2800
    },
    {
      "epoch": 1.35,
      "learning_rate": 1.0973871733966748e-05,
      "loss": 0.212,
      "step": 2850
    },
    {
      "epoch": 1.38,
      "learning_rate": 1.081551860649248e-05,
      "loss": 0.2076,
      "step": 2900
    },
    {
      "epoch": 1.4,
      "learning_rate": 1.0657165479018212e-05,
      "loss": 0.1836,
      "step": 2950
    },
    {
      "epoch": 1.43,
      "learning_rate": 1.0498812351543943e-05,
      "loss": 0.1934,
      "step": 3000
    },
    {
      "epoch": 1.43,
      "eval_accuracy": 0.8876146788990825,
      "eval_loss": 0.32173678278923035,
      "eval_runtime": 1.9731,
      "eval_samples_per_second": 441.938,
      "eval_steps_per_second": 27.875,
      "step": 3000
    },
    {
      "epoch": 1.45,
      "learning_rate": 1.0340459224069676e-05,
      "loss": 0.191,
      "step": 3050
    },
    {
      "epoch": 1.47,
      "learning_rate": 1.0182106096595407e-05,
      "loss": 0.1861,
      "step": 3100
    },
    {
      "epoch": 1.5,
      "learning_rate": 1.0023752969121141e-05,
      "loss": 0.2138,
      "step": 3150
    },
    {
      "epoch": 1.52,
      "learning_rate": 9.865399841646872e-06,
      "loss": 0.1771,
      "step": 3200
    },
    {
      "epoch": 1.54,
      "learning_rate": 9.707046714172605e-06,
      "loss": 0.2223,
      "step": 3250
    },
    {
      "epoch": 1.57,
      "learning_rate": 9.548693586698338e-06,
      "loss": 0.2112,
      "step": 3300
    },
    {
      "epoch": 1.59,
      "learning_rate": 9.39034045922407e-06,
      "loss": 0.1988,
      "step": 3350
    },
    {
      "epoch": 1.62,
      "learning_rate": 9.231987331749803e-06,
      "loss": 0.2039,
      "step": 3400
    },
    {
      "epoch": 1.64,
      "learning_rate": 9.073634204275536e-06,
      "loss": 0.188,
      "step": 3450
    },
    {
      "epoch": 1.66,
      "learning_rate": 8.915281076801267e-06,
      "loss": 0.1697,
      "step": 3500
    },
    {
      "epoch": 1.66,
      "eval_accuracy": 0.8876146788990825,
      "eval_loss": 0.3626970052719116,
      "eval_runtime": 1.9706,
      "eval_samples_per_second": 442.514,
      "eval_steps_per_second": 27.911,
      "step": 3500
    },
    {
      "epoch": 1.69,
      "learning_rate": 8.756927949327e-06,
      "loss": 0.194,
      "step": 3550
    },
    {
      "epoch": 1.71,
      "learning_rate": 8.598574821852733e-06,
      "loss": 0.1905,
      "step": 3600
    },
    {
      "epoch": 1.73,
      "learning_rate": 8.440221694378464e-06,
      "loss": 0.1873,
      "step": 3650
    },
    {
      "epoch": 1.76,
      "learning_rate": 8.281868566904197e-06,
      "loss": 0.2015,
      "step": 3700
    },
    {
      "epoch": 1.78,
      "learning_rate": 8.12351543942993e-06,
      "loss": 0.1779,
      "step": 3750
    },
    {
      "epoch": 1.81,
      "learning_rate": 7.965162311955662e-06,
      "loss": 0.1796,
      "step": 3800
    },
    {
      "epoch": 1.83,
      "learning_rate": 7.806809184481395e-06,
      "loss": 0.1825,
      "step": 3850
    },
    {
      "epoch": 1.85,
      "learning_rate": 7.648456057007126e-06,
      "loss": 0.181,
      "step": 3900
    },
    {
      "epoch": 1.88,
      "learning_rate": 7.490102929532859e-06,
      "loss": 0.1814,
      "step": 3950
    },
    {
      "epoch": 1.9,
      "learning_rate": 7.331749802058591e-06,
      "loss": 0.1757,
      "step": 4000
    },
    {
      "epoch": 1.9,
      "eval_accuracy": 0.8899082568807339,
      "eval_loss": 0.33660009503364563,
      "eval_runtime": 1.9723,
      "eval_samples_per_second": 442.128,
      "eval_steps_per_second": 27.887,
      "step": 4000
    },
    {
      "epoch": 1.92,
      "learning_rate": 7.173396674584323e-06,
      "loss": 0.1808,
      "step": 4050
    },
    {
      "epoch": 1.95,
      "learning_rate": 7.015043547110056e-06,
      "loss": 0.1825,
      "step": 4100
    },
    {
      "epoch": 1.97,
      "learning_rate": 6.856690419635789e-06,
      "loss": 0.2018,
      "step": 4150
    },
    {
      "epoch": 2.0,
      "learning_rate": 6.698337292161521e-06,
      "loss": 0.158,
      "step": 4200
    },
    {
      "epoch": 2.02,
      "learning_rate": 6.539984164687253e-06,
      "loss": 0.1539,
      "step": 4250
    },
    {
      "epoch": 2.04,
      "learning_rate": 6.3816310372129854e-06,
      "loss": 0.1381,
      "step": 4300
    },
    {
      "epoch": 2.07,
      "learning_rate": 6.223277909738717e-06,
      "loss": 0.1471,
      "step": 4350
    },
    {
      "epoch": 2.09,
      "learning_rate": 6.06492478226445e-06,
      "loss": 0.137,
      "step": 4400
    },
    {
      "epoch": 2.11,
      "learning_rate": 5.906571654790183e-06,
      "loss": 0.1604,
      "step": 4450
    },
    {
      "epoch": 2.14,
      "learning_rate": 5.748218527315916e-06,
      "loss": 0.1328,
      "step": 4500
    },
    {
      "epoch": 2.14,
      "eval_accuracy": 0.8876146788990825,
      "eval_loss": 0.42656058073043823,
      "eval_runtime": 1.9787,
      "eval_samples_per_second": 440.687,
      "eval_steps_per_second": 27.796,
      "step": 4500
    },
    {
      "epoch": 2.16,
      "learning_rate": 5.5898653998416475e-06,
      "loss": 0.1213,
      "step": 4550
    },
    {
      "epoch": 2.19,
      "learning_rate": 5.431512272367379e-06,
      "loss": 0.1861,
      "step": 4600
    },
    {
      "epoch": 2.21,
      "learning_rate": 5.273159144893112e-06,
      "loss": 0.1769,
      "step": 4650
    },
    {
      "epoch": 2.23,
      "learning_rate": 5.114806017418844e-06,
      "loss": 0.1544,
      "step": 4700
    },
    {
      "epoch": 2.26,
      "learning_rate": 4.956452889944577e-06,
      "loss": 0.1325,
      "step": 4750
    },
    {
      "epoch": 2.28,
      "learning_rate": 4.798099762470309e-06,
      "loss": 0.1441,
      "step": 4800
    },
    {
      "epoch": 2.3,
      "learning_rate": 4.6397466349960415e-06,
      "loss": 0.1411,
      "step": 4850
    },
    {
      "epoch": 2.33,
      "learning_rate": 4.481393507521774e-06,
      "loss": 0.1465,
      "step": 4900
    },
    {
      "epoch": 2.35,
      "learning_rate": 4.323040380047506e-06,
      "loss": 0.1559,
      "step": 4950
    },
    {
      "epoch": 2.38,
      "learning_rate": 4.164687252573238e-06,
      "loss": 0.1475,
      "step": 5000
    },
    {
      "epoch": 2.38,
      "eval_accuracy": 0.893348623853211,
      "eval_loss": 0.3737166225910187,
      "eval_runtime": 1.9731,
      "eval_samples_per_second": 441.952,
      "eval_steps_per_second": 27.875,
      "step": 5000
    },
    {
      "epoch": 2.4,
      "learning_rate": 4.006334125098972e-06,
      "loss": 0.1504,
      "step": 5050
    },
    {
      "epoch": 2.42,
      "learning_rate": 3.8479809976247036e-06,
      "loss": 0.1497,
      "step": 5100
    },
    {
      "epoch": 2.45,
      "learning_rate": 3.6896278701504355e-06,
      "loss": 0.1434,
      "step": 5150
    },
    {
      "epoch": 2.47,
      "learning_rate": 3.531274742676168e-06,
      "loss": 0.15,
      "step": 5200
    },
    {
      "epoch": 2.49,
      "learning_rate": 3.3729216152019006e-06,
      "loss": 0.1408,
      "step": 5250
    },
    {
      "epoch": 2.52,
      "learning_rate": 3.214568487727633e-06,
      "loss": 0.1207,
      "step": 5300
    },
    {
      "epoch": 2.54,
      "learning_rate": 3.0562153602533652e-06,
      "loss": 0.1482,
      "step": 5350
    },
    {
      "epoch": 2.57,
      "learning_rate": 2.897862232779098e-06,
      "loss": 0.1478,
      "step": 5400
    },
    {
      "epoch": 2.59,
      "learning_rate": 2.73950910530483e-06,
      "loss": 0.1413,
      "step": 5450
    },
    {
      "epoch": 2.61,
      "learning_rate": 2.581155977830562e-06,
      "loss": 0.1574,
      "step": 5500
    },
    {
      "epoch": 2.61,
      "eval_accuracy": 0.8910550458715596,
      "eval_loss": 0.3887576162815094,
      "eval_runtime": 1.9592,
      "eval_samples_per_second": 445.071,
      "eval_steps_per_second": 28.072,
      "step": 5500
    },
    {
      "epoch": 2.64,
      "learning_rate": 2.422802850356295e-06,
      "loss": 0.1543,
      "step": 5550
    },
    {
      "epoch": 2.66,
      "learning_rate": 2.264449722882027e-06,
      "loss": 0.1419,
      "step": 5600
    },
    {
      "epoch": 2.68,
      "learning_rate": 2.1060965954077596e-06,
      "loss": 0.142,
      "step": 5650
    },
    {
      "epoch": 2.71,
      "learning_rate": 1.947743467933492e-06,
      "loss": 0.1444,
      "step": 5700
    },
    {
      "epoch": 2.73,
      "learning_rate": 1.789390340459224e-06,
      "loss": 0.1202,
      "step": 5750
    },
    {
      "epoch": 2.76,
      "learning_rate": 1.6310372129849566e-06,
      "loss": 0.1473,
      "step": 5800
    },
    {
      "epoch": 2.78,
      "learning_rate": 1.472684085510689e-06,
      "loss": 0.1453,
      "step": 5850
    },
    {
      "epoch": 2.8,
      "learning_rate": 1.3143309580364213e-06,
      "loss": 0.1293,
      "step": 5900
    },
    {
      "epoch": 2.83,
      "learning_rate": 1.1559778305621536e-06,
      "loss": 0.1273,
      "step": 5950
    },
    {
      "epoch": 2.85,
      "learning_rate": 9.976247030878861e-07,
      "loss": 0.1548,
      "step": 6000
    },
    {
      "epoch": 2.85,
      "eval_accuracy": 0.8864678899082569,
      "eval_loss": 0.40631914138793945,
      "eval_runtime": 1.98,
      "eval_samples_per_second": 440.401,
      "eval_steps_per_second": 27.778,
      "step": 6000
    },
    {
      "epoch": 2.87,
      "learning_rate": 8.392715756136184e-07,
      "loss": 0.1528,
      "step": 6050
    },
    {
      "epoch": 2.9,
      "learning_rate": 6.809184481393508e-07,
      "loss": 0.1205,
      "step": 6100
    },
    {
      "epoch": 2.92,
      "learning_rate": 5.225653206650831e-07,
      "loss": 0.1327,
      "step": 6150
    },
    {
      "epoch": 2.95,
      "learning_rate": 3.642121931908155e-07,
      "loss": 0.1665,
      "step": 6200
    },
    {
      "epoch": 2.97,
      "learning_rate": 2.0585906571654793e-07,
      "loss": 0.1528,
      "step": 6250
    },
    {
      "epoch": 2.99,
      "learning_rate": 4.750593824228029e-08,
      "loss": 0.1353,
      "step": 6300
    },
    {
      "epoch": 3.0,
      "step": 6315,
      "total_flos": 1.016719948492212e+16,
      "train_loss": 0.2152629395090014,
      "train_runtime": 1626.1548,
      "train_samples_per_second": 124.248,
      "train_steps_per_second": 3.883
    }
  ],
  "logging_steps": 50,
  "max_steps": 6315,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "total_flos": 1.016719948492212e+16,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}