{
  "best_metric": 3.1790146827697754,
  "best_model_checkpoint": "./output/models/rotating-head-gp-gpt2-medium-wikitext/checkpoint-10500",
  "epoch": 6.0,
  "eval_steps": 500,
  "global_step": 10692,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.05611672278338945,
      "grad_norm": 1.7026984691619873,
      "learning_rate": 1.1223344556677892e-05,
      "loss": 8.993,
      "step": 100
    },
    {
      "epoch": 0.1122334455667789,
      "grad_norm": 1.4564013481140137,
      "learning_rate": 2.2446689113355783e-05,
      "loss": 7.3838,
      "step": 200
    },
    {
      "epoch": 0.16835016835016836,
      "grad_norm": 1.6578171253204346,
      "learning_rate": 3.3670033670033675e-05,
      "loss": 6.5796,
      "step": 300
    },
    {
      "epoch": 0.2244668911335578,
      "grad_norm": 2.0035881996154785,
      "learning_rate": 4.4893378226711566e-05,
      "loss": 6.1995,
      "step": 400
    },
    {
      "epoch": 0.28058361391694725,
      "grad_norm": 1.3597742319107056,
      "learning_rate": 5.611672278338945e-05,
      "loss": 5.9062,
      "step": 500
    },
    {
      "epoch": 0.28058361391694725,
      "eval_accuracy": 0.22335966032779805,
      "eval_bleu": 0.0492964548196352,
      "eval_loss": 5.746989727020264,
      "eval_perplexity": 313.2462827607658,
      "eval_runtime": 12.1113,
      "eval_samples_per_second": 94.209,
      "eval_steps_per_second": 1.486,
      "step": 500
    },
    {
      "epoch": 0.3367003367003367,
      "grad_norm": 1.2420822381973267,
      "learning_rate": 6.734006734006735e-05,
      "loss": 5.6583,
      "step": 600
    },
    {
      "epoch": 0.39281705948372614,
      "grad_norm": 1.7358133792877197,
      "learning_rate": 7.856341189674523e-05,
      "loss": 5.4354,
      "step": 700
    },
    {
      "epoch": 0.4489337822671156,
      "grad_norm": 1.1559091806411743,
      "learning_rate": 8.978675645342313e-05,
      "loss": 5.252,
      "step": 800
    },
    {
      "epoch": 0.5050505050505051,
      "grad_norm": 1.1872116327285767,
      "learning_rate": 9.988776655443322e-05,
      "loss": 5.0452,
      "step": 900
    },
    {
      "epoch": 0.5611672278338945,
      "grad_norm": 1.0503571033477783,
      "learning_rate": 9.864072827035791e-05,
      "loss": 4.8598,
      "step": 1000
    },
    {
      "epoch": 0.5611672278338945,
      "eval_accuracy": 0.28109999374594663,
      "eval_bleu": 0.06981793198064049,
      "eval_loss": 4.74280309677124,
      "eval_perplexity": 114.75542181664323,
      "eval_runtime": 12.1357,
      "eval_samples_per_second": 94.02,
      "eval_steps_per_second": 1.483,
      "step": 1000
    },
    {
      "epoch": 0.6172839506172839,
      "grad_norm": 0.9331828951835632,
      "learning_rate": 9.73936899862826e-05,
      "loss": 4.7116,
      "step": 1100
    },
    {
      "epoch": 0.6734006734006734,
      "grad_norm": 0.8568651676177979,
      "learning_rate": 9.614665170220725e-05,
      "loss": 4.5871,
      "step": 1200
    },
    {
      "epoch": 0.7295173961840629,
      "grad_norm": 0.7943041920661926,
      "learning_rate": 9.489961341813194e-05,
      "loss": 4.48,
      "step": 1300
    },
    {
      "epoch": 0.7856341189674523,
      "grad_norm": 0.929481029510498,
      "learning_rate": 9.365257513405662e-05,
      "loss": 4.3936,
      "step": 1400
    },
    {
      "epoch": 0.8417508417508418,
      "grad_norm": 0.9010036587715149,
      "learning_rate": 9.24055368499813e-05,
      "loss": 4.3025,
      "step": 1500
    },
    {
      "epoch": 0.8417508417508418,
      "eval_accuracy": 0.31702567503082046,
      "eval_bleu": 0.08337643356737867,
      "eval_loss": 4.23293399810791,
      "eval_perplexity": 68.91914446331879,
      "eval_runtime": 12.1629,
      "eval_samples_per_second": 93.809,
      "eval_steps_per_second": 1.48,
      "step": 1500
    },
    {
      "epoch": 0.8978675645342312,
      "grad_norm": 0.8791123628616333,
      "learning_rate": 9.115849856590598e-05,
      "loss": 4.2455,
      "step": 1600
    },
    {
      "epoch": 0.9539842873176206,
      "grad_norm": 0.938328742980957,
      "learning_rate": 8.991146028183066e-05,
      "loss": 4.1605,
      "step": 1700
    },
    {
      "epoch": 1.0101010101010102,
      "grad_norm": 1.0388487577438354,
      "learning_rate": 8.866442199775533e-05,
      "loss": 4.1011,
      "step": 1800
    },
    {
      "epoch": 1.0662177328843996,
      "grad_norm": 0.9663709998130798,
      "learning_rate": 8.741738371368002e-05,
      "loss": 4.0148,
      "step": 1900
    },
    {
      "epoch": 1.122334455667789,
      "grad_norm": 0.8401734828948975,
      "learning_rate": 8.617034542960469e-05,
      "loss": 3.9635,
      "step": 2000
    },
    {
      "epoch": 1.122334455667789,
      "eval_accuracy": 0.34535482328872397,
      "eval_bleu": 0.09320946829041966,
      "eval_loss": 3.9290566444396973,
      "eval_perplexity": 50.85897693319932,
      "eval_runtime": 12.2142,
      "eval_samples_per_second": 93.416,
      "eval_steps_per_second": 1.474,
      "step": 2000
    },
    {
      "epoch": 1.1784511784511784,
      "grad_norm": 1.1741111278533936,
      "learning_rate": 8.492330714552937e-05,
      "loss": 3.9219,
      "step": 2100
    },
    {
      "epoch": 1.2345679012345678,
      "grad_norm": 1.1291131973266602,
      "learning_rate": 8.367626886145406e-05,
      "loss": 3.876,
      "step": 2200
    },
    {
      "epoch": 1.2906846240179575,
      "grad_norm": 1.0407253503799438,
      "learning_rate": 8.242923057737873e-05,
      "loss": 3.8557,
      "step": 2300
    },
    {
      "epoch": 1.3468013468013469,
      "grad_norm": 1.0766143798828125,
      "learning_rate": 8.11821922933034e-05,
      "loss": 3.8141,
      "step": 2400
    },
    {
      "epoch": 1.4029180695847363,
      "grad_norm": 1.42284095287323,
      "learning_rate": 7.993515400922809e-05,
      "loss": 3.7769,
      "step": 2500
    },
    {
      "epoch": 1.4029180695847363,
      "eval_accuracy": 0.36358581717774274,
      "eval_bleu": 0.10197996736893164,
      "eval_loss": 3.7426531314849854,
      "eval_perplexity": 42.209829965401774,
      "eval_runtime": 12.1768,
      "eval_samples_per_second": 93.703,
      "eval_steps_per_second": 1.478,
      "step": 2500
    },
    {
      "epoch": 1.4590347923681257,
      "grad_norm": 1.3685775995254517,
      "learning_rate": 7.868811572515277e-05,
      "loss": 3.7592,
      "step": 2600
    },
    {
      "epoch": 1.5151515151515151,
      "grad_norm": 1.6901907920837402,
      "learning_rate": 7.744107744107744e-05,
      "loss": 3.7311,
      "step": 2700
    },
    {
      "epoch": 1.5712682379349046,
      "grad_norm": 1.2826731204986572,
      "learning_rate": 7.619403915700213e-05,
      "loss": 3.6989,
      "step": 2800
    },
    {
      "epoch": 1.627384960718294,
      "grad_norm": 1.1707065105438232,
      "learning_rate": 7.49470008729268e-05,
      "loss": 3.6858,
      "step": 2900
    },
    {
      "epoch": 1.6835016835016834,
      "grad_norm": 1.8538202047348022,
      "learning_rate": 7.369996258885148e-05,
      "loss": 3.6738,
      "step": 3000
    },
    {
      "epoch": 1.6835016835016834,
      "eval_accuracy": 0.3753888436255347,
      "eval_bleu": 0.1065937514211335,
      "eval_loss": 3.6224589347839355,
      "eval_perplexity": 37.429491437089155,
      "eval_runtime": 12.1708,
      "eval_samples_per_second": 93.749,
      "eval_steps_per_second": 1.479,
      "step": 3000
    },
    {
      "epoch": 1.7396184062850728,
      "grad_norm": 1.2428492307662964,
      "learning_rate": 7.245292430477615e-05,
      "loss": 3.6346,
      "step": 3100
    },
    {
      "epoch": 1.7957351290684624,
      "grad_norm": 1.5533177852630615,
      "learning_rate": 7.120588602070084e-05,
      "loss": 3.6228,
      "step": 3200
    },
    {
      "epoch": 1.8518518518518519,
      "grad_norm": 1.538769006729126,
      "learning_rate": 6.995884773662552e-05,
      "loss": 3.5964,
      "step": 3300
    },
    {
      "epoch": 1.9079685746352413,
      "grad_norm": 0.9837027788162231,
      "learning_rate": 6.871180945255021e-05,
      "loss": 3.5916,
      "step": 3400
    },
    {
      "epoch": 1.964085297418631,
      "grad_norm": 1.3107187747955322,
      "learning_rate": 6.746477116847487e-05,
      "loss": 3.5744,
      "step": 3500
    },
    {
      "epoch": 1.964085297418631,
      "eval_accuracy": 0.38449063305584186,
      "eval_bleu": 0.11184212819864767,
      "eval_loss": 3.5325236320495605,
      "eval_perplexity": 34.21019270752698,
      "eval_runtime": 12.1937,
      "eval_samples_per_second": 93.573,
      "eval_steps_per_second": 1.476,
      "step": 3500
    },
    {
      "epoch": 2.0202020202020203,
      "grad_norm": 1.3989722728729248,
      "learning_rate": 6.621773288439955e-05,
      "loss": 3.5381,
      "step": 3600
    },
    {
      "epoch": 2.0763187429854097,
      "grad_norm": 1.5801029205322266,
      "learning_rate": 6.497069460032424e-05,
      "loss": 3.4865,
      "step": 3700
    },
    {
      "epoch": 2.132435465768799,
      "grad_norm": 1.2122889757156372,
      "learning_rate": 6.372365631624892e-05,
      "loss": 3.4665,
      "step": 3800
    },
    {
      "epoch": 2.1885521885521886,
      "grad_norm": 2.2837603092193604,
      "learning_rate": 6.247661803217359e-05,
      "loss": 3.4675,
      "step": 3900
    },
    {
      "epoch": 2.244668911335578,
      "grad_norm": 1.1615939140319824,
      "learning_rate": 6.122957974809826e-05,
      "loss": 3.456,
      "step": 4000
    },
    {
      "epoch": 2.244668911335578,
      "eval_accuracy": 0.3902143769549271,
      "eval_bleu": 0.11388067737811994,
      "eval_loss": 3.4704020023345947,
      "eval_perplexity": 32.149664087333434,
      "eval_runtime": 12.121,
      "eval_samples_per_second": 94.134,
      "eval_steps_per_second": 1.485,
      "step": 4000
    },
    {
      "epoch": 2.3007856341189674,
      "grad_norm": 0.9502741694450378,
      "learning_rate": 5.998254146402295e-05,
      "loss": 3.4467,
      "step": 4100
    },
    {
      "epoch": 2.356902356902357,
      "grad_norm": 1.7536747455596924,
      "learning_rate": 5.8735503179947625e-05,
      "loss": 3.4298,
      "step": 4200
    },
    {
      "epoch": 2.4130190796857462,
      "grad_norm": 1.4674713611602783,
      "learning_rate": 5.748846489587231e-05,
      "loss": 3.4221,
      "step": 4300
    },
    {
      "epoch": 2.4691358024691357,
      "grad_norm": 2.0394678115844727,
      "learning_rate": 5.624142661179699e-05,
      "loss": 3.4072,
      "step": 4400
    },
    {
      "epoch": 2.525252525252525,
      "grad_norm": 2.8717079162597656,
      "learning_rate": 5.4994388327721666e-05,
      "loss": 3.3972,
      "step": 4500
    },
    {
      "epoch": 2.525252525252525,
      "eval_accuracy": 0.3955106177548291,
      "eval_bleu": 0.12298741482321522,
      "eval_loss": 3.4189839363098145,
      "eval_perplexity": 30.53837032278329,
      "eval_runtime": 12.1708,
      "eval_samples_per_second": 93.749,
      "eval_steps_per_second": 1.479,
      "step": 4500
    },
    {
      "epoch": 2.581369248035915,
      "grad_norm": 2.2189624309539795,
      "learning_rate": 5.374735004364634e-05,
      "loss": 3.4006,
      "step": 4600
    },
    {
      "epoch": 2.637485970819304,
      "grad_norm": 1.444754719734192,
      "learning_rate": 5.250031175957102e-05,
      "loss": 3.3886,
      "step": 4700
    },
    {
      "epoch": 2.6936026936026938,
      "grad_norm": 1.8333204984664917,
      "learning_rate": 5.12532734754957e-05,
      "loss": 3.3813,
      "step": 4800
    },
    {
      "epoch": 2.749719416386083,
      "grad_norm": 2.1033811569213867,
      "learning_rate": 5.000623519142038e-05,
      "loss": 3.372,
      "step": 4900
    },
    {
      "epoch": 2.8058361391694726,
      "grad_norm": 1.8956849575042725,
      "learning_rate": 4.8759196907345056e-05,
      "loss": 3.3654,
      "step": 5000
    },
    {
      "epoch": 2.8058361391694726,
      "eval_accuracy": 0.40071690299277873,
      "eval_bleu": 0.12304297750670024,
      "eval_loss": 3.368644952774048,
      "eval_perplexity": 29.039150964630583,
      "eval_runtime": 12.2072,
      "eval_samples_per_second": 93.469,
      "eval_steps_per_second": 1.475,
      "step": 5000
    },
    {
      "epoch": 2.861952861952862,
      "grad_norm": 2.0555260181427,
      "learning_rate": 4.751215862326974e-05,
      "loss": 3.3622,
      "step": 5100
    },
    {
      "epoch": 2.9180695847362514,
      "grad_norm": 1.1492657661437988,
      "learning_rate": 4.626512033919442e-05,
      "loss": 3.3413,
      "step": 5200
    },
    {
      "epoch": 2.974186307519641,
      "grad_norm": 2.6185925006866455,
      "learning_rate": 4.5018082055119096e-05,
      "loss": 3.3452,
      "step": 5300
    },
    {
      "epoch": 3.0303030303030303,
      "grad_norm": 1.4890856742858887,
      "learning_rate": 4.3771043771043774e-05,
      "loss": 3.2908,
      "step": 5400
    },
    {
      "epoch": 3.0864197530864197,
      "grad_norm": 2.316535711288452,
      "learning_rate": 4.252400548696845e-05,
      "loss": 3.247,
      "step": 5500
    },
    {
      "epoch": 3.0864197530864197,
      "eval_accuracy": 0.40426800589080425,
      "eval_bleu": 0.1247222065482489,
      "eval_loss": 3.3328051567077637,
      "eval_perplexity": 28.016823154790686,
      "eval_runtime": 12.2011,
      "eval_samples_per_second": 93.516,
      "eval_steps_per_second": 1.475,
      "step": 5500
    },
    {
      "epoch": 3.142536475869809,
      "grad_norm": 1.5401073694229126,
      "learning_rate": 4.127696720289313e-05,
      "loss": 3.253,
      "step": 5600
    },
    {
      "epoch": 3.1986531986531985,
      "grad_norm": 1.508957028388977,
      "learning_rate": 4.002992891881781e-05,
      "loss": 3.2477,
      "step": 5700
    },
    {
      "epoch": 3.254769921436588,
      "grad_norm": 1.5511479377746582,
      "learning_rate": 3.8782890634742486e-05,
      "loss": 3.2456,
      "step": 5800
    },
    {
      "epoch": 3.3108866442199774,
      "grad_norm": 1.5875085592269897,
      "learning_rate": 3.7535852350667164e-05,
      "loss": 3.2472,
      "step": 5900
    },
    {
      "epoch": 3.3670033670033668,
      "grad_norm": 1.1592992544174194,
      "learning_rate": 3.628881406659185e-05,
      "loss": 3.2403,
      "step": 6000
    },
    {
      "epoch": 3.3670033670033668,
      "eval_accuracy": 0.40832457337503847,
      "eval_bleu": 0.129799477764246,
      "eval_loss": 3.298476457595825,
      "eval_perplexity": 27.07136311627636,
      "eval_runtime": 12.1774,
      "eval_samples_per_second": 93.699,
      "eval_steps_per_second": 1.478,
      "step": 6000
    },
    {
      "epoch": 3.4231200897867566,
      "grad_norm": 1.2478611469268799,
      "learning_rate": 3.504177578251652e-05,
      "loss": 3.2187,
      "step": 6100
    },
    {
      "epoch": 3.479236812570146,
      "grad_norm": 1.8688626289367676,
      "learning_rate": 3.3794737498441205e-05,
      "loss": 3.2313,
      "step": 6200
    },
    {
      "epoch": 3.5353535353535355,
      "grad_norm": 1.8166719675064087,
      "learning_rate": 3.254769921436588e-05,
      "loss": 3.2192,
      "step": 6300
    },
    {
      "epoch": 3.591470258136925,
      "grad_norm": 1.6677237749099731,
      "learning_rate": 3.130066093029056e-05,
      "loss": 3.221,
      "step": 6400
    },
    {
      "epoch": 3.6475869809203143,
      "grad_norm": 1.4927235841751099,
      "learning_rate": 3.0053622646215242e-05,
      "loss": 3.2167,
      "step": 6500
    },
    {
      "epoch": 3.6475869809203143,
      "eval_accuracy": 0.4111534616185319,
      "eval_bleu": 0.1288262427188711,
      "eval_loss": 3.269272565841675,
      "eval_perplexity": 26.292206536048766,
      "eval_runtime": 12.1618,
      "eval_samples_per_second": 93.819,
      "eval_steps_per_second": 1.48,
      "step": 6500
    },
    {
      "epoch": 3.7037037037037037,
      "grad_norm": 1.9368300437927246,
      "learning_rate": 2.880658436213992e-05,
      "loss": 3.2084,
      "step": 6600
    },
    {
      "epoch": 3.759820426487093,
      "grad_norm": 1.1655679941177368,
      "learning_rate": 2.7559546078064598e-05,
      "loss": 3.1997,
      "step": 6700
    },
    {
      "epoch": 3.8159371492704826,
      "grad_norm": 1.2166575193405151,
      "learning_rate": 2.6312507793989276e-05,
      "loss": 3.1889,
      "step": 6800
    },
    {
      "epoch": 3.872053872053872,
      "grad_norm": 1.2495133876800537,
      "learning_rate": 2.5065469509913957e-05,
      "loss": 3.1814,
      "step": 6900
    },
    {
      "epoch": 3.9281705948372614,
      "grad_norm": 1.4042552709579468,
      "learning_rate": 2.3818431225838632e-05,
      "loss": 3.1903,
      "step": 7000
    },
    {
      "epoch": 3.9281705948372614,
      "eval_accuracy": 0.41343918961176035,
      "eval_bleu": 0.13053470151276936,
      "eval_loss": 3.2455883026123047,
      "eval_perplexity": 25.67681135292233,
      "eval_runtime": 12.1742,
      "eval_samples_per_second": 93.723,
      "eval_steps_per_second": 1.479,
      "step": 7000
    },
    {
      "epoch": 3.984287317620651,
      "grad_norm": 1.472741723060608,
      "learning_rate": 2.2571392941763313e-05,
      "loss": 3.1863,
      "step": 7100
    },
    {
      "epoch": 4.040404040404041,
      "grad_norm": 1.0730637311935425,
      "learning_rate": 2.132435465768799e-05,
      "loss": 3.1197,
      "step": 7200
    },
    {
      "epoch": 4.09652076318743,
      "grad_norm": 1.275009036064148,
      "learning_rate": 2.007731637361267e-05,
      "loss": 3.1142,
      "step": 7300
    },
    {
      "epoch": 4.1526374859708195,
      "grad_norm": 1.156083583831787,
      "learning_rate": 1.883027808953735e-05,
      "loss": 3.1174,
      "step": 7400
    },
    {
      "epoch": 4.2087542087542085,
      "grad_norm": 1.0556427240371704,
      "learning_rate": 1.758323980546203e-05,
      "loss": 3.1212,
      "step": 7500
    },
    {
      "epoch": 4.2087542087542085,
      "eval_accuracy": 0.4161310027132311,
      "eval_bleu": 0.1325477390562919,
      "eval_loss": 3.2261738777160645,
      "eval_perplexity": 25.1831187134555,
      "eval_runtime": 12.1902,
      "eval_samples_per_second": 93.6,
      "eval_steps_per_second": 1.477,
      "step": 7500
    },
    {
      "epoch": 4.264870931537598,
      "grad_norm": 2.7228903770446777,
      "learning_rate": 1.6336201521386706e-05,
      "loss": 3.1073,
      "step": 7600
    },
    {
      "epoch": 4.320987654320987,
      "grad_norm": 1.526665210723877,
      "learning_rate": 1.5089163237311384e-05,
      "loss": 3.0936,
      "step": 7700
    },
    {
      "epoch": 4.377104377104377,
      "grad_norm": 1.0413861274719238,
      "learning_rate": 1.3842124953236066e-05,
      "loss": 3.104,
      "step": 7800
    },
    {
      "epoch": 4.433221099887767,
      "grad_norm": 1.394884467124939,
      "learning_rate": 1.2595086669160744e-05,
      "loss": 3.1018,
      "step": 7900
    },
    {
      "epoch": 4.489337822671156,
      "grad_norm": 1.2792820930480957,
      "learning_rate": 1.1348048385085423e-05,
      "loss": 3.0816,
      "step": 8000
    },
    {
      "epoch": 4.489337822671156,
      "eval_accuracy": 0.4176114142470762,
      "eval_bleu": 0.13073000122523024,
      "eval_loss": 3.2127764225006104,
      "eval_perplexity": 24.847979030639497,
      "eval_runtime": 12.2049,
      "eval_samples_per_second": 93.487,
      "eval_steps_per_second": 1.475,
      "step": 8000
    },
    {
      "epoch": 4.545454545454545,
      "grad_norm": 1.173828363418579,
      "learning_rate": 1.0101010101010101e-05,
      "loss": 3.0993,
      "step": 8100
    },
    {
      "epoch": 4.601571268237935,
      "grad_norm": 1.09369957447052,
      "learning_rate": 8.853971816934781e-06,
      "loss": 3.0905,
      "step": 8200
    },
    {
      "epoch": 4.657687991021325,
      "grad_norm": 1.0657522678375244,
      "learning_rate": 7.606933532859459e-06,
      "loss": 3.097,
      "step": 8300
    },
    {
      "epoch": 4.713804713804714,
      "grad_norm": 1.1146267652511597,
      "learning_rate": 6.359895248784138e-06,
      "loss": 3.0896,
      "step": 8400
    },
    {
      "epoch": 4.7699214365881035,
      "grad_norm": 2.1886322498321533,
      "learning_rate": 5.112856964708817e-06,
      "loss": 3.0917,
      "step": 8500
    },
    {
      "epoch": 4.7699214365881035,
      "eval_accuracy": 0.4195613081423491,
      "eval_bleu": 0.13385511134755296,
      "eval_loss": 3.1984846591949463,
      "eval_perplexity": 24.49538320533492,
      "eval_runtime": 12.1678,
      "eval_samples_per_second": 93.772,
      "eval_steps_per_second": 1.479,
      "step": 8500
    },
    {
      "epoch": 4.8260381593714925,
      "grad_norm": 1.1058974266052246,
      "learning_rate": 3.865818680633495e-06,
      "loss": 3.0822,
      "step": 8600
    },
    {
      "epoch": 4.882154882154882,
      "grad_norm": 1.3942530155181885,
      "learning_rate": 2.6187803965581742e-06,
      "loss": 3.0923,
      "step": 8700
    },
    {
      "epoch": 4.938271604938271,
      "grad_norm": 1.3995310068130493,
      "learning_rate": 1.3717421124828533e-06,
      "loss": 3.088,
      "step": 8800
    },
    {
      "epoch": 4.994388327721661,
      "grad_norm": 1.1968387365341187,
      "learning_rate": 1.2470382840753213e-07,
      "loss": 3.0789,
      "step": 8900
    },
    {
      "epoch": 5.05050505050505,
      "grad_norm": 1.2899984121322632,
      "learning_rate": 1.7584701725213056e-05,
      "loss": 3.0562,
      "step": 9000
    },
    {
      "epoch": 5.05050505050505,
      "eval_accuracy": 0.4185418117735553,
      "eval_bleu": 0.13262877395418665,
      "eval_loss": 3.2048630714416504,
      "eval_perplexity": 24.652124205511342,
      "eval_runtime": 12.0913,
      "eval_samples_per_second": 94.365,
      "eval_steps_per_second": 1.489,
      "step": 9000
    },
    {
      "epoch": 5.10662177328844,
      "grad_norm": 1.1415753364562988,
      "learning_rate": 1.654541675327375e-05,
      "loss": 3.0602,
      "step": 9100
    },
    {
      "epoch": 5.162738496071829,
      "grad_norm": 1.3813053369522095,
      "learning_rate": 1.5506131781334445e-05,
      "loss": 3.0627,
      "step": 9200
    },
    {
      "epoch": 5.218855218855219,
      "grad_norm": 1.9901399612426758,
      "learning_rate": 1.4466846809395137e-05,
      "loss": 3.0647,
      "step": 9300
    },
    {
      "epoch": 5.274971941638609,
      "grad_norm": 1.109480857849121,
      "learning_rate": 1.3427561837455832e-05,
      "loss": 3.0496,
      "step": 9400
    },
    {
      "epoch": 5.331088664421998,
      "grad_norm": 1.1226487159729004,
      "learning_rate": 1.2388276865516525e-05,
      "loss": 3.0683,
      "step": 9500
    },
    {
      "epoch": 5.331088664421998,
      "eval_accuracy": 0.4195261826372058,
      "eval_bleu": 0.1307104242557084,
      "eval_loss": 3.197028875350952,
      "eval_perplexity": 24.459749166234648,
      "eval_runtime": 12.1714,
      "eval_samples_per_second": 93.745,
      "eval_steps_per_second": 1.479,
      "step": 9500
    },
    {
      "epoch": 5.3872053872053876,
      "grad_norm": 1.8179054260253906,
      "learning_rate": 1.134899189357722e-05,
      "loss": 3.0641,
      "step": 9600
    },
    {
      "epoch": 5.4433221099887765,
      "grad_norm": 1.5777114629745483,
      "learning_rate": 1.0309706921637914e-05,
      "loss": 3.0545,
      "step": 9700
    },
    {
      "epoch": 5.499438832772166,
      "grad_norm": 2.161355972290039,
      "learning_rate": 9.270421949698607e-06,
      "loss": 3.057,
      "step": 9800
    },
    {
      "epoch": 5.555555555555555,
      "grad_norm": 2.167677640914917,
      "learning_rate": 8.231136977759301e-06,
      "loss": 3.0532,
      "step": 9900
    },
    {
      "epoch": 5.611672278338945,
      "grad_norm": 0.995482325553894,
      "learning_rate": 7.1918520058199965e-06,
      "loss": 3.0502,
      "step": 10000
    },
    {
      "epoch": 5.611672278338945,
      "eval_accuracy": 0.4209234923661997,
      "eval_bleu": 0.13307821076833265,
      "eval_loss": 3.1857192516326904,
      "eval_perplexity": 24.18467702013368,
      "eval_runtime": 12.2734,
      "eval_samples_per_second": 92.965,
      "eval_steps_per_second": 1.467,
      "step": 10000
    },
    {
      "epoch": 5.667789001122334,
      "grad_norm": 1.5987473726272583,
      "learning_rate": 6.152567033880691e-06,
      "loss": 3.0525,
      "step": 10100
    },
    {
      "epoch": 5.723905723905724,
      "grad_norm": 1.1128613948822021,
      "learning_rate": 5.113282061941385e-06,
      "loss": 3.0491,
      "step": 10200
    },
    {
      "epoch": 5.780022446689113,
      "grad_norm": 1.4610397815704346,
      "learning_rate": 4.073997090002079e-06,
      "loss": 3.0498,
      "step": 10300
    },
    {
      "epoch": 5.836139169472503,
      "grad_norm": 1.4574775695800781,
      "learning_rate": 3.0347121180627732e-06,
      "loss": 3.0475,
      "step": 10400
    },
    {
      "epoch": 5.892255892255893,
      "grad_norm": 1.4116820096969604,
      "learning_rate": 1.9954271461234673e-06,
      "loss": 3.0469,
      "step": 10500
    },
    {
      "epoch": 5.892255892255893,
      "eval_accuracy": 0.42165513093674584,
      "eval_bleu": 0.13088786550672932,
      "eval_loss": 3.1790146827697754,
      "eval_perplexity": 24.023071540521666,
      "eval_runtime": 12.1539,
      "eval_samples_per_second": 93.88,
      "eval_steps_per_second": 1.481,
      "step": 10500
    },
    {
      "epoch": 5.948372615039282,
      "grad_norm": 1.1517071723937988,
      "learning_rate": 9.561421741841614e-07,
      "loss": 3.0538,
      "step": 10600
    },
    {
      "epoch": 6.0,
      "step": 10692,
      "total_flos": 1.270447321217237e+18,
      "train_loss": 0.0,
      "train_runtime": 0.7883,
      "train_samples_per_second": 867659.215,
      "train_steps_per_second": 13563.597
    }
  ],
  "logging_steps": 100,
  "max_steps": 10692,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 6,
  "save_steps": 500,
  "stateful_callbacks": {
    "EarlyStoppingCallback": {
      "args": {
        "early_stopping_patience": 2,
        "early_stopping_threshold": 0.0
      },
      "attributes": {
        "early_stopping_patience_counter": 0
      }
    },
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 1.270447321217237e+18,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}