{
  "best_global_step": 25500,
  "best_metric": 0.21131116151809692,
  "best_model_checkpoint": "reverse_model/checkpoint-25500",
  "epoch": 3.7155762785953663,
  "eval_steps": 500,
  "global_step": 25500,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.07285443683520326,
      "grad_norm": 0.7519411444664001,
      "learning_rate": 0.00019818228180096167,
      "loss": 0.974,
      "step": 500
    },
    {
      "epoch": 0.07285443683520326,
      "eval_loss": 0.6119500398635864,
      "eval_runtime": 0.5563,
      "eval_samples_per_second": 179.775,
      "eval_steps_per_second": 23.371,
      "step": 500
    },
    {
      "epoch": 0.14570887367040652,
      "grad_norm": 0.8617602586746216,
      "learning_rate": 0.0001963609208800816,
      "loss": 0.6234,
      "step": 1000
    },
    {
      "epoch": 0.14570887367040652,
      "eval_loss": 0.5182287096977234,
      "eval_runtime": 0.5766,
      "eval_samples_per_second": 173.435,
      "eval_steps_per_second": 22.546,
      "step": 1000
    },
    {
      "epoch": 0.2185633105056098,
      "grad_norm": 0.4657430648803711,
      "learning_rate": 0.0001945395599592015,
      "loss": 0.5373,
      "step": 1500
    },
    {
      "epoch": 0.2185633105056098,
      "eval_loss": 0.4636226296424866,
      "eval_runtime": 0.5903,
      "eval_samples_per_second": 169.403,
      "eval_steps_per_second": 22.022,
      "step": 1500
    },
    {
      "epoch": 0.29141774734081305,
      "grad_norm": 0.5456737875938416,
      "learning_rate": 0.00019271819903832145,
      "loss": 0.4952,
      "step": 2000
    },
    {
      "epoch": 0.29141774734081305,
      "eval_loss": 0.4364851415157318,
      "eval_runtime": 0.5613,
      "eval_samples_per_second": 178.171,
      "eval_steps_per_second": 23.162,
      "step": 2000
    },
    {
      "epoch": 0.3642721841760163,
      "grad_norm": 0.8705533146858215,
      "learning_rate": 0.00019089683811744136,
      "loss": 0.4634,
      "step": 2500
    },
    {
      "epoch": 0.3642721841760163,
      "eval_loss": 0.4068869650363922,
      "eval_runtime": 0.5819,
      "eval_samples_per_second": 171.863,
      "eval_steps_per_second": 22.342,
      "step": 2500
    },
    {
      "epoch": 0.4371266210112196,
      "grad_norm": 0.45825353264808655,
      "learning_rate": 0.0001890754771965613,
      "loss": 0.4429,
      "step": 3000
    },
    {
      "epoch": 0.4371266210112196,
      "eval_loss": 0.395874947309494,
      "eval_runtime": 0.5825,
      "eval_samples_per_second": 171.678,
      "eval_steps_per_second": 22.318,
      "step": 3000
    },
    {
      "epoch": 0.5099810578464229,
      "grad_norm": 0.5927444100379944,
      "learning_rate": 0.0001872541162756812,
      "loss": 0.4223,
      "step": 3500
    },
    {
      "epoch": 0.5099810578464229,
      "eval_loss": 0.36890122294425964,
      "eval_runtime": 0.5584,
      "eval_samples_per_second": 179.096,
      "eval_steps_per_second": 23.282,
      "step": 3500
    },
    {
      "epoch": 0.5828354946816261,
      "grad_norm": 0.6521668434143066,
      "learning_rate": 0.0001854327553548011,
      "loss": 0.4066,
      "step": 4000
    },
    {
      "epoch": 0.5828354946816261,
      "eval_loss": 0.3580659031867981,
      "eval_runtime": 0.5849,
      "eval_samples_per_second": 170.956,
      "eval_steps_per_second": 22.224,
      "step": 4000
    },
    {
      "epoch": 0.6556899315168294,
      "grad_norm": 0.5500112175941467,
      "learning_rate": 0.00018361139443392105,
      "loss": 0.3909,
      "step": 4500
    },
    {
      "epoch": 0.6556899315168294,
      "eval_loss": 0.3474609851837158,
      "eval_runtime": 0.569,
      "eval_samples_per_second": 175.738,
      "eval_steps_per_second": 22.846,
      "step": 4500
    },
    {
      "epoch": 0.7285443683520326,
      "grad_norm": 0.703709065914154,
      "learning_rate": 0.00018179003351304095,
      "loss": 0.3806,
      "step": 5000
    },
    {
      "epoch": 0.7285443683520326,
      "eval_loss": 0.33880433440208435,
      "eval_runtime": 0.5577,
      "eval_samples_per_second": 179.313,
      "eval_steps_per_second": 23.311,
      "step": 5000
    },
    {
      "epoch": 0.8013988051872359,
      "grad_norm": 0.568647027015686,
      "learning_rate": 0.0001799686725921609,
      "loss": 0.3731,
      "step": 5500
    },
    {
      "epoch": 0.8013988051872359,
      "eval_loss": 0.33678069710731506,
      "eval_runtime": 0.5788,
      "eval_samples_per_second": 172.771,
      "eval_steps_per_second": 22.46,
      "step": 5500
    },
    {
      "epoch": 0.8742532420224391,
      "grad_norm": 0.6542627811431885,
      "learning_rate": 0.00017814731167128077,
      "loss": 0.3597,
      "step": 6000
    },
    {
      "epoch": 0.8742532420224391,
      "eval_loss": 0.3265901207923889,
      "eval_runtime": 0.5582,
      "eval_samples_per_second": 179.158,
      "eval_steps_per_second": 23.291,
      "step": 6000
    },
    {
      "epoch": 0.9471076788576425,
      "grad_norm": 0.6129189729690552,
      "learning_rate": 0.0001763259507504007,
      "loss": 0.3555,
      "step": 6500
    },
    {
      "epoch": 0.9471076788576425,
      "eval_loss": 0.3173937499523163,
      "eval_runtime": 0.5933,
      "eval_samples_per_second": 168.543,
      "eval_steps_per_second": 21.911,
      "step": 6500
    },
    {
      "epoch": 1.0199621156928458,
      "grad_norm": 0.46865177154541016,
      "learning_rate": 0.00017450458982952062,
      "loss": 0.3471,
      "step": 7000
    },
    {
      "epoch": 1.0199621156928458,
      "eval_loss": 0.3176809549331665,
      "eval_runtime": 0.5602,
      "eval_samples_per_second": 178.494,
      "eval_steps_per_second": 23.204,
      "step": 7000
    },
    {
      "epoch": 1.0928165525280489,
      "grad_norm": 0.5402314066886902,
      "learning_rate": 0.00017268322890864055,
      "loss": 0.3377,
      "step": 7500
    },
    {
      "epoch": 1.0928165525280489,
      "eval_loss": 0.30362746119499207,
      "eval_runtime": 0.5623,
      "eval_samples_per_second": 177.838,
      "eval_steps_per_second": 23.119,
      "step": 7500
    },
    {
      "epoch": 1.1656709893632522,
      "grad_norm": 0.4397026300430298,
      "learning_rate": 0.00017086186798776046,
      "loss": 0.3327,
      "step": 8000
    },
    {
      "epoch": 1.1656709893632522,
      "eval_loss": 0.2984870970249176,
      "eval_runtime": 0.5634,
      "eval_samples_per_second": 177.483,
      "eval_steps_per_second": 23.073,
      "step": 8000
    },
    {
      "epoch": 1.2385254261984555,
      "grad_norm": 0.4877306818962097,
      "learning_rate": 0.00016904050706688037,
      "loss": 0.3257,
      "step": 8500
    },
    {
      "epoch": 1.2385254261984555,
      "eval_loss": 0.29171615839004517,
      "eval_runtime": 0.579,
      "eval_samples_per_second": 172.707,
      "eval_steps_per_second": 22.452,
      "step": 8500
    },
    {
      "epoch": 1.3113798630336588,
      "grad_norm": 1.0982270240783691,
      "learning_rate": 0.0001672191461460003,
      "loss": 0.32,
      "step": 9000
    },
    {
      "epoch": 1.3113798630336588,
      "eval_loss": 0.2923184633255005,
      "eval_runtime": 0.558,
      "eval_samples_per_second": 179.22,
      "eval_steps_per_second": 23.299,
      "step": 9000
    },
    {
      "epoch": 1.384234299868862,
      "grad_norm": 0.6584481000900269,
      "learning_rate": 0.0001653977852251202,
      "loss": 0.3129,
      "step": 9500
    },
    {
      "epoch": 1.384234299868862,
      "eval_loss": 0.2869529128074646,
      "eval_runtime": 0.5651,
      "eval_samples_per_second": 176.96,
      "eval_steps_per_second": 23.005,
      "step": 9500
    },
    {
      "epoch": 1.4570887367040652,
      "grad_norm": 0.5571127533912659,
      "learning_rate": 0.00016357642430424015,
      "loss": 0.3178,
      "step": 10000
    },
    {
      "epoch": 1.4570887367040652,
      "eval_loss": 0.2796230614185333,
      "eval_runtime": 0.5645,
      "eval_samples_per_second": 177.159,
      "eval_steps_per_second": 23.031,
      "step": 10000
    },
    {
      "epoch": 1.5299431735392686,
      "grad_norm": 0.33073556423187256,
      "learning_rate": 0.00016175506338336006,
      "loss": 0.3054,
      "step": 10500
    },
    {
      "epoch": 1.5299431735392686,
      "eval_loss": 0.27576252818107605,
      "eval_runtime": 0.5776,
      "eval_samples_per_second": 173.121,
      "eval_steps_per_second": 22.506,
      "step": 10500
    },
    {
      "epoch": 1.6027976103744717,
      "grad_norm": 0.45740246772766113,
      "learning_rate": 0.00015993370246247996,
      "loss": 0.307,
      "step": 11000
    },
    {
      "epoch": 1.6027976103744717,
      "eval_loss": 0.27131548523902893,
      "eval_runtime": 0.5666,
      "eval_samples_per_second": 176.496,
      "eval_steps_per_second": 22.945,
      "step": 11000
    },
    {
      "epoch": 1.6756520472096752,
      "grad_norm": 0.45748448371887207,
      "learning_rate": 0.00015811234154159987,
      "loss": 0.3015,
      "step": 11500
    },
    {
      "epoch": 1.6756520472096752,
      "eval_loss": 0.26808932423591614,
      "eval_runtime": 0.5676,
      "eval_samples_per_second": 176.195,
      "eval_steps_per_second": 22.905,
      "step": 11500
    },
    {
      "epoch": 1.7485064840448783,
      "grad_norm": 0.4469503164291382,
      "learning_rate": 0.0001562909806207198,
      "loss": 0.301,
      "step": 12000
    },
    {
      "epoch": 1.7485064840448783,
      "eval_loss": 0.2650498151779175,
      "eval_runtime": 0.5572,
      "eval_samples_per_second": 179.466,
      "eval_steps_per_second": 23.331,
      "step": 12000
    },
    {
      "epoch": 1.8213609208800816,
      "grad_norm": 0.6140857338905334,
      "learning_rate": 0.00015446961969983972,
      "loss": 0.2948,
      "step": 12500
    },
    {
      "epoch": 1.8213609208800816,
      "eval_loss": 0.2665635645389557,
      "eval_runtime": 0.5602,
      "eval_samples_per_second": 178.493,
      "eval_steps_per_second": 23.204,
      "step": 12500
    },
    {
      "epoch": 1.894215357715285,
      "grad_norm": 0.4431038200855255,
      "learning_rate": 0.00015264825877895965,
      "loss": 0.2929,
      "step": 13000
    },
    {
      "epoch": 1.894215357715285,
      "eval_loss": 0.26208823919296265,
      "eval_runtime": 0.562,
      "eval_samples_per_second": 177.925,
      "eval_steps_per_second": 23.13,
      "step": 13000
    },
    {
      "epoch": 1.967069794550488,
      "grad_norm": 0.5034199953079224,
      "learning_rate": 0.0001508268978580796,
      "loss": 0.2844,
      "step": 13500
    },
    {
      "epoch": 1.967069794550488,
      "eval_loss": 0.2574382722377777,
      "eval_runtime": 0.5819,
      "eval_samples_per_second": 171.837,
      "eval_steps_per_second": 22.339,
      "step": 13500
    },
    {
      "epoch": 2.0399242313856916,
      "grad_norm": 0.5565065741539001,
      "learning_rate": 0.00014900553693719947,
      "loss": 0.2807,
      "step": 14000
    },
    {
      "epoch": 2.0399242313856916,
      "eval_loss": 0.25828686356544495,
      "eval_runtime": 0.5827,
      "eval_samples_per_second": 171.603,
      "eval_steps_per_second": 22.308,
      "step": 14000
    },
    {
      "epoch": 2.1127786682208947,
      "grad_norm": 0.5862753987312317,
      "learning_rate": 0.0001471841760163194,
      "loss": 0.2806,
      "step": 14500
    },
    {
      "epoch": 2.1127786682208947,
      "eval_loss": 0.2531309425830841,
      "eval_runtime": 0.5905,
      "eval_samples_per_second": 169.362,
      "eval_steps_per_second": 22.017,
      "step": 14500
    },
    {
      "epoch": 2.1856331050560978,
      "grad_norm": 0.7702651619911194,
      "learning_rate": 0.0001453628150954393,
      "loss": 0.2776,
      "step": 15000
    },
    {
      "epoch": 2.1856331050560978,
      "eval_loss": 0.24827995896339417,
      "eval_runtime": 0.5626,
      "eval_samples_per_second": 177.742,
      "eval_steps_per_second": 23.106,
      "step": 15000
    },
    {
      "epoch": 2.2584875418913013,
      "grad_norm": 0.4496975541114807,
      "learning_rate": 0.00014354145417455925,
      "loss": 0.2729,
      "step": 15500
    },
    {
      "epoch": 2.2584875418913013,
      "eval_loss": 0.24621199071407318,
      "eval_runtime": 0.5772,
      "eval_samples_per_second": 173.237,
      "eval_steps_per_second": 22.521,
      "step": 15500
    },
    {
      "epoch": 2.3313419787265044,
      "grad_norm": 0.5896193981170654,
      "learning_rate": 0.00014172009325367916,
      "loss": 0.2718,
      "step": 16000
    },
    {
      "epoch": 2.3313419787265044,
      "eval_loss": 0.24523521959781647,
      "eval_runtime": 0.5819,
      "eval_samples_per_second": 171.859,
      "eval_steps_per_second": 22.342,
      "step": 16000
    },
    {
      "epoch": 2.4041964155617075,
      "grad_norm": 0.411600798368454,
      "learning_rate": 0.00013989873233279907,
      "loss": 0.2718,
      "step": 16500
    },
    {
      "epoch": 2.4041964155617075,
      "eval_loss": 0.24163128435611725,
      "eval_runtime": 0.5597,
      "eval_samples_per_second": 178.657,
      "eval_steps_per_second": 23.225,
      "step": 16500
    },
    {
      "epoch": 2.477050852396911,
      "grad_norm": 0.5009840130805969,
      "learning_rate": 0.00013807737141191897,
      "loss": 0.2652,
      "step": 17000
    },
    {
      "epoch": 2.477050852396911,
      "eval_loss": 0.24147550761699677,
      "eval_runtime": 0.5809,
      "eval_samples_per_second": 172.154,
      "eval_steps_per_second": 22.38,
      "step": 17000
    },
    {
      "epoch": 2.549905289232114,
      "grad_norm": 0.5353007912635803,
      "learning_rate": 0.0001362560104910389,
      "loss": 0.2628,
      "step": 17500
    },
    {
      "epoch": 2.549905289232114,
      "eval_loss": 0.2369464635848999,
      "eval_runtime": 0.5912,
      "eval_samples_per_second": 169.158,
      "eval_steps_per_second": 21.991,
      "step": 17500
    },
    {
      "epoch": 2.6227597260673177,
      "grad_norm": 0.4573606848716736,
      "learning_rate": 0.00013443464957015885,
      "loss": 0.2609,
      "step": 18000
    },
    {
      "epoch": 2.6227597260673177,
      "eval_loss": 0.23909151554107666,
      "eval_runtime": 0.5596,
      "eval_samples_per_second": 178.696,
      "eval_steps_per_second": 23.23,
      "step": 18000
    },
    {
      "epoch": 2.6956141629025208,
      "grad_norm": 0.4674642086029053,
      "learning_rate": 0.00013261328864927875,
      "loss": 0.257,
      "step": 18500
    },
    {
      "epoch": 2.6956141629025208,
      "eval_loss": 0.2337809056043625,
      "eval_runtime": 0.5605,
      "eval_samples_per_second": 178.402,
      "eval_steps_per_second": 23.192,
      "step": 18500
    },
    {
      "epoch": 2.768468599737724,
      "grad_norm": 0.43507474660873413,
      "learning_rate": 0.00013079192772839866,
      "loss": 0.2566,
      "step": 19000
    },
    {
      "epoch": 2.768468599737724,
      "eval_loss": 0.2308950424194336,
      "eval_runtime": 0.561,
      "eval_samples_per_second": 178.259,
      "eval_steps_per_second": 23.174,
      "step": 19000
    },
    {
      "epoch": 2.8413230365729274,
      "grad_norm": 0.4606495797634125,
      "learning_rate": 0.00012897056680751857,
      "loss": 0.2607,
      "step": 19500
    },
    {
      "epoch": 2.8413230365729274,
      "eval_loss": 0.23102878034114838,
      "eval_runtime": 0.5745,
      "eval_samples_per_second": 174.057,
      "eval_steps_per_second": 22.627,
      "step": 19500
    },
    {
      "epoch": 2.9141774734081305,
      "grad_norm": 0.686039388179779,
      "learning_rate": 0.0001271492058866385,
      "loss": 0.2562,
      "step": 20000
    },
    {
      "epoch": 2.9141774734081305,
      "eval_loss": 0.22536581754684448,
      "eval_runtime": 0.5619,
      "eval_samples_per_second": 177.974,
      "eval_steps_per_second": 23.137,
      "step": 20000
    },
    {
      "epoch": 2.987031910243334,
      "grad_norm": 0.48106732964515686,
      "learning_rate": 0.00012532784496575841,
      "loss": 0.2524,
      "step": 20500
    },
    {
      "epoch": 2.987031910243334,
      "eval_loss": 0.22328069806098938,
      "eval_runtime": 0.562,
      "eval_samples_per_second": 177.946,
      "eval_steps_per_second": 23.133,
      "step": 20500
    },
    {
      "epoch": 3.059886347078537,
      "grad_norm": 0.44265252351760864,
      "learning_rate": 0.00012350648404487835,
      "loss": 0.2469,
      "step": 21000
    },
    {
      "epoch": 3.059886347078537,
      "eval_loss": 0.22362683713436127,
      "eval_runtime": 0.5835,
      "eval_samples_per_second": 171.391,
      "eval_steps_per_second": 22.281,
      "step": 21000
    },
    {
      "epoch": 3.1327407839137402,
      "grad_norm": 0.4698319435119629,
      "learning_rate": 0.00012168512312399827,
      "loss": 0.2521,
      "step": 21500
    },
    {
      "epoch": 3.1327407839137402,
      "eval_loss": 0.22086407244205475,
      "eval_runtime": 0.5707,
      "eval_samples_per_second": 175.226,
      "eval_steps_per_second": 22.779,
      "step": 21500
    },
    {
      "epoch": 3.2055952207489438,
      "grad_norm": 0.4953310191631317,
      "learning_rate": 0.00011986376220311817,
      "loss": 0.2478,
      "step": 22000
    },
    {
      "epoch": 3.2055952207489438,
      "eval_loss": 0.22542959451675415,
      "eval_runtime": 0.5593,
      "eval_samples_per_second": 178.801,
      "eval_steps_per_second": 23.244,
      "step": 22000
    },
    {
      "epoch": 3.278449657584147,
      "grad_norm": 0.921518087387085,
      "learning_rate": 0.00011804240128223809,
      "loss": 0.2447,
      "step": 22500
    },
    {
      "epoch": 3.278449657584147,
      "eval_loss": 0.22003282606601715,
      "eval_runtime": 0.5616,
      "eval_samples_per_second": 178.07,
      "eval_steps_per_second": 23.149,
      "step": 22500
    },
    {
      "epoch": 3.35130409441935,
      "grad_norm": 0.5857972502708435,
      "learning_rate": 0.00011622104036135801,
      "loss": 0.2432,
      "step": 23000
    },
    {
      "epoch": 3.35130409441935,
      "eval_loss": 0.21950890123844147,
      "eval_runtime": 0.5587,
      "eval_samples_per_second": 178.975,
      "eval_steps_per_second": 23.267,
      "step": 23000
    },
    {
      "epoch": 3.4241585312545535,
      "grad_norm": 0.5046322345733643,
      "learning_rate": 0.00011439967944047793,
      "loss": 0.2394,
      "step": 23500
    },
    {
      "epoch": 3.4241585312545535,
      "eval_loss": 0.21759502589702606,
      "eval_runtime": 0.5757,
      "eval_samples_per_second": 173.711,
      "eval_steps_per_second": 22.582,
      "step": 23500
    },
    {
      "epoch": 3.4970129680897566,
      "grad_norm": 0.6441205739974976,
      "learning_rate": 0.00011257831851959786,
      "loss": 0.2371,
      "step": 24000
    },
    {
      "epoch": 3.4970129680897566,
      "eval_loss": 0.21436944603919983,
      "eval_runtime": 0.5779,
      "eval_samples_per_second": 173.027,
      "eval_steps_per_second": 22.494,
      "step": 24000
    },
    {
      "epoch": 3.5698674049249597,
      "grad_norm": 0.5626524090766907,
      "learning_rate": 0.00011075695759871776,
      "loss": 0.2391,
      "step": 24500
    },
    {
      "epoch": 3.5698674049249597,
      "eval_loss": 0.21467125415802002,
      "eval_runtime": 0.5829,
      "eval_samples_per_second": 171.542,
      "eval_steps_per_second": 22.3,
      "step": 24500
    },
    {
      "epoch": 3.6427218417601632,
      "grad_norm": 0.6218989491462708,
      "learning_rate": 0.00010893559667783769,
      "loss": 0.235,
      "step": 25000
    },
    {
      "epoch": 3.6427218417601632,
      "eval_loss": 0.21365521848201752,
      "eval_runtime": 0.5617,
      "eval_samples_per_second": 178.033,
      "eval_steps_per_second": 23.144,
      "step": 25000
    },
    {
      "epoch": 3.7155762785953663,
      "grad_norm": 0.39173460006713867,
      "learning_rate": 0.00010711423575695761,
      "loss": 0.2364,
      "step": 25500
    },
    {
      "epoch": 3.7155762785953663,
      "eval_loss": 0.21131116151809692,
      "eval_runtime": 0.5528,
      "eval_samples_per_second": 180.905,
      "eval_steps_per_second": 23.518,
      "step": 25500
    }
  ],
  "logging_steps": 500,
  "max_steps": 54904,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 8,
  "save_steps": 500,
  "stateful_callbacks": {
    "EarlyStoppingCallback": {
      "args": {
        "early_stopping_patience": 3,
        "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": false
      },
      "attributes": {}
    }
  },
  "total_flos": 6902431875072000.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}