{
"best_metric": 2.435030460357666,
"best_model_checkpoint": "zhtw-en/checkpoint-92500",
"epoch": 3.0,
"eval_steps": 2500,
"global_step": 93276,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.016081307088640164,
"grad_norm": 8.550692558288574,
"learning_rate": 4.919593464556799e-05,
"loss": 3.4707,
"num_input_tokens_seen": 301408,
"step": 500
},
{
"epoch": 0.03216261417728033,
"grad_norm": 9.999076843261719,
"learning_rate": 4.8391869291135987e-05,
"loss": 3.3664,
"num_input_tokens_seen": 597672,
"step": 1000
},
{
"epoch": 0.048243921265920496,
"grad_norm": 8.839406967163086,
"learning_rate": 4.7587803936703975e-05,
"loss": 3.3089,
"num_input_tokens_seen": 896520,
"step": 1500
},
{
"epoch": 0.06432522835456066,
"grad_norm": 7.105090618133545,
"learning_rate": 4.678373858227197e-05,
"loss": 3.267,
"num_input_tokens_seen": 1194832,
"step": 2000
},
{
"epoch": 0.08040653544320082,
"grad_norm": 7.992733001708984,
"learning_rate": 4.597967322783996e-05,
"loss": 3.2254,
"num_input_tokens_seen": 1493088,
"step": 2500
},
{
"epoch": 0.08040653544320082,
"eval_loss": 2.910461664199829,
"eval_runtime": 2.5005,
"eval_samples_per_second": 999.788,
"eval_steps_per_second": 125.173,
"num_input_tokens_seen": 1493088,
"step": 2500
},
{
"epoch": 0.09648784253184099,
"grad_norm": 8.68657112121582,
"learning_rate": 4.5175607873407955e-05,
"loss": 3.1691,
"num_input_tokens_seen": 1793976,
"step": 3000
},
{
"epoch": 0.11256914962048115,
"grad_norm": 7.246800899505615,
"learning_rate": 4.4371542518975943e-05,
"loss": 3.1685,
"num_input_tokens_seen": 2095352,
"step": 3500
},
{
"epoch": 0.1286504567091213,
"grad_norm": 9.030860900878906,
"learning_rate": 4.356747716454393e-05,
"loss": 3.1333,
"num_input_tokens_seen": 2393856,
"step": 4000
},
{
"epoch": 0.14473176379776148,
"grad_norm": 7.463845252990723,
"learning_rate": 4.276341181011193e-05,
"loss": 3.1295,
"num_input_tokens_seen": 2694496,
"step": 4500
},
{
"epoch": 0.16081307088640165,
"grad_norm": 8.482089042663574,
"learning_rate": 4.195934645567992e-05,
"loss": 3.0946,
"num_input_tokens_seen": 2990968,
"step": 5000
},
{
"epoch": 0.16081307088640165,
"eval_loss": 2.830476999282837,
"eval_runtime": 2.5201,
"eval_samples_per_second": 992.009,
"eval_steps_per_second": 124.199,
"num_input_tokens_seen": 2990968,
"step": 5000
},
{
"epoch": 0.17689437797504182,
"grad_norm": 8.570518493652344,
"learning_rate": 4.115528110124791e-05,
"loss": 3.112,
"num_input_tokens_seen": 3289488,
"step": 5500
},
{
"epoch": 0.19297568506368198,
"grad_norm": 9.759325981140137,
"learning_rate": 4.03512157468159e-05,
"loss": 3.0933,
"num_input_tokens_seen": 3590264,
"step": 6000
},
{
"epoch": 0.20905699215232215,
"grad_norm": 6.518988609313965,
"learning_rate": 3.9547150392383896e-05,
"loss": 3.0858,
"num_input_tokens_seen": 3885160,
"step": 6500
},
{
"epoch": 0.2251382992409623,
"grad_norm": 6.913475036621094,
"learning_rate": 3.8743085037951885e-05,
"loss": 3.0543,
"num_input_tokens_seen": 4184600,
"step": 7000
},
{
"epoch": 0.24121960632960246,
"grad_norm": 8.485562324523926,
"learning_rate": 3.793901968351988e-05,
"loss": 3.0473,
"num_input_tokens_seen": 4477792,
"step": 7500
},
{
"epoch": 0.24121960632960246,
"eval_loss": 2.773728847503662,
"eval_runtime": 2.5738,
"eval_samples_per_second": 971.323,
"eval_steps_per_second": 121.61,
"num_input_tokens_seen": 4477792,
"step": 7500
},
{
"epoch": 0.2573009134182426,
"grad_norm": 7.89262056350708,
"learning_rate": 3.713495432908787e-05,
"loss": 3.029,
"num_input_tokens_seen": 4779520,
"step": 8000
},
{
"epoch": 0.2733822205068828,
"grad_norm": 6.879751205444336,
"learning_rate": 3.6330888974655864e-05,
"loss": 3.0127,
"num_input_tokens_seen": 5078952,
"step": 8500
},
{
"epoch": 0.28946352759552296,
"grad_norm": 8.109273910522461,
"learning_rate": 3.552682362022385e-05,
"loss": 3.0078,
"num_input_tokens_seen": 5376128,
"step": 9000
},
{
"epoch": 0.3055448346841631,
"grad_norm": 8.074146270751953,
"learning_rate": 3.472275826579184e-05,
"loss": 2.9988,
"num_input_tokens_seen": 5671664,
"step": 9500
},
{
"epoch": 0.3216261417728033,
"grad_norm": 6.523529529571533,
"learning_rate": 3.391869291135984e-05,
"loss": 2.9633,
"num_input_tokens_seen": 5967560,
"step": 10000
},
{
"epoch": 0.3216261417728033,
"eval_loss": 2.7306864261627197,
"eval_runtime": 2.505,
"eval_samples_per_second": 997.986,
"eval_steps_per_second": 124.948,
"num_input_tokens_seen": 5967560,
"step": 10000
},
{
"epoch": 0.33770744886144344,
"grad_norm": 6.974269866943359,
"learning_rate": 3.311462755692783e-05,
"loss": 2.9732,
"num_input_tokens_seen": 6265312,
"step": 10500
},
{
"epoch": 0.35378875595008363,
"grad_norm": 7.644798278808594,
"learning_rate": 3.231056220249582e-05,
"loss": 2.9729,
"num_input_tokens_seen": 6563632,
"step": 11000
},
{
"epoch": 0.3698700630387238,
"grad_norm": 7.96437406539917,
"learning_rate": 3.150649684806381e-05,
"loss": 2.9484,
"num_input_tokens_seen": 6865528,
"step": 11500
},
{
"epoch": 0.38595137012736397,
"grad_norm": 7.939519882202148,
"learning_rate": 3.0702431493631805e-05,
"loss": 2.9387,
"num_input_tokens_seen": 7165632,
"step": 12000
},
{
"epoch": 0.4020326772160041,
"grad_norm": 7.698306083679199,
"learning_rate": 2.9898366139199797e-05,
"loss": 2.9355,
"num_input_tokens_seen": 7463192,
"step": 12500
},
{
"epoch": 0.4020326772160041,
"eval_loss": 2.684298515319824,
"eval_runtime": 2.5214,
"eval_samples_per_second": 991.52,
"eval_steps_per_second": 124.138,
"num_input_tokens_seen": 7463192,
"step": 12500
},
{
"epoch": 0.4181139843046443,
"grad_norm": 7.21583890914917,
"learning_rate": 2.9094300784767786e-05,
"loss": 2.9418,
"num_input_tokens_seen": 7758024,
"step": 13000
},
{
"epoch": 0.43419529139328444,
"grad_norm": 7.767180919647217,
"learning_rate": 2.8290235430335778e-05,
"loss": 2.923,
"num_input_tokens_seen": 8052032,
"step": 13500
},
{
"epoch": 0.4502765984819246,
"grad_norm": 7.057159423828125,
"learning_rate": 2.7486170075903773e-05,
"loss": 2.9016,
"num_input_tokens_seen": 8347768,
"step": 14000
},
{
"epoch": 0.4663579055705648,
"grad_norm": 7.320003032684326,
"learning_rate": 2.6682104721471762e-05,
"loss": 2.885,
"num_input_tokens_seen": 8646192,
"step": 14500
},
{
"epoch": 0.4824392126592049,
"grad_norm": 7.630561828613281,
"learning_rate": 2.587803936703975e-05,
"loss": 2.9076,
"num_input_tokens_seen": 8950264,
"step": 15000
},
{
"epoch": 0.4824392126592049,
"eval_loss": 2.658709764480591,
"eval_runtime": 2.6566,
"eval_samples_per_second": 941.035,
"eval_steps_per_second": 117.818,
"num_input_tokens_seen": 8950264,
"step": 15000
},
{
"epoch": 0.4985205197478451,
"grad_norm": 6.413602828979492,
"learning_rate": 2.507397401260775e-05,
"loss": 2.8818,
"num_input_tokens_seen": 9247272,
"step": 15500
},
{
"epoch": 0.5146018268364853,
"grad_norm": 6.833747863769531,
"learning_rate": 2.426990865817574e-05,
"loss": 2.8998,
"num_input_tokens_seen": 9547432,
"step": 16000
},
{
"epoch": 0.5306831339251254,
"grad_norm": 7.930506706237793,
"learning_rate": 2.3465843303743727e-05,
"loss": 2.8543,
"num_input_tokens_seen": 9844072,
"step": 16500
},
{
"epoch": 0.5467644410137656,
"grad_norm": 6.841344356536865,
"learning_rate": 2.2661777949311722e-05,
"loss": 2.8669,
"num_input_tokens_seen": 10142344,
"step": 17000
},
{
"epoch": 0.5628457481024057,
"grad_norm": 6.899343967437744,
"learning_rate": 2.185771259487971e-05,
"loss": 2.8714,
"num_input_tokens_seen": 10443344,
"step": 17500
},
{
"epoch": 0.5628457481024057,
"eval_loss": 2.6303601264953613,
"eval_runtime": 2.5354,
"eval_samples_per_second": 986.046,
"eval_steps_per_second": 123.453,
"num_input_tokens_seen": 10443344,
"step": 17500
},
{
"epoch": 0.5789270551910459,
"grad_norm": 6.256689071655273,
"learning_rate": 2.1053647240447703e-05,
"loss": 2.8418,
"num_input_tokens_seen": 10744312,
"step": 18000
},
{
"epoch": 0.5950083622796861,
"grad_norm": 7.627821445465088,
"learning_rate": 2.0249581886015695e-05,
"loss": 2.8462,
"num_input_tokens_seen": 11048208,
"step": 18500
},
{
"epoch": 0.6110896693683262,
"grad_norm": 7.331953525543213,
"learning_rate": 1.9445516531583687e-05,
"loss": 2.8345,
"num_input_tokens_seen": 11347880,
"step": 19000
},
{
"epoch": 0.6271709764569664,
"grad_norm": 6.463207244873047,
"learning_rate": 1.864145117715168e-05,
"loss": 2.8531,
"num_input_tokens_seen": 11650144,
"step": 19500
},
{
"epoch": 0.6432522835456066,
"grad_norm": 7.423746109008789,
"learning_rate": 1.783738582271967e-05,
"loss": 2.8716,
"num_input_tokens_seen": 11951096,
"step": 20000
},
{
"epoch": 0.6432522835456066,
"eval_loss": 2.6024744510650635,
"eval_runtime": 2.5182,
"eval_samples_per_second": 992.755,
"eval_steps_per_second": 124.293,
"num_input_tokens_seen": 11951096,
"step": 20000
},
{
"epoch": 0.6593335906342468,
"grad_norm": 7.352589130401611,
"learning_rate": 1.7033320468287664e-05,
"loss": 2.8243,
"num_input_tokens_seen": 12252592,
"step": 20500
},
{
"epoch": 0.6754148977228869,
"grad_norm": 7.22981071472168,
"learning_rate": 1.6229255113855656e-05,
"loss": 2.8454,
"num_input_tokens_seen": 12546792,
"step": 21000
},
{
"epoch": 0.6914962048115271,
"grad_norm": 6.819567680358887,
"learning_rate": 1.5425189759423648e-05,
"loss": 2.8047,
"num_input_tokens_seen": 12838728,
"step": 21500
},
{
"epoch": 0.7075775119001673,
"grad_norm": 8.716426849365234,
"learning_rate": 1.4621124404991638e-05,
"loss": 2.8144,
"num_input_tokens_seen": 13137688,
"step": 22000
},
{
"epoch": 0.7236588189888075,
"grad_norm": 7.324875831604004,
"learning_rate": 1.381705905055963e-05,
"loss": 2.7989,
"num_input_tokens_seen": 13432464,
"step": 22500
},
{
"epoch": 0.7236588189888075,
"eval_loss": 2.5822224617004395,
"eval_runtime": 2.5158,
"eval_samples_per_second": 993.705,
"eval_steps_per_second": 124.412,
"num_input_tokens_seen": 13432464,
"step": 22500
},
{
"epoch": 0.7397401260774475,
"grad_norm": 7.962778568267822,
"learning_rate": 1.301299369612762e-05,
"loss": 2.7653,
"num_input_tokens_seen": 13730024,
"step": 23000
},
{
"epoch": 0.7558214331660877,
"grad_norm": 6.807019233703613,
"learning_rate": 1.2208928341695614e-05,
"loss": 2.7933,
"num_input_tokens_seen": 14026400,
"step": 23500
},
{
"epoch": 0.7719027402547279,
"grad_norm": 8.716556549072266,
"learning_rate": 1.1404862987263605e-05,
"loss": 2.7988,
"num_input_tokens_seen": 14326608,
"step": 24000
},
{
"epoch": 0.787984047343368,
"grad_norm": 7.388988018035889,
"learning_rate": 1.0600797632831597e-05,
"loss": 2.7928,
"num_input_tokens_seen": 14623864,
"step": 24500
},
{
"epoch": 0.8040653544320082,
"grad_norm": 7.011099815368652,
"learning_rate": 9.796732278399589e-06,
"loss": 2.7941,
"num_input_tokens_seen": 14919424,
"step": 25000
},
{
"epoch": 0.8040653544320082,
"eval_loss": 2.5630149841308594,
"eval_runtime": 2.5601,
"eval_samples_per_second": 976.534,
"eval_steps_per_second": 122.262,
"num_input_tokens_seen": 14919424,
"step": 25000
},
{
"epoch": 0.8201466615206484,
"grad_norm": 6.740393161773682,
"learning_rate": 8.992666923967581e-06,
"loss": 2.8089,
"num_input_tokens_seen": 15216136,
"step": 25500
},
{
"epoch": 0.8362279686092886,
"grad_norm": 7.124479293823242,
"learning_rate": 8.188601569535573e-06,
"loss": 2.7704,
"num_input_tokens_seen": 15515592,
"step": 26000
},
{
"epoch": 0.8523092756979287,
"grad_norm": 7.781102180480957,
"learning_rate": 7.384536215103564e-06,
"loss": 2.8022,
"num_input_tokens_seen": 15818560,
"step": 26500
},
{
"epoch": 0.8683905827865689,
"grad_norm": 6.861135005950928,
"learning_rate": 6.580470860671556e-06,
"loss": 2.7891,
"num_input_tokens_seen": 16114056,
"step": 27000
},
{
"epoch": 0.8844718898752091,
"grad_norm": 7.128973484039307,
"learning_rate": 5.776405506239547e-06,
"loss": 2.7692,
"num_input_tokens_seen": 16415080,
"step": 27500
},
{
"epoch": 0.8844718898752091,
"eval_loss": 2.5496785640716553,
"eval_runtime": 2.6422,
"eval_samples_per_second": 946.175,
"eval_steps_per_second": 118.461,
"num_input_tokens_seen": 16415080,
"step": 27500
},
{
"epoch": 0.9005531969638492,
"grad_norm": 8.560084342956543,
"learning_rate": 4.9723401518075395e-06,
"loss": 2.7627,
"num_input_tokens_seen": 16711136,
"step": 28000
},
{
"epoch": 0.9166345040524894,
"grad_norm": 7.5000224113464355,
"learning_rate": 4.168274797375531e-06,
"loss": 2.7687,
"num_input_tokens_seen": 17005880,
"step": 28500
},
{
"epoch": 0.9327158111411296,
"grad_norm": 6.699025630950928,
"learning_rate": 3.3642094429435228e-06,
"loss": 2.779,
"num_input_tokens_seen": 17307064,
"step": 29000
},
{
"epoch": 0.9487971182297698,
"grad_norm": 6.6417131423950195,
"learning_rate": 2.560144088511514e-06,
"loss": 2.7493,
"num_input_tokens_seen": 17602296,
"step": 29500
},
{
"epoch": 0.9648784253184098,
"grad_norm": 6.775792121887207,
"learning_rate": 1.756078734079506e-06,
"loss": 2.757,
"num_input_tokens_seen": 17897832,
"step": 30000
},
{
"epoch": 0.9648784253184098,
"eval_loss": 2.5388031005859375,
"eval_runtime": 2.6123,
"eval_samples_per_second": 957.018,
"eval_steps_per_second": 119.819,
"num_input_tokens_seen": 17897832,
"step": 30000
},
{
"epoch": 0.98095973240705,
"grad_norm": 7.619235038757324,
"learning_rate": 9.520133796474978e-07,
"loss": 2.7433,
"num_input_tokens_seen": 18195568,
"step": 30500
},
{
"epoch": 0.9970410394956902,
"grad_norm": 6.682379722595215,
"learning_rate": 1.479480252154895e-07,
"loss": 2.7266,
"num_input_tokens_seen": 18491904,
"step": 31000
},
{
"epoch": 1.0131223465843304,
"grad_norm": 6.5948309898376465,
"learning_rate": 3.311462755692783e-05,
"loss": 2.6645,
"num_input_tokens_seen": 18790628,
"step": 31500
},
{
"epoch": 1.0292036536729705,
"grad_norm": 6.920671463012695,
"learning_rate": 3.2846605772117164e-05,
"loss": 2.6881,
"num_input_tokens_seen": 19090780,
"step": 32000
},
{
"epoch": 1.0452849607616108,
"grad_norm": 6.296219348907471,
"learning_rate": 3.257858398730649e-05,
"loss": 2.7024,
"num_input_tokens_seen": 19384812,
"step": 32500
},
{
"epoch": 1.0452849607616108,
"eval_loss": 2.6005640029907227,
"eval_runtime": 2.5084,
"eval_samples_per_second": 996.636,
"eval_steps_per_second": 124.779,
"num_input_tokens_seen": 19384812,
"step": 32500
},
{
"epoch": 1.061366267850251,
"grad_norm": 7.068648815155029,
"learning_rate": 3.231056220249582e-05,
"loss": 2.6939,
"num_input_tokens_seen": 19683060,
"step": 33000
},
{
"epoch": 1.077447574938891,
"grad_norm": 5.753154754638672,
"learning_rate": 3.204254041768515e-05,
"loss": 2.6977,
"num_input_tokens_seen": 19979196,
"step": 33500
},
{
"epoch": 1.0935288820275313,
"grad_norm": 8.155505180358887,
"learning_rate": 3.1774518632874485e-05,
"loss": 2.7048,
"num_input_tokens_seen": 20278524,
"step": 34000
},
{
"epoch": 1.1096101891161714,
"grad_norm": 7.031659126281738,
"learning_rate": 3.150649684806381e-05,
"loss": 2.7237,
"num_input_tokens_seen": 20577572,
"step": 34500
},
{
"epoch": 1.1256914962048115,
"grad_norm": 7.90298318862915,
"learning_rate": 3.123847506325314e-05,
"loss": 2.7248,
"num_input_tokens_seen": 20876844,
"step": 35000
},
{
"epoch": 1.1256914962048115,
"eval_loss": 2.6041972637176514,
"eval_runtime": 2.7564,
"eval_samples_per_second": 906.969,
"eval_steps_per_second": 113.552,
"num_input_tokens_seen": 20876844,
"step": 35000
},
{
"epoch": 1.1417728032934518,
"grad_norm": 6.368433475494385,
"learning_rate": 3.0970453278442473e-05,
"loss": 2.7246,
"num_input_tokens_seen": 21179820,
"step": 35500
},
{
"epoch": 1.1578541103820919,
"grad_norm": 7.143220901489258,
"learning_rate": 3.0702431493631805e-05,
"loss": 2.7211,
"num_input_tokens_seen": 21476172,
"step": 36000
},
{
"epoch": 1.173935417470732,
"grad_norm": 7.216341972351074,
"learning_rate": 3.0434409708821134e-05,
"loss": 2.7088,
"num_input_tokens_seen": 21774292,
"step": 36500
},
{
"epoch": 1.1900167245593722,
"grad_norm": 6.958596706390381,
"learning_rate": 3.0166387924010465e-05,
"loss": 2.7166,
"num_input_tokens_seen": 22070908,
"step": 37000
},
{
"epoch": 1.2060980316480123,
"grad_norm": 7.161530494689941,
"learning_rate": 2.9898366139199797e-05,
"loss": 2.6764,
"num_input_tokens_seen": 22372340,
"step": 37500
},
{
"epoch": 1.2060980316480123,
"eval_loss": 2.5923423767089844,
"eval_runtime": 2.6849,
"eval_samples_per_second": 931.145,
"eval_steps_per_second": 116.579,
"num_input_tokens_seen": 22372340,
"step": 37500
},
{
"epoch": 1.2221793387366526,
"grad_norm": 7.448612213134766,
"learning_rate": 2.963034435438913e-05,
"loss": 2.7098,
"num_input_tokens_seen": 22672932,
"step": 38000
},
{
"epoch": 1.2382606458252927,
"grad_norm": 8.339189529418945,
"learning_rate": 2.9362322569578454e-05,
"loss": 2.702,
"num_input_tokens_seen": 22971844,
"step": 38500
},
{
"epoch": 1.2543419529139328,
"grad_norm": 6.795124053955078,
"learning_rate": 2.9094300784767786e-05,
"loss": 2.7007,
"num_input_tokens_seen": 23266964,
"step": 39000
},
{
"epoch": 1.2704232600025729,
"grad_norm": 6.3036298751831055,
"learning_rate": 2.8826278999957118e-05,
"loss": 2.71,
"num_input_tokens_seen": 23564068,
"step": 39500
},
{
"epoch": 1.2865045670912132,
"grad_norm": 8.75069808959961,
"learning_rate": 2.855825721514645e-05,
"loss": 2.6854,
"num_input_tokens_seen": 23866100,
"step": 40000
},
{
"epoch": 1.2865045670912132,
"eval_loss": 2.5792863368988037,
"eval_runtime": 2.5776,
"eval_samples_per_second": 969.878,
"eval_steps_per_second": 121.429,
"num_input_tokens_seen": 23866100,
"step": 40000
},
{
"epoch": 1.3025858741798533,
"grad_norm": 6.966170310974121,
"learning_rate": 2.8290235430335778e-05,
"loss": 2.697,
"num_input_tokens_seen": 24162356,
"step": 40500
},
{
"epoch": 1.3186671812684936,
"grad_norm": 7.854964733123779,
"learning_rate": 2.802221364552511e-05,
"loss": 2.6954,
"num_input_tokens_seen": 24458980,
"step": 41000
},
{
"epoch": 1.3347484883571337,
"grad_norm": 7.1461944580078125,
"learning_rate": 2.775419186071444e-05,
"loss": 2.6839,
"num_input_tokens_seen": 24757828,
"step": 41500
},
{
"epoch": 1.3508297954457738,
"grad_norm": 8.25295639038086,
"learning_rate": 2.7486170075903773e-05,
"loss": 2.7035,
"num_input_tokens_seen": 25052236,
"step": 42000
},
{
"epoch": 1.366911102534414,
"grad_norm": 6.336223602294922,
"learning_rate": 2.7218148291093105e-05,
"loss": 2.683,
"num_input_tokens_seen": 25348084,
"step": 42500
},
{
"epoch": 1.366911102534414,
"eval_loss": 2.5722219944000244,
"eval_runtime": 2.7384,
"eval_samples_per_second": 912.958,
"eval_steps_per_second": 114.302,
"num_input_tokens_seen": 25348084,
"step": 42500
},
{
"epoch": 1.3829924096230541,
"grad_norm": 9.477555274963379,
"learning_rate": 2.695012650628243e-05,
"loss": 2.6877,
"num_input_tokens_seen": 25642372,
"step": 43000
},
{
"epoch": 1.3990737167116944,
"grad_norm": 8.233431816101074,
"learning_rate": 2.6682104721471762e-05,
"loss": 2.6927,
"num_input_tokens_seen": 25939652,
"step": 43500
},
{
"epoch": 1.4151550238003345,
"grad_norm": 5.860446929931641,
"learning_rate": 2.6414082936661094e-05,
"loss": 2.6819,
"num_input_tokens_seen": 26248940,
"step": 44000
},
{
"epoch": 1.4312363308889746,
"grad_norm": 6.748124599456787,
"learning_rate": 2.6146061151850426e-05,
"loss": 2.6893,
"num_input_tokens_seen": 26552860,
"step": 44500
},
{
"epoch": 1.4473176379776147,
"grad_norm": 6.038182258605957,
"learning_rate": 2.587803936703975e-05,
"loss": 2.6871,
"num_input_tokens_seen": 26854100,
"step": 45000
},
{
"epoch": 1.4473176379776147,
"eval_loss": 2.5538456439971924,
"eval_runtime": 2.6078,
"eval_samples_per_second": 958.67,
"eval_steps_per_second": 120.026,
"num_input_tokens_seen": 26854100,
"step": 45000
},
{
"epoch": 1.463398945066255,
"grad_norm": 7.815784454345703,
"learning_rate": 2.5610017582229086e-05,
"loss": 2.6709,
"num_input_tokens_seen": 27148148,
"step": 45500
},
{
"epoch": 1.479480252154895,
"grad_norm": 7.8851094245910645,
"learning_rate": 2.5341995797418418e-05,
"loss": 2.6698,
"num_input_tokens_seen": 27445020,
"step": 46000
},
{
"epoch": 1.4955615592435354,
"grad_norm": 7.389246940612793,
"learning_rate": 2.507397401260775e-05,
"loss": 2.6787,
"num_input_tokens_seen": 27742908,
"step": 46500
},
{
"epoch": 1.5116428663321755,
"grad_norm": 7.621913909912109,
"learning_rate": 2.4805952227797078e-05,
"loss": 2.6713,
"num_input_tokens_seen": 28037284,
"step": 47000
},
{
"epoch": 1.5277241734208156,
"grad_norm": 7.889066219329834,
"learning_rate": 2.4537930442986407e-05,
"loss": 2.6551,
"num_input_tokens_seen": 28332612,
"step": 47500
},
{
"epoch": 1.5277241734208156,
"eval_loss": 2.5442593097686768,
"eval_runtime": 2.6341,
"eval_samples_per_second": 949.086,
"eval_steps_per_second": 118.826,
"num_input_tokens_seen": 28332612,
"step": 47500
},
{
"epoch": 1.5438054805094557,
"grad_norm": 7.912906646728516,
"learning_rate": 2.426990865817574e-05,
"loss": 2.6881,
"num_input_tokens_seen": 28630948,
"step": 48000
},
{
"epoch": 1.559886787598096,
"grad_norm": 6.370878219604492,
"learning_rate": 2.4001886873365067e-05,
"loss": 2.6424,
"num_input_tokens_seen": 28928732,
"step": 48500
},
{
"epoch": 1.5759680946867363,
"grad_norm": 7.0892653465271,
"learning_rate": 2.37338650885544e-05,
"loss": 2.6626,
"num_input_tokens_seen": 29224436,
"step": 49000
},
{
"epoch": 1.5920494017753763,
"grad_norm": 6.357864856719971,
"learning_rate": 2.3465843303743727e-05,
"loss": 2.6546,
"num_input_tokens_seen": 29520148,
"step": 49500
},
{
"epoch": 1.6081307088640164,
"grad_norm": 8.4866943359375,
"learning_rate": 2.319782151893306e-05,
"loss": 2.661,
"num_input_tokens_seen": 29822156,
"step": 50000
},
{
"epoch": 1.6081307088640164,
"eval_loss": 2.527804374694824,
"eval_runtime": 2.6771,
"eval_samples_per_second": 933.855,
"eval_steps_per_second": 116.919,
"num_input_tokens_seen": 29822156,
"step": 50000
},
{
"epoch": 1.6242120159526565,
"grad_norm": 6.843733787536621,
"learning_rate": 2.292979973412239e-05,
"loss": 2.6521,
"num_input_tokens_seen": 30122052,
"step": 50500
},
{
"epoch": 1.6402933230412968,
"grad_norm": 6.88835334777832,
"learning_rate": 2.2661777949311722e-05,
"loss": 2.6614,
"num_input_tokens_seen": 30422196,
"step": 51000
},
{
"epoch": 1.656374630129937,
"grad_norm": 5.855214595794678,
"learning_rate": 2.239375616450105e-05,
"loss": 2.6454,
"num_input_tokens_seen": 30722660,
"step": 51500
},
{
"epoch": 1.6724559372185772,
"grad_norm": 6.58035135269165,
"learning_rate": 2.2125734379690383e-05,
"loss": 2.6524,
"num_input_tokens_seen": 31018220,
"step": 52000
},
{
"epoch": 1.6885372443072173,
"grad_norm": 6.767495155334473,
"learning_rate": 2.185771259487971e-05,
"loss": 2.6497,
"num_input_tokens_seen": 31319476,
"step": 52500
},
{
"epoch": 1.6885372443072173,
"eval_loss": 2.526638984680176,
"eval_runtime": 2.5213,
"eval_samples_per_second": 991.545,
"eval_steps_per_second": 124.141,
"num_input_tokens_seen": 31319476,
"step": 52500
},
{
"epoch": 1.7046185513958574,
"grad_norm": 7.022729873657227,
"learning_rate": 2.1589690810069043e-05,
"loss": 2.6437,
"num_input_tokens_seen": 31621308,
"step": 53000
},
{
"epoch": 1.7206998584844975,
"grad_norm": 6.241069793701172,
"learning_rate": 2.132166902525837e-05,
"loss": 2.6447,
"num_input_tokens_seen": 31917460,
"step": 53500
},
{
"epoch": 1.7367811655731378,
"grad_norm": 7.7204084396362305,
"learning_rate": 2.1053647240447703e-05,
"loss": 2.6448,
"num_input_tokens_seen": 32217596,
"step": 54000
},
{
"epoch": 1.752862472661778,
"grad_norm": 6.703210830688477,
"learning_rate": 2.0785625455637035e-05,
"loss": 2.6366,
"num_input_tokens_seen": 32513884,
"step": 54500
},
{
"epoch": 1.7689437797504182,
"grad_norm": 6.371466159820557,
"learning_rate": 2.0517603670826367e-05,
"loss": 2.6281,
"num_input_tokens_seen": 32813220,
"step": 55000
},
{
"epoch": 1.7689437797504182,
"eval_loss": 2.5115973949432373,
"eval_runtime": 2.5216,
"eval_samples_per_second": 991.443,
"eval_steps_per_second": 124.129,
"num_input_tokens_seen": 32813220,
"step": 55000
},
{
"epoch": 1.7850250868390583,
"grad_norm": 7.277946949005127,
"learning_rate": 2.0249581886015695e-05,
"loss": 2.6536,
"num_input_tokens_seen": 33110188,
"step": 55500
},
{
"epoch": 1.8011063939276983,
"grad_norm": 7.93104887008667,
"learning_rate": 1.9981560101205027e-05,
"loss": 2.5981,
"num_input_tokens_seen": 33405980,
"step": 56000
},
{
"epoch": 1.8171877010163386,
"grad_norm": 7.782486438751221,
"learning_rate": 1.971353831639436e-05,
"loss": 2.635,
"num_input_tokens_seen": 33700596,
"step": 56500
},
{
"epoch": 1.8332690081049787,
"grad_norm": 8.59358024597168,
"learning_rate": 1.9445516531583687e-05,
"loss": 2.6269,
"num_input_tokens_seen": 33997724,
"step": 57000
},
{
"epoch": 1.849350315193619,
"grad_norm": 6.669950485229492,
"learning_rate": 1.917749474677302e-05,
"loss": 2.6067,
"num_input_tokens_seen": 34298052,
"step": 57500
},
{
"epoch": 1.849350315193619,
"eval_loss": 2.5047078132629395,
"eval_runtime": 2.5169,
"eval_samples_per_second": 993.288,
"eval_steps_per_second": 124.36,
"num_input_tokens_seen": 34298052,
"step": 57500
},
{
"epoch": 1.8654316222822591,
"grad_norm": 6.265903949737549,
"learning_rate": 1.8909472961962348e-05,
"loss": 2.5966,
"num_input_tokens_seen": 34593980,
"step": 58000
},
{
"epoch": 1.8815129293708992,
"grad_norm": 7.943974018096924,
"learning_rate": 1.864145117715168e-05,
"loss": 2.6303,
"num_input_tokens_seen": 34894428,
"step": 58500
},
{
"epoch": 1.8975942364595393,
"grad_norm": 8.290629386901855,
"learning_rate": 1.837342939234101e-05,
"loss": 2.6303,
"num_input_tokens_seen": 35193236,
"step": 59000
},
{
"epoch": 1.9136755435481796,
"grad_norm": 7.974947929382324,
"learning_rate": 1.8105407607530343e-05,
"loss": 2.6272,
"num_input_tokens_seen": 35486796,
"step": 59500
},
{
"epoch": 1.92975685063682,
"grad_norm": 5.827637195587158,
"learning_rate": 1.783738582271967e-05,
"loss": 2.6112,
"num_input_tokens_seen": 35783604,
"step": 60000
},
{
"epoch": 1.92975685063682,
"eval_loss": 2.4935405254364014,
"eval_runtime": 2.5107,
"eval_samples_per_second": 995.741,
"eval_steps_per_second": 124.667,
"num_input_tokens_seen": 35783604,
"step": 60000
},
{
"epoch": 1.94583815772546,
"grad_norm": 6.535378456115723,
"learning_rate": 1.7569364037909003e-05,
"loss": 2.6135,
"num_input_tokens_seen": 36086620,
"step": 60500
},
{
"epoch": 1.9619194648141,
"grad_norm": 6.398725986480713,
"learning_rate": 1.7301342253098332e-05,
"loss": 2.602,
"num_input_tokens_seen": 36386140,
"step": 61000
},
{
"epoch": 1.9780007719027402,
"grad_norm": 6.332113265991211,
"learning_rate": 1.7033320468287664e-05,
"loss": 2.6258,
"num_input_tokens_seen": 36684308,
"step": 61500
},
{
"epoch": 1.9940820789913805,
"grad_norm": 7.8002753257751465,
"learning_rate": 1.6765298683476992e-05,
"loss": 2.6226,
"num_input_tokens_seen": 36984724,
"step": 62000
},
{
"epoch": 2.0101633860800208,
"grad_norm": 6.9457011222839355,
"learning_rate": 1.6497276898666324e-05,
"loss": 2.5207,
"num_input_tokens_seen": 37281092,
"step": 62500
},
{
"epoch": 2.0101633860800208,
"eval_loss": 2.4945950508117676,
"eval_runtime": 2.5094,
"eval_samples_per_second": 996.26,
"eval_steps_per_second": 124.732,
"num_input_tokens_seen": 37281092,
"step": 62500
},
{
"epoch": 2.026244693168661,
"grad_norm": 7.541498184204102,
"learning_rate": 1.6229255113855656e-05,
"loss": 2.4728,
"num_input_tokens_seen": 37582300,
"step": 63000
},
{
"epoch": 2.042326000257301,
"grad_norm": 6.7798027992248535,
"learning_rate": 1.5961233329044987e-05,
"loss": 2.4539,
"num_input_tokens_seen": 37880828,
"step": 63500
},
{
"epoch": 2.058407307345941,
"grad_norm": 7.033351898193359,
"learning_rate": 1.5693211544234316e-05,
"loss": 2.4467,
"num_input_tokens_seen": 38181276,
"step": 64000
},
{
"epoch": 2.074488614434581,
"grad_norm": 6.487890720367432,
"learning_rate": 1.5425189759423648e-05,
"loss": 2.4764,
"num_input_tokens_seen": 38473348,
"step": 64500
},
{
"epoch": 2.0905699215232216,
"grad_norm": 6.955127716064453,
"learning_rate": 1.5157167974612976e-05,
"loss": 2.4799,
"num_input_tokens_seen": 38768588,
"step": 65000
},
{
"epoch": 2.0905699215232216,
"eval_loss": 2.491555690765381,
"eval_runtime": 2.5076,
"eval_samples_per_second": 996.967,
"eval_steps_per_second": 124.82,
"num_input_tokens_seen": 38768588,
"step": 65000
},
{
"epoch": 2.1066512286118617,
"grad_norm": 6.78762674331665,
"learning_rate": 1.4889146189802308e-05,
"loss": 2.4726,
"num_input_tokens_seen": 39067460,
"step": 65500
},
{
"epoch": 2.122732535700502,
"grad_norm": 7.199331283569336,
"learning_rate": 1.4621124404991638e-05,
"loss": 2.4562,
"num_input_tokens_seen": 39360244,
"step": 66000
},
{
"epoch": 2.138813842789142,
"grad_norm": 7.353775501251221,
"learning_rate": 1.435310262018097e-05,
"loss": 2.4629,
"num_input_tokens_seen": 39660020,
"step": 66500
},
{
"epoch": 2.154895149877782,
"grad_norm": 6.827337265014648,
"learning_rate": 1.4085080835370298e-05,
"loss": 2.4817,
"num_input_tokens_seen": 39960476,
"step": 67000
},
{
"epoch": 2.170976456966422,
"grad_norm": 6.532020092010498,
"learning_rate": 1.381705905055963e-05,
"loss": 2.4727,
"num_input_tokens_seen": 40252972,
"step": 67500
},
{
"epoch": 2.170976456966422,
"eval_loss": 2.4865615367889404,
"eval_runtime": 2.5372,
"eval_samples_per_second": 985.335,
"eval_steps_per_second": 123.364,
"num_input_tokens_seen": 40252972,
"step": 67500
},
{
"epoch": 2.1870577640550626,
"grad_norm": 6.601158142089844,
"learning_rate": 1.354903726574896e-05,
"loss": 2.4666,
"num_input_tokens_seen": 40553732,
"step": 68000
},
{
"epoch": 2.2031390711437027,
"grad_norm": 7.200645446777344,
"learning_rate": 1.3281015480938292e-05,
"loss": 2.4657,
"num_input_tokens_seen": 40851196,
"step": 68500
},
{
"epoch": 2.2192203782323427,
"grad_norm": 8.067240715026855,
"learning_rate": 1.301299369612762e-05,
"loss": 2.4801,
"num_input_tokens_seen": 41149276,
"step": 69000
},
{
"epoch": 2.235301685320983,
"grad_norm": 7.724194526672363,
"learning_rate": 1.2744971911316952e-05,
"loss": 2.4766,
"num_input_tokens_seen": 41448540,
"step": 69500
},
{
"epoch": 2.251382992409623,
"grad_norm": 6.999200344085693,
"learning_rate": 1.2476950126506282e-05,
"loss": 2.4719,
"num_input_tokens_seen": 41746300,
"step": 70000
},
{
"epoch": 2.251382992409623,
"eval_loss": 2.476020097732544,
"eval_runtime": 2.5166,
"eval_samples_per_second": 993.407,
"eval_steps_per_second": 124.375,
"num_input_tokens_seen": 41746300,
"step": 70000
},
{
"epoch": 2.267464299498263,
"grad_norm": 6.666884899139404,
"learning_rate": 1.2208928341695614e-05,
"loss": 2.4771,
"num_input_tokens_seen": 42042532,
"step": 70500
},
{
"epoch": 2.2835456065869035,
"grad_norm": 8.354509353637695,
"learning_rate": 1.1940906556884944e-05,
"loss": 2.4679,
"num_input_tokens_seen": 42341628,
"step": 71000
},
{
"epoch": 2.2996269136755436,
"grad_norm": 8.39284610748291,
"learning_rate": 1.1672884772074275e-05,
"loss": 2.4597,
"num_input_tokens_seen": 42642948,
"step": 71500
},
{
"epoch": 2.3157082207641837,
"grad_norm": 7.233700275421143,
"learning_rate": 1.1404862987263605e-05,
"loss": 2.4592,
"num_input_tokens_seen": 42941820,
"step": 72000
},
{
"epoch": 2.331789527852824,
"grad_norm": 7.843503475189209,
"learning_rate": 1.1136841202452935e-05,
"loss": 2.4738,
"num_input_tokens_seen": 43241188,
"step": 72500
},
{
"epoch": 2.331789527852824,
"eval_loss": 2.47127103805542,
"eval_runtime": 2.5277,
"eval_samples_per_second": 989.05,
"eval_steps_per_second": 123.829,
"num_input_tokens_seen": 43241188,
"step": 72500
},
{
"epoch": 2.347870834941464,
"grad_norm": 6.37482213973999,
"learning_rate": 1.0868819417642267e-05,
"loss": 2.4576,
"num_input_tokens_seen": 43535900,
"step": 73000
},
{
"epoch": 2.3639521420301044,
"grad_norm": 6.642532825469971,
"learning_rate": 1.0600797632831597e-05,
"loss": 2.467,
"num_input_tokens_seen": 43833516,
"step": 73500
},
{
"epoch": 2.3800334491187445,
"grad_norm": 6.606197357177734,
"learning_rate": 1.0332775848020927e-05,
"loss": 2.4752,
"num_input_tokens_seen": 44134084,
"step": 74000
},
{
"epoch": 2.3961147562073846,
"grad_norm": 6.338978290557861,
"learning_rate": 1.0064754063210257e-05,
"loss": 2.4473,
"num_input_tokens_seen": 44432540,
"step": 74500
},
{
"epoch": 2.4121960632960247,
"grad_norm": 7.172792434692383,
"learning_rate": 9.796732278399589e-06,
"loss": 2.4629,
"num_input_tokens_seen": 44730244,
"step": 75000
},
{
"epoch": 2.4121960632960247,
"eval_loss": 2.4629955291748047,
"eval_runtime": 2.521,
"eval_samples_per_second": 991.68,
"eval_steps_per_second": 124.158,
"num_input_tokens_seen": 44730244,
"step": 75000
},
{
"epoch": 2.4282773703846647,
"grad_norm": 6.65930700302124,
"learning_rate": 9.528710493588919e-06,
"loss": 2.4512,
"num_input_tokens_seen": 45031884,
"step": 75500
},
{
"epoch": 2.4443586774733053,
"grad_norm": 7.209745407104492,
"learning_rate": 9.260688708778249e-06,
"loss": 2.4557,
"num_input_tokens_seen": 45334924,
"step": 76000
},
{
"epoch": 2.4604399845619453,
"grad_norm": 6.847073078155518,
"learning_rate": 8.992666923967581e-06,
"loss": 2.4512,
"num_input_tokens_seen": 45633532,
"step": 76500
},
{
"epoch": 2.4765212916505854,
"grad_norm": 7.705162525177002,
"learning_rate": 8.724645139156911e-06,
"loss": 2.4568,
"num_input_tokens_seen": 45933804,
"step": 77000
},
{
"epoch": 2.4926025987392255,
"grad_norm": 7.5681962966918945,
"learning_rate": 8.456623354346243e-06,
"loss": 2.4524,
"num_input_tokens_seen": 46231060,
"step": 77500
},
{
"epoch": 2.4926025987392255,
"eval_loss": 2.457481861114502,
"eval_runtime": 2.5921,
"eval_samples_per_second": 964.467,
"eval_steps_per_second": 120.751,
"num_input_tokens_seen": 46231060,
"step": 77500
},
{
"epoch": 2.5086839058278656,
"grad_norm": 6.8857269287109375,
"learning_rate": 8.188601569535573e-06,
"loss": 2.4622,
"num_input_tokens_seen": 46525772,
"step": 78000
},
{
"epoch": 2.524765212916506,
"grad_norm": 6.347681522369385,
"learning_rate": 7.920579784724903e-06,
"loss": 2.4528,
"num_input_tokens_seen": 46822532,
"step": 78500
},
{
"epoch": 2.5408465200051458,
"grad_norm": 6.935575008392334,
"learning_rate": 7.652557999914233e-06,
"loss": 2.4414,
"num_input_tokens_seen": 47122964,
"step": 79000
},
{
"epoch": 2.5569278270937863,
"grad_norm": 6.603360652923584,
"learning_rate": 7.384536215103564e-06,
"loss": 2.4655,
"num_input_tokens_seen": 47423300,
"step": 79500
},
{
"epoch": 2.5730091341824264,
"grad_norm": 7.182071208953857,
"learning_rate": 7.116514430292895e-06,
"loss": 2.435,
"num_input_tokens_seen": 47718964,
"step": 80000
},
{
"epoch": 2.5730091341824264,
"eval_loss": 2.455320358276367,
"eval_runtime": 2.5065,
"eval_samples_per_second": 997.4,
"eval_steps_per_second": 124.874,
"num_input_tokens_seen": 47718964,
"step": 80000
},
{
"epoch": 2.5890904412710665,
"grad_norm": 7.3647260665893555,
"learning_rate": 6.848492645482225e-06,
"loss": 2.4356,
"num_input_tokens_seen": 48015996,
"step": 80500
},
{
"epoch": 2.6051717483597066,
"grad_norm": 7.950341701507568,
"learning_rate": 6.580470860671556e-06,
"loss": 2.4453,
"num_input_tokens_seen": 48316420,
"step": 81000
},
{
"epoch": 2.6212530554483466,
"grad_norm": 6.016787052154541,
"learning_rate": 6.312449075860886e-06,
"loss": 2.45,
"num_input_tokens_seen": 48611452,
"step": 81500
},
{
"epoch": 2.637334362536987,
"grad_norm": 7.281980514526367,
"learning_rate": 6.044427291050217e-06,
"loss": 2.4687,
"num_input_tokens_seen": 48913668,
"step": 82000
},
{
"epoch": 2.6534156696256272,
"grad_norm": 6.644787311553955,
"learning_rate": 5.776405506239547e-06,
"loss": 2.4621,
"num_input_tokens_seen": 49209724,
"step": 82500
},
{
"epoch": 2.6534156696256272,
"eval_loss": 2.4475488662719727,
"eval_runtime": 2.517,
"eval_samples_per_second": 993.239,
"eval_steps_per_second": 124.354,
"num_input_tokens_seen": 49209724,
"step": 82500
},
{
"epoch": 2.6694969767142673,
"grad_norm": 6.181220054626465,
"learning_rate": 5.508383721428878e-06,
"loss": 2.4343,
"num_input_tokens_seen": 49505772,
"step": 83000
},
{
"epoch": 2.6855782838029074,
"grad_norm": 6.418393135070801,
"learning_rate": 5.2403619366182085e-06,
"loss": 2.4329,
"num_input_tokens_seen": 49809956,
"step": 83500
},
{
"epoch": 2.7016595908915475,
"grad_norm": 6.279716491699219,
"learning_rate": 4.9723401518075395e-06,
"loss": 2.4481,
"num_input_tokens_seen": 50112060,
"step": 84000
},
{
"epoch": 2.717740897980188,
"grad_norm": 6.502873420715332,
"learning_rate": 4.70431836699687e-06,
"loss": 2.4464,
"num_input_tokens_seen": 50414356,
"step": 84500
},
{
"epoch": 2.733822205068828,
"grad_norm": 6.15990686416626,
"learning_rate": 4.4362965821862e-06,
"loss": 2.4492,
"num_input_tokens_seen": 50712980,
"step": 85000
},
{
"epoch": 2.733822205068828,
"eval_loss": 2.4440150260925293,
"eval_runtime": 2.529,
"eval_samples_per_second": 988.544,
"eval_steps_per_second": 123.766,
"num_input_tokens_seen": 50712980,
"step": 85000
},
{
"epoch": 2.749903512157468,
"grad_norm": 6.876352310180664,
"learning_rate": 4.168274797375531e-06,
"loss": 2.4514,
"num_input_tokens_seen": 51012460,
"step": 85500
},
{
"epoch": 2.7659848192461083,
"grad_norm": 7.305426597595215,
"learning_rate": 3.900253012564861e-06,
"loss": 2.4317,
"num_input_tokens_seen": 51308524,
"step": 86000
},
{
"epoch": 2.7820661263347484,
"grad_norm": 6.460892677307129,
"learning_rate": 3.632231227754192e-06,
"loss": 2.4559,
"num_input_tokens_seen": 51610700,
"step": 86500
},
{
"epoch": 2.798147433423389,
"grad_norm": 8.062651634216309,
"learning_rate": 3.3642094429435228e-06,
"loss": 2.4535,
"num_input_tokens_seen": 51910236,
"step": 87000
},
{
"epoch": 2.814228740512029,
"grad_norm": 7.140311241149902,
"learning_rate": 3.0961876581328533e-06,
"loss": 2.4536,
"num_input_tokens_seen": 52204380,
"step": 87500
},
{
"epoch": 2.814228740512029,
"eval_loss": 2.4393906593322754,
"eval_runtime": 2.5312,
"eval_samples_per_second": 987.685,
"eval_steps_per_second": 123.658,
"num_input_tokens_seen": 52204380,
"step": 87500
},
{
"epoch": 2.830310047600669,
"grad_norm": 6.569787502288818,
"learning_rate": 2.8281658733221834e-06,
"loss": 2.4379,
"num_input_tokens_seen": 52504668,
"step": 88000
},
{
"epoch": 2.846391354689309,
"grad_norm": 7.735711097717285,
"learning_rate": 2.560144088511514e-06,
"loss": 2.4239,
"num_input_tokens_seen": 52798740,
"step": 88500
},
{
"epoch": 2.8624726617779492,
"grad_norm": 7.504124641418457,
"learning_rate": 2.292122303700845e-06,
"loss": 2.4427,
"num_input_tokens_seen": 53097716,
"step": 89000
},
{
"epoch": 2.8785539688665893,
"grad_norm": 6.647756099700928,
"learning_rate": 2.0241005188901755e-06,
"loss": 2.4682,
"num_input_tokens_seen": 53397564,
"step": 89500
},
{
"epoch": 2.8946352759552294,
"grad_norm": 6.640815734863281,
"learning_rate": 1.756078734079506e-06,
"loss": 2.4148,
"num_input_tokens_seen": 53695620,
"step": 90000
},
{
"epoch": 2.8946352759552294,
"eval_loss": 2.43597674369812,
"eval_runtime": 2.5128,
"eval_samples_per_second": 994.891,
"eval_steps_per_second": 124.56,
"num_input_tokens_seen": 53695620,
"step": 90000
},
{
"epoch": 2.91071658304387,
"grad_norm": 7.346447467803955,
"learning_rate": 1.4880569492688366e-06,
"loss": 2.4352,
"num_input_tokens_seen": 53991180,
"step": 90500
},
{
"epoch": 2.92679789013251,
"grad_norm": 6.777767658233643,
"learning_rate": 1.2200351644581672e-06,
"loss": 2.4664,
"num_input_tokens_seen": 54288348,
"step": 91000
},
{
"epoch": 2.94287919722115,
"grad_norm": 6.908254623413086,
"learning_rate": 9.520133796474978e-07,
"loss": 2.4474,
"num_input_tokens_seen": 54590740,
"step": 91500
},
{
"epoch": 2.95896050430979,
"grad_norm": 7.04544734954834,
"learning_rate": 6.839915948368284e-07,
"loss": 2.4554,
"num_input_tokens_seen": 54889220,
"step": 92000
},
{
"epoch": 2.9750418113984303,
"grad_norm": 9.98161792755127,
"learning_rate": 4.159698100261589e-07,
"loss": 2.4243,
"num_input_tokens_seen": 55190020,
"step": 92500
},
{
"epoch": 2.9750418113984303,
"eval_loss": 2.435030460357666,
"eval_runtime": 2.5128,
"eval_samples_per_second": 994.908,
"eval_steps_per_second": 124.563,
"num_input_tokens_seen": 55190020,
"step": 92500
},
{
"epoch": 2.991123118487071,
"grad_norm": 6.586206912994385,
"learning_rate": 1.479480252154895e-07,
"loss": 2.44,
"num_input_tokens_seen": 55490868,
"step": 93000
},
{
"epoch": 3.0,
"num_input_tokens_seen": 55653732,
"step": 93276,
"total_flos": 1.4738832163602432e+16,
"train_loss": 1.7073542784139526,
"train_runtime": 2504.3889,
"train_samples_per_second": 297.955,
"train_steps_per_second": 37.245,
"train_tokens_per_second": 22216.164
}
],
"logging_steps": 500,
"max_steps": 93276,
"num_input_tokens_seen": 55653732,
"num_train_epochs": 3,
"save_steps": 2500,
"stateful_callbacks": {
"TrainerControl": {
"args": {
"should_epoch_stop": false,
"should_evaluate": false,
"should_log": false,
"should_save": true,
"should_training_stop": true
},
"attributes": {}
}
},
"total_flos": 1.4738832163602432e+16,
"train_batch_size": 8,
"trial_name": null,
"trial_params": null
}