imdatta0's picture
End of training
8b1b653 verified
raw
history blame
26.1 kB
{
"best_metric": 0.506678581237793,
"best_model_checkpoint": "/home/datta0/models/lora_final/Meta-Llama-3-8B_metamath_reverse/checkpoint-611",
"epoch": 0.9898744430943702,
"eval_steps": 13,
"global_step": 611,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.0016200891049007696,
"grad_norm": 21.833539962768555,
"learning_rate": 2.3076923076923076e-05,
"loss": 1.2178,
"step": 1
},
{
"epoch": 0.011340623734305387,
"grad_norm": 4.27054500579834,
"learning_rate": 0.00016153846153846153,
"loss": 0.8667,
"step": 7
},
{
"epoch": 0.021061158363710003,
"eval_loss": 0.7332164645195007,
"eval_runtime": 12.1685,
"eval_samples_per_second": 41.09,
"eval_steps_per_second": 5.177,
"step": 13
},
{
"epoch": 0.022681247468610773,
"grad_norm": 4.137913703918457,
"learning_rate": 0.0002999979709808197,
"loss": 0.6948,
"step": 14
},
{
"epoch": 0.03402187120291616,
"grad_norm": 6.889961242675781,
"learning_rate": 0.0002998701612152596,
"loss": 0.7009,
"step": 21
},
{
"epoch": 0.042122316727420006,
"eval_loss": 0.7322337627410889,
"eval_runtime": 98.0914,
"eval_samples_per_second": 5.097,
"eval_steps_per_second": 0.642,
"step": 26
},
{
"epoch": 0.04536249493722155,
"grad_norm": 7.484898567199707,
"learning_rate": 0.0002995437011859465,
"loss": 0.7139,
"step": 28
},
{
"epoch": 0.056703118671526935,
"grad_norm": 9.079093933105469,
"learning_rate": 0.00029901902360990936,
"loss": 0.7161,
"step": 35
},
{
"epoch": 0.06318347509113001,
"eval_loss": 0.7228567004203796,
"eval_runtime": 99.5149,
"eval_samples_per_second": 5.024,
"eval_steps_per_second": 0.633,
"step": 39
},
{
"epoch": 0.06804374240583232,
"grad_norm": 4.264255523681641,
"learning_rate": 0.00029829682393805085,
"loss": 0.6852,
"step": 42
},
{
"epoch": 0.0793843661401377,
"grad_norm": 4.362518787384033,
"learning_rate": 0.0002973780594333385,
"loss": 0.6909,
"step": 49
},
{
"epoch": 0.08424463345484001,
"eval_loss": 0.7190279364585876,
"eval_runtime": 12.1598,
"eval_samples_per_second": 41.119,
"eval_steps_per_second": 5.181,
"step": 52
},
{
"epoch": 0.0907249898744431,
"grad_norm": 5.616891384124756,
"learning_rate": 0.00029626394790197025,
"loss": 0.6897,
"step": 56
},
{
"epoch": 0.10206561360874848,
"grad_norm": 7.568899154663086,
"learning_rate": 0.00029495596607919305,
"loss": 0.6541,
"step": 63
},
{
"epoch": 0.10530579181855002,
"eval_loss": 0.708296000957489,
"eval_runtime": 108.5306,
"eval_samples_per_second": 4.607,
"eval_steps_per_second": 0.58,
"step": 65
},
{
"epoch": 0.11340623734305387,
"grad_norm": 6.362358570098877,
"learning_rate": 0.00029345584767191685,
"loss": 0.6792,
"step": 70
},
{
"epoch": 0.12474686107735926,
"grad_norm": 5.232390403747559,
"learning_rate": 0.0002917655810607161,
"loss": 0.6704,
"step": 77
},
{
"epoch": 0.12636695018226002,
"eval_loss": 0.7013765573501587,
"eval_runtime": 107.5502,
"eval_samples_per_second": 4.649,
"eval_steps_per_second": 0.586,
"step": 78
},
{
"epoch": 0.13608748481166463,
"grad_norm": 5.135895252227783,
"learning_rate": 0.0002898874066642667,
"loss": 0.6633,
"step": 84
},
{
"epoch": 0.14742810854597002,
"grad_norm": 6.573320388793945,
"learning_rate": 0.00028782381396971003,
"loss": 0.6806,
"step": 91
},
{
"epoch": 0.14742810854597002,
"eval_loss": 0.699935257434845,
"eval_runtime": 12.1729,
"eval_samples_per_second": 41.075,
"eval_steps_per_second": 5.175,
"step": 91
},
{
"epoch": 0.1587687322802754,
"grad_norm": 6.816903591156006,
"learning_rate": 0.00028557753823288173,
"loss": 0.6735,
"step": 98
},
{
"epoch": 0.16848926690968002,
"eval_loss": 0.6932142376899719,
"eval_runtime": 108.078,
"eval_samples_per_second": 4.626,
"eval_steps_per_second": 0.583,
"step": 104
},
{
"epoch": 0.1701093560145808,
"grad_norm": 5.952498435974121,
"learning_rate": 0.0002831515568527781,
"loss": 0.6837,
"step": 105
},
{
"epoch": 0.1814499797488862,
"grad_norm": 7.350447177886963,
"learning_rate": 0.00028054908542506627,
"loss": 0.6509,
"step": 112
},
{
"epoch": 0.18955042527339003,
"eval_loss": 0.6961794495582581,
"eval_runtime": 119.5602,
"eval_samples_per_second": 4.182,
"eval_steps_per_second": 0.527,
"step": 117
},
{
"epoch": 0.19279060348319157,
"grad_norm": 5.018365859985352,
"learning_rate": 0.00027777357347986823,
"loss": 0.6661,
"step": 119
},
{
"epoch": 0.20413122721749696,
"grad_norm": 6.165168285369873,
"learning_rate": 0.00027482869990946986,
"loss": 0.6537,
"step": 126
},
{
"epoch": 0.21061158363710003,
"eval_loss": 0.6906590461730957,
"eval_runtime": 12.1745,
"eval_samples_per_second": 41.069,
"eval_steps_per_second": 5.175,
"step": 130
},
{
"epoch": 0.21547185095180235,
"grad_norm": 7.098977088928223,
"learning_rate": 0.0002717183680920135,
"loss": 0.6461,
"step": 133
},
{
"epoch": 0.22681247468610774,
"grad_norm": 4.91437292098999,
"learning_rate": 0.00026844670071763906,
"loss": 0.6508,
"step": 140
},
{
"epoch": 0.23167274200081003,
"eval_loss": 0.6892094612121582,
"eval_runtime": 99.6868,
"eval_samples_per_second": 5.016,
"eval_steps_per_second": 0.632,
"step": 143
},
{
"epoch": 0.23815309842041313,
"grad_norm": 5.432809352874756,
"learning_rate": 0.00026501803432393037,
"loss": 0.6696,
"step": 147
},
{
"epoch": 0.24949372215471852,
"grad_norm": 5.0243706703186035,
"learning_rate": 0.00026143691354791145,
"loss": 0.6594,
"step": 154
},
{
"epoch": 0.25273390036452004,
"eval_loss": 0.6816443204879761,
"eval_runtime": 102.7811,
"eval_samples_per_second": 4.865,
"eval_steps_per_second": 0.613,
"step": 156
},
{
"epoch": 0.2608343458890239,
"grad_norm": 4.282688617706299,
"learning_rate": 0.00025770808510220956,
"loss": 0.6503,
"step": 161
},
{
"epoch": 0.27217496962332927,
"grad_norm": 5.6248884201049805,
"learning_rate": 0.00025383649148337105,
"loss": 0.6534,
"step": 168
},
{
"epoch": 0.27379505872823007,
"eval_loss": 0.6733805537223816,
"eval_runtime": 12.1401,
"eval_samples_per_second": 41.186,
"eval_steps_per_second": 5.189,
"step": 169
},
{
"epoch": 0.28351559335763465,
"grad_norm": 6.263326168060303,
"learning_rate": 0.0002498272644206695,
"loss": 0.657,
"step": 175
},
{
"epoch": 0.29485621709194004,
"grad_norm": 5.819485187530518,
"learning_rate": 0.0002456857180740884,
"loss": 0.6559,
"step": 182
},
{
"epoch": 0.29485621709194004,
"eval_loss": 0.6744294166564941,
"eval_runtime": 93.9309,
"eval_samples_per_second": 5.323,
"eval_steps_per_second": 0.671,
"step": 182
},
{
"epoch": 0.30619684082624543,
"grad_norm": 4.828428268432617,
"learning_rate": 0.0002414173419904956,
"loss": 0.6391,
"step": 189
},
{
"epoch": 0.3159173754556501,
"eval_loss": 0.6739293932914734,
"eval_runtime": 94.7015,
"eval_samples_per_second": 5.28,
"eval_steps_per_second": 0.665,
"step": 195
},
{
"epoch": 0.3175374645605508,
"grad_norm": 4.6714348793029785,
"learning_rate": 0.00023702779382734566,
"loss": 0.6434,
"step": 196
},
{
"epoch": 0.3288780882948562,
"grad_norm": 4.983964920043945,
"learning_rate": 0.0002325228918535541,
"loss": 0.6115,
"step": 203
},
{
"epoch": 0.33697853381936005,
"eval_loss": 0.6627940535545349,
"eval_runtime": 12.1956,
"eval_samples_per_second": 40.998,
"eval_steps_per_second": 5.166,
"step": 208
},
{
"epoch": 0.3402187120291616,
"grad_norm": 4.924707412719727,
"learning_rate": 0.00022790860723748442,
"loss": 0.6406,
"step": 210
},
{
"epoch": 0.351559335763467,
"grad_norm": 5.098822116851807,
"learning_rate": 0.00022319105613226921,
"loss": 0.6261,
"step": 217
},
{
"epoch": 0.3580396921830701,
"eval_loss": 0.6547751426696777,
"eval_runtime": 95.155,
"eval_samples_per_second": 5.255,
"eval_steps_per_second": 0.662,
"step": 221
},
{
"epoch": 0.3628999594977724,
"grad_norm": 4.204422950744629,
"learning_rate": 0.00021837649156895706,
"loss": 0.6338,
"step": 224
},
{
"epoch": 0.37424058323207776,
"grad_norm": 6.743808746337891,
"learning_rate": 0.00021347129516822945,
"loss": 0.6288,
"step": 231
},
{
"epoch": 0.37910085054678005,
"eval_loss": 0.6544941067695618,
"eval_runtime": 94.1977,
"eval_samples_per_second": 5.308,
"eval_steps_per_second": 0.669,
"step": 234
},
{
"epoch": 0.38558120696638315,
"grad_norm": 4.939682960510254,
"learning_rate": 0.00020848196868167505,
"loss": 0.641,
"step": 238
},
{
"epoch": 0.39692183070068854,
"grad_norm": 3.6235039234161377,
"learning_rate": 0.000203415125373832,
"loss": 0.6377,
"step": 245
},
{
"epoch": 0.4001620089104901,
"eval_loss": 0.6509546041488647,
"eval_runtime": 12.1816,
"eval_samples_per_second": 41.046,
"eval_steps_per_second": 5.172,
"step": 247
},
{
"epoch": 0.4082624544349939,
"grad_norm": 3.9135453701019287,
"learning_rate": 0.00019827748125642242,
"loss": 0.6295,
"step": 252
},
{
"epoch": 0.4196030781692993,
"grad_norm": 4.376216411590576,
"learning_rate": 0.0001930758461863965,
"loss": 0.6106,
"step": 259
},
{
"epoch": 0.42122316727420006,
"eval_loss": 0.6465097069740295,
"eval_runtime": 95.6315,
"eval_samples_per_second": 5.228,
"eval_steps_per_second": 0.659,
"step": 260
},
{
"epoch": 0.4309437019036047,
"grad_norm": 4.2812418937683105,
"learning_rate": 0.0001878171148395872,
"loss": 0.6208,
"step": 266
},
{
"epoch": 0.4422843256379101,
"grad_norm": 5.637602806091309,
"learning_rate": 0.00018250825757193848,
"loss": 0.6203,
"step": 273
},
{
"epoch": 0.4422843256379101,
"eval_loss": 0.637697160243988,
"eval_runtime": 96.3956,
"eval_samples_per_second": 5.187,
"eval_steps_per_second": 0.654,
"step": 273
},
{
"epoch": 0.4536249493722155,
"grad_norm": 3.6267099380493164,
"learning_rate": 0.0001771563111804211,
"loss": 0.6196,
"step": 280
},
{
"epoch": 0.46334548400162007,
"eval_loss": 0.6275562047958374,
"eval_runtime": 12.1998,
"eval_samples_per_second": 40.984,
"eval_steps_per_second": 5.164,
"step": 286
},
{
"epoch": 0.46496557310652087,
"grad_norm": 5.152669906616211,
"learning_rate": 0.0001717683695758819,
"loss": 0.6019,
"step": 287
},
{
"epoch": 0.47630619684082626,
"grad_norm": 3.877650022506714,
"learning_rate": 0.00016635157438018983,
"loss": 0.6146,
"step": 294
},
{
"epoch": 0.4844066423653301,
"eval_loss": 0.6215513944625854,
"eval_runtime": 109.2195,
"eval_samples_per_second": 4.578,
"eval_steps_per_second": 0.577,
"step": 299
},
{
"epoch": 0.48764682057513165,
"grad_norm": 4.388294696807861,
"learning_rate": 0.0001609131054601416,
"loss": 0.6115,
"step": 301
},
{
"epoch": 0.49898744430943703,
"grad_norm": 3.800142765045166,
"learning_rate": 0.00015546017141067432,
"loss": 0.5931,
"step": 308
},
{
"epoch": 0.5054678007290401,
"eval_loss": 0.6187200546264648,
"eval_runtime": 102.485,
"eval_samples_per_second": 4.879,
"eval_steps_per_second": 0.615,
"step": 312
},
{
"epoch": 0.5103280680437424,
"grad_norm": 3.4495720863342285,
"learning_rate": 0.00015,
"loss": 0.6109,
"step": 315
},
{
"epoch": 0.5216686917780478,
"grad_norm": 4.938626289367676,
"learning_rate": 0.0001445398285893257,
"loss": 0.5926,
"step": 322
},
{
"epoch": 0.5265289590927501,
"eval_loss": 0.605811357498169,
"eval_runtime": 12.2293,
"eval_samples_per_second": 40.885,
"eval_steps_per_second": 5.152,
"step": 325
},
{
"epoch": 0.5330093155123532,
"grad_norm": 2.700883626937866,
"learning_rate": 0.0001390868945398584,
"loss": 0.5906,
"step": 329
},
{
"epoch": 0.5443499392466585,
"grad_norm": 2.808101177215576,
"learning_rate": 0.00013364842561981014,
"loss": 0.5807,
"step": 336
},
{
"epoch": 0.5475901174564601,
"eval_loss": 0.6017518043518066,
"eval_runtime": 107.7874,
"eval_samples_per_second": 4.639,
"eval_steps_per_second": 0.584,
"step": 338
},
{
"epoch": 0.555690562980964,
"grad_norm": 5.031184196472168,
"learning_rate": 0.00012823163042411807,
"loss": 0.5699,
"step": 343
},
{
"epoch": 0.5670311867152693,
"grad_norm": 2.7506072521209717,
"learning_rate": 0.0001228436888195789,
"loss": 0.5738,
"step": 350
},
{
"epoch": 0.56865127582017,
"eval_loss": 0.591549813747406,
"eval_runtime": 108.9536,
"eval_samples_per_second": 4.589,
"eval_steps_per_second": 0.578,
"step": 351
},
{
"epoch": 0.5783718104495748,
"grad_norm": 3.25809645652771,
"learning_rate": 0.00011749174242806152,
"loss": 0.5777,
"step": 357
},
{
"epoch": 0.5897124341838801,
"grad_norm": 3.419900417327881,
"learning_rate": 0.00011218288516041279,
"loss": 0.5509,
"step": 364
},
{
"epoch": 0.5897124341838801,
"eval_loss": 0.5852319002151489,
"eval_runtime": 12.2275,
"eval_samples_per_second": 40.892,
"eval_steps_per_second": 5.152,
"step": 364
},
{
"epoch": 0.6010530579181855,
"grad_norm": 3.074368953704834,
"learning_rate": 0.00010692415381360349,
"loss": 0.5641,
"step": 371
},
{
"epoch": 0.6107735925475901,
"eval_loss": 0.5814855694770813,
"eval_runtime": 97.6377,
"eval_samples_per_second": 5.121,
"eval_steps_per_second": 0.645,
"step": 377
},
{
"epoch": 0.6123936816524909,
"grad_norm": 3.7298901081085205,
"learning_rate": 0.00010172251874357757,
"loss": 0.5582,
"step": 378
},
{
"epoch": 0.6237343053867963,
"grad_norm": 3.000365972518921,
"learning_rate": 9.658487462616794e-05,
"loss": 0.5606,
"step": 385
},
{
"epoch": 0.6318347509113001,
"eval_loss": 0.5722550749778748,
"eval_runtime": 108.1224,
"eval_samples_per_second": 4.624,
"eval_steps_per_second": 0.583,
"step": 390
},
{
"epoch": 0.6350749291211016,
"grad_norm": 3.4912805557250977,
"learning_rate": 9.151803131832493e-05,
"loss": 0.5679,
"step": 392
},
{
"epoch": 0.6464155528554071,
"grad_norm": 3.3884501457214355,
"learning_rate": 8.652870483177049e-05,
"loss": 0.5478,
"step": 399
},
{
"epoch": 0.6528959092750102,
"eval_loss": 0.5653434991836548,
"eval_runtime": 12.2742,
"eval_samples_per_second": 40.736,
"eval_steps_per_second": 5.133,
"step": 403
},
{
"epoch": 0.6577561765897124,
"grad_norm": 2.968888282775879,
"learning_rate": 8.162350843104291e-05,
"loss": 0.5543,
"step": 406
},
{
"epoch": 0.6690968003240179,
"grad_norm": 2.6131746768951416,
"learning_rate": 7.680894386773072e-05,
"loss": 0.5451,
"step": 413
},
{
"epoch": 0.6739570676387201,
"eval_loss": 0.5613173246383667,
"eval_runtime": 96.25,
"eval_samples_per_second": 5.195,
"eval_steps_per_second": 0.655,
"step": 416
},
{
"epoch": 0.6804374240583232,
"grad_norm": 3.4300389289855957,
"learning_rate": 7.209139276251558e-05,
"loss": 0.5393,
"step": 420
},
{
"epoch": 0.6917780477926286,
"grad_norm": 2.889118194580078,
"learning_rate": 6.747710814644589e-05,
"loss": 0.5362,
"step": 427
},
{
"epoch": 0.6950182260024301,
"eval_loss": 0.5555723905563354,
"eval_runtime": 105.9062,
"eval_samples_per_second": 4.721,
"eval_steps_per_second": 0.595,
"step": 429
},
{
"epoch": 0.703118671526934,
"grad_norm": 3.436547040939331,
"learning_rate": 6.297220617265435e-05,
"loss": 0.5238,
"step": 434
},
{
"epoch": 0.7144592952612394,
"grad_norm": 2.8543200492858887,
"learning_rate": 5.858265800950438e-05,
"loss": 0.5328,
"step": 441
},
{
"epoch": 0.7160793843661402,
"eval_loss": 0.5473873019218445,
"eval_runtime": 12.2208,
"eval_samples_per_second": 40.914,
"eval_steps_per_second": 5.155,
"step": 442
},
{
"epoch": 0.7257999189955447,
"grad_norm": 2.4065663814544678,
"learning_rate": 5.4314281925911634e-05,
"loss": 0.527,
"step": 448
},
{
"epoch": 0.7371405427298502,
"grad_norm": 2.7027878761291504,
"learning_rate": 5.0172735579330526e-05,
"loss": 0.5185,
"step": 455
},
{
"epoch": 0.7371405427298502,
"eval_loss": 0.5413048267364502,
"eval_runtime": 95.8607,
"eval_samples_per_second": 5.216,
"eval_steps_per_second": 0.657,
"step": 455
},
{
"epoch": 0.7484811664641555,
"grad_norm": 3.0348236560821533,
"learning_rate": 4.616350851662895e-05,
"loss": 0.5127,
"step": 462
},
{
"epoch": 0.7582017010935601,
"eval_loss": 0.5359441041946411,
"eval_runtime": 104.7029,
"eval_samples_per_second": 4.775,
"eval_steps_per_second": 0.602,
"step": 468
},
{
"epoch": 0.759821790198461,
"grad_norm": 3.366955280303955,
"learning_rate": 4.229191489779047e-05,
"loss": 0.5387,
"step": 469
},
{
"epoch": 0.7711624139327663,
"grad_norm": 3.5870349407196045,
"learning_rate": 3.8563086452088506e-05,
"loss": 0.5036,
"step": 476
},
{
"epoch": 0.7792628594572701,
"eval_loss": 0.5299127101898193,
"eval_runtime": 12.1891,
"eval_samples_per_second": 41.02,
"eval_steps_per_second": 5.169,
"step": 481
},
{
"epoch": 0.7825030376670717,
"grad_norm": 2.6109859943389893,
"learning_rate": 3.498196567606959e-05,
"loss": 0.5213,
"step": 483
},
{
"epoch": 0.7938436614013771,
"grad_norm": 3.2474308013916016,
"learning_rate": 3.1553299282360966e-05,
"loss": 0.4922,
"step": 490
},
{
"epoch": 0.8003240178209802,
"eval_loss": 0.5264901518821716,
"eval_runtime": 102.9216,
"eval_samples_per_second": 4.858,
"eval_steps_per_second": 0.612,
"step": 494
},
{
"epoch": 0.8051842851356824,
"grad_norm": 2.540072202682495,
"learning_rate": 2.828163190798644e-05,
"loss": 0.511,
"step": 497
},
{
"epoch": 0.8165249088699879,
"grad_norm": 3.375274419784546,
"learning_rate": 2.5171300090530106e-05,
"loss": 0.5246,
"step": 504
},
{
"epoch": 0.8213851761846902,
"eval_loss": 0.5218872427940369,
"eval_runtime": 104.3935,
"eval_samples_per_second": 4.79,
"eval_steps_per_second": 0.603,
"step": 507
},
{
"epoch": 0.8278655326042932,
"grad_norm": 2.5530292987823486,
"learning_rate": 2.2226426520131734e-05,
"loss": 0.5028,
"step": 511
},
{
"epoch": 0.8392061563385986,
"grad_norm": 2.5905301570892334,
"learning_rate": 1.9450914574933725e-05,
"loss": 0.5088,
"step": 518
},
{
"epoch": 0.8424463345484001,
"eval_loss": 0.517468273639679,
"eval_runtime": 12.2173,
"eval_samples_per_second": 40.925,
"eval_steps_per_second": 5.157,
"step": 520
},
{
"epoch": 0.850546780072904,
"grad_norm": 2.506103992462158,
"learning_rate": 1.6848443147221828e-05,
"loss": 0.4833,
"step": 525
},
{
"epoch": 0.8618874038072094,
"grad_norm": 2.6942226886749268,
"learning_rate": 1.4422461767118233e-05,
"loss": 0.4908,
"step": 532
},
{
"epoch": 0.8635074929121102,
"eval_loss": 0.5149521231651306,
"eval_runtime": 109.6256,
"eval_samples_per_second": 4.561,
"eval_steps_per_second": 0.575,
"step": 533
},
{
"epoch": 0.8732280275415147,
"grad_norm": 3.0847156047821045,
"learning_rate": 1.2176186030289936e-05,
"loss": 0.4903,
"step": 539
},
{
"epoch": 0.8845686512758202,
"grad_norm": 3.1814467906951904,
"learning_rate": 1.011259333573326e-05,
"loss": 0.5091,
"step": 546
},
{
"epoch": 0.8845686512758202,
"eval_loss": 0.5119792222976685,
"eval_runtime": 106.4399,
"eval_samples_per_second": 4.697,
"eval_steps_per_second": 0.592,
"step": 546
},
{
"epoch": 0.8959092750101255,
"grad_norm": 2.7887933254241943,
"learning_rate": 8.234418939283866e-06,
"loss": 0.4902,
"step": 553
},
{
"epoch": 0.9056298096395302,
"eval_loss": 0.5096355676651001,
"eval_runtime": 12.246,
"eval_samples_per_second": 40.83,
"eval_steps_per_second": 5.145,
"step": 559
},
{
"epoch": 0.907249898744431,
"grad_norm": 2.728210210800171,
"learning_rate": 6.544152328083152e-06,
"loss": 0.5004,
"step": 560
},
{
"epoch": 0.9185905224787363,
"grad_norm": 2.694986581802368,
"learning_rate": 5.044033920806933e-06,
"loss": 0.4865,
"step": 567
},
{
"epoch": 0.9266909680032401,
"eval_loss": 0.5082899332046509,
"eval_runtime": 108.3733,
"eval_samples_per_second": 4.614,
"eval_steps_per_second": 0.581,
"step": 572
},
{
"epoch": 0.9299311462130417,
"grad_norm": 2.781261920928955,
"learning_rate": 3.7360520980297514e-06,
"loss": 0.4896,
"step": 574
},
{
"epoch": 0.9412717699473471,
"grad_norm": 2.8047616481781006,
"learning_rate": 2.6219405666614402e-06,
"loss": 0.5007,
"step": 581
},
{
"epoch": 0.9477521263669502,
"eval_loss": 0.5072164535522461,
"eval_runtime": 109.1206,
"eval_samples_per_second": 4.582,
"eval_steps_per_second": 0.577,
"step": 585
},
{
"epoch": 0.9526123936816525,
"grad_norm": 2.2356209754943848,
"learning_rate": 1.7031760619491353e-06,
"loss": 0.4764,
"step": 588
},
{
"epoch": 0.9639530174159578,
"grad_norm": 2.597925901412964,
"learning_rate": 9.809763900905875e-07,
"loss": 0.5001,
"step": 595
},
{
"epoch": 0.9688132847306602,
"eval_loss": 0.5067591071128845,
"eval_runtime": 12.2032,
"eval_samples_per_second": 40.973,
"eval_steps_per_second": 5.163,
"step": 598
},
{
"epoch": 0.9752936411502633,
"grad_norm": 2.403350353240967,
"learning_rate": 4.562988140535073e-07,
"loss": 0.4896,
"step": 602
},
{
"epoch": 0.9866342648845686,
"grad_norm": 2.5980334281921387,
"learning_rate": 1.298387847403437e-07,
"loss": 0.4989,
"step": 609
},
{
"epoch": 0.9898744430943702,
"eval_loss": 0.506678581237793,
"eval_runtime": 104.7374,
"eval_samples_per_second": 4.774,
"eval_steps_per_second": 0.602,
"step": 611
}
],
"logging_steps": 7,
"max_steps": 617,
"num_input_tokens_seen": 0,
"num_train_epochs": 1,
"save_steps": 13,
"stateful_callbacks": {
"TrainerControl": {
"args": {
"should_epoch_stop": false,
"should_evaluate": false,
"should_log": false,
"should_save": true,
"should_training_stop": false
},
"attributes": {}
}
},
"total_flos": 7.394413366880502e+17,
"train_batch_size": 8,
"trial_name": null,
"trial_params": null
}