{
"best_metric": 0.6744288802146912,
"best_model_checkpoint": "models/llama-3.2-3b-sft-2/checkpoint-10000",
"epoch": 1.3452613170108294,
"eval_steps": 500,
"global_step": 10000,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.00013452613170108295,
"grad_norm": 5.089283466339111,
"learning_rate": 5e-09,
"loss": 1.0115,
"step": 1
},
{
"epoch": 0.006726306585054147,
"grad_norm": 3.6531453132629395,
"learning_rate": 2.5000000000000004e-07,
"loss": 1.0329,
"step": 50
},
{
"epoch": 0.013452613170108294,
"grad_norm": 1.954504132270813,
"learning_rate": 5.000000000000001e-07,
"loss": 0.9628,
"step": 100
},
{
"epoch": 0.02017891975516244,
"grad_norm": 1.6202716827392578,
"learning_rate": 7.5e-07,
"loss": 0.882,
"step": 150
},
{
"epoch": 0.02690522634021659,
"grad_norm": 1.5833114385604858,
"learning_rate": 1.0000000000000002e-06,
"loss": 0.8489,
"step": 200
},
{
"epoch": 0.033631532925270734,
"grad_norm": 1.6729789972305298,
"learning_rate": 1.25e-06,
"loss": 0.8464,
"step": 250
},
{
"epoch": 0.04035783951032488,
"grad_norm": 1.551665186882019,
"learning_rate": 1.5e-06,
"loss": 0.8205,
"step": 300
},
{
"epoch": 0.047084146095379024,
"grad_norm": 1.6382861137390137,
"learning_rate": 1.75e-06,
"loss": 0.7994,
"step": 350
},
{
"epoch": 0.05381045268043318,
"grad_norm": 1.8948196172714233,
"learning_rate": 2.0000000000000003e-06,
"loss": 0.7942,
"step": 400
},
{
"epoch": 0.06053675926548732,
"grad_norm": 1.5274723768234253,
"learning_rate": 2.25e-06,
"loss": 0.7843,
"step": 450
},
{
"epoch": 0.06726306585054147,
"grad_norm": 2.083868980407715,
"learning_rate": 2.5e-06,
"loss": 0.7792,
"step": 500
},
{
"epoch": 0.06726306585054147,
"eval_loss": 0.772584855556488,
"eval_runtime": 412.1529,
"eval_samples_per_second": 121.501,
"eval_steps_per_second": 1.9,
"step": 500
},
{
"epoch": 0.07398937243559561,
"grad_norm": 1.4476754665374756,
"learning_rate": 2.7500000000000004e-06,
"loss": 0.7598,
"step": 550
},
{
"epoch": 0.08071567902064976,
"grad_norm": 1.5177018642425537,
"learning_rate": 3e-06,
"loss": 0.7672,
"step": 600
},
{
"epoch": 0.0874419856057039,
"grad_norm": 1.5025829076766968,
"learning_rate": 3.2500000000000002e-06,
"loss": 0.7442,
"step": 650
},
{
"epoch": 0.09416829219075805,
"grad_norm": 1.6595162153244019,
"learning_rate": 3.5e-06,
"loss": 0.7562,
"step": 700
},
{
"epoch": 0.10089459877581221,
"grad_norm": 1.4664040803909302,
"learning_rate": 3.7500000000000005e-06,
"loss": 0.7485,
"step": 750
},
{
"epoch": 0.10762090536086635,
"grad_norm": 1.5404261350631714,
"learning_rate": 4.000000000000001e-06,
"loss": 0.7482,
"step": 800
},
{
"epoch": 0.1143472119459205,
"grad_norm": 1.553759217262268,
"learning_rate": 4.25e-06,
"loss": 0.7475,
"step": 850
},
{
"epoch": 0.12107351853097464,
"grad_norm": 1.5157835483551025,
"learning_rate": 4.5e-06,
"loss": 0.7403,
"step": 900
},
{
"epoch": 0.12779982511602878,
"grad_norm": 1.4276628494262695,
"learning_rate": 4.75e-06,
"loss": 0.7392,
"step": 950
},
{
"epoch": 0.13452613170108293,
"grad_norm": 1.5712127685546875,
"learning_rate": 5e-06,
"loss": 0.7496,
"step": 1000
},
{
"epoch": 0.13452613170108293,
"eval_loss": 0.7444220185279846,
"eval_runtime": 412.2013,
"eval_samples_per_second": 121.487,
"eval_steps_per_second": 1.9,
"step": 1000
},
{
"epoch": 0.1412524382861371,
"grad_norm": 1.3969111442565918,
"learning_rate": 4.9996192378909785e-06,
"loss": 0.7548,
"step": 1050
},
{
"epoch": 0.14797874487119123,
"grad_norm": 1.4484981298446655,
"learning_rate": 4.99847706754774e-06,
"loss": 0.73,
"step": 1100
},
{
"epoch": 0.15470505145624538,
"grad_norm": 1.4184221029281616,
"learning_rate": 4.9965738368864345e-06,
"loss": 0.734,
"step": 1150
},
{
"epoch": 0.16143135804129952,
"grad_norm": 1.4049128293991089,
"learning_rate": 4.993910125649561e-06,
"loss": 0.7295,
"step": 1200
},
{
"epoch": 0.16815766462635368,
"grad_norm": 1.3905514478683472,
"learning_rate": 4.990486745229364e-06,
"loss": 0.7405,
"step": 1250
},
{
"epoch": 0.1748839712114078,
"grad_norm": 1.5445802211761475,
"learning_rate": 4.986304738420684e-06,
"loss": 0.7198,
"step": 1300
},
{
"epoch": 0.18161027779646197,
"grad_norm": 1.4158939123153687,
"learning_rate": 4.981365379103306e-06,
"loss": 0.7355,
"step": 1350
},
{
"epoch": 0.1883365843815161,
"grad_norm": 1.3896397352218628,
"learning_rate": 4.975670171853926e-06,
"loss": 0.7168,
"step": 1400
},
{
"epoch": 0.19506289096657026,
"grad_norm": 1.5642178058624268,
"learning_rate": 4.9692208514878445e-06,
"loss": 0.7369,
"step": 1450
},
{
"epoch": 0.20178919755162442,
"grad_norm": 1.287514328956604,
"learning_rate": 4.962019382530521e-06,
"loss": 0.7243,
"step": 1500
},
{
"epoch": 0.20178919755162442,
"eval_loss": 0.7295696139335632,
"eval_runtime": 408.0925,
"eval_samples_per_second": 122.71,
"eval_steps_per_second": 1.919,
"step": 1500
},
{
"epoch": 0.20851550413667855,
"grad_norm": 1.380998969078064,
"learning_rate": 4.9540679586191605e-06,
"loss": 0.7309,
"step": 1550
},
{
"epoch": 0.2152418107217327,
"grad_norm": 1.4188984632492065,
"learning_rate": 4.9453690018345144e-06,
"loss": 0.7217,
"step": 1600
},
{
"epoch": 0.22196811730678684,
"grad_norm": 1.306766390800476,
"learning_rate": 4.935925161963089e-06,
"loss": 0.7336,
"step": 1650
},
{
"epoch": 0.228694423891841,
"grad_norm": 1.460458517074585,
"learning_rate": 4.925739315689991e-06,
"loss": 0.7258,
"step": 1700
},
{
"epoch": 0.23542073047689513,
"grad_norm": 1.4665378332138062,
"learning_rate": 4.914814565722671e-06,
"loss": 0.7249,
"step": 1750
},
{
"epoch": 0.2421470370619493,
"grad_norm": 1.5657004117965698,
"learning_rate": 4.903154239845798e-06,
"loss": 0.7253,
"step": 1800
},
{
"epoch": 0.24887334364700342,
"grad_norm": 1.3820271492004395,
"learning_rate": 4.890761889907589e-06,
"loss": 0.7151,
"step": 1850
},
{
"epoch": 0.25559965023205755,
"grad_norm": 1.4721177816390991,
"learning_rate": 4.8776412907378845e-06,
"loss": 0.7251,
"step": 1900
},
{
"epoch": 0.26232595681711174,
"grad_norm": 1.4119197130203247,
"learning_rate": 4.863796438998293e-06,
"loss": 0.7188,
"step": 1950
},
{
"epoch": 0.26905226340216587,
"grad_norm": 1.3871151208877563,
"learning_rate": 4.849231551964771e-06,
"loss": 0.7178,
"step": 2000
},
{
"epoch": 0.26905226340216587,
"eval_loss": 0.7197288870811462,
"eval_runtime": 407.583,
"eval_samples_per_second": 122.863,
"eval_steps_per_second": 1.921,
"step": 2000
},
{
"epoch": 0.27577856998722,
"grad_norm": 1.4082683324813843,
"learning_rate": 4.833951066243004e-06,
"loss": 0.7255,
"step": 2050
},
{
"epoch": 0.2825048765722742,
"grad_norm": 1.3477662801742554,
"learning_rate": 4.817959636416969e-06,
"loss": 0.7141,
"step": 2100
},
{
"epoch": 0.2892311831573283,
"grad_norm": 1.316319227218628,
"learning_rate": 4.801262133631101e-06,
"loss": 0.7146,
"step": 2150
},
{
"epoch": 0.29595748974238245,
"grad_norm": 1.3264553546905518,
"learning_rate": 4.783863644106502e-06,
"loss": 0.7176,
"step": 2200
},
{
"epoch": 0.3026837963274366,
"grad_norm": 1.4020483493804932,
"learning_rate": 4.765769467591626e-06,
"loss": 0.7082,
"step": 2250
},
{
"epoch": 0.30941010291249077,
"grad_norm": 1.4155000448226929,
"learning_rate": 4.746985115747918e-06,
"loss": 0.7133,
"step": 2300
},
{
"epoch": 0.3161364094975449,
"grad_norm": 1.3636876344680786,
"learning_rate": 4.72751631047092e-06,
"loss": 0.721,
"step": 2350
},
{
"epoch": 0.32286271608259903,
"grad_norm": 1.3558250665664673,
"learning_rate": 4.707368982147318e-06,
"loss": 0.7056,
"step": 2400
},
{
"epoch": 0.32958902266765316,
"grad_norm": 1.3763400316238403,
"learning_rate": 4.68654926784849e-06,
"loss": 0.7157,
"step": 2450
},
{
"epoch": 0.33631532925270735,
"grad_norm": 1.4673148393630981,
"learning_rate": 4.665063509461098e-06,
"loss": 0.7077,
"step": 2500
},
{
"epoch": 0.33631532925270735,
"eval_loss": 0.7126947045326233,
"eval_runtime": 407.7534,
"eval_samples_per_second": 122.812,
"eval_steps_per_second": 1.92,
"step": 2500
},
{
"epoch": 0.3430416358377615,
"grad_norm": 1.5427167415618896,
"learning_rate": 4.642918251755281e-06,
"loss": 0.7216,
"step": 2550
},
{
"epoch": 0.3497679424228156,
"grad_norm": 1.4927399158477783,
"learning_rate": 4.620120240391065e-06,
"loss": 0.7096,
"step": 2600
},
{
"epoch": 0.3564942490078698,
"grad_norm": 1.3618338108062744,
"learning_rate": 4.596676419863561e-06,
"loss": 0.7106,
"step": 2650
},
{
"epoch": 0.36322055559292393,
"grad_norm": 1.3076339960098267,
"learning_rate": 4.572593931387604e-06,
"loss": 0.7213,
"step": 2700
},
{
"epoch": 0.36994686217797806,
"grad_norm": 1.2195158004760742,
"learning_rate": 4.54788011072248e-06,
"loss": 0.7102,
"step": 2750
},
{
"epoch": 0.3766731687630322,
"grad_norm": 1.341899037361145,
"learning_rate": 4.522542485937369e-06,
"loss": 0.7021,
"step": 2800
},
{
"epoch": 0.3833994753480864,
"grad_norm": 1.4170920848846436,
"learning_rate": 4.496588775118232e-06,
"loss": 0.6942,
"step": 2850
},
{
"epoch": 0.3901257819331405,
"grad_norm": 1.502806305885315,
"learning_rate": 4.470026884016805e-06,
"loss": 0.6947,
"step": 2900
},
{
"epoch": 0.39685208851819465,
"grad_norm": 1.3311142921447754,
"learning_rate": 4.442864903642428e-06,
"loss": 0.6926,
"step": 2950
},
{
"epoch": 0.40357839510324883,
"grad_norm": 1.3881827592849731,
"learning_rate": 4.415111107797445e-06,
"loss": 0.6992,
"step": 3000
},
{
"epoch": 0.40357839510324883,
"eval_loss": 0.7066433429718018,
"eval_runtime": 407.4989,
"eval_samples_per_second": 122.889,
"eval_steps_per_second": 1.921,
"step": 3000
},
{
"epoch": 0.41030470168830296,
"grad_norm": 1.4951732158660889,
"learning_rate": 4.386773950556931e-06,
"loss": 0.6967,
"step": 3050
},
{
"epoch": 0.4170310082733571,
"grad_norm": 1.3532462120056152,
"learning_rate": 4.357862063693486e-06,
"loss": 0.7056,
"step": 3100
},
{
"epoch": 0.4237573148584112,
"grad_norm": 1.3927810192108154,
"learning_rate": 4.328384254047927e-06,
"loss": 0.6962,
"step": 3150
},
{
"epoch": 0.4304836214434654,
"grad_norm": 1.2891749143600464,
"learning_rate": 4.2983495008466285e-06,
"loss": 0.7193,
"step": 3200
},
{
"epoch": 0.43720992802851955,
"grad_norm": 1.3432807922363281,
"learning_rate": 4.267766952966369e-06,
"loss": 0.7008,
"step": 3250
},
{
"epoch": 0.4439362346135737,
"grad_norm": 1.5565185546875,
"learning_rate": 4.236645926147493e-06,
"loss": 0.6939,
"step": 3300
},
{
"epoch": 0.4506625411986278,
"grad_norm": 1.3859912157058716,
"learning_rate": 4.204995900156247e-06,
"loss": 0.6953,
"step": 3350
},
{
"epoch": 0.457388847783682,
"grad_norm": 1.358071208000183,
"learning_rate": 4.172826515897146e-06,
"loss": 0.6847,
"step": 3400
},
{
"epoch": 0.4641151543687361,
"grad_norm": 1.3290272951126099,
"learning_rate": 4.140147572476269e-06,
"loss": 0.7069,
"step": 3450
},
{
"epoch": 0.47084146095379026,
"grad_norm": 1.387556552886963,
"learning_rate": 4.106969024216348e-06,
"loss": 0.6992,
"step": 3500
},
{
"epoch": 0.47084146095379026,
"eval_loss": 0.7012136578559875,
"eval_runtime": 407.5913,
"eval_samples_per_second": 122.861,
"eval_steps_per_second": 1.921,
"step": 3500
},
{
"epoch": 0.47756776753884445,
"grad_norm": 1.4065874814987183,
"learning_rate": 4.073300977624594e-06,
"loss": 0.7044,
"step": 3550
},
{
"epoch": 0.4842940741238986,
"grad_norm": 1.3837950229644775,
"learning_rate": 4.039153688314146e-06,
"loss": 0.7002,
"step": 3600
},
{
"epoch": 0.4910203807089527,
"grad_norm": 1.345460057258606,
"learning_rate": 4.0045375578801216e-06,
"loss": 0.692,
"step": 3650
},
{
"epoch": 0.49774668729400684,
"grad_norm": 1.3991773128509521,
"learning_rate": 3.969463130731183e-06,
"loss": 0.691,
"step": 3700
},
{
"epoch": 0.504472993879061,
"grad_norm": 1.4494290351867676,
"learning_rate": 3.933941090877615e-06,
"loss": 0.6975,
"step": 3750
},
{
"epoch": 0.5111993004641151,
"grad_norm": 1.4690207242965698,
"learning_rate": 3.897982258676867e-06,
"loss": 0.6951,
"step": 3800
},
{
"epoch": 0.5179256070491693,
"grad_norm": 1.2572777271270752,
"learning_rate": 3.861597587537568e-06,
"loss": 0.6997,
"step": 3850
},
{
"epoch": 0.5246519136342235,
"grad_norm": 1.416960597038269,
"learning_rate": 3.824798160583012e-06,
"loss": 0.6883,
"step": 3900
},
{
"epoch": 0.5313782202192776,
"grad_norm": 1.3925755023956299,
"learning_rate": 3.787595187275136e-06,
"loss": 0.6894,
"step": 3950
},
{
"epoch": 0.5381045268043317,
"grad_norm": 1.2715755701065063,
"learning_rate": 3.7500000000000005e-06,
"loss": 0.6945,
"step": 4000
},
{
"epoch": 0.5381045268043317,
"eval_loss": 0.6965380907058716,
"eval_runtime": 407.4593,
"eval_samples_per_second": 122.901,
"eval_steps_per_second": 1.922,
"step": 4000
},
{
"epoch": 0.5448308333893859,
"grad_norm": 1.1935741901397705,
"learning_rate": 3.7120240506158433e-06,
"loss": 0.6899,
"step": 4050
},
{
"epoch": 0.55155713997444,
"grad_norm": 1.3253408670425415,
"learning_rate": 3.6736789069647273e-06,
"loss": 0.6964,
"step": 4100
},
{
"epoch": 0.5582834465594941,
"grad_norm": 1.4265397787094116,
"learning_rate": 3.634976249348867e-06,
"loss": 0.6867,
"step": 4150
},
{
"epoch": 0.5650097531445484,
"grad_norm": 1.3048572540283203,
"learning_rate": 3.595927866972694e-06,
"loss": 0.6959,
"step": 4200
},
{
"epoch": 0.5717360597296025,
"grad_norm": 1.495254397392273,
"learning_rate": 3.556545654351749e-06,
"loss": 0.6841,
"step": 4250
},
{
"epoch": 0.5784623663146566,
"grad_norm": 1.3568543195724487,
"learning_rate": 3.516841607689501e-06,
"loss": 0.6911,
"step": 4300
},
{
"epoch": 0.5851886728997108,
"grad_norm": 1.274553656578064,
"learning_rate": 3.476827821223184e-06,
"loss": 0.6819,
"step": 4350
},
{
"epoch": 0.5919149794847649,
"grad_norm": 1.3335059881210327,
"learning_rate": 3.436516483539781e-06,
"loss": 0.7018,
"step": 4400
},
{
"epoch": 0.598641286069819,
"grad_norm": 1.265262246131897,
"learning_rate": 3.39591987386325e-06,
"loss": 0.6821,
"step": 4450
},
{
"epoch": 0.6053675926548732,
"grad_norm": 1.3332948684692383,
"learning_rate": 3.3550503583141726e-06,
"loss": 0.6879,
"step": 4500
},
{
"epoch": 0.6053675926548732,
"eval_loss": 0.6919859647750854,
"eval_runtime": 407.9173,
"eval_samples_per_second": 122.763,
"eval_steps_per_second": 1.92,
"step": 4500
},
{
"epoch": 0.6120938992399274,
"grad_norm": 1.310651421546936,
"learning_rate": 3.313920386142892e-06,
"loss": 0.6857,
"step": 4550
},
{
"epoch": 0.6188202058249815,
"grad_norm": 1.2892873287200928,
"learning_rate": 3.272542485937369e-06,
"loss": 0.6812,
"step": 4600
},
{
"epoch": 0.6255465124100357,
"grad_norm": 1.2226920127868652,
"learning_rate": 3.230929261806842e-06,
"loss": 0.6774,
"step": 4650
},
{
"epoch": 0.6322728189950898,
"grad_norm": 1.2619507312774658,
"learning_rate": 3.189093389542498e-06,
"loss": 0.6832,
"step": 4700
},
{
"epoch": 0.6389991255801439,
"grad_norm": 1.4155713319778442,
"learning_rate": 3.147047612756302e-06,
"loss": 0.6897,
"step": 4750
},
{
"epoch": 0.6457254321651981,
"grad_norm": 1.3170417547225952,
"learning_rate": 3.1048047389991693e-06,
"loss": 0.6962,
"step": 4800
},
{
"epoch": 0.6524517387502522,
"grad_norm": 1.3795112371444702,
"learning_rate": 3.062377635859663e-06,
"loss": 0.6838,
"step": 4850
},
{
"epoch": 0.6591780453353063,
"grad_norm": 1.3435876369476318,
"learning_rate": 3.019779227044398e-06,
"loss": 0.6813,
"step": 4900
},
{
"epoch": 0.6659043519203606,
"grad_norm": 1.4538263082504272,
"learning_rate": 2.9770224884413625e-06,
"loss": 0.6941,
"step": 4950
},
{
"epoch": 0.6726306585054147,
"grad_norm": 1.4418572187423706,
"learning_rate": 2.9341204441673267e-06,
"loss": 0.6901,
"step": 5000
},
{
"epoch": 0.6726306585054147,
"eval_loss": 0.6878845691680908,
"eval_runtime": 407.8759,
"eval_samples_per_second": 122.775,
"eval_steps_per_second": 1.92,
"step": 5000
},
{
"epoch": 0.6793569650904688,
"grad_norm": 1.3608489036560059,
"learning_rate": 2.8910861626005774e-06,
"loss": 0.6822,
"step": 5050
},
{
"epoch": 0.686083271675523,
"grad_norm": 1.2994650602340698,
"learning_rate": 2.847932752400164e-06,
"loss": 0.6762,
"step": 5100
},
{
"epoch": 0.6928095782605771,
"grad_norm": 1.565314531326294,
"learning_rate": 2.804673358512869e-06,
"loss": 0.686,
"step": 5150
},
{
"epoch": 0.6995358848456312,
"grad_norm": 1.3825584650039673,
"learning_rate": 2.761321158169134e-06,
"loss": 0.6872,
"step": 5200
},
{
"epoch": 0.7062621914306854,
"grad_norm": 1.3237903118133545,
"learning_rate": 2.717889356869146e-06,
"loss": 0.6873,
"step": 5250
},
{
"epoch": 0.7129884980157396,
"grad_norm": 1.4130380153656006,
"learning_rate": 2.6743911843603134e-06,
"loss": 0.6849,
"step": 5300
},
{
"epoch": 0.7197148046007937,
"grad_norm": 1.3584030866622925,
"learning_rate": 2.6308398906073603e-06,
"loss": 0.6914,
"step": 5350
},
{
"epoch": 0.7264411111858479,
"grad_norm": 1.4434551000595093,
"learning_rate": 2.587248741756253e-06,
"loss": 0.6815,
"step": 5400
},
{
"epoch": 0.733167417770902,
"grad_norm": 1.343894362449646,
"learning_rate": 2.543631016093209e-06,
"loss": 0.6739,
"step": 5450
},
{
"epoch": 0.7398937243559561,
"grad_norm": 1.5112074613571167,
"learning_rate": 2.5e-06,
"loss": 0.6759,
"step": 5500
},
{
"epoch": 0.7398937243559561,
"eval_loss": 0.6843944787979126,
"eval_runtime": 407.3845,
"eval_samples_per_second": 122.923,
"eval_steps_per_second": 1.922,
"step": 5500
},
{
"epoch": 0.7466200309410103,
"grad_norm": 1.471646785736084,
"learning_rate": 2.4563689839067913e-06,
"loss": 0.6914,
"step": 5550
},
{
"epoch": 0.7533463375260644,
"grad_norm": 1.3042908906936646,
"learning_rate": 2.4127512582437486e-06,
"loss": 0.6813,
"step": 5600
},
{
"epoch": 0.7600726441111186,
"grad_norm": 1.2429543733596802,
"learning_rate": 2.3691601093926406e-06,
"loss": 0.6757,
"step": 5650
},
{
"epoch": 0.7667989506961728,
"grad_norm": 1.300581932067871,
"learning_rate": 2.325608815639687e-06,
"loss": 0.6861,
"step": 5700
},
{
"epoch": 0.7735252572812269,
"grad_norm": 1.3551619052886963,
"learning_rate": 2.2821106431308546e-06,
"loss": 0.677,
"step": 5750
},
{
"epoch": 0.780251563866281,
"grad_norm": 1.2355047464370728,
"learning_rate": 2.238678841830867e-06,
"loss": 0.6728,
"step": 5800
},
{
"epoch": 0.7869778704513352,
"grad_norm": 1.3812388181686401,
"learning_rate": 2.195326641487132e-06,
"loss": 0.6807,
"step": 5850
},
{
"epoch": 0.7937041770363893,
"grad_norm": 1.3730298280715942,
"learning_rate": 2.1520672475998374e-06,
"loss": 0.6835,
"step": 5900
},
{
"epoch": 0.8004304836214434,
"grad_norm": 1.3047642707824707,
"learning_rate": 2.1089138373994226e-06,
"loss": 0.695,
"step": 5950
},
{
"epoch": 0.8071567902064977,
"grad_norm": 1.2433747053146362,
"learning_rate": 2.0658795558326745e-06,
"loss": 0.6752,
"step": 6000
},
{
"epoch": 0.8071567902064977,
"eval_loss": 0.681221604347229,
"eval_runtime": 407.827,
"eval_samples_per_second": 122.79,
"eval_steps_per_second": 1.92,
"step": 6000
},
{
"epoch": 0.8138830967915518,
"grad_norm": 1.245276927947998,
"learning_rate": 2.022977511558638e-06,
"loss": 0.6822,
"step": 6050
},
{
"epoch": 0.8206094033766059,
"grad_norm": 1.306039810180664,
"learning_rate": 1.9802207729556023e-06,
"loss": 0.6902,
"step": 6100
},
{
"epoch": 0.8273357099616601,
"grad_norm": 1.3029533624649048,
"learning_rate": 1.937622364140338e-06,
"loss": 0.69,
"step": 6150
},
{
"epoch": 0.8340620165467142,
"grad_norm": 1.2798274755477905,
"learning_rate": 1.895195261000831e-06,
"loss": 0.6791,
"step": 6200
},
{
"epoch": 0.8407883231317683,
"grad_norm": 1.3335996866226196,
"learning_rate": 1.852952387243698e-06,
"loss": 0.6807,
"step": 6250
},
{
"epoch": 0.8475146297168225,
"grad_norm": 1.3937443494796753,
"learning_rate": 1.8109066104575023e-06,
"loss": 0.6817,
"step": 6300
},
{
"epoch": 0.8542409363018766,
"grad_norm": 1.3417977094650269,
"learning_rate": 1.7690707381931585e-06,
"loss": 0.6824,
"step": 6350
},
{
"epoch": 0.8609672428869308,
"grad_norm": 1.2701561450958252,
"learning_rate": 1.7274575140626318e-06,
"loss": 0.674,
"step": 6400
},
{
"epoch": 0.867693549471985,
"grad_norm": 1.2082581520080566,
"learning_rate": 1.686079613857109e-06,
"loss": 0.6725,
"step": 6450
},
{
"epoch": 0.8744198560570391,
"grad_norm": 1.4414860010147095,
"learning_rate": 1.6449496416858285e-06,
"loss": 0.6826,
"step": 6500
},
{
"epoch": 0.8744198560570391,
"eval_loss": 0.6783373951911926,
"eval_runtime": 407.3438,
"eval_samples_per_second": 122.935,
"eval_steps_per_second": 1.922,
"step": 6500
},
{
"epoch": 0.8811461626420932,
"grad_norm": 1.2324413061141968,
"learning_rate": 1.6040801261367494e-06,
"loss": 0.6682,
"step": 6550
},
{
"epoch": 0.8878724692271474,
"grad_norm": 1.2953166961669922,
"learning_rate": 1.56348351646022e-06,
"loss": 0.6811,
"step": 6600
},
{
"epoch": 0.8945987758122015,
"grad_norm": 1.2646501064300537,
"learning_rate": 1.5231721787768162e-06,
"loss": 0.6813,
"step": 6650
},
{
"epoch": 0.9013250823972556,
"grad_norm": 1.4454089403152466,
"learning_rate": 1.4831583923105e-06,
"loss": 0.6744,
"step": 6700
},
{
"epoch": 0.9080513889823099,
"grad_norm": 1.3868876695632935,
"learning_rate": 1.443454345648252e-06,
"loss": 0.6764,
"step": 6750
},
{
"epoch": 0.914777695567364,
"grad_norm": 1.2375882863998413,
"learning_rate": 1.4040721330273063e-06,
"loss": 0.678,
"step": 6800
},
{
"epoch": 0.9215040021524181,
"grad_norm": 1.3029659986495972,
"learning_rate": 1.3650237506511333e-06,
"loss": 0.6736,
"step": 6850
},
{
"epoch": 0.9282303087374723,
"grad_norm": 1.3237438201904297,
"learning_rate": 1.3263210930352737e-06,
"loss": 0.6769,
"step": 6900
},
{
"epoch": 0.9349566153225264,
"grad_norm": 1.3647958040237427,
"learning_rate": 1.2879759493841577e-06,
"loss": 0.6663,
"step": 6950
},
{
"epoch": 0.9416829219075805,
"grad_norm": 1.2964695692062378,
"learning_rate": 1.2500000000000007e-06,
"loss": 0.6804,
"step": 7000
},
{
"epoch": 0.9416829219075805,
"eval_loss": 0.6757956147193909,
"eval_runtime": 407.4829,
"eval_samples_per_second": 122.894,
"eval_steps_per_second": 1.922,
"step": 7000
},
{
"epoch": 0.9484092284926346,
"grad_norm": 1.444977045059204,
"learning_rate": 1.2124048127248644e-06,
"loss": 0.6816,
"step": 7050
},
{
"epoch": 0.9551355350776889,
"grad_norm": 1.262220025062561,
"learning_rate": 1.1752018394169882e-06,
"loss": 0.6687,
"step": 7100
},
{
"epoch": 0.961861841662743,
"grad_norm": 1.2450040578842163,
"learning_rate": 1.1384024124624324e-06,
"loss": 0.6738,
"step": 7150
},
{
"epoch": 0.9685881482477972,
"grad_norm": 1.31246817111969,
"learning_rate": 1.1020177413231334e-06,
"loss": 0.6664,
"step": 7200
},
{
"epoch": 0.9753144548328513,
"grad_norm": 1.307934284210205,
"learning_rate": 1.0660589091223854e-06,
"loss": 0.6845,
"step": 7250
},
{
"epoch": 0.9820407614179054,
"grad_norm": 1.2466742992401123,
"learning_rate": 1.0305368692688175e-06,
"loss": 0.6732,
"step": 7300
},
{
"epoch": 0.9887670680029595,
"grad_norm": 1.3563283681869507,
"learning_rate": 9.95462442119879e-07,
"loss": 0.6683,
"step": 7350
},
{
"epoch": 0.9954933745880137,
"grad_norm": 1.344889760017395,
"learning_rate": 9.608463116858544e-07,
"loss": 0.6794,
"step": 7400
},
{
"epoch": 1.002219681173068,
"grad_norm": 1.2191548347473145,
"learning_rate": 9.266990223754069e-07,
"loss": 0.6551,
"step": 7450
},
{
"epoch": 1.008945987758122,
"grad_norm": 1.4573768377304077,
"learning_rate": 8.930309757836517e-07,
"loss": 0.6131,
"step": 7500
},
{
"epoch": 1.008945987758122,
"eval_loss": 0.6763768196105957,
"eval_runtime": 407.3062,
"eval_samples_per_second": 122.947,
"eval_steps_per_second": 1.922,
"step": 7500
},
{
"epoch": 1.0156722943431762,
"grad_norm": 1.4787367582321167,
"learning_rate": 8.598524275237321e-07,
"loss": 0.5996,
"step": 7550
},
{
"epoch": 1.0223986009282302,
"grad_norm": 1.2250146865844727,
"learning_rate": 8.271734841028553e-07,
"loss": 0.6111,
"step": 7600
},
{
"epoch": 1.0291249075132844,
"grad_norm": 1.2959485054016113,
"learning_rate": 7.950040998437541e-07,
"loss": 0.5994,
"step": 7650
},
{
"epoch": 1.0358512140983387,
"grad_norm": 1.3388257026672363,
"learning_rate": 7.633540738525066e-07,
"loss": 0.6092,
"step": 7700
},
{
"epoch": 1.0425775206833927,
"grad_norm": 1.3289120197296143,
"learning_rate": 7.322330470336314e-07,
"loss": 0.6154,
"step": 7750
},
{
"epoch": 1.049303827268447,
"grad_norm": 1.4141819477081299,
"learning_rate": 7.016504991533727e-07,
"loss": 0.6132,
"step": 7800
},
{
"epoch": 1.056030133853501,
"grad_norm": 1.368107557296753,
"learning_rate": 6.716157459520739e-07,
"loss": 0.6147,
"step": 7850
},
{
"epoch": 1.0627564404385552,
"grad_norm": 1.3877215385437012,
"learning_rate": 6.421379363065142e-07,
"loss": 0.6234,
"step": 7900
},
{
"epoch": 1.0694827470236092,
"grad_norm": 1.3689961433410645,
"learning_rate": 6.1322604944307e-07,
"loss": 0.6041,
"step": 7950
},
{
"epoch": 1.0762090536086635,
"grad_norm": 1.4440808296203613,
"learning_rate": 5.848888922025553e-07,
"loss": 0.6012,
"step": 8000
},
{
"epoch": 1.0762090536086635,
"eval_loss": 0.6758261322975159,
"eval_runtime": 407.4199,
"eval_samples_per_second": 122.912,
"eval_steps_per_second": 1.922,
"step": 8000
},
{
"epoch": 1.0829353601937177,
"grad_norm": 1.3043277263641357,
"learning_rate": 5.571350963575728e-07,
"loss": 0.6173,
"step": 8050
},
{
"epoch": 1.0896616667787717,
"grad_norm": 1.3491568565368652,
"learning_rate": 5.299731159831953e-07,
"loss": 0.6101,
"step": 8100
},
{
"epoch": 1.096387973363826,
"grad_norm": 1.3337068557739258,
"learning_rate": 5.034112248817685e-07,
"loss": 0.6157,
"step": 8150
},
{
"epoch": 1.10311427994888,
"grad_norm": 1.4220951795578003,
"learning_rate": 4.774575140626317e-07,
"loss": 0.613,
"step": 8200
},
{
"epoch": 1.1098405865339342,
"grad_norm": 1.3012259006500244,
"learning_rate": 4.5211988927752026e-07,
"loss": 0.6025,
"step": 8250
},
{
"epoch": 1.1165668931189883,
"grad_norm": 1.3978270292282104,
"learning_rate": 4.27406068612396e-07,
"loss": 0.6086,
"step": 8300
},
{
"epoch": 1.1232931997040425,
"grad_norm": 1.381862998008728,
"learning_rate": 4.033235801364402e-07,
"loss": 0.6044,
"step": 8350
},
{
"epoch": 1.1300195062890968,
"grad_norm": 1.3832786083221436,
"learning_rate": 3.798797596089351e-07,
"loss": 0.6117,
"step": 8400
},
{
"epoch": 1.1367458128741508,
"grad_norm": 1.3328065872192383,
"learning_rate": 3.5708174824471947e-07,
"loss": 0.6021,
"step": 8450
},
{
"epoch": 1.143472119459205,
"grad_norm": 1.5573444366455078,
"learning_rate": 3.3493649053890325e-07,
"loss": 0.6136,
"step": 8500
},
{
"epoch": 1.143472119459205,
"eval_loss": 0.675076425075531,
"eval_runtime": 407.3808,
"eval_samples_per_second": 122.924,
"eval_steps_per_second": 1.922,
"step": 8500
},
{
"epoch": 1.150198426044259,
"grad_norm": 1.3546710014343262,
"learning_rate": 3.134507321515107e-07,
"loss": 0.6107,
"step": 8550
},
{
"epoch": 1.1569247326293133,
"grad_norm": 1.2918901443481445,
"learning_rate": 2.9263101785268253e-07,
"loss": 0.6085,
"step": 8600
},
{
"epoch": 1.1636510392143673,
"grad_norm": 1.3548731803894043,
"learning_rate": 2.7248368952908055e-07,
"loss": 0.6033,
"step": 8650
},
{
"epoch": 1.1703773457994215,
"grad_norm": 1.4083459377288818,
"learning_rate": 2.53014884252083e-07,
"loss": 0.623,
"step": 8700
},
{
"epoch": 1.1771036523844756,
"grad_norm": 1.3091658353805542,
"learning_rate": 2.3423053240837518e-07,
"loss": 0.6088,
"step": 8750
},
{
"epoch": 1.1838299589695298,
"grad_norm": 1.394978404045105,
"learning_rate": 2.1613635589349756e-07,
"loss": 0.6069,
"step": 8800
},
{
"epoch": 1.190556265554584,
"grad_norm": 1.3812838792800903,
"learning_rate": 1.9873786636889908e-07,
"loss": 0.6087,
"step": 8850
},
{
"epoch": 1.197282572139638,
"grad_norm": 1.3317595720291138,
"learning_rate": 1.8204036358303173e-07,
"loss": 0.6059,
"step": 8900
},
{
"epoch": 1.2040088787246923,
"grad_norm": 1.2979373931884766,
"learning_rate": 1.6604893375699594e-07,
"loss": 0.5978,
"step": 8950
},
{
"epoch": 1.2107351853097463,
"grad_norm": 1.427206039428711,
"learning_rate": 1.507684480352292e-07,
"loss": 0.6127,
"step": 9000
},
{
"epoch": 1.2107351853097463,
"eval_loss": 0.6747262477874756,
"eval_runtime": 407.1101,
"eval_samples_per_second": 123.006,
"eval_steps_per_second": 1.923,
"step": 9000
},
{
"epoch": 1.2174614918948006,
"grad_norm": 1.2737078666687012,
"learning_rate": 1.362035610017079e-07,
"loss": 0.6171,
"step": 9050
},
{
"epoch": 1.2241877984798548,
"grad_norm": 1.4532371759414673,
"learning_rate": 1.223587092621162e-07,
"loss": 0.6093,
"step": 9100
},
{
"epoch": 1.2309141050649088,
"grad_norm": 1.4686241149902344,
"learning_rate": 1.0923811009241142e-07,
"loss": 0.6073,
"step": 9150
},
{
"epoch": 1.237640411649963,
"grad_norm": 1.3653959035873413,
"learning_rate": 9.684576015420277e-08,
"loss": 0.6131,
"step": 9200
},
{
"epoch": 1.244366718235017,
"grad_norm": 1.3024178743362427,
"learning_rate": 8.518543427732951e-08,
"loss": 0.6162,
"step": 9250
},
{
"epoch": 1.2510930248200713,
"grad_norm": 1.3427560329437256,
"learning_rate": 7.426068431000883e-08,
"loss": 0.6033,
"step": 9300
},
{
"epoch": 1.2578193314051254,
"grad_norm": 1.4388611316680908,
"learning_rate": 6.407483803691216e-08,
"loss": 0.6104,
"step": 9350
},
{
"epoch": 1.2645456379901796,
"grad_norm": 1.610689401626587,
"learning_rate": 5.463099816548578e-08,
"loss": 0.612,
"step": 9400
},
{
"epoch": 1.2712719445752336,
"grad_norm": 1.4857087135314941,
"learning_rate": 4.593204138084006e-08,
"loss": 0.6069,
"step": 9450
},
{
"epoch": 1.2779982511602879,
"grad_norm": 1.3649048805236816,
"learning_rate": 3.798061746947995e-08,
"loss": 0.6076,
"step": 9500
},
{
"epoch": 1.2779982511602879,
"eval_loss": 0.6744668483734131,
"eval_runtime": 407.221,
"eval_samples_per_second": 122.973,
"eval_steps_per_second": 1.923,
"step": 9500
},
{
"epoch": 1.284724557745342,
"grad_norm": 1.2920268774032593,
"learning_rate": 3.077914851215585e-08,
"loss": 0.6195,
"step": 9550
},
{
"epoch": 1.2914508643303961,
"grad_norm": 1.3456374406814575,
"learning_rate": 2.4329828146074096e-08,
"loss": 0.5986,
"step": 9600
},
{
"epoch": 1.2981771709154504,
"grad_norm": 1.340410828590393,
"learning_rate": 1.8634620896695044e-08,
"loss": 0.6146,
"step": 9650
},
{
"epoch": 1.3049034775005044,
"grad_norm": 1.4664082527160645,
"learning_rate": 1.3695261579316776e-08,
"loss": 0.6148,
"step": 9700
},
{
"epoch": 1.3116297840855586,
"grad_norm": 1.3726171255111694,
"learning_rate": 9.513254770636138e-09,
"loss": 0.6112,
"step": 9750
},
{
"epoch": 1.3183560906706129,
"grad_norm": 1.3762577772140503,
"learning_rate": 6.089874350439507e-09,
"loss": 0.6066,
"step": 9800
},
{
"epoch": 1.325082397255667,
"grad_norm": 1.3914607763290405,
"learning_rate": 3.4261631135654174e-09,
"loss": 0.6114,
"step": 9850
},
{
"epoch": 1.3318087038407211,
"grad_norm": 1.2674617767333984,
"learning_rate": 1.5229324522605949e-09,
"loss": 0.6165,
"step": 9900
},
{
"epoch": 1.3385350104257752,
"grad_norm": 1.2815463542938232,
"learning_rate": 3.8076210902182607e-10,
"loss": 0.6101,
"step": 9950
},
{
"epoch": 1.3452613170108294,
"grad_norm": 1.4005900621414185,
"learning_rate": 0.0,
"loss": 0.6033,
"step": 10000
},
{
"epoch": 1.3452613170108294,
"eval_loss": 0.6744288802146912,
"eval_runtime": 407.1493,
"eval_samples_per_second": 122.994,
"eval_steps_per_second": 1.923,
"step": 10000
},
{
"epoch": 1.3452613170108294,
"step": 10000,
"total_flos": 1119551134941184.0,
"train_loss": 0.6844510423541069,
"train_runtime": 79339.5992,
"train_samples_per_second": 16.133,
"train_steps_per_second": 0.126
}
],
"logging_steps": 50,
"max_steps": 10000,
"num_input_tokens_seen": 0,
"num_train_epochs": 2,
"save_steps": 500,
"stateful_callbacks": {
"TrainerControl": {
"args": {
"should_epoch_stop": false,
"should_evaluate": false,
"should_log": false,
"should_save": true,
"should_training_stop": true
},
"attributes": {}
}
},
"total_flos": 1119551134941184.0,
"train_batch_size": 8,
"trial_name": null,
"trial_params": null
}