GaetanMichelet's picture
Model save
6ad0939 verified
{
"best_metric": 0.922440230846405,
"best_model_checkpoint": "data/Llama-31-8B_task-1_60-samples_config-3_full/checkpoint-241",
"epoch": 48.869565217391305,
"eval_steps": 500,
"global_step": 281,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.17391304347826086,
"grad_norm": 2.0927231311798096,
"learning_rate": 1.3333333333333336e-07,
"loss": 2.45,
"step": 1
},
{
"epoch": 0.34782608695652173,
"grad_norm": 2.0467705726623535,
"learning_rate": 2.666666666666667e-07,
"loss": 2.5615,
"step": 2
},
{
"epoch": 0.6956521739130435,
"grad_norm": 1.8534517288208008,
"learning_rate": 5.333333333333335e-07,
"loss": 2.5395,
"step": 4
},
{
"epoch": 0.8695652173913043,
"eval_loss": 2.414898633956909,
"eval_runtime": 4.8246,
"eval_samples_per_second": 2.487,
"eval_steps_per_second": 2.487,
"step": 5
},
{
"epoch": 1.0434782608695652,
"grad_norm": 1.889623761177063,
"learning_rate": 8.000000000000001e-07,
"loss": 2.4961,
"step": 6
},
{
"epoch": 1.391304347826087,
"grad_norm": 2.059065103530884,
"learning_rate": 1.066666666666667e-06,
"loss": 2.5344,
"step": 8
},
{
"epoch": 1.7391304347826086,
"grad_norm": 1.8490760326385498,
"learning_rate": 1.3333333333333334e-06,
"loss": 2.4512,
"step": 10
},
{
"epoch": 1.9130434782608696,
"eval_loss": 2.3973469734191895,
"eval_runtime": 4.826,
"eval_samples_per_second": 2.487,
"eval_steps_per_second": 2.487,
"step": 11
},
{
"epoch": 2.0869565217391304,
"grad_norm": 1.871017336845398,
"learning_rate": 1.6000000000000001e-06,
"loss": 2.5467,
"step": 12
},
{
"epoch": 2.4347826086956523,
"grad_norm": 1.7037708759307861,
"learning_rate": 1.8666666666666669e-06,
"loss": 2.5079,
"step": 14
},
{
"epoch": 2.782608695652174,
"grad_norm": 1.658904790878296,
"learning_rate": 2.133333333333334e-06,
"loss": 2.4419,
"step": 16
},
{
"epoch": 2.9565217391304346,
"eval_loss": 2.372079372406006,
"eval_runtime": 4.8248,
"eval_samples_per_second": 2.487,
"eval_steps_per_second": 2.487,
"step": 17
},
{
"epoch": 3.130434782608696,
"grad_norm": 1.583447813987732,
"learning_rate": 2.4000000000000003e-06,
"loss": 2.4868,
"step": 18
},
{
"epoch": 3.4782608695652173,
"grad_norm": 1.6701875925064087,
"learning_rate": 2.666666666666667e-06,
"loss": 2.459,
"step": 20
},
{
"epoch": 3.8260869565217392,
"grad_norm": 1.580588936805725,
"learning_rate": 2.9333333333333338e-06,
"loss": 2.3921,
"step": 22
},
{
"epoch": 4.0,
"eval_loss": 2.3361399173736572,
"eval_runtime": 4.8269,
"eval_samples_per_second": 2.486,
"eval_steps_per_second": 2.486,
"step": 23
},
{
"epoch": 4.173913043478261,
"grad_norm": 1.8148592710494995,
"learning_rate": 3.2000000000000003e-06,
"loss": 2.5221,
"step": 24
},
{
"epoch": 4.521739130434782,
"grad_norm": 1.5363281965255737,
"learning_rate": 3.4666666666666672e-06,
"loss": 2.4424,
"step": 26
},
{
"epoch": 4.869565217391305,
"grad_norm": 1.7005550861358643,
"learning_rate": 3.7333333333333337e-06,
"loss": 2.3357,
"step": 28
},
{
"epoch": 4.869565217391305,
"eval_loss": 2.295424461364746,
"eval_runtime": 4.8298,
"eval_samples_per_second": 2.485,
"eval_steps_per_second": 2.485,
"step": 28
},
{
"epoch": 5.217391304347826,
"grad_norm": 1.866983413696289,
"learning_rate": 4.000000000000001e-06,
"loss": 2.3658,
"step": 30
},
{
"epoch": 5.565217391304348,
"grad_norm": 1.9320333003997803,
"learning_rate": 4.266666666666668e-06,
"loss": 2.3898,
"step": 32
},
{
"epoch": 5.913043478260869,
"grad_norm": 1.7484076023101807,
"learning_rate": 4.533333333333334e-06,
"loss": 2.3559,
"step": 34
},
{
"epoch": 5.913043478260869,
"eval_loss": 2.2286934852600098,
"eval_runtime": 4.8271,
"eval_samples_per_second": 2.486,
"eval_steps_per_second": 2.486,
"step": 34
},
{
"epoch": 6.260869565217392,
"grad_norm": 1.423509955406189,
"learning_rate": 4.800000000000001e-06,
"loss": 2.2329,
"step": 36
},
{
"epoch": 6.608695652173913,
"grad_norm": 1.3438361883163452,
"learning_rate": 5.0666666666666676e-06,
"loss": 2.2697,
"step": 38
},
{
"epoch": 6.956521739130435,
"grad_norm": 1.4381420612335205,
"learning_rate": 5.333333333333334e-06,
"loss": 2.2622,
"step": 40
},
{
"epoch": 6.956521739130435,
"eval_loss": 2.165445566177368,
"eval_runtime": 4.8282,
"eval_samples_per_second": 2.485,
"eval_steps_per_second": 2.485,
"step": 40
},
{
"epoch": 7.304347826086957,
"grad_norm": 1.5358691215515137,
"learning_rate": 5.600000000000001e-06,
"loss": 2.2903,
"step": 42
},
{
"epoch": 7.6521739130434785,
"grad_norm": 1.7526369094848633,
"learning_rate": 5.8666666666666675e-06,
"loss": 2.1646,
"step": 44
},
{
"epoch": 8.0,
"grad_norm": 1.6337189674377441,
"learning_rate": 6.133333333333334e-06,
"loss": 2.186,
"step": 46
},
{
"epoch": 8.0,
"eval_loss": 2.0752172470092773,
"eval_runtime": 4.8264,
"eval_samples_per_second": 2.486,
"eval_steps_per_second": 2.486,
"step": 46
},
{
"epoch": 8.347826086956522,
"grad_norm": 1.175416350364685,
"learning_rate": 6.4000000000000006e-06,
"loss": 2.1512,
"step": 48
},
{
"epoch": 8.695652173913043,
"grad_norm": 1.0512149333953857,
"learning_rate": 6.666666666666667e-06,
"loss": 2.0842,
"step": 50
},
{
"epoch": 8.869565217391305,
"eval_loss": 2.0000388622283936,
"eval_runtime": 4.8255,
"eval_samples_per_second": 2.487,
"eval_steps_per_second": 2.487,
"step": 51
},
{
"epoch": 9.043478260869565,
"grad_norm": 1.13035249710083,
"learning_rate": 6.9333333333333344e-06,
"loss": 2.0837,
"step": 52
},
{
"epoch": 9.391304347826088,
"grad_norm": 1.1123379468917847,
"learning_rate": 7.2000000000000005e-06,
"loss": 2.0165,
"step": 54
},
{
"epoch": 9.73913043478261,
"grad_norm": 1.0869874954223633,
"learning_rate": 7.4666666666666675e-06,
"loss": 2.0522,
"step": 56
},
{
"epoch": 9.91304347826087,
"eval_loss": 1.895993709564209,
"eval_runtime": 4.8268,
"eval_samples_per_second": 2.486,
"eval_steps_per_second": 2.486,
"step": 57
},
{
"epoch": 10.08695652173913,
"grad_norm": 0.8969458341598511,
"learning_rate": 7.733333333333334e-06,
"loss": 1.9308,
"step": 58
},
{
"epoch": 10.434782608695652,
"grad_norm": 0.9428701400756836,
"learning_rate": 8.000000000000001e-06,
"loss": 1.9244,
"step": 60
},
{
"epoch": 10.782608695652174,
"grad_norm": 0.8733699321746826,
"learning_rate": 8.266666666666667e-06,
"loss": 1.911,
"step": 62
},
{
"epoch": 10.956521739130435,
"eval_loss": 1.7942091226577759,
"eval_runtime": 4.8257,
"eval_samples_per_second": 2.487,
"eval_steps_per_second": 2.487,
"step": 63
},
{
"epoch": 11.130434782608695,
"grad_norm": 0.8451066613197327,
"learning_rate": 8.533333333333335e-06,
"loss": 1.8571,
"step": 64
},
{
"epoch": 11.478260869565217,
"grad_norm": 0.8815340995788574,
"learning_rate": 8.8e-06,
"loss": 1.7567,
"step": 66
},
{
"epoch": 11.826086956521738,
"grad_norm": 0.9107153415679932,
"learning_rate": 9.066666666666667e-06,
"loss": 1.8076,
"step": 68
},
{
"epoch": 12.0,
"eval_loss": 1.6759809255599976,
"eval_runtime": 4.8213,
"eval_samples_per_second": 2.489,
"eval_steps_per_second": 2.489,
"step": 69
},
{
"epoch": 12.173913043478262,
"grad_norm": 0.9571322202682495,
"learning_rate": 9.333333333333334e-06,
"loss": 1.6998,
"step": 70
},
{
"epoch": 12.521739130434783,
"grad_norm": 0.9073280096054077,
"learning_rate": 9.600000000000001e-06,
"loss": 1.701,
"step": 72
},
{
"epoch": 12.869565217391305,
"grad_norm": 1.23154878616333,
"learning_rate": 9.866666666666668e-06,
"loss": 1.659,
"step": 74
},
{
"epoch": 12.869565217391305,
"eval_loss": 1.564476490020752,
"eval_runtime": 4.83,
"eval_samples_per_second": 2.484,
"eval_steps_per_second": 2.484,
"step": 74
},
{
"epoch": 13.217391304347826,
"grad_norm": 1.1613293886184692,
"learning_rate": 9.999945845889795e-06,
"loss": 1.5798,
"step": 76
},
{
"epoch": 13.565217391304348,
"grad_norm": 1.0568084716796875,
"learning_rate": 9.999512620046523e-06,
"loss": 1.4833,
"step": 78
},
{
"epoch": 13.91304347826087,
"grad_norm": 0.9477815628051758,
"learning_rate": 9.99864620589731e-06,
"loss": 1.5002,
"step": 80
},
{
"epoch": 13.91304347826087,
"eval_loss": 1.4213947057724,
"eval_runtime": 4.834,
"eval_samples_per_second": 2.482,
"eval_steps_per_second": 2.482,
"step": 80
},
{
"epoch": 14.26086956521739,
"grad_norm": 1.0447523593902588,
"learning_rate": 9.99734667851357e-06,
"loss": 1.4439,
"step": 82
},
{
"epoch": 14.608695652173914,
"grad_norm": 0.9249804019927979,
"learning_rate": 9.995614150494293e-06,
"loss": 1.3666,
"step": 84
},
{
"epoch": 14.956521739130435,
"grad_norm": 1.0063517093658447,
"learning_rate": 9.993448771956285e-06,
"loss": 1.309,
"step": 86
},
{
"epoch": 14.956521739130435,
"eval_loss": 1.294012188911438,
"eval_runtime": 4.8245,
"eval_samples_per_second": 2.487,
"eval_steps_per_second": 2.487,
"step": 86
},
{
"epoch": 15.304347826086957,
"grad_norm": 1.0014103651046753,
"learning_rate": 9.99085073052117e-06,
"loss": 1.3033,
"step": 88
},
{
"epoch": 15.652173913043478,
"grad_norm": 0.9277029037475586,
"learning_rate": 9.987820251299121e-06,
"loss": 1.2597,
"step": 90
},
{
"epoch": 16.0,
"grad_norm": 0.8152615427970886,
"learning_rate": 9.984357596869369e-06,
"loss": 1.2079,
"step": 92
},
{
"epoch": 16.0,
"eval_loss": 1.183652400970459,
"eval_runtime": 4.8266,
"eval_samples_per_second": 2.486,
"eval_steps_per_second": 2.486,
"step": 92
},
{
"epoch": 16.347826086956523,
"grad_norm": 0.7255731821060181,
"learning_rate": 9.980463067257437e-06,
"loss": 1.1647,
"step": 94
},
{
"epoch": 16.695652173913043,
"grad_norm": 0.755571186542511,
"learning_rate": 9.976136999909156e-06,
"loss": 1.1738,
"step": 96
},
{
"epoch": 16.869565217391305,
"eval_loss": 1.1230496168136597,
"eval_runtime": 4.8268,
"eval_samples_per_second": 2.486,
"eval_steps_per_second": 2.486,
"step": 97
},
{
"epoch": 17.043478260869566,
"grad_norm": 0.7182919383049011,
"learning_rate": 9.971379769661422e-06,
"loss": 1.122,
"step": 98
},
{
"epoch": 17.391304347826086,
"grad_norm": 0.7512508034706116,
"learning_rate": 9.966191788709716e-06,
"loss": 1.131,
"step": 100
},
{
"epoch": 17.73913043478261,
"grad_norm": 0.6947119235992432,
"learning_rate": 9.960573506572391e-06,
"loss": 1.0304,
"step": 102
},
{
"epoch": 17.91304347826087,
"eval_loss": 1.0780998468399048,
"eval_runtime": 4.8327,
"eval_samples_per_second": 2.483,
"eval_steps_per_second": 2.483,
"step": 103
},
{
"epoch": 18.08695652173913,
"grad_norm": 0.6211488246917725,
"learning_rate": 9.95452541005172e-06,
"loss": 1.0439,
"step": 104
},
{
"epoch": 18.434782608695652,
"grad_norm": 0.6497947573661804,
"learning_rate": 9.948048023191728e-06,
"loss": 1.0278,
"step": 106
},
{
"epoch": 18.782608695652176,
"grad_norm": 0.5914028882980347,
"learning_rate": 9.941141907232766e-06,
"loss": 1.0485,
"step": 108
},
{
"epoch": 18.956521739130434,
"eval_loss": 1.0459355115890503,
"eval_runtime": 4.8455,
"eval_samples_per_second": 2.477,
"eval_steps_per_second": 2.477,
"step": 109
},
{
"epoch": 19.130434782608695,
"grad_norm": 0.6026037931442261,
"learning_rate": 9.933807660562898e-06,
"loss": 1.0755,
"step": 110
},
{
"epoch": 19.47826086956522,
"grad_norm": 0.5290880799293518,
"learning_rate": 9.926045918666045e-06,
"loss": 1.0554,
"step": 112
},
{
"epoch": 19.82608695652174,
"grad_norm": 0.5054655075073242,
"learning_rate": 9.91785735406693e-06,
"loss": 0.9687,
"step": 114
},
{
"epoch": 20.0,
"eval_loss": 1.0257662534713745,
"eval_runtime": 4.8234,
"eval_samples_per_second": 2.488,
"eval_steps_per_second": 2.488,
"step": 115
},
{
"epoch": 20.17391304347826,
"grad_norm": 0.5547760725021362,
"learning_rate": 9.909242676272797e-06,
"loss": 0.9625,
"step": 116
},
{
"epoch": 20.52173913043478,
"grad_norm": 0.5321418642997742,
"learning_rate": 9.90020263171194e-06,
"loss": 1.0313,
"step": 118
},
{
"epoch": 20.869565217391305,
"grad_norm": 0.4589158296585083,
"learning_rate": 9.890738003669029e-06,
"loss": 0.9883,
"step": 120
},
{
"epoch": 20.869565217391305,
"eval_loss": 1.0146735906600952,
"eval_runtime": 4.8436,
"eval_samples_per_second": 2.478,
"eval_steps_per_second": 2.478,
"step": 120
},
{
"epoch": 21.217391304347824,
"grad_norm": 0.47460901737213135,
"learning_rate": 9.880849612217238e-06,
"loss": 0.9519,
"step": 122
},
{
"epoch": 21.565217391304348,
"grad_norm": 0.4670533835887909,
"learning_rate": 9.870538314147194e-06,
"loss": 1.0113,
"step": 124
},
{
"epoch": 21.91304347826087,
"grad_norm": 0.5945753455162048,
"learning_rate": 9.859805002892733e-06,
"loss": 0.974,
"step": 126
},
{
"epoch": 21.91304347826087,
"eval_loss": 1.00125253200531,
"eval_runtime": 4.8426,
"eval_samples_per_second": 2.478,
"eval_steps_per_second": 2.478,
"step": 126
},
{
"epoch": 22.26086956521739,
"grad_norm": 0.45630648732185364,
"learning_rate": 9.84865060845349e-06,
"loss": 0.9679,
"step": 128
},
{
"epoch": 22.608695652173914,
"grad_norm": 0.4838283360004425,
"learning_rate": 9.83707609731432e-06,
"loss": 0.9776,
"step": 130
},
{
"epoch": 22.956521739130434,
"grad_norm": 0.44602882862091064,
"learning_rate": 9.825082472361558e-06,
"loss": 0.9397,
"step": 132
},
{
"epoch": 22.956521739130434,
"eval_loss": 0.9904754757881165,
"eval_runtime": 4.8474,
"eval_samples_per_second": 2.476,
"eval_steps_per_second": 2.476,
"step": 132
},
{
"epoch": 23.304347826086957,
"grad_norm": 0.5826734304428101,
"learning_rate": 9.812670772796113e-06,
"loss": 0.9607,
"step": 134
},
{
"epoch": 23.652173913043477,
"grad_norm": 0.4837598204612732,
"learning_rate": 9.799842074043438e-06,
"loss": 0.949,
"step": 136
},
{
"epoch": 24.0,
"grad_norm": 0.6160168051719666,
"learning_rate": 9.786597487660336e-06,
"loss": 0.9522,
"step": 138
},
{
"epoch": 24.0,
"eval_loss": 0.9816201329231262,
"eval_runtime": 4.84,
"eval_samples_per_second": 2.479,
"eval_steps_per_second": 2.479,
"step": 138
},
{
"epoch": 24.347826086956523,
"grad_norm": 0.5032287836074829,
"learning_rate": 9.77293816123866e-06,
"loss": 0.9474,
"step": 140
},
{
"epoch": 24.695652173913043,
"grad_norm": 0.5077723860740662,
"learning_rate": 9.75886527830587e-06,
"loss": 0.9115,
"step": 142
},
{
"epoch": 24.869565217391305,
"eval_loss": 0.9739218354225159,
"eval_runtime": 4.8407,
"eval_samples_per_second": 2.479,
"eval_steps_per_second": 2.479,
"step": 143
},
{
"epoch": 25.043478260869566,
"grad_norm": 0.49520373344421387,
"learning_rate": 9.744380058222483e-06,
"loss": 0.9696,
"step": 144
},
{
"epoch": 25.391304347826086,
"grad_norm": 0.6057612299919128,
"learning_rate": 9.729483756076436e-06,
"loss": 0.9363,
"step": 146
},
{
"epoch": 25.73913043478261,
"grad_norm": 0.5979700088500977,
"learning_rate": 9.714177662574316e-06,
"loss": 0.9412,
"step": 148
},
{
"epoch": 25.91304347826087,
"eval_loss": 0.9668495655059814,
"eval_runtime": 4.8393,
"eval_samples_per_second": 2.48,
"eval_steps_per_second": 2.48,
"step": 149
},
{
"epoch": 26.08695652173913,
"grad_norm": 0.5768654942512512,
"learning_rate": 9.698463103929542e-06,
"loss": 0.9277,
"step": 150
},
{
"epoch": 26.434782608695652,
"grad_norm": 0.5760251879692078,
"learning_rate": 9.682341441747446e-06,
"loss": 0.8601,
"step": 152
},
{
"epoch": 26.782608695652176,
"grad_norm": 0.569760799407959,
"learning_rate": 9.665814072907293e-06,
"loss": 0.9168,
"step": 154
},
{
"epoch": 26.956521739130434,
"eval_loss": 0.9610000252723694,
"eval_runtime": 4.8499,
"eval_samples_per_second": 2.474,
"eval_steps_per_second": 2.474,
"step": 155
},
{
"epoch": 27.130434782608695,
"grad_norm": 0.5672867894172668,
"learning_rate": 9.648882429441258e-06,
"loss": 0.9273,
"step": 156
},
{
"epoch": 27.47826086956522,
"grad_norm": 0.5984346866607666,
"learning_rate": 9.63154797841033e-06,
"loss": 0.9134,
"step": 158
},
{
"epoch": 27.82608695652174,
"grad_norm": 0.5996063947677612,
"learning_rate": 9.613812221777212e-06,
"loss": 0.9461,
"step": 160
},
{
"epoch": 28.0,
"eval_loss": 0.9547294974327087,
"eval_runtime": 4.8363,
"eval_samples_per_second": 2.481,
"eval_steps_per_second": 2.481,
"step": 161
},
{
"epoch": 28.17391304347826,
"grad_norm": 0.5729570984840393,
"learning_rate": 9.595676696276173e-06,
"loss": 0.8601,
"step": 162
},
{
"epoch": 28.52173913043478,
"grad_norm": 0.6671696305274963,
"learning_rate": 9.577142973279896e-06,
"loss": 0.9214,
"step": 164
},
{
"epoch": 28.869565217391305,
"grad_norm": 0.6552798748016357,
"learning_rate": 9.55821265866333e-06,
"loss": 0.8579,
"step": 166
},
{
"epoch": 28.869565217391305,
"eval_loss": 0.9499099850654602,
"eval_runtime": 4.8467,
"eval_samples_per_second": 2.476,
"eval_steps_per_second": 2.476,
"step": 166
},
{
"epoch": 29.217391304347824,
"grad_norm": 0.6634982824325562,
"learning_rate": 9.538887392664544e-06,
"loss": 0.869,
"step": 168
},
{
"epoch": 29.565217391304348,
"grad_norm": 0.6576085090637207,
"learning_rate": 9.519168849742603e-06,
"loss": 0.9128,
"step": 170
},
{
"epoch": 29.91304347826087,
"grad_norm": 0.6544057130813599,
"learning_rate": 9.499058738432492e-06,
"loss": 0.8857,
"step": 172
},
{
"epoch": 29.91304347826087,
"eval_loss": 0.9454047083854675,
"eval_runtime": 4.8567,
"eval_samples_per_second": 2.471,
"eval_steps_per_second": 2.471,
"step": 172
},
{
"epoch": 30.26086956521739,
"grad_norm": 0.7192568778991699,
"learning_rate": 9.478558801197065e-06,
"loss": 0.8608,
"step": 174
},
{
"epoch": 30.608695652173914,
"grad_norm": 0.8773357272148132,
"learning_rate": 9.457670814276083e-06,
"loss": 0.8683,
"step": 176
},
{
"epoch": 30.956521739130434,
"grad_norm": 0.6999642252922058,
"learning_rate": 9.436396587532297e-06,
"loss": 0.8465,
"step": 178
},
{
"epoch": 30.956521739130434,
"eval_loss": 0.9405279755592346,
"eval_runtime": 4.8418,
"eval_samples_per_second": 2.478,
"eval_steps_per_second": 2.478,
"step": 178
},
{
"epoch": 31.304347826086957,
"grad_norm": 0.8239638805389404,
"learning_rate": 9.414737964294636e-06,
"loss": 0.911,
"step": 180
},
{
"epoch": 31.652173913043477,
"grad_norm": 0.672042965888977,
"learning_rate": 9.392696821198488e-06,
"loss": 0.8287,
"step": 182
},
{
"epoch": 32.0,
"grad_norm": 0.8173245191574097,
"learning_rate": 9.370275068023097e-06,
"loss": 0.8681,
"step": 184
},
{
"epoch": 32.0,
"eval_loss": 0.939335286617279,
"eval_runtime": 4.8393,
"eval_samples_per_second": 2.48,
"eval_steps_per_second": 2.48,
"step": 184
},
{
"epoch": 32.34782608695652,
"grad_norm": 0.7714847326278687,
"learning_rate": 9.347474647526095e-06,
"loss": 0.8525,
"step": 186
},
{
"epoch": 32.69565217391305,
"grad_norm": 0.7664777040481567,
"learning_rate": 9.324297535275156e-06,
"loss": 0.8257,
"step": 188
},
{
"epoch": 32.869565217391305,
"eval_loss": 0.9344443678855896,
"eval_runtime": 4.8421,
"eval_samples_per_second": 2.478,
"eval_steps_per_second": 2.478,
"step": 189
},
{
"epoch": 33.04347826086956,
"grad_norm": 0.9124040007591248,
"learning_rate": 9.30074573947683e-06,
"loss": 0.8519,
"step": 190
},
{
"epoch": 33.391304347826086,
"grad_norm": 0.8003515005111694,
"learning_rate": 9.276821300802535e-06,
"loss": 0.8863,
"step": 192
},
{
"epoch": 33.73913043478261,
"grad_norm": 0.8585532903671265,
"learning_rate": 9.25252629221175e-06,
"loss": 0.8425,
"step": 194
},
{
"epoch": 33.91304347826087,
"eval_loss": 0.9336052536964417,
"eval_runtime": 4.8393,
"eval_samples_per_second": 2.48,
"eval_steps_per_second": 2.48,
"step": 195
},
{
"epoch": 34.08695652173913,
"grad_norm": 0.8088707327842712,
"learning_rate": 9.227862818772392e-06,
"loss": 0.7927,
"step": 196
},
{
"epoch": 34.43478260869565,
"grad_norm": 1.031660795211792,
"learning_rate": 9.202833017478421e-06,
"loss": 0.8326,
"step": 198
},
{
"epoch": 34.78260869565217,
"grad_norm": 0.9502612352371216,
"learning_rate": 9.177439057064684e-06,
"loss": 0.8405,
"step": 200
},
{
"epoch": 34.95652173913044,
"eval_loss": 0.9280831217765808,
"eval_runtime": 4.8396,
"eval_samples_per_second": 2.48,
"eval_steps_per_second": 2.48,
"step": 201
},
{
"epoch": 35.130434782608695,
"grad_norm": 0.836633026599884,
"learning_rate": 9.151683137818989e-06,
"loss": 0.8087,
"step": 202
},
{
"epoch": 35.47826086956522,
"grad_norm": 0.8722138404846191,
"learning_rate": 9.125567491391476e-06,
"loss": 0.8281,
"step": 204
},
{
"epoch": 35.82608695652174,
"grad_norm": 0.9598701596260071,
"learning_rate": 9.099094380601244e-06,
"loss": 0.8101,
"step": 206
},
{
"epoch": 36.0,
"eval_loss": 0.9282541275024414,
"eval_runtime": 4.8408,
"eval_samples_per_second": 2.479,
"eval_steps_per_second": 2.479,
"step": 207
},
{
"epoch": 36.17391304347826,
"grad_norm": 0.9185284972190857,
"learning_rate": 9.072266099240286e-06,
"loss": 0.8047,
"step": 208
},
{
"epoch": 36.52173913043478,
"grad_norm": 0.9987873435020447,
"learning_rate": 9.045084971874738e-06,
"loss": 0.8261,
"step": 210
},
{
"epoch": 36.869565217391305,
"grad_norm": 0.9007511138916016,
"learning_rate": 9.017553353643479e-06,
"loss": 0.7808,
"step": 212
},
{
"epoch": 36.869565217391305,
"eval_loss": 0.9258975982666016,
"eval_runtime": 4.8444,
"eval_samples_per_second": 2.477,
"eval_steps_per_second": 2.477,
"step": 212
},
{
"epoch": 37.21739130434783,
"grad_norm": 0.9600070714950562,
"learning_rate": 8.989673630054044e-06,
"loss": 0.8273,
"step": 214
},
{
"epoch": 37.56521739130435,
"grad_norm": 0.8776330351829529,
"learning_rate": 8.961448216775955e-06,
"loss": 0.7458,
"step": 216
},
{
"epoch": 37.91304347826087,
"grad_norm": 1.0257493257522583,
"learning_rate": 8.932879559431392e-06,
"loss": 0.7971,
"step": 218
},
{
"epoch": 37.91304347826087,
"eval_loss": 0.9259226322174072,
"eval_runtime": 4.8247,
"eval_samples_per_second": 2.487,
"eval_steps_per_second": 2.487,
"step": 218
},
{
"epoch": 38.26086956521739,
"grad_norm": 0.9936304092407227,
"learning_rate": 8.903970133383297e-06,
"loss": 0.8089,
"step": 220
},
{
"epoch": 38.608695652173914,
"grad_norm": 0.9844217896461487,
"learning_rate": 8.874722443520898e-06,
"loss": 0.7994,
"step": 222
},
{
"epoch": 38.95652173913044,
"grad_norm": 1.0099091529846191,
"learning_rate": 8.845139024042664e-06,
"loss": 0.7766,
"step": 224
},
{
"epoch": 38.95652173913044,
"eval_loss": 0.9234599471092224,
"eval_runtime": 4.8243,
"eval_samples_per_second": 2.487,
"eval_steps_per_second": 2.487,
"step": 224
},
{
"epoch": 39.30434782608695,
"grad_norm": 0.9933149814605713,
"learning_rate": 8.815222438236726e-06,
"loss": 0.7425,
"step": 226
},
{
"epoch": 39.65217391304348,
"grad_norm": 0.9792707562446594,
"learning_rate": 8.784975278258783e-06,
"loss": 0.7805,
"step": 228
},
{
"epoch": 40.0,
"grad_norm": 1.1036958694458008,
"learning_rate": 8.754400164907496e-06,
"loss": 0.7748,
"step": 230
},
{
"epoch": 40.0,
"eval_loss": 0.9245247840881348,
"eval_runtime": 4.8317,
"eval_samples_per_second": 2.484,
"eval_steps_per_second": 2.484,
"step": 230
},
{
"epoch": 40.34782608695652,
"grad_norm": 1.0168761014938354,
"learning_rate": 8.723499747397415e-06,
"loss": 0.7756,
"step": 232
},
{
"epoch": 40.69565217391305,
"grad_norm": 1.0455681085586548,
"learning_rate": 8.692276703129421e-06,
"loss": 0.7476,
"step": 234
},
{
"epoch": 40.869565217391305,
"eval_loss": 0.9252880215644836,
"eval_runtime": 4.8248,
"eval_samples_per_second": 2.487,
"eval_steps_per_second": 2.487,
"step": 235
},
{
"epoch": 41.04347826086956,
"grad_norm": 1.2682240009307861,
"learning_rate": 8.660733737458751e-06,
"loss": 0.7633,
"step": 236
},
{
"epoch": 41.391304347826086,
"grad_norm": 1.054377555847168,
"learning_rate": 8.628873583460593e-06,
"loss": 0.7528,
"step": 238
},
{
"epoch": 41.73913043478261,
"grad_norm": 1.2782171964645386,
"learning_rate": 8.596699001693257e-06,
"loss": 0.7007,
"step": 240
},
{
"epoch": 41.91304347826087,
"eval_loss": 0.922440230846405,
"eval_runtime": 4.8272,
"eval_samples_per_second": 2.486,
"eval_steps_per_second": 2.486,
"step": 241
},
{
"epoch": 42.08695652173913,
"grad_norm": 1.09280526638031,
"learning_rate": 8.564212779959003e-06,
"loss": 0.7865,
"step": 242
},
{
"epoch": 42.43478260869565,
"grad_norm": 1.2800565958023071,
"learning_rate": 8.531417733062476e-06,
"loss": 0.742,
"step": 244
},
{
"epoch": 42.78260869565217,
"grad_norm": 1.2969675064086914,
"learning_rate": 8.498316702566828e-06,
"loss": 0.741,
"step": 246
},
{
"epoch": 42.95652173913044,
"eval_loss": 0.9261074066162109,
"eval_runtime": 4.8284,
"eval_samples_per_second": 2.485,
"eval_steps_per_second": 2.485,
"step": 247
},
{
"epoch": 43.130434782608695,
"grad_norm": 1.0573549270629883,
"learning_rate": 8.464912556547486e-06,
"loss": 0.7027,
"step": 248
},
{
"epoch": 43.47826086956522,
"grad_norm": 1.2901216745376587,
"learning_rate": 8.43120818934367e-06,
"loss": 0.7325,
"step": 250
},
{
"epoch": 43.82608695652174,
"grad_norm": 1.110465407371521,
"learning_rate": 8.397206521307584e-06,
"loss": 0.7371,
"step": 252
},
{
"epoch": 44.0,
"eval_loss": 0.9239481091499329,
"eval_runtime": 4.8243,
"eval_samples_per_second": 2.487,
"eval_steps_per_second": 2.487,
"step": 253
},
{
"epoch": 44.17391304347826,
"grad_norm": 1.1729587316513062,
"learning_rate": 8.362910498551402e-06,
"loss": 0.7156,
"step": 254
},
{
"epoch": 44.52173913043478,
"grad_norm": 1.1838419437408447,
"learning_rate": 8.328323092691985e-06,
"loss": 0.712,
"step": 256
},
{
"epoch": 44.869565217391305,
"grad_norm": 1.4539400339126587,
"learning_rate": 8.293447300593402e-06,
"loss": 0.7239,
"step": 258
},
{
"epoch": 44.869565217391305,
"eval_loss": 0.9323117136955261,
"eval_runtime": 4.8254,
"eval_samples_per_second": 2.487,
"eval_steps_per_second": 2.487,
"step": 258
},
{
"epoch": 45.21739130434783,
"grad_norm": 1.2814104557037354,
"learning_rate": 8.258286144107277e-06,
"loss": 0.729,
"step": 260
},
{
"epoch": 45.56521739130435,
"grad_norm": 1.1663718223571777,
"learning_rate": 8.222842669810936e-06,
"loss": 0.6689,
"step": 262
},
{
"epoch": 45.91304347826087,
"grad_norm": 1.2235040664672852,
"learning_rate": 8.18711994874345e-06,
"loss": 0.671,
"step": 264
},
{
"epoch": 45.91304347826087,
"eval_loss": 0.9268998503684998,
"eval_runtime": 4.8329,
"eval_samples_per_second": 2.483,
"eval_steps_per_second": 2.483,
"step": 264
},
{
"epoch": 46.26086956521739,
"grad_norm": 1.2671639919281006,
"learning_rate": 8.151121076139534e-06,
"loss": 0.7247,
"step": 266
},
{
"epoch": 46.608695652173914,
"grad_norm": 1.2693313360214233,
"learning_rate": 8.11484917116136e-06,
"loss": 0.6697,
"step": 268
},
{
"epoch": 46.95652173913044,
"grad_norm": 1.452881932258606,
"learning_rate": 8.078307376628292e-06,
"loss": 0.7312,
"step": 270
},
{
"epoch": 46.95652173913044,
"eval_loss": 0.9332544207572937,
"eval_runtime": 4.8317,
"eval_samples_per_second": 2.484,
"eval_steps_per_second": 2.484,
"step": 270
},
{
"epoch": 47.30434782608695,
"grad_norm": 1.3550682067871094,
"learning_rate": 8.041498858744572e-06,
"loss": 0.7067,
"step": 272
},
{
"epoch": 47.65217391304348,
"grad_norm": 1.3532733917236328,
"learning_rate": 8.004426806824985e-06,
"loss": 0.6586,
"step": 274
},
{
"epoch": 48.0,
"grad_norm": 1.4842863082885742,
"learning_rate": 7.967094433018508e-06,
"loss": 0.6826,
"step": 276
},
{
"epoch": 48.0,
"eval_loss": 0.934523344039917,
"eval_runtime": 4.8197,
"eval_samples_per_second": 2.49,
"eval_steps_per_second": 2.49,
"step": 276
},
{
"epoch": 48.34782608695652,
"grad_norm": 1.209937334060669,
"learning_rate": 7.929504972030003e-06,
"loss": 0.6882,
"step": 278
},
{
"epoch": 48.69565217391305,
"grad_norm": 1.5903738737106323,
"learning_rate": 7.891661680839932e-06,
"loss": 0.6472,
"step": 280
},
{
"epoch": 48.869565217391305,
"eval_loss": 0.9393181204795837,
"eval_runtime": 4.8289,
"eval_samples_per_second": 2.485,
"eval_steps_per_second": 2.485,
"step": 281
},
{
"epoch": 48.869565217391305,
"step": 281,
"total_flos": 3.3738755183476736e+16,
"train_loss": 1.2498505263990354,
"train_runtime": 2644.3439,
"train_samples_per_second": 2.609,
"train_steps_per_second": 0.284
}
],
"logging_steps": 2,
"max_steps": 750,
"num_input_tokens_seen": 0,
"num_train_epochs": 150,
"save_steps": 25,
"stateful_callbacks": {
"EarlyStoppingCallback": {
"args": {
"early_stopping_patience": 7,
"early_stopping_threshold": 0.0
},
"attributes": {
"early_stopping_patience_counter": 0
}
},
"TrainerControl": {
"args": {
"should_epoch_stop": false,
"should_evaluate": false,
"should_log": false,
"should_save": true,
"should_training_stop": true
},
"attributes": {}
}
},
"total_flos": 3.3738755183476736e+16,
"train_batch_size": 1,
"trial_name": null,
"trial_params": null
}