imdatta0's picture
End of training
8b1b653 verified
raw
history blame
22.3 kB
{
"best_metric": 1.265087604522705,
"best_model_checkpoint": "/home/datta0/models/lora_final/Meta-Llama-3-8B_magiccoder_ortho/checkpoint-148",
"epoch": 0.9975708502024292,
"eval_steps": 4,
"global_step": 154,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.006477732793522267,
"grad_norm": 4.835215091705322,
"learning_rate": 7.5e-05,
"loss": 1.3887,
"step": 1
},
{
"epoch": 0.012955465587044534,
"grad_norm": 5.698163032531738,
"learning_rate": 0.00015,
"loss": 1.5047,
"step": 2
},
{
"epoch": 0.025910931174089068,
"grad_norm": 3.3888368606567383,
"learning_rate": 0.0003,
"loss": 1.2463,
"step": 4
},
{
"epoch": 0.025910931174089068,
"eval_loss": 1.403336763381958,
"eval_runtime": 26.4071,
"eval_samples_per_second": 18.745,
"eval_steps_per_second": 2.348,
"step": 4
},
{
"epoch": 0.038866396761133605,
"grad_norm": 4.008037567138672,
"learning_rate": 0.00029986842451482874,
"loss": 1.4275,
"step": 6
},
{
"epoch": 0.051821862348178135,
"grad_norm": 3.188356637954712,
"learning_rate": 0.0002994739288874256,
"loss": 1.3247,
"step": 8
},
{
"epoch": 0.051821862348178135,
"eval_loss": 1.3321858644485474,
"eval_runtime": 26.4909,
"eval_samples_per_second": 18.686,
"eval_steps_per_second": 2.34,
"step": 8
},
{
"epoch": 0.06477732793522267,
"grad_norm": 3.098771572113037,
"learning_rate": 0.0002988172051971717,
"loss": 1.2629,
"step": 10
},
{
"epoch": 0.07773279352226721,
"grad_norm": 2.1870386600494385,
"learning_rate": 0.0002978994055605757,
"loss": 1.2964,
"step": 12
},
{
"epoch": 0.07773279352226721,
"eval_loss": 1.3196189403533936,
"eval_runtime": 26.5166,
"eval_samples_per_second": 18.668,
"eval_steps_per_second": 2.338,
"step": 12
},
{
"epoch": 0.09068825910931175,
"grad_norm": 3.890054225921631,
"learning_rate": 0.0002967221401100708,
"loss": 1.3716,
"step": 14
},
{
"epoch": 0.10364372469635627,
"grad_norm": 2.7932441234588623,
"learning_rate": 0.00029528747416929463,
"loss": 1.3289,
"step": 16
},
{
"epoch": 0.10364372469635627,
"eval_loss": 1.322401762008667,
"eval_runtime": 26.4113,
"eval_samples_per_second": 18.742,
"eval_steps_per_second": 2.347,
"step": 16
},
{
"epoch": 0.11659919028340081,
"grad_norm": 2.2833335399627686,
"learning_rate": 0.00029359792462981004,
"loss": 1.2691,
"step": 18
},
{
"epoch": 0.12955465587044535,
"grad_norm": 2.577002763748169,
"learning_rate": 0.00029165645553562214,
"loss": 1.3483,
"step": 20
},
{
"epoch": 0.12955465587044535,
"eval_loss": 1.3099651336669922,
"eval_runtime": 26.4006,
"eval_samples_per_second": 18.75,
"eval_steps_per_second": 2.348,
"step": 20
},
{
"epoch": 0.14251012145748987,
"grad_norm": 2.350156784057617,
"learning_rate": 0.00028946647288323766,
"loss": 1.3168,
"step": 22
},
{
"epoch": 0.15546558704453442,
"grad_norm": 2.7228236198425293,
"learning_rate": 0.0002870318186463901,
"loss": 1.2491,
"step": 24
},
{
"epoch": 0.15546558704453442,
"eval_loss": 1.3193615674972534,
"eval_runtime": 26.2281,
"eval_samples_per_second": 18.873,
"eval_steps_per_second": 2.364,
"step": 24
},
{
"epoch": 0.16842105263157894,
"grad_norm": 2.790332078933716,
"learning_rate": 0.0002843567640359119,
"loss": 1.249,
"step": 26
},
{
"epoch": 0.1813765182186235,
"grad_norm": 2.422020196914673,
"learning_rate": 0.0002814460020065795,
"loss": 1.3072,
"step": 28
},
{
"epoch": 0.1813765182186235,
"eval_loss": 1.3144476413726807,
"eval_runtime": 63.9332,
"eval_samples_per_second": 7.742,
"eval_steps_per_second": 0.97,
"step": 28
},
{
"epoch": 0.19433198380566802,
"grad_norm": 1.7604172229766846,
"learning_rate": 0.000278304639024076,
"loss": 1.3043,
"step": 30
},
{
"epoch": 0.20728744939271254,
"grad_norm": 1.9333935976028442,
"learning_rate": 0.00027493818610651487,
"loss": 1.2865,
"step": 32
},
{
"epoch": 0.20728744939271254,
"eval_loss": 1.3123451471328735,
"eval_runtime": 65.1552,
"eval_samples_per_second": 7.597,
"eval_steps_per_second": 0.952,
"step": 32
},
{
"epoch": 0.2202429149797571,
"grad_norm": 2.5974128246307373,
"learning_rate": 0.0002713525491562421,
"loss": 1.3347,
"step": 34
},
{
"epoch": 0.23319838056680162,
"grad_norm": 2.198709726333618,
"learning_rate": 0.00026755401859887595,
"loss": 1.3102,
"step": 36
},
{
"epoch": 0.23319838056680162,
"eval_loss": 1.3170647621154785,
"eval_runtime": 62.3447,
"eval_samples_per_second": 7.94,
"eval_steps_per_second": 0.994,
"step": 36
},
{
"epoch": 0.24615384615384617,
"grad_norm": 2.3173534870147705,
"learning_rate": 0.00026354925834776345,
"loss": 1.2321,
"step": 38
},
{
"epoch": 0.2591093117408907,
"grad_norm": 3.5600650310516357,
"learning_rate": 0.0002593452941132117,
"loss": 1.3752,
"step": 40
},
{
"epoch": 0.2591093117408907,
"eval_loss": 1.3158267736434937,
"eval_runtime": 62.0865,
"eval_samples_per_second": 7.973,
"eval_steps_per_second": 0.999,
"step": 40
},
{
"epoch": 0.2720647773279352,
"grad_norm": 2.1025493144989014,
"learning_rate": 0.0002549495010770048,
"loss": 1.3306,
"step": 42
},
{
"epoch": 0.28502024291497974,
"grad_norm": 2.076892137527466,
"learning_rate": 0.0002503695909538287,
"loss": 1.3244,
"step": 44
},
{
"epoch": 0.28502024291497974,
"eval_loss": 1.3114184141159058,
"eval_runtime": 62.2069,
"eval_samples_per_second": 7.957,
"eval_steps_per_second": 0.997,
"step": 44
},
{
"epoch": 0.2979757085020243,
"grad_norm": 2.493210792541504,
"learning_rate": 0.0002456135984623034,
"loss": 1.2883,
"step": 46
},
{
"epoch": 0.31093117408906884,
"grad_norm": 1.9219117164611816,
"learning_rate": 0.00024068986722935624,
"loss": 1.2311,
"step": 48
},
{
"epoch": 0.31093117408906884,
"eval_loss": 1.3117666244506836,
"eval_runtime": 62.4372,
"eval_samples_per_second": 7.928,
"eval_steps_per_second": 0.993,
"step": 48
},
{
"epoch": 0.32388663967611336,
"grad_norm": 2.240204095840454,
"learning_rate": 0.00023560703515266478,
"loss": 1.3414,
"step": 50
},
{
"epoch": 0.3368421052631579,
"grad_norm": 2.00461483001709,
"learning_rate": 0.00023037401924684946,
"loss": 1.2911,
"step": 52
},
{
"epoch": 0.3368421052631579,
"eval_loss": 1.3134574890136719,
"eval_runtime": 62.0904,
"eval_samples_per_second": 7.972,
"eval_steps_per_second": 0.999,
"step": 52
},
{
"epoch": 0.3497975708502024,
"grad_norm": 1.787941575050354,
"learning_rate": 0.000225,
"loss": 1.2649,
"step": 54
},
{
"epoch": 0.362753036437247,
"grad_norm": 2.0428550243377686,
"learning_rate": 0.00021949440526797926,
"loss": 1.3409,
"step": 56
},
{
"epoch": 0.362753036437247,
"eval_loss": 1.3078958988189697,
"eval_runtime": 62.252,
"eval_samples_per_second": 7.952,
"eval_steps_per_second": 0.996,
"step": 56
},
{
"epoch": 0.3757085020242915,
"grad_norm": 1.9100394248962402,
"learning_rate": 0.00021386689373476087,
"loss": 1.2472,
"step": 58
},
{
"epoch": 0.38866396761133604,
"grad_norm": 1.7520307302474976,
"learning_rate": 0.00020812733796781542,
"loss": 1.3069,
"step": 60
},
{
"epoch": 0.38866396761133604,
"eval_loss": 1.3043341636657715,
"eval_runtime": 62.3545,
"eval_samples_per_second": 7.938,
"eval_steps_per_second": 0.994,
"step": 60
},
{
"epoch": 0.40161943319838056,
"grad_norm": 2.4810237884521484,
"learning_rate": 0.00020228580709827227,
"loss": 1.2735,
"step": 62
},
{
"epoch": 0.4145748987854251,
"grad_norm": 2.136749744415283,
"learning_rate": 0.0001963525491562421,
"loss": 1.362,
"step": 64
},
{
"epoch": 0.4145748987854251,
"eval_loss": 1.3125,
"eval_runtime": 62.3299,
"eval_samples_per_second": 7.942,
"eval_steps_per_second": 0.995,
"step": 64
},
{
"epoch": 0.42753036437246966,
"grad_norm": 2.069849967956543,
"learning_rate": 0.00019033797309228983,
"loss": 1.2799,
"step": 66
},
{
"epoch": 0.4404858299595142,
"grad_norm": 1.8651994466781616,
"learning_rate": 0.00018425263051659836,
"loss": 1.2206,
"step": 68
},
{
"epoch": 0.4404858299595142,
"eval_loss": 1.3050696849822998,
"eval_runtime": 26.5576,
"eval_samples_per_second": 18.639,
"eval_steps_per_second": 2.335,
"step": 68
},
{
"epoch": 0.4534412955465587,
"grad_norm": 1.9996711015701294,
"learning_rate": 0.0001781071971878587,
"loss": 1.2901,
"step": 70
},
{
"epoch": 0.46639676113360323,
"grad_norm": 1.663189172744751,
"learning_rate": 0.00017191245428436173,
"loss": 1.2838,
"step": 72
},
{
"epoch": 0.46639676113360323,
"eval_loss": 1.2985526323318481,
"eval_runtime": 26.5419,
"eval_samples_per_second": 18.65,
"eval_steps_per_second": 2.336,
"step": 72
},
{
"epoch": 0.47935222672064776,
"grad_norm": 1.6311547756195068,
"learning_rate": 0.000165679269490148,
"loss": 1.2626,
"step": 74
},
{
"epoch": 0.49230769230769234,
"grad_norm": 1.993211269378662,
"learning_rate": 0.000159418577929397,
"loss": 1.2348,
"step": 76
},
{
"epoch": 0.49230769230769234,
"eval_loss": 1.3072912693023682,
"eval_runtime": 26.5378,
"eval_samples_per_second": 18.653,
"eval_steps_per_second": 2.336,
"step": 76
},
{
"epoch": 0.5052631578947369,
"grad_norm": 1.840496301651001,
"learning_rate": 0.00015314136298250354,
"loss": 1.3011,
"step": 78
},
{
"epoch": 0.5182186234817814,
"grad_norm": 1.9565261602401733,
"learning_rate": 0.00014685863701749646,
"loss": 1.3171,
"step": 80
},
{
"epoch": 0.5182186234817814,
"eval_loss": 1.2922303676605225,
"eval_runtime": 26.4704,
"eval_samples_per_second": 18.7,
"eval_steps_per_second": 2.342,
"step": 80
},
{
"epoch": 0.5311740890688259,
"grad_norm": 1.8801251649856567,
"learning_rate": 0.000140581422070603,
"loss": 1.3119,
"step": 82
},
{
"epoch": 0.5441295546558704,
"grad_norm": 1.784315824508667,
"learning_rate": 0.000134320730509852,
"loss": 1.2556,
"step": 84
},
{
"epoch": 0.5441295546558704,
"eval_loss": 1.296497106552124,
"eval_runtime": 26.4835,
"eval_samples_per_second": 18.691,
"eval_steps_per_second": 2.341,
"step": 84
},
{
"epoch": 0.557085020242915,
"grad_norm": 2.0044546127319336,
"learning_rate": 0.00012808754571563827,
"loss": 1.2803,
"step": 86
},
{
"epoch": 0.5700404858299595,
"grad_norm": 2.1428680419921875,
"learning_rate": 0.00012189280281214126,
"loss": 1.2803,
"step": 88
},
{
"epoch": 0.5700404858299595,
"eval_loss": 1.291109561920166,
"eval_runtime": 26.3248,
"eval_samples_per_second": 18.804,
"eval_steps_per_second": 2.355,
"step": 88
},
{
"epoch": 0.582995951417004,
"grad_norm": 1.4893293380737305,
"learning_rate": 0.00011574736948340163,
"loss": 1.2737,
"step": 90
},
{
"epoch": 0.5959514170040486,
"grad_norm": 1.5611094236373901,
"learning_rate": 0.00010966202690771014,
"loss": 1.3796,
"step": 92
},
{
"epoch": 0.5959514170040486,
"eval_loss": 1.2853686809539795,
"eval_runtime": 26.1611,
"eval_samples_per_second": 18.921,
"eval_steps_per_second": 2.37,
"step": 92
},
{
"epoch": 0.6089068825910932,
"grad_norm": 1.776159405708313,
"learning_rate": 0.0001036474508437579,
"loss": 1.218,
"step": 94
},
{
"epoch": 0.6218623481781377,
"grad_norm": 1.4612406492233276,
"learning_rate": 9.771419290172773e-05,
"loss": 1.2047,
"step": 96
},
{
"epoch": 0.6218623481781377,
"eval_loss": 1.2870514392852783,
"eval_runtime": 61.6553,
"eval_samples_per_second": 8.029,
"eval_steps_per_second": 1.006,
"step": 96
},
{
"epoch": 0.6348178137651822,
"grad_norm": 2.135633707046509,
"learning_rate": 9.187266203218456e-05,
"loss": 1.325,
"step": 98
},
{
"epoch": 0.6477732793522267,
"grad_norm": 1.6593787670135498,
"learning_rate": 8.613310626523909e-05,
"loss": 1.2821,
"step": 100
},
{
"epoch": 0.6477732793522267,
"eval_loss": 1.2865735292434692,
"eval_runtime": 62.3493,
"eval_samples_per_second": 7.939,
"eval_steps_per_second": 0.994,
"step": 100
},
{
"epoch": 0.6607287449392713,
"grad_norm": 1.7937084436416626,
"learning_rate": 8.050559473202077e-05,
"loss": 1.3041,
"step": 102
},
{
"epoch": 0.6736842105263158,
"grad_norm": 1.7992043495178223,
"learning_rate": 7.500000000000002e-05,
"loss": 1.2012,
"step": 104
},
{
"epoch": 0.6736842105263158,
"eval_loss": 1.2837588787078857,
"eval_runtime": 62.4051,
"eval_samples_per_second": 7.932,
"eval_steps_per_second": 0.994,
"step": 104
},
{
"epoch": 0.6866396761133603,
"grad_norm": 1.7913408279418945,
"learning_rate": 6.962598075315046e-05,
"loss": 1.3132,
"step": 106
},
{
"epoch": 0.6995951417004048,
"grad_norm": 1.61056387424469,
"learning_rate": 6.439296484733525e-05,
"loss": 1.2116,
"step": 108
},
{
"epoch": 0.6995951417004048,
"eval_loss": 1.2799283266067505,
"eval_runtime": 62.0965,
"eval_samples_per_second": 7.971,
"eval_steps_per_second": 0.998,
"step": 108
},
{
"epoch": 0.7125506072874493,
"grad_norm": 1.418599247932434,
"learning_rate": 5.931013277064377e-05,
"loss": 1.232,
"step": 110
},
{
"epoch": 0.725506072874494,
"grad_norm": 1.4885016679763794,
"learning_rate": 5.4386401537696536e-05,
"loss": 1.23,
"step": 112
},
{
"epoch": 0.725506072874494,
"eval_loss": 1.2750165462493896,
"eval_runtime": 61.731,
"eval_samples_per_second": 8.019,
"eval_steps_per_second": 1.004,
"step": 112
},
{
"epoch": 0.7384615384615385,
"grad_norm": 1.9909800291061401,
"learning_rate": 4.963040904617131e-05,
"loss": 1.2355,
"step": 114
},
{
"epoch": 0.751417004048583,
"grad_norm": 1.4655224084854126,
"learning_rate": 4.5050498922995166e-05,
"loss": 1.2679,
"step": 116
},
{
"epoch": 0.751417004048583,
"eval_loss": 1.2715320587158203,
"eval_runtime": 61.9746,
"eval_samples_per_second": 7.987,
"eval_steps_per_second": 1.0,
"step": 116
},
{
"epoch": 0.7643724696356275,
"grad_norm": 1.7242943048477173,
"learning_rate": 4.06547058867883e-05,
"loss": 1.1916,
"step": 118
},
{
"epoch": 0.7773279352226721,
"grad_norm": 1.4455363750457764,
"learning_rate": 3.645074165223655e-05,
"loss": 1.2573,
"step": 120
},
{
"epoch": 0.7773279352226721,
"eval_loss": 1.2714204788208008,
"eval_runtime": 61.4581,
"eval_samples_per_second": 8.054,
"eval_steps_per_second": 1.009,
"step": 120
},
{
"epoch": 0.7902834008097166,
"grad_norm": 1.4860985279083252,
"learning_rate": 3.2445981401124035e-05,
"loss": 1.232,
"step": 122
},
{
"epoch": 0.8032388663967611,
"grad_norm": 1.4141103029251099,
"learning_rate": 2.8647450843757897e-05,
"loss": 1.2802,
"step": 124
},
{
"epoch": 0.8032388663967611,
"eval_loss": 1.2692365646362305,
"eval_runtime": 61.9164,
"eval_samples_per_second": 7.995,
"eval_steps_per_second": 1.001,
"step": 124
},
{
"epoch": 0.8161943319838056,
"grad_norm": 1.710113763809204,
"learning_rate": 2.5061813893485085e-05,
"loss": 1.3136,
"step": 126
},
{
"epoch": 0.8291497975708502,
"grad_norm": 1.509994626045227,
"learning_rate": 2.169536097592401e-05,
"loss": 1.2772,
"step": 128
},
{
"epoch": 0.8291497975708502,
"eval_loss": 1.2681289911270142,
"eval_runtime": 62.1335,
"eval_samples_per_second": 7.967,
"eval_steps_per_second": 0.998,
"step": 128
},
{
"epoch": 0.8421052631578947,
"grad_norm": 1.5493632555007935,
"learning_rate": 1.8553997993420495e-05,
"loss": 1.2752,
"step": 130
},
{
"epoch": 0.8550607287449393,
"grad_norm": 1.5303385257720947,
"learning_rate": 1.5643235964088064e-05,
"loss": 1.2594,
"step": 132
},
{
"epoch": 0.8550607287449393,
"eval_loss": 1.2668787240982056,
"eval_runtime": 62.3532,
"eval_samples_per_second": 7.939,
"eval_steps_per_second": 0.994,
"step": 132
},
{
"epoch": 0.8680161943319838,
"grad_norm": 1.4996320009231567,
"learning_rate": 1.2968181353609852e-05,
"loss": 1.2546,
"step": 134
},
{
"epoch": 0.8809716599190284,
"grad_norm": 1.4413329362869263,
"learning_rate": 1.0533527116762296e-05,
"loss": 1.218,
"step": 136
},
{
"epoch": 0.8809716599190284,
"eval_loss": 1.266615390777588,
"eval_runtime": 26.5779,
"eval_samples_per_second": 18.625,
"eval_steps_per_second": 2.333,
"step": 136
},
{
"epoch": 0.8939271255060729,
"grad_norm": 1.725917100906372,
"learning_rate": 8.343544464377849e-06,
"loss": 1.2472,
"step": 138
},
{
"epoch": 0.9068825910931174,
"grad_norm": 1.4963750839233398,
"learning_rate": 6.402075370189913e-06,
"loss": 1.2391,
"step": 140
},
{
"epoch": 0.9068825910931174,
"eval_loss": 1.265771746635437,
"eval_runtime": 26.5434,
"eval_samples_per_second": 18.649,
"eval_steps_per_second": 2.336,
"step": 140
},
{
"epoch": 0.9198380566801619,
"grad_norm": 1.4187538623809814,
"learning_rate": 4.712525830705338e-06,
"loss": 1.2176,
"step": 142
},
{
"epoch": 0.9327935222672065,
"grad_norm": 1.6049160957336426,
"learning_rate": 3.2778598899291465e-06,
"loss": 1.2084,
"step": 144
},
{
"epoch": 0.9327935222672065,
"eval_loss": 1.2656446695327759,
"eval_runtime": 26.5096,
"eval_samples_per_second": 18.672,
"eval_steps_per_second": 2.339,
"step": 144
},
{
"epoch": 0.945748987854251,
"grad_norm": 1.41164231300354,
"learning_rate": 2.100594439424269e-06,
"loss": 1.2457,
"step": 146
},
{
"epoch": 0.9587044534412955,
"grad_norm": 1.5624706745147705,
"learning_rate": 1.1827948028283352e-06,
"loss": 1.2245,
"step": 148
},
{
"epoch": 0.9587044534412955,
"eval_loss": 1.265087604522705,
"eval_runtime": 26.4922,
"eval_samples_per_second": 18.685,
"eval_steps_per_second": 2.34,
"step": 148
},
{
"epoch": 0.97165991902834,
"grad_norm": 1.3811231851577759,
"learning_rate": 5.260711125743444e-07,
"loss": 1.2052,
"step": 150
},
{
"epoch": 0.9846153846153847,
"grad_norm": 1.5876771211624146,
"learning_rate": 1.315754851712425e-07,
"loss": 1.1995,
"step": 152
},
{
"epoch": 0.9846153846153847,
"eval_loss": 1.2650998830795288,
"eval_runtime": 26.377,
"eval_samples_per_second": 18.766,
"eval_steps_per_second": 2.351,
"step": 152
},
{
"epoch": 0.9975708502024292,
"grad_norm": 1.6923319101333618,
"learning_rate": 0.0,
"loss": 1.2372,
"step": 154
}
],
"logging_steps": 2,
"max_steps": 154,
"num_input_tokens_seen": 0,
"num_train_epochs": 1,
"save_steps": 4,
"stateful_callbacks": {
"TrainerControl": {
"args": {
"should_epoch_stop": false,
"should_evaluate": false,
"should_log": false,
"should_save": true,
"should_training_stop": true
},
"attributes": {}
}
},
"total_flos": 4.1926442566916506e+17,
"train_batch_size": 8,
"trial_name": null,
"trial_params": null
}