1046 lines
28 KiB
JSON
1046 lines
28 KiB
JSON
{
|
|
"best_global_step": null,
|
|
"best_metric": null,
|
|
"best_model_checkpoint": null,
|
|
"epoch": 0.44385264092321347,
|
|
"eval_steps": 500,
|
|
"global_step": 500,
|
|
"is_hyper_param_search": false,
|
|
"is_local_process_zero": true,
|
|
"is_world_process_zero": true,
|
|
"log_history": [
|
|
{
|
|
"entropy": 10.692031192779542,
|
|
"epoch": 0.004438526409232135,
|
|
"grad_norm": 14.75,
|
|
"learning_rate": 2e-06,
|
|
"loss": 10.8207,
|
|
"mean_token_accuracy": 0.000128369708545506,
|
|
"num_tokens": 9810.0,
|
|
"step": 5
|
|
},
|
|
{
|
|
"entropy": 10.691952133178711,
|
|
"epoch": 0.00887705281846427,
|
|
"grad_norm": 12.9375,
|
|
"learning_rate": 4.5e-06,
|
|
"loss": 10.7613,
|
|
"mean_token_accuracy": 0.0,
|
|
"num_tokens": 20538.0,
|
|
"step": 10
|
|
},
|
|
{
|
|
"entropy": 10.690592384338379,
|
|
"epoch": 0.013315579227696404,
|
|
"grad_norm": 9.9375,
|
|
"learning_rate": 7e-06,
|
|
"loss": 10.5054,
|
|
"mean_token_accuracy": 0.03474730136804283,
|
|
"num_tokens": 30258.0,
|
|
"step": 15
|
|
},
|
|
{
|
|
"entropy": 10.66368646621704,
|
|
"epoch": 0.01775410563692854,
|
|
"grad_norm": 5.875,
|
|
"learning_rate": 9.5e-06,
|
|
"loss": 10.1947,
|
|
"mean_token_accuracy": 0.04976645708084106,
|
|
"num_tokens": 41249.0,
|
|
"step": 20
|
|
},
|
|
{
|
|
"entropy": 10.518089294433594,
|
|
"epoch": 0.022192632046160676,
|
|
"grad_norm": 3.609375,
|
|
"learning_rate": 1.2e-05,
|
|
"loss": 9.9,
|
|
"mean_token_accuracy": 0.053335465490818024,
|
|
"num_tokens": 51738.0,
|
|
"step": 25
|
|
},
|
|
{
|
|
"entropy": 10.489847755432129,
|
|
"epoch": 0.02663115845539281,
|
|
"grad_norm": 3.0625,
|
|
"learning_rate": 1.4500000000000002e-05,
|
|
"loss": 9.8161,
|
|
"mean_token_accuracy": 0.05364362597465515,
|
|
"num_tokens": 62920.0,
|
|
"step": 30
|
|
},
|
|
{
|
|
"entropy": 10.506165885925293,
|
|
"epoch": 0.031069684864624945,
|
|
"grad_norm": 2.859375,
|
|
"learning_rate": 1.7000000000000003e-05,
|
|
"loss": 9.7103,
|
|
"mean_token_accuracy": 0.05455525256693363,
|
|
"num_tokens": 73163.0,
|
|
"step": 35
|
|
},
|
|
{
|
|
"entropy": 10.478147506713867,
|
|
"epoch": 0.03550821127385708,
|
|
"grad_norm": 2.40625,
|
|
"learning_rate": 1.95e-05,
|
|
"loss": 9.682,
|
|
"mean_token_accuracy": 0.05231944099068642,
|
|
"num_tokens": 84180.0,
|
|
"step": 40
|
|
},
|
|
{
|
|
"entropy": 10.461441326141358,
|
|
"epoch": 0.03994673768308921,
|
|
"grad_norm": 2.359375,
|
|
"learning_rate": 2.2e-05,
|
|
"loss": 9.6388,
|
|
"mean_token_accuracy": 0.05221981666982174,
|
|
"num_tokens": 94108.0,
|
|
"step": 45
|
|
},
|
|
{
|
|
"entropy": 10.476018333435059,
|
|
"epoch": 0.04438526409232135,
|
|
"grad_norm": 3.140625,
|
|
"learning_rate": 2.4500000000000003e-05,
|
|
"loss": 9.6253,
|
|
"mean_token_accuracy": 0.05243444815278053,
|
|
"num_tokens": 104696.0,
|
|
"step": 50
|
|
},
|
|
{
|
|
"entropy": 10.478155612945557,
|
|
"epoch": 0.048823790501553485,
|
|
"grad_norm": 2.375,
|
|
"learning_rate": 2.7e-05,
|
|
"loss": 9.5235,
|
|
"mean_token_accuracy": 0.0584161002188921,
|
|
"num_tokens": 114568.0,
|
|
"step": 55
|
|
},
|
|
{
|
|
"entropy": 10.39725103378296,
|
|
"epoch": 0.05326231691078562,
|
|
"grad_norm": 2.125,
|
|
"learning_rate": 2.95e-05,
|
|
"loss": 9.4713,
|
|
"mean_token_accuracy": 0.064088149741292,
|
|
"num_tokens": 126251.0,
|
|
"step": 60
|
|
},
|
|
{
|
|
"entropy": 10.248920154571532,
|
|
"epoch": 0.05770084332001776,
|
|
"grad_norm": 2.109375,
|
|
"learning_rate": 3.2e-05,
|
|
"loss": 9.3504,
|
|
"mean_token_accuracy": 0.06839247606694698,
|
|
"num_tokens": 135612.0,
|
|
"step": 65
|
|
},
|
|
{
|
|
"entropy": 10.320736598968505,
|
|
"epoch": 0.06213936972924989,
|
|
"grad_norm": 2.078125,
|
|
"learning_rate": 3.4500000000000005e-05,
|
|
"loss": 9.4004,
|
|
"mean_token_accuracy": 0.0615400142967701,
|
|
"num_tokens": 145962.0,
|
|
"step": 70
|
|
},
|
|
{
|
|
"entropy": 10.323652744293213,
|
|
"epoch": 0.06657789613848203,
|
|
"grad_norm": 1.890625,
|
|
"learning_rate": 3.7e-05,
|
|
"loss": 9.1911,
|
|
"mean_token_accuracy": 0.07103513963520527,
|
|
"num_tokens": 155587.0,
|
|
"step": 75
|
|
},
|
|
{
|
|
"entropy": 10.258409976959229,
|
|
"epoch": 0.07101642254771416,
|
|
"grad_norm": 2.109375,
|
|
"learning_rate": 3.95e-05,
|
|
"loss": 9.1343,
|
|
"mean_token_accuracy": 0.06803948283195496,
|
|
"num_tokens": 165657.0,
|
|
"step": 80
|
|
},
|
|
{
|
|
"entropy": 10.24950189590454,
|
|
"epoch": 0.0754549489569463,
|
|
"grad_norm": 2.34375,
|
|
"learning_rate": 4.2000000000000004e-05,
|
|
"loss": 9.1277,
|
|
"mean_token_accuracy": 0.06729609742760659,
|
|
"num_tokens": 175844.0,
|
|
"step": 85
|
|
},
|
|
{
|
|
"entropy": 10.224775981903075,
|
|
"epoch": 0.07989347536617843,
|
|
"grad_norm": 2.046875,
|
|
"learning_rate": 4.45e-05,
|
|
"loss": 8.9736,
|
|
"mean_token_accuracy": 0.06970795653760434,
|
|
"num_tokens": 185453.0,
|
|
"step": 90
|
|
},
|
|
{
|
|
"entropy": 10.128818893432618,
|
|
"epoch": 0.08433200177541056,
|
|
"grad_norm": 1.8046875,
|
|
"learning_rate": 4.7000000000000004e-05,
|
|
"loss": 8.9469,
|
|
"mean_token_accuracy": 0.07015660107135772,
|
|
"num_tokens": 195296.0,
|
|
"step": 95
|
|
},
|
|
{
|
|
"entropy": 10.158192539215088,
|
|
"epoch": 0.0887705281846427,
|
|
"grad_norm": 1.7578125,
|
|
"learning_rate": 4.9500000000000004e-05,
|
|
"loss": 8.8594,
|
|
"mean_token_accuracy": 0.0717735793441534,
|
|
"num_tokens": 204491.0,
|
|
"step": 100
|
|
},
|
|
{
|
|
"entropy": 10.063592338562012,
|
|
"epoch": 0.09320905459387484,
|
|
"grad_norm": 1.578125,
|
|
"learning_rate": 5.2e-05,
|
|
"loss": 8.8275,
|
|
"mean_token_accuracy": 0.06594390086829663,
|
|
"num_tokens": 214916.0,
|
|
"step": 105
|
|
},
|
|
{
|
|
"entropy": 10.0551118850708,
|
|
"epoch": 0.09764758100310697,
|
|
"grad_norm": 1.640625,
|
|
"learning_rate": 5.45e-05,
|
|
"loss": 8.6661,
|
|
"mean_token_accuracy": 0.07398808561265469,
|
|
"num_tokens": 226660.0,
|
|
"step": 110
|
|
},
|
|
{
|
|
"entropy": 9.929056739807129,
|
|
"epoch": 0.1020861074123391,
|
|
"grad_norm": 2.203125,
|
|
"learning_rate": 5.7e-05,
|
|
"loss": 8.6191,
|
|
"mean_token_accuracy": 0.06771207004785537,
|
|
"num_tokens": 236554.0,
|
|
"step": 115
|
|
},
|
|
{
|
|
"entropy": 9.87773323059082,
|
|
"epoch": 0.10652463382157124,
|
|
"grad_norm": 1.390625,
|
|
"learning_rate": 5.9499999999999996e-05,
|
|
"loss": 8.5466,
|
|
"mean_token_accuracy": 0.07039828561246395,
|
|
"num_tokens": 247388.0,
|
|
"step": 120
|
|
},
|
|
{
|
|
"entropy": 9.78554401397705,
|
|
"epoch": 0.11096316023080337,
|
|
"grad_norm": 1.34375,
|
|
"learning_rate": 6.2e-05,
|
|
"loss": 8.3693,
|
|
"mean_token_accuracy": 0.07505594864487648,
|
|
"num_tokens": 257404.0,
|
|
"step": 125
|
|
},
|
|
{
|
|
"entropy": 9.569751739501953,
|
|
"epoch": 0.11540168664003551,
|
|
"grad_norm": 1.5859375,
|
|
"learning_rate": 6.450000000000001e-05,
|
|
"loss": 8.366,
|
|
"mean_token_accuracy": 0.0721758782863617,
|
|
"num_tokens": 268024.0,
|
|
"step": 130
|
|
},
|
|
{
|
|
"entropy": 9.49385757446289,
|
|
"epoch": 0.11984021304926765,
|
|
"grad_norm": 1.7890625,
|
|
"learning_rate": 6.7e-05,
|
|
"loss": 8.3116,
|
|
"mean_token_accuracy": 0.07231119871139527,
|
|
"num_tokens": 278477.0,
|
|
"step": 135
|
|
},
|
|
{
|
|
"entropy": 9.405902671813966,
|
|
"epoch": 0.12427873945849978,
|
|
"grad_norm": 1.2109375,
|
|
"learning_rate": 6.950000000000001e-05,
|
|
"loss": 8.2427,
|
|
"mean_token_accuracy": 0.06903640553355217,
|
|
"num_tokens": 289974.0,
|
|
"step": 140
|
|
},
|
|
{
|
|
"entropy": 9.16763277053833,
|
|
"epoch": 0.1287172658677319,
|
|
"grad_norm": 1.1953125,
|
|
"learning_rate": 7.2e-05,
|
|
"loss": 8.1839,
|
|
"mean_token_accuracy": 0.06725569888949394,
|
|
"num_tokens": 300150.0,
|
|
"step": 145
|
|
},
|
|
{
|
|
"entropy": 9.045907497406006,
|
|
"epoch": 0.13315579227696406,
|
|
"grad_norm": 1.2578125,
|
|
"learning_rate": 7.45e-05,
|
|
"loss": 8.0648,
|
|
"mean_token_accuracy": 0.07113164104521275,
|
|
"num_tokens": 310482.0,
|
|
"step": 150
|
|
},
|
|
{
|
|
"entropy": 8.951902961730957,
|
|
"epoch": 0.13759431868619618,
|
|
"grad_norm": 1.1875,
|
|
"learning_rate": 7.7e-05,
|
|
"loss": 8.1043,
|
|
"mean_token_accuracy": 0.07190654650330544,
|
|
"num_tokens": 321350.0,
|
|
"step": 155
|
|
},
|
|
{
|
|
"entropy": 8.734932899475098,
|
|
"epoch": 0.14203284509542832,
|
|
"grad_norm": 1.25,
|
|
"learning_rate": 7.950000000000001e-05,
|
|
"loss": 7.9902,
|
|
"mean_token_accuracy": 0.07260116301476956,
|
|
"num_tokens": 332009.0,
|
|
"step": 160
|
|
},
|
|
{
|
|
"entropy": 8.618322277069092,
|
|
"epoch": 0.14647137150466044,
|
|
"grad_norm": 1.4140625,
|
|
"learning_rate": 8.2e-05,
|
|
"loss": 8.0046,
|
|
"mean_token_accuracy": 0.07010005228221416,
|
|
"num_tokens": 341980.0,
|
|
"step": 165
|
|
},
|
|
{
|
|
"entropy": 8.576034545898438,
|
|
"epoch": 0.1509098979138926,
|
|
"grad_norm": 1.609375,
|
|
"learning_rate": 8.450000000000001e-05,
|
|
"loss": 7.9965,
|
|
"mean_token_accuracy": 0.06660077758133412,
|
|
"num_tokens": 354249.0,
|
|
"step": 170
|
|
},
|
|
{
|
|
"entropy": 8.524139022827148,
|
|
"epoch": 0.15534842432312473,
|
|
"grad_norm": 1.1875,
|
|
"learning_rate": 8.7e-05,
|
|
"loss": 7.9076,
|
|
"mean_token_accuracy": 0.07078476324677467,
|
|
"num_tokens": 364295.0,
|
|
"step": 175
|
|
},
|
|
{
|
|
"entropy": 8.328704929351806,
|
|
"epoch": 0.15978695073235685,
|
|
"grad_norm": 1.09375,
|
|
"learning_rate": 8.95e-05,
|
|
"loss": 7.8464,
|
|
"mean_token_accuracy": 0.0794072538614273,
|
|
"num_tokens": 373601.0,
|
|
"step": 180
|
|
},
|
|
{
|
|
"entropy": 8.382467269897461,
|
|
"epoch": 0.164225477141589,
|
|
"grad_norm": 1.8515625,
|
|
"learning_rate": 9.2e-05,
|
|
"loss": 7.8583,
|
|
"mean_token_accuracy": 0.07436131834983825,
|
|
"num_tokens": 384260.0,
|
|
"step": 185
|
|
},
|
|
{
|
|
"entropy": 8.316840362548827,
|
|
"epoch": 0.16866400355082112,
|
|
"grad_norm": 1.390625,
|
|
"learning_rate": 9.45e-05,
|
|
"loss": 7.9058,
|
|
"mean_token_accuracy": 0.06711366176605224,
|
|
"num_tokens": 394964.0,
|
|
"step": 190
|
|
},
|
|
{
|
|
"entropy": 8.204895782470704,
|
|
"epoch": 0.17310252996005326,
|
|
"grad_norm": 1.1484375,
|
|
"learning_rate": 9.7e-05,
|
|
"loss": 7.8384,
|
|
"mean_token_accuracy": 0.07343141734600067,
|
|
"num_tokens": 404470.0,
|
|
"step": 195
|
|
},
|
|
{
|
|
"entropy": 8.270458793640136,
|
|
"epoch": 0.1775410563692854,
|
|
"grad_norm": 1.0078125,
|
|
"learning_rate": 9.95e-05,
|
|
"loss": 7.9041,
|
|
"mean_token_accuracy": 0.07145134881138801,
|
|
"num_tokens": 414822.0,
|
|
"step": 200
|
|
},
|
|
{
|
|
"entropy": 8.344293403625489,
|
|
"epoch": 0.18197958277851753,
|
|
"grad_norm": 1.2578125,
|
|
"learning_rate": 0.000102,
|
|
"loss": 7.8228,
|
|
"mean_token_accuracy": 0.07206394150853157,
|
|
"num_tokens": 424367.0,
|
|
"step": 205
|
|
},
|
|
{
|
|
"entropy": 8.169469261169434,
|
|
"epoch": 0.18641810918774968,
|
|
"grad_norm": 1.4296875,
|
|
"learning_rate": 0.00010449999999999999,
|
|
"loss": 7.8063,
|
|
"mean_token_accuracy": 0.06796807795763016,
|
|
"num_tokens": 434911.0,
|
|
"step": 210
|
|
},
|
|
{
|
|
"entropy": 8.127049207687378,
|
|
"epoch": 0.1908566355969818,
|
|
"grad_norm": 1.0703125,
|
|
"learning_rate": 0.000107,
|
|
"loss": 7.7417,
|
|
"mean_token_accuracy": 0.08103418126702308,
|
|
"num_tokens": 445215.0,
|
|
"step": 215
|
|
},
|
|
{
|
|
"entropy": 8.216034698486329,
|
|
"epoch": 0.19529516200621394,
|
|
"grad_norm": 1.1875,
|
|
"learning_rate": 0.0001095,
|
|
"loss": 7.7919,
|
|
"mean_token_accuracy": 0.07842708677053452,
|
|
"num_tokens": 454786.0,
|
|
"step": 220
|
|
},
|
|
{
|
|
"entropy": 8.140088844299317,
|
|
"epoch": 0.19973368841544606,
|
|
"grad_norm": 1.4765625,
|
|
"learning_rate": 0.000112,
|
|
"loss": 7.8219,
|
|
"mean_token_accuracy": 0.07302889823913575,
|
|
"num_tokens": 465185.0,
|
|
"step": 225
|
|
},
|
|
{
|
|
"entropy": 8.021256399154662,
|
|
"epoch": 0.2041722148246782,
|
|
"grad_norm": 1.2890625,
|
|
"learning_rate": 0.0001145,
|
|
"loss": 7.7156,
|
|
"mean_token_accuracy": 0.08027704544365406,
|
|
"num_tokens": 475235.0,
|
|
"step": 230
|
|
},
|
|
{
|
|
"entropy": 8.17042875289917,
|
|
"epoch": 0.20861074123391035,
|
|
"grad_norm": 1.453125,
|
|
"learning_rate": 0.00011700000000000001,
|
|
"loss": 7.7317,
|
|
"mean_token_accuracy": 0.07805200070142745,
|
|
"num_tokens": 485408.0,
|
|
"step": 235
|
|
},
|
|
{
|
|
"entropy": 8.040202140808105,
|
|
"epoch": 0.21304926764314247,
|
|
"grad_norm": 1.28125,
|
|
"learning_rate": 0.00011949999999999999,
|
|
"loss": 7.7331,
|
|
"mean_token_accuracy": 0.0744746584445238,
|
|
"num_tokens": 495570.0,
|
|
"step": 240
|
|
},
|
|
{
|
|
"entropy": 8.125858116149903,
|
|
"epoch": 0.21748779405237462,
|
|
"grad_norm": 1.1953125,
|
|
"learning_rate": 0.000122,
|
|
"loss": 7.7248,
|
|
"mean_token_accuracy": 0.07628845125436783,
|
|
"num_tokens": 506103.0,
|
|
"step": 245
|
|
},
|
|
{
|
|
"entropy": 8.083836793899536,
|
|
"epoch": 0.22192632046160674,
|
|
"grad_norm": 0.94921875,
|
|
"learning_rate": 0.0001245,
|
|
"loss": 7.6989,
|
|
"mean_token_accuracy": 0.07588708400726318,
|
|
"num_tokens": 516416.0,
|
|
"step": 250
|
|
},
|
|
{
|
|
"entropy": 8.041877603530883,
|
|
"epoch": 0.22636484687083888,
|
|
"grad_norm": 1.0546875,
|
|
"learning_rate": 0.000127,
|
|
"loss": 7.7498,
|
|
"mean_token_accuracy": 0.0751910775899887,
|
|
"num_tokens": 525841.0,
|
|
"step": 255
|
|
},
|
|
{
|
|
"entropy": 8.231132793426514,
|
|
"epoch": 0.23080337328007103,
|
|
"grad_norm": 1.0390625,
|
|
"learning_rate": 0.0001295,
|
|
"loss": 7.7681,
|
|
"mean_token_accuracy": 0.07214248739182949,
|
|
"num_tokens": 536426.0,
|
|
"step": 260
|
|
},
|
|
{
|
|
"entropy": 8.028632116317748,
|
|
"epoch": 0.23524189968930315,
|
|
"grad_norm": 1.0703125,
|
|
"learning_rate": 0.000132,
|
|
"loss": 7.7229,
|
|
"mean_token_accuracy": 0.07526009976863861,
|
|
"num_tokens": 545852.0,
|
|
"step": 265
|
|
},
|
|
{
|
|
"entropy": 7.998026943206787,
|
|
"epoch": 0.2396804260985353,
|
|
"grad_norm": 1.015625,
|
|
"learning_rate": 0.00013450000000000002,
|
|
"loss": 7.6466,
|
|
"mean_token_accuracy": 0.07897478640079499,
|
|
"num_tokens": 555222.0,
|
|
"step": 270
|
|
},
|
|
{
|
|
"entropy": 7.9018083095550535,
|
|
"epoch": 0.2441189525077674,
|
|
"grad_norm": 1.046875,
|
|
"learning_rate": 0.00013700000000000002,
|
|
"loss": 7.6683,
|
|
"mean_token_accuracy": 0.0769478440284729,
|
|
"num_tokens": 566311.0,
|
|
"step": 275
|
|
},
|
|
{
|
|
"entropy": 8.067350673675538,
|
|
"epoch": 0.24855747891699956,
|
|
"grad_norm": 0.91015625,
|
|
"learning_rate": 0.0001395,
|
|
"loss": 7.6718,
|
|
"mean_token_accuracy": 0.07673554085195064,
|
|
"num_tokens": 576883.0,
|
|
"step": 280
|
|
},
|
|
{
|
|
"entropy": 8.010624122619628,
|
|
"epoch": 0.2529960053262317,
|
|
"grad_norm": 1.265625,
|
|
"learning_rate": 0.00014199999999999998,
|
|
"loss": 7.8136,
|
|
"mean_token_accuracy": 0.07454071268439293,
|
|
"num_tokens": 587387.0,
|
|
"step": 285
|
|
},
|
|
{
|
|
"entropy": 8.04631519317627,
|
|
"epoch": 0.2574345317354638,
|
|
"grad_norm": 0.99609375,
|
|
"learning_rate": 0.0001445,
|
|
"loss": 7.6914,
|
|
"mean_token_accuracy": 0.07921014353632927,
|
|
"num_tokens": 597229.0,
|
|
"step": 290
|
|
},
|
|
{
|
|
"entropy": 8.029859972000121,
|
|
"epoch": 0.26187305814469597,
|
|
"grad_norm": 1.0390625,
|
|
"learning_rate": 0.000147,
|
|
"loss": 7.6793,
|
|
"mean_token_accuracy": 0.0733891949057579,
|
|
"num_tokens": 606997.0,
|
|
"step": 295
|
|
},
|
|
{
|
|
"entropy": 7.9327795028686525,
|
|
"epoch": 0.2663115845539281,
|
|
"grad_norm": 1.2890625,
|
|
"learning_rate": 0.0001495,
|
|
"loss": 7.6346,
|
|
"mean_token_accuracy": 0.07976229339838029,
|
|
"num_tokens": 617316.0,
|
|
"step": 300
|
|
},
|
|
{
|
|
"entropy": 7.945201730728149,
|
|
"epoch": 0.2707501109631602,
|
|
"grad_norm": 1.125,
|
|
"learning_rate": 0.000152,
|
|
"loss": 7.6262,
|
|
"mean_token_accuracy": 0.07671116665005684,
|
|
"num_tokens": 626494.0,
|
|
"step": 305
|
|
},
|
|
{
|
|
"entropy": 7.911632776260376,
|
|
"epoch": 0.27518863737239235,
|
|
"grad_norm": 1.296875,
|
|
"learning_rate": 0.00015450000000000001,
|
|
"loss": 7.6211,
|
|
"mean_token_accuracy": 0.07656443342566491,
|
|
"num_tokens": 636816.0,
|
|
"step": 310
|
|
},
|
|
{
|
|
"entropy": 7.9385381698608395,
|
|
"epoch": 0.2796271637816245,
|
|
"grad_norm": 0.92578125,
|
|
"learning_rate": 0.000157,
|
|
"loss": 7.5921,
|
|
"mean_token_accuracy": 0.07875077873468399,
|
|
"num_tokens": 647520.0,
|
|
"step": 315
|
|
},
|
|
{
|
|
"entropy": 7.8889247417449955,
|
|
"epoch": 0.28406569019085665,
|
|
"grad_norm": 1.296875,
|
|
"learning_rate": 0.0001595,
|
|
"loss": 7.5615,
|
|
"mean_token_accuracy": 0.0802665926516056,
|
|
"num_tokens": 656772.0,
|
|
"step": 320
|
|
},
|
|
{
|
|
"entropy": 7.87529935836792,
|
|
"epoch": 0.2885042166000888,
|
|
"grad_norm": 1.0390625,
|
|
"learning_rate": 0.000162,
|
|
"loss": 7.6269,
|
|
"mean_token_accuracy": 0.07935804799199105,
|
|
"num_tokens": 666570.0,
|
|
"step": 325
|
|
},
|
|
{
|
|
"entropy": 7.889472913742066,
|
|
"epoch": 0.2929427430093209,
|
|
"grad_norm": 0.984375,
|
|
"learning_rate": 0.00016450000000000001,
|
|
"loss": 7.6449,
|
|
"mean_token_accuracy": 0.07449713721871376,
|
|
"num_tokens": 676873.0,
|
|
"step": 330
|
|
},
|
|
{
|
|
"entropy": 7.901484632492066,
|
|
"epoch": 0.29738126941855303,
|
|
"grad_norm": 1.0859375,
|
|
"learning_rate": 0.00016700000000000002,
|
|
"loss": 7.556,
|
|
"mean_token_accuracy": 0.0819828025996685,
|
|
"num_tokens": 687322.0,
|
|
"step": 335
|
|
},
|
|
{
|
|
"entropy": 7.938011121749878,
|
|
"epoch": 0.3018197958277852,
|
|
"grad_norm": 1.0859375,
|
|
"learning_rate": 0.00016950000000000003,
|
|
"loss": 7.6304,
|
|
"mean_token_accuracy": 0.07710144743323326,
|
|
"num_tokens": 697657.0,
|
|
"step": 340
|
|
},
|
|
{
|
|
"entropy": 7.953681612014771,
|
|
"epoch": 0.3062583222370173,
|
|
"grad_norm": 1.2265625,
|
|
"learning_rate": 0.00017199999999999998,
|
|
"loss": 7.5355,
|
|
"mean_token_accuracy": 0.07943713515996934,
|
|
"num_tokens": 707713.0,
|
|
"step": 345
|
|
},
|
|
{
|
|
"entropy": 7.818657636642456,
|
|
"epoch": 0.31069684864624947,
|
|
"grad_norm": 1.0546875,
|
|
"learning_rate": 0.00017449999999999999,
|
|
"loss": 7.5792,
|
|
"mean_token_accuracy": 0.08700495660305023,
|
|
"num_tokens": 719438.0,
|
|
"step": 350
|
|
},
|
|
{
|
|
"entropy": 7.922421979904175,
|
|
"epoch": 0.31513537505548156,
|
|
"grad_norm": 1.109375,
|
|
"learning_rate": 0.000177,
|
|
"loss": 7.5211,
|
|
"mean_token_accuracy": 0.0777974933385849,
|
|
"num_tokens": 728991.0,
|
|
"step": 355
|
|
},
|
|
{
|
|
"entropy": 7.7726140975952145,
|
|
"epoch": 0.3195739014647137,
|
|
"grad_norm": 1.7421875,
|
|
"learning_rate": 0.0001795,
|
|
"loss": 7.4375,
|
|
"mean_token_accuracy": 0.09254143610596657,
|
|
"num_tokens": 738276.0,
|
|
"step": 360
|
|
},
|
|
{
|
|
"entropy": 7.8144529342651365,
|
|
"epoch": 0.32401242787394585,
|
|
"grad_norm": 1.109375,
|
|
"learning_rate": 0.000182,
|
|
"loss": 7.537,
|
|
"mean_token_accuracy": 0.07724328823387623,
|
|
"num_tokens": 748405.0,
|
|
"step": 365
|
|
},
|
|
{
|
|
"entropy": 7.8304845809936525,
|
|
"epoch": 0.328450954283178,
|
|
"grad_norm": 1.015625,
|
|
"learning_rate": 0.0001845,
|
|
"loss": 7.5365,
|
|
"mean_token_accuracy": 0.08004771247506141,
|
|
"num_tokens": 759111.0,
|
|
"step": 370
|
|
},
|
|
{
|
|
"entropy": 7.755885601043701,
|
|
"epoch": 0.33288948069241014,
|
|
"grad_norm": 1.203125,
|
|
"learning_rate": 0.000187,
|
|
"loss": 7.4431,
|
|
"mean_token_accuracy": 0.08885042890906333,
|
|
"num_tokens": 770315.0,
|
|
"step": 375
|
|
},
|
|
{
|
|
"entropy": 7.815682411193848,
|
|
"epoch": 0.33732800710164224,
|
|
"grad_norm": 1.078125,
|
|
"learning_rate": 0.0001895,
|
|
"loss": 7.5761,
|
|
"mean_token_accuracy": 0.07824730798602104,
|
|
"num_tokens": 781390.0,
|
|
"step": 380
|
|
},
|
|
{
|
|
"entropy": 7.868446302413941,
|
|
"epoch": 0.3417665335108744,
|
|
"grad_norm": 1.0546875,
|
|
"learning_rate": 0.000192,
|
|
"loss": 7.4659,
|
|
"mean_token_accuracy": 0.08218091987073421,
|
|
"num_tokens": 790710.0,
|
|
"step": 385
|
|
},
|
|
{
|
|
"entropy": 7.664791536331177,
|
|
"epoch": 0.34620505992010653,
|
|
"grad_norm": 1.0625,
|
|
"learning_rate": 0.0001945,
|
|
"loss": 7.4344,
|
|
"mean_token_accuracy": 0.08516876846551895,
|
|
"num_tokens": 800475.0,
|
|
"step": 390
|
|
},
|
|
{
|
|
"entropy": 7.776907110214234,
|
|
"epoch": 0.3506435863293387,
|
|
"grad_norm": 1.109375,
|
|
"learning_rate": 0.00019700000000000002,
|
|
"loss": 7.5457,
|
|
"mean_token_accuracy": 0.07697484940290451,
|
|
"num_tokens": 810394.0,
|
|
"step": 395
|
|
},
|
|
{
|
|
"entropy": 7.851606655120849,
|
|
"epoch": 0.3550821127385708,
|
|
"grad_norm": 1.2421875,
|
|
"learning_rate": 0.00019950000000000002,
|
|
"loss": 7.5259,
|
|
"mean_token_accuracy": 0.07952482551336289,
|
|
"num_tokens": 820493.0,
|
|
"step": 400
|
|
},
|
|
{
|
|
"entropy": 7.881973552703857,
|
|
"epoch": 0.3595206391478029,
|
|
"grad_norm": 1.1796875,
|
|
"learning_rate": 0.000202,
|
|
"loss": 7.5896,
|
|
"mean_token_accuracy": 0.07090779766440392,
|
|
"num_tokens": 831570.0,
|
|
"step": 405
|
|
},
|
|
{
|
|
"entropy": 7.75118236541748,
|
|
"epoch": 0.36395916555703506,
|
|
"grad_norm": 1.09375,
|
|
"learning_rate": 0.00020449999999999998,
|
|
"loss": 7.5822,
|
|
"mean_token_accuracy": 0.07436199747025966,
|
|
"num_tokens": 841851.0,
|
|
"step": 410
|
|
},
|
|
{
|
|
"entropy": 7.8130326747894285,
|
|
"epoch": 0.3683976919662672,
|
|
"grad_norm": 1.2109375,
|
|
"learning_rate": 0.000207,
|
|
"loss": 7.4771,
|
|
"mean_token_accuracy": 0.08917279317975044,
|
|
"num_tokens": 851221.0,
|
|
"step": 415
|
|
},
|
|
{
|
|
"entropy": 7.784592771530152,
|
|
"epoch": 0.37283621837549935,
|
|
"grad_norm": 1.0703125,
|
|
"learning_rate": 0.0002095,
|
|
"loss": 7.5136,
|
|
"mean_token_accuracy": 0.08482674360275269,
|
|
"num_tokens": 861876.0,
|
|
"step": 420
|
|
},
|
|
{
|
|
"entropy": 7.806199550628662,
|
|
"epoch": 0.37727474478473144,
|
|
"grad_norm": 1.2265625,
|
|
"learning_rate": 0.000212,
|
|
"loss": 7.5012,
|
|
"mean_token_accuracy": 0.08520847856998444,
|
|
"num_tokens": 871742.0,
|
|
"step": 425
|
|
},
|
|
{
|
|
"entropy": 7.63223876953125,
|
|
"epoch": 0.3817132711939636,
|
|
"grad_norm": 1.0625,
|
|
"learning_rate": 0.0002145,
|
|
"loss": 7.4356,
|
|
"mean_token_accuracy": 0.08614917695522309,
|
|
"num_tokens": 881968.0,
|
|
"step": 430
|
|
},
|
|
{
|
|
"entropy": 7.708009338378906,
|
|
"epoch": 0.38615179760319573,
|
|
"grad_norm": 1.0546875,
|
|
"learning_rate": 0.00021700000000000002,
|
|
"loss": 7.3441,
|
|
"mean_token_accuracy": 0.0903160773217678,
|
|
"num_tokens": 892142.0,
|
|
"step": 435
|
|
},
|
|
{
|
|
"entropy": 7.681766939163208,
|
|
"epoch": 0.3905903240124279,
|
|
"grad_norm": 1.15625,
|
|
"learning_rate": 0.0002195,
|
|
"loss": 7.408,
|
|
"mean_token_accuracy": 0.09188471287488938,
|
|
"num_tokens": 901992.0,
|
|
"step": 440
|
|
},
|
|
{
|
|
"entropy": 7.797896862030029,
|
|
"epoch": 0.39502885042166,
|
|
"grad_norm": 1.1171875,
|
|
"learning_rate": 0.000222,
|
|
"loss": 7.5382,
|
|
"mean_token_accuracy": 0.07850045077502728,
|
|
"num_tokens": 913438.0,
|
|
"step": 445
|
|
},
|
|
{
|
|
"entropy": 7.55986213684082,
|
|
"epoch": 0.3994673768308921,
|
|
"grad_norm": 2.109375,
|
|
"learning_rate": 0.0002245,
|
|
"loss": 7.273,
|
|
"mean_token_accuracy": 0.10574427619576454,
|
|
"num_tokens": 923879.0,
|
|
"step": 450
|
|
},
|
|
{
|
|
"entropy": 7.674528026580811,
|
|
"epoch": 0.40390590324012426,
|
|
"grad_norm": 1.1015625,
|
|
"learning_rate": 0.00022700000000000002,
|
|
"loss": 7.4341,
|
|
"mean_token_accuracy": 0.08904042169451713,
|
|
"num_tokens": 933954.0,
|
|
"step": 455
|
|
},
|
|
{
|
|
"entropy": 7.707700920104981,
|
|
"epoch": 0.4083444296493564,
|
|
"grad_norm": 1.15625,
|
|
"learning_rate": 0.00022950000000000002,
|
|
"loss": 7.4156,
|
|
"mean_token_accuracy": 0.0867860458791256,
|
|
"num_tokens": 942992.0,
|
|
"step": 460
|
|
},
|
|
{
|
|
"entropy": 7.656433868408203,
|
|
"epoch": 0.41278295605858856,
|
|
"grad_norm": 1.25,
|
|
"learning_rate": 0.00023200000000000003,
|
|
"loss": 7.4528,
|
|
"mean_token_accuracy": 0.08972205594182014,
|
|
"num_tokens": 953822.0,
|
|
"step": 465
|
|
},
|
|
{
|
|
"entropy": 7.627706384658813,
|
|
"epoch": 0.4172214824678207,
|
|
"grad_norm": 1.1015625,
|
|
"learning_rate": 0.00023449999999999998,
|
|
"loss": 7.3472,
|
|
"mean_token_accuracy": 0.08931803703308105,
|
|
"num_tokens": 964475.0,
|
|
"step": 470
|
|
},
|
|
{
|
|
"entropy": 7.642127990722656,
|
|
"epoch": 0.4216600088770528,
|
|
"grad_norm": 1.0859375,
|
|
"learning_rate": 0.000237,
|
|
"loss": 7.434,
|
|
"mean_token_accuracy": 0.0859565369784832,
|
|
"num_tokens": 974535.0,
|
|
"step": 475
|
|
},
|
|
{
|
|
"entropy": 7.666657447814941,
|
|
"epoch": 0.42609853528628494,
|
|
"grad_norm": 1.0703125,
|
|
"learning_rate": 0.0002395,
|
|
"loss": 7.3944,
|
|
"mean_token_accuracy": 0.08582191392779351,
|
|
"num_tokens": 984674.0,
|
|
"step": 480
|
|
},
|
|
{
|
|
"entropy": 7.544086360931397,
|
|
"epoch": 0.4305370616955171,
|
|
"grad_norm": 1.171875,
|
|
"learning_rate": 0.000242,
|
|
"loss": 7.37,
|
|
"mean_token_accuracy": 0.09044271633028984,
|
|
"num_tokens": 995248.0,
|
|
"step": 485
|
|
},
|
|
{
|
|
"entropy": 7.742295646667481,
|
|
"epoch": 0.43497558810474923,
|
|
"grad_norm": 1.2890625,
|
|
"learning_rate": 0.0002445,
|
|
"loss": 7.3632,
|
|
"mean_token_accuracy": 0.0895160473883152,
|
|
"num_tokens": 1005547.0,
|
|
"step": 490
|
|
},
|
|
{
|
|
"entropy": 7.550811624526977,
|
|
"epoch": 0.4394141145139814,
|
|
"grad_norm": 1.171875,
|
|
"learning_rate": 0.000247,
|
|
"loss": 7.3413,
|
|
"mean_token_accuracy": 0.08750802129507065,
|
|
"num_tokens": 1015498.0,
|
|
"step": 495
|
|
},
|
|
{
|
|
"entropy": 7.609345579147339,
|
|
"epoch": 0.44385264092321347,
|
|
"grad_norm": 1.1015625,
|
|
"learning_rate": 0.0002495,
|
|
"loss": 7.4055,
|
|
"mean_token_accuracy": 0.08584300205111503,
|
|
"num_tokens": 1025420.0,
|
|
"step": 500
|
|
},
|
|
{
|
|
"epoch": 0.44385264092321347,
|
|
"eval_entropy": 7.669124412989673,
|
|
"eval_loss": 7.446435451507568,
|
|
"eval_mean_token_accuracy": 0.08312557260738132,
|
|
"eval_num_tokens": 1025420.0,
|
|
"eval_runtime": 2.7854,
|
|
"eval_samples_per_second": 1208.804,
|
|
"eval_steps_per_second": 151.145,
|
|
"step": 500
|
|
}
|
|
],
|
|
"logging_steps": 5,
|
|
"max_steps": 11260,
|
|
"num_input_tokens_seen": 0,
|
|
"num_train_epochs": 10,
|
|
"save_steps": 500,
|
|
"stateful_callbacks": {
|
|
"TrainerControl": {
|
|
"args": {
|
|
"should_epoch_stop": false,
|
|
"should_evaluate": false,
|
|
"should_log": false,
|
|
"should_save": true,
|
|
"should_training_stop": false
|
|
},
|
|
"attributes": {}
|
|
}
|
|
},
|
|
"total_flos": 1555912396800000.0,
|
|
"train_batch_size": 16,
|
|
"trial_name": null,
|
|
"trial_params": null
|
|
}
|