{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 0.9991673605328892,
"eval_steps": 50,
"global_step": 450,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.011101859561476548,
"grad_norm": 1.140625,
"learning_rate": 2.222222222222222e-06,
"loss": 1.0928,
"mean_token_accuracy": 0.717196784696785,
"step": 5
},
{
"epoch": 0.022203719122953096,
"grad_norm": 1.1015625,
"learning_rate": 4.444444444444444e-06,
"loss": 1.1024,
"mean_token_accuracy": 0.7158291392437733,
"step": 10
},
{
"epoch": 0.03330557868442964,
"grad_norm": 1.03125,
"learning_rate": 6.666666666666667e-06,
"loss": 1.0782,
"mean_token_accuracy": 0.7194658119658119,
"step": 15
},
{
"epoch": 0.04440743824590619,
"grad_norm": 0.96484375,
"learning_rate": 8.888888888888888e-06,
"loss": 1.0777,
"mean_token_accuracy": 0.7197588522588524,
"step": 20
},
{
"epoch": 0.055509297807382736,
"grad_norm": 0.7734375,
"learning_rate": 1.1111111111111113e-05,
"loss": 1.0614,
"mean_token_accuracy": 0.7216748066748067,
"step": 25
},
{
"epoch": 0.06661115736885928,
"grad_norm": 0.6796875,
"learning_rate": 1.3333333333333333e-05,
"loss": 1.059,
"mean_token_accuracy": 0.719067969067969,
"step": 30
},
{
"epoch": 0.07771301693033583,
"grad_norm": 0.6171875,
"learning_rate": 1.555555555555556e-05,
"loss": 1.0328,
"mean_token_accuracy": 0.7237962962962966,
"step": 35
},
{
"epoch": 0.08881487649181238,
"grad_norm": 0.55859375,
"learning_rate": 1.7777777777777777e-05,
"loss": 1.014,
"mean_token_accuracy": 0.7259676434676435,
"step": 40
},
{
"epoch": 0.09991673605328892,
"grad_norm": 0.46484375,
"learning_rate": 2e-05,
"loss": 1.0066,
"mean_token_accuracy": 0.725960520960521,
"step": 45
},
{
"epoch": 0.11101859561476547,
"grad_norm": 0.4140625,
"learning_rate": 1.9992479525042305e-05,
"loss": 0.9537,
"mean_token_accuracy": 0.7386558811558809,
"step": 50
},
{
"epoch": 0.11101859561476547,
"eval_loss": 0.9796226024627686,
"eval_mean_token_accuracy": 0.7271781595311009,
"eval_runtime": 9.7194,
"eval_samples_per_second": 13.272,
"eval_steps_per_second": 3.395,
"step": 50
},
{
"epoch": 0.12212045517624202,
"grad_norm": 0.359375,
"learning_rate": 1.996992941167792e-05,
"loss": 0.9602,
"mean_token_accuracy": 0.7342826617826618,
"step": 55
},
{
"epoch": 0.13322231473771856,
"grad_norm": 0.326171875,
"learning_rate": 1.9932383577419432e-05,
"loss": 0.916,
"mean_token_accuracy": 0.74437307661244,
"step": 60
},
{
"epoch": 0.1443241742991951,
"grad_norm": 0.3515625,
"learning_rate": 1.9879898494768093e-05,
"loss": 0.909,
"mean_token_accuracy": 0.7461548636548637,
"step": 65
},
{
"epoch": 0.15542603386067166,
"grad_norm": 0.3125,
"learning_rate": 1.9812553106273848e-05,
"loss": 0.9079,
"mean_token_accuracy": 0.74519129019129,
"step": 70
},
{
"epoch": 0.16652789342214822,
"grad_norm": 0.333984375,
"learning_rate": 1.973044870579824e-05,
"loss": 0.9071,
"mean_token_accuracy": 0.7446357346357346,
"step": 75
},
{
"epoch": 0.17762975298362477,
"grad_norm": 0.30078125,
"learning_rate": 1.9633708786158803e-05,
"loss": 0.8953,
"mean_token_accuracy": 0.746184371184371,
"step": 80
},
{
"epoch": 0.18873161254510132,
"grad_norm": 0.275390625,
"learning_rate": 1.9522478853384154e-05,
"loss": 0.8582,
"mean_token_accuracy": 0.7562728937728938,
"step": 85
},
{
"epoch": 0.19983347210657784,
"grad_norm": 0.296875,
"learning_rate": 1.9396926207859085e-05,
"loss": 0.8719,
"mean_token_accuracy": 0.7528663003663003,
"step": 90
},
{
"epoch": 0.2109353316680544,
"grad_norm": 0.271484375,
"learning_rate": 1.9257239692688907e-05,
"loss": 0.8735,
"mean_token_accuracy": 0.7512274542096703,
"step": 95
},
{
"epoch": 0.22203719122953094,
"grad_norm": 0.2734375,
"learning_rate": 1.9103629409661468e-05,
"loss": 0.8656,
"mean_token_accuracy": 0.7536752136752136,
"step": 100
},
{
"epoch": 0.22203719122953094,
"eval_loss": 0.9016062617301941,
"eval_mean_token_accuracy": 0.7405376758317936,
"eval_runtime": 9.7188,
"eval_samples_per_second": 13.273,
"eval_steps_per_second": 3.395,
"step": 100
},
{
"epoch": 0.2331390507910075,
"grad_norm": 0.2734375,
"learning_rate": 1.8936326403234125e-05,
"loss": 0.8687,
"mean_token_accuracy": 0.7531064306064306,
"step": 105
},
{
"epoch": 0.24424091035248405,
"grad_norm": 0.267578125,
"learning_rate": 1.8755582313020912e-05,
"loss": 0.8548,
"mean_token_accuracy": 0.75515422674015,
"step": 110
},
{
"epoch": 0.2553427699139606,
"grad_norm": 0.26953125,
"learning_rate": 1.8561668995302668e-05,
"loss": 0.8543,
"mean_token_accuracy": 0.7542501017501018,
"step": 115
},
{
"epoch": 0.2664446294754371,
"grad_norm": 0.26953125,
"learning_rate": 1.8354878114129368e-05,
"loss": 0.8373,
"mean_token_accuracy": 0.7589896214896215,
"step": 120
},
{
"epoch": 0.2775464890369137,
"grad_norm": 0.265625,
"learning_rate": 1.8135520702629677e-05,
"loss": 0.8489,
"mean_token_accuracy": 0.7568091168091166,
"step": 125
},
{
"epoch": 0.2886483485983902,
"grad_norm": 0.267578125,
"learning_rate": 1.7903926695187595e-05,
"loss": 0.8557,
"mean_token_accuracy": 0.7542979242979242,
"step": 130
},
{
"epoch": 0.2997502081598668,
"grad_norm": 0.26171875,
"learning_rate": 1.766044443118978e-05,
"loss": 0.8329,
"mean_token_accuracy": 0.7595054945054944,
"step": 135
},
{
"epoch": 0.31085206772134333,
"grad_norm": 0.2490234375,
"learning_rate": 1.740544013109005e-05,
"loss": 0.874,
"mean_token_accuracy": 0.748472730972731,
"step": 140
},
{
"epoch": 0.32195392728281985,
"grad_norm": 0.255859375,
"learning_rate": 1.7139297345578992e-05,
"loss": 0.8295,
"mean_token_accuracy": 0.7612077737077737,
"step": 145
},
{
"epoch": 0.33305578684429643,
"grad_norm": 0.251953125,
"learning_rate": 1.686241637868734e-05,
"loss": 0.8092,
"mean_token_accuracy": 0.7669424094424093,
"step": 150
},
{
"epoch": 0.33305578684429643,
"eval_loss": 0.8762778639793396,
"eval_mean_token_accuracy": 0.7449022654905009,
"eval_runtime": 9.7018,
"eval_samples_per_second": 13.296,
"eval_steps_per_second": 3.401,
"step": 150
},
{
"epoch": 0.34415764640577295,
"grad_norm": 0.251953125,
"learning_rate": 1.657521368569064e-05,
"loss": 0.8355,
"mean_token_accuracy": 0.7576159951159951,
"step": 155
},
{
"epoch": 0.35525950596724953,
"grad_norm": 0.2578125,
"learning_rate": 1.627812124672099e-05,
"loss": 0.8398,
"mean_token_accuracy": 0.7564778061483249,
"step": 160
},
{
"epoch": 0.36636136552872606,
"grad_norm": 0.251953125,
"learning_rate": 1.5971585917027864e-05,
"loss": 0.8576,
"mean_token_accuracy": 0.7524267399267397,
"step": 165
},
{
"epoch": 0.37746322509020264,
"grad_norm": 0.2578125,
"learning_rate": 1.5656068754865388e-05,
"loss": 0.8254,
"mean_token_accuracy": 0.760271918721826,
"step": 170
},
{
"epoch": 0.38856508465167916,
"grad_norm": 0.255859375,
"learning_rate": 1.5332044328016916e-05,
"loss": 0.8502,
"mean_token_accuracy": 0.7546072446072445,
"step": 175
},
{
"epoch": 0.3996669442131557,
"grad_norm": 0.25390625,
"learning_rate": 1.5000000000000002e-05,
"loss": 0.8418,
"mean_token_accuracy": 0.7569129019129018,
"step": 180
},
{
"epoch": 0.41076880377463226,
"grad_norm": 0.255859375,
"learning_rate": 1.4660435197025391e-05,
"loss": 0.8346,
"mean_token_accuracy": 0.7576149776149775,
"step": 185
},
{
"epoch": 0.4218706633361088,
"grad_norm": 0.24609375,
"learning_rate": 1.4313860656812537e-05,
"loss": 0.8478,
"mean_token_accuracy": 0.7535816035816034,
"step": 190
},
{
"epoch": 0.43297252289758537,
"grad_norm": 0.259765625,
"learning_rate": 1.396079766039157e-05,
"loss": 0.8107,
"mean_token_accuracy": 0.7654395604395604,
"step": 195
},
{
"epoch": 0.4440743824590619,
"grad_norm": 0.251953125,
"learning_rate": 1.3601777248047105e-05,
"loss": 0.8124,
"mean_token_accuracy": 0.763194953194953,
"step": 200
},
{
"epoch": 0.4440743824590619,
"eval_loss": 0.8637130260467529,
"eval_mean_token_accuracy": 0.7472996502408268,
"eval_runtime": 9.6971,
"eval_samples_per_second": 13.303,
"eval_steps_per_second": 3.403,
"step": 200
},
{
"epoch": 0.4551762420205384,
"grad_norm": 0.25390625,
"learning_rate": 1.3237339420583213e-05,
"loss": 0.8375,
"mean_token_accuracy": 0.757132682132682,
"step": 205
},
{
"epoch": 0.466278101582015,
"grad_norm": 0.251953125,
"learning_rate": 1.2868032327110904e-05,
"loss": 0.8577,
"mean_token_accuracy": 0.750907610907611,
"step": 210
},
{
"epoch": 0.4773799611434915,
"grad_norm": 0.2431640625,
"learning_rate": 1.2494411440579814e-05,
"loss": 0.82,
"mean_token_accuracy": 0.7625814000814001,
"step": 215
},
{
"epoch": 0.4884818207049681,
"grad_norm": 0.25390625,
"learning_rate": 1.211703872229411e-05,
"loss": 0.8099,
"mean_token_accuracy": 0.7646743996743997,
"step": 220
},
{
"epoch": 0.4995836802664446,
"grad_norm": 0.2490234375,
"learning_rate": 1.1736481776669307e-05,
"loss": 0.8195,
"mean_token_accuracy": 0.7627518372346456,
"step": 225
},
{
"epoch": 0.5106855398279212,
"grad_norm": 0.2490234375,
"learning_rate": 1.1353312997501313e-05,
"loss": 0.8248,
"mean_token_accuracy": 0.7611874236874236,
"step": 230
},
{
"epoch": 0.5217873993893978,
"grad_norm": 0.23828125,
"learning_rate": 1.0968108707031792e-05,
"loss": 0.8174,
"mean_token_accuracy": 0.7629131054131055,
"step": 235
},
{
"epoch": 0.5328892589508742,
"grad_norm": 0.244140625,
"learning_rate": 1.0581448289104759e-05,
"loss": 0.8105,
"mean_token_accuracy": 0.7650734632542214,
"step": 240
},
{
"epoch": 0.5439911185123508,
"grad_norm": 0.244140625,
"learning_rate": 1.0193913317718245e-05,
"loss": 0.81,
"mean_token_accuracy": 0.7640618640618639,
"step": 245
},
{
"epoch": 0.5550929780738274,
"grad_norm": 0.2421875,
"learning_rate": 9.806086682281759e-06,
"loss": 0.8205,
"mean_token_accuracy": 0.7610907610907611,
"step": 250
},
{
"epoch": 0.5550929780738274,
"eval_loss": 0.8572535514831543,
"eval_mean_token_accuracy": 0.748435551376728,
"eval_runtime": 9.7051,
"eval_samples_per_second": 13.292,
"eval_steps_per_second": 3.4,
"step": 250
},
{
"epoch": 0.5661948376353039,
"grad_norm": 0.255859375,
"learning_rate": 9.418551710895243e-06,
"loss": 0.8235,
"mean_token_accuracy": 0.7599084249084251,
"step": 255
},
{
"epoch": 0.5772966971967805,
"grad_norm": 0.251953125,
"learning_rate": 9.03189129296821e-06,
"loss": 0.8286,
"mean_token_accuracy": 0.7585877085877086,
"step": 260
},
{
"epoch": 0.588398556758257,
"grad_norm": 0.259765625,
"learning_rate": 8.646687002498692e-06,
"loss": 0.8256,
"mean_token_accuracy": 0.7592775742775744,
"step": 265
},
{
"epoch": 0.5995004163197336,
"grad_norm": 0.251953125,
"learning_rate": 8.263518223330698e-06,
"loss": 0.8217,
"mean_token_accuracy": 0.7614479039479042,
"step": 270
},
{
"epoch": 0.6106022758812101,
"grad_norm": 0.2431640625,
"learning_rate": 7.882961277705897e-06,
"loss": 0.8354,
"mean_token_accuracy": 0.7570135306548853,
"step": 275
},
{
"epoch": 0.6217041354426867,
"grad_norm": 0.2490234375,
"learning_rate": 7.505588559420188e-06,
"loss": 0.8222,
"mean_token_accuracy": 0.7614336589336589,
"step": 280
},
{
"epoch": 0.6328059950041632,
"grad_norm": 0.244140625,
"learning_rate": 7.131967672889101e-06,
"loss": 0.8579,
"mean_token_accuracy": 0.7512146494805758,
"step": 285
},
{
"epoch": 0.6439078545656397,
"grad_norm": 0.25390625,
"learning_rate": 6.762660579416791e-06,
"loss": 0.8253,
"mean_token_accuracy": 0.7587220187220187,
"step": 290
},
{
"epoch": 0.6550097141271163,
"grad_norm": 0.2412109375,
"learning_rate": 6.3982227519528986e-06,
"loss": 0.8294,
"mean_token_accuracy": 0.7600905599174735,
"step": 295
},
{
"epoch": 0.6661115736885929,
"grad_norm": 0.2470703125,
"learning_rate": 6.039202339608432e-06,
"loss": 0.8129,
"mean_token_accuracy": 0.763727106227106,
"step": 300
},
{
"epoch": 0.6661115736885929,
"eval_loss": 0.854372501373291,
"eval_mean_token_accuracy": 0.7488219841161018,
"eval_runtime": 9.716,
"eval_samples_per_second": 13.277,
"eval_steps_per_second": 3.396,
"step": 300
},
{
"epoch": 0.6772134332500694,
"grad_norm": 0.2431640625,
"learning_rate": 5.686139343187468e-06,
"loss": 0.8089,
"mean_token_accuracy": 0.7635978835978834,
"step": 305
},
{
"epoch": 0.6883152928115459,
"grad_norm": 0.2421875,
"learning_rate": 5.339564802974615e-06,
"loss": 0.8162,
"mean_token_accuracy": 0.7625620675620676,
"step": 310
},
{
"epoch": 0.6994171523730225,
"grad_norm": 0.25,
"learning_rate": 5.000000000000003e-06,
"loss": 0.818,
"mean_token_accuracy": 0.7617165242165239,
"step": 315
},
{
"epoch": 0.7105190119344991,
"grad_norm": 0.251953125,
"learning_rate": 4.66795567198309e-06,
"loss": 0.8164,
"mean_token_accuracy": 0.7618395600557701,
"step": 320
},
{
"epoch": 0.7216208714959755,
"grad_norm": 0.248046875,
"learning_rate": 4.343931245134616e-06,
"loss": 0.8391,
"mean_token_accuracy": 0.7558740333740334,
"step": 325
},
{
"epoch": 0.7327227310574521,
"grad_norm": 0.251953125,
"learning_rate": 4.028414082972141e-06,
"loss": 0.8288,
"mean_token_accuracy": 0.7581603581603582,
"step": 330
},
{
"epoch": 0.7438245906189287,
"grad_norm": 0.2470703125,
"learning_rate": 3.7218787532790167e-06,
"loss": 0.8266,
"mean_token_accuracy": 0.758972323972324,
"step": 335
},
{
"epoch": 0.7549264501804053,
"grad_norm": 0.23828125,
"learning_rate": 3.424786314309365e-06,
"loss": 0.8169,
"mean_token_accuracy": 0.7615511468202194,
"step": 340
},
{
"epoch": 0.7660283097418817,
"grad_norm": 0.248046875,
"learning_rate": 3.1375836213126653e-06,
"loss": 0.8207,
"mean_token_accuracy": 0.7615923890923892,
"step": 345
},
{
"epoch": 0.7771301693033583,
"grad_norm": 0.23828125,
"learning_rate": 2.8607026544210115e-06,
"loss": 0.8251,
"mean_token_accuracy": 0.7595614570614571,
"step": 350
},
{
"epoch": 0.7771301693033583,
"eval_loss": 0.8534859418869019,
"eval_mean_token_accuracy": 0.7491665197547551,
"eval_runtime": 9.7117,
"eval_samples_per_second": 13.283,
"eval_steps_per_second": 3.398,
"step": 350
},
{
"epoch": 0.7882320288648349,
"grad_norm": 0.2314453125,
"learning_rate": 2.594559868909956e-06,
"loss": 0.789,
"mean_token_accuracy": 0.7690638990638987,
"step": 355
},
{
"epoch": 0.7993338884263114,
"grad_norm": 0.2490234375,
"learning_rate": 2.339555568810221e-06,
"loss": 0.8195,
"mean_token_accuracy": 0.76007733007733,
"step": 360
},
{
"epoch": 0.810435747987788,
"grad_norm": 0.2578125,
"learning_rate": 2.0960733048124082e-06,
"loss": 0.8358,
"mean_token_accuracy": 0.7567429792429792,
"step": 365
},
{
"epoch": 0.8215376075492645,
"grad_norm": 0.234375,
"learning_rate": 1.8644792973703252e-06,
"loss": 0.797,
"mean_token_accuracy": 0.7672700447700449,
"step": 370
},
{
"epoch": 0.832639467110741,
"grad_norm": 0.2333984375,
"learning_rate": 1.6451218858706374e-06,
"loss": 0.8269,
"mean_token_accuracy": 0.7588715913715915,
"step": 375
},
{
"epoch": 0.8437413266722176,
"grad_norm": 0.244140625,
"learning_rate": 1.4383310046973365e-06,
"loss": 0.8243,
"mean_token_accuracy": 0.7594324261790015,
"step": 380
},
{
"epoch": 0.8548431862336942,
"grad_norm": 0.2412109375,
"learning_rate": 1.2444176869790925e-06,
"loss": 0.8229,
"mean_token_accuracy": 0.7609747659747659,
"step": 385
},
{
"epoch": 0.8659450457951707,
"grad_norm": 0.24609375,
"learning_rate": 1.0636735967658785e-06,
"loss": 0.804,
"mean_token_accuracy": 0.7651831501831503,
"step": 390
},
{
"epoch": 0.8770469053566472,
"grad_norm": 0.2470703125,
"learning_rate": 8.963705903385344e-07,
"loss": 0.8512,
"mean_token_accuracy": 0.7523127798127797,
"step": 395
},
{
"epoch": 0.8881487649181238,
"grad_norm": 0.259765625,
"learning_rate": 7.427603073110967e-07,
"loss": 0.8437,
"mean_token_accuracy": 0.7555453805453806,
"step": 400
},
{
"epoch": 0.8881487649181238,
"eval_loss": 0.8532679080963135,
"eval_mean_token_accuracy": 0.748890434184552,
"eval_runtime": 9.7147,
"eval_samples_per_second": 13.279,
"eval_steps_per_second": 3.397,
"step": 400
},
{
"epoch": 0.8992506244796004,
"grad_norm": 0.2490234375,
"learning_rate": 6.030737921409169e-07,
"loss": 0.8155,
"mean_token_accuracy": 0.7616107041107041,
"step": 405
},
{
"epoch": 0.9103524840410768,
"grad_norm": 0.25390625,
"learning_rate": 4.775211466158469e-07,
"loss": 0.8045,
"mean_token_accuracy": 0.7646845746845747,
"step": 410
},
{
"epoch": 0.9214543436025534,
"grad_norm": 0.2412109375,
"learning_rate": 3.662912138411967e-07,
"loss": 0.848,
"mean_token_accuracy": 0.7535592185592186,
"step": 415
},
{
"epoch": 0.93255620316403,
"grad_norm": 0.2412109375,
"learning_rate": 2.6955129420176193e-07,
"loss": 0.796,
"mean_token_accuracy": 0.7675030525030523,
"step": 420
},
{
"epoch": 0.9436580627255066,
"grad_norm": 0.2451171875,
"learning_rate": 1.874468937261531e-07,
"loss": 0.8322,
"mean_token_accuracy": 0.7581685062193277,
"step": 425
},
{
"epoch": 0.954759922286983,
"grad_norm": 0.265625,
"learning_rate": 1.201015052319099e-07,
"loss": 0.8023,
"mean_token_accuracy": 0.7657336182336185,
"step": 430
},
{
"epoch": 0.9658617818484596,
"grad_norm": 0.23828125,
"learning_rate": 6.761642258056977e-08,
"loss": 0.8094,
"mean_token_accuracy": 0.7642338217338217,
"step": 435
},
{
"epoch": 0.9769636414099362,
"grad_norm": 0.2392578125,
"learning_rate": 3.0070588322079765e-08,
"loss": 0.8206,
"mean_token_accuracy": 0.7607519332519332,
"step": 440
},
{
"epoch": 0.9880655009714127,
"grad_norm": 0.2490234375,
"learning_rate": 7.520474957699586e-09,
"loss": 0.8151,
"mean_token_accuracy": 0.7614092389092388,
"step": 445
},
{
"epoch": 0.9991673605328892,
"grad_norm": 0.23828125,
"learning_rate": 0.0,
"loss": 0.7954,
"mean_token_accuracy": 0.7673087098087098,
"step": 450
},
{
"epoch": 0.9991673605328892,
"eval_loss": 0.8533338904380798,
"eval_mean_token_accuracy": 0.748884884179002,
"eval_runtime": 9.7042,
"eval_samples_per_second": 13.293,
"eval_steps_per_second": 3.401,
"step": 450
},
{
"epoch": 0.9991673605328892,
"step": 450,
"total_flos": 6.955833048956928e+17,
"train_loss": 0.8593901687198215,
"train_runtime": 5693.7016,
"train_samples_per_second": 3.797,
"train_steps_per_second": 0.079
}
],
"logging_steps": 5,
"max_steps": 450,
"num_input_tokens_seen": 0,
"num_train_epochs": 1,
"save_steps": 500,
"stateful_callbacks": {
"TrainerControl": {
"args": {
"should_epoch_stop": false,
"should_evaluate": false,
"should_log": false,
"should_save": false,
"should_training_stop": false
},
"attributes": {}
}
},
"total_flos": 6.955833048956928e+17,
"train_batch_size": 6,
"trial_name": null,
"trial_params": null
}