| { | |
| "best_metric": 1.7295050621032715, | |
| "best_model_checkpoint": "./llama3-itihasa/checkpoint-4000", | |
| "epoch": 1.9474196689386563, | |
| "eval_steps": 500, | |
| "global_step": 4000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.024342745861733205, | |
| "grad_norm": 0.798042893409729, | |
| "learning_rate": 0.00012165450121654502, | |
| "loss": 2.421, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.04868549172346641, | |
| "grad_norm": 0.6517059803009033, | |
| "learning_rate": 0.00024330900243309004, | |
| "loss": 2.0325, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.0730282375851996, | |
| "grad_norm": 1.047722339630127, | |
| "learning_rate": 0.00036496350364963507, | |
| "loss": 2.0045, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.09737098344693282, | |
| "grad_norm": 0.724214494228363, | |
| "learning_rate": 0.00048661800486618007, | |
| "loss": 1.9665, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.12171372930866602, | |
| "grad_norm": 0.7182360887527466, | |
| "learning_rate": 0.0006082725060827251, | |
| "loss": 1.9531, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.1460564751703992, | |
| "grad_norm": 0.741189181804657, | |
| "learning_rate": 0.0007299270072992701, | |
| "loss": 1.9566, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.17039922103213243, | |
| "grad_norm": 0.7193257212638855, | |
| "learning_rate": 0.0008515815085158151, | |
| "loss": 1.9536, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.19474196689386564, | |
| "grad_norm": 0.7618579864501953, | |
| "learning_rate": 0.0009732360097323601, | |
| "loss": 1.9642, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.21908471275559882, | |
| "grad_norm": 0.8242681622505188, | |
| "learning_rate": 0.0009894509061401135, | |
| "loss": 1.9899, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.24342745861733203, | |
| "grad_norm": 1.0397591590881348, | |
| "learning_rate": 0.0009759264268325669, | |
| "loss": 1.9819, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.24342745861733203, | |
| "eval_loss": 1.9741202592849731, | |
| "eval_runtime": 452.9914, | |
| "eval_samples_per_second": 3.022, | |
| "eval_steps_per_second": 0.757, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.26777020447906524, | |
| "grad_norm": 0.904543936252594, | |
| "learning_rate": 0.0009624019475250203, | |
| "loss": 1.9513, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 0.2921129503407984, | |
| "grad_norm": 0.8555233478546143, | |
| "learning_rate": 0.0009488774682174737, | |
| "loss": 1.9543, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.31645569620253167, | |
| "grad_norm": 0.7415252923965454, | |
| "learning_rate": 0.000935352988909927, | |
| "loss": 1.9621, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 0.34079844206426485, | |
| "grad_norm": 0.7195931077003479, | |
| "learning_rate": 0.0009218285096023803, | |
| "loss": 1.9414, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.36514118792599803, | |
| "grad_norm": 0.9943011999130249, | |
| "learning_rate": 0.0009083040302948337, | |
| "loss": 1.9594, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 0.3894839337877313, | |
| "grad_norm": 0.7514945268630981, | |
| "learning_rate": 0.000894779550987287, | |
| "loss": 1.9278, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.41382667964946446, | |
| "grad_norm": 0.7884072661399841, | |
| "learning_rate": 0.0008812550716797404, | |
| "loss": 1.9345, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 0.43816942551119764, | |
| "grad_norm": 0.7612814903259277, | |
| "learning_rate": 0.0008677305923721937, | |
| "loss": 1.9081, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.4625121713729309, | |
| "grad_norm": 0.7805455327033997, | |
| "learning_rate": 0.000854206113064647, | |
| "loss": 1.9188, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 0.48685491723466406, | |
| "grad_norm": 0.7169475555419922, | |
| "learning_rate": 0.0008406816337571004, | |
| "loss": 1.8903, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.48685491723466406, | |
| "eval_loss": 1.919243335723877, | |
| "eval_runtime": 452.9281, | |
| "eval_samples_per_second": 3.023, | |
| "eval_steps_per_second": 0.757, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.5111976630963972, | |
| "grad_norm": 0.7700228095054626, | |
| "learning_rate": 0.0008271571544495538, | |
| "loss": 1.8996, | |
| "step": 1050 | |
| }, | |
| { | |
| "epoch": 0.5355404089581305, | |
| "grad_norm": 0.8165948390960693, | |
| "learning_rate": 0.000813632675142007, | |
| "loss": 1.9178, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 0.5598831548198637, | |
| "grad_norm": 0.6771554350852966, | |
| "learning_rate": 0.0008001081958344604, | |
| "loss": 1.9012, | |
| "step": 1150 | |
| }, | |
| { | |
| "epoch": 0.5842259006815969, | |
| "grad_norm": 0.7270589470863342, | |
| "learning_rate": 0.0007865837165269137, | |
| "loss": 1.8971, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 0.6085686465433301, | |
| "grad_norm": 0.7590579390525818, | |
| "learning_rate": 0.0007730592372193671, | |
| "loss": 1.8842, | |
| "step": 1250 | |
| }, | |
| { | |
| "epoch": 0.6329113924050633, | |
| "grad_norm": 0.642905592918396, | |
| "learning_rate": 0.0007595347579118204, | |
| "loss": 1.8661, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 0.6572541382667965, | |
| "grad_norm": 0.7350940704345703, | |
| "learning_rate": 0.0007460102786042737, | |
| "loss": 1.8792, | |
| "step": 1350 | |
| }, | |
| { | |
| "epoch": 0.6815968841285297, | |
| "grad_norm": 0.7721772193908691, | |
| "learning_rate": 0.0007324857992967271, | |
| "loss": 1.8627, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 0.7059396299902629, | |
| "grad_norm": 0.7251712083816528, | |
| "learning_rate": 0.0007189613199891805, | |
| "loss": 1.8715, | |
| "step": 1450 | |
| }, | |
| { | |
| "epoch": 0.7302823758519961, | |
| "grad_norm": 0.6874309778213501, | |
| "learning_rate": 0.0007054368406816337, | |
| "loss": 1.8491, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.7302823758519961, | |
| "eval_loss": 1.8761951923370361, | |
| "eval_runtime": 452.8961, | |
| "eval_samples_per_second": 3.023, | |
| "eval_steps_per_second": 0.757, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.7546251217137293, | |
| "grad_norm": 0.7240548729896545, | |
| "learning_rate": 0.000691912361374087, | |
| "loss": 1.8585, | |
| "step": 1550 | |
| }, | |
| { | |
| "epoch": 0.7789678675754625, | |
| "grad_norm": 0.6931944489479065, | |
| "learning_rate": 0.0006783878820665405, | |
| "loss": 1.8623, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 0.8033106134371957, | |
| "grad_norm": 0.710396409034729, | |
| "learning_rate": 0.0006648634027589938, | |
| "loss": 1.8467, | |
| "step": 1650 | |
| }, | |
| { | |
| "epoch": 0.8276533592989289, | |
| "grad_norm": 0.7441169619560242, | |
| "learning_rate": 0.0006513389234514471, | |
| "loss": 1.857, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 0.8519961051606622, | |
| "grad_norm": 0.752155065536499, | |
| "learning_rate": 0.0006378144441439005, | |
| "loss": 1.8404, | |
| "step": 1750 | |
| }, | |
| { | |
| "epoch": 0.8763388510223953, | |
| "grad_norm": 0.6920920014381409, | |
| "learning_rate": 0.0006242899648363538, | |
| "loss": 1.8331, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 0.9006815968841285, | |
| "grad_norm": 0.6669381260871887, | |
| "learning_rate": 0.0006107654855288071, | |
| "loss": 1.8466, | |
| "step": 1850 | |
| }, | |
| { | |
| "epoch": 0.9250243427458618, | |
| "grad_norm": 0.6877081394195557, | |
| "learning_rate": 0.0005972410062212606, | |
| "loss": 1.829, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 0.9493670886075949, | |
| "grad_norm": 0.7491869926452637, | |
| "learning_rate": 0.0005837165269137138, | |
| "loss": 1.8331, | |
| "step": 1950 | |
| }, | |
| { | |
| "epoch": 0.9737098344693281, | |
| "grad_norm": 0.7513737082481384, | |
| "learning_rate": 0.0005701920476061671, | |
| "loss": 1.8106, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.9737098344693281, | |
| "eval_loss": 1.8369135856628418, | |
| "eval_runtime": 452.8951, | |
| "eval_samples_per_second": 3.023, | |
| "eval_steps_per_second": 0.757, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.9980525803310614, | |
| "grad_norm": 0.7290380001068115, | |
| "learning_rate": 0.0005566675682986206, | |
| "loss": 1.8111, | |
| "step": 2050 | |
| }, | |
| { | |
| "epoch": 1.0223953261927945, | |
| "grad_norm": 0.6029945611953735, | |
| "learning_rate": 0.0005431430889910739, | |
| "loss": 1.7597, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 1.0467380720545278, | |
| "grad_norm": 0.6136703491210938, | |
| "learning_rate": 0.0005296186096835271, | |
| "loss": 1.7358, | |
| "step": 2150 | |
| }, | |
| { | |
| "epoch": 1.071080817916261, | |
| "grad_norm": 0.6262673139572144, | |
| "learning_rate": 0.0005160941303759806, | |
| "loss": 1.742, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 1.095423563777994, | |
| "grad_norm": 0.7081790566444397, | |
| "learning_rate": 0.0005025696510684339, | |
| "loss": 1.7594, | |
| "step": 2250 | |
| }, | |
| { | |
| "epoch": 1.1197663096397275, | |
| "grad_norm": 0.6639809608459473, | |
| "learning_rate": 0.0004890451717608872, | |
| "loss": 1.7558, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 1.1441090555014606, | |
| "grad_norm": 0.7874799370765686, | |
| "learning_rate": 0.00047552069245334054, | |
| "loss": 1.7365, | |
| "step": 2350 | |
| }, | |
| { | |
| "epoch": 1.1684518013631937, | |
| "grad_norm": 0.6873753666877747, | |
| "learning_rate": 0.0004619962131457939, | |
| "loss": 1.7391, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 1.192794547224927, | |
| "grad_norm": 0.7038766741752625, | |
| "learning_rate": 0.00044847173383824726, | |
| "loss": 1.7223, | |
| "step": 2450 | |
| }, | |
| { | |
| "epoch": 1.2171372930866602, | |
| "grad_norm": 0.719927191734314, | |
| "learning_rate": 0.00043494725453070053, | |
| "loss": 1.7183, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 1.2171372930866602, | |
| "eval_loss": 1.80846107006073, | |
| "eval_runtime": 452.923, | |
| "eval_samples_per_second": 3.023, | |
| "eval_steps_per_second": 0.757, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 1.2414800389483933, | |
| "grad_norm": 0.62213134765625, | |
| "learning_rate": 0.0004214227752231539, | |
| "loss": 1.7484, | |
| "step": 2550 | |
| }, | |
| { | |
| "epoch": 1.2658227848101267, | |
| "grad_norm": 0.6441882252693176, | |
| "learning_rate": 0.00040789829591560725, | |
| "loss": 1.7351, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 1.2901655306718598, | |
| "grad_norm": 0.6303529739379883, | |
| "learning_rate": 0.0003943738166080606, | |
| "loss": 1.7175, | |
| "step": 2650 | |
| }, | |
| { | |
| "epoch": 1.314508276533593, | |
| "grad_norm": 0.6477542519569397, | |
| "learning_rate": 0.00038084933730051396, | |
| "loss": 1.7431, | |
| "step": 2700 | |
| }, | |
| { | |
| "epoch": 1.3388510223953263, | |
| "grad_norm": 0.6949708461761475, | |
| "learning_rate": 0.0003673248579929673, | |
| "loss": 1.7145, | |
| "step": 2750 | |
| }, | |
| { | |
| "epoch": 1.3631937682570594, | |
| "grad_norm": 0.681567907333374, | |
| "learning_rate": 0.0003538003786854206, | |
| "loss": 1.7041, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 1.3875365141187925, | |
| "grad_norm": 0.6230831146240234, | |
| "learning_rate": 0.00034027589937787395, | |
| "loss": 1.7028, | |
| "step": 2850 | |
| }, | |
| { | |
| "epoch": 1.4118792599805259, | |
| "grad_norm": 0.6681589484214783, | |
| "learning_rate": 0.00032675142007032733, | |
| "loss": 1.7034, | |
| "step": 2900 | |
| }, | |
| { | |
| "epoch": 1.436222005842259, | |
| "grad_norm": 0.5797551274299622, | |
| "learning_rate": 0.0003132269407627806, | |
| "loss": 1.705, | |
| "step": 2950 | |
| }, | |
| { | |
| "epoch": 1.4605647517039921, | |
| "grad_norm": 0.6354831457138062, | |
| "learning_rate": 0.000299702461455234, | |
| "loss": 1.7073, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 1.4605647517039921, | |
| "eval_loss": 1.7770596742630005, | |
| "eval_runtime": 452.7916, | |
| "eval_samples_per_second": 3.023, | |
| "eval_steps_per_second": 0.758, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 1.4849074975657255, | |
| "grad_norm": 0.6392273306846619, | |
| "learning_rate": 0.0002861779821476873, | |
| "loss": 1.6905, | |
| "step": 3050 | |
| }, | |
| { | |
| "epoch": 1.5092502434274586, | |
| "grad_norm": 0.6308421492576599, | |
| "learning_rate": 0.00027265350284014065, | |
| "loss": 1.6881, | |
| "step": 3100 | |
| }, | |
| { | |
| "epoch": 1.533592989289192, | |
| "grad_norm": 0.6558692455291748, | |
| "learning_rate": 0.000259129023532594, | |
| "loss": 1.6928, | |
| "step": 3150 | |
| }, | |
| { | |
| "epoch": 1.557935735150925, | |
| "grad_norm": 0.5978163480758667, | |
| "learning_rate": 0.0002456045442250473, | |
| "loss": 1.6833, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 1.5822784810126582, | |
| "grad_norm": 0.6277039051055908, | |
| "learning_rate": 0.0002320800649175007, | |
| "loss": 1.6819, | |
| "step": 3250 | |
| }, | |
| { | |
| "epoch": 1.6066212268743914, | |
| "grad_norm": 0.6343510746955872, | |
| "learning_rate": 0.00021855558560995403, | |
| "loss": 1.6876, | |
| "step": 3300 | |
| }, | |
| { | |
| "epoch": 1.6309639727361245, | |
| "grad_norm": 0.5587642192840576, | |
| "learning_rate": 0.00020503110630240738, | |
| "loss": 1.6881, | |
| "step": 3350 | |
| }, | |
| { | |
| "epoch": 1.6553067185978578, | |
| "grad_norm": 0.6643992066383362, | |
| "learning_rate": 0.0001915066269948607, | |
| "loss": 1.6897, | |
| "step": 3400 | |
| }, | |
| { | |
| "epoch": 1.6796494644595912, | |
| "grad_norm": 0.6108680367469788, | |
| "learning_rate": 0.00017798214768731404, | |
| "loss": 1.6712, | |
| "step": 3450 | |
| }, | |
| { | |
| "epoch": 1.7039922103213243, | |
| "grad_norm": 0.5982262492179871, | |
| "learning_rate": 0.0001644576683797674, | |
| "loss": 1.6737, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 1.7039922103213243, | |
| "eval_loss": 1.7497289180755615, | |
| "eval_runtime": 452.9263, | |
| "eval_samples_per_second": 3.023, | |
| "eval_steps_per_second": 0.757, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 1.7283349561830574, | |
| "grad_norm": 0.5886790156364441, | |
| "learning_rate": 0.00015093318907222073, | |
| "loss": 1.6783, | |
| "step": 3550 | |
| }, | |
| { | |
| "epoch": 1.7526777020447906, | |
| "grad_norm": 0.6495505571365356, | |
| "learning_rate": 0.00013740870976467406, | |
| "loss": 1.6753, | |
| "step": 3600 | |
| }, | |
| { | |
| "epoch": 1.7770204479065237, | |
| "grad_norm": 0.5986368656158447, | |
| "learning_rate": 0.00012388423045712742, | |
| "loss": 1.6758, | |
| "step": 3650 | |
| }, | |
| { | |
| "epoch": 1.801363193768257, | |
| "grad_norm": 0.5811861753463745, | |
| "learning_rate": 0.00011035975114958075, | |
| "loss": 1.66, | |
| "step": 3700 | |
| }, | |
| { | |
| "epoch": 1.8257059396299904, | |
| "grad_norm": 0.5493278503417969, | |
| "learning_rate": 9.683527184203409e-05, | |
| "loss": 1.6574, | |
| "step": 3750 | |
| }, | |
| { | |
| "epoch": 1.8500486854917235, | |
| "grad_norm": 0.558631181716919, | |
| "learning_rate": 8.331079253448742e-05, | |
| "loss": 1.68, | |
| "step": 3800 | |
| }, | |
| { | |
| "epoch": 1.8743914313534566, | |
| "grad_norm": 0.6028789281845093, | |
| "learning_rate": 6.978631322694076e-05, | |
| "loss": 1.661, | |
| "step": 3850 | |
| }, | |
| { | |
| "epoch": 1.8987341772151898, | |
| "grad_norm": 0.5945218801498413, | |
| "learning_rate": 5.6261833919394105e-05, | |
| "loss": 1.6498, | |
| "step": 3900 | |
| }, | |
| { | |
| "epoch": 1.9230769230769231, | |
| "grad_norm": 0.5483628511428833, | |
| "learning_rate": 4.273735461184745e-05, | |
| "loss": 1.6497, | |
| "step": 3950 | |
| }, | |
| { | |
| "epoch": 1.9474196689386563, | |
| "grad_norm": 0.5737510323524475, | |
| "learning_rate": 2.9212875304300785e-05, | |
| "loss": 1.6584, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 1.9474196689386563, | |
| "eval_loss": 1.7295050621032715, | |
| "eval_runtime": 452.9477, | |
| "eval_samples_per_second": 3.022, | |
| "eval_steps_per_second": 0.757, | |
| "step": 4000 | |
| } | |
| ], | |
| "logging_steps": 50, | |
| "max_steps": 4108, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 2, | |
| "save_steps": 1000, | |
| "total_flos": 7.390810477481165e+17, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |