{"wall": 95.272, "eval_loss": 10.906768560409546, "eval_ppl": 54544.31640625, "step": 0} {"wall": 419.777, "step": 10, "loss": 10.60250965754191, "lm_loss": 10.60250965754191, "mtp_loss": 0.0, "lr": 0.0026077586206896554, "grad_norm": 41.252960205078125, "tokens": 2949120, "adamw_lr": 4.172413793103448e-05, "mem_alloc_gb": 14.685915470123291, "mem_reserv_gb": 79.091796875, "loss_smoothed": 10.776294565200805, "seq_len": 4096} {"wall": 624.401, "step": 20, "loss": 10.126710494359335, "lm_loss": 10.126710494359335, "mtp_loss": 0.0, "lr": 0.002715517241379311, "grad_norm": 47.74290084838867, "tokens": 5898240, "adamw_lr": 4.344827586206897e-05, "mem_alloc_gb": 14.685915470123291, "mem_reserv_gb": 79.091796875, "loss_smoothed": 10.3451526761055, "seq_len": 4096} {"wall": 828.915, "step": 30, "loss": 9.625130295753479, "lm_loss": 9.625130295753479, "mtp_loss": 0.0, "lr": 0.0028232758620689658, "grad_norm": 44.3715705871582, "tokens": 8847360, "adamw_lr": 4.5172413793103454e-05, "mem_alloc_gb": 14.685915470123291, "mem_reserv_gb": 79.091796875, "loss_smoothed": 9.832994333902995, "seq_len": 4096} {"wall": 1033.589, "step": 40, "loss": 9.33853797117869, "lm_loss": 9.33853797117869, "mtp_loss": 0.0, "lr": 0.002931034482758621, "grad_norm": 40.05640411376953, "tokens": 11796480, "adamw_lr": 4.689655172413794e-05, "mem_alloc_gb": 14.685915470123291, "mem_reserv_gb": 79.091796875, "loss_smoothed": 9.472525890668233, "seq_len": 4096} {"wall": 1238.411, "step": 50, "loss": 9.0480348666509, "lm_loss": 9.0480348666509, "mtp_loss": 0.0, "lr": 0.003038793103448276, "grad_norm": 33.76705551147461, "tokens": 14745600, "adamw_lr": 4.862068965517242e-05, "mem_alloc_gb": 14.685915470123291, "mem_reserv_gb": 79.091796875, "loss_smoothed": 9.20002806186676, "seq_len": 4096} {"wall": 1442.854, "step": 60, "loss": 8.854629317919413, "lm_loss": 8.854629317919413, "mtp_loss": 0.0, "lr": 0.003146551724137931, "grad_norm": 19.726198196411133, "tokens": 17694720, "adamw_lr": 5.03448275862069e-05, "mem_alloc_gb": 14.685915470123291, "mem_reserv_gb": 79.091796875, "loss_smoothed": 8.887987152735393, "seq_len": 4096} {"wall": 1647.638, "step": 70, "loss": 8.517783562342325, "lm_loss": 8.517783562342325, "mtp_loss": 0.0, "lr": 0.0032543103448275865, "grad_norm": 8.524994850158691, "tokens": 20643840, "adamw_lr": 5.2068965517241384e-05, "mem_alloc_gb": 14.685915470123291, "mem_reserv_gb": 79.091796875, "loss_smoothed": 8.663269186019898, "seq_len": 4096} {"wall": 1852.149, "step": 80, "loss": 8.281542678674063, "lm_loss": 8.281542678674063, "mtp_loss": 0.0, "lr": 0.0033620689655172414, "grad_norm": 4.6403913497924805, "tokens": 23592960, "adamw_lr": 5.379310344827586e-05, "mem_alloc_gb": 14.685915470123291, "mem_reserv_gb": 79.091796875, "loss_smoothed": 8.442335818211237, "seq_len": 4096} {"wall": 2056.764, "step": 90, "loss": 8.10840779542923, "lm_loss": 8.10840779542923, "mtp_loss": 0.0, "lr": 0.003469827586206897, "grad_norm": 4.253584861755371, "tokens": 26542080, "adamw_lr": 5.551724137931035e-05, "mem_alloc_gb": 14.685915470123291, "mem_reserv_gb": 79.091796875, "loss_smoothed": 8.193158115943273, "seq_len": 4096} {"wall": 2263.641, "eval_loss": 7.587542653083801, "eval_ppl": 1973.4578857421875, "step": 100} {"wall": 2263.643, "step": 100, "loss": 7.6973557472229, "lm_loss": 7.6973557472229, "mtp_loss": 0.0, "lr": 0.003577586206896552, "grad_norm": 3.319485902786255, "tokens": 29491200, "adamw_lr": 5.7241379310344835e-05, "mem_alloc_gb": 14.685915470123291, "mem_reserv_gb": 79.091796875, "loss_smoothed": 7.927917273839315, "seq_len": 4096} {"wall": 2483.344, "step": 110, "loss": 7.373127996921539, "lm_loss": 7.373127996921539, "mtp_loss": 0.0, "lr": 0.0036853448275862076, "grad_norm": 3.933872699737549, "tokens": 32440320, "adamw_lr": 5.896551724137932e-05, "mem_alloc_gb": 14.685915470123291, "mem_reserv_gb": 79.091796875, "loss_smoothed": 7.5819320778052015, "seq_len": 4096} {"wall": 2687.851, "step": 120, "loss": 6.970027923583984, "lm_loss": 6.970027923583984, "mtp_loss": 0.0, "lr": 0.003793103448275862, "grad_norm": 2.7088065147399902, "tokens": 35389440, "adamw_lr": 6.068965517241379e-05, "mem_alloc_gb": 14.685915470123291, "mem_reserv_gb": 79.091796875, "loss_smoothed": 7.176820496718089, "seq_len": 4096} {"wall": 2892.671, "step": 130, "loss": 6.61169562737147, "lm_loss": 6.61169562737147, "mtp_loss": 0.0, "lr": 0.0039008620689655175, "grad_norm": 2.1832492351531982, "tokens": 38338560, "adamw_lr": 6.241379310344829e-05, "mem_alloc_gb": 14.685915470123291, "mem_reserv_gb": 79.091796875, "loss_smoothed": 6.7989161014556885, "seq_len": 4096} {"wall": 3097.007, "step": 140, "loss": 6.321819464365642, "lm_loss": 6.321819464365642, "mtp_loss": 0.0, "lr": 0.0040086206896551724, "grad_norm": 2.087441921234131, "tokens": 41287680, "adamw_lr": 6.413793103448276e-05, "mem_alloc_gb": 14.685915470123291, "mem_reserv_gb": 79.091796875, "loss_smoothed": 6.478065655628841, "seq_len": 4096} {"wall": 3301.756, "step": 150, "loss": 5.81686665614446, "lm_loss": 5.81686665614446, "mtp_loss": 0.0, "lr": 0.004116379310344829, "grad_norm": 2.1521246433258057, "tokens": 44236800, "adamw_lr": 6.586206896551726e-05, "mem_alloc_gb": 14.685915470123291, "mem_reserv_gb": 79.091796875, "loss_smoothed": 6.217310984929402, "seq_len": 4096} {"wall": 3506.621, "step": 160, "loss": 5.797270357608795, "lm_loss": 5.797270357608795, "mtp_loss": 0.0, "lr": 0.004224137931034483, "grad_norm": 1.5286247730255127, "tokens": 47185920, "adamw_lr": 6.758620689655173e-05, "mem_alloc_gb": 14.685915470123291, "mem_reserv_gb": 79.091796875, "loss_smoothed": 5.923025580247243, "seq_len": 4096} {"wall": 3710.934, "step": 170, "loss": 5.618978142738342, "lm_loss": 5.618978142738342, "mtp_loss": 0.0, "lr": 0.004331896551724138, "grad_norm": 14.21914005279541, "tokens": 50135040, "adamw_lr": 6.93103448275862e-05, "mem_alloc_gb": 14.685915470123291, "mem_reserv_gb": 79.091796875, "loss_smoothed": 5.749992540478706, "seq_len": 4096} {"wall": 3915.971, "step": 180, "loss": 5.459208190441132, "lm_loss": 5.459208190441132, "mtp_loss": 0.0, "lr": 0.004439655172413794, "grad_norm": 6.371452331542969, "tokens": 53084160, "adamw_lr": 7.10344827586207e-05, "mem_alloc_gb": 14.685915470123291, "mem_reserv_gb": 79.091796875, "loss_smoothed": 5.48839204609394, "seq_len": 4096} {"wall": 4120.633, "step": 190, "loss": 5.325698763132095, "lm_loss": 5.325698763132095, "mtp_loss": 0.0, "lr": 0.0045474137931034485, "grad_norm": 1.407970905303955, "tokens": 56033280, "adamw_lr": 7.275862068965517e-05, "mem_alloc_gb": 14.685915470123291, "mem_reserv_gb": 79.091796875, "loss_smoothed": 5.2632093598445255, "seq_len": 4096} {"wall": 4327.692, "eval_loss": 4.719881325960159, "eval_ppl": 112.15496826171875, "step": 200} {"wall": 4327.694, "step": 200, "loss": 5.145787676175435, "lm_loss": 5.145787676175435, "mtp_loss": 0.0, "lr": 0.004655172413793104, "grad_norm": 1.3314249515533447, "tokens": 58982400, "adamw_lr": 7.448275862068966e-05, "mem_alloc_gb": 14.685915470123291, "mem_reserv_gb": 79.091796875, "loss_smoothed": 5.076351340611775, "seq_len": 4096} {"wall": 4618.35, "step": 210, "loss": 4.9645954966545105, "lm_loss": 4.9645954966545105, "mtp_loss": 0.0, "lr": 0.004762931034482759, "grad_norm": 1.3545795679092407, "tokens": 61931520, "adamw_lr": 7.620689655172415e-05, "mem_alloc_gb": 14.685915470123291, "mem_reserv_gb": 79.091796875, "loss_smoothed": 4.979698797067007, "seq_len": 4096} {"wall": 4822.894, "step": 220, "loss": 4.709302713473638, "lm_loss": 4.709302713473638, "mtp_loss": 0.0, "lr": 0.004870689655172414, "grad_norm": 1.1894505023956299, "tokens": 64880640, "adamw_lr": 7.793103448275862e-05, "mem_alloc_gb": 14.685915470123291, "mem_reserv_gb": 79.091796875, "loss_smoothed": 4.869353184103966, "seq_len": 4096} {"wall": 5027.582, "step": 230, "loss": 4.778348823388417, "lm_loss": 4.778348823388417, "mtp_loss": 0.0, "lr": 0.004978448275862069, "grad_norm": 1.0662766695022583, "tokens": 67829760, "adamw_lr": 7.965517241379312e-05, "mem_alloc_gb": 14.685915470123291, "mem_reserv_gb": 79.091796875, "loss_smoothed": 4.7800374041001, "seq_len": 4096} {"wall": 5232.536, "step": 240, "loss": 4.432990938425064, "lm_loss": 4.432990938425064, "mtp_loss": 0.0, "lr": 0.005086206896551725, "grad_norm": 1.0933345556259155, "tokens": 70778880, "adamw_lr": 8.13793103448276e-05, "mem_alloc_gb": 14.685915470123291, "mem_reserv_gb": 79.091796875, "loss_smoothed": 4.605582780639331, "seq_len": 4096} {"wall": 5437.161, "step": 250, "loss": 4.554456979036331, "lm_loss": 4.554456979036331, "mtp_loss": 0.0, "lr": 0.00519396551724138, "grad_norm": 1.2142772674560547, "tokens": 73728000, "adamw_lr": 8.310344827586208e-05, "mem_alloc_gb": 14.685915470123291, "mem_reserv_gb": 79.091796875, "loss_smoothed": 4.5874644378821055, "seq_len": 4096} {"wall": 5641.952, "step": 260, "loss": 4.511315256357193, "lm_loss": 4.511315256357193, "mtp_loss": 0.0, "lr": 0.005301724137931035, "grad_norm": 0.931605875492096, "tokens": 76677120, "adamw_lr": 8.482758620689656e-05, "mem_alloc_gb": 14.685915470123291, "mem_reserv_gb": 79.091796875, "loss_smoothed": 4.549328143397966, "seq_len": 4096} {"wall": 5846.754, "step": 270, "loss": 4.403887708981832, "lm_loss": 4.403887708981832, "mtp_loss": 0.0, "lr": 0.005409482758620691, "grad_norm": 0.8883979916572571, "tokens": 79626240, "adamw_lr": 8.655172413793105e-05, "mem_alloc_gb": 14.685915470123291, "mem_reserv_gb": 79.091796875, "loss_smoothed": 4.415097549557686, "seq_len": 4096} {"wall": 6051.271, "step": 280, "loss": 4.378030329942703, "lm_loss": 4.378030329942703, "mtp_loss": 0.0, "lr": 0.005517241379310345, "grad_norm": 1.074326515197754, "tokens": 82575360, "adamw_lr": 8.827586206896552e-05, "mem_alloc_gb": 14.685915470123291, "mem_reserv_gb": 79.091796875, "loss_smoothed": 4.378048166632652, "seq_len": 4096} {"wall": 6255.901, "step": 290, "loss": 4.304773251215617, "lm_loss": 4.304773251215617, "mtp_loss": 0.0, "lr": 0.005625000000000001, "grad_norm": 0.8975216746330261, "tokens": 85524480, "adamw_lr": 9e-05, "mem_alloc_gb": 14.685915470123291, "mem_reserv_gb": 79.091796875, "loss_smoothed": 4.208633749683698, "seq_len": 4096} {"wall": 6462.918, "eval_loss": 3.842194080352783, "eval_ppl": 46.62766647338867, "step": 300} {"wall": 6462.92, "step": 300, "loss": 4.22456615169843, "lm_loss": 4.22456615169843, "mtp_loss": 0.0, "lr": 0.005732758620689656, "grad_norm": 0.7665477395057678, "tokens": 88473600, "adamw_lr": 9.172413793103449e-05, "mem_alloc_gb": 14.685915470123291, "mem_reserv_gb": 79.091796875, "loss_smoothed": 4.235406318306923, "seq_len": 4096} {"wall": 6744.051, "step": 310, "loss": 4.228705475727717, "lm_loss": 4.228705475727717, "mtp_loss": 0.0, "lr": 0.0058405172413793114, "grad_norm": 0.8244591951370239, "tokens": 91422720, "adamw_lr": 9.344827586206898e-05, "mem_alloc_gb": 14.685915470123291, "mem_reserv_gb": 79.091796875, "loss_smoothed": 4.2227811465660725, "seq_len": 4096} {"wall": 6948.465, "step": 320, "loss": 4.114003260930379, "lm_loss": 4.114003260930379, "mtp_loss": 0.0, "lr": 0.005948275862068966, "grad_norm": 0.7613635659217834, "tokens": 94371840, "adamw_lr": 9.517241379310345e-05, "mem_alloc_gb": 14.685915470123291, "mem_reserv_gb": 79.091796875, "loss_smoothed": 4.149946423371633, "seq_len": 4096} {"wall": 7153.457, "step": 330, "loss": 4.0217796464761095, "lm_loss": 4.0217796464761095, "mtp_loss": 0.0, "lr": 0.006056034482758621, "grad_norm": 0.7248853445053101, "tokens": 97320960, "adamw_lr": 9.689655172413794e-05, "mem_alloc_gb": 14.685915470123291, "mem_reserv_gb": 79.091796875, "loss_smoothed": 4.073039283355078, "seq_len": 4096} {"wall": 7358.308, "step": 340, "loss": 4.103031307458878, "lm_loss": 4.103031307458878, "mtp_loss": 0.0, "lr": 0.006163793103448276, "grad_norm": 0.6125715970993042, "tokens": 100270080, "adamw_lr": 9.862068965517242e-05, "mem_alloc_gb": 14.685915470123291, "mem_reserv_gb": 79.091796875, "loss_smoothed": 4.057874144117037, "seq_len": 4096} {"wall": 7563.063, "step": 350, "loss": 4.049249897400538, "lm_loss": 4.049249897400538, "mtp_loss": 0.0, "lr": 0.00627155172413793, "grad_norm": 0.627769410610199, "tokens": 103219200, "adamw_lr": 0.0001003448275862069, "mem_alloc_gb": 14.685915470123291, "mem_reserv_gb": 79.091796875, "loss_smoothed": 3.9558626304070152, "seq_len": 4096} {"wall": 7767.709, "step": 360, "loss": 3.890793025493622, "lm_loss": 3.890793025493622, "mtp_loss": 0.0, "lr": 0.006379310344827587, "grad_norm": 0.6004687547683716, "tokens": 106168320, "adamw_lr": 0.0001020689655172414, "mem_alloc_gb": 14.685915470123291, "mem_reserv_gb": 79.091796875, "loss_smoothed": 3.9234980950752893, "seq_len": 4096} {"wall": 7972.801, "step": 370, "loss": 3.9312441448370614, "lm_loss": 3.9312441448370614, "mtp_loss": 0.0, "lr": 0.006487068965517243, "grad_norm": 0.6114732623100281, "tokens": 109117440, "adamw_lr": 0.00010379310344827588, "mem_alloc_gb": 14.685915470123291, "mem_reserv_gb": 79.091796875, "loss_smoothed": 3.8404794096946717, "seq_len": 4096} {"wall": 8177.568, "step": 380, "loss": 3.82335631052653, "lm_loss": 3.82335631052653, "mtp_loss": 0.0, "lr": 0.006594827586206897, "grad_norm": 0.5476613640785217, "tokens": 112066560, "adamw_lr": 0.00010551724137931037, "mem_alloc_gb": 14.685915470123291, "mem_reserv_gb": 79.091796875, "loss_smoothed": 3.751047362883886, "seq_len": 4096} {"wall": 8382.197, "step": 390, "loss": 3.6748522917429605, "lm_loss": 3.6748522917429605, "mtp_loss": 0.0, "lr": 0.006702586206896552, "grad_norm": 0.5107900500297546, "tokens": 115015680, "adamw_lr": 0.00010724137931034484, "mem_alloc_gb": 14.685915470123291, "mem_reserv_gb": 79.091796875, "loss_smoothed": 3.772439342737198, "seq_len": 4096} {"wall": 8589.383, "eval_loss": 3.356147885322571, "eval_ppl": 28.67850112915039, "step": 400} {"wall": 8589.385, "step": 400, "loss": 3.9156430065631866, "lm_loss": 3.9156430065631866, "mtp_loss": 0.0, "lr": 0.006810344827586207, "grad_norm": 0.5330420732498169, "tokens": 117964800, "adamw_lr": 0.00010896551724137931, "mem_alloc_gb": 14.685915470123291, "mem_reserv_gb": 79.091796875, "loss_smoothed": 3.758845458428065, "seq_len": 4096} {"wall": 136.519, "eval_loss": 3.356149345636368, "eval_ppl": 28.67854881286621, "step": 400} {"wall": 868.791, "step": 410, "loss": 3.5486598114172616, "lm_loss": 3.5486598114172616, "mtp_loss": 0.0, "lr": 0.006918103448275862, "grad_norm": 0.5229614973068237, "tokens": 120913920, "adamw_lr": 0.00011068965517241378, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.302734375, "loss_smoothed": 3.6149593581755957, "seq_len": 4096} {"wall": 1074.358, "step": 420, "loss": 3.567764788866043, "lm_loss": 3.567764788866043, "mtp_loss": 0.0, "lr": 0.007025862068965518, "grad_norm": 0.5050688982009888, "tokens": 123863040, "adamw_lr": 0.00011241379310344828, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.302734375, "loss_smoothed": 3.6559927205244698, "seq_len": 4096} {"wall": 1280.197, "step": 430, "loss": 3.7853246331214905, "lm_loss": 3.7853246331214905, "mtp_loss": 0.0, "lr": 0.007133620689655173, "grad_norm": 0.5664331912994385, "tokens": 126812160, "adamw_lr": 0.00011413793103448275, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.302734375, "loss_smoothed": 3.619067614277204, "seq_len": 4096} {"wall": 1485.806, "step": 440, "loss": 3.3555537263552346, "lm_loss": 3.3555537263552346, "mtp_loss": 0.0, "lr": 0.007241379310344829, "grad_norm": 0.46991798281669617, "tokens": 129761280, "adamw_lr": 0.00011586206896551725, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.302734375, "loss_smoothed": 3.5059588780005777, "seq_len": 4096} {"wall": 1691.193, "step": 450, "loss": 3.467200458049774, "lm_loss": 3.467200458049774, "mtp_loss": 0.0, "lr": 0.007349137931034483, "grad_norm": 0.4909840226173401, "tokens": 132710400, "adamw_lr": 0.00011758620689655173, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.302734375, "loss_smoothed": 3.485131781796614, "seq_len": 4096} {"wall": 1896.439, "step": 460, "loss": 3.424856702486674, "lm_loss": 3.424856702486674, "mtp_loss": 0.0, "lr": 0.007456896551724138, "grad_norm": 0.5101647973060608, "tokens": 135659520, "adamw_lr": 0.00011931034482758621, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.302734375, "loss_smoothed": 3.5235658168792723, "seq_len": 4096} {"wall": 2102.053, "step": 470, "loss": 3.296060790618261, "lm_loss": 3.296060790618261, "mtp_loss": 0.0, "lr": 0.007564655172413794, "grad_norm": 0.415724515914917, "tokens": 138608640, "adamw_lr": 0.00012103448275862071, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.302734375, "loss_smoothed": 3.3816489075620972, "seq_len": 4096} {"wall": 2307.043, "step": 480, "loss": 3.1637478371461234, "lm_loss": 3.1637478371461234, "mtp_loss": 0.0, "lr": 0.007672413793103449, "grad_norm": 0.3800559639930725, "tokens": 141557760, "adamw_lr": 0.00012275862068965518, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.302734375, "loss_smoothed": 3.3671953986088434, "seq_len": 4096} {"wall": 2512.202, "step": 490, "loss": 3.460976928472519, "lm_loss": 3.460976928472519, "mtp_loss": 0.0, "lr": 0.007780172413793105, "grad_norm": 0.49153071641921997, "tokens": 144506880, "adamw_lr": 0.00012448275862068968, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.302734375, "loss_smoothed": 3.361513742307822, "seq_len": 4096} {"wall": 2720.005, "eval_loss": 2.8888417780399323, "eval_ppl": 17.972482681274414, "step": 500} {"wall": 2720.007, "step": 500, "loss": 3.235230634609858, "lm_loss": 3.235230634609858, "mtp_loss": 0.0, "lr": 0.00788793103448276, "grad_norm": 0.42190101742744446, "tokens": 147456000, "adamw_lr": 0.00012620689655172415, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.302734375, "loss_smoothed": 3.278893948098024, "seq_len": 4096} {"wall": 2943.204, "step": 510, "loss": 3.1318486432234445, "lm_loss": 3.1318486432234445, "mtp_loss": 0.0, "lr": 0.007995689655172415, "grad_norm": 0.4522649347782135, "tokens": 150405120, "adamw_lr": 0.00012793103448275863, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 3.2176832939187685, "seq_len": 4096} {"wall": 3148.343, "step": 520, "loss": 3.122708668311437, "lm_loss": 3.122708668311437, "mtp_loss": 0.0, "lr": 0.008103448275862069, "grad_norm": 0.4709608256816864, "tokens": 153354240, "adamw_lr": 0.0001296551724137931, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 3.1653188318014145, "seq_len": 4096} {"wall": 3352.508, "step": 530, "loss": 3.16356893380483, "lm_loss": 3.16356893380483, "mtp_loss": 0.0, "lr": 0.008211206896551724, "grad_norm": 0.43616124987602234, "tokens": 156303360, "adamw_lr": 0.00013137931034482757, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 3.178527111311754, "seq_len": 4096} {"wall": 3557.286, "step": 540, "loss": 3.1014984995126724, "lm_loss": 3.1014984995126724, "mtp_loss": 0.0, "lr": 0.008318965517241381, "grad_norm": 0.41523876786231995, "tokens": 159252480, "adamw_lr": 0.0001331034482758621, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 3.10849542170763, "seq_len": 4096} {"wall": 3761.986, "step": 550, "loss": 3.0115936597188315, "lm_loss": 3.0115936597188315, "mtp_loss": 0.0, "lr": 0.008426724137931036, "grad_norm": 0.43522706627845764, "tokens": 162201600, "adamw_lr": 0.00013482758620689657, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 3.062070642908414, "seq_len": 4096} {"wall": 3966.706, "step": 560, "loss": 3.1142669717470803, "lm_loss": 3.1142669717470803, "mtp_loss": 0.0, "lr": 0.00853448275862069, "grad_norm": 0.4074176251888275, "tokens": 165150720, "adamw_lr": 0.00013655172413793104, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 3.1016694217920304, "seq_len": 4096} {"wall": 4171.54, "step": 570, "loss": 2.9883016645908356, "lm_loss": 2.9883016645908356, "mtp_loss": 0.0, "lr": 0.008642241379310345, "grad_norm": 0.42330268025398254, "tokens": 168099840, "adamw_lr": 0.00013827586206896552, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 3.089569028218587, "seq_len": 4096} {"wall": 4376.073, "step": 580, "loss": 3.0862867732842765, "lm_loss": 3.0862867732842765, "mtp_loss": 0.0, "lr": 0.008749999999999999, "grad_norm": 0.47565507888793945, "tokens": 171048960, "adamw_lr": 0.00014, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.9748789851864177, "seq_len": 4096} {"wall": 4581.344, "step": 590, "loss": 2.9889077146848044, "lm_loss": 2.9889077146848044, "mtp_loss": 0.0, "lr": 0.008857758620689654, "grad_norm": 0.33694231510162354, "tokens": 173998080, "adamw_lr": 0.00014172413793103446, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 3.003328869243463, "seq_len": 4096} {"wall": 4788.656, "eval_loss": 2.6420803517103195, "eval_ppl": 14.042386054992676, "step": 600} {"wall": 4788.658, "step": 600, "loss": 2.9617048303286233, "lm_loss": 2.9617048303286233, "mtp_loss": 0.0, "lr": 0.008965517241379312, "grad_norm": 0.379936546087265, "tokens": 176947200, "adamw_lr": 0.000143448275862069, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 3.0106960877776148, "seq_len": 4096} {"wall": 5073.092, "step": 610, "loss": 2.788805936773618, "lm_loss": 2.788805936773618, "mtp_loss": 0.0, "lr": 0.009073275862068967, "grad_norm": 0.39765995740890503, "tokens": 179896320, "adamw_lr": 0.00014517241379310346, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.498046875, "loss_smoothed": 3.0054008866349857, "seq_len": 4096} {"wall": 5277.912, "step": 620, "loss": 3.1092912952105203, "lm_loss": 3.1092912952105203, "mtp_loss": 0.0, "lr": 0.00918103448275862, "grad_norm": 0.3888259530067444, "tokens": 182845440, "adamw_lr": 0.00014689655172413793, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.498046875, "loss_smoothed": 2.9856473858157795, "seq_len": 4096} {"wall": 5482.585, "step": 630, "loss": 3.0365436176458993, "lm_loss": 3.0365436176458993, "mtp_loss": 0.0, "lr": 0.009288793103448276, "grad_norm": 0.3814540505409241, "tokens": 185794560, "adamw_lr": 0.00014862068965517243, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.498046875, "loss_smoothed": 2.975585889319579, "seq_len": 4096} {"wall": 5687.237, "step": 640, "loss": 2.861163849631945, "lm_loss": 2.861163849631945, "mtp_loss": 0.0, "lr": 0.009396551724137931, "grad_norm": 0.33615854382514954, "tokens": 188743680, "adamw_lr": 0.0001503448275862069, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.498046875, "loss_smoothed": 2.9465123549103738, "seq_len": 4096} {"wall": 5891.815, "step": 650, "loss": 2.881339266896248, "lm_loss": 2.881339266896248, "mtp_loss": 0.0, "lr": 0.009504310344827589, "grad_norm": 0.389344185590744, "tokens": 191692800, "adamw_lr": 0.0001520689655172414, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.498046875, "loss_smoothed": 2.91913234492143, "seq_len": 4096} {"wall": 6096.181, "step": 660, "loss": 2.903037662307421, "lm_loss": 2.903037662307421, "mtp_loss": 0.0, "lr": 0.009612068965517244, "grad_norm": 0.34364646673202515, "tokens": 194641920, "adamw_lr": 0.00015379310344827587, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.498046875, "loss_smoothed": 2.8855203971266747, "seq_len": 4096} {"wall": 6300.507, "step": 670, "loss": 3.0850104987621307, "lm_loss": 3.0850104987621307, "mtp_loss": 0.0, "lr": 0.009719827586206898, "grad_norm": 0.38163837790489197, "tokens": 197591040, "adamw_lr": 0.00015551724137931037, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.498046875, "loss_smoothed": 2.919249752163887, "seq_len": 4096} {"wall": 6504.892, "step": 680, "loss": 2.8839345475037894, "lm_loss": 2.8839345475037894, "mtp_loss": 0.0, "lr": 0.009827586206896553, "grad_norm": 0.35990768671035767, "tokens": 200540160, "adamw_lr": 0.00015724137931034485, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.498046875, "loss_smoothed": 2.915369973083337, "seq_len": 4096} {"wall": 6709.386, "step": 690, "loss": 2.7870888113975525, "lm_loss": 2.7870888113975525, "mtp_loss": 0.0, "lr": 0.009935344827586207, "grad_norm": 0.37112730741500854, "tokens": 203489280, "adamw_lr": 0.00015896551724137932, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.498046875, "loss_smoothed": 2.839839977522691, "seq_len": 4096} {"wall": 6916.488, "eval_loss": 2.5078581273555756, "eval_ppl": 12.27860164642334, "step": 700} {"wall": 6916.49, "step": 700, "loss": 2.849542031685511, "lm_loss": 2.849542031685511, "mtp_loss": 0.0, "lr": 0.010043103448275862, "grad_norm": 0.35465747117996216, "tokens": 206438400, "adamw_lr": 0.0001606896551724138, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.498046875, "loss_smoothed": 2.871141530573368, "seq_len": 4096} {"wall": 7197.635, "step": 710, "loss": 2.7673010925451913, "lm_loss": 2.7673010925451913, "mtp_loss": 0.0, "lr": 0.010150862068965517, "grad_norm": 0.3415488004684448, "tokens": 209387520, "adamw_lr": 0.00016241379310344826, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.8325316990415255, "seq_len": 4096} {"wall": 7402.797, "step": 720, "loss": 2.7131109535694122, "lm_loss": 2.7131109535694122, "mtp_loss": 0.0, "lr": 0.010258620689655175, "grad_norm": 0.31251952052116394, "tokens": 212336640, "adamw_lr": 0.0001641379310344828, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.8363250931104025, "seq_len": 4096} {"wall": 7607.825, "step": 730, "loss": 2.8091430912415185, "lm_loss": 2.8091430912415185, "mtp_loss": 0.0, "lr": 0.010366379310344828, "grad_norm": 0.33058351278305054, "tokens": 215285760, "adamw_lr": 0.00016586206896551726, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.8281921764214832, "seq_len": 4096} {"wall": 7812.995, "step": 740, "loss": 2.8163169076045356, "lm_loss": 2.8163169076045356, "mtp_loss": 0.0, "lr": 0.010474137931034484, "grad_norm": 0.2965877652168274, "tokens": 218234880, "adamw_lr": 0.00016758620689655173, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.880216523508231, "seq_len": 4096} {"wall": 8018.263, "step": 750, "loss": 2.816475788752238, "lm_loss": 2.816475788752238, "mtp_loss": 0.0, "lr": 0.010581896551724139, "grad_norm": 0.3142307996749878, "tokens": 221184000, "adamw_lr": 0.0001693103448275862, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.8358156606554985, "seq_len": 4096} {"wall": 8223.439, "step": 760, "loss": 2.8265002369880676, "lm_loss": 2.8265002369880676, "mtp_loss": 0.0, "lr": 0.010689655172413796, "grad_norm": 0.3451839089393616, "tokens": 224133120, "adamw_lr": 0.00017103448275862073, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.800556102891763, "seq_len": 4096} {"wall": 8428.693, "step": 770, "loss": 2.7186859051386514, "lm_loss": 2.7186859051386514, "mtp_loss": 0.0, "lr": 0.01079741379310345, "grad_norm": 0.3134319484233856, "tokens": 227082240, "adamw_lr": 0.0001727586206896552, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.724161379535993, "seq_len": 4096} {"wall": 8634.039, "step": 780, "loss": 2.6894926180442176, "lm_loss": 2.6894926180442176, "mtp_loss": 0.0, "lr": 0.010905172413793105, "grad_norm": 0.3232871890068054, "tokens": 230031360, "adamw_lr": 0.00017448275862068968, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.8012067899107933, "seq_len": 4096} {"wall": 8839.278, "step": 790, "loss": 2.7878577609856925, "lm_loss": 2.7878577609856925, "mtp_loss": 0.0, "lr": 0.01101293103448276, "grad_norm": 0.30463919043540955, "tokens": 232980480, "adamw_lr": 0.00017620689655172415, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.8301200682918233, "seq_len": 4096} {"wall": 9046.021, "eval_loss": 2.4274289309978485, "eval_ppl": 11.329715728759766, "step": 800} {"wall": 9046.023, "step": 800, "loss": 2.8215343356132507, "lm_loss": 2.8215343356132507, "mtp_loss": 0.0, "lr": 0.011120689655172414, "grad_norm": 0.29079872369766235, "tokens": 235929600, "adamw_lr": 0.00017793103448275862, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.7649593005577726, "seq_len": 4096} {"wall": 9324.837, "step": 810, "loss": 2.7860269049803414, "lm_loss": 2.7860269049803414, "mtp_loss": 0.0, "lr": 0.01122844827586207, "grad_norm": 0.3061423897743225, "tokens": 238878720, "adamw_lr": 0.0001796551724137931, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.722043778002262, "seq_len": 4096} {"wall": 9529.188, "step": 820, "loss": 2.686206897099813, "lm_loss": 2.686206897099813, "mtp_loss": 0.0, "lr": 0.011336206896551723, "grad_norm": 0.31777650117874146, "tokens": 241827840, "adamw_lr": 0.00018137931034482757, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.751613004008929, "seq_len": 4096} {"wall": 9733.814, "step": 830, "loss": 2.713415731986364, "lm_loss": 2.713415731986364, "mtp_loss": 0.0, "lr": 0.011443965517241382, "grad_norm": 0.3034782111644745, "tokens": 244776960, "adamw_lr": 0.0001831034482758621, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.74873983412981, "seq_len": 4096} {"wall": 9938.294, "step": 840, "loss": 2.738632877667745, "lm_loss": 2.738632877667745, "mtp_loss": 0.0, "lr": 0.011551724137931036, "grad_norm": 0.32327279448509216, "tokens": 247726080, "adamw_lr": 0.00018482758620689657, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.76094165345033, "seq_len": 4096} {"wall": 10142.873, "step": 850, "loss": 2.7500869731108346, "lm_loss": 2.7500869731108346, "mtp_loss": 0.0, "lr": 0.011659482758620691, "grad_norm": 0.2844919264316559, "tokens": 250675200, "adamw_lr": 0.00018655172413793104, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.740406103928884, "seq_len": 4096} {"wall": 10347.602, "step": 860, "loss": 2.8121034701665244, "lm_loss": 2.8121034701665244, "mtp_loss": 0.0, "lr": 0.011767241379310345, "grad_norm": 0.3241842985153198, "tokens": 253624320, "adamw_lr": 0.00018827586206896554, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.754697799682617, "seq_len": 4096} {"wall": 10552.32, "step": 870, "loss": 2.734913617372513, "lm_loss": 2.734913617372513, "mtp_loss": 0.0, "lr": 0.011875, "grad_norm": 0.30280420184135437, "tokens": 256573440, "adamw_lr": 0.00019, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.7626943563421564, "seq_len": 4096} {"wall": 10757.234, "step": 880, "loss": 2.682573070128759, "lm_loss": 2.682573070128759, "mtp_loss": 0.0, "lr": 0.011982758620689657, "grad_norm": 0.2953279912471771, "tokens": 259522560, "adamw_lr": 0.0001917241379310345, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.7208181127905844, "seq_len": 4096} {"wall": 10961.73, "step": 890, "loss": 2.7665851215521493, "lm_loss": 2.7665851215521493, "mtp_loss": 0.0, "lr": 0.012090517241379313, "grad_norm": 0.3175070285797119, "tokens": 262471680, "adamw_lr": 0.00019344827586206898, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.70145142575105, "seq_len": 4096} {"wall": 11169.053, "eval_loss": 2.3512655794620514, "eval_ppl": 10.498849868774414, "step": 900} {"wall": 11169.054, "step": 900, "loss": 2.732331628600756, "lm_loss": 2.732331628600756, "mtp_loss": 0.0, "lr": 0.012198275862068966, "grad_norm": 0.3632154166698456, "tokens": 265420800, "adamw_lr": 0.00019517241379310348, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.706065988043944, "seq_len": 4096} {"wall": 11451.862, "step": 910, "loss": 2.5951233903566995, "lm_loss": 2.5951233903566995, "mtp_loss": 0.0, "lr": 0.012306034482758622, "grad_norm": 0.3383670449256897, "tokens": 268369920, "adamw_lr": 0.00019689655172413795, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.6307262937227884, "seq_len": 4096} {"wall": 11656.409, "step": 920, "loss": 2.5566190679868064, "lm_loss": 2.5566190679868064, "mtp_loss": 0.0, "lr": 0.012413793103448277, "grad_norm": 0.2626509666442871, "tokens": 271319040, "adamw_lr": 0.00019862068965517243, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.608734048406283, "seq_len": 4096} {"wall": 11861.396, "step": 930, "loss": 2.504589964946111, "lm_loss": 2.504589964946111, "mtp_loss": 0.0, "lr": 0.01252155172413793, "grad_norm": 0.3104836940765381, "tokens": 274268160, "adamw_lr": 0.0002003448275862069, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.644425466656685, "seq_len": 4096} {"wall": 12066.298, "step": 940, "loss": 2.6940058022737503, "lm_loss": 2.6940058022737503, "mtp_loss": 0.0, "lr": 0.012629310344827588, "grad_norm": 0.30289584398269653, "tokens": 277217280, "adamw_lr": 0.00020206896551724143, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.695872774720192, "seq_len": 4096} {"wall": 12271.379, "step": 950, "loss": 2.7344687283039093, "lm_loss": 2.7344687283039093, "mtp_loss": 0.0, "lr": 0.012737068965517243, "grad_norm": 0.3127801716327667, "tokens": 280166400, "adamw_lr": 0.0002037931034482759, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.6639696230491006, "seq_len": 4096} {"wall": 12475.967, "step": 960, "loss": 2.6944516549507775, "lm_loss": 2.6944516549507775, "mtp_loss": 0.0, "lr": 0.012844827586206899, "grad_norm": 0.289466917514801, "tokens": 283115520, "adamw_lr": 0.00020551724137931037, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.6690850034356117, "seq_len": 4096} {"wall": 12680.735, "step": 970, "loss": 2.666459952791532, "lm_loss": 2.666459952791532, "mtp_loss": 0.0, "lr": 0.012952586206896552, "grad_norm": 0.2581908404827118, "tokens": 286064640, "adamw_lr": 0.00020724137931034484, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.624989042679469, "seq_len": 4096} {"wall": 12885.252, "step": 980, "loss": 2.6627366741498313, "lm_loss": 2.6627366741498313, "mtp_loss": 0.0, "lr": 0.013060344827586208, "grad_norm": 0.3300381600856781, "tokens": 289013760, "adamw_lr": 0.00020896551724137931, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.639344044526418, "seq_len": 4096} {"wall": 13089.822, "step": 990, "loss": 2.4889247020085654, "lm_loss": 2.4889247020085654, "mtp_loss": 0.0, "lr": 0.013168103448275861, "grad_norm": 0.2994905114173889, "tokens": 291962880, "adamw_lr": 0.0002106896551724138, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.655226083099842, "seq_len": 4096} {"wall": 13296.905, "eval_loss": 2.305895909667015, "eval_ppl": 10.033162117004395, "step": 1000} {"wall": 13296.906, "step": 1000, "loss": 2.569704612096151, "lm_loss": 2.569704612096151, "mtp_loss": 0.0, "lr": 0.013275862068965517, "grad_norm": 0.33411943912506104, "tokens": 294912000, "adamw_lr": 0.00021241379310344826, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.679789461195469, "seq_len": 4096} {"wall": 13575.663, "step": 1010, "loss": 2.7994199047485986, "lm_loss": 2.7994199047485986, "mtp_loss": 0.0, "lr": 0.013383620689655174, "grad_norm": 0.3851888179779053, "tokens": 297861120, "adamw_lr": 0.00021413793103448279, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.6395157029231386, "seq_len": 4096} {"wall": 13780.018, "step": 1020, "loss": 2.6564633597930274, "lm_loss": 2.6564633597930274, "mtp_loss": 0.0, "lr": 0.01349137931034483, "grad_norm": 0.28680774569511414, "tokens": 300810240, "adamw_lr": 0.00021586206896551726, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.6213230828444165, "seq_len": 4096} {"wall": 13984.573, "step": 1030, "loss": 2.6717295994361243, "lm_loss": 2.6717295994361243, "mtp_loss": 0.0, "lr": 0.013599137931034483, "grad_norm": 0.28287261724472046, "tokens": 303759360, "adamw_lr": 0.00021758620689655173, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.5854913567503295, "seq_len": 4096} {"wall": 14189.396, "step": 1040, "loss": 2.7166887621084848, "lm_loss": 2.7166887621084848, "mtp_loss": 0.0, "lr": 0.013706896551724138, "grad_norm": 0.33239999413490295, "tokens": 306708480, "adamw_lr": 0.0002193103448275862, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.6636414354046183, "seq_len": 4096} {"wall": 14393.805, "step": 1050, "loss": 2.64456777771314, "lm_loss": 2.64456777771314, "mtp_loss": 0.0, "lr": 0.013814655172413795, "grad_norm": 0.2744719088077545, "tokens": 309657600, "adamw_lr": 0.00022103448275862073, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.6546419352293014, "seq_len": 4096} {"wall": 14598.492, "step": 1060, "loss": 2.7042347689469657, "lm_loss": 2.7042347689469657, "mtp_loss": 0.0, "lr": 0.013922413793103447, "grad_norm": 0.28318214416503906, "tokens": 312606720, "adamw_lr": 0.00022275862068965517, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.6145413244764013, "seq_len": 4096} {"wall": 14803.13, "step": 1070, "loss": 2.5346389561891556, "lm_loss": 2.5346389561891556, "mtp_loss": 0.0, "lr": 0.014030172413793104, "grad_norm": 0.294711172580719, "tokens": 315555840, "adamw_lr": 0.00022448275862068967, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.5959721595048904, "seq_len": 4096} {"wall": 15007.621, "step": 1080, "loss": 2.518240839242935, "lm_loss": 2.518240839242935, "mtp_loss": 0.0, "lr": 0.01413793103448276, "grad_norm": 0.33644044399261475, "tokens": 318504960, "adamw_lr": 0.00022620689655172415, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.624595999221007, "seq_len": 4096} {"wall": 15212.098, "step": 1090, "loss": 2.6243305852015815, "lm_loss": 2.6243305852015815, "mtp_loss": 0.0, "lr": 0.014245689655172415, "grad_norm": 0.2975899875164032, "tokens": 321454080, "adamw_lr": 0.00022793103448275862, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.595784265299638, "seq_len": 4096} {"wall": 15419.171, "eval_loss": 2.2586387246847153, "eval_ppl": 9.570052146911621, "step": 1100} {"wall": 15419.173, "step": 1100, "loss": 2.6468052119016647, "lm_loss": 2.6468052119016647, "mtp_loss": 0.0, "lr": 0.014353448275862069, "grad_norm": 0.2975345551967621, "tokens": 324403200, "adamw_lr": 0.00022965517241379312, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.596099112927914, "seq_len": 4096} {"wall": 15699.348, "step": 1110, "loss": 2.6576004972060523, "lm_loss": 2.6576004972060523, "mtp_loss": 0.0, "lr": 0.014461206896551726, "grad_norm": 0.34916356205940247, "tokens": 327352320, "adamw_lr": 0.00023137931034482762, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.5572761043906214, "seq_len": 4096} {"wall": 15904.23, "step": 1120, "loss": 2.491793086131414, "lm_loss": 2.491793086131414, "mtp_loss": 0.0, "lr": 0.014568965517241378, "grad_norm": 0.2819589078426361, "tokens": 330301440, "adamw_lr": 0.00023310344827586206, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.5953410794337595, "seq_len": 4096} {"wall": 16108.997, "step": 1130, "loss": 2.5870670775572457, "lm_loss": 2.5870670775572457, "mtp_loss": 0.0, "lr": 0.014676724137931037, "grad_norm": 0.3304673433303833, "tokens": 333250560, "adamw_lr": 0.0002348275862068966, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.566261485715707, "seq_len": 4096} {"wall": 16313.251, "step": 1140, "loss": 2.579717164238294, "lm_loss": 2.579717164238294, "mtp_loss": 0.0, "lr": 0.01478448275862069, "grad_norm": 0.27444571256637573, "tokens": 336199680, "adamw_lr": 0.00023655172413793106, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.562514987587929, "seq_len": 4096} {"wall": 16517.859, "step": 1150, "loss": 2.485587919751803, "lm_loss": 2.485587919751803, "mtp_loss": 0.0, "lr": 0.014892241379310346, "grad_norm": 0.2674020230770111, "tokens": 339148800, "adamw_lr": 0.00023827586206896553, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.542663616935412, "seq_len": 4096} {"wall": 16722.381, "step": 1160, "loss": 2.620291213194529, "lm_loss": 2.620291213194529, "mtp_loss": 0.0, "lr": 0.015, "grad_norm": 0.31663674116134644, "tokens": 342097920, "adamw_lr": 0.00024, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.6010019014279044, "seq_len": 4096} {"wall": 16926.836, "step": 1170, "loss": 2.6702882051467896, "lm_loss": 2.6702882051467896, "mtp_loss": 0.0, "lr": 0.015107758620689655, "grad_norm": 0.31884798407554626, "tokens": 345047040, "adamw_lr": 0.00024172413793103448, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.598666233321031, "seq_len": 4096} {"wall": 17131.512, "step": 1180, "loss": 2.569907238086065, "lm_loss": 2.569907238086065, "mtp_loss": 0.0, "lr": 0.01521551724137931, "grad_norm": 0.31876519322395325, "tokens": 347996160, "adamw_lr": 0.00024344827586206895, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.52620187997818, "seq_len": 4096} {"wall": 17336.218, "step": 1190, "loss": 2.6323324888944626, "lm_loss": 2.6323324888944626, "mtp_loss": 0.0, "lr": 0.015323275862068967, "grad_norm": 0.3236926794052124, "tokens": 350945280, "adamw_lr": 0.00024517241379310345, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.556960013508797, "seq_len": 4096} {"wall": 17543.372, "eval_loss": 2.227431818842888, "eval_ppl": 9.276012420654297, "step": 1200} {"wall": 17543.374, "step": 1200, "loss": 2.5096692194541297, "lm_loss": 2.5096692194541297, "mtp_loss": 0.0, "lr": 0.015431034482758621, "grad_norm": 0.3337138593196869, "tokens": 353894400, "adamw_lr": 0.0002468965517241379, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.5599994915227096, "seq_len": 4096} {"wall": 17825.909, "step": 1210, "loss": 2.468213746945063, "lm_loss": 2.468213746945063, "mtp_loss": 0.0, "lr": 0.015538793103448276, "grad_norm": 0.34320297837257385, "tokens": 356843520, "adamw_lr": 0.00024862068965517245, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.5527267336845396, "seq_len": 4096} {"wall": 18030.699, "step": 1220, "loss": 2.657329112291336, "lm_loss": 2.657329112291336, "mtp_loss": 0.0, "lr": 0.015646551724137935, "grad_norm": 0.2916238605976105, "tokens": 359792640, "adamw_lr": 0.0002503448275862069, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.5742235511541365, "seq_len": 4096} {"wall": 18235.398, "step": 1230, "loss": 2.556588053703308, "lm_loss": 2.556588053703308, "mtp_loss": 0.0, "lr": 0.015754310344827587, "grad_norm": 0.3516257107257843, "tokens": 362741760, "adamw_lr": 0.0002520689655172414, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.548284167051315, "seq_len": 4096} {"wall": 18440.183, "step": 1240, "loss": 2.5085714062054953, "lm_loss": 2.5085714062054953, "mtp_loss": 0.0, "lr": 0.015862068965517243, "grad_norm": 0.29000964760780334, "tokens": 365690880, "adamw_lr": 0.00025379310344827587, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.538485497236252, "seq_len": 4096} {"wall": 18644.987, "step": 1250, "loss": 2.59797802567482, "lm_loss": 2.59797802567482, "mtp_loss": 0.0, "lr": 0.015969827586206898, "grad_norm": 0.3371170461177826, "tokens": 368640000, "adamw_lr": 0.0002555172413793104, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.562560077011585, "seq_len": 4096} {"wall": 18849.725, "step": 1260, "loss": 2.548169622818629, "lm_loss": 2.548169622818629, "mtp_loss": 0.0, "lr": 0.016077586206896553, "grad_norm": 0.2513684332370758, "tokens": 371589120, "adamw_lr": 0.00025724137931034486, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.5271636406580607, "seq_len": 4096} {"wall": 19054.527, "step": 1270, "loss": 2.5857877880334854, "lm_loss": 2.5857877880334854, "mtp_loss": 0.0, "lr": 0.01618534482758621, "grad_norm": 0.27380484342575073, "tokens": 374538240, "adamw_lr": 0.00025896551724137934, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.5146716373662157, "seq_len": 4096} {"wall": 19259.474, "step": 1280, "loss": 2.5204806278149285, "lm_loss": 2.5204806278149285, "mtp_loss": 0.0, "lr": 0.01629310344827586, "grad_norm": 0.28453198075294495, "tokens": 377487360, "adamw_lr": 0.0002606896551724138, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.5180041725436846, "seq_len": 4096} {"wall": 19464.137, "step": 1290, "loss": 2.6409872074921927, "lm_loss": 2.6409872074921927, "mtp_loss": 0.0, "lr": 0.016400862068965516, "grad_norm": 0.31916344165802, "tokens": 380436480, "adamw_lr": 0.0002624137931034483, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.5248389199376104, "seq_len": 4096} {"wall": 19671.192, "eval_loss": 2.200078308582306, "eval_ppl": 9.02571964263916, "step": 1300} {"wall": 19671.194, "step": 1300, "loss": 2.4465708633263907, "lm_loss": 2.4465708633263907, "mtp_loss": 0.0, "lr": 0.016508620689655175, "grad_norm": 0.3228042721748352, "tokens": 383385600, "adamw_lr": 0.0002641379310344828, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.5161517615119613, "seq_len": 4096} {"wall": 19955.379, "step": 1310, "loss": 2.512823904554049, "lm_loss": 2.512823904554049, "mtp_loss": 0.0, "lr": 0.016616379310344827, "grad_norm": 0.3049370348453522, "tokens": 386334720, "adamw_lr": 0.0002658620689655172, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.5407624433437985, "seq_len": 4096} {"wall": 20160.096, "step": 1320, "loss": 2.464399511615435, "lm_loss": 2.464399511615435, "mtp_loss": 0.0, "lr": 0.016724137931034482, "grad_norm": 0.281416118144989, "tokens": 389283840, "adamw_lr": 0.00026758620689655175, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.5283413151899974, "seq_len": 4096} {"wall": 20364.107, "step": 1330, "loss": 2.416213184595108, "lm_loss": 2.416213184595108, "mtp_loss": 0.0, "lr": 0.016831896551724138, "grad_norm": 0.25306636095046997, "tokens": 392232960, "adamw_lr": 0.0002693103448275862, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.4812298600872356, "seq_len": 4096} {"wall": 20568.446, "step": 1340, "loss": 2.3890107373396554, "lm_loss": 2.3890107373396554, "mtp_loss": 0.0, "lr": 0.016939655172413793, "grad_norm": 0.28795409202575684, "tokens": 395182080, "adamw_lr": 0.0002710344827586207, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.4943172191580136, "seq_len": 4096} {"wall": 20773.098, "step": 1350, "loss": 2.4561042189598083, "lm_loss": 2.4561042189598083, "mtp_loss": 0.0, "lr": 0.017047413793103452, "grad_norm": 0.33338987827301025, "tokens": 398131200, "adamw_lr": 0.0002727586206896552, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.5348877536753815, "seq_len": 4096} {"wall": 20977.483, "step": 1360, "loss": 2.5115426182746887, "lm_loss": 2.5115426182746887, "mtp_loss": 0.0, "lr": 0.017155172413793104, "grad_norm": 0.32184603810310364, "tokens": 401080320, "adamw_lr": 0.00027448275862068964, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.527845093111197, "seq_len": 4096} {"wall": 21182.218, "step": 1370, "loss": 2.5992491046587625, "lm_loss": 2.5992491046587625, "mtp_loss": 0.0, "lr": 0.01726293103448276, "grad_norm": 0.2853281795978546, "tokens": 404029440, "adamw_lr": 0.00027620689655172417, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.4930308495958644, "seq_len": 4096} {"wall": 21387.025, "step": 1380, "loss": 2.4695956309636435, "lm_loss": 2.4695956309636435, "mtp_loss": 0.0, "lr": 0.017370689655172415, "grad_norm": 0.2857533395290375, "tokens": 406978560, "adamw_lr": 0.00027793103448275864, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.4878001367052396, "seq_len": 4096} {"wall": 21591.152, "step": 1390, "loss": 2.47264655927817, "lm_loss": 2.47264655927817, "mtp_loss": 0.0, "lr": 0.01747844827586207, "grad_norm": 0.3927861750125885, "tokens": 409927680, "adamw_lr": 0.0002796551724137931, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.4452354292074836, "seq_len": 4096} {"wall": 21798.089, "eval_loss": 2.174726814031601, "eval_ppl": 8.799779891967773, "step": 1400} {"wall": 21798.09, "step": 1400, "loss": 2.4667510191599527, "lm_loss": 2.4667510191599527, "mtp_loss": 0.0, "lr": 0.017586206896551725, "grad_norm": 0.2765067219734192, "tokens": 412876800, "adamw_lr": 0.0002813793103448276, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.4564414133628207, "seq_len": 4096} {"wall": 22084.799, "step": 1410, "loss": 2.4670983155568442, "lm_loss": 2.4670983155568442, "mtp_loss": 0.0, "lr": 0.01769396551724138, "grad_norm": 0.3350383937358856, "tokens": 415825920, "adamw_lr": 0.0002831034482758621, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.512902276714643, "seq_len": 4096} {"wall": 22289.138, "step": 1420, "loss": 2.412215749422709, "lm_loss": 2.412215749422709, "mtp_loss": 0.0, "lr": 0.017801724137931033, "grad_norm": 0.2698183059692383, "tokens": 418775040, "adamw_lr": 0.00028482758620689653, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.4934741581479707, "seq_len": 4096} {"wall": 22493.749, "step": 1430, "loss": 2.5578766961892447, "lm_loss": 2.5578766961892447, "mtp_loss": 0.0, "lr": 0.01790948275862069, "grad_norm": 0.3312450647354126, "tokens": 421724160, "adamw_lr": 0.00028655172413793106, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.4876419951518374, "seq_len": 4096} {"wall": 22698.327, "step": 1440, "loss": 2.6316196024417877, "lm_loss": 2.6316196024417877, "mtp_loss": 0.0, "lr": 0.018017241379310347, "grad_norm": 0.2883259952068329, "tokens": 424673280, "adamw_lr": 0.00028827586206896553, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.5281749799847604, "seq_len": 4096} {"wall": 22902.891, "step": 1450, "loss": 2.606349582473437, "lm_loss": 2.606349582473437, "mtp_loss": 0.0, "lr": 0.018125, "grad_norm": 0.3152984380722046, "tokens": 427622400, "adamw_lr": 0.00029, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.512746781607469, "seq_len": 4096} {"wall": 23107.465, "step": 1460, "loss": 2.421518271168073, "lm_loss": 2.421518271168073, "mtp_loss": 0.0, "lr": 0.018232758620689654, "grad_norm": 0.31367528438568115, "tokens": 430571520, "adamw_lr": 0.0002917241379310345, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.503869617978732, "seq_len": 4096} {"wall": 23312.408, "step": 1470, "loss": 2.6243596225976944, "lm_loss": 2.6243596225976944, "mtp_loss": 0.0, "lr": 0.01834051724137931, "grad_norm": 0.3219667375087738, "tokens": 433520640, "adamw_lr": 0.00029344827586206895, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.5017490943272906, "seq_len": 4096} {"wall": 23516.783, "step": 1480, "loss": 2.463107963403066, "lm_loss": 2.463107963403066, "mtp_loss": 0.0, "lr": 0.01844827586206897, "grad_norm": 0.30643293261528015, "tokens": 436469760, "adamw_lr": 0.00029517241379310347, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.505280234416326, "seq_len": 4096} {"wall": 23721.143, "step": 1490, "loss": 2.444238339861234, "lm_loss": 2.444238339861234, "mtp_loss": 0.0, "lr": 0.01855603448275862, "grad_norm": 0.29442548751831055, "tokens": 439418880, "adamw_lr": 0.00029689655172413795, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.4906118457516033, "seq_len": 4096} {"wall": 23927.861, "eval_loss": 2.1461108326911926, "eval_ppl": 8.551534652709961, "step": 1500} {"wall": 23927.862, "step": 1500, "loss": 2.603661889831225, "lm_loss": 2.603661889831225, "mtp_loss": 0.0, "lr": 0.018663793103448276, "grad_norm": 0.32476934790611267, "tokens": 442368000, "adamw_lr": 0.0002986206896551724, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.505725318690141, "seq_len": 4096} {"wall": 24210.733, "step": 1510, "loss": 2.5625812262296677, "lm_loss": 2.5625812262296677, "mtp_loss": 0.0, "lr": 0.01877155172413793, "grad_norm": 0.2633122205734253, "tokens": 445317120, "adamw_lr": 0.0003003448275862069, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.4752353241046268, "seq_len": 4096} {"wall": 24415.39, "step": 1520, "loss": 2.5294721722602844, "lm_loss": 2.5294721722602844, "mtp_loss": 0.0, "lr": 0.01887931034482759, "grad_norm": 0.2783941328525543, "tokens": 448266240, "adamw_lr": 0.0003020689655172414, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.472114773094654, "seq_len": 4096} {"wall": 24619.96, "step": 1530, "loss": 2.481573596596718, "lm_loss": 2.481573596596718, "mtp_loss": 0.0, "lr": 0.018987068965517242, "grad_norm": 0.3470197916030884, "tokens": 451215360, "adamw_lr": 0.00030379310344827583, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.494137130926053, "seq_len": 4096} {"wall": 24824.449, "step": 1540, "loss": 2.5164048870404563, "lm_loss": 2.5164048870404563, "mtp_loss": 0.0, "lr": 0.019094827586206897, "grad_norm": 0.29719462990760803, "tokens": 454164480, "adamw_lr": 0.00030551724137931036, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.479942007859548, "seq_len": 4096} {"wall": 25029.261, "step": 1550, "loss": 2.439524089296659, "lm_loss": 2.439524089296659, "mtp_loss": 0.0, "lr": 0.019202586206896553, "grad_norm": 0.3048279881477356, "tokens": 457113600, "adamw_lr": 0.00030724137931034483, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.457093199590842, "seq_len": 4096} {"wall": 25233.959, "step": 1560, "loss": 2.45836678147316, "lm_loss": 2.45836678147316, "mtp_loss": 0.0, "lr": 0.019310344827586208, "grad_norm": 0.30734628438949585, "tokens": 460062720, "adamw_lr": 0.0003089655172413793, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.464548046390216, "seq_len": 4096} {"wall": 25438.295, "step": 1570, "loss": 2.443247675895691, "lm_loss": 2.443247675895691, "mtp_loss": 0.0, "lr": 0.019418103448275863, "grad_norm": 0.4995042681694031, "tokens": 463011840, "adamw_lr": 0.0003106896551724138, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.504032717148463, "seq_len": 4096} {"wall": 25642.661, "step": 1580, "loss": 2.584405834476153, "lm_loss": 2.584405834476153, "mtp_loss": 0.0, "lr": 0.01952586206896552, "grad_norm": 0.2984749674797058, "tokens": 465960960, "adamw_lr": 0.00031241379310344825, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.474234933654467, "seq_len": 4096} {"wall": 25847.202, "step": 1590, "loss": 2.4689024488131204, "lm_loss": 2.4689024488131204, "mtp_loss": 0.0, "lr": 0.01963362068965517, "grad_norm": 0.40233686566352844, "tokens": 468910080, "adamw_lr": 0.0003141379310344827, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.4690493543942766, "seq_len": 4096} {"wall": 26054.309, "eval_loss": 2.132255792617798, "eval_ppl": 8.433870315551758, "step": 1600} {"wall": 26054.311, "step": 1600, "loss": 2.533935005466143, "lm_loss": 2.533935005466143, "mtp_loss": 0.0, "lr": 0.01974137931034483, "grad_norm": 0.3007613718509674, "tokens": 471859200, "adamw_lr": 0.00031586206896551725, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.469442205131054, "seq_len": 4096} {"wall": 26336.375, "step": 1610, "loss": 2.4872783968846, "lm_loss": 2.4872783968846, "mtp_loss": 0.0, "lr": 0.019849137931034485, "grad_norm": 0.2523801624774933, "tokens": 474808320, "adamw_lr": 0.0003175862068965517, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.442336176335812, "seq_len": 4096} {"wall": 26541.498, "step": 1620, "loss": 2.5586009472608566, "lm_loss": 2.5586009472608566, "mtp_loss": 0.0, "lr": 0.01995689655172414, "grad_norm": 0.281892865896225, "tokens": 477757440, "adamw_lr": 0.0003193103448275862, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.476579446842273, "seq_len": 4096} {"wall": 26746.629, "step": 1630, "loss": 2.2772057031591735, "lm_loss": 2.2772057031591735, "mtp_loss": 0.0, "lr": 0.020064655172413796, "grad_norm": 0.35026177763938904, "tokens": 480706560, "adamw_lr": 0.0003210344827586207, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.444815018028021, "seq_len": 4096} {"wall": 26951.292, "step": 1640, "loss": 2.4300494442383447, "lm_loss": 2.4300494442383447, "mtp_loss": 0.0, "lr": 0.020172413793103448, "grad_norm": 0.3597499132156372, "tokens": 483655680, "adamw_lr": 0.0003227586206896552, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.4382774968942007, "seq_len": 4096} {"wall": 27155.917, "step": 1650, "loss": 2.4066772560278573, "lm_loss": 2.4066772560278573, "mtp_loss": 0.0, "lr": 0.020280172413793107, "grad_norm": 0.2845275402069092, "tokens": 486604800, "adamw_lr": 0.00032448275862068967, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.4630458876490593, "seq_len": 4096} {"wall": 27360.565, "step": 1660, "loss": 2.543087879816691, "lm_loss": 2.543087879816691, "mtp_loss": 0.0, "lr": 0.020387931034482762, "grad_norm": 0.2975473999977112, "tokens": 489553920, "adamw_lr": 0.00032620689655172414, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.430437622219324, "seq_len": 4096} {"wall": 27565.111, "step": 1670, "loss": 2.475115900238355, "lm_loss": 2.475115900238355, "mtp_loss": 0.0, "lr": 0.020495689655172414, "grad_norm": 0.29108354449272156, "tokens": 492503040, "adamw_lr": 0.0003279310344827586, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.4243010674913728, "seq_len": 4096} {"wall": 27769.869, "step": 1680, "loss": 2.44895305732886, "lm_loss": 2.44895305732886, "mtp_loss": 0.0, "lr": 0.02060344827586207, "grad_norm": 0.2755866050720215, "tokens": 495452160, "adamw_lr": 0.00032965517241379314, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.4317899117867148, "seq_len": 4096} {"wall": 27975.028, "step": 1690, "loss": 2.5950539211432138, "lm_loss": 2.5950539211432138, "mtp_loss": 0.0, "lr": 0.020711206896551725, "grad_norm": 0.2927231192588806, "tokens": 498401280, "adamw_lr": 0.0003313793103448276, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.4921760166684783, "seq_len": 4096} {"wall": 28181.671, "eval_loss": 2.1012617647647858, "eval_ppl": 8.176481246948242, "step": 1700} {"wall": 28181.673, "step": 1700, "loss": 2.3400975416103997, "lm_loss": 2.3400975416103997, "mtp_loss": 0.0, "lr": 0.02081896551724138, "grad_norm": 0.3008866310119629, "tokens": 501350400, "adamw_lr": 0.0003331034482758621, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.419304740180572, "seq_len": 4096} {"wall": 28466.447, "step": 1710, "loss": 2.3195971995592117, "lm_loss": 2.3195971995592117, "mtp_loss": 0.0, "lr": 0.020926724137931035, "grad_norm": 0.24794039130210876, "tokens": 504299520, "adamw_lr": 0.0003348275862068966, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.378471448024114, "seq_len": 4096} {"wall": 28670.986, "step": 1720, "loss": 2.402149553100268, "lm_loss": 2.402149553100268, "mtp_loss": 0.0, "lr": 0.021034482758620687, "grad_norm": 0.3566400110721588, "tokens": 507248640, "adamw_lr": 0.000336551724137931, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.422120800614357, "seq_len": 4096} {"wall": 28875.37, "step": 1730, "loss": 2.3449217726786933, "lm_loss": 2.3449217726786933, "mtp_loss": 0.0, "lr": 0.021142241379310346, "grad_norm": 0.2989150881767273, "tokens": 510197760, "adamw_lr": 0.00033827586206896555, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.429736902564764, "seq_len": 4096} {"wall": 29080.229, "step": 1740, "loss": 2.53921739757061, "lm_loss": 2.53921739757061, "mtp_loss": 0.0, "lr": 0.02125, "grad_norm": 0.3000688850879669, "tokens": 513146880, "adamw_lr": 0.00034, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.422532702734073, "seq_len": 4096} {"wall": 29284.951, "step": 1750, "loss": 2.5040428886810937, "lm_loss": 2.5040428886810937, "mtp_loss": 0.0, "lr": 0.021357758620689657, "grad_norm": 0.2876095473766327, "tokens": 516096000, "adamw_lr": 0.0003417241379310345, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.4062749835352104, "seq_len": 4096} {"wall": 29489.546, "step": 1760, "loss": 2.4210762977600098, "lm_loss": 2.4210762977600098, "mtp_loss": 0.0, "lr": 0.021465517241379312, "grad_norm": 0.3076370358467102, "tokens": 519045120, "adamw_lr": 0.000343448275862069, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.4711631228526434, "seq_len": 4096} {"wall": 29694.04, "step": 1770, "loss": 2.430409918228785, "lm_loss": 2.430409918228785, "mtp_loss": 0.0, "lr": 0.021573275862068968, "grad_norm": 0.28631284832954407, "tokens": 521994240, "adamw_lr": 0.0003451724137931035, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.4159024978677435, "seq_len": 4096} {"wall": 29898.549, "step": 1780, "loss": 2.4111603250106177, "lm_loss": 2.4111603250106177, "mtp_loss": 0.0, "lr": 0.021681034482758623, "grad_norm": 0.3473292589187622, "tokens": 524943360, "adamw_lr": 0.00034689655172413797, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.4271213283141453, "seq_len": 4096} {"wall": 30103.346, "step": 1790, "loss": 2.4423790127038956, "lm_loss": 2.4423790127038956, "mtp_loss": 0.0, "lr": 0.02178879310344828, "grad_norm": 0.30307552218437195, "tokens": 527892480, "adamw_lr": 0.00034862068965517244, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.3916766859591005, "seq_len": 4096} {"wall": 30310.67, "eval_loss": 2.09358286857605, "eval_ppl": 8.113934516906738, "step": 1800} {"wall": 30310.672, "step": 1800, "loss": 2.4677269707123437, "lm_loss": 2.4677269707123437, "mtp_loss": 0.0, "lr": 0.02189655172413793, "grad_norm": 0.3158144950866699, "tokens": 530841600, "adamw_lr": 0.0003503448275862069, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.398469946285089, "seq_len": 4096} {"wall": 30592.567, "step": 1810, "loss": 2.3220304697752, "lm_loss": 2.3220304697752, "mtp_loss": 0.0, "lr": 0.022004310344827586, "grad_norm": 0.3215239942073822, "tokens": 533790720, "adamw_lr": 0.0003520689655172414, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.443376849095027, "seq_len": 4096} {"wall": 30797.43, "step": 1820, "loss": 2.460940515001615, "lm_loss": 2.460940515001615, "mtp_loss": 0.0, "lr": 0.022112068965517245, "grad_norm": 0.3083120584487915, "tokens": 536739840, "adamw_lr": 0.0003537931034482759, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.442970803876718, "seq_len": 4096} {"wall": 31001.704, "step": 1830, "loss": 2.2702773064374924, "lm_loss": 2.2702773064374924, "mtp_loss": 0.0, "lr": 0.022219827586206897, "grad_norm": 0.29031550884246826, "tokens": 539688960, "adamw_lr": 0.00035551724137931033, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.4167617097496987, "seq_len": 4096} {"wall": 31205.865, "step": 1840, "loss": 2.61529349287351, "lm_loss": 2.61529349287351, "mtp_loss": 0.0, "lr": 0.022327586206896552, "grad_norm": 0.30613332986831665, "tokens": 542638080, "adamw_lr": 0.00035724137931034486, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.413722278177738, "seq_len": 4096} {"wall": 31410.123, "step": 1850, "loss": 2.418633371591568, "lm_loss": 2.418633371591568, "mtp_loss": 0.0, "lr": 0.022435344827586207, "grad_norm": 0.30521056056022644, "tokens": 545587200, "adamw_lr": 0.00035896551724137933, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.4863791649540263, "seq_len": 4096} {"wall": 31614.388, "step": 1860, "loss": 2.4171287765105567, "lm_loss": 2.4171287765105567, "mtp_loss": 0.0, "lr": 0.022543103448275863, "grad_norm": 0.288765549659729, "tokens": 548536320, "adamw_lr": 0.0003606896551724138, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.4360587686300277, "seq_len": 4096} {"wall": 31818.696, "step": 1870, "loss": 2.3654319643974304, "lm_loss": 2.3654319643974304, "mtp_loss": 0.0, "lr": 0.022650862068965518, "grad_norm": 0.284858763217926, "tokens": 551485440, "adamw_lr": 0.0003624137931034483, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.3979890699187916, "seq_len": 4096} {"wall": 32023.165, "step": 1880, "loss": 2.3196586072444916, "lm_loss": 2.3196586072444916, "mtp_loss": 0.0, "lr": 0.022758620689655173, "grad_norm": 0.2654084861278534, "tokens": 554434560, "adamw_lr": 0.0003641379310344828, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.3999721097449465, "seq_len": 4096} {"wall": 32227.702, "step": 1890, "loss": 2.301830679178238, "lm_loss": 2.301830679178238, "mtp_loss": 0.0, "lr": 0.02286637931034483, "grad_norm": 0.3266041874885559, "tokens": 557383680, "adamw_lr": 0.00036586206896551727, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.423158970226844, "seq_len": 4096} {"wall": 32434.625, "eval_loss": 2.082899361848831, "eval_ppl": 8.0277099609375, "step": 1900} {"wall": 32434.627, "step": 1900, "loss": 2.349966511130333, "lm_loss": 2.349966511130333, "mtp_loss": 0.0, "lr": 0.022974137931034484, "grad_norm": 0.2889018654823303, "tokens": 560332800, "adamw_lr": 0.00036758620689655174, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.372437023123106, "seq_len": 4096} {"wall": 32724.596, "step": 1910, "loss": 2.4146891931692758, "lm_loss": 2.4146891931692758, "mtp_loss": 0.0, "lr": 0.02308189655172414, "grad_norm": 0.2923961281776428, "tokens": 563281920, "adamw_lr": 0.0003693103448275862, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.437975592414538, "seq_len": 4096} {"wall": 32928.543, "step": 1920, "loss": 2.459726005792618, "lm_loss": 2.459726005792618, "mtp_loss": 0.0, "lr": 0.023189655172413795, "grad_norm": 0.344122052192688, "tokens": 566231040, "adamw_lr": 0.0003710344827586207, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.4443977842728297, "seq_len": 4096} {"wall": 33132.99, "step": 1930, "loss": 2.361880893508593, "lm_loss": 2.361880893508593, "mtp_loss": 0.0, "lr": 0.02329741379310345, "grad_norm": 0.25869354605674744, "tokens": 569180160, "adamw_lr": 0.0003727586206896552, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.435405395925045, "seq_len": 4096} {"wall": 33337.904, "step": 1940, "loss": 2.424991230169932, "lm_loss": 2.424991230169932, "mtp_loss": 0.0, "lr": 0.023405172413793102, "grad_norm": 0.34141379594802856, "tokens": 572129280, "adamw_lr": 0.00037448275862068963, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.4409229442477227, "seq_len": 4096} {"wall": 33542.273, "step": 1950, "loss": 2.526063601175944, "lm_loss": 2.526063601175944, "mtp_loss": 0.0, "lr": 0.02351293103448276, "grad_norm": 0.3155788779258728, "tokens": 575078400, "adamw_lr": 0.00037620689655172416, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.43130637605985, "seq_len": 4096} {"wall": 33746.517, "step": 1960, "loss": 2.4948665698369346, "lm_loss": 2.4948665698369346, "mtp_loss": 0.0, "lr": 0.023620689655172417, "grad_norm": 0.33685585856437683, "tokens": 578027520, "adamw_lr": 0.00037793103448275863, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.416972962021828, "seq_len": 4096} {"wall": 33951.095, "step": 1970, "loss": 2.3471201757589975, "lm_loss": 2.3471201757589975, "mtp_loss": 0.0, "lr": 0.02372844827586207, "grad_norm": 0.34831371903419495, "tokens": 580976640, "adamw_lr": 0.0003796551724137931, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.4117079282800358, "seq_len": 4096} {"wall": 34155.708, "step": 1980, "loss": 2.2366568793853125, "lm_loss": 2.2366568793853125, "mtp_loss": 0.0, "lr": 0.023836206896551724, "grad_norm": 0.29188409447669983, "tokens": 583925760, "adamw_lr": 0.0003813793103448276, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.3956563368439676, "seq_len": 4096} {"wall": 34359.789, "step": 1990, "loss": 2.2812112967173257, "lm_loss": 2.2812112967173257, "mtp_loss": 0.0, "lr": 0.023943965517241383, "grad_norm": 0.2790980339050293, "tokens": 586874880, "adamw_lr": 0.0003831034482758621, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.3865343615412713, "seq_len": 4096} {"wall": 34566.882, "eval_loss": 2.0821542739868164, "eval_ppl": 8.02173137664795, "step": 2000} {"wall": 34566.884, "step": 2000, "loss": 2.3030193199714026, "lm_loss": 2.3030193199714026, "mtp_loss": 0.0, "lr": 0.024051724137931035, "grad_norm": 0.321010559797287, "tokens": 589824000, "adamw_lr": 0.0003848275862068965, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.396643881003062, "seq_len": 4096} {"wall": 34848.469, "step": 2010, "loss": 2.3700279196103415, "lm_loss": 2.3700279196103415, "mtp_loss": 0.0, "lr": 0.02415948275862069, "grad_norm": 0.3494803011417389, "tokens": 592773120, "adamw_lr": 0.00038655172413793105, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.4113460943102836, "seq_len": 4096} {"wall": 35052.437, "step": 2020, "loss": 2.4243109275897345, "lm_loss": 2.4243109275897345, "mtp_loss": 0.0, "lr": 0.024267241379310345, "grad_norm": 0.30334383249282837, "tokens": 595722240, "adamw_lr": 0.0003882758620689655, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.4035588825742407, "seq_len": 4096} {"wall": 35256.871, "step": 2030, "loss": 2.5345825403928757, "lm_loss": 2.5345825403928757, "mtp_loss": 0.0, "lr": 0.024375, "grad_norm": 0.35708311200141907, "tokens": 598671360, "adamw_lr": 0.00039, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.4201062068343164, "seq_len": 4096} {"wall": 35461.329, "step": 2040, "loss": 2.4152799397706985, "lm_loss": 2.4152799397706985, "mtp_loss": 0.0, "lr": 0.02448275862068966, "grad_norm": 0.27107542753219604, "tokens": 601620480, "adamw_lr": 0.0003917241379310345, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.4279615705211954, "seq_len": 4096} {"wall": 35665.501, "step": 2050, "loss": 2.4163536032040915, "lm_loss": 2.4163536032040915, "mtp_loss": 0.0, "lr": 0.02459051724137931, "grad_norm": 0.3091508746147156, "tokens": 604569600, "adamw_lr": 0.00039344827586206894, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.420803375542164, "seq_len": 4096} {"wall": 35870.288, "step": 2060, "loss": 2.457728256781896, "lm_loss": 2.457728256781896, "mtp_loss": 0.0, "lr": 0.024698275862068967, "grad_norm": 0.33579692244529724, "tokens": 607518720, "adamw_lr": 0.00039517241379310346, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.410672289133072, "seq_len": 4096} {"wall": 36074.964, "step": 2070, "loss": 2.3755461325248084, "lm_loss": 2.3755461325248084, "mtp_loss": 0.0, "lr": 0.024806034482758622, "grad_norm": 0.3201757073402405, "tokens": 610467840, "adamw_lr": 0.00039689655172413794, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.366001576681932, "seq_len": 4096} {"wall": 36279.384, "step": 2080, "loss": 2.2992865989605584, "lm_loss": 2.2992865989605584, "mtp_loss": 0.0, "lr": 0.024913793103448278, "grad_norm": 0.27776074409484863, "tokens": 613416960, "adamw_lr": 0.0003986206896551724, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.4182087923089663, "seq_len": 4096} {"wall": 36483.852, "step": 2090, "loss": 2.4489999413490295, "lm_loss": 2.4489999413490295, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.321524053812027, "tokens": 616366080, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.4214704759418963, "seq_len": 4096} {"wall": 36690.669, "eval_loss": 2.096211865544319, "eval_ppl": 8.135293960571289, "step": 2100} {"wall": 36690.671, "step": 2100, "loss": 2.5555553883314133, "lm_loss": 2.5555553883314133, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.30124107003211975, "tokens": 619315200, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.4348988299568495, "seq_len": 4096} {"wall": 36975.941, "step": 2110, "loss": 2.434971039493879, "lm_loss": 2.434971039493879, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.32397380471229553, "tokens": 622264320, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.4002451529105504, "seq_len": 4096} {"wall": 37180.773, "step": 2120, "loss": 2.3529858539501824, "lm_loss": 2.3529858539501824, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.293853223323822, "tokens": 625213440, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.37815696845452, "seq_len": 4096} {"wall": 37385.574, "step": 2130, "loss": 2.4117353409528732, "lm_loss": 2.4117353409528732, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.29792821407318115, "tokens": 628162560, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.4537356346845627, "seq_len": 4096} {"wall": 37590.281, "step": 2140, "loss": 2.4829519043366113, "lm_loss": 2.4829519043366113, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.33825138211250305, "tokens": 631111680, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.4335977166891096, "seq_len": 4096} {"wall": 37794.926, "step": 2150, "loss": 2.3943870017925897, "lm_loss": 2.3943870017925897, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2787977457046509, "tokens": 634060800, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.405546272794406, "seq_len": 4096} {"wall": 37999.527, "step": 2160, "loss": 2.499101087450981, "lm_loss": 2.499101087450981, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3435911238193512, "tokens": 637009920, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.446065755188465, "seq_len": 4096} {"wall": 38204.457, "step": 2170, "loss": 2.492062916358312, "lm_loss": 2.492062916358312, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3451925814151764, "tokens": 639959040, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.4808782761295634, "seq_len": 4096} {"wall": 38408.893, "step": 2180, "loss": 2.384795665740967, "lm_loss": 2.384795665740967, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.288972944021225, "tokens": 642908160, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.398649915804466, "seq_len": 4096} {"wall": 38613.49, "step": 2190, "loss": 2.3532824367284775, "lm_loss": 2.3532824367284775, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2802496552467346, "tokens": 645857280, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.407958609362443, "seq_len": 4096} {"wall": 38820.678, "eval_loss": 2.0954036563634872, "eval_ppl": 8.128721237182617, "step": 2200} {"wall": 38820.68, "step": 2200, "loss": 2.550136998295784, "lm_loss": 2.550136998295784, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3746642470359802, "tokens": 648806400, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.407705707848072, "seq_len": 4096} {"wall": 39109.574, "step": 2210, "loss": 2.43685973683993, "lm_loss": 2.43685973683993, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3076431453227997, "tokens": 651755520, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.4046191816528637, "seq_len": 4096} {"wall": 39314.159, "step": 2220, "loss": 2.4166626632213593, "lm_loss": 2.4166626632213593, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.33470240235328674, "tokens": 654704640, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.4166905616720515, "seq_len": 4096} {"wall": 39518.941, "step": 2230, "loss": 2.488762393593788, "lm_loss": 2.488762393593788, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.27093520760536194, "tokens": 657653760, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.422620540112257, "seq_len": 4096} {"wall": 39723.278, "step": 2240, "loss": 2.3475877344608307, "lm_loss": 2.3475877344608307, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.32109999656677246, "tokens": 660602880, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.393320864935716, "seq_len": 4096} {"wall": 39927.688, "step": 2250, "loss": 2.3667423079411187, "lm_loss": 2.3667423079411187, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3092903196811676, "tokens": 663552000, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.4057193035880724, "seq_len": 4096} {"wall": 40132.645, "step": 2260, "loss": 2.3872450590133667, "lm_loss": 2.3872450590133667, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2997111678123474, "tokens": 666501120, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.4234238957365353, "seq_len": 4096} {"wall": 40336.885, "step": 2270, "loss": 2.4444533735513687, "lm_loss": 2.4444533735513687, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.6872879862785339, "tokens": 669450240, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.423800218850374, "seq_len": 4096} {"wall": 40541.37, "step": 2280, "loss": 2.4483343015114465, "lm_loss": 2.4483343015114465, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.331506609916687, "tokens": 672399360, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.36546676158905, "seq_len": 4096} {"wall": 40746.638, "step": 2290, "loss": 2.4783935298522315, "lm_loss": 2.4783935298522315, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.33166226744651794, "tokens": 675348480, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.4246818482875825, "seq_len": 4096} {"wall": 40953.435, "eval_loss": 2.0909867733716965, "eval_ppl": 8.092896461486816, "step": 2300} {"wall": 40953.437, "step": 2300, "loss": 2.3069461534420648, "lm_loss": 2.3069461534420648, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3211821913719177, "tokens": 678297600, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.374314498404662, "seq_len": 4096} {"wall": 162.242, "eval_loss": 2.090891808271408, "eval_ppl": 8.09212875366211, "step": 2300} {"wall": 889.143, "step": 2310, "loss": 2.2706132183472314, "lm_loss": 2.2706132183472314, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.28068679571151733, "tokens": 681246720, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.302734375, "loss_smoothed": 2.385465321441491, "seq_len": 4096} {"wall": 1094.571, "step": 2320, "loss": 2.489685446023941, "lm_loss": 2.489685446023941, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3078932762145996, "tokens": 684195840, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.302734375, "loss_smoothed": 2.387843908369541, "seq_len": 4096} {"wall": 1300.373, "step": 2330, "loss": 2.5411554723978043, "lm_loss": 2.5411554723978043, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.32691243290901184, "tokens": 687144960, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.302734375, "loss_smoothed": 2.43782646805048, "seq_len": 4096} {"wall": 1506.163, "step": 2340, "loss": 2.4947382311026254, "lm_loss": 2.4947382311026254, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.32675880193710327, "tokens": 690094080, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.302734375, "loss_smoothed": 2.394199187060197, "seq_len": 4096} {"wall": 1711.721, "step": 2350, "loss": 2.450049708286921, "lm_loss": 2.450049708286921, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.30903688073158264, "tokens": 693043200, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.302734375, "loss_smoothed": 2.404519961774349, "seq_len": 4096} {"wall": 1916.815, "step": 2360, "loss": 2.3869230399529138, "lm_loss": 2.3869230399529138, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.4353170394897461, "tokens": 695992320, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.302734375, "loss_smoothed": 2.39206033796072, "seq_len": 4096} {"wall": 2122.369, "step": 2370, "loss": 2.410594458381335, "lm_loss": 2.410594458381335, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2802353799343109, "tokens": 698941440, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.302734375, "loss_smoothed": 2.4252615099151926, "seq_len": 4096} {"wall": 2327.221, "step": 2380, "loss": 2.4271848797798157, "lm_loss": 2.4271848797798157, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.32658785581588745, "tokens": 701890560, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.302734375, "loss_smoothed": 2.3758485411604244, "seq_len": 4096} {"wall": 2532.265, "step": 2390, "loss": 2.4123382220665612, "lm_loss": 2.4123382220665612, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3463912308216095, "tokens": 704839680, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.302734375, "loss_smoothed": 2.3773457229137422, "seq_len": 4096} {"wall": 2739.884, "eval_loss": 2.0789602398872375, "eval_ppl": 7.996150016784668, "step": 2400} {"wall": 2739.886, "step": 2400, "loss": 2.3266218354304633, "lm_loss": 2.3266218354304633, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2958207130432129, "tokens": 707788800, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.302734375, "loss_smoothed": 2.3711629832784338, "seq_len": 4096} {"wall": 2957.932, "step": 2410, "loss": 2.3676431477069855, "lm_loss": 2.3676431477069855, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3131493031978607, "tokens": 710737920, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.407296201090018, "seq_len": 4096} {"wall": 3163.025, "step": 2420, "loss": 2.3523744493722916, "lm_loss": 2.3523744493722916, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3055857717990875, "tokens": 713687040, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.3410366661846638, "seq_len": 4096} {"wall": 3368.64, "step": 2430, "loss": 2.4736185520887375, "lm_loss": 2.4736185520887375, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2714441418647766, "tokens": 716636160, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.424522533019384, "seq_len": 4096} {"wall": 3573.938, "step": 2440, "loss": 2.527541155616442, "lm_loss": 2.527541155616442, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3350970149040222, "tokens": 719585280, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.408432550728321, "seq_len": 4096} {"wall": 3779.639, "step": 2450, "loss": 2.5754251132408776, "lm_loss": 2.5754251132408776, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.30403029918670654, "tokens": 722534400, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.412079118192196, "seq_len": 4096} {"wall": 3985.662, "step": 2460, "loss": 2.3990454971790314, "lm_loss": 2.3990454971790314, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3041885197162628, "tokens": 725483520, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.424768832822641, "seq_len": 4096} {"wall": 4191.317, "step": 2470, "loss": 2.312309349576632, "lm_loss": 2.312309349576632, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3475656807422638, "tokens": 728432640, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.390899403144916, "seq_len": 4096} {"wall": 4397.68, "step": 2480, "loss": 2.3463843911886215, "lm_loss": 2.3463843911886215, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2822207510471344, "tokens": 731381760, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.384663338462512, "seq_len": 4096} {"wall": 4603.267, "step": 2490, "loss": 2.307203029592832, "lm_loss": 2.307203029592832, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.393951952457428, "tokens": 734330880, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.3811200072367984, "seq_len": 4096} {"wall": 4811.6, "eval_loss": 2.075627326965332, "eval_ppl": 7.969544410705566, "step": 2500} {"wall": 4811.603, "step": 2500, "loss": 2.449363276362419, "lm_loss": 2.449363276362419, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3427805006504059, "tokens": 737280000, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.4044167454044025, "seq_len": 4096} {"wall": 5096.669, "step": 2510, "loss": 2.4028195589780807, "lm_loss": 2.4028195589780807, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.29855310916900635, "tokens": 740229120, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.498046875, "loss_smoothed": 2.4142402703563373, "seq_len": 4096} {"wall": 5302.491, "step": 2520, "loss": 2.3409312466780343, "lm_loss": 2.3409312466780343, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.26909878849983215, "tokens": 743178240, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.498046875, "loss_smoothed": 2.391053719321887, "seq_len": 4096} {"wall": 5508.597, "step": 2530, "loss": 2.4916159907976785, "lm_loss": 2.4916159907976785, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3083780109882355, "tokens": 746127360, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.498046875, "loss_smoothed": 2.382126798232396, "seq_len": 4096} {"wall": 5713.971, "step": 2540, "loss": 2.434667855501175, "lm_loss": 2.434667855501175, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3472559452056885, "tokens": 749076480, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.498046875, "loss_smoothed": 2.410599689185619, "seq_len": 4096} {"wall": 5919.725, "step": 2550, "loss": 2.1796451409657798, "lm_loss": 2.1796451409657798, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.30701443552970886, "tokens": 752025600, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.498046875, "loss_smoothed": 2.399655569096406, "seq_len": 4096} {"wall": 6125.766, "step": 2560, "loss": 2.477734699845314, "lm_loss": 2.477734699845314, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.5095551609992981, "tokens": 754974720, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.498046875, "loss_smoothed": 2.4253319792449473, "seq_len": 4096} {"wall": 6331.155, "step": 2570, "loss": 2.225933546821276, "lm_loss": 2.225933546821276, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2783104479312897, "tokens": 757923840, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.498046875, "loss_smoothed": 2.4121640731890994, "seq_len": 4096} {"wall": 6537.038, "step": 2580, "loss": 2.364851012825966, "lm_loss": 2.364851012825966, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3034749925136566, "tokens": 760872960, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.498046875, "loss_smoothed": 2.411417755981286, "seq_len": 4096} {"wall": 6742.1, "step": 2590, "loss": 2.4402101635932922, "lm_loss": 2.4402101635932922, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.30848437547683716, "tokens": 763822080, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.498046875, "loss_smoothed": 2.3923341557383537, "seq_len": 4096} {"wall": 6949.889, "eval_loss": 2.0711669325828552, "eval_ppl": 7.93407678604126, "step": 2600} {"wall": 6949.89, "step": 2600, "loss": 2.3969432512919107, "lm_loss": 2.3969432512919107, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.292197048664093, "tokens": 766771200, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.498046875, "loss_smoothed": 2.4261590123176573, "seq_len": 4096} {"wall": 7227.127, "step": 2610, "loss": 2.3251306265592575, "lm_loss": 2.3251306265592575, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2713758945465088, "tokens": 769720320, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.402927601834138, "seq_len": 4096} {"wall": 7432.707, "step": 2620, "loss": 2.331801394621531, "lm_loss": 2.331801394621531, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2876489460468292, "tokens": 772669440, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.4198328693707785, "seq_len": 4096} {"wall": 7638.252, "step": 2630, "loss": 2.3886264910300574, "lm_loss": 2.3886264910300574, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.33785319328308105, "tokens": 775618560, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.3898462072014808, "seq_len": 4096} {"wall": 7843.832, "step": 2640, "loss": 2.4271352291107178, "lm_loss": 2.4271352291107178, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.335702508687973, "tokens": 778567680, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.384073333442211, "seq_len": 4096} {"wall": 8049.281, "step": 2650, "loss": 2.4418605218331018, "lm_loss": 2.4418605218331018, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.378488153219223, "tokens": 781516800, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.381879515449206, "seq_len": 4096} {"wall": 8254.78, "step": 2660, "loss": 2.3277864108482995, "lm_loss": 2.3277864108482995, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.29394999146461487, "tokens": 784465920, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.3898412068684896, "seq_len": 4096} {"wall": 8460.272, "step": 2670, "loss": 2.4613911310831704, "lm_loss": 2.4613911310831704, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.32561132311820984, "tokens": 787415040, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.4174506296714147, "seq_len": 4096} {"wall": 8665.813, "step": 2680, "loss": 2.409371773401896, "lm_loss": 2.409371773401896, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.35828545689582825, "tokens": 790364160, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.420901363591353, "seq_len": 4096} {"wall": 8871.702, "step": 2690, "loss": 2.340522920091947, "lm_loss": 2.340522920091947, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.32498157024383545, "tokens": 793313280, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.3903381129105883, "seq_len": 4096} {"wall": 9079.285, "eval_loss": 2.060535654425621, "eval_ppl": 7.8501739501953125, "step": 2700} {"wall": 9079.287, "step": 2700, "loss": 2.4610305577516556, "lm_loss": 2.4610305577516556, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.27578380703926086, "tokens": 796262400, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.393895389636358, "seq_len": 4096} {"wall": 9359.125, "step": 2710, "loss": 2.3517491767803826, "lm_loss": 2.3517491767803826, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3003229796886444, "tokens": 799211520, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.40359793951114, "seq_len": 4096} {"wall": 9564.239, "step": 2720, "loss": 2.2036605527003608, "lm_loss": 2.2036605527003608, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.32080376148223877, "tokens": 802160640, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.3536525055766107, "seq_len": 4096} {"wall": 9769.596, "step": 2730, "loss": 2.339831148584684, "lm_loss": 2.339831148584684, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3330256938934326, "tokens": 805109760, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.3738512525955833, "seq_len": 4096} {"wall": 9975.262, "step": 2740, "loss": 2.276591271162033, "lm_loss": 2.276591271162033, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.33070310950279236, "tokens": 808058880, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.355255114038785, "seq_len": 4096} {"wall": 10180.301, "step": 2750, "loss": 2.3849065055449805, "lm_loss": 2.3849065055449805, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.28472045063972473, "tokens": 811008000, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.3822258924444517, "seq_len": 4096} {"wall": 10385.671, "step": 2760, "loss": 2.4954823007186255, "lm_loss": 2.4954823007186255, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.35184967517852783, "tokens": 813957120, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.383971809844176, "seq_len": 4096} {"wall": 10591.38, "step": 2770, "loss": 2.3958478470643363, "lm_loss": 2.3958478470643363, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3394794464111328, "tokens": 816906240, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.4022816496590775, "seq_len": 4096} {"wall": 10796.385, "step": 2780, "loss": 2.521375517050425, "lm_loss": 2.521375517050425, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.37089869379997253, "tokens": 819855360, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.3985438148180642, "seq_len": 4096} {"wall": 11001.749, "step": 2790, "loss": 2.401612957318624, "lm_loss": 2.401612957318624, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3365861475467682, "tokens": 822804480, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.3752407665054003, "seq_len": 4096} {"wall": 11209.87, "eval_loss": 2.0609982311725616, "eval_ppl": 7.8538055419921875, "step": 2800} {"wall": 11209.872, "step": 2800, "loss": 2.2842209140459695, "lm_loss": 2.2842209140459695, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.34520211815834045, "tokens": 825753600, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.391816285749276, "seq_len": 4096} {"wall": 11488.232, "step": 2810, "loss": 2.232790986696879, "lm_loss": 2.232790986696879, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.28880175948143005, "tokens": 828702720, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.3788080026706058, "seq_len": 4096} {"wall": 11693.498, "step": 2820, "loss": 2.4666858116785684, "lm_loss": 2.4666858116785684, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3601905405521393, "tokens": 831651840, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.406011632581552, "seq_len": 4096} {"wall": 11898.748, "step": 2830, "loss": 2.3082680652538934, "lm_loss": 2.3082680652538934, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.31325554847717285, "tokens": 834600960, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.3629752049843473, "seq_len": 4096} {"wall": 12104.004, "step": 2840, "loss": 2.3504309554894767, "lm_loss": 2.3504309554894767, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.36122673749923706, "tokens": 837550080, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.369897318383058, "seq_len": 4096} {"wall": 12309.307, "step": 2850, "loss": 2.2860397001107535, "lm_loss": 2.2860397001107535, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.33700090646743774, "tokens": 840499200, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.3361700157324474, "seq_len": 4096} {"wall": 12514.584, "step": 2860, "loss": 2.431338498989741, "lm_loss": 2.431338498989741, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.27545809745788574, "tokens": 843448320, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.3762092262506487, "seq_len": 4096} {"wall": 12719.844, "step": 2870, "loss": 2.4216273923714957, "lm_loss": 2.4216273923714957, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.323720782995224, "tokens": 846397440, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.3392245170970756, "seq_len": 4096} {"wall": 12925.113, "step": 2880, "loss": 2.310641830166181, "lm_loss": 2.310641830166181, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3895651698112488, "tokens": 849346560, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.368773040672143, "seq_len": 4096} {"wall": 13130.337, "step": 2890, "loss": 2.4803576519091926, "lm_loss": 2.4803576519091926, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.28973284363746643, "tokens": 852295680, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.3592667351166408, "seq_len": 4096} {"wall": 13338.006, "eval_loss": 2.0456356704235077, "eval_ppl": 7.734073638916016, "step": 2900} {"wall": 13338.008, "step": 2900, "loss": 2.4372294346491494, "lm_loss": 2.4372294346491494, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3169550597667694, "tokens": 855244800, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.3592675333221753, "seq_len": 4096} {"wall": 13614.79, "step": 2910, "loss": 2.3777088820934296, "lm_loss": 2.3777088820934296, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2525681257247925, "tokens": 858193920, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.417883912225564, "seq_len": 4096} {"wall": 13820.166, "step": 2920, "loss": 2.3323667695124946, "lm_loss": 2.3323667695124946, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3485579788684845, "tokens": 861143040, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.330772883196672, "seq_len": 4096} {"wall": 14025.187, "step": 2930, "loss": 2.13866226375103, "lm_loss": 2.13866226375103, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2981953024864197, "tokens": 864092160, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.3474927400549253, "seq_len": 4096} {"wall": 14230.845, "step": 2940, "loss": 2.21536752084891, "lm_loss": 2.21536752084891, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.30154988169670105, "tokens": 867041280, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.399258427321911, "seq_len": 4096} {"wall": 14436.274, "step": 2950, "loss": 2.32833224038283, "lm_loss": 2.32833224038283, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.28879597783088684, "tokens": 869990400, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.3975314756234485, "seq_len": 4096} {"wall": 14641.657, "step": 2960, "loss": 2.3620364516973495, "lm_loss": 2.3620364516973495, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.31192049384117126, "tokens": 872939520, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.364880831539631, "seq_len": 4096} {"wall": 14847.13, "step": 2970, "loss": 2.386471246679624, "lm_loss": 2.386471246679624, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3044072389602661, "tokens": 875888640, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.3544550155599913, "seq_len": 4096} {"wall": 15052.542, "step": 2980, "loss": 2.504521355032921, "lm_loss": 2.504521355032921, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.35281020402908325, "tokens": 878837760, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.359475727379322, "seq_len": 4096} {"wall": 15257.961, "step": 2990, "loss": 2.344013497233391, "lm_loss": 2.344013497233391, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.26311078667640686, "tokens": 881786880, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.332331005235513, "seq_len": 4096} {"wall": 15465.931, "eval_loss": 2.035447135567665, "eval_ppl": 7.655674457550049, "step": 3000} {"wall": 15465.932, "step": 3000, "loss": 2.4098018060127893, "lm_loss": 2.4098018060127893, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.27695131301879883, "tokens": 884736000, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.3035868500669796, "seq_len": 4096} {"wall": 15747.364, "step": 3010, "loss": 2.1593175580104194, "lm_loss": 2.1593175580104194, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.29180899262428284, "tokens": 887685120, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.3650080129504203, "seq_len": 4096} {"wall": 15953.036, "step": 3020, "loss": 2.385241463780403, "lm_loss": 2.385241463780403, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.30845385789871216, "tokens": 890634240, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.3436903660496076, "seq_len": 4096} {"wall": 16158.683, "step": 3030, "loss": 2.321258614460627, "lm_loss": 2.321258614460627, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2814440429210663, "tokens": 893583360, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.289335835476716, "seq_len": 4096} {"wall": 16364.364, "step": 3040, "loss": 2.364855572581291, "lm_loss": 2.364855572581291, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3311496078968048, "tokens": 896532480, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.381845504542192, "seq_len": 4096} {"wall": 16570.095, "step": 3050, "loss": 2.240134318669637, "lm_loss": 2.240134318669637, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3348594009876251, "tokens": 899481600, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.3312477104365827, "seq_len": 4096} {"wall": 16775.767, "step": 3060, "loss": 2.2396445026000342, "lm_loss": 2.2396445026000342, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.30847710371017456, "tokens": 902430720, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.3783456129332383, "seq_len": 4096} {"wall": 16981.511, "step": 3070, "loss": 2.486742759744326, "lm_loss": 2.486742759744326, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3033945858478546, "tokens": 905379840, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.3467153762777646, "seq_len": 4096} {"wall": 17187.228, "step": 3080, "loss": 2.325472816824913, "lm_loss": 2.325472816824913, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.23671194911003113, "tokens": 908328960, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.323434173564116, "seq_len": 4096} {"wall": 17392.823, "step": 3090, "loss": 2.399363632003466, "lm_loss": 2.399363632003466, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.31685158610343933, "tokens": 911278080, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.3898762116829557, "seq_len": 4096} {"wall": 17600.923, "eval_loss": 2.0321144610643387, "eval_ppl": 7.6302032470703125, "step": 3100} {"wall": 17600.925, "step": 3100, "loss": 2.4235828717549643, "lm_loss": 2.4235828717549643, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.27814245223999023, "tokens": 914227200, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.360673382381598, "seq_len": 4096} {"wall": 17884.353, "step": 3110, "loss": 2.302197113633156, "lm_loss": 2.302197113633156, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.31770089268684387, "tokens": 917176320, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.365491496026516, "seq_len": 4096} {"wall": 18089.54, "step": 3120, "loss": 2.363096127907435, "lm_loss": 2.363096127907435, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2824629545211792, "tokens": 920125440, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.358985222876072, "seq_len": 4096} {"wall": 18295.091, "step": 3130, "loss": 2.3429997265338898, "lm_loss": 2.3429997265338898, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3491456210613251, "tokens": 923074560, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.386794222891331, "seq_len": 4096} {"wall": 18500.937, "step": 3140, "loss": 2.2874504923820496, "lm_loss": 2.2874504923820496, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3435392379760742, "tokens": 926023680, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.3095869498948254, "seq_len": 4096} {"wall": 18706.158, "step": 3150, "loss": 2.3590458780527115, "lm_loss": 2.3590458780527115, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3202435076236725, "tokens": 928972800, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.347825280576944, "seq_len": 4096} {"wall": 18911.665, "step": 3160, "loss": 2.4065480679273605, "lm_loss": 2.4065480679273605, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.36662420630455017, "tokens": 931921920, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.377879308909178, "seq_len": 4096} {"wall": 19117.479, "step": 3170, "loss": 2.3739805022875466, "lm_loss": 2.3739805022875466, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3090190589427948, "tokens": 934871040, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.3546909819046657, "seq_len": 4096} {"wall": 19322.733, "step": 3180, "loss": 2.2691655308008194, "lm_loss": 2.2691655308008194, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2956485152244568, "tokens": 937820160, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.3093202084302904, "seq_len": 4096} {"wall": 19528.288, "step": 3190, "loss": 2.312223027149836, "lm_loss": 2.312223027149836, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3092699348926544, "tokens": 940769280, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.3608381976683934, "seq_len": 4096} {"wall": 19736.529, "eval_loss": 2.0319669246673584, "eval_ppl": 7.629077434539795, "step": 3200} {"wall": 19736.53, "step": 3200, "loss": 2.318228860696157, "lm_loss": 2.318228860696157, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.31854087114334106, "tokens": 943718400, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.3901913111408555, "seq_len": 4096} {"wall": 20019.564, "step": 3210, "loss": 2.32607501745224, "lm_loss": 2.32607501745224, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.33806124329566956, "tokens": 946667520, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.348513483007749, "seq_len": 4096} {"wall": 20224.981, "step": 3220, "loss": 2.4387807150681815, "lm_loss": 2.4387807150681815, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3114839196205139, "tokens": 949616640, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.377580572664738, "seq_len": 4096} {"wall": 20430.549, "step": 3230, "loss": 2.334783817330996, "lm_loss": 2.334783817330996, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.29818636178970337, "tokens": 952565760, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.35209970275561, "seq_len": 4096} {"wall": 20636.189, "step": 3240, "loss": 2.4279608726501465, "lm_loss": 2.4279608726501465, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3271477222442627, "tokens": 955514880, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.362203133106232, "seq_len": 4096} {"wall": 20841.987, "step": 3250, "loss": 2.4272336860497794, "lm_loss": 2.4272336860497794, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3461313247680664, "tokens": 958464000, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.4086137898266315, "seq_len": 4096} {"wall": 21048.236, "step": 3260, "loss": 2.3334175248940787, "lm_loss": 2.3334175248940787, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2951754629611969, "tokens": 961413120, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.360353251794974, "seq_len": 4096} {"wall": 21254.38, "step": 3270, "loss": 2.2359056919813156, "lm_loss": 2.2359056919813156, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.36090636253356934, "tokens": 964362240, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.333083864549796, "seq_len": 4096} {"wall": 21460.477, "step": 3280, "loss": 2.322725291053454, "lm_loss": 2.322725291053454, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3210364878177643, "tokens": 967311360, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.3213381613294284, "seq_len": 4096} {"wall": 21666.588, "step": 3290, "loss": 2.335369735956192, "lm_loss": 2.335369735956192, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.27157264947891235, "tokens": 970260480, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.3461535945534706, "seq_len": 4096} {"wall": 21874.927, "eval_loss": 2.0156021416187286, "eval_ppl": 7.505245208740234, "step": 3300} {"wall": 21874.929, "step": 3300, "loss": 2.2324905345837274, "lm_loss": 2.2324905345837274, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.26551809906959534, "tokens": 973209600, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.312051945924759, "seq_len": 4096} {"wall": 22155.21, "step": 3310, "loss": 2.3968216329813004, "lm_loss": 2.3968216329813004, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3036367893218994, "tokens": 976158720, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.3194799080491064, "seq_len": 4096} {"wall": 22360.885, "step": 3320, "loss": 2.3335942874352136, "lm_loss": 2.3335942874352136, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.28027021884918213, "tokens": 979107840, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.3491282522678376, "seq_len": 4096} {"wall": 22566.382, "step": 3330, "loss": 2.317744399110476, "lm_loss": 2.317744399110476, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3634820282459259, "tokens": 982056960, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.3179487590988477, "seq_len": 4096} {"wall": 22771.83, "step": 3340, "loss": 2.1853387355804443, "lm_loss": 2.1853387355804443, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3103221654891968, "tokens": 985006080, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.3139469032486284, "seq_len": 4096} {"wall": 22977.348, "step": 3350, "loss": 2.2853725949923196, "lm_loss": 2.2853725949923196, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.26520073413848877, "tokens": 987955200, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.3181910892327626, "seq_len": 4096} {"wall": 23182.717, "step": 3360, "loss": 2.2502723385890326, "lm_loss": 2.2502723385890326, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 1.914910912513733, "tokens": 990904320, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.3562310591340063, "seq_len": 4096} {"wall": 23388.106, "step": 3370, "loss": 2.379084919889768, "lm_loss": 2.379084919889768, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3482917249202728, "tokens": 993853440, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.338215977946917, "seq_len": 4096} {"wall": 23593.47, "step": 3380, "loss": 2.210047880808512, "lm_loss": 2.210047880808512, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.28022828698158264, "tokens": 996802560, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.338292007644971, "seq_len": 4096} {"wall": 23798.761, "step": 3390, "loss": 2.3839308669169745, "lm_loss": 2.3839308669169745, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3019014000892639, "tokens": 999751680, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.3054998757938545, "seq_len": 4096} {"wall": 24006.512, "eval_loss": 2.0306787937879562, "eval_ppl": 7.619256019592285, "step": 3400} {"wall": 24006.514, "step": 3400, "loss": 2.3980544010798135, "lm_loss": 2.3980544010798135, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2622263729572296, "tokens": 1002700800, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.364986236145099, "seq_len": 4096} {"wall": 24289.671, "step": 3410, "loss": 2.4128857403993607, "lm_loss": 2.4128857403993607, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.329630047082901, "tokens": 1005649920, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.3232215697566665, "seq_len": 4096} {"wall": 24494.76, "step": 3420, "loss": 2.267313078045845, "lm_loss": 2.267313078045845, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3010055422782898, "tokens": 1008599040, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.294047260781129, "seq_len": 4096} {"wall": 24700.275, "step": 3430, "loss": 2.385045592983564, "lm_loss": 2.385045592983564, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.32748404145240784, "tokens": 1011548160, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.337817900876204, "seq_len": 4096} {"wall": 24906.335, "step": 3440, "loss": 2.3113924711942673, "lm_loss": 2.3113924711942673, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2911752164363861, "tokens": 1014497280, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.3161376759409906, "seq_len": 4096} {"wall": 25111.74, "step": 3450, "loss": 2.2950114955504737, "lm_loss": 2.2950114955504737, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3076159954071045, "tokens": 1017446400, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.3246580233176553, "seq_len": 4096} {"wall": 25317.296, "step": 3460, "loss": 2.297581826647123, "lm_loss": 2.297581826647123, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.27815988659858704, "tokens": 1020395520, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.3300435073673724, "seq_len": 4096} {"wall": 25523.165, "step": 3470, "loss": 2.3048068831364312, "lm_loss": 2.3048068831364312, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3263028562068939, "tokens": 1023344640, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.3093763013680775, "seq_len": 4096} {"wall": 25728.318, "step": 3480, "loss": 2.285806248585383, "lm_loss": 2.285806248585383, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2940136790275574, "tokens": 1026293760, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.316733593742053, "seq_len": 4096} {"wall": 25933.807, "step": 3490, "loss": 2.3412947356700897, "lm_loss": 2.3412947356700897, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.316491037607193, "tokens": 1029242880, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.3207935964067774, "seq_len": 4096} {"wall": 26141.989, "eval_loss": 2.011107102036476, "eval_ppl": 7.471585273742676, "step": 3500} {"wall": 26141.991, "step": 3500, "loss": 2.4466043561697006, "lm_loss": 2.4466043561697006, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.34116217494010925, "tokens": 1032192000, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.3215810686349867, "seq_len": 4096} {"wall": 26419.801, "step": 3510, "loss": 2.3045889188845954, "lm_loss": 2.3045889188845954, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3338564336299896, "tokens": 1035141120, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.314079688489437, "seq_len": 4096} {"wall": 26625.626, "step": 3520, "loss": 2.4232520759105682, "lm_loss": 2.4232520759105682, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.27849486470222473, "tokens": 1038090240, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.3411509002248447, "seq_len": 4096} {"wall": 26830.809, "step": 3530, "loss": 2.3898125290870667, "lm_loss": 2.3898125290870667, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.30553489923477173, "tokens": 1041039360, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.361477561791738, "seq_len": 4096} {"wall": 27036.749, "step": 3540, "loss": 2.247066323955854, "lm_loss": 2.247066323955854, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3131866455078125, "tokens": 1043988480, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.329711863398552, "seq_len": 4096} {"wall": 27242.247, "step": 3550, "loss": 2.3632258077462516, "lm_loss": 2.3632258077462516, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.26814278960227966, "tokens": 1046937600, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.3492312133312225, "seq_len": 4096} {"wall": 27447.845, "step": 3560, "loss": 2.2994448468089104, "lm_loss": 2.2994448468089104, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2972908318042755, "tokens": 1049886720, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.335952853659789, "seq_len": 4096} {"wall": 27653.376, "step": 3570, "loss": 2.3210372428099313, "lm_loss": 2.3210372428099313, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.35684874653816223, "tokens": 1052835840, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.331839783489704, "seq_len": 4096} {"wall": 27858.917, "step": 3580, "loss": 2.411817267537117, "lm_loss": 2.411817267537117, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3381896913051605, "tokens": 1055784960, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.3215047483642897, "seq_len": 4096} {"wall": 28064.92, "step": 3590, "loss": 2.310844490925471, "lm_loss": 2.310844490925471, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.32259708642959595, "tokens": 1058734080, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.2697311341762543, "seq_len": 4096} {"wall": 28272.7, "eval_loss": 2.0063460916280746, "eval_ppl": 7.43609619140625, "step": 3600} {"wall": 28272.702, "step": 3600, "loss": 2.2496517101923623, "lm_loss": 2.2496517101923623, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3343440890312195, "tokens": 1061683200, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.334844469030698, "seq_len": 4096} {"wall": 28551.61, "step": 3610, "loss": 2.0911045918862023, "lm_loss": 2.0911045918862023, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.27868035435676575, "tokens": 1064632320, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.291697395592928, "seq_len": 4096} {"wall": 28756.999, "step": 3620, "loss": 2.3160340785980225, "lm_loss": 2.3160340785980225, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.33949148654937744, "tokens": 1067581440, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.297313145796458, "seq_len": 4096} {"wall": 28962.36, "step": 3630, "loss": 2.236493100722631, "lm_loss": 2.236493100722631, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2840043306350708, "tokens": 1070530560, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.356958057979743, "seq_len": 4096} {"wall": 29167.777, "step": 3640, "loss": 2.2913524309794107, "lm_loss": 2.2913524309794107, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.40449824929237366, "tokens": 1073479680, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.32279443119963, "seq_len": 4096} {"wall": 29373.174, "step": 3650, "loss": 2.194408029317856, "lm_loss": 2.194408029317856, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3063080906867981, "tokens": 1076428800, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.3146276965737345, "seq_len": 4096} {"wall": 29578.636, "step": 3660, "loss": 2.1758248656988144, "lm_loss": 2.1758248656988144, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2901221215724945, "tokens": 1079377920, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.3045253391067186, "seq_len": 4096} {"wall": 29784.068, "step": 3670, "loss": 2.3572417249282203, "lm_loss": 2.3572417249282203, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.30918076634407043, "tokens": 1082327040, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.3018980041146277, "seq_len": 4096} {"wall": 29989.413, "step": 3680, "loss": 2.205589940150579, "lm_loss": 2.205589940150579, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3288930654525757, "tokens": 1085276160, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.2841677019993467, "seq_len": 4096} {"wall": 30194.725, "step": 3690, "loss": 2.3887674063444138, "lm_loss": 2.3887674063444138, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3618822693824768, "tokens": 1088225280, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.300629032899936, "seq_len": 4096} {"wall": 30402.542, "eval_loss": 2.010743170976639, "eval_ppl": 7.468865871429443, "step": 3700} {"wall": 30402.543, "step": 3700, "loss": 2.2586391866207123, "lm_loss": 2.2586391866207123, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3355817198753357, "tokens": 1091174400, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.3770429149270056, "seq_len": 4096} {"wall": 30681.002, "step": 3710, "loss": 2.403900866707166, "lm_loss": 2.403900866707166, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2781667411327362, "tokens": 1094123520, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.3284270733594896, "seq_len": 4096} {"wall": 30885.906, "step": 3720, "loss": 2.2879284024238586, "lm_loss": 2.2879284024238586, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3097768723964691, "tokens": 1097072640, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.3012727064390974, "seq_len": 4096} {"wall": 31091.102, "step": 3730, "loss": 2.2642348557710648, "lm_loss": 2.2642348557710648, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3742814064025879, "tokens": 1100021760, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.3234059090415635, "seq_len": 4096} {"wall": 31296.298, "step": 3740, "loss": 2.3903710693120956, "lm_loss": 2.3903710693120956, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.26004132628440857, "tokens": 1102970880, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.318677811821302, "seq_len": 4096} {"wall": 31501.494, "step": 3750, "loss": 2.2488015592098236, "lm_loss": 2.2488015592098236, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.4414287209510803, "tokens": 1105920000, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.2900486782193186, "seq_len": 4096} {"wall": 31706.664, "step": 3760, "loss": 2.369227876265844, "lm_loss": 2.369227876265844, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2679476737976074, "tokens": 1108869120, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.352008824547132, "seq_len": 4096} {"wall": 31912.129, "step": 3770, "loss": 2.252730848888556, "lm_loss": 2.252730848888556, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.32552221417427063, "tokens": 1111818240, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.2884594105184077, "seq_len": 4096} {"wall": 32117.008, "step": 3780, "loss": 2.331302141149839, "lm_loss": 2.331302141149839, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3281020224094391, "tokens": 1114767360, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.325238207976023, "seq_len": 4096} {"wall": 32322.194, "step": 3790, "loss": 2.4461147487163544, "lm_loss": 2.4461147487163544, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.37889859080314636, "tokens": 1117716480, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.3200571427742642, "seq_len": 4096} {"wall": 32530.026, "eval_loss": 1.9987515062093735, "eval_ppl": 7.379836559295654, "step": 3800} {"wall": 32530.028, "step": 3800, "loss": 2.2688764880100885, "lm_loss": 2.2688764880100885, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.30903297662734985, "tokens": 1120665600, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.3344395865996677, "seq_len": 4096} {"wall": 32804.533, "step": 3810, "loss": 2.3529752691586814, "lm_loss": 2.3529752691586814, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2967120409011841, "tokens": 1123614720, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.2423869922757147, "seq_len": 4096} {"wall": 33009.643, "step": 3820, "loss": 2.2011287758747735, "lm_loss": 2.2011287758747735, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.32086268067359924, "tokens": 1126563840, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.2532337730129557, "seq_len": 4096} {"wall": 33214.784, "step": 3830, "loss": 2.337669238448143, "lm_loss": 2.337669238448143, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.29129090905189514, "tokens": 1129512960, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.2918283368150396, "seq_len": 4096} {"wall": 33419.919, "step": 3840, "loss": 2.3673308193683624, "lm_loss": 2.3673308193683624, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.27737194299697876, "tokens": 1132462080, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.322488239655892, "seq_len": 4096} {"wall": 33625.025, "step": 3850, "loss": 2.337844451268514, "lm_loss": 2.337844451268514, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.32450243830680847, "tokens": 1135411200, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.309301597376664, "seq_len": 4096} {"wall": 33830.089, "step": 3860, "loss": 2.332978660861651, "lm_loss": 2.332978660861651, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2789473533630371, "tokens": 1138360320, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.3504971514145536, "seq_len": 4096} {"wall": 34035.163, "step": 3870, "loss": 2.2675019651651382, "lm_loss": 2.2675019651651382, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3064415752887726, "tokens": 1141309440, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.328681294123332, "seq_len": 4096} {"wall": 34239.938, "step": 3880, "loss": 2.1951493422190347, "lm_loss": 2.1951493422190347, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3021073043346405, "tokens": 1144258560, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.3221607722342013, "seq_len": 4096} {"wall": 34445.375, "step": 3890, "loss": 2.353191795448462, "lm_loss": 2.353191795448462, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.33257433772087097, "tokens": 1147207680, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.3352434587975344, "seq_len": 4096} {"wall": 34652.964, "eval_loss": 1.9969928562641144, "eval_ppl": 7.3668694496154785, "step": 3900} {"wall": 34652.966, "step": 3900, "loss": 2.308309624592463, "lm_loss": 2.308309624592463, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3228517174720764, "tokens": 1150156800, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.3274450048804285, "seq_len": 4096} {"wall": 162.639, "eval_loss": 1.9969928562641144, "eval_ppl": 7.3668694496154785, "step": 3900} {"wall": 900.884, "step": 3910, "loss": 2.2970247666041055, "lm_loss": 2.2970247666041055, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.28618764877319336, "tokens": 1153105920, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.302734375, "loss_smoothed": 2.328149360915025, "seq_len": 4096} {"wall": 1106.242, "step": 3920, "loss": 2.3002373526493707, "lm_loss": 2.3002373526493707, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3822069764137268, "tokens": 1156055040, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.302734375, "loss_smoothed": 2.304272034515937, "seq_len": 4096} {"wall": 1311.526, "step": 3930, "loss": 2.2208579232295356, "lm_loss": 2.2208579232295356, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.28942281007766724, "tokens": 1159004160, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.302734375, "loss_smoothed": 2.2894992878039675, "seq_len": 4096} {"wall": 1516.808, "step": 3940, "loss": 2.353952313462893, "lm_loss": 2.353952313462893, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3709816336631775, "tokens": 1161953280, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.302734375, "loss_smoothed": 2.3333519677321117, "seq_len": 4096} {"wall": 1722.16, "step": 3950, "loss": 2.1431847661733627, "lm_loss": 2.1431847661733627, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.27518323063850403, "tokens": 1164902400, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.302734375, "loss_smoothed": 2.3192242463429773, "seq_len": 4096} {"wall": 1927.595, "step": 3960, "loss": 2.307695910334587, "lm_loss": 2.307695910334587, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3134920597076416, "tokens": 1167851520, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.302734375, "loss_smoothed": 2.274075897037983, "seq_len": 4096} {"wall": 2133.338, "step": 3970, "loss": 2.2303503851095834, "lm_loss": 2.2303503851095834, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3274407386779785, "tokens": 1170800640, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.302734375, "loss_smoothed": 2.3588085139791173, "seq_len": 4096} {"wall": 2338.197, "step": 3980, "loss": 2.4381078531344733, "lm_loss": 2.4381078531344733, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3028520345687866, "tokens": 1173749760, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.302734375, "loss_smoothed": 2.2957155113418897, "seq_len": 4096} {"wall": 2543.284, "step": 3990, "loss": 2.217991669972738, "lm_loss": 2.217991669972738, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2682332396507263, "tokens": 1176698880, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.302734375, "loss_smoothed": 2.299174203723669, "seq_len": 4096} {"wall": 2750.749, "eval_loss": 1.9981351494789124, "eval_ppl": 7.375288963317871, "step": 4000} {"wall": 2750.751, "step": 4000, "loss": 2.426855911811193, "lm_loss": 2.426855911811193, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.34014368057250977, "tokens": 1179648000, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.302734375, "loss_smoothed": 2.321574518084526, "seq_len": 4096} {"wall": 2975.777, "step": 4010, "loss": 2.348714237411817, "lm_loss": 2.348714237411817, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.36971408128738403, "tokens": 1182597120, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.3424620414773623, "seq_len": 4096} {"wall": 3180.878, "step": 4020, "loss": 2.2494720866282782, "lm_loss": 2.2494720866282782, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.25057125091552734, "tokens": 1185546240, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.291462533672651, "seq_len": 4096} {"wall": 3386.622, "step": 4030, "loss": 2.333390181263288, "lm_loss": 2.333390181263288, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3168596625328064, "tokens": 1188495360, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.3422334901988506, "seq_len": 4096} {"wall": 3591.325, "step": 4040, "loss": 2.217664048075676, "lm_loss": 2.217664048075676, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2778990566730499, "tokens": 1191444480, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.295905868709087, "seq_len": 4096} {"wall": 3796.515, "step": 4050, "loss": 2.4154373009999595, "lm_loss": 2.4154373009999595, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.33762702345848083, "tokens": 1194393600, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.318433098991712, "seq_len": 4096} {"wall": 4001.957, "step": 4060, "loss": 2.296932319800059, "lm_loss": 2.296932319800059, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2829109728336334, "tokens": 1197342720, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.2962942828734714, "seq_len": 4096} {"wall": 4206.755, "step": 4070, "loss": 2.4795156121253967, "lm_loss": 2.4795156121253967, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.304851770401001, "tokens": 1200291840, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.3468339939912157, "seq_len": 4096} {"wall": 4411.905, "step": 4080, "loss": 2.35933751364549, "lm_loss": 2.35933751364549, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3406641185283661, "tokens": 1203240960, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.3392301802833875, "seq_len": 4096} {"wall": 4617.343, "step": 4090, "loss": 2.3776279240846634, "lm_loss": 2.3776279240846634, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.39710351824760437, "tokens": 1206190080, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.314330453177293, "seq_len": 4096} {"wall": 4824.467, "eval_loss": 1.987112283706665, "eval_ppl": 7.29443883895874, "step": 4100} {"wall": 4824.47, "step": 4100, "loss": 2.2996339003245034, "lm_loss": 2.2996339003245034, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.341933935880661, "tokens": 1209139200, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.3185249653955298, "seq_len": 4096} {"wall": 5110.106, "step": 4110, "loss": 2.310012459754944, "lm_loss": 2.310012459754944, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.28288760781288147, "tokens": 1212088320, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.498046875, "loss_smoothed": 2.313581250111262, "seq_len": 4096} {"wall": 5315.358, "step": 4120, "loss": 2.486856589714686, "lm_loss": 2.486856589714686, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.38917276263237, "tokens": 1215037440, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.498046875, "loss_smoothed": 2.271176219979922, "seq_len": 4096} {"wall": 5520.489, "step": 4130, "loss": 2.3013138324022293, "lm_loss": 2.3013138324022293, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3129495084285736, "tokens": 1217986560, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.498046875, "loss_smoothed": 2.275946230192979, "seq_len": 4096} {"wall": 5725.802, "step": 4140, "loss": 2.327399119734764, "lm_loss": 2.327399119734764, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2986436188220978, "tokens": 1220935680, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.498046875, "loss_smoothed": 2.287241642177105, "seq_len": 4096} {"wall": 5930.149, "step": 4150, "loss": 2.343650499979655, "lm_loss": 2.343650499979655, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2999875545501709, "tokens": 1223884800, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.498046875, "loss_smoothed": 2.338930256664753, "seq_len": 4096} {"wall": 6135.44, "step": 4160, "loss": 2.3739077796538672, "lm_loss": 2.3739077796538672, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.34482377767562866, "tokens": 1226833920, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.498046875, "loss_smoothed": 2.3536537299553553, "seq_len": 4096} {"wall": 6340.202, "step": 4170, "loss": 2.38982621828715, "lm_loss": 2.38982621828715, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3107546865940094, "tokens": 1229783040, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.498046875, "loss_smoothed": 2.3025260299444197, "seq_len": 4096} {"wall": 6545.865, "step": 4180, "loss": 2.2623259872198105, "lm_loss": 2.2623259872198105, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3683205246925354, "tokens": 1232732160, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.498046875, "loss_smoothed": 2.31588570078214, "seq_len": 4096} {"wall": 6750.584, "step": 4190, "loss": 2.3310880760351815, "lm_loss": 2.3310880760351815, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3517949879169464, "tokens": 1235681280, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.498046875, "loss_smoothed": 2.2686798031131423, "seq_len": 4096} {"wall": 6957.761, "eval_loss": 1.9774283170700073, "eval_ppl": 7.224140644073486, "step": 4200} {"wall": 6957.763, "step": 4200, "loss": 2.141810064514478, "lm_loss": 2.141810064514478, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3614646792411804, "tokens": 1238630400, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.498046875, "loss_smoothed": 2.317626625796159, "seq_len": 4096} {"wall": 7236.444, "step": 4210, "loss": 2.2448582698901496, "lm_loss": 2.2448582698901496, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2581794857978821, "tokens": 1241579520, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.3075753505031265, "seq_len": 4096} {"wall": 7441.194, "step": 4220, "loss": 2.301096553603808, "lm_loss": 2.301096553603808, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.28373509645462036, "tokens": 1244528640, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.279228546967109, "seq_len": 4096} {"wall": 7645.967, "step": 4230, "loss": 2.405691390236219, "lm_loss": 2.405691390236219, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3027753531932831, "tokens": 1247477760, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.3400250382721426, "seq_len": 4096} {"wall": 7850.738, "step": 4240, "loss": 2.243665431936582, "lm_loss": 2.243665431936582, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.30280712246894836, "tokens": 1250426880, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.329560128847758, "seq_len": 4096} {"wall": 8055.664, "step": 4250, "loss": 2.268511558572451, "lm_loss": 2.268511558572451, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3328533172607422, "tokens": 1253376000, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.292477312187354, "seq_len": 4096} {"wall": 8260.432, "step": 4260, "loss": 2.3092797100543976, "lm_loss": 2.3092797100543976, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2755608856678009, "tokens": 1256325120, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.3010345175862312, "seq_len": 4096} {"wall": 8465.173, "step": 4270, "loss": 2.202350233991941, "lm_loss": 2.202350233991941, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.37042689323425293, "tokens": 1259274240, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.314405974249045, "seq_len": 4096} {"wall": 8670.072, "step": 4280, "loss": 2.2184703052043915, "lm_loss": 2.2184703052043915, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.24145135283470154, "tokens": 1262223360, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.2615089163184168, "seq_len": 4096} {"wall": 8874.963, "step": 4290, "loss": 2.4344564974308014, "lm_loss": 2.4344564974308014, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.35482439398765564, "tokens": 1265172480, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.307204947868983, "seq_len": 4096} {"wall": 9082.181, "eval_loss": 1.9796053171157837, "eval_ppl": 7.239884853363037, "step": 4300} {"wall": 9082.183, "step": 4300, "loss": 2.3339691360791526, "lm_loss": 2.3339691360791526, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.32626670598983765, "tokens": 1268121600, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.3273246005177497, "seq_len": 4096} {"wall": 9363.941, "step": 4310, "loss": 2.3369947175184884, "lm_loss": 2.3369947175184884, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3067188560962677, "tokens": 1271070720, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.2817342191934586, "seq_len": 4096} {"wall": 9568.161, "step": 4320, "loss": 2.200664073228836, "lm_loss": 2.200664073228836, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.29374730587005615, "tokens": 1274019840, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.2985720127820968, "seq_len": 4096} {"wall": 9772.866, "step": 4330, "loss": 2.3013075639804206, "lm_loss": 2.3013075639804206, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2830943763256073, "tokens": 1276968960, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.2904707312583925, "seq_len": 4096} {"wall": 9978.006, "step": 4340, "loss": 2.338807021578153, "lm_loss": 2.338807021578153, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3013685941696167, "tokens": 1279918080, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.284991158048312, "seq_len": 4096} {"wall": 10182.369, "step": 4350, "loss": 2.277171572049459, "lm_loss": 2.277171572049459, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3245026171207428, "tokens": 1282867200, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.274509550631046, "seq_len": 4096} {"wall": 10387.075, "step": 4360, "loss": 2.146371175845464, "lm_loss": 2.146371175845464, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3139601945877075, "tokens": 1285816320, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.258426134785016, "seq_len": 4096} {"wall": 10592.14, "step": 4370, "loss": 2.1007209867239, "lm_loss": 2.1007209867239, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.32639554142951965, "tokens": 1288765440, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.288391343255838, "seq_len": 4096} {"wall": 10796.419, "step": 4380, "loss": 2.2130464439590773, "lm_loss": 2.2130464439590773, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.37530139088630676, "tokens": 1291714560, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.2654297180473804, "seq_len": 4096} {"wall": 11001.165, "step": 4390, "loss": 2.2409280141194663, "lm_loss": 2.2409280141194663, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.27733272314071655, "tokens": 1294663680, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.2898504242300985, "seq_len": 4096} {"wall": 11208.554, "eval_loss": 1.9847777634859085, "eval_ppl": 7.277430057525635, "step": 4400} {"wall": 11208.556, "step": 4400, "loss": 2.2825101912021637, "lm_loss": 2.2825101912021637, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.314958393573761, "tokens": 1297612800, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.284888881444931, "seq_len": 4096} {"wall": 11486.688, "step": 4410, "loss": 2.20896016061306, "lm_loss": 2.20896016061306, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2943080961704254, "tokens": 1300561920, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.340261648843686, "seq_len": 4096} {"wall": 11691.398, "step": 4420, "loss": 2.2930777023235955, "lm_loss": 2.2930777023235955, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3294093906879425, "tokens": 1303511040, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.267504049340884, "seq_len": 4096} {"wall": 11896.172, "step": 4430, "loss": 2.225581561525663, "lm_loss": 2.225581561525663, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.25968122482299805, "tokens": 1306460160, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.259104266266028, "seq_len": 4096} {"wall": 12100.744, "step": 4440, "loss": 2.323175832629204, "lm_loss": 2.323175832629204, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.4412066638469696, "tokens": 1309409280, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.298807110140721, "seq_len": 4096} {"wall": 12305.374, "step": 4450, "loss": 2.3767090837160745, "lm_loss": 2.3767090837160745, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.28138744831085205, "tokens": 1312358400, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.332330220937729, "seq_len": 4096} {"wall": 12510.074, "step": 4460, "loss": 2.2736753722031913, "lm_loss": 2.2736753722031913, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.331178218126297, "tokens": 1315307520, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.331635396182537, "seq_len": 4096} {"wall": 12714.722, "step": 4470, "loss": 2.315573344628016, "lm_loss": 2.315573344628016, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.26067423820495605, "tokens": 1318256640, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.27554689347744, "seq_len": 4096} {"wall": 12919.441, "step": 4480, "loss": 2.4669148673613868, "lm_loss": 2.4669148673613868, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3410515785217285, "tokens": 1321205760, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.33084904178977, "seq_len": 4096} {"wall": 13124.065, "step": 4490, "loss": 2.297509794433912, "lm_loss": 2.297509794433912, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2706671953201294, "tokens": 1324154880, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.279651053249836, "seq_len": 4096} {"wall": 13331.054, "eval_loss": 1.9709259867668152, "eval_ppl": 7.1773200035095215, "step": 4500} {"wall": 13331.056, "step": 4500, "loss": 2.27086632947127, "lm_loss": 2.27086632947127, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.24234984815120697, "tokens": 1327104000, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.243646354724963, "seq_len": 4096} {"wall": 13617.154, "step": 4510, "loss": 2.131101412077745, "lm_loss": 2.131101412077745, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3005063533782959, "tokens": 1330053120, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.27009051690499, "seq_len": 4096} {"wall": 13821.866, "step": 4520, "loss": 2.4902512580156326, "lm_loss": 2.4902512580156326, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3612169027328491, "tokens": 1333002240, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.3045506204168005, "seq_len": 4096} {"wall": 14026.505, "step": 4530, "loss": 2.316556672255198, "lm_loss": 2.316556672255198, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.397555947303772, "tokens": 1335951360, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.267861818273862, "seq_len": 4096} {"wall": 14231.239, "step": 4540, "loss": 2.314505939682325, "lm_loss": 2.314505939682325, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.29728788137435913, "tokens": 1338900480, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.2996699551741284, "seq_len": 4096} {"wall": 14435.961, "step": 4550, "loss": 2.2151933958133063, "lm_loss": 2.2151933958133063, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3291861414909363, "tokens": 1341849600, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.283464442193508, "seq_len": 4096} {"wall": 14640.9, "step": 4560, "loss": 2.362608184417089, "lm_loss": 2.362608184417089, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.24498479068279266, "tokens": 1344798720, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.3126017461220423, "seq_len": 4096} {"wall": 14845.838, "step": 4570, "loss": 2.347856988509496, "lm_loss": 2.347856988509496, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.29637953639030457, "tokens": 1347747840, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.292075352370739, "seq_len": 4096} {"wall": 15050.565, "step": 4580, "loss": 2.1808175841967263, "lm_loss": 2.1808175841967263, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3319665491580963, "tokens": 1350696960, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.2439490576585133, "seq_len": 4096} {"wall": 15255.186, "step": 4590, "loss": 2.3909295201301575, "lm_loss": 2.3909295201301575, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2726655602455139, "tokens": 1353646080, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.3212194979190826, "seq_len": 4096} {"wall": 15462.148, "eval_loss": 1.9651268869638443, "eval_ppl": 7.135818004608154, "step": 4600} {"wall": 15462.149, "step": 4600, "loss": 2.208125332991282, "lm_loss": 2.208125332991282, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.31542259454727173, "tokens": 1356595200, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.2941080758968986, "seq_len": 4096} {"wall": 15744.799, "step": 4610, "loss": 2.270738889773687, "lm_loss": 2.270738889773687, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3585429787635803, "tokens": 1359544320, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.2912612405916053, "seq_len": 4096} {"wall": 15949.351, "step": 4620, "loss": 2.3177720655997596, "lm_loss": 2.3177720655997596, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.25746476650238037, "tokens": 1362493440, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.2694014499584836, "seq_len": 4096} {"wall": 16154.15, "step": 4630, "loss": 2.3556737502415976, "lm_loss": 2.3556737502415976, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.32054707407951355, "tokens": 1365442560, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.2991502021749817, "seq_len": 4096} {"wall": 16359.368, "step": 4640, "loss": 2.2462972899278006, "lm_loss": 2.2462972899278006, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3473624587059021, "tokens": 1368391680, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.2425760703782243, "seq_len": 4096} {"wall": 16563.819, "step": 4650, "loss": 2.129750445485115, "lm_loss": 2.129750445485115, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.26345112919807434, "tokens": 1371340800, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.2798673431078593, "seq_len": 4096} {"wall": 16768.686, "step": 4660, "loss": 2.376389756798744, "lm_loss": 2.376389756798744, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.27532312273979187, "tokens": 1374289920, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.318821482360363, "seq_len": 4096} {"wall": 16973.894, "step": 4670, "loss": 2.2938588857650757, "lm_loss": 2.2938588857650757, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.362019419670105, "tokens": 1377239040, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.3184630244970323, "seq_len": 4096} {"wall": 17178.449, "step": 4680, "loss": 2.2531345188617706, "lm_loss": 2.2531345188617706, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.5082509517669678, "tokens": 1380188160, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.3117672125498454, "seq_len": 4096} {"wall": 17383.267, "step": 4690, "loss": 2.3535376290480294, "lm_loss": 2.3535376290480294, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2700042128562927, "tokens": 1383137280, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.307564248144627, "seq_len": 4096} {"wall": 17590.789, "eval_loss": 1.9631298631429672, "eval_ppl": 7.12158203125, "step": 4700} {"wall": 17590.791, "step": 4700, "loss": 2.3643401712179184, "lm_loss": 2.3643401712179184, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.303374707698822, "tokens": 1386086400, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.2994798600673674, "seq_len": 4096} {"wall": 17869.854, "step": 4710, "loss": 2.2511280874411264, "lm_loss": 2.2511280874411264, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3186286687850952, "tokens": 1389035520, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.213846592356761, "seq_len": 4096} {"wall": 18075.124, "step": 4720, "loss": 2.3231609016656876, "lm_loss": 2.3231609016656876, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2936219871044159, "tokens": 1391984640, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.2720833870271844, "seq_len": 4096} {"wall": 18279.594, "step": 4730, "loss": 2.3596453815698624, "lm_loss": 2.3596453815698624, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.37524718046188354, "tokens": 1394933760, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.2914987126986186, "seq_len": 4096} {"wall": 18484.816, "step": 4740, "loss": 2.062475269039472, "lm_loss": 2.062475269039472, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.31554219126701355, "tokens": 1397882880, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.27299583380421, "seq_len": 4096} {"wall": 18689.69, "step": 4750, "loss": 2.3163120796283088, "lm_loss": 2.3163120796283088, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3055061101913452, "tokens": 1400832000, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.307262877126535, "seq_len": 4096} {"wall": 18894.414, "step": 4760, "loss": 2.4468911786874137, "lm_loss": 2.4468911786874137, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.35434451699256897, "tokens": 1403781120, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.2925923496484755, "seq_len": 4096} {"wall": 19099.368, "step": 4770, "loss": 2.366644556323687, "lm_loss": 2.366644556323687, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.39857685565948486, "tokens": 1406730240, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.297435703873634, "seq_len": 4096} {"wall": 19303.936, "step": 4780, "loss": 2.0470461696386337, "lm_loss": 2.0470461696386337, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3311469256877899, "tokens": 1409679360, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.281114610284567, "seq_len": 4096} {"wall": 19508.961, "step": 4790, "loss": 2.271361584464709, "lm_loss": 2.271361584464709, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3281214237213135, "tokens": 1412628480, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.2756183676421644, "seq_len": 4096} {"wall": 19716.378, "eval_loss": 1.9629793018102646, "eval_ppl": 7.120509624481201, "step": 4800} {"wall": 19716.38, "step": 4800, "loss": 2.2418206930160522, "lm_loss": 2.2418206930160522, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3446308672428131, "tokens": 1415577600, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.298887424170971, "seq_len": 4096} {"wall": 20031.639, "step": 4810, "loss": 2.3236943582693734, "lm_loss": 2.3236943582693734, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2800176441669464, "tokens": 1418526720, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.286710145572821, "seq_len": 4096} {"wall": 20236.64, "step": 4820, "loss": 2.271898572643598, "lm_loss": 2.271898572643598, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.32931801676750183, "tokens": 1421475840, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.2481924961010615, "seq_len": 4096} {"wall": 20441.586, "step": 4830, "loss": 2.313989594578743, "lm_loss": 2.313989594578743, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3280819058418274, "tokens": 1424424960, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.270102508366108, "seq_len": 4096} {"wall": 20646.523, "step": 4840, "loss": 2.2701780796051025, "lm_loss": 2.2701780796051025, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3570149540901184, "tokens": 1427374080, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.2834046572446822, "seq_len": 4096} {"wall": 20851.42, "step": 4850, "loss": 2.3548500587542853, "lm_loss": 2.3548500587542853, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.32407310605049133, "tokens": 1430323200, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.2669774770736693, "seq_len": 4096} {"wall": 21056.204, "step": 4860, "loss": 2.2389998684326806, "lm_loss": 2.2389998684326806, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.30002567172050476, "tokens": 1433272320, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.2795269772410394, "seq_len": 4096} {"wall": 21260.931, "step": 4870, "loss": 2.231810763478279, "lm_loss": 2.231810763478279, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.30261144042015076, "tokens": 1436221440, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.261749358475208, "seq_len": 4096} {"wall": 21465.637, "step": 4880, "loss": 2.261833061774572, "lm_loss": 2.261833061774572, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3508816063404083, "tokens": 1439170560, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.269101319462061, "seq_len": 4096} {"wall": 21670.501, "step": 4890, "loss": 2.2455941538016, "lm_loss": 2.2455941538016, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.28434398770332336, "tokens": 1442119680, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.244691454867522, "seq_len": 4096} {"wall": 21877.869, "eval_loss": 1.9590819329023361, "eval_ppl": 7.0928120613098145, "step": 4900} {"wall": 21877.871, "step": 4900, "loss": 2.2640264431635537, "lm_loss": 2.2640264431635537, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3638264536857605, "tokens": 1445068800, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.2847884585460028, "seq_len": 4096} {"wall": 22162.691, "step": 4910, "loss": 2.3596068074305854, "lm_loss": 2.3596068074305854, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3998192548751831, "tokens": 1448017920, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.267326138416926, "seq_len": 4096} {"wall": 22367.275, "step": 4920, "loss": 2.327595353126526, "lm_loss": 2.327595353126526, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3187769949436188, "tokens": 1450967040, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.331216502189636, "seq_len": 4096} {"wall": 22572.078, "step": 4930, "loss": 2.3061987658341727, "lm_loss": 2.3061987658341727, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.28579220175743103, "tokens": 1453916160, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.2767448564370474, "seq_len": 4096} {"wall": 22776.845, "step": 4940, "loss": 2.341548591852188, "lm_loss": 2.341548591852188, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.29190605878829956, "tokens": 1456865280, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.2653190826376277, "seq_len": 4096} {"wall": 22981.532, "step": 4950, "loss": 2.3177539159854255, "lm_loss": 2.3177539159854255, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3562741279602051, "tokens": 1459814400, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.2571915959318476, "seq_len": 4096} {"wall": 23186.126, "step": 4960, "loss": 2.2187898655732474, "lm_loss": 2.2187898655732474, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3415212333202362, "tokens": 1462763520, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.290767337878545, "seq_len": 4096} {"wall": 23391.32, "step": 4970, "loss": 2.3006073037783303, "lm_loss": 2.3006073037783303, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3637368679046631, "tokens": 1465712640, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.3109663585821787, "seq_len": 4096} {"wall": 23595.69, "step": 4980, "loss": 2.304606889684995, "lm_loss": 2.304606889684995, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2722596228122711, "tokens": 1468661760, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.3073580980300905, "seq_len": 4096} {"wall": 23800.619, "step": 4990, "loss": 2.2544881999492645, "lm_loss": 2.2544881999492645, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.28143420815467834, "tokens": 1471610880, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.255283458530903, "seq_len": 4096} {"wall": 24008.125, "eval_loss": 1.9555416703224182, "eval_ppl": 7.067746162414551, "step": 5000} {"wall": 24008.128, "step": 5000, "loss": 2.2230978136261306, "lm_loss": 2.2230978136261306, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3508321940898895, "tokens": 1474560000, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.2531963412960367, "seq_len": 4096} {"wall": 24294.972, "step": 5010, "loss": 2.3102116187413535, "lm_loss": 2.3102116187413535, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.29947978258132935, "tokens": 1477509120, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.2887885361909865, "seq_len": 4096} {"wall": 24499.839, "step": 5020, "loss": 2.278143306573232, "lm_loss": 2.278143306573232, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.393230676651001, "tokens": 1480458240, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.2850762968262037, "seq_len": 4096} {"wall": 24704.584, "step": 5030, "loss": 2.315869520107905, "lm_loss": 2.315869520107905, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.32923322916030884, "tokens": 1483407360, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.310396496454875, "seq_len": 4096} {"wall": 24909.436, "step": 5040, "loss": 2.3162387162446976, "lm_loss": 2.3162387162446976, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2948680818080902, "tokens": 1486356480, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.2621110320091247, "seq_len": 4096} {"wall": 25114.313, "step": 5050, "loss": 2.3244013537963233, "lm_loss": 2.3244013537963233, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2922755181789398, "tokens": 1489305600, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.301947422325611, "seq_len": 4096} {"wall": 25319.244, "step": 5060, "loss": 2.1470124572515488, "lm_loss": 2.1470124572515488, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.36055758595466614, "tokens": 1492254720, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.2576745520035426, "seq_len": 4096} {"wall": 25524.084, "step": 5070, "loss": 2.246529991428057, "lm_loss": 2.246529991428057, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.31829559803009033, "tokens": 1495203840, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.2980069269736605, "seq_len": 4096} {"wall": 25728.726, "step": 5080, "loss": 2.243485748767853, "lm_loss": 2.243485748767853, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3129276633262634, "tokens": 1498152960, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.296546672781308, "seq_len": 4096} {"wall": 25933.96, "step": 5090, "loss": 2.2671431551376977, "lm_loss": 2.2671431551376977, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.36652693152427673, "tokens": 1501102080, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.2564182594418525, "seq_len": 4096} {"wall": 26141.152, "eval_loss": 1.949966922402382, "eval_ppl": 7.028454780578613, "step": 5100} {"wall": 26141.154, "step": 5100, "loss": 2.0832044035196304, "lm_loss": 2.0832044035196304, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2999263107776642, "tokens": 1504051200, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.254081284503142, "seq_len": 4096} {"wall": 26423.414, "step": 5110, "loss": 2.3617481688658395, "lm_loss": 2.3617481688658395, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.30344003438949585, "tokens": 1507000320, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.3075728476047517, "seq_len": 4096} {"wall": 26628.927, "step": 5120, "loss": 2.12732886771361, "lm_loss": 2.12732886771361, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.36513012647628784, "tokens": 1509949440, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.257479375849168, "seq_len": 4096} {"wall": 26833.863, "step": 5130, "loss": 2.229914272824923, "lm_loss": 2.229914272824923, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3160487413406372, "tokens": 1512898560, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.2695607746640842, "seq_len": 4096} {"wall": 27038.919, "step": 5140, "loss": 2.166354392965635, "lm_loss": 2.166354392965635, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.32388269901275635, "tokens": 1515847680, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.254609042654435, "seq_len": 4096} {"wall": 27243.809, "step": 5150, "loss": 2.177501544356346, "lm_loss": 2.177501544356346, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.32071152329444885, "tokens": 1518796800, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.27394925703605, "seq_len": 4096} {"wall": 27448.581, "step": 5160, "loss": 2.341016471385956, "lm_loss": 2.341016471385956, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2898924946784973, "tokens": 1521745920, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.29994565397501, "seq_len": 4096} {"wall": 27653.939, "step": 5170, "loss": 2.3178790559371314, "lm_loss": 2.3178790559371314, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2937655746936798, "tokens": 1524695040, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.3041624734799067, "seq_len": 4096} {"wall": 27858.501, "step": 5180, "loss": 2.2113875498374305, "lm_loss": 2.2113875498374305, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.30921125411987305, "tokens": 1527644160, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.2838189755876863, "seq_len": 4096} {"wall": 28063.606, "step": 5190, "loss": 2.282326782743136, "lm_loss": 2.282326782743136, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2910389304161072, "tokens": 1530593280, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.2473667845129968, "seq_len": 4096} {"wall": 28270.991, "eval_loss": 1.956422746181488, "eval_ppl": 7.073976993560791, "step": 5200} {"wall": 28270.993, "step": 5200, "loss": 2.388671045502027, "lm_loss": 2.388671045502027, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.34994256496429443, "tokens": 1533542400, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.252350437392791, "seq_len": 4096} {"wall": 28557.454, "step": 5210, "loss": 2.2846509416898093, "lm_loss": 2.2846509416898093, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.32474949955940247, "tokens": 1536491520, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.2969700212279958, "seq_len": 4096} {"wall": 28762.384, "step": 5220, "loss": 2.4461072186628976, "lm_loss": 2.4461072186628976, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.30273956060409546, "tokens": 1539440640, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.3077606678009035, "seq_len": 4096} {"wall": 28967.467, "step": 5230, "loss": 2.2413840343554816, "lm_loss": 2.2413840343554816, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3045918643474579, "tokens": 1542389760, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.309528946876526, "seq_len": 4096} {"wall": 29172.779, "step": 5240, "loss": 2.1372035294771194, "lm_loss": 2.1372035294771194, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2837091386318207, "tokens": 1545338880, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.2607031057278317, "seq_len": 4096} {"wall": 29377.526, "step": 5250, "loss": 2.2270501802364984, "lm_loss": 2.2270501802364984, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3194490671157837, "tokens": 1548288000, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.2842275063196817, "seq_len": 4096} {"wall": 29582.66, "step": 5260, "loss": 2.273283218344053, "lm_loss": 2.273283218344053, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2855227589607239, "tokens": 1551237120, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.2740125000476836, "seq_len": 4096} {"wall": 29788.048, "step": 5270, "loss": 2.239202246069908, "lm_loss": 2.239202246069908, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.28823283314704895, "tokens": 1554186240, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.2675339353581268, "seq_len": 4096} {"wall": 29992.631, "step": 5280, "loss": 2.276614248752594, "lm_loss": 2.276614248752594, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.29925844073295593, "tokens": 1557135360, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.2555651957790053, "seq_len": 4096} {"wall": 30197.928, "step": 5290, "loss": 2.314302305380503, "lm_loss": 2.314302305380503, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2631337642669678, "tokens": 1560084480, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.2850370720028876, "seq_len": 4096} {"wall": 30405.405, "eval_loss": 1.9428090900182724, "eval_ppl": 6.978326320648193, "step": 5300} {"wall": 30405.407, "step": 5300, "loss": 2.2413981705904007, "lm_loss": 2.2413981705904007, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.294560968875885, "tokens": 1563033600, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.2678203428785007, "seq_len": 4096} {"wall": 155.941, "eval_loss": 1.9428090900182724, "eval_ppl": 6.978326320648193, "step": 5300} {"wall": 882.661, "step": 5310, "loss": 2.372272198398908, "lm_loss": 2.372272198398908, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.34610500931739807, "tokens": 1565982720, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.302734375, "loss_smoothed": 2.273585519194603, "seq_len": 4096} {"wall": 1087.671, "step": 5320, "loss": 2.247748772303263, "lm_loss": 2.247748772303263, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3196762204170227, "tokens": 1568931840, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.302734375, "loss_smoothed": 2.296545861164729, "seq_len": 4096} {"wall": 1292.683, "step": 5330, "loss": 2.253061900536219, "lm_loss": 2.253061900536219, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.41952165961265564, "tokens": 1571880960, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.302734375, "loss_smoothed": 2.275045096625884, "seq_len": 4096} {"wall": 96.065, "eval_loss": 1.9428210854530334, "eval_ppl": 6.978409290313721, "step": 5300} {"wall": 419.215, "step": 5310, "loss": 2.372256984313329, "lm_loss": 2.372256984313329, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.33244484663009644, "tokens": 1565982720, "adamw_lr": 0.0004, "muon/grad_norm": 0.2637771666049957, "adam/grad_norm": 0.20229938626289368, "muon/param_norm": 1969.43212890625, "muon/update_ratio": 0.010183842852711678, "muon/update_ratio_max": 0.0794452428817749, "adam/param_norm": 695.318115234375, "adam/update_ratio": 0.0011236814316362143, "adam/update_ratio_max": 0.003251532092690468, "ademamix/cos_fast_slow": 0.5302163362503052, "ademamix/slow_fast_norm_ratio": 0.16601823270320892, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.2735201850533486, "seq_len": 4096} {"wall": 623.916, "step": 5320, "loss": 2.24907253185908, "lm_loss": 2.24907253185908, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3260827660560608, "tokens": 1568931840, "adamw_lr": 0.0004, "muon/grad_norm": 0.2576025724411011, "adam/grad_norm": 0.1995786428451538, "muon/param_norm": 1969.3326416015625, "muon/update_ratio": 0.010252803564071655, "muon/update_ratio_max": 0.08413223177194595, "adam/param_norm": 695.3658447265625, "adam/update_ratio": 0.0011273027630522847, "adam/update_ratio_max": 0.003241038415580988, "ademamix/cos_fast_slow": 0.5175214409828186, "ademamix/slow_fast_norm_ratio": 0.17250385880470276, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.29597794637084, "seq_len": 4096} {"wall": 828.573, "step": 5330, "loss": 2.2508311917384467, "lm_loss": 2.2508311917384467, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.39744457602500916, "tokens": 1571880960, "adamw_lr": 0.0004, "muon/grad_norm": 0.2778123617172241, "adam/grad_norm": 0.2844775319099426, "muon/param_norm": 1969.3134765625, "muon/update_ratio": 0.009937628172338009, "muon/update_ratio_max": 0.08225643634796143, "adam/param_norm": 695.4111328125, "adam/update_ratio": 0.0011162377195432782, "adam/update_ratio_max": 0.003137043211609125, "ademamix/cos_fast_slow": 0.5085721611976624, "ademamix/slow_fast_norm_ratio": 0.14075295627117157, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.2747400475045043, "seq_len": 4096} {"wall": 1033.001, "step": 5340, "loss": 2.4671517511208854, "lm_loss": 2.4671517511208854, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.4533902108669281, "tokens": 1574830080, "adamw_lr": 0.0004, "muon/grad_norm": 0.33291760087013245, "adam/grad_norm": 0.3073519170284271, "muon/param_norm": 1969.1517333984375, "muon/update_ratio": 0.010187295265495777, "muon/update_ratio_max": 0.08188702911138535, "adam/param_norm": 695.45751953125, "adam/update_ratio": 0.0011266371002420783, "adam/update_ratio_max": 0.0031165776308625937, "ademamix/cos_fast_slow": 0.21507175266742706, "ademamix/slow_fast_norm_ratio": 0.12876468896865845, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.269735698898633, "seq_len": 4096} {"wall": 1237.436, "step": 5350, "loss": 2.156657879551252, "lm_loss": 2.156657879551252, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2977261245250702, "tokens": 1577779200, "adamw_lr": 0.0004, "muon/grad_norm": 0.22413595020771027, "adam/grad_norm": 0.19573231041431427, "muon/param_norm": 1969.4041748046875, "muon/update_ratio": 0.010293408297002316, "muon/update_ratio_max": 0.08147377520799637, "adam/param_norm": 695.502685546875, "adam/update_ratio": 0.001117289299145341, "adam/update_ratio_max": 0.003023200435563922, "ademamix/cos_fast_slow": 0.4997387230396271, "ademamix/slow_fast_norm_ratio": 0.1748049557209015, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.2636762291193007, "seq_len": 4096} {"wall": 1441.556, "step": 5360, "loss": 2.3303068727254868, "lm_loss": 2.3303068727254868, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.34511205554008484, "tokens": 1580728320, "adamw_lr": 0.0004, "muon/grad_norm": 0.28470152616500854, "adam/grad_norm": 0.19488368928432465, "muon/param_norm": 1969.7391357421875, "muon/update_ratio": 0.010251199826598167, "muon/update_ratio_max": 0.08206494152545929, "adam/param_norm": 695.5497436523438, "adam/update_ratio": 0.0011140058049932122, "adam/update_ratio_max": 0.0031563383527100086, "ademamix/cos_fast_slow": 0.24248993396759033, "ademamix/slow_fast_norm_ratio": 0.16627462208271027, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.288683488468329, "seq_len": 4096} {"wall": 1646.033, "step": 5370, "loss": 2.256022463242213, "lm_loss": 2.256022463242213, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.30304381251335144, "tokens": 1583677440, "adamw_lr": 0.0004, "muon/grad_norm": 0.22826938331127167, "adam/grad_norm": 0.19947406649589539, "muon/param_norm": 1969.9288330078125, "muon/update_ratio": 0.010297025553882122, "muon/update_ratio_max": 0.08472290635108948, "adam/param_norm": 695.5968017578125, "adam/update_ratio": 0.0011286120861768723, "adam/update_ratio_max": 0.0030807999428361654, "ademamix/cos_fast_slow": 0.26589149236679077, "ademamix/slow_fast_norm_ratio": 0.16994644701480865, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.2921296377976734, "seq_len": 4096} {"wall": 1850.577, "step": 5380, "loss": 2.1272258361180625, "lm_loss": 2.1272258361180625, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2753531038761139, "tokens": 1586626560, "adamw_lr": 0.0004, "muon/grad_norm": 0.2226887196302414, "adam/grad_norm": 0.16136009991168976, "muon/param_norm": 1970.252685546875, "muon/update_ratio": 0.010313499718904495, "muon/update_ratio_max": 0.0809883251786232, "adam/param_norm": 695.6432495117188, "adam/update_ratio": 0.0011120670242235065, "adam/update_ratio_max": 0.0032242226880043745, "ademamix/cos_fast_slow": 0.4056258499622345, "ademamix/slow_fast_norm_ratio": 0.1765325516462326, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.281516209244728, "seq_len": 4096} {"wall": 2054.978, "step": 5390, "loss": 2.2249905094504356, "lm_loss": 2.2249905094504356, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3003338575363159, "tokens": 1589575680, "adamw_lr": 0.0004, "muon/grad_norm": 0.24034728109836578, "adam/grad_norm": 0.18040414154529572, "muon/param_norm": 1970.5440673828125, "muon/update_ratio": 0.010236591100692749, "muon/update_ratio_max": 0.08359873294830322, "adam/param_norm": 695.689697265625, "adam/update_ratio": 0.0011117731919512153, "adam/update_ratio_max": 0.0030448390170931816, "ademamix/cos_fast_slow": 0.22462746500968933, "ademamix/slow_fast_norm_ratio": 0.17166540026664734, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.239544476817051, "seq_len": 4096} {"wall": 2261.834, "eval_loss": 1.9499294012784958, "eval_ppl": 7.028191089630127, "step": 5400} {"wall": 2261.835, "step": 5400, "loss": 2.313798169294993, "lm_loss": 2.313798169294993, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.34091758728027344, "tokens": 1592524800, "adamw_lr": 0.0004, "muon/grad_norm": 0.2563403248786926, "adam/grad_norm": 0.22499938309192657, "muon/param_norm": 1970.4403076171875, "muon/update_ratio": 0.010240163654088974, "muon/update_ratio_max": 0.08607205748558044, "adam/param_norm": 695.7344360351562, "adam/update_ratio": 0.001123092952184379, "adam/update_ratio_max": 0.00309833069331944, "ademamix/cos_fast_slow": 0.5816023945808411, "ademamix/slow_fast_norm_ratio": 0.16103221476078033, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.323751750588417, "seq_len": 4096} {"wall": 2481.713, "step": 5410, "loss": 2.3017111122608185, "lm_loss": 2.3017111122608185, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3058151304721832, "tokens": 1595473920, "adamw_lr": 0.0004, "muon/grad_norm": 0.2164243757724762, "adam/grad_norm": 0.2163732498884201, "muon/param_norm": 1970.3133544921875, "muon/update_ratio": 0.010299636982381344, "muon/update_ratio_max": 0.08278069645166397, "adam/param_norm": 695.7794799804688, "adam/update_ratio": 0.001115086255595088, "adam/update_ratio_max": 0.003048037877306342, "ademamix/cos_fast_slow": 0.47066619992256165, "ademamix/slow_fast_norm_ratio": 0.16152113676071167, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.439453125, "loss_smoothed": 2.3132891789078713, "seq_len": 4096} {"wall": 2686.113, "step": 5420, "loss": 2.2985716263453164, "lm_loss": 2.2985716263453164, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3214900493621826, "tokens": 1598423040, "adamw_lr": 0.0004, "muon/grad_norm": 0.23579144477844238, "adam/grad_norm": 0.21886169910430908, "muon/param_norm": 1969.97998046875, "muon/update_ratio": 0.01010146178305149, "muon/update_ratio_max": 0.08441507071256638, "adam/param_norm": 695.8265991210938, "adam/update_ratio": 0.0011130376951768994, "adam/update_ratio_max": 0.003278829623013735, "ademamix/cos_fast_slow": 0.20509637892246246, "ademamix/slow_fast_norm_ratio": 0.15671826899051666, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.439453125, "loss_smoothed": 2.24413092409571, "seq_len": 4096} {"wall": 2890.512, "step": 5430, "loss": 2.19836962223053, "lm_loss": 2.19836962223053, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2968391180038452, "tokens": 1601372160, "adamw_lr": 0.0004, "muon/grad_norm": 0.20866931974887848, "adam/grad_norm": 0.21081401407718658, "muon/param_norm": 1969.837646484375, "muon/update_ratio": 0.010278910398483276, "muon/update_ratio_max": 0.08437313884496689, "adam/param_norm": 695.8718872070312, "adam/update_ratio": 0.0011135345557704568, "adam/update_ratio_max": 0.003227286273613572, "ademamix/cos_fast_slow": 0.4819231927394867, "ademamix/slow_fast_norm_ratio": 0.16564220190048218, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.439453125, "loss_smoothed": 2.297496990362803, "seq_len": 4096} {"wall": 3094.94, "step": 5440, "loss": 2.233083496491114, "lm_loss": 2.233083496491114, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.29637816548347473, "tokens": 1604321280, "adamw_lr": 0.0004, "muon/grad_norm": 0.21287930011749268, "adam/grad_norm": 0.20592518150806427, "muon/param_norm": 1969.7340087890625, "muon/update_ratio": 0.010328234173357487, "muon/update_ratio_max": 0.0820070430636406, "adam/param_norm": 695.9191284179688, "adam/update_ratio": 0.001124063041061163, "adam/update_ratio_max": 0.003071852959692478, "ademamix/cos_fast_slow": 0.420539915561676, "ademamix/slow_fast_norm_ratio": 0.15207350254058838, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.439453125, "loss_smoothed": 2.2497033988436064, "seq_len": 4096} {"wall": 3299.362, "step": 5450, "loss": 2.1736045529445014, "lm_loss": 2.1736045529445014, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3406440019607544, "tokens": 1607270400, "adamw_lr": 0.0004, "muon/grad_norm": 0.244330495595932, "adam/grad_norm": 0.23780222237110138, "muon/param_norm": 1969.43994140625, "muon/update_ratio": 0.010283703915774822, "muon/update_ratio_max": 0.07943857461214066, "adam/param_norm": 695.9644775390625, "adam/update_ratio": 0.0011210019001737237, "adam/update_ratio_max": 0.0031612308230251074, "ademamix/cos_fast_slow": 0.36596477031707764, "ademamix/slow_fast_norm_ratio": 0.15633593499660492, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.439453125, "loss_smoothed": 2.238020167251428, "seq_len": 4096} {"wall": 3503.785, "step": 5460, "loss": 2.2086146424214044, "lm_loss": 2.2086146424214044, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3154474198818207, "tokens": 1610219520, "adamw_lr": 0.0004, "muon/grad_norm": 0.2412889450788498, "adam/grad_norm": 0.20344871282577515, "muon/param_norm": 1969.3514404296875, "muon/update_ratio": 0.010310153476893902, "muon/update_ratio_max": 0.08074367046356201, "adam/param_norm": 696.0106201171875, "adam/update_ratio": 0.0011206435738131404, "adam/update_ratio_max": 0.003015088615939021, "ademamix/cos_fast_slow": 0.4094582200050354, "ademamix/slow_fast_norm_ratio": 0.16857026517391205, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.439453125, "loss_smoothed": 2.2360974689324697, "seq_len": 4096} {"wall": 3708.193, "step": 5470, "loss": 2.4144091606140137, "lm_loss": 2.4144091606140137, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.33396509289741516, "tokens": 1613168640, "adamw_lr": 0.0004, "muon/grad_norm": 0.25998619198799133, "adam/grad_norm": 0.20920903980731964, "muon/param_norm": 1968.935302734375, "muon/update_ratio": 0.010142957791686058, "muon/update_ratio_max": 0.08224966377019882, "adam/param_norm": 696.0562133789062, "adam/update_ratio": 0.0011200261069461703, "adam/update_ratio_max": 0.003173190401867032, "ademamix/cos_fast_slow": 0.5822194218635559, "ademamix/slow_fast_norm_ratio": 0.15061236917972565, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.439453125, "loss_smoothed": 2.2638705357909203, "seq_len": 4096} {"wall": 3912.648, "step": 5480, "loss": 2.2245878676573434, "lm_loss": 2.2245878676573434, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.27545103430747986, "tokens": 1616117760, "adamw_lr": 0.0004, "muon/grad_norm": 0.22008077800273895, "adam/grad_norm": 0.16516700387001038, "muon/param_norm": 1968.0556640625, "muon/update_ratio": 0.010250302962958813, "muon/update_ratio_max": 0.07769356667995453, "adam/param_norm": 696.1036987304688, "adam/update_ratio": 0.0011281163897365332, "adam/update_ratio_max": 0.0031258794479072094, "ademamix/cos_fast_slow": 0.4860627353191376, "ademamix/slow_fast_norm_ratio": 0.16100892424583435, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.439453125, "loss_smoothed": 2.262600672245026, "seq_len": 4096} {"wall": 4117.039, "step": 5490, "loss": 2.155974934498469, "lm_loss": 2.155974934498469, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.271736741065979, "tokens": 1619066880, "adamw_lr": 0.0004, "muon/grad_norm": 0.1989639699459076, "adam/grad_norm": 0.1852157711982727, "muon/param_norm": 1967.8172607421875, "muon/update_ratio": 0.01029290072619915, "muon/update_ratio_max": 0.07930497080087662, "adam/param_norm": 696.1507568359375, "adam/update_ratio": 0.0011331317946314812, "adam/update_ratio_max": 0.0030579555314034224, "ademamix/cos_fast_slow": 0.5186712145805359, "ademamix/slow_fast_norm_ratio": 0.16891370713710785, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.439453125, "loss_smoothed": 2.316227321823438, "seq_len": 4096} {"wall": 4323.807, "eval_loss": 1.944024071097374, "eval_ppl": 6.986810207366943, "step": 5500} {"wall": 4323.809, "step": 5500, "loss": 2.236482545733452, "lm_loss": 2.236482545733452, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2972224950790405, "tokens": 1622016000, "adamw_lr": 0.0004, "muon/grad_norm": 0.24585746228694916, "adam/grad_norm": 0.16701139509677887, "muon/param_norm": 1968.0477294921875, "muon/update_ratio": 0.01030783075839281, "muon/update_ratio_max": 0.08224654942750931, "adam/param_norm": 696.1966552734375, "adam/update_ratio": 0.0011157947592437267, "adam/update_ratio_max": 0.0030607853550463915, "ademamix/cos_fast_slow": 0.524472177028656, "ademamix/slow_fast_norm_ratio": 0.16906718909740448, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.439453125, "loss_smoothed": 2.228307566046715, "seq_len": 4096} {"wall": 4609.929, "step": 5510, "loss": 2.2719117403030396, "lm_loss": 2.2719117403030396, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3583320677280426, "tokens": 1624965120, "adamw_lr": 0.0004, "muon/grad_norm": 0.23929984867572784, "adam/grad_norm": 0.26691311597824097, "muon/param_norm": 1967.9844970703125, "muon/update_ratio": 0.010315482504665852, "muon/update_ratio_max": 0.0805426687002182, "adam/param_norm": 696.2437133789062, "adam/update_ratio": 0.0011278439778834581, "adam/update_ratio_max": 0.003072819672524929, "ademamix/cos_fast_slow": 0.21214412152767181, "ademamix/slow_fast_norm_ratio": 0.1409134566783905, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 88.033203125, "loss_smoothed": 2.289713143805663, "seq_len": 4096} {"wall": 4814.556, "step": 5520, "loss": 2.1691219558318457, "lm_loss": 2.1691219558318457, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2580789625644684, "tokens": 1627914240, "adamw_lr": 0.0004, "muon/grad_norm": 0.20351186394691467, "adam/grad_norm": 0.15908032655715942, "muon/param_norm": 1967.948974609375, "muon/update_ratio": 0.01035364530980587, "muon/update_ratio_max": 0.08211566507816315, "adam/param_norm": 696.291259765625, "adam/update_ratio": 0.001125409617088735, "adam/update_ratio_max": 0.0030266568064689636, "ademamix/cos_fast_slow": 0.33989080786705017, "ademamix/slow_fast_norm_ratio": 0.15994413197040558, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 88.033203125, "loss_smoothed": 2.2573828821380935, "seq_len": 4096} {"wall": 5019.379, "step": 5530, "loss": 2.1835455944140754, "lm_loss": 2.1835455944140754, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.4467841386795044, "tokens": 1630863360, "adamw_lr": 0.0004, "muon/grad_norm": 0.3345143795013428, "adam/grad_norm": 0.29561468958854675, "muon/param_norm": 1967.7838134765625, "muon/update_ratio": 0.010045343078672886, "muon/update_ratio_max": 0.0796356052160263, "adam/param_norm": 696.3392944335938, "adam/update_ratio": 0.0011109107872471213, "adam/update_ratio_max": 0.0033101020380854607, "ademamix/cos_fast_slow": 0.19613361358642578, "ademamix/slow_fast_norm_ratio": 0.12594591081142426, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 88.033203125, "loss_smoothed": 2.2436126805841923, "seq_len": 4096} {"wall": 5223.971, "step": 5540, "loss": 2.118728756904602, "lm_loss": 2.118728756904602, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.32327884435653687, "tokens": 1633812480, "adamw_lr": 0.0004, "muon/grad_norm": 0.23333264887332916, "adam/grad_norm": 0.2240203619003296, "muon/param_norm": 1967.572509765625, "muon/update_ratio": 0.010271619074046612, "muon/update_ratio_max": 0.08410954475402832, "adam/param_norm": 696.3853149414062, "adam/update_ratio": 0.0011120045091956854, "adam/update_ratio_max": 0.0030730979051440954, "ademamix/cos_fast_slow": 0.37238505482673645, "ademamix/slow_fast_norm_ratio": 0.1519046574831009, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 88.033203125, "loss_smoothed": 2.2221349264184633, "seq_len": 4096} {"wall": 5428.612, "step": 5550, "loss": 2.242256502310435, "lm_loss": 2.242256502310435, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.37069761753082275, "tokens": 1636761600, "adamw_lr": 0.0004, "muon/grad_norm": 0.3066508173942566, "adam/grad_norm": 0.20869450271129608, "muon/param_norm": 1967.57568359375, "muon/update_ratio": 0.010251633822917938, "muon/update_ratio_max": 0.0809985026717186, "adam/param_norm": 696.4319458007812, "adam/update_ratio": 0.0011276568984612823, "adam/update_ratio_max": 0.003021470969542861, "ademamix/cos_fast_slow": 0.24872992932796478, "ademamix/slow_fast_norm_ratio": 0.16354098916053772, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 88.033203125, "loss_smoothed": 2.2034117783109344, "seq_len": 4096} {"wall": 5633.556, "step": 5560, "loss": 2.2236773719390235, "lm_loss": 2.2236773719390235, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3451436460018158, "tokens": 1639710720, "adamw_lr": 0.0004, "muon/grad_norm": 0.24068135023117065, "adam/grad_norm": 0.24703209102153778, "muon/param_norm": 1967.3193359375, "muon/update_ratio": 0.010230221785604954, "muon/update_ratio_max": 0.0806552916765213, "adam/param_norm": 696.4782104492188, "adam/update_ratio": 0.0011164102470502257, "adam/update_ratio_max": 0.003117655636742711, "ademamix/cos_fast_slow": 0.513984739780426, "ademamix/slow_fast_norm_ratio": 0.14760161936283112, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 88.033203125, "loss_smoothed": 2.2703213155269624, "seq_len": 4096} {"wall": 5838.183, "step": 5570, "loss": 2.2457821468512216, "lm_loss": 2.2457821468512216, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3068672716617584, "tokens": 1642659840, "adamw_lr": 0.0004, "muon/grad_norm": 0.23383836448192596, "adam/grad_norm": 0.19898200035095215, "muon/param_norm": 1967.3223876953125, "muon/update_ratio": 0.010278234258294106, "muon/update_ratio_max": 0.07971265912055969, "adam/param_norm": 696.5258178710938, "adam/update_ratio": 0.0011112758656963706, "adam/update_ratio_max": 0.002898183884099126, "ademamix/cos_fast_slow": 0.44787582755088806, "ademamix/slow_fast_norm_ratio": 0.1670921891927719, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 88.033203125, "loss_smoothed": 2.2347231189409893, "seq_len": 4096} {"wall": 6042.836, "step": 5580, "loss": 2.2299736539522805, "lm_loss": 2.2299736539522805, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.362669438123703, "tokens": 1645608960, "adamw_lr": 0.0004, "muon/grad_norm": 0.29084286093711853, "adam/grad_norm": 0.2164289653301239, "muon/param_norm": 1967.374755859375, "muon/update_ratio": 0.01022836472839117, "muon/update_ratio_max": 0.07783735543489456, "adam/param_norm": 696.5724487304688, "adam/update_ratio": 0.0011133708758279681, "adam/update_ratio_max": 0.002911350456997752, "ademamix/cos_fast_slow": 0.2862420380115509, "ademamix/slow_fast_norm_ratio": 0.17190991342067719, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 88.033203125, "loss_smoothed": 2.2240090211232504, "seq_len": 4096} {"wall": 6247.466, "step": 5590, "loss": 2.208316912253698, "lm_loss": 2.208316912253698, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.29572051763534546, "tokens": 1648558080, "adamw_lr": 0.0004, "muon/grad_norm": 0.19887587428092957, "adam/grad_norm": 0.2191377431154251, "muon/param_norm": 1966.1524658203125, "muon/update_ratio": 0.01003174763172865, "muon/update_ratio_max": 0.0804748460650444, "adam/param_norm": 696.6178588867188, "adam/update_ratio": 0.0011151881190016866, "adam/update_ratio_max": 0.0032312849070876837, "ademamix/cos_fast_slow": 0.3758978545665741, "ademamix/slow_fast_norm_ratio": 0.1419163942337036, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 88.033203125, "loss_smoothed": 2.265436845769485, "seq_len": 4096} {"wall": 6454.528, "eval_loss": 1.944543719291687, "eval_ppl": 6.99044132232666, "step": 5600} {"wall": 6454.53, "step": 5600, "loss": 2.2460116147994995, "lm_loss": 2.2460116147994995, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.39517736434936523, "tokens": 1651507200, "adamw_lr": 0.0004, "muon/grad_norm": 0.27585792541503906, "adam/grad_norm": 0.2823046147823334, "muon/param_norm": 1965.02978515625, "muon/update_ratio": 0.010082019492983818, "muon/update_ratio_max": 0.08082237839698792, "adam/param_norm": 696.6642456054688, "adam/update_ratio": 0.0011183571768924594, "adam/update_ratio_max": 0.003012295812368393, "ademamix/cos_fast_slow": 0.6314868927001953, "ademamix/slow_fast_norm_ratio": 0.1442595273256302, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 88.033203125, "loss_smoothed": 2.267383317897717, "seq_len": 4096} {"wall": 6731.265, "step": 5610, "loss": 2.2537953505913415, "lm_loss": 2.2537953505913415, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.35469767451286316, "tokens": 1654456320, "adamw_lr": 0.0004, "muon/grad_norm": 0.22880101203918457, "adam/grad_norm": 0.270736426115036, "muon/param_norm": 1960.1488037109375, "muon/update_ratio": 0.008810391649603844, "muon/update_ratio_max": 0.08171475678682327, "adam/param_norm": 696.711181640625, "adam/update_ratio": 0.0011230793315917253, "adam/update_ratio_max": 0.002974295523017645, "ademamix/cos_fast_slow": 0.5742713212966919, "ademamix/slow_fast_norm_ratio": 0.12590175867080688, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.380859375, "loss_smoothed": 2.25365641862154, "seq_len": 4096} {"wall": 6936.067, "step": 5620, "loss": 2.2989485462506614, "lm_loss": 2.2989485462506614, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.30234500765800476, "tokens": 1657405440, "adamw_lr": 0.0004, "muon/grad_norm": 0.22109739482402802, "adam/grad_norm": 0.20594650506973267, "muon/param_norm": 1957.4769287109375, "muon/update_ratio": 0.009493990801274776, "muon/update_ratio_max": 0.08263343572616577, "adam/param_norm": 696.7584228515625, "adam/update_ratio": 0.0011376513866707683, "adam/update_ratio_max": 0.002959176432341337, "ademamix/cos_fast_slow": 0.3965778648853302, "ademamix/slow_fast_norm_ratio": 0.18538588285446167, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.380859375, "loss_smoothed": 2.238643138607343, "seq_len": 4096} {"wall": 7140.829, "step": 5630, "loss": 2.2729491541783013, "lm_loss": 2.2729491541783013, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.41919341683387756, "tokens": 1660354560, "adamw_lr": 0.0004, "muon/grad_norm": 0.3148612678050995, "adam/grad_norm": 0.27640268206596375, "muon/param_norm": 1956.0050048828125, "muon/update_ratio": 0.00982239842414856, "muon/update_ratio_max": 0.07785438746213913, "adam/param_norm": 696.8040771484375, "adam/update_ratio": 0.0011211910750716925, "adam/update_ratio_max": 0.00294088339433074, "ademamix/cos_fast_slow": 0.1866423636674881, "ademamix/slow_fast_norm_ratio": 0.17141322791576385, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.380859375, "loss_smoothed": 2.301579337567091, "seq_len": 4096} {"wall": 7345.789, "step": 5640, "loss": 2.340511292219162, "lm_loss": 2.340511292219162, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3514616787433624, "tokens": 1663303680, "adamw_lr": 0.0004, "muon/grad_norm": 0.26679837703704834, "adam/grad_norm": 0.2289075255393982, "muon/param_norm": 1955.04931640625, "muon/update_ratio": 0.010110458359122276, "muon/update_ratio_max": 0.08520230650901794, "adam/param_norm": 696.8497924804688, "adam/update_ratio": 0.0011244190391153097, "adam/update_ratio_max": 0.003319285809993744, "ademamix/cos_fast_slow": 0.3692624866962433, "ademamix/slow_fast_norm_ratio": 0.1849445402622223, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.380859375, "loss_smoothed": 2.223074858387311, "seq_len": 4096} {"wall": 7550.799, "step": 5650, "loss": 2.339863896369934, "lm_loss": 2.339863896369934, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.33909404277801514, "tokens": 1666252800, "adamw_lr": 0.0004, "muon/grad_norm": 0.23081064224243164, "adam/grad_norm": 0.24816948175430298, "muon/param_norm": 1954.90625, "muon/update_ratio": 0.010260111652314663, "muon/update_ratio_max": 0.08742941170930862, "adam/param_norm": 696.8955078125, "adam/update_ratio": 0.0011132125509902835, "adam/update_ratio_max": 0.0029031222220510244, "ademamix/cos_fast_slow": 0.18231835961341858, "ademamix/slow_fast_norm_ratio": 0.16826286911964417, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.380859375, "loss_smoothed": 2.2805423557758333, "seq_len": 4096} {"wall": 7755.539, "step": 5660, "loss": 2.1912790536880493, "lm_loss": 2.1912790536880493, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.30023086071014404, "tokens": 1669201920, "adamw_lr": 0.0004, "muon/grad_norm": 0.23436620831489563, "adam/grad_norm": 0.18777796626091003, "muon/param_norm": 1955.38525390625, "muon/update_ratio": 0.010340576060116291, "muon/update_ratio_max": 0.08383893966674805, "adam/param_norm": 696.9417114257812, "adam/update_ratio": 0.0011120293056592345, "adam/update_ratio_max": 0.0028895558789372444, "ademamix/cos_fast_slow": 0.38670721650123596, "ademamix/slow_fast_norm_ratio": 0.19125083088874817, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.380859375, "loss_smoothed": 2.265355168779691, "seq_len": 4096} {"wall": 7960.357, "step": 5670, "loss": 2.2816834499438605, "lm_loss": 2.2816834499438605, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3504064083099365, "tokens": 1672151040, "adamw_lr": 0.0004, "muon/grad_norm": 0.2675439119338989, "adam/grad_norm": 0.2267179638147354, "muon/param_norm": 1956.099365234375, "muon/update_ratio": 0.010268856771290302, "muon/update_ratio_max": 0.08223985880613327, "adam/param_norm": 696.9871826171875, "adam/update_ratio": 0.001126046641729772, "adam/update_ratio_max": 0.002925092587247491, "ademamix/cos_fast_slow": 0.5129193663597107, "ademamix/slow_fast_norm_ratio": 0.17429184913635254, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.380859375, "loss_smoothed": 2.2299599843720594, "seq_len": 4096} {"wall": 8165.421, "step": 5680, "loss": 2.303608402609825, "lm_loss": 2.303608402609825, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3867301642894745, "tokens": 1675100160, "adamw_lr": 0.0004, "muon/grad_norm": 0.2980870306491852, "adam/grad_norm": 0.24682196974754333, "muon/param_norm": 1957.1204833984375, "muon/update_ratio": 0.010263808071613312, "muon/update_ratio_max": 0.08276588469743729, "adam/param_norm": 697.0349731445312, "adam/update_ratio": 0.0011294892756268382, "adam/update_ratio_max": 0.0029330055695027113, "ademamix/cos_fast_slow": 0.22249707579612732, "ademamix/slow_fast_norm_ratio": 0.16796769201755524, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.380859375, "loss_smoothed": 2.2594774797558785, "seq_len": 4096} {"wall": 8370.446, "step": 5690, "loss": 2.3007793178160987, "lm_loss": 2.3007793178160987, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2857891917228699, "tokens": 1678049280, "adamw_lr": 0.0004, "muon/grad_norm": 0.22006288170814514, "adam/grad_norm": 0.18237067759037018, "muon/param_norm": 1957.7513427734375, "muon/update_ratio": 0.010198017582297325, "muon/update_ratio_max": 0.08482172340154648, "adam/param_norm": 697.0801391601562, "adam/update_ratio": 0.0011297538876533508, "adam/update_ratio_max": 0.0029321948532015085, "ademamix/cos_fast_slow": 0.3070315718650818, "ademamix/slow_fast_norm_ratio": 0.17573925852775574, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.380859375, "loss_smoothed": 2.2505522377789022, "seq_len": 4096} {"wall": 8577.724, "eval_loss": 1.9394226968288422, "eval_ppl": 6.954735279083252, "step": 5700} {"wall": 8577.726, "step": 5700, "loss": 2.124493812521299, "lm_loss": 2.124493812521299, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3546895384788513, "tokens": 1680998400, "adamw_lr": 0.0004, "muon/grad_norm": 0.2719769775867462, "adam/grad_norm": 0.2272709459066391, "muon/param_norm": 1958.4400634765625, "muon/update_ratio": 0.010280964896082878, "muon/update_ratio_max": 0.08602257072925568, "adam/param_norm": 697.1248168945312, "adam/update_ratio": 0.0010963744716718793, "adam/update_ratio_max": 0.0028447993099689484, "ademamix/cos_fast_slow": 0.5505945086479187, "ademamix/slow_fast_norm_ratio": 0.16787000000476837, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.380859375, "loss_smoothed": 2.2712034836411474, "seq_len": 4096} {"wall": 8868.101, "step": 5710, "loss": 2.1822577069203057, "lm_loss": 2.1822577069203057, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.29985320568084717, "tokens": 1683947520, "adamw_lr": 0.0004, "muon/grad_norm": 0.22188033163547516, "adam/grad_norm": 0.20112426578998566, "muon/param_norm": 1958.784423828125, "muon/update_ratio": 0.010313084349036217, "muon/update_ratio_max": 0.08234080672264099, "adam/param_norm": 697.1689453125, "adam/update_ratio": 0.0011076208902522922, "adam/update_ratio_max": 0.0028728048782795668, "ademamix/cos_fast_slow": 0.4666985869407654, "ademamix/slow_fast_norm_ratio": 0.17416100203990936, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.556640625, "loss_smoothed": 2.2579903095960616, "seq_len": 4096} {"wall": 9072.796, "step": 5720, "loss": 2.224769244591395, "lm_loss": 2.224769244591395, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3506685197353363, "tokens": 1686896640, "adamw_lr": 0.0004, "muon/grad_norm": 0.25861138105392456, "adam/grad_norm": 0.23714163899421692, "muon/param_norm": 1958.9923095703125, "muon/update_ratio": 0.010305069386959076, "muon/update_ratio_max": 0.07814550399780273, "adam/param_norm": 697.2144165039062, "adam/update_ratio": 0.0011203978210687637, "adam/update_ratio_max": 0.002905034925788641, "ademamix/cos_fast_slow": 0.5129470825195312, "ademamix/slow_fast_norm_ratio": 0.17625372111797333, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.556640625, "loss_smoothed": 2.2606705938776335, "seq_len": 4096} {"wall": 9277.393, "step": 5730, "loss": 2.3075571258862815, "lm_loss": 2.3075571258862815, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2808573544025421, "tokens": 1689845760, "adamw_lr": 0.0004, "muon/grad_norm": 0.2261584848165512, "adam/grad_norm": 0.16626527905464172, "muon/param_norm": 1959.1790771484375, "muon/update_ratio": 0.010366812348365784, "muon/update_ratio_max": 0.08223807066679001, "adam/param_norm": 697.260986328125, "adam/update_ratio": 0.0011317284079268575, "adam/update_ratio_max": 0.0029765202198177576, "ademamix/cos_fast_slow": 0.4367903172969818, "ademamix/slow_fast_norm_ratio": 0.17775841057300568, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.556640625, "loss_smoothed": 2.2787339741985004, "seq_len": 4096} {"wall": 9482.054, "step": 5740, "loss": 2.2095649341742196, "lm_loss": 2.2095649341742196, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.31372275948524475, "tokens": 1692794880, "adamw_lr": 0.0004, "muon/grad_norm": 0.22391830384731293, "adam/grad_norm": 0.21971860527992249, "muon/param_norm": 1959.544677734375, "muon/update_ratio": 0.010365348309278488, "muon/update_ratio_max": 0.07793962210416794, "adam/param_norm": 697.3072509765625, "adam/update_ratio": 0.001133216661401093, "adam/update_ratio_max": 0.0029361406341195107, "ademamix/cos_fast_slow": 0.5661042332649231, "ademamix/slow_fast_norm_ratio": 0.15822896361351013, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.556640625, "loss_smoothed": 2.2743177225192386, "seq_len": 4096} {"wall": 9686.759, "step": 5750, "loss": 2.3440762162208557, "lm_loss": 2.3440762162208557, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.5148034691810608, "tokens": 1695744000, "adamw_lr": 0.0004, "muon/grad_norm": 0.4314412772655487, "adam/grad_norm": 0.28049975633621216, "muon/param_norm": 1960.0509033203125, "muon/update_ratio": 0.010251975618302822, "muon/update_ratio_max": 0.07856142520904541, "adam/param_norm": 697.3529663085938, "adam/update_ratio": 0.0011151545913890004, "adam/update_ratio_max": 0.0028881479520350695, "ademamix/cos_fast_slow": 0.07238154113292694, "ademamix/slow_fast_norm_ratio": 0.13179033994674683, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.556640625, "loss_smoothed": 2.2899985467394193, "seq_len": 4096} {"wall": 9891.421, "step": 5760, "loss": 2.3016868034998574, "lm_loss": 2.3016868034998574, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.34878358244895935, "tokens": 1698693120, "adamw_lr": 0.0004, "muon/grad_norm": 0.2375461310148239, "adam/grad_norm": 0.2545308470726013, "muon/param_norm": 1960.2613525390625, "muon/update_ratio": 0.010292181745171547, "muon/update_ratio_max": 0.07821682840585709, "adam/param_norm": 697.3971557617188, "adam/update_ratio": 0.0011179446009919047, "adam/update_ratio_max": 0.0028942495118826628, "ademamix/cos_fast_slow": 0.2410404086112976, "ademamix/slow_fast_norm_ratio": 0.16934262216091156, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.556640625, "loss_smoothed": 2.216635566453139, "seq_len": 4096} {"wall": 10096.224, "step": 5770, "loss": 2.10786784440279, "lm_loss": 2.10786784440279, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.284125953912735, "tokens": 1701642240, "adamw_lr": 0.0004, "muon/grad_norm": 0.22450153529644012, "adam/grad_norm": 0.1736041158437729, "muon/param_norm": 1960.6962890625, "muon/update_ratio": 0.010333254933357239, "muon/update_ratio_max": 0.07669412344694138, "adam/param_norm": 697.4395141601562, "adam/update_ratio": 0.001109373988583684, "adam/update_ratio_max": 0.002870927331969142, "ademamix/cos_fast_slow": 0.44967928528785706, "ademamix/slow_fast_norm_ratio": 0.17183423042297363, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.556640625, "loss_smoothed": 2.264224243412415, "seq_len": 4096} {"wall": 10301.137, "step": 5780, "loss": 2.263913827637831, "lm_loss": 2.263913827637831, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.31778010725975037, "tokens": 1704591360, "adamw_lr": 0.0004, "muon/grad_norm": 0.2711409330368042, "adam/grad_norm": 0.16572126746177673, "muon/param_norm": 1960.978271484375, "muon/update_ratio": 0.010267606005072594, "muon/update_ratio_max": 0.07968208938837051, "adam/param_norm": 697.4840698242188, "adam/update_ratio": 0.0011186213232576847, "adam/update_ratio_max": 0.002893861848860979, "ademamix/cos_fast_slow": 0.4911496043205261, "ademamix/slow_fast_norm_ratio": 0.16538484394550323, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.556640625, "loss_smoothed": 2.2290821192165216, "seq_len": 4096} {"wall": 10505.766, "step": 5790, "loss": 2.2749491035938263, "lm_loss": 2.2749491035938263, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.34144875407218933, "tokens": 1707540480, "adamw_lr": 0.0004, "muon/grad_norm": 0.2731405794620514, "adam/grad_norm": 0.20485827326774597, "muon/param_norm": 1961.17041015625, "muon/update_ratio": 0.010337342508137226, "muon/update_ratio_max": 0.0834478810429573, "adam/param_norm": 697.5283203125, "adam/update_ratio": 0.001135237980633974, "adam/update_ratio_max": 0.0029358002357184887, "ademamix/cos_fast_slow": 0.17520937323570251, "ademamix/slow_fast_norm_ratio": 0.16599081456661224, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.556640625, "loss_smoothed": 2.262734924256802, "seq_len": 4096} {"wall": 10712.819, "eval_loss": 1.9349409639835358, "eval_ppl": 6.923635005950928, "step": 5800} {"wall": 10712.82, "step": 5800, "loss": 2.2370287477970123, "lm_loss": 2.2370287477970123, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3262898921966553, "tokens": 1710489600, "adamw_lr": 0.0004, "muon/grad_norm": 0.24707722663879395, "adam/grad_norm": 0.21291320025920868, "muon/param_norm": 1961.2274169921875, "muon/update_ratio": 0.010352106764912605, "muon/update_ratio_max": 0.08484240621328354, "adam/param_norm": 697.57080078125, "adam/update_ratio": 0.001117818639613688, "adam/update_ratio_max": 0.0028898376040160656, "ademamix/cos_fast_slow": 0.19296576082706451, "ademamix/slow_fast_norm_ratio": 0.15001660585403442, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.556640625, "loss_smoothed": 2.2719001044829685, "seq_len": 4096} {"wall": 10985.918, "step": 5810, "loss": 2.303060531616211, "lm_loss": 2.303060531616211, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.34567341208457947, "tokens": 1713438720, "adamw_lr": 0.0004, "muon/grad_norm": 0.25115150213241577, "adam/grad_norm": 0.23757526278495789, "muon/param_norm": 1961.19775390625, "muon/update_ratio": 0.010264627635478973, "muon/update_ratio_max": 0.08078572899103165, "adam/param_norm": 697.6155395507812, "adam/update_ratio": 0.0011235300917178392, "adam/update_ratio_max": 0.002905461471527815, "ademamix/cos_fast_slow": 0.5224853157997131, "ademamix/slow_fast_norm_ratio": 0.16078758239746094, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 87.193359375, "loss_smoothed": 2.2824714546402296, "seq_len": 4096} {"wall": 11190.63, "step": 5820, "loss": 2.1943548570076623, "lm_loss": 2.1943548570076623, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.30024266242980957, "tokens": 1716387840, "adamw_lr": 0.0004, "muon/grad_norm": 0.22697091102600098, "adam/grad_norm": 0.19665803015232086, "muon/param_norm": 1961.2720947265625, "muon/update_ratio": 0.010331337340176105, "muon/update_ratio_max": 0.0750245600938797, "adam/param_norm": 697.6603393554688, "adam/update_ratio": 0.0011093139182776213, "adam/update_ratio_max": 0.0028656015638262033, "ademamix/cos_fast_slow": 0.24610216915607452, "ademamix/slow_fast_norm_ratio": 0.16661721467971802, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 87.193359375, "loss_smoothed": 2.219525705774625, "seq_len": 4096} {"wall": 11395.566, "step": 5830, "loss": 2.24636880060037, "lm_loss": 2.24636880060037, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3187268078327179, "tokens": 1719336960, "adamw_lr": 0.0004, "muon/grad_norm": 0.259511262178421, "adam/grad_norm": 0.1853506863117218, "muon/param_norm": 1961.30224609375, "muon/update_ratio": 0.010323241353034973, "muon/update_ratio_max": 0.07382664829492569, "adam/param_norm": 697.7044677734375, "adam/update_ratio": 0.001142418826930225, "adam/update_ratio_max": 0.0029501637909561396, "ademamix/cos_fast_slow": 0.3841795325279236, "ademamix/slow_fast_norm_ratio": 0.16630733013153076, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 87.193359375, "loss_smoothed": 2.231609829515219, "seq_len": 4096} {"wall": 11600.279, "step": 5840, "loss": 2.2189669062693915, "lm_loss": 2.2189669062693915, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3201228976249695, "tokens": 1722286080, "adamw_lr": 0.0004, "muon/grad_norm": 0.2415689378976822, "adam/grad_norm": 0.21035544574260712, "muon/param_norm": 1960.0599365234375, "muon/update_ratio": 0.009585062973201275, "muon/update_ratio_max": 0.07526121288537979, "adam/param_norm": 697.7491455078125, "adam/update_ratio": 0.0011203483445569873, "adam/update_ratio_max": 0.003555299947038293, "ademamix/cos_fast_slow": 0.5516106486320496, "ademamix/slow_fast_norm_ratio": 0.15342888236045837, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 87.193359375, "loss_smoothed": 2.2598476956288023, "seq_len": 4096} {"wall": 11805.032, "step": 5850, "loss": 2.2075602263212204, "lm_loss": 2.2075602263212204, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3698040246963501, "tokens": 1725235200, "adamw_lr": 0.0004, "muon/grad_norm": 0.2476743459701538, "adam/grad_norm": 0.27433857321739197, "muon/param_norm": 1958.391845703125, "muon/update_ratio": 0.009942441247403622, "muon/update_ratio_max": 0.08167313784360886, "adam/param_norm": 697.7933349609375, "adam/update_ratio": 0.001115982187911868, "adam/update_ratio_max": 0.002879542764276266, "ademamix/cos_fast_slow": 0.15015718340873718, "ademamix/slow_fast_norm_ratio": 0.13922081887722015, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 87.193359375, "loss_smoothed": 2.2171953034897642, "seq_len": 4096} {"wall": 12010.081, "step": 5860, "loss": 2.322172616918882, "lm_loss": 2.322172616918882, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.26923057436943054, "tokens": 1728184320, "adamw_lr": 0.0004, "muon/grad_norm": 0.21960121393203735, "adam/grad_norm": 0.1556635946035385, "muon/param_norm": 1957.63037109375, "muon/update_ratio": 0.010176142677664757, "muon/update_ratio_max": 0.08426222950220108, "adam/param_norm": 697.8367309570312, "adam/update_ratio": 0.001127598574385047, "adam/update_ratio_max": 0.002908532740548253, "ademamix/cos_fast_slow": 0.34694162011146545, "ademamix/slow_fast_norm_ratio": 0.1624803990125656, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 87.193359375, "loss_smoothed": 2.246958513557911, "seq_len": 4096} {"wall": 12214.745, "step": 5870, "loss": 2.3605535378058753, "lm_loss": 2.3605535378058753, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3675636649131775, "tokens": 1731133440, "adamw_lr": 0.0004, "muon/grad_norm": 0.23950083553791046, "adam/grad_norm": 0.2796747386455536, "muon/param_norm": 1957.259033203125, "muon/update_ratio": 0.010196234099566936, "muon/update_ratio_max": 0.08323720097541809, "adam/param_norm": 697.8818969726562, "adam/update_ratio": 0.0011302258353680372, "adam/update_ratio_max": 0.0029141053091734648, "ademamix/cos_fast_slow": 0.6258048415184021, "ademamix/slow_fast_norm_ratio": 0.14568805694580078, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 87.193359375, "loss_smoothed": 2.2804117441177367, "seq_len": 4096} {"wall": 12419.452, "step": 5880, "loss": 2.2986473192771277, "lm_loss": 2.2986473192771277, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2936614155769348, "tokens": 1734082560, "adamw_lr": 0.0004, "muon/grad_norm": 0.2535794675350189, "adam/grad_norm": 0.1480819135904312, "muon/param_norm": 1957.4183349609375, "muon/update_ratio": 0.01028257142752409, "muon/update_ratio_max": 0.08132968842983246, "adam/param_norm": 697.9259033203125, "adam/update_ratio": 0.0011122125433757901, "adam/update_ratio_max": 0.0028666972648352385, "ademamix/cos_fast_slow": 0.31815823912620544, "ademamix/slow_fast_norm_ratio": 0.1757218986749649, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 87.193359375, "loss_smoothed": 2.2482129380106928, "seq_len": 4096} {"wall": 12624.152, "step": 5890, "loss": 2.1483212411403656, "lm_loss": 2.1483212411403656, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2904950678348541, "tokens": 1737031680, "adamw_lr": 0.0004, "muon/grad_norm": 0.23047517240047455, "adam/grad_norm": 0.17677058279514313, "muon/param_norm": 1957.574951171875, "muon/update_ratio": 0.01029623206704855, "muon/update_ratio_max": 0.08260158449411392, "adam/param_norm": 697.9694213867188, "adam/update_ratio": 0.0011003342224285007, "adam/update_ratio_max": 0.003027750179171562, "ademamix/cos_fast_slow": 0.2503216862678528, "ademamix/slow_fast_norm_ratio": 0.1671723872423172, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 87.193359375, "loss_smoothed": 2.2393326359490553, "seq_len": 4096} {"wall": 12831.328, "eval_loss": 1.9324708133935928, "eval_ppl": 6.906553745269775, "step": 5900} {"wall": 12831.33, "step": 5900, "loss": 2.3586809982856116, "lm_loss": 2.3586809982856116, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.37005797028541565, "tokens": 1739980800, "adamw_lr": 0.0004, "muon/grad_norm": 0.2698573172092438, "adam/grad_norm": 0.2532430589199066, "muon/param_norm": 1957.7957763671875, "muon/update_ratio": 0.010286875069141388, "muon/update_ratio_max": 0.08289773762226105, "adam/param_norm": 698.0123291015625, "adam/update_ratio": 0.0011206023627892137, "adam/update_ratio_max": 0.0028904436621814966, "ademamix/cos_fast_slow": 0.47986242175102234, "ademamix/slow_fast_norm_ratio": 0.15836834907531738, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 87.193359375, "loss_smoothed": 2.231179748227199, "seq_len": 4096} {"wall": 13113.754, "step": 5910, "loss": 2.2055609126885733, "lm_loss": 2.2055609126885733, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.28878188133239746, "tokens": 1742929920, "adamw_lr": 0.0004, "muon/grad_norm": 0.23011422157287598, "adam/grad_norm": 0.17417845129966736, "muon/param_norm": 1958.0009765625, "muon/update_ratio": 0.010297446511685848, "muon/update_ratio_max": 0.08021941781044006, "adam/param_norm": 698.0552368164062, "adam/update_ratio": 0.001103859511204064, "adam/update_ratio_max": 0.0028862617909908295, "ademamix/cos_fast_slow": 0.45967602729797363, "ademamix/slow_fast_norm_ratio": 0.16650079190731049, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.244140625, "loss_smoothed": 2.2298761459688348, "seq_len": 4096} {"wall": 13318.154, "step": 5920, "loss": 2.219648336370786, "lm_loss": 2.219648336370786, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.38173267245292664, "tokens": 1745879040, "adamw_lr": 0.0004, "muon/grad_norm": 0.28290700912475586, "adam/grad_norm": 0.2572937607765198, "muon/param_norm": 1958.057861328125, "muon/update_ratio": 0.010254119522869587, "muon/update_ratio_max": 0.08309980481863022, "adam/param_norm": 698.099365234375, "adam/update_ratio": 0.0011186957126483321, "adam/update_ratio_max": 0.002879336243495345, "ademamix/cos_fast_slow": 0.5969088673591614, "ademamix/slow_fast_norm_ratio": 0.15265417098999023, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.244140625, "loss_smoothed": 2.281313646833102, "seq_len": 4096} {"wall": 13522.626, "step": 5930, "loss": 2.2438504099845886, "lm_loss": 2.2438504099845886, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.40444329380989075, "tokens": 1748828160, "adamw_lr": 0.0004, "muon/grad_norm": 0.2949620187282562, "adam/grad_norm": 0.27660471200942993, "muon/param_norm": 1958.21484375, "muon/update_ratio": 0.010237615555524826, "muon/update_ratio_max": 0.0834108367562294, "adam/param_norm": 698.143798828125, "adam/update_ratio": 0.0011225000489503145, "adam/update_ratio_max": 0.002887993585318327, "ademamix/cos_fast_slow": 0.19116760790348053, "ademamix/slow_fast_norm_ratio": 0.13801348209381104, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.244140625, "loss_smoothed": 2.2268637428681055, "seq_len": 4096} {"wall": 13727.376, "step": 5940, "loss": 2.225400075316429, "lm_loss": 2.225400075316429, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3011232018470764, "tokens": 1751777280, "adamw_lr": 0.0004, "muon/grad_norm": 0.2277860939502716, "adam/grad_norm": 0.1967998594045639, "muon/param_norm": 1958.422607421875, "muon/update_ratio": 0.010349342599511147, "muon/update_ratio_max": 0.08231285214424133, "adam/param_norm": 698.1870727539062, "adam/update_ratio": 0.0011046412400901318, "adam/update_ratio_max": 0.002840991597622633, "ademamix/cos_fast_slow": 0.2800775170326233, "ademamix/slow_fast_norm_ratio": 0.1669846475124359, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.244140625, "loss_smoothed": 2.211532687644164, "seq_len": 4096} {"wall": 13931.867, "step": 5950, "loss": 2.2867107639710107, "lm_loss": 2.2867107639710107, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.40245795249938965, "tokens": 1754726400, "adamw_lr": 0.0004, "muon/grad_norm": 0.277922123670578, "adam/grad_norm": 0.29183506965637207, "muon/param_norm": 1958.245849609375, "muon/update_ratio": 0.01022028923034668, "muon/update_ratio_max": 0.08247654885053635, "adam/param_norm": 698.2305908203125, "adam/update_ratio": 0.0011238851584494114, "adam/update_ratio_max": 0.002895947080105543, "ademamix/cos_fast_slow": 0.6066774129867554, "ademamix/slow_fast_norm_ratio": 0.14457879960536957, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.244140625, "loss_smoothed": 2.2538359542687734, "seq_len": 4096} {"wall": 14136.327, "step": 5960, "loss": 2.3697602351506553, "lm_loss": 2.3697602351506553, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.27590835094451904, "tokens": 1757675520, "adamw_lr": 0.0004, "muon/grad_norm": 0.22080093622207642, "adam/grad_norm": 0.16596132516860962, "muon/param_norm": 1957.9136962890625, "muon/update_ratio": 0.01030499953776598, "muon/update_ratio_max": 0.08091413974761963, "adam/param_norm": 698.2753295898438, "adam/update_ratio": 0.0011210816446691751, "adam/update_ratio_max": 0.0028811078518629074, "ademamix/cos_fast_slow": 0.5119785666465759, "ademamix/slow_fast_norm_ratio": 0.1623537838459015, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.244140625, "loss_smoothed": 2.295413743456205, "seq_len": 4096} {"wall": 14341.06, "step": 5970, "loss": 2.32583021124204, "lm_loss": 2.32583021124204, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.33112168312072754, "tokens": 1760624640, "adamw_lr": 0.0004, "muon/grad_norm": 0.2615351974964142, "adam/grad_norm": 0.20362049341201782, "muon/param_norm": 1955.4140625, "muon/update_ratio": 0.009708903729915619, "muon/update_ratio_max": 0.08076354116201401, "adam/param_norm": 698.3199462890625, "adam/update_ratio": 0.0011178873246535659, "adam/update_ratio_max": 0.00298844906501472, "ademamix/cos_fast_slow": 0.5908347964286804, "ademamix/slow_fast_norm_ratio": 0.14496798813343048, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.244140625, "loss_smoothed": 2.2878866657614707, "seq_len": 4096} {"wall": 14545.553, "step": 5980, "loss": 2.1736278931299844, "lm_loss": 2.1736278931299844, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3260287344455719, "tokens": 1763573760, "adamw_lr": 0.0004, "muon/grad_norm": 0.2785196006298065, "adam/grad_norm": 0.16947613656520844, "muon/param_norm": 1954.1912841796875, "muon/update_ratio": 0.009986368007957935, "muon/update_ratio_max": 0.07746081054210663, "adam/param_norm": 698.364990234375, "adam/update_ratio": 0.0011168888304382563, "adam/update_ratio_max": 0.002890239702537656, "ademamix/cos_fast_slow": 0.5237269997596741, "ademamix/slow_fast_norm_ratio": 0.15922047197818756, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.244140625, "loss_smoothed": 2.2866457156836986, "seq_len": 4096} {"wall": 14750.143, "step": 5990, "loss": 2.226038008928299, "lm_loss": 2.226038008928299, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.27617478370666504, "tokens": 1766522880, "adamw_lr": 0.0004, "muon/grad_norm": 0.21557916700839996, "adam/grad_norm": 0.17272959649562836, "muon/param_norm": 1953.589111328125, "muon/update_ratio": 0.01025746762752533, "muon/update_ratio_max": 0.07988626509904861, "adam/param_norm": 698.4105834960938, "adam/update_ratio": 0.0011227523209527135, "adam/update_ratio_max": 0.002915265504270792, "ademamix/cos_fast_slow": 0.23425480723381042, "ademamix/slow_fast_norm_ratio": 0.16110967099666595, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.244140625, "loss_smoothed": 2.2280841673413914, "seq_len": 4096} {"wall": 14957.062, "eval_loss": 1.9318017214536667, "eval_ppl": 6.9019341468811035, "step": 6000} {"wall": 14957.064, "step": 6000, "loss": 2.062079593539238, "lm_loss": 2.062079593539238, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.29608505964279175, "tokens": 1769472000, "adamw_lr": 0.0004, "muon/grad_norm": 0.20768769085407257, "adam/grad_norm": 0.21131788194179535, "muon/param_norm": 1953.567626953125, "muon/update_ratio": 0.010358658619225025, "muon/update_ratio_max": 0.08323647081851959, "adam/param_norm": 698.454833984375, "adam/update_ratio": 0.0011053149355575442, "adam/update_ratio_max": 0.002874156692996621, "ademamix/cos_fast_slow": 0.371584415435791, "ademamix/slow_fast_norm_ratio": 0.16271360218524933, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.244140625, "loss_smoothed": 2.2324359493951005, "seq_len": 4096} {"wall": 15235.586, "step": 6010, "loss": 2.1331530461708703, "lm_loss": 2.1331530461708703, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.32235532999038696, "tokens": 1772421120, "adamw_lr": 0.0004, "muon/grad_norm": 0.242487832903862, "adam/grad_norm": 0.21248573064804077, "muon/param_norm": 1953.699462890625, "muon/update_ratio": 0.010357791557908058, "muon/update_ratio_max": 0.08395528793334961, "adam/param_norm": 698.5003662109375, "adam/update_ratio": 0.0011218215804547071, "adam/update_ratio_max": 0.0029067194554954767, "ademamix/cos_fast_slow": 0.5068437457084656, "ademamix/slow_fast_norm_ratio": 0.15931259095668793, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.263671875, "loss_smoothed": 2.255220989137888, "seq_len": 4096} {"wall": 15440.28, "step": 6020, "loss": 2.2394294887781143, "lm_loss": 2.2394294887781143, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3370193839073181, "tokens": 1775370240, "adamw_lr": 0.0004, "muon/grad_norm": 0.25425174832344055, "adam/grad_norm": 0.2210906744003296, "muon/param_norm": 1954.02001953125, "muon/update_ratio": 0.010294556617736816, "muon/update_ratio_max": 0.08453502506017685, "adam/param_norm": 698.5448608398438, "adam/update_ratio": 0.0011228062212467194, "adam/update_ratio_max": 0.002879127860069275, "ademamix/cos_fast_slow": 0.15695609152317047, "ademamix/slow_fast_norm_ratio": 0.152213916182518, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.263671875, "loss_smoothed": 2.220130104819934, "seq_len": 4096} {"wall": 15644.843, "step": 6030, "loss": 2.2447179754575095, "lm_loss": 2.2447179754575095, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.29360806941986084, "tokens": 1778319360, "adamw_lr": 0.0004, "muon/grad_norm": 0.23798254132270813, "adam/grad_norm": 0.1717427521944046, "muon/param_norm": 1954.570068359375, "muon/update_ratio": 0.010347931645810604, "muon/update_ratio_max": 0.08478594571352005, "adam/param_norm": 698.5875244140625, "adam/update_ratio": 0.0011039008386433125, "adam/update_ratio_max": 0.0028296285308897495, "ademamix/cos_fast_slow": 0.39853355288505554, "ademamix/slow_fast_norm_ratio": 0.16728325188159943, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.263671875, "loss_smoothed": 2.235915906727314, "seq_len": 4096} {"wall": 15849.429, "step": 6040, "loss": 2.201822896798452, "lm_loss": 2.201822896798452, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3377438485622406, "tokens": 1781268480, "adamw_lr": 0.0004, "muon/grad_norm": 0.272680401802063, "adam/grad_norm": 0.19930826127529144, "muon/param_norm": 1955.01171875, "muon/update_ratio": 0.010341116227209568, "muon/update_ratio_max": 0.08705153316259384, "adam/param_norm": 698.6315307617188, "adam/update_ratio": 0.001119251479394734, "adam/update_ratio_max": 0.002868113573640585, "ademamix/cos_fast_slow": 0.48490241169929504, "ademamix/slow_fast_norm_ratio": 0.16165956854820251, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.263671875, "loss_smoothed": 2.2298132926225662, "seq_len": 4096} {"wall": 16054.027, "step": 6050, "loss": 2.249732494354248, "lm_loss": 2.249732494354248, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.37681668996810913, "tokens": 1784217600, "adamw_lr": 0.0004, "muon/grad_norm": 0.2797986567020416, "adam/grad_norm": 0.25264352560043335, "muon/param_norm": 1955.1700439453125, "muon/update_ratio": 0.010261745192110538, "muon/update_ratio_max": 0.08473273366689682, "adam/param_norm": 698.6755981445312, "adam/update_ratio": 0.001128186471760273, "adam/update_ratio_max": 0.0028899051249027252, "ademamix/cos_fast_slow": 0.1901739239692688, "ademamix/slow_fast_norm_ratio": 0.14146970212459564, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.263671875, "loss_smoothed": 2.2540185332298277, "seq_len": 4096} {"wall": 16258.861, "step": 6060, "loss": 2.244700556000074, "lm_loss": 2.244700556000074, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.317087322473526, "tokens": 1787166720, "adamw_lr": 0.0004, "muon/grad_norm": 0.24070793390274048, "adam/grad_norm": 0.20649270713329315, "muon/param_norm": 1955.3070068359375, "muon/update_ratio": 0.010345629416406155, "muon/update_ratio_max": 0.07936807721853256, "adam/param_norm": 698.7196044921875, "adam/update_ratio": 0.0011171407531946898, "adam/update_ratio_max": 0.0028607286512851715, "ademamix/cos_fast_slow": 0.21609508991241455, "ademamix/slow_fast_norm_ratio": 0.14629100263118744, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.263671875, "loss_smoothed": 2.2836011037230493, "seq_len": 4096} {"wall": 16463.492, "step": 6070, "loss": 2.1025436967611313, "lm_loss": 2.1025436967611313, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2663581371307373, "tokens": 1790115840, "adamw_lr": 0.0004, "muon/grad_norm": 0.21962213516235352, "adam/grad_norm": 0.15127670764923096, "muon/param_norm": 1955.71630859375, "muon/update_ratio": 0.010388745926320553, "muon/update_ratio_max": 0.08108915388584137, "adam/param_norm": 698.7623901367188, "adam/update_ratio": 0.0011108495527878404, "adam/update_ratio_max": 0.002843724098056555, "ademamix/cos_fast_slow": 0.3228417932987213, "ademamix/slow_fast_norm_ratio": 0.17130722105503082, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.263671875, "loss_smoothed": 2.2179596215486526, "seq_len": 4096} {"wall": 16668.091, "step": 6080, "loss": 2.261456787586212, "lm_loss": 2.261456787586212, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3623914420604706, "tokens": 1793064960, "adamw_lr": 0.0004, "muon/grad_norm": 0.260525643825531, "adam/grad_norm": 0.25159668922424316, "muon/param_norm": 1956.0084228515625, "muon/update_ratio": 0.010291163809597492, "muon/update_ratio_max": 0.08501596748828888, "adam/param_norm": 698.8068237304688, "adam/update_ratio": 0.0011192490346729755, "adam/update_ratio_max": 0.002864283276721835, "ademamix/cos_fast_slow": 0.1783948689699173, "ademamix/slow_fast_norm_ratio": 0.13982820510864258, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.263671875, "loss_smoothed": 2.2488150348265963, "seq_len": 4096} {"wall": 16872.974, "step": 6090, "loss": 2.214984342455864, "lm_loss": 2.214984342455864, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3018248379230499, "tokens": 1796014080, "adamw_lr": 0.0004, "muon/grad_norm": 0.22417595982551575, "adam/grad_norm": 0.2022089958190918, "muon/param_norm": 1956.2000732421875, "muon/update_ratio": 0.01035806629806757, "muon/update_ratio_max": 0.08175025135278702, "adam/param_norm": 698.8509521484375, "adam/update_ratio": 0.0011270492104813457, "adam/update_ratio_max": 0.0028831304516643286, "ademamix/cos_fast_slow": 0.5210108160972595, "ademamix/slow_fast_norm_ratio": 0.15537802875041962, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.263671875, "loss_smoothed": 2.2172315632303556, "seq_len": 4096} {"wall": 17080.106, "eval_loss": 1.925626054406166, "eval_ppl": 6.859441757202148, "step": 6100} {"wall": 17080.107, "step": 6100, "loss": 2.2519205758968988, "lm_loss": 2.2519205758968988, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2748687267303467, "tokens": 1798963200, "adamw_lr": 0.0004, "muon/grad_norm": 0.2127678543329239, "adam/grad_norm": 0.1739804893732071, "muon/param_norm": 1955.8883056640625, "muon/update_ratio": 0.010271537117660046, "muon/update_ratio_max": 0.0820690467953682, "adam/param_norm": 698.8948974609375, "adam/update_ratio": 0.0011094511719420552, "adam/update_ratio_max": 0.00283672078512609, "ademamix/cos_fast_slow": 0.4529344439506531, "ademamix/slow_fast_norm_ratio": 0.14923037588596344, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.263671875, "loss_smoothed": 2.2392301360766096, "seq_len": 4096} {"wall": 17363.583, "step": 6110, "loss": 2.374813531835874, "lm_loss": 2.374813531835874, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3433511555194855, "tokens": 1801912320, "adamw_lr": 0.0004, "muon/grad_norm": 0.27018532156944275, "adam/grad_norm": 0.21202914416790009, "muon/param_norm": 1955.4952392578125, "muon/update_ratio": 0.010236796922981739, "muon/update_ratio_max": 0.07927681505680084, "adam/param_norm": 698.940185546875, "adam/update_ratio": 0.0011136091779917479, "adam/update_ratio_max": 0.0028467413503676653, "ademamix/cos_fast_slow": 0.09790332615375519, "ademamix/slow_fast_norm_ratio": 0.1313459426164627, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 87.564453125, "loss_smoothed": 2.2505930890639623, "seq_len": 4096} {"wall": 17568.167, "step": 6120, "loss": 2.3613444765408835, "lm_loss": 2.3613444765408835, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2933277487754822, "tokens": 1804861440, "adamw_lr": 0.0004, "muon/grad_norm": 0.22959181666374207, "adam/grad_norm": 0.18270286917686462, "muon/param_norm": 1955.6204833984375, "muon/update_ratio": 0.010317509062588215, "muon/update_ratio_max": 0.082325279712677, "adam/param_norm": 698.9830932617188, "adam/update_ratio": 0.0011223199544474483, "adam/update_ratio_max": 0.002868054434657097, "ademamix/cos_fast_slow": 0.4647497236728668, "ademamix/slow_fast_norm_ratio": 0.1609926074743271, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 87.564453125, "loss_smoothed": 2.2495647897322972, "seq_len": 4096} {"wall": 17772.675, "step": 6130, "loss": 2.290799140930176, "lm_loss": 2.290799140930176, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.33593952655792236, "tokens": 1807810560, "adamw_lr": 0.0004, "muon/grad_norm": 0.23264241218566895, "adam/grad_norm": 0.24208828806877136, "muon/param_norm": 1955.9512939453125, "muon/update_ratio": 0.010339279659092426, "muon/update_ratio_max": 0.08021200448274612, "adam/param_norm": 699.0249633789062, "adam/update_ratio": 0.0011115293018519878, "adam/update_ratio_max": 0.0028393901884555817, "ademamix/cos_fast_slow": 0.47447848320007324, "ademamix/slow_fast_norm_ratio": 0.15206077694892883, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 87.564453125, "loss_smoothed": 2.2312280719478923, "seq_len": 4096} {"wall": 17977.134, "step": 6140, "loss": 2.3184603403011956, "lm_loss": 2.3184603403011956, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.30506739020347595, "tokens": 1810759680, "adamw_lr": 0.0004, "muon/grad_norm": 0.23318122327327728, "adam/grad_norm": 0.19624388217926025, "muon/param_norm": 1956.1353759765625, "muon/update_ratio": 0.010302377864718437, "muon/update_ratio_max": 0.08291497081518173, "adam/param_norm": 699.0684204101562, "adam/update_ratio": 0.0011128764599561691, "adam/update_ratio_max": 0.0028417950961738825, "ademamix/cos_fast_slow": 0.44276532530784607, "ademamix/slow_fast_norm_ratio": 0.14576607942581177, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 87.564453125, "loss_smoothed": 2.2258429234226544, "seq_len": 4096} {"wall": 18181.631, "step": 6150, "loss": 2.360014388958613, "lm_loss": 2.360014388958613, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.4114752411842346, "tokens": 1813708800, "adamw_lr": 0.0004, "muon/grad_norm": 0.28116291761398315, "adam/grad_norm": 0.3005664646625519, "muon/param_norm": 1956.106689453125, "muon/update_ratio": 0.0102534806355834, "muon/update_ratio_max": 0.08151459693908691, "adam/param_norm": 699.11279296875, "adam/update_ratio": 0.0011170421494171023, "adam/update_ratio_max": 0.0028514110017567873, "ademamix/cos_fast_slow": 0.6618120074272156, "ademamix/slow_fast_norm_ratio": 0.1296454221010208, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 87.564453125, "loss_smoothed": 2.3093857710560166, "seq_len": 4096} {"wall": 18385.952, "step": 6160, "loss": 2.1014452824989953, "lm_loss": 2.1014452824989953, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.250838965177536, "tokens": 1816657920, "adamw_lr": 0.0004, "muon/grad_norm": 0.20289993286132812, "adam/grad_norm": 0.14749956130981445, "muon/param_norm": 1956.1082763671875, "muon/update_ratio": 0.01036540325731039, "muon/update_ratio_max": 0.07754017412662506, "adam/param_norm": 699.15869140625, "adam/update_ratio": 0.001111677149310708, "adam/update_ratio_max": 0.002836640225723386, "ademamix/cos_fast_slow": 0.31734171509742737, "ademamix/slow_fast_norm_ratio": 0.16363388299942017, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 87.564453125, "loss_smoothed": 2.2008904568850993, "seq_len": 4096} {"wall": 18590.549, "step": 6170, "loss": 2.2225137253602347, "lm_loss": 2.2225137253602347, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3677746653556824, "tokens": 1819607040, "adamw_lr": 0.0004, "muon/grad_norm": 0.3153975009918213, "adam/grad_norm": 0.18911528587341309, "muon/param_norm": 1956.0252685546875, "muon/update_ratio": 0.01033799722790718, "muon/update_ratio_max": 0.07997202128171921, "adam/param_norm": 699.2051391601562, "adam/update_ratio": 0.0011253304546698928, "adam/update_ratio_max": 0.0028705799486488104, "ademamix/cos_fast_slow": 0.3036539554595947, "ademamix/slow_fast_norm_ratio": 0.15311573445796967, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 87.564453125, "loss_smoothed": 2.205646491299073, "seq_len": 4096} {"wall": 18795.065, "step": 6180, "loss": 2.246986518303553, "lm_loss": 2.246986518303553, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.33628207445144653, "tokens": 1822556160, "adamw_lr": 0.0004, "muon/grad_norm": 0.24487118422985077, "adam/grad_norm": 0.2310411036014557, "muon/param_norm": 1955.911865234375, "muon/update_ratio": 0.010326686315238476, "muon/update_ratio_max": 0.08164804428815842, "adam/param_norm": 699.2510986328125, "adam/update_ratio": 0.001117671257816255, "adam/update_ratio_max": 0.0028498407918959856, "ademamix/cos_fast_slow": 0.5050949454307556, "ademamix/slow_fast_norm_ratio": 0.15078115463256836, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 87.564453125, "loss_smoothed": 2.2416029453277586, "seq_len": 4096} {"wall": 18999.632, "step": 6190, "loss": 2.33236933251222, "lm_loss": 2.33236933251222, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.38990458846092224, "tokens": 1825505280, "adamw_lr": 0.0004, "muon/grad_norm": 0.2791048586368561, "adam/grad_norm": 0.273113876581192, "muon/param_norm": 1955.7347412109375, "muon/update_ratio": 0.010265461169183254, "muon/update_ratio_max": 0.08196459710597992, "adam/param_norm": 699.2977294921875, "adam/update_ratio": 0.0011148623889312148, "adam/update_ratio_max": 0.0028415846172720194, "ademamix/cos_fast_slow": 0.5667711496353149, "ademamix/slow_fast_norm_ratio": 0.13103143870830536, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 87.564453125, "loss_smoothed": 2.234791656335195, "seq_len": 4096} {"wall": 19206.47, "eval_loss": 1.9236205518245697, "eval_ppl": 6.845699310302734, "step": 6200} {"wall": 19206.472, "step": 6200, "loss": 2.2566911379496255, "lm_loss": 2.2566911379496255, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2880415916442871, "tokens": 1828454400, "adamw_lr": 0.0004, "muon/grad_norm": 0.20020832121372223, "adam/grad_norm": 0.20674479007720947, "muon/param_norm": 1956.1141357421875, "muon/update_ratio": 0.010388936847448349, "muon/update_ratio_max": 0.07936105877161026, "adam/param_norm": 699.3434448242188, "adam/update_ratio": 0.001103681861422956, "adam/update_ratio_max": 0.0028121231589466333, "ademamix/cos_fast_slow": 0.4702085554599762, "ademamix/slow_fast_norm_ratio": 0.14971701800823212, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 87.564453125, "loss_smoothed": 2.2818642680843673, "seq_len": 4096} {"wall": 19485.444, "step": 6210, "loss": 2.3260751416285834, "lm_loss": 2.3260751416285834, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.37086471915245056, "tokens": 1831403520, "adamw_lr": 0.0004, "muon/grad_norm": 0.25098246335983276, "adam/grad_norm": 0.2728824019432068, "muon/param_norm": 1956.169677734375, "muon/update_ratio": 0.010281946510076523, "muon/update_ratio_max": 0.08323168009519577, "adam/param_norm": 699.387939453125, "adam/update_ratio": 0.0011286414228379726, "adam/update_ratio_max": 0.002874854253605008, "ademamix/cos_fast_slow": 0.5114361047744751, "ademamix/slow_fast_norm_ratio": 0.14328540861606598, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.341796875, "loss_smoothed": 2.239574437836806, "seq_len": 4096} {"wall": 19690.305, "step": 6220, "loss": 2.156107574701309, "lm_loss": 2.156107574701309, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3609350323677063, "tokens": 1834352640, "adamw_lr": 0.0004, "muon/grad_norm": 0.27766552567481995, "adam/grad_norm": 0.23090913891792297, "muon/param_norm": 1956.172119140625, "muon/update_ratio": 0.010257291607558727, "muon/update_ratio_max": 0.08044469356536865, "adam/param_norm": 699.43310546875, "adam/update_ratio": 0.001115213381126523, "adam/update_ratio_max": 0.0028395382687449455, "ademamix/cos_fast_slow": 0.19957253336906433, "ademamix/slow_fast_norm_ratio": 0.13790060579776764, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.341796875, "loss_smoothed": 2.180080750087897, "seq_len": 4096} {"wall": 19895.108, "step": 6230, "loss": 2.158857688307762, "lm_loss": 2.158857688307762, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.27542904019355774, "tokens": 1837301760, "adamw_lr": 0.0004, "muon/grad_norm": 0.2207660973072052, "adam/grad_norm": 0.1642150729894638, "muon/param_norm": 1956.0498046875, "muon/update_ratio": 0.010318338871002197, "muon/update_ratio_max": 0.07980788499116898, "adam/param_norm": 699.4747924804688, "adam/update_ratio": 0.0011215441627427936, "adam/update_ratio_max": 0.002937884069979191, "ademamix/cos_fast_slow": 0.2795415222644806, "ademamix/slow_fast_norm_ratio": 0.158836230635643, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.341796875, "loss_smoothed": 2.2236168816685677, "seq_len": 4096} {"wall": 20100.27, "step": 6240, "loss": 2.323360567291578, "lm_loss": 2.323360567291578, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3662949204444885, "tokens": 1840250880, "adamw_lr": 0.0004, "muon/grad_norm": 0.2662179470062256, "adam/grad_norm": 0.2523513436317444, "muon/param_norm": 1955.7806396484375, "muon/update_ratio": 0.010293223895132542, "muon/update_ratio_max": 0.0807361826300621, "adam/param_norm": 699.51806640625, "adam/update_ratio": 0.0011190376244485378, "adam/update_ratio_max": 0.0028473057318478823, "ademamix/cos_fast_slow": 0.5411744117736816, "ademamix/slow_fast_norm_ratio": 0.13206009566783905, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.341796875, "loss_smoothed": 2.23103737582763, "seq_len": 4096} {"wall": 20305.062, "step": 6250, "loss": 2.2911230673392615, "lm_loss": 2.2911230673392615, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3130630552768707, "tokens": 1843200000, "adamw_lr": 0.0004, "muon/grad_norm": 0.2691566050052643, "adam/grad_norm": 0.1597924530506134, "muon/param_norm": 1954.9365234375, "muon/update_ratio": 0.010058706626296043, "muon/update_ratio_max": 0.08255128562450409, "adam/param_norm": 699.5625610351562, "adam/update_ratio": 0.0011204537004232407, "adam/update_ratio_max": 0.0028684830758720636, "ademamix/cos_fast_slow": 0.34921830892562866, "ademamix/slow_fast_norm_ratio": 0.14495940506458282, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.341796875, "loss_smoothed": 2.2829130043586097, "seq_len": 4096} {"wall": 20509.85, "step": 6260, "loss": 2.206106429298719, "lm_loss": 2.206106429298719, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.35361993312835693, "tokens": 1846149120, "adamw_lr": 0.0004, "muon/grad_norm": 0.28767555952072144, "adam/grad_norm": 0.20600610971450806, "muon/param_norm": 1954.5341796875, "muon/update_ratio": 0.010243361815810204, "muon/update_ratio_max": 0.07836788147687912, "adam/param_norm": 699.60791015625, "adam/update_ratio": 0.0011171449441462755, "adam/update_ratio_max": 0.0028404484037309885, "ademamix/cos_fast_slow": 0.2631348967552185, "ademamix/slow_fast_norm_ratio": 0.1552371084690094, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.341796875, "loss_smoothed": 2.2821837936838465, "seq_len": 4096} {"wall": 20714.653, "step": 6270, "loss": 2.315541779001554, "lm_loss": 2.315541779001554, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3494569957256317, "tokens": 1849098240, "adamw_lr": 0.0004, "muon/grad_norm": 0.2894538342952728, "adam/grad_norm": 0.19566556811332703, "muon/param_norm": 1954.53515625, "muon/update_ratio": 0.010226573795080185, "muon/update_ratio_max": 0.08061345666646957, "adam/param_norm": 699.6513671875, "adam/update_ratio": 0.0011310772970318794, "adam/update_ratio_max": 0.0028747913893312216, "ademamix/cos_fast_slow": 0.5980537533760071, "ademamix/slow_fast_norm_ratio": 0.14319461584091187, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.341796875, "loss_smoothed": 2.2520127308865385, "seq_len": 4096} {"wall": 20919.57, "step": 6280, "loss": 2.2903530498345694, "lm_loss": 2.2903530498345694, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.31648728251457214, "tokens": 1852047360, "adamw_lr": 0.0004, "muon/grad_norm": 0.2544022500514984, "adam/grad_norm": 0.18884903192520142, "muon/param_norm": 1953.7418212890625, "muon/update_ratio": 0.010213728062808514, "muon/update_ratio_max": 0.07928505539894104, "adam/param_norm": 699.6952514648438, "adam/update_ratio": 0.001126488670706749, "adam/update_ratio_max": 0.0028618774376809597, "ademamix/cos_fast_slow": 0.48036032915115356, "ademamix/slow_fast_norm_ratio": 0.1536177098751068, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.341796875, "loss_smoothed": 2.233883027235667, "seq_len": 4096} {"wall": 21124.363, "step": 6290, "loss": 2.181689520676931, "lm_loss": 2.181689520676931, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.27833279967308044, "tokens": 1854996480, "adamw_lr": 0.0004, "muon/grad_norm": 0.24823671579360962, "adam/grad_norm": 0.1260010451078415, "muon/param_norm": 1953.4991455078125, "muon/update_ratio": 0.010366413742303848, "muon/update_ratio_max": 0.0790846198797226, "adam/param_norm": 699.7388916015625, "adam/update_ratio": 0.0011107808677479625, "adam/update_ratio_max": 0.0028209977317601442, "ademamix/cos_fast_slow": 0.38242846727371216, "ademamix/slow_fast_norm_ratio": 0.16635935008525848, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.341796875, "loss_smoothed": 2.267614348232746, "seq_len": 4096} {"wall": 21331.534, "eval_loss": 1.929088056087494, "eval_ppl": 6.883230686187744, "step": 6300} {"wall": 21331.536, "step": 6300, "loss": 2.323396513859431, "lm_loss": 2.323396513859431, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.329689621925354, "tokens": 1857945600, "adamw_lr": 0.0004, "muon/grad_norm": 0.24759897589683533, "adam/grad_norm": 0.21788480877876282, "muon/param_norm": 1953.8642578125, "muon/update_ratio": 0.010330836288630962, "muon/update_ratio_max": 0.0811643972992897, "adam/param_norm": 699.7820434570312, "adam/update_ratio": 0.001107303542084992, "adam/update_ratio_max": 0.00281114736571908, "ademamix/cos_fast_slow": 0.1525723934173584, "ademamix/slow_fast_norm_ratio": 0.14072929322719574, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.341796875, "loss_smoothed": 2.2307470445831616, "seq_len": 4096} {"wall": 21613.318, "step": 6310, "loss": 2.212636947631836, "lm_loss": 2.212636947631836, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.33087021112442017, "tokens": 1860894720, "adamw_lr": 0.0004, "muon/grad_norm": 0.2336714267730713, "adam/grad_norm": 0.23457981646060944, "muon/param_norm": 1954.6822509765625, "muon/update_ratio": 0.010342254303395748, "muon/update_ratio_max": 0.07938449084758759, "adam/param_norm": 699.8239135742188, "adam/update_ratio": 0.0011075726943090558, "adam/update_ratio_max": 0.0028113576117902994, "ademamix/cos_fast_slow": 0.4048003852367401, "ademamix/slow_fast_norm_ratio": 0.14514796435832977, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.302734375, "loss_smoothed": 2.253888102869193, "seq_len": 4096} {"wall": 21818.033, "step": 6320, "loss": 2.2221476634343467, "lm_loss": 2.2221476634343467, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.33810627460479736, "tokens": 1863843840, "adamw_lr": 0.0004, "muon/grad_norm": 0.25132423639297485, "adam/grad_norm": 0.22669845819473267, "muon/param_norm": 1955.1104736328125, "muon/update_ratio": 0.010278159752488136, "muon/update_ratio_max": 0.07953768223524094, "adam/param_norm": 699.8662109375, "adam/update_ratio": 0.0011188954813405871, "adam/update_ratio_max": 0.002838991116732359, "ademamix/cos_fast_slow": 0.5901783108711243, "ademamix/slow_fast_norm_ratio": 0.14336231350898743, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.302734375, "loss_smoothed": 2.267107897500197, "seq_len": 4096} {"wall": 22022.8, "step": 6330, "loss": 2.404520963629087, "lm_loss": 2.404520963629087, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.33830127120018005, "tokens": 1866792960, "adamw_lr": 0.0004, "muon/grad_norm": 0.25561609864234924, "adam/grad_norm": 0.22149302065372467, "muon/param_norm": 1954.7567138671875, "muon/update_ratio": 0.009825387969613075, "muon/update_ratio_max": 0.08107440173625946, "adam/param_norm": 699.91015625, "adam/update_ratio": 0.0011077964445576072, "adam/update_ratio_max": 0.0028090712148696184, "ademamix/cos_fast_slow": 0.6508665084838867, "ademamix/slow_fast_norm_ratio": 0.0671086311340332, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.302734375, "loss_smoothed": 2.2863911027709642, "seq_len": 4096} {"wall": 22227.524, "step": 6340, "loss": 2.3328912804524102, "lm_loss": 2.3328912804524102, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.28181421756744385, "tokens": 1869742080, "adamw_lr": 0.0004, "muon/grad_norm": 0.2261882722377777, "adam/grad_norm": 0.1685514748096466, "muon/param_norm": 1954.1192626953125, "muon/update_ratio": 0.010161695070564747, "muon/update_ratio_max": 0.0798039510846138, "adam/param_norm": 699.9526977539062, "adam/update_ratio": 0.0011166386539116502, "adam/update_ratio_max": 0.002831135643646121, "ademamix/cos_fast_slow": 0.675818145275116, "ademamix/slow_fast_norm_ratio": 0.1378822922706604, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.302734375, "loss_smoothed": 2.26581021224459, "seq_len": 4096} {"wall": 22432.208, "step": 6350, "loss": 2.2908369849125543, "lm_loss": 2.2908369849125543, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.27612125873565674, "tokens": 1872691200, "adamw_lr": 0.0004, "muon/grad_norm": 0.21390917897224426, "adam/grad_norm": 0.17506946623325348, "muon/param_norm": 1951.6668701171875, "muon/update_ratio": 0.009743849746882915, "muon/update_ratio_max": 0.08161631971597672, "adam/param_norm": 699.99560546875, "adam/update_ratio": 0.0011182405287399888, "adam/update_ratio_max": 0.0028335386887192726, "ademamix/cos_fast_slow": 0.5845619440078735, "ademamix/slow_fast_norm_ratio": 0.16784609854221344, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.302734375, "loss_smoothed": 2.292127536733945, "seq_len": 4096} {"wall": 22637.288, "step": 6360, "loss": 2.2614586800336838, "lm_loss": 2.2614586800336838, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2523075342178345, "tokens": 1875640320, "adamw_lr": 0.0004, "muon/grad_norm": 0.20503555238246918, "adam/grad_norm": 0.14654532074928284, "muon/param_norm": 1950.763671875, "muon/update_ratio": 0.010077692568302155, "muon/update_ratio_max": 0.08139175921678543, "adam/param_norm": 700.0391845703125, "adam/update_ratio": 0.0011170883662998676, "adam/update_ratio_max": 0.0028299125842750072, "ademamix/cos_fast_slow": 0.4570518136024475, "ademamix/slow_fast_norm_ratio": 0.1929081231355667, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.302734375, "loss_smoothed": 2.2391013110677402, "seq_len": 4096} {"wall": 22842.096, "step": 6370, "loss": 2.2223020792007446, "lm_loss": 2.2223020792007446, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.36376768350601196, "tokens": 1878589440, "adamw_lr": 0.0004, "muon/grad_norm": 0.26204246282577515, "adam/grad_norm": 0.2527657151222229, "muon/param_norm": 1950.7745361328125, "muon/update_ratio": 0.010120955295860767, "muon/update_ratio_max": 0.07882802188396454, "adam/param_norm": 700.0833740234375, "adam/update_ratio": 0.001102737383916974, "adam/update_ratio_max": 0.002792748622596264, "ademamix/cos_fast_slow": 0.527883768081665, "ademamix/slow_fast_norm_ratio": 0.17704245448112488, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.302734375, "loss_smoothed": 2.193653742472331, "seq_len": 4096} {"wall": 23046.825, "step": 6380, "loss": 2.153037905693054, "lm_loss": 2.153037905693054, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.28906264901161194, "tokens": 1881538560, "adamw_lr": 0.0004, "muon/grad_norm": 0.22952596843242645, "adam/grad_norm": 0.17530760169029236, "muon/param_norm": 1950.0869140625, "muon/update_ratio": 0.010255328379571438, "muon/update_ratio_max": 0.08134446293115616, "adam/param_norm": 700.1265258789062, "adam/update_ratio": 0.0011113766813650727, "adam/update_ratio_max": 0.0029445148538798094, "ademamix/cos_fast_slow": 0.2959023714065552, "ademamix/slow_fast_norm_ratio": 0.16380605101585388, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.302734375, "loss_smoothed": 2.1921593164404234, "seq_len": 4096} {"wall": 23251.585, "step": 6390, "loss": 2.226190303762754, "lm_loss": 2.226190303762754, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.31065812706947327, "tokens": 1884487680, "adamw_lr": 0.0004, "muon/grad_norm": 0.2371910959482193, "adam/grad_norm": 0.2005554735660553, "muon/param_norm": 1949.9534912109375, "muon/update_ratio": 0.010301030240952969, "muon/update_ratio_max": 0.08151722699403763, "adam/param_norm": 700.1683959960938, "adam/update_ratio": 0.0011132736690342426, "adam/update_ratio_max": 0.0028178379870951176, "ademamix/cos_fast_slow": 0.24401715397834778, "ademamix/slow_fast_norm_ratio": 0.18079787492752075, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.302734375, "loss_smoothed": 2.2329645432531833, "seq_len": 4096} {"wall": 23458.609, "eval_loss": 1.9205068498849869, "eval_ppl": 6.824416637420654, "step": 6400} {"wall": 23458.611, "step": 6400, "loss": 2.267684275905291, "lm_loss": 2.267684275905291, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3823656737804413, "tokens": 1887436800, "adamw_lr": 0.0004, "muon/grad_norm": 0.2846863269805908, "adam/grad_norm": 0.2553001344203949, "muon/param_norm": 1950.4227294921875, "muon/update_ratio": 0.01024044118821621, "muon/update_ratio_max": 0.08440664410591125, "adam/param_norm": 700.2108154296875, "adam/update_ratio": 0.001112140715122223, "adam/update_ratio_max": 0.0028138672932982445, "ademamix/cos_fast_slow": 0.1720370352268219, "ademamix/slow_fast_norm_ratio": 0.1605573296546936, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.302734375, "loss_smoothed": 2.240089636047681, "seq_len": 4096} {"wall": 23741.692, "step": 6410, "loss": 2.2411272128423056, "lm_loss": 2.2411272128423056, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.36523008346557617, "tokens": 1890385920, "adamw_lr": 0.0004, "muon/grad_norm": 0.2796265184879303, "adam/grad_norm": 0.2349158376455307, "muon/param_norm": 1950.8743896484375, "muon/update_ratio": 0.01030921470373869, "muon/update_ratio_max": 0.08427929878234863, "adam/param_norm": 700.2538452148438, "adam/update_ratio": 0.0011218483559787273, "adam/update_ratio_max": 0.002837809966877103, "ademamix/cos_fast_slow": 0.07102619856595993, "ademamix/slow_fast_norm_ratio": 0.16069623827934265, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.380859375, "loss_smoothed": 2.2450652812918026, "seq_len": 4096} {"wall": 23946.337, "step": 6420, "loss": 2.221293797095617, "lm_loss": 2.221293797095617, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3445623815059662, "tokens": 1893335040, "adamw_lr": 0.0004, "muon/grad_norm": 0.25398752093315125, "adam/grad_norm": 0.2331966608762741, "muon/param_norm": 1950.9451904296875, "muon/update_ratio": 0.010237759910523891, "muon/update_ratio_max": 0.08206656575202942, "adam/param_norm": 700.2963256835938, "adam/update_ratio": 0.0011290429392829537, "adam/update_ratio_max": 0.0028550149872899055, "ademamix/cos_fast_slow": 0.421455055475235, "ademamix/slow_fast_norm_ratio": 0.1637251228094101, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.380859375, "loss_smoothed": 2.245338726540407, "seq_len": 4096} {"wall": 24151.038, "step": 6430, "loss": 2.259443407257398, "lm_loss": 2.259443407257398, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3318830728530884, "tokens": 1896284160, "adamw_lr": 0.0004, "muon/grad_norm": 0.23592284321784973, "adam/grad_norm": 0.233305886387825, "muon/param_norm": 1951.4256591796875, "muon/update_ratio": 0.010348211042582989, "muon/update_ratio_max": 0.08368594944477081, "adam/param_norm": 700.3384399414062, "adam/update_ratio": 0.001115428633056581, "adam/update_ratio_max": 0.0028197276405990124, "ademamix/cos_fast_slow": 0.37342503666877747, "ademamix/slow_fast_norm_ratio": 0.15630404651165009, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.380859375, "loss_smoothed": 2.24480841755867, "seq_len": 4096} {"wall": 24355.838, "step": 6440, "loss": 2.4173267831405005, "lm_loss": 2.4173267831405005, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.41841521859169006, "tokens": 1899233280, "adamw_lr": 0.0004, "muon/grad_norm": 0.283340185880661, "adam/grad_norm": 0.30730628967285156, "muon/param_norm": 1951.950927734375, "muon/update_ratio": 0.010291858576238155, "muon/update_ratio_max": 0.0859140157699585, "adam/param_norm": 700.3806762695312, "adam/update_ratio": 0.0011016519274562597, "adam/update_ratio_max": 0.002783988369628787, "ademamix/cos_fast_slow": 0.1808496117591858, "ademamix/slow_fast_norm_ratio": 0.14434805512428284, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.380859375, "loss_smoothed": 2.2456188040475054, "seq_len": 4096} {"wall": 24560.529, "step": 6450, "loss": 2.2748230596383414, "lm_loss": 2.2748230596383414, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2584631145000458, "tokens": 1902182400, "adamw_lr": 0.0004, "muon/grad_norm": 0.20403370261192322, "adam/grad_norm": 0.15889745950698853, "muon/param_norm": 1952.3118896484375, "muon/update_ratio": 0.010367008857429028, "muon/update_ratio_max": 0.08535680919885635, "adam/param_norm": 700.421875, "adam/update_ratio": 0.0011126945028081536, "adam/update_ratio_max": 0.002811063313856721, "ademamix/cos_fast_slow": 0.3063642382621765, "ademamix/slow_fast_norm_ratio": 0.17705996334552765, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.380859375, "loss_smoothed": 2.2776488453149795, "seq_len": 4096} {"wall": 24765.292, "step": 6460, "loss": 2.3400511294603348, "lm_loss": 2.3400511294603348, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3492790758609772, "tokens": 1905131520, "adamw_lr": 0.0004, "muon/grad_norm": 0.2606285810470581, "adam/grad_norm": 0.23227344453334808, "muon/param_norm": 1952.712646484375, "muon/update_ratio": 0.010308601893484592, "muon/update_ratio_max": 0.08083640038967133, "adam/param_norm": 700.4639892578125, "adam/update_ratio": 0.0011166456388309598, "adam/update_ratio_max": 0.0028202133253216743, "ademamix/cos_fast_slow": 0.18750087916851044, "ademamix/slow_fast_norm_ratio": 0.16674210131168365, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.380859375, "loss_smoothed": 2.2597472424308456, "seq_len": 4096} {"wall": 24970.281, "step": 6470, "loss": 2.242026920119921, "lm_loss": 2.242026920119921, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.31616947054862976, "tokens": 1908080640, "adamw_lr": 0.0004, "muon/grad_norm": 0.26052895188331604, "adam/grad_norm": 0.178891122341156, "muon/param_norm": 1953.00341796875, "muon/update_ratio": 0.010216101072728634, "muon/update_ratio_max": 0.08226653933525085, "adam/param_norm": 700.5057373046875, "adam/update_ratio": 0.0010973591124638915, "adam/update_ratio_max": 0.0027704350650310516, "ademamix/cos_fast_slow": 0.17947719991207123, "ademamix/slow_fast_norm_ratio": 0.16326504945755005, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.380859375, "loss_smoothed": 2.2461003934343653, "seq_len": 4096} {"wall": 25174.985, "step": 6480, "loss": 2.1949508264660835, "lm_loss": 2.1949508264660835, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2905825674533844, "tokens": 1911029760, "adamw_lr": 0.0004, "muon/grad_norm": 0.21360325813293457, "adam/grad_norm": 0.19739967584609985, "muon/param_norm": 1953.1094970703125, "muon/update_ratio": 0.010379319079220295, "muon/update_ratio_max": 0.08574764430522919, "adam/param_norm": 700.548828125, "adam/update_ratio": 0.001103991991840303, "adam/update_ratio_max": 0.0027863362338393927, "ademamix/cos_fast_slow": 0.25985056161880493, "ademamix/slow_fast_norm_ratio": 0.1689351350069046, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.380859375, "loss_smoothed": 2.219179874161879, "seq_len": 4096} {"wall": 25379.714, "step": 6490, "loss": 2.221061497926712, "lm_loss": 2.221061497926712, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3471120297908783, "tokens": 1913978880, "adamw_lr": 0.0004, "muon/grad_norm": 0.23724134266376495, "adam/grad_norm": 0.2536752223968506, "muon/param_norm": 1953.2586669921875, "muon/update_ratio": 0.0102668646723032, "muon/update_ratio_max": 0.0839681550860405, "adam/param_norm": 700.5905151367188, "adam/update_ratio": 0.0011157215340062976, "adam/update_ratio_max": 0.0028150882571935654, "ademamix/cos_fast_slow": 0.4607994258403778, "ademamix/slow_fast_norm_ratio": 0.16386756300926208, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.380859375, "loss_smoothed": 2.271878810226917, "seq_len": 4096} {"wall": 25587.058, "eval_loss": 1.9228394627571106, "eval_ppl": 6.840353488922119, "step": 6500} {"wall": 25587.059, "step": 6500, "loss": 2.3179061114788055, "lm_loss": 2.3179061114788055, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3365240693092346, "tokens": 1916928000, "adamw_lr": 0.0004, "muon/grad_norm": 0.24711529910564423, "adam/grad_norm": 0.22856071591377258, "muon/param_norm": 1953.066162109375, "muon/update_ratio": 0.010238440707325935, "muon/update_ratio_max": 0.08089371770620346, "adam/param_norm": 700.6329345703125, "adam/update_ratio": 0.0011110452469438314, "adam/update_ratio_max": 0.0028022057376801968, "ademamix/cos_fast_slow": 0.23770250380039215, "ademamix/slow_fast_norm_ratio": 0.17036643624305725, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.380859375, "loss_smoothed": 2.264437239865462, "seq_len": 4096} {"wall": 25862.375, "step": 6510, "loss": 2.2367916256189346, "lm_loss": 2.2367916256189346, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3332391083240509, "tokens": 1919877120, "adamw_lr": 0.0004, "muon/grad_norm": 0.22853243350982666, "adam/grad_norm": 0.24230821430683136, "muon/param_norm": 1953.0570068359375, "muon/update_ratio": 0.010321295820176601, "muon/update_ratio_max": 0.07987526804208755, "adam/param_norm": 700.675048828125, "adam/update_ratio": 0.001117696869187057, "adam/update_ratio_max": 0.0028182929381728172, "ademamix/cos_fast_slow": 0.4929220974445343, "ademamix/slow_fast_norm_ratio": 0.1537931114435196, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.044921875, "loss_smoothed": 2.2347532843550044, "seq_len": 4096} {"wall": 26066.971, "step": 6520, "loss": 2.3312488993008933, "lm_loss": 2.3312488993008933, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2953052222728729, "tokens": 1922826240, "adamw_lr": 0.0004, "muon/grad_norm": 0.2129904180765152, "adam/grad_norm": 0.2045159935951233, "muon/param_norm": 1953.1292724609375, "muon/update_ratio": 0.010384977795183659, "muon/update_ratio_max": 0.08403770625591278, "adam/param_norm": 700.7173461914062, "adam/update_ratio": 0.001101656467653811, "adam/update_ratio_max": 0.0027768018189817667, "ademamix/cos_fast_slow": 0.4925941228866577, "ademamix/slow_fast_norm_ratio": 0.16148161888122559, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.044921875, "loss_smoothed": 2.233817575375239, "seq_len": 4096} {"wall": 26271.399, "step": 6530, "loss": 2.115580673019091, "lm_loss": 2.115580673019091, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3085646629333496, "tokens": 1925775360, "adamw_lr": 0.0004, "muon/grad_norm": 0.2241160273551941, "adam/grad_norm": 0.2127189338207245, "muon/param_norm": 1953.3314208984375, "muon/update_ratio": 0.010339335538446903, "muon/update_ratio_max": 0.08405354619026184, "adam/param_norm": 700.7599487304688, "adam/update_ratio": 0.0011061035329475999, "adam/update_ratio_max": 0.0027870279736816883, "ademamix/cos_fast_slow": 0.24463102221488953, "ademamix/slow_fast_norm_ratio": 0.1553502380847931, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.044921875, "loss_smoothed": 2.21441923280557, "seq_len": 4096} {"wall": 26476.237, "step": 6540, "loss": 2.1069567451874414, "lm_loss": 2.1069567451874414, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.30606284737586975, "tokens": 1928724480, "adamw_lr": 0.0004, "muon/grad_norm": 0.22203214466571808, "adam/grad_norm": 0.21100600063800812, "muon/param_norm": 1951.2532958984375, "muon/update_ratio": 0.009848175570368767, "muon/update_ratio_max": 0.08469591289758682, "adam/param_norm": 700.8011474609375, "adam/update_ratio": 0.0011095417430624366, "adam/update_ratio_max": 0.0027944259345531464, "ademamix/cos_fast_slow": 0.2585083842277527, "ademamix/slow_fast_norm_ratio": 0.1616944670677185, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.044921875, "loss_smoothed": 2.214960476756096, "seq_len": 4096} {"wall": 26680.933, "step": 6550, "loss": 2.2154037108023963, "lm_loss": 2.2154037108023963, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.34615358710289, "tokens": 1931673600, "adamw_lr": 0.0004, "muon/grad_norm": 0.27107685804367065, "adam/grad_norm": 0.21535992622375488, "muon/param_norm": 1950.3023681640625, "muon/update_ratio": 0.010052620433270931, "muon/update_ratio_max": 0.08421050757169724, "adam/param_norm": 700.8413696289062, "adam/update_ratio": 0.0010999938240274787, "adam/update_ratio_max": 0.002769554266706109, "ademamix/cos_fast_slow": 0.5106268525123596, "ademamix/slow_fast_norm_ratio": 0.16530658304691315, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.044921875, "loss_smoothed": 2.201563313603401, "seq_len": 4096} {"wall": 26885.544, "step": 6560, "loss": 2.3403042256832123, "lm_loss": 2.3403042256832123, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.34471938014030457, "tokens": 1934622720, "adamw_lr": 0.0004, "muon/grad_norm": 0.2455650269985199, "adam/grad_norm": 0.24150420725345612, "muon/param_norm": 1949.6156005859375, "muon/update_ratio": 0.010199197567999363, "muon/update_ratio_max": 0.0824364647269249, "adam/param_norm": 700.8839111328125, "adam/update_ratio": 0.0011490981560200453, "adam/update_ratio_max": 0.0028925701044499874, "ademamix/cos_fast_slow": 0.468128502368927, "ademamix/slow_fast_norm_ratio": 0.14640094339847565, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.044921875, "loss_smoothed": 2.2301745608448984, "seq_len": 4096} {"wall": 27090.275, "step": 6570, "loss": 2.0598992655674615, "lm_loss": 2.0598992655674615, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.37936311960220337, "tokens": 1937571840, "adamw_lr": 0.0004, "muon/grad_norm": 0.28647175431251526, "adam/grad_norm": 0.24854272603988647, "muon/param_norm": 1949.4036865234375, "muon/update_ratio": 0.010315202176570892, "muon/update_ratio_max": 0.08011268079280853, "adam/param_norm": 700.9241333007812, "adam/update_ratio": 0.0010996436467394233, "adam/update_ratio_max": 0.002767023863270879, "ademamix/cos_fast_slow": 0.5866585373878479, "ademamix/slow_fast_norm_ratio": 0.1455063670873642, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.044921875, "loss_smoothed": 2.228823459148407, "seq_len": 4096} {"wall": 27294.955, "step": 6580, "loss": 2.2811574985583625, "lm_loss": 2.2811574985583625, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3283892869949341, "tokens": 1940520960, "adamw_lr": 0.0004, "muon/grad_norm": 0.24387918412685394, "adam/grad_norm": 0.22019745409488678, "muon/param_norm": 1949.1927490234375, "muon/update_ratio": 0.010328586213290691, "muon/update_ratio_max": 0.08000657707452774, "adam/param_norm": 700.9642944335938, "adam/update_ratio": 0.001137384446337819, "adam/update_ratio_max": 0.0028612667229026556, "ademamix/cos_fast_slow": 0.3785156309604645, "ademamix/slow_fast_norm_ratio": 0.1543944776058197, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.044921875, "loss_smoothed": 2.259411644935608, "seq_len": 4096} {"wall": 27499.68, "step": 6590, "loss": 2.2362391650676727, "lm_loss": 2.2362391650676727, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3025195300579071, "tokens": 1943470080, "adamw_lr": 0.0004, "muon/grad_norm": 0.22113439440727234, "adam/grad_norm": 0.20608927309513092, "muon/param_norm": 1949.21142578125, "muon/update_ratio": 0.010309556499123573, "muon/update_ratio_max": 0.08014589548110962, "adam/param_norm": 701.0049438476562, "adam/update_ratio": 0.001112494501285255, "adam/update_ratio_max": 0.0027977151330560446, "ademamix/cos_fast_slow": 0.5135946869850159, "ademamix/slow_fast_norm_ratio": 0.1596246063709259, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.044921875, "loss_smoothed": 2.230343161026637, "seq_len": 4096} {"wall": 27706.68, "eval_loss": 1.9187065660953522, "eval_ppl": 6.812141418457031, "step": 6600} {"wall": 27706.682, "step": 6600, "loss": 2.307321603099505, "lm_loss": 2.307321603099505, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.289691299200058, "tokens": 1946419200, "adamw_lr": 0.0004, "muon/grad_norm": 0.20301194489002228, "adam/grad_norm": 0.20692376792430878, "muon/param_norm": 1949.4393310546875, "muon/update_ratio": 0.010361744090914726, "muon/update_ratio_max": 0.07667863368988037, "adam/param_norm": 701.0458984375, "adam/update_ratio": 0.0011172100203111768, "adam/update_ratio_max": 0.002808681922033429, "ademamix/cos_fast_slow": 0.533511757850647, "ademamix/slow_fast_norm_ratio": 0.14246781170368195, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.044921875, "loss_smoothed": 2.248861824721098, "seq_len": 4096} {"wall": 27982.513, "step": 6610, "loss": 2.152559700111548, "lm_loss": 2.152559700111548, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2862686216831207, "tokens": 1949368320, "adamw_lr": 0.0004, "muon/grad_norm": 0.20701155066490173, "adam/grad_norm": 0.19802244007587433, "muon/param_norm": 1949.7264404296875, "muon/update_ratio": 0.010369428433477879, "muon/update_ratio_max": 0.07812632620334625, "adam/param_norm": 701.088623046875, "adam/update_ratio": 0.0011191640514880419, "adam/update_ratio_max": 0.0028127715922892094, "ademamix/cos_fast_slow": 0.5450750589370728, "ademamix/slow_fast_norm_ratio": 0.15721578896045685, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.256189309805632, "seq_len": 4096} {"wall": 28187.147, "step": 6620, "loss": 2.0938139458497367, "lm_loss": 2.0938139458497367, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3393917679786682, "tokens": 1952317440, "adamw_lr": 0.0004, "muon/grad_norm": 0.27658528089523315, "adam/grad_norm": 0.19673678278923035, "muon/param_norm": 1949.9248046875, "muon/update_ratio": 0.010276627726852894, "muon/update_ratio_max": 0.08180832862854004, "adam/param_norm": 701.129638671875, "adam/update_ratio": 0.0011108143953606486, "adam/update_ratio_max": 0.0027907979674637318, "ademamix/cos_fast_slow": 0.3869442343711853, "ademamix/slow_fast_norm_ratio": 0.16413810849189758, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.2454533775647483, "seq_len": 4096} {"wall": 28391.827, "step": 6630, "loss": 2.248965968688329, "lm_loss": 2.248965968688329, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3115096390247345, "tokens": 1955266560, "adamw_lr": 0.0004, "muon/grad_norm": 0.24987736344337463, "adam/grad_norm": 0.18570576608181, "muon/param_norm": 1950.2322998046875, "muon/update_ratio": 0.010325368493795395, "muon/update_ratio_max": 0.08428924530744553, "adam/param_norm": 701.1724853515625, "adam/update_ratio": 0.001108647556975484, "adam/update_ratio_max": 0.0027845401782542467, "ademamix/cos_fast_slow": 0.2305898815393448, "ademamix/slow_fast_norm_ratio": 0.15488474071025848, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.2445970286925636, "seq_len": 4096} {"wall": 28596.712, "step": 6640, "loss": 2.265627697110176, "lm_loss": 2.265627697110176, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2919914126396179, "tokens": 1958215680, "adamw_lr": 0.0004, "muon/grad_norm": 0.25901249051094055, "adam/grad_norm": 0.1347082406282425, "muon/param_norm": 1950.3609619140625, "muon/update_ratio": 0.010281355120241642, "muon/update_ratio_max": 0.08625113219022751, "adam/param_norm": 701.2138671875, "adam/update_ratio": 0.0011140728602185845, "adam/update_ratio_max": 0.002797226421535015, "ademamix/cos_fast_slow": 0.27097660303115845, "ademamix/slow_fast_norm_ratio": 0.1640031337738037, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.272806216776371, "seq_len": 4096} {"wall": 28801.319, "step": 6650, "loss": 2.1222305546204248, "lm_loss": 2.1222305546204248, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.32393738627433777, "tokens": 1961164800, "adamw_lr": 0.0004, "muon/grad_norm": 0.27404892444610596, "adam/grad_norm": 0.17216980457305908, "muon/param_norm": 1950.4161376953125, "muon/update_ratio": 0.010307654738426208, "muon/update_ratio_max": 0.08606132864952087, "adam/param_norm": 701.2543334960938, "adam/update_ratio": 0.001114494283683598, "adam/update_ratio_max": 0.002797492314130068, "ademamix/cos_fast_slow": 0.5030252933502197, "ademamix/slow_fast_norm_ratio": 0.15962642431259155, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.262218136092027, "seq_len": 4096} {"wall": 29005.956, "step": 6660, "loss": 2.3100067377090454, "lm_loss": 2.3100067377090454, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.32582712173461914, "tokens": 1964113920, "adamw_lr": 0.0004, "muon/grad_norm": 0.2454175353050232, "adam/grad_norm": 0.21421903371810913, "muon/param_norm": 1950.026123046875, "muon/update_ratio": 0.010270917788147926, "muon/update_ratio_max": 0.08764102309942245, "adam/param_norm": 701.2963256835938, "adam/update_ratio": 0.0011104029836133122, "adam/update_ratio_max": 0.002786277560517192, "ademamix/cos_fast_slow": 0.14631114900112152, "ademamix/slow_fast_norm_ratio": 0.1452324241399765, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.214663826674223, "seq_len": 4096} {"wall": 29210.889, "step": 6670, "loss": 2.306340590119362, "lm_loss": 2.306340590119362, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.31416353583335876, "tokens": 1967063040, "adamw_lr": 0.0004, "muon/grad_norm": 0.24620312452316284, "adam/grad_norm": 0.19480010867118835, "muon/param_norm": 1949.841064453125, "muon/update_ratio": 0.010264829732477665, "muon/update_ratio_max": 0.0850042924284935, "adam/param_norm": 701.3382568359375, "adam/update_ratio": 0.0011082662967965007, "adam/update_ratio_max": 0.0027800730895251036, "ademamix/cos_fast_slow": 0.3146420121192932, "ademamix/slow_fast_norm_ratio": 0.16414007544517517, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.259697379668554, "seq_len": 4096} {"wall": 29415.553, "step": 6680, "loss": 2.283428226908048, "lm_loss": 2.283428226908048, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.37605050206184387, "tokens": 1970012160, "adamw_lr": 0.0004, "muon/grad_norm": 0.2906859815120697, "adam/grad_norm": 0.23898735642433167, "muon/param_norm": 1949.62109375, "muon/update_ratio": 0.010289969854056835, "muon/update_ratio_max": 0.08646565675735474, "adam/param_norm": 701.3810424804688, "adam/update_ratio": 0.00113700854126364, "adam/update_ratio_max": 0.002851297380402684, "ademamix/cos_fast_slow": 0.5534237623214722, "ademamix/slow_fast_norm_ratio": 0.14021123945713043, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.237253954509894, "seq_len": 4096} {"wall": 29620.207, "step": 6690, "loss": 2.2340379556020102, "lm_loss": 2.2340379556020102, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.356855571269989, "tokens": 1972961280, "adamw_lr": 0.0004, "muon/grad_norm": 0.2884644567966461, "adam/grad_norm": 0.20991672575473785, "muon/param_norm": 1949.25341796875, "muon/update_ratio": 0.010245623998343945, "muon/update_ratio_max": 0.08161525428295135, "adam/param_norm": 701.4234619140625, "adam/update_ratio": 0.0011171414516866207, "adam/update_ratio_max": 0.002800246700644493, "ademamix/cos_fast_slow": 0.39416277408599854, "ademamix/slow_fast_norm_ratio": 0.14628858864307404, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.248449696103732, "seq_len": 4096} {"wall": 29827.406, "eval_loss": 1.9155887216329575, "eval_ppl": 6.790935516357422, "step": 6700} {"wall": 29827.408, "step": 6700, "loss": 2.1664728273948035, "lm_loss": 2.1664728273948035, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.27146103978157043, "tokens": 1975910400, "adamw_lr": 0.0004, "muon/grad_norm": 0.2237558215856552, "adam/grad_norm": 0.15373089909553528, "muon/param_norm": 1949.3021240234375, "muon/update_ratio": 0.010332305915653706, "muon/update_ratio_max": 0.08008655160665512, "adam/param_norm": 701.4664306640625, "adam/update_ratio": 0.0011076590744778514, "adam/update_ratio_max": 0.002775833010673523, "ademamix/cos_fast_slow": 0.4004875719547272, "ademamix/slow_fast_norm_ratio": 0.1606684774160385, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.2110290532310803, "seq_len": 4096} {"wall": 30106.255, "step": 6710, "loss": 2.371864154934883, "lm_loss": 2.371864154934883, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3766610026359558, "tokens": 1978859520, "adamw_lr": 0.0004, "muon/grad_norm": 0.27381059527397156, "adam/grad_norm": 0.25896817445755005, "muon/param_norm": 1949.322998046875, "muon/update_ratio": 0.01026150956749916, "muon/update_ratio_max": 0.08326840400695801, "adam/param_norm": 701.5094604492188, "adam/update_ratio": 0.0011127280304208398, "adam/update_ratio_max": 0.002787669887766242, "ademamix/cos_fast_slow": 0.6120182871818542, "ademamix/slow_fast_norm_ratio": 0.1402052938938141, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 89.751953125, "loss_smoothed": 2.2316236048936844, "seq_len": 4096} {"wall": 30310.795, "step": 6720, "loss": 2.4031543831030526, "lm_loss": 2.4031543831030526, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3174404203891754, "tokens": 1981808640, "adamw_lr": 0.0004, "muon/grad_norm": 0.2461768388748169, "adam/grad_norm": 0.2007117122411728, "muon/param_norm": 1949.226318359375, "muon/update_ratio": 0.010261580348014832, "muon/update_ratio_max": 0.08149789273738861, "adam/param_norm": 701.5518188476562, "adam/update_ratio": 0.0011170272482559085, "adam/update_ratio_max": 0.0027975148987025023, "ademamix/cos_fast_slow": 0.47888869047164917, "ademamix/slow_fast_norm_ratio": 0.15497104823589325, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 89.751953125, "loss_smoothed": 2.2522437304258345, "seq_len": 4096} {"wall": 30515.358, "step": 6730, "loss": 2.330992599328359, "lm_loss": 2.330992599328359, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3949853181838989, "tokens": 1984757760, "adamw_lr": 0.0004, "muon/grad_norm": 0.27945375442504883, "adam/grad_norm": 0.27900904417037964, "muon/param_norm": 1949.1702880859375, "muon/update_ratio": 0.010281480848789215, "muon/update_ratio_max": 0.08198480308055878, "adam/param_norm": 701.5945434570312, "adam/update_ratio": 0.0011090505868196487, "adam/update_ratio_max": 0.00277654523961246, "ademamix/cos_fast_slow": 0.1412503868341446, "ademamix/slow_fast_norm_ratio": 0.1391686648130417, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 89.751953125, "loss_smoothed": 2.2636740446090697, "seq_len": 4096} {"wall": 30719.912, "step": 6740, "loss": 2.2686607440312705, "lm_loss": 2.2686607440312705, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.27583634853363037, "tokens": 1987706880, "adamw_lr": 0.0004, "muon/grad_norm": 0.19492340087890625, "adam/grad_norm": 0.1950884908437729, "muon/param_norm": 1949.09033203125, "muon/update_ratio": 0.01039155200123787, "muon/update_ratio_max": 0.08202744275331497, "adam/param_norm": 701.63623046875, "adam/update_ratio": 0.0011025716084986925, "adam/update_ratio_max": 0.002759371418505907, "ademamix/cos_fast_slow": 0.2852190136909485, "ademamix/slow_fast_norm_ratio": 0.1456570029258728, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 89.751953125, "loss_smoothed": 2.2017784833908083, "seq_len": 4096} {"wall": 30924.84, "step": 6750, "loss": 2.338958660761515, "lm_loss": 2.338958660761515, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3223281800746918, "tokens": 1990656000, "adamw_lr": 0.0004, "muon/grad_norm": 0.24088114500045776, "adam/grad_norm": 0.2140970379114151, "muon/param_norm": 1949.1810302734375, "muon/update_ratio": 0.010311265476047993, "muon/update_ratio_max": 0.08134087920188904, "adam/param_norm": 701.6786499023438, "adam/update_ratio": 0.001110972836613655, "adam/update_ratio_max": 0.002779576927423477, "ademamix/cos_fast_slow": 0.527484118938446, "ademamix/slow_fast_norm_ratio": 0.1520114541053772, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 89.751953125, "loss_smoothed": 2.2548789526025454, "seq_len": 4096} {"wall": 31129.403, "step": 6760, "loss": 2.0790655439098678, "lm_loss": 2.0790655439098678, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2700743079185486, "tokens": 1993605120, "adamw_lr": 0.0004, "muon/grad_norm": 0.20555780827999115, "adam/grad_norm": 0.17544572055339813, "muon/param_norm": 1949.281005859375, "muon/update_ratio": 0.010363091714680195, "muon/update_ratio_max": 0.08193741738796234, "adam/param_norm": 701.7203369140625, "adam/update_ratio": 0.0011029880261048675, "adam/update_ratio_max": 0.002758709480985999, "ademamix/cos_fast_slow": 0.17699815332889557, "ademamix/slow_fast_norm_ratio": 0.15572209656238556, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 89.751953125, "loss_smoothed": 2.1952039482692878, "seq_len": 4096} {"wall": 31333.966, "step": 6770, "loss": 2.3631841093301773, "lm_loss": 2.3631841093301773, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3332101106643677, "tokens": 1996554240, "adamw_lr": 0.0004, "muon/grad_norm": 0.24941571056842804, "adam/grad_norm": 0.22104188799858093, "muon/param_norm": 1949.2430419921875, "muon/update_ratio": 0.010333023965358734, "muon/update_ratio_max": 0.07869474589824677, "adam/param_norm": 701.761962890625, "adam/update_ratio": 0.0011097689857706428, "adam/update_ratio_max": 0.0027749279979616404, "ademamix/cos_fast_slow": 0.3908745050430298, "ademamix/slow_fast_norm_ratio": 0.14196375012397766, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 89.751953125, "loss_smoothed": 2.2806566705306373, "seq_len": 4096} {"wall": 31538.716, "step": 6780, "loss": 2.1283636341492334, "lm_loss": 2.1283636341492334, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.347936749458313, "tokens": 1999503360, "adamw_lr": 0.0004, "muon/grad_norm": 0.27485305070877075, "adam/grad_norm": 0.21319136023521423, "muon/param_norm": 1948.989990234375, "muon/update_ratio": 0.010294186882674694, "muon/update_ratio_max": 0.08372591435909271, "adam/param_norm": 701.804931640625, "adam/update_ratio": 0.0011060674441978335, "adam/update_ratio_max": 0.002764587290585041, "ademamix/cos_fast_slow": 0.19210349023342133, "ademamix/slow_fast_norm_ratio": 0.14753973484039307, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 89.751953125, "loss_smoothed": 2.224427942434947, "seq_len": 4096} {"wall": 31743.295, "step": 6790, "loss": 2.1515299628178277, "lm_loss": 2.1515299628178277, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.35422876477241516, "tokens": 2002452480, "adamw_lr": 0.0004, "muon/grad_norm": 0.22638842463493347, "adam/grad_norm": 0.2730989158153534, "muon/param_norm": 1948.8507080078125, "muon/update_ratio": 0.01031632162630558, "muon/update_ratio_max": 0.08447130024433136, "adam/param_norm": 701.8458862304688, "adam/update_ratio": 0.0010990748414769769, "adam/update_ratio_max": 0.002746270038187504, "ademamix/cos_fast_slow": 0.30580317974090576, "ademamix/slow_fast_norm_ratio": 0.1446378082036972, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 89.751953125, "loss_smoothed": 2.206200750668844, "seq_len": 4096} {"wall": 31950.269, "eval_loss": 1.913491502404213, "eval_ppl": 6.776708126068115, "step": 6800} {"wall": 31950.271, "step": 6800, "loss": 2.1396388908227286, "lm_loss": 2.1396388908227286, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.33098116517066956, "tokens": 2005401600, "adamw_lr": 0.0004, "muon/grad_norm": 0.2641364634037018, "adam/grad_norm": 0.19937513768672943, "muon/param_norm": 1948.7379150390625, "muon/update_ratio": 0.010270449332892895, "muon/update_ratio_max": 0.08361620455980301, "adam/param_norm": 701.8875122070312, "adam/update_ratio": 0.0011046964209526777, "adam/update_ratio_max": 0.0027593837585300207, "ademamix/cos_fast_slow": 0.47565963864326477, "ademamix/slow_fast_norm_ratio": 0.15884001553058624, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 89.751953125, "loss_smoothed": 2.217946869631608, "seq_len": 4096} {"wall": 32228.615, "step": 6810, "loss": 2.1628893117109933, "lm_loss": 2.1628893117109933, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.28050073981285095, "tokens": 2008350720, "adamw_lr": 0.0004, "muon/grad_norm": 0.229431614279747, "adam/grad_norm": 0.1609959900379181, "muon/param_norm": 1948.4776611328125, "muon/update_ratio": 0.010331143625080585, "muon/update_ratio_max": 0.08270955830812454, "adam/param_norm": 701.9290771484375, "adam/update_ratio": 0.0011082985438406467, "adam/update_ratio_max": 0.0027675526216626167, "ademamix/cos_fast_slow": 0.3594348132610321, "ademamix/slow_fast_norm_ratio": 0.15841339528560638, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.341796875, "loss_smoothed": 2.270363899072011, "seq_len": 4096} {"wall": 32433.376, "step": 6820, "loss": 2.17500572403272, "lm_loss": 2.17500572403272, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.39754965901374817, "tokens": 2011299840, "adamw_lr": 0.0004, "muon/grad_norm": 0.3131704330444336, "adam/grad_norm": 0.2445729523897171, "muon/param_norm": 1948.3155517578125, "muon/update_ratio": 0.010218004696071148, "muon/update_ratio_max": 0.08611337095499039, "adam/param_norm": 701.971923828125, "adam/update_ratio": 0.0011123217409476638, "adam/update_ratio_max": 0.002776786684989929, "ademamix/cos_fast_slow": 0.18154190480709076, "ademamix/slow_fast_norm_ratio": 0.13580749928951263, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.341796875, "loss_smoothed": 2.2155225316683453, "seq_len": 4096} {"wall": 32638.131, "step": 6830, "loss": 2.376835823059082, "lm_loss": 2.376835823059082, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.26890480518341064, "tokens": 2014248960, "adamw_lr": 0.0004, "muon/grad_norm": 0.22715161740779877, "adam/grad_norm": 0.14424480497837067, "muon/param_norm": 1948.212646484375, "muon/update_ratio": 0.010327733121812344, "muon/update_ratio_max": 0.08505846560001373, "adam/param_norm": 702.01171875, "adam/update_ratio": 0.0011156986001878977, "adam/update_ratio_max": 0.0027843895368278027, "ademamix/cos_fast_slow": 0.24335148930549622, "ademamix/slow_fast_norm_ratio": 0.1582682877779007, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.341796875, "loss_smoothed": 2.2220754355192183, "seq_len": 4096} {"wall": 32842.882, "step": 6840, "loss": 2.3290578524271646, "lm_loss": 2.3290578524271646, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3783022165298462, "tokens": 2017198080, "adamw_lr": 0.0004, "muon/grad_norm": 0.24718008935451508, "adam/grad_norm": 0.2866678237915039, "muon/param_norm": 1947.8238525390625, "muon/update_ratio": 0.01027605403214693, "muon/update_ratio_max": 0.0823146402835846, "adam/param_norm": 702.0526733398438, "adam/update_ratio": 0.0011042352998629212, "adam/update_ratio_max": 0.002754753455519676, "ademamix/cos_fast_slow": 0.5371944904327393, "ademamix/slow_fast_norm_ratio": 0.13627585768699646, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.341796875, "loss_smoothed": 2.2361458743611973, "seq_len": 4096} {"wall": 33047.624, "step": 6850, "loss": 2.3116477032502494, "lm_loss": 2.3116477032502494, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.28293630480766296, "tokens": 2020147200, "adamw_lr": 0.0004, "muon/grad_norm": 0.2334618717432022, "adam/grad_norm": 0.15955384075641632, "muon/param_norm": 1947.183837890625, "muon/update_ratio": 0.010262615978717804, "muon/update_ratio_max": 0.07992710173130035, "adam/param_norm": 702.0950927734375, "adam/update_ratio": 0.0011099227704107761, "adam/update_ratio_max": 0.0027681428473442793, "ademamix/cos_fast_slow": 0.3929647207260132, "ademamix/slow_fast_norm_ratio": 0.1542685478925705, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.341796875, "loss_smoothed": 2.2481398741404215, "seq_len": 4096} {"wall": 33252.387, "step": 6860, "loss": 2.2220066264271736, "lm_loss": 2.2220066264271736, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.39483603835105896, "tokens": 2023096320, "adamw_lr": 0.0004, "muon/grad_norm": 0.2797917425632477, "adam/grad_norm": 0.2785637378692627, "muon/param_norm": 1947.1192626953125, "muon/update_ratio": 0.010347153060138226, "muon/update_ratio_max": 0.07395171374082565, "adam/param_norm": 702.1375732421875, "adam/update_ratio": 0.0011202814057469368, "adam/update_ratio_max": 0.002793220104649663, "ademamix/cos_fast_slow": 0.5855908393859863, "ademamix/slow_fast_norm_ratio": 0.13468147814273834, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.341796875, "loss_smoothed": 2.229454872260491, "seq_len": 4096} {"wall": 33457.126, "step": 6870, "loss": 2.3460857023795447, "lm_loss": 2.3460857023795447, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3174861669540405, "tokens": 2026045440, "adamw_lr": 0.0004, "muon/grad_norm": 0.2726294696331024, "adam/grad_norm": 0.16315330564975739, "muon/param_norm": 1947.2916259765625, "muon/update_ratio": 0.010296939872205257, "muon/update_ratio_max": 0.07538900524377823, "adam/param_norm": 702.1780395507812, "adam/update_ratio": 0.001122707617469132, "adam/update_ratio_max": 0.002798423869535327, "ademamix/cos_fast_slow": 0.19087354838848114, "ademamix/slow_fast_norm_ratio": 0.14967676997184753, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.341796875, "loss_smoothed": 2.248389422893524, "seq_len": 4096} {"wall": 33662.089, "step": 6880, "loss": 2.156269093354543, "lm_loss": 2.156269093354543, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3769347667694092, "tokens": 2028994560, "adamw_lr": 0.0004, "muon/grad_norm": 0.26921316981315613, "adam/grad_norm": 0.2632756233215332, "muon/param_norm": 1947.1705322265625, "muon/update_ratio": 0.01030984241515398, "muon/update_ratio_max": 0.07867112755775452, "adam/param_norm": 702.2191772460938, "adam/update_ratio": 0.0011104810982942581, "adam/update_ratio_max": 0.0027669561095535755, "ademamix/cos_fast_slow": 0.5867698788642883, "ademamix/slow_fast_norm_ratio": 0.14377416670322418, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.341796875, "loss_smoothed": 2.2374528060356775, "seq_len": 4096} {"wall": 33866.595, "step": 6890, "loss": 2.225855549176534, "lm_loss": 2.225855549176534, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2527085244655609, "tokens": 2031943680, "adamw_lr": 0.0004, "muon/grad_norm": 0.2005714774131775, "adam/grad_norm": 0.15408176183700562, "muon/param_norm": 1944.9931640625, "muon/update_ratio": 0.009759166277945042, "muon/update_ratio_max": 0.07884804904460907, "adam/param_norm": 702.2606811523438, "adam/update_ratio": 0.0011195512488484383, "adam/update_ratio_max": 0.002787061035633087, "ademamix/cos_fast_slow": 0.41479742527008057, "ademamix/slow_fast_norm_ratio": 0.15238215029239655, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.341796875, "loss_smoothed": 2.2253022161622846, "seq_len": 4096} {"wall": 34073.773, "eval_loss": 1.920089140534401, "eval_ppl": 6.821566581726074, "step": 6900} {"wall": 34073.774, "step": 6900, "loss": 2.2206980933745704, "lm_loss": 2.2206980933745704, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.358114629983902, "tokens": 2034892800, "adamw_lr": 0.0004, "muon/grad_norm": 0.28810256719589233, "adam/grad_norm": 0.21327680349349976, "muon/param_norm": 1943.7581787109375, "muon/update_ratio": 0.010063649155199528, "muon/update_ratio_max": 0.07929740101099014, "adam/param_norm": 702.301513671875, "adam/update_ratio": 0.0011251905234530568, "adam/update_ratio_max": 0.002800649031996727, "ademamix/cos_fast_slow": 0.1435338705778122, "ademamix/slow_fast_norm_ratio": 0.13675746321678162, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.341796875, "loss_smoothed": 2.244120192031066, "seq_len": 4096} {"wall": 34355.912, "step": 6910, "loss": 2.197949687639872, "lm_loss": 2.197949687639872, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2986474335193634, "tokens": 2037841920, "adamw_lr": 0.0004, "muon/grad_norm": 0.2408483326435089, "adam/grad_norm": 0.17630897462368011, "muon/param_norm": 1943.2100830078125, "muon/update_ratio": 0.010269303806126118, "muon/update_ratio_max": 0.08036928623914719, "adam/param_norm": 702.3391723632812, "adam/update_ratio": 0.0010989048751071095, "adam/update_ratio_max": 0.0027347065042704344, "ademamix/cos_fast_slow": 0.3736755847930908, "ademamix/slow_fast_norm_ratio": 0.15601548552513123, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 89.126953125, "loss_smoothed": 2.199149176726739, "seq_len": 4096} {"wall": 34560.975, "step": 6920, "loss": 2.316284457842509, "lm_loss": 2.316284457842509, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3273453712463379, "tokens": 2040791040, "adamw_lr": 0.0004, "muon/grad_norm": 0.2559764087200165, "adam/grad_norm": 0.2038538157939911, "muon/param_norm": 1943.1573486328125, "muon/update_ratio": 0.010303116403520107, "muon/update_ratio_max": 0.08037768304347992, "adam/param_norm": 702.379150390625, "adam/update_ratio": 0.0011057067895308137, "adam/update_ratio_max": 0.002750984625890851, "ademamix/cos_fast_slow": 0.4108177125453949, "ademamix/slow_fast_norm_ratio": 0.15577292442321777, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 89.126953125, "loss_smoothed": 2.2050613686442375, "seq_len": 4096} {"wall": 34765.741, "step": 6930, "loss": 2.2565966099500656, "lm_loss": 2.2565966099500656, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3174152374267578, "tokens": 2043740160, "adamw_lr": 0.0004, "muon/grad_norm": 0.22502706944942474, "adam/grad_norm": 0.224061518907547, "muon/param_norm": 1943.275634765625, "muon/update_ratio": 0.010417424142360687, "muon/update_ratio_max": 0.08446099609136581, "adam/param_norm": 702.4188842773438, "adam/update_ratio": 0.001116134226322174, "adam/update_ratio_max": 0.0027761636301875114, "ademamix/cos_fast_slow": 0.21084360778331757, "ademamix/slow_fast_norm_ratio": 0.14453481137752533, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 89.126953125, "loss_smoothed": 2.2617536172270776, "seq_len": 4096} {"wall": 34970.499, "step": 6940, "loss": 2.324753443400065, "lm_loss": 2.324753443400065, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3121359050273895, "tokens": 2046689280, "adamw_lr": 0.0004, "muon/grad_norm": 0.21589724719524384, "adam/grad_norm": 0.22574874758720398, "muon/param_norm": 1943.6585693359375, "muon/update_ratio": 0.010393986478447914, "muon/update_ratio_max": 0.08356275409460068, "adam/param_norm": 702.4591064453125, "adam/update_ratio": 0.0011006523855030537, "adam/update_ratio_max": 0.0027366771828383207, "ademamix/cos_fast_slow": 0.13609398901462555, "ademamix/slow_fast_norm_ratio": 0.1438298225402832, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 89.126953125, "loss_smoothed": 2.2073157037297886, "seq_len": 4096} {"wall": 35175.58, "step": 6950, "loss": 2.1838913410902023, "lm_loss": 2.1838913410902023, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.43538907170295715, "tokens": 2049638400, "adamw_lr": 0.0004, "muon/grad_norm": 0.3120400309562683, "adam/grad_norm": 0.3030409812927246, "muon/param_norm": 1943.994384765625, "muon/update_ratio": 0.010280564427375793, "muon/update_ratio_max": 0.08322355151176453, "adam/param_norm": 702.499267578125, "adam/update_ratio": 0.0011024519335478544, "adam/update_ratio_max": 0.0027409435715526342, "ademamix/cos_fast_slow": 0.5729215741157532, "ademamix/slow_fast_norm_ratio": 0.13427890837192535, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 89.126953125, "loss_smoothed": 2.202219717701276, "seq_len": 4096} {"wall": 35380.416, "step": 6960, "loss": 2.272816389799118, "lm_loss": 2.272816389799118, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.29504069685935974, "tokens": 2052587520, "adamw_lr": 0.0004, "muon/grad_norm": 0.2556132674217224, "adam/grad_norm": 0.14709776639938354, "muon/param_norm": 1944.4072265625, "muon/update_ratio": 0.010339300148189068, "muon/update_ratio_max": 0.08563563972711563, "adam/param_norm": 702.5396118164062, "adam/update_ratio": 0.0010998136131092906, "adam/update_ratio_max": 0.0027336168568581343, "ademamix/cos_fast_slow": 0.40882742404937744, "ademamix/slow_fast_norm_ratio": 0.15540175139904022, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 89.126953125, "loss_smoothed": 2.222871380050977, "seq_len": 4096} {"wall": 35585.199, "step": 6970, "loss": 2.248067999879519, "lm_loss": 2.248067999879519, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.4087880849838257, "tokens": 2055536640, "adamw_lr": 0.0004, "muon/grad_norm": 0.32983022928237915, "adam/grad_norm": 0.24183671176433563, "muon/param_norm": 1944.527099609375, "muon/update_ratio": 0.010262854397296906, "muon/update_ratio_max": 0.08738169074058533, "adam/param_norm": 702.5807495117188, "adam/update_ratio": 0.0011089947074651718, "adam/update_ratio_max": 0.0027555376291275024, "ademamix/cos_fast_slow": 0.08036869764328003, "ademamix/slow_fast_norm_ratio": 0.13230542838573456, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 89.126953125, "loss_smoothed": 2.19844280531009, "seq_len": 4096} {"wall": 35789.973, "step": 6980, "loss": 2.1094255298376083, "lm_loss": 2.1094255298376083, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.29899856448173523, "tokens": 2058485760, "adamw_lr": 0.0004, "muon/grad_norm": 0.23347611725330353, "adam/grad_norm": 0.18628057837486267, "muon/param_norm": 1944.4857177734375, "muon/update_ratio": 0.010353140532970428, "muon/update_ratio_max": 0.08743353188037872, "adam/param_norm": 702.6190185546875, "adam/update_ratio": 0.0011199393775314093, "adam/update_ratio_max": 0.0027821045368909836, "ademamix/cos_fast_slow": 0.23879878222942352, "ademamix/slow_fast_norm_ratio": 0.15172715485095978, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 89.126953125, "loss_smoothed": 2.2420775453249613, "seq_len": 4096} {"wall": 35994.758, "step": 6990, "loss": 2.1537207762400308, "lm_loss": 2.1537207762400308, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.5124521851539612, "tokens": 2061434880, "adamw_lr": 0.0004, "muon/grad_norm": 0.4269431233406067, "adam/grad_norm": 0.2839125692844391, "muon/param_norm": 1944.5789794921875, "muon/update_ratio": 0.010289720259606838, "muon/update_ratio_max": 0.08006539940834045, "adam/param_norm": 702.6586303710938, "adam/update_ratio": 0.0011068473104387522, "adam/update_ratio_max": 0.002748787635937333, "ademamix/cos_fast_slow": 0.07174038887023926, "ademamix/slow_fast_norm_ratio": 0.12179861217737198, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 89.126953125, "loss_smoothed": 2.2688795069853467, "seq_len": 4096} {"wall": 36201.883, "eval_loss": 1.9132684171199799, "eval_ppl": 6.7751970291137695, "step": 7000} {"wall": 36201.885, "step": 7000, "loss": 2.258311246832212, "lm_loss": 2.258311246832212, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3235454559326172, "tokens": 2064384000, "adamw_lr": 0.0004, "muon/grad_norm": 0.2525714039802551, "adam/grad_norm": 0.20193210244178772, "muon/param_norm": 1944.685302734375, "muon/update_ratio": 0.010313635692000389, "muon/update_ratio_max": 0.082757368683815, "adam/param_norm": 702.69921875, "adam/update_ratio": 0.0011325891828164458, "adam/update_ratio_max": 0.002811908023431897, "ademamix/cos_fast_slow": 0.44731980562210083, "ademamix/slow_fast_norm_ratio": 0.15585926175117493, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 89.126953125, "loss_smoothed": 2.2386674133439857, "seq_len": 4096} {"wall": 36482.499, "step": 7010, "loss": 2.130632370710373, "lm_loss": 2.130632370710373, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.305874764919281, "tokens": 2067333120, "adamw_lr": 0.0004, "muon/grad_norm": 0.25101223587989807, "adam/grad_norm": 0.1745021790266037, "muon/param_norm": 1944.831787109375, "muon/update_ratio": 0.010358244180679321, "muon/update_ratio_max": 0.08186646550893784, "adam/param_norm": 702.7408447265625, "adam/update_ratio": 0.0011216539423912764, "adam/update_ratio_max": 0.0027839220128953457, "ademamix/cos_fast_slow": 0.44494226574897766, "ademamix/slow_fast_norm_ratio": 0.14771783351898193, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 87.193359375, "loss_smoothed": 2.2124877870082855, "seq_len": 4096} {"wall": 36687.258, "step": 7020, "loss": 2.266058752934138, "lm_loss": 2.266058752934138, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2974391281604767, "tokens": 2070282240, "adamw_lr": 0.0004, "muon/grad_norm": 0.24568460881710052, "adam/grad_norm": 0.16726528108119965, "muon/param_norm": 1944.7681884765625, "muon/update_ratio": 0.010389716364443302, "muon/update_ratio_max": 0.08490773290395737, "adam/param_norm": 702.7808227539062, "adam/update_ratio": 0.0011064419522881508, "adam/update_ratio_max": 0.002745269564911723, "ademamix/cos_fast_slow": 0.46142762899398804, "ademamix/slow_fast_norm_ratio": 0.1445845514535904, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 87.193359375, "loss_smoothed": 2.2821429898341496, "seq_len": 4096} {"wall": 36892.041, "step": 7030, "loss": 2.08332567413648, "lm_loss": 2.08332567413648, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.27940380573272705, "tokens": 2073231360, "adamw_lr": 0.0004, "muon/grad_norm": 0.2297559231519699, "adam/grad_norm": 0.15874995291233063, "muon/param_norm": 1944.96337890625, "muon/update_ratio": 0.010379434563219547, "muon/update_ratio_max": 0.0828583836555481, "adam/param_norm": 702.8212280273438, "adam/update_ratio": 0.0010989606380462646, "adam/update_ratio_max": 0.002725896891206503, "ademamix/cos_fast_slow": 0.2483680695295334, "ademamix/slow_fast_norm_ratio": 0.15465904772281647, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 87.193359375, "loss_smoothed": 2.2187946580350397, "seq_len": 4096} {"wall": 37097.087, "step": 7040, "loss": 2.2175692468881607, "lm_loss": 2.2175692468881607, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3350965976715088, "tokens": 2076180480, "adamw_lr": 0.0004, "muon/grad_norm": 0.2326658070087433, "adam/grad_norm": 0.2413560152053833, "muon/param_norm": 1945.017333984375, "muon/update_ratio": 0.010391722433269024, "muon/update_ratio_max": 0.07674433290958405, "adam/param_norm": 702.8609619140625, "adam/update_ratio": 0.001118830987252295, "adam/update_ratio_max": 0.002774410182610154, "ademamix/cos_fast_slow": 0.4916623830795288, "ademamix/slow_fast_norm_ratio": 0.14137592911720276, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 87.193359375, "loss_smoothed": 2.2132288478314877, "seq_len": 4096} {"wall": 37301.943, "step": 7050, "loss": 2.1538579364617667, "lm_loss": 2.1538579364617667, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2909638583660126, "tokens": 2079129600, "adamw_lr": 0.0004, "muon/grad_norm": 0.22528015077114105, "adam/grad_norm": 0.1843811422586441, "muon/param_norm": 1942.8721923828125, "muon/update_ratio": 0.008696387521922588, "muon/update_ratio_max": 0.08094626665115356, "adam/param_norm": 702.9002685546875, "adam/update_ratio": 0.001105319825001061, "adam/update_ratio_max": 0.003829737426713109, "ademamix/cos_fast_slow": 0.6717765927314758, "ademamix/slow_fast_norm_ratio": 0.0722811222076416, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 87.193359375, "loss_smoothed": 2.231220548351606, "seq_len": 4096} {"wall": 37506.968, "step": 7060, "loss": 2.1781855324904122, "lm_loss": 2.1781855324904122, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.29763761162757874, "tokens": 2082078720, "adamw_lr": 0.0004, "muon/grad_norm": 0.2106885015964508, "adam/grad_norm": 0.21080094575881958, "muon/param_norm": 1938.6497802734375, "muon/update_ratio": 0.00923758652061224, "muon/update_ratio_max": 0.08305899053812027, "adam/param_norm": 702.9398193359375, "adam/update_ratio": 0.0011020569363608956, "adam/update_ratio_max": 0.0027268650010228157, "ademamix/cos_fast_slow": 0.6664668917655945, "ademamix/slow_fast_norm_ratio": 0.13741451501846313, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 87.193359375, "loss_smoothed": 2.2518135343988734, "seq_len": 4096} {"wall": 37712.133, "step": 7070, "loss": 2.0883895258108773, "lm_loss": 2.0883895258108773, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.376005619764328, "tokens": 2085027840, "adamw_lr": 0.0004, "muon/grad_norm": 0.29341813921928406, "adam/grad_norm": 0.2353830486536026, "muon/param_norm": 1936.4388427734375, "muon/update_ratio": 0.00968173984438181, "muon/update_ratio_max": 0.08169040083885193, "adam/param_norm": 702.980224609375, "adam/update_ratio": 0.001103787450119853, "adam/update_ratio_max": 0.002731399377807975, "ademamix/cos_fast_slow": 0.5943299531936646, "ademamix/slow_fast_norm_ratio": 0.1611930876970291, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 87.193359375, "loss_smoothed": 2.203413685659567, "seq_len": 4096} {"wall": 37917.242, "step": 7080, "loss": 2.3586447139581046, "lm_loss": 2.3586447139581046, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3157270848751068, "tokens": 2087976960, "adamw_lr": 0.0004, "muon/grad_norm": 0.2849082350730896, "adam/grad_norm": 0.1351531445980072, "muon/param_norm": 1935.266357421875, "muon/update_ratio": 0.009958971291780472, "muon/update_ratio_max": 0.08537560701370239, "adam/param_norm": 703.0197143554688, "adam/update_ratio": 0.0011316131567582488, "adam/update_ratio_max": 0.0028002196922898293, "ademamix/cos_fast_slow": 0.41270002722740173, "ademamix/slow_fast_norm_ratio": 0.1807824969291687, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 87.193359375, "loss_smoothed": 2.2749661087989805, "seq_len": 4096} {"wall": 38123.378, "step": 7090, "loss": 2.3431515097618103, "lm_loss": 2.3431515097618103, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.35349026322364807, "tokens": 2090926080, "adamw_lr": 0.0004, "muon/grad_norm": 0.2672734558582306, "adam/grad_norm": 0.23101240396499634, "muon/param_norm": 1934.9898681640625, "muon/update_ratio": 0.010192738845944405, "muon/update_ratio_max": 0.08472900092601776, "adam/param_norm": 703.05908203125, "adam/update_ratio": 0.0011098675895482302, "adam/update_ratio_max": 0.0027460982091724873, "ademamix/cos_fast_slow": 0.4847829043865204, "ademamix/slow_fast_norm_ratio": 0.16377447545528412, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 87.193359375, "loss_smoothed": 2.267962817847729, "seq_len": 4096} {"wall": 38331.954, "eval_loss": 1.913412258028984, "eval_ppl": 6.7761712074279785, "step": 7100} {"wall": 38331.956, "step": 7100, "loss": 2.2551298985878625, "lm_loss": 2.2551298985878625, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3961523175239563, "tokens": 2093875200, "adamw_lr": 0.0004, "muon/grad_norm": 0.28893762826919556, "adam/grad_norm": 0.2706913352012634, "muon/param_norm": 1935.6365966796875, "muon/update_ratio": 0.010265960358083248, "muon/update_ratio_max": 0.08107627183198929, "adam/param_norm": 703.098876953125, "adam/update_ratio": 0.0011033584596589208, "adam/update_ratio_max": 0.0027297688648104668, "ademamix/cos_fast_slow": 0.4723471701145172, "ademamix/slow_fast_norm_ratio": 0.1545729786157608, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 87.193359375, "loss_smoothed": 2.1922246190408865, "seq_len": 4096} {"wall": 38616.104, "step": 7110, "loss": 2.146254926919937, "lm_loss": 2.146254926919937, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.32693564891815186, "tokens": 2096824320, "adamw_lr": 0.0004, "muon/grad_norm": 0.27200034260749817, "adam/grad_norm": 0.18099771440029144, "muon/param_norm": 1936.438232421875, "muon/update_ratio": 0.010374738834798336, "muon/update_ratio_max": 0.08286671340465546, "adam/param_norm": 703.1383666992188, "adam/update_ratio": 0.001113034668378532, "adam/update_ratio_max": 0.002753526670858264, "ademamix/cos_fast_slow": 0.34179073572158813, "ademamix/slow_fast_norm_ratio": 0.1745539903640747, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 86.998046875, "loss_smoothed": 2.2125641251603763, "seq_len": 4096} {"wall": 38820.942, "step": 7120, "loss": 2.208982835213343, "lm_loss": 2.208982835213343, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.29664671421051025, "tokens": 2099773440, "adamw_lr": 0.0004, "muon/grad_norm": 0.23739418387413025, "adam/grad_norm": 0.17806054651737213, "muon/param_norm": 1937.1259765625, "muon/update_ratio": 0.010308446362614632, "muon/update_ratio_max": 0.08106248080730438, "adam/param_norm": 703.1776123046875, "adam/update_ratio": 0.0011181153822690248, "adam/update_ratio_max": 0.0027654317673295736, "ademamix/cos_fast_slow": 0.1308314949274063, "ademamix/slow_fast_norm_ratio": 0.1688465178012848, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 86.998046875, "loss_smoothed": 2.2387433690329392, "seq_len": 4096} {"wall": 39025.866, "step": 7130, "loss": 2.3229711999495826, "lm_loss": 2.3229711999495826, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.30694156885147095, "tokens": 2102722560, "adamw_lr": 0.0004, "muon/grad_norm": 0.2419779747724533, "adam/grad_norm": 0.18927760422229767, "muon/param_norm": 1937.80322265625, "muon/update_ratio": 0.010372728109359741, "muon/update_ratio_max": 0.08661907911300659, "adam/param_norm": 703.2172241210938, "adam/update_ratio": 0.0011017043143510818, "adam/update_ratio_max": 0.0027244158554822206, "ademamix/cos_fast_slow": 0.2062850296497345, "ademamix/slow_fast_norm_ratio": 0.17441976070404053, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 86.998046875, "loss_smoothed": 2.23619814068079, "seq_len": 4096} {"wall": 39230.946, "step": 7140, "loss": 2.3360040187835693, "lm_loss": 2.3360040187835693, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3183291554450989, "tokens": 2105671680, "adamw_lr": 0.0004, "muon/grad_norm": 0.24205194413661957, "adam/grad_norm": 0.20660743117332458, "muon/param_norm": 1938.1934814453125, "muon/update_ratio": 0.010341639630496502, "muon/update_ratio_max": 0.0845983475446701, "adam/param_norm": 703.2564697265625, "adam/update_ratio": 0.0011076928349211812, "adam/update_ratio_max": 0.0027386187575757504, "ademamix/cos_fast_slow": 0.13973161578178406, "ademamix/slow_fast_norm_ratio": 0.15974080562591553, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 86.998046875, "loss_smoothed": 2.198451068500678, "seq_len": 4096} {"wall": 39436.056, "step": 7150, "loss": 2.2250578751166663, "lm_loss": 2.2250578751166663, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.36783212423324585, "tokens": 2108620800, "adamw_lr": 0.0004, "muon/grad_norm": 0.2645391523838043, "adam/grad_norm": 0.25607484579086304, "muon/param_norm": 1939.1033935546875, "muon/update_ratio": 0.010359910316765308, "muon/update_ratio_max": 0.08377829194068909, "adam/param_norm": 703.295166015625, "adam/update_ratio": 0.001098374486900866, "adam/update_ratio_max": 0.002714886562898755, "ademamix/cos_fast_slow": 0.5181430578231812, "ademamix/slow_fast_norm_ratio": 0.15527091920375824, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 86.998046875, "loss_smoothed": 2.1902010530233382, "seq_len": 4096} {"wall": 39640.945, "step": 7160, "loss": 2.238897035519282, "lm_loss": 2.238897035519282, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3327997326850891, "tokens": 2111569920, "adamw_lr": 0.0004, "muon/grad_norm": 0.25278908014297485, "adam/grad_norm": 0.2162550985813141, "muon/param_norm": 1939.9241943359375, "muon/update_ratio": 0.010370458476245403, "muon/update_ratio_max": 0.08241190016269684, "adam/param_norm": 703.3339233398438, "adam/update_ratio": 0.0011077072704210877, "adam/update_ratio_max": 0.0027372518088668585, "ademamix/cos_fast_slow": 0.4155324101448059, "ademamix/slow_fast_norm_ratio": 0.15867112576961517, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 86.998046875, "loss_smoothed": 2.218229531496763, "seq_len": 4096} {"wall": 39845.784, "step": 7170, "loss": 2.296092987060547, "lm_loss": 2.296092987060547, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.29080620408058167, "tokens": 2114519040, "adamw_lr": 0.0004, "muon/grad_norm": 0.23902618885040283, "adam/grad_norm": 0.16558825969696045, "muon/param_norm": 1940.3323974609375, "muon/update_ratio": 0.010329633951187134, "muon/update_ratio_max": 0.0847652330994606, "adam/param_norm": 703.37255859375, "adam/update_ratio": 0.0011045689461752772, "adam/update_ratio_max": 0.0027285986579954624, "ademamix/cos_fast_slow": 0.20260852575302124, "ademamix/slow_fast_norm_ratio": 0.16485446691513062, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 86.998046875, "loss_smoothed": 2.3001925602555273, "seq_len": 4096} {"wall": 40051.07, "step": 7180, "loss": 2.2127148509025574, "lm_loss": 2.2127148509025574, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3149455785751343, "tokens": 2117468160, "adamw_lr": 0.0004, "muon/grad_norm": 0.2157479226589203, "adam/grad_norm": 0.22931477427482605, "muon/param_norm": 1940.410888671875, "muon/update_ratio": 0.01037109736353159, "muon/update_ratio_max": 0.08278517425060272, "adam/param_norm": 703.41357421875, "adam/update_ratio": 0.0010995727498084307, "adam/update_ratio_max": 0.002715323818847537, "ademamix/cos_fast_slow": 0.4886539876461029, "ademamix/slow_fast_norm_ratio": 0.1435643881559372, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 86.998046875, "loss_smoothed": 2.1503451990584534, "seq_len": 4096} {"wall": 40255.975, "step": 7190, "loss": 2.208519327143828, "lm_loss": 2.208519327143828, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3396632671356201, "tokens": 2120417280, "adamw_lr": 0.0004, "muon/grad_norm": 0.2572692930698395, "adam/grad_norm": 0.22219517827033997, "muon/param_norm": 1940.748046875, "muon/update_ratio": 0.010280540212988853, "muon/update_ratio_max": 0.08274061977863312, "adam/param_norm": 703.4539794921875, "adam/update_ratio": 0.001100590219721198, "adam/update_ratio_max": 0.002717256546020508, "ademamix/cos_fast_slow": 0.19853557646274567, "ademamix/slow_fast_norm_ratio": 0.14843077957630157, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 86.998046875, "loss_smoothed": 2.1960854942599934, "seq_len": 4096} {"wall": 40463.299, "eval_loss": 1.899511456489563, "eval_ppl": 6.682629108428955, "step": 7200} {"wall": 40463.301, "step": 7200, "loss": 2.3936877051989236, "lm_loss": 2.3936877051989236, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.33779776096343994, "tokens": 2123366400, "adamw_lr": 0.0004, "muon/grad_norm": 0.269137442111969, "adam/grad_norm": 0.20412927865982056, "muon/param_norm": 1940.89306640625, "muon/update_ratio": 0.01032662857323885, "muon/update_ratio_max": 0.07882080227136612, "adam/param_norm": 703.4942016601562, "adam/update_ratio": 0.0011091063497588038, "adam/update_ratio_max": 0.0027374238707125187, "ademamix/cos_fast_slow": 0.26110315322875977, "ademamix/slow_fast_norm_ratio": 0.15337790548801422, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 86.998046875, "loss_smoothed": 2.2821626265843706, "seq_len": 4096} {"wall": 158.4, "eval_loss": 1.8995585590600967, "eval_ppl": 6.682943820953369, "step": 7200} {"wall": 892.381, "step": 7210, "loss": 2.2340605705976486, "lm_loss": 2.2340605705976486, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.40077805519104004, "tokens": 2126315520, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.302734375, "loss_smoothed": 2.245195053021113, "seq_len": 4096} {"wall": 1097.884, "step": 7220, "loss": 2.1312559098005295, "lm_loss": 2.1312559098005295, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2823336124420166, "tokens": 2129264640, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.302734375, "loss_smoothed": 2.24267407009999, "seq_len": 4096} {"wall": 1303.333, "step": 7230, "loss": 2.2536853005488715, "lm_loss": 2.2536853005488715, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3660520017147064, "tokens": 2132213760, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.302734375, "loss_smoothed": 2.1824468585352106, "seq_len": 4096} {"wall": 1508.593, "step": 7240, "loss": 2.3294680068890252, "lm_loss": 2.3294680068890252, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.24897290766239166, "tokens": 2135162880, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.302734375, "loss_smoothed": 2.230551139762004, "seq_len": 4096} {"wall": 1714.051, "step": 7250, "loss": 2.1645777374505997, "lm_loss": 2.1645777374505997, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.5797768235206604, "tokens": 2138112000, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.302734375, "loss_smoothed": 2.252466354270776, "seq_len": 4096} {"wall": 1919.382, "step": 7260, "loss": 2.3493137458960214, "lm_loss": 2.3493137458960214, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3170466125011444, "tokens": 2141061120, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.302734375, "loss_smoothed": 2.22153766900301, "seq_len": 4096} {"wall": 2125.367, "step": 7270, "loss": 2.228617404898008, "lm_loss": 2.228617404898008, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3620855212211609, "tokens": 2144010240, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.302734375, "loss_smoothed": 2.229142802457015, "seq_len": 4096} {"wall": 2330.442, "step": 7280, "loss": 2.1343268752098083, "lm_loss": 2.1343268752098083, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3332114815711975, "tokens": 2146959360, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.302734375, "loss_smoothed": 2.2388343468308447, "seq_len": 4096} {"wall": 2535.744, "step": 7290, "loss": 2.2821491956710815, "lm_loss": 2.2821491956710815, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.33235281705856323, "tokens": 2149908480, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.302734375, "loss_smoothed": 2.2132319231828053, "seq_len": 4096} {"wall": 2743.493, "eval_loss": 1.9102170765399933, "eval_ppl": 6.754554748535156, "step": 7300} {"wall": 2743.494, "step": 7300, "loss": 2.2169979264338813, "lm_loss": 2.2169979264338813, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.28496015071868896, "tokens": 2152857600, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.302734375, "loss_smoothed": 2.2011576225360234, "seq_len": 4096} {"wall": 2962.36, "step": 7310, "loss": 2.330208937327067, "lm_loss": 2.330208937327067, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2925664186477661, "tokens": 2155806720, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.2687580565611523, "seq_len": 4096} {"wall": 3168.235, "step": 7320, "loss": 2.267301062742869, "lm_loss": 2.267301062742869, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.35944196581840515, "tokens": 2158755840, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.2001725653807322, "seq_len": 4096} {"wall": 3373.42, "step": 7330, "loss": 2.3385067731142044, "lm_loss": 2.3385067731142044, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.30972275137901306, "tokens": 2161704960, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.221622159083684, "seq_len": 4096} {"wall": 3578.869, "step": 7340, "loss": 2.351225902636846, "lm_loss": 2.351225902636846, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3389928936958313, "tokens": 2164654080, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.2542589376370112, "seq_len": 4096} {"wall": 3784.737, "step": 7350, "loss": 2.2483836313088736, "lm_loss": 2.2483836313088736, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3582954704761505, "tokens": 2167603200, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.270545894900958, "seq_len": 4096} {"wall": 3989.825, "step": 7360, "loss": 2.2672678331534066, "lm_loss": 2.2672678331534066, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3806975483894348, "tokens": 2170552320, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.219811411698659, "seq_len": 4096} {"wall": 4195.34, "step": 7370, "loss": 2.213309958577156, "lm_loss": 2.213309958577156, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.29287073016166687, "tokens": 2173501440, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.244069251169761, "seq_len": 4096} {"wall": 4401.168, "step": 7380, "loss": 2.2784912238518396, "lm_loss": 2.2784912238518396, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.318633109331131, "tokens": 2176450560, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.1897954230507213, "seq_len": 4096} {"wall": 4606.485, "step": 7390, "loss": 2.2171241293350854, "lm_loss": 2.2171241293350854, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3484538495540619, "tokens": 2179399680, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.228401086727778, "seq_len": 4096} {"wall": 4814.343, "eval_loss": 1.906006470322609, "eval_ppl": 6.7261738777160645, "step": 7400} {"wall": 4814.345, "step": 7400, "loss": 2.185451557238897, "lm_loss": 2.185451557238897, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3062881827354431, "tokens": 2182348800, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.2244600216547648, "seq_len": 4096} {"wall": 5100.139, "step": 7410, "loss": 2.329829196135203, "lm_loss": 2.329829196135203, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3201523423194885, "tokens": 2185297920, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.498046875, "loss_smoothed": 2.2566741198301314, "seq_len": 4096} {"wall": 5305.555, "step": 7420, "loss": 2.173454225063324, "lm_loss": 2.173454225063324, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.32110509276390076, "tokens": 2188247040, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.498046875, "loss_smoothed": 2.180071218063434, "seq_len": 4096} {"wall": 5511.452, "step": 7430, "loss": 2.0616615787148476, "lm_loss": 2.0616615787148476, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.24810375273227692, "tokens": 2191196160, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.498046875, "loss_smoothed": 2.2278392501175404, "seq_len": 4096} {"wall": 5716.202, "step": 7440, "loss": 2.127371077736219, "lm_loss": 2.127371077736219, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2983195185661316, "tokens": 2194145280, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.498046875, "loss_smoothed": 2.1678169553478557, "seq_len": 4096} {"wall": 5920.868, "step": 7450, "loss": 2.1583589812119803, "lm_loss": 2.1583589812119803, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3288244307041168, "tokens": 2197094400, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.498046875, "loss_smoothed": 2.2407224165896578, "seq_len": 4096} {"wall": 6125.797, "step": 7460, "loss": 2.15178815027078, "lm_loss": 2.15178815027078, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3441295921802521, "tokens": 2200043520, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.498046875, "loss_smoothed": 2.2203321866691113, "seq_len": 4096} {"wall": 6330.3, "step": 7470, "loss": 2.1315069595972695, "lm_loss": 2.1315069595972695, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3508169651031494, "tokens": 2202992640, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.498046875, "loss_smoothed": 2.265157192448775, "seq_len": 4096} {"wall": 6535.212, "step": 7480, "loss": 2.391707330942154, "lm_loss": 2.391707330942154, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3084454834461212, "tokens": 2205941760, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.498046875, "loss_smoothed": 2.186328380803267, "seq_len": 4096} {"wall": 6740.268, "step": 7490, "loss": 2.152279863754908, "lm_loss": 2.152279863754908, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3159566819667816, "tokens": 2208890880, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.498046875, "loss_smoothed": 2.205030154188474, "seq_len": 4096} {"wall": 6947.098, "eval_loss": 1.904066413640976, "eval_ppl": 6.713137626647949, "step": 7500} {"wall": 6947.1, "step": 7500, "loss": 2.268050566315651, "lm_loss": 2.268050566315651, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.31010136008262634, "tokens": 2211840000, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.498046875, "loss_smoothed": 2.227519405881564, "seq_len": 4096} {"wall": 7229.589, "step": 7510, "loss": 2.2065874338150024, "lm_loss": 2.2065874338150024, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.33978474140167236, "tokens": 2214789120, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.202607789138953, "seq_len": 4096} {"wall": 7434.265, "step": 7520, "loss": 2.280862679084142, "lm_loss": 2.280862679084142, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2641179859638214, "tokens": 2217738240, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.2242303557693957, "seq_len": 4096} {"wall": 7638.993, "step": 7530, "loss": 2.0638744135697684, "lm_loss": 2.0638744135697684, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.34064438939094543, "tokens": 2220687360, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.1932508826255797, "seq_len": 4096} {"wall": 7843.935, "step": 7540, "loss": 2.258356382449468, "lm_loss": 2.258356382449468, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3244863450527191, "tokens": 2223636480, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.1596314653754236, "seq_len": 4096} {"wall": 8048.28, "step": 7550, "loss": 2.256014421582222, "lm_loss": 2.256014421582222, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.36427590250968933, "tokens": 2226585600, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.220189055800438, "seq_len": 4096} {"wall": 8252.945, "step": 7560, "loss": 2.2816530664761863, "lm_loss": 2.2816530664761863, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3668995499610901, "tokens": 2229534720, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.2056768750151, "seq_len": 4096} {"wall": 8458.297, "step": 7570, "loss": 2.1709067448973656, "lm_loss": 2.1709067448973656, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.29076918959617615, "tokens": 2232483840, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.2029625174899894, "seq_len": 4096} {"wall": 8662.757, "step": 7580, "loss": 2.24979467689991, "lm_loss": 2.24979467689991, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.37173184752464294, "tokens": 2235432960, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.2057685454686484, "seq_len": 4096} {"wall": 8867.914, "step": 7590, "loss": 2.27835022409757, "lm_loss": 2.27835022409757, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3151431977748871, "tokens": 2238382080, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.182744794587294, "seq_len": 4096} {"wall": 9075.259, "eval_loss": 1.906115546822548, "eval_ppl": 6.726907730102539, "step": 7600} {"wall": 9075.26, "step": 7600, "loss": 2.1286584635575614, "lm_loss": 2.1286584635575614, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2787507176399231, "tokens": 2241331200, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.2174641465147333, "seq_len": 4096} {"wall": 9362.99, "step": 7610, "loss": 2.235388000806173, "lm_loss": 2.235388000806173, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.29029297828674316, "tokens": 2244280320, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.2313093240062396, "seq_len": 4096} {"wall": 9567.804, "step": 7620, "loss": 2.128022685647011, "lm_loss": 2.128022685647011, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.33465319871902466, "tokens": 2247229440, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.1540014391144116, "seq_len": 4096} {"wall": 9772.593, "step": 7630, "loss": 2.2664321114619574, "lm_loss": 2.2664321114619574, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3112320005893707, "tokens": 2250178560, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.203535044689973, "seq_len": 4096} {"wall": 9977.346, "step": 7640, "loss": 2.2655963053305945, "lm_loss": 2.2655963053305945, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3280060887336731, "tokens": 2253127680, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.197291707744201, "seq_len": 4096} {"wall": 10182.022, "step": 7650, "loss": 2.1786059687534967, "lm_loss": 2.1786059687534967, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3665941655635834, "tokens": 2256076800, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.244910891354084, "seq_len": 4096} {"wall": 10387.303, "step": 7660, "loss": 2.2124266922473907, "lm_loss": 2.2124266922473907, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.31704291701316833, "tokens": 2259025920, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.217182046175003, "seq_len": 4096} {"wall": 10591.964, "step": 7670, "loss": 2.356703912218412, "lm_loss": 2.356703912218412, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.27564117312431335, "tokens": 2261975040, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.24526683712999, "seq_len": 4096} {"wall": 10796.946, "step": 7680, "loss": 2.2067540933688483, "lm_loss": 2.2067540933688483, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.28841763734817505, "tokens": 2264924160, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.2240700883169966, "seq_len": 4096} {"wall": 11001.545, "step": 7690, "loss": 2.228425160050392, "lm_loss": 2.228425160050392, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3071777820587158, "tokens": 2267873280, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.208002592374881, "seq_len": 4096} {"wall": 11208.947, "eval_loss": 1.9024571925401688, "eval_ppl": 6.702343463897705, "step": 7700} {"wall": 11208.949, "step": 7700, "loss": 2.3295548806587854, "lm_loss": 2.3295548806587854, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3380230963230133, "tokens": 2270822400, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.2238015733659267, "seq_len": 4096} {"wall": 11494.844, "step": 7710, "loss": 2.248975788553556, "lm_loss": 2.248975788553556, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3082026541233063, "tokens": 2273771520, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.199691420048475, "seq_len": 4096} {"wall": 11700.069, "step": 7720, "loss": 2.1555013383428254, "lm_loss": 2.1555013383428254, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3110292851924896, "tokens": 2276720640, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.229215592890978, "seq_len": 4096} {"wall": 11904.403, "step": 7730, "loss": 2.1926390727361045, "lm_loss": 2.1926390727361045, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.31910625100135803, "tokens": 2279669760, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.2334785483777524, "seq_len": 4096} {"wall": 12109.406, "step": 7740, "loss": 2.2431828677654266, "lm_loss": 2.2431828677654266, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3617793321609497, "tokens": 2282618880, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.2332383846243222, "seq_len": 4096} {"wall": 12314.249, "step": 7750, "loss": 2.2641634990771613, "lm_loss": 2.2641634990771613, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2675546109676361, "tokens": 2285568000, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.240728139380614, "seq_len": 4096} {"wall": 12519.103, "step": 7760, "loss": 2.2328336040178933, "lm_loss": 2.2328336040178933, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.33760982751846313, "tokens": 2288517120, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.2212015648682915, "seq_len": 4096} {"wall": 12723.909, "step": 7770, "loss": 2.3036185751358667, "lm_loss": 2.3036185751358667, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.34003689885139465, "tokens": 2291466240, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.228162195533514, "seq_len": 4096} {"wall": 12928.392, "step": 7780, "loss": 2.369109263022741, "lm_loss": 2.369109263022741, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3338720202445984, "tokens": 2294415360, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.1726746576527756, "seq_len": 4096} {"wall": 13133.554, "step": 7790, "loss": 2.334624578555425, "lm_loss": 2.334624578555425, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.31701698899269104, "tokens": 2297364480, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.221100699901581, "seq_len": 4096} {"wall": 13340.651, "eval_loss": 1.9029482007026672, "eval_ppl": 6.705634593963623, "step": 7800} {"wall": 13340.653, "step": 7800, "loss": 2.1653951605161033, "lm_loss": 2.1653951605161033, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3741092383861542, "tokens": 2300313600, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.1631858296692372, "seq_len": 4096} {"wall": 13619.854, "step": 7810, "loss": 2.154275973637899, "lm_loss": 2.154275973637899, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2739344537258148, "tokens": 2303262720, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.1842840063075224, "seq_len": 4096} {"wall": 13824.729, "step": 7820, "loss": 2.277797758579254, "lm_loss": 2.277797758579254, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.292782187461853, "tokens": 2306211840, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.2118501385052998, "seq_len": 4096} {"wall": 14029.662, "step": 7830, "loss": 2.2013813902934394, "lm_loss": 2.2013813902934394, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.283364474773407, "tokens": 2309160960, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.2121288627386093, "seq_len": 4096} {"wall": 14234.41, "step": 7840, "loss": 2.190535932779312, "lm_loss": 2.190535932779312, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.27794167399406433, "tokens": 2312110080, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.2195437361796695, "seq_len": 4096} {"wall": 14439.706, "step": 7850, "loss": 2.3625869949658713, "lm_loss": 2.3625869949658713, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.36203011870384216, "tokens": 2315059200, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.2027997756997744, "seq_len": 4096} {"wall": 14644.16, "step": 7860, "loss": 2.265399679541588, "lm_loss": 2.265399679541588, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.40670856833457947, "tokens": 2318008320, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.2369289708634215, "seq_len": 4096} {"wall": 14848.842, "step": 7870, "loss": 2.1562751680612564, "lm_loss": 2.1562751680612564, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.24357803165912628, "tokens": 2320957440, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.2111367285251617, "seq_len": 4096} {"wall": 15054.017, "step": 7880, "loss": 2.064914765457312, "lm_loss": 2.064914765457312, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2768966257572174, "tokens": 2323906560, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.1809512073795, "seq_len": 4096} {"wall": 15258.508, "step": 7890, "loss": 2.1932215293248496, "lm_loss": 2.1932215293248496, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.5038187503814697, "tokens": 2326855680, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.1874153057734174, "seq_len": 4096} {"wall": 15465.586, "eval_loss": 1.8994162827730179, "eval_ppl": 6.681993007659912, "step": 7900} {"wall": 15465.587, "step": 7900, "loss": 2.3041215737660727, "lm_loss": 2.3041215737660727, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.30242452025413513, "tokens": 2329804800, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.2263425553838414, "seq_len": 4096} {"wall": 15756.125, "step": 7910, "loss": 2.298175980647405, "lm_loss": 2.298175980647405, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.43881699442863464, "tokens": 2332753920, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.198708090931177, "seq_len": 4096} {"wall": 15960.578, "step": 7920, "loss": 2.278415476282438, "lm_loss": 2.278415476282438, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.25365638732910156, "tokens": 2335703040, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.2165106385946274, "seq_len": 4096} {"wall": 16166.2, "step": 7930, "loss": 2.3049920797348022, "lm_loss": 2.3049920797348022, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3911150097846985, "tokens": 2338652160, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.229645757873853, "seq_len": 4096} {"wall": 16371.42, "step": 7940, "loss": 2.2513194481531777, "lm_loss": 2.2513194481531777, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2990930676460266, "tokens": 2341601280, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.203076369563738, "seq_len": 4096} {"wall": 16575.944, "step": 7950, "loss": 2.4793144861857095, "lm_loss": 2.4793144861857095, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3786318898200989, "tokens": 2344550400, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.2500917226076127, "seq_len": 4096} {"wall": 16780.876, "step": 7960, "loss": 2.1652946869532266, "lm_loss": 2.1652946869532266, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.30846846103668213, "tokens": 2347499520, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.1871015548706056, "seq_len": 4096} {"wall": 16985.748, "step": 7970, "loss": 2.0576858818531036, "lm_loss": 2.0576858818531036, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.32670900225639343, "tokens": 2350448640, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.1916867822408674, "seq_len": 4096} {"wall": 17190.669, "step": 7980, "loss": 2.2885589947303138, "lm_loss": 2.2885589947303138, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.715192437171936, "tokens": 2353397760, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.232943900426229, "seq_len": 4096} {"wall": 17395.607, "step": 7990, "loss": 2.1097533206144967, "lm_loss": 2.1097533206144967, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3186236321926117, "tokens": 2356346880, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.2170216903090476, "seq_len": 4096} {"wall": 17603.019, "eval_loss": 1.8997206091880798, "eval_ppl": 6.684027194976807, "step": 8000} {"wall": 17603.02, "step": 8000, "loss": 2.303067954878012, "lm_loss": 2.303067954878012, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.35067158937454224, "tokens": 2359296000, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.2130718879401683, "seq_len": 4096} {"wall": 17885.858, "step": 8010, "loss": 2.173337683081627, "lm_loss": 2.173337683081627, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3558432459831238, "tokens": 2362245120, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.21005629748106, "seq_len": 4096} {"wall": 18090.891, "step": 8020, "loss": 2.259723966320356, "lm_loss": 2.259723966320356, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.43335434794425964, "tokens": 2365194240, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.1756410161654154, "seq_len": 4096} {"wall": 18295.945, "step": 8030, "loss": 2.101552282770475, "lm_loss": 2.101552282770475, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3495882451534271, "tokens": 2368143360, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.1816205769777297, "seq_len": 4096} {"wall": 18500.891, "step": 8040, "loss": 2.199344297250112, "lm_loss": 2.199344297250112, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3041631281375885, "tokens": 2371092480, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.2151820468405883, "seq_len": 4096} {"wall": 18705.93, "step": 8050, "loss": 2.088580640653769, "lm_loss": 2.088580640653769, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3355260491371155, "tokens": 2374041600, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.2036061875522135, "seq_len": 4096} {"wall": 18910.717, "step": 8060, "loss": 2.2525749852259955, "lm_loss": 2.2525749852259955, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.26860693097114563, "tokens": 2376990720, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.194067202011744, "seq_len": 4096} {"wall": 19116.081, "step": 8070, "loss": 2.1440604527791343, "lm_loss": 2.1440604527791343, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3616364598274231, "tokens": 2379939840, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.2079728662967684, "seq_len": 4096} {"wall": 19320.507, "step": 8080, "loss": 2.103204980492592, "lm_loss": 2.103204980492592, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.457001268863678, "tokens": 2382888960, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.2370830257733663, "seq_len": 4096} {"wall": 19525.208, "step": 8090, "loss": 2.275044451157252, "lm_loss": 2.275044451157252, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.28139054775238037, "tokens": 2385838080, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.2232179060578345, "seq_len": 4096} {"wall": 19732.332, "eval_loss": 1.89013309776783, "eval_ppl": 6.620250225067139, "step": 8100} {"wall": 19732.334, "step": 8100, "loss": 2.176281303167343, "lm_loss": 2.176281303167343, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3302142322063446, "tokens": 2388787200, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.1944095517198248, "seq_len": 4096} {"wall": 20016.961, "step": 8110, "loss": 2.1791651969154677, "lm_loss": 2.1791651969154677, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.33590859174728394, "tokens": 2391736320, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.233439705520868, "seq_len": 4096} {"wall": 20221.566, "step": 8120, "loss": 2.2926071733236313, "lm_loss": 2.2926071733236313, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.283729612827301, "tokens": 2394685440, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.2048228954275446, "seq_len": 4096} {"wall": 20426.157, "step": 8130, "loss": 2.1754858692487082, "lm_loss": 2.1754858692487082, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.46738749742507935, "tokens": 2397634560, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.1833214064439135, "seq_len": 4096} {"wall": 20630.744, "step": 8140, "loss": 2.279161721467972, "lm_loss": 2.279161721467972, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3081234097480774, "tokens": 2400583680, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.2046298759678997, "seq_len": 4096} {"wall": 20835.38, "step": 8150, "loss": 2.0842377493778863, "lm_loss": 2.0842377493778863, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.28167468309402466, "tokens": 2403532800, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.2233631243308385, "seq_len": 4096} {"wall": 21039.994, "step": 8160, "loss": 2.2968273013830185, "lm_loss": 2.2968273013830185, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.36065006256103516, "tokens": 2406481920, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.2249596108992895, "seq_len": 4096} {"wall": 21244.738, "step": 8170, "loss": 2.233495349685351, "lm_loss": 2.233495349685351, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2922065854072571, "tokens": 2409431040, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.221995280931393, "seq_len": 4096} {"wall": 21449.468, "step": 8180, "loss": 2.154076879223188, "lm_loss": 2.154076879223188, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.31586316227912903, "tokens": 2412380160, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.2155892764528593, "seq_len": 4096} {"wall": 21654.067, "step": 8190, "loss": 2.2173466831445694, "lm_loss": 2.2173466831445694, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3109112083911896, "tokens": 2415329280, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.19790043408672, "seq_len": 4096} {"wall": 21861.131, "eval_loss": 1.8878356367349625, "eval_ppl": 6.605057239532471, "step": 8200} {"wall": 21861.133, "step": 8200, "loss": 2.1101492991050086, "lm_loss": 2.1101492991050086, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2938058376312256, "tokens": 2418278400, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.2320437436302507, "seq_len": 4096} {"wall": 22143.575, "step": 8210, "loss": 2.099737763404846, "lm_loss": 2.099737763404846, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.32662296295166016, "tokens": 2421227520, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.1919822255770365, "seq_len": 4096} {"wall": 22348.095, "step": 8220, "loss": 2.2428107062975564, "lm_loss": 2.2428107062975564, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.33412572741508484, "tokens": 2424176640, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.2284502521157266, "seq_len": 4096} {"wall": 22552.643, "step": 8230, "loss": 2.1347451955080032, "lm_loss": 2.1347451955080032, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3441259562969208, "tokens": 2427125760, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.1948978185653685, "seq_len": 4096} {"wall": 22757.165, "step": 8240, "loss": 2.2299968947966895, "lm_loss": 2.2299968947966895, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.4217263162136078, "tokens": 2430074880, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.223567947248618, "seq_len": 4096} {"wall": 22961.761, "step": 8250, "loss": 2.230174094438553, "lm_loss": 2.230174094438553, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3245074450969696, "tokens": 2433024000, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.2159902637203532, "seq_len": 4096} {"wall": 23166.398, "step": 8260, "loss": 2.1850279370943704, "lm_loss": 2.1850279370943704, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.31393611431121826, "tokens": 2435973120, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.2211281597614287, "seq_len": 4096} {"wall": 23370.999, "step": 8270, "loss": 2.226835866769155, "lm_loss": 2.226835866769155, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3375342786312103, "tokens": 2438922240, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.2262100651860237, "seq_len": 4096} {"wall": 23575.637, "step": 8280, "loss": 2.243646298845609, "lm_loss": 2.243646298845609, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.28769829869270325, "tokens": 2441871360, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.2637660359342893, "seq_len": 4096} {"wall": 23780.365, "step": 8290, "loss": 2.1150326281785965, "lm_loss": 2.1150326281785965, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2924090027809143, "tokens": 2444820480, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.197126042097807, "seq_len": 4096} {"wall": 23987.575, "eval_loss": 1.8920727968215942, "eval_ppl": 6.633103370666504, "step": 8300} {"wall": 23987.576, "step": 8300, "loss": 2.2416761815547943, "lm_loss": 2.2416761815547943, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3252115249633789, "tokens": 2447769600, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.18785262281696, "seq_len": 4096} {"wall": 24269.011, "step": 8310, "loss": 2.209514165918032, "lm_loss": 2.209514165918032, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.32168102264404297, "tokens": 2450718720, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.163073041041692, "seq_len": 4096} {"wall": 24473.751, "step": 8320, "loss": 2.2327010333538055, "lm_loss": 2.2327010333538055, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.422590047121048, "tokens": 2453667840, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.214720828831196, "seq_len": 4096} {"wall": 24678.47, "step": 8330, "loss": 2.1731815735499063, "lm_loss": 2.1731815735499063, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3025880455970764, "tokens": 2456616960, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.216018046935399, "seq_len": 4096} {"wall": 24883.228, "step": 8340, "loss": 2.221416582663854, "lm_loss": 2.221416582663854, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.36143144965171814, "tokens": 2459566080, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.240000936140617, "seq_len": 4096} {"wall": 25088.034, "step": 8350, "loss": 2.1699275275071463, "lm_loss": 2.1699275275071463, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3927488923072815, "tokens": 2462515200, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.200775723656019, "seq_len": 4096} {"wall": 25292.68, "step": 8360, "loss": 2.2916423777739205, "lm_loss": 2.2916423777739205, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3392316401004791, "tokens": 2465464320, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.2040772013366223, "seq_len": 4096} {"wall": 25497.376, "step": 8370, "loss": 2.2827947487433753, "lm_loss": 2.2827947487433753, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.28866517543792725, "tokens": 2468413440, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.221640496701002, "seq_len": 4096} {"wall": 25702.043, "step": 8380, "loss": 2.214064215620359, "lm_loss": 2.214064215620359, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3201902508735657, "tokens": 2471362560, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.23091605330507, "seq_len": 4096} {"wall": 25906.841, "step": 8390, "loss": 2.210369442900022, "lm_loss": 2.210369442900022, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3134042024612427, "tokens": 2474311680, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.2110944263637067, "seq_len": 4096} {"wall": 26113.92, "eval_loss": 1.8919975757598877, "eval_ppl": 6.632604598999023, "step": 8400} {"wall": 26113.922, "step": 8400, "loss": 2.203824132680893, "lm_loss": 2.203824132680893, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3938198387622833, "tokens": 2477260800, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.2051547969381016, "seq_len": 4096} {"wall": 26401.416, "step": 8410, "loss": 2.135632356007894, "lm_loss": 2.135632356007894, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3589159846305847, "tokens": 2480209920, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.167037724951903, "seq_len": 4096} {"wall": 26606.335, "step": 8420, "loss": 2.2329188883304596, "lm_loss": 2.2329188883304596, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3600963056087494, "tokens": 2483159040, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.2018901688357193, "seq_len": 4096} {"wall": 26810.882, "step": 8430, "loss": 2.2141149739424386, "lm_loss": 2.2141149739424386, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3195876181125641, "tokens": 2486108160, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.190939672042926, "seq_len": 4096} {"wall": 27016.154, "step": 8440, "loss": 2.297335142890612, "lm_loss": 2.297335142890612, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3711977005004883, "tokens": 2489057280, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.2393693988521894, "seq_len": 4096} {"wall": 27221.412, "step": 8450, "loss": 2.139629448453585, "lm_loss": 2.139629448453585, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.29836270213127136, "tokens": 2492006400, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.2113716582457226, "seq_len": 4096} {"wall": 27428.136, "step": 8460, "loss": 2.3431309312582016, "lm_loss": 2.3431309312582016, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.4377794861793518, "tokens": 2494955520, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.2221706504623095, "seq_len": 4096} {"wall": 27635.215, "step": 8470, "loss": 2.104428539673487, "lm_loss": 2.104428539673487, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3082326054573059, "tokens": 2497904640, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.2262180715799333, "seq_len": 4096} {"wall": 27842.344, "step": 8480, "loss": 2.241530492901802, "lm_loss": 2.241530492901802, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.5883384346961975, "tokens": 2500853760, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.1979792063434918, "seq_len": 4096} {"wall": 28049.248, "step": 8490, "loss": 2.179541235168775, "lm_loss": 2.179541235168775, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.25921809673309326, "tokens": 2503802880, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.2095045154293382, "seq_len": 4096} {"wall": 28258.722, "eval_loss": 1.8956854492425919, "eval_ppl": 6.65710973739624, "step": 8500} {"wall": 28258.724, "step": 8500, "loss": 2.1457214057445526, "lm_loss": 2.1457214057445526, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.37356603145599365, "tokens": 2506752000, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.1923113877574605, "seq_len": 4096} {"wall": 28544.836, "step": 8510, "loss": 2.206611692905426, "lm_loss": 2.206611692905426, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3154886066913605, "tokens": 2509701120, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.2054184041917324, "seq_len": 4096} {"wall": 28751.771, "step": 8520, "loss": 2.157962133487066, "lm_loss": 2.157962133487066, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2731880843639374, "tokens": 2512650240, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.2296022934218245, "seq_len": 4096} {"wall": 28958.795, "step": 8530, "loss": 2.311821088194847, "lm_loss": 2.311821088194847, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.40301331877708435, "tokens": 2515599360, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.2309940472245215, "seq_len": 4096} {"wall": 29165.049, "step": 8540, "loss": 2.2992339034875235, "lm_loss": 2.2992339034875235, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.28307414054870605, "tokens": 2518548480, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.229489574829737, "seq_len": 4096} {"wall": 29370.01, "step": 8550, "loss": 2.2430338164170585, "lm_loss": 2.2430338164170585, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.4137013852596283, "tokens": 2521497600, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.2357658676803114, "seq_len": 4096} {"wall": 29574.854, "step": 8560, "loss": 2.317216078440348, "lm_loss": 2.317216078440348, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3779999017715454, "tokens": 2524446720, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.2216518759727477, "seq_len": 4096} {"wall": 29779.701, "step": 8570, "loss": 2.203330338001251, "lm_loss": 2.203330338001251, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3287440836429596, "tokens": 2527395840, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.241067658613125, "seq_len": 4096} {"wall": 29984.717, "step": 8580, "loss": 2.2416037122408548, "lm_loss": 2.2416037122408548, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.33285120129585266, "tokens": 2530344960, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.2116117457548774, "seq_len": 4096} {"wall": 30189.803, "step": 8590, "loss": 2.183014616370201, "lm_loss": 2.183014616370201, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.30776819586753845, "tokens": 2533294080, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.2101633292933305, "seq_len": 4096} {"wall": 30397.48, "eval_loss": 1.89206163585186, "eval_ppl": 6.633028984069824, "step": 8600} {"wall": 30397.482, "step": 8600, "loss": 2.310994486014048, "lm_loss": 2.310994486014048, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3985544443130493, "tokens": 2536243200, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 85.845703125, "loss_smoothed": 2.2216722791393595, "seq_len": 4096} {"wall": 30684.367, "step": 8610, "loss": 2.2535609354575477, "lm_loss": 2.2535609354575477, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.364594042301178, "tokens": 2539192320, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.2338745596508187, "seq_len": 4096} {"wall": 30889.014, "step": 8620, "loss": 2.120684484640757, "lm_loss": 2.120684484640757, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.28498223423957825, "tokens": 2542141440, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.172184731811285, "seq_len": 4096} {"wall": 31093.718, "step": 8630, "loss": 2.0673815508683524, "lm_loss": 2.0673815508683524, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2929283380508423, "tokens": 2545090560, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.1953128563861055, "seq_len": 4096} {"wall": 31298.668, "step": 8640, "loss": 2.292102908094724, "lm_loss": 2.292102908094724, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.39476194977760315, "tokens": 2548039680, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.1892951779067515, "seq_len": 4096} {"wall": 31503.068, "step": 8650, "loss": 2.1931192775567374, "lm_loss": 2.1931192775567374, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.38146036863327026, "tokens": 2550988800, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.2392854005098344, "seq_len": 4096} {"wall": 31708.158, "step": 8660, "loss": 2.257858162124952, "lm_loss": 2.257858162124952, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.31853005290031433, "tokens": 2553937920, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.2366736869017285, "seq_len": 4096} {"wall": 31912.839, "step": 8670, "loss": 2.2477586567401886, "lm_loss": 2.2477586567401886, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.30155667662620544, "tokens": 2556887040, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.258374109864235, "seq_len": 4096} {"wall": 32117.304, "step": 8680, "loss": 2.2482252518335977, "lm_loss": 2.2482252518335977, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3704703450202942, "tokens": 2559836160, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.2089474156498907, "seq_len": 4096} {"wall": 32322.356, "step": 8690, "loss": 2.1704713702201843, "lm_loss": 2.1704713702201843, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3447979688644409, "tokens": 2562785280, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.1816768700877827, "seq_len": 4096} {"wall": 32529.823, "eval_loss": 1.8958604037761688, "eval_ppl": 6.658275127410889, "step": 8700} {"wall": 32529.824, "step": 8700, "loss": 2.1704570253690085, "lm_loss": 2.1704570253690085, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3942050635814667, "tokens": 2565734400, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.478515625, "loss_smoothed": 2.1936829134821894, "seq_len": 4096} {"wall": 32814.225, "step": 8710, "loss": 2.0723077158133187, "lm_loss": 2.0723077158133187, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.28119415044784546, "tokens": 2568683520, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.2010496241350967, "seq_len": 4096} {"wall": 33019.528, "step": 8720, "loss": 2.0825964957475662, "lm_loss": 2.0825964957475662, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3575156331062317, "tokens": 2571632640, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.1611937788625557, "seq_len": 4096} {"wall": 33224.569, "step": 8730, "loss": 2.137632891535759, "lm_loss": 2.137632891535759, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.29383498430252075, "tokens": 2574581760, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.1692957900464536, "seq_len": 4096} {"wall": 33429.582, "step": 8740, "loss": 2.1754573037226996, "lm_loss": 2.1754573037226996, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2945241630077362, "tokens": 2577530880, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.1990611369411153, "seq_len": 4096} {"wall": 33634.352, "step": 8750, "loss": 2.258285711208979, "lm_loss": 2.258285711208979, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3128923773765564, "tokens": 2580480000, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.206527413179477, "seq_len": 4096} {"wall": 33839.104, "step": 8760, "loss": 2.110237861673037, "lm_loss": 2.110237861673037, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3499571681022644, "tokens": 2583429120, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.2059565082192423, "seq_len": 4096} {"wall": 34044.29, "step": 8770, "loss": 2.168568432331085, "lm_loss": 2.168568432331085, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.29228320717811584, "tokens": 2586378240, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.2106131722529727, "seq_len": 4096} {"wall": 34249.34, "step": 8780, "loss": 2.072224482893944, "lm_loss": 2.072224482893944, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3084263503551483, "tokens": 2589327360, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.206298779944579, "seq_len": 4096} {"wall": 34454.598, "step": 8790, "loss": 2.2927756309509277, "lm_loss": 2.2927756309509277, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3539814054965973, "tokens": 2592276480, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.2009719659884768, "seq_len": 4096} {"wall": 34661.656, "eval_loss": 1.8887925297021866, "eval_ppl": 6.611380577087402, "step": 8800} {"wall": 34661.658, "step": 8800, "loss": 2.2004939317703247, "lm_loss": 2.2004939317703247, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3101682960987091, "tokens": 2595225600, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.869140625, "loss_smoothed": 2.2188954735795656, "seq_len": 4096} {"wall": 34944.688, "step": 8810, "loss": 2.1803054213523865, "lm_loss": 2.1803054213523865, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.330489844083786, "tokens": 2598174720, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.191805800795555, "seq_len": 4096} {"wall": 35149.588, "step": 8820, "loss": 2.3282599498828254, "lm_loss": 2.3282599498828254, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.34954729676246643, "tokens": 2601123840, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.1516020223498344, "seq_len": 4096} {"wall": 35354.235, "step": 8830, "loss": 2.2395586421092353, "lm_loss": 2.2395586421092353, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.27429354190826416, "tokens": 2604072960, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.2267720282077788, "seq_len": 4096} {"wall": 35559.169, "step": 8840, "loss": 2.4195551723241806, "lm_loss": 2.4195551723241806, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 1.9352396726608276, "tokens": 2607022080, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.2075491530199844, "seq_len": 4096} {"wall": 35764.047, "step": 8850, "loss": 2.2244799832503, "lm_loss": 2.2244799832503, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.36368095874786377, "tokens": 2609971200, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.207415170719226, "seq_len": 4096} {"wall": 35968.897, "step": 8860, "loss": 2.1508665631214776, "lm_loss": 2.1508665631214776, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.24858562648296356, "tokens": 2612920320, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.2017408832907677, "seq_len": 4096} {"wall": 36173.525, "step": 8870, "loss": 2.25496898094813, "lm_loss": 2.25496898094813, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.4418840706348419, "tokens": 2615869440, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.1886405989527704, "seq_len": 4096} {"wall": 36378.477, "step": 8880, "loss": 2.1527448147535324, "lm_loss": 2.1527448147535324, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.31772464513778687, "tokens": 2618818560, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.205388567596674, "seq_len": 4096} {"wall": 36583.095, "step": 8890, "loss": 2.342189739147822, "lm_loss": 2.342189739147822, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.42404118180274963, "tokens": 2621767680, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.2292632661759852, "seq_len": 4096} {"wall": 36790.351, "eval_loss": 1.8898823708295822, "eval_ppl": 6.618589878082275, "step": 8900} {"wall": 36790.353, "step": 8900, "loss": 2.204180359840393, "lm_loss": 2.204180359840393, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3638288974761963, "tokens": 2624716800, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.322265625, "loss_smoothed": 2.199283715089162, "seq_len": 4096} {"wall": 37071.788, "step": 8910, "loss": 2.2513590157032013, "lm_loss": 2.2513590157032013, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.282009094953537, "tokens": 2627665920, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.177339280645053, "seq_len": 4096} {"wall": 37277.148, "step": 8920, "loss": 2.1943815598885217, "lm_loss": 2.1943815598885217, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.35300424695014954, "tokens": 2630615040, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.227088125050068, "seq_len": 4096} {"wall": 37481.872, "step": 8930, "loss": 2.339835430185, "lm_loss": 2.339835430185, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.4013414978981018, "tokens": 2633564160, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.2545548727114992, "seq_len": 4096} {"wall": 37687.119, "step": 8940, "loss": 2.1854360749324164, "lm_loss": 2.1854360749324164, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.37080591917037964, "tokens": 2636513280, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.2144772226611775, "seq_len": 4096} {"wall": 37892.0, "step": 8950, "loss": 2.3257524917523065, "lm_loss": 2.3257524917523065, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3271777927875519, "tokens": 2639462400, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.2009986266493797, "seq_len": 4096} {"wall": 38096.781, "step": 8960, "loss": 2.2108110835154853, "lm_loss": 2.2108110835154853, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.37348687648773193, "tokens": 2642411520, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.2010288799802464, "seq_len": 4096} {"wall": 38301.675, "step": 8970, "loss": 2.162339915831884, "lm_loss": 2.162339915831884, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.2698018550872803, "tokens": 2645360640, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.1893862441182135, "seq_len": 4096} {"wall": 38506.315, "step": 8980, "loss": 2.221137121319771, "lm_loss": 2.221137121319771, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3344132900238037, "tokens": 2648309760, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.253730969627698, "seq_len": 4096} {"wall": 38712.204, "step": 8990, "loss": 2.375624646743139, "lm_loss": 2.375624646743139, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.34155896306037903, "tokens": 2651258880, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.175163394212723, "seq_len": 4096} {"wall": 38919.824, "eval_loss": 1.8912978321313858, "eval_ppl": 6.627964973449707, "step": 9000} {"wall": 38919.826, "step": 9000, "loss": 2.1462546487649283, "lm_loss": 2.1462546487649283, "mtp_loss": 0.0, "lr": 0.025, "grad_norm": 0.3457578122615814, "tokens": 2654208000, "adamw_lr": 0.0004, "mem_alloc_gb": 19.520028591156006, "mem_reserv_gb": 84.361328125, "loss_smoothed": 2.2180759822328886, "seq_len": 4096}