diff --git "a/trainer_state.json" "b/trainer_state.json" new file mode 100644--- /dev/null +++ "b/trainer_state.json" @@ -0,0 +1,5975 @@ +{ + "best_global_step": null, + "best_metric": null, + "best_model_checkpoint": null, + "epoch": 2.0, + "eval_steps": 59, + "global_step": 586, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.003418803418803419, + "grad_norm": 0.7900828123092651, + "learning_rate": 0.0, + "loss": 1.4756, + "num_input_tokens_seen": 16096, + "step": 1, + "train_runtime": 4.4851, + "train_tokens_per_second": 3588.771 + }, + { + "epoch": 0.006837606837606838, + "grad_norm": 0.8550273180007935, + "learning_rate": 4e-05, + "loss": 1.4762, + "num_input_tokens_seen": 31264, + "step": 2, + "train_runtime": 6.6431, + "train_tokens_per_second": 4706.205 + }, + { + "epoch": 0.010256410256410256, + "grad_norm": 0.7560146450996399, + "learning_rate": 8e-05, + "loss": 1.3375, + "num_input_tokens_seen": 45584, + "step": 3, + "train_runtime": 8.7724, + "train_tokens_per_second": 5196.291 + }, + { + "epoch": 0.013675213675213675, + "grad_norm": 0.6870028376579285, + "learning_rate": 0.00012, + "loss": 1.0514, + "num_input_tokens_seen": 75072, + "step": 4, + "train_runtime": 13.0746, + "train_tokens_per_second": 5741.815 + }, + { + "epoch": 0.017094017094017096, + "grad_norm": 1.0243306159973145, + "learning_rate": 0.00016, + "loss": 1.1457, + "num_input_tokens_seen": 93296, + "step": 5, + "train_runtime": 15.6571, + "train_tokens_per_second": 5958.706 + }, + { + "epoch": 0.020512820512820513, + "grad_norm": 0.9480018019676208, + "learning_rate": 0.0002, + "loss": 0.8195, + "num_input_tokens_seen": 108880, + "step": 6, + "train_runtime": 17.9212, + "train_tokens_per_second": 6075.488 + }, + { + "epoch": 0.023931623931623933, + "grad_norm": 0.6540018320083618, + "learning_rate": 0.00019965576592082616, + "loss": 0.6109, + "num_input_tokens_seen": 124688, + "step": 7, + "train_runtime": 20.1568, + "train_tokens_per_second": 6185.907 + }, + { + "epoch": 0.02735042735042735, + "grad_norm": 0.28220847249031067, + "learning_rate": 0.00019931153184165235, + "loss": 0.5465, + "num_input_tokens_seen": 141552, + "step": 8, + "train_runtime": 22.5656, + "train_tokens_per_second": 6272.909 + }, + { + "epoch": 0.03076923076923077, + "grad_norm": 0.18244771659374237, + "learning_rate": 0.0001989672977624785, + "loss": 0.4196, + "num_input_tokens_seen": 156656, + "step": 9, + "train_runtime": 24.7703, + "train_tokens_per_second": 6324.342 + }, + { + "epoch": 0.03418803418803419, + "grad_norm": 0.19386130571365356, + "learning_rate": 0.00019862306368330465, + "loss": 0.4848, + "num_input_tokens_seen": 175424, + "step": 10, + "train_runtime": 27.4449, + "train_tokens_per_second": 6391.856 + }, + { + "epoch": 0.037606837606837605, + "grad_norm": 0.21845291554927826, + "learning_rate": 0.0001982788296041308, + "loss": 0.4163, + "num_input_tokens_seen": 191904, + "step": 11, + "train_runtime": 29.8233, + "train_tokens_per_second": 6434.711 + }, + { + "epoch": 0.041025641025641026, + "grad_norm": 0.2227926403284073, + "learning_rate": 0.000197934595524957, + "loss": 0.5241, + "num_input_tokens_seen": 204480, + "step": 12, + "train_runtime": 31.6741, + "train_tokens_per_second": 6455.753 + }, + { + "epoch": 0.044444444444444446, + "grad_norm": 0.14958512783050537, + "learning_rate": 0.00019759036144578314, + "loss": 0.3789, + "num_input_tokens_seen": 222928, + "step": 13, + "train_runtime": 34.2727, + "train_tokens_per_second": 6504.542 + }, + { + "epoch": 0.04786324786324787, + "grad_norm": 0.12511517107486725, + "learning_rate": 0.0001972461273666093, + "loss": 0.4255, + "num_input_tokens_seen": 243248, + "step": 14, + "train_runtime": 37.1309, + "train_tokens_per_second": 6551.089 + }, + { + "epoch": 0.05128205128205128, + "grad_norm": 0.12351791560649872, + "learning_rate": 0.00019690189328743548, + "loss": 0.4773, + "num_input_tokens_seen": 259984, + "step": 15, + "train_runtime": 39.5774, + "train_tokens_per_second": 6569.005 + }, + { + "epoch": 0.0547008547008547, + "grad_norm": 0.15381421148777008, + "learning_rate": 0.00019655765920826164, + "loss": 0.572, + "num_input_tokens_seen": 273488, + "step": 16, + "train_runtime": 41.5599, + "train_tokens_per_second": 6580.574 + }, + { + "epoch": 0.05811965811965812, + "grad_norm": 0.12012167274951935, + "learning_rate": 0.0001962134251290878, + "loss": 0.4824, + "num_input_tokens_seen": 289072, + "step": 17, + "train_runtime": 43.7917, + "train_tokens_per_second": 6601.074 + }, + { + "epoch": 0.06153846153846154, + "grad_norm": 0.12363950163125992, + "learning_rate": 0.00019586919104991394, + "loss": 0.4198, + "num_input_tokens_seen": 303408, + "step": 18, + "train_runtime": 45.8994, + "train_tokens_per_second": 6610.276 + }, + { + "epoch": 0.06495726495726496, + "grad_norm": 0.13709434866905212, + "learning_rate": 0.00019552495697074013, + "loss": 0.5045, + "num_input_tokens_seen": 321088, + "step": 19, + "train_runtime": 48.4273, + "train_tokens_per_second": 6630.309 + }, + { + "epoch": 0.06837606837606838, + "grad_norm": 0.11719892919063568, + "learning_rate": 0.00019518072289156628, + "loss": 0.444, + "num_input_tokens_seen": 339024, + "step": 20, + "train_runtime": 51.0105, + "train_tokens_per_second": 6646.155 + }, + { + "epoch": 0.07179487179487179, + "grad_norm": 0.1195915937423706, + "learning_rate": 0.00019483648881239243, + "loss": 0.4404, + "num_input_tokens_seen": 355648, + "step": 21, + "train_runtime": 53.4008, + "train_tokens_per_second": 6659.98 + }, + { + "epoch": 0.07521367521367521, + "grad_norm": 0.12735435366630554, + "learning_rate": 0.0001944922547332186, + "loss": 0.4253, + "num_input_tokens_seen": 377104, + "step": 22, + "train_runtime": 58.0087, + "train_tokens_per_second": 6500.816 + }, + { + "epoch": 0.07863247863247863, + "grad_norm": 0.13012833893299103, + "learning_rate": 0.00019414802065404477, + "loss": 0.4549, + "num_input_tokens_seen": 388496, + "step": 23, + "train_runtime": 59.7134, + "train_tokens_per_second": 6506.008 + }, + { + "epoch": 0.08205128205128205, + "grad_norm": 0.1187860295176506, + "learning_rate": 0.00019380378657487093, + "loss": 0.4447, + "num_input_tokens_seen": 405136, + "step": 24, + "train_runtime": 62.1253, + "train_tokens_per_second": 6521.268 + }, + { + "epoch": 0.08547008547008547, + "grad_norm": 0.12317577004432678, + "learning_rate": 0.0001934595524956971, + "loss": 0.4326, + "num_input_tokens_seen": 421088, + "step": 25, + "train_runtime": 64.4069, + "train_tokens_per_second": 6537.935 + }, + { + "epoch": 0.08888888888888889, + "grad_norm": 0.12564784288406372, + "learning_rate": 0.00019311531841652323, + "loss": 0.4521, + "num_input_tokens_seen": 435328, + "step": 26, + "train_runtime": 66.5043, + "train_tokens_per_second": 6545.86 + }, + { + "epoch": 0.09230769230769231, + "grad_norm": 0.10685139149427414, + "learning_rate": 0.00019277108433734942, + "loss": 0.4642, + "num_input_tokens_seen": 451968, + "step": 27, + "train_runtime": 68.9138, + "train_tokens_per_second": 6558.455 + }, + { + "epoch": 0.09572649572649573, + "grad_norm": 0.11094007641077042, + "learning_rate": 0.00019242685025817557, + "loss": 0.4877, + "num_input_tokens_seen": 464928, + "step": 28, + "train_runtime": 70.8316, + "train_tokens_per_second": 6563.853 + }, + { + "epoch": 0.09914529914529914, + "grad_norm": 0.1094178631901741, + "learning_rate": 0.00019208261617900175, + "loss": 0.5, + "num_input_tokens_seen": 479280, + "step": 29, + "train_runtime": 72.9759, + "train_tokens_per_second": 6567.645 + }, + { + "epoch": 0.10256410256410256, + "grad_norm": 0.09002909064292908, + "learning_rate": 0.00019173838209982788, + "loss": 0.3389, + "num_input_tokens_seen": 496320, + "step": 30, + "train_runtime": 75.4421, + "train_tokens_per_second": 6578.822 + }, + { + "epoch": 0.10598290598290598, + "grad_norm": 0.10585018992424011, + "learning_rate": 0.00019139414802065406, + "loss": 0.4548, + "num_input_tokens_seen": 511296, + "step": 31, + "train_runtime": 78.2326, + "train_tokens_per_second": 6535.59 + }, + { + "epoch": 0.1094017094017094, + "grad_norm": 0.09583117812871933, + "learning_rate": 0.00019104991394148021, + "loss": 0.4003, + "num_input_tokens_seen": 529008, + "step": 32, + "train_runtime": 80.7656, + "train_tokens_per_second": 6549.918 + }, + { + "epoch": 0.11282051282051282, + "grad_norm": 0.1044701486825943, + "learning_rate": 0.0001907056798623064, + "loss": 0.4673, + "num_input_tokens_seen": 544384, + "step": 33, + "train_runtime": 83.0622, + "train_tokens_per_second": 6553.935 + }, + { + "epoch": 0.11623931623931624, + "grad_norm": 0.10936785489320755, + "learning_rate": 0.00019036144578313252, + "loss": 0.4197, + "num_input_tokens_seen": 562096, + "step": 34, + "train_runtime": 85.6158, + "train_tokens_per_second": 6565.333 + }, + { + "epoch": 0.11965811965811966, + "grad_norm": 0.08397576957941055, + "learning_rate": 0.0001900172117039587, + "loss": 0.3889, + "num_input_tokens_seen": 580704, + "step": 35, + "train_runtime": 88.3072, + "train_tokens_per_second": 6575.955 + }, + { + "epoch": 0.12307692307692308, + "grad_norm": 0.10643361508846283, + "learning_rate": 0.00018967297762478486, + "loss": 0.474, + "num_input_tokens_seen": 596128, + "step": 36, + "train_runtime": 90.568, + "train_tokens_per_second": 6582.102 + }, + { + "epoch": 0.1264957264957265, + "grad_norm": 0.11568325012922287, + "learning_rate": 0.00018932874354561104, + "loss": 0.4805, + "num_input_tokens_seen": 609488, + "step": 37, + "train_runtime": 92.5585, + "train_tokens_per_second": 6584.897 + }, + { + "epoch": 0.12991452991452992, + "grad_norm": 0.1005433201789856, + "learning_rate": 0.00018898450946643717, + "loss": 0.3679, + "num_input_tokens_seen": 624592, + "step": 38, + "train_runtime": 94.7955, + "train_tokens_per_second": 6588.837 + }, + { + "epoch": 0.13333333333333333, + "grad_norm": 0.0977381095290184, + "learning_rate": 0.00018864027538726335, + "loss": 0.4141, + "num_input_tokens_seen": 639840, + "step": 39, + "train_runtime": 97.0423, + "train_tokens_per_second": 6593.411 + }, + { + "epoch": 0.13675213675213677, + "grad_norm": 0.11257043480873108, + "learning_rate": 0.0001882960413080895, + "loss": 0.5211, + "num_input_tokens_seen": 652048, + "step": 40, + "train_runtime": 98.9078, + "train_tokens_per_second": 6592.481 + }, + { + "epoch": 0.14017094017094017, + "grad_norm": 0.10306330025196075, + "learning_rate": 0.00018795180722891569, + "loss": 0.3752, + "num_input_tokens_seen": 670112, + "step": 41, + "train_runtime": 101.5158, + "train_tokens_per_second": 6601.063 + }, + { + "epoch": 0.14358974358974358, + "grad_norm": 0.07424470037221909, + "learning_rate": 0.00018760757314974184, + "loss": 0.2781, + "num_input_tokens_seen": 705168, + "step": 42, + "train_runtime": 106.5454, + "train_tokens_per_second": 6618.476 + }, + { + "epoch": 0.147008547008547, + "grad_norm": 0.09572459757328033, + "learning_rate": 0.000187263339070568, + "loss": 0.5014, + "num_input_tokens_seen": 720320, + "step": 43, + "train_runtime": 108.8037, + "train_tokens_per_second": 6620.362 + }, + { + "epoch": 0.15042735042735042, + "grad_norm": 0.09543762356042862, + "learning_rate": 0.00018691910499139415, + "loss": 0.4411, + "num_input_tokens_seen": 736032, + "step": 44, + "train_runtime": 111.1227, + "train_tokens_per_second": 6623.596 + }, + { + "epoch": 0.15384615384615385, + "grad_norm": 0.09599870443344116, + "learning_rate": 0.00018657487091222033, + "loss": 0.3435, + "num_input_tokens_seen": 750576, + "step": 45, + "train_runtime": 113.3125, + "train_tokens_per_second": 6623.948 + }, + { + "epoch": 0.15726495726495726, + "grad_norm": 0.0982552245259285, + "learning_rate": 0.00018623063683304649, + "loss": 0.5206, + "num_input_tokens_seen": 768944, + "step": 46, + "train_runtime": 115.975, + "train_tokens_per_second": 6630.256 + }, + { + "epoch": 0.1606837606837607, + "grad_norm": 0.10187075287103653, + "learning_rate": 0.00018588640275387264, + "loss": 0.4408, + "num_input_tokens_seen": 784144, + "step": 47, + "train_runtime": 118.2082, + "train_tokens_per_second": 6633.584 + }, + { + "epoch": 0.1641025641025641, + "grad_norm": 0.0859491378068924, + "learning_rate": 0.0001855421686746988, + "loss": 0.4154, + "num_input_tokens_seen": 804272, + "step": 48, + "train_runtime": 121.1095, + "train_tokens_per_second": 6640.864 + }, + { + "epoch": 0.1675213675213675, + "grad_norm": 0.08325184136629105, + "learning_rate": 0.00018519793459552498, + "loss": 0.4557, + "num_input_tokens_seen": 818224, + "step": 49, + "train_runtime": 123.1709, + "train_tokens_per_second": 6642.997 + }, + { + "epoch": 0.17094017094017094, + "grad_norm": 0.09160371869802475, + "learning_rate": 0.00018485370051635113, + "loss": 0.3976, + "num_input_tokens_seen": 837600, + "step": 50, + "train_runtime": 125.9707, + "train_tokens_per_second": 6649.167 + }, + { + "epoch": 0.17435897435897435, + "grad_norm": 0.08669862151145935, + "learning_rate": 0.00018450946643717729, + "loss": 0.4597, + "num_input_tokens_seen": 855360, + "step": 51, + "train_runtime": 128.5702, + "train_tokens_per_second": 6652.866 + }, + { + "epoch": 0.17777777777777778, + "grad_norm": 0.10008913278579712, + "learning_rate": 0.00018416523235800344, + "loss": 0.4608, + "num_input_tokens_seen": 869472, + "step": 52, + "train_runtime": 130.6754, + "train_tokens_per_second": 6653.679 + }, + { + "epoch": 0.1811965811965812, + "grad_norm": 0.08692440390586853, + "learning_rate": 0.00018382099827882962, + "loss": 0.4229, + "num_input_tokens_seen": 887936, + "step": 53, + "train_runtime": 133.3545, + "train_tokens_per_second": 6658.462 + }, + { + "epoch": 0.18461538461538463, + "grad_norm": 0.12537947297096252, + "learning_rate": 0.00018347676419965578, + "loss": 0.5767, + "num_input_tokens_seen": 902240, + "step": 54, + "train_runtime": 135.5092, + "train_tokens_per_second": 6658.146 + }, + { + "epoch": 0.18803418803418803, + "grad_norm": 0.08537557721138, + "learning_rate": 0.00018313253012048193, + "loss": 0.4519, + "num_input_tokens_seen": 920688, + "step": 55, + "train_runtime": 138.1817, + "train_tokens_per_second": 6662.881 + }, + { + "epoch": 0.19145299145299147, + "grad_norm": 0.09388847649097443, + "learning_rate": 0.00018278829604130808, + "loss": 0.4083, + "num_input_tokens_seen": 937024, + "step": 56, + "train_runtime": 140.574, + "train_tokens_per_second": 6665.697 + }, + { + "epoch": 0.19487179487179487, + "grad_norm": 0.12476610392332077, + "learning_rate": 0.00018244406196213427, + "loss": 0.457, + "num_input_tokens_seen": 949568, + "step": 57, + "train_runtime": 142.4521, + "train_tokens_per_second": 6665.875 + }, + { + "epoch": 0.19829059829059828, + "grad_norm": 0.11227541416883469, + "learning_rate": 0.00018209982788296042, + "loss": 0.4666, + "num_input_tokens_seen": 964000, + "step": 58, + "train_runtime": 144.5997, + "train_tokens_per_second": 6666.681 + }, + { + "epoch": 0.20170940170940171, + "grad_norm": 0.10149527341127396, + "learning_rate": 0.0001817555938037866, + "loss": 0.473, + "num_input_tokens_seen": 976560, + "step": 59, + "train_runtime": 146.4793, + "train_tokens_per_second": 6666.883 + }, + { + "epoch": 0.20170940170940171, + "eval_loss": 0.4301406145095825, + "eval_runtime": 11.8663, + "eval_samples_per_second": 84.104, + "eval_steps_per_second": 5.309, + "num_input_tokens_seen": 976560, + "step": 59 + }, + { + "epoch": 0.20512820512820512, + "grad_norm": 0.0855872854590416, + "learning_rate": 0.00018141135972461273, + "loss": 0.4248, + "num_input_tokens_seen": 992944, + "step": 60, + "train_runtime": 160.7566, + "train_tokens_per_second": 6176.693 + }, + { + "epoch": 0.20854700854700856, + "grad_norm": 0.11866864562034607, + "learning_rate": 0.0001810671256454389, + "loss": 0.537, + "num_input_tokens_seen": 1005920, + "step": 61, + "train_runtime": 163.2145, + "train_tokens_per_second": 6163.179 + }, + { + "epoch": 0.21196581196581196, + "grad_norm": 0.0955129861831665, + "learning_rate": 0.00018072289156626507, + "loss": 0.3983, + "num_input_tokens_seen": 1022704, + "step": 62, + "train_runtime": 165.6518, + "train_tokens_per_second": 6173.817 + }, + { + "epoch": 0.2153846153846154, + "grad_norm": 0.11176180094480515, + "learning_rate": 0.00018037865748709125, + "loss": 0.4392, + "num_input_tokens_seen": 1038928, + "step": 63, + "train_runtime": 168.0352, + "train_tokens_per_second": 6182.8 + }, + { + "epoch": 0.2188034188034188, + "grad_norm": 0.08749086409807205, + "learning_rate": 0.00018003442340791737, + "loss": 0.412, + "num_input_tokens_seen": 1057504, + "step": 64, + "train_runtime": 170.7322, + "train_tokens_per_second": 6193.934 + }, + { + "epoch": 0.2222222222222222, + "grad_norm": 0.10065372288227081, + "learning_rate": 0.00017969018932874356, + "loss": 0.4496, + "num_input_tokens_seen": 1071952, + "step": 65, + "train_runtime": 172.8869, + "train_tokens_per_second": 6200.307 + }, + { + "epoch": 0.22564102564102564, + "grad_norm": 0.09161035716533661, + "learning_rate": 0.0001793459552495697, + "loss": 0.4995, + "num_input_tokens_seen": 1091408, + "step": 66, + "train_runtime": 175.7012, + "train_tokens_per_second": 6211.727 + }, + { + "epoch": 0.22905982905982905, + "grad_norm": 0.08862069249153137, + "learning_rate": 0.0001790017211703959, + "loss": 0.4157, + "num_input_tokens_seen": 1107040, + "step": 67, + "train_runtime": 178.0171, + "train_tokens_per_second": 6218.728 + }, + { + "epoch": 0.23247863247863249, + "grad_norm": 0.09428606927394867, + "learning_rate": 0.00017865748709122202, + "loss": 0.4845, + "num_input_tokens_seen": 1121712, + "step": 68, + "train_runtime": 180.2346, + "train_tokens_per_second": 6223.624 + }, + { + "epoch": 0.2358974358974359, + "grad_norm": 0.1100679412484169, + "learning_rate": 0.0001783132530120482, + "loss": 0.4706, + "num_input_tokens_seen": 1135664, + "step": 69, + "train_runtime": 182.3342, + "train_tokens_per_second": 6228.477 + }, + { + "epoch": 0.23931623931623933, + "grad_norm": 0.10081252455711365, + "learning_rate": 0.00017796901893287436, + "loss": 0.4073, + "num_input_tokens_seen": 1154224, + "step": 70, + "train_runtime": 185.0219, + "train_tokens_per_second": 6238.31 + }, + { + "epoch": 0.24273504273504273, + "grad_norm": 0.08933307230472565, + "learning_rate": 0.00017762478485370054, + "loss": 0.4457, + "num_input_tokens_seen": 1171424, + "step": 71, + "train_runtime": 187.5554, + "train_tokens_per_second": 6245.75 + }, + { + "epoch": 0.24615384615384617, + "grad_norm": 0.0866895243525505, + "learning_rate": 0.00017728055077452666, + "loss": 0.3439, + "num_input_tokens_seen": 1189504, + "step": 72, + "train_runtime": 190.1853, + "train_tokens_per_second": 6254.449 + }, + { + "epoch": 0.24957264957264957, + "grad_norm": 0.09268927574157715, + "learning_rate": 0.00017693631669535285, + "loss": 0.4454, + "num_input_tokens_seen": 1207600, + "step": 73, + "train_runtime": 192.8217, + "train_tokens_per_second": 6262.782 + }, + { + "epoch": 0.252991452991453, + "grad_norm": 0.08207046240568161, + "learning_rate": 0.000176592082616179, + "loss": 0.4721, + "num_input_tokens_seen": 1227552, + "step": 74, + "train_runtime": 195.7293, + "train_tokens_per_second": 6271.682 + }, + { + "epoch": 0.2564102564102564, + "grad_norm": 0.08926599472761154, + "learning_rate": 0.00017624784853700518, + "loss": 0.5064, + "num_input_tokens_seen": 1245984, + "step": 75, + "train_runtime": 198.439, + "train_tokens_per_second": 6278.928 + }, + { + "epoch": 0.25982905982905985, + "grad_norm": 0.10555540025234222, + "learning_rate": 0.00017590361445783134, + "loss": 0.4264, + "num_input_tokens_seen": 1259536, + "step": 76, + "train_runtime": 200.4904, + "train_tokens_per_second": 6282.277 + }, + { + "epoch": 0.26324786324786326, + "grad_norm": 0.09617225080728531, + "learning_rate": 0.0001755593803786575, + "loss": 0.4686, + "num_input_tokens_seen": 1278768, + "step": 77, + "train_runtime": 203.2722, + "train_tokens_per_second": 6290.914 + }, + { + "epoch": 0.26666666666666666, + "grad_norm": 0.12434384226799011, + "learning_rate": 0.00017521514629948364, + "loss": 0.5939, + "num_input_tokens_seen": 1292928, + "step": 78, + "train_runtime": 205.3845, + "train_tokens_per_second": 6295.159 + }, + { + "epoch": 0.27008547008547007, + "grad_norm": 0.1012965589761734, + "learning_rate": 0.00017487091222030983, + "loss": 0.403, + "num_input_tokens_seen": 1310160, + "step": 79, + "train_runtime": 207.8738, + "train_tokens_per_second": 6302.671 + }, + { + "epoch": 0.27350427350427353, + "grad_norm": 0.08980307728052139, + "learning_rate": 0.00017452667814113598, + "loss": 0.3688, + "num_input_tokens_seen": 1328608, + "step": 80, + "train_runtime": 210.5562, + "train_tokens_per_second": 6309.993 + }, + { + "epoch": 0.27692307692307694, + "grad_norm": 0.09891822189092636, + "learning_rate": 0.00017418244406196214, + "loss": 0.4751, + "num_input_tokens_seen": 1346016, + "step": 81, + "train_runtime": 213.0979, + "train_tokens_per_second": 6316.422 + }, + { + "epoch": 0.28034188034188035, + "grad_norm": 0.10093598067760468, + "learning_rate": 0.0001738382099827883, + "loss": 0.5395, + "num_input_tokens_seen": 1359184, + "step": 82, + "train_runtime": 215.139, + "train_tokens_per_second": 6317.702 + }, + { + "epoch": 0.28376068376068375, + "grad_norm": 0.09215115010738373, + "learning_rate": 0.00017349397590361447, + "loss": 0.4067, + "num_input_tokens_seen": 1374384, + "step": 83, + "train_runtime": 217.3721, + "train_tokens_per_second": 6322.726 + }, + { + "epoch": 0.28717948717948716, + "grad_norm": 0.09286817163228989, + "learning_rate": 0.00017314974182444063, + "loss": 0.4327, + "num_input_tokens_seen": 1395312, + "step": 84, + "train_runtime": 220.3981, + "train_tokens_per_second": 6330.871 + }, + { + "epoch": 0.2905982905982906, + "grad_norm": 0.10269108414649963, + "learning_rate": 0.00017280550774526678, + "loss": 0.4137, + "num_input_tokens_seen": 1410480, + "step": 85, + "train_runtime": 222.6432, + "train_tokens_per_second": 6335.159 + }, + { + "epoch": 0.294017094017094, + "grad_norm": 0.11593605577945709, + "learning_rate": 0.00017246127366609296, + "loss": 0.488, + "num_input_tokens_seen": 1423008, + "step": 86, + "train_runtime": 224.5645, + "train_tokens_per_second": 6336.746 + }, + { + "epoch": 0.29743589743589743, + "grad_norm": 0.09121434390544891, + "learning_rate": 0.00017211703958691912, + "loss": 0.4372, + "num_input_tokens_seen": 1443008, + "step": 87, + "train_runtime": 227.4516, + "train_tokens_per_second": 6344.241 + }, + { + "epoch": 0.30085470085470084, + "grad_norm": 0.11658863723278046, + "learning_rate": 0.00017177280550774527, + "loss": 0.4129, + "num_input_tokens_seen": 1457936, + "step": 88, + "train_runtime": 229.6543, + "train_tokens_per_second": 6348.393 + }, + { + "epoch": 0.30427350427350425, + "grad_norm": 0.1077234074473381, + "learning_rate": 0.00017142857142857143, + "loss": 0.4311, + "num_input_tokens_seen": 1471536, + "step": 89, + "train_runtime": 231.6871, + "train_tokens_per_second": 6351.395 + }, + { + "epoch": 0.3076923076923077, + "grad_norm": 0.10258413106203079, + "learning_rate": 0.0001710843373493976, + "loss": 0.4856, + "num_input_tokens_seen": 1486080, + "step": 90, + "train_runtime": 233.8589, + "train_tokens_per_second": 6354.6 + }, + { + "epoch": 0.3111111111111111, + "grad_norm": 0.09048861265182495, + "learning_rate": 0.00017074010327022376, + "loss": 0.4573, + "num_input_tokens_seen": 1501360, + "step": 91, + "train_runtime": 236.6697, + "train_tokens_per_second": 6343.694 + }, + { + "epoch": 0.3145299145299145, + "grad_norm": 0.09616488963365555, + "learning_rate": 0.00017039586919104992, + "loss": 0.3533, + "num_input_tokens_seen": 1521744, + "step": 92, + "train_runtime": 239.5612, + "train_tokens_per_second": 6352.215 + }, + { + "epoch": 0.31794871794871793, + "grad_norm": 0.10119186341762543, + "learning_rate": 0.00017005163511187607, + "loss": 0.4687, + "num_input_tokens_seen": 1538640, + "step": 93, + "train_runtime": 242.0203, + "train_tokens_per_second": 6357.484 + }, + { + "epoch": 0.3213675213675214, + "grad_norm": 0.103726826608181, + "learning_rate": 0.00016970740103270225, + "loss": 0.43, + "num_input_tokens_seen": 1553200, + "step": 94, + "train_runtime": 244.1818, + "train_tokens_per_second": 6360.835 + }, + { + "epoch": 0.3247863247863248, + "grad_norm": 0.09369926899671555, + "learning_rate": 0.0001693631669535284, + "loss": 0.4155, + "num_input_tokens_seen": 1571952, + "step": 95, + "train_runtime": 246.9205, + "train_tokens_per_second": 6366.227 + }, + { + "epoch": 0.3282051282051282, + "grad_norm": 0.09061973541975021, + "learning_rate": 0.0001690189328743546, + "loss": 0.4044, + "num_input_tokens_seen": 1584416, + "step": 96, + "train_runtime": 248.812, + "train_tokens_per_second": 6367.924 + }, + { + "epoch": 0.3316239316239316, + "grad_norm": 0.08632508665323257, + "learning_rate": 0.00016867469879518074, + "loss": 0.405, + "num_input_tokens_seen": 1604672, + "step": 97, + "train_runtime": 251.7603, + "train_tokens_per_second": 6373.808 + }, + { + "epoch": 0.335042735042735, + "grad_norm": 0.08924511075019836, + "learning_rate": 0.0001683304647160069, + "loss": 0.4357, + "num_input_tokens_seen": 1622608, + "step": 98, + "train_runtime": 254.3946, + "train_tokens_per_second": 6378.311 + }, + { + "epoch": 0.3384615384615385, + "grad_norm": 0.10018568485975266, + "learning_rate": 0.00016798623063683305, + "loss": 0.4007, + "num_input_tokens_seen": 1636752, + "step": 99, + "train_runtime": 256.4999, + "train_tokens_per_second": 6381.101 + }, + { + "epoch": 0.3418803418803419, + "grad_norm": 0.08820787072181702, + "learning_rate": 0.00016764199655765923, + "loss": 0.3778, + "num_input_tokens_seen": 1649408, + "step": 100, + "train_runtime": 258.3868, + "train_tokens_per_second": 6383.484 + }, + { + "epoch": 0.3452991452991453, + "grad_norm": 0.10735058784484863, + "learning_rate": 0.0001672977624784854, + "loss": 0.5828, + "num_input_tokens_seen": 1665776, + "step": 101, + "train_runtime": 260.7883, + "train_tokens_per_second": 6387.466 + }, + { + "epoch": 0.3487179487179487, + "grad_norm": 0.09399764984846115, + "learning_rate": 0.00016695352839931154, + "loss": 0.4238, + "num_input_tokens_seen": 1688816, + "step": 102, + "train_runtime": 264.0814, + "train_tokens_per_second": 6395.059 + }, + { + "epoch": 0.35213675213675216, + "grad_norm": 0.0975056141614914, + "learning_rate": 0.0001666092943201377, + "loss": 0.4374, + "num_input_tokens_seen": 1703456, + "step": 103, + "train_runtime": 266.2978, + "train_tokens_per_second": 6396.807 + }, + { + "epoch": 0.35555555555555557, + "grad_norm": 0.09030232578516006, + "learning_rate": 0.00016626506024096388, + "loss": 0.448, + "num_input_tokens_seen": 1724528, + "step": 104, + "train_runtime": 269.3165, + "train_tokens_per_second": 6403.35 + }, + { + "epoch": 0.358974358974359, + "grad_norm": 0.10417474806308746, + "learning_rate": 0.00016592082616179003, + "loss": 0.4109, + "num_input_tokens_seen": 1740768, + "step": 105, + "train_runtime": 271.6966, + "train_tokens_per_second": 6407.029 + }, + { + "epoch": 0.3623931623931624, + "grad_norm": 0.10844124108552933, + "learning_rate": 0.00016557659208261619, + "loss": 0.4165, + "num_input_tokens_seen": 1753632, + "step": 106, + "train_runtime": 273.6236, + "train_tokens_per_second": 6408.921 + }, + { + "epoch": 0.3658119658119658, + "grad_norm": 0.08827276527881622, + "learning_rate": 0.00016523235800344234, + "loss": 0.4233, + "num_input_tokens_seen": 1768480, + "step": 107, + "train_runtime": 275.8419, + "train_tokens_per_second": 6411.209 + }, + { + "epoch": 0.36923076923076925, + "grad_norm": 0.10917767882347107, + "learning_rate": 0.00016488812392426852, + "loss": 0.4624, + "num_input_tokens_seen": 1782800, + "step": 108, + "train_runtime": 277.9694, + "train_tokens_per_second": 6413.656 + }, + { + "epoch": 0.37264957264957266, + "grad_norm": 0.10163897275924683, + "learning_rate": 0.00016454388984509468, + "loss": 0.4553, + "num_input_tokens_seen": 1799008, + "step": 109, + "train_runtime": 280.3434, + "train_tokens_per_second": 6417.158 + }, + { + "epoch": 0.37606837606837606, + "grad_norm": 0.10832331329584122, + "learning_rate": 0.00016419965576592083, + "loss": 0.359, + "num_input_tokens_seen": 1811792, + "step": 110, + "train_runtime": 282.2797, + "train_tokens_per_second": 6418.429 + }, + { + "epoch": 0.37948717948717947, + "grad_norm": 0.08609069138765335, + "learning_rate": 0.00016385542168674699, + "loss": 0.4809, + "num_input_tokens_seen": 1828512, + "step": 111, + "train_runtime": 284.7476, + "train_tokens_per_second": 6421.518 + }, + { + "epoch": 0.38290598290598293, + "grad_norm": 0.11904721707105637, + "learning_rate": 0.00016351118760757317, + "loss": 0.4375, + "num_input_tokens_seen": 1849616, + "step": 112, + "train_runtime": 287.8004, + "train_tokens_per_second": 6426.732 + }, + { + "epoch": 0.38632478632478634, + "grad_norm": 0.09531164169311523, + "learning_rate": 0.00016316695352839932, + "loss": 0.4484, + "num_input_tokens_seen": 1866480, + "step": 113, + "train_runtime": 290.2693, + "train_tokens_per_second": 6430.167 + }, + { + "epoch": 0.38974358974358975, + "grad_norm": 0.13019715249538422, + "learning_rate": 0.0001628227194492255, + "loss": 0.4285, + "num_input_tokens_seen": 1881008, + "step": 114, + "train_runtime": 292.3998, + "train_tokens_per_second": 6433.0 + }, + { + "epoch": 0.39316239316239315, + "grad_norm": 0.1122983917593956, + "learning_rate": 0.00016247848537005163, + "loss": 0.5555, + "num_input_tokens_seen": 1893264, + "step": 115, + "train_runtime": 294.2603, + "train_tokens_per_second": 6433.977 + }, + { + "epoch": 0.39658119658119656, + "grad_norm": 0.0886179730296135, + "learning_rate": 0.0001621342512908778, + "loss": 0.4052, + "num_input_tokens_seen": 1911168, + "step": 116, + "train_runtime": 296.8399, + "train_tokens_per_second": 6438.379 + }, + { + "epoch": 0.4, + "grad_norm": 0.11588186770677567, + "learning_rate": 0.00016179001721170397, + "loss": 0.4881, + "num_input_tokens_seen": 1922320, + "step": 117, + "train_runtime": 298.5272, + "train_tokens_per_second": 6439.346 + }, + { + "epoch": 0.40341880341880343, + "grad_norm": 0.09329588711261749, + "learning_rate": 0.00016144578313253015, + "loss": 0.392, + "num_input_tokens_seen": 1935888, + "step": 118, + "train_runtime": 300.5442, + "train_tokens_per_second": 6441.276 + }, + { + "epoch": 0.40341880341880343, + "eval_loss": 0.4257439076900482, + "eval_runtime": 11.8891, + "eval_samples_per_second": 83.943, + "eval_steps_per_second": 5.299, + "num_input_tokens_seen": 1935888, + "step": 118 + }, + { + "epoch": 0.40683760683760684, + "grad_norm": 0.09967512637376785, + "learning_rate": 0.00016110154905335628, + "loss": 0.4716, + "num_input_tokens_seen": 1949360, + "step": 119, + "train_runtime": 314.4663, + "train_tokens_per_second": 6198.947 + }, + { + "epoch": 0.41025641025641024, + "grad_norm": 0.08704519271850586, + "learning_rate": 0.00016075731497418246, + "loss": 0.3573, + "num_input_tokens_seen": 1965152, + "step": 120, + "train_runtime": 316.7907, + "train_tokens_per_second": 6203.314 + }, + { + "epoch": 0.41367521367521365, + "grad_norm": 0.09572669863700867, + "learning_rate": 0.0001604130808950086, + "loss": 0.3742, + "num_input_tokens_seen": 1981632, + "step": 121, + "train_runtime": 319.7181, + "train_tokens_per_second": 6198.06 + }, + { + "epoch": 0.4170940170940171, + "grad_norm": 0.09697537869215012, + "learning_rate": 0.0001600688468158348, + "loss": 0.4811, + "num_input_tokens_seen": 1999488, + "step": 122, + "train_runtime": 322.3203, + "train_tokens_per_second": 6203.42 + }, + { + "epoch": 0.4205128205128205, + "grad_norm": 0.10042616724967957, + "learning_rate": 0.00015972461273666092, + "loss": 0.4146, + "num_input_tokens_seen": 2015360, + "step": 123, + "train_runtime": 324.66, + "train_tokens_per_second": 6207.601 + }, + { + "epoch": 0.4239316239316239, + "grad_norm": 0.09732874482870102, + "learning_rate": 0.0001593803786574871, + "loss": 0.3647, + "num_input_tokens_seen": 2029840, + "step": 124, + "train_runtime": 326.8173, + "train_tokens_per_second": 6210.931 + }, + { + "epoch": 0.42735042735042733, + "grad_norm": 0.0899026021361351, + "learning_rate": 0.00015903614457831326, + "loss": 0.3614, + "num_input_tokens_seen": 2043888, + "step": 125, + "train_runtime": 328.9033, + "train_tokens_per_second": 6214.251 + }, + { + "epoch": 0.4307692307692308, + "grad_norm": 0.1127215027809143, + "learning_rate": 0.00015869191049913944, + "loss": 0.521, + "num_input_tokens_seen": 2056784, + "step": 126, + "train_runtime": 330.847, + "train_tokens_per_second": 6216.723 + }, + { + "epoch": 0.4341880341880342, + "grad_norm": 0.10900437831878662, + "learning_rate": 0.00015834767641996557, + "loss": 0.5004, + "num_input_tokens_seen": 2073152, + "step": 127, + "train_runtime": 333.236, + "train_tokens_per_second": 6221.272 + }, + { + "epoch": 0.4376068376068376, + "grad_norm": 0.10284293442964554, + "learning_rate": 0.00015800344234079175, + "loss": 0.4758, + "num_input_tokens_seen": 2086832, + "step": 128, + "train_runtime": 335.3261, + "train_tokens_per_second": 6223.291 + }, + { + "epoch": 0.441025641025641, + "grad_norm": 0.08526846766471863, + "learning_rate": 0.0001576592082616179, + "loss": 0.3601, + "num_input_tokens_seen": 2102704, + "step": 129, + "train_runtime": 337.6773, + "train_tokens_per_second": 6226.962 + }, + { + "epoch": 0.4444444444444444, + "grad_norm": 0.0935521200299263, + "learning_rate": 0.00015731497418244408, + "loss": 0.4191, + "num_input_tokens_seen": 2120672, + "step": 130, + "train_runtime": 340.2901, + "train_tokens_per_second": 6231.953 + }, + { + "epoch": 0.4478632478632479, + "grad_norm": 0.09808618575334549, + "learning_rate": 0.00015697074010327024, + "loss": 0.432, + "num_input_tokens_seen": 2134528, + "step": 131, + "train_runtime": 342.374, + "train_tokens_per_second": 6234.492 + }, + { + "epoch": 0.4512820512820513, + "grad_norm": 0.08662474900484085, + "learning_rate": 0.0001566265060240964, + "loss": 0.4311, + "num_input_tokens_seen": 2150768, + "step": 132, + "train_runtime": 344.7887, + "train_tokens_per_second": 6237.931 + }, + { + "epoch": 0.4547008547008547, + "grad_norm": 0.09085840731859207, + "learning_rate": 0.00015628227194492255, + "loss": 0.4223, + "num_input_tokens_seen": 2171536, + "step": 133, + "train_runtime": 347.7754, + "train_tokens_per_second": 6244.076 + }, + { + "epoch": 0.4581196581196581, + "grad_norm": 0.1134534478187561, + "learning_rate": 0.00015593803786574873, + "loss": 0.4995, + "num_input_tokens_seen": 2187392, + "step": 134, + "train_runtime": 350.1232, + "train_tokens_per_second": 6247.493 + }, + { + "epoch": 0.46153846153846156, + "grad_norm": 0.0914321318268776, + "learning_rate": 0.00015559380378657488, + "loss": 0.3414, + "num_input_tokens_seen": 2208736, + "step": 135, + "train_runtime": 353.1913, + "train_tokens_per_second": 6253.653 + }, + { + "epoch": 0.46495726495726497, + "grad_norm": 0.09697174280881882, + "learning_rate": 0.00015524956970740104, + "loss": 0.4192, + "num_input_tokens_seen": 2221520, + "step": 136, + "train_runtime": 355.1021, + "train_tokens_per_second": 6256.004 + }, + { + "epoch": 0.4683760683760684, + "grad_norm": 0.09903775900602341, + "learning_rate": 0.0001549053356282272, + "loss": 0.4107, + "num_input_tokens_seen": 2234736, + "step": 137, + "train_runtime": 357.0576, + "train_tokens_per_second": 6258.755 + }, + { + "epoch": 0.4717948717948718, + "grad_norm": 0.10500401258468628, + "learning_rate": 0.00015456110154905337, + "loss": 0.4507, + "num_input_tokens_seen": 2246432, + "step": 138, + "train_runtime": 358.8165, + "train_tokens_per_second": 6260.671 + }, + { + "epoch": 0.4752136752136752, + "grad_norm": 0.10770396888256073, + "learning_rate": 0.00015421686746987953, + "loss": 0.4844, + "num_input_tokens_seen": 2264496, + "step": 139, + "train_runtime": 361.4046, + "train_tokens_per_second": 6265.82 + }, + { + "epoch": 0.47863247863247865, + "grad_norm": 0.09635472297668457, + "learning_rate": 0.00015387263339070568, + "loss": 0.3587, + "num_input_tokens_seen": 2279664, + "step": 140, + "train_runtime": 363.662, + "train_tokens_per_second": 6268.634 + }, + { + "epoch": 0.48205128205128206, + "grad_norm": 0.10260126739740372, + "learning_rate": 0.00015352839931153184, + "loss": 0.4402, + "num_input_tokens_seen": 2294976, + "step": 141, + "train_runtime": 365.9416, + "train_tokens_per_second": 6271.427 + }, + { + "epoch": 0.48547008547008547, + "grad_norm": 0.11078032851219177, + "learning_rate": 0.00015318416523235802, + "loss": 0.4819, + "num_input_tokens_seen": 2310256, + "step": 142, + "train_runtime": 368.2151, + "train_tokens_per_second": 6274.202 + }, + { + "epoch": 0.4888888888888889, + "grad_norm": 0.08946855366230011, + "learning_rate": 0.00015283993115318417, + "loss": 0.3455, + "num_input_tokens_seen": 2332160, + "step": 143, + "train_runtime": 371.3782, + "train_tokens_per_second": 6279.745 + }, + { + "epoch": 0.49230769230769234, + "grad_norm": 0.089847132563591, + "learning_rate": 0.00015249569707401033, + "loss": 0.3624, + "num_input_tokens_seen": 2346720, + "step": 144, + "train_runtime": 373.5395, + "train_tokens_per_second": 6282.389 + }, + { + "epoch": 0.49572649572649574, + "grad_norm": 0.10344398766756058, + "learning_rate": 0.00015215146299483648, + "loss": 0.4373, + "num_input_tokens_seen": 2360992, + "step": 145, + "train_runtime": 375.6328, + "train_tokens_per_second": 6285.372 + }, + { + "epoch": 0.49914529914529915, + "grad_norm": 0.10121628642082214, + "learning_rate": 0.00015180722891566266, + "loss": 0.4471, + "num_input_tokens_seen": 2375680, + "step": 146, + "train_runtime": 377.7982, + "train_tokens_per_second": 6288.226 + }, + { + "epoch": 0.5025641025641026, + "grad_norm": 0.08698102086782455, + "learning_rate": 0.00015146299483648882, + "loss": 0.3908, + "num_input_tokens_seen": 2397024, + "step": 147, + "train_runtime": 380.8567, + "train_tokens_per_second": 6293.768 + }, + { + "epoch": 0.505982905982906, + "grad_norm": 0.08766576647758484, + "learning_rate": 0.000151118760757315, + "loss": 0.3881, + "num_input_tokens_seen": 2414128, + "step": 148, + "train_runtime": 383.3371, + "train_tokens_per_second": 6297.663 + }, + { + "epoch": 0.5094017094017094, + "grad_norm": 0.10594574362039566, + "learning_rate": 0.00015077452667814113, + "loss": 0.3554, + "num_input_tokens_seen": 2430736, + "step": 149, + "train_runtime": 385.7797, + "train_tokens_per_second": 6300.84 + }, + { + "epoch": 0.5128205128205128, + "grad_norm": 0.10977370291948318, + "learning_rate": 0.0001504302925989673, + "loss": 0.4906, + "num_input_tokens_seen": 2449856, + "step": 150, + "train_runtime": 388.5377, + "train_tokens_per_second": 6305.324 + }, + { + "epoch": 0.5162393162393163, + "grad_norm": 0.09074017405509949, + "learning_rate": 0.00015008605851979346, + "loss": 0.4157, + "num_input_tokens_seen": 2467024, + "step": 151, + "train_runtime": 391.5399, + "train_tokens_per_second": 6300.824 + }, + { + "epoch": 0.5196581196581197, + "grad_norm": 0.08705534040927887, + "learning_rate": 0.00014974182444061964, + "loss": 0.3707, + "num_input_tokens_seen": 2485488, + "step": 152, + "train_runtime": 394.1859, + "train_tokens_per_second": 6305.37 + }, + { + "epoch": 0.5230769230769231, + "grad_norm": 0.09123992174863815, + "learning_rate": 0.00014939759036144577, + "loss": 0.4008, + "num_input_tokens_seen": 2501296, + "step": 153, + "train_runtime": 396.5327, + "train_tokens_per_second": 6307.918 + }, + { + "epoch": 0.5264957264957265, + "grad_norm": 0.07104610651731491, + "learning_rate": 0.00014905335628227195, + "loss": 0.3379, + "num_input_tokens_seen": 2529040, + "step": 154, + "train_runtime": 400.4284, + "train_tokens_per_second": 6315.835 + }, + { + "epoch": 0.5299145299145299, + "grad_norm": 0.11433000862598419, + "learning_rate": 0.0001487091222030981, + "loss": 0.5056, + "num_input_tokens_seen": 2544496, + "step": 155, + "train_runtime": 402.6965, + "train_tokens_per_second": 6318.645 + }, + { + "epoch": 0.5333333333333333, + "grad_norm": 0.12911392748355865, + "learning_rate": 0.0001483648881239243, + "loss": 0.4726, + "num_input_tokens_seen": 2558704, + "step": 156, + "train_runtime": 404.78, + "train_tokens_per_second": 6321.221 + }, + { + "epoch": 0.5367521367521367, + "grad_norm": 0.11388979107141495, + "learning_rate": 0.00014802065404475042, + "loss": 0.44, + "num_input_tokens_seen": 2571648, + "step": 157, + "train_runtime": 406.7058, + "train_tokens_per_second": 6323.115 + }, + { + "epoch": 0.5401709401709401, + "grad_norm": 0.1020691767334938, + "learning_rate": 0.0001476764199655766, + "loss": 0.5144, + "num_input_tokens_seen": 2585280, + "step": 158, + "train_runtime": 408.7686, + "train_tokens_per_second": 6324.556 + }, + { + "epoch": 0.5435897435897435, + "grad_norm": 0.11266439408063889, + "learning_rate": 0.00014733218588640275, + "loss": 0.4898, + "num_input_tokens_seen": 2602560, + "step": 159, + "train_runtime": 411.2897, + "train_tokens_per_second": 6327.802 + }, + { + "epoch": 0.5470085470085471, + "grad_norm": 0.09511423856019974, + "learning_rate": 0.00014698795180722893, + "loss": 0.4601, + "num_input_tokens_seen": 2617520, + "step": 160, + "train_runtime": 413.5019, + "train_tokens_per_second": 6330.127 + }, + { + "epoch": 0.5504273504273505, + "grad_norm": 0.08593633025884628, + "learning_rate": 0.0001466437177280551, + "loss": 0.3578, + "num_input_tokens_seen": 2638032, + "step": 161, + "train_runtime": 416.4729, + "train_tokens_per_second": 6334.222 + }, + { + "epoch": 0.5538461538461539, + "grad_norm": 0.10547275096178055, + "learning_rate": 0.00014629948364888124, + "loss": 0.4053, + "num_input_tokens_seen": 2652912, + "step": 162, + "train_runtime": 418.6689, + "train_tokens_per_second": 6336.54 + }, + { + "epoch": 0.5572649572649573, + "grad_norm": 0.11780176311731339, + "learning_rate": 0.0001459552495697074, + "loss": 0.4034, + "num_input_tokens_seen": 2664944, + "step": 163, + "train_runtime": 420.4837, + "train_tokens_per_second": 6337.806 + }, + { + "epoch": 0.5606837606837607, + "grad_norm": 0.098305843770504, + "learning_rate": 0.00014561101549053358, + "loss": 0.4804, + "num_input_tokens_seen": 2680720, + "step": 164, + "train_runtime": 422.8103, + "train_tokens_per_second": 6340.243 + }, + { + "epoch": 0.5641025641025641, + "grad_norm": 0.08514019846916199, + "learning_rate": 0.00014526678141135973, + "loss": 0.4623, + "num_input_tokens_seen": 2693696, + "step": 165, + "train_runtime": 424.733, + "train_tokens_per_second": 6342.093 + }, + { + "epoch": 0.5675213675213675, + "grad_norm": 0.10156818479299545, + "learning_rate": 0.0001449225473321859, + "loss": 0.4374, + "num_input_tokens_seen": 2706672, + "step": 166, + "train_runtime": 426.6739, + "train_tokens_per_second": 6343.655 + }, + { + "epoch": 0.5709401709401709, + "grad_norm": 0.09817566722631454, + "learning_rate": 0.00014457831325301204, + "loss": 0.3713, + "num_input_tokens_seen": 2717856, + "step": 167, + "train_runtime": 428.3605, + "train_tokens_per_second": 6344.787 + }, + { + "epoch": 0.5743589743589743, + "grad_norm": 0.08898711949586868, + "learning_rate": 0.00014423407917383822, + "loss": 0.3958, + "num_input_tokens_seen": 2738048, + "step": 168, + "train_runtime": 431.2652, + "train_tokens_per_second": 6348.873 + }, + { + "epoch": 0.5777777777777777, + "grad_norm": 0.10815781354904175, + "learning_rate": 0.00014388984509466438, + "loss": 0.4212, + "num_input_tokens_seen": 2751264, + "step": 169, + "train_runtime": 433.2567, + "train_tokens_per_second": 6350.193 + }, + { + "epoch": 0.5811965811965812, + "grad_norm": 0.0858953520655632, + "learning_rate": 0.00014354561101549053, + "loss": 0.4252, + "num_input_tokens_seen": 2768352, + "step": 170, + "train_runtime": 435.7397, + "train_tokens_per_second": 6353.224 + }, + { + "epoch": 0.5846153846153846, + "grad_norm": 0.1065724790096283, + "learning_rate": 0.0001432013769363167, + "loss": 0.4376, + "num_input_tokens_seen": 2783696, + "step": 171, + "train_runtime": 438.0174, + "train_tokens_per_second": 6355.217 + }, + { + "epoch": 0.588034188034188, + "grad_norm": 0.0936359241604805, + "learning_rate": 0.00014285714285714287, + "loss": 0.4641, + "num_input_tokens_seen": 2797712, + "step": 172, + "train_runtime": 440.1373, + "train_tokens_per_second": 6356.452 + }, + { + "epoch": 0.5914529914529915, + "grad_norm": 0.10138653963804245, + "learning_rate": 0.00014251290877796902, + "loss": 0.4442, + "num_input_tokens_seen": 2812432, + "step": 173, + "train_runtime": 442.3579, + "train_tokens_per_second": 6357.821 + }, + { + "epoch": 0.5948717948717949, + "grad_norm": 0.10231815278530121, + "learning_rate": 0.00014216867469879518, + "loss": 0.4113, + "num_input_tokens_seen": 2824544, + "step": 174, + "train_runtime": 444.2041, + "train_tokens_per_second": 6358.663 + }, + { + "epoch": 0.5982905982905983, + "grad_norm": 0.10455156117677689, + "learning_rate": 0.00014182444061962136, + "loss": 0.5237, + "num_input_tokens_seen": 2841296, + "step": 175, + "train_runtime": 446.6226, + "train_tokens_per_second": 6361.738 + }, + { + "epoch": 0.6017094017094017, + "grad_norm": 0.09039057046175003, + "learning_rate": 0.0001414802065404475, + "loss": 0.4316, + "num_input_tokens_seen": 2854304, + "step": 176, + "train_runtime": 448.5655, + "train_tokens_per_second": 6363.182 + }, + { + "epoch": 0.6051282051282051, + "grad_norm": 0.09800916165113449, + "learning_rate": 0.00014113597246127367, + "loss": 0.4883, + "num_input_tokens_seen": 2871312, + "step": 177, + "train_runtime": 451.0616, + "train_tokens_per_second": 6365.676 + }, + { + "epoch": 0.6051282051282051, + "eval_loss": 0.4250277280807495, + "eval_runtime": 11.8943, + "eval_samples_per_second": 83.906, + "eval_steps_per_second": 5.297, + "num_input_tokens_seen": 2871312, + "step": 177 + }, + { + "epoch": 0.6085470085470085, + "grad_norm": 0.1056000217795372, + "learning_rate": 0.00014079173838209982, + "loss": 0.3984, + "num_input_tokens_seen": 2882896, + "step": 178, + "train_runtime": 464.7282, + "train_tokens_per_second": 6203.403 + }, + { + "epoch": 0.611965811965812, + "grad_norm": 0.10022575408220291, + "learning_rate": 0.000140447504302926, + "loss": 0.3613, + "num_input_tokens_seen": 2897328, + "step": 179, + "train_runtime": 466.8908, + "train_tokens_per_second": 6205.58 + }, + { + "epoch": 0.6153846153846154, + "grad_norm": 0.10060451924800873, + "learning_rate": 0.00014010327022375216, + "loss": 0.505, + "num_input_tokens_seen": 2910544, + "step": 180, + "train_runtime": 468.834, + "train_tokens_per_second": 6208.049 + }, + { + "epoch": 0.6188034188034188, + "grad_norm": 0.10652410238981247, + "learning_rate": 0.00013975903614457834, + "loss": 0.456, + "num_input_tokens_seen": 2921712, + "step": 181, + "train_runtime": 471.0911, + "train_tokens_per_second": 6202.01 + }, + { + "epoch": 0.6222222222222222, + "grad_norm": 0.1084052249789238, + "learning_rate": 0.00013941480206540447, + "loss": 0.3833, + "num_input_tokens_seen": 2932832, + "step": 182, + "train_runtime": 472.7877, + "train_tokens_per_second": 6203.274 + }, + { + "epoch": 0.6256410256410256, + "grad_norm": 0.09704000502824783, + "learning_rate": 0.00013907056798623065, + "loss": 0.4889, + "num_input_tokens_seen": 2947600, + "step": 183, + "train_runtime": 474.9298, + "train_tokens_per_second": 6206.391 + }, + { + "epoch": 0.629059829059829, + "grad_norm": 0.09996459633111954, + "learning_rate": 0.0001387263339070568, + "loss": 0.3787, + "num_input_tokens_seen": 2963088, + "step": 184, + "train_runtime": 477.2542, + "train_tokens_per_second": 6208.616 + }, + { + "epoch": 0.6324786324786325, + "grad_norm": 0.09264053404331207, + "learning_rate": 0.00013838209982788298, + "loss": 0.3498, + "num_input_tokens_seen": 2975088, + "step": 185, + "train_runtime": 479.0775, + "train_tokens_per_second": 6210.035 + }, + { + "epoch": 0.6358974358974359, + "grad_norm": 0.09588699042797089, + "learning_rate": 0.00013803786574870914, + "loss": 0.485, + "num_input_tokens_seen": 2988944, + "step": 186, + "train_runtime": 481.1581, + "train_tokens_per_second": 6211.978 + }, + { + "epoch": 0.6393162393162393, + "grad_norm": 0.09032841771841049, + "learning_rate": 0.0001376936316695353, + "loss": 0.4083, + "num_input_tokens_seen": 3006016, + "step": 187, + "train_runtime": 483.6506, + "train_tokens_per_second": 6215.264 + }, + { + "epoch": 0.6427350427350428, + "grad_norm": 0.11055252701044083, + "learning_rate": 0.00013734939759036145, + "loss": 0.4339, + "num_input_tokens_seen": 3018096, + "step": 188, + "train_runtime": 485.4309, + "train_tokens_per_second": 6217.355 + }, + { + "epoch": 0.6461538461538462, + "grad_norm": 0.12127861380577087, + "learning_rate": 0.00013700516351118763, + "loss": 0.4601, + "num_input_tokens_seen": 3031280, + "step": 189, + "train_runtime": 487.4282, + "train_tokens_per_second": 6218.926 + }, + { + "epoch": 0.6495726495726496, + "grad_norm": 0.10579859465360641, + "learning_rate": 0.00013666092943201378, + "loss": 0.4519, + "num_input_tokens_seen": 3046560, + "step": 190, + "train_runtime": 489.6819, + "train_tokens_per_second": 6221.508 + }, + { + "epoch": 0.652991452991453, + "grad_norm": 0.1086781919002533, + "learning_rate": 0.00013631669535283994, + "loss": 0.4843, + "num_input_tokens_seen": 3061024, + "step": 191, + "train_runtime": 491.8462, + "train_tokens_per_second": 6223.539 + }, + { + "epoch": 0.6564102564102564, + "grad_norm": 0.09214667230844498, + "learning_rate": 0.0001359724612736661, + "loss": 0.43, + "num_input_tokens_seen": 3076848, + "step": 192, + "train_runtime": 494.2062, + "train_tokens_per_second": 6225.839 + }, + { + "epoch": 0.6598290598290598, + "grad_norm": 0.11928390711545944, + "learning_rate": 0.00013562822719449227, + "loss": 0.5062, + "num_input_tokens_seen": 3092208, + "step": 193, + "train_runtime": 496.4458, + "train_tokens_per_second": 6228.692 + }, + { + "epoch": 0.6632478632478632, + "grad_norm": 0.09398146718740463, + "learning_rate": 0.00013528399311531843, + "loss": 0.372, + "num_input_tokens_seen": 3103168, + "step": 194, + "train_runtime": 498.1054, + "train_tokens_per_second": 6229.943 + }, + { + "epoch": 0.6666666666666666, + "grad_norm": 0.08932854235172272, + "learning_rate": 0.00013493975903614458, + "loss": 0.4515, + "num_input_tokens_seen": 3120512, + "step": 195, + "train_runtime": 500.6536, + "train_tokens_per_second": 6232.877 + }, + { + "epoch": 0.67008547008547, + "grad_norm": 0.09448716044425964, + "learning_rate": 0.00013459552495697074, + "loss": 0.4346, + "num_input_tokens_seen": 3136544, + "step": 196, + "train_runtime": 503.0211, + "train_tokens_per_second": 6235.412 + }, + { + "epoch": 0.6735042735042736, + "grad_norm": 0.09740068763494492, + "learning_rate": 0.00013425129087779692, + "loss": 0.3846, + "num_input_tokens_seen": 3152720, + "step": 197, + "train_runtime": 505.3917, + "train_tokens_per_second": 6238.172 + }, + { + "epoch": 0.676923076923077, + "grad_norm": 0.10378974676132202, + "learning_rate": 0.00013390705679862307, + "loss": 0.4808, + "num_input_tokens_seen": 3167392, + "step": 198, + "train_runtime": 507.5697, + "train_tokens_per_second": 6240.31 + }, + { + "epoch": 0.6803418803418804, + "grad_norm": 0.08869824558496475, + "learning_rate": 0.00013356282271944923, + "loss": 0.387, + "num_input_tokens_seen": 3186496, + "step": 199, + "train_runtime": 510.3483, + "train_tokens_per_second": 6243.767 + }, + { + "epoch": 0.6837606837606838, + "grad_norm": 0.09974192827939987, + "learning_rate": 0.00013321858864027538, + "loss": 0.4512, + "num_input_tokens_seen": 3205344, + "step": 200, + "train_runtime": 513.0801, + "train_tokens_per_second": 6247.259 + }, + { + "epoch": 0.6871794871794872, + "grad_norm": 0.08839447796344757, + "learning_rate": 0.00013287435456110156, + "loss": 0.4027, + "num_input_tokens_seen": 3222720, + "step": 201, + "train_runtime": 515.6265, + "train_tokens_per_second": 6250.106 + }, + { + "epoch": 0.6905982905982906, + "grad_norm": 0.07892576605081558, + "learning_rate": 0.00013253012048192772, + "loss": 0.3836, + "num_input_tokens_seen": 3247392, + "step": 202, + "train_runtime": 519.1427, + "train_tokens_per_second": 6255.297 + }, + { + "epoch": 0.694017094017094, + "grad_norm": 0.09967231005430222, + "learning_rate": 0.0001321858864027539, + "loss": 0.6602, + "num_input_tokens_seen": 3263344, + "step": 203, + "train_runtime": 521.4876, + "train_tokens_per_second": 6257.76 + }, + { + "epoch": 0.6974358974358974, + "grad_norm": 0.09928127378225327, + "learning_rate": 0.00013184165232358003, + "loss": 0.3557, + "num_input_tokens_seen": 3275472, + "step": 204, + "train_runtime": 523.3642, + "train_tokens_per_second": 6258.494 + }, + { + "epoch": 0.7008547008547008, + "grad_norm": 0.10789816826581955, + "learning_rate": 0.0001314974182444062, + "loss": 0.3937, + "num_input_tokens_seen": 3293904, + "step": 205, + "train_runtime": 526.0362, + "train_tokens_per_second": 6261.744 + }, + { + "epoch": 0.7042735042735043, + "grad_norm": 0.10802796483039856, + "learning_rate": 0.00013115318416523236, + "loss": 0.4472, + "num_input_tokens_seen": 3309584, + "step": 206, + "train_runtime": 528.3293, + "train_tokens_per_second": 6264.245 + }, + { + "epoch": 0.7076923076923077, + "grad_norm": 0.09764997661113739, + "learning_rate": 0.00013080895008605854, + "loss": 0.457, + "num_input_tokens_seen": 3323776, + "step": 207, + "train_runtime": 530.4461, + "train_tokens_per_second": 6266.002 + }, + { + "epoch": 0.7111111111111111, + "grad_norm": 0.10141371190547943, + "learning_rate": 0.00013046471600688467, + "loss": 0.4036, + "num_input_tokens_seen": 3339600, + "step": 208, + "train_runtime": 532.7571, + "train_tokens_per_second": 6268.523 + }, + { + "epoch": 0.7145299145299145, + "grad_norm": 0.11183507740497589, + "learning_rate": 0.00013012048192771085, + "loss": 0.4412, + "num_input_tokens_seen": 3353280, + "step": 209, + "train_runtime": 534.8217, + "train_tokens_per_second": 6269.902 + }, + { + "epoch": 0.717948717948718, + "grad_norm": 0.1051330491900444, + "learning_rate": 0.000129776247848537, + "loss": 0.4337, + "num_input_tokens_seen": 3369296, + "step": 210, + "train_runtime": 537.1671, + "train_tokens_per_second": 6272.342 + }, + { + "epoch": 0.7213675213675214, + "grad_norm": 0.10551027208566666, + "learning_rate": 0.0001294320137693632, + "loss": 0.4042, + "num_input_tokens_seen": 3384000, + "step": 211, + "train_runtime": 539.924, + "train_tokens_per_second": 6267.548 + }, + { + "epoch": 0.7247863247863248, + "grad_norm": 0.10018742829561234, + "learning_rate": 0.00012908777969018932, + "loss": 0.4694, + "num_input_tokens_seen": 3399088, + "step": 212, + "train_runtime": 542.1589, + "train_tokens_per_second": 6269.542 + }, + { + "epoch": 0.7282051282051282, + "grad_norm": 0.12323636561632156, + "learning_rate": 0.0001287435456110155, + "loss": 0.4411, + "num_input_tokens_seen": 3411392, + "step": 213, + "train_runtime": 544.0078, + "train_tokens_per_second": 6270.851 + }, + { + "epoch": 0.7316239316239316, + "grad_norm": 0.11548493057489395, + "learning_rate": 0.00012839931153184165, + "loss": 0.5167, + "num_input_tokens_seen": 3428416, + "step": 214, + "train_runtime": 546.4748, + "train_tokens_per_second": 6273.695 + }, + { + "epoch": 0.7350427350427351, + "grad_norm": 0.0862911120057106, + "learning_rate": 0.00012805507745266783, + "loss": 0.3712, + "num_input_tokens_seen": 3459568, + "step": 215, + "train_runtime": 553.9819, + "train_tokens_per_second": 6244.912 + }, + { + "epoch": 0.7384615384615385, + "grad_norm": 0.1001635193824768, + "learning_rate": 0.00012771084337349396, + "loss": 0.394, + "num_input_tokens_seen": 3474128, + "step": 216, + "train_runtime": 556.1513, + "train_tokens_per_second": 6246.732 + }, + { + "epoch": 0.7418803418803419, + "grad_norm": 0.10096625983715057, + "learning_rate": 0.00012736660929432014, + "loss": 0.4354, + "num_input_tokens_seen": 3488032, + "step": 217, + "train_runtime": 558.2102, + "train_tokens_per_second": 6248.6 + }, + { + "epoch": 0.7452991452991453, + "grad_norm": 0.10238556563854218, + "learning_rate": 0.0001270223752151463, + "loss": 0.4276, + "num_input_tokens_seen": 3504688, + "step": 218, + "train_runtime": 560.6522, + "train_tokens_per_second": 6251.091 + }, + { + "epoch": 0.7487179487179487, + "grad_norm": 0.09864645451307297, + "learning_rate": 0.00012667814113597248, + "loss": 0.4098, + "num_input_tokens_seen": 3521088, + "step": 219, + "train_runtime": 563.1091, + "train_tokens_per_second": 6252.941 + }, + { + "epoch": 0.7521367521367521, + "grad_norm": 0.10077870637178421, + "learning_rate": 0.00012633390705679863, + "loss": 0.4802, + "num_input_tokens_seen": 3537040, + "step": 220, + "train_runtime": 565.4347, + "train_tokens_per_second": 6255.435 + }, + { + "epoch": 0.7555555555555555, + "grad_norm": 0.10053377598524094, + "learning_rate": 0.0001259896729776248, + "loss": 0.4794, + "num_input_tokens_seen": 3551456, + "step": 221, + "train_runtime": 567.5603, + "train_tokens_per_second": 6257.407 + }, + { + "epoch": 0.7589743589743589, + "grad_norm": 0.10047437995672226, + "learning_rate": 0.00012564543889845094, + "loss": 0.4938, + "num_input_tokens_seen": 3567616, + "step": 222, + "train_runtime": 569.9285, + "train_tokens_per_second": 6259.761 + }, + { + "epoch": 0.7623931623931623, + "grad_norm": 0.07970099151134491, + "learning_rate": 0.00012530120481927712, + "loss": 0.3474, + "num_input_tokens_seen": 3588144, + "step": 223, + "train_runtime": 572.8933, + "train_tokens_per_second": 6263.198 + }, + { + "epoch": 0.7658119658119659, + "grad_norm": 0.10995665192604065, + "learning_rate": 0.00012495697074010328, + "loss": 0.3795, + "num_input_tokens_seen": 3600144, + "step": 224, + "train_runtime": 574.6969, + "train_tokens_per_second": 6264.422 + }, + { + "epoch": 0.7692307692307693, + "grad_norm": 0.12067970633506775, + "learning_rate": 0.00012461273666092943, + "loss": 0.4804, + "num_input_tokens_seen": 3621280, + "step": 225, + "train_runtime": 577.7205, + "train_tokens_per_second": 6268.222 + }, + { + "epoch": 0.7726495726495727, + "grad_norm": 0.11939004063606262, + "learning_rate": 0.0001242685025817556, + "loss": 0.4533, + "num_input_tokens_seen": 3634160, + "step": 226, + "train_runtime": 579.6747, + "train_tokens_per_second": 6269.31 + }, + { + "epoch": 0.7760683760683761, + "grad_norm": 0.10946697741746902, + "learning_rate": 0.00012392426850258177, + "loss": 0.4324, + "num_input_tokens_seen": 3645536, + "step": 227, + "train_runtime": 581.3907, + "train_tokens_per_second": 6270.372 + }, + { + "epoch": 0.7794871794871795, + "grad_norm": 0.09853693842887878, + "learning_rate": 0.00012358003442340792, + "loss": 0.4728, + "num_input_tokens_seen": 3660048, + "step": 228, + "train_runtime": 583.5443, + "train_tokens_per_second": 6272.1 + }, + { + "epoch": 0.7829059829059829, + "grad_norm": 0.1062566265463829, + "learning_rate": 0.00012323580034423408, + "loss": 0.5393, + "num_input_tokens_seen": 3672816, + "step": 229, + "train_runtime": 585.4621, + "train_tokens_per_second": 6273.363 + }, + { + "epoch": 0.7863247863247863, + "grad_norm": 0.09363020956516266, + "learning_rate": 0.00012289156626506023, + "loss": 0.4152, + "num_input_tokens_seen": 3696128, + "step": 230, + "train_runtime": 588.8137, + "train_tokens_per_second": 6277.245 + }, + { + "epoch": 0.7897435897435897, + "grad_norm": 0.09704622626304626, + "learning_rate": 0.00012254733218588641, + "loss": 0.5139, + "num_input_tokens_seen": 3710160, + "step": 231, + "train_runtime": 590.8993, + "train_tokens_per_second": 6278.836 + }, + { + "epoch": 0.7931623931623931, + "grad_norm": 0.09564655274152756, + "learning_rate": 0.00012220309810671257, + "loss": 0.4335, + "num_input_tokens_seen": 3723536, + "step": 232, + "train_runtime": 592.9271, + "train_tokens_per_second": 6279.922 + }, + { + "epoch": 0.7965811965811965, + "grad_norm": 0.08687002956867218, + "learning_rate": 0.00012185886402753872, + "loss": 0.3445, + "num_input_tokens_seen": 3738592, + "step": 233, + "train_runtime": 595.173, + "train_tokens_per_second": 6281.522 + }, + { + "epoch": 0.8, + "grad_norm": 0.09096026420593262, + "learning_rate": 0.00012151462994836489, + "loss": 0.5372, + "num_input_tokens_seen": 3755936, + "step": 234, + "train_runtime": 597.7017, + "train_tokens_per_second": 6283.964 + }, + { + "epoch": 0.8034188034188035, + "grad_norm": 0.09353629499673843, + "learning_rate": 0.00012117039586919106, + "loss": 0.3614, + "num_input_tokens_seen": 3770768, + "step": 235, + "train_runtime": 599.8797, + "train_tokens_per_second": 6285.873 + }, + { + "epoch": 0.8068376068376069, + "grad_norm": 0.0944899320602417, + "learning_rate": 0.00012082616179001723, + "loss": 0.3849, + "num_input_tokens_seen": 3795312, + "step": 236, + "train_runtime": 603.3944, + "train_tokens_per_second": 6289.935 + }, + { + "epoch": 0.8068376068376069, + "eval_loss": 0.42357537150382996, + "eval_runtime": 11.8678, + "eval_samples_per_second": 84.093, + "eval_steps_per_second": 5.308, + "num_input_tokens_seen": 3795312, + "step": 236 + }, + { + "epoch": 0.8102564102564103, + "grad_norm": 0.10128959268331528, + "learning_rate": 0.0001204819277108434, + "loss": 0.4026, + "num_input_tokens_seen": 3806656, + "step": 237, + "train_runtime": 616.9818, + "train_tokens_per_second": 6169.803 + }, + { + "epoch": 0.8136752136752137, + "grad_norm": 0.09387616068124771, + "learning_rate": 0.00012013769363166954, + "loss": 0.4182, + "num_input_tokens_seen": 3822144, + "step": 238, + "train_runtime": 619.269, + "train_tokens_per_second": 6172.026 + }, + { + "epoch": 0.8170940170940171, + "grad_norm": 0.09902401268482208, + "learning_rate": 0.0001197934595524957, + "loss": 0.4232, + "num_input_tokens_seen": 3842016, + "step": 239, + "train_runtime": 622.1266, + "train_tokens_per_second": 6175.617 + }, + { + "epoch": 0.8205128205128205, + "grad_norm": 0.09925245493650436, + "learning_rate": 0.00011944922547332187, + "loss": 0.4478, + "num_input_tokens_seen": 3854752, + "step": 240, + "train_runtime": 624.0402, + "train_tokens_per_second": 6177.089 + }, + { + "epoch": 0.8239316239316239, + "grad_norm": 0.11560021340847015, + "learning_rate": 0.00011910499139414804, + "loss": 0.4808, + "num_input_tokens_seen": 3866496, + "step": 241, + "train_runtime": 626.2473, + "train_tokens_per_second": 6174.072 + }, + { + "epoch": 0.8273504273504273, + "grad_norm": 0.08798137307167053, + "learning_rate": 0.00011876075731497418, + "loss": 0.4486, + "num_input_tokens_seen": 3885600, + "step": 242, + "train_runtime": 629.0654, + "train_tokens_per_second": 6176.782 + }, + { + "epoch": 0.8307692307692308, + "grad_norm": 0.09302930533885956, + "learning_rate": 0.00011841652323580035, + "loss": 0.436, + "num_input_tokens_seen": 3905008, + "step": 243, + "train_runtime": 631.8771, + "train_tokens_per_second": 6180.012 + }, + { + "epoch": 0.8341880341880342, + "grad_norm": 0.1044159084558487, + "learning_rate": 0.00011807228915662652, + "loss": 0.4986, + "num_input_tokens_seen": 3917472, + "step": 244, + "train_runtime": 633.7768, + "train_tokens_per_second": 6181.154 + }, + { + "epoch": 0.8376068376068376, + "grad_norm": 0.10373450815677643, + "learning_rate": 0.00011772805507745268, + "loss": 0.5037, + "num_input_tokens_seen": 3933664, + "step": 245, + "train_runtime": 636.199, + "train_tokens_per_second": 6183.072 + }, + { + "epoch": 0.841025641025641, + "grad_norm": 0.08972188085317612, + "learning_rate": 0.00011738382099827883, + "loss": 0.4908, + "num_input_tokens_seen": 3952288, + "step": 246, + "train_runtime": 638.8866, + "train_tokens_per_second": 6186.212 + }, + { + "epoch": 0.8444444444444444, + "grad_norm": 0.0838053897023201, + "learning_rate": 0.000117039586919105, + "loss": 0.3793, + "num_input_tokens_seen": 3970672, + "step": 247, + "train_runtime": 641.5476, + "train_tokens_per_second": 6189.209 + }, + { + "epoch": 0.8478632478632478, + "grad_norm": 0.06703872978687286, + "learning_rate": 0.00011669535283993116, + "loss": 0.3059, + "num_input_tokens_seen": 3999184, + "step": 248, + "train_runtime": 645.5767, + "train_tokens_per_second": 6194.747 + }, + { + "epoch": 0.8512820512820513, + "grad_norm": 0.08314735442399979, + "learning_rate": 0.00011635111876075733, + "loss": 0.4058, + "num_input_tokens_seen": 4013280, + "step": 249, + "train_runtime": 647.6748, + "train_tokens_per_second": 6196.443 + }, + { + "epoch": 0.8547008547008547, + "grad_norm": 0.08154241740703583, + "learning_rate": 0.00011600688468158347, + "loss": 0.3805, + "num_input_tokens_seen": 4031920, + "step": 250, + "train_runtime": 650.4038, + "train_tokens_per_second": 6199.102 + }, + { + "epoch": 0.8581196581196581, + "grad_norm": 0.09133270382881165, + "learning_rate": 0.00011566265060240964, + "loss": 0.3681, + "num_input_tokens_seen": 4053744, + "step": 251, + "train_runtime": 653.5797, + "train_tokens_per_second": 6202.371 + }, + { + "epoch": 0.8615384615384616, + "grad_norm": 0.10926900058984756, + "learning_rate": 0.0001153184165232358, + "loss": 0.4441, + "num_input_tokens_seen": 4065536, + "step": 252, + "train_runtime": 655.3821, + "train_tokens_per_second": 6203.307 + }, + { + "epoch": 0.864957264957265, + "grad_norm": 0.08510483056306839, + "learning_rate": 0.00011497418244406197, + "loss": 0.4944, + "num_input_tokens_seen": 4083376, + "step": 253, + "train_runtime": 657.9531, + "train_tokens_per_second": 6206.181 + }, + { + "epoch": 0.8683760683760684, + "grad_norm": 0.09709641337394714, + "learning_rate": 0.00011462994836488814, + "loss": 0.481, + "num_input_tokens_seen": 4098848, + "step": 254, + "train_runtime": 660.2372, + "train_tokens_per_second": 6208.145 + }, + { + "epoch": 0.8717948717948718, + "grad_norm": 0.08926745504140854, + "learning_rate": 0.00011428571428571428, + "loss": 0.4043, + "num_input_tokens_seen": 4113904, + "step": 255, + "train_runtime": 662.4724, + "train_tokens_per_second": 6209.925 + }, + { + "epoch": 0.8752136752136752, + "grad_norm": 0.09374278038740158, + "learning_rate": 0.00011394148020654045, + "loss": 0.4502, + "num_input_tokens_seen": 4131168, + "step": 256, + "train_runtime": 664.9653, + "train_tokens_per_second": 6212.607 + }, + { + "epoch": 0.8786324786324786, + "grad_norm": 0.09459812939167023, + "learning_rate": 0.00011359724612736662, + "loss": 0.4439, + "num_input_tokens_seen": 4149904, + "step": 257, + "train_runtime": 667.6682, + "train_tokens_per_second": 6215.519 + }, + { + "epoch": 0.882051282051282, + "grad_norm": 0.1052205041050911, + "learning_rate": 0.00011325301204819279, + "loss": 0.4385, + "num_input_tokens_seen": 4166832, + "step": 258, + "train_runtime": 670.0958, + "train_tokens_per_second": 6218.263 + }, + { + "epoch": 0.8854700854700854, + "grad_norm": 0.09694038331508636, + "learning_rate": 0.00011290877796901893, + "loss": 0.375, + "num_input_tokens_seen": 4184976, + "step": 259, + "train_runtime": 672.74, + "train_tokens_per_second": 6220.793 + }, + { + "epoch": 0.8888888888888888, + "grad_norm": 0.11208639293909073, + "learning_rate": 0.0001125645438898451, + "loss": 0.4517, + "num_input_tokens_seen": 4199888, + "step": 260, + "train_runtime": 674.9288, + "train_tokens_per_second": 6222.713 + }, + { + "epoch": 0.8923076923076924, + "grad_norm": 0.09207024425268173, + "learning_rate": 0.00011222030981067126, + "loss": 0.4132, + "num_input_tokens_seen": 4215664, + "step": 261, + "train_runtime": 677.2813, + "train_tokens_per_second": 6224.392 + }, + { + "epoch": 0.8957264957264958, + "grad_norm": 0.0904022753238678, + "learning_rate": 0.00011187607573149743, + "loss": 0.3322, + "num_input_tokens_seen": 4230080, + "step": 262, + "train_runtime": 679.4101, + "train_tokens_per_second": 6226.107 + }, + { + "epoch": 0.8991452991452992, + "grad_norm": 0.12087982147932053, + "learning_rate": 0.00011153184165232357, + "loss": 0.4281, + "num_input_tokens_seen": 4241440, + "step": 263, + "train_runtime": 681.1509, + "train_tokens_per_second": 6226.873 + }, + { + "epoch": 0.9025641025641026, + "grad_norm": 0.12484072893857956, + "learning_rate": 0.00011118760757314974, + "loss": 0.5186, + "num_input_tokens_seen": 4252976, + "step": 264, + "train_runtime": 682.9063, + "train_tokens_per_second": 6227.759 + }, + { + "epoch": 0.905982905982906, + "grad_norm": 0.08252517133951187, + "learning_rate": 0.00011084337349397591, + "loss": 0.4049, + "num_input_tokens_seen": 4269312, + "step": 265, + "train_runtime": 685.2856, + "train_tokens_per_second": 6229.975 + }, + { + "epoch": 0.9094017094017094, + "grad_norm": 0.11546450853347778, + "learning_rate": 0.00011049913941480208, + "loss": 0.4628, + "num_input_tokens_seen": 4284544, + "step": 266, + "train_runtime": 687.5331, + "train_tokens_per_second": 6231.764 + }, + { + "epoch": 0.9128205128205128, + "grad_norm": 0.11925678700208664, + "learning_rate": 0.00011015490533562822, + "loss": 0.4275, + "num_input_tokens_seen": 4296208, + "step": 267, + "train_runtime": 689.3281, + "train_tokens_per_second": 6232.457 + }, + { + "epoch": 0.9162393162393162, + "grad_norm": 0.09587883204221725, + "learning_rate": 0.00010981067125645439, + "loss": 0.4098, + "num_input_tokens_seen": 4313872, + "step": 268, + "train_runtime": 691.9285, + "train_tokens_per_second": 6234.563 + }, + { + "epoch": 0.9196581196581196, + "grad_norm": 0.08406773209571838, + "learning_rate": 0.00010946643717728055, + "loss": 0.3972, + "num_input_tokens_seen": 4335664, + "step": 269, + "train_runtime": 695.0743, + "train_tokens_per_second": 6237.699 + }, + { + "epoch": 0.9230769230769231, + "grad_norm": 0.1020239070057869, + "learning_rate": 0.00010912220309810672, + "loss": 0.3777, + "num_input_tokens_seen": 4353360, + "step": 270, + "train_runtime": 697.6355, + "train_tokens_per_second": 6240.165 + }, + { + "epoch": 0.9264957264957265, + "grad_norm": 0.11209993064403534, + "learning_rate": 0.00010877796901893289, + "loss": 0.4114, + "num_input_tokens_seen": 4374544, + "step": 271, + "train_runtime": 701.2128, + "train_tokens_per_second": 6238.54 + }, + { + "epoch": 0.9299145299145299, + "grad_norm": 0.11567652225494385, + "learning_rate": 0.00010843373493975903, + "loss": 0.4051, + "num_input_tokens_seen": 4391456, + "step": 272, + "train_runtime": 703.6794, + "train_tokens_per_second": 6240.706 + }, + { + "epoch": 0.9333333333333333, + "grad_norm": 0.11092627048492432, + "learning_rate": 0.0001080895008605852, + "loss": 0.4194, + "num_input_tokens_seen": 4404608, + "step": 273, + "train_runtime": 705.6779, + "train_tokens_per_second": 6241.669 + }, + { + "epoch": 0.9367521367521368, + "grad_norm": 0.10917766392230988, + "learning_rate": 0.00010774526678141137, + "loss": 0.4475, + "num_input_tokens_seen": 4416592, + "step": 274, + "train_runtime": 707.4805, + "train_tokens_per_second": 6242.705 + }, + { + "epoch": 0.9401709401709402, + "grad_norm": 0.09348605573177338, + "learning_rate": 0.00010740103270223754, + "loss": 0.4913, + "num_input_tokens_seen": 4433728, + "step": 275, + "train_runtime": 709.9546, + "train_tokens_per_second": 6245.086 + }, + { + "epoch": 0.9435897435897436, + "grad_norm": 0.10354448109865189, + "learning_rate": 0.00010705679862306368, + "loss": 0.4018, + "num_input_tokens_seen": 4448224, + "step": 276, + "train_runtime": 712.1021, + "train_tokens_per_second": 6246.61 + }, + { + "epoch": 0.947008547008547, + "grad_norm": 0.09892738610506058, + "learning_rate": 0.00010671256454388984, + "loss": 0.4477, + "num_input_tokens_seen": 4467024, + "step": 277, + "train_runtime": 714.8588, + "train_tokens_per_second": 6248.82 + }, + { + "epoch": 0.9504273504273504, + "grad_norm": 0.11075231432914734, + "learning_rate": 0.00010636833046471601, + "loss": 0.3678, + "num_input_tokens_seen": 4481232, + "step": 278, + "train_runtime": 716.9881, + "train_tokens_per_second": 6250.078 + }, + { + "epoch": 0.9538461538461539, + "grad_norm": 0.10120780020952225, + "learning_rate": 0.00010602409638554218, + "loss": 0.4281, + "num_input_tokens_seen": 4500528, + "step": 279, + "train_runtime": 719.7698, + "train_tokens_per_second": 6252.733 + }, + { + "epoch": 0.9572649572649573, + "grad_norm": 0.09937260299921036, + "learning_rate": 0.00010567986230636832, + "loss": 0.3882, + "num_input_tokens_seen": 4525136, + "step": 280, + "train_runtime": 723.3072, + "train_tokens_per_second": 6256.175 + }, + { + "epoch": 0.9606837606837607, + "grad_norm": 0.08478062599897385, + "learning_rate": 0.00010533562822719449, + "loss": 0.3849, + "num_input_tokens_seen": 4541200, + "step": 281, + "train_runtime": 725.6747, + "train_tokens_per_second": 6257.9 + }, + { + "epoch": 0.9641025641025641, + "grad_norm": 0.09681829810142517, + "learning_rate": 0.00010499139414802066, + "loss": 0.3051, + "num_input_tokens_seen": 4578816, + "step": 282, + "train_runtime": 731.2044, + "train_tokens_per_second": 6262.019 + }, + { + "epoch": 0.9675213675213675, + "grad_norm": 0.09028521925210953, + "learning_rate": 0.00010464716006884682, + "loss": 0.4796, + "num_input_tokens_seen": 4596608, + "step": 283, + "train_runtime": 733.7729, + "train_tokens_per_second": 6264.347 + }, + { + "epoch": 0.9709401709401709, + "grad_norm": 0.09063593298196793, + "learning_rate": 0.00010430292598967298, + "loss": 0.347, + "num_input_tokens_seen": 4607632, + "step": 284, + "train_runtime": 735.4673, + "train_tokens_per_second": 6264.904 + }, + { + "epoch": 0.9743589743589743, + "grad_norm": 0.09063643962144852, + "learning_rate": 0.00010395869191049913, + "loss": 0.2912, + "num_input_tokens_seen": 4631200, + "step": 285, + "train_runtime": 738.8336, + "train_tokens_per_second": 6268.258 + }, + { + "epoch": 0.9777777777777777, + "grad_norm": 0.11136946082115173, + "learning_rate": 0.0001036144578313253, + "loss": 0.4875, + "num_input_tokens_seen": 4645856, + "step": 286, + "train_runtime": 741.0112, + "train_tokens_per_second": 6269.616 + }, + { + "epoch": 0.9811965811965812, + "grad_norm": 0.09543482214212418, + "learning_rate": 0.00010327022375215147, + "loss": 0.4009, + "num_input_tokens_seen": 4662528, + "step": 287, + "train_runtime": 743.4774, + "train_tokens_per_second": 6271.244 + }, + { + "epoch": 0.9846153846153847, + "grad_norm": 0.11340586096048355, + "learning_rate": 0.00010292598967297764, + "loss": 0.4523, + "num_input_tokens_seen": 4677744, + "step": 288, + "train_runtime": 745.7291, + "train_tokens_per_second": 6272.712 + }, + { + "epoch": 0.9880341880341881, + "grad_norm": 0.09249569475650787, + "learning_rate": 0.00010258175559380379, + "loss": 0.3638, + "num_input_tokens_seen": 4694512, + "step": 289, + "train_runtime": 748.2102, + "train_tokens_per_second": 6274.323 + }, + { + "epoch": 0.9914529914529915, + "grad_norm": 0.11417587846517563, + "learning_rate": 0.00010223752151462995, + "loss": 0.513, + "num_input_tokens_seen": 4710224, + "step": 290, + "train_runtime": 750.5071, + "train_tokens_per_second": 6276.055 + }, + { + "epoch": 0.9948717948717949, + "grad_norm": 0.09056784212589264, + "learning_rate": 0.00010189328743545611, + "loss": 0.4324, + "num_input_tokens_seen": 4727296, + "step": 291, + "train_runtime": 752.9922, + "train_tokens_per_second": 6278.014 + }, + { + "epoch": 0.9982905982905983, + "grad_norm": 0.09821169078350067, + "learning_rate": 0.00010154905335628228, + "loss": 0.3605, + "num_input_tokens_seen": 4741520, + "step": 292, + "train_runtime": 755.0666, + "train_tokens_per_second": 6279.605 + }, + { + "epoch": 1.0, + "grad_norm": 0.13449996709823608, + "learning_rate": 0.00010120481927710844, + "loss": 0.3949, + "num_input_tokens_seen": 4751278, + "step": 293, + "train_runtime": 756.491, + "train_tokens_per_second": 6280.68 + }, + { + "epoch": 1.0034188034188034, + "grad_norm": 0.09440767765045166, + "learning_rate": 0.0001008605851979346, + "loss": 0.3857, + "num_input_tokens_seen": 4764718, + "step": 294, + "train_runtime": 758.5245, + "train_tokens_per_second": 6281.561 + }, + { + "epoch": 1.0068376068376068, + "grad_norm": 0.08331061899662018, + "learning_rate": 0.00010051635111876076, + "loss": 0.3654, + "num_input_tokens_seen": 4782670, + "step": 295, + "train_runtime": 761.0921, + "train_tokens_per_second": 6283.957 + }, + { + "epoch": 1.0068376068376068, + "eval_loss": 0.421943724155426, + "eval_runtime": 11.8792, + "eval_samples_per_second": 84.013, + "eval_steps_per_second": 5.303, + "num_input_tokens_seen": 4782670, + "step": 295 + }, + { + "epoch": 1.0102564102564102, + "grad_norm": 0.09061148762702942, + "learning_rate": 0.00010017211703958693, + "loss": 0.3673, + "num_input_tokens_seen": 4795310, + "step": 296, + "train_runtime": 774.907, + "train_tokens_per_second": 6188.239 + }, + { + "epoch": 1.0136752136752136, + "grad_norm": 0.09946513175964355, + "learning_rate": 9.982788296041308e-05, + "loss": 0.3978, + "num_input_tokens_seen": 4810798, + "step": 297, + "train_runtime": 777.235, + "train_tokens_per_second": 6189.631 + }, + { + "epoch": 1.017094017094017, + "grad_norm": 0.10077892988920212, + "learning_rate": 9.948364888123925e-05, + "loss": 0.3721, + "num_input_tokens_seen": 4821918, + "step": 298, + "train_runtime": 778.9429, + "train_tokens_per_second": 6190.336 + }, + { + "epoch": 1.0205128205128204, + "grad_norm": 0.09540440887212753, + "learning_rate": 9.91394148020654e-05, + "loss": 0.4804, + "num_input_tokens_seen": 4839806, + "step": 299, + "train_runtime": 781.5496, + "train_tokens_per_second": 6192.577 + }, + { + "epoch": 1.0239316239316238, + "grad_norm": 0.09200432151556015, + "learning_rate": 9.879518072289157e-05, + "loss": 0.4185, + "num_input_tokens_seen": 4855374, + "step": 300, + "train_runtime": 783.8509, + "train_tokens_per_second": 6194.257 + }, + { + "epoch": 1.0273504273504273, + "grad_norm": 0.09225641936063766, + "learning_rate": 9.845094664371774e-05, + "loss": 0.3448, + "num_input_tokens_seen": 4875646, + "step": 301, + "train_runtime": 787.3218, + "train_tokens_per_second": 6192.698 + }, + { + "epoch": 1.0307692307692307, + "grad_norm": 0.09093517065048218, + "learning_rate": 9.81067125645439e-05, + "loss": 0.3735, + "num_input_tokens_seen": 4889006, + "step": 302, + "train_runtime": 789.3263, + "train_tokens_per_second": 6193.897 + }, + { + "epoch": 1.0341880341880343, + "grad_norm": 0.1052882969379425, + "learning_rate": 9.776247848537006e-05, + "loss": 0.4364, + "num_input_tokens_seen": 4902686, + "step": 303, + "train_runtime": 791.403, + "train_tokens_per_second": 6194.93 + }, + { + "epoch": 1.0376068376068377, + "grad_norm": 0.09175313264131546, + "learning_rate": 9.741824440619622e-05, + "loss": 0.4034, + "num_input_tokens_seen": 4922078, + "step": 304, + "train_runtime": 794.1808, + "train_tokens_per_second": 6197.679 + }, + { + "epoch": 1.041025641025641, + "grad_norm": 0.09906546771526337, + "learning_rate": 9.707401032702239e-05, + "loss": 0.448, + "num_input_tokens_seen": 4943358, + "step": 305, + "train_runtime": 797.2477, + "train_tokens_per_second": 6200.53 + }, + { + "epoch": 1.0444444444444445, + "grad_norm": 0.10816071182489395, + "learning_rate": 9.672977624784855e-05, + "loss": 0.3971, + "num_input_tokens_seen": 4956078, + "step": 306, + "train_runtime": 799.1676, + "train_tokens_per_second": 6201.55 + }, + { + "epoch": 1.047863247863248, + "grad_norm": 0.09161638468503952, + "learning_rate": 9.638554216867471e-05, + "loss": 0.3514, + "num_input_tokens_seen": 4981710, + "step": 307, + "train_runtime": 802.8012, + "train_tokens_per_second": 6205.409 + }, + { + "epoch": 1.0512820512820513, + "grad_norm": 0.09889056533575058, + "learning_rate": 9.604130808950088e-05, + "loss": 0.3105, + "num_input_tokens_seen": 5001214, + "step": 308, + "train_runtime": 805.5994, + "train_tokens_per_second": 6208.066 + }, + { + "epoch": 1.0547008547008547, + "grad_norm": 0.1137273982167244, + "learning_rate": 9.569707401032703e-05, + "loss": 0.3571, + "num_input_tokens_seen": 5017694, + "step": 309, + "train_runtime": 808.0045, + "train_tokens_per_second": 6209.982 + }, + { + "epoch": 1.0581196581196581, + "grad_norm": 0.09778351336717606, + "learning_rate": 9.53528399311532e-05, + "loss": 0.3879, + "num_input_tokens_seen": 5035758, + "step": 310, + "train_runtime": 810.6114, + "train_tokens_per_second": 6212.296 + }, + { + "epoch": 1.0615384615384615, + "grad_norm": 0.1118144690990448, + "learning_rate": 9.500860585197935e-05, + "loss": 0.3965, + "num_input_tokens_seen": 5052750, + "step": 311, + "train_runtime": 813.0944, + "train_tokens_per_second": 6214.223 + }, + { + "epoch": 1.064957264957265, + "grad_norm": 0.11538024991750717, + "learning_rate": 9.466437177280552e-05, + "loss": 0.4422, + "num_input_tokens_seen": 5067630, + "step": 312, + "train_runtime": 815.286, + "train_tokens_per_second": 6215.77 + }, + { + "epoch": 1.0683760683760684, + "grad_norm": 0.10927511751651764, + "learning_rate": 9.432013769363168e-05, + "loss": 0.4174, + "num_input_tokens_seen": 5083150, + "step": 313, + "train_runtime": 817.5852, + "train_tokens_per_second": 6217.272 + }, + { + "epoch": 1.0717948717948718, + "grad_norm": 0.10767415165901184, + "learning_rate": 9.397590361445784e-05, + "loss": 0.3619, + "num_input_tokens_seen": 5104078, + "step": 314, + "train_runtime": 820.6095, + "train_tokens_per_second": 6219.862 + }, + { + "epoch": 1.0752136752136752, + "grad_norm": 0.10561515390872955, + "learning_rate": 9.3631669535284e-05, + "loss": 0.4125, + "num_input_tokens_seen": 5118014, + "step": 315, + "train_runtime": 822.6736, + "train_tokens_per_second": 6221.197 + }, + { + "epoch": 1.0786324786324786, + "grad_norm": 0.10327659547328949, + "learning_rate": 9.328743545611017e-05, + "loss": 0.4252, + "num_input_tokens_seen": 5137150, + "step": 316, + "train_runtime": 825.4032, + "train_tokens_per_second": 6223.807 + }, + { + "epoch": 1.082051282051282, + "grad_norm": 0.1158551275730133, + "learning_rate": 9.294320137693632e-05, + "loss": 0.3862, + "num_input_tokens_seen": 5152190, + "step": 317, + "train_runtime": 827.6678, + "train_tokens_per_second": 6224.949 + }, + { + "epoch": 1.0854700854700854, + "grad_norm": 0.11891324818134308, + "learning_rate": 9.259896729776249e-05, + "loss": 0.4514, + "num_input_tokens_seen": 5165790, + "step": 318, + "train_runtime": 829.6973, + "train_tokens_per_second": 6226.114 + }, + { + "epoch": 1.0888888888888888, + "grad_norm": 0.09706568717956543, + "learning_rate": 9.225473321858864e-05, + "loss": 0.4038, + "num_input_tokens_seen": 5187630, + "step": 319, + "train_runtime": 832.8397, + "train_tokens_per_second": 6228.846 + }, + { + "epoch": 1.0923076923076924, + "grad_norm": 0.1240311861038208, + "learning_rate": 9.191049913941481e-05, + "loss": 0.4434, + "num_input_tokens_seen": 5201454, + "step": 320, + "train_runtime": 834.8979, + "train_tokens_per_second": 6230.048 + }, + { + "epoch": 1.0957264957264958, + "grad_norm": 0.10608533769845963, + "learning_rate": 9.156626506024096e-05, + "loss": 0.3612, + "num_input_tokens_seen": 5217502, + "step": 321, + "train_runtime": 837.243, + "train_tokens_per_second": 6231.766 + }, + { + "epoch": 1.0991452991452992, + "grad_norm": 0.11198005080223083, + "learning_rate": 9.122203098106713e-05, + "loss": 0.3461, + "num_input_tokens_seen": 5233502, + "step": 322, + "train_runtime": 839.607, + "train_tokens_per_second": 6233.276 + }, + { + "epoch": 1.1025641025641026, + "grad_norm": 0.11071202903985977, + "learning_rate": 9.08777969018933e-05, + "loss": 0.3657, + "num_input_tokens_seen": 5247166, + "step": 323, + "train_runtime": 841.6904, + "train_tokens_per_second": 6234.081 + }, + { + "epoch": 1.105982905982906, + "grad_norm": 0.128941148519516, + "learning_rate": 9.053356282271946e-05, + "loss": 0.3202, + "num_input_tokens_seen": 5264302, + "step": 324, + "train_runtime": 844.1905, + "train_tokens_per_second": 6235.917 + }, + { + "epoch": 1.1094017094017095, + "grad_norm": 0.10935687273740768, + "learning_rate": 9.018932874354562e-05, + "loss": 0.3335, + "num_input_tokens_seen": 5285150, + "step": 325, + "train_runtime": 847.1788, + "train_tokens_per_second": 6238.53 + }, + { + "epoch": 1.1128205128205129, + "grad_norm": 0.11587445437908173, + "learning_rate": 8.984509466437178e-05, + "loss": 0.3548, + "num_input_tokens_seen": 5303454, + "step": 326, + "train_runtime": 849.8521, + "train_tokens_per_second": 6240.443 + }, + { + "epoch": 1.1162393162393163, + "grad_norm": 0.14800626039505005, + "learning_rate": 8.950086058519795e-05, + "loss": 0.4223, + "num_input_tokens_seen": 5313854, + "step": 327, + "train_runtime": 851.4701, + "train_tokens_per_second": 6240.8 + }, + { + "epoch": 1.1196581196581197, + "grad_norm": 0.12176591157913208, + "learning_rate": 8.91566265060241e-05, + "loss": 0.3805, + "num_input_tokens_seen": 5327342, + "step": 328, + "train_runtime": 853.4843, + "train_tokens_per_second": 6241.875 + }, + { + "epoch": 1.123076923076923, + "grad_norm": 0.1277189403772354, + "learning_rate": 8.881239242685027e-05, + "loss": 0.4071, + "num_input_tokens_seen": 5343678, + "step": 329, + "train_runtime": 855.8801, + "train_tokens_per_second": 6243.489 + }, + { + "epoch": 1.1264957264957265, + "grad_norm": 0.10131321847438812, + "learning_rate": 8.846815834767642e-05, + "loss": 0.2666, + "num_input_tokens_seen": 5364910, + "step": 330, + "train_runtime": 858.9596, + "train_tokens_per_second": 6245.823 + }, + { + "epoch": 1.12991452991453, + "grad_norm": 0.12116528302431107, + "learning_rate": 8.812392426850259e-05, + "loss": 0.4034, + "num_input_tokens_seen": 5379150, + "step": 331, + "train_runtime": 861.6047, + "train_tokens_per_second": 6243.176 + }, + { + "epoch": 1.1333333333333333, + "grad_norm": 0.13882531225681305, + "learning_rate": 8.777969018932875e-05, + "loss": 0.3558, + "num_input_tokens_seen": 5392910, + "step": 332, + "train_runtime": 863.675, + "train_tokens_per_second": 6244.142 + }, + { + "epoch": 1.1367521367521367, + "grad_norm": 0.10566709190607071, + "learning_rate": 8.743545611015491e-05, + "loss": 0.3374, + "num_input_tokens_seen": 5410798, + "step": 333, + "train_runtime": 866.263, + "train_tokens_per_second": 6246.137 + }, + { + "epoch": 1.1401709401709401, + "grad_norm": 0.1450272500514984, + "learning_rate": 8.709122203098107e-05, + "loss": 0.3957, + "num_input_tokens_seen": 5425198, + "step": 334, + "train_runtime": 868.4297, + "train_tokens_per_second": 6247.136 + }, + { + "epoch": 1.1435897435897435, + "grad_norm": 0.13562506437301636, + "learning_rate": 8.674698795180724e-05, + "loss": 0.4037, + "num_input_tokens_seen": 5439374, + "step": 335, + "train_runtime": 870.5364, + "train_tokens_per_second": 6248.301 + }, + { + "epoch": 1.147008547008547, + "grad_norm": 0.12343757599592209, + "learning_rate": 8.640275387263339e-05, + "loss": 0.4106, + "num_input_tokens_seen": 5456574, + "step": 336, + "train_runtime": 873.0308, + "train_tokens_per_second": 6250.151 + }, + { + "epoch": 1.1504273504273503, + "grad_norm": 0.12856614589691162, + "learning_rate": 8.605851979345956e-05, + "loss": 0.4445, + "num_input_tokens_seen": 5475102, + "step": 337, + "train_runtime": 875.7122, + "train_tokens_per_second": 6252.17 + }, + { + "epoch": 1.1538461538461537, + "grad_norm": 0.13173913955688477, + "learning_rate": 8.571428571428571e-05, + "loss": 0.412, + "num_input_tokens_seen": 5488254, + "step": 338, + "train_runtime": 877.6958, + "train_tokens_per_second": 6253.025 + }, + { + "epoch": 1.1572649572649572, + "grad_norm": 0.13166651129722595, + "learning_rate": 8.537005163511188e-05, + "loss": 0.3414, + "num_input_tokens_seen": 5503694, + "step": 339, + "train_runtime": 879.9753, + "train_tokens_per_second": 6254.373 + }, + { + "epoch": 1.1606837606837608, + "grad_norm": 0.14514580368995667, + "learning_rate": 8.502581755593804e-05, + "loss": 0.4121, + "num_input_tokens_seen": 5518158, + "step": 340, + "train_runtime": 882.0821, + "train_tokens_per_second": 6255.833 + }, + { + "epoch": 1.1641025641025642, + "grad_norm": 0.13414417207241058, + "learning_rate": 8.46815834767642e-05, + "loss": 0.3738, + "num_input_tokens_seen": 5531150, + "step": 341, + "train_runtime": 884.0358, + "train_tokens_per_second": 6256.704 + }, + { + "epoch": 1.1675213675213676, + "grad_norm": 0.13111338019371033, + "learning_rate": 8.433734939759037e-05, + "loss": 0.4629, + "num_input_tokens_seen": 5547870, + "step": 342, + "train_runtime": 886.4774, + "train_tokens_per_second": 6258.332 + }, + { + "epoch": 1.170940170940171, + "grad_norm": 0.13821318745613098, + "learning_rate": 8.399311531841653e-05, + "loss": 0.3793, + "num_input_tokens_seen": 5560814, + "step": 343, + "train_runtime": 888.4364, + "train_tokens_per_second": 6259.102 + }, + { + "epoch": 1.1743589743589744, + "grad_norm": 0.133073627948761, + "learning_rate": 8.36488812392427e-05, + "loss": 0.3765, + "num_input_tokens_seen": 5573470, + "step": 344, + "train_runtime": 890.3654, + "train_tokens_per_second": 6259.756 + }, + { + "epoch": 1.1777777777777778, + "grad_norm": 0.14837369322776794, + "learning_rate": 8.330464716006885e-05, + "loss": 0.4345, + "num_input_tokens_seen": 5589358, + "step": 345, + "train_runtime": 892.7259, + "train_tokens_per_second": 6261.001 + }, + { + "epoch": 1.1811965811965812, + "grad_norm": 0.11554093658924103, + "learning_rate": 8.296041308089502e-05, + "loss": 0.3454, + "num_input_tokens_seen": 5608286, + "step": 346, + "train_runtime": 895.4735, + "train_tokens_per_second": 6262.928 + }, + { + "epoch": 1.1846153846153846, + "grad_norm": 0.13670602440834045, + "learning_rate": 8.261617900172117e-05, + "loss": 0.3641, + "num_input_tokens_seen": 5624222, + "step": 347, + "train_runtime": 897.8188, + "train_tokens_per_second": 6264.317 + }, + { + "epoch": 1.188034188034188, + "grad_norm": 0.11645640432834625, + "learning_rate": 8.227194492254734e-05, + "loss": 0.3086, + "num_input_tokens_seen": 5644302, + "step": 348, + "train_runtime": 900.68, + "train_tokens_per_second": 6266.711 + }, + { + "epoch": 1.1914529914529914, + "grad_norm": 0.13790756464004517, + "learning_rate": 8.192771084337349e-05, + "loss": 0.3614, + "num_input_tokens_seen": 5660814, + "step": 349, + "train_runtime": 903.1107, + "train_tokens_per_second": 6268.129 + }, + { + "epoch": 1.1948717948717948, + "grad_norm": 0.12048908323049545, + "learning_rate": 8.158347676419966e-05, + "loss": 0.3175, + "num_input_tokens_seen": 5678238, + "step": 350, + "train_runtime": 905.6459, + "train_tokens_per_second": 6269.822 + }, + { + "epoch": 1.1982905982905983, + "grad_norm": 0.1469748467206955, + "learning_rate": 8.123924268502582e-05, + "loss": 0.402, + "num_input_tokens_seen": 5690766, + "step": 351, + "train_runtime": 907.5181, + "train_tokens_per_second": 6270.692 + }, + { + "epoch": 1.2017094017094017, + "grad_norm": 0.15761221945285797, + "learning_rate": 8.089500860585198e-05, + "loss": 0.4114, + "num_input_tokens_seen": 5704126, + "step": 352, + "train_runtime": 909.528, + "train_tokens_per_second": 6271.523 + }, + { + "epoch": 1.205128205128205, + "grad_norm": 0.12943622469902039, + "learning_rate": 8.055077452667814e-05, + "loss": 0.3823, + "num_input_tokens_seen": 5721070, + "step": 353, + "train_runtime": 912.0326, + "train_tokens_per_second": 6272.879 + }, + { + "epoch": 1.2085470085470085, + "grad_norm": 0.14878448843955994, + "learning_rate": 8.02065404475043e-05, + "loss": 0.4004, + "num_input_tokens_seen": 5738638, + "step": 354, + "train_runtime": 914.5921, + "train_tokens_per_second": 6274.532 + }, + { + "epoch": 1.2085470085470085, + "eval_loss": 0.4262903034687042, + "eval_runtime": 11.8731, + "eval_samples_per_second": 84.056, + "eval_steps_per_second": 5.306, + "num_input_tokens_seen": 5738638, + "step": 354 + }, + { + "epoch": 1.2119658119658119, + "grad_norm": 0.15027886629104614, + "learning_rate": 7.986230636833046e-05, + "loss": 0.352, + "num_input_tokens_seen": 5755054, + "step": 355, + "train_runtime": 928.8681, + "train_tokens_per_second": 6195.771 + }, + { + "epoch": 1.2153846153846155, + "grad_norm": 0.13495990633964539, + "learning_rate": 7.951807228915663e-05, + "loss": 0.3563, + "num_input_tokens_seen": 5767742, + "step": 356, + "train_runtime": 930.7674, + "train_tokens_per_second": 6196.76 + }, + { + "epoch": 1.218803418803419, + "grad_norm": 0.1535864919424057, + "learning_rate": 7.917383820998278e-05, + "loss": 0.3608, + "num_input_tokens_seen": 5787614, + "step": 357, + "train_runtime": 933.6119, + "train_tokens_per_second": 6199.164 + }, + { + "epoch": 1.2222222222222223, + "grad_norm": 0.13963304460048676, + "learning_rate": 7.882960413080895e-05, + "loss": 0.4206, + "num_input_tokens_seen": 5802558, + "step": 358, + "train_runtime": 935.8349, + "train_tokens_per_second": 6200.407 + }, + { + "epoch": 1.2256410256410257, + "grad_norm": 0.163704514503479, + "learning_rate": 7.848537005163512e-05, + "loss": 0.391, + "num_input_tokens_seen": 5816382, + "step": 359, + "train_runtime": 937.9276, + "train_tokens_per_second": 6201.312 + }, + { + "epoch": 1.2290598290598291, + "grad_norm": 0.14221028983592987, + "learning_rate": 7.814113597246127e-05, + "loss": 0.3529, + "num_input_tokens_seen": 5833550, + "step": 360, + "train_runtime": 940.4404, + "train_tokens_per_second": 6202.998 + }, + { + "epoch": 1.2324786324786325, + "grad_norm": 0.14981484413146973, + "learning_rate": 7.779690189328744e-05, + "loss": 0.4097, + "num_input_tokens_seen": 5847742, + "step": 361, + "train_runtime": 943.0512, + "train_tokens_per_second": 6200.874 + }, + { + "epoch": 1.235897435897436, + "grad_norm": 0.15353192389011383, + "learning_rate": 7.74526678141136e-05, + "loss": 0.4207, + "num_input_tokens_seen": 5861422, + "step": 362, + "train_runtime": 945.0925, + "train_tokens_per_second": 6201.956 + }, + { + "epoch": 1.2393162393162394, + "grad_norm": 0.1498330533504486, + "learning_rate": 7.710843373493976e-05, + "loss": 0.4371, + "num_input_tokens_seen": 5879550, + "step": 363, + "train_runtime": 947.7032, + "train_tokens_per_second": 6203.999 + }, + { + "epoch": 1.2427350427350428, + "grad_norm": 0.16379210352897644, + "learning_rate": 7.676419965576592e-05, + "loss": 0.4001, + "num_input_tokens_seen": 5890414, + "step": 364, + "train_runtime": 949.3619, + "train_tokens_per_second": 6204.603 + }, + { + "epoch": 1.2461538461538462, + "grad_norm": 0.14771924912929535, + "learning_rate": 7.641996557659209e-05, + "loss": 0.4091, + "num_input_tokens_seen": 5903870, + "step": 365, + "train_runtime": 951.3898, + "train_tokens_per_second": 6205.522 + }, + { + "epoch": 1.2495726495726496, + "grad_norm": 0.1314336210489273, + "learning_rate": 7.607573149741824e-05, + "loss": 0.3686, + "num_input_tokens_seen": 5922702, + "step": 366, + "train_runtime": 954.0994, + "train_tokens_per_second": 6207.636 + }, + { + "epoch": 1.252991452991453, + "grad_norm": 0.14639326930046082, + "learning_rate": 7.573149741824441e-05, + "loss": 0.391, + "num_input_tokens_seen": 5943566, + "step": 367, + "train_runtime": 957.0573, + "train_tokens_per_second": 6210.251 + }, + { + "epoch": 1.2564102564102564, + "grad_norm": 0.14717791974544525, + "learning_rate": 7.538726333907056e-05, + "loss": 0.4646, + "num_input_tokens_seen": 5960878, + "step": 368, + "train_runtime": 959.6021, + "train_tokens_per_second": 6211.823 + }, + { + "epoch": 1.2598290598290598, + "grad_norm": 0.14937731623649597, + "learning_rate": 7.504302925989673e-05, + "loss": 0.448, + "num_input_tokens_seen": 5976654, + "step": 369, + "train_runtime": 961.9381, + "train_tokens_per_second": 6213.138 + }, + { + "epoch": 1.2632478632478632, + "grad_norm": 0.15321378409862518, + "learning_rate": 7.469879518072289e-05, + "loss": 0.4242, + "num_input_tokens_seen": 5989406, + "step": 370, + "train_runtime": 963.873, + "train_tokens_per_second": 6213.896 + }, + { + "epoch": 1.2666666666666666, + "grad_norm": 0.13933706283569336, + "learning_rate": 7.435456110154905e-05, + "loss": 0.4543, + "num_input_tokens_seen": 6007710, + "step": 371, + "train_runtime": 966.5773, + "train_tokens_per_second": 6215.447 + }, + { + "epoch": 1.27008547008547, + "grad_norm": 0.1422203630208969, + "learning_rate": 7.401032702237521e-05, + "loss": 0.3656, + "num_input_tokens_seen": 6021982, + "step": 372, + "train_runtime": 968.6875, + "train_tokens_per_second": 6216.641 + }, + { + "epoch": 1.2735042735042734, + "grad_norm": 0.15311431884765625, + "learning_rate": 7.366609294320138e-05, + "loss": 0.5369, + "num_input_tokens_seen": 6036606, + "step": 373, + "train_runtime": 970.8599, + "train_tokens_per_second": 6217.793 + }, + { + "epoch": 1.2769230769230768, + "grad_norm": 0.15368784964084625, + "learning_rate": 7.332185886402754e-05, + "loss": 0.4691, + "num_input_tokens_seen": 6050686, + "step": 374, + "train_runtime": 972.9604, + "train_tokens_per_second": 6218.841 + }, + { + "epoch": 1.2803418803418802, + "grad_norm": 0.128590390086174, + "learning_rate": 7.29776247848537e-05, + "loss": 0.3645, + "num_input_tokens_seen": 6071006, + "step": 375, + "train_runtime": 975.8603, + "train_tokens_per_second": 6221.183 + }, + { + "epoch": 1.2837606837606836, + "grad_norm": 0.16918911039829254, + "learning_rate": 7.263339070567987e-05, + "loss": 0.4516, + "num_input_tokens_seen": 6083838, + "step": 376, + "train_runtime": 977.7866, + "train_tokens_per_second": 6222.051 + }, + { + "epoch": 1.287179487179487, + "grad_norm": 0.1317807137966156, + "learning_rate": 7.228915662650602e-05, + "loss": 0.3979, + "num_input_tokens_seen": 6103870, + "step": 377, + "train_runtime": 980.6218, + "train_tokens_per_second": 6224.489 + }, + { + "epoch": 1.2905982905982907, + "grad_norm": 0.15844884514808655, + "learning_rate": 7.194492254733219e-05, + "loss": 0.4077, + "num_input_tokens_seen": 6119342, + "step": 378, + "train_runtime": 982.9217, + "train_tokens_per_second": 6225.666 + }, + { + "epoch": 1.294017094017094, + "grad_norm": 0.13094189763069153, + "learning_rate": 7.160068846815836e-05, + "loss": 0.3317, + "num_input_tokens_seen": 6135582, + "step": 379, + "train_runtime": 985.318, + "train_tokens_per_second": 6227.007 + }, + { + "epoch": 1.2974358974358975, + "grad_norm": 0.1428907960653305, + "learning_rate": 7.125645438898451e-05, + "loss": 0.3667, + "num_input_tokens_seen": 6153166, + "step": 380, + "train_runtime": 987.8695, + "train_tokens_per_second": 6228.724 + }, + { + "epoch": 1.300854700854701, + "grad_norm": 0.1479697972536087, + "learning_rate": 7.091222030981068e-05, + "loss": 0.4703, + "num_input_tokens_seen": 6167790, + "step": 381, + "train_runtime": 990.0311, + "train_tokens_per_second": 6229.895 + }, + { + "epoch": 1.3042735042735043, + "grad_norm": 0.1379973441362381, + "learning_rate": 7.056798623063683e-05, + "loss": 0.305, + "num_input_tokens_seen": 6188846, + "step": 382, + "train_runtime": 993.0644, + "train_tokens_per_second": 6232.069 + }, + { + "epoch": 1.3076923076923077, + "grad_norm": 0.15161173045635223, + "learning_rate": 7.0223752151463e-05, + "loss": 0.3361, + "num_input_tokens_seen": 6202542, + "step": 383, + "train_runtime": 995.1105, + "train_tokens_per_second": 6233.018 + }, + { + "epoch": 1.3111111111111111, + "grad_norm": 0.14290477335453033, + "learning_rate": 6.987951807228917e-05, + "loss": 0.3268, + "num_input_tokens_seen": 6220366, + "step": 384, + "train_runtime": 997.717, + "train_tokens_per_second": 6234.599 + }, + { + "epoch": 1.3145299145299145, + "grad_norm": 0.1553800255060196, + "learning_rate": 6.953528399311532e-05, + "loss": 0.4104, + "num_input_tokens_seen": 6237422, + "step": 385, + "train_runtime": 1000.1787, + "train_tokens_per_second": 6236.307 + }, + { + "epoch": 1.317948717948718, + "grad_norm": 0.15926386415958405, + "learning_rate": 6.919104991394149e-05, + "loss": 0.3809, + "num_input_tokens_seen": 6249070, + "step": 386, + "train_runtime": 1001.9601, + "train_tokens_per_second": 6236.845 + }, + { + "epoch": 1.3213675213675213, + "grad_norm": 0.15950970351696014, + "learning_rate": 6.884681583476765e-05, + "loss": 0.3429, + "num_input_tokens_seen": 6265198, + "step": 387, + "train_runtime": 1004.333, + "train_tokens_per_second": 6238.168 + }, + { + "epoch": 1.3247863247863247, + "grad_norm": 0.15029311180114746, + "learning_rate": 6.850258175559381e-05, + "loss": 0.3975, + "num_input_tokens_seen": 6281966, + "step": 388, + "train_runtime": 1006.7802, + "train_tokens_per_second": 6239.66 + }, + { + "epoch": 1.3282051282051281, + "grad_norm": 0.16369454562664032, + "learning_rate": 6.815834767641997e-05, + "loss": 0.3495, + "num_input_tokens_seen": 6298718, + "step": 389, + "train_runtime": 1009.2265, + "train_tokens_per_second": 6241.134 + }, + { + "epoch": 1.3316239316239316, + "grad_norm": 0.1701713353395462, + "learning_rate": 6.781411359724614e-05, + "loss": 0.4102, + "num_input_tokens_seen": 6312510, + "step": 390, + "train_runtime": 1011.2977, + "train_tokens_per_second": 6241.99 + }, + { + "epoch": 1.335042735042735, + "grad_norm": 0.15310458838939667, + "learning_rate": 6.746987951807229e-05, + "loss": 0.3758, + "num_input_tokens_seen": 6329998, + "step": 391, + "train_runtime": 1014.3612, + "train_tokens_per_second": 6240.379 + }, + { + "epoch": 1.3384615384615386, + "grad_norm": 0.1793118566274643, + "learning_rate": 6.712564543889846e-05, + "loss": 0.4579, + "num_input_tokens_seen": 6344158, + "step": 392, + "train_runtime": 1016.4559, + "train_tokens_per_second": 6241.449 + }, + { + "epoch": 1.341880341880342, + "grad_norm": 0.1364370435476303, + "learning_rate": 6.678141135972461e-05, + "loss": 0.3032, + "num_input_tokens_seen": 6362286, + "step": 393, + "train_runtime": 1019.0898, + "train_tokens_per_second": 6243.106 + }, + { + "epoch": 1.3452991452991454, + "grad_norm": 0.19111692905426025, + "learning_rate": 6.643717728055078e-05, + "loss": 0.3836, + "num_input_tokens_seen": 6374334, + "step": 394, + "train_runtime": 1020.8971, + "train_tokens_per_second": 6243.856 + }, + { + "epoch": 1.3487179487179488, + "grad_norm": 0.13976141810417175, + "learning_rate": 6.609294320137695e-05, + "loss": 0.3489, + "num_input_tokens_seen": 6392286, + "step": 395, + "train_runtime": 1023.5184, + "train_tokens_per_second": 6245.404 + }, + { + "epoch": 1.3521367521367522, + "grad_norm": 0.1717049479484558, + "learning_rate": 6.57487091222031e-05, + "loss": 0.4546, + "num_input_tokens_seen": 6407102, + "step": 396, + "train_runtime": 1025.7201, + "train_tokens_per_second": 6246.443 + }, + { + "epoch": 1.3555555555555556, + "grad_norm": 0.15389494597911835, + "learning_rate": 6.540447504302927e-05, + "loss": 0.3984, + "num_input_tokens_seen": 6424190, + "step": 397, + "train_runtime": 1028.1999, + "train_tokens_per_second": 6247.997 + }, + { + "epoch": 1.358974358974359, + "grad_norm": 0.15128876268863678, + "learning_rate": 6.506024096385543e-05, + "loss": 0.4107, + "num_input_tokens_seen": 6438926, + "step": 398, + "train_runtime": 1030.39, + "train_tokens_per_second": 6249.018 + }, + { + "epoch": 1.3623931623931624, + "grad_norm": 0.1720062643289566, + "learning_rate": 6.47160068846816e-05, + "loss": 0.4398, + "num_input_tokens_seen": 6454206, + "step": 399, + "train_runtime": 1032.6825, + "train_tokens_per_second": 6249.942 + }, + { + "epoch": 1.3658119658119658, + "grad_norm": 0.16892306506633759, + "learning_rate": 6.437177280550775e-05, + "loss": 0.4768, + "num_input_tokens_seen": 6470446, + "step": 400, + "train_runtime": 1035.0677, + "train_tokens_per_second": 6251.23 + }, + { + "epoch": 1.3692307692307693, + "grad_norm": 0.16459856927394867, + "learning_rate": 6.402753872633392e-05, + "loss": 0.394, + "num_input_tokens_seen": 6485566, + "step": 401, + "train_runtime": 1037.3161, + "train_tokens_per_second": 6252.256 + }, + { + "epoch": 1.3726495726495727, + "grad_norm": 0.146686390042305, + "learning_rate": 6.368330464716007e-05, + "loss": 0.4074, + "num_input_tokens_seen": 6500878, + "step": 402, + "train_runtime": 1039.5797, + "train_tokens_per_second": 6253.371 + }, + { + "epoch": 1.376068376068376, + "grad_norm": 0.13347510993480682, + "learning_rate": 6.333907056798624e-05, + "loss": 0.3718, + "num_input_tokens_seen": 6527726, + "step": 403, + "train_runtime": 1043.3774, + "train_tokens_per_second": 6256.342 + }, + { + "epoch": 1.3794871794871795, + "grad_norm": 0.15149444341659546, + "learning_rate": 6.29948364888124e-05, + "loss": 0.4889, + "num_input_tokens_seen": 6549054, + "step": 404, + "train_runtime": 1046.4485, + "train_tokens_per_second": 6258.363 + }, + { + "epoch": 1.3829059829059829, + "grad_norm": 0.15119557082653046, + "learning_rate": 6.265060240963856e-05, + "loss": 0.3562, + "num_input_tokens_seen": 6565374, + "step": 405, + "train_runtime": 1048.8497, + "train_tokens_per_second": 6259.595 + }, + { + "epoch": 1.3863247863247863, + "grad_norm": 0.1899130493402481, + "learning_rate": 6.230636833046472e-05, + "loss": 0.4359, + "num_input_tokens_seen": 6577854, + "step": 406, + "train_runtime": 1050.7387, + "train_tokens_per_second": 6260.219 + }, + { + "epoch": 1.3897435897435897, + "grad_norm": 0.15261997282505035, + "learning_rate": 6.196213425129088e-05, + "loss": 0.3755, + "num_input_tokens_seen": 6595182, + "step": 407, + "train_runtime": 1053.2481, + "train_tokens_per_second": 6261.755 + }, + { + "epoch": 1.393162393162393, + "grad_norm": 0.1569528728723526, + "learning_rate": 6.161790017211704e-05, + "loss": 0.4357, + "num_input_tokens_seen": 6609166, + "step": 408, + "train_runtime": 1055.3193, + "train_tokens_per_second": 6262.717 + }, + { + "epoch": 1.3965811965811965, + "grad_norm": 0.15785826742649078, + "learning_rate": 6.127366609294321e-05, + "loss": 0.3605, + "num_input_tokens_seen": 6622254, + "step": 409, + "train_runtime": 1057.2748, + "train_tokens_per_second": 6263.513 + }, + { + "epoch": 1.4, + "grad_norm": 0.1593894064426422, + "learning_rate": 6.092943201376936e-05, + "loss": 0.461, + "num_input_tokens_seen": 6637758, + "step": 410, + "train_runtime": 1059.5389, + "train_tokens_per_second": 6264.761 + }, + { + "epoch": 1.4034188034188033, + "grad_norm": 0.1655980795621872, + "learning_rate": 6.058519793459553e-05, + "loss": 0.3856, + "num_input_tokens_seen": 6653086, + "step": 411, + "train_runtime": 1061.8036, + "train_tokens_per_second": 6265.835 + }, + { + "epoch": 1.4068376068376067, + "grad_norm": 0.16413439810276031, + "learning_rate": 6.02409638554217e-05, + "loss": 0.3601, + "num_input_tokens_seen": 6671230, + "step": 412, + "train_runtime": 1064.4168, + "train_tokens_per_second": 6267.498 + }, + { + "epoch": 1.4102564102564101, + "grad_norm": 0.17654147744178772, + "learning_rate": 5.989672977624785e-05, + "loss": 0.4383, + "num_input_tokens_seen": 6685294, + "step": 413, + "train_runtime": 1066.4829, + "train_tokens_per_second": 6268.543 + }, + { + "epoch": 1.4102564102564101, + "eval_loss": 0.42687925696372986, + "eval_runtime": 11.8448, + "eval_samples_per_second": 84.257, + "eval_steps_per_second": 5.319, + "num_input_tokens_seen": 6685294, + "step": 413 + }, + { + "epoch": 1.4136752136752135, + "grad_norm": 0.14143113791942596, + "learning_rate": 5.955249569707402e-05, + "loss": 0.3417, + "num_input_tokens_seen": 6706110, + "step": 414, + "train_runtime": 1081.3024, + "train_tokens_per_second": 6201.882 + }, + { + "epoch": 1.4170940170940172, + "grad_norm": 0.17738325893878937, + "learning_rate": 5.9208261617900174e-05, + "loss": 0.4049, + "num_input_tokens_seen": 6719390, + "step": 415, + "train_runtime": 1083.3114, + "train_tokens_per_second": 6202.64 + }, + { + "epoch": 1.4205128205128206, + "grad_norm": 0.15967565774917603, + "learning_rate": 5.886402753872634e-05, + "loss": 0.322, + "num_input_tokens_seen": 6732926, + "step": 416, + "train_runtime": 1085.3328, + "train_tokens_per_second": 6203.559 + }, + { + "epoch": 1.423931623931624, + "grad_norm": 0.17965038120746613, + "learning_rate": 5.85197934595525e-05, + "loss": 0.4321, + "num_input_tokens_seen": 6748014, + "step": 417, + "train_runtime": 1087.5395, + "train_tokens_per_second": 6204.845 + }, + { + "epoch": 1.4273504273504274, + "grad_norm": 0.17158453166484833, + "learning_rate": 5.8175559380378665e-05, + "loss": 0.3844, + "num_input_tokens_seen": 6762782, + "step": 418, + "train_runtime": 1089.7402, + "train_tokens_per_second": 6205.866 + }, + { + "epoch": 1.4307692307692308, + "grad_norm": 0.17010599374771118, + "learning_rate": 5.783132530120482e-05, + "loss": 0.3944, + "num_input_tokens_seen": 6775838, + "step": 419, + "train_runtime": 1091.687, + "train_tokens_per_second": 6206.759 + }, + { + "epoch": 1.4341880341880342, + "grad_norm": 0.14824308454990387, + "learning_rate": 5.748709122203099e-05, + "loss": 0.3437, + "num_input_tokens_seen": 6794014, + "step": 420, + "train_runtime": 1094.3074, + "train_tokens_per_second": 6208.506 + }, + { + "epoch": 1.4376068376068376, + "grad_norm": 0.1721956878900528, + "learning_rate": 5.714285714285714e-05, + "loss": 0.3813, + "num_input_tokens_seen": 6809502, + "step": 421, + "train_runtime": 1097.1486, + "train_tokens_per_second": 6206.545 + }, + { + "epoch": 1.441025641025641, + "grad_norm": 0.17868974804878235, + "learning_rate": 5.679862306368331e-05, + "loss": 0.439, + "num_input_tokens_seen": 6823038, + "step": 422, + "train_runtime": 1099.1662, + "train_tokens_per_second": 6207.467 + }, + { + "epoch": 1.4444444444444444, + "grad_norm": 0.15935160219669342, + "learning_rate": 5.6454388984509464e-05, + "loss": 0.387, + "num_input_tokens_seen": 6841582, + "step": 423, + "train_runtime": 1101.8557, + "train_tokens_per_second": 6209.145 + }, + { + "epoch": 1.4478632478632478, + "grad_norm": 0.1758459508419037, + "learning_rate": 5.611015490533563e-05, + "loss": 0.5356, + "num_input_tokens_seen": 6855310, + "step": 424, + "train_runtime": 1103.8914, + "train_tokens_per_second": 6210.131 + }, + { + "epoch": 1.4512820512820512, + "grad_norm": 0.16828133165836334, + "learning_rate": 5.576592082616179e-05, + "loss": 0.3835, + "num_input_tokens_seen": 6875486, + "step": 425, + "train_runtime": 1106.7831, + "train_tokens_per_second": 6212.135 + }, + { + "epoch": 1.4547008547008546, + "grad_norm": 0.16434825956821442, + "learning_rate": 5.5421686746987955e-05, + "loss": 0.3765, + "num_input_tokens_seen": 6889502, + "step": 426, + "train_runtime": 1108.8374, + "train_tokens_per_second": 6213.266 + }, + { + "epoch": 1.458119658119658, + "grad_norm": 0.16002990305423737, + "learning_rate": 5.507745266781411e-05, + "loss": 0.4121, + "num_input_tokens_seen": 6908046, + "step": 427, + "train_runtime": 1111.5373, + "train_tokens_per_second": 6214.858 + }, + { + "epoch": 1.4615384615384617, + "grad_norm": 0.17183275520801544, + "learning_rate": 5.473321858864028e-05, + "loss": 0.439, + "num_input_tokens_seen": 6924190, + "step": 428, + "train_runtime": 1113.8957, + "train_tokens_per_second": 6216.192 + }, + { + "epoch": 1.464957264957265, + "grad_norm": 0.18027374148368835, + "learning_rate": 5.4388984509466445e-05, + "loss": 0.4895, + "num_input_tokens_seen": 6939070, + "step": 429, + "train_runtime": 1116.1017, + "train_tokens_per_second": 6217.238 + }, + { + "epoch": 1.4683760683760685, + "grad_norm": 0.15044456720352173, + "learning_rate": 5.40447504302926e-05, + "loss": 0.4429, + "num_input_tokens_seen": 6956478, + "step": 430, + "train_runtime": 1118.6746, + "train_tokens_per_second": 6218.5 + }, + { + "epoch": 1.471794871794872, + "grad_norm": 0.15003447234630585, + "learning_rate": 5.370051635111877e-05, + "loss": 0.5628, + "num_input_tokens_seen": 6973246, + "step": 431, + "train_runtime": 1121.1665, + "train_tokens_per_second": 6219.635 + }, + { + "epoch": 1.4752136752136753, + "grad_norm": 0.1591937392950058, + "learning_rate": 5.335628227194492e-05, + "loss": 0.3721, + "num_input_tokens_seen": 6986990, + "step": 432, + "train_runtime": 1123.2293, + "train_tokens_per_second": 6220.448 + }, + { + "epoch": 1.4786324786324787, + "grad_norm": 0.18640592694282532, + "learning_rate": 5.301204819277109e-05, + "loss": 0.3516, + "num_input_tokens_seen": 7001950, + "step": 433, + "train_runtime": 1125.4499, + "train_tokens_per_second": 6221.468 + }, + { + "epoch": 1.4820512820512821, + "grad_norm": 0.17443552613258362, + "learning_rate": 5.2667814113597244e-05, + "loss": 0.4784, + "num_input_tokens_seen": 7016318, + "step": 434, + "train_runtime": 1127.6255, + "train_tokens_per_second": 6222.206 + }, + { + "epoch": 1.4854700854700855, + "grad_norm": 0.1685023009777069, + "learning_rate": 5.232358003442341e-05, + "loss": 0.4227, + "num_input_tokens_seen": 7029358, + "step": 435, + "train_runtime": 1129.5922, + "train_tokens_per_second": 6222.916 + }, + { + "epoch": 1.488888888888889, + "grad_norm": 0.17190369963645935, + "learning_rate": 5.197934595524957e-05, + "loss": 0.3718, + "num_input_tokens_seen": 7046350, + "step": 436, + "train_runtime": 1132.1043, + "train_tokens_per_second": 6224.117 + }, + { + "epoch": 1.4923076923076923, + "grad_norm": 0.16871225833892822, + "learning_rate": 5.1635111876075735e-05, + "loss": 0.4095, + "num_input_tokens_seen": 7060286, + "step": 437, + "train_runtime": 1134.2429, + "train_tokens_per_second": 6224.669 + }, + { + "epoch": 1.4957264957264957, + "grad_norm": 0.15742996335029602, + "learning_rate": 5.1290877796901896e-05, + "loss": 0.3377, + "num_input_tokens_seen": 7077550, + "step": 438, + "train_runtime": 1136.7468, + "train_tokens_per_second": 6226.145 + }, + { + "epoch": 1.4991452991452991, + "grad_norm": 0.1806265264749527, + "learning_rate": 5.094664371772806e-05, + "loss": 0.4356, + "num_input_tokens_seen": 7096110, + "step": 439, + "train_runtime": 1139.4331, + "train_tokens_per_second": 6227.755 + }, + { + "epoch": 1.5025641025641026, + "grad_norm": 0.1986326277256012, + "learning_rate": 5.060240963855422e-05, + "loss": 0.421, + "num_input_tokens_seen": 7110462, + "step": 440, + "train_runtime": 1141.5636, + "train_tokens_per_second": 6228.704 + }, + { + "epoch": 1.505982905982906, + "grad_norm": 0.18825186789035797, + "learning_rate": 5.025817555938038e-05, + "loss": 0.3866, + "num_input_tokens_seen": 7123438, + "step": 441, + "train_runtime": 1143.5279, + "train_tokens_per_second": 6229.352 + }, + { + "epoch": 1.5094017094017094, + "grad_norm": 0.15836316347122192, + "learning_rate": 4.991394148020654e-05, + "loss": 0.3759, + "num_input_tokens_seen": 7139678, + "step": 442, + "train_runtime": 1145.9334, + "train_tokens_per_second": 6230.448 + }, + { + "epoch": 1.5128205128205128, + "grad_norm": 0.1643362194299698, + "learning_rate": 4.95697074010327e-05, + "loss": 0.3715, + "num_input_tokens_seen": 7152798, + "step": 443, + "train_runtime": 1147.927, + "train_tokens_per_second": 6231.056 + }, + { + "epoch": 1.5162393162393162, + "grad_norm": 0.18112708628177643, + "learning_rate": 4.922547332185887e-05, + "loss": 0.5004, + "num_input_tokens_seen": 7166718, + "step": 444, + "train_runtime": 1149.9906, + "train_tokens_per_second": 6231.98 + }, + { + "epoch": 1.5196581196581196, + "grad_norm": 0.17923805117607117, + "learning_rate": 4.888123924268503e-05, + "loss": 0.3967, + "num_input_tokens_seen": 7180718, + "step": 445, + "train_runtime": 1152.0495, + "train_tokens_per_second": 6232.994 + }, + { + "epoch": 1.523076923076923, + "grad_norm": 0.15098166465759277, + "learning_rate": 4.853700516351119e-05, + "loss": 0.379, + "num_input_tokens_seen": 7198382, + "step": 446, + "train_runtime": 1154.5661, + "train_tokens_per_second": 6234.708 + }, + { + "epoch": 1.5264957264957264, + "grad_norm": 0.17328427731990814, + "learning_rate": 4.8192771084337354e-05, + "loss": 0.3807, + "num_input_tokens_seen": 7213006, + "step": 447, + "train_runtime": 1156.7041, + "train_tokens_per_second": 6235.827 + }, + { + "epoch": 1.5299145299145298, + "grad_norm": 0.13623268902301788, + "learning_rate": 4.7848537005163515e-05, + "loss": 0.3011, + "num_input_tokens_seen": 7243582, + "step": 448, + "train_runtime": 1161.0841, + "train_tokens_per_second": 6238.637 + }, + { + "epoch": 1.5333333333333332, + "grad_norm": 0.23146113753318787, + "learning_rate": 4.7504302925989676e-05, + "loss": 0.4946, + "num_input_tokens_seen": 7257710, + "step": 449, + "train_runtime": 1163.1872, + "train_tokens_per_second": 6239.503 + }, + { + "epoch": 1.5367521367521366, + "grad_norm": 0.1464138627052307, + "learning_rate": 4.716006884681584e-05, + "loss": 0.3956, + "num_input_tokens_seen": 7275646, + "step": 450, + "train_runtime": 1165.763, + "train_tokens_per_second": 6241.102 + }, + { + "epoch": 1.54017094017094, + "grad_norm": 0.13081535696983337, + "learning_rate": 4.6815834767642e-05, + "loss": 0.351, + "num_input_tokens_seen": 7295710, + "step": 451, + "train_runtime": 1169.156, + "train_tokens_per_second": 6240.151 + }, + { + "epoch": 1.5435897435897434, + "grad_norm": 0.16835983097553253, + "learning_rate": 4.647160068846816e-05, + "loss": 0.4823, + "num_input_tokens_seen": 7314574, + "step": 452, + "train_runtime": 1171.8934, + "train_tokens_per_second": 6241.672 + }, + { + "epoch": 1.547008547008547, + "grad_norm": 0.12163835763931274, + "learning_rate": 4.612736660929432e-05, + "loss": 0.2982, + "num_input_tokens_seen": 7343022, + "step": 453, + "train_runtime": 1175.8796, + "train_tokens_per_second": 6244.706 + }, + { + "epoch": 1.5504273504273505, + "grad_norm": 0.14073219895362854, + "learning_rate": 4.578313253012048e-05, + "loss": 0.2964, + "num_input_tokens_seen": 7363630, + "step": 454, + "train_runtime": 1178.8416, + "train_tokens_per_second": 6246.497 + }, + { + "epoch": 1.5538461538461539, + "grad_norm": 0.1757187843322754, + "learning_rate": 4.543889845094665e-05, + "loss": 0.4486, + "num_input_tokens_seen": 7378830, + "step": 455, + "train_runtime": 1181.0635, + "train_tokens_per_second": 6247.615 + }, + { + "epoch": 1.5572649572649573, + "grad_norm": 0.16719017922878265, + "learning_rate": 4.509466437177281e-05, + "loss": 0.342, + "num_input_tokens_seen": 7396958, + "step": 456, + "train_runtime": 1183.6752, + "train_tokens_per_second": 6249.145 + }, + { + "epoch": 1.5606837606837607, + "grad_norm": 0.15160873532295227, + "learning_rate": 4.475043029259897e-05, + "loss": 0.3305, + "num_input_tokens_seen": 7419918, + "step": 457, + "train_runtime": 1186.9428, + "train_tokens_per_second": 6251.285 + }, + { + "epoch": 1.564102564102564, + "grad_norm": 0.16827233135700226, + "learning_rate": 4.4406196213425134e-05, + "loss": 0.3478, + "num_input_tokens_seen": 7438590, + "step": 458, + "train_runtime": 1189.6616, + "train_tokens_per_second": 6252.694 + }, + { + "epoch": 1.5675213675213675, + "grad_norm": 0.15692555904388428, + "learning_rate": 4.4061962134251295e-05, + "loss": 0.3317, + "num_input_tokens_seen": 7457614, + "step": 459, + "train_runtime": 1192.4245, + "train_tokens_per_second": 6254.16 + }, + { + "epoch": 1.570940170940171, + "grad_norm": 0.16111750900745392, + "learning_rate": 4.371772805507746e-05, + "loss": 0.3559, + "num_input_tokens_seen": 7478254, + "step": 460, + "train_runtime": 1195.4195, + "train_tokens_per_second": 6255.757 + }, + { + "epoch": 1.5743589743589743, + "grad_norm": 0.1745642125606537, + "learning_rate": 4.337349397590362e-05, + "loss": 0.3515, + "num_input_tokens_seen": 7491422, + "step": 461, + "train_runtime": 1197.3704, + "train_tokens_per_second": 6256.562 + }, + { + "epoch": 1.5777777777777777, + "grad_norm": 0.1923861801624298, + "learning_rate": 4.302925989672978e-05, + "loss": 0.4927, + "num_input_tokens_seen": 7508174, + "step": 462, + "train_runtime": 1199.809, + "train_tokens_per_second": 6257.808 + }, + { + "epoch": 1.5811965811965814, + "grad_norm": 0.19131140410900116, + "learning_rate": 4.268502581755594e-05, + "loss": 0.4331, + "num_input_tokens_seen": 7520590, + "step": 463, + "train_runtime": 1201.6752, + "train_tokens_per_second": 6258.422 + }, + { + "epoch": 1.5846153846153848, + "grad_norm": 0.16123589873313904, + "learning_rate": 4.23407917383821e-05, + "loss": 0.3504, + "num_input_tokens_seen": 7534430, + "step": 464, + "train_runtime": 1203.7756, + "train_tokens_per_second": 6258.999 + }, + { + "epoch": 1.5880341880341882, + "grad_norm": 0.1816190630197525, + "learning_rate": 4.199655765920826e-05, + "loss": 0.349, + "num_input_tokens_seen": 7550974, + "step": 465, + "train_runtime": 1206.2032, + "train_tokens_per_second": 6260.117 + }, + { + "epoch": 1.5914529914529916, + "grad_norm": 0.1618788093328476, + "learning_rate": 4.1652323580034424e-05, + "loss": 0.3963, + "num_input_tokens_seen": 7568494, + "step": 466, + "train_runtime": 1208.719, + "train_tokens_per_second": 6261.583 + }, + { + "epoch": 1.594871794871795, + "grad_norm": 0.1682547926902771, + "learning_rate": 4.1308089500860585e-05, + "loss": 0.4128, + "num_input_tokens_seen": 7584510, + "step": 467, + "train_runtime": 1211.076, + "train_tokens_per_second": 6262.621 + }, + { + "epoch": 1.5982905982905984, + "grad_norm": 0.17071352899074554, + "learning_rate": 4.0963855421686746e-05, + "loss": 0.4007, + "num_input_tokens_seen": 7601806, + "step": 468, + "train_runtime": 1213.648, + "train_tokens_per_second": 6263.6 + }, + { + "epoch": 1.6017094017094018, + "grad_norm": 0.19961658120155334, + "learning_rate": 4.061962134251291e-05, + "loss": 0.3831, + "num_input_tokens_seen": 7615134, + "step": 469, + "train_runtime": 1215.6143, + "train_tokens_per_second": 6264.433 + }, + { + "epoch": 1.6051282051282052, + "grad_norm": 0.15891629457473755, + "learning_rate": 4.027538726333907e-05, + "loss": 0.3579, + "num_input_tokens_seen": 7632446, + "step": 470, + "train_runtime": 1218.1415, + "train_tokens_per_second": 6265.648 + }, + { + "epoch": 1.6085470085470086, + "grad_norm": 0.18139025568962097, + "learning_rate": 3.993115318416523e-05, + "loss": 0.4023, + "num_input_tokens_seen": 7647038, + "step": 471, + "train_runtime": 1220.3009, + "train_tokens_per_second": 6266.519 + }, + { + "epoch": 1.611965811965812, + "grad_norm": 0.1726229041814804, + "learning_rate": 3.958691910499139e-05, + "loss": 0.3495, + "num_input_tokens_seen": 7664846, + "step": 472, + "train_runtime": 1222.889, + "train_tokens_per_second": 6267.818 + }, + { + "epoch": 1.611965811965812, + "eval_loss": 0.4264853894710541, + "eval_runtime": 11.8946, + "eval_samples_per_second": 83.904, + "eval_steps_per_second": 5.297, + "num_input_tokens_seen": 7664846, + "step": 472 + }, + { + "epoch": 1.6153846153846154, + "grad_norm": 0.2046596258878708, + "learning_rate": 3.924268502581756e-05, + "loss": 0.4611, + "num_input_tokens_seen": 7678558, + "step": 473, + "train_runtime": 1236.8201, + "train_tokens_per_second": 6208.306 + }, + { + "epoch": 1.6188034188034188, + "grad_norm": 0.18339784443378448, + "learning_rate": 3.889845094664372e-05, + "loss": 0.4687, + "num_input_tokens_seen": 7692542, + "step": 474, + "train_runtime": 1238.9304, + "train_tokens_per_second": 6209.019 + }, + { + "epoch": 1.6222222222222222, + "grad_norm": 0.15463566780090332, + "learning_rate": 3.855421686746988e-05, + "loss": 0.3533, + "num_input_tokens_seen": 7706094, + "step": 475, + "train_runtime": 1240.9949, + "train_tokens_per_second": 6209.61 + }, + { + "epoch": 1.6256410256410256, + "grad_norm": 0.18534722924232483, + "learning_rate": 3.820998278829604e-05, + "loss": 0.429, + "num_input_tokens_seen": 7723006, + "step": 476, + "train_runtime": 1243.4537, + "train_tokens_per_second": 6210.932 + }, + { + "epoch": 1.629059829059829, + "grad_norm": 0.16153299808502197, + "learning_rate": 3.7865748709122204e-05, + "loss": 0.368, + "num_input_tokens_seen": 7739230, + "step": 477, + "train_runtime": 1245.8498, + "train_tokens_per_second": 6212.009 + }, + { + "epoch": 1.6324786324786325, + "grad_norm": 0.18664556741714478, + "learning_rate": 3.7521514629948365e-05, + "loss": 0.3884, + "num_input_tokens_seen": 7753966, + "step": 478, + "train_runtime": 1248.0407, + "train_tokens_per_second": 6212.911 + }, + { + "epoch": 1.6358974358974359, + "grad_norm": 0.14830684661865234, + "learning_rate": 3.717728055077453e-05, + "loss": 0.3094, + "num_input_tokens_seen": 7771758, + "step": 479, + "train_runtime": 1250.6256, + "train_tokens_per_second": 6214.296 + }, + { + "epoch": 1.6393162393162393, + "grad_norm": 0.1431615799665451, + "learning_rate": 3.683304647160069e-05, + "loss": 0.3434, + "num_input_tokens_seen": 7788030, + "step": 480, + "train_runtime": 1253.0226, + "train_tokens_per_second": 6215.395 + }, + { + "epoch": 1.6427350427350427, + "grad_norm": 0.17345507442951202, + "learning_rate": 3.648881239242685e-05, + "loss": 0.4109, + "num_input_tokens_seen": 7805374, + "step": 481, + "train_runtime": 1256.1793, + "train_tokens_per_second": 6213.583 + }, + { + "epoch": 1.646153846153846, + "grad_norm": 0.18331271409988403, + "learning_rate": 3.614457831325301e-05, + "loss": 0.483, + "num_input_tokens_seen": 7821278, + "step": 482, + "train_runtime": 1258.5131, + "train_tokens_per_second": 6214.697 + }, + { + "epoch": 1.6495726495726495, + "grad_norm": 0.14575603604316711, + "learning_rate": 3.580034423407918e-05, + "loss": 0.3499, + "num_input_tokens_seen": 7838750, + "step": 483, + "train_runtime": 1261.0972, + "train_tokens_per_second": 6215.817 + }, + { + "epoch": 1.652991452991453, + "grad_norm": 0.16931626200675964, + "learning_rate": 3.545611015490534e-05, + "loss": 0.364, + "num_input_tokens_seen": 7855742, + "step": 484, + "train_runtime": 1263.5759, + "train_tokens_per_second": 6217.072 + }, + { + "epoch": 1.6564102564102563, + "grad_norm": 0.17259550094604492, + "learning_rate": 3.51118760757315e-05, + "loss": 0.4564, + "num_input_tokens_seen": 7874494, + "step": 485, + "train_runtime": 1266.3175, + "train_tokens_per_second": 6218.42 + }, + { + "epoch": 1.6598290598290597, + "grad_norm": 0.17083777487277985, + "learning_rate": 3.476764199655766e-05, + "loss": 0.3809, + "num_input_tokens_seen": 7890430, + "step": 486, + "train_runtime": 1268.6327, + "train_tokens_per_second": 6219.633 + }, + { + "epoch": 1.6632478632478631, + "grad_norm": 0.1671985685825348, + "learning_rate": 3.442340791738382e-05, + "loss": 0.3755, + "num_input_tokens_seen": 7906142, + "step": 487, + "train_runtime": 1270.96, + "train_tokens_per_second": 6220.606 + }, + { + "epoch": 1.6666666666666665, + "grad_norm": 0.18116803467273712, + "learning_rate": 3.4079173838209984e-05, + "loss": 0.3913, + "num_input_tokens_seen": 7922366, + "step": 488, + "train_runtime": 1273.3633, + "train_tokens_per_second": 6221.607 + }, + { + "epoch": 1.67008547008547, + "grad_norm": 0.18932628631591797, + "learning_rate": 3.3734939759036146e-05, + "loss": 0.4293, + "num_input_tokens_seen": 7934942, + "step": 489, + "train_runtime": 1275.2793, + "train_tokens_per_second": 6222.121 + }, + { + "epoch": 1.6735042735042736, + "grad_norm": 0.17077142000198364, + "learning_rate": 3.339070567986231e-05, + "loss": 0.3818, + "num_input_tokens_seen": 7951486, + "step": 490, + "train_runtime": 1277.6946, + "train_tokens_per_second": 6223.307 + }, + { + "epoch": 1.676923076923077, + "grad_norm": 0.1565760374069214, + "learning_rate": 3.3046471600688475e-05, + "loss": 0.4047, + "num_input_tokens_seen": 7970478, + "step": 491, + "train_runtime": 1280.4332, + "train_tokens_per_second": 6224.829 + }, + { + "epoch": 1.6803418803418804, + "grad_norm": 0.17805105447769165, + "learning_rate": 3.2702237521514636e-05, + "loss": 0.3757, + "num_input_tokens_seen": 7984814, + "step": 492, + "train_runtime": 1282.5395, + "train_tokens_per_second": 6225.784 + }, + { + "epoch": 1.6837606837606838, + "grad_norm": 0.19390779733657837, + "learning_rate": 3.23580034423408e-05, + "loss": 0.3895, + "num_input_tokens_seen": 7998286, + "step": 493, + "train_runtime": 1284.5377, + "train_tokens_per_second": 6226.587 + }, + { + "epoch": 1.6871794871794872, + "grad_norm": 0.16203835606575012, + "learning_rate": 3.201376936316696e-05, + "loss": 0.3933, + "num_input_tokens_seen": 8022334, + "step": 494, + "train_runtime": 1287.9546, + "train_tokens_per_second": 6228.74 + }, + { + "epoch": 1.6905982905982906, + "grad_norm": 0.164924755692482, + "learning_rate": 3.166953528399312e-05, + "loss": 0.367, + "num_input_tokens_seen": 8043294, + "step": 495, + "train_runtime": 1290.9515, + "train_tokens_per_second": 6230.516 + }, + { + "epoch": 1.694017094017094, + "grad_norm": 0.187575563788414, + "learning_rate": 3.132530120481928e-05, + "loss": 0.4334, + "num_input_tokens_seen": 8057310, + "step": 496, + "train_runtime": 1293.0552, + "train_tokens_per_second": 6231.219 + }, + { + "epoch": 1.6974358974358974, + "grad_norm": 0.1323373168706894, + "learning_rate": 3.098106712564544e-05, + "loss": 0.3116, + "num_input_tokens_seen": 8083774, + "step": 497, + "train_runtime": 1296.8416, + "train_tokens_per_second": 6233.432 + }, + { + "epoch": 1.7008547008547008, + "grad_norm": 0.20211701095104218, + "learning_rate": 3.0636833046471604e-05, + "loss": 0.4324, + "num_input_tokens_seen": 8097038, + "step": 498, + "train_runtime": 1298.8399, + "train_tokens_per_second": 6234.054 + }, + { + "epoch": 1.7042735042735044, + "grad_norm": 0.17863459885120392, + "learning_rate": 3.0292598967297765e-05, + "loss": 0.2992, + "num_input_tokens_seen": 8114750, + "step": 499, + "train_runtime": 1301.44, + "train_tokens_per_second": 6235.208 + }, + { + "epoch": 1.7076923076923078, + "grad_norm": 0.19749584794044495, + "learning_rate": 2.9948364888123926e-05, + "loss": 0.4242, + "num_input_tokens_seen": 8126750, + "step": 500, + "train_runtime": 1303.2477, + "train_tokens_per_second": 6235.768 + }, + { + "epoch": 1.7111111111111112, + "grad_norm": 0.171902135014534, + "learning_rate": 2.9604130808950087e-05, + "loss": 0.3748, + "num_input_tokens_seen": 8144766, + "step": 501, + "train_runtime": 1305.8965, + "train_tokens_per_second": 6236.915 + }, + { + "epoch": 1.7145299145299147, + "grad_norm": 0.15472909808158875, + "learning_rate": 2.925989672977625e-05, + "loss": 0.3657, + "num_input_tokens_seen": 8164254, + "step": 502, + "train_runtime": 1308.7049, + "train_tokens_per_second": 6238.423 + }, + { + "epoch": 1.717948717948718, + "grad_norm": 0.1825917810201645, + "learning_rate": 2.891566265060241e-05, + "loss": 0.3551, + "num_input_tokens_seen": 8178510, + "step": 503, + "train_runtime": 1310.798, + "train_tokens_per_second": 6239.337 + }, + { + "epoch": 1.7213675213675215, + "grad_norm": 0.1650628000497818, + "learning_rate": 2.857142857142857e-05, + "loss": 0.3871, + "num_input_tokens_seen": 8196702, + "step": 504, + "train_runtime": 1313.4498, + "train_tokens_per_second": 6240.59 + }, + { + "epoch": 1.7247863247863249, + "grad_norm": 0.17389492690563202, + "learning_rate": 2.8227194492254732e-05, + "loss": 0.3691, + "num_input_tokens_seen": 8216254, + "step": 505, + "train_runtime": 1316.2774, + "train_tokens_per_second": 6242.038 + }, + { + "epoch": 1.7282051282051283, + "grad_norm": 0.17346595227718353, + "learning_rate": 2.7882960413080893e-05, + "loss": 0.4057, + "num_input_tokens_seen": 8233758, + "step": 506, + "train_runtime": 1318.8695, + "train_tokens_per_second": 6243.042 + }, + { + "epoch": 1.7316239316239317, + "grad_norm": 0.15873438119888306, + "learning_rate": 2.7538726333907055e-05, + "loss": 0.332, + "num_input_tokens_seen": 8247550, + "step": 507, + "train_runtime": 1320.9274, + "train_tokens_per_second": 6243.757 + }, + { + "epoch": 1.735042735042735, + "grad_norm": 0.17913693189620972, + "learning_rate": 2.7194492254733223e-05, + "loss": 0.3435, + "num_input_tokens_seen": 8263054, + "step": 508, + "train_runtime": 1323.2116, + "train_tokens_per_second": 6244.696 + }, + { + "epoch": 1.7384615384615385, + "grad_norm": 0.15651476383209229, + "learning_rate": 2.6850258175559384e-05, + "loss": 0.3343, + "num_input_tokens_seen": 8278046, + "step": 509, + "train_runtime": 1325.4232, + "train_tokens_per_second": 6245.587 + }, + { + "epoch": 1.741880341880342, + "grad_norm": 0.17296011745929718, + "learning_rate": 2.6506024096385545e-05, + "loss": 0.4069, + "num_input_tokens_seen": 8295214, + "step": 510, + "train_runtime": 1327.9451, + "train_tokens_per_second": 6246.654 + }, + { + "epoch": 1.7452991452991453, + "grad_norm": 0.18028153479099274, + "learning_rate": 2.6161790017211706e-05, + "loss": 0.3585, + "num_input_tokens_seen": 8312942, + "step": 511, + "train_runtime": 1331.0641, + "train_tokens_per_second": 6245.336 + }, + { + "epoch": 1.7487179487179487, + "grad_norm": 0.18217986822128296, + "learning_rate": 2.5817555938037867e-05, + "loss": 0.3445, + "num_input_tokens_seen": 8329982, + "step": 512, + "train_runtime": 1333.576, + "train_tokens_per_second": 6246.35 + }, + { + "epoch": 1.7521367521367521, + "grad_norm": 0.17503562569618225, + "learning_rate": 2.547332185886403e-05, + "loss": 0.3677, + "num_input_tokens_seen": 8341998, + "step": 513, + "train_runtime": 1335.4087, + "train_tokens_per_second": 6246.775 + }, + { + "epoch": 1.7555555555555555, + "grad_norm": 0.12160934507846832, + "learning_rate": 2.512908777969019e-05, + "loss": 0.282, + "num_input_tokens_seen": 8371790, + "step": 514, + "train_runtime": 1339.6445, + "train_tokens_per_second": 6249.262 + }, + { + "epoch": 1.758974358974359, + "grad_norm": 0.1700914204120636, + "learning_rate": 2.478485370051635e-05, + "loss": 0.4006, + "num_input_tokens_seen": 8390766, + "step": 515, + "train_runtime": 1342.3857, + "train_tokens_per_second": 6250.637 + }, + { + "epoch": 1.7623931623931623, + "grad_norm": 0.18512286245822906, + "learning_rate": 2.4440619621342516e-05, + "loss": 0.4371, + "num_input_tokens_seen": 8401710, + "step": 516, + "train_runtime": 1344.1099, + "train_tokens_per_second": 6250.761 + }, + { + "epoch": 1.7658119658119658, + "grad_norm": 0.1791275441646576, + "learning_rate": 2.4096385542168677e-05, + "loss": 0.3738, + "num_input_tokens_seen": 8418446, + "step": 517, + "train_runtime": 1346.5498, + "train_tokens_per_second": 6251.864 + }, + { + "epoch": 1.7692307692307692, + "grad_norm": 0.18900056183338165, + "learning_rate": 2.3752151462994838e-05, + "loss": 0.4529, + "num_input_tokens_seen": 8430590, + "step": 518, + "train_runtime": 1348.3721, + "train_tokens_per_second": 6252.421 + }, + { + "epoch": 1.7726495726495726, + "grad_norm": 0.1776425540447235, + "learning_rate": 2.3407917383821e-05, + "loss": 0.3513, + "num_input_tokens_seen": 8450318, + "step": 519, + "train_runtime": 1351.2218, + "train_tokens_per_second": 6253.835 + }, + { + "epoch": 1.776068376068376, + "grad_norm": 0.18933919072151184, + "learning_rate": 2.306368330464716e-05, + "loss": 0.3775, + "num_input_tokens_seen": 8464190, + "step": 520, + "train_runtime": 1353.2852, + "train_tokens_per_second": 6254.55 + }, + { + "epoch": 1.7794871794871794, + "grad_norm": 0.1620868742465973, + "learning_rate": 2.2719449225473325e-05, + "loss": 0.4034, + "num_input_tokens_seen": 8483038, + "step": 521, + "train_runtime": 1356.0363, + "train_tokens_per_second": 6255.76 + }, + { + "epoch": 1.7829059829059828, + "grad_norm": 0.16931356489658356, + "learning_rate": 2.2375215146299486e-05, + "loss": 0.3733, + "num_input_tokens_seen": 8499278, + "step": 522, + "train_runtime": 1358.4322, + "train_tokens_per_second": 6256.682 + }, + { + "epoch": 1.7863247863247862, + "grad_norm": 0.1474018692970276, + "learning_rate": 2.2030981067125648e-05, + "loss": 0.3131, + "num_input_tokens_seen": 8518766, + "step": 523, + "train_runtime": 1361.2451, + "train_tokens_per_second": 6258.069 + }, + { + "epoch": 1.7897435897435896, + "grad_norm": 0.16736820340156555, + "learning_rate": 2.168674698795181e-05, + "loss": 0.3634, + "num_input_tokens_seen": 8534430, + "step": 524, + "train_runtime": 1363.5049, + "train_tokens_per_second": 6259.186 + }, + { + "epoch": 1.793162393162393, + "grad_norm": 0.17213431000709534, + "learning_rate": 2.134251290877797e-05, + "loss": 0.3632, + "num_input_tokens_seen": 8546862, + "step": 525, + "train_runtime": 1365.3454, + "train_tokens_per_second": 6259.853 + }, + { + "epoch": 1.7965811965811964, + "grad_norm": 0.19592760503292084, + "learning_rate": 2.099827882960413e-05, + "loss": 0.381, + "num_input_tokens_seen": 8561358, + "step": 526, + "train_runtime": 1367.5092, + "train_tokens_per_second": 6260.549 + }, + { + "epoch": 1.8, + "grad_norm": 0.1715427190065384, + "learning_rate": 2.0654044750430293e-05, + "loss": 0.4199, + "num_input_tokens_seen": 8581310, + "step": 527, + "train_runtime": 1370.3773, + "train_tokens_per_second": 6262.005 + }, + { + "epoch": 1.8034188034188035, + "grad_norm": 0.194061741232872, + "learning_rate": 2.0309810671256454e-05, + "loss": 0.3965, + "num_input_tokens_seen": 8592766, + "step": 528, + "train_runtime": 1372.1254, + "train_tokens_per_second": 6262.376 + }, + { + "epoch": 1.8068376068376069, + "grad_norm": 0.1930086761713028, + "learning_rate": 1.9965576592082615e-05, + "loss": 0.3976, + "num_input_tokens_seen": 8607166, + "step": 529, + "train_runtime": 1374.2507, + "train_tokens_per_second": 6263.17 + }, + { + "epoch": 1.8102564102564103, + "grad_norm": 0.16003040969371796, + "learning_rate": 1.962134251290878e-05, + "loss": 0.3367, + "num_input_tokens_seen": 8622478, + "step": 530, + "train_runtime": 1376.5089, + "train_tokens_per_second": 6264.019 + }, + { + "epoch": 1.8136752136752137, + "grad_norm": 0.19035999476909637, + "learning_rate": 1.927710843373494e-05, + "loss": 0.5509, + "num_input_tokens_seen": 8637918, + "step": 531, + "train_runtime": 1378.7679, + "train_tokens_per_second": 6264.954 + }, + { + "epoch": 1.8136752136752137, + "eval_loss": 0.4248267114162445, + "eval_runtime": 11.8853, + "eval_samples_per_second": 83.969, + "eval_steps_per_second": 5.301, + "num_input_tokens_seen": 8637918, + "step": 531 + }, + { + "epoch": 1.817094017094017, + "grad_norm": 0.17047119140625, + "learning_rate": 1.8932874354561102e-05, + "loss": 0.3973, + "num_input_tokens_seen": 8656606, + "step": 532, + "train_runtime": 1393.3816, + "train_tokens_per_second": 6212.66 + }, + { + "epoch": 1.8205128205128205, + "grad_norm": 0.1665705293416977, + "learning_rate": 1.8588640275387263e-05, + "loss": 0.4401, + "num_input_tokens_seen": 8675454, + "step": 533, + "train_runtime": 1396.1479, + "train_tokens_per_second": 6213.85 + }, + { + "epoch": 1.823931623931624, + "grad_norm": 0.18982985615730286, + "learning_rate": 1.8244406196213425e-05, + "loss": 0.4021, + "num_input_tokens_seen": 8690142, + "step": 534, + "train_runtime": 1398.3083, + "train_tokens_per_second": 6214.754 + }, + { + "epoch": 1.8273504273504273, + "grad_norm": 0.16096247732639313, + "learning_rate": 1.790017211703959e-05, + "loss": 0.4033, + "num_input_tokens_seen": 8706558, + "step": 535, + "train_runtime": 1400.6804, + "train_tokens_per_second": 6215.949 + }, + { + "epoch": 1.830769230769231, + "grad_norm": 0.1925681084394455, + "learning_rate": 1.755593803786575e-05, + "loss": 0.4394, + "num_input_tokens_seen": 8722270, + "step": 536, + "train_runtime": 1402.9974, + "train_tokens_per_second": 6216.882 + }, + { + "epoch": 1.8341880341880343, + "grad_norm": 0.18180759251117706, + "learning_rate": 1.721170395869191e-05, + "loss": 0.4643, + "num_input_tokens_seen": 8737982, + "step": 537, + "train_runtime": 1405.3075, + "train_tokens_per_second": 6217.844 + }, + { + "epoch": 1.8376068376068377, + "grad_norm": 0.1800447255373001, + "learning_rate": 1.6867469879518073e-05, + "loss": 0.6214, + "num_input_tokens_seen": 8753838, + "step": 538, + "train_runtime": 1407.6382, + "train_tokens_per_second": 6218.812 + }, + { + "epoch": 1.8410256410256411, + "grad_norm": 0.18196427822113037, + "learning_rate": 1.6523235800344237e-05, + "loss": 0.3515, + "num_input_tokens_seen": 8767294, + "step": 539, + "train_runtime": 1409.6646, + "train_tokens_per_second": 6219.418 + }, + { + "epoch": 1.8444444444444446, + "grad_norm": 0.1552770733833313, + "learning_rate": 1.61790017211704e-05, + "loss": 0.3492, + "num_input_tokens_seen": 8790974, + "step": 540, + "train_runtime": 1413.0808, + "train_tokens_per_second": 6221.14 + }, + { + "epoch": 1.847863247863248, + "grad_norm": 0.17243799567222595, + "learning_rate": 1.583476764199656e-05, + "loss": 0.4604, + "num_input_tokens_seen": 8802782, + "step": 541, + "train_runtime": 1415.3829, + "train_tokens_per_second": 6219.364 + }, + { + "epoch": 1.8512820512820514, + "grad_norm": 0.1819770187139511, + "learning_rate": 1.549053356282272e-05, + "loss": 0.4274, + "num_input_tokens_seen": 8818334, + "step": 542, + "train_runtime": 1417.6847, + "train_tokens_per_second": 6220.237 + }, + { + "epoch": 1.8547008547008548, + "grad_norm": 0.17073139548301697, + "learning_rate": 1.5146299483648882e-05, + "loss": 0.4203, + "num_input_tokens_seen": 8837166, + "step": 543, + "train_runtime": 1420.4111, + "train_tokens_per_second": 6221.555 + }, + { + "epoch": 1.8581196581196582, + "grad_norm": 0.1646145135164261, + "learning_rate": 1.4802065404475044e-05, + "loss": 0.3531, + "num_input_tokens_seen": 8854318, + "step": 544, + "train_runtime": 1422.8958, + "train_tokens_per_second": 6222.745 + }, + { + "epoch": 1.8615384615384616, + "grad_norm": 0.21891261637210846, + "learning_rate": 1.4457831325301205e-05, + "loss": 0.4048, + "num_input_tokens_seen": 8867102, + "step": 545, + "train_runtime": 1424.813, + "train_tokens_per_second": 6223.345 + }, + { + "epoch": 1.864957264957265, + "grad_norm": 0.2045115977525711, + "learning_rate": 1.4113597246127366e-05, + "loss": 0.5021, + "num_input_tokens_seen": 8880990, + "step": 546, + "train_runtime": 1426.8816, + "train_tokens_per_second": 6224.055 + }, + { + "epoch": 1.8683760683760684, + "grad_norm": 0.18179452419281006, + "learning_rate": 1.3769363166953527e-05, + "loss": 0.4367, + "num_input_tokens_seen": 8895422, + "step": 547, + "train_runtime": 1429.0398, + "train_tokens_per_second": 6224.754 + }, + { + "epoch": 1.8717948717948718, + "grad_norm": 0.1890801042318344, + "learning_rate": 1.3425129087779692e-05, + "loss": 0.3696, + "num_input_tokens_seen": 8908718, + "step": 548, + "train_runtime": 1431.0057, + "train_tokens_per_second": 6225.495 + }, + { + "epoch": 1.8752136752136752, + "grad_norm": 0.1882188469171524, + "learning_rate": 1.3080895008605853e-05, + "loss": 0.4208, + "num_input_tokens_seen": 8923086, + "step": 549, + "train_runtime": 1433.137, + "train_tokens_per_second": 6226.262 + }, + { + "epoch": 1.8786324786324786, + "grad_norm": 0.168639674782753, + "learning_rate": 1.2736660929432014e-05, + "loss": 0.3938, + "num_input_tokens_seen": 8941022, + "step": 550, + "train_runtime": 1435.7111, + "train_tokens_per_second": 6227.591 + }, + { + "epoch": 1.882051282051282, + "grad_norm": 0.21191436052322388, + "learning_rate": 1.2392426850258176e-05, + "loss": 0.3797, + "num_input_tokens_seen": 8951518, + "step": 551, + "train_runtime": 1437.3232, + "train_tokens_per_second": 6227.909 + }, + { + "epoch": 1.8854700854700854, + "grad_norm": 0.17271479964256287, + "learning_rate": 1.2048192771084338e-05, + "loss": 0.3924, + "num_input_tokens_seen": 8967406, + "step": 552, + "train_runtime": 1439.697, + "train_tokens_per_second": 6228.676 + }, + { + "epoch": 1.8888888888888888, + "grad_norm": 0.17815494537353516, + "learning_rate": 1.17039586919105e-05, + "loss": 0.3984, + "num_input_tokens_seen": 8980910, + "step": 553, + "train_runtime": 1441.6466, + "train_tokens_per_second": 6229.619 + }, + { + "epoch": 1.8923076923076922, + "grad_norm": 0.18840882182121277, + "learning_rate": 1.1359724612736663e-05, + "loss": 0.4698, + "num_input_tokens_seen": 8996702, + "step": 554, + "train_runtime": 1444.0033, + "train_tokens_per_second": 6230.389 + }, + { + "epoch": 1.8957264957264957, + "grad_norm": 0.17246229946613312, + "learning_rate": 1.1015490533562824e-05, + "loss": 0.3294, + "num_input_tokens_seen": 9014414, + "step": 555, + "train_runtime": 1446.5743, + "train_tokens_per_second": 6231.56 + }, + { + "epoch": 1.899145299145299, + "grad_norm": 0.20323026180267334, + "learning_rate": 1.0671256454388985e-05, + "loss": 0.4267, + "num_input_tokens_seen": 9031550, + "step": 556, + "train_runtime": 1449.0794, + "train_tokens_per_second": 6232.612 + }, + { + "epoch": 1.9025641025641025, + "grad_norm": 0.17592619359493256, + "learning_rate": 1.0327022375215146e-05, + "loss": 0.3815, + "num_input_tokens_seen": 9048238, + "step": 557, + "train_runtime": 1451.4988, + "train_tokens_per_second": 6233.721 + }, + { + "epoch": 1.9059829059829059, + "grad_norm": 0.16363990306854248, + "learning_rate": 9.982788296041308e-06, + "loss": 0.3811, + "num_input_tokens_seen": 9066494, + "step": 558, + "train_runtime": 1454.1349, + "train_tokens_per_second": 6234.975 + }, + { + "epoch": 1.9094017094017093, + "grad_norm": 0.1940564066171646, + "learning_rate": 9.63855421686747e-06, + "loss": 0.3237, + "num_input_tokens_seen": 9079822, + "step": 559, + "train_runtime": 1456.1609, + "train_tokens_per_second": 6235.453 + }, + { + "epoch": 1.9128205128205127, + "grad_norm": 0.2036203294992447, + "learning_rate": 9.294320137693632e-06, + "loss": 0.4353, + "num_input_tokens_seen": 9091598, + "step": 560, + "train_runtime": 1457.9853, + "train_tokens_per_second": 6235.727 + }, + { + "epoch": 1.916239316239316, + "grad_norm": 0.18901148438453674, + "learning_rate": 8.950086058519795e-06, + "loss": 0.4001, + "num_input_tokens_seen": 9104990, + "step": 561, + "train_runtime": 1460.0107, + "train_tokens_per_second": 6236.249 + }, + { + "epoch": 1.9196581196581195, + "grad_norm": 0.18928927183151245, + "learning_rate": 8.605851979345956e-06, + "loss": 0.3813, + "num_input_tokens_seen": 9119310, + "step": 562, + "train_runtime": 1462.1634, + "train_tokens_per_second": 6236.861 + }, + { + "epoch": 1.9230769230769231, + "grad_norm": 0.1698116660118103, + "learning_rate": 8.261617900172119e-06, + "loss": 0.3993, + "num_input_tokens_seen": 9134702, + "step": 563, + "train_runtime": 1464.4317, + "train_tokens_per_second": 6237.711 + }, + { + "epoch": 1.9264957264957265, + "grad_norm": 0.1814728081226349, + "learning_rate": 7.91738382099828e-06, + "loss": 0.4208, + "num_input_tokens_seen": 9147822, + "step": 564, + "train_runtime": 1466.388, + "train_tokens_per_second": 6238.336 + }, + { + "epoch": 1.92991452991453, + "grad_norm": 0.18567024171352386, + "learning_rate": 7.573149741824441e-06, + "loss": 0.4371, + "num_input_tokens_seen": 9164110, + "step": 565, + "train_runtime": 1468.7747, + "train_tokens_per_second": 6239.289 + }, + { + "epoch": 1.9333333333333333, + "grad_norm": 0.2137572318315506, + "learning_rate": 7.228915662650602e-06, + "loss": 0.3899, + "num_input_tokens_seen": 9174894, + "step": 566, + "train_runtime": 1470.4523, + "train_tokens_per_second": 6239.505 + }, + { + "epoch": 1.9367521367521368, + "grad_norm": 0.15320739150047302, + "learning_rate": 6.884681583476764e-06, + "loss": 0.3531, + "num_input_tokens_seen": 9191886, + "step": 567, + "train_runtime": 1472.9549, + "train_tokens_per_second": 6240.44 + }, + { + "epoch": 1.9401709401709402, + "grad_norm": 0.21417830884456635, + "learning_rate": 6.5404475043029266e-06, + "loss": 0.4213, + "num_input_tokens_seen": 9203742, + "step": 568, + "train_runtime": 1474.745, + "train_tokens_per_second": 6240.904 + }, + { + "epoch": 1.9435897435897436, + "grad_norm": 0.1843339502811432, + "learning_rate": 6.196213425129088e-06, + "loss": 0.374, + "num_input_tokens_seen": 9216110, + "step": 569, + "train_runtime": 1476.6219, + "train_tokens_per_second": 6241.347 + }, + { + "epoch": 1.947008547008547, + "grad_norm": 0.16858874261379242, + "learning_rate": 5.85197934595525e-06, + "loss": 0.3967, + "num_input_tokens_seen": 9232750, + "step": 570, + "train_runtime": 1479.0428, + "train_tokens_per_second": 6242.382 + }, + { + "epoch": 1.9504273504273504, + "grad_norm": 0.21565116941928864, + "learning_rate": 5.507745266781412e-06, + "loss": 0.4639, + "num_input_tokens_seen": 9244110, + "step": 571, + "train_runtime": 1481.2963, + "train_tokens_per_second": 6240.554 + }, + { + "epoch": 1.953846153846154, + "grad_norm": 0.17212358117103577, + "learning_rate": 5.163511187607573e-06, + "loss": 0.4541, + "num_input_tokens_seen": 9260670, + "step": 572, + "train_runtime": 1483.7253, + "train_tokens_per_second": 6241.499 + }, + { + "epoch": 1.9572649572649574, + "grad_norm": 0.16098545491695404, + "learning_rate": 4.819277108433735e-06, + "loss": 0.3808, + "num_input_tokens_seen": 9277182, + "step": 573, + "train_runtime": 1486.1266, + "train_tokens_per_second": 6242.525 + }, + { + "epoch": 1.9606837606837608, + "grad_norm": 0.1885039359331131, + "learning_rate": 4.475043029259897e-06, + "loss": 0.4711, + "num_input_tokens_seen": 9291486, + "step": 574, + "train_runtime": 1488.2625, + "train_tokens_per_second": 6243.177 + }, + { + "epoch": 1.9641025641025642, + "grad_norm": 0.17435158789157867, + "learning_rate": 4.130808950086059e-06, + "loss": 0.3834, + "num_input_tokens_seen": 9307182, + "step": 575, + "train_runtime": 1490.6012, + "train_tokens_per_second": 6243.911 + }, + { + "epoch": 1.9675213675213676, + "grad_norm": 0.17134054005146027, + "learning_rate": 3.7865748709122206e-06, + "loss": 0.3521, + "num_input_tokens_seen": 9324318, + "step": 576, + "train_runtime": 1493.1064, + "train_tokens_per_second": 6244.912 + }, + { + "epoch": 1.970940170940171, + "grad_norm": 0.18249361217021942, + "learning_rate": 3.442340791738382e-06, + "loss": 0.3362, + "num_input_tokens_seen": 9338622, + "step": 577, + "train_runtime": 1495.2539, + "train_tokens_per_second": 6245.509 + }, + { + "epoch": 1.9743589743589745, + "grad_norm": 0.17978602647781372, + "learning_rate": 3.098106712564544e-06, + "loss": 0.4274, + "num_input_tokens_seen": 9353918, + "step": 578, + "train_runtime": 1497.5034, + "train_tokens_per_second": 6246.342 + }, + { + "epoch": 1.9777777777777779, + "grad_norm": 0.14672812819480896, + "learning_rate": 2.753872633390706e-06, + "loss": 0.314, + "num_input_tokens_seen": 9390878, + "step": 579, + "train_runtime": 1502.7758, + "train_tokens_per_second": 6249.021 + }, + { + "epoch": 1.9811965811965813, + "grad_norm": 0.19798055291175842, + "learning_rate": 2.4096385542168676e-06, + "loss": 0.4314, + "num_input_tokens_seen": 9404654, + "step": 580, + "train_runtime": 1504.8242, + "train_tokens_per_second": 6249.67 + }, + { + "epoch": 1.9846153846153847, + "grad_norm": 0.1839577704668045, + "learning_rate": 2.0654044750430297e-06, + "loss": 0.346, + "num_input_tokens_seen": 9421838, + "step": 581, + "train_runtime": 1507.3329, + "train_tokens_per_second": 6250.668 + }, + { + "epoch": 1.988034188034188, + "grad_norm": 0.1459440439939499, + "learning_rate": 1.721170395869191e-06, + "loss": 0.302, + "num_input_tokens_seen": 9450766, + "step": 582, + "train_runtime": 1511.4289, + "train_tokens_per_second": 6252.868 + }, + { + "epoch": 1.9914529914529915, + "grad_norm": 0.1679002195596695, + "learning_rate": 1.376936316695353e-06, + "loss": 0.3794, + "num_input_tokens_seen": 9467966, + "step": 583, + "train_runtime": 1513.9495, + "train_tokens_per_second": 6253.819 + }, + { + "epoch": 1.994871794871795, + "grad_norm": 0.18191100656986237, + "learning_rate": 1.0327022375215148e-06, + "loss": 0.4131, + "num_input_tokens_seen": 9482622, + "step": 584, + "train_runtime": 1516.0908, + "train_tokens_per_second": 6254.653 + }, + { + "epoch": 1.9982905982905983, + "grad_norm": 0.17011763155460358, + "learning_rate": 6.884681583476765e-07, + "loss": 0.3366, + "num_input_tokens_seen": 9497966, + "step": 585, + "train_runtime": 1518.3683, + "train_tokens_per_second": 6255.377 + }, + { + "epoch": 2.0, + "grad_norm": 0.23105928301811218, + "learning_rate": 3.4423407917383825e-07, + "loss": 0.3455, + "num_input_tokens_seen": 9505226, + "step": 586, + "train_runtime": 1519.4393, + "train_tokens_per_second": 6255.746 + } + ], + "logging_steps": 1, + "max_steps": 586, + "num_input_tokens_seen": 9505226, + "num_train_epochs": 2, + "save_steps": 30, + "stateful_callbacks": { + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": true + }, + "attributes": {} + } + }, + "total_flos": 4.055493629648671e+17, + "train_batch_size": 16, + "trial_name": null, + "trial_params": null +}