{ "best_global_step": 10500, "best_metric": 2.502063751220703, "best_model_checkpoint": "../runs/karpathy/fineweb-edu-100b-shuffle/HuggingFaceTB/SmolLM2-360M/linear_adamw_wd1e-03_8x1024_mem32_bs64_hf_armt_dmem64/run_61/checkpoint-10500", "epoch": 0.22, "eval_steps": 100, "global_step": 11000, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 0.0005, "grad_norm": 7.414668730380984, "learning_rate": 4.8e-08, "loss": 2.58, "step": 25 }, { "epoch": 0.001, "grad_norm": 9.160375364793794, "learning_rate": 9.8e-08, "loss": 2.5669, "step": 50 }, { "epoch": 0.0015, "grad_norm": 9.022014619463027, "learning_rate": 1.4800000000000003e-07, "loss": 2.5685, "step": 75 }, { "epoch": 0.002, "grad_norm": 8.08180658385987, "learning_rate": 1.9800000000000003e-07, "loss": 2.574, "step": 100 }, { "epoch": 0.002, "eval_loss": 2.5819575786590576, "eval_runtime": 31.3642, "eval_samples_per_second": 3.38, "eval_steps_per_second": 0.861, "step": 100 }, { "epoch": 0.0025, "grad_norm": 8.148585386546005, "learning_rate": 2.48e-07, "loss": 2.559, "step": 125 }, { "epoch": 0.003, "grad_norm": 8.48517035589218, "learning_rate": 2.9800000000000005e-07, "loss": 2.5705, "step": 150 }, { "epoch": 0.0035, "grad_norm": 7.4647297827229675, "learning_rate": 3.48e-07, "loss": 2.5569, "step": 175 }, { "epoch": 0.004, "grad_norm": 7.5892309502710456, "learning_rate": 3.9800000000000004e-07, "loss": 2.5654, "step": 200 }, { "epoch": 0.004, "eval_loss": 2.5757665634155273, "eval_runtime": 31.4592, "eval_samples_per_second": 3.369, "eval_steps_per_second": 0.858, "step": 200 }, { "epoch": 0.0045, "grad_norm": 8.513594577252011, "learning_rate": 4.4800000000000004e-07, "loss": 2.5702, "step": 225 }, { "epoch": 0.005, "grad_norm": 7.36618260817592, "learning_rate": 4.98e-07, "loss": 2.5616, "step": 250 }, { "epoch": 0.0055, "grad_norm": 7.979652660110063, "learning_rate": 5.480000000000001e-07, "loss": 2.5724, "step": 275 }, { "epoch": 0.006, "grad_norm": 8.286710116248681, "learning_rate": 5.98e-07, "loss": 2.5645, "step": 300 }, { "epoch": 0.006, "eval_loss": 2.573997735977173, "eval_runtime": 31.4489, "eval_samples_per_second": 3.371, "eval_steps_per_second": 0.859, "step": 300 }, { "epoch": 0.0065, "grad_norm": 7.793115280849132, "learning_rate": 6.48e-07, "loss": 2.5561, "step": 325 }, { "epoch": 0.007, "grad_norm": 7.4565249755148395, "learning_rate": 6.98e-07, "loss": 2.5559, "step": 350 }, { "epoch": 0.0075, "grad_norm": 7.162170017009099, "learning_rate": 7.480000000000001e-07, "loss": 2.5606, "step": 375 }, { "epoch": 0.008, "grad_norm": 7.634262924490496, "learning_rate": 7.98e-07, "loss": 2.5584, "step": 400 }, { "epoch": 0.008, "eval_loss": 2.571639060974121, "eval_runtime": 31.4335, "eval_samples_per_second": 3.372, "eval_steps_per_second": 0.859, "step": 400 }, { "epoch": 0.0085, "grad_norm": 6.741882313360166, "learning_rate": 8.480000000000001e-07, "loss": 2.5489, "step": 425 }, { "epoch": 0.009, "grad_norm": 7.0353863053832875, "learning_rate": 8.980000000000001e-07, "loss": 2.5607, "step": 450 }, { "epoch": 0.0095, "grad_norm": 6.545029332470873, "learning_rate": 9.480000000000001e-07, "loss": 2.5606, "step": 475 }, { "epoch": 0.01, "grad_norm": 5.856492237540935, "learning_rate": 9.98e-07, "loss": 2.5558, "step": 500 }, { "epoch": 0.01, "eval_loss": 2.568985939025879, "eval_runtime": 31.4693, "eval_samples_per_second": 3.368, "eval_steps_per_second": 0.858, "step": 500 }, { "epoch": 0.0105, "grad_norm": 6.010839203358671, "learning_rate": 1.0480000000000002e-06, "loss": 2.5606, "step": 525 }, { "epoch": 0.011, "grad_norm": 5.622365258702099, "learning_rate": 1.0980000000000001e-06, "loss": 2.5663, "step": 550 }, { "epoch": 0.0115, "grad_norm": 4.998479332021628, "learning_rate": 1.148e-06, "loss": 2.5498, "step": 575 }, { "epoch": 0.012, "grad_norm": 4.938142978417324, "learning_rate": 1.1980000000000002e-06, "loss": 2.5522, "step": 600 }, { "epoch": 0.012, "eval_loss": 2.5657429695129395, "eval_runtime": 31.567, "eval_samples_per_second": 3.358, "eval_steps_per_second": 0.855, "step": 600 }, { "epoch": 0.0125, "grad_norm": 4.643553586272799, "learning_rate": 1.248e-06, "loss": 2.562, "step": 625 }, { "epoch": 0.013, "grad_norm": 4.370764832622782, "learning_rate": 1.2980000000000001e-06, "loss": 2.5524, "step": 650 }, { "epoch": 0.0135, "grad_norm": 4.4782950314103624, "learning_rate": 1.348e-06, "loss": 2.5461, "step": 675 }, { "epoch": 0.014, "grad_norm": 3.88097541690361, "learning_rate": 1.3980000000000002e-06, "loss": 2.5589, "step": 700 }, { "epoch": 0.014, "eval_loss": 2.5625, "eval_runtime": 31.6097, "eval_samples_per_second": 3.353, "eval_steps_per_second": 0.854, "step": 700 }, { "epoch": 0.0145, "grad_norm": 4.6364438893944415, "learning_rate": 1.4480000000000002e-06, "loss": 2.544, "step": 725 }, { "epoch": 0.015, "grad_norm": 4.087341405976335, "learning_rate": 1.498e-06, "loss": 2.544, "step": 750 }, { "epoch": 0.0155, "grad_norm": 5.895841058437022, "learning_rate": 1.548e-06, "loss": 2.5425, "step": 775 }, { "epoch": 0.016, "grad_norm": 4.993061404246151, "learning_rate": 1.5980000000000002e-06, "loss": 2.5487, "step": 800 }, { "epoch": 0.016, "eval_loss": 2.561025857925415, "eval_runtime": 31.518, "eval_samples_per_second": 3.363, "eval_steps_per_second": 0.857, "step": 800 }, { "epoch": 0.0165, "grad_norm": 8.716353678903719, "learning_rate": 1.6480000000000001e-06, "loss": 2.5416, "step": 825 }, { "epoch": 0.017, "grad_norm": 5.529513150034155, "learning_rate": 1.6980000000000003e-06, "loss": 2.5524, "step": 850 }, { "epoch": 0.0175, "grad_norm": 3.5181432936318693, "learning_rate": 1.7480000000000002e-06, "loss": 2.5316, "step": 875 }, { "epoch": 0.018, "grad_norm": 3.6272596212263326, "learning_rate": 1.798e-06, "loss": 2.5482, "step": 900 }, { "epoch": 0.018, "eval_loss": 2.561025857925415, "eval_runtime": 31.5635, "eval_samples_per_second": 3.358, "eval_steps_per_second": 0.855, "step": 900 }, { "epoch": 0.0185, "grad_norm": 3.4634269778278206, "learning_rate": 1.8480000000000001e-06, "loss": 2.5437, "step": 925 }, { "epoch": 0.019, "grad_norm": 3.2744959130191753, "learning_rate": 1.898e-06, "loss": 2.5373, "step": 950 }, { "epoch": 0.0195, "grad_norm": 4.866747798407841, "learning_rate": 1.9480000000000002e-06, "loss": 2.5481, "step": 975 }, { "epoch": 0.02, "grad_norm": 6.493920664330759, "learning_rate": 1.998e-06, "loss": 2.5572, "step": 1000 }, { "epoch": 0.02, "eval_loss": 2.558077812194824, "eval_runtime": 32.5314, "eval_samples_per_second": 3.258, "eval_steps_per_second": 0.83, "step": 1000 }, { "epoch": 0.0205, "grad_norm": 3.4376334682400023, "learning_rate": 2.048e-06, "loss": 2.5467, "step": 1025 }, { "epoch": 0.021, "grad_norm": 5.716848161977786, "learning_rate": 2.098e-06, "loss": 2.5453, "step": 1050 }, { "epoch": 0.0215, "grad_norm": 2.427114399660508, "learning_rate": 2.148e-06, "loss": 2.5587, "step": 1075 }, { "epoch": 0.022, "grad_norm": 2.841648670814877, "learning_rate": 2.198e-06, "loss": 2.5391, "step": 1100 }, { "epoch": 0.022, "eval_loss": 2.5557193756103516, "eval_runtime": 31.5806, "eval_samples_per_second": 3.356, "eval_steps_per_second": 0.855, "step": 1100 }, { "epoch": 0.0225, "grad_norm": 2.7671559538325217, "learning_rate": 2.2480000000000003e-06, "loss": 2.5473, "step": 1125 }, { "epoch": 0.023, "grad_norm": 2.762586633385912, "learning_rate": 2.2980000000000003e-06, "loss": 2.5266, "step": 1150 }, { "epoch": 0.0235, "grad_norm": 1.8996614708410953, "learning_rate": 2.3480000000000002e-06, "loss": 2.5383, "step": 1175 }, { "epoch": 0.024, "grad_norm": 2.757053334920087, "learning_rate": 2.398e-06, "loss": 2.5446, "step": 1200 }, { "epoch": 0.024, "eval_loss": 2.5548348426818848, "eval_runtime": 33.6988, "eval_samples_per_second": 3.146, "eval_steps_per_second": 0.801, "step": 1200 }, { "epoch": 0.0245, "grad_norm": 4.33045144646476, "learning_rate": 2.448e-06, "loss": 2.5483, "step": 1225 }, { "epoch": 0.025, "grad_norm": 3.138318361162452, "learning_rate": 2.498e-06, "loss": 2.5422, "step": 1250 }, { "epoch": 0.0255, "grad_norm": 2.2101804256377315, "learning_rate": 2.5480000000000004e-06, "loss": 2.5459, "step": 1275 }, { "epoch": 0.026, "grad_norm": 6.96587026664173, "learning_rate": 2.598e-06, "loss": 2.5482, "step": 1300 }, { "epoch": 0.026, "eval_loss": 2.5545401573181152, "eval_runtime": 33.7014, "eval_samples_per_second": 3.145, "eval_steps_per_second": 0.801, "step": 1300 }, { "epoch": 0.0265, "grad_norm": 3.7254283111121023, "learning_rate": 2.648e-06, "loss": 2.5358, "step": 1325 }, { "epoch": 0.027, "grad_norm": 5.786847124865984, "learning_rate": 2.6980000000000003e-06, "loss": 2.5522, "step": 1350 }, { "epoch": 0.0275, "grad_norm": 3.443648998727752, "learning_rate": 2.748e-06, "loss": 2.5485, "step": 1375 }, { "epoch": 0.028, "grad_norm": 2.8078549307686744, "learning_rate": 2.798e-06, "loss": 2.5449, "step": 1400 }, { "epoch": 0.028, "eval_loss": 2.5536556243896484, "eval_runtime": 33.6843, "eval_samples_per_second": 3.147, "eval_steps_per_second": 0.802, "step": 1400 }, { "epoch": 0.0285, "grad_norm": 2.2231663508054695, "learning_rate": 2.848e-06, "loss": 2.5429, "step": 1425 }, { "epoch": 0.029, "grad_norm": 2.017011692108824, "learning_rate": 2.8980000000000005e-06, "loss": 2.5358, "step": 1450 }, { "epoch": 0.0295, "grad_norm": 3.473206409631469, "learning_rate": 2.9480000000000004e-06, "loss": 2.5446, "step": 1475 }, { "epoch": 0.03, "grad_norm": 3.9724966312793453, "learning_rate": 2.9980000000000003e-06, "loss": 2.5271, "step": 1500 }, { "epoch": 0.03, "eval_loss": 2.5515918731689453, "eval_runtime": 33.4333, "eval_samples_per_second": 3.17, "eval_steps_per_second": 0.808, "step": 1500 }, { "epoch": 0.0305, "grad_norm": 3.794429177775492, "learning_rate": 3.0480000000000003e-06, "loss": 2.5373, "step": 1525 }, { "epoch": 0.031, "grad_norm": 3.541018772674368, "learning_rate": 3.0980000000000007e-06, "loss": 2.5428, "step": 1550 }, { "epoch": 0.0315, "grad_norm": 1.8297000018140448, "learning_rate": 3.1480000000000006e-06, "loss": 2.5461, "step": 1575 }, { "epoch": 0.032, "grad_norm": 1.8847742274784853, "learning_rate": 3.198e-06, "loss": 2.5356, "step": 1600 }, { "epoch": 0.032, "eval_loss": 2.551297187805176, "eval_runtime": 31.7073, "eval_samples_per_second": 3.343, "eval_steps_per_second": 0.852, "step": 1600 }, { "epoch": 0.0325, "grad_norm": 2.0428744582128626, "learning_rate": 3.248e-06, "loss": 2.539, "step": 1625 }, { "epoch": 0.033, "grad_norm": 3.923062191693087, "learning_rate": 3.298e-06, "loss": 2.5451, "step": 1650 }, { "epoch": 0.0335, "grad_norm": 2.5701642609905475, "learning_rate": 3.348e-06, "loss": 2.5312, "step": 1675 }, { "epoch": 0.034, "grad_norm": 1.372625402655975, "learning_rate": 3.3980000000000003e-06, "loss": 2.5293, "step": 1700 }, { "epoch": 0.034, "eval_loss": 2.5492334365844727, "eval_runtime": 31.7196, "eval_samples_per_second": 3.342, "eval_steps_per_second": 0.851, "step": 1700 }, { "epoch": 0.0345, "grad_norm": 2.9992155036204764, "learning_rate": 3.4480000000000003e-06, "loss": 2.5471, "step": 1725 }, { "epoch": 0.035, "grad_norm": 7.065660583217205, "learning_rate": 3.4980000000000002e-06, "loss": 2.5254, "step": 1750 }, { "epoch": 0.0355, "grad_norm": 7.132721212830254, "learning_rate": 3.548e-06, "loss": 2.5413, "step": 1775 }, { "epoch": 0.036, "grad_norm": 2.530655358425179, "learning_rate": 3.5980000000000005e-06, "loss": 2.5387, "step": 1800 }, { "epoch": 0.036, "eval_loss": 2.547759532928467, "eval_runtime": 31.5128, "eval_samples_per_second": 3.364, "eval_steps_per_second": 0.857, "step": 1800 }, { "epoch": 0.0365, "grad_norm": 3.2608379101064826, "learning_rate": 3.6480000000000005e-06, "loss": 2.5403, "step": 1825 }, { "epoch": 0.037, "grad_norm": 57.229529386597854, "learning_rate": 3.6980000000000004e-06, "loss": 2.5343, "step": 1850 }, { "epoch": 0.0375, "grad_norm": 13.211892844738115, "learning_rate": 3.7480000000000004e-06, "loss": 2.5293, "step": 1875 }, { "epoch": 0.038, "grad_norm": 21.43989601963285, "learning_rate": 3.7980000000000007e-06, "loss": 2.5426, "step": 1900 }, { "epoch": 0.038, "eval_loss": 2.5456957817077637, "eval_runtime": 31.6472, "eval_samples_per_second": 3.349, "eval_steps_per_second": 0.853, "step": 1900 }, { "epoch": 0.0385, "grad_norm": 9.895019592079299, "learning_rate": 3.848e-06, "loss": 2.5382, "step": 1925 }, { "epoch": 0.039, "grad_norm": 10.302555220341693, "learning_rate": 3.898e-06, "loss": 2.5218, "step": 1950 }, { "epoch": 0.0395, "grad_norm": 13.079667784859808, "learning_rate": 3.948e-06, "loss": 2.5297, "step": 1975 }, { "epoch": 0.04, "grad_norm": 3.625926882991876, "learning_rate": 3.9980000000000005e-06, "loss": 2.5442, "step": 2000 }, { "epoch": 0.04, "eval_loss": 2.544811248779297, "eval_runtime": 31.7739, "eval_samples_per_second": 3.336, "eval_steps_per_second": 0.85, "step": 2000 }, { "epoch": 0.0405, "grad_norm": 5.046118481185164, "learning_rate": 4.048e-06, "loss": 2.5309, "step": 2025 }, { "epoch": 0.041, "grad_norm": 3.7507696805849324, "learning_rate": 4.098e-06, "loss": 2.5367, "step": 2050 }, { "epoch": 0.0415, "grad_norm": 2.16344700764137, "learning_rate": 4.148000000000001e-06, "loss": 2.5284, "step": 2075 }, { "epoch": 0.042, "grad_norm": 5.977563588437311, "learning_rate": 4.198e-06, "loss": 2.5364, "step": 2100 }, { "epoch": 0.042, "eval_loss": 2.543337345123291, "eval_runtime": 31.6423, "eval_samples_per_second": 3.35, "eval_steps_per_second": 0.853, "step": 2100 }, { "epoch": 0.0425, "grad_norm": 3.2783494537813276, "learning_rate": 4.248000000000001e-06, "loss": 2.5404, "step": 2125 }, { "epoch": 0.043, "grad_norm": 7.304771067898564, "learning_rate": 4.298e-06, "loss": 2.526, "step": 2150 }, { "epoch": 0.0435, "grad_norm": 1.470780395904616, "learning_rate": 4.3480000000000006e-06, "loss": 2.5279, "step": 2175 }, { "epoch": 0.044, "grad_norm": 4.0481167806519, "learning_rate": 4.398000000000001e-06, "loss": 2.5243, "step": 2200 }, { "epoch": 0.044, "eval_loss": 2.541273593902588, "eval_runtime": 31.6256, "eval_samples_per_second": 3.352, "eval_steps_per_second": 0.854, "step": 2200 }, { "epoch": 0.0445, "grad_norm": 3.078344902427094, "learning_rate": 4.4480000000000004e-06, "loss": 2.5298, "step": 2225 }, { "epoch": 0.045, "grad_norm": 2.424401506170523, "learning_rate": 4.498e-06, "loss": 2.5334, "step": 2250 }, { "epoch": 0.0455, "grad_norm": 6.696232827922732, "learning_rate": 4.548e-06, "loss": 2.5342, "step": 2275 }, { "epoch": 0.046, "grad_norm": 1.6221976812615748, "learning_rate": 4.598e-06, "loss": 2.5254, "step": 2300 }, { "epoch": 0.046, "eval_loss": 2.5389151573181152, "eval_runtime": 31.6396, "eval_samples_per_second": 3.35, "eval_steps_per_second": 0.853, "step": 2300 }, { "epoch": 0.0465, "grad_norm": 4.59240027677146, "learning_rate": 4.648e-06, "loss": 2.5264, "step": 2325 }, { "epoch": 0.047, "grad_norm": 3.1036878087546205, "learning_rate": 4.698000000000001e-06, "loss": 2.5356, "step": 2350 }, { "epoch": 0.0475, "grad_norm": 2.698795487781583, "learning_rate": 4.748e-06, "loss": 2.5251, "step": 2375 }, { "epoch": 0.048, "grad_norm": 4.022902376629088, "learning_rate": 4.7980000000000005e-06, "loss": 2.5211, "step": 2400 }, { "epoch": 0.048, "eval_loss": 2.5371463298797607, "eval_runtime": 31.6111, "eval_samples_per_second": 3.353, "eval_steps_per_second": 0.854, "step": 2400 }, { "epoch": 0.0485, "grad_norm": 1.9230869645311859, "learning_rate": 4.848000000000001e-06, "loss": 2.5138, "step": 2425 }, { "epoch": 0.049, "grad_norm": 4.234899191448228, "learning_rate": 4.898e-06, "loss": 2.5342, "step": 2450 }, { "epoch": 0.0495, "grad_norm": 4.877820273874513, "learning_rate": 4.948000000000001e-06, "loss": 2.5306, "step": 2475 }, { "epoch": 0.05, "grad_norm": 1.2038033583067678, "learning_rate": 4.998e-06, "loss": 2.5299, "step": 2500 }, { "epoch": 0.05, "eval_loss": 2.535377264022827, "eval_runtime": 31.7053, "eval_samples_per_second": 3.343, "eval_steps_per_second": 0.852, "step": 2500 }, { "epoch": 0.0505, "grad_norm": 1.457847344913734, "learning_rate": 5.048000000000001e-06, "loss": 2.5284, "step": 2525 }, { "epoch": 0.051, "grad_norm": 3.3976354659650205, "learning_rate": 5.098000000000001e-06, "loss": 2.5255, "step": 2550 }, { "epoch": 0.0515, "grad_norm": 1.3795475908793975, "learning_rate": 5.1480000000000005e-06, "loss": 2.5298, "step": 2575 }, { "epoch": 0.052, "grad_norm": 2.503585009963638, "learning_rate": 5.198000000000001e-06, "loss": 2.5329, "step": 2600 }, { "epoch": 0.052, "eval_loss": 2.533313751220703, "eval_runtime": 31.7208, "eval_samples_per_second": 3.342, "eval_steps_per_second": 0.851, "step": 2600 }, { "epoch": 0.0525, "grad_norm": 1.5928697147211628, "learning_rate": 5.248000000000001e-06, "loss": 2.5207, "step": 2625 }, { "epoch": 0.053, "grad_norm": 4.142079930582361, "learning_rate": 5.298000000000001e-06, "loss": 2.5276, "step": 2650 }, { "epoch": 0.0535, "grad_norm": 1.6506514288651493, "learning_rate": 5.348000000000001e-06, "loss": 2.5351, "step": 2675 }, { "epoch": 0.054, "grad_norm": 0.9935806352551346, "learning_rate": 5.398e-06, "loss": 2.5136, "step": 2700 }, { "epoch": 0.054, "eval_loss": 2.5315449237823486, "eval_runtime": 31.6455, "eval_samples_per_second": 3.35, "eval_steps_per_second": 0.853, "step": 2700 }, { "epoch": 0.0545, "grad_norm": 1.0414498471656755, "learning_rate": 5.448e-06, "loss": 2.5174, "step": 2725 }, { "epoch": 0.055, "grad_norm": 1.6440932231809144, "learning_rate": 5.498e-06, "loss": 2.5241, "step": 2750 }, { "epoch": 0.0555, "grad_norm": 1.2671657905301041, "learning_rate": 5.548e-06, "loss": 2.5214, "step": 2775 }, { "epoch": 0.056, "grad_norm": 2.193388679999868, "learning_rate": 5.5980000000000004e-06, "loss": 2.5177, "step": 2800 }, { "epoch": 0.056, "eval_loss": 2.530365467071533, "eval_runtime": 31.6494, "eval_samples_per_second": 3.349, "eval_steps_per_second": 0.853, "step": 2800 }, { "epoch": 0.0565, "grad_norm": 1.1100755835403464, "learning_rate": 5.648e-06, "loss": 2.513, "step": 2825 }, { "epoch": 0.057, "grad_norm": 1.012687003610205, "learning_rate": 5.698e-06, "loss": 2.5171, "step": 2850 }, { "epoch": 0.0575, "grad_norm": 3.590123409049871, "learning_rate": 5.748e-06, "loss": 2.5185, "step": 2875 }, { "epoch": 0.058, "grad_norm": 0.9965504720980352, "learning_rate": 5.798e-06, "loss": 2.5261, "step": 2900 }, { "epoch": 0.058, "eval_loss": 2.529186248779297, "eval_runtime": 31.7426, "eval_samples_per_second": 3.339, "eval_steps_per_second": 0.851, "step": 2900 }, { "epoch": 0.0585, "grad_norm": 10.858302444645757, "learning_rate": 5.848000000000001e-06, "loss": 2.5175, "step": 2925 }, { "epoch": 0.059, "grad_norm": 1.4839413654149125, "learning_rate": 5.898e-06, "loss": 2.5272, "step": 2950 }, { "epoch": 0.0595, "grad_norm": 4.153709792295477, "learning_rate": 5.9480000000000005e-06, "loss": 2.5095, "step": 2975 }, { "epoch": 0.06, "grad_norm": 0.9099122371246413, "learning_rate": 5.998000000000001e-06, "loss": 2.5152, "step": 3000 }, { "epoch": 0.06, "eval_loss": 2.527712345123291, "eval_runtime": 31.6424, "eval_samples_per_second": 3.35, "eval_steps_per_second": 0.853, "step": 3000 }, { "epoch": 0.0605, "grad_norm": 1.8746080139582246, "learning_rate": 6.048e-06, "loss": 2.5109, "step": 3025 }, { "epoch": 0.061, "grad_norm": 0.9932284615776802, "learning_rate": 6.098000000000001e-06, "loss": 2.5241, "step": 3050 }, { "epoch": 0.0615, "grad_norm": 1.065058944784537, "learning_rate": 6.148e-06, "loss": 2.5234, "step": 3075 }, { "epoch": 0.062, "grad_norm": 4.2213264971539495, "learning_rate": 6.198000000000001e-06, "loss": 2.5161, "step": 3100 }, { "epoch": 0.062, "eval_loss": 2.527712345123291, "eval_runtime": 31.6811, "eval_samples_per_second": 3.346, "eval_steps_per_second": 0.852, "step": 3100 }, { "epoch": 0.0625, "grad_norm": 1.0142156249845284, "learning_rate": 6.248000000000001e-06, "loss": 2.5092, "step": 3125 }, { "epoch": 0.063, "grad_norm": 5.27143251635753, "learning_rate": 6.2980000000000005e-06, "loss": 2.5204, "step": 3150 }, { "epoch": 0.0635, "grad_norm": 4.064896667035173, "learning_rate": 6.348000000000001e-06, "loss": 2.5166, "step": 3175 }, { "epoch": 0.064, "grad_norm": 2.3110502080484188, "learning_rate": 6.398000000000001e-06, "loss": 2.5179, "step": 3200 }, { "epoch": 0.064, "eval_loss": 2.5265331268310547, "eval_runtime": 31.9236, "eval_samples_per_second": 3.32, "eval_steps_per_second": 0.846, "step": 3200 }, { "epoch": 0.0645, "grad_norm": 1.3859981336713079, "learning_rate": 6.448000000000001e-06, "loss": 2.5086, "step": 3225 }, { "epoch": 0.065, "grad_norm": 1.29981278376012, "learning_rate": 6.498000000000001e-06, "loss": 2.5156, "step": 3250 }, { "epoch": 0.0655, "grad_norm": 1.212377675038532, "learning_rate": 6.548000000000001e-06, "loss": 2.5085, "step": 3275 }, { "epoch": 0.066, "grad_norm": 0.8179357929230128, "learning_rate": 6.598000000000001e-06, "loss": 2.5151, "step": 3300 }, { "epoch": 0.066, "eval_loss": 2.526238203048706, "eval_runtime": 31.6057, "eval_samples_per_second": 3.354, "eval_steps_per_second": 0.854, "step": 3300 }, { "epoch": 0.0665, "grad_norm": 0.8510661723413057, "learning_rate": 6.648e-06, "loss": 2.501, "step": 3325 }, { "epoch": 0.067, "grad_norm": 0.8276927200382412, "learning_rate": 6.698e-06, "loss": 2.5142, "step": 3350 }, { "epoch": 0.0675, "grad_norm": 0.9378604337812182, "learning_rate": 6.7480000000000004e-06, "loss": 2.5186, "step": 3375 }, { "epoch": 0.068, "grad_norm": 0.9361692538817794, "learning_rate": 6.798e-06, "loss": 2.5172, "step": 3400 }, { "epoch": 0.068, "eval_loss": 2.5250589847564697, "eval_runtime": 31.6061, "eval_samples_per_second": 3.354, "eval_steps_per_second": 0.854, "step": 3400 }, { "epoch": 0.0685, "grad_norm": 2.3422890150387055, "learning_rate": 6.848e-06, "loss": 2.5228, "step": 3425 }, { "epoch": 0.069, "grad_norm": 1.4884269701808963, "learning_rate": 6.898e-06, "loss": 2.5202, "step": 3450 }, { "epoch": 0.0695, "grad_norm": 2.1790981658446458, "learning_rate": 6.948e-06, "loss": 2.5093, "step": 3475 }, { "epoch": 0.07, "grad_norm": 2.0704908228926886, "learning_rate": 6.998000000000001e-06, "loss": 2.5139, "step": 3500 }, { "epoch": 0.07, "eval_loss": 2.524764060974121, "eval_runtime": 31.5652, "eval_samples_per_second": 3.358, "eval_steps_per_second": 0.855, "step": 3500 }, { "epoch": 0.0705, "grad_norm": 1.7179296688504122, "learning_rate": 7.048e-06, "loss": 2.5233, "step": 3525 }, { "epoch": 0.071, "grad_norm": 1.0022504705786135, "learning_rate": 7.0980000000000005e-06, "loss": 2.5239, "step": 3550 }, { "epoch": 0.0715, "grad_norm": 0.8330455350318481, "learning_rate": 7.148000000000001e-06, "loss": 2.5138, "step": 3575 }, { "epoch": 0.072, "grad_norm": 1.4134926601254947, "learning_rate": 7.198e-06, "loss": 2.5164, "step": 3600 }, { "epoch": 0.072, "eval_loss": 2.524174451828003, "eval_runtime": 31.6742, "eval_samples_per_second": 3.347, "eval_steps_per_second": 0.852, "step": 3600 }, { "epoch": 0.0725, "grad_norm": 6.079377219850451, "learning_rate": 7.248000000000001e-06, "loss": 2.5286, "step": 3625 }, { "epoch": 0.073, "grad_norm": 1.2135601583656899, "learning_rate": 7.298e-06, "loss": 2.522, "step": 3650 }, { "epoch": 0.0735, "grad_norm": 1.1679985497748089, "learning_rate": 7.348000000000001e-06, "loss": 2.5177, "step": 3675 }, { "epoch": 0.074, "grad_norm": 1.7943286008630555, "learning_rate": 7.398000000000001e-06, "loss": 2.496, "step": 3700 }, { "epoch": 0.074, "eval_loss": 2.5235848426818848, "eval_runtime": 31.7011, "eval_samples_per_second": 3.344, "eval_steps_per_second": 0.852, "step": 3700 }, { "epoch": 0.0745, "grad_norm": 3.9796277292747666, "learning_rate": 7.4480000000000005e-06, "loss": 2.5169, "step": 3725 }, { "epoch": 0.075, "grad_norm": 1.380394250398271, "learning_rate": 7.498000000000001e-06, "loss": 2.5094, "step": 3750 }, { "epoch": 0.0755, "grad_norm": 3.0313765413416895, "learning_rate": 7.548000000000001e-06, "loss": 2.5055, "step": 3775 }, { "epoch": 0.076, "grad_norm": 2.245387638839618, "learning_rate": 7.598000000000001e-06, "loss": 2.5172, "step": 3800 }, { "epoch": 0.076, "eval_loss": 2.5229952335357666, "eval_runtime": 31.6283, "eval_samples_per_second": 3.351, "eval_steps_per_second": 0.854, "step": 3800 }, { "epoch": 0.0765, "grad_norm": 1.706692381304059, "learning_rate": 7.648e-06, "loss": 2.5128, "step": 3825 }, { "epoch": 0.077, "grad_norm": 1.085621140871541, "learning_rate": 7.698000000000002e-06, "loss": 2.5104, "step": 3850 }, { "epoch": 0.0775, "grad_norm": 1.240137406582716, "learning_rate": 7.748000000000001e-06, "loss": 2.5221, "step": 3875 }, { "epoch": 0.078, "grad_norm": 1.4727017842198826, "learning_rate": 7.798e-06, "loss": 2.5002, "step": 3900 }, { "epoch": 0.078, "eval_loss": 2.5224056243896484, "eval_runtime": 31.6561, "eval_samples_per_second": 3.348, "eval_steps_per_second": 0.853, "step": 3900 }, { "epoch": 0.0785, "grad_norm": 0.8287061589077178, "learning_rate": 7.848000000000002e-06, "loss": 2.5188, "step": 3925 }, { "epoch": 0.079, "grad_norm": 1.9412593547366452, "learning_rate": 7.898e-06, "loss": 2.5128, "step": 3950 }, { "epoch": 0.0795, "grad_norm": 1.507826764306867, "learning_rate": 7.948e-06, "loss": 2.5157, "step": 3975 }, { "epoch": 0.08, "grad_norm": 1.0070408907220643, "learning_rate": 7.998e-06, "loss": 2.5136, "step": 4000 }, { "epoch": 0.08, "eval_loss": 2.5203418731689453, "eval_runtime": 31.6704, "eval_samples_per_second": 3.347, "eval_steps_per_second": 0.853, "step": 4000 }, { "epoch": 0.0805, "grad_norm": 0.9139528882890205, "learning_rate": 8.048e-06, "loss": 2.5153, "step": 4025 }, { "epoch": 0.081, "grad_norm": 3.1234207366182027, "learning_rate": 8.098000000000001e-06, "loss": 2.5125, "step": 4050 }, { "epoch": 0.0815, "grad_norm": 0.8261934904482937, "learning_rate": 8.148e-06, "loss": 2.5197, "step": 4075 }, { "epoch": 0.082, "grad_norm": 0.8407881378613186, "learning_rate": 8.198e-06, "loss": 2.5184, "step": 4100 }, { "epoch": 0.082, "eval_loss": 2.5209317207336426, "eval_runtime": 31.816, "eval_samples_per_second": 3.332, "eval_steps_per_second": 0.849, "step": 4100 }, { "epoch": 0.0825, "grad_norm": 0.6865188622763857, "learning_rate": 8.248e-06, "loss": 2.5108, "step": 4125 }, { "epoch": 0.083, "grad_norm": 0.8440391843752413, "learning_rate": 8.298000000000001e-06, "loss": 2.504, "step": 4150 }, { "epoch": 0.0835, "grad_norm": 1.2406936940712257, "learning_rate": 8.348e-06, "loss": 2.5128, "step": 4175 }, { "epoch": 0.084, "grad_norm": 1.0065152284825731, "learning_rate": 8.398e-06, "loss": 2.511, "step": 4200 }, { "epoch": 0.084, "eval_loss": 2.5203418731689453, "eval_runtime": 31.6615, "eval_samples_per_second": 3.348, "eval_steps_per_second": 0.853, "step": 4200 }, { "epoch": 0.0845, "grad_norm": 1.037069923815588, "learning_rate": 8.448000000000001e-06, "loss": 2.5088, "step": 4225 }, { "epoch": 0.085, "grad_norm": 2.4303667066934254, "learning_rate": 8.498e-06, "loss": 2.4997, "step": 4250 }, { "epoch": 0.0855, "grad_norm": 1.4724222217702825, "learning_rate": 8.548e-06, "loss": 2.5073, "step": 4275 }, { "epoch": 0.086, "grad_norm": 9.444493964030588, "learning_rate": 8.598000000000001e-06, "loss": 2.5124, "step": 4300 }, { "epoch": 0.086, "eval_loss": 2.5194575786590576, "eval_runtime": 31.6406, "eval_samples_per_second": 3.35, "eval_steps_per_second": 0.853, "step": 4300 }, { "epoch": 0.0865, "grad_norm": 2.218623091874727, "learning_rate": 8.648000000000001e-06, "loss": 2.5, "step": 4325 }, { "epoch": 0.087, "grad_norm": 1.1325241529581473, "learning_rate": 8.698e-06, "loss": 2.5131, "step": 4350 }, { "epoch": 0.0875, "grad_norm": 1.4061959818148488, "learning_rate": 8.748000000000002e-06, "loss": 2.5078, "step": 4375 }, { "epoch": 0.088, "grad_norm": 0.7924670159627136, "learning_rate": 8.798000000000001e-06, "loss": 2.5081, "step": 4400 }, { "epoch": 0.088, "eval_loss": 2.519162654876709, "eval_runtime": 31.5649, "eval_samples_per_second": 3.358, "eval_steps_per_second": 0.855, "step": 4400 }, { "epoch": 0.0885, "grad_norm": 1.7741039093023965, "learning_rate": 8.848e-06, "loss": 2.5067, "step": 4425 }, { "epoch": 0.089, "grad_norm": 0.766073391828734, "learning_rate": 8.898000000000002e-06, "loss": 2.5029, "step": 4450 }, { "epoch": 0.0895, "grad_norm": 0.7195151746831012, "learning_rate": 8.948000000000001e-06, "loss": 2.5085, "step": 4475 }, { "epoch": 0.09, "grad_norm": 0.8599409201158774, "learning_rate": 8.998000000000001e-06, "loss": 2.5071, "step": 4500 }, { "epoch": 0.09, "eval_loss": 2.5194575786590576, "eval_runtime": 31.6267, "eval_samples_per_second": 3.352, "eval_steps_per_second": 0.854, "step": 4500 }, { "epoch": 0.0905, "grad_norm": 0.713893441525958, "learning_rate": 9.048e-06, "loss": 2.5072, "step": 4525 }, { "epoch": 0.091, "grad_norm": 0.6233173190909058, "learning_rate": 9.098000000000002e-06, "loss": 2.5019, "step": 4550 }, { "epoch": 0.0915, "grad_norm": 0.6469529621475963, "learning_rate": 9.148e-06, "loss": 2.5044, "step": 4575 }, { "epoch": 0.092, "grad_norm": 0.6608368808707902, "learning_rate": 9.198e-06, "loss": 2.4945, "step": 4600 }, { "epoch": 0.092, "eval_loss": 2.518573045730591, "eval_runtime": 31.6806, "eval_samples_per_second": 3.346, "eval_steps_per_second": 0.852, "step": 4600 }, { "epoch": 0.0925, "grad_norm": 0.8106332715320632, "learning_rate": 9.248e-06, "loss": 2.5067, "step": 4625 }, { "epoch": 0.093, "grad_norm": 0.613334869116551, "learning_rate": 9.298e-06, "loss": 2.514, "step": 4650 }, { "epoch": 0.0935, "grad_norm": 0.6303730281536211, "learning_rate": 9.348000000000001e-06, "loss": 2.5083, "step": 4675 }, { "epoch": 0.094, "grad_norm": 0.6322414031260126, "learning_rate": 9.398e-06, "loss": 2.5068, "step": 4700 }, { "epoch": 0.094, "eval_loss": 2.517688751220703, "eval_runtime": 31.6555, "eval_samples_per_second": 3.349, "eval_steps_per_second": 0.853, "step": 4700 }, { "epoch": 0.0945, "grad_norm": 0.7141142061627065, "learning_rate": 9.448e-06, "loss": 2.505, "step": 4725 }, { "epoch": 0.095, "grad_norm": 0.7359902069087727, "learning_rate": 9.498000000000001e-06, "loss": 2.5064, "step": 4750 }, { "epoch": 0.0955, "grad_norm": 2.8939342680384064, "learning_rate": 9.548e-06, "loss": 2.4955, "step": 4775 }, { "epoch": 0.096, "grad_norm": 0.6142018458398064, "learning_rate": 9.598e-06, "loss": 2.5262, "step": 4800 }, { "epoch": 0.096, "eval_loss": 2.5173938274383545, "eval_runtime": 31.8932, "eval_samples_per_second": 3.324, "eval_steps_per_second": 0.847, "step": 4800 }, { "epoch": 0.0965, "grad_norm": 0.7591886413422146, "learning_rate": 9.648000000000001e-06, "loss": 2.5039, "step": 4825 }, { "epoch": 0.097, "grad_norm": 0.8741359664470049, "learning_rate": 9.698000000000001e-06, "loss": 2.5101, "step": 4850 }, { "epoch": 0.0975, "grad_norm": 0.6058451377431769, "learning_rate": 9.748e-06, "loss": 2.508, "step": 4875 }, { "epoch": 0.098, "grad_norm": 0.6676967508556061, "learning_rate": 9.798e-06, "loss": 2.5105, "step": 4900 }, { "epoch": 0.098, "eval_loss": 2.5173938274383545, "eval_runtime": 31.6434, "eval_samples_per_second": 3.35, "eval_steps_per_second": 0.853, "step": 4900 }, { "epoch": 0.0985, "grad_norm": 0.6658518907375024, "learning_rate": 9.848000000000001e-06, "loss": 2.5086, "step": 4925 }, { "epoch": 0.099, "grad_norm": 0.6210642351345088, "learning_rate": 9.898e-06, "loss": 2.5088, "step": 4950 }, { "epoch": 0.0995, "grad_norm": 0.5889895299848711, "learning_rate": 9.948e-06, "loss": 2.5125, "step": 4975 }, { "epoch": 0.1, "grad_norm": 0.601939432781195, "learning_rate": 9.998000000000002e-06, "loss": 2.507, "step": 5000 }, { "epoch": 0.1, "eval_loss": 2.5168042182922363, "eval_runtime": 31.7525, "eval_samples_per_second": 3.338, "eval_steps_per_second": 0.85, "step": 5000 }, { "epoch": 0.1005, "grad_norm": 0.5863277906860563, "learning_rate": 9.994666666666668e-06, "loss": 2.5049, "step": 5025 }, { "epoch": 0.101, "grad_norm": 0.5907628491061302, "learning_rate": 9.989111111111111e-06, "loss": 2.5081, "step": 5050 }, { "epoch": 0.1015, "grad_norm": 0.6052892176154054, "learning_rate": 9.983555555555556e-06, "loss": 2.492, "step": 5075 }, { "epoch": 0.102, "grad_norm": 0.60615426444314, "learning_rate": 9.978000000000002e-06, "loss": 2.5117, "step": 5100 }, { "epoch": 0.102, "eval_loss": 2.5159199237823486, "eval_runtime": 31.6096, "eval_samples_per_second": 3.353, "eval_steps_per_second": 0.854, "step": 5100 }, { "epoch": 0.1025, "grad_norm": 0.9921486289898994, "learning_rate": 9.972444444444445e-06, "loss": 2.5042, "step": 5125 }, { "epoch": 0.103, "grad_norm": 0.5438905467342706, "learning_rate": 9.966888888888889e-06, "loss": 2.5186, "step": 5150 }, { "epoch": 0.1035, "grad_norm": 0.6671923807124565, "learning_rate": 9.961333333333334e-06, "loss": 2.4965, "step": 5175 }, { "epoch": 0.104, "grad_norm": 0.5927621920818558, "learning_rate": 9.95577777777778e-06, "loss": 2.5151, "step": 5200 }, { "epoch": 0.104, "eval_loss": 2.516214609146118, "eval_runtime": 33.4192, "eval_samples_per_second": 3.172, "eval_steps_per_second": 0.808, "step": 5200 }, { "epoch": 0.1045, "grad_norm": 0.5887168586235929, "learning_rate": 9.950222222222223e-06, "loss": 2.5022, "step": 5225 }, { "epoch": 0.105, "grad_norm": 0.601347414578016, "learning_rate": 9.944666666666668e-06, "loss": 2.5033, "step": 5250 }, { "epoch": 0.1055, "grad_norm": 1.1147364476570238, "learning_rate": 9.939111111111112e-06, "loss": 2.5028, "step": 5275 }, { "epoch": 0.106, "grad_norm": 0.6327491084423742, "learning_rate": 9.933555555555557e-06, "loss": 2.5015, "step": 5300 }, { "epoch": 0.106, "eval_loss": 2.515625, "eval_runtime": 33.2149, "eval_samples_per_second": 3.191, "eval_steps_per_second": 0.813, "step": 5300 }, { "epoch": 0.1065, "grad_norm": 0.5782577240017591, "learning_rate": 9.928e-06, "loss": 2.5027, "step": 5325 }, { "epoch": 0.107, "grad_norm": 0.5793288292949659, "learning_rate": 9.922444444444446e-06, "loss": 2.5211, "step": 5350 }, { "epoch": 0.1075, "grad_norm": 0.5590969465700796, "learning_rate": 9.91688888888889e-06, "loss": 2.5011, "step": 5375 }, { "epoch": 0.108, "grad_norm": 0.5479364900193122, "learning_rate": 9.911333333333335e-06, "loss": 2.5017, "step": 5400 }, { "epoch": 0.108, "eval_loss": 2.515035390853882, "eval_runtime": 33.559, "eval_samples_per_second": 3.159, "eval_steps_per_second": 0.805, "step": 5400 }, { "epoch": 0.1085, "grad_norm": 0.6071449747188377, "learning_rate": 9.905777777777778e-06, "loss": 2.5082, "step": 5425 }, { "epoch": 0.109, "grad_norm": 0.586787674324507, "learning_rate": 9.900222222222223e-06, "loss": 2.4976, "step": 5450 }, { "epoch": 0.1095, "grad_norm": 0.5518198424351226, "learning_rate": 9.894666666666669e-06, "loss": 2.5015, "step": 5475 }, { "epoch": 0.11, "grad_norm": 0.6202159170702171, "learning_rate": 9.889111111111112e-06, "loss": 2.5112, "step": 5500 }, { "epoch": 0.11, "eval_loss": 2.515035390853882, "eval_runtime": 33.5553, "eval_samples_per_second": 3.159, "eval_steps_per_second": 0.805, "step": 5500 }, { "epoch": 0.1105, "grad_norm": 0.6453434661163482, "learning_rate": 9.883555555555556e-06, "loss": 2.4963, "step": 5525 }, { "epoch": 0.111, "grad_norm": 0.5456728245790678, "learning_rate": 9.878000000000001e-06, "loss": 2.503, "step": 5550 }, { "epoch": 0.1115, "grad_norm": 0.5360784869281232, "learning_rate": 9.872444444444446e-06, "loss": 2.502, "step": 5575 }, { "epoch": 0.112, "grad_norm": 0.5546380785080989, "learning_rate": 9.86688888888889e-06, "loss": 2.5047, "step": 5600 }, { "epoch": 0.112, "eval_loss": 2.5138561725616455, "eval_runtime": 31.7558, "eval_samples_per_second": 3.338, "eval_steps_per_second": 0.85, "step": 5600 }, { "epoch": 0.1125, "grad_norm": 0.5677229074415441, "learning_rate": 9.861333333333333e-06, "loss": 2.5084, "step": 5625 }, { "epoch": 0.113, "grad_norm": 0.5665828460218333, "learning_rate": 9.855777777777779e-06, "loss": 2.5006, "step": 5650 }, { "epoch": 0.1135, "grad_norm": 0.6231915277731267, "learning_rate": 9.850222222222224e-06, "loss": 2.5019, "step": 5675 }, { "epoch": 0.114, "grad_norm": 0.5386821317170616, "learning_rate": 9.844666666666667e-06, "loss": 2.4936, "step": 5700 }, { "epoch": 0.114, "eval_loss": 2.512676954269409, "eval_runtime": 31.6592, "eval_samples_per_second": 3.348, "eval_steps_per_second": 0.853, "step": 5700 }, { "epoch": 0.1145, "grad_norm": 0.5333458931081642, "learning_rate": 9.839111111111111e-06, "loss": 2.5128, "step": 5725 }, { "epoch": 0.115, "grad_norm": 0.6634412967350025, "learning_rate": 9.833555555555556e-06, "loss": 2.4935, "step": 5750 }, { "epoch": 0.1155, "grad_norm": 0.6495713608767936, "learning_rate": 9.828000000000001e-06, "loss": 2.5097, "step": 5775 }, { "epoch": 0.116, "grad_norm": 0.6359186009058505, "learning_rate": 9.822444444444445e-06, "loss": 2.5047, "step": 5800 }, { "epoch": 0.116, "eval_loss": 2.5129716396331787, "eval_runtime": 31.5608, "eval_samples_per_second": 3.359, "eval_steps_per_second": 0.855, "step": 5800 }, { "epoch": 0.1165, "grad_norm": 0.6239988101253886, "learning_rate": 9.81688888888889e-06, "loss": 2.4978, "step": 5825 }, { "epoch": 0.117, "grad_norm": 0.5942643505285292, "learning_rate": 9.811333333333334e-06, "loss": 2.4944, "step": 5850 }, { "epoch": 0.1175, "grad_norm": 0.5335363645218594, "learning_rate": 9.805777777777779e-06, "loss": 2.4962, "step": 5875 }, { "epoch": 0.118, "grad_norm": 0.5494019049427782, "learning_rate": 9.800222222222223e-06, "loss": 2.5034, "step": 5900 }, { "epoch": 0.118, "eval_loss": 2.512087345123291, "eval_runtime": 31.5964, "eval_samples_per_second": 3.355, "eval_steps_per_second": 0.855, "step": 5900 }, { "epoch": 0.1185, "grad_norm": 0.5628262336899048, "learning_rate": 9.794666666666668e-06, "loss": 2.498, "step": 5925 }, { "epoch": 0.119, "grad_norm": 0.5309875914486921, "learning_rate": 9.789111111111111e-06, "loss": 2.5058, "step": 5950 }, { "epoch": 0.1195, "grad_norm": 0.5098048565176445, "learning_rate": 9.783555555555557e-06, "loss": 2.5019, "step": 5975 }, { "epoch": 0.12, "grad_norm": 0.5525705733015823, "learning_rate": 9.778e-06, "loss": 2.5073, "step": 6000 }, { "epoch": 0.12, "eval_loss": 2.5123820304870605, "eval_runtime": 31.6524, "eval_samples_per_second": 3.349, "eval_steps_per_second": 0.853, "step": 6000 }, { "epoch": 0.1205, "grad_norm": 0.667754049742565, "learning_rate": 9.772444444444445e-06, "loss": 2.5045, "step": 6025 }, { "epoch": 0.121, "grad_norm": 0.5496899380142036, "learning_rate": 9.76688888888889e-06, "loss": 2.498, "step": 6050 }, { "epoch": 0.1215, "grad_norm": 0.5366932619820161, "learning_rate": 9.761333333333334e-06, "loss": 2.5047, "step": 6075 }, { "epoch": 0.122, "grad_norm": 0.5307187417069992, "learning_rate": 9.755777777777778e-06, "loss": 2.5068, "step": 6100 }, { "epoch": 0.122, "eval_loss": 2.511202812194824, "eval_runtime": 31.6436, "eval_samples_per_second": 3.35, "eval_steps_per_second": 0.853, "step": 6100 }, { "epoch": 0.1225, "grad_norm": 0.5120264018317957, "learning_rate": 9.750222222222223e-06, "loss": 2.5015, "step": 6125 }, { "epoch": 0.123, "grad_norm": 0.5282121644036795, "learning_rate": 9.744666666666668e-06, "loss": 2.507, "step": 6150 }, { "epoch": 0.1235, "grad_norm": 0.634830465577626, "learning_rate": 9.739111111111112e-06, "loss": 2.5014, "step": 6175 }, { "epoch": 0.124, "grad_norm": 0.514470702860497, "learning_rate": 9.733555555555555e-06, "loss": 2.4965, "step": 6200 }, { "epoch": 0.124, "eval_loss": 2.511202812194824, "eval_runtime": 31.6088, "eval_samples_per_second": 3.353, "eval_steps_per_second": 0.854, "step": 6200 }, { "epoch": 0.1245, "grad_norm": 0.4997175642406598, "learning_rate": 9.728e-06, "loss": 2.4948, "step": 6225 }, { "epoch": 0.125, "grad_norm": 0.5497609531752035, "learning_rate": 9.722444444444446e-06, "loss": 2.503, "step": 6250 }, { "epoch": 0.1255, "grad_norm": 0.6144070155717539, "learning_rate": 9.71688888888889e-06, "loss": 2.4977, "step": 6275 }, { "epoch": 0.126, "grad_norm": 0.5191303677449087, "learning_rate": 9.711333333333333e-06, "loss": 2.4975, "step": 6300 }, { "epoch": 0.126, "eval_loss": 2.5103182792663574, "eval_runtime": 31.7187, "eval_samples_per_second": 3.342, "eval_steps_per_second": 0.851, "step": 6300 }, { "epoch": 0.1265, "grad_norm": 0.6330521999309859, "learning_rate": 9.705777777777778e-06, "loss": 2.4995, "step": 6325 }, { "epoch": 0.127, "grad_norm": 0.5462430526003252, "learning_rate": 9.700222222222224e-06, "loss": 2.5047, "step": 6350 }, { "epoch": 0.1275, "grad_norm": 0.9329321931683348, "learning_rate": 9.694666666666667e-06, "loss": 2.5096, "step": 6375 }, { "epoch": 0.128, "grad_norm": 0.49836723352839235, "learning_rate": 9.68911111111111e-06, "loss": 2.4995, "step": 6400 }, { "epoch": 0.128, "eval_loss": 2.510023593902588, "eval_runtime": 31.6077, "eval_samples_per_second": 3.354, "eval_steps_per_second": 0.854, "step": 6400 }, { "epoch": 0.1285, "grad_norm": 0.5413116858337882, "learning_rate": 9.683555555555556e-06, "loss": 2.4997, "step": 6425 }, { "epoch": 0.129, "grad_norm": 0.5192425593604233, "learning_rate": 9.678000000000001e-06, "loss": 2.4985, "step": 6450 }, { "epoch": 0.1295, "grad_norm": 0.543683383393961, "learning_rate": 9.672444444444445e-06, "loss": 2.4905, "step": 6475 }, { "epoch": 0.13, "grad_norm": 0.5120901133193708, "learning_rate": 9.66688888888889e-06, "loss": 2.4981, "step": 6500 }, { "epoch": 0.13, "eval_loss": 2.5097286701202393, "eval_runtime": 31.6399, "eval_samples_per_second": 3.35, "eval_steps_per_second": 0.853, "step": 6500 }, { "epoch": 0.1305, "grad_norm": 0.6094958786556292, "learning_rate": 9.661333333333334e-06, "loss": 2.5072, "step": 6525 }, { "epoch": 0.131, "grad_norm": 0.5072948460719591, "learning_rate": 9.655777777777779e-06, "loss": 2.4927, "step": 6550 }, { "epoch": 0.1315, "grad_norm": 0.623957969467135, "learning_rate": 9.650222222222222e-06, "loss": 2.5075, "step": 6575 }, { "epoch": 0.132, "grad_norm": 0.4720667981244025, "learning_rate": 9.644666666666668e-06, "loss": 2.5004, "step": 6600 }, { "epoch": 0.132, "eval_loss": 2.5094339847564697, "eval_runtime": 31.6371, "eval_samples_per_second": 3.35, "eval_steps_per_second": 0.853, "step": 6600 }, { "epoch": 0.1325, "grad_norm": 0.6015585416072105, "learning_rate": 9.639111111111113e-06, "loss": 2.5031, "step": 6625 }, { "epoch": 0.133, "grad_norm": 0.5046082226722687, "learning_rate": 9.633555555555556e-06, "loss": 2.4909, "step": 6650 }, { "epoch": 0.1335, "grad_norm": 0.48445442127855576, "learning_rate": 9.628e-06, "loss": 2.5076, "step": 6675 }, { "epoch": 0.134, "grad_norm": 0.560939985354956, "learning_rate": 9.622444444444445e-06, "loss": 2.5048, "step": 6700 }, { "epoch": 0.134, "eval_loss": 2.5088443756103516, "eval_runtime": 31.6114, "eval_samples_per_second": 3.353, "eval_steps_per_second": 0.854, "step": 6700 }, { "epoch": 0.1345, "grad_norm": 0.4688522723722928, "learning_rate": 9.61688888888889e-06, "loss": 2.4975, "step": 6725 }, { "epoch": 0.135, "grad_norm": 0.485594338083258, "learning_rate": 9.611333333333334e-06, "loss": 2.4913, "step": 6750 }, { "epoch": 0.1355, "grad_norm": 0.5040897875178989, "learning_rate": 9.605777777777778e-06, "loss": 2.4943, "step": 6775 }, { "epoch": 0.136, "grad_norm": 0.5221567713674383, "learning_rate": 9.600222222222223e-06, "loss": 2.5075, "step": 6800 }, { "epoch": 0.136, "eval_loss": 2.5088443756103516, "eval_runtime": 31.6782, "eval_samples_per_second": 3.346, "eval_steps_per_second": 0.852, "step": 6800 }, { "epoch": 0.1365, "grad_norm": 0.5082542008947744, "learning_rate": 9.594666666666668e-06, "loss": 2.4978, "step": 6825 }, { "epoch": 0.137, "grad_norm": 0.5524384559495555, "learning_rate": 9.589111111111112e-06, "loss": 2.4972, "step": 6850 }, { "epoch": 0.1375, "grad_norm": 0.7161574482434298, "learning_rate": 9.583555555555555e-06, "loss": 2.4975, "step": 6875 }, { "epoch": 0.138, "grad_norm": 0.5149515733981297, "learning_rate": 9.578e-06, "loss": 2.4907, "step": 6900 }, { "epoch": 0.138, "eval_loss": 2.508549451828003, "eval_runtime": 31.6469, "eval_samples_per_second": 3.349, "eval_steps_per_second": 0.853, "step": 6900 }, { "epoch": 0.1385, "grad_norm": 0.49397824038426996, "learning_rate": 9.572444444444446e-06, "loss": 2.5014, "step": 6925 }, { "epoch": 0.139, "grad_norm": 0.6097004085544548, "learning_rate": 9.56688888888889e-06, "loss": 2.4954, "step": 6950 }, { "epoch": 0.1395, "grad_norm": 0.49894922737806585, "learning_rate": 9.561333333333333e-06, "loss": 2.4985, "step": 6975 }, { "epoch": 0.14, "grad_norm": 0.5218723910214107, "learning_rate": 9.555777777777778e-06, "loss": 2.5004, "step": 7000 }, { "epoch": 0.14, "eval_loss": 2.508549451828003, "eval_runtime": 31.5947, "eval_samples_per_second": 3.355, "eval_steps_per_second": 0.855, "step": 7000 }, { "epoch": 0.1405, "grad_norm": 0.563499806545184, "learning_rate": 9.550222222222223e-06, "loss": 2.4977, "step": 7025 }, { "epoch": 0.141, "grad_norm": 0.5009758538726304, "learning_rate": 9.544666666666667e-06, "loss": 2.4952, "step": 7050 }, { "epoch": 0.1415, "grad_norm": 0.4508864187267148, "learning_rate": 9.539111111111112e-06, "loss": 2.508, "step": 7075 }, { "epoch": 0.142, "grad_norm": 0.4912390956522406, "learning_rate": 9.533555555555556e-06, "loss": 2.4966, "step": 7100 }, { "epoch": 0.142, "eval_loss": 2.5082547664642334, "eval_runtime": 33.4326, "eval_samples_per_second": 3.171, "eval_steps_per_second": 0.808, "step": 7100 }, { "epoch": 0.1425, "grad_norm": 0.49279878561988544, "learning_rate": 9.528000000000001e-06, "loss": 2.4968, "step": 7125 }, { "epoch": 0.143, "grad_norm": 0.46664155679935654, "learning_rate": 9.522444444444444e-06, "loss": 2.4895, "step": 7150 }, { "epoch": 0.1435, "grad_norm": 0.5124511409909606, "learning_rate": 9.51688888888889e-06, "loss": 2.4915, "step": 7175 }, { "epoch": 0.144, "grad_norm": 0.47556388802117383, "learning_rate": 9.511333333333335e-06, "loss": 2.4936, "step": 7200 }, { "epoch": 0.144, "eval_loss": 2.5082547664642334, "eval_runtime": 33.449, "eval_samples_per_second": 3.169, "eval_steps_per_second": 0.807, "step": 7200 }, { "epoch": 0.1445, "grad_norm": 0.48653320400655387, "learning_rate": 9.505777777777779e-06, "loss": 2.5018, "step": 7225 }, { "epoch": 0.145, "grad_norm": 0.5047213279832499, "learning_rate": 9.500222222222222e-06, "loss": 2.4988, "step": 7250 }, { "epoch": 0.1455, "grad_norm": 0.5400569539883815, "learning_rate": 9.494666666666667e-06, "loss": 2.506, "step": 7275 }, { "epoch": 0.146, "grad_norm": 0.4719790037079782, "learning_rate": 9.489111111111113e-06, "loss": 2.5052, "step": 7300 }, { "epoch": 0.146, "eval_loss": 2.5082547664642334, "eval_runtime": 32.1166, "eval_samples_per_second": 3.3, "eval_steps_per_second": 0.841, "step": 7300 }, { "epoch": 0.1465, "grad_norm": 0.4614427746650018, "learning_rate": 9.483555555555556e-06, "loss": 2.4915, "step": 7325 }, { "epoch": 0.147, "grad_norm": 0.445520399417109, "learning_rate": 9.478e-06, "loss": 2.4854, "step": 7350 }, { "epoch": 0.1475, "grad_norm": 0.5186998152774688, "learning_rate": 9.472444444444445e-06, "loss": 2.5034, "step": 7375 }, { "epoch": 0.148, "grad_norm": 0.5220314227255131, "learning_rate": 9.46688888888889e-06, "loss": 2.5027, "step": 7400 }, { "epoch": 0.148, "eval_loss": 2.5079598426818848, "eval_runtime": 31.6827, "eval_samples_per_second": 3.346, "eval_steps_per_second": 0.852, "step": 7400 }, { "epoch": 0.1485, "grad_norm": 0.4801927645172527, "learning_rate": 9.461333333333334e-06, "loss": 2.4944, "step": 7425 }, { "epoch": 0.149, "grad_norm": 0.5083289818755181, "learning_rate": 9.455777777777777e-06, "loss": 2.4968, "step": 7450 }, { "epoch": 0.1495, "grad_norm": 0.5279857420863452, "learning_rate": 9.450222222222223e-06, "loss": 2.4943, "step": 7475 }, { "epoch": 0.15, "grad_norm": 0.46632473469022884, "learning_rate": 9.444666666666668e-06, "loss": 2.4888, "step": 7500 }, { "epoch": 0.15, "eval_loss": 2.5073702335357666, "eval_runtime": 31.6038, "eval_samples_per_second": 3.354, "eval_steps_per_second": 0.854, "step": 7500 }, { "epoch": 0.1505, "grad_norm": 0.49040227635264855, "learning_rate": 9.439111111111111e-06, "loss": 2.501, "step": 7525 }, { "epoch": 0.151, "grad_norm": 0.47705017863255805, "learning_rate": 9.433555555555557e-06, "loss": 2.496, "step": 7550 }, { "epoch": 0.1515, "grad_norm": 0.5043629708475141, "learning_rate": 9.428e-06, "loss": 2.5044, "step": 7575 }, { "epoch": 0.152, "grad_norm": 0.5031322287130929, "learning_rate": 9.422444444444445e-06, "loss": 2.5059, "step": 7600 }, { "epoch": 0.152, "eval_loss": 2.5067806243896484, "eval_runtime": 31.7068, "eval_samples_per_second": 3.343, "eval_steps_per_second": 0.852, "step": 7600 }, { "epoch": 0.1525, "grad_norm": 0.5134164370121488, "learning_rate": 9.41688888888889e-06, "loss": 2.5024, "step": 7625 }, { "epoch": 0.153, "grad_norm": 0.4623704562763065, "learning_rate": 9.411333333333334e-06, "loss": 2.501, "step": 7650 }, { "epoch": 0.1535, "grad_norm": 0.49559420602071524, "learning_rate": 9.405777777777778e-06, "loss": 2.5093, "step": 7675 }, { "epoch": 0.154, "grad_norm": 0.5243071177219503, "learning_rate": 9.400222222222223e-06, "loss": 2.4945, "step": 7700 }, { "epoch": 0.154, "eval_loss": 2.5073702335357666, "eval_runtime": 31.6912, "eval_samples_per_second": 3.345, "eval_steps_per_second": 0.852, "step": 7700 }, { "epoch": 0.1545, "grad_norm": 0.5073458932892907, "learning_rate": 9.394666666666668e-06, "loss": 2.4941, "step": 7725 }, { "epoch": 0.155, "grad_norm": 0.5788636950869732, "learning_rate": 9.389111111111112e-06, "loss": 2.4922, "step": 7750 }, { "epoch": 0.1555, "grad_norm": 0.4512531421146986, "learning_rate": 9.383555555555557e-06, "loss": 2.4765, "step": 7775 }, { "epoch": 0.156, "grad_norm": 0.49522295115785536, "learning_rate": 9.378e-06, "loss": 2.5012, "step": 7800 }, { "epoch": 0.156, "eval_loss": 2.507075548171997, "eval_runtime": 31.6414, "eval_samples_per_second": 3.35, "eval_steps_per_second": 0.853, "step": 7800 }, { "epoch": 0.1565, "grad_norm": 0.47848211440373783, "learning_rate": 9.372444444444446e-06, "loss": 2.5, "step": 7825 }, { "epoch": 0.157, "grad_norm": 0.4640110840807158, "learning_rate": 9.36688888888889e-06, "loss": 2.4955, "step": 7850 }, { "epoch": 0.1575, "grad_norm": 0.5396945826650712, "learning_rate": 9.361333333333335e-06, "loss": 2.4878, "step": 7875 }, { "epoch": 0.158, "grad_norm": 0.46462373787773387, "learning_rate": 9.355777777777778e-06, "loss": 2.4996, "step": 7900 }, { "epoch": 0.158, "eval_loss": 2.507075548171997, "eval_runtime": 31.7182, "eval_samples_per_second": 3.342, "eval_steps_per_second": 0.851, "step": 7900 }, { "epoch": 0.1585, "grad_norm": 0.47428789515219244, "learning_rate": 9.350222222222224e-06, "loss": 2.5048, "step": 7925 }, { "epoch": 0.159, "grad_norm": 0.5086252587406799, "learning_rate": 9.344666666666667e-06, "loss": 2.495, "step": 7950 }, { "epoch": 0.1595, "grad_norm": 0.5063443688127869, "learning_rate": 9.339111111111112e-06, "loss": 2.4941, "step": 7975 }, { "epoch": 0.16, "grad_norm": 0.5048581245780791, "learning_rate": 9.333555555555558e-06, "loss": 2.501, "step": 8000 }, { "epoch": 0.16, "eval_loss": 2.5067806243896484, "eval_runtime": 31.6901, "eval_samples_per_second": 3.345, "eval_steps_per_second": 0.852, "step": 8000 }, { "epoch": 0.1605, "grad_norm": 0.5643292039022151, "learning_rate": 9.328000000000001e-06, "loss": 2.5023, "step": 8025 }, { "epoch": 0.161, "grad_norm": 0.4727327148593234, "learning_rate": 9.322444444444445e-06, "loss": 2.4994, "step": 8050 }, { "epoch": 0.1615, "grad_norm": 0.45408437785378586, "learning_rate": 9.31688888888889e-06, "loss": 2.4915, "step": 8075 }, { "epoch": 0.162, "grad_norm": 0.458839332576604, "learning_rate": 9.311333333333335e-06, "loss": 2.4971, "step": 8100 }, { "epoch": 0.162, "eval_loss": 2.5067806243896484, "eval_runtime": 31.6662, "eval_samples_per_second": 3.347, "eval_steps_per_second": 0.853, "step": 8100 }, { "epoch": 0.1625, "grad_norm": 0.4639797537335715, "learning_rate": 9.305777777777779e-06, "loss": 2.4961, "step": 8125 }, { "epoch": 0.163, "grad_norm": 0.46286744081281633, "learning_rate": 9.300222222222222e-06, "loss": 2.491, "step": 8150 }, { "epoch": 0.1635, "grad_norm": 0.4921652771211608, "learning_rate": 9.294666666666668e-06, "loss": 2.4885, "step": 8175 }, { "epoch": 0.164, "grad_norm": 0.4593448087813921, "learning_rate": 9.289111111111113e-06, "loss": 2.4976, "step": 8200 }, { "epoch": 0.164, "eval_loss": 2.5061910152435303, "eval_runtime": 31.7395, "eval_samples_per_second": 3.34, "eval_steps_per_second": 0.851, "step": 8200 }, { "epoch": 0.1645, "grad_norm": 0.4714716672374406, "learning_rate": 9.283555555555556e-06, "loss": 2.4899, "step": 8225 }, { "epoch": 0.165, "grad_norm": 0.46793052489676595, "learning_rate": 9.278e-06, "loss": 2.4929, "step": 8250 }, { "epoch": 0.1655, "grad_norm": 0.488049297756729, "learning_rate": 9.272444444444445e-06, "loss": 2.4972, "step": 8275 }, { "epoch": 0.166, "grad_norm": 0.43609720972184396, "learning_rate": 9.26688888888889e-06, "loss": 2.4902, "step": 8300 }, { "epoch": 0.166, "eval_loss": 2.5061910152435303, "eval_runtime": 31.61, "eval_samples_per_second": 3.353, "eval_steps_per_second": 0.854, "step": 8300 }, { "epoch": 0.1665, "grad_norm": 0.51494306682902, "learning_rate": 9.261333333333334e-06, "loss": 2.5017, "step": 8325 }, { "epoch": 0.167, "grad_norm": 0.43687456366636745, "learning_rate": 9.25577777777778e-06, "loss": 2.4842, "step": 8350 }, { "epoch": 0.1675, "grad_norm": 0.4398983603701802, "learning_rate": 9.250222222222223e-06, "loss": 2.4883, "step": 8375 }, { "epoch": 0.168, "grad_norm": 0.44766592452770393, "learning_rate": 9.244666666666668e-06, "loss": 2.4924, "step": 8400 }, { "epoch": 0.168, "eval_loss": 2.5058963298797607, "eval_runtime": 31.6191, "eval_samples_per_second": 3.352, "eval_steps_per_second": 0.854, "step": 8400 }, { "epoch": 0.1685, "grad_norm": 0.5420691250593338, "learning_rate": 9.239111111111112e-06, "loss": 2.498, "step": 8425 }, { "epoch": 0.169, "grad_norm": 0.486342998144357, "learning_rate": 9.233555555555557e-06, "loss": 2.4963, "step": 8450 }, { "epoch": 0.1695, "grad_norm": 0.47737125328195895, "learning_rate": 9.228e-06, "loss": 2.4911, "step": 8475 }, { "epoch": 0.17, "grad_norm": 0.45349557813678815, "learning_rate": 9.222444444444446e-06, "loss": 2.4875, "step": 8500 }, { "epoch": 0.17, "eval_loss": 2.505601406097412, "eval_runtime": 31.689, "eval_samples_per_second": 3.345, "eval_steps_per_second": 0.852, "step": 8500 }, { "epoch": 0.1705, "grad_norm": 0.4796880606629406, "learning_rate": 9.21688888888889e-06, "loss": 2.497, "step": 8525 }, { "epoch": 0.171, "grad_norm": 0.47434802884790866, "learning_rate": 9.211333333333334e-06, "loss": 2.4868, "step": 8550 }, { "epoch": 0.1715, "grad_norm": 0.45594445468586775, "learning_rate": 9.20577777777778e-06, "loss": 2.4946, "step": 8575 }, { "epoch": 0.172, "grad_norm": 0.47122307827537663, "learning_rate": 9.200222222222223e-06, "loss": 2.4921, "step": 8600 }, { "epoch": 0.172, "eval_loss": 2.5053067207336426, "eval_runtime": 31.6154, "eval_samples_per_second": 3.353, "eval_steps_per_second": 0.854, "step": 8600 }, { "epoch": 0.1725, "grad_norm": 0.49346597298990835, "learning_rate": 9.194666666666667e-06, "loss": 2.4904, "step": 8625 }, { "epoch": 0.173, "grad_norm": 0.44191548376332535, "learning_rate": 9.189111111111112e-06, "loss": 2.4889, "step": 8650 }, { "epoch": 0.1735, "grad_norm": 0.4584560431283127, "learning_rate": 9.183555555555557e-06, "loss": 2.4946, "step": 8675 }, { "epoch": 0.174, "grad_norm": 0.4583103938160734, "learning_rate": 9.178000000000001e-06, "loss": 2.497, "step": 8700 }, { "epoch": 0.174, "eval_loss": 2.505011796951294, "eval_runtime": 33.293, "eval_samples_per_second": 3.184, "eval_steps_per_second": 0.811, "step": 8700 }, { "epoch": 0.1745, "grad_norm": 0.4282508677893212, "learning_rate": 9.172444444444444e-06, "loss": 2.4905, "step": 8725 }, { "epoch": 0.175, "grad_norm": 0.4799435951689765, "learning_rate": 9.16688888888889e-06, "loss": 2.4945, "step": 8750 }, { "epoch": 0.1755, "grad_norm": 0.47295693237440145, "learning_rate": 9.161333333333335e-06, "loss": 2.4872, "step": 8775 }, { "epoch": 0.176, "grad_norm": 0.4293829472868077, "learning_rate": 9.155777777777779e-06, "loss": 2.491, "step": 8800 }, { "epoch": 0.176, "eval_loss": 2.5047168731689453, "eval_runtime": 33.529, "eval_samples_per_second": 3.161, "eval_steps_per_second": 0.805, "step": 8800 }, { "epoch": 0.1765, "grad_norm": 0.4948989086561893, "learning_rate": 9.150222222222222e-06, "loss": 2.498, "step": 8825 }, { "epoch": 0.177, "grad_norm": 0.46905017505428037, "learning_rate": 9.144666666666667e-06, "loss": 2.4872, "step": 8850 }, { "epoch": 0.1775, "grad_norm": 0.4706274420773099, "learning_rate": 9.139111111111113e-06, "loss": 2.4879, "step": 8875 }, { "epoch": 0.178, "grad_norm": 0.439079741715738, "learning_rate": 9.133555555555556e-06, "loss": 2.4836, "step": 8900 }, { "epoch": 0.178, "eval_loss": 2.504422187805176, "eval_runtime": 31.6179, "eval_samples_per_second": 3.353, "eval_steps_per_second": 0.854, "step": 8900 }, { "epoch": 0.1785, "grad_norm": 0.5257446524841846, "learning_rate": 9.128e-06, "loss": 2.477, "step": 8925 }, { "epoch": 0.179, "grad_norm": 0.495973361699502, "learning_rate": 9.122444444444445e-06, "loss": 2.492, "step": 8950 }, { "epoch": 0.1795, "grad_norm": 0.44404234473403686, "learning_rate": 9.11688888888889e-06, "loss": 2.4966, "step": 8975 }, { "epoch": 0.18, "grad_norm": 0.48259834253773454, "learning_rate": 9.111333333333334e-06, "loss": 2.4999, "step": 9000 }, { "epoch": 0.18, "eval_loss": 2.5047168731689453, "eval_runtime": 31.5925, "eval_samples_per_second": 3.355, "eval_steps_per_second": 0.855, "step": 9000 }, { "epoch": 0.1805, "grad_norm": 0.4421775299979092, "learning_rate": 9.105777777777779e-06, "loss": 2.5015, "step": 9025 }, { "epoch": 0.181, "grad_norm": 0.4634033140014565, "learning_rate": 9.100222222222223e-06, "loss": 2.5039, "step": 9050 }, { "epoch": 0.1815, "grad_norm": 0.46159115578461074, "learning_rate": 9.094666666666668e-06, "loss": 2.5008, "step": 9075 }, { "epoch": 0.182, "grad_norm": 0.449669813485736, "learning_rate": 9.089111111111111e-06, "loss": 2.5009, "step": 9100 }, { "epoch": 0.182, "eval_loss": 2.504422187805176, "eval_runtime": 31.7112, "eval_samples_per_second": 3.343, "eval_steps_per_second": 0.851, "step": 9100 }, { "epoch": 0.1825, "grad_norm": 0.4624450881526358, "learning_rate": 9.083555555555557e-06, "loss": 2.4878, "step": 9125 }, { "epoch": 0.183, "grad_norm": 0.4759040871891137, "learning_rate": 9.078000000000002e-06, "loss": 2.4949, "step": 9150 }, { "epoch": 0.1835, "grad_norm": 0.4565317959105483, "learning_rate": 9.072444444444445e-06, "loss": 2.5008, "step": 9175 }, { "epoch": 0.184, "grad_norm": 0.6027882705089656, "learning_rate": 9.066888888888889e-06, "loss": 2.4831, "step": 9200 }, { "epoch": 0.184, "eval_loss": 2.5038325786590576, "eval_runtime": 31.8272, "eval_samples_per_second": 3.33, "eval_steps_per_second": 0.848, "step": 9200 }, { "epoch": 0.1845, "grad_norm": 0.4672088128196563, "learning_rate": 9.061333333333334e-06, "loss": 2.4929, "step": 9225 }, { "epoch": 0.185, "grad_norm": 0.425744444651012, "learning_rate": 9.05577777777778e-06, "loss": 2.4979, "step": 9250 }, { "epoch": 0.1855, "grad_norm": 0.4127999381509735, "learning_rate": 9.050222222222223e-06, "loss": 2.5065, "step": 9275 }, { "epoch": 0.186, "grad_norm": 0.45683795031899443, "learning_rate": 9.044666666666667e-06, "loss": 2.4971, "step": 9300 }, { "epoch": 0.186, "eval_loss": 2.504127264022827, "eval_runtime": 31.6049, "eval_samples_per_second": 3.354, "eval_steps_per_second": 0.854, "step": 9300 }, { "epoch": 0.1865, "grad_norm": 0.47360766082322153, "learning_rate": 9.039111111111112e-06, "loss": 2.491, "step": 9325 }, { "epoch": 0.187, "grad_norm": 0.5526247307925498, "learning_rate": 9.033555555555557e-06, "loss": 2.4934, "step": 9350 }, { "epoch": 0.1875, "grad_norm": 0.4670755026657148, "learning_rate": 9.028e-06, "loss": 2.5025, "step": 9375 }, { "epoch": 0.188, "grad_norm": 0.45668031478164683, "learning_rate": 9.022444444444444e-06, "loss": 2.4892, "step": 9400 }, { "epoch": 0.188, "eval_loss": 2.5038325786590576, "eval_runtime": 31.8115, "eval_samples_per_second": 3.332, "eval_steps_per_second": 0.849, "step": 9400 }, { "epoch": 0.1885, "grad_norm": 0.4393528481721735, "learning_rate": 9.01688888888889e-06, "loss": 2.4993, "step": 9425 }, { "epoch": 0.189, "grad_norm": 0.49694977951670527, "learning_rate": 9.011333333333335e-06, "loss": 2.486, "step": 9450 }, { "epoch": 0.1895, "grad_norm": 0.45862304441087315, "learning_rate": 9.005777777777778e-06, "loss": 2.484, "step": 9475 }, { "epoch": 0.19, "grad_norm": 0.4403368479751325, "learning_rate": 9.000222222222222e-06, "loss": 2.4839, "step": 9500 }, { "epoch": 0.19, "eval_loss": 2.5032429695129395, "eval_runtime": 31.7192, "eval_samples_per_second": 3.342, "eval_steps_per_second": 0.851, "step": 9500 }, { "epoch": 0.1905, "grad_norm": 0.48763976397960673, "learning_rate": 8.994666666666667e-06, "loss": 2.4946, "step": 9525 }, { "epoch": 0.191, "grad_norm": 0.46329003827956744, "learning_rate": 8.989111111111112e-06, "loss": 2.5064, "step": 9550 }, { "epoch": 0.1915, "grad_norm": 0.3980817630959744, "learning_rate": 8.983555555555556e-06, "loss": 2.4932, "step": 9575 }, { "epoch": 0.192, "grad_norm": 0.5009684981679573, "learning_rate": 8.978000000000001e-06, "loss": 2.4963, "step": 9600 }, { "epoch": 0.192, "eval_loss": 2.503537654876709, "eval_runtime": 31.5776, "eval_samples_per_second": 3.357, "eval_steps_per_second": 0.855, "step": 9600 }, { "epoch": 0.1925, "grad_norm": 0.43627846543007903, "learning_rate": 8.972444444444445e-06, "loss": 2.4799, "step": 9625 }, { "epoch": 0.193, "grad_norm": 0.4341220630570628, "learning_rate": 8.96688888888889e-06, "loss": 2.4939, "step": 9650 }, { "epoch": 0.1935, "grad_norm": 0.5248960220872864, "learning_rate": 8.961333333333333e-06, "loss": 2.4983, "step": 9675 }, { "epoch": 0.194, "grad_norm": 0.5205641262868196, "learning_rate": 8.955777777777779e-06, "loss": 2.4992, "step": 9700 }, { "epoch": 0.194, "eval_loss": 2.5032429695129395, "eval_runtime": 31.6005, "eval_samples_per_second": 3.354, "eval_steps_per_second": 0.854, "step": 9700 }, { "epoch": 0.1945, "grad_norm": 0.42707875009123103, "learning_rate": 8.950222222222224e-06, "loss": 2.5017, "step": 9725 }, { "epoch": 0.195, "grad_norm": 0.4648079275311894, "learning_rate": 8.944666666666668e-06, "loss": 2.4998, "step": 9750 }, { "epoch": 0.1955, "grad_norm": 0.43005920693607586, "learning_rate": 8.939111111111111e-06, "loss": 2.4971, "step": 9775 }, { "epoch": 0.196, "grad_norm": 0.45585903163769387, "learning_rate": 8.933555555555556e-06, "loss": 2.493, "step": 9800 }, { "epoch": 0.196, "eval_loss": 2.502948045730591, "eval_runtime": 31.632, "eval_samples_per_second": 3.351, "eval_steps_per_second": 0.854, "step": 9800 }, { "epoch": 0.1965, "grad_norm": 0.4266057927636531, "learning_rate": 8.928000000000002e-06, "loss": 2.5011, "step": 9825 }, { "epoch": 0.197, "grad_norm": 0.43778187408753183, "learning_rate": 8.922444444444445e-06, "loss": 2.4951, "step": 9850 }, { "epoch": 0.1975, "grad_norm": 0.5338200834169559, "learning_rate": 8.916888888888889e-06, "loss": 2.4896, "step": 9875 }, { "epoch": 0.198, "grad_norm": 0.46823595750617164, "learning_rate": 8.911333333333334e-06, "loss": 2.4845, "step": 9900 }, { "epoch": 0.198, "eval_loss": 2.5023584365844727, "eval_runtime": 31.6857, "eval_samples_per_second": 3.345, "eval_steps_per_second": 0.852, "step": 9900 }, { "epoch": 0.1985, "grad_norm": 0.4570575398172366, "learning_rate": 8.90577777777778e-06, "loss": 2.4864, "step": 9925 }, { "epoch": 0.199, "grad_norm": 0.42166301716080085, "learning_rate": 8.900222222222223e-06, "loss": 2.4931, "step": 9950 }, { "epoch": 0.1995, "grad_norm": 0.47358354147378096, "learning_rate": 8.894666666666666e-06, "loss": 2.4953, "step": 9975 }, { "epoch": 0.2, "grad_norm": 0.4390332031938933, "learning_rate": 8.889111111111112e-06, "loss": 2.4855, "step": 10000 }, { "epoch": 0.2, "eval_loss": 2.502948045730591, "eval_runtime": 31.6124, "eval_samples_per_second": 3.353, "eval_steps_per_second": 0.854, "step": 10000 }, { "epoch": 0.2005, "grad_norm": 0.4366599313768622, "learning_rate": 8.883555555555557e-06, "loss": 2.5065, "step": 10025 }, { "epoch": 0.201, "grad_norm": 0.41140266099356065, "learning_rate": 8.878e-06, "loss": 2.5008, "step": 10050 }, { "epoch": 0.2015, "grad_norm": 0.43905042554021645, "learning_rate": 8.872444444444444e-06, "loss": 2.4883, "step": 10075 }, { "epoch": 0.202, "grad_norm": 0.43926747314424747, "learning_rate": 8.86688888888889e-06, "loss": 2.4858, "step": 10100 }, { "epoch": 0.202, "eval_loss": 2.5026533603668213, "eval_runtime": 33.7529, "eval_samples_per_second": 3.14, "eval_steps_per_second": 0.8, "step": 10100 }, { "epoch": 0.2025, "grad_norm": 0.4573967588072801, "learning_rate": 8.861333333333334e-06, "loss": 2.4946, "step": 10125 }, { "epoch": 0.203, "grad_norm": 0.44102158236984695, "learning_rate": 8.855777777777778e-06, "loss": 2.4959, "step": 10150 }, { "epoch": 0.2035, "grad_norm": 0.4244034611509627, "learning_rate": 8.850222222222223e-06, "loss": 2.4882, "step": 10175 }, { "epoch": 0.204, "grad_norm": 0.4292353602313773, "learning_rate": 8.844666666666667e-06, "loss": 2.475, "step": 10200 }, { "epoch": 0.204, "eval_loss": 2.5026533603668213, "eval_runtime": 33.6895, "eval_samples_per_second": 3.146, "eval_steps_per_second": 0.801, "step": 10200 }, { "epoch": 0.2045, "grad_norm": 0.43537766637284453, "learning_rate": 8.839111111111112e-06, "loss": 2.4919, "step": 10225 }, { "epoch": 0.205, "grad_norm": 0.42814122374245395, "learning_rate": 8.833555555555556e-06, "loss": 2.4937, "step": 10250 }, { "epoch": 0.2055, "grad_norm": 0.4112227358131473, "learning_rate": 8.828000000000001e-06, "loss": 2.4898, "step": 10275 }, { "epoch": 0.206, "grad_norm": 0.4351870367519067, "learning_rate": 8.822444444444446e-06, "loss": 2.4881, "step": 10300 }, { "epoch": 0.206, "eval_loss": 2.5023584365844727, "eval_runtime": 33.435, "eval_samples_per_second": 3.17, "eval_steps_per_second": 0.808, "step": 10300 }, { "epoch": 0.2065, "grad_norm": 0.4263308787171397, "learning_rate": 8.81688888888889e-06, "loss": 2.475, "step": 10325 }, { "epoch": 0.207, "grad_norm": 0.45579928417932, "learning_rate": 8.811333333333333e-06, "loss": 2.4943, "step": 10350 }, { "epoch": 0.2075, "grad_norm": 0.4280704481397019, "learning_rate": 8.805777777777778e-06, "loss": 2.4845, "step": 10375 }, { "epoch": 0.208, "grad_norm": 0.4670479060805604, "learning_rate": 8.800222222222224e-06, "loss": 2.4903, "step": 10400 }, { "epoch": 0.208, "eval_loss": 2.5023584365844727, "eval_runtime": 31.6522, "eval_samples_per_second": 3.349, "eval_steps_per_second": 0.853, "step": 10400 }, { "epoch": 0.2085, "grad_norm": 0.43912910279855394, "learning_rate": 8.794666666666667e-06, "loss": 2.4975, "step": 10425 }, { "epoch": 0.209, "grad_norm": 0.4494490825084985, "learning_rate": 8.78911111111111e-06, "loss": 2.4879, "step": 10450 }, { "epoch": 0.2095, "grad_norm": 0.42491374212906596, "learning_rate": 8.783555555555556e-06, "loss": 2.4964, "step": 10475 }, { "epoch": 0.21, "grad_norm": 0.39781460942112506, "learning_rate": 8.778000000000001e-06, "loss": 2.4927, "step": 10500 }, { "epoch": 0.21, "eval_loss": 2.502063751220703, "eval_runtime": 31.7621, "eval_samples_per_second": 3.337, "eval_steps_per_second": 0.85, "step": 10500 }, { "epoch": 0.2105, "grad_norm": 0.43424351284899937, "learning_rate": 8.772444444444445e-06, "loss": 2.4928, "step": 10525 }, { "epoch": 0.211, "grad_norm": 0.4295006342308409, "learning_rate": 8.766888888888888e-06, "loss": 2.487, "step": 10550 }, { "epoch": 0.2115, "grad_norm": 0.4237014246862784, "learning_rate": 8.761333333333334e-06, "loss": 2.4897, "step": 10575 }, { "epoch": 0.212, "grad_norm": 0.45856730861474515, "learning_rate": 8.755777777777779e-06, "loss": 2.4928, "step": 10600 }, { "epoch": 0.212, "eval_loss": 2.502063751220703, "eval_runtime": 31.7252, "eval_samples_per_second": 3.341, "eval_steps_per_second": 0.851, "step": 10600 }, { "epoch": 0.2125, "grad_norm": 0.41191624399106863, "learning_rate": 8.750222222222223e-06, "loss": 2.4876, "step": 10625 }, { "epoch": 0.213, "grad_norm": 0.46542602416489315, "learning_rate": 8.744666666666666e-06, "loss": 2.5034, "step": 10650 }, { "epoch": 0.2135, "grad_norm": 0.7915220564442529, "learning_rate": 8.739111111111111e-06, "loss": 2.4834, "step": 10675 }, { "epoch": 0.214, "grad_norm": 0.42259767066868636, "learning_rate": 8.733555555555557e-06, "loss": 2.4854, "step": 10700 }, { "epoch": 0.214, "eval_loss": 2.502063751220703, "eval_runtime": 31.7022, "eval_samples_per_second": 3.344, "eval_steps_per_second": 0.852, "step": 10700 }, { "epoch": 0.2145, "grad_norm": 0.41123050537861, "learning_rate": 8.728e-06, "loss": 2.4938, "step": 10725 }, { "epoch": 0.215, "grad_norm": 0.45525972565172595, "learning_rate": 8.722444444444445e-06, "loss": 2.4852, "step": 10750 }, { "epoch": 0.2155, "grad_norm": 0.41607212965523915, "learning_rate": 8.716888888888889e-06, "loss": 2.474, "step": 10775 }, { "epoch": 0.216, "grad_norm": 0.39792115019783625, "learning_rate": 8.711333333333334e-06, "loss": 2.5009, "step": 10800 }, { "epoch": 0.216, "eval_loss": 2.502063751220703, "eval_runtime": 31.7144, "eval_samples_per_second": 3.342, "eval_steps_per_second": 0.851, "step": 10800 }, { "epoch": 0.2165, "grad_norm": 0.4585655221780486, "learning_rate": 8.705777777777778e-06, "loss": 2.4916, "step": 10825 }, { "epoch": 0.217, "grad_norm": 0.4080298748776425, "learning_rate": 8.700222222222223e-06, "loss": 2.4902, "step": 10850 }, { "epoch": 0.2175, "grad_norm": 0.46957233489171346, "learning_rate": 8.694666666666668e-06, "loss": 2.4822, "step": 10875 }, { "epoch": 0.218, "grad_norm": 0.41478691206301865, "learning_rate": 8.689111111111112e-06, "loss": 2.4824, "step": 10900 }, { "epoch": 0.218, "eval_loss": 2.502063751220703, "eval_runtime": 31.8439, "eval_samples_per_second": 3.329, "eval_steps_per_second": 0.848, "step": 10900 }, { "epoch": 0.2185, "grad_norm": 0.42928030409496243, "learning_rate": 8.683555555555555e-06, "loss": 2.4853, "step": 10925 }, { "epoch": 0.219, "grad_norm": 0.41463223470042737, "learning_rate": 8.678e-06, "loss": 2.4804, "step": 10950 }, { "epoch": 0.2195, "grad_norm": 0.42126484160991745, "learning_rate": 8.672444444444446e-06, "loss": 2.489, "step": 10975 }, { "epoch": 0.22, "grad_norm": 0.4431696815658413, "learning_rate": 8.66688888888889e-06, "loss": 2.4897, "step": 11000 }, { "epoch": 0.22, "eval_loss": 2.502063751220703, "eval_runtime": 31.7255, "eval_samples_per_second": 3.341, "eval_steps_per_second": 0.851, "step": 11000 } ], "logging_steps": 25, "max_steps": 50000, "num_input_tokens_seen": 0, "num_train_epochs": 9223372036854775807, "save_steps": 500, "stateful_callbacks": { "TrainerControl": { "args": { "should_epoch_stop": false, "should_evaluate": false, "should_log": false, "should_save": true, "should_training_stop": false }, "attributes": {} } }, "total_flos": 1.1301896471634248e+19, "train_batch_size": 2, "trial_name": null, "trial_params": null }