| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 7.168458781362007, | |
| "eval_steps": 2000, | |
| "global_step": 10000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.07, | |
| "grad_norm": 2.5333876609802246, | |
| "learning_rate": 9.9e-07, | |
| "loss": 0.9981, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.14, | |
| "grad_norm": 2.1556735038757324, | |
| "learning_rate": 9.9e-07, | |
| "loss": 0.7461, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.22, | |
| "grad_norm": 2.5188281536102295, | |
| "learning_rate": 9.7989898989899e-07, | |
| "loss": 0.6707, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "grad_norm": 2.559631824493408, | |
| "learning_rate": 9.697979797979798e-07, | |
| "loss": 0.6281, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.36, | |
| "grad_norm": 3.3940675258636475, | |
| "learning_rate": 9.596969696969696e-07, | |
| "loss": 0.6437, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.43, | |
| "grad_norm": 2.0749378204345703, | |
| "learning_rate": 9.495959595959595e-07, | |
| "loss": 0.6319, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "grad_norm": 2.3027114868164062, | |
| "learning_rate": 9.394949494949495e-07, | |
| "loss": 0.6477, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.57, | |
| "grad_norm": 1.6429717540740967, | |
| "learning_rate": 9.293939393939394e-07, | |
| "loss": 0.6564, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.65, | |
| "grad_norm": 2.391716718673706, | |
| "learning_rate": 9.192929292929292e-07, | |
| "loss": 0.6263, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.72, | |
| "grad_norm": 2.9794795513153076, | |
| "learning_rate": 9.091919191919192e-07, | |
| "loss": 0.6422, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.79, | |
| "grad_norm": 2.853337049484253, | |
| "learning_rate": 8.99090909090909e-07, | |
| "loss": 0.6043, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 0.86, | |
| "grad_norm": 2.453305244445801, | |
| "learning_rate": 8.88989898989899e-07, | |
| "loss": 0.5914, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 0.93, | |
| "grad_norm": 2.768860340118408, | |
| "learning_rate": 8.788888888888889e-07, | |
| "loss": 0.6194, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 4.065279483795166, | |
| "learning_rate": 8.687878787878787e-07, | |
| "loss": 0.6038, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 1.08, | |
| "grad_norm": 3.8275232315063477, | |
| "learning_rate": 8.586868686868687e-07, | |
| "loss": 0.6038, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 1.15, | |
| "grad_norm": 2.6090586185455322, | |
| "learning_rate": 8.485858585858586e-07, | |
| "loss": 0.592, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 1.22, | |
| "grad_norm": 1.9870526790618896, | |
| "learning_rate": 8.384848484848484e-07, | |
| "loss": 0.5885, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 1.29, | |
| "grad_norm": 2.2942471504211426, | |
| "learning_rate": 8.283838383838383e-07, | |
| "loss": 0.622, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 1.36, | |
| "grad_norm": 2.0904691219329834, | |
| "learning_rate": 8.182828282828283e-07, | |
| "loss": 0.5857, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 1.43, | |
| "grad_norm": 2.909604072570801, | |
| "learning_rate": 8.081818181818182e-07, | |
| "loss": 0.5841, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 1.43, | |
| "eval_loss": 0.7236512303352356, | |
| "eval_runtime": 192.8239, | |
| "eval_samples_per_second": 5.186, | |
| "eval_steps_per_second": 0.648, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 1.51, | |
| "grad_norm": 2.645550012588501, | |
| "learning_rate": 7.98080808080808e-07, | |
| "loss": 0.6013, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 1.58, | |
| "grad_norm": 2.5540435314178467, | |
| "learning_rate": 7.88080808080808e-07, | |
| "loss": 0.6049, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 1.65, | |
| "grad_norm": 2.3379604816436768, | |
| "learning_rate": 7.779797979797979e-07, | |
| "loss": 0.5794, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 1.72, | |
| "grad_norm": 2.4884698390960693, | |
| "learning_rate": 7.678787878787878e-07, | |
| "loss": 0.5832, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 1.79, | |
| "grad_norm": 2.4760003089904785, | |
| "learning_rate": 7.577777777777778e-07, | |
| "loss": 0.5852, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 1.86, | |
| "grad_norm": 2.334510087966919, | |
| "learning_rate": 7.476767676767676e-07, | |
| "loss": 0.5825, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 1.94, | |
| "grad_norm": 3.3062024116516113, | |
| "learning_rate": 7.375757575757575e-07, | |
| "loss": 0.5739, | |
| "step": 2700 | |
| }, | |
| { | |
| "epoch": 2.01, | |
| "grad_norm": 3.0897557735443115, | |
| "learning_rate": 7.274747474747475e-07, | |
| "loss": 0.5823, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 2.08, | |
| "grad_norm": 3.0633504390716553, | |
| "learning_rate": 7.173737373737373e-07, | |
| "loss": 0.5625, | |
| "step": 2900 | |
| }, | |
| { | |
| "epoch": 2.15, | |
| "grad_norm": 3.3426527976989746, | |
| "learning_rate": 7.072727272727272e-07, | |
| "loss": 0.5484, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 2.22, | |
| "grad_norm": 2.7486729621887207, | |
| "learning_rate": 6.971717171717172e-07, | |
| "loss": 0.5853, | |
| "step": 3100 | |
| }, | |
| { | |
| "epoch": 2.29, | |
| "grad_norm": 2.5889084339141846, | |
| "learning_rate": 6.87070707070707e-07, | |
| "loss": 0.568, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 2.37, | |
| "grad_norm": 2.474393606185913, | |
| "learning_rate": 6.76969696969697e-07, | |
| "loss": 0.5699, | |
| "step": 3300 | |
| }, | |
| { | |
| "epoch": 2.44, | |
| "grad_norm": 2.99786114692688, | |
| "learning_rate": 6.668686868686869e-07, | |
| "loss": 0.5723, | |
| "step": 3400 | |
| }, | |
| { | |
| "epoch": 2.51, | |
| "grad_norm": 2.888652801513672, | |
| "learning_rate": 6.567676767676767e-07, | |
| "loss": 0.5942, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 2.58, | |
| "grad_norm": 3.20930552482605, | |
| "learning_rate": 6.466666666666666e-07, | |
| "loss": 0.5827, | |
| "step": 3600 | |
| }, | |
| { | |
| "epoch": 2.65, | |
| "grad_norm": 2.312199592590332, | |
| "learning_rate": 6.365656565656565e-07, | |
| "loss": 0.5644, | |
| "step": 3700 | |
| }, | |
| { | |
| "epoch": 2.72, | |
| "grad_norm": 3.1894469261169434, | |
| "learning_rate": 6.264646464646465e-07, | |
| "loss": 0.5409, | |
| "step": 3800 | |
| }, | |
| { | |
| "epoch": 2.8, | |
| "grad_norm": 2.740865468978882, | |
| "learning_rate": 6.163636363636363e-07, | |
| "loss": 0.5518, | |
| "step": 3900 | |
| }, | |
| { | |
| "epoch": 2.87, | |
| "grad_norm": 4.003296375274658, | |
| "learning_rate": 6.062626262626262e-07, | |
| "loss": 0.5796, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 2.87, | |
| "eval_loss": 0.7056564688682556, | |
| "eval_runtime": 192.8797, | |
| "eval_samples_per_second": 5.185, | |
| "eval_steps_per_second": 0.648, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 2.94, | |
| "grad_norm": 3.06562876701355, | |
| "learning_rate": 5.961616161616161e-07, | |
| "loss": 0.58, | |
| "step": 4100 | |
| }, | |
| { | |
| "epoch": 3.01, | |
| "grad_norm": 8.636664390563965, | |
| "learning_rate": 5.860606060606061e-07, | |
| "loss": 0.552, | |
| "step": 4200 | |
| }, | |
| { | |
| "epoch": 3.08, | |
| "grad_norm": 3.2252180576324463, | |
| "learning_rate": 5.759595959595959e-07, | |
| "loss": 0.5515, | |
| "step": 4300 | |
| }, | |
| { | |
| "epoch": 3.15, | |
| "grad_norm": 3.46083664894104, | |
| "learning_rate": 5.658585858585858e-07, | |
| "loss": 0.55, | |
| "step": 4400 | |
| }, | |
| { | |
| "epoch": 3.23, | |
| "grad_norm": 2.788499116897583, | |
| "learning_rate": 5.558585858585858e-07, | |
| "loss": 0.561, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 3.3, | |
| "grad_norm": 4.106410026550293, | |
| "learning_rate": 5.457575757575758e-07, | |
| "loss": 0.546, | |
| "step": 4600 | |
| }, | |
| { | |
| "epoch": 3.37, | |
| "grad_norm": 2.6162779331207275, | |
| "learning_rate": 5.356565656565656e-07, | |
| "loss": 0.556, | |
| "step": 4700 | |
| }, | |
| { | |
| "epoch": 3.44, | |
| "grad_norm": 3.473173141479492, | |
| "learning_rate": 5.255555555555555e-07, | |
| "loss": 0.552, | |
| "step": 4800 | |
| }, | |
| { | |
| "epoch": 3.51, | |
| "grad_norm": 3.0796926021575928, | |
| "learning_rate": 5.154545454545455e-07, | |
| "loss": 0.538, | |
| "step": 4900 | |
| }, | |
| { | |
| "epoch": 3.58, | |
| "grad_norm": 2.6169357299804688, | |
| "learning_rate": 5.053535353535353e-07, | |
| "loss": 0.5527, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 3.66, | |
| "grad_norm": 2.964797019958496, | |
| "learning_rate": 4.952525252525253e-07, | |
| "loss": 0.5603, | |
| "step": 5100 | |
| }, | |
| { | |
| "epoch": 3.73, | |
| "grad_norm": 3.1007957458496094, | |
| "learning_rate": 4.851515151515152e-07, | |
| "loss": 0.5326, | |
| "step": 5200 | |
| }, | |
| { | |
| "epoch": 3.8, | |
| "grad_norm": 2.996995449066162, | |
| "learning_rate": 4.75050505050505e-07, | |
| "loss": 0.5446, | |
| "step": 5300 | |
| }, | |
| { | |
| "epoch": 3.87, | |
| "grad_norm": 3.0833990573883057, | |
| "learning_rate": 4.6494949494949493e-07, | |
| "loss": 0.5612, | |
| "step": 5400 | |
| }, | |
| { | |
| "epoch": 3.94, | |
| "grad_norm": 3.361348867416382, | |
| "learning_rate": 4.548484848484848e-07, | |
| "loss": 0.5771, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 4.01, | |
| "grad_norm": 3.4631500244140625, | |
| "learning_rate": 4.447474747474747e-07, | |
| "loss": 0.5531, | |
| "step": 5600 | |
| }, | |
| { | |
| "epoch": 4.09, | |
| "grad_norm": 3.6133594512939453, | |
| "learning_rate": 4.3464646464646463e-07, | |
| "loss": 0.529, | |
| "step": 5700 | |
| }, | |
| { | |
| "epoch": 4.16, | |
| "grad_norm": 3.092081308364868, | |
| "learning_rate": 4.2454545454545454e-07, | |
| "loss": 0.5346, | |
| "step": 5800 | |
| }, | |
| { | |
| "epoch": 4.23, | |
| "grad_norm": 3.862914800643921, | |
| "learning_rate": 4.144444444444444e-07, | |
| "loss": 0.5271, | |
| "step": 5900 | |
| }, | |
| { | |
| "epoch": 4.3, | |
| "grad_norm": 3.39168643951416, | |
| "learning_rate": 4.043434343434343e-07, | |
| "loss": 0.5466, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 4.3, | |
| "eval_loss": 0.7162824869155884, | |
| "eval_runtime": 192.3954, | |
| "eval_samples_per_second": 5.198, | |
| "eval_steps_per_second": 0.65, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 4.37, | |
| "grad_norm": 2.3356008529663086, | |
| "learning_rate": 3.942424242424242e-07, | |
| "loss": 0.5182, | |
| "step": 6100 | |
| }, | |
| { | |
| "epoch": 4.44, | |
| "grad_norm": 4.625268459320068, | |
| "learning_rate": 3.841414141414141e-07, | |
| "loss": 0.5124, | |
| "step": 6200 | |
| }, | |
| { | |
| "epoch": 4.52, | |
| "grad_norm": 3.602734327316284, | |
| "learning_rate": 3.7404040404040407e-07, | |
| "loss": 0.5356, | |
| "step": 6300 | |
| }, | |
| { | |
| "epoch": 4.59, | |
| "grad_norm": 3.235563039779663, | |
| "learning_rate": 3.6393939393939393e-07, | |
| "loss": 0.5481, | |
| "step": 6400 | |
| }, | |
| { | |
| "epoch": 4.66, | |
| "grad_norm": 3.244750738143921, | |
| "learning_rate": 3.539393939393939e-07, | |
| "loss": 0.5375, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 4.73, | |
| "grad_norm": 4.223140716552734, | |
| "learning_rate": 3.4383838383838384e-07, | |
| "loss": 0.5418, | |
| "step": 6600 | |
| }, | |
| { | |
| "epoch": 4.8, | |
| "grad_norm": 3.9079456329345703, | |
| "learning_rate": 3.337373737373737e-07, | |
| "loss": 0.5575, | |
| "step": 6700 | |
| }, | |
| { | |
| "epoch": 4.87, | |
| "grad_norm": 3.0635576248168945, | |
| "learning_rate": 3.236363636363636e-07, | |
| "loss": 0.5314, | |
| "step": 6800 | |
| }, | |
| { | |
| "epoch": 4.95, | |
| "grad_norm": 3.8982629776000977, | |
| "learning_rate": 3.1353535353535353e-07, | |
| "loss": 0.5493, | |
| "step": 6900 | |
| }, | |
| { | |
| "epoch": 5.02, | |
| "grad_norm": 3.750946521759033, | |
| "learning_rate": 3.0343434343434345e-07, | |
| "loss": 0.5356, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 5.09, | |
| "grad_norm": 3.491605520248413, | |
| "learning_rate": 2.933333333333333e-07, | |
| "loss": 0.5302, | |
| "step": 7100 | |
| }, | |
| { | |
| "epoch": 5.16, | |
| "grad_norm": 2.702629566192627, | |
| "learning_rate": 2.8323232323232323e-07, | |
| "loss": 0.5144, | |
| "step": 7200 | |
| }, | |
| { | |
| "epoch": 5.23, | |
| "grad_norm": 4.217553615570068, | |
| "learning_rate": 2.731313131313131e-07, | |
| "loss": 0.5309, | |
| "step": 7300 | |
| }, | |
| { | |
| "epoch": 5.3, | |
| "grad_norm": 2.7974092960357666, | |
| "learning_rate": 2.63030303030303e-07, | |
| "loss": 0.519, | |
| "step": 7400 | |
| }, | |
| { | |
| "epoch": 5.38, | |
| "grad_norm": 4.0323686599731445, | |
| "learning_rate": 2.529292929292929e-07, | |
| "loss": 0.5513, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 5.45, | |
| "grad_norm": 4.013027191162109, | |
| "learning_rate": 2.4282828282828284e-07, | |
| "loss": 0.5331, | |
| "step": 7600 | |
| }, | |
| { | |
| "epoch": 5.52, | |
| "grad_norm": 3.470280647277832, | |
| "learning_rate": 2.3272727272727273e-07, | |
| "loss": 0.5252, | |
| "step": 7700 | |
| }, | |
| { | |
| "epoch": 5.59, | |
| "grad_norm": 3.5019471645355225, | |
| "learning_rate": 2.2262626262626262e-07, | |
| "loss": 0.5203, | |
| "step": 7800 | |
| }, | |
| { | |
| "epoch": 5.66, | |
| "grad_norm": 3.3975048065185547, | |
| "learning_rate": 2.1252525252525254e-07, | |
| "loss": 0.5255, | |
| "step": 7900 | |
| }, | |
| { | |
| "epoch": 5.73, | |
| "grad_norm": 3.226830244064331, | |
| "learning_rate": 2.0242424242424242e-07, | |
| "loss": 0.5264, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 5.73, | |
| "eval_loss": 0.7162704467773438, | |
| "eval_runtime": 209.7748, | |
| "eval_samples_per_second": 4.767, | |
| "eval_steps_per_second": 0.596, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 5.81, | |
| "grad_norm": 4.459393501281738, | |
| "learning_rate": 1.9232323232323231e-07, | |
| "loss": 0.5171, | |
| "step": 8100 | |
| }, | |
| { | |
| "epoch": 5.88, | |
| "grad_norm": 3.706793785095215, | |
| "learning_rate": 1.8222222222222223e-07, | |
| "loss": 0.5259, | |
| "step": 8200 | |
| }, | |
| { | |
| "epoch": 5.95, | |
| "grad_norm": 2.9756672382354736, | |
| "learning_rate": 1.7212121212121212e-07, | |
| "loss": 0.5112, | |
| "step": 8300 | |
| }, | |
| { | |
| "epoch": 6.02, | |
| "grad_norm": 3.595360517501831, | |
| "learning_rate": 1.62020202020202e-07, | |
| "loss": 0.5353, | |
| "step": 8400 | |
| }, | |
| { | |
| "epoch": 6.09, | |
| "grad_norm": 4.781970977783203, | |
| "learning_rate": 1.5191919191919193e-07, | |
| "loss": 0.5223, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 6.16, | |
| "grad_norm": 3.7664666175842285, | |
| "learning_rate": 1.4191919191919192e-07, | |
| "loss": 0.5084, | |
| "step": 8600 | |
| }, | |
| { | |
| "epoch": 6.24, | |
| "grad_norm": 3.0045299530029297, | |
| "learning_rate": 1.318181818181818e-07, | |
| "loss": 0.5109, | |
| "step": 8700 | |
| }, | |
| { | |
| "epoch": 6.31, | |
| "grad_norm": 4.879513740539551, | |
| "learning_rate": 1.2171717171717172e-07, | |
| "loss": 0.5252, | |
| "step": 8800 | |
| }, | |
| { | |
| "epoch": 6.38, | |
| "grad_norm": 3.5674266815185547, | |
| "learning_rate": 1.1161616161616161e-07, | |
| "loss": 0.5164, | |
| "step": 8900 | |
| }, | |
| { | |
| "epoch": 6.45, | |
| "grad_norm": 4.9154887199401855, | |
| "learning_rate": 1.0151515151515151e-07, | |
| "loss": 0.5171, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 6.52, | |
| "grad_norm": 5.594084739685059, | |
| "learning_rate": 9.141414141414142e-08, | |
| "loss": 0.5178, | |
| "step": 9100 | |
| }, | |
| { | |
| "epoch": 6.59, | |
| "grad_norm": 5.077495574951172, | |
| "learning_rate": 8.13131313131313e-08, | |
| "loss": 0.5043, | |
| "step": 9200 | |
| }, | |
| { | |
| "epoch": 6.67, | |
| "grad_norm": 4.188448429107666, | |
| "learning_rate": 7.121212121212121e-08, | |
| "loss": 0.5152, | |
| "step": 9300 | |
| }, | |
| { | |
| "epoch": 6.74, | |
| "grad_norm": 3.8687632083892822, | |
| "learning_rate": 6.111111111111111e-08, | |
| "loss": 0.5132, | |
| "step": 9400 | |
| }, | |
| { | |
| "epoch": 6.81, | |
| "grad_norm": 4.494914531707764, | |
| "learning_rate": 5.101010101010101e-08, | |
| "loss": 0.5068, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 6.88, | |
| "grad_norm": 6.755849361419678, | |
| "learning_rate": 4.090909090909091e-08, | |
| "loss": 0.5168, | |
| "step": 9600 | |
| }, | |
| { | |
| "epoch": 6.95, | |
| "grad_norm": 4.19561767578125, | |
| "learning_rate": 3.0808080808080806e-08, | |
| "loss": 0.5308, | |
| "step": 9700 | |
| }, | |
| { | |
| "epoch": 7.03, | |
| "grad_norm": 11.913464546203613, | |
| "learning_rate": 2.070707070707071e-08, | |
| "loss": 0.5295, | |
| "step": 9800 | |
| }, | |
| { | |
| "epoch": 7.1, | |
| "grad_norm": 4.664416313171387, | |
| "learning_rate": 1.0606060606060607e-08, | |
| "loss": 0.5051, | |
| "step": 9900 | |
| }, | |
| { | |
| "epoch": 7.17, | |
| "grad_norm": 4.764692783355713, | |
| "learning_rate": 5.050505050505051e-10, | |
| "loss": 0.5101, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 7.17, | |
| "eval_loss": 0.7244647145271301, | |
| "eval_runtime": 192.8667, | |
| "eval_samples_per_second": 5.185, | |
| "eval_steps_per_second": 0.648, | |
| "step": 10000 | |
| } | |
| ], | |
| "logging_steps": 100, | |
| "max_steps": 10000, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 8, | |
| "save_steps": 2000, | |
| "total_flos": 1.884652011396268e+18, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |