| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 3.5855145213338115, | |
| "eval_steps": 2000, | |
| "global_step": 10000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.04, | |
| "grad_norm": 2.5916264057159424, | |
| "learning_rate": 1e-06, | |
| "loss": 1.0776, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.07, | |
| "grad_norm": 3.6381211280822754, | |
| "learning_rate": 9.898989898989898e-07, | |
| "loss": 0.8345, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.11, | |
| "grad_norm": 3.9932849407196045, | |
| "learning_rate": 9.7989898989899e-07, | |
| "loss": 0.6863, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.14, | |
| "grad_norm": 2.8963606357574463, | |
| "learning_rate": 9.697979797979798e-07, | |
| "loss": 0.6659, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.18, | |
| "grad_norm": 4.000125885009766, | |
| "learning_rate": 9.598989898989899e-07, | |
| "loss": 0.6552, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.22, | |
| "grad_norm": 3.7767927646636963, | |
| "learning_rate": 9.497979797979798e-07, | |
| "loss": 0.6583, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "grad_norm": 3.204336166381836, | |
| "learning_rate": 9.396969696969696e-07, | |
| "loss": 0.6229, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "grad_norm": 3.798582077026367, | |
| "learning_rate": 9.295959595959596e-07, | |
| "loss": 0.6138, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "grad_norm": 3.92914080619812, | |
| "learning_rate": 9.194949494949495e-07, | |
| "loss": 0.6266, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.36, | |
| "grad_norm": 3.982314109802246, | |
| "learning_rate": 9.093939393939394e-07, | |
| "loss": 0.6311, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.39, | |
| "grad_norm": 3.385096311569214, | |
| "learning_rate": 8.992929292929292e-07, | |
| "loss": 0.6141, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 0.43, | |
| "grad_norm": 3.015211343765259, | |
| "learning_rate": 8.891919191919191e-07, | |
| "loss": 0.6369, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 0.47, | |
| "grad_norm": 5.285658359527588, | |
| "learning_rate": 8.790909090909091e-07, | |
| "loss": 0.6442, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "grad_norm": 4.69854736328125, | |
| "learning_rate": 8.68989898989899e-07, | |
| "loss": 0.6419, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 0.54, | |
| "grad_norm": 2.95841121673584, | |
| "learning_rate": 8.58989898989899e-07, | |
| "loss": 0.6575, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.57, | |
| "grad_norm": 2.4997267723083496, | |
| "learning_rate": 8.488888888888888e-07, | |
| "loss": 0.6346, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 0.61, | |
| "grad_norm": 4.126640319824219, | |
| "learning_rate": 8.387878787878787e-07, | |
| "loss": 0.6005, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 0.65, | |
| "grad_norm": 3.040174961090088, | |
| "learning_rate": 8.286868686868687e-07, | |
| "loss": 0.6434, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 0.68, | |
| "grad_norm": 4.870083332061768, | |
| "learning_rate": 8.185858585858586e-07, | |
| "loss": 0.6202, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 0.72, | |
| "grad_norm": 4.307583808898926, | |
| "learning_rate": 8.084848484848484e-07, | |
| "loss": 0.6552, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.72, | |
| "eval_loss": 0.746256411075592, | |
| "eval_runtime": 199.9457, | |
| "eval_samples_per_second": 5.001, | |
| "eval_steps_per_second": 1.25, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "grad_norm": 4.793485164642334, | |
| "learning_rate": 7.983838383838384e-07, | |
| "loss": 0.6, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 0.79, | |
| "grad_norm": 5.155008316040039, | |
| "learning_rate": 7.882828282828282e-07, | |
| "loss": 0.5989, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 0.82, | |
| "grad_norm": 3.4713668823242188, | |
| "learning_rate": 7.781818181818182e-07, | |
| "loss": 0.589, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 0.86, | |
| "grad_norm": 4.512465476989746, | |
| "learning_rate": 7.68080808080808e-07, | |
| "loss": 0.5949, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 0.9, | |
| "grad_norm": 2.854623556137085, | |
| "learning_rate": 7.579797979797979e-07, | |
| "loss": 0.6103, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 0.93, | |
| "grad_norm": 6.94374418258667, | |
| "learning_rate": 7.478787878787879e-07, | |
| "loss": 0.6236, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 0.97, | |
| "grad_norm": 5.974214553833008, | |
| "learning_rate": 7.377777777777777e-07, | |
| "loss": 0.5915, | |
| "step": 2700 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 4.472714424133301, | |
| "learning_rate": 7.276767676767677e-07, | |
| "loss": 0.6023, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 1.04, | |
| "grad_norm": 5.157464504241943, | |
| "learning_rate": 7.175757575757575e-07, | |
| "loss": 0.6137, | |
| "step": 2900 | |
| }, | |
| { | |
| "epoch": 1.08, | |
| "grad_norm": 5.155125617980957, | |
| "learning_rate": 7.074747474747474e-07, | |
| "loss": 0.5723, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 1.11, | |
| "grad_norm": 2.872635841369629, | |
| "learning_rate": 6.973737373737374e-07, | |
| "loss": 0.5835, | |
| "step": 3100 | |
| }, | |
| { | |
| "epoch": 1.15, | |
| "grad_norm": 4.128993034362793, | |
| "learning_rate": 6.872727272727273e-07, | |
| "loss": 0.5777, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 1.18, | |
| "grad_norm": 3.2823545932769775, | |
| "learning_rate": 6.771717171717171e-07, | |
| "loss": 0.5875, | |
| "step": 3300 | |
| }, | |
| { | |
| "epoch": 1.22, | |
| "grad_norm": 3.7050082683563232, | |
| "learning_rate": 6.67070707070707e-07, | |
| "loss": 0.5901, | |
| "step": 3400 | |
| }, | |
| { | |
| "epoch": 1.25, | |
| "grad_norm": 5.737078666687012, | |
| "learning_rate": 6.56969696969697e-07, | |
| "loss": 0.5913, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 1.29, | |
| "grad_norm": 2.425701141357422, | |
| "learning_rate": 6.468686868686868e-07, | |
| "loss": 0.5759, | |
| "step": 3600 | |
| }, | |
| { | |
| "epoch": 1.33, | |
| "grad_norm": 3.5726475715637207, | |
| "learning_rate": 6.367676767676767e-07, | |
| "loss": 0.5976, | |
| "step": 3700 | |
| }, | |
| { | |
| "epoch": 1.36, | |
| "grad_norm": 5.098260879516602, | |
| "learning_rate": 6.266666666666667e-07, | |
| "loss": 0.5851, | |
| "step": 3800 | |
| }, | |
| { | |
| "epoch": 1.4, | |
| "grad_norm": 4.929934501647949, | |
| "learning_rate": 6.165656565656565e-07, | |
| "loss": 0.6006, | |
| "step": 3900 | |
| }, | |
| { | |
| "epoch": 1.43, | |
| "grad_norm": 4.153616428375244, | |
| "learning_rate": 6.064646464646465e-07, | |
| "loss": 0.5933, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 1.43, | |
| "eval_loss": 0.7265416383743286, | |
| "eval_runtime": 200.0654, | |
| "eval_samples_per_second": 4.998, | |
| "eval_steps_per_second": 1.25, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 1.47, | |
| "grad_norm": 4.094910144805908, | |
| "learning_rate": 5.963636363636363e-07, | |
| "loss": 0.6005, | |
| "step": 4100 | |
| }, | |
| { | |
| "epoch": 1.51, | |
| "grad_norm": 3.998939275741577, | |
| "learning_rate": 5.862626262626262e-07, | |
| "loss": 0.5747, | |
| "step": 4200 | |
| }, | |
| { | |
| "epoch": 1.54, | |
| "grad_norm": 3.7930455207824707, | |
| "learning_rate": 5.761616161616162e-07, | |
| "loss": 0.5879, | |
| "step": 4300 | |
| }, | |
| { | |
| "epoch": 1.58, | |
| "grad_norm": 2.8265774250030518, | |
| "learning_rate": 5.660606060606061e-07, | |
| "loss": 0.5867, | |
| "step": 4400 | |
| }, | |
| { | |
| "epoch": 1.61, | |
| "grad_norm": 5.012646198272705, | |
| "learning_rate": 5.559595959595959e-07, | |
| "loss": 0.5882, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 1.65, | |
| "grad_norm": 4.642397880554199, | |
| "learning_rate": 5.458585858585858e-07, | |
| "loss": 0.5623, | |
| "step": 4600 | |
| }, | |
| { | |
| "epoch": 1.69, | |
| "grad_norm": 4.442890167236328, | |
| "learning_rate": 5.357575757575758e-07, | |
| "loss": 0.6009, | |
| "step": 4700 | |
| }, | |
| { | |
| "epoch": 1.72, | |
| "grad_norm": 4.405778884887695, | |
| "learning_rate": 5.256565656565657e-07, | |
| "loss": 0.5962, | |
| "step": 4800 | |
| }, | |
| { | |
| "epoch": 1.76, | |
| "grad_norm": 4.690769672393799, | |
| "learning_rate": 5.155555555555555e-07, | |
| "loss": 0.5562, | |
| "step": 4900 | |
| }, | |
| { | |
| "epoch": 1.79, | |
| "grad_norm": 6.986325263977051, | |
| "learning_rate": 5.055555555555555e-07, | |
| "loss": 0.5812, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 1.83, | |
| "grad_norm": 3.599179983139038, | |
| "learning_rate": 4.954545454545454e-07, | |
| "loss": 0.5907, | |
| "step": 5100 | |
| }, | |
| { | |
| "epoch": 1.86, | |
| "grad_norm": 4.54085111618042, | |
| "learning_rate": 4.853535353535353e-07, | |
| "loss": 0.5971, | |
| "step": 5200 | |
| }, | |
| { | |
| "epoch": 1.9, | |
| "grad_norm": 4.154399871826172, | |
| "learning_rate": 4.752525252525252e-07, | |
| "loss": 0.5774, | |
| "step": 5300 | |
| }, | |
| { | |
| "epoch": 1.94, | |
| "grad_norm": 4.425224781036377, | |
| "learning_rate": 4.6515151515151513e-07, | |
| "loss": 0.5625, | |
| "step": 5400 | |
| }, | |
| { | |
| "epoch": 1.97, | |
| "grad_norm": 4.346124172210693, | |
| "learning_rate": 4.55050505050505e-07, | |
| "loss": 0.5645, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 2.01, | |
| "grad_norm": 4.570713043212891, | |
| "learning_rate": 4.449494949494949e-07, | |
| "loss": 0.587, | |
| "step": 5600 | |
| }, | |
| { | |
| "epoch": 2.04, | |
| "grad_norm": 3.669879674911499, | |
| "learning_rate": 4.3484848484848483e-07, | |
| "loss": 0.5609, | |
| "step": 5700 | |
| }, | |
| { | |
| "epoch": 2.08, | |
| "grad_norm": 5.310702800750732, | |
| "learning_rate": 4.2474747474747474e-07, | |
| "loss": 0.5487, | |
| "step": 5800 | |
| }, | |
| { | |
| "epoch": 2.12, | |
| "grad_norm": 10.60392951965332, | |
| "learning_rate": 4.1464646464646466e-07, | |
| "loss": 0.5603, | |
| "step": 5900 | |
| }, | |
| { | |
| "epoch": 2.15, | |
| "grad_norm": 3.6110637187957764, | |
| "learning_rate": 4.045454545454545e-07, | |
| "loss": 0.596, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 2.15, | |
| "eval_loss": 0.7302116751670837, | |
| "eval_runtime": 199.3626, | |
| "eval_samples_per_second": 5.016, | |
| "eval_steps_per_second": 1.254, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 2.19, | |
| "grad_norm": 5.180839538574219, | |
| "learning_rate": 3.9444444444444444e-07, | |
| "loss": 0.5581, | |
| "step": 6100 | |
| }, | |
| { | |
| "epoch": 2.22, | |
| "grad_norm": 6.653825283050537, | |
| "learning_rate": 3.843434343434343e-07, | |
| "loss": 0.5536, | |
| "step": 6200 | |
| }, | |
| { | |
| "epoch": 2.26, | |
| "grad_norm": 5.317089080810547, | |
| "learning_rate": 3.7424242424242427e-07, | |
| "loss": 0.5769, | |
| "step": 6300 | |
| }, | |
| { | |
| "epoch": 2.29, | |
| "grad_norm": 5.519111633300781, | |
| "learning_rate": 3.6414141414141413e-07, | |
| "loss": 0.5359, | |
| "step": 6400 | |
| }, | |
| { | |
| "epoch": 2.33, | |
| "grad_norm": 4.846569061279297, | |
| "learning_rate": 3.5404040404040405e-07, | |
| "loss": 0.5857, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 2.37, | |
| "grad_norm": 4.538639545440674, | |
| "learning_rate": 3.439393939393939e-07, | |
| "loss": 0.5757, | |
| "step": 6600 | |
| }, | |
| { | |
| "epoch": 2.4, | |
| "grad_norm": 5.268352031707764, | |
| "learning_rate": 3.3383838383838383e-07, | |
| "loss": 0.5481, | |
| "step": 6700 | |
| }, | |
| { | |
| "epoch": 2.44, | |
| "grad_norm": 4.480013370513916, | |
| "learning_rate": 3.237373737373737e-07, | |
| "loss": 0.5682, | |
| "step": 6800 | |
| }, | |
| { | |
| "epoch": 2.47, | |
| "grad_norm": 5.44348669052124, | |
| "learning_rate": 3.1363636363636366e-07, | |
| "loss": 0.5796, | |
| "step": 6900 | |
| }, | |
| { | |
| "epoch": 2.51, | |
| "grad_norm": 6.241185665130615, | |
| "learning_rate": 3.035353535353535e-07, | |
| "loss": 0.5931, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 2.55, | |
| "grad_norm": 5.187110424041748, | |
| "learning_rate": 2.9343434343434344e-07, | |
| "loss": 0.5696, | |
| "step": 7100 | |
| }, | |
| { | |
| "epoch": 2.58, | |
| "grad_norm": 6.06135368347168, | |
| "learning_rate": 2.833333333333333e-07, | |
| "loss": 0.5466, | |
| "step": 7200 | |
| }, | |
| { | |
| "epoch": 2.62, | |
| "grad_norm": 5.4311628341674805, | |
| "learning_rate": 2.732323232323232e-07, | |
| "loss": 0.5409, | |
| "step": 7300 | |
| }, | |
| { | |
| "epoch": 2.65, | |
| "grad_norm": 5.2975263595581055, | |
| "learning_rate": 2.631313131313131e-07, | |
| "loss": 0.5577, | |
| "step": 7400 | |
| }, | |
| { | |
| "epoch": 2.69, | |
| "grad_norm": 5.464794158935547, | |
| "learning_rate": 2.5303030303030305e-07, | |
| "loss": 0.584, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 2.72, | |
| "grad_norm": 5.726445198059082, | |
| "learning_rate": 2.429292929292929e-07, | |
| "loss": 0.5689, | |
| "step": 7600 | |
| }, | |
| { | |
| "epoch": 2.76, | |
| "grad_norm": 4.5637736320495605, | |
| "learning_rate": 2.3282828282828283e-07, | |
| "loss": 0.5704, | |
| "step": 7700 | |
| }, | |
| { | |
| "epoch": 2.8, | |
| "grad_norm": 5.696423530578613, | |
| "learning_rate": 2.2272727272727272e-07, | |
| "loss": 0.5689, | |
| "step": 7800 | |
| }, | |
| { | |
| "epoch": 2.83, | |
| "grad_norm": 5.203188896179199, | |
| "learning_rate": 2.1262626262626264e-07, | |
| "loss": 0.554, | |
| "step": 7900 | |
| }, | |
| { | |
| "epoch": 2.87, | |
| "grad_norm": 5.128230571746826, | |
| "learning_rate": 2.0252525252525253e-07, | |
| "loss": 0.5412, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 2.87, | |
| "eval_loss": 0.7274453043937683, | |
| "eval_runtime": 199.1299, | |
| "eval_samples_per_second": 5.022, | |
| "eval_steps_per_second": 1.255, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 2.9, | |
| "grad_norm": 5.771433353424072, | |
| "learning_rate": 1.9242424242424241e-07, | |
| "loss": 0.5662, | |
| "step": 8100 | |
| }, | |
| { | |
| "epoch": 2.94, | |
| "grad_norm": 4.031991004943848, | |
| "learning_rate": 1.8232323232323233e-07, | |
| "loss": 0.553, | |
| "step": 8200 | |
| }, | |
| { | |
| "epoch": 2.98, | |
| "grad_norm": 5.155855655670166, | |
| "learning_rate": 1.7222222222222222e-07, | |
| "loss": 0.5825, | |
| "step": 8300 | |
| }, | |
| { | |
| "epoch": 3.01, | |
| "grad_norm": 6.224823951721191, | |
| "learning_rate": 1.621212121212121e-07, | |
| "loss": 0.5891, | |
| "step": 8400 | |
| }, | |
| { | |
| "epoch": 3.05, | |
| "grad_norm": 5.529131889343262, | |
| "learning_rate": 1.5202020202020203e-07, | |
| "loss": 0.5318, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 3.08, | |
| "grad_norm": 4.826006889343262, | |
| "learning_rate": 1.4191919191919192e-07, | |
| "loss": 0.5817, | |
| "step": 8600 | |
| }, | |
| { | |
| "epoch": 3.12, | |
| "grad_norm": 4.968801498413086, | |
| "learning_rate": 1.318181818181818e-07, | |
| "loss": 0.5448, | |
| "step": 8700 | |
| }, | |
| { | |
| "epoch": 3.16, | |
| "grad_norm": 5.9340715408325195, | |
| "learning_rate": 1.2171717171717172e-07, | |
| "loss": 0.5697, | |
| "step": 8800 | |
| }, | |
| { | |
| "epoch": 3.19, | |
| "grad_norm": 6.170198440551758, | |
| "learning_rate": 1.1161616161616161e-07, | |
| "loss": 0.5539, | |
| "step": 8900 | |
| }, | |
| { | |
| "epoch": 3.23, | |
| "grad_norm": 3.5836563110351562, | |
| "learning_rate": 1.0151515151515151e-07, | |
| "loss": 0.5514, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 3.26, | |
| "grad_norm": 5.903813362121582, | |
| "learning_rate": 9.141414141414142e-08, | |
| "loss": 0.556, | |
| "step": 9100 | |
| }, | |
| { | |
| "epoch": 3.3, | |
| "grad_norm": 3.9017932415008545, | |
| "learning_rate": 8.13131313131313e-08, | |
| "loss": 0.5558, | |
| "step": 9200 | |
| }, | |
| { | |
| "epoch": 3.33, | |
| "grad_norm": 6.613101959228516, | |
| "learning_rate": 7.121212121212121e-08, | |
| "loss": 0.5476, | |
| "step": 9300 | |
| }, | |
| { | |
| "epoch": 3.37, | |
| "grad_norm": 6.56988525390625, | |
| "learning_rate": 6.121212121212121e-08, | |
| "loss": 0.5665, | |
| "step": 9400 | |
| }, | |
| { | |
| "epoch": 3.41, | |
| "grad_norm": 6.609213829040527, | |
| "learning_rate": 5.1111111111111114e-08, | |
| "loss": 0.5468, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 3.44, | |
| "grad_norm": 5.318997859954834, | |
| "learning_rate": 4.101010101010101e-08, | |
| "loss": 0.571, | |
| "step": 9600 | |
| }, | |
| { | |
| "epoch": 3.48, | |
| "grad_norm": 4.767919540405273, | |
| "learning_rate": 3.090909090909091e-08, | |
| "loss": 0.5523, | |
| "step": 9700 | |
| }, | |
| { | |
| "epoch": 3.51, | |
| "grad_norm": 5.016121864318848, | |
| "learning_rate": 2.0808080808080806e-08, | |
| "loss": 0.5431, | |
| "step": 9800 | |
| }, | |
| { | |
| "epoch": 3.55, | |
| "grad_norm": 5.048821926116943, | |
| "learning_rate": 1.0707070707070706e-08, | |
| "loss": 0.5418, | |
| "step": 9900 | |
| }, | |
| { | |
| "epoch": 3.59, | |
| "grad_norm": 4.946585655212402, | |
| "learning_rate": 6.06060606060606e-10, | |
| "loss": 0.5754, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 3.59, | |
| "eval_loss": 0.7410796880722046, | |
| "eval_runtime": 204.7897, | |
| "eval_samples_per_second": 4.883, | |
| "eval_steps_per_second": 1.221, | |
| "step": 10000 | |
| } | |
| ], | |
| "logging_steps": 100, | |
| "max_steps": 10000, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 4, | |
| "save_steps": 2000, | |
| "total_flos": 9.426795635681526e+17, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |