{ "best_metric": null, "best_model_checkpoint": null, "epoch": 0.11466574934067195, "eval_steps": 500, "global_step": 500, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 0.0, "grad_norm": 0.0, "learning_rate": 1.3333333333333334e-06, "loss": 0.0, "step": 1 }, { "epoch": 0.0, "grad_norm": 0.0, "learning_rate": 2.666666666666667e-06, "loss": 0.0, "step": 2 }, { "epoch": 0.0, "grad_norm": 0.0, "learning_rate": 4.000000000000001e-06, "loss": 0.0, "step": 3 }, { "epoch": 0.0, "grad_norm": 0.0, "learning_rate": 5.333333333333334e-06, "loss": 0.0, "step": 4 }, { "epoch": 0.0, "grad_norm": 0.0, "learning_rate": 6.666666666666667e-06, "loss": 0.0, "step": 5 }, { "epoch": 0.0, "grad_norm": 0.0, "learning_rate": 8.000000000000001e-06, "loss": 0.0, "step": 6 }, { "epoch": 0.0, "grad_norm": 0.0, "learning_rate": 9.333333333333334e-06, "loss": 0.0, "step": 7 }, { "epoch": 0.0, "grad_norm": 0.0, "learning_rate": 1.0666666666666667e-05, "loss": 0.0, "step": 8 }, { "epoch": 0.0, "grad_norm": 0.0, "learning_rate": 1.2e-05, "loss": 0.0, "step": 9 }, { "epoch": 0.0, "grad_norm": 0.0, "learning_rate": 1.3333333333333333e-05, "loss": 0.0, "step": 10 }, { "epoch": 0.0, "grad_norm": 0.0, "learning_rate": 1.4666666666666666e-05, "loss": 0.0, "step": 11 }, { "epoch": 0.0, "grad_norm": 0.0, "learning_rate": 1.6000000000000003e-05, "loss": 0.0, "step": 12 }, { "epoch": 0.0, "grad_norm": 0.0, "learning_rate": 1.7333333333333336e-05, "loss": 0.0, "step": 13 }, { "epoch": 0.0, "grad_norm": 0.0, "learning_rate": 1.866666666666667e-05, "loss": 0.0, "step": 14 }, { "epoch": 0.0, "grad_norm": 0.0, "learning_rate": 2e-05, "loss": 0.0, "step": 15 }, { "epoch": 0.0, "grad_norm": 0.0, "learning_rate": 1.999979021001399e-05, "loss": 0.0, "step": 16 }, { "epoch": 0.0, "grad_norm": 0.0, "learning_rate": 1.999916084885832e-05, "loss": 0.0, "step": 17 }, { "epoch": 0.0, "grad_norm": 0.0, "learning_rate": 1.9998111942939727e-05, "loss": 0.0, "step": 18 }, { "epoch": 0.0, "grad_norm": 0.0, "learning_rate": 1.9996643536268202e-05, "loss": 0.0, "step": 19 }, { "epoch": 0.0, "grad_norm": 0.0, "learning_rate": 1.9994755690455154e-05, "loss": 0.0, "step": 20 }, { "epoch": 0.0, "grad_norm": 0.0, "learning_rate": 1.99924484847108e-05, "loss": 0.0, "step": 21 }, { "epoch": 0.01, "grad_norm": 0.0, "learning_rate": 1.998972201584088e-05, "loss": 0.0, "step": 22 }, { "epoch": 0.01, "grad_norm": 0.0, "learning_rate": 1.9986576398242566e-05, "loss": 0.0, "step": 23 }, { "epoch": 0.01, "grad_norm": 0.0, "learning_rate": 1.9983011763899674e-05, "loss": 0.0, "step": 24 }, { "epoch": 0.01, "grad_norm": 0.0, "learning_rate": 1.997902826237712e-05, "loss": 0.0, "step": 25 }, { "epoch": 0.01, "grad_norm": 0.0, "learning_rate": 1.997462606081465e-05, "loss": 0.0, "step": 26 }, { "epoch": 0.01, "grad_norm": 0.0, "learning_rate": 1.9969805343919822e-05, "loss": 0.0, "step": 27 }, { "epoch": 0.01, "grad_norm": 0.0, "learning_rate": 1.9964566313960265e-05, "loss": 0.0, "step": 28 }, { "epoch": 0.01, "grad_norm": 0.0, "learning_rate": 1.995890919075519e-05, "loss": 0.0, "step": 29 }, { "epoch": 0.01, "grad_norm": 0.0, "learning_rate": 1.995283421166614e-05, "loss": 0.0, "step": 30 }, { "epoch": 0.01, "grad_norm": 0.0, "learning_rate": 1.9946341631587086e-05, "loss": 0.0, "step": 31 }, { "epoch": 0.01, "grad_norm": 0.0, "learning_rate": 1.9939431722933678e-05, "loss": 0.0, "step": 32 }, { "epoch": 0.01, "grad_norm": 0.0, "learning_rate": 1.9932104775631847e-05, "loss": 0.0, "step": 33 }, { "epoch": 0.01, "grad_norm": 0.0, "learning_rate": 1.9924361097105624e-05, "loss": 0.0, "step": 34 }, { "epoch": 0.01, "grad_norm": 0.0, "learning_rate": 1.9916201012264255e-05, "loss": 0.0, "step": 35 }, { "epoch": 0.01, "grad_norm": 0.0, "learning_rate": 1.990762486348855e-05, "loss": 0.0, "step": 36 }, { "epoch": 0.01, "grad_norm": 0.0, "learning_rate": 1.989863301061654e-05, "loss": 0.0, "step": 37 }, { "epoch": 0.01, "grad_norm": 0.0, "learning_rate": 1.9889225830928365e-05, "loss": 0.0, "step": 38 }, { "epoch": 0.01, "grad_norm": 0.0, "learning_rate": 1.987940371913044e-05, "loss": 0.0, "step": 39 }, { "epoch": 0.01, "grad_norm": 0.0, "learning_rate": 1.9869167087338908e-05, "loss": 0.0, "step": 40 }, { "epoch": 0.01, "grad_norm": 0.0, "learning_rate": 1.9858516365062334e-05, "loss": 0.0, "step": 41 }, { "epoch": 0.01, "grad_norm": 0.0, "learning_rate": 1.9847451999183692e-05, "loss": 0.0, "step": 42 }, { "epoch": 0.01, "grad_norm": 0.0, "learning_rate": 1.9835974453941623e-05, "loss": 0.0, "step": 43 }, { "epoch": 0.01, "grad_norm": 0.0, "learning_rate": 1.9824084210910924e-05, "loss": 0.0, "step": 44 }, { "epoch": 0.01, "grad_norm": 0.0, "learning_rate": 1.9811781768982392e-05, "loss": 0.0, "step": 45 }, { "epoch": 0.01, "grad_norm": 0.0, "learning_rate": 1.9799067644341844e-05, "loss": 0.0, "step": 46 }, { "epoch": 0.01, "grad_norm": 0.0, "learning_rate": 1.978594237044849e-05, "loss": 0.0, "step": 47 }, { "epoch": 0.01, "grad_norm": 0.0, "learning_rate": 1.977240649801253e-05, "loss": 0.0, "step": 48 }, { "epoch": 0.01, "grad_norm": 0.0, "learning_rate": 1.9758460594972068e-05, "loss": 0.0, "step": 49 }, { "epoch": 0.01, "grad_norm": 0.0, "learning_rate": 1.9744105246469264e-05, "loss": 0.0, "step": 50 }, { "epoch": 0.01, "grad_norm": 0.0, "learning_rate": 1.9729341054825783e-05, "loss": 0.0, "step": 51 }, { "epoch": 0.01, "grad_norm": 0.0, "learning_rate": 1.9714168639517543e-05, "loss": 0.0, "step": 52 }, { "epoch": 0.01, "grad_norm": 0.0, "learning_rate": 1.9698588637148705e-05, "loss": 0.0, "step": 53 }, { "epoch": 0.01, "grad_norm": 0.0, "learning_rate": 1.9682601701424958e-05, "loss": 0.0, "step": 54 }, { "epoch": 0.01, "grad_norm": 0.0, "learning_rate": 1.9666208503126115e-05, "loss": 0.0, "step": 55 }, { "epoch": 0.01, "grad_norm": 0.0, "learning_rate": 1.9649409730077934e-05, "loss": 0.0, "step": 56 }, { "epoch": 0.01, "grad_norm": 0.0, "learning_rate": 1.9632206087123296e-05, "loss": 0.0, "step": 57 }, { "epoch": 0.01, "grad_norm": 0.0, "learning_rate": 1.9614598296092603e-05, "loss": 0.0, "step": 58 }, { "epoch": 0.01, "grad_norm": 0.0, "learning_rate": 1.9596587095773496e-05, "loss": 0.0, "step": 59 }, { "epoch": 0.01, "grad_norm": 0.0, "learning_rate": 1.957817324187987e-05, "loss": 0.0, "step": 60 }, { "epoch": 0.01, "grad_norm": 0.0, "learning_rate": 1.9559357507020163e-05, "loss": 0.0, "step": 61 }, { "epoch": 0.01, "grad_norm": 0.0, "learning_rate": 1.9540140680664915e-05, "loss": 0.0, "step": 62 }, { "epoch": 0.01, "grad_norm": 0.0, "learning_rate": 1.952052356911368e-05, "loss": 0.0, "step": 63 }, { "epoch": 0.01, "grad_norm": 0.0, "learning_rate": 1.950050699546116e-05, "loss": 0.0, "step": 64 }, { "epoch": 0.01, "grad_norm": 0.0, "learning_rate": 1.9480091799562706e-05, "loss": 0.0, "step": 65 }, { "epoch": 0.02, "grad_norm": 0.0, "learning_rate": 1.9459278837999048e-05, "loss": 0.0, "step": 66 }, { "epoch": 0.02, "grad_norm": 0.0, "learning_rate": 1.9438068984040366e-05, "loss": 0.0, "step": 67 }, { "epoch": 0.02, "grad_norm": 0.0, "learning_rate": 1.9416463127609655e-05, "loss": 0.0, "step": 68 }, { "epoch": 0.02, "grad_norm": 0.0, "learning_rate": 1.9394462175245382e-05, "loss": 0.0, "step": 69 }, { "epoch": 0.02, "grad_norm": 0.0, "learning_rate": 1.937206705006344e-05, "loss": 0.0, "step": 70 }, { "epoch": 0.02, "grad_norm": 0.0, "learning_rate": 1.9349278691718426e-05, "loss": 0.0, "step": 71 }, { "epoch": 0.02, "grad_norm": 0.0, "learning_rate": 1.9326098056364224e-05, "loss": 0.0, "step": 72 }, { "epoch": 0.02, "grad_norm": 0.0, "learning_rate": 1.9302526116613863e-05, "loss": 0.0, "step": 73 }, { "epoch": 0.02, "grad_norm": 0.0, "learning_rate": 1.9278563861498726e-05, "loss": 0.0, "step": 74 }, { "epoch": 0.02, "grad_norm": 0.0, "learning_rate": 1.9254212296427043e-05, "loss": 0.0, "step": 75 }, { "epoch": 0.02, "grad_norm": 0.0, "learning_rate": 1.922947244314172e-05, "loss": 0.0, "step": 76 }, { "epoch": 0.02, "grad_norm": 0.0, "learning_rate": 1.9204345339677442e-05, "loss": 0.0, "step": 77 }, { "epoch": 0.02, "grad_norm": 0.0, "learning_rate": 1.9178832040317153e-05, "loss": 0.0, "step": 78 }, { "epoch": 0.02, "grad_norm": 0.0, "learning_rate": 1.91529336155478e-05, "loss": 0.0, "step": 79 }, { "epoch": 0.02, "grad_norm": 0.0, "learning_rate": 1.9126651152015404e-05, "loss": 0.0, "step": 80 }, { "epoch": 0.02, "grad_norm": 0.0, "learning_rate": 1.9099985752479505e-05, "loss": 0.0, "step": 81 }, { "epoch": 0.02, "grad_norm": 0.0, "learning_rate": 1.9072938535766864e-05, "loss": 0.0, "step": 82 }, { "epoch": 0.02, "grad_norm": 0.0, "learning_rate": 1.904551063672452e-05, "loss": 0.0, "step": 83 }, { "epoch": 0.02, "grad_norm": 0.0, "learning_rate": 1.9017703206172187e-05, "loss": 0.0, "step": 84 }, { "epoch": 0.02, "grad_norm": 0.0, "learning_rate": 1.8989517410853956e-05, "loss": 0.0, "step": 85 }, { "epoch": 0.02, "grad_norm": 0.0, "learning_rate": 1.896095443338935e-05, "loss": 0.0, "step": 86 }, { "epoch": 0.02, "grad_norm": 0.0, "learning_rate": 1.8932015472223692e-05, "loss": 0.0, "step": 87 }, { "epoch": 0.02, "grad_norm": 0.0, "learning_rate": 1.8902701741577844e-05, "loss": 0.0, "step": 88 }, { "epoch": 0.02, "grad_norm": 0.0, "learning_rate": 1.8873014471397225e-05, "loss": 0.0, "step": 89 }, { "epoch": 0.02, "grad_norm": 0.0, "learning_rate": 1.8842954907300236e-05, "loss": 0.0, "step": 90 }, { "epoch": 0.02, "grad_norm": 0.0, "learning_rate": 1.881252431052599e-05, "loss": 0.0, "step": 91 }, { "epoch": 0.02, "grad_norm": 0.0, "learning_rate": 1.8781723957881374e-05, "loss": 0.0, "step": 92 }, { "epoch": 0.02, "grad_norm": 0.0, "learning_rate": 1.87505551416875e-05, "loss": 0.0, "step": 93 }, { "epoch": 0.02, "grad_norm": 0.0, "learning_rate": 1.871901916972547e-05, "loss": 0.0, "step": 94 }, { "epoch": 0.02, "grad_norm": 0.0, "learning_rate": 1.8687117365181514e-05, "loss": 0.0, "step": 95 }, { "epoch": 0.02, "grad_norm": 0.0, "learning_rate": 1.865485106659145e-05, "loss": 0.0, "step": 96 }, { "epoch": 0.02, "grad_norm": 0.0, "learning_rate": 1.862222162778454e-05, "loss": 0.0, "step": 97 }, { "epoch": 0.02, "grad_norm": 0.0, "learning_rate": 1.85892304178267e-05, "loss": 0.0, "step": 98 }, { "epoch": 0.02, "grad_norm": 0.0, "learning_rate": 1.8555878820963014e-05, "loss": 0.0, "step": 99 }, { "epoch": 0.02, "grad_norm": 0.0, "learning_rate": 1.8522168236559693e-05, "loss": 0.0, "step": 100 }, { "epoch": 0.02, "grad_norm": 0.0, "learning_rate": 1.8488100079045345e-05, "loss": 0.0, "step": 101 }, { "epoch": 0.02, "grad_norm": 0.0, "learning_rate": 1.8453675777851627e-05, "loss": 0.0, "step": 102 }, { "epoch": 0.02, "grad_norm": 0.0, "learning_rate": 1.8418896777353272e-05, "loss": 0.0, "step": 103 }, { "epoch": 0.02, "grad_norm": 0.0, "learning_rate": 1.8383764536807486e-05, "loss": 0.0, "step": 104 }, { "epoch": 0.02, "grad_norm": 0.0, "learning_rate": 1.8348280530292712e-05, "loss": 0.0, "step": 105 }, { "epoch": 0.02, "grad_norm": 0.0, "learning_rate": 1.831244624664681e-05, "loss": 0.0, "step": 106 }, { "epoch": 0.02, "grad_norm": 0.0, "learning_rate": 1.827626318940454e-05, "loss": 0.0, "step": 107 }, { "epoch": 0.02, "grad_norm": 0.0, "learning_rate": 1.8239732876734525e-05, "loss": 0.0, "step": 108 }, { "epoch": 0.02, "grad_norm": 0.0, "learning_rate": 1.8202856841375517e-05, "loss": 0.0, "step": 109 }, { "epoch": 0.03, "grad_norm": 0.0, "learning_rate": 1.816563663057211e-05, "loss": 0.0, "step": 110 }, { "epoch": 0.03, "grad_norm": 0.0, "learning_rate": 1.81280738060098e-05, "loss": 0.0, "step": 111 }, { "epoch": 0.03, "grad_norm": 0.0, "learning_rate": 1.8090169943749477e-05, "loss": 0.0, "step": 112 }, { "epoch": 0.03, "grad_norm": 0.0, "learning_rate": 1.8051926634161282e-05, "loss": 0.0, "step": 113 }, { "epoch": 0.03, "grad_norm": 0.0, "learning_rate": 1.8013345481857903e-05, "loss": 0.0, "step": 114 }, { "epoch": 0.03, "grad_norm": 0.0, "learning_rate": 1.797442810562721e-05, "loss": 0.0, "step": 115 }, { "epoch": 0.03, "grad_norm": 0.0, "learning_rate": 1.793517613836437e-05, "loss": 0.0, "step": 116 }, { "epoch": 0.03, "grad_norm": 0.0, "learning_rate": 1.7895591227003316e-05, "loss": 0.0, "step": 117 }, { "epoch": 0.03, "grad_norm": 0.0, "learning_rate": 1.7855675032447648e-05, "loss": 0.0, "step": 118 }, { "epoch": 0.03, "grad_norm": 0.0, "learning_rate": 1.7815429229500946e-05, "loss": 0.0, "step": 119 }, { "epoch": 0.03, "grad_norm": 0.0, "learning_rate": 1.7774855506796497e-05, "loss": 0.0, "step": 120 }, { "epoch": 0.03, "grad_norm": 0.0, "learning_rate": 1.7733955566726438e-05, "loss": 0.0, "step": 121 }, { "epoch": 0.03, "grad_norm": 0.0, "learning_rate": 1.7692731125370355e-05, "loss": 0.0, "step": 122 }, { "epoch": 0.03, "grad_norm": 0.0, "learning_rate": 1.7651183912423228e-05, "loss": 0.0, "step": 123 }, { "epoch": 0.03, "grad_norm": 0.0, "learning_rate": 1.7609315671122912e-05, "loss": 0.0, "step": 124 }, { "epoch": 0.03, "grad_norm": 0.0, "learning_rate": 1.7567128158176955e-05, "loss": 0.0, "step": 125 }, { "epoch": 0.03, "grad_norm": 0.0, "learning_rate": 1.7524623143688905e-05, "loss": 0.0, "step": 126 }, { "epoch": 0.03, "grad_norm": 0.0, "learning_rate": 1.748180241108404e-05, "loss": 0.0, "step": 127 }, { "epoch": 0.03, "grad_norm": 0.0, "learning_rate": 1.7438667757034547e-05, "loss": 0.0, "step": 128 }, { "epoch": 0.03, "grad_norm": 0.0, "learning_rate": 1.739522099138411e-05, "loss": 0.0, "step": 129 }, { "epoch": 0.03, "grad_norm": 0.0, "learning_rate": 1.7351463937072008e-05, "loss": 0.0, "step": 130 }, { "epoch": 0.03, "grad_norm": 0.0, "learning_rate": 1.7307398430056595e-05, "loss": 0.0, "step": 131 }, { "epoch": 0.03, "grad_norm": 0.0, "learning_rate": 1.72630263192383e-05, "loss": 0.0, "step": 132 }, { "epoch": 0.03, "grad_norm": 0.0, "learning_rate": 1.7218349466382024e-05, "loss": 0.0, "step": 133 }, { "epoch": 0.03, "grad_norm": 0.0, "learning_rate": 1.7173369746039026e-05, "loss": 0.0, "step": 134 }, { "epoch": 0.03, "grad_norm": 0.0, "learning_rate": 1.7128089045468294e-05, "loss": 0.0, "step": 135 }, { "epoch": 0.03, "grad_norm": 0.0, "learning_rate": 1.7082509264557333e-05, "loss": 0.0, "step": 136 }, { "epoch": 0.03, "grad_norm": 0.0, "learning_rate": 1.7036632315742464e-05, "loss": 0.0, "step": 137 }, { "epoch": 0.03, "grad_norm": 0.0, "learning_rate": 1.6990460123928577e-05, "loss": 0.0, "step": 138 }, { "epoch": 0.03, "grad_norm": 0.0, "learning_rate": 1.6943994626408365e-05, "loss": 0.0, "step": 139 }, { "epoch": 0.03, "grad_norm": 0.0, "learning_rate": 1.6897237772781046e-05, "loss": 0.0, "step": 140 }, { "epoch": 0.03, "grad_norm": 0.0, "learning_rate": 1.6850191524870548e-05, "loss": 0.0, "step": 141 }, { "epoch": 0.03, "grad_norm": 0.0, "learning_rate": 1.6802857856643214e-05, "loss": 0.0, "step": 142 }, { "epoch": 0.03, "grad_norm": 0.0, "learning_rate": 1.6755238754124965e-05, "loss": 0.0, "step": 143 }, { "epoch": 0.03, "grad_norm": 0.0, "learning_rate": 1.6707336215317968e-05, "loss": 0.0, "step": 144 }, { "epoch": 0.03, "grad_norm": 0.0, "learning_rate": 1.665915225011681e-05, "loss": 0.0, "step": 145 }, { "epoch": 0.03, "grad_norm": 0.0, "learning_rate": 1.6610688880224178e-05, "loss": 0.0, "step": 146 }, { "epoch": 0.03, "grad_norm": 0.0, "learning_rate": 1.6561948139065997e-05, "loss": 0.0, "step": 147 }, { "epoch": 0.03, "grad_norm": 0.0, "learning_rate": 1.6512932071706153e-05, "loss": 0.0, "step": 148 }, { "epoch": 0.03, "grad_norm": 0.0, "learning_rate": 1.646364273476067e-05, "loss": 0.0, "step": 149 }, { "epoch": 0.03, "grad_norm": 0.0, "learning_rate": 1.6414082196311402e-05, "loss": 0.0, "step": 150 }, { "epoch": 0.03, "grad_norm": 0.0, "learning_rate": 1.6364252535819284e-05, "loss": 0.0, "step": 151 }, { "epoch": 0.03, "grad_norm": 0.0, "learning_rate": 1.6314155844037074e-05, "loss": 0.0, "step": 152 }, { "epoch": 0.04, "grad_norm": 0.0, "learning_rate": 1.626379422292162e-05, "loss": 0.0, "step": 153 }, { "epoch": 0.04, "grad_norm": 0.0, "learning_rate": 1.6213169785545688e-05, "loss": 0.0, "step": 154 }, { "epoch": 0.04, "grad_norm": 0.0, "learning_rate": 1.6162284656009276e-05, "loss": 0.0, "step": 155 }, { "epoch": 0.04, "grad_norm": 0.0, "learning_rate": 1.6111140969350504e-05, "loss": 0.0, "step": 156 }, { "epoch": 0.04, "grad_norm": 0.0, "learning_rate": 1.6059740871456035e-05, "loss": 0.0, "step": 157 }, { "epoch": 0.04, "grad_norm": 0.0, "learning_rate": 1.6008086518971037e-05, "loss": 0.0, "step": 158 }, { "epoch": 0.04, "grad_norm": 0.0, "learning_rate": 1.5956180079208684e-05, "loss": 0.0, "step": 159 }, { "epoch": 0.04, "grad_norm": 0.0, "learning_rate": 1.5904023730059227e-05, "loss": 0.0, "step": 160 }, { "epoch": 0.04, "grad_norm": 0.0, "learning_rate": 1.5851619659898623e-05, "loss": 0.0, "step": 161 }, { "epoch": 0.04, "grad_norm": 0.0, "learning_rate": 1.57989700674967e-05, "loss": 0.0, "step": 162 }, { "epoch": 0.04, "grad_norm": 0.0, "learning_rate": 1.5746077161924905e-05, "loss": 0.0, "step": 163 }, { "epoch": 0.04, "grad_norm": 0.0, "learning_rate": 1.5692943162463628e-05, "loss": 0.0, "step": 164 }, { "epoch": 0.04, "grad_norm": 0.0, "learning_rate": 1.5639570298509067e-05, "loss": 0.0, "step": 165 }, { "epoch": 0.04, "grad_norm": 0.0, "learning_rate": 1.5585960809479698e-05, "loss": 0.0, "step": 166 }, { "epoch": 0.04, "grad_norm": 0.0, "learning_rate": 1.5532116944722308e-05, "loss": 0.0, "step": 167 }, { "epoch": 0.04, "grad_norm": 0.0, "learning_rate": 1.547804096341763e-05, "loss": 0.0, "step": 168 }, { "epoch": 0.04, "grad_norm": 0.0, "learning_rate": 1.5423735134485537e-05, "loss": 0.0, "step": 169 }, { "epoch": 0.04, "grad_norm": 0.0, "learning_rate": 1.536920173648984e-05, "loss": 0.0, "step": 170 }, { "epoch": 0.04, "grad_norm": 0.0, "learning_rate": 1.5314443057542703e-05, "loss": 0.0, "step": 171 }, { "epoch": 0.04, "grad_norm": 0.0, "learning_rate": 1.5259461395208628e-05, "loss": 0.0, "step": 172 }, { "epoch": 0.04, "grad_norm": 0.0, "learning_rate": 1.5204259056408046e-05, "loss": 0.0, "step": 173 }, { "epoch": 0.04, "grad_norm": 0.0, "learning_rate": 1.5148838357320537e-05, "loss": 0.0, "step": 174 }, { "epoch": 0.04, "grad_norm": 0.0, "learning_rate": 1.5093201623287631e-05, "loss": 0.0, "step": 175 }, { "epoch": 0.04, "grad_norm": 0.0, "learning_rate": 1.5037351188715265e-05, "loss": 0.0, "step": 176 }, { "epoch": 0.04, "grad_norm": 0.0, "learning_rate": 1.4981289396975818e-05, "loss": 0.0, "step": 177 }, { "epoch": 0.04, "grad_norm": 0.0, "learning_rate": 1.4925018600309784e-05, "loss": 0.0, "step": 178 }, { "epoch": 0.04, "grad_norm": 0.0, "learning_rate": 1.4868541159727097e-05, "loss": 0.0, "step": 179 }, { "epoch": 0.04, "grad_norm": 0.0, "learning_rate": 1.4811859444908053e-05, "loss": 0.0, "step": 180 }, { "epoch": 0.04, "grad_norm": 0.0, "learning_rate": 1.4754975834103877e-05, "loss": 0.0, "step": 181 }, { "epoch": 0.04, "grad_norm": 0.0, "learning_rate": 1.4697892714036959e-05, "loss": 0.0, "step": 182 }, { "epoch": 0.04, "grad_norm": 0.0, "learning_rate": 1.4640612479800686e-05, "loss": 0.0, "step": 183 }, { "epoch": 0.04, "grad_norm": 0.0, "learning_rate": 1.4583137534758968e-05, "loss": 0.0, "step": 184 }, { "epoch": 0.04, "grad_norm": 0.0, "learning_rate": 1.4525470290445392e-05, "loss": 0.0, "step": 185 }, { "epoch": 0.04, "grad_norm": 0.0, "learning_rate": 1.4467613166462024e-05, "loss": 0.0, "step": 186 }, { "epoch": 0.04, "grad_norm": 0.0, "learning_rate": 1.4409568590377918e-05, "loss": 0.0, "step": 187 }, { "epoch": 0.04, "grad_norm": 0.0, "learning_rate": 1.4351338997627233e-05, "loss": 0.0, "step": 188 }, { "epoch": 0.04, "grad_norm": 0.0, "learning_rate": 1.429292683140706e-05, "loss": 0.0, "step": 189 }, { "epoch": 0.04, "grad_norm": 0.0, "learning_rate": 1.4234334542574906e-05, "loss": 0.0, "step": 190 }, { "epoch": 0.04, "grad_norm": 0.0, "learning_rate": 1.4175564589545853e-05, "loss": 0.0, "step": 191 }, { "epoch": 0.04, "grad_norm": 0.0, "learning_rate": 1.411661943818944e-05, "loss": 0.0, "step": 192 }, { "epoch": 0.04, "grad_norm": 0.0, "learning_rate": 1.4057501561726157e-05, "loss": 0.0, "step": 193 }, { "epoch": 0.04, "grad_norm": 0.0, "learning_rate": 1.3998213440623691e-05, "loss": 0.0, "step": 194 }, { "epoch": 0.04, "grad_norm": 0.0, "learning_rate": 1.3938757562492873e-05, "loss": 0.0, "step": 195 }, { "epoch": 0.04, "grad_norm": 0.0, "learning_rate": 1.3879136421983265e-05, "loss": 0.0, "step": 196 }, { "epoch": 0.05, "grad_norm": 0.0, "learning_rate": 1.3819352520678519e-05, "loss": 0.0, "step": 197 }, { "epoch": 0.05, "grad_norm": 0.0, "learning_rate": 1.3759408366991391e-05, "loss": 0.0, "step": 198 }, { "epoch": 0.05, "grad_norm": 0.0, "learning_rate": 1.3699306476058523e-05, "loss": 0.0, "step": 199 }, { "epoch": 0.05, "grad_norm": 0.0, "learning_rate": 1.3639049369634878e-05, "loss": 0.0, "step": 200 }, { "epoch": 0.05, "grad_norm": 0.0, "learning_rate": 1.357863957598796e-05, "loss": 0.0, "step": 201 }, { "epoch": 0.05, "grad_norm": 0.0, "learning_rate": 1.3518079629791725e-05, "loss": 0.0, "step": 202 }, { "epoch": 0.05, "grad_norm": 0.0, "learning_rate": 1.345737207202023e-05, "loss": 0.0, "step": 203 }, { "epoch": 0.05, "grad_norm": 0.0, "learning_rate": 1.3396519449841006e-05, "loss": 0.0, "step": 204 }, { "epoch": 0.05, "grad_norm": 0.0, "learning_rate": 1.3335524316508208e-05, "loss": 0.0, "step": 205 }, { "epoch": 0.05, "grad_norm": 0.0, "learning_rate": 1.3274389231255466e-05, "loss": 0.0, "step": 206 }, { "epoch": 0.05, "grad_norm": 0.0, "learning_rate": 1.3213116759188525e-05, "loss": 0.0, "step": 207 }, { "epoch": 0.05, "grad_norm": 0.0, "learning_rate": 1.3151709471177589e-05, "loss": 0.0, "step": 208 }, { "epoch": 0.05, "grad_norm": 0.0, "learning_rate": 1.3090169943749475e-05, "loss": 0.0, "step": 209 }, { "epoch": 0.05, "grad_norm": 0.0, "learning_rate": 1.3028500758979507e-05, "loss": 0.0, "step": 210 }, { "epoch": 0.05, "grad_norm": 0.0, "learning_rate": 1.296670450438317e-05, "loss": 0.0, "step": 211 }, { "epoch": 0.05, "grad_norm": 0.0, "learning_rate": 1.2904783772807534e-05, "loss": 0.0, "step": 212 }, { "epoch": 0.05, "grad_norm": 0.0, "learning_rate": 1.2842741162322487e-05, "loss": 0.0, "step": 213 }, { "epoch": 0.05, "grad_norm": 0.0, "learning_rate": 1.2780579276111702e-05, "loss": 0.0, "step": 214 }, { "epoch": 0.05, "grad_norm": 0.0, "learning_rate": 1.2718300722363431e-05, "loss": 0.0, "step": 215 }, { "epoch": 0.05, "grad_norm": 0.0, "learning_rate": 1.2655908114161053e-05, "loss": 0.0, "step": 216 }, { "epoch": 0.05, "grad_norm": 0.0, "learning_rate": 1.2593404069373452e-05, "loss": 0.0, "step": 217 }, { "epoch": 0.05, "grad_norm": 0.0, "learning_rate": 1.2530791210545163e-05, "loss": 0.0, "step": 218 }, { "epoch": 0.05, "grad_norm": 0.0, "learning_rate": 1.2468072164786342e-05, "loss": 0.0, "step": 219 }, { "epoch": 0.05, "grad_norm": 0.0, "learning_rate": 1.2405249563662539e-05, "loss": 0.0, "step": 220 }, { "epoch": 0.05, "grad_norm": 0.0, "learning_rate": 1.2342326043084268e-05, "loss": 0.0, "step": 221 }, { "epoch": 0.05, "grad_norm": 0.0, "learning_rate": 1.2279304243196438e-05, "loss": 0.0, "step": 222 }, { "epoch": 0.05, "grad_norm": 0.0, "learning_rate": 1.2216186808267544e-05, "loss": 0.0, "step": 223 }, { "epoch": 0.05, "grad_norm": 0.0, "learning_rate": 1.215297638657875e-05, "loss": 0.0, "step": 224 }, { "epoch": 0.05, "grad_norm": 0.0, "learning_rate": 1.2089675630312755e-05, "loss": 0.0, "step": 225 }, { "epoch": 0.05, "grad_norm": 0.0, "learning_rate": 1.2026287195442503e-05, "loss": 0.0, "step": 226 }, { "epoch": 0.05, "grad_norm": 0.0, "learning_rate": 1.1962813741619777e-05, "loss": 0.0, "step": 227 }, { "epoch": 0.05, "grad_norm": 0.0, "learning_rate": 1.189925793206357e-05, "loss": 0.0, "step": 228 }, { "epoch": 0.05, "grad_norm": 0.0, "learning_rate": 1.1835622433448361e-05, "loss": 0.0, "step": 229 }, { "epoch": 0.05, "grad_norm": 0.0, "learning_rate": 1.177190991579223e-05, "loss": 0.0, "step": 230 }, { "epoch": 0.05, "grad_norm": 0.0, "learning_rate": 1.1708123052344803e-05, "loss": 0.0, "step": 231 }, { "epoch": 0.05, "grad_norm": 0.0, "learning_rate": 1.164426451947513e-05, "loss": 0.0, "step": 232 }, { "epoch": 0.05, "grad_norm": 0.0, "learning_rate": 1.1580336996559343e-05, "loss": 0.0, "step": 233 }, { "epoch": 0.05, "grad_norm": 0.0, "learning_rate": 1.151634316586828e-05, "loss": 0.0, "step": 234 }, { "epoch": 0.05, "grad_norm": 0.0, "learning_rate": 1.1452285712454905e-05, "loss": 0.0, "step": 235 }, { "epoch": 0.05, "grad_norm": 0.0, "learning_rate": 1.138816732404167e-05, "loss": 0.0, "step": 236 }, { "epoch": 0.05, "grad_norm": 0.0, "learning_rate": 1.1323990690907734e-05, "loss": 0.0, "step": 237 }, { "epoch": 0.05, "grad_norm": 0.0, "learning_rate": 1.1259758505776092e-05, "loss": 0.0, "step": 238 }, { "epoch": 0.05, "grad_norm": 0.0, "learning_rate": 1.119547346370059e-05, "loss": 0.0, "step": 239 }, { "epoch": 0.06, "grad_norm": 0.0, "learning_rate": 1.1131138261952845e-05, "loss": 0.0, "step": 240 }, { "epoch": 0.06, "grad_norm": 0.0, "learning_rate": 1.1066755599909065e-05, "loss": 0.0, "step": 241 }, { "epoch": 0.06, "grad_norm": 0.0, "learning_rate": 1.1002328178936813e-05, "loss": 0.0, "step": 242 }, { "epoch": 0.06, "grad_norm": 0.0, "learning_rate": 1.0937858702281631e-05, "loss": 0.0, "step": 243 }, { "epoch": 0.06, "grad_norm": 0.0, "learning_rate": 1.087334987495364e-05, "loss": 0.0, "step": 244 }, { "epoch": 0.06, "grad_norm": 0.0, "learning_rate": 1.0808804403614044e-05, "loss": 0.0, "step": 245 }, { "epoch": 0.06, "grad_norm": 0.0, "learning_rate": 1.0744224996461541e-05, "loss": 0.0, "step": 246 }, { "epoch": 0.06, "grad_norm": 0.0, "learning_rate": 1.0679614363118718e-05, "loss": 0.0, "step": 247 }, { "epoch": 0.06, "grad_norm": 0.0, "learning_rate": 1.061497521451835e-05, "loss": 0.0, "step": 248 }, { "epoch": 0.06, "grad_norm": 0.0, "learning_rate": 1.055031026278965e-05, "loss": 0.0, "step": 249 }, { "epoch": 0.06, "grad_norm": 0.0, "learning_rate": 1.0485622221144485e-05, "loss": 0.0, "step": 250 }, { "epoch": 0.06, "grad_norm": 0.0, "learning_rate": 1.0420913803763522e-05, "loss": 0.0, "step": 251 }, { "epoch": 0.06, "grad_norm": 0.0, "learning_rate": 1.0356187725682359e-05, "loss": 0.0, "step": 252 }, { "epoch": 0.06, "grad_norm": 0.0, "learning_rate": 1.0291446702677598e-05, "loss": 0.0, "step": 253 }, { "epoch": 0.06, "grad_norm": 0.0, "learning_rate": 1.02266934511529e-05, "loss": 0.0, "step": 254 }, { "epoch": 0.06, "grad_norm": 0.0, "learning_rate": 1.0161930688025018e-05, "loss": 0.0, "step": 255 }, { "epoch": 0.06, "grad_norm": 0.0, "learning_rate": 1.0097161130609774e-05, "loss": 0.0, "step": 256 }, { "epoch": 0.06, "grad_norm": 0.0, "learning_rate": 1.003238749650809e-05, "loss": 0.0, "step": 257 }, { "epoch": 0.06, "grad_norm": 0.0, "learning_rate": 9.967612503491915e-06, "loss": 0.0, "step": 258 }, { "epoch": 0.06, "grad_norm": 0.0, "learning_rate": 9.90283886939023e-06, "loss": 0.0, "step": 259 }, { "epoch": 0.06, "grad_norm": 0.0, "learning_rate": 9.838069311974986e-06, "loss": 0.0, "step": 260 }, { "epoch": 0.06, "grad_norm": 0.0, "learning_rate": 9.773306548847102e-06, "loss": 0.0, "step": 261 }, { "epoch": 0.06, "grad_norm": 0.0, "learning_rate": 9.708553297322407e-06, "loss": 0.0, "step": 262 }, { "epoch": 0.06, "grad_norm": 0.0, "learning_rate": 9.643812274317644e-06, "loss": 0.0, "step": 263 }, { "epoch": 0.06, "grad_norm": 0.0, "learning_rate": 9.579086196236483e-06, "loss": 0.0, "step": 264 }, { "epoch": 0.06, "grad_norm": 0.0, "learning_rate": 9.514377778855521e-06, "loss": 0.0, "step": 265 }, { "epoch": 0.06, "grad_norm": 0.0, "learning_rate": 9.449689737210352e-06, "loss": 0.0, "step": 266 }, { "epoch": 0.06, "grad_norm": 0.0, "learning_rate": 9.385024785481653e-06, "loss": 0.0, "step": 267 }, { "epoch": 0.06, "grad_norm": 0.0, "learning_rate": 9.320385636881283e-06, "loss": 0.0, "step": 268 }, { "epoch": 0.06, "grad_norm": 0.0, "learning_rate": 9.255775003538462e-06, "loss": 0.0, "step": 269 }, { "epoch": 0.06, "grad_norm": 0.0, "learning_rate": 9.19119559638596e-06, "loss": 0.0, "step": 270 }, { "epoch": 0.06, "grad_norm": 0.0, "learning_rate": 9.126650125046361e-06, "loss": 0.0, "step": 271 }, { "epoch": 0.06, "grad_norm": 0.0, "learning_rate": 9.062141297718372e-06, "loss": 0.0, "step": 272 }, { "epoch": 0.06, "grad_norm": 0.0, "learning_rate": 8.99767182106319e-06, "loss": 0.0, "step": 273 }, { "epoch": 0.06, "grad_norm": 0.0, "learning_rate": 8.933244400090937e-06, "loss": 0.0, "step": 274 }, { "epoch": 0.06, "grad_norm": 0.0, "learning_rate": 8.868861738047158e-06, "loss": 0.0, "step": 275 }, { "epoch": 0.06, "grad_norm": 0.0, "learning_rate": 8.804526536299413e-06, "loss": 0.0, "step": 276 }, { "epoch": 0.06, "grad_norm": 0.0, "learning_rate": 8.740241494223911e-06, "loss": 0.0, "step": 277 }, { "epoch": 0.06, "grad_norm": 0.0, "learning_rate": 8.676009309092273e-06, "loss": 0.0, "step": 278 }, { "epoch": 0.06, "grad_norm": 0.0, "learning_rate": 8.611832675958335e-06, "loss": 0.0, "step": 279 }, { "epoch": 0.06, "grad_norm": 0.0, "learning_rate": 8.5477142875451e-06, "loss": 0.0, "step": 280 }, { "epoch": 0.06, "grad_norm": 0.0, "learning_rate": 8.48365683413172e-06, "loss": 0.0, "step": 281 }, { "epoch": 0.06, "grad_norm": 0.0, "learning_rate": 8.419663003440657e-06, "loss": 0.0, "step": 282 }, { "epoch": 0.06, "grad_norm": 0.0, "learning_rate": 8.355735480524874e-06, "loss": 0.0, "step": 283 }, { "epoch": 0.07, "grad_norm": 0.0, "learning_rate": 8.291876947655197e-06, "loss": 0.0, "step": 284 }, { "epoch": 0.07, "grad_norm": 0.0, "learning_rate": 8.228090084207773e-06, "loss": 0.0, "step": 285 }, { "epoch": 0.07, "grad_norm": 0.0, "learning_rate": 8.16437756655164e-06, "loss": 0.0, "step": 286 }, { "epoch": 0.07, "grad_norm": 0.0, "learning_rate": 8.100742067936432e-06, "loss": 0.0, "step": 287 }, { "epoch": 0.07, "grad_norm": 0.0, "learning_rate": 8.037186258380226e-06, "loss": 0.0, "step": 288 }, { "epoch": 0.07, "grad_norm": 0.0, "learning_rate": 7.9737128045575e-06, "loss": 0.0, "step": 289 }, { "epoch": 0.07, "grad_norm": 0.0, "learning_rate": 7.91032436968725e-06, "loss": 0.0, "step": 290 }, { "epoch": 0.07, "grad_norm": 0.0, "learning_rate": 7.847023613421251e-06, "loss": 0.0, "step": 291 }, { "epoch": 0.07, "grad_norm": 0.0, "learning_rate": 7.78381319173246e-06, "loss": 0.0, "step": 292 }, { "epoch": 0.07, "grad_norm": 0.0, "learning_rate": 7.720695756803569e-06, "loss": 0.0, "step": 293 }, { "epoch": 0.07, "grad_norm": 0.0, "learning_rate": 7.657673956915735e-06, "loss": 0.0, "step": 294 }, { "epoch": 0.07, "grad_norm": 0.0, "learning_rate": 7.594750436337467e-06, "loss": 0.0, "step": 295 }, { "epoch": 0.07, "grad_norm": 0.0, "learning_rate": 7.531927835213657e-06, "loss": 0.0, "step": 296 }, { "epoch": 0.07, "grad_norm": 0.0, "learning_rate": 7.469208789454838e-06, "loss": 0.0, "step": 297 }, { "epoch": 0.07, "grad_norm": 0.0, "learning_rate": 7.40659593062655e-06, "loss": 0.0, "step": 298 }, { "epoch": 0.07, "grad_norm": 0.0, "learning_rate": 7.344091885838949e-06, "loss": 0.0, "step": 299 }, { "epoch": 0.07, "grad_norm": 0.0, "learning_rate": 7.2816992776365714e-06, "loss": 0.0, "step": 300 }, { "epoch": 0.07, "grad_norm": 0.0, "learning_rate": 7.219420723888301e-06, "loss": 0.0, "step": 301 }, { "epoch": 0.07, "grad_norm": 0.0, "learning_rate": 7.157258837677514e-06, "loss": 0.0, "step": 302 }, { "epoch": 0.07, "grad_norm": 0.0, "learning_rate": 7.095216227192467e-06, "loss": 0.0, "step": 303 }, { "epoch": 0.07, "grad_norm": 0.0, "learning_rate": 7.033295495616834e-06, "loss": 0.0, "step": 304 }, { "epoch": 0.07, "grad_norm": 0.0, "learning_rate": 6.971499241020495e-06, "loss": 0.0, "step": 305 }, { "epoch": 0.07, "grad_norm": 0.0, "learning_rate": 6.909830056250527e-06, "loss": 0.0, "step": 306 }, { "epoch": 0.07, "grad_norm": 0.0, "learning_rate": 6.848290528822417e-06, "loss": 0.0, "step": 307 }, { "epoch": 0.07, "grad_norm": 0.0, "learning_rate": 6.786883240811479e-06, "loss": 0.0, "step": 308 }, { "epoch": 0.07, "grad_norm": 0.0, "learning_rate": 6.725610768744535e-06, "loss": 0.0, "step": 309 }, { "epoch": 0.07, "grad_norm": 0.0, "learning_rate": 6.664475683491797e-06, "loss": 0.0, "step": 310 }, { "epoch": 0.07, "grad_norm": 0.0, "learning_rate": 6.603480550158995e-06, "loss": 0.0, "step": 311 }, { "epoch": 0.07, "grad_norm": 0.0, "learning_rate": 6.542627927979772e-06, "loss": 0.0, "step": 312 }, { "epoch": 0.07, "grad_norm": 0.0, "learning_rate": 6.481920370208274e-06, "loss": 0.0, "step": 313 }, { "epoch": 0.07, "grad_norm": 0.0, "learning_rate": 6.421360424012039e-06, "loss": 0.0, "step": 314 }, { "epoch": 0.07, "grad_norm": 0.0, "learning_rate": 6.360950630365126e-06, "loss": 0.0, "step": 315 }, { "epoch": 0.07, "grad_norm": 0.0, "learning_rate": 6.300693523941481e-06, "loss": 0.0, "step": 316 }, { "epoch": 0.07, "grad_norm": 0.0, "learning_rate": 6.2405916330086106e-06, "loss": 0.0, "step": 317 }, { "epoch": 0.07, "grad_norm": 0.0, "learning_rate": 6.180647479321484e-06, "loss": 0.0, "step": 318 }, { "epoch": 0.07, "grad_norm": 0.0, "learning_rate": 6.120863578016736e-06, "loss": 0.0, "step": 319 }, { "epoch": 0.07, "grad_norm": 0.0, "learning_rate": 6.061242437507131e-06, "loss": 0.0, "step": 320 }, { "epoch": 0.07, "grad_norm": 0.0, "learning_rate": 6.00178655937631e-06, "loss": 0.0, "step": 321 }, { "epoch": 0.07, "grad_norm": 0.0, "learning_rate": 5.942498438273849e-06, "loss": 0.0, "step": 322 }, { "epoch": 0.07, "grad_norm": 0.0, "learning_rate": 5.8833805618105635e-06, "loss": 0.0, "step": 323 }, { "epoch": 0.07, "grad_norm": 0.0, "learning_rate": 5.82443541045415e-06, "loss": 0.0, "step": 324 }, { "epoch": 0.07, "grad_norm": 0.0, "learning_rate": 5.765665457425102e-06, "loss": 0.0, "step": 325 }, { "epoch": 0.07, "grad_norm": 0.0, "learning_rate": 5.707073168592943e-06, "loss": 0.0, "step": 326 }, { "epoch": 0.07, "grad_norm": 0.0, "learning_rate": 5.648661002372769e-06, "loss": 0.0, "step": 327 }, { "epoch": 0.08, "grad_norm": 0.0, "learning_rate": 5.590431409622081e-06, "loss": 0.0, "step": 328 }, { "epoch": 0.08, "grad_norm": 0.0, "learning_rate": 5.5323868335379775e-06, "loss": 0.0, "step": 329 }, { "epoch": 0.08, "grad_norm": 0.0, "learning_rate": 5.4745297095546125e-06, "loss": 0.0, "step": 330 }, { "epoch": 0.08, "grad_norm": 0.0, "learning_rate": 5.416862465241033e-06, "loss": 0.0, "step": 331 }, { "epoch": 0.08, "grad_norm": 0.0, "learning_rate": 5.359387520199317e-06, "loss": 0.0, "step": 332 }, { "epoch": 0.08, "grad_norm": 0.0, "learning_rate": 5.302107285963045e-06, "loss": 0.0, "step": 333 }, { "epoch": 0.08, "grad_norm": 0.0, "learning_rate": 5.245024165896126e-06, "loss": 0.0, "step": 334 }, { "epoch": 0.08, "grad_norm": 0.0, "learning_rate": 5.18814055509195e-06, "loss": 0.0, "step": 335 }, { "epoch": 0.08, "grad_norm": 0.0, "learning_rate": 5.131458840272905e-06, "loss": 0.0, "step": 336 }, { "epoch": 0.08, "grad_norm": 0.0, "learning_rate": 5.074981399690219e-06, "loss": 0.0, "step": 337 }, { "epoch": 0.08, "grad_norm": 0.0, "learning_rate": 5.018710603024187e-06, "loss": 0.0, "step": 338 }, { "epoch": 0.08, "grad_norm": 0.0, "learning_rate": 4.9626488112847384e-06, "loss": 0.0, "step": 339 }, { "epoch": 0.08, "grad_norm": 0.0, "learning_rate": 4.9067983767123736e-06, "loss": 0.0, "step": 340 }, { "epoch": 0.08, "grad_norm": 0.0, "learning_rate": 4.851161642679466e-06, "loss": 0.0, "step": 341 }, { "epoch": 0.08, "grad_norm": 0.0, "learning_rate": 4.795740943591955e-06, "loss": 0.0, "step": 342 }, { "epoch": 0.08, "grad_norm": 0.0, "learning_rate": 4.740538604791371e-06, "loss": 0.0, "step": 343 }, { "epoch": 0.08, "grad_norm": 0.0, "learning_rate": 4.685556942457296e-06, "loss": 0.0, "step": 344 }, { "epoch": 0.08, "grad_norm": 0.0, "learning_rate": 4.630798263510162e-06, "loss": 0.0, "step": 345 }, { "epoch": 0.08, "grad_norm": 0.0, "learning_rate": 4.576264865514467e-06, "loss": 0.0, "step": 346 }, { "epoch": 0.08, "grad_norm": 0.0, "learning_rate": 4.521959036582372e-06, "loss": 0.0, "step": 347 }, { "epoch": 0.08, "grad_norm": 0.0, "learning_rate": 4.467883055277696e-06, "loss": 0.0, "step": 348 }, { "epoch": 0.08, "grad_norm": 0.0, "learning_rate": 4.414039190520308e-06, "loss": 0.0, "step": 349 }, { "epoch": 0.08, "grad_norm": 0.0, "learning_rate": 4.360429701490935e-06, "loss": 0.0, "step": 350 }, { "epoch": 0.08, "grad_norm": 0.0, "learning_rate": 4.307056837536373e-06, "loss": 0.0, "step": 351 }, { "epoch": 0.08, "grad_norm": 0.0, "learning_rate": 4.2539228380750955e-06, "loss": 0.0, "step": 352 }, { "epoch": 0.08, "grad_norm": 0.0, "learning_rate": 4.201029932503303e-06, "loss": 0.0, "step": 353 }, { "epoch": 0.08, "grad_norm": 0.0, "learning_rate": 4.14838034010138e-06, "loss": 0.0, "step": 354 }, { "epoch": 0.08, "grad_norm": 0.0, "learning_rate": 4.095976269940777e-06, "loss": 0.0, "step": 355 }, { "epoch": 0.08, "grad_norm": 0.0, "learning_rate": 4.043819920791322e-06, "loss": 0.0, "step": 356 }, { "epoch": 0.08, "grad_norm": 0.0, "learning_rate": 3.991913481028965e-06, "loss": 0.0, "step": 357 }, { "epoch": 0.08, "grad_norm": 0.0, "learning_rate": 3.940259128543967e-06, "loss": 0.0, "step": 358 }, { "epoch": 0.08, "grad_norm": 0.0, "learning_rate": 3.888859030649498e-06, "loss": 0.0, "step": 359 }, { "epoch": 0.08, "grad_norm": 0.0, "learning_rate": 3.837715343990727e-06, "loss": 0.0, "step": 360 }, { "epoch": 0.08, "grad_norm": 0.0, "learning_rate": 3.7868302144543146e-06, "loss": 0.0, "step": 361 }, { "epoch": 0.08, "grad_norm": 0.0, "learning_rate": 3.736205777078381e-06, "loss": 0.0, "step": 362 }, { "epoch": 0.08, "grad_norm": 0.0, "learning_rate": 3.685844155962931e-06, "loss": 0.0, "step": 363 }, { "epoch": 0.08, "grad_norm": 0.0, "learning_rate": 3.63574746418072e-06, "loss": 0.0, "step": 364 }, { "epoch": 0.08, "grad_norm": 0.0, "learning_rate": 3.585917803688603e-06, "loss": 0.0, "step": 365 }, { "epoch": 0.08, "grad_norm": 0.0, "learning_rate": 3.536357265239333e-06, "loss": 0.0, "step": 366 }, { "epoch": 0.08, "grad_norm": 0.0, "learning_rate": 3.487067928293848e-06, "loss": 0.0, "step": 367 }, { "epoch": 0.08, "grad_norm": 0.0, "learning_rate": 3.4380518609340076e-06, "loss": 0.0, "step": 368 }, { "epoch": 0.08, "grad_norm": 0.0, "learning_rate": 3.3893111197758276e-06, "loss": 0.0, "step": 369 }, { "epoch": 0.08, "grad_norm": 0.0, "learning_rate": 3.3408477498831917e-06, "loss": 0.0, "step": 370 }, { "epoch": 0.09, "grad_norm": 0.0, "learning_rate": 3.2926637846820366e-06, "loss": 0.0, "step": 371 }, { "epoch": 0.09, "grad_norm": 0.0, "learning_rate": 3.2447612458750365e-06, "loss": 0.0, "step": 372 }, { "epoch": 0.09, "grad_norm": 0.0, "learning_rate": 3.197142143356787e-06, "loss": 0.0, "step": 373 }, { "epoch": 0.09, "grad_norm": 0.0, "learning_rate": 3.1498084751294523e-06, "loss": 0.0, "step": 374 }, { "epoch": 0.09, "grad_norm": 0.0, "learning_rate": 3.1027622272189572e-06, "loss": 0.0, "step": 375 }, { "epoch": 0.09, "grad_norm": 0.0, "learning_rate": 3.0560053735916372e-06, "loss": 0.0, "step": 376 }, { "epoch": 0.09, "grad_norm": 0.0, "learning_rate": 3.009539876071427e-06, "loss": 0.0, "step": 377 }, { "epoch": 0.09, "grad_norm": 0.0, "learning_rate": 2.9633676842575386e-06, "loss": 0.0, "step": 378 }, { "epoch": 0.09, "grad_norm": 0.0, "learning_rate": 2.9174907354426696e-06, "loss": 0.0, "step": 379 }, { "epoch": 0.09, "grad_norm": 0.0, "learning_rate": 2.8719109545317102e-06, "loss": 0.0, "step": 380 }, { "epoch": 0.09, "grad_norm": 0.0, "learning_rate": 2.8266302539609747e-06, "loss": 0.0, "step": 381 }, { "epoch": 0.09, "grad_norm": 0.0, "learning_rate": 2.78165053361798e-06, "loss": 0.0, "step": 382 }, { "epoch": 0.09, "grad_norm": 0.0, "learning_rate": 2.736973680761702e-06, "loss": 0.0, "step": 383 }, { "epoch": 0.09, "grad_norm": 0.0, "learning_rate": 2.692601569943407e-06, "loss": 0.0, "step": 384 }, { "epoch": 0.09, "grad_norm": 0.0, "learning_rate": 2.648536062927999e-06, "loss": 0.0, "step": 385 }, { "epoch": 0.09, "grad_norm": 0.0, "learning_rate": 2.604779008615895e-06, "loss": 0.0, "step": 386 }, { "epoch": 0.09, "grad_norm": 0.0, "learning_rate": 2.5613322429654573e-06, "loss": 0.0, "step": 387 }, { "epoch": 0.09, "grad_norm": 0.0, "learning_rate": 2.5181975889159615e-06, "loss": 0.0, "step": 388 }, { "epoch": 0.09, "grad_norm": 0.0, "learning_rate": 2.475376856311097e-06, "loss": 0.0, "step": 389 }, { "epoch": 0.09, "grad_norm": 0.0, "learning_rate": 2.432871841823047e-06, "loss": 0.0, "step": 390 }, { "epoch": 0.09, "grad_norm": 0.0, "learning_rate": 2.390684328877089e-06, "loss": 0.0, "step": 391 }, { "epoch": 0.09, "grad_norm": 0.0, "learning_rate": 2.3488160875767717e-06, "loss": 0.0, "step": 392 }, { "epoch": 0.09, "grad_norm": 0.0, "learning_rate": 2.307268874629649e-06, "loss": 0.0, "step": 393 }, { "epoch": 0.09, "grad_norm": 0.0, "learning_rate": 2.266044433273562e-06, "loss": 0.0, "step": 394 }, { "epoch": 0.09, "grad_norm": 0.0, "learning_rate": 2.2251444932035094e-06, "loss": 0.0, "step": 395 }, { "epoch": 0.09, "grad_norm": 0.0, "learning_rate": 2.184570770499056e-06, "loss": 0.0, "step": 396 }, { "epoch": 0.09, "grad_norm": 0.0, "learning_rate": 2.1443249675523536e-06, "loss": 0.0, "step": 397 }, { "epoch": 0.09, "grad_norm": 0.0, "learning_rate": 2.1044087729966856e-06, "loss": 0.0, "step": 398 }, { "epoch": 0.09, "grad_norm": 0.0, "learning_rate": 2.064823861635633e-06, "loss": 0.0, "step": 399 }, { "epoch": 0.09, "grad_norm": 0.0, "learning_rate": 2.025571894372794e-06, "loss": 0.0, "step": 400 }, { "epoch": 0.09, "grad_norm": 0.0, "learning_rate": 1.9866545181421016e-06, "loss": 0.0, "step": 401 }, { "epoch": 0.09, "grad_norm": 0.0, "learning_rate": 1.9480733658387175e-06, "loss": 0.0, "step": 402 }, { "epoch": 0.09, "grad_norm": 0.0, "learning_rate": 1.9098300562505266e-06, "loss": 0.0, "step": 403 }, { "epoch": 0.09, "grad_norm": 0.0, "learning_rate": 1.8719261939902023e-06, "loss": 0.0, "step": 404 }, { "epoch": 0.09, "grad_norm": 0.0, "learning_rate": 1.8343633694278895e-06, "loss": 0.0, "step": 405 }, { "epoch": 0.09, "grad_norm": 0.0, "learning_rate": 1.7971431586244814e-06, "loss": 0.0, "step": 406 }, { "epoch": 0.09, "grad_norm": 0.0, "learning_rate": 1.7602671232654755e-06, "loss": 0.0, "step": 407 }, { "epoch": 0.09, "grad_norm": 0.0, "learning_rate": 1.723736810595461e-06, "loss": 0.0, "step": 408 }, { "epoch": 0.09, "grad_norm": 0.0, "learning_rate": 1.687553753353195e-06, "loss": 0.0, "step": 409 }, { "epoch": 0.09, "grad_norm": 0.0, "learning_rate": 1.6517194697072903e-06, "loss": 0.0, "step": 410 }, { "epoch": 0.09, "grad_norm": 0.0, "learning_rate": 1.6162354631925203e-06, "loss": 0.0, "step": 411 }, { "epoch": 0.09, "grad_norm": 0.0, "learning_rate": 1.5811032226467304e-06, "loss": 0.0, "step": 412 }, { "epoch": 0.09, "grad_norm": 0.0, "learning_rate": 1.5463242221483742e-06, "loss": 0.0, "step": 413 }, { "epoch": 0.09, "grad_norm": 0.0, "learning_rate": 1.511899920954656e-06, "loss": 0.0, "step": 414 }, { "epoch": 0.1, "grad_norm": 0.0, "learning_rate": 1.4778317634403082e-06, "loss": 0.0, "step": 415 }, { "epoch": 0.1, "grad_norm": 0.0, "learning_rate": 1.4441211790369892e-06, "loss": 0.0, "step": 416 }, { "epoch": 0.1, "grad_norm": 0.0, "learning_rate": 1.4107695821733026e-06, "loss": 0.0, "step": 417 }, { "epoch": 0.1, "grad_norm": 0.0, "learning_rate": 1.3777783722154603e-06, "loss": 0.0, "step": 418 }, { "epoch": 0.1, "grad_norm": 0.0, "learning_rate": 1.3451489334085555e-06, "loss": 0.0, "step": 419 }, { "epoch": 0.1, "grad_norm": 0.0, "learning_rate": 1.3128826348184886e-06, "loss": 0.0, "step": 420 }, { "epoch": 0.1, "grad_norm": 0.0, "learning_rate": 1.2809808302745298e-06, "loss": 0.0, "step": 421 }, { "epoch": 0.1, "grad_norm": 0.0, "learning_rate": 1.249444858312502e-06, "loss": 0.0, "step": 422 }, { "epoch": 0.1, "grad_norm": 0.0, "learning_rate": 1.218276042118629e-06, "loss": 0.0, "step": 423 }, { "epoch": 0.1, "grad_norm": 0.0, "learning_rate": 1.1874756894740137e-06, "loss": 0.0, "step": 424 }, { "epoch": 0.1, "grad_norm": 0.0, "learning_rate": 1.1570450926997657e-06, "loss": 0.0, "step": 425 }, { "epoch": 0.1, "grad_norm": 0.0, "learning_rate": 1.1269855286027798e-06, "loss": 0.0, "step": 426 }, { "epoch": 0.1, "grad_norm": 0.0, "learning_rate": 1.0972982584221592e-06, "loss": 0.0, "step": 427 }, { "epoch": 0.1, "grad_norm": 0.0, "learning_rate": 1.067984527776309e-06, "loss": 0.0, "step": 428 }, { "epoch": 0.1, "grad_norm": 0.0, "learning_rate": 1.0390455666106547e-06, "loss": 0.0, "step": 429 }, { "epoch": 0.1, "grad_norm": 0.0, "learning_rate": 1.010482589146048e-06, "loss": 0.0, "step": 430 }, { "epoch": 0.1, "grad_norm": 0.0, "learning_rate": 9.822967938278172e-07, "loss": 0.0, "step": 431 }, { "epoch": 0.1, "grad_norm": 0.0, "learning_rate": 9.544893632754816e-07, "loss": 0.0, "step": 432 }, { "epoch": 0.1, "grad_norm": 0.0, "learning_rate": 9.270614642331377e-07, "loss": 0.0, "step": 433 }, { "epoch": 0.1, "grad_norm": 0.0, "learning_rate": 9.000142475204965e-07, "loss": 0.0, "step": 434 }, { "epoch": 0.1, "grad_norm": 0.0, "learning_rate": 8.733488479845997e-07, "loss": 0.0, "step": 435 }, { "epoch": 0.1, "grad_norm": 0.0, "learning_rate": 8.470663844522053e-07, "loss": 0.0, "step": 436 }, { "epoch": 0.1, "grad_norm": 0.0, "learning_rate": 8.211679596828481e-07, "loss": 0.0, "step": 437 }, { "epoch": 0.1, "grad_norm": 0.0, "learning_rate": 7.956546603225601e-07, "loss": 0.0, "step": 438 }, { "epoch": 0.1, "grad_norm": 0.0, "learning_rate": 7.705275568582848e-07, "loss": 0.0, "step": 439 }, { "epoch": 0.1, "grad_norm": 0.0, "learning_rate": 7.457877035729588e-07, "loss": 0.0, "step": 440 }, { "epoch": 0.1, "grad_norm": 0.0, "learning_rate": 7.21436138501278e-07, "loss": 0.0, "step": 441 }, { "epoch": 0.1, "grad_norm": 0.0, "learning_rate": 6.974738833861383e-07, "loss": 0.0, "step": 442 }, { "epoch": 0.1, "grad_norm": 0.0, "learning_rate": 6.739019436357774e-07, "loss": 0.0, "step": 443 }, { "epoch": 0.1, "grad_norm": 0.0, "learning_rate": 6.507213082815745e-07, "loss": 0.0, "step": 444 }, { "epoch": 0.1, "grad_norm": 0.0, "learning_rate": 6.279329499365649e-07, "loss": 0.0, "step": 445 }, { "epoch": 0.1, "grad_norm": 0.0, "learning_rate": 6.055378247546217e-07, "loss": 0.0, "step": 446 }, { "epoch": 0.1, "grad_norm": 0.0, "learning_rate": 5.835368723903456e-07, "loss": 0.0, "step": 447 }, { "epoch": 0.1, "grad_norm": 0.0, "learning_rate": 5.619310159596358e-07, "loss": 0.0, "step": 448 }, { "epoch": 0.1, "grad_norm": 0.0, "learning_rate": 5.407211620009545e-07, "loss": 0.0, "step": 449 }, { "epoch": 0.1, "grad_norm": 0.0, "learning_rate": 5.199082004372958e-07, "loss": 0.0, "step": 450 }, { "epoch": 0.1, "grad_norm": 0.0, "learning_rate": 4.994930045388414e-07, "loss": 0.0, "step": 451 }, { "epoch": 0.1, "grad_norm": 0.0, "learning_rate": 4.794764308863242e-07, "loss": 0.0, "step": 452 }, { "epoch": 0.1, "grad_norm": 0.0, "learning_rate": 4.5985931933508757e-07, "loss": 0.0, "step": 453 }, { "epoch": 0.1, "grad_norm": 0.0, "learning_rate": 4.406424929798403e-07, "loss": 0.0, "step": 454 }, { "epoch": 0.1, "grad_norm": 0.0, "learning_rate": 4.218267581201296e-07, "loss": 0.0, "step": 455 }, { "epoch": 0.1, "grad_norm": 0.0, "learning_rate": 4.034129042265067e-07, "loss": 0.0, "step": 456 }, { "epoch": 0.1, "grad_norm": 0.0, "learning_rate": 3.8540170390740097e-07, "loss": 0.0, "step": 457 }, { "epoch": 0.11, "grad_norm": 0.0, "learning_rate": 3.67793912876705e-07, "loss": 0.0, "step": 458 }, { "epoch": 0.11, "grad_norm": 0.0, "learning_rate": 3.5059026992206645e-07, "loss": 0.0, "step": 459 }, { "epoch": 0.11, "grad_norm": 0.0, "learning_rate": 3.3379149687388866e-07, "loss": 0.0, "step": 460 }, { "epoch": 0.11, "grad_norm": 0.0, "learning_rate": 3.1739829857504235e-07, "loss": 0.0, "step": 461 }, { "epoch": 0.11, "grad_norm": 0.0, "learning_rate": 3.0141136285129825e-07, "loss": 0.0, "step": 462 }, { "epoch": 0.11, "grad_norm": 0.0, "learning_rate": 2.8583136048245697e-07, "loss": 0.0, "step": 463 }, { "epoch": 0.11, "grad_norm": 0.0, "learning_rate": 2.706589451742181e-07, "loss": 0.0, "step": 464 }, { "epoch": 0.11, "grad_norm": 0.0, "learning_rate": 2.5589475353073987e-07, "loss": 0.0, "step": 465 }, { "epoch": 0.11, "grad_norm": 0.0, "learning_rate": 2.4153940502793185e-07, "loss": 0.0, "step": 466 }, { "epoch": 0.11, "grad_norm": 0.0, "learning_rate": 2.2759350198746978e-07, "loss": 0.0, "step": 467 }, { "epoch": 0.11, "grad_norm": 0.0, "learning_rate": 2.1405762955151178e-07, "loss": 0.0, "step": 468 }, { "epoch": 0.11, "grad_norm": 0.0, "learning_rate": 2.009323556581566e-07, "loss": 0.0, "step": 469 }, { "epoch": 0.11, "grad_norm": 0.0, "learning_rate": 1.8821823101760949e-07, "loss": 0.0, "step": 470 }, { "epoch": 0.11, "grad_norm": 0.0, "learning_rate": 1.7591578908907724e-07, "loss": 0.0, "step": 471 }, { "epoch": 0.11, "grad_norm": 0.0, "learning_rate": 1.6402554605838173e-07, "loss": 0.0, "step": 472 }, { "epoch": 0.11, "grad_norm": 0.0, "learning_rate": 1.5254800081630828e-07, "loss": 0.0, "step": 473 }, { "epoch": 0.11, "grad_norm": 0.0, "learning_rate": 1.4148363493766803e-07, "loss": 0.0, "step": 474 }, { "epoch": 0.11, "grad_norm": 0.0, "learning_rate": 1.30832912661093e-07, "loss": 0.0, "step": 475 }, { "epoch": 0.11, "grad_norm": 0.0, "learning_rate": 1.2059628086956044e-07, "loss": 0.0, "step": 476 }, { "epoch": 0.11, "grad_norm": 0.0, "learning_rate": 1.1077416907163573e-07, "loss": 0.0, "step": 477 }, { "epoch": 0.11, "grad_norm": 0.0, "learning_rate": 1.0136698938346012e-07, "loss": 0.0, "step": 478 }, { "epoch": 0.11, "grad_norm": 0.0, "learning_rate": 9.237513651145224e-08, "loss": 0.0, "step": 479 }, { "epoch": 0.11, "grad_norm": 0.0, "learning_rate": 8.379898773574924e-08, "loss": 0.0, "step": 480 }, { "epoch": 0.11, "grad_norm": 0.0, "learning_rate": 7.563890289437825e-08, "loss": 0.0, "step": 481 }, { "epoch": 0.11, "grad_norm": 0.0, "learning_rate": 6.78952243681541e-08, "loss": 0.0, "step": 482 }, { "epoch": 0.11, "grad_norm": 0.0, "learning_rate": 6.056827706632185e-08, "loss": 0.0, "step": 483 }, { "epoch": 0.11, "grad_norm": 0.0, "learning_rate": 5.365836841291439e-08, "loss": 0.0, "step": 484 }, { "epoch": 0.11, "grad_norm": 0.0, "learning_rate": 4.716578833386054e-08, "loss": 0.0, "step": 485 }, { "epoch": 0.11, "grad_norm": 0.0, "learning_rate": 4.109080924481479e-08, "loss": 0.0, "step": 486 }, { "epoch": 0.11, "grad_norm": 0.0, "learning_rate": 3.543368603973529e-08, "loss": 0.0, "step": 487 }, { "epoch": 0.11, "grad_norm": 0.0, "learning_rate": 3.019465608018024e-08, "loss": 0.0, "step": 488 }, { "epoch": 0.11, "grad_norm": 0.0, "learning_rate": 2.537393918535358e-08, "loss": 0.0, "step": 489 }, { "epoch": 0.11, "grad_norm": 0.0, "learning_rate": 2.0971737622883515e-08, "loss": 0.0, "step": 490 }, { "epoch": 0.11, "grad_norm": 0.0, "learning_rate": 1.698823610032929e-08, "loss": 0.0, "step": 491 }, { "epoch": 0.11, "grad_norm": 0.0, "learning_rate": 1.3423601757436289e-08, "loss": 0.0, "step": 492 }, { "epoch": 0.11, "grad_norm": 0.0, "learning_rate": 1.0277984159122734e-08, "loss": 0.0, "step": 493 }, { "epoch": 0.11, "grad_norm": 0.0, "learning_rate": 7.551515289203615e-09, "loss": 0.0, "step": 494 }, { "epoch": 0.11, "grad_norm": 0.0, "learning_rate": 5.2443095448506674e-09, "loss": 0.0, "step": 495 }, { "epoch": 0.11, "grad_norm": 0.0, "learning_rate": 3.3564637317984318e-09, "loss": 0.0, "step": 496 }, { "epoch": 0.11, "grad_norm": 0.0, "learning_rate": 1.888057060274173e-09, "loss": 0.0, "step": 497 }, { "epoch": 0.11, "grad_norm": 0.0, "learning_rate": 8.391511416816489e-10, "loss": 0.0, "step": 498 }, { "epoch": 0.11, "grad_norm": 0.0, "learning_rate": 2.0978998601206558e-10, "loss": 0.0, "step": 499 }, { "epoch": 0.11, "grad_norm": 0.0, "learning_rate": 0.0, "loss": 0.0, "step": 500 }, { "epoch": 0.11, "step": 500, "total_flos": 8031892119552.0, "train_loss": 0.0, "train_runtime": 40433.9748, "train_samples_per_second": 1.583, "train_steps_per_second": 0.012 } ], "logging_steps": 1.0, "max_steps": 500, "num_input_tokens_seen": 0, "num_train_epochs": 1, "save_steps": 5000, "total_flos": 8031892119552.0, "train_batch_size": 32, "trial_name": null, "trial_params": null }