| { | |
| "best_metric": 2.8477718830108643, | |
| "best_model_checkpoint": "miner_id_24/checkpoint-150", | |
| "epoch": 0.11417697431018078, | |
| "eval_steps": 50, | |
| "global_step": 150, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0007611798287345385, | |
| "grad_norm": 62.14710235595703, | |
| "learning_rate": 7e-06, | |
| "loss": 11.9374, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.0007611798287345385, | |
| "eval_loss": 3.837273597717285, | |
| "eval_runtime": 157.4082, | |
| "eval_samples_per_second": 14.059, | |
| "eval_steps_per_second": 3.52, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.001522359657469077, | |
| "grad_norm": 32.307308197021484, | |
| "learning_rate": 1.4e-05, | |
| "loss": 12.6976, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.002283539486203616, | |
| "grad_norm": 26.411426544189453, | |
| "learning_rate": 2.1e-05, | |
| "loss": 13.2102, | |
| "step": 3 | |
| }, | |
| { | |
| "epoch": 0.003044719314938154, | |
| "grad_norm": 26.69647789001465, | |
| "learning_rate": 2.8e-05, | |
| "loss": 12.9265, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.003805899143672693, | |
| "grad_norm": 21.253070831298828, | |
| "learning_rate": 3.5e-05, | |
| "loss": 12.6845, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.004567078972407232, | |
| "grad_norm": 18.088212966918945, | |
| "learning_rate": 4.2e-05, | |
| "loss": 12.0568, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.00532825880114177, | |
| "grad_norm": 17.1400089263916, | |
| "learning_rate": 4.899999999999999e-05, | |
| "loss": 11.9566, | |
| "step": 7 | |
| }, | |
| { | |
| "epoch": 0.006089438629876308, | |
| "grad_norm": 17.086406707763672, | |
| "learning_rate": 5.6e-05, | |
| "loss": 11.8525, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.0068506184586108465, | |
| "grad_norm": 16.366493225097656, | |
| "learning_rate": 6.3e-05, | |
| "loss": 11.2577, | |
| "step": 9 | |
| }, | |
| { | |
| "epoch": 0.007611798287345386, | |
| "grad_norm": 14.400430679321289, | |
| "learning_rate": 7e-05, | |
| "loss": 11.5039, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.008372978116079923, | |
| "grad_norm": 14.400092124938965, | |
| "learning_rate": 6.999521567473641e-05, | |
| "loss": 10.997, | |
| "step": 11 | |
| }, | |
| { | |
| "epoch": 0.009134157944814463, | |
| "grad_norm": 17.782190322875977, | |
| "learning_rate": 6.998086400693241e-05, | |
| "loss": 11.153, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 0.009895337773549001, | |
| "grad_norm": 15.323582649230957, | |
| "learning_rate": 6.995694892019065e-05, | |
| "loss": 11.3713, | |
| "step": 13 | |
| }, | |
| { | |
| "epoch": 0.01065651760228354, | |
| "grad_norm": 18.582822799682617, | |
| "learning_rate": 6.99234769526571e-05, | |
| "loss": 11.6407, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 0.011417697431018078, | |
| "grad_norm": 18.919496536254883, | |
| "learning_rate": 6.988045725523343e-05, | |
| "loss": 11.4483, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.012178877259752616, | |
| "grad_norm": 15.953317642211914, | |
| "learning_rate": 6.982790158907539e-05, | |
| "loss": 11.4306, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 0.012940057088487155, | |
| "grad_norm": 34.4675178527832, | |
| "learning_rate": 6.976582432237733e-05, | |
| "loss": 11.3819, | |
| "step": 17 | |
| }, | |
| { | |
| "epoch": 0.013701236917221693, | |
| "grad_norm": 16.293533325195312, | |
| "learning_rate": 6.969424242644413e-05, | |
| "loss": 11.3279, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 0.014462416745956231, | |
| "grad_norm": 14.284806251525879, | |
| "learning_rate": 6.961317547105138e-05, | |
| "loss": 10.9124, | |
| "step": 19 | |
| }, | |
| { | |
| "epoch": 0.015223596574690771, | |
| "grad_norm": 16.047115325927734, | |
| "learning_rate": 6.952264561909527e-05, | |
| "loss": 11.1305, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.01598477640342531, | |
| "grad_norm": 16.61277961730957, | |
| "learning_rate": 6.942267762053337e-05, | |
| "loss": 11.2785, | |
| "step": 21 | |
| }, | |
| { | |
| "epoch": 0.016745956232159846, | |
| "grad_norm": 16.367738723754883, | |
| "learning_rate": 6.931329880561832e-05, | |
| "loss": 11.1815, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 0.017507136060894386, | |
| "grad_norm": 16.200246810913086, | |
| "learning_rate": 6.919453907742597e-05, | |
| "loss": 11.2894, | |
| "step": 23 | |
| }, | |
| { | |
| "epoch": 0.018268315889628926, | |
| "grad_norm": 15.94240665435791, | |
| "learning_rate": 6.90664309036802e-05, | |
| "loss": 10.8494, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 0.019029495718363463, | |
| "grad_norm": 15.30321216583252, | |
| "learning_rate": 6.892900930787656e-05, | |
| "loss": 10.9201, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.019790675547098003, | |
| "grad_norm": 18.01148223876953, | |
| "learning_rate": 6.87823118597072e-05, | |
| "loss": 11.3518, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 0.02055185537583254, | |
| "grad_norm": 16.691192626953125, | |
| "learning_rate": 6.862637866478969e-05, | |
| "loss": 10.8088, | |
| "step": 27 | |
| }, | |
| { | |
| "epoch": 0.02131303520456708, | |
| "grad_norm": 16.822267532348633, | |
| "learning_rate": 6.846125235370252e-05, | |
| "loss": 11.7738, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 0.022074215033301616, | |
| "grad_norm": 17.422344207763672, | |
| "learning_rate": 6.828697807033038e-05, | |
| "loss": 11.9039, | |
| "step": 29 | |
| }, | |
| { | |
| "epoch": 0.022835394862036156, | |
| "grad_norm": 15.227424621582031, | |
| "learning_rate": 6.81036034595222e-05, | |
| "loss": 10.5233, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.023596574690770696, | |
| "grad_norm": 16.58725929260254, | |
| "learning_rate": 6.791117865406564e-05, | |
| "loss": 10.2275, | |
| "step": 31 | |
| }, | |
| { | |
| "epoch": 0.024357754519505233, | |
| "grad_norm": 15.96583080291748, | |
| "learning_rate": 6.770975626098112e-05, | |
| "loss": 10.0431, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 0.025118934348239773, | |
| "grad_norm": 16.036426544189453, | |
| "learning_rate": 6.749939134713974e-05, | |
| "loss": 9.8158, | |
| "step": 33 | |
| }, | |
| { | |
| "epoch": 0.02588011417697431, | |
| "grad_norm": 15.925104141235352, | |
| "learning_rate": 6.728014142420846e-05, | |
| "loss": 9.9834, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 0.02664129400570885, | |
| "grad_norm": 17.323747634887695, | |
| "learning_rate": 6.7052066432927e-05, | |
| "loss": 10.1752, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.027402473834443386, | |
| "grad_norm": 17.286712646484375, | |
| "learning_rate": 6.681522872672069e-05, | |
| "loss": 10.1717, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 0.028163653663177926, | |
| "grad_norm": 21.032140731811523, | |
| "learning_rate": 6.656969305465356e-05, | |
| "loss": 9.4869, | |
| "step": 37 | |
| }, | |
| { | |
| "epoch": 0.028924833491912463, | |
| "grad_norm": 18.7946834564209, | |
| "learning_rate": 6.631552654372672e-05, | |
| "loss": 10.5547, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 0.029686013320647003, | |
| "grad_norm": 19.529447555541992, | |
| "learning_rate": 6.60527986805264e-05, | |
| "loss": 10.4017, | |
| "step": 39 | |
| }, | |
| { | |
| "epoch": 0.030447193149381543, | |
| "grad_norm": 18.256792068481445, | |
| "learning_rate": 6.578158129222711e-05, | |
| "loss": 9.9321, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.03120837297811608, | |
| "grad_norm": 19.373693466186523, | |
| "learning_rate": 6.550194852695469e-05, | |
| "loss": 10.2322, | |
| "step": 41 | |
| }, | |
| { | |
| "epoch": 0.03196955280685062, | |
| "grad_norm": 19.30985450744629, | |
| "learning_rate": 6.521397683351509e-05, | |
| "loss": 10.274, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 0.03273073263558516, | |
| "grad_norm": 19.986684799194336, | |
| "learning_rate": 6.491774494049386e-05, | |
| "loss": 9.9905, | |
| "step": 43 | |
| }, | |
| { | |
| "epoch": 0.03349191246431969, | |
| "grad_norm": 20.237152099609375, | |
| "learning_rate": 6.461333383473272e-05, | |
| "loss": 10.3248, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 0.03425309229305423, | |
| "grad_norm": 16.88189125061035, | |
| "learning_rate": 6.430082673918849e-05, | |
| "loss": 9.3653, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 0.03501427212178877, | |
| "grad_norm": 25.35614776611328, | |
| "learning_rate": 6.398030909018069e-05, | |
| "loss": 10.3682, | |
| "step": 46 | |
| }, | |
| { | |
| "epoch": 0.03577545195052331, | |
| "grad_norm": 22.58070945739746, | |
| "learning_rate": 6.365186851403423e-05, | |
| "loss": 10.2078, | |
| "step": 47 | |
| }, | |
| { | |
| "epoch": 0.03653663177925785, | |
| "grad_norm": 18.730236053466797, | |
| "learning_rate": 6.331559480312315e-05, | |
| "loss": 9.5991, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 0.037297811607992386, | |
| "grad_norm": 24.788707733154297, | |
| "learning_rate": 6.297157989132236e-05, | |
| "loss": 9.0081, | |
| "step": 49 | |
| }, | |
| { | |
| "epoch": 0.038058991436726926, | |
| "grad_norm": 37.831748962402344, | |
| "learning_rate": 6.261991782887377e-05, | |
| "loss": 8.2863, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.038058991436726926, | |
| "eval_loss": 3.2690162658691406, | |
| "eval_runtime": 159.7447, | |
| "eval_samples_per_second": 13.853, | |
| "eval_steps_per_second": 3.468, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.038820171265461466, | |
| "grad_norm": 40.48371887207031, | |
| "learning_rate": 6.226070475667393e-05, | |
| "loss": 12.8986, | |
| "step": 51 | |
| }, | |
| { | |
| "epoch": 0.039581351094196006, | |
| "grad_norm": 28.599153518676758, | |
| "learning_rate": 6.189403887999006e-05, | |
| "loss": 12.1586, | |
| "step": 52 | |
| }, | |
| { | |
| "epoch": 0.04034253092293054, | |
| "grad_norm": 16.86196517944336, | |
| "learning_rate": 6.152002044161171e-05, | |
| "loss": 11.3232, | |
| "step": 53 | |
| }, | |
| { | |
| "epoch": 0.04110371075166508, | |
| "grad_norm": 13.314998626708984, | |
| "learning_rate": 6.113875169444539e-05, | |
| "loss": 11.3059, | |
| "step": 54 | |
| }, | |
| { | |
| "epoch": 0.04186489058039962, | |
| "grad_norm": 12.314648628234863, | |
| "learning_rate": 6.0750336873559605e-05, | |
| "loss": 11.3874, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 0.04262607040913416, | |
| "grad_norm": 16.70978355407715, | |
| "learning_rate": 6.035488216768811e-05, | |
| "loss": 11.2571, | |
| "step": 56 | |
| }, | |
| { | |
| "epoch": 0.0433872502378687, | |
| "grad_norm": 12.304032325744629, | |
| "learning_rate": 5.9952495690198894e-05, | |
| "loss": 10.8975, | |
| "step": 57 | |
| }, | |
| { | |
| "epoch": 0.04414843006660323, | |
| "grad_norm": 12.1074800491333, | |
| "learning_rate": 5.954328744953709e-05, | |
| "loss": 10.5291, | |
| "step": 58 | |
| }, | |
| { | |
| "epoch": 0.04490960989533777, | |
| "grad_norm": 11.482349395751953, | |
| "learning_rate": 5.91273693191498e-05, | |
| "loss": 10.8876, | |
| "step": 59 | |
| }, | |
| { | |
| "epoch": 0.04567078972407231, | |
| "grad_norm": 11.817330360412598, | |
| "learning_rate": 5.870485500690094e-05, | |
| "loss": 10.9615, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.04643196955280685, | |
| "grad_norm": 11.868997573852539, | |
| "learning_rate": 5.827586002398468e-05, | |
| "loss": 10.8039, | |
| "step": 61 | |
| }, | |
| { | |
| "epoch": 0.04719314938154139, | |
| "grad_norm": 11.931234359741211, | |
| "learning_rate": 5.784050165334589e-05, | |
| "loss": 10.3456, | |
| "step": 62 | |
| }, | |
| { | |
| "epoch": 0.047954329210275926, | |
| "grad_norm": 12.019116401672363, | |
| "learning_rate": 5.739889891761608e-05, | |
| "loss": 10.3588, | |
| "step": 63 | |
| }, | |
| { | |
| "epoch": 0.048715509039010466, | |
| "grad_norm": 12.904118537902832, | |
| "learning_rate": 5.6951172546573794e-05, | |
| "loss": 10.4684, | |
| "step": 64 | |
| }, | |
| { | |
| "epoch": 0.049476688867745006, | |
| "grad_norm": 11.970799446105957, | |
| "learning_rate": 5.6497444944138376e-05, | |
| "loss": 10.8854, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 0.050237868696479546, | |
| "grad_norm": 12.086244583129883, | |
| "learning_rate": 5.603784015490587e-05, | |
| "loss": 10.4556, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 0.05099904852521408, | |
| "grad_norm": 12.689579010009766, | |
| "learning_rate": 5.557248383023655e-05, | |
| "loss": 11.5082, | |
| "step": 67 | |
| }, | |
| { | |
| "epoch": 0.05176022835394862, | |
| "grad_norm": 12.876420974731445, | |
| "learning_rate": 5.510150319390302e-05, | |
| "loss": 10.8254, | |
| "step": 68 | |
| }, | |
| { | |
| "epoch": 0.05252140818268316, | |
| "grad_norm": 12.876321792602539, | |
| "learning_rate": 5.4625027007308546e-05, | |
| "loss": 10.6497, | |
| "step": 69 | |
| }, | |
| { | |
| "epoch": 0.0532825880114177, | |
| "grad_norm": 12.555021286010742, | |
| "learning_rate": 5.414318553428494e-05, | |
| "loss": 10.7097, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.05404376784015224, | |
| "grad_norm": 14.044441223144531, | |
| "learning_rate": 5.3656110505479776e-05, | |
| "loss": 11.4296, | |
| "step": 71 | |
| }, | |
| { | |
| "epoch": 0.05480494766888677, | |
| "grad_norm": 12.55296802520752, | |
| "learning_rate": 5.316393508234253e-05, | |
| "loss": 10.8332, | |
| "step": 72 | |
| }, | |
| { | |
| "epoch": 0.05556612749762131, | |
| "grad_norm": 16.553260803222656, | |
| "learning_rate": 5.266679382071953e-05, | |
| "loss": 10.2235, | |
| "step": 73 | |
| }, | |
| { | |
| "epoch": 0.05632730732635585, | |
| "grad_norm": 12.925923347473145, | |
| "learning_rate": 5.216482263406778e-05, | |
| "loss": 11.0243, | |
| "step": 74 | |
| }, | |
| { | |
| "epoch": 0.05708848715509039, | |
| "grad_norm": 13.123712539672852, | |
| "learning_rate": 5.1658158756297576e-05, | |
| "loss": 10.5786, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.057849666983824925, | |
| "grad_norm": 15.51366901397705, | |
| "learning_rate": 5.114694070425407e-05, | |
| "loss": 10.7227, | |
| "step": 76 | |
| }, | |
| { | |
| "epoch": 0.058610846812559465, | |
| "grad_norm": 13.080121994018555, | |
| "learning_rate": 5.063130823984823e-05, | |
| "loss": 10.228, | |
| "step": 77 | |
| }, | |
| { | |
| "epoch": 0.059372026641294005, | |
| "grad_norm": 13.711491584777832, | |
| "learning_rate": 5.011140233184724e-05, | |
| "loss": 10.5938, | |
| "step": 78 | |
| }, | |
| { | |
| "epoch": 0.060133206470028545, | |
| "grad_norm": 15.03261947631836, | |
| "learning_rate": 4.958736511733516e-05, | |
| "loss": 10.9826, | |
| "step": 79 | |
| }, | |
| { | |
| "epoch": 0.060894386298763085, | |
| "grad_norm": 12.706965446472168, | |
| "learning_rate": 4.905933986285393e-05, | |
| "loss": 9.9897, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.06165556612749762, | |
| "grad_norm": 16.334020614624023, | |
| "learning_rate": 4.8527470925235824e-05, | |
| "loss": 11.1604, | |
| "step": 81 | |
| }, | |
| { | |
| "epoch": 0.06241674595623216, | |
| "grad_norm": 15.609319686889648, | |
| "learning_rate": 4.799190371213772e-05, | |
| "loss": 11.1633, | |
| "step": 82 | |
| }, | |
| { | |
| "epoch": 0.0631779257849667, | |
| "grad_norm": 14.876952171325684, | |
| "learning_rate": 4.745278464228808e-05, | |
| "loss": 10.7648, | |
| "step": 83 | |
| }, | |
| { | |
| "epoch": 0.06393910561370124, | |
| "grad_norm": 15.908063888549805, | |
| "learning_rate": 4.69102611054575e-05, | |
| "loss": 10.5009, | |
| "step": 84 | |
| }, | |
| { | |
| "epoch": 0.06470028544243578, | |
| "grad_norm": 15.217329025268555, | |
| "learning_rate": 4.6364481422163926e-05, | |
| "loss": 10.9934, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 0.06546146527117032, | |
| "grad_norm": 15.298270225524902, | |
| "learning_rate": 4.581559480312316e-05, | |
| "loss": 10.359, | |
| "step": 86 | |
| }, | |
| { | |
| "epoch": 0.06622264509990486, | |
| "grad_norm": 14.394475936889648, | |
| "learning_rate": 4.526375130845627e-05, | |
| "loss": 10.089, | |
| "step": 87 | |
| }, | |
| { | |
| "epoch": 0.06698382492863938, | |
| "grad_norm": 15.546106338500977, | |
| "learning_rate": 4.4709101806664554e-05, | |
| "loss": 10.7069, | |
| "step": 88 | |
| }, | |
| { | |
| "epoch": 0.06774500475737393, | |
| "grad_norm": 16.768985748291016, | |
| "learning_rate": 4.4151797933383685e-05, | |
| "loss": 10.2604, | |
| "step": 89 | |
| }, | |
| { | |
| "epoch": 0.06850618458610847, | |
| "grad_norm": 15.508210182189941, | |
| "learning_rate": 4.359199204992797e-05, | |
| "loss": 10.2447, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.069267364414843, | |
| "grad_norm": 15.253734588623047, | |
| "learning_rate": 4.30298372016363e-05, | |
| "loss": 9.6666, | |
| "step": 91 | |
| }, | |
| { | |
| "epoch": 0.07002854424357755, | |
| "grad_norm": 17.911222457885742, | |
| "learning_rate": 4.246548707603114e-05, | |
| "loss": 10.0422, | |
| "step": 92 | |
| }, | |
| { | |
| "epoch": 0.07078972407231209, | |
| "grad_norm": 17.074037551879883, | |
| "learning_rate": 4.1899095960801805e-05, | |
| "loss": 9.4583, | |
| "step": 93 | |
| }, | |
| { | |
| "epoch": 0.07155090390104663, | |
| "grad_norm": 16.951196670532227, | |
| "learning_rate": 4.133081870162385e-05, | |
| "loss": 9.2461, | |
| "step": 94 | |
| }, | |
| { | |
| "epoch": 0.07231208372978117, | |
| "grad_norm": 16.778108596801758, | |
| "learning_rate": 4.076081065982569e-05, | |
| "loss": 8.8192, | |
| "step": 95 | |
| }, | |
| { | |
| "epoch": 0.0730732635585157, | |
| "grad_norm": 19.881837844848633, | |
| "learning_rate": 4.018922766991447e-05, | |
| "loss": 8.6392, | |
| "step": 96 | |
| }, | |
| { | |
| "epoch": 0.07383444338725023, | |
| "grad_norm": 24.723371505737305, | |
| "learning_rate": 3.961622599697241e-05, | |
| "loss": 9.6325, | |
| "step": 97 | |
| }, | |
| { | |
| "epoch": 0.07459562321598477, | |
| "grad_norm": 21.536834716796875, | |
| "learning_rate": 3.9041962293935516e-05, | |
| "loss": 8.0487, | |
| "step": 98 | |
| }, | |
| { | |
| "epoch": 0.07535680304471931, | |
| "grad_norm": 21.819488525390625, | |
| "learning_rate": 3.84665935587662e-05, | |
| "loss": 8.5147, | |
| "step": 99 | |
| }, | |
| { | |
| "epoch": 0.07611798287345385, | |
| "grad_norm": 28.30029296875, | |
| "learning_rate": 3.7890277091531636e-05, | |
| "loss": 5.6556, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.07611798287345385, | |
| "eval_loss": 3.2521181106567383, | |
| "eval_runtime": 159.753, | |
| "eval_samples_per_second": 13.853, | |
| "eval_steps_per_second": 3.468, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.07687916270218839, | |
| "grad_norm": 40.049400329589844, | |
| "learning_rate": 3.7313170451399475e-05, | |
| "loss": 11.8576, | |
| "step": 101 | |
| }, | |
| { | |
| "epoch": 0.07764034253092293, | |
| "grad_norm": 33.80797576904297, | |
| "learning_rate": 3.673543141356278e-05, | |
| "loss": 12.1796, | |
| "step": 102 | |
| }, | |
| { | |
| "epoch": 0.07840152235965747, | |
| "grad_norm": 45.65495300292969, | |
| "learning_rate": 3.6157217926105783e-05, | |
| "loss": 11.9832, | |
| "step": 103 | |
| }, | |
| { | |
| "epoch": 0.07916270218839201, | |
| "grad_norm": 35.831092834472656, | |
| "learning_rate": 3.557868806682255e-05, | |
| "loss": 12.129, | |
| "step": 104 | |
| }, | |
| { | |
| "epoch": 0.07992388201712655, | |
| "grad_norm": 19.742515563964844, | |
| "learning_rate": 3.5e-05, | |
| "loss": 11.3112, | |
| "step": 105 | |
| }, | |
| { | |
| "epoch": 0.08068506184586108, | |
| "grad_norm": 12.872544288635254, | |
| "learning_rate": 3.442131193317745e-05, | |
| "loss": 10.8966, | |
| "step": 106 | |
| }, | |
| { | |
| "epoch": 0.08144624167459562, | |
| "grad_norm": 16.163219451904297, | |
| "learning_rate": 3.384278207389421e-05, | |
| "loss": 10.8695, | |
| "step": 107 | |
| }, | |
| { | |
| "epoch": 0.08220742150333016, | |
| "grad_norm": 15.760530471801758, | |
| "learning_rate": 3.3264568586437216e-05, | |
| "loss": 11.1322, | |
| "step": 108 | |
| }, | |
| { | |
| "epoch": 0.0829686013320647, | |
| "grad_norm": 12.737210273742676, | |
| "learning_rate": 3.268682954860052e-05, | |
| "loss": 11.1206, | |
| "step": 109 | |
| }, | |
| { | |
| "epoch": 0.08372978116079924, | |
| "grad_norm": 12.96513843536377, | |
| "learning_rate": 3.210972290846837e-05, | |
| "loss": 10.6524, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.08449096098953378, | |
| "grad_norm": 15.09691333770752, | |
| "learning_rate": 3.15334064412338e-05, | |
| "loss": 10.8931, | |
| "step": 111 | |
| }, | |
| { | |
| "epoch": 0.08525214081826832, | |
| "grad_norm": 12.536844253540039, | |
| "learning_rate": 3.0958037706064485e-05, | |
| "loss": 10.828, | |
| "step": 112 | |
| }, | |
| { | |
| "epoch": 0.08601332064700286, | |
| "grad_norm": 12.665741920471191, | |
| "learning_rate": 3.038377400302758e-05, | |
| "loss": 11.0088, | |
| "step": 113 | |
| }, | |
| { | |
| "epoch": 0.0867745004757374, | |
| "grad_norm": 11.625788688659668, | |
| "learning_rate": 2.9810772330085524e-05, | |
| "loss": 10.7514, | |
| "step": 114 | |
| }, | |
| { | |
| "epoch": 0.08753568030447194, | |
| "grad_norm": 13.06114673614502, | |
| "learning_rate": 2.9239189340174306e-05, | |
| "loss": 10.7538, | |
| "step": 115 | |
| }, | |
| { | |
| "epoch": 0.08829686013320646, | |
| "grad_norm": 13.02463150024414, | |
| "learning_rate": 2.8669181298376163e-05, | |
| "loss": 10.5935, | |
| "step": 116 | |
| }, | |
| { | |
| "epoch": 0.089058039961941, | |
| "grad_norm": 11.050668716430664, | |
| "learning_rate": 2.8100904039198193e-05, | |
| "loss": 10.4296, | |
| "step": 117 | |
| }, | |
| { | |
| "epoch": 0.08981921979067554, | |
| "grad_norm": 11.642556190490723, | |
| "learning_rate": 2.7534512923968863e-05, | |
| "loss": 10.5926, | |
| "step": 118 | |
| }, | |
| { | |
| "epoch": 0.09058039961941008, | |
| "grad_norm": 12.23384952545166, | |
| "learning_rate": 2.6970162798363695e-05, | |
| "loss": 10.8684, | |
| "step": 119 | |
| }, | |
| { | |
| "epoch": 0.09134157944814462, | |
| "grad_norm": 12.222782135009766, | |
| "learning_rate": 2.640800795007203e-05, | |
| "loss": 11.1264, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.09210275927687916, | |
| "grad_norm": 11.30429458618164, | |
| "learning_rate": 2.5848202066616305e-05, | |
| "loss": 10.1724, | |
| "step": 121 | |
| }, | |
| { | |
| "epoch": 0.0928639391056137, | |
| "grad_norm": 14.014923095703125, | |
| "learning_rate": 2.5290898193335446e-05, | |
| "loss": 11.1533, | |
| "step": 122 | |
| }, | |
| { | |
| "epoch": 0.09362511893434824, | |
| "grad_norm": 14.214231491088867, | |
| "learning_rate": 2.4736248691543736e-05, | |
| "loss": 10.6027, | |
| "step": 123 | |
| }, | |
| { | |
| "epoch": 0.09438629876308278, | |
| "grad_norm": 13.18128776550293, | |
| "learning_rate": 2.4184405196876842e-05, | |
| "loss": 10.7151, | |
| "step": 124 | |
| }, | |
| { | |
| "epoch": 0.09514747859181731, | |
| "grad_norm": 12.126294136047363, | |
| "learning_rate": 2.363551857783608e-05, | |
| "loss": 10.0795, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 0.09590865842055185, | |
| "grad_norm": 12.301307678222656, | |
| "learning_rate": 2.308973889454249e-05, | |
| "loss": 10.0897, | |
| "step": 126 | |
| }, | |
| { | |
| "epoch": 0.09666983824928639, | |
| "grad_norm": 12.24648666381836, | |
| "learning_rate": 2.2547215357711918e-05, | |
| "loss": 10.0434, | |
| "step": 127 | |
| }, | |
| { | |
| "epoch": 0.09743101807802093, | |
| "grad_norm": 13.534128189086914, | |
| "learning_rate": 2.2008096287862266e-05, | |
| "loss": 10.5216, | |
| "step": 128 | |
| }, | |
| { | |
| "epoch": 0.09819219790675547, | |
| "grad_norm": 13.032715797424316, | |
| "learning_rate": 2.1472529074764177e-05, | |
| "loss": 10.751, | |
| "step": 129 | |
| }, | |
| { | |
| "epoch": 0.09895337773549001, | |
| "grad_norm": 13.36697006225586, | |
| "learning_rate": 2.0940660137146074e-05, | |
| "loss": 10.4866, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.09971455756422455, | |
| "grad_norm": 14.086920738220215, | |
| "learning_rate": 2.041263488266484e-05, | |
| "loss": 10.4804, | |
| "step": 131 | |
| }, | |
| { | |
| "epoch": 0.10047573739295909, | |
| "grad_norm": 13.180987358093262, | |
| "learning_rate": 1.988859766815275e-05, | |
| "loss": 9.7492, | |
| "step": 132 | |
| }, | |
| { | |
| "epoch": 0.10123691722169363, | |
| "grad_norm": 14.282618522644043, | |
| "learning_rate": 1.9368691760151773e-05, | |
| "loss": 11.3742, | |
| "step": 133 | |
| }, | |
| { | |
| "epoch": 0.10199809705042816, | |
| "grad_norm": 14.527554512023926, | |
| "learning_rate": 1.885305929574593e-05, | |
| "loss": 10.8801, | |
| "step": 134 | |
| }, | |
| { | |
| "epoch": 0.1027592768791627, | |
| "grad_norm": 12.420266151428223, | |
| "learning_rate": 1.8341841243702424e-05, | |
| "loss": 9.6681, | |
| "step": 135 | |
| }, | |
| { | |
| "epoch": 0.10352045670789724, | |
| "grad_norm": 15.356304168701172, | |
| "learning_rate": 1.7835177365932225e-05, | |
| "loss": 10.354, | |
| "step": 136 | |
| }, | |
| { | |
| "epoch": 0.10428163653663178, | |
| "grad_norm": 13.979833602905273, | |
| "learning_rate": 1.7333206179280478e-05, | |
| "loss": 10.2454, | |
| "step": 137 | |
| }, | |
| { | |
| "epoch": 0.10504281636536632, | |
| "grad_norm": 13.443976402282715, | |
| "learning_rate": 1.6836064917657478e-05, | |
| "loss": 9.7525, | |
| "step": 138 | |
| }, | |
| { | |
| "epoch": 0.10580399619410086, | |
| "grad_norm": 13.744975090026855, | |
| "learning_rate": 1.6343889494520224e-05, | |
| "loss": 9.7207, | |
| "step": 139 | |
| }, | |
| { | |
| "epoch": 0.1065651760228354, | |
| "grad_norm": 15.326041221618652, | |
| "learning_rate": 1.5856814465715064e-05, | |
| "loss": 9.8803, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.10732635585156994, | |
| "grad_norm": 14.626502990722656, | |
| "learning_rate": 1.5374972992691458e-05, | |
| "loss": 9.7187, | |
| "step": 141 | |
| }, | |
| { | |
| "epoch": 0.10808753568030448, | |
| "grad_norm": 15.12524700164795, | |
| "learning_rate": 1.4898496806096974e-05, | |
| "loss": 9.5225, | |
| "step": 142 | |
| }, | |
| { | |
| "epoch": 0.108848715509039, | |
| "grad_norm": 15.535211563110352, | |
| "learning_rate": 1.4427516169763444e-05, | |
| "loss": 9.0, | |
| "step": 143 | |
| }, | |
| { | |
| "epoch": 0.10960989533777354, | |
| "grad_norm": 16.60105323791504, | |
| "learning_rate": 1.396215984509412e-05, | |
| "loss": 9.1394, | |
| "step": 144 | |
| }, | |
| { | |
| "epoch": 0.11037107516650808, | |
| "grad_norm": 14.278752326965332, | |
| "learning_rate": 1.3502555055861625e-05, | |
| "loss": 8.6505, | |
| "step": 145 | |
| }, | |
| { | |
| "epoch": 0.11113225499524262, | |
| "grad_norm": 17.42091941833496, | |
| "learning_rate": 1.3048827453426203e-05, | |
| "loss": 8.7583, | |
| "step": 146 | |
| }, | |
| { | |
| "epoch": 0.11189343482397716, | |
| "grad_norm": 18.650836944580078, | |
| "learning_rate": 1.2601101082383917e-05, | |
| "loss": 9.2693, | |
| "step": 147 | |
| }, | |
| { | |
| "epoch": 0.1126546146527117, | |
| "grad_norm": 19.8348331451416, | |
| "learning_rate": 1.2159498346654094e-05, | |
| "loss": 8.5282, | |
| "step": 148 | |
| }, | |
| { | |
| "epoch": 0.11341579448144624, | |
| "grad_norm": 22.303525924682617, | |
| "learning_rate": 1.1724139976015306e-05, | |
| "loss": 8.885, | |
| "step": 149 | |
| }, | |
| { | |
| "epoch": 0.11417697431018078, | |
| "grad_norm": 27.490812301635742, | |
| "learning_rate": 1.1295144993099068e-05, | |
| "loss": 8.4182, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.11417697431018078, | |
| "eval_loss": 2.8477718830108643, | |
| "eval_runtime": 159.7361, | |
| "eval_samples_per_second": 13.854, | |
| "eval_steps_per_second": 3.468, | |
| "step": 150 | |
| } | |
| ], | |
| "logging_steps": 1, | |
| "max_steps": 200, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 50, | |
| "stateful_callbacks": { | |
| "EarlyStoppingCallback": { | |
| "args": { | |
| "early_stopping_patience": 4, | |
| "early_stopping_threshold": 0.0 | |
| }, | |
| "attributes": { | |
| "early_stopping_patience_counter": 0 | |
| } | |
| }, | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 2.1510037676281037e+17, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |