| { | |
| "best_metric": 1.2121704816818237, | |
| "best_model_checkpoint": "miner_id_24/checkpoint-150", | |
| "epoch": 0.5145797598627787, | |
| "eval_steps": 50, | |
| "global_step": 150, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.003430531732418525, | |
| "grad_norm": 0.11610166728496552, | |
| "learning_rate": 3.3333333333333333e-06, | |
| "loss": 1.4495, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.003430531732418525, | |
| "eval_loss": 1.635466456413269, | |
| "eval_runtime": 13.7969, | |
| "eval_samples_per_second": 35.588, | |
| "eval_steps_per_second": 17.83, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.00686106346483705, | |
| "grad_norm": 0.11409541964530945, | |
| "learning_rate": 6.666666666666667e-06, | |
| "loss": 1.5905, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.010291595197255575, | |
| "grad_norm": 0.12520195543766022, | |
| "learning_rate": 1e-05, | |
| "loss": 1.5542, | |
| "step": 3 | |
| }, | |
| { | |
| "epoch": 0.0137221269296741, | |
| "grad_norm": 0.12021903693675995, | |
| "learning_rate": 1.3333333333333333e-05, | |
| "loss": 1.5811, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.017152658662092625, | |
| "grad_norm": 0.13395224511623383, | |
| "learning_rate": 1.6666666666666667e-05, | |
| "loss": 1.6203, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.02058319039451115, | |
| "grad_norm": 0.14789403975009918, | |
| "learning_rate": 2e-05, | |
| "loss": 1.5317, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.024013722126929673, | |
| "grad_norm": 0.14486268162727356, | |
| "learning_rate": 2.3333333333333336e-05, | |
| "loss": 1.545, | |
| "step": 7 | |
| }, | |
| { | |
| "epoch": 0.0274442538593482, | |
| "grad_norm": 0.1345217078924179, | |
| "learning_rate": 2.6666666666666667e-05, | |
| "loss": 1.5916, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.030874785591766724, | |
| "grad_norm": 0.15623602271080017, | |
| "learning_rate": 3e-05, | |
| "loss": 1.6245, | |
| "step": 9 | |
| }, | |
| { | |
| "epoch": 0.03430531732418525, | |
| "grad_norm": 0.14663076400756836, | |
| "learning_rate": 3.3333333333333335e-05, | |
| "loss": 1.5943, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.03773584905660377, | |
| "grad_norm": 0.1576271653175354, | |
| "learning_rate": 3.6666666666666666e-05, | |
| "loss": 1.6079, | |
| "step": 11 | |
| }, | |
| { | |
| "epoch": 0.0411663807890223, | |
| "grad_norm": 0.14678916335105896, | |
| "learning_rate": 4e-05, | |
| "loss": 1.5798, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 0.044596912521440824, | |
| "grad_norm": 0.16557946801185608, | |
| "learning_rate": 4.3333333333333334e-05, | |
| "loss": 1.5852, | |
| "step": 13 | |
| }, | |
| { | |
| "epoch": 0.048027444253859346, | |
| "grad_norm": 0.14881809055805206, | |
| "learning_rate": 4.666666666666667e-05, | |
| "loss": 1.5622, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 0.051457975986277875, | |
| "grad_norm": 0.16381366550922394, | |
| "learning_rate": 5e-05, | |
| "loss": 1.6237, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.0548885077186964, | |
| "grad_norm": 0.16224238276481628, | |
| "learning_rate": 5.333333333333333e-05, | |
| "loss": 1.6018, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 0.058319039451114926, | |
| "grad_norm": 0.15393707156181335, | |
| "learning_rate": 5.666666666666667e-05, | |
| "loss": 1.5426, | |
| "step": 17 | |
| }, | |
| { | |
| "epoch": 0.06174957118353345, | |
| "grad_norm": 0.16475345194339752, | |
| "learning_rate": 6e-05, | |
| "loss": 1.6154, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 0.06518010291595197, | |
| "grad_norm": 0.1683150976896286, | |
| "learning_rate": 6.333333333333333e-05, | |
| "loss": 1.5662, | |
| "step": 19 | |
| }, | |
| { | |
| "epoch": 0.0686106346483705, | |
| "grad_norm": 0.17348520457744598, | |
| "learning_rate": 6.666666666666667e-05, | |
| "loss": 1.5324, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.07204116638078903, | |
| "grad_norm": 0.19107702374458313, | |
| "learning_rate": 7e-05, | |
| "loss": 1.5642, | |
| "step": 21 | |
| }, | |
| { | |
| "epoch": 0.07547169811320754, | |
| "grad_norm": 0.18596619367599487, | |
| "learning_rate": 7.333333333333333e-05, | |
| "loss": 1.5147, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 0.07890222984562607, | |
| "grad_norm": 0.16649138927459717, | |
| "learning_rate": 7.666666666666667e-05, | |
| "loss": 1.5331, | |
| "step": 23 | |
| }, | |
| { | |
| "epoch": 0.0823327615780446, | |
| "grad_norm": 0.18072673678398132, | |
| "learning_rate": 8e-05, | |
| "loss": 1.5107, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 0.08576329331046312, | |
| "grad_norm": 0.19641919434070587, | |
| "learning_rate": 8.333333333333334e-05, | |
| "loss": 1.5476, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.08919382504288165, | |
| "grad_norm": 0.19105511903762817, | |
| "learning_rate": 8.666666666666667e-05, | |
| "loss": 1.4737, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 0.09262435677530018, | |
| "grad_norm": 0.19515928626060486, | |
| "learning_rate": 9e-05, | |
| "loss": 1.5276, | |
| "step": 27 | |
| }, | |
| { | |
| "epoch": 0.09605488850771869, | |
| "grad_norm": 0.2127508521080017, | |
| "learning_rate": 9.333333333333334e-05, | |
| "loss": 1.4991, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 0.09948542024013722, | |
| "grad_norm": 0.21000954508781433, | |
| "learning_rate": 9.666666666666667e-05, | |
| "loss": 1.4888, | |
| "step": 29 | |
| }, | |
| { | |
| "epoch": 0.10291595197255575, | |
| "grad_norm": 0.23589254915714264, | |
| "learning_rate": 0.0001, | |
| "loss": 1.4491, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.10634648370497427, | |
| "grad_norm": 0.21637442708015442, | |
| "learning_rate": 9.999146252290264e-05, | |
| "loss": 1.5132, | |
| "step": 31 | |
| }, | |
| { | |
| "epoch": 0.1097770154373928, | |
| "grad_norm": 0.209445059299469, | |
| "learning_rate": 9.996585300715116e-05, | |
| "loss": 1.3777, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 0.11320754716981132, | |
| "grad_norm": 0.21992652118206024, | |
| "learning_rate": 9.99231801983717e-05, | |
| "loss": 1.3847, | |
| "step": 33 | |
| }, | |
| { | |
| "epoch": 0.11663807890222985, | |
| "grad_norm": 0.2562519311904907, | |
| "learning_rate": 9.986345866928941e-05, | |
| "loss": 1.3906, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 0.12006861063464837, | |
| "grad_norm": 0.2565302848815918, | |
| "learning_rate": 9.978670881475172e-05, | |
| "loss": 1.408, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.1234991423670669, | |
| "grad_norm": 0.26474887132644653, | |
| "learning_rate": 9.96929568447637e-05, | |
| "loss": 1.4223, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 0.1269296740994854, | |
| "grad_norm": 0.24229450523853302, | |
| "learning_rate": 9.958223477553714e-05, | |
| "loss": 1.4247, | |
| "step": 37 | |
| }, | |
| { | |
| "epoch": 0.13036020583190394, | |
| "grad_norm": 0.2645803987979889, | |
| "learning_rate": 9.94545804185573e-05, | |
| "loss": 1.3998, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 0.13379073756432247, | |
| "grad_norm": 0.2674286663532257, | |
| "learning_rate": 9.931003736767013e-05, | |
| "loss": 1.422, | |
| "step": 39 | |
| }, | |
| { | |
| "epoch": 0.137221269296741, | |
| "grad_norm": 0.30323857069015503, | |
| "learning_rate": 9.91486549841951e-05, | |
| "loss": 1.3878, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.14065180102915953, | |
| "grad_norm": 0.28205499053001404, | |
| "learning_rate": 9.89704883800683e-05, | |
| "loss": 1.4102, | |
| "step": 41 | |
| }, | |
| { | |
| "epoch": 0.14408233276157806, | |
| "grad_norm": 0.2820606827735901, | |
| "learning_rate": 9.877559839902184e-05, | |
| "loss": 1.3916, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 0.14751286449399656, | |
| "grad_norm": 0.32162535190582275, | |
| "learning_rate": 9.85640515958057e-05, | |
| "loss": 1.4053, | |
| "step": 43 | |
| }, | |
| { | |
| "epoch": 0.1509433962264151, | |
| "grad_norm": 0.36674660444259644, | |
| "learning_rate": 9.833592021345937e-05, | |
| "loss": 1.3878, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 0.15437392795883362, | |
| "grad_norm": 0.3291763663291931, | |
| "learning_rate": 9.809128215864097e-05, | |
| "loss": 1.3465, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 0.15780445969125215, | |
| "grad_norm": 0.35642141103744507, | |
| "learning_rate": 9.783022097502204e-05, | |
| "loss": 1.4211, | |
| "step": 46 | |
| }, | |
| { | |
| "epoch": 0.16123499142367068, | |
| "grad_norm": 0.35451963543891907, | |
| "learning_rate": 9.755282581475769e-05, | |
| "loss": 1.456, | |
| "step": 47 | |
| }, | |
| { | |
| "epoch": 0.1646655231560892, | |
| "grad_norm": 0.36406123638153076, | |
| "learning_rate": 9.725919140804099e-05, | |
| "loss": 1.3712, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 0.1680960548885077, | |
| "grad_norm": 0.4121352434158325, | |
| "learning_rate": 9.694941803075283e-05, | |
| "loss": 1.3851, | |
| "step": 49 | |
| }, | |
| { | |
| "epoch": 0.17152658662092624, | |
| "grad_norm": 0.45500677824020386, | |
| "learning_rate": 9.662361147021779e-05, | |
| "loss": 1.3785, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.17152658662092624, | |
| "eval_loss": 1.3733711242675781, | |
| "eval_runtime": 13.893, | |
| "eval_samples_per_second": 35.342, | |
| "eval_steps_per_second": 17.707, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.17495711835334476, | |
| "grad_norm": 0.46119561791419983, | |
| "learning_rate": 9.628188298907782e-05, | |
| "loss": 1.4779, | |
| "step": 51 | |
| }, | |
| { | |
| "epoch": 0.1783876500857633, | |
| "grad_norm": 0.3245432376861572, | |
| "learning_rate": 9.592434928729616e-05, | |
| "loss": 1.5325, | |
| "step": 52 | |
| }, | |
| { | |
| "epoch": 0.18181818181818182, | |
| "grad_norm": 0.29628750681877136, | |
| "learning_rate": 9.555113246230442e-05, | |
| "loss": 1.5435, | |
| "step": 53 | |
| }, | |
| { | |
| "epoch": 0.18524871355060035, | |
| "grad_norm": 0.31039223074913025, | |
| "learning_rate": 9.516235996730645e-05, | |
| "loss": 1.4436, | |
| "step": 54 | |
| }, | |
| { | |
| "epoch": 0.18867924528301888, | |
| "grad_norm": 0.30162546038627625, | |
| "learning_rate": 9.475816456775313e-05, | |
| "loss": 1.4301, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 0.19210977701543738, | |
| "grad_norm": 0.2942914366722107, | |
| "learning_rate": 9.43386842960031e-05, | |
| "loss": 1.3478, | |
| "step": 56 | |
| }, | |
| { | |
| "epoch": 0.1955403087478559, | |
| "grad_norm": 0.23570454120635986, | |
| "learning_rate": 9.39040624041849e-05, | |
| "loss": 1.4735, | |
| "step": 57 | |
| }, | |
| { | |
| "epoch": 0.19897084048027444, | |
| "grad_norm": 0.18676193058490753, | |
| "learning_rate": 9.345444731527642e-05, | |
| "loss": 1.4547, | |
| "step": 58 | |
| }, | |
| { | |
| "epoch": 0.20240137221269297, | |
| "grad_norm": 0.19960841536521912, | |
| "learning_rate": 9.298999257241863e-05, | |
| "loss": 1.414, | |
| "step": 59 | |
| }, | |
| { | |
| "epoch": 0.2058319039451115, | |
| "grad_norm": 0.2147475779056549, | |
| "learning_rate": 9.251085678648072e-05, | |
| "loss": 1.3576, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.20926243567753003, | |
| "grad_norm": 0.2323458343744278, | |
| "learning_rate": 9.201720358189464e-05, | |
| "loss": 1.3581, | |
| "step": 61 | |
| }, | |
| { | |
| "epoch": 0.21269296740994853, | |
| "grad_norm": 0.2327861487865448, | |
| "learning_rate": 9.150920154077754e-05, | |
| "loss": 1.4149, | |
| "step": 62 | |
| }, | |
| { | |
| "epoch": 0.21612349914236706, | |
| "grad_norm": 0.20968282222747803, | |
| "learning_rate": 9.098702414536107e-05, | |
| "loss": 1.3564, | |
| "step": 63 | |
| }, | |
| { | |
| "epoch": 0.2195540308747856, | |
| "grad_norm": 0.18935877084732056, | |
| "learning_rate": 9.045084971874738e-05, | |
| "loss": 1.3218, | |
| "step": 64 | |
| }, | |
| { | |
| "epoch": 0.22298456260720412, | |
| "grad_norm": 0.17681340873241425, | |
| "learning_rate": 8.9900861364012e-05, | |
| "loss": 1.3779, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 0.22641509433962265, | |
| "grad_norm": 0.19004489481449127, | |
| "learning_rate": 8.933724690167417e-05, | |
| "loss": 1.3508, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 0.22984562607204118, | |
| "grad_norm": 0.2089233696460724, | |
| "learning_rate": 8.876019880555649e-05, | |
| "loss": 1.3112, | |
| "step": 67 | |
| }, | |
| { | |
| "epoch": 0.2332761578044597, | |
| "grad_norm": 0.2069028615951538, | |
| "learning_rate": 8.816991413705516e-05, | |
| "loss": 1.3396, | |
| "step": 68 | |
| }, | |
| { | |
| "epoch": 0.2367066895368782, | |
| "grad_norm": 0.2308463752269745, | |
| "learning_rate": 8.756659447784368e-05, | |
| "loss": 1.3472, | |
| "step": 69 | |
| }, | |
| { | |
| "epoch": 0.24013722126929674, | |
| "grad_norm": 0.22154006361961365, | |
| "learning_rate": 8.695044586103296e-05, | |
| "loss": 1.3622, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.24356775300171526, | |
| "grad_norm": 0.2146824598312378, | |
| "learning_rate": 8.632167870081121e-05, | |
| "loss": 1.374, | |
| "step": 71 | |
| }, | |
| { | |
| "epoch": 0.2469982847341338, | |
| "grad_norm": 0.21277782320976257, | |
| "learning_rate": 8.568050772058762e-05, | |
| "loss": 1.3181, | |
| "step": 72 | |
| }, | |
| { | |
| "epoch": 0.2504288164665523, | |
| "grad_norm": 0.21313494443893433, | |
| "learning_rate": 8.502715187966455e-05, | |
| "loss": 1.2741, | |
| "step": 73 | |
| }, | |
| { | |
| "epoch": 0.2538593481989708, | |
| "grad_norm": 0.2278512716293335, | |
| "learning_rate": 8.436183429846313e-05, | |
| "loss": 1.2932, | |
| "step": 74 | |
| }, | |
| { | |
| "epoch": 0.25728987993138935, | |
| "grad_norm": 0.23364584147930145, | |
| "learning_rate": 8.368478218232787e-05, | |
| "loss": 1.3354, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.2607204116638079, | |
| "grad_norm": 0.24717149138450623, | |
| "learning_rate": 8.299622674393614e-05, | |
| "loss": 1.3353, | |
| "step": 76 | |
| }, | |
| { | |
| "epoch": 0.2641509433962264, | |
| "grad_norm": 0.23784001171588898, | |
| "learning_rate": 8.229640312433937e-05, | |
| "loss": 1.3005, | |
| "step": 77 | |
| }, | |
| { | |
| "epoch": 0.26758147512864494, | |
| "grad_norm": 0.2664402723312378, | |
| "learning_rate": 8.158555031266254e-05, | |
| "loss": 1.318, | |
| "step": 78 | |
| }, | |
| { | |
| "epoch": 0.27101200686106347, | |
| "grad_norm": 0.22661960124969482, | |
| "learning_rate": 8.086391106448965e-05, | |
| "loss": 1.3024, | |
| "step": 79 | |
| }, | |
| { | |
| "epoch": 0.274442538593482, | |
| "grad_norm": 0.21936969459056854, | |
| "learning_rate": 8.013173181896283e-05, | |
| "loss": 1.2807, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.27787307032590053, | |
| "grad_norm": 0.27401024103164673, | |
| "learning_rate": 7.938926261462366e-05, | |
| "loss": 1.2482, | |
| "step": 81 | |
| }, | |
| { | |
| "epoch": 0.28130360205831906, | |
| "grad_norm": 0.2559613287448883, | |
| "learning_rate": 7.863675700402526e-05, | |
| "loss": 1.322, | |
| "step": 82 | |
| }, | |
| { | |
| "epoch": 0.2847341337907376, | |
| "grad_norm": 0.24693572521209717, | |
| "learning_rate": 7.787447196714427e-05, | |
| "loss": 1.2792, | |
| "step": 83 | |
| }, | |
| { | |
| "epoch": 0.2881646655231561, | |
| "grad_norm": 0.2803778350353241, | |
| "learning_rate": 7.710266782362247e-05, | |
| "loss": 1.3206, | |
| "step": 84 | |
| }, | |
| { | |
| "epoch": 0.2915951972555746, | |
| "grad_norm": 0.2616997957229614, | |
| "learning_rate": 7.63216081438678e-05, | |
| "loss": 1.2982, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 0.2950257289879931, | |
| "grad_norm": 0.274962842464447, | |
| "learning_rate": 7.553155965904535e-05, | |
| "loss": 1.2987, | |
| "step": 86 | |
| }, | |
| { | |
| "epoch": 0.29845626072041165, | |
| "grad_norm": 0.3058188557624817, | |
| "learning_rate": 7.473279216998895e-05, | |
| "loss": 1.3222, | |
| "step": 87 | |
| }, | |
| { | |
| "epoch": 0.3018867924528302, | |
| "grad_norm": 0.2801516354084015, | |
| "learning_rate": 7.392557845506432e-05, | |
| "loss": 1.2495, | |
| "step": 88 | |
| }, | |
| { | |
| "epoch": 0.3053173241852487, | |
| "grad_norm": 0.2805141508579254, | |
| "learning_rate": 7.311019417701566e-05, | |
| "loss": 1.2647, | |
| "step": 89 | |
| }, | |
| { | |
| "epoch": 0.30874785591766724, | |
| "grad_norm": 0.30509936809539795, | |
| "learning_rate": 7.228691778882693e-05, | |
| "loss": 1.2689, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.31217838765008576, | |
| "grad_norm": 0.29796192049980164, | |
| "learning_rate": 7.145603043863045e-05, | |
| "loss": 1.2896, | |
| "step": 91 | |
| }, | |
| { | |
| "epoch": 0.3156089193825043, | |
| "grad_norm": 0.31429192423820496, | |
| "learning_rate": 7.061781587369519e-05, | |
| "loss": 1.2563, | |
| "step": 92 | |
| }, | |
| { | |
| "epoch": 0.3190394511149228, | |
| "grad_norm": 0.3246179521083832, | |
| "learning_rate": 6.977256034352712e-05, | |
| "loss": 1.2631, | |
| "step": 93 | |
| }, | |
| { | |
| "epoch": 0.32246998284734135, | |
| "grad_norm": 0.32547494769096375, | |
| "learning_rate": 6.892055250211552e-05, | |
| "loss": 1.2476, | |
| "step": 94 | |
| }, | |
| { | |
| "epoch": 0.3259005145797599, | |
| "grad_norm": 0.3263494670391083, | |
| "learning_rate": 6.806208330935766e-05, | |
| "loss": 1.2056, | |
| "step": 95 | |
| }, | |
| { | |
| "epoch": 0.3293310463121784, | |
| "grad_norm": 0.3600931763648987, | |
| "learning_rate": 6.719744593169641e-05, | |
| "loss": 1.237, | |
| "step": 96 | |
| }, | |
| { | |
| "epoch": 0.33276157804459694, | |
| "grad_norm": 0.36960792541503906, | |
| "learning_rate": 6.632693564200416e-05, | |
| "loss": 1.1656, | |
| "step": 97 | |
| }, | |
| { | |
| "epoch": 0.3361921097770154, | |
| "grad_norm": 0.39583271741867065, | |
| "learning_rate": 6.545084971874738e-05, | |
| "loss": 1.2885, | |
| "step": 98 | |
| }, | |
| { | |
| "epoch": 0.33962264150943394, | |
| "grad_norm": 0.42984476685523987, | |
| "learning_rate": 6.456948734446624e-05, | |
| "loss": 1.2469, | |
| "step": 99 | |
| }, | |
| { | |
| "epoch": 0.34305317324185247, | |
| "grad_norm": 0.5518860816955566, | |
| "learning_rate": 6.368314950360415e-05, | |
| "loss": 1.3106, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.34305317324185247, | |
| "eval_loss": 1.2660377025604248, | |
| "eval_runtime": 13.8359, | |
| "eval_samples_per_second": 35.487, | |
| "eval_steps_per_second": 17.78, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.346483704974271, | |
| "grad_norm": 0.46960487961769104, | |
| "learning_rate": 6.279213887972179e-05, | |
| "loss": 1.3384, | |
| "step": 101 | |
| }, | |
| { | |
| "epoch": 0.34991423670668953, | |
| "grad_norm": 0.4278581142425537, | |
| "learning_rate": 6.189675975213094e-05, | |
| "loss": 1.5108, | |
| "step": 102 | |
| }, | |
| { | |
| "epoch": 0.35334476843910806, | |
| "grad_norm": 0.3242987394332886, | |
| "learning_rate": 6.099731789198344e-05, | |
| "loss": 1.4282, | |
| "step": 103 | |
| }, | |
| { | |
| "epoch": 0.3567753001715266, | |
| "grad_norm": 0.2774030864238739, | |
| "learning_rate": 6.009412045785051e-05, | |
| "loss": 1.3264, | |
| "step": 104 | |
| }, | |
| { | |
| "epoch": 0.3602058319039451, | |
| "grad_norm": 0.28905799984931946, | |
| "learning_rate": 5.918747589082853e-05, | |
| "loss": 1.3673, | |
| "step": 105 | |
| }, | |
| { | |
| "epoch": 0.36363636363636365, | |
| "grad_norm": 0.286334753036499, | |
| "learning_rate": 5.82776938092065e-05, | |
| "loss": 1.3391, | |
| "step": 106 | |
| }, | |
| { | |
| "epoch": 0.3670668953687822, | |
| "grad_norm": 0.28761571645736694, | |
| "learning_rate": 5.736508490273188e-05, | |
| "loss": 1.3136, | |
| "step": 107 | |
| }, | |
| { | |
| "epoch": 0.3704974271012007, | |
| "grad_norm": 0.23601886630058289, | |
| "learning_rate": 5.644996082651017e-05, | |
| "loss": 1.4489, | |
| "step": 108 | |
| }, | |
| { | |
| "epoch": 0.37392795883361923, | |
| "grad_norm": 0.26613518595695496, | |
| "learning_rate": 5.553263409457504e-05, | |
| "loss": 1.2763, | |
| "step": 109 | |
| }, | |
| { | |
| "epoch": 0.37735849056603776, | |
| "grad_norm": 0.24539977312088013, | |
| "learning_rate": 5.4613417973165106e-05, | |
| "loss": 1.2729, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.38078902229845624, | |
| "grad_norm": 0.23666392266750336, | |
| "learning_rate": 5.3692626373743706e-05, | |
| "loss": 1.3449, | |
| "step": 111 | |
| }, | |
| { | |
| "epoch": 0.38421955403087477, | |
| "grad_norm": 0.2275010645389557, | |
| "learning_rate": 5.27705737457985e-05, | |
| "loss": 1.2318, | |
| "step": 112 | |
| }, | |
| { | |
| "epoch": 0.3876500857632933, | |
| "grad_norm": 0.24695216119289398, | |
| "learning_rate": 5.184757496945726e-05, | |
| "loss": 1.3218, | |
| "step": 113 | |
| }, | |
| { | |
| "epoch": 0.3910806174957118, | |
| "grad_norm": 0.25049513578414917, | |
| "learning_rate": 5.092394524795649e-05, | |
| "loss": 1.3459, | |
| "step": 114 | |
| }, | |
| { | |
| "epoch": 0.39451114922813035, | |
| "grad_norm": 0.2567203640937805, | |
| "learning_rate": 5e-05, | |
| "loss": 1.2598, | |
| "step": 115 | |
| }, | |
| { | |
| "epoch": 0.3979416809605489, | |
| "grad_norm": 0.22766773402690887, | |
| "learning_rate": 4.907605475204352e-05, | |
| "loss": 1.217, | |
| "step": 116 | |
| }, | |
| { | |
| "epoch": 0.4013722126929674, | |
| "grad_norm": 0.24498119950294495, | |
| "learning_rate": 4.8152425030542766e-05, | |
| "loss": 1.2414, | |
| "step": 117 | |
| }, | |
| { | |
| "epoch": 0.40480274442538594, | |
| "grad_norm": 0.2613143026828766, | |
| "learning_rate": 4.72294262542015e-05, | |
| "loss": 1.2934, | |
| "step": 118 | |
| }, | |
| { | |
| "epoch": 0.40823327615780447, | |
| "grad_norm": 0.2593250870704651, | |
| "learning_rate": 4.6307373626256306e-05, | |
| "loss": 1.2766, | |
| "step": 119 | |
| }, | |
| { | |
| "epoch": 0.411663807890223, | |
| "grad_norm": 0.23566867411136627, | |
| "learning_rate": 4.5386582026834906e-05, | |
| "loss": 1.221, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.41509433962264153, | |
| "grad_norm": 0.2302766591310501, | |
| "learning_rate": 4.446736590542497e-05, | |
| "loss": 1.2444, | |
| "step": 121 | |
| }, | |
| { | |
| "epoch": 0.41852487135506006, | |
| "grad_norm": 0.22572791576385498, | |
| "learning_rate": 4.3550039173489845e-05, | |
| "loss": 1.2369, | |
| "step": 122 | |
| }, | |
| { | |
| "epoch": 0.4219554030874786, | |
| "grad_norm": 0.2492562234401703, | |
| "learning_rate": 4.2634915097268115e-05, | |
| "loss": 1.1794, | |
| "step": 123 | |
| }, | |
| { | |
| "epoch": 0.42538593481989706, | |
| "grad_norm": 0.26300421357154846, | |
| "learning_rate": 4.1722306190793495e-05, | |
| "loss": 1.2073, | |
| "step": 124 | |
| }, | |
| { | |
| "epoch": 0.4288164665523156, | |
| "grad_norm": 0.25651299953460693, | |
| "learning_rate": 4.0812524109171476e-05, | |
| "loss": 1.2113, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 0.4322469982847341, | |
| "grad_norm": 0.25439882278442383, | |
| "learning_rate": 3.99058795421495e-05, | |
| "loss": 1.3366, | |
| "step": 126 | |
| }, | |
| { | |
| "epoch": 0.43567753001715265, | |
| "grad_norm": 0.25761932134628296, | |
| "learning_rate": 3.9002682108016585e-05, | |
| "loss": 1.2507, | |
| "step": 127 | |
| }, | |
| { | |
| "epoch": 0.4391080617495712, | |
| "grad_norm": 0.25570839643478394, | |
| "learning_rate": 3.8103240247869075e-05, | |
| "loss": 1.2225, | |
| "step": 128 | |
| }, | |
| { | |
| "epoch": 0.4425385934819897, | |
| "grad_norm": 0.28426146507263184, | |
| "learning_rate": 3.720786112027822e-05, | |
| "loss": 1.261, | |
| "step": 129 | |
| }, | |
| { | |
| "epoch": 0.44596912521440824, | |
| "grad_norm": 0.2929293215274811, | |
| "learning_rate": 3.631685049639586e-05, | |
| "loss": 1.1955, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.44939965694682676, | |
| "grad_norm": 0.26186859607696533, | |
| "learning_rate": 3.543051265553377e-05, | |
| "loss": 1.2179, | |
| "step": 131 | |
| }, | |
| { | |
| "epoch": 0.4528301886792453, | |
| "grad_norm": 0.2666632831096649, | |
| "learning_rate": 3.4549150281252636e-05, | |
| "loss": 1.1792, | |
| "step": 132 | |
| }, | |
| { | |
| "epoch": 0.4562607204116638, | |
| "grad_norm": 0.30045852065086365, | |
| "learning_rate": 3.367306435799584e-05, | |
| "loss": 1.2296, | |
| "step": 133 | |
| }, | |
| { | |
| "epoch": 0.45969125214408235, | |
| "grad_norm": 0.2867948114871979, | |
| "learning_rate": 3.2802554068303596e-05, | |
| "loss": 1.166, | |
| "step": 134 | |
| }, | |
| { | |
| "epoch": 0.4631217838765009, | |
| "grad_norm": 0.3068961501121521, | |
| "learning_rate": 3.1937916690642356e-05, | |
| "loss": 1.2112, | |
| "step": 135 | |
| }, | |
| { | |
| "epoch": 0.4665523156089194, | |
| "grad_norm": 0.31013593077659607, | |
| "learning_rate": 3.107944749788449e-05, | |
| "loss": 1.2374, | |
| "step": 136 | |
| }, | |
| { | |
| "epoch": 0.4699828473413379, | |
| "grad_norm": 0.3226790726184845, | |
| "learning_rate": 3.0227439656472877e-05, | |
| "loss": 1.2555, | |
| "step": 137 | |
| }, | |
| { | |
| "epoch": 0.4734133790737564, | |
| "grad_norm": 0.30750301480293274, | |
| "learning_rate": 2.9382184126304834e-05, | |
| "loss": 1.2291, | |
| "step": 138 | |
| }, | |
| { | |
| "epoch": 0.47684391080617494, | |
| "grad_norm": 0.3350450396537781, | |
| "learning_rate": 2.8543969561369556e-05, | |
| "loss": 1.2717, | |
| "step": 139 | |
| }, | |
| { | |
| "epoch": 0.48027444253859347, | |
| "grad_norm": 0.33156758546829224, | |
| "learning_rate": 2.771308221117309e-05, | |
| "loss": 1.2334, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.483704974271012, | |
| "grad_norm": 0.3310351073741913, | |
| "learning_rate": 2.688980582298435e-05, | |
| "loss": 1.1668, | |
| "step": 141 | |
| }, | |
| { | |
| "epoch": 0.48713550600343053, | |
| "grad_norm": 0.3672282099723816, | |
| "learning_rate": 2.607442154493568e-05, | |
| "loss": 1.1896, | |
| "step": 142 | |
| }, | |
| { | |
| "epoch": 0.49056603773584906, | |
| "grad_norm": 0.38358932733535767, | |
| "learning_rate": 2.5267207830011068e-05, | |
| "loss": 1.2141, | |
| "step": 143 | |
| }, | |
| { | |
| "epoch": 0.4939965694682676, | |
| "grad_norm": 0.3809894025325775, | |
| "learning_rate": 2.446844034095466e-05, | |
| "loss": 1.2576, | |
| "step": 144 | |
| }, | |
| { | |
| "epoch": 0.4974271012006861, | |
| "grad_norm": 0.3670910894870758, | |
| "learning_rate": 2.3678391856132204e-05, | |
| "loss": 1.199, | |
| "step": 145 | |
| }, | |
| { | |
| "epoch": 0.5008576329331046, | |
| "grad_norm": 0.4124976098537445, | |
| "learning_rate": 2.2897332176377528e-05, | |
| "loss": 1.169, | |
| "step": 146 | |
| }, | |
| { | |
| "epoch": 0.5042881646655232, | |
| "grad_norm": 0.4204322099685669, | |
| "learning_rate": 2.2125528032855724e-05, | |
| "loss": 1.1429, | |
| "step": 147 | |
| }, | |
| { | |
| "epoch": 0.5077186963979416, | |
| "grad_norm": 0.4493495523929596, | |
| "learning_rate": 2.136324299597474e-05, | |
| "loss": 1.2079, | |
| "step": 148 | |
| }, | |
| { | |
| "epoch": 0.5111492281303602, | |
| "grad_norm": 0.5025396943092346, | |
| "learning_rate": 2.061073738537635e-05, | |
| "loss": 1.3403, | |
| "step": 149 | |
| }, | |
| { | |
| "epoch": 0.5145797598627787, | |
| "grad_norm": 0.6223259568214417, | |
| "learning_rate": 1.9868268181037185e-05, | |
| "loss": 1.3662, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.5145797598627787, | |
| "eval_loss": 1.2121704816818237, | |
| "eval_runtime": 13.7988, | |
| "eval_samples_per_second": 35.583, | |
| "eval_steps_per_second": 17.828, | |
| "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": 5, | |
| "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": 1.080131307700224e+16, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |