| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 3.0, | |
| "eval_steps": 500, | |
| "global_step": 29073, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.05159426271798576, | |
| "grad_norm": 7.558670997619629, | |
| "learning_rate": 2.5736389684813753e-06, | |
| "loss": 0.7519, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.10318852543597153, | |
| "grad_norm": 6.3421125411987305, | |
| "learning_rate": 5.15243553008596e-06, | |
| "loss": 0.3291, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.15478278815395727, | |
| "grad_norm": 7.725200176239014, | |
| "learning_rate": 7.720916905444127e-06, | |
| "loss": 0.2716, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.20637705087194305, | |
| "grad_norm": 6.253057479858398, | |
| "learning_rate": 8.917008196721311e-06, | |
| "loss": 0.2444, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.2579713135899288, | |
| "grad_norm": 8.090571403503418, | |
| "learning_rate": 8.752341920374708e-06, | |
| "loss": 0.2057, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 0.30956557630791454, | |
| "grad_norm": 5.437821865081787, | |
| "learning_rate": 8.587675644028104e-06, | |
| "loss": 0.1938, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.3611598390259003, | |
| "grad_norm": 3.594510316848755, | |
| "learning_rate": 8.423009367681498e-06, | |
| "loss": 0.1888, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 0.4127541017438861, | |
| "grad_norm": 1.5619558095932007, | |
| "learning_rate": 8.258343091334895e-06, | |
| "loss": 0.1681, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.46434836446187183, | |
| "grad_norm": 5.857831001281738, | |
| "learning_rate": 8.093676814988291e-06, | |
| "loss": 0.1581, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 0.5159426271798576, | |
| "grad_norm": 2.4653804302215576, | |
| "learning_rate": 7.929010538641687e-06, | |
| "loss": 0.1469, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.5675368898978433, | |
| "grad_norm": 8.388437271118164, | |
| "learning_rate": 7.764673594847776e-06, | |
| "loss": 0.1416, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 0.6191311526158291, | |
| "grad_norm": 0.739997386932373, | |
| "learning_rate": 7.600007318501171e-06, | |
| "loss": 0.143, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 0.6707254153338149, | |
| "grad_norm": 4.071037292480469, | |
| "learning_rate": 7.435341042154568e-06, | |
| "loss": 0.1309, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 0.7223196780518006, | |
| "grad_norm": 1.3754676580429077, | |
| "learning_rate": 7.270674765807963e-06, | |
| "loss": 0.1311, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 0.7739139407697864, | |
| "grad_norm": 7.085391998291016, | |
| "learning_rate": 7.1060084894613585e-06, | |
| "loss": 0.1194, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 0.8255082034877722, | |
| "grad_norm": 3.828031539916992, | |
| "learning_rate": 6.941342213114754e-06, | |
| "loss": 0.1128, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 0.8771024662057579, | |
| "grad_norm": 6.162450790405273, | |
| "learning_rate": 6.77667593676815e-06, | |
| "loss": 0.115, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 0.9286967289237437, | |
| "grad_norm": 3.1122217178344727, | |
| "learning_rate": 6.612009660421546e-06, | |
| "loss": 0.1105, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 0.9802909916417294, | |
| "grad_norm": 9.108379364013672, | |
| "learning_rate": 6.447343384074942e-06, | |
| "loss": 0.1054, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_dev_accuracy": 0.9035151515151515, | |
| "eval_dev_loss": 0.38524121046066284, | |
| "eval_dev_runtime": 24.3219, | |
| "eval_dev_samples_per_second": 678.401, | |
| "eval_dev_steps_per_second": 42.431, | |
| "step": 9691 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_test_accuracy": 0.9203636363636364, | |
| "eval_test_loss": 0.31113502383232117, | |
| "eval_test_runtime": 24.5354, | |
| "eval_test_samples_per_second": 672.497, | |
| "eval_test_steps_per_second": 42.062, | |
| "step": 9691 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_verified_test_accuracy": 0.963639301874596, | |
| "eval_verified_test_loss": 0.13967397809028625, | |
| "eval_verified_test_runtime": 18.4171, | |
| "eval_verified_test_samples_per_second": 671.986, | |
| "eval_verified_test_steps_per_second": 42.026, | |
| "step": 9691 | |
| }, | |
| { | |
| "epoch": 1.0318852543597152, | |
| "grad_norm": 0.7660722136497498, | |
| "learning_rate": 6.2826771077283374e-06, | |
| "loss": 0.095, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 1.083479517077701, | |
| "grad_norm": 4.1145124435424805, | |
| "learning_rate": 6.118010831381733e-06, | |
| "loss": 0.0881, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 1.1350737797956867, | |
| "grad_norm": 0.3238352835178375, | |
| "learning_rate": 5.953344555035129e-06, | |
| "loss": 0.0771, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 1.1866680425136724, | |
| "grad_norm": 5.4124908447265625, | |
| "learning_rate": 5.788678278688525e-06, | |
| "loss": 0.0841, | |
| "step": 11500 | |
| }, | |
| { | |
| "epoch": 1.2382623052316581, | |
| "grad_norm": 13.157797813415527, | |
| "learning_rate": 5.624012002341921e-06, | |
| "loss": 0.0861, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 1.289856567949644, | |
| "grad_norm": 1.3548617362976074, | |
| "learning_rate": 5.4593457259953155e-06, | |
| "loss": 0.0852, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 1.3414508306676298, | |
| "grad_norm": 2.0701634883880615, | |
| "learning_rate": 5.294679449648712e-06, | |
| "loss": 0.088, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 1.3930450933856156, | |
| "grad_norm": 0.24185556173324585, | |
| "learning_rate": 5.1303425058548e-06, | |
| "loss": 0.076, | |
| "step": 13500 | |
| }, | |
| { | |
| "epoch": 1.4446393561036013, | |
| "grad_norm": 1.3316065073013306, | |
| "learning_rate": 4.965676229508197e-06, | |
| "loss": 0.0769, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 1.496233618821587, | |
| "grad_norm": 2.571403741836548, | |
| "learning_rate": 4.801339285714286e-06, | |
| "loss": 0.0775, | |
| "step": 14500 | |
| }, | |
| { | |
| "epoch": 1.5478278815395727, | |
| "grad_norm": 2.0419647693634033, | |
| "learning_rate": 4.636673009367682e-06, | |
| "loss": 0.0757, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 1.5994221442575585, | |
| "grad_norm": 4.6739912033081055, | |
| "learning_rate": 4.472006733021077e-06, | |
| "loss": 0.0775, | |
| "step": 15500 | |
| }, | |
| { | |
| "epoch": 1.6510164069755442, | |
| "grad_norm": 7.1135053634643555, | |
| "learning_rate": 4.307340456674473e-06, | |
| "loss": 0.0728, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 1.70261066969353, | |
| "grad_norm": 4.576988220214844, | |
| "learning_rate": 4.142674180327869e-06, | |
| "loss": 0.0774, | |
| "step": 16500 | |
| }, | |
| { | |
| "epoch": 1.7542049324115159, | |
| "grad_norm": 0.0673130601644516, | |
| "learning_rate": 3.978337236533958e-06, | |
| "loss": 0.0636, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 1.8057991951295016, | |
| "grad_norm": 8.489474296569824, | |
| "learning_rate": 3.8136709601873536e-06, | |
| "loss": 0.0692, | |
| "step": 17500 | |
| }, | |
| { | |
| "epoch": 1.8573934578474873, | |
| "grad_norm": 13.07606315612793, | |
| "learning_rate": 3.6490046838407495e-06, | |
| "loss": 0.0689, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 1.9089877205654733, | |
| "grad_norm": 0.5132156610488892, | |
| "learning_rate": 3.4843384074941454e-06, | |
| "loss": 0.0668, | |
| "step": 18500 | |
| }, | |
| { | |
| "epoch": 1.960581983283459, | |
| "grad_norm": 6.281818389892578, | |
| "learning_rate": 3.3196721311475413e-06, | |
| "loss": 0.0705, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_dev_accuracy": 0.8983636363636364, | |
| "eval_dev_loss": 0.5292544364929199, | |
| "eval_dev_runtime": 24.2247, | |
| "eval_dev_samples_per_second": 681.122, | |
| "eval_dev_steps_per_second": 42.601, | |
| "step": 19382 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_test_accuracy": 0.9152727272727272, | |
| "eval_test_loss": 0.4211724102497101, | |
| "eval_test_runtime": 24.492, | |
| "eval_test_samples_per_second": 673.688, | |
| "eval_test_steps_per_second": 42.136, | |
| "step": 19382 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_verified_test_accuracy": 0.9569327731092437, | |
| "eval_verified_test_loss": 0.1964138001203537, | |
| "eval_verified_test_runtime": 18.3946, | |
| "eval_verified_test_samples_per_second": 672.807, | |
| "eval_verified_test_steps_per_second": 42.078, | |
| "step": 19382 | |
| }, | |
| { | |
| "epoch": 2.0121762460014447, | |
| "grad_norm": 0.1685587465763092, | |
| "learning_rate": 3.15533518735363e-06, | |
| "loss": 0.0604, | |
| "step": 19500 | |
| }, | |
| { | |
| "epoch": 2.0637705087194305, | |
| "grad_norm": 3.583531141281128, | |
| "learning_rate": 2.990668911007026e-06, | |
| "loss": 0.0486, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 2.115364771437416, | |
| "grad_norm": 2.067493200302124, | |
| "learning_rate": 2.8260026346604216e-06, | |
| "loss": 0.0446, | |
| "step": 20500 | |
| }, | |
| { | |
| "epoch": 2.166959034155402, | |
| "grad_norm": 0.8109681010246277, | |
| "learning_rate": 2.6616656908665106e-06, | |
| "loss": 0.043, | |
| "step": 21000 | |
| }, | |
| { | |
| "epoch": 2.2185532968733876, | |
| "grad_norm": 1.798444151878357, | |
| "learning_rate": 2.4969994145199065e-06, | |
| "loss": 0.051, | |
| "step": 21500 | |
| }, | |
| { | |
| "epoch": 2.2701475595913734, | |
| "grad_norm": 0.02468039281666279, | |
| "learning_rate": 2.332333138173302e-06, | |
| "loss": 0.0475, | |
| "step": 22000 | |
| }, | |
| { | |
| "epoch": 2.321741822309359, | |
| "grad_norm": 6.017990589141846, | |
| "learning_rate": 2.167666861826698e-06, | |
| "loss": 0.0473, | |
| "step": 22500 | |
| }, | |
| { | |
| "epoch": 2.373336085027345, | |
| "grad_norm": 0.12943503260612488, | |
| "learning_rate": 2.0030005854800936e-06, | |
| "loss": 0.0469, | |
| "step": 23000 | |
| }, | |
| { | |
| "epoch": 2.4249303477453306, | |
| "grad_norm": 0.710267961025238, | |
| "learning_rate": 1.8386636416861827e-06, | |
| "loss": 0.0482, | |
| "step": 23500 | |
| }, | |
| { | |
| "epoch": 2.4765246104633163, | |
| "grad_norm": 4.638703346252441, | |
| "learning_rate": 1.6739973653395785e-06, | |
| "loss": 0.0554, | |
| "step": 24000 | |
| }, | |
| { | |
| "epoch": 2.5281188731813025, | |
| "grad_norm": 1.667284607887268, | |
| "learning_rate": 1.5096604215456673e-06, | |
| "loss": 0.042, | |
| "step": 24500 | |
| }, | |
| { | |
| "epoch": 2.579713135899288, | |
| "grad_norm": 2.6058883666992188, | |
| "learning_rate": 1.3449941451990634e-06, | |
| "loss": 0.0487, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 2.631307398617274, | |
| "grad_norm": 1.2941348552703857, | |
| "learning_rate": 1.180327868852459e-06, | |
| "loss": 0.0482, | |
| "step": 25500 | |
| }, | |
| { | |
| "epoch": 2.6829016613352596, | |
| "grad_norm": 0.13713909685611725, | |
| "learning_rate": 1.0156615925058547e-06, | |
| "loss": 0.047, | |
| "step": 26000 | |
| }, | |
| { | |
| "epoch": 2.7344959240532454, | |
| "grad_norm": 6.215357303619385, | |
| "learning_rate": 8.509953161592506e-07, | |
| "loss": 0.0408, | |
| "step": 26500 | |
| }, | |
| { | |
| "epoch": 2.786090186771231, | |
| "grad_norm": 0.08112023025751114, | |
| "learning_rate": 6.863290398126464e-07, | |
| "loss": 0.0419, | |
| "step": 27000 | |
| }, | |
| { | |
| "epoch": 2.837684449489217, | |
| "grad_norm": 1.3868869543075562, | |
| "learning_rate": 5.216627634660422e-07, | |
| "loss": 0.0449, | |
| "step": 27500 | |
| }, | |
| { | |
| "epoch": 2.8892787122072026, | |
| "grad_norm": 0.07935940474271774, | |
| "learning_rate": 3.56996487119438e-07, | |
| "loss": 0.048, | |
| "step": 28000 | |
| }, | |
| { | |
| "epoch": 2.9408729749251883, | |
| "grad_norm": 0.05485483631491661, | |
| "learning_rate": 1.9233021077283373e-07, | |
| "loss": 0.0393, | |
| "step": 28500 | |
| }, | |
| { | |
| "epoch": 2.992467237643174, | |
| "grad_norm": 4.398229598999023, | |
| "learning_rate": 2.7663934426229508e-08, | |
| "loss": 0.0446, | |
| "step": 29000 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_dev_accuracy": 0.8943030303030303, | |
| "eval_dev_loss": 0.6738013625144958, | |
| "eval_dev_runtime": 24.3063, | |
| "eval_dev_samples_per_second": 678.836, | |
| "eval_dev_steps_per_second": 42.458, | |
| "step": 29073 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_test_accuracy": 0.9123030303030303, | |
| "eval_test_loss": 0.5465541481971741, | |
| "eval_test_runtime": 24.5738, | |
| "eval_test_samples_per_second": 671.448, | |
| "eval_test_steps_per_second": 41.996, | |
| "step": 29073 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_verified_test_accuracy": 0.9536199095022625, | |
| "eval_verified_test_loss": 0.27511948347091675, | |
| "eval_verified_test_runtime": 18.4409, | |
| "eval_verified_test_samples_per_second": 671.117, | |
| "eval_verified_test_steps_per_second": 41.972, | |
| "step": 29073 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 29073, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "save_steps": 500, | |
| "total_flos": 5.067736824879187e+16, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |