{ "best_metric": 2.412879228591919, "best_model_checkpoint": "miner_id_24/checkpoint-50", "epoch": 0.3318125259228536, "eval_steps": 25, "global_step": 50, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 0.006636250518457071, "grad_norm": 23.848411560058594, "learning_rate": 5e-05, "loss": 96.9709, "step": 1 }, { "epoch": 0.006636250518457071, "eval_loss": 3.6564836502075195, "eval_runtime": 5.5662, "eval_samples_per_second": 8.983, "eval_steps_per_second": 2.336, "step": 1 }, { "epoch": 0.013272501036914143, "grad_norm": 26.234432220458984, "learning_rate": 0.0001, "loss": 98.9038, "step": 2 }, { "epoch": 0.019908751555371217, "grad_norm": 32.293235778808594, "learning_rate": 9.990365154573717e-05, "loss": 103.2514, "step": 3 }, { "epoch": 0.026545002073828285, "grad_norm": 46.88783645629883, "learning_rate": 9.961501876182148e-05, "loss": 103.0289, "step": 4 }, { "epoch": 0.03318125259228536, "grad_norm": 74.84331512451172, "learning_rate": 9.913533761814537e-05, "loss": 101.7903, "step": 5 }, { "epoch": 0.03981750311074243, "grad_norm": 51.89060592651367, "learning_rate": 9.846666218300807e-05, "loss": 97.8142, "step": 6 }, { "epoch": 0.046453753629199505, "grad_norm": 69.3984603881836, "learning_rate": 9.761185582727977e-05, "loss": 94.8137, "step": 7 }, { "epoch": 0.05309000414765657, "grad_norm": 78.8531265258789, "learning_rate": 9.657457896300791e-05, "loss": 89.8656, "step": 8 }, { "epoch": 0.05972625466611364, "grad_norm": 82.24502563476562, "learning_rate": 9.535927336897098e-05, "loss": 90.7878, "step": 9 }, { "epoch": 0.06636250518457072, "grad_norm": 100.51170349121094, "learning_rate": 9.397114317029975e-05, "loss": 107.3313, "step": 10 }, { "epoch": 0.07299875570302779, "grad_norm": 73.24173736572266, "learning_rate": 9.241613255361455e-05, "loss": 116.1533, "step": 11 }, { "epoch": 0.07963500622148487, "grad_norm": 70.09320068359375, "learning_rate": 9.070090031310558e-05, "loss": 118.7542, "step": 12 }, { "epoch": 0.08627125673994193, "grad_norm": 60.52884292602539, "learning_rate": 8.883279133655399e-05, "loss": 110.9796, "step": 13 }, { "epoch": 0.09290750725839901, "grad_norm": 35.291526794433594, "learning_rate": 8.681980515339464e-05, "loss": 90.8626, "step": 14 }, { "epoch": 0.09954375777685608, "grad_norm": 36.884613037109375, "learning_rate": 8.467056167950311e-05, "loss": 80.5083, "step": 15 }, { "epoch": 0.10618000829531314, "grad_norm": 44.371177673339844, "learning_rate": 8.239426430539243e-05, "loss": 83.4735, "step": 16 }, { "epoch": 0.11281625881377022, "grad_norm": 55.198856353759766, "learning_rate": 8.000066048588211e-05, "loss": 80.5926, "step": 17 }, { "epoch": 0.11945250933222729, "grad_norm": 34.99528884887695, "learning_rate": 7.75e-05, "loss": 79.5121, "step": 18 }, { "epoch": 0.12608875985068435, "grad_norm": 43.458160400390625, "learning_rate": 7.490299105985507e-05, "loss": 76.3123, "step": 19 }, { "epoch": 0.13272501036914144, "grad_norm": 30.95506477355957, "learning_rate": 7.222075445642904e-05, "loss": 76.066, "step": 20 }, { "epoch": 0.1393612608875985, "grad_norm": 40.55569076538086, "learning_rate": 6.946477593864228e-05, "loss": 76.2038, "step": 21 }, { "epoch": 0.14599751140605557, "grad_norm": 102.33631134033203, "learning_rate": 6.664685702961344e-05, "loss": 91.2752, "step": 22 }, { "epoch": 0.15263376192451264, "grad_norm": 120.65931701660156, "learning_rate": 6.377906449072578e-05, "loss": 100.8825, "step": 23 }, { "epoch": 0.15927001244296973, "grad_norm": 111.62467193603516, "learning_rate": 6.087367864990233e-05, "loss": 99.1154, "step": 24 }, { "epoch": 0.1659062629614268, "grad_norm": 99.9602279663086, "learning_rate": 5.794314081535644e-05, "loss": 122.6857, "step": 25 }, { "epoch": 0.1659062629614268, "eval_loss": 2.5546910762786865, "eval_runtime": 5.5506, "eval_samples_per_second": 9.008, "eval_steps_per_second": 2.342, "step": 25 }, { "epoch": 0.17254251347988386, "grad_norm": 24.010421752929688, "learning_rate": 5.500000000000001e-05, "loss": 89.4733, "step": 26 }, { "epoch": 0.17917876399834093, "grad_norm": 16.442790985107422, "learning_rate": 5.205685918464356e-05, "loss": 79.1688, "step": 27 }, { "epoch": 0.18581501451679802, "grad_norm": 23.459074020385742, "learning_rate": 4.912632135009769e-05, "loss": 75.9344, "step": 28 }, { "epoch": 0.1924512650352551, "grad_norm": 23.219343185424805, "learning_rate": 4.6220935509274235e-05, "loss": 75.0841, "step": 29 }, { "epoch": 0.19908751555371215, "grad_norm": 36.46348571777344, "learning_rate": 4.3353142970386564e-05, "loss": 71.8608, "step": 30 }, { "epoch": 0.20572376607216922, "grad_norm": 24.08379364013672, "learning_rate": 4.053522406135775e-05, "loss": 72.6857, "step": 31 }, { "epoch": 0.21236001659062628, "grad_norm": 20.700775146484375, "learning_rate": 3.777924554357096e-05, "loss": 73.6169, "step": 32 }, { "epoch": 0.21899626710908338, "grad_norm": 14.344165802001953, "learning_rate": 3.509700894014496e-05, "loss": 72.8962, "step": 33 }, { "epoch": 0.22563251762754044, "grad_norm": 19.467073440551758, "learning_rate": 3.250000000000001e-05, "loss": 75.4409, "step": 34 }, { "epoch": 0.2322687681459975, "grad_norm": 35.137939453125, "learning_rate": 2.9999339514117912e-05, "loss": 93.4653, "step": 35 }, { "epoch": 0.23890501866445457, "grad_norm": 45.6280517578125, "learning_rate": 2.760573569460757e-05, "loss": 88.3274, "step": 36 }, { "epoch": 0.24554126918291166, "grad_norm": 47.0994873046875, "learning_rate": 2.53294383204969e-05, "loss": 98.4404, "step": 37 }, { "epoch": 0.2521775197013687, "grad_norm": 56.962154388427734, "learning_rate": 2.3180194846605367e-05, "loss": 107.2939, "step": 38 }, { "epoch": 0.2588137702198258, "grad_norm": 14.456352233886719, "learning_rate": 2.1167208663446025e-05, "loss": 81.9888, "step": 39 }, { "epoch": 0.2654500207382829, "grad_norm": 11.632476806640625, "learning_rate": 1.9299099686894423e-05, "loss": 72.7757, "step": 40 }, { "epoch": 0.27208627125673995, "grad_norm": 13.044448852539062, "learning_rate": 1.758386744638546e-05, "loss": 75.0724, "step": 41 }, { "epoch": 0.278722521775197, "grad_norm": 12.491575241088867, "learning_rate": 1.602885682970026e-05, "loss": 74.1388, "step": 42 }, { "epoch": 0.2853587722936541, "grad_norm": 12.832867622375488, "learning_rate": 1.464072663102903e-05, "loss": 72.2979, "step": 43 }, { "epoch": 0.29199502281211115, "grad_norm": 12.466416358947754, "learning_rate": 1.3425421036992098e-05, "loss": 72.1957, "step": 44 }, { "epoch": 0.2986312733305682, "grad_norm": 13.381525993347168, "learning_rate": 1.2388144172720251e-05, "loss": 71.8714, "step": 45 }, { "epoch": 0.3052675238490253, "grad_norm": 17.53035545349121, "learning_rate": 1.1533337816991932e-05, "loss": 71.4497, "step": 46 }, { "epoch": 0.31190377436748234, "grad_norm": 25.668964385986328, "learning_rate": 1.0864662381854632e-05, "loss": 86.7853, "step": 47 }, { "epoch": 0.31854002488593947, "grad_norm": 40.41082763671875, "learning_rate": 1.0384981238178534e-05, "loss": 88.0605, "step": 48 }, { "epoch": 0.32517627540439653, "grad_norm": 40.9230842590332, "learning_rate": 1.0096348454262845e-05, "loss": 85.4458, "step": 49 }, { "epoch": 0.3318125259228536, "grad_norm": 53.349674224853516, "learning_rate": 1e-05, "loss": 109.816, "step": 50 }, { "epoch": 0.3318125259228536, "eval_loss": 2.412879228591919, "eval_runtime": 5.5758, "eval_samples_per_second": 8.967, "eval_steps_per_second": 2.331, "step": 50 } ], "logging_steps": 1, "max_steps": 50, "num_input_tokens_seen": 0, "num_train_epochs": 1, "save_steps": 25, "stateful_callbacks": { "EarlyStoppingCallback": { "args": { "early_stopping_patience": 1, "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": true }, "attributes": {} } }, "total_flos": 5.259156315439104e+17, "train_batch_size": 1, "trial_name": null, "trial_params": null }