| { | |
| "best_metric": 10.949830055236816, | |
| "best_model_checkpoint": "miner_id_24/checkpoint-75", | |
| "epoch": 0.645682001614205, | |
| "eval_steps": 25, | |
| "global_step": 75, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.008609093354856066, | |
| "grad_norm": 10.867487907409668, | |
| "learning_rate": 3.3333333333333335e-05, | |
| "loss": 355.0044, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.008609093354856066, | |
| "eval_loss": 11.096755027770996, | |
| "eval_runtime": 0.6034, | |
| "eval_samples_per_second": 82.868, | |
| "eval_steps_per_second": 21.546, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.017218186709712133, | |
| "grad_norm": 11.303132057189941, | |
| "learning_rate": 6.666666666666667e-05, | |
| "loss": 354.975, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.0258272800645682, | |
| "grad_norm": 13.843127250671387, | |
| "learning_rate": 0.0001, | |
| "loss": 355.0412, | |
| "step": 3 | |
| }, | |
| { | |
| "epoch": 0.034436373419424265, | |
| "grad_norm": 11.937524795532227, | |
| "learning_rate": 9.99571699711836e-05, | |
| "loss": 354.924, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.04304546677428033, | |
| "grad_norm": 12.751911163330078, | |
| "learning_rate": 9.982876141412856e-05, | |
| "loss": 354.8232, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.0516545601291364, | |
| "grad_norm": 12.354517936706543, | |
| "learning_rate": 9.961501876182148e-05, | |
| "loss": 354.8041, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.060263653483992465, | |
| "grad_norm": 13.037909507751465, | |
| "learning_rate": 9.931634888554937e-05, | |
| "loss": 354.611, | |
| "step": 7 | |
| }, | |
| { | |
| "epoch": 0.06887274683884853, | |
| "grad_norm": 13.028841972351074, | |
| "learning_rate": 9.893332032039701e-05, | |
| "loss": 354.4688, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.0774818401937046, | |
| "grad_norm": 14.73702335357666, | |
| "learning_rate": 9.846666218300807e-05, | |
| "loss": 354.3231, | |
| "step": 9 | |
| }, | |
| { | |
| "epoch": 0.08609093354856066, | |
| "grad_norm": 17.07733917236328, | |
| "learning_rate": 9.791726278367022e-05, | |
| "loss": 354.056, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.09470002690341674, | |
| "grad_norm": 21.84981918334961, | |
| "learning_rate": 9.728616793536588e-05, | |
| "loss": 353.632, | |
| "step": 11 | |
| }, | |
| { | |
| "epoch": 0.1033091202582728, | |
| "grad_norm": 25.133949279785156, | |
| "learning_rate": 9.657457896300791e-05, | |
| "loss": 353.3881, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 0.11191821361312887, | |
| "grad_norm": 18.78001594543457, | |
| "learning_rate": 9.578385041664925e-05, | |
| "loss": 353.659, | |
| "step": 13 | |
| }, | |
| { | |
| "epoch": 0.12052730696798493, | |
| "grad_norm": 10.563817024230957, | |
| "learning_rate": 9.491548749301997e-05, | |
| "loss": 353.9691, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 0.129136400322841, | |
| "grad_norm": 11.821118354797363, | |
| "learning_rate": 9.397114317029975e-05, | |
| "loss": 353.8424, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.13774549367769706, | |
| "grad_norm": 11.502336502075195, | |
| "learning_rate": 9.295261506157986e-05, | |
| "loss": 353.7357, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 0.14635458703255314, | |
| "grad_norm": 10.819986343383789, | |
| "learning_rate": 9.186184199300464e-05, | |
| "loss": 353.6895, | |
| "step": 17 | |
| }, | |
| { | |
| "epoch": 0.1549636803874092, | |
| "grad_norm": 11.171977996826172, | |
| "learning_rate": 9.070090031310558e-05, | |
| "loss": 353.5294, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 0.16357277374226525, | |
| "grad_norm": 11.021985054016113, | |
| "learning_rate": 8.947199994035401e-05, | |
| "loss": 353.4442, | |
| "step": 19 | |
| }, | |
| { | |
| "epoch": 0.17218186709712133, | |
| "grad_norm": 11.427451133728027, | |
| "learning_rate": 8.817748015645558e-05, | |
| "loss": 353.2932, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.1807909604519774, | |
| "grad_norm": 12.078248977661133, | |
| "learning_rate": 8.681980515339464e-05, | |
| "loss": 353.0774, | |
| "step": 21 | |
| }, | |
| { | |
| "epoch": 0.18940005380683347, | |
| "grad_norm": 13.572750091552734, | |
| "learning_rate": 8.540155934270471e-05, | |
| "loss": 352.7897, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 0.19800914716168955, | |
| "grad_norm": 16.139827728271484, | |
| "learning_rate": 8.392544243589427e-05, | |
| "loss": 352.1232, | |
| "step": 23 | |
| }, | |
| { | |
| "epoch": 0.2066182405165456, | |
| "grad_norm": 19.074623107910156, | |
| "learning_rate": 8.239426430539243e-05, | |
| "loss": 351.4342, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 0.21522733387140167, | |
| "grad_norm": 22.747406005859375, | |
| "learning_rate": 8.081093963579707e-05, | |
| "loss": 350.7261, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.21522733387140167, | |
| "eval_loss": 11.010053634643555, | |
| "eval_runtime": 0.0769, | |
| "eval_samples_per_second": 650.534, | |
| "eval_steps_per_second": 169.139, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.22383642722625774, | |
| "grad_norm": 8.548608779907227, | |
| "learning_rate": 7.917848237560709e-05, | |
| "loss": 353.2629, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 0.2324455205811138, | |
| "grad_norm": 9.44759750366211, | |
| "learning_rate": 7.75e-05, | |
| "loss": 352.9843, | |
| "step": 27 | |
| }, | |
| { | |
| "epoch": 0.24105461393596986, | |
| "grad_norm": 10.709016799926758, | |
| "learning_rate": 7.577868759557654e-05, | |
| "loss": 352.6682, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 0.24966370729082593, | |
| "grad_norm": 9.533336639404297, | |
| "learning_rate": 7.401782177833148e-05, | |
| "loss": 352.8022, | |
| "step": 29 | |
| }, | |
| { | |
| "epoch": 0.258272800645682, | |
| "grad_norm": 9.468610763549805, | |
| "learning_rate": 7.222075445642904e-05, | |
| "loss": 352.7345, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.2668818940005381, | |
| "grad_norm": 9.217055320739746, | |
| "learning_rate": 7.03909064496551e-05, | |
| "loss": 352.7065, | |
| "step": 31 | |
| }, | |
| { | |
| "epoch": 0.2754909873553941, | |
| "grad_norm": 9.265793800354004, | |
| "learning_rate": 6.853176097769229e-05, | |
| "loss": 352.6331, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 0.2841000807102502, | |
| "grad_norm": 9.615739822387695, | |
| "learning_rate": 6.664685702961344e-05, | |
| "loss": 352.4907, | |
| "step": 33 | |
| }, | |
| { | |
| "epoch": 0.29270917406510627, | |
| "grad_norm": 10.331357955932617, | |
| "learning_rate": 6.473978262721463e-05, | |
| "loss": 352.2647, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 0.3013182674199623, | |
| "grad_norm": 12.123662948608398, | |
| "learning_rate": 6.281416799501188e-05, | |
| "loss": 351.5476, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.3099273607748184, | |
| "grad_norm": 14.970151901245117, | |
| "learning_rate": 6.087367864990233e-05, | |
| "loss": 350.6174, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 0.31853645412967446, | |
| "grad_norm": 18.231613159179688, | |
| "learning_rate": 5.8922008423644624e-05, | |
| "loss": 349.7336, | |
| "step": 37 | |
| }, | |
| { | |
| "epoch": 0.3271455474845305, | |
| "grad_norm": 14.60504150390625, | |
| "learning_rate": 5.696287243144013e-05, | |
| "loss": 350.7061, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 0.3357546408393866, | |
| "grad_norm": 9.001364707946777, | |
| "learning_rate": 5.500000000000001e-05, | |
| "loss": 352.3064, | |
| "step": 39 | |
| }, | |
| { | |
| "epoch": 0.34436373419424265, | |
| "grad_norm": 9.125486373901367, | |
| "learning_rate": 5.303712756855988e-05, | |
| "loss": 352.185, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.35297282754909876, | |
| "grad_norm": 8.79714584350586, | |
| "learning_rate": 5.107799157635538e-05, | |
| "loss": 352.1189, | |
| "step": 41 | |
| }, | |
| { | |
| "epoch": 0.3615819209039548, | |
| "grad_norm": 8.923425674438477, | |
| "learning_rate": 4.912632135009769e-05, | |
| "loss": 352.1105, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 0.37019101425881085, | |
| "grad_norm": 8.289681434631348, | |
| "learning_rate": 4.718583200498814e-05, | |
| "loss": 352.2823, | |
| "step": 43 | |
| }, | |
| { | |
| "epoch": 0.37880010761366695, | |
| "grad_norm": 8.199746131896973, | |
| "learning_rate": 4.526021737278538e-05, | |
| "loss": 352.2198, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 0.387409200968523, | |
| "grad_norm": 8.637060165405273, | |
| "learning_rate": 4.3353142970386564e-05, | |
| "loss": 352.026, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 0.3960182943233791, | |
| "grad_norm": 9.83723258972168, | |
| "learning_rate": 4.146823902230772e-05, | |
| "loss": 351.563, | |
| "step": 46 | |
| }, | |
| { | |
| "epoch": 0.40462738767823514, | |
| "grad_norm": 10.942987442016602, | |
| "learning_rate": 3.960909355034491e-05, | |
| "loss": 351.0425, | |
| "step": 47 | |
| }, | |
| { | |
| "epoch": 0.4132364810330912, | |
| "grad_norm": 13.6652193069458, | |
| "learning_rate": 3.777924554357096e-05, | |
| "loss": 350.1031, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 0.4218455743879473, | |
| "grad_norm": 16.773061752319336, | |
| "learning_rate": 3.598217822166854e-05, | |
| "loss": 349.162, | |
| "step": 49 | |
| }, | |
| { | |
| "epoch": 0.43045466774280333, | |
| "grad_norm": 20.8967227935791, | |
| "learning_rate": 3.422131240442349e-05, | |
| "loss": 348.3621, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.43045466774280333, | |
| "eval_loss": 10.965777397155762, | |
| "eval_runtime": 0.0782, | |
| "eval_samples_per_second": 639.534, | |
| "eval_steps_per_second": 166.279, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.4390637610976594, | |
| "grad_norm": 7.474067687988281, | |
| "learning_rate": 3.250000000000001e-05, | |
| "loss": 352.3553, | |
| "step": 51 | |
| }, | |
| { | |
| "epoch": 0.4476728544525155, | |
| "grad_norm": 8.300982475280762, | |
| "learning_rate": 3.082151762439293e-05, | |
| "loss": 351.9735, | |
| "step": 52 | |
| }, | |
| { | |
| "epoch": 0.4562819478073715, | |
| "grad_norm": 8.844514846801758, | |
| "learning_rate": 2.9189060364202943e-05, | |
| "loss": 351.6582, | |
| "step": 53 | |
| }, | |
| { | |
| "epoch": 0.4648910411622276, | |
| "grad_norm": 8.853071212768555, | |
| "learning_rate": 2.760573569460757e-05, | |
| "loss": 351.5546, | |
| "step": 54 | |
| }, | |
| { | |
| "epoch": 0.47350013451708367, | |
| "grad_norm": 8.490742683410645, | |
| "learning_rate": 2.6074557564105727e-05, | |
| "loss": 351.6604, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 0.4821092278719397, | |
| "grad_norm": 7.409176349639893, | |
| "learning_rate": 2.459844065729529e-05, | |
| "loss": 352.1426, | |
| "step": 56 | |
| }, | |
| { | |
| "epoch": 0.4907183212267958, | |
| "grad_norm": 7.821872234344482, | |
| "learning_rate": 2.3180194846605367e-05, | |
| "loss": 351.8066, | |
| "step": 57 | |
| }, | |
| { | |
| "epoch": 0.49932741458165186, | |
| "grad_norm": 8.349543571472168, | |
| "learning_rate": 2.1822519843544424e-05, | |
| "loss": 351.5433, | |
| "step": 58 | |
| }, | |
| { | |
| "epoch": 0.5079365079365079, | |
| "grad_norm": 9.260781288146973, | |
| "learning_rate": 2.0528000059645997e-05, | |
| "loss": 351.208, | |
| "step": 59 | |
| }, | |
| { | |
| "epoch": 0.516545601291364, | |
| "grad_norm": 10.735580444335938, | |
| "learning_rate": 1.9299099686894423e-05, | |
| "loss": 350.6387, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.5251546946462201, | |
| "grad_norm": 13.846049308776855, | |
| "learning_rate": 1.8138158006995364e-05, | |
| "loss": 349.4801, | |
| "step": 61 | |
| }, | |
| { | |
| "epoch": 0.5337637880010762, | |
| "grad_norm": 17.65458869934082, | |
| "learning_rate": 1.7047384938420154e-05, | |
| "loss": 348.2975, | |
| "step": 62 | |
| }, | |
| { | |
| "epoch": 0.5423728813559322, | |
| "grad_norm": 13.394047737121582, | |
| "learning_rate": 1.602885682970026e-05, | |
| "loss": 349.9447, | |
| "step": 63 | |
| }, | |
| { | |
| "epoch": 0.5509819747107882, | |
| "grad_norm": 7.800135612487793, | |
| "learning_rate": 1.5084512506980026e-05, | |
| "loss": 351.9247, | |
| "step": 64 | |
| }, | |
| { | |
| "epoch": 0.5595910680656443, | |
| "grad_norm": 8.641666412353516, | |
| "learning_rate": 1.4216149583350754e-05, | |
| "loss": 351.5002, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 0.5682001614205004, | |
| "grad_norm": 8.432456970214844, | |
| "learning_rate": 1.3425421036992098e-05, | |
| "loss": 351.3875, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 0.5768092547753565, | |
| "grad_norm": 9.028701782226562, | |
| "learning_rate": 1.2713832064634126e-05, | |
| "loss": 351.1097, | |
| "step": 67 | |
| }, | |
| { | |
| "epoch": 0.5854183481302125, | |
| "grad_norm": 8.943015098571777, | |
| "learning_rate": 1.2082737216329794e-05, | |
| "loss": 351.1015, | |
| "step": 68 | |
| }, | |
| { | |
| "epoch": 0.5940274414850686, | |
| "grad_norm": 7.793323516845703, | |
| "learning_rate": 1.1533337816991932e-05, | |
| "loss": 351.5153, | |
| "step": 69 | |
| }, | |
| { | |
| "epoch": 0.6026365348399246, | |
| "grad_norm": 8.137232780456543, | |
| "learning_rate": 1.1066679679603e-05, | |
| "loss": 351.4627, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.6112456281947808, | |
| "grad_norm": 8.868036270141602, | |
| "learning_rate": 1.0683651114450641e-05, | |
| "loss": 351.1549, | |
| "step": 71 | |
| }, | |
| { | |
| "epoch": 0.6198547215496368, | |
| "grad_norm": 10.001080513000488, | |
| "learning_rate": 1.0384981238178534e-05, | |
| "loss": 350.6647, | |
| "step": 72 | |
| }, | |
| { | |
| "epoch": 0.6284638149044929, | |
| "grad_norm": 12.413795471191406, | |
| "learning_rate": 1.017123858587145e-05, | |
| "loss": 349.7495, | |
| "step": 73 | |
| }, | |
| { | |
| "epoch": 0.6370729082593489, | |
| "grad_norm": 16.752674102783203, | |
| "learning_rate": 1.00428300288164e-05, | |
| "loss": 348.3455, | |
| "step": 74 | |
| }, | |
| { | |
| "epoch": 0.645682001614205, | |
| "grad_norm": 20.989471435546875, | |
| "learning_rate": 1e-05, | |
| "loss": 347.5739, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.645682001614205, | |
| "eval_loss": 10.949830055236816, | |
| "eval_runtime": 0.0779, | |
| "eval_samples_per_second": 641.476, | |
| "eval_steps_per_second": 166.784, | |
| "step": 75 | |
| } | |
| ], | |
| "logging_steps": 1, | |
| "max_steps": 75, | |
| "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": 6323739033600.0, | |
| "train_batch_size": 1, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |