| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.6944444444444444, | |
| "eval_steps": 500, | |
| "global_step": 100, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.01, | |
| "learning_rate": 4.000000000000001e-06, | |
| "loss": 2.065, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.01, | |
| "learning_rate": 8.000000000000001e-06, | |
| "loss": 2.0734, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.02, | |
| "learning_rate": 1.2e-05, | |
| "loss": 2.0027, | |
| "step": 3 | |
| }, | |
| { | |
| "epoch": 0.03, | |
| "learning_rate": 1.6000000000000003e-05, | |
| "loss": 1.8999, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.03, | |
| "learning_rate": 2e-05, | |
| "loss": 2.0094, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.04, | |
| "learning_rate": 1.999744599547812e-05, | |
| "loss": 2.0019, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.05, | |
| "learning_rate": 1.9989785286500294e-05, | |
| "loss": 1.979, | |
| "step": 7 | |
| }, | |
| { | |
| "epoch": 0.06, | |
| "learning_rate": 1.99770217861636e-05, | |
| "loss": 1.9602, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.06, | |
| "learning_rate": 1.9959162014075553e-05, | |
| "loss": 1.886, | |
| "step": 9 | |
| }, | |
| { | |
| "epoch": 0.07, | |
| "learning_rate": 1.9936215093023884e-05, | |
| "loss": 1.9153, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "learning_rate": 1.990819274431662e-05, | |
| "loss": 1.9168, | |
| "step": 11 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "learning_rate": 1.9875109281794828e-05, | |
| "loss": 1.899, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 0.09, | |
| "learning_rate": 1.9836981604521077e-05, | |
| "loss": 1.884, | |
| "step": 13 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "learning_rate": 1.9793829188147406e-05, | |
| "loss": 1.8718, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "learning_rate": 1.974567407496712e-05, | |
| "loss": 1.8484, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.11, | |
| "learning_rate": 1.9692540862655587e-05, | |
| "loss": 1.8129, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 0.12, | |
| "learning_rate": 1.9634456691705705e-05, | |
| "loss": 1.8475, | |
| "step": 17 | |
| }, | |
| { | |
| "epoch": 0.12, | |
| "learning_rate": 1.9571451231564523e-05, | |
| "loss": 1.8583, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "learning_rate": 1.9503556665478066e-05, | |
| "loss": 1.8149, | |
| "step": 19 | |
| }, | |
| { | |
| "epoch": 0.14, | |
| "learning_rate": 1.9430807674052092e-05, | |
| "loss": 1.8637, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "learning_rate": 1.9353241417537216e-05, | |
| "loss": 1.8277, | |
| "step": 21 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "learning_rate": 1.9270897516847406e-05, | |
| "loss": 1.8263, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "learning_rate": 1.9183818033321612e-05, | |
| "loss": 1.8175, | |
| "step": 23 | |
| }, | |
| { | |
| "epoch": 0.17, | |
| "learning_rate": 1.9092047447238775e-05, | |
| "loss": 1.8434, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 0.17, | |
| "learning_rate": 1.899563263509725e-05, | |
| "loss": 1.8388, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.18, | |
| "learning_rate": 1.8894622845670282e-05, | |
| "loss": 1.8461, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 0.19, | |
| "learning_rate": 1.878906967484966e-05, | |
| "loss": 1.7963, | |
| "step": 27 | |
| }, | |
| { | |
| "epoch": 0.19, | |
| "learning_rate": 1.86790270392905e-05, | |
| "loss": 1.7976, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "learning_rate": 1.856455114887056e-05, | |
| "loss": 1.805, | |
| "step": 29 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "learning_rate": 1.8445700477978207e-05, | |
| "loss": 1.8068, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.22, | |
| "learning_rate": 1.8322535735643604e-05, | |
| "loss": 1.793, | |
| "step": 31 | |
| }, | |
| { | |
| "epoch": 0.22, | |
| "learning_rate": 1.8195119834528535e-05, | |
| "loss": 1.8285, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 0.23, | |
| "learning_rate": 1.8063517858790517e-05, | |
| "loss": 1.8104, | |
| "step": 33 | |
| }, | |
| { | |
| "epoch": 0.24, | |
| "learning_rate": 1.792779703083777e-05, | |
| "loss": 1.8032, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 0.24, | |
| "learning_rate": 1.778802667699196e-05, | |
| "loss": 1.7777, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "learning_rate": 1.764427819207624e-05, | |
| "loss": 1.7989, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "learning_rate": 1.7496625002946702e-05, | |
| "loss": 1.7938, | |
| "step": 37 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "learning_rate": 1.734514253098589e-05, | |
| "loss": 1.8199, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 0.27, | |
| "learning_rate": 1.7189908153577473e-05, | |
| "loss": 1.7558, | |
| "step": 39 | |
| }, | |
| { | |
| "epoch": 0.28, | |
| "learning_rate": 1.7031001164581828e-05, | |
| "loss": 1.763, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.28, | |
| "learning_rate": 1.6868502733832647e-05, | |
| "loss": 1.8022, | |
| "step": 41 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "learning_rate": 1.670249586567531e-05, | |
| "loss": 1.7876, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 0.3, | |
| "learning_rate": 1.6533065356568206e-05, | |
| "loss": 1.7981, | |
| "step": 43 | |
| }, | |
| { | |
| "epoch": 0.31, | |
| "learning_rate": 1.636029775176862e-05, | |
| "loss": 1.7688, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 0.31, | |
| "learning_rate": 1.618428130112533e-05, | |
| "loss": 1.7708, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "learning_rate": 1.6005105914000508e-05, | |
| "loss": 1.7715, | |
| "step": 46 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "learning_rate": 1.5822863113343934e-05, | |
| "loss": 1.7705, | |
| "step": 47 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "learning_rate": 1.5637645988943008e-05, | |
| "loss": 1.7814, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 0.34, | |
| "learning_rate": 1.544954914987238e-05, | |
| "loss": 1.7909, | |
| "step": 49 | |
| }, | |
| { | |
| "epoch": 0.35, | |
| "learning_rate": 1.5258668676167548e-05, | |
| "loss": 1.7873, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.35, | |
| "learning_rate": 1.5065102069747117e-05, | |
| "loss": 1.7985, | |
| "step": 51 | |
| }, | |
| { | |
| "epoch": 0.36, | |
| "learning_rate": 1.48689482046087e-05, | |
| "loss": 1.7832, | |
| "step": 52 | |
| }, | |
| { | |
| "epoch": 0.37, | |
| "learning_rate": 1.467030727632401e-05, | |
| "loss": 1.7282, | |
| "step": 53 | |
| }, | |
| { | |
| "epoch": 0.38, | |
| "learning_rate": 1.4469280750858854e-05, | |
| "loss": 1.7772, | |
| "step": 54 | |
| }, | |
| { | |
| "epoch": 0.38, | |
| "learning_rate": 1.4265971312744252e-05, | |
| "loss": 1.8001, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 0.39, | |
| "learning_rate": 1.4060482812625055e-05, | |
| "loss": 1.8, | |
| "step": 56 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "learning_rate": 1.3852920214212966e-05, | |
| "loss": 1.7753, | |
| "step": 57 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "learning_rate": 1.3643389540670963e-05, | |
| "loss": 1.7986, | |
| "step": 58 | |
| }, | |
| { | |
| "epoch": 0.41, | |
| "learning_rate": 1.3431997820456592e-05, | |
| "loss": 1.7699, | |
| "step": 59 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "learning_rate": 1.3218853032651719e-05, | |
| "loss": 1.7632, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "learning_rate": 1.3004064051806712e-05, | |
| "loss": 1.7823, | |
| "step": 61 | |
| }, | |
| { | |
| "epoch": 0.43, | |
| "learning_rate": 1.2787740592327232e-05, | |
| "loss": 1.7726, | |
| "step": 62 | |
| }, | |
| { | |
| "epoch": 0.44, | |
| "learning_rate": 1.2569993152432028e-05, | |
| "loss": 1.7821, | |
| "step": 63 | |
| }, | |
| { | |
| "epoch": 0.44, | |
| "learning_rate": 1.2350932957710322e-05, | |
| "loss": 1.7243, | |
| "step": 64 | |
| }, | |
| { | |
| "epoch": 0.45, | |
| "learning_rate": 1.2130671904307692e-05, | |
| "loss": 1.7454, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "learning_rate": 1.1909322501769407e-05, | |
| "loss": 1.779, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 0.47, | |
| "learning_rate": 1.1686997815570473e-05, | |
| "loss": 1.7664, | |
| "step": 67 | |
| }, | |
| { | |
| "epoch": 0.47, | |
| "learning_rate": 1.1463811409361667e-05, | |
| "loss": 1.7372, | |
| "step": 68 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "learning_rate": 1.1239877286961123e-05, | |
| "loss": 1.7553, | |
| "step": 69 | |
| }, | |
| { | |
| "epoch": 0.49, | |
| "learning_rate": 1.1015309834121083e-05, | |
| "loss": 1.7569, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.49, | |
| "learning_rate": 1.079022376009955e-05, | |
| "loss": 1.6978, | |
| "step": 71 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "learning_rate": 1.05647340390667e-05, | |
| "loss": 1.7539, | |
| "step": 72 | |
| }, | |
| { | |
| "epoch": 0.51, | |
| "learning_rate": 1.0338955851375962e-05, | |
| "loss": 1.7037, | |
| "step": 73 | |
| }, | |
| { | |
| "epoch": 0.51, | |
| "learning_rate": 1.01130045247298e-05, | |
| "loss": 1.7486, | |
| "step": 74 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "learning_rate": 9.886995475270205e-06, | |
| "loss": 1.7398, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.53, | |
| "learning_rate": 9.661044148624038e-06, | |
| "loss": 1.7263, | |
| "step": 76 | |
| }, | |
| { | |
| "epoch": 0.53, | |
| "learning_rate": 9.435265960933304e-06, | |
| "loss": 1.6939, | |
| "step": 77 | |
| }, | |
| { | |
| "epoch": 0.54, | |
| "learning_rate": 9.209776239900453e-06, | |
| "loss": 1.7249, | |
| "step": 78 | |
| }, | |
| { | |
| "epoch": 0.55, | |
| "learning_rate": 8.98469016587892e-06, | |
| "loss": 1.704, | |
| "step": 79 | |
| }, | |
| { | |
| "epoch": 0.56, | |
| "learning_rate": 8.76012271303888e-06, | |
| "loss": 1.7639, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.56, | |
| "learning_rate": 8.536188590638334e-06, | |
| "loss": 1.7387, | |
| "step": 81 | |
| }, | |
| { | |
| "epoch": 0.57, | |
| "learning_rate": 8.313002184429529e-06, | |
| "loss": 1.7466, | |
| "step": 82 | |
| }, | |
| { | |
| "epoch": 0.58, | |
| "learning_rate": 8.090677498230598e-06, | |
| "loss": 1.7444, | |
| "step": 83 | |
| }, | |
| { | |
| "epoch": 0.58, | |
| "learning_rate": 7.869328095692313e-06, | |
| "loss": 1.77, | |
| "step": 84 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "learning_rate": 7.649067042289681e-06, | |
| "loss": 1.7414, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 0.6, | |
| "learning_rate": 7.430006847567972e-06, | |
| "loss": 1.7505, | |
| "step": 86 | |
| }, | |
| { | |
| "epoch": 0.6, | |
| "learning_rate": 7.2122594076727705e-06, | |
| "loss": 1.7738, | |
| "step": 87 | |
| }, | |
| { | |
| "epoch": 0.61, | |
| "learning_rate": 6.995935948193294e-06, | |
| "loss": 1.754, | |
| "step": 88 | |
| }, | |
| { | |
| "epoch": 0.62, | |
| "learning_rate": 6.781146967348283e-06, | |
| "loss": 1.6997, | |
| "step": 89 | |
| }, | |
| { | |
| "epoch": 0.62, | |
| "learning_rate": 6.568002179543409e-06, | |
| "loss": 1.7367, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.63, | |
| "learning_rate": 6.356610459329038e-06, | |
| "loss": 1.7594, | |
| "step": 91 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "learning_rate": 6.147079785787038e-06, | |
| "loss": 1.7171, | |
| "step": 92 | |
| }, | |
| { | |
| "epoch": 0.65, | |
| "learning_rate": 5.93951718737495e-06, | |
| "loss": 1.7117, | |
| "step": 93 | |
| }, | |
| { | |
| "epoch": 0.65, | |
| "learning_rate": 5.7340286872557515e-06, | |
| "loss": 1.7322, | |
| "step": 94 | |
| }, | |
| { | |
| "epoch": 0.66, | |
| "learning_rate": 5.530719249141148e-06, | |
| "loss": 1.7317, | |
| "step": 95 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "learning_rate": 5.329692723675994e-06, | |
| "loss": 1.6866, | |
| "step": 96 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "learning_rate": 5.131051795391302e-06, | |
| "loss": 1.7128, | |
| "step": 97 | |
| }, | |
| { | |
| "epoch": 0.68, | |
| "learning_rate": 4.934897930252887e-06, | |
| "loss": 1.7262, | |
| "step": 98 | |
| }, | |
| { | |
| "epoch": 0.69, | |
| "learning_rate": 4.7413313238324556e-06, | |
| "loss": 1.7017, | |
| "step": 99 | |
| }, | |
| { | |
| "epoch": 0.69, | |
| "learning_rate": 4.550450850127626e-06, | |
| "loss": 1.6793, | |
| "step": 100 | |
| } | |
| ], | |
| "logging_steps": 1.0, | |
| "max_steps": 144, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 100, | |
| "total_flos": 0.0, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |