| { | |
| "best_metric": 0.9380890727043152, | |
| "best_model_checkpoint": "data/Gemma-2-2B_task-3_180-samples_config-2_full_auto/checkpoint-76", | |
| "epoch": 16.0, | |
| "eval_steps": 500, | |
| "global_step": 136, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.11764705882352941, | |
| "grad_norm": 0.43080374598503113, | |
| "learning_rate": 2.5e-06, | |
| "loss": 1.3882, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.23529411764705882, | |
| "grad_norm": 0.4670499265193939, | |
| "learning_rate": 5e-06, | |
| "loss": 1.4021, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.47058823529411764, | |
| "grad_norm": 0.4263714551925659, | |
| "learning_rate": 1e-05, | |
| "loss": 1.4051, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.7058823529411765, | |
| "grad_norm": 0.40158483386039734, | |
| "learning_rate": 1.5e-05, | |
| "loss": 1.3466, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.9411764705882353, | |
| "grad_norm": 0.39395830035209656, | |
| "learning_rate": 2e-05, | |
| "loss": 1.3868, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.9411764705882353, | |
| "eval_loss": 1.3681319952011108, | |
| "eval_runtime": 11.3614, | |
| "eval_samples_per_second": 3.169, | |
| "eval_steps_per_second": 3.169, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 1.1764705882352942, | |
| "grad_norm": 0.43305495381355286, | |
| "learning_rate": 2.5e-05, | |
| "loss": 1.3965, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 1.4117647058823528, | |
| "grad_norm": 0.37267065048217773, | |
| "learning_rate": 3e-05, | |
| "loss": 1.3279, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 1.6470588235294117, | |
| "grad_norm": 0.33385807275772095, | |
| "learning_rate": 3.5e-05, | |
| "loss": 1.3028, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 1.8823529411764706, | |
| "grad_norm": 0.31279081106185913, | |
| "learning_rate": 4e-05, | |
| "loss": 1.2861, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_loss": 1.233823299407959, | |
| "eval_runtime": 11.3535, | |
| "eval_samples_per_second": 3.171, | |
| "eval_steps_per_second": 3.171, | |
| "step": 17 | |
| }, | |
| { | |
| "epoch": 2.1176470588235294, | |
| "grad_norm": 0.25079676508903503, | |
| "learning_rate": 4.5e-05, | |
| "loss": 1.2528, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 2.3529411764705883, | |
| "grad_norm": 0.22452540695667267, | |
| "learning_rate": 5e-05, | |
| "loss": 1.1858, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 2.588235294117647, | |
| "grad_norm": 0.24543669819831848, | |
| "learning_rate": 5.500000000000001e-05, | |
| "loss": 1.1775, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 2.8235294117647056, | |
| "grad_norm": 0.26882806420326233, | |
| "learning_rate": 6e-05, | |
| "loss": 1.1878, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 2.9411764705882355, | |
| "eval_loss": 1.1353175640106201, | |
| "eval_runtime": 11.3572, | |
| "eval_samples_per_second": 3.17, | |
| "eval_steps_per_second": 3.17, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 3.0588235294117645, | |
| "grad_norm": 0.2506559193134308, | |
| "learning_rate": 6.500000000000001e-05, | |
| "loss": 1.1169, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 3.2941176470588234, | |
| "grad_norm": 0.22496089339256287, | |
| "learning_rate": 7e-05, | |
| "loss": 1.0775, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 3.5294117647058822, | |
| "grad_norm": 0.2545152008533478, | |
| "learning_rate": 7.500000000000001e-05, | |
| "loss": 1.048, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 3.764705882352941, | |
| "grad_norm": 0.4344964027404785, | |
| "learning_rate": 8e-05, | |
| "loss": 1.0626, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.22770516574382782, | |
| "learning_rate": 8.5e-05, | |
| "loss": 0.9929, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_loss": 1.0231775045394897, | |
| "eval_runtime": 11.3499, | |
| "eval_samples_per_second": 3.172, | |
| "eval_steps_per_second": 3.172, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 4.235294117647059, | |
| "grad_norm": 0.22388432919979095, | |
| "learning_rate": 9e-05, | |
| "loss": 0.9831, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 4.470588235294118, | |
| "grad_norm": 0.18862313032150269, | |
| "learning_rate": 9.5e-05, | |
| "loss": 0.9578, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 4.705882352941177, | |
| "grad_norm": 0.15573561191558838, | |
| "learning_rate": 0.0001, | |
| "loss": 0.9595, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 4.9411764705882355, | |
| "grad_norm": 0.1683398336172104, | |
| "learning_rate": 9.999238475781957e-05, | |
| "loss": 0.9221, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 4.9411764705882355, | |
| "eval_loss": 0.9788100719451904, | |
| "eval_runtime": 11.3499, | |
| "eval_samples_per_second": 3.172, | |
| "eval_steps_per_second": 3.172, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 5.176470588235294, | |
| "grad_norm": 0.16426807641983032, | |
| "learning_rate": 9.99695413509548e-05, | |
| "loss": 0.8977, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 5.411764705882353, | |
| "grad_norm": 0.15270468592643738, | |
| "learning_rate": 9.99314767377287e-05, | |
| "loss": 0.9233, | |
| "step": 46 | |
| }, | |
| { | |
| "epoch": 5.647058823529412, | |
| "grad_norm": 0.14316344261169434, | |
| "learning_rate": 9.987820251299122e-05, | |
| "loss": 0.898, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 5.882352941176471, | |
| "grad_norm": 0.1501656472682953, | |
| "learning_rate": 9.980973490458728e-05, | |
| "loss": 0.9315, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_loss": 0.9562962651252747, | |
| "eval_runtime": 11.3445, | |
| "eval_samples_per_second": 3.173, | |
| "eval_steps_per_second": 3.173, | |
| "step": 51 | |
| }, | |
| { | |
| "epoch": 6.117647058823529, | |
| "grad_norm": 0.13924811780452728, | |
| "learning_rate": 9.972609476841367e-05, | |
| "loss": 0.8689, | |
| "step": 52 | |
| }, | |
| { | |
| "epoch": 6.352941176470588, | |
| "grad_norm": 0.142573282122612, | |
| "learning_rate": 9.962730758206611e-05, | |
| "loss": 0.8584, | |
| "step": 54 | |
| }, | |
| { | |
| "epoch": 6.588235294117647, | |
| "grad_norm": 0.14089959859848022, | |
| "learning_rate": 9.951340343707852e-05, | |
| "loss": 0.8839, | |
| "step": 56 | |
| }, | |
| { | |
| "epoch": 6.823529411764706, | |
| "grad_norm": 0.14376065135002136, | |
| "learning_rate": 9.938441702975689e-05, | |
| "loss": 0.9202, | |
| "step": 58 | |
| }, | |
| { | |
| "epoch": 6.9411764705882355, | |
| "eval_loss": 0.9455355405807495, | |
| "eval_runtime": 11.3536, | |
| "eval_samples_per_second": 3.171, | |
| "eval_steps_per_second": 3.171, | |
| "step": 59 | |
| }, | |
| { | |
| "epoch": 7.0588235294117645, | |
| "grad_norm": 0.14621061086654663, | |
| "learning_rate": 9.924038765061042e-05, | |
| "loss": 0.8807, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 7.294117647058823, | |
| "grad_norm": 0.15416990220546722, | |
| "learning_rate": 9.908135917238321e-05, | |
| "loss": 0.8629, | |
| "step": 62 | |
| }, | |
| { | |
| "epoch": 7.529411764705882, | |
| "grad_norm": 0.15954792499542236, | |
| "learning_rate": 9.890738003669029e-05, | |
| "loss": 0.8322, | |
| "step": 64 | |
| }, | |
| { | |
| "epoch": 7.764705882352941, | |
| "grad_norm": 0.1517380326986313, | |
| "learning_rate": 9.871850323926177e-05, | |
| "loss": 0.8311, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "grad_norm": 0.18680177628993988, | |
| "learning_rate": 9.851478631379982e-05, | |
| "loss": 0.8604, | |
| "step": 68 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_loss": 0.9398467540740967, | |
| "eval_runtime": 11.3714, | |
| "eval_samples_per_second": 3.166, | |
| "eval_steps_per_second": 3.166, | |
| "step": 68 | |
| }, | |
| { | |
| "epoch": 8.235294117647058, | |
| "grad_norm": 0.16186895966529846, | |
| "learning_rate": 9.829629131445342e-05, | |
| "loss": 0.8201, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 8.470588235294118, | |
| "grad_norm": 0.17347827553749084, | |
| "learning_rate": 9.806308479691595e-05, | |
| "loss": 0.8493, | |
| "step": 72 | |
| }, | |
| { | |
| "epoch": 8.705882352941176, | |
| "grad_norm": 0.16852523386478424, | |
| "learning_rate": 9.781523779815179e-05, | |
| "loss": 0.8367, | |
| "step": 74 | |
| }, | |
| { | |
| "epoch": 8.941176470588236, | |
| "grad_norm": 0.19314789772033691, | |
| "learning_rate": 9.755282581475769e-05, | |
| "loss": 0.7955, | |
| "step": 76 | |
| }, | |
| { | |
| "epoch": 8.941176470588236, | |
| "eval_loss": 0.9380890727043152, | |
| "eval_runtime": 11.3476, | |
| "eval_samples_per_second": 3.172, | |
| "eval_steps_per_second": 3.172, | |
| "step": 76 | |
| }, | |
| { | |
| "epoch": 9.176470588235293, | |
| "grad_norm": 0.17392347753047943, | |
| "learning_rate": 9.727592877996585e-05, | |
| "loss": 0.8045, | |
| "step": 78 | |
| }, | |
| { | |
| "epoch": 9.411764705882353, | |
| "grad_norm": 0.18482913076877594, | |
| "learning_rate": 9.698463103929542e-05, | |
| "loss": 0.8057, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 9.647058823529411, | |
| "grad_norm": 0.20443104207515717, | |
| "learning_rate": 9.667902132486009e-05, | |
| "loss": 0.7994, | |
| "step": 82 | |
| }, | |
| { | |
| "epoch": 9.882352941176471, | |
| "grad_norm": 0.21509741246700287, | |
| "learning_rate": 9.635919272833938e-05, | |
| "loss": 0.7864, | |
| "step": 84 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_loss": 0.9430419206619263, | |
| "eval_runtime": 11.3488, | |
| "eval_samples_per_second": 3.172, | |
| "eval_steps_per_second": 3.172, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 10.117647058823529, | |
| "grad_norm": 0.2130606323480606, | |
| "learning_rate": 9.602524267262203e-05, | |
| "loss": 0.7651, | |
| "step": 86 | |
| }, | |
| { | |
| "epoch": 10.352941176470589, | |
| "grad_norm": 0.2297953963279724, | |
| "learning_rate": 9.567727288213005e-05, | |
| "loss": 0.7833, | |
| "step": 88 | |
| }, | |
| { | |
| "epoch": 10.588235294117647, | |
| "grad_norm": 0.2459597885608673, | |
| "learning_rate": 9.53153893518325e-05, | |
| "loss": 0.7477, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 10.823529411764707, | |
| "grad_norm": 0.2760981321334839, | |
| "learning_rate": 9.493970231495835e-05, | |
| "loss": 0.7442, | |
| "step": 92 | |
| }, | |
| { | |
| "epoch": 10.941176470588236, | |
| "eval_loss": 0.9520426988601685, | |
| "eval_runtime": 11.3461, | |
| "eval_samples_per_second": 3.173, | |
| "eval_steps_per_second": 3.173, | |
| "step": 93 | |
| }, | |
| { | |
| "epoch": 11.058823529411764, | |
| "grad_norm": 0.2601296007633209, | |
| "learning_rate": 9.45503262094184e-05, | |
| "loss": 0.7805, | |
| "step": 94 | |
| }, | |
| { | |
| "epoch": 11.294117647058824, | |
| "grad_norm": 0.2620711028575897, | |
| "learning_rate": 9.414737964294636e-05, | |
| "loss": 0.7653, | |
| "step": 96 | |
| }, | |
| { | |
| "epoch": 11.529411764705882, | |
| "grad_norm": 0.41495072841644287, | |
| "learning_rate": 9.373098535696979e-05, | |
| "loss": 0.6947, | |
| "step": 98 | |
| }, | |
| { | |
| "epoch": 11.764705882352942, | |
| "grad_norm": 0.3301529586315155, | |
| "learning_rate": 9.330127018922194e-05, | |
| "loss": 0.7291, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "grad_norm": 0.3447425067424774, | |
| "learning_rate": 9.285836503510562e-05, | |
| "loss": 0.6892, | |
| "step": 102 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "eval_loss": 0.9796677827835083, | |
| "eval_runtime": 11.3558, | |
| "eval_samples_per_second": 3.17, | |
| "eval_steps_per_second": 3.17, | |
| "step": 102 | |
| }, | |
| { | |
| "epoch": 12.235294117647058, | |
| "grad_norm": 0.3976879119873047, | |
| "learning_rate": 9.24024048078213e-05, | |
| "loss": 0.6707, | |
| "step": 104 | |
| }, | |
| { | |
| "epoch": 12.470588235294118, | |
| "grad_norm": 0.44007667899131775, | |
| "learning_rate": 9.193352839727121e-05, | |
| "loss": 0.6681, | |
| "step": 106 | |
| }, | |
| { | |
| "epoch": 12.705882352941176, | |
| "grad_norm": 0.45111826062202454, | |
| "learning_rate": 9.145187862775209e-05, | |
| "loss": 0.6864, | |
| "step": 108 | |
| }, | |
| { | |
| "epoch": 12.941176470588236, | |
| "grad_norm": 0.5192270278930664, | |
| "learning_rate": 9.09576022144496e-05, | |
| "loss": 0.6833, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 12.941176470588236, | |
| "eval_loss": 0.9951361417770386, | |
| "eval_runtime": 11.3578, | |
| "eval_samples_per_second": 3.17, | |
| "eval_steps_per_second": 3.17, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 13.176470588235293, | |
| "grad_norm": 0.44144096970558167, | |
| "learning_rate": 9.045084971874738e-05, | |
| "loss": 0.6609, | |
| "step": 112 | |
| }, | |
| { | |
| "epoch": 13.411764705882353, | |
| "grad_norm": 0.5179834961891174, | |
| "learning_rate": 8.993177550236464e-05, | |
| "loss": 0.5923, | |
| "step": 114 | |
| }, | |
| { | |
| "epoch": 13.647058823529411, | |
| "grad_norm": 0.5081561803817749, | |
| "learning_rate": 8.940053768033609e-05, | |
| "loss": 0.6329, | |
| "step": 116 | |
| }, | |
| { | |
| "epoch": 13.882352941176471, | |
| "grad_norm": 0.5672034025192261, | |
| "learning_rate": 8.885729807284856e-05, | |
| "loss": 0.649, | |
| "step": 118 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "eval_loss": 1.0528039932250977, | |
| "eval_runtime": 11.3545, | |
| "eval_samples_per_second": 3.171, | |
| "eval_steps_per_second": 3.171, | |
| "step": 119 | |
| }, | |
| { | |
| "epoch": 14.117647058823529, | |
| "grad_norm": 0.5097953677177429, | |
| "learning_rate": 8.83022221559489e-05, | |
| "loss": 0.6355, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 14.352941176470589, | |
| "grad_norm": 0.6940440535545349, | |
| "learning_rate": 8.773547901113862e-05, | |
| "loss": 0.5308, | |
| "step": 122 | |
| }, | |
| { | |
| "epoch": 14.588235294117647, | |
| "grad_norm": 0.6824241280555725, | |
| "learning_rate": 8.715724127386972e-05, | |
| "loss": 0.6125, | |
| "step": 124 | |
| }, | |
| { | |
| "epoch": 14.823529411764707, | |
| "grad_norm": 0.7449726462364197, | |
| "learning_rate": 8.656768508095853e-05, | |
| "loss": 0.5557, | |
| "step": 126 | |
| }, | |
| { | |
| "epoch": 14.941176470588236, | |
| "eval_loss": 1.1158599853515625, | |
| "eval_runtime": 11.3464, | |
| "eval_samples_per_second": 3.173, | |
| "eval_steps_per_second": 3.173, | |
| "step": 127 | |
| }, | |
| { | |
| "epoch": 15.058823529411764, | |
| "grad_norm": 0.6661369800567627, | |
| "learning_rate": 8.596699001693255e-05, | |
| "loss": 0.6004, | |
| "step": 128 | |
| }, | |
| { | |
| "epoch": 15.294117647058824, | |
| "grad_norm": 0.7347066402435303, | |
| "learning_rate": 8.535533905932738e-05, | |
| "loss": 0.5369, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 15.529411764705882, | |
| "grad_norm": 0.7058692574501038, | |
| "learning_rate": 8.473291852294987e-05, | |
| "loss": 0.5507, | |
| "step": 132 | |
| }, | |
| { | |
| "epoch": 15.764705882352942, | |
| "grad_norm": 0.8206327557563782, | |
| "learning_rate": 8.409991800312493e-05, | |
| "loss": 0.5566, | |
| "step": 134 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "grad_norm": 0.7464694380760193, | |
| "learning_rate": 8.345653031794292e-05, | |
| "loss": 0.4908, | |
| "step": 136 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "eval_loss": 1.1621650457382202, | |
| "eval_runtime": 11.3513, | |
| "eval_samples_per_second": 3.171, | |
| "eval_steps_per_second": 3.171, | |
| "step": 136 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "step": 136, | |
| "total_flos": 5.900668991910707e+16, | |
| "train_loss": 0.8783913802574662, | |
| "train_runtime": 2482.4714, | |
| "train_samples_per_second": 2.739, | |
| "train_steps_per_second": 0.161 | |
| } | |
| ], | |
| "logging_steps": 2, | |
| "max_steps": 400, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 50, | |
| "save_steps": 25, | |
| "stateful_callbacks": { | |
| "EarlyStoppingCallback": { | |
| "args": { | |
| "early_stopping_patience": 7, | |
| "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": 5.900668991910707e+16, | |
| "train_batch_size": 1, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |