| { | |
| "best_metric": 1.0351691246032715, | |
| "best_model_checkpoint": "data/Llama-31-8B_task-2_180-samples_config-2_full/checkpoint-85", | |
| "epoch": 16.941176470588236, | |
| "eval_steps": 500, | |
| "global_step": 144, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.11764705882352941, | |
| "grad_norm": 0.5214905142784119, | |
| "learning_rate": 2.5e-06, | |
| "loss": 1.5966, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.23529411764705882, | |
| "grad_norm": 0.5420028567314148, | |
| "learning_rate": 5e-06, | |
| "loss": 1.6122, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.47058823529411764, | |
| "grad_norm": 0.4707539975643158, | |
| "learning_rate": 1e-05, | |
| "loss": 1.6094, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.7058823529411765, | |
| "grad_norm": 0.4507639706134796, | |
| "learning_rate": 1.5e-05, | |
| "loss": 1.5539, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.9411764705882353, | |
| "grad_norm": 0.44413137435913086, | |
| "learning_rate": 2e-05, | |
| "loss": 1.5563, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.9411764705882353, | |
| "eval_loss": 1.557669758796692, | |
| "eval_runtime": 37.0524, | |
| "eval_samples_per_second": 0.972, | |
| "eval_steps_per_second": 0.972, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 1.1764705882352942, | |
| "grad_norm": 0.502502977848053, | |
| "learning_rate": 2.5e-05, | |
| "loss": 1.5536, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 1.4117647058823528, | |
| "grad_norm": 0.381038099527359, | |
| "learning_rate": 3e-05, | |
| "loss": 1.5016, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 1.6470588235294117, | |
| "grad_norm": 0.3993341326713562, | |
| "learning_rate": 3.5e-05, | |
| "loss": 1.4584, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 1.8823529411764706, | |
| "grad_norm": 0.33516234159469604, | |
| "learning_rate": 4e-05, | |
| "loss": 1.4499, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_loss": 1.4146790504455566, | |
| "eval_runtime": 37.0216, | |
| "eval_samples_per_second": 0.972, | |
| "eval_steps_per_second": 0.972, | |
| "step": 17 | |
| }, | |
| { | |
| "epoch": 2.1176470588235294, | |
| "grad_norm": 0.2883775234222412, | |
| "learning_rate": 4.5e-05, | |
| "loss": 1.4232, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 2.3529411764705883, | |
| "grad_norm": 0.23242510855197906, | |
| "learning_rate": 5e-05, | |
| "loss": 1.3566, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 2.588235294117647, | |
| "grad_norm": 0.24405810236930847, | |
| "learning_rate": 5.500000000000001e-05, | |
| "loss": 1.3283, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 2.8235294117647056, | |
| "grad_norm": 0.2586018145084381, | |
| "learning_rate": 6e-05, | |
| "loss": 1.3435, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 2.9411764705882355, | |
| "eval_loss": 1.2967584133148193, | |
| "eval_runtime": 37.0078, | |
| "eval_samples_per_second": 0.973, | |
| "eval_steps_per_second": 0.973, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 3.0588235294117645, | |
| "grad_norm": 0.2536551356315613, | |
| "learning_rate": 6.500000000000001e-05, | |
| "loss": 1.271, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 3.2941176470588234, | |
| "grad_norm": 0.20180310308933258, | |
| "learning_rate": 7e-05, | |
| "loss": 1.2373, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 3.5294117647058822, | |
| "grad_norm": 0.23023776710033417, | |
| "learning_rate": 7.500000000000001e-05, | |
| "loss": 1.1788, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 3.764705882352941, | |
| "grad_norm": 0.3016901910305023, | |
| "learning_rate": 8e-05, | |
| "loss": 1.1924, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.2720400094985962, | |
| "learning_rate": 8.5e-05, | |
| "loss": 1.0893, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_loss": 1.1421051025390625, | |
| "eval_runtime": 36.9984, | |
| "eval_samples_per_second": 0.973, | |
| "eval_steps_per_second": 0.973, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 4.235294117647059, | |
| "grad_norm": 0.2179834544658661, | |
| "learning_rate": 9e-05, | |
| "loss": 1.0829, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 4.470588235294118, | |
| "grad_norm": 0.17470435798168182, | |
| "learning_rate": 9.5e-05, | |
| "loss": 1.0623, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 4.705882352941177, | |
| "grad_norm": 0.1350124180316925, | |
| "learning_rate": 0.0001, | |
| "loss": 1.0629, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 4.9411764705882355, | |
| "grad_norm": 0.157378152012825, | |
| "learning_rate": 9.999238475781957e-05, | |
| "loss": 1.0304, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 4.9411764705882355, | |
| "eval_loss": 1.0948100090026855, | |
| "eval_runtime": 36.9995, | |
| "eval_samples_per_second": 0.973, | |
| "eval_steps_per_second": 0.973, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 5.176470588235294, | |
| "grad_norm": 0.12661626935005188, | |
| "learning_rate": 9.99695413509548e-05, | |
| "loss": 0.9981, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 5.411764705882353, | |
| "grad_norm": 0.13132977485656738, | |
| "learning_rate": 9.99314767377287e-05, | |
| "loss": 1.0463, | |
| "step": 46 | |
| }, | |
| { | |
| "epoch": 5.647058823529412, | |
| "grad_norm": 0.13775759935379028, | |
| "learning_rate": 9.987820251299122e-05, | |
| "loss": 1.0092, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 5.882352941176471, | |
| "grad_norm": 0.11698019504547119, | |
| "learning_rate": 9.980973490458728e-05, | |
| "loss": 1.0336, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_loss": 1.0724878311157227, | |
| "eval_runtime": 37.0084, | |
| "eval_samples_per_second": 0.973, | |
| "eval_steps_per_second": 0.973, | |
| "step": 51 | |
| }, | |
| { | |
| "epoch": 6.117647058823529, | |
| "grad_norm": 0.12450984865427017, | |
| "learning_rate": 9.972609476841367e-05, | |
| "loss": 0.9865, | |
| "step": 52 | |
| }, | |
| { | |
| "epoch": 6.352941176470588, | |
| "grad_norm": 0.13155825436115265, | |
| "learning_rate": 9.962730758206611e-05, | |
| "loss": 0.9607, | |
| "step": 54 | |
| }, | |
| { | |
| "epoch": 6.588235294117647, | |
| "grad_norm": 0.14671258628368378, | |
| "learning_rate": 9.951340343707852e-05, | |
| "loss": 0.9987, | |
| "step": 56 | |
| }, | |
| { | |
| "epoch": 6.823529411764706, | |
| "grad_norm": 0.12457507848739624, | |
| "learning_rate": 9.938441702975689e-05, | |
| "loss": 1.0393, | |
| "step": 58 | |
| }, | |
| { | |
| "epoch": 6.9411764705882355, | |
| "eval_loss": 1.0579314231872559, | |
| "eval_runtime": 36.9948, | |
| "eval_samples_per_second": 0.973, | |
| "eval_steps_per_second": 0.973, | |
| "step": 59 | |
| }, | |
| { | |
| "epoch": 7.0588235294117645, | |
| "grad_norm": 0.13238783180713654, | |
| "learning_rate": 9.924038765061042e-05, | |
| "loss": 0.994, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 7.294117647058823, | |
| "grad_norm": 0.1260843425989151, | |
| "learning_rate": 9.908135917238321e-05, | |
| "loss": 0.98, | |
| "step": 62 | |
| }, | |
| { | |
| "epoch": 7.529411764705882, | |
| "grad_norm": 0.13164746761322021, | |
| "learning_rate": 9.890738003669029e-05, | |
| "loss": 0.9447, | |
| "step": 64 | |
| }, | |
| { | |
| "epoch": 7.764705882352941, | |
| "grad_norm": 0.13616593182086945, | |
| "learning_rate": 9.871850323926177e-05, | |
| "loss": 0.9456, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "grad_norm": 0.15812847018241882, | |
| "learning_rate": 9.851478631379982e-05, | |
| "loss": 0.9727, | |
| "step": 68 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_loss": 1.0466517210006714, | |
| "eval_runtime": 37.0, | |
| "eval_samples_per_second": 0.973, | |
| "eval_steps_per_second": 0.973, | |
| "step": 68 | |
| }, | |
| { | |
| "epoch": 8.235294117647058, | |
| "grad_norm": 0.14172162115573883, | |
| "learning_rate": 9.829629131445342e-05, | |
| "loss": 0.9399, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 8.470588235294118, | |
| "grad_norm": 0.15690505504608154, | |
| "learning_rate": 9.806308479691595e-05, | |
| "loss": 0.9677, | |
| "step": 72 | |
| }, | |
| { | |
| "epoch": 8.705882352941176, | |
| "grad_norm": 0.15911728143692017, | |
| "learning_rate": 9.781523779815179e-05, | |
| "loss": 0.9473, | |
| "step": 74 | |
| }, | |
| { | |
| "epoch": 8.941176470588236, | |
| "grad_norm": 0.1991043984889984, | |
| "learning_rate": 9.755282581475769e-05, | |
| "loss": 0.9068, | |
| "step": 76 | |
| }, | |
| { | |
| "epoch": 8.941176470588236, | |
| "eval_loss": 1.039350986480713, | |
| "eval_runtime": 37.0233, | |
| "eval_samples_per_second": 0.972, | |
| "eval_steps_per_second": 0.972, | |
| "step": 76 | |
| }, | |
| { | |
| "epoch": 9.176470588235293, | |
| "grad_norm": 0.17068785429000854, | |
| "learning_rate": 9.727592877996585e-05, | |
| "loss": 0.9201, | |
| "step": 78 | |
| }, | |
| { | |
| "epoch": 9.411764705882353, | |
| "grad_norm": 0.16773968935012817, | |
| "learning_rate": 9.698463103929542e-05, | |
| "loss": 0.925, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 9.647058823529411, | |
| "grad_norm": 0.17334376275539398, | |
| "learning_rate": 9.667902132486009e-05, | |
| "loss": 0.9152, | |
| "step": 82 | |
| }, | |
| { | |
| "epoch": 9.882352941176471, | |
| "grad_norm": 0.18037906289100647, | |
| "learning_rate": 9.635919272833938e-05, | |
| "loss": 0.9015, | |
| "step": 84 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_loss": 1.0351691246032715, | |
| "eval_runtime": 37.0191, | |
| "eval_samples_per_second": 0.972, | |
| "eval_steps_per_second": 0.972, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 10.117647058823529, | |
| "grad_norm": 0.17189693450927734, | |
| "learning_rate": 9.602524267262203e-05, | |
| "loss": 0.8763, | |
| "step": 86 | |
| }, | |
| { | |
| "epoch": 10.352941176470589, | |
| "grad_norm": 0.18831922113895416, | |
| "learning_rate": 9.567727288213005e-05, | |
| "loss": 0.9003, | |
| "step": 88 | |
| }, | |
| { | |
| "epoch": 10.588235294117647, | |
| "grad_norm": 0.18492361903190613, | |
| "learning_rate": 9.53153893518325e-05, | |
| "loss": 0.8576, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 10.823529411764707, | |
| "grad_norm": 0.22255094349384308, | |
| "learning_rate": 9.493970231495835e-05, | |
| "loss": 0.8647, | |
| "step": 92 | |
| }, | |
| { | |
| "epoch": 10.941176470588236, | |
| "eval_loss": 1.037386417388916, | |
| "eval_runtime": 37.0133, | |
| "eval_samples_per_second": 0.973, | |
| "eval_steps_per_second": 0.973, | |
| "step": 93 | |
| }, | |
| { | |
| "epoch": 11.058823529411764, | |
| "grad_norm": 0.2781403064727783, | |
| "learning_rate": 9.45503262094184e-05, | |
| "loss": 0.904, | |
| "step": 94 | |
| }, | |
| { | |
| "epoch": 11.294117647058824, | |
| "grad_norm": 0.2218756228685379, | |
| "learning_rate": 9.414737964294636e-05, | |
| "loss": 0.8968, | |
| "step": 96 | |
| }, | |
| { | |
| "epoch": 11.529411764705882, | |
| "grad_norm": 0.2906467020511627, | |
| "learning_rate": 9.373098535696979e-05, | |
| "loss": 0.8088, | |
| "step": 98 | |
| }, | |
| { | |
| "epoch": 11.764705882352942, | |
| "grad_norm": 0.2616478502750397, | |
| "learning_rate": 9.330127018922194e-05, | |
| "loss": 0.8433, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "grad_norm": 0.26752302050590515, | |
| "learning_rate": 9.285836503510562e-05, | |
| "loss": 0.798, | |
| "step": 102 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "eval_loss": 1.052980899810791, | |
| "eval_runtime": 37.0134, | |
| "eval_samples_per_second": 0.973, | |
| "eval_steps_per_second": 0.973, | |
| "step": 102 | |
| }, | |
| { | |
| "epoch": 12.235294117647058, | |
| "grad_norm": 0.29832029342651367, | |
| "learning_rate": 9.24024048078213e-05, | |
| "loss": 0.7826, | |
| "step": 104 | |
| }, | |
| { | |
| "epoch": 12.470588235294118, | |
| "grad_norm": 0.35154664516448975, | |
| "learning_rate": 9.193352839727121e-05, | |
| "loss": 0.7799, | |
| "step": 106 | |
| }, | |
| { | |
| "epoch": 12.705882352941176, | |
| "grad_norm": 0.35771840810775757, | |
| "learning_rate": 9.145187862775209e-05, | |
| "loss": 0.8146, | |
| "step": 108 | |
| }, | |
| { | |
| "epoch": 12.941176470588236, | |
| "grad_norm": 0.3104204833507538, | |
| "learning_rate": 9.09576022144496e-05, | |
| "loss": 0.795, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 12.941176470588236, | |
| "eval_loss": 1.066689133644104, | |
| "eval_runtime": 37.0122, | |
| "eval_samples_per_second": 0.973, | |
| "eval_steps_per_second": 0.973, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 13.176470588235293, | |
| "grad_norm": 0.3114722967147827, | |
| "learning_rate": 9.045084971874738e-05, | |
| "loss": 0.7704, | |
| "step": 112 | |
| }, | |
| { | |
| "epoch": 13.411764705882353, | |
| "grad_norm": 0.39343976974487305, | |
| "learning_rate": 8.993177550236464e-05, | |
| "loss": 0.7026, | |
| "step": 114 | |
| }, | |
| { | |
| "epoch": 13.647058823529411, | |
| "grad_norm": 0.5083499550819397, | |
| "learning_rate": 8.940053768033609e-05, | |
| "loss": 0.737, | |
| "step": 116 | |
| }, | |
| { | |
| "epoch": 13.882352941176471, | |
| "grad_norm": 0.39240530133247375, | |
| "learning_rate": 8.885729807284856e-05, | |
| "loss": 0.7562, | |
| "step": 118 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "eval_loss": 1.09353506565094, | |
| "eval_runtime": 37.0188, | |
| "eval_samples_per_second": 0.972, | |
| "eval_steps_per_second": 0.972, | |
| "step": 119 | |
| }, | |
| { | |
| "epoch": 14.117647058823529, | |
| "grad_norm": 0.4597809612751007, | |
| "learning_rate": 8.83022221559489e-05, | |
| "loss": 0.7452, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 14.352941176470589, | |
| "grad_norm": 0.48050206899642944, | |
| "learning_rate": 8.773547901113862e-05, | |
| "loss": 0.6235, | |
| "step": 122 | |
| }, | |
| { | |
| "epoch": 14.588235294117647, | |
| "grad_norm": 0.45742419362068176, | |
| "learning_rate": 8.715724127386972e-05, | |
| "loss": 0.713, | |
| "step": 124 | |
| }, | |
| { | |
| "epoch": 14.823529411764707, | |
| "grad_norm": 0.5329111218452454, | |
| "learning_rate": 8.656768508095853e-05, | |
| "loss": 0.6454, | |
| "step": 126 | |
| }, | |
| { | |
| "epoch": 14.941176470588236, | |
| "eval_loss": 1.1195636987686157, | |
| "eval_runtime": 37.0104, | |
| "eval_samples_per_second": 0.973, | |
| "eval_steps_per_second": 0.973, | |
| "step": 127 | |
| }, | |
| { | |
| "epoch": 15.058823529411764, | |
| "grad_norm": 0.602240800857544, | |
| "learning_rate": 8.596699001693255e-05, | |
| "loss": 0.6981, | |
| "step": 128 | |
| }, | |
| { | |
| "epoch": 15.294117647058824, | |
| "grad_norm": 1.2815865278244019, | |
| "learning_rate": 8.535533905932738e-05, | |
| "loss": 0.6211, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 15.529411764705882, | |
| "grad_norm": 0.7014145255088806, | |
| "learning_rate": 8.473291852294987e-05, | |
| "loss": 0.6437, | |
| "step": 132 | |
| }, | |
| { | |
| "epoch": 15.764705882352942, | |
| "grad_norm": 0.7518697381019592, | |
| "learning_rate": 8.409991800312493e-05, | |
| "loss": 0.6435, | |
| "step": 134 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "grad_norm": 0.6456379294395447, | |
| "learning_rate": 8.345653031794292e-05, | |
| "loss": 0.5665, | |
| "step": 136 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "eval_loss": 1.1875462532043457, | |
| "eval_runtime": 37.0165, | |
| "eval_samples_per_second": 0.973, | |
| "eval_steps_per_second": 0.973, | |
| "step": 136 | |
| }, | |
| { | |
| "epoch": 16.235294117647058, | |
| "grad_norm": 2.2997963428497314, | |
| "learning_rate": 8.280295144952536e-05, | |
| "loss": 0.5285, | |
| "step": 138 | |
| }, | |
| { | |
| "epoch": 16.470588235294116, | |
| "grad_norm": 0.7830283045768738, | |
| "learning_rate": 8.213938048432697e-05, | |
| "loss": 0.6144, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 16.705882352941178, | |
| "grad_norm": 0.6990455389022827, | |
| "learning_rate": 8.146601955249188e-05, | |
| "loss": 0.5474, | |
| "step": 142 | |
| }, | |
| { | |
| "epoch": 16.941176470588236, | |
| "grad_norm": 1.0758552551269531, | |
| "learning_rate": 8.07830737662829e-05, | |
| "loss": 0.565, | |
| "step": 144 | |
| }, | |
| { | |
| "epoch": 16.941176470588236, | |
| "eval_loss": 1.2293667793273926, | |
| "eval_runtime": 37.0081, | |
| "eval_samples_per_second": 0.973, | |
| "eval_steps_per_second": 0.973, | |
| "step": 144 | |
| }, | |
| { | |
| "epoch": 16.941176470588236, | |
| "step": 144, | |
| "total_flos": 2.2129957988519117e+17, | |
| "train_loss": 0.9785721999489598, | |
| "train_runtime": 7905.4252, | |
| "train_samples_per_second": 0.86, | |
| "train_steps_per_second": 0.051 | |
| } | |
| ], | |
| "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": 2.2129957988519117e+17, | |
| "train_batch_size": 1, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |