| { | |
| "best_metric": 0.8710008263587952, | |
| "best_model_checkpoint": "data/Gemma-2-2B_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": 0.4646625816822052, | |
| "learning_rate": 2.5e-06, | |
| "loss": 1.3348, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.23529411764705882, | |
| "grad_norm": 0.4941255748271942, | |
| "learning_rate": 5e-06, | |
| "loss": 1.3615, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.47058823529411764, | |
| "grad_norm": 0.4547555446624756, | |
| "learning_rate": 1e-05, | |
| "loss": 1.3619, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.7058823529411765, | |
| "grad_norm": 0.4238773584365845, | |
| "learning_rate": 1.5e-05, | |
| "loss": 1.2933, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.9411764705882353, | |
| "grad_norm": 0.41738811135292053, | |
| "learning_rate": 2e-05, | |
| "loss": 1.3182, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.9411764705882353, | |
| "eval_loss": 1.314428687095642, | |
| "eval_runtime": 13.9271, | |
| "eval_samples_per_second": 2.585, | |
| "eval_steps_per_second": 2.585, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 1.1764705882352942, | |
| "grad_norm": 0.46454063057899475, | |
| "learning_rate": 2.5e-05, | |
| "loss": 1.3438, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 1.4117647058823528, | |
| "grad_norm": 0.3903467655181885, | |
| "learning_rate": 3e-05, | |
| "loss": 1.2762, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 1.6470588235294117, | |
| "grad_norm": 0.35459548234939575, | |
| "learning_rate": 3.5e-05, | |
| "loss": 1.2376, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 1.8823529411764706, | |
| "grad_norm": 0.3327643573284149, | |
| "learning_rate": 4e-05, | |
| "loss": 1.2253, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_loss": 1.1693224906921387, | |
| "eval_runtime": 13.922, | |
| "eval_samples_per_second": 2.586, | |
| "eval_steps_per_second": 2.586, | |
| "step": 17 | |
| }, | |
| { | |
| "epoch": 2.1176470588235294, | |
| "grad_norm": 0.26664429903030396, | |
| "learning_rate": 4.5e-05, | |
| "loss": 1.1858, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 2.3529411764705883, | |
| "grad_norm": 0.22536298632621765, | |
| "learning_rate": 5e-05, | |
| "loss": 1.1146, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 2.588235294117647, | |
| "grad_norm": 0.29759112000465393, | |
| "learning_rate": 5.500000000000001e-05, | |
| "loss": 1.1207, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 2.8235294117647056, | |
| "grad_norm": 0.2725456655025482, | |
| "learning_rate": 6e-05, | |
| "loss": 1.1196, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 2.9411764705882355, | |
| "eval_loss": 1.0760149955749512, | |
| "eval_runtime": 13.9219, | |
| "eval_samples_per_second": 2.586, | |
| "eval_steps_per_second": 2.586, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 3.0588235294117645, | |
| "grad_norm": 0.2560402452945709, | |
| "learning_rate": 6.500000000000001e-05, | |
| "loss": 1.052, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 3.2941176470588234, | |
| "grad_norm": 0.21541963517665863, | |
| "learning_rate": 7e-05, | |
| "loss": 1.023, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 3.5294117647058822, | |
| "grad_norm": 0.23459596931934357, | |
| "learning_rate": 7.500000000000001e-05, | |
| "loss": 1.0138, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 3.764705882352941, | |
| "grad_norm": 0.2933233678340912, | |
| "learning_rate": 8e-05, | |
| "loss": 1.0071, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.2300327569246292, | |
| "learning_rate": 8.5e-05, | |
| "loss": 0.9289, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_loss": 0.9661836624145508, | |
| "eval_runtime": 13.9237, | |
| "eval_samples_per_second": 2.586, | |
| "eval_steps_per_second": 2.586, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 4.235294117647059, | |
| "grad_norm": 0.209079310297966, | |
| "learning_rate": 9e-05, | |
| "loss": 0.9432, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 4.470588235294118, | |
| "grad_norm": 0.168331116437912, | |
| "learning_rate": 9.5e-05, | |
| "loss": 0.9028, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 4.705882352941177, | |
| "grad_norm": 0.14146201312541962, | |
| "learning_rate": 0.0001, | |
| "loss": 0.8951, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 4.9411764705882355, | |
| "grad_norm": 0.15306784212589264, | |
| "learning_rate": 9.999238475781957e-05, | |
| "loss": 0.8623, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 4.9411764705882355, | |
| "eval_loss": 0.9211580753326416, | |
| "eval_runtime": 13.9243, | |
| "eval_samples_per_second": 2.585, | |
| "eval_steps_per_second": 2.585, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 5.176470588235294, | |
| "grad_norm": 0.1463106870651245, | |
| "learning_rate": 9.99695413509548e-05, | |
| "loss": 0.8416, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 5.411764705882353, | |
| "grad_norm": 0.14740553498268127, | |
| "learning_rate": 9.99314767377287e-05, | |
| "loss": 0.8785, | |
| "step": 46 | |
| }, | |
| { | |
| "epoch": 5.647058823529412, | |
| "grad_norm": 0.1291574090719223, | |
| "learning_rate": 9.987820251299122e-05, | |
| "loss": 0.8335, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 5.882352941176471, | |
| "grad_norm": 0.14685429632663727, | |
| "learning_rate": 9.980973490458728e-05, | |
| "loss": 0.8786, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_loss": 0.8972748517990112, | |
| "eval_runtime": 13.9294, | |
| "eval_samples_per_second": 2.584, | |
| "eval_steps_per_second": 2.584, | |
| "step": 51 | |
| }, | |
| { | |
| "epoch": 6.117647058823529, | |
| "grad_norm": 0.13178876042366028, | |
| "learning_rate": 9.972609476841367e-05, | |
| "loss": 0.8167, | |
| "step": 52 | |
| }, | |
| { | |
| "epoch": 6.352941176470588, | |
| "grad_norm": 0.12248564511537552, | |
| "learning_rate": 9.962730758206611e-05, | |
| "loss": 0.8085, | |
| "step": 54 | |
| }, | |
| { | |
| "epoch": 6.588235294117647, | |
| "grad_norm": 0.12658923864364624, | |
| "learning_rate": 9.951340343707852e-05, | |
| "loss": 0.8299, | |
| "step": 56 | |
| }, | |
| { | |
| "epoch": 6.823529411764706, | |
| "grad_norm": 0.12179239839315414, | |
| "learning_rate": 9.938441702975689e-05, | |
| "loss": 0.861, | |
| "step": 58 | |
| }, | |
| { | |
| "epoch": 6.9411764705882355, | |
| "eval_loss": 0.8844776749610901, | |
| "eval_runtime": 13.9268, | |
| "eval_samples_per_second": 2.585, | |
| "eval_steps_per_second": 2.585, | |
| "step": 59 | |
| }, | |
| { | |
| "epoch": 7.0588235294117645, | |
| "grad_norm": 0.13848571479320526, | |
| "learning_rate": 9.924038765061042e-05, | |
| "loss": 0.8375, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 7.294117647058823, | |
| "grad_norm": 0.13029764592647552, | |
| "learning_rate": 9.908135917238321e-05, | |
| "loss": 0.8097, | |
| "step": 62 | |
| }, | |
| { | |
| "epoch": 7.529411764705882, | |
| "grad_norm": 0.1324029564857483, | |
| "learning_rate": 9.890738003669029e-05, | |
| "loss": 0.7823, | |
| "step": 64 | |
| }, | |
| { | |
| "epoch": 7.764705882352941, | |
| "grad_norm": 0.13409382104873657, | |
| "learning_rate": 9.871850323926177e-05, | |
| "loss": 0.774, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "grad_norm": 0.1507219523191452, | |
| "learning_rate": 9.851478631379982e-05, | |
| "loss": 0.8112, | |
| "step": 68 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_loss": 0.8751078248023987, | |
| "eval_runtime": 13.9264, | |
| "eval_samples_per_second": 2.585, | |
| "eval_steps_per_second": 2.585, | |
| "step": 68 | |
| }, | |
| { | |
| "epoch": 8.235294117647058, | |
| "grad_norm": 0.13311730325222015, | |
| "learning_rate": 9.829629131445342e-05, | |
| "loss": 0.7713, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 8.470588235294118, | |
| "grad_norm": 0.1371852308511734, | |
| "learning_rate": 9.806308479691595e-05, | |
| "loss": 0.7967, | |
| "step": 72 | |
| }, | |
| { | |
| "epoch": 8.705882352941176, | |
| "grad_norm": 0.14112600684165955, | |
| "learning_rate": 9.781523779815179e-05, | |
| "loss": 0.7909, | |
| "step": 74 | |
| }, | |
| { | |
| "epoch": 8.941176470588236, | |
| "grad_norm": 0.16504345834255219, | |
| "learning_rate": 9.755282581475769e-05, | |
| "loss": 0.7491, | |
| "step": 76 | |
| }, | |
| { | |
| "epoch": 8.941176470588236, | |
| "eval_loss": 0.8714024424552917, | |
| "eval_runtime": 13.9243, | |
| "eval_samples_per_second": 2.585, | |
| "eval_steps_per_second": 2.585, | |
| "step": 76 | |
| }, | |
| { | |
| "epoch": 9.176470588235293, | |
| "grad_norm": 0.17292000353336334, | |
| "learning_rate": 9.727592877996585e-05, | |
| "loss": 0.7597, | |
| "step": 78 | |
| }, | |
| { | |
| "epoch": 9.411764705882353, | |
| "grad_norm": 0.1559344381093979, | |
| "learning_rate": 9.698463103929542e-05, | |
| "loss": 0.7566, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 9.647058823529411, | |
| "grad_norm": 0.17069514095783234, | |
| "learning_rate": 9.667902132486009e-05, | |
| "loss": 0.7534, | |
| "step": 82 | |
| }, | |
| { | |
| "epoch": 9.882352941176471, | |
| "grad_norm": 0.1825936883687973, | |
| "learning_rate": 9.635919272833938e-05, | |
| "loss": 0.7439, | |
| "step": 84 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_loss": 0.8710008263587952, | |
| "eval_runtime": 13.9329, | |
| "eval_samples_per_second": 2.584, | |
| "eval_steps_per_second": 2.584, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 10.117647058823529, | |
| "grad_norm": 0.17276407778263092, | |
| "learning_rate": 9.602524267262203e-05, | |
| "loss": 0.7186, | |
| "step": 86 | |
| }, | |
| { | |
| "epoch": 10.352941176470589, | |
| "grad_norm": 0.20144543051719666, | |
| "learning_rate": 9.567727288213005e-05, | |
| "loss": 0.736, | |
| "step": 88 | |
| }, | |
| { | |
| "epoch": 10.588235294117647, | |
| "grad_norm": 0.19740185141563416, | |
| "learning_rate": 9.53153893518325e-05, | |
| "loss": 0.7088, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 10.823529411764707, | |
| "grad_norm": 0.22668859362602234, | |
| "learning_rate": 9.493970231495835e-05, | |
| "loss": 0.7136, | |
| "step": 92 | |
| }, | |
| { | |
| "epoch": 10.941176470588236, | |
| "eval_loss": 0.8754350543022156, | |
| "eval_runtime": 13.9268, | |
| "eval_samples_per_second": 2.585, | |
| "eval_steps_per_second": 2.585, | |
| "step": 93 | |
| }, | |
| { | |
| "epoch": 11.058823529411764, | |
| "grad_norm": 0.20473405718803406, | |
| "learning_rate": 9.45503262094184e-05, | |
| "loss": 0.7447, | |
| "step": 94 | |
| }, | |
| { | |
| "epoch": 11.294117647058824, | |
| "grad_norm": 0.21799054741859436, | |
| "learning_rate": 9.414737964294636e-05, | |
| "loss": 0.7287, | |
| "step": 96 | |
| }, | |
| { | |
| "epoch": 11.529411764705882, | |
| "grad_norm": 0.2471453845500946, | |
| "learning_rate": 9.373098535696979e-05, | |
| "loss": 0.6675, | |
| "step": 98 | |
| }, | |
| { | |
| "epoch": 11.764705882352942, | |
| "grad_norm": 0.2599719762802124, | |
| "learning_rate": 9.330127018922194e-05, | |
| "loss": 0.6934, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "grad_norm": 0.2612287104129791, | |
| "learning_rate": 9.285836503510562e-05, | |
| "loss": 0.6546, | |
| "step": 102 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "eval_loss": 0.8866230845451355, | |
| "eval_runtime": 13.9305, | |
| "eval_samples_per_second": 2.584, | |
| "eval_steps_per_second": 2.584, | |
| "step": 102 | |
| }, | |
| { | |
| "epoch": 12.235294117647058, | |
| "grad_norm": 0.3070856034755707, | |
| "learning_rate": 9.24024048078213e-05, | |
| "loss": 0.6527, | |
| "step": 104 | |
| }, | |
| { | |
| "epoch": 12.470588235294118, | |
| "grad_norm": 0.3168947100639343, | |
| "learning_rate": 9.193352839727121e-05, | |
| "loss": 0.6387, | |
| "step": 106 | |
| }, | |
| { | |
| "epoch": 12.705882352941176, | |
| "grad_norm": 0.3805824816226959, | |
| "learning_rate": 9.145187862775209e-05, | |
| "loss": 0.666, | |
| "step": 108 | |
| }, | |
| { | |
| "epoch": 12.941176470588236, | |
| "grad_norm": 0.42959412932395935, | |
| "learning_rate": 9.09576022144496e-05, | |
| "loss": 0.6539, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 12.941176470588236, | |
| "eval_loss": 0.9093228578567505, | |
| "eval_runtime": 13.9213, | |
| "eval_samples_per_second": 2.586, | |
| "eval_steps_per_second": 2.586, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 13.176470588235293, | |
| "grad_norm": 0.3179619312286377, | |
| "learning_rate": 9.045084971874738e-05, | |
| "loss": 0.6358, | |
| "step": 112 | |
| }, | |
| { | |
| "epoch": 13.411764705882353, | |
| "grad_norm": 0.37940555810928345, | |
| "learning_rate": 8.993177550236464e-05, | |
| "loss": 0.5786, | |
| "step": 114 | |
| }, | |
| { | |
| "epoch": 13.647058823529411, | |
| "grad_norm": 0.3830829858779907, | |
| "learning_rate": 8.940053768033609e-05, | |
| "loss": 0.6127, | |
| "step": 116 | |
| }, | |
| { | |
| "epoch": 13.882352941176471, | |
| "grad_norm": 0.42292138934135437, | |
| "learning_rate": 8.885729807284856e-05, | |
| "loss": 0.632, | |
| "step": 118 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "eval_loss": 0.9347273707389832, | |
| "eval_runtime": 13.9419, | |
| "eval_samples_per_second": 2.582, | |
| "eval_steps_per_second": 2.582, | |
| "step": 119 | |
| }, | |
| { | |
| "epoch": 14.117647058823529, | |
| "grad_norm": 0.41069409251213074, | |
| "learning_rate": 8.83022221559489e-05, | |
| "loss": 0.626, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 14.352941176470589, | |
| "grad_norm": 0.528698742389679, | |
| "learning_rate": 8.773547901113862e-05, | |
| "loss": 0.5267, | |
| "step": 122 | |
| }, | |
| { | |
| "epoch": 14.588235294117647, | |
| "grad_norm": 0.4924652874469757, | |
| "learning_rate": 8.715724127386972e-05, | |
| "loss": 0.5981, | |
| "step": 124 | |
| }, | |
| { | |
| "epoch": 14.823529411764707, | |
| "grad_norm": 0.5208938121795654, | |
| "learning_rate": 8.656768508095853e-05, | |
| "loss": 0.543, | |
| "step": 126 | |
| }, | |
| { | |
| "epoch": 14.941176470588236, | |
| "eval_loss": 0.9885026216506958, | |
| "eval_runtime": 13.9217, | |
| "eval_samples_per_second": 2.586, | |
| "eval_steps_per_second": 2.586, | |
| "step": 127 | |
| }, | |
| { | |
| "epoch": 15.058823529411764, | |
| "grad_norm": 0.5211675763130188, | |
| "learning_rate": 8.596699001693255e-05, | |
| "loss": 0.5916, | |
| "step": 128 | |
| }, | |
| { | |
| "epoch": 15.294117647058824, | |
| "grad_norm": 0.5175803899765015, | |
| "learning_rate": 8.535533905932738e-05, | |
| "loss": 0.5308, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 15.529411764705882, | |
| "grad_norm": 0.55540931224823, | |
| "learning_rate": 8.473291852294987e-05, | |
| "loss": 0.551, | |
| "step": 132 | |
| }, | |
| { | |
| "epoch": 15.764705882352942, | |
| "grad_norm": 0.527597188949585, | |
| "learning_rate": 8.409991800312493e-05, | |
| "loss": 0.5514, | |
| "step": 134 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "grad_norm": 0.5720416307449341, | |
| "learning_rate": 8.345653031794292e-05, | |
| "loss": 0.4869, | |
| "step": 136 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "eval_loss": 1.0261539220809937, | |
| "eval_runtime": 13.9314, | |
| "eval_samples_per_second": 2.584, | |
| "eval_steps_per_second": 2.584, | |
| "step": 136 | |
| }, | |
| { | |
| "epoch": 16.235294117647058, | |
| "grad_norm": 0.7312718033790588, | |
| "learning_rate": 8.280295144952536e-05, | |
| "loss": 0.4512, | |
| "step": 138 | |
| }, | |
| { | |
| "epoch": 16.470588235294116, | |
| "grad_norm": 0.7295913100242615, | |
| "learning_rate": 8.213938048432697e-05, | |
| "loss": 0.5245, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 16.705882352941178, | |
| "grad_norm": 0.7654659152030945, | |
| "learning_rate": 8.146601955249188e-05, | |
| "loss": 0.4661, | |
| "step": 142 | |
| }, | |
| { | |
| "epoch": 16.941176470588236, | |
| "grad_norm": 0.6638887524604797, | |
| "learning_rate": 8.07830737662829e-05, | |
| "loss": 0.4926, | |
| "step": 144 | |
| }, | |
| { | |
| "epoch": 16.941176470588236, | |
| "eval_loss": 1.043527603149414, | |
| "eval_runtime": 13.9302, | |
| "eval_samples_per_second": 2.584, | |
| "eval_steps_per_second": 2.584, | |
| "step": 144 | |
| }, | |
| { | |
| "epoch": 16.941176470588236, | |
| "step": 144, | |
| "total_flos": 7.574414567879475e+16, | |
| "train_loss": 0.8188091479241848, | |
| "train_runtime": 3221.0584, | |
| "train_samples_per_second": 2.111, | |
| "train_steps_per_second": 0.124 | |
| } | |
| ], | |
| "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": 7.574414567879475e+16, | |
| "train_batch_size": 1, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |