| { | |
| "best_metric": 0.7848126292228699, | |
| "best_model_checkpoint": "data/Llama-31-8B_task-1_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": 1.65293550491333, | |
| "learning_rate": 2.5e-06, | |
| "loss": 2.1767, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.23529411764705882, | |
| "grad_norm": 1.7422088384628296, | |
| "learning_rate": 5e-06, | |
| "loss": 2.1217, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.47058823529411764, | |
| "grad_norm": 1.5451767444610596, | |
| "learning_rate": 1e-05, | |
| "loss": 2.1526, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.7058823529411765, | |
| "grad_norm": 1.4295381307601929, | |
| "learning_rate": 1.5e-05, | |
| "loss": 2.1431, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.9411764705882353, | |
| "grad_norm": 1.4298073053359985, | |
| "learning_rate": 2e-05, | |
| "loss": 2.1079, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.9411764705882353, | |
| "eval_loss": 2.043940544128418, | |
| "eval_runtime": 14.8539, | |
| "eval_samples_per_second": 2.424, | |
| "eval_steps_per_second": 2.424, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 1.1764705882352942, | |
| "grad_norm": 1.2632858753204346, | |
| "learning_rate": 2.5e-05, | |
| "loss": 2.0029, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 1.4117647058823528, | |
| "grad_norm": 1.0134398937225342, | |
| "learning_rate": 3e-05, | |
| "loss": 1.9339, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 1.6470588235294117, | |
| "grad_norm": 0.9206991791725159, | |
| "learning_rate": 3.5e-05, | |
| "loss": 1.8533, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 1.8823529411764706, | |
| "grad_norm": 1.641518473625183, | |
| "learning_rate": 4e-05, | |
| "loss": 1.7735, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_loss": 1.6643422842025757, | |
| "eval_runtime": 14.8482, | |
| "eval_samples_per_second": 2.425, | |
| "eval_steps_per_second": 2.425, | |
| "step": 17 | |
| }, | |
| { | |
| "epoch": 2.1176470588235294, | |
| "grad_norm": 0.6680443286895752, | |
| "learning_rate": 4.5e-05, | |
| "loss": 1.6775, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 2.3529411764705883, | |
| "grad_norm": 0.679865300655365, | |
| "learning_rate": 5e-05, | |
| "loss": 1.5899, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 2.588235294117647, | |
| "grad_norm": 0.8137246966362, | |
| "learning_rate": 5.500000000000001e-05, | |
| "loss": 1.4771, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 2.8235294117647056, | |
| "grad_norm": 0.8322336673736572, | |
| "learning_rate": 6e-05, | |
| "loss": 1.4106, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 2.9411764705882355, | |
| "eval_loss": 1.232216238975525, | |
| "eval_runtime": 14.8645, | |
| "eval_samples_per_second": 2.422, | |
| "eval_steps_per_second": 2.422, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 3.0588235294117645, | |
| "grad_norm": 0.9232141375541687, | |
| "learning_rate": 6.500000000000001e-05, | |
| "loss": 1.2258, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 3.2941176470588234, | |
| "grad_norm": 0.82151198387146, | |
| "learning_rate": 7e-05, | |
| "loss": 1.1418, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 3.5294117647058822, | |
| "grad_norm": 0.5457940697669983, | |
| "learning_rate": 7.500000000000001e-05, | |
| "loss": 1.0264, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 3.764705882352941, | |
| "grad_norm": 0.4534192979335785, | |
| "learning_rate": 8e-05, | |
| "loss": 0.9651, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.41437003016471863, | |
| "learning_rate": 8.5e-05, | |
| "loss": 0.901, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_loss": 0.9258820414543152, | |
| "eval_runtime": 14.8541, | |
| "eval_samples_per_second": 2.424, | |
| "eval_steps_per_second": 2.424, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 4.235294117647059, | |
| "grad_norm": 0.345889687538147, | |
| "learning_rate": 9e-05, | |
| "loss": 0.8852, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 4.470588235294118, | |
| "grad_norm": 0.33972397446632385, | |
| "learning_rate": 9.5e-05, | |
| "loss": 0.8785, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 4.705882352941177, | |
| "grad_norm": 0.3219912350177765, | |
| "learning_rate": 0.0001, | |
| "loss": 0.8862, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 4.9411764705882355, | |
| "grad_norm": 0.3695300817489624, | |
| "learning_rate": 9.999238475781957e-05, | |
| "loss": 0.8657, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 4.9411764705882355, | |
| "eval_loss": 0.8770194053649902, | |
| "eval_runtime": 14.8502, | |
| "eval_samples_per_second": 2.424, | |
| "eval_steps_per_second": 2.424, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 5.176470588235294, | |
| "grad_norm": 0.317543625831604, | |
| "learning_rate": 9.99695413509548e-05, | |
| "loss": 0.843, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 5.411764705882353, | |
| "grad_norm": 0.3217869997024536, | |
| "learning_rate": 9.99314767377287e-05, | |
| "loss": 0.8024, | |
| "step": 46 | |
| }, | |
| { | |
| "epoch": 5.647058823529412, | |
| "grad_norm": 0.3718527555465698, | |
| "learning_rate": 9.987820251299122e-05, | |
| "loss": 0.7786, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 5.882352941176471, | |
| "grad_norm": 0.41980597376823425, | |
| "learning_rate": 9.980973490458728e-05, | |
| "loss": 0.846, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_loss": 0.8422229290008545, | |
| "eval_runtime": 14.848, | |
| "eval_samples_per_second": 2.425, | |
| "eval_steps_per_second": 2.425, | |
| "step": 51 | |
| }, | |
| { | |
| "epoch": 6.117647058823529, | |
| "grad_norm": 0.4026230573654175, | |
| "learning_rate": 9.972609476841367e-05, | |
| "loss": 0.7872, | |
| "step": 52 | |
| }, | |
| { | |
| "epoch": 6.352941176470588, | |
| "grad_norm": 0.41930103302001953, | |
| "learning_rate": 9.962730758206611e-05, | |
| "loss": 0.7829, | |
| "step": 54 | |
| }, | |
| { | |
| "epoch": 6.588235294117647, | |
| "grad_norm": 0.47130709886550903, | |
| "learning_rate": 9.951340343707852e-05, | |
| "loss": 0.7723, | |
| "step": 56 | |
| }, | |
| { | |
| "epoch": 6.823529411764706, | |
| "grad_norm": 0.5386373400688171, | |
| "learning_rate": 9.938441702975689e-05, | |
| "loss": 0.7582, | |
| "step": 58 | |
| }, | |
| { | |
| "epoch": 6.9411764705882355, | |
| "eval_loss": 0.8152710795402527, | |
| "eval_runtime": 14.8487, | |
| "eval_samples_per_second": 2.424, | |
| "eval_steps_per_second": 2.424, | |
| "step": 59 | |
| }, | |
| { | |
| "epoch": 7.0588235294117645, | |
| "grad_norm": 0.5676649212837219, | |
| "learning_rate": 9.924038765061042e-05, | |
| "loss": 0.7721, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 7.294117647058823, | |
| "grad_norm": 0.6037997007369995, | |
| "learning_rate": 9.908135917238321e-05, | |
| "loss": 0.7302, | |
| "step": 62 | |
| }, | |
| { | |
| "epoch": 7.529411764705882, | |
| "grad_norm": 0.4915821850299835, | |
| "learning_rate": 9.890738003669029e-05, | |
| "loss": 0.687, | |
| "step": 64 | |
| }, | |
| { | |
| "epoch": 7.764705882352941, | |
| "grad_norm": 0.36889201402664185, | |
| "learning_rate": 9.871850323926177e-05, | |
| "loss": 0.6989, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "grad_norm": 0.3445635735988617, | |
| "learning_rate": 9.851478631379982e-05, | |
| "loss": 0.7203, | |
| "step": 68 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_loss": 0.7905827760696411, | |
| "eval_runtime": 14.8822, | |
| "eval_samples_per_second": 2.419, | |
| "eval_steps_per_second": 2.419, | |
| "step": 68 | |
| }, | |
| { | |
| "epoch": 8.235294117647058, | |
| "grad_norm": 0.3439154326915741, | |
| "learning_rate": 9.829629131445342e-05, | |
| "loss": 0.6672, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 8.470588235294118, | |
| "grad_norm": 0.39092251658439636, | |
| "learning_rate": 9.806308479691595e-05, | |
| "loss": 0.6864, | |
| "step": 72 | |
| }, | |
| { | |
| "epoch": 8.705882352941176, | |
| "grad_norm": 0.4097364842891693, | |
| "learning_rate": 9.781523779815179e-05, | |
| "loss": 0.6149, | |
| "step": 74 | |
| }, | |
| { | |
| "epoch": 8.941176470588236, | |
| "grad_norm": 0.5072527527809143, | |
| "learning_rate": 9.755282581475769e-05, | |
| "loss": 0.6598, | |
| "step": 76 | |
| }, | |
| { | |
| "epoch": 8.941176470588236, | |
| "eval_loss": 0.7848126292228699, | |
| "eval_runtime": 14.8542, | |
| "eval_samples_per_second": 2.424, | |
| "eval_steps_per_second": 2.424, | |
| "step": 76 | |
| }, | |
| { | |
| "epoch": 9.176470588235293, | |
| "grad_norm": 0.39041876792907715, | |
| "learning_rate": 9.727592877996585e-05, | |
| "loss": 0.6054, | |
| "step": 78 | |
| }, | |
| { | |
| "epoch": 9.411764705882353, | |
| "grad_norm": 0.4171600341796875, | |
| "learning_rate": 9.698463103929542e-05, | |
| "loss": 0.5743, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 9.647058823529411, | |
| "grad_norm": 0.4799637496471405, | |
| "learning_rate": 9.667902132486009e-05, | |
| "loss": 0.5992, | |
| "step": 82 | |
| }, | |
| { | |
| "epoch": 9.882352941176471, | |
| "grad_norm": 0.5218389630317688, | |
| "learning_rate": 9.635919272833938e-05, | |
| "loss": 0.5979, | |
| "step": 84 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_loss": 0.7953908443450928, | |
| "eval_runtime": 14.8439, | |
| "eval_samples_per_second": 2.425, | |
| "eval_steps_per_second": 2.425, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 10.117647058823529, | |
| "grad_norm": 0.5198573470115662, | |
| "learning_rate": 9.602524267262203e-05, | |
| "loss": 0.5292, | |
| "step": 86 | |
| }, | |
| { | |
| "epoch": 10.352941176470589, | |
| "grad_norm": 0.5616230964660645, | |
| "learning_rate": 9.567727288213005e-05, | |
| "loss": 0.5099, | |
| "step": 88 | |
| }, | |
| { | |
| "epoch": 10.588235294117647, | |
| "grad_norm": 0.5473162531852722, | |
| "learning_rate": 9.53153893518325e-05, | |
| "loss": 0.5352, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 10.823529411764707, | |
| "grad_norm": 0.5906912684440613, | |
| "learning_rate": 9.493970231495835e-05, | |
| "loss": 0.5547, | |
| "step": 92 | |
| }, | |
| { | |
| "epoch": 10.941176470588236, | |
| "eval_loss": 0.80945885181427, | |
| "eval_runtime": 14.8718, | |
| "eval_samples_per_second": 2.421, | |
| "eval_steps_per_second": 2.421, | |
| "step": 93 | |
| }, | |
| { | |
| "epoch": 11.058823529411764, | |
| "grad_norm": 0.5553572177886963, | |
| "learning_rate": 9.45503262094184e-05, | |
| "loss": 0.5297, | |
| "step": 94 | |
| }, | |
| { | |
| "epoch": 11.294117647058824, | |
| "grad_norm": 0.5985352993011475, | |
| "learning_rate": 9.414737964294636e-05, | |
| "loss": 0.4793, | |
| "step": 96 | |
| }, | |
| { | |
| "epoch": 11.529411764705882, | |
| "grad_norm": 0.6815150380134583, | |
| "learning_rate": 9.373098535696979e-05, | |
| "loss": 0.464, | |
| "step": 98 | |
| }, | |
| { | |
| "epoch": 11.764705882352942, | |
| "grad_norm": 0.784904956817627, | |
| "learning_rate": 9.330127018922194e-05, | |
| "loss": 0.4358, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "grad_norm": 0.7796445488929749, | |
| "learning_rate": 9.285836503510562e-05, | |
| "loss": 0.4007, | |
| "step": 102 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "eval_loss": 0.8623189926147461, | |
| "eval_runtime": 14.8555, | |
| "eval_samples_per_second": 2.423, | |
| "eval_steps_per_second": 2.423, | |
| "step": 102 | |
| }, | |
| { | |
| "epoch": 12.235294117647058, | |
| "grad_norm": 0.8457428216934204, | |
| "learning_rate": 9.24024048078213e-05, | |
| "loss": 0.4048, | |
| "step": 104 | |
| }, | |
| { | |
| "epoch": 12.470588235294118, | |
| "grad_norm": 0.9290648102760315, | |
| "learning_rate": 9.193352839727121e-05, | |
| "loss": 0.3823, | |
| "step": 106 | |
| }, | |
| { | |
| "epoch": 12.705882352941176, | |
| "grad_norm": 0.9386951923370361, | |
| "learning_rate": 9.145187862775209e-05, | |
| "loss": 0.3356, | |
| "step": 108 | |
| }, | |
| { | |
| "epoch": 12.941176470588236, | |
| "grad_norm": 0.966197669506073, | |
| "learning_rate": 9.09576022144496e-05, | |
| "loss": 0.3489, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 12.941176470588236, | |
| "eval_loss": 0.9626535773277283, | |
| "eval_runtime": 14.855, | |
| "eval_samples_per_second": 2.423, | |
| "eval_steps_per_second": 2.423, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 13.176470588235293, | |
| "grad_norm": 0.9536658525466919, | |
| "learning_rate": 9.045084971874738e-05, | |
| "loss": 0.3107, | |
| "step": 112 | |
| }, | |
| { | |
| "epoch": 13.411764705882353, | |
| "grad_norm": 1.2646312713623047, | |
| "learning_rate": 8.993177550236464e-05, | |
| "loss": 0.2621, | |
| "step": 114 | |
| }, | |
| { | |
| "epoch": 13.647058823529411, | |
| "grad_norm": 0.9954571723937988, | |
| "learning_rate": 8.940053768033609e-05, | |
| "loss": 0.2778, | |
| "step": 116 | |
| }, | |
| { | |
| "epoch": 13.882352941176471, | |
| "grad_norm": 0.8930367231369019, | |
| "learning_rate": 8.885729807284856e-05, | |
| "loss": 0.2894, | |
| "step": 118 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "eval_loss": 1.0530530214309692, | |
| "eval_runtime": 14.8495, | |
| "eval_samples_per_second": 2.424, | |
| "eval_steps_per_second": 2.424, | |
| "step": 119 | |
| }, | |
| { | |
| "epoch": 14.117647058823529, | |
| "grad_norm": 0.8195274472236633, | |
| "learning_rate": 8.83022221559489e-05, | |
| "loss": 0.2601, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 14.352941176470589, | |
| "grad_norm": 1.1883975267410278, | |
| "learning_rate": 8.773547901113862e-05, | |
| "loss": 0.2052, | |
| "step": 122 | |
| }, | |
| { | |
| "epoch": 14.588235294117647, | |
| "grad_norm": 1.118826150894165, | |
| "learning_rate": 8.715724127386972e-05, | |
| "loss": 0.2123, | |
| "step": 124 | |
| }, | |
| { | |
| "epoch": 14.823529411764707, | |
| "grad_norm": 0.9769270420074463, | |
| "learning_rate": 8.656768508095853e-05, | |
| "loss": 0.1972, | |
| "step": 126 | |
| }, | |
| { | |
| "epoch": 14.941176470588236, | |
| "eval_loss": 1.121673345565796, | |
| "eval_runtime": 14.8573, | |
| "eval_samples_per_second": 2.423, | |
| "eval_steps_per_second": 2.423, | |
| "step": 127 | |
| }, | |
| { | |
| "epoch": 15.058823529411764, | |
| "grad_norm": 1.0091590881347656, | |
| "learning_rate": 8.596699001693255e-05, | |
| "loss": 0.1966, | |
| "step": 128 | |
| }, | |
| { | |
| "epoch": 15.294117647058824, | |
| "grad_norm": 1.139027714729309, | |
| "learning_rate": 8.535533905932738e-05, | |
| "loss": 0.1565, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 15.529411764705882, | |
| "grad_norm": 1.2662736177444458, | |
| "learning_rate": 8.473291852294987e-05, | |
| "loss": 0.156, | |
| "step": 132 | |
| }, | |
| { | |
| "epoch": 15.764705882352942, | |
| "grad_norm": 0.8412072658538818, | |
| "learning_rate": 8.409991800312493e-05, | |
| "loss": 0.1451, | |
| "step": 134 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "grad_norm": 1.1424492597579956, | |
| "learning_rate": 8.345653031794292e-05, | |
| "loss": 0.1682, | |
| "step": 136 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "eval_loss": 1.2315659523010254, | |
| "eval_runtime": 14.8492, | |
| "eval_samples_per_second": 2.424, | |
| "eval_steps_per_second": 2.424, | |
| "step": 136 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "step": 136, | |
| "total_flos": 3.69233824275497e+16, | |
| "train_loss": 0.80556961368112, | |
| "train_runtime": 2622.3315, | |
| "train_samples_per_second": 2.593, | |
| "train_steps_per_second": 0.153 | |
| } | |
| ], | |
| "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": 3.69233824275497e+16, | |
| "train_batch_size": 1, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |