| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 9.803921568627452, | |
| "eval_steps": 500, | |
| "global_step": 250, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0392156862745098, | |
| "grad_norm": 536.0, | |
| "learning_rate": 8.000000000000001e-06, | |
| "loss": 51.7984, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.19607843137254902, | |
| "grad_norm": 314.0, | |
| "learning_rate": 4e-05, | |
| "loss": 43.035, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.39215686274509803, | |
| "grad_norm": 45.25, | |
| "learning_rate": 8e-05, | |
| "loss": 28.3835, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.5882352941176471, | |
| "grad_norm": 11.5, | |
| "learning_rate": 0.00012, | |
| "loss": 21.6194, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.7843137254901961, | |
| "grad_norm": 10.0625, | |
| "learning_rate": 0.00016, | |
| "loss": 19.3838, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.9803921568627451, | |
| "grad_norm": 37.75, | |
| "learning_rate": 0.0002, | |
| "loss": 16.725, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.9803921568627451, | |
| "eval_loss": 6.552042484283447, | |
| "eval_runtime": 2.0523, | |
| "eval_samples_per_second": 4.873, | |
| "eval_steps_per_second": 0.975, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 1.1764705882352942, | |
| "grad_norm": 10.875, | |
| "learning_rate": 0.00019975640502598244, | |
| "loss": 7.143, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 1.3725490196078431, | |
| "grad_norm": 7.75, | |
| "learning_rate": 0.00019902680687415705, | |
| "loss": 2.2222, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 1.5686274509803921, | |
| "grad_norm": 2.453125, | |
| "learning_rate": 0.00019781476007338058, | |
| "loss": 1.7799, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 1.7647058823529411, | |
| "grad_norm": 4.34375, | |
| "learning_rate": 0.0001961261695938319, | |
| "loss": 1.5965, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 1.9607843137254903, | |
| "grad_norm": 2.4375, | |
| "learning_rate": 0.00019396926207859084, | |
| "loss": 1.5122, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_loss": 2.6625783443450928, | |
| "eval_runtime": 2.0517, | |
| "eval_samples_per_second": 4.874, | |
| "eval_steps_per_second": 0.975, | |
| "step": 51 | |
| }, | |
| { | |
| "epoch": 2.156862745098039, | |
| "grad_norm": 1.7421875, | |
| "learning_rate": 0.0001913545457642601, | |
| "loss": 1.3478, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 2.3529411764705883, | |
| "grad_norm": 2.046875, | |
| "learning_rate": 0.00018829475928589271, | |
| "loss": 1.26, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 2.549019607843137, | |
| "grad_norm": 3.8125, | |
| "learning_rate": 0.0001848048096156426, | |
| "loss": 1.2315, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 2.7450980392156863, | |
| "grad_norm": 4.46875, | |
| "learning_rate": 0.00018090169943749476, | |
| "loss": 1.1548, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 2.9411764705882355, | |
| "grad_norm": 1.5, | |
| "learning_rate": 0.0001766044443118978, | |
| "loss": 1.1154, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 2.980392156862745, | |
| "eval_loss": 2.591742753982544, | |
| "eval_runtime": 2.0491, | |
| "eval_samples_per_second": 4.88, | |
| "eval_steps_per_second": 0.976, | |
| "step": 76 | |
| }, | |
| { | |
| "epoch": 3.1372549019607843, | |
| "grad_norm": 1.9375, | |
| "learning_rate": 0.0001719339800338651, | |
| "loss": 1.0096, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 3.3333333333333335, | |
| "grad_norm": 3.40625, | |
| "learning_rate": 0.00016691306063588583, | |
| "loss": 0.9379, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 3.5294117647058822, | |
| "grad_norm": 1.65625, | |
| "learning_rate": 0.0001615661475325658, | |
| "loss": 0.9333, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 3.7254901960784315, | |
| "grad_norm": 3.109375, | |
| "learning_rate": 0.0001559192903470747, | |
| "loss": 0.9114, | |
| "step": 95 | |
| }, | |
| { | |
| "epoch": 3.9215686274509802, | |
| "grad_norm": 11.0, | |
| "learning_rate": 0.00015000000000000001, | |
| "loss": 0.9204, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_loss": 2.6570026874542236, | |
| "eval_runtime": 2.0401, | |
| "eval_samples_per_second": 4.902, | |
| "eval_steps_per_second": 0.98, | |
| "step": 102 | |
| }, | |
| { | |
| "epoch": 4.117647058823529, | |
| "grad_norm": 1.4296875, | |
| "learning_rate": 0.00014383711467890774, | |
| "loss": 0.8556, | |
| "step": 105 | |
| }, | |
| { | |
| "epoch": 4.313725490196078, | |
| "grad_norm": 1.3984375, | |
| "learning_rate": 0.00013746065934159123, | |
| "loss": 0.786, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 4.509803921568627, | |
| "grad_norm": 1.0703125, | |
| "learning_rate": 0.00013090169943749476, | |
| "loss": 0.797, | |
| "step": 115 | |
| }, | |
| { | |
| "epoch": 4.705882352941177, | |
| "grad_norm": 1.9609375, | |
| "learning_rate": 0.00012419218955996676, | |
| "loss": 0.7689, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 4.901960784313726, | |
| "grad_norm": 1.4296875, | |
| "learning_rate": 0.00011736481776669306, | |
| "loss": 0.779, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 4.980392156862745, | |
| "eval_loss": 2.749800205230713, | |
| "eval_runtime": 2.0472, | |
| "eval_samples_per_second": 4.885, | |
| "eval_steps_per_second": 0.977, | |
| "step": 127 | |
| }, | |
| { | |
| "epoch": 5.098039215686274, | |
| "grad_norm": 15.25, | |
| "learning_rate": 0.00011045284632676536, | |
| "loss": 0.6889, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 5.294117647058823, | |
| "grad_norm": 1.28125, | |
| "learning_rate": 0.00010348994967025012, | |
| "loss": 0.6232, | |
| "step": 135 | |
| }, | |
| { | |
| "epoch": 5.490196078431373, | |
| "grad_norm": 1.359375, | |
| "learning_rate": 9.651005032974994e-05, | |
| "loss": 0.6134, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 5.686274509803922, | |
| "grad_norm": 1.25, | |
| "learning_rate": 8.954715367323468e-05, | |
| "loss": 0.612, | |
| "step": 145 | |
| }, | |
| { | |
| "epoch": 5.882352941176471, | |
| "grad_norm": 1.0390625, | |
| "learning_rate": 8.263518223330697e-05, | |
| "loss": 0.6207, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_loss": 2.9976024627685547, | |
| "eval_runtime": 2.0445, | |
| "eval_samples_per_second": 4.891, | |
| "eval_steps_per_second": 0.978, | |
| "step": 153 | |
| }, | |
| { | |
| "epoch": 6.078431372549019, | |
| "grad_norm": 1.1328125, | |
| "learning_rate": 7.580781044003324e-05, | |
| "loss": 0.5836, | |
| "step": 155 | |
| }, | |
| { | |
| "epoch": 6.2745098039215685, | |
| "grad_norm": 1.203125, | |
| "learning_rate": 6.909830056250527e-05, | |
| "loss": 0.5047, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 6.470588235294118, | |
| "grad_norm": 1.4765625, | |
| "learning_rate": 6.25393406584088e-05, | |
| "loss": 0.4862, | |
| "step": 165 | |
| }, | |
| { | |
| "epoch": 6.666666666666667, | |
| "grad_norm": 1.21875, | |
| "learning_rate": 5.616288532109225e-05, | |
| "loss": 0.4953, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 6.862745098039216, | |
| "grad_norm": 1.1484375, | |
| "learning_rate": 5.000000000000002e-05, | |
| "loss": 0.4762, | |
| "step": 175 | |
| }, | |
| { | |
| "epoch": 6.980392156862745, | |
| "eval_loss": 3.466813325881958, | |
| "eval_runtime": 2.0565, | |
| "eval_samples_per_second": 4.863, | |
| "eval_steps_per_second": 0.973, | |
| "step": 178 | |
| }, | |
| { | |
| "epoch": 7.0588235294117645, | |
| "grad_norm": 1.203125, | |
| "learning_rate": 4.4080709652925336e-05, | |
| "loss": 0.4669, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 7.254901960784314, | |
| "grad_norm": 1.4140625, | |
| "learning_rate": 3.843385246743417e-05, | |
| "loss": 0.3974, | |
| "step": 185 | |
| }, | |
| { | |
| "epoch": 7.450980392156863, | |
| "grad_norm": 1.0625, | |
| "learning_rate": 3.308693936411421e-05, | |
| "loss": 0.3956, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 7.647058823529412, | |
| "grad_norm": 1.0703125, | |
| "learning_rate": 2.8066019966134904e-05, | |
| "loss": 0.386, | |
| "step": 195 | |
| }, | |
| { | |
| "epoch": 7.8431372549019605, | |
| "grad_norm": 1.1328125, | |
| "learning_rate": 2.339555568810221e-05, | |
| "loss": 0.3908, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_loss": 3.824570417404175, | |
| "eval_runtime": 2.0426, | |
| "eval_samples_per_second": 4.896, | |
| "eval_steps_per_second": 0.979, | |
| "step": 204 | |
| }, | |
| { | |
| "epoch": 8.03921568627451, | |
| "grad_norm": 0.984375, | |
| "learning_rate": 1.9098300562505266e-05, | |
| "loss": 0.3721, | |
| "step": 205 | |
| }, | |
| { | |
| "epoch": 8.235294117647058, | |
| "grad_norm": 0.96875, | |
| "learning_rate": 1.5195190384357404e-05, | |
| "loss": 0.3342, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 8.431372549019608, | |
| "grad_norm": 1.046875, | |
| "learning_rate": 1.1705240714107302e-05, | |
| "loss": 0.346, | |
| "step": 215 | |
| }, | |
| { | |
| "epoch": 8.627450980392156, | |
| "grad_norm": 1.046875, | |
| "learning_rate": 8.645454235739903e-06, | |
| "loss": 0.3381, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 8.823529411764707, | |
| "grad_norm": 1.046875, | |
| "learning_rate": 6.030737921409169e-06, | |
| "loss": 0.3418, | |
| "step": 225 | |
| }, | |
| { | |
| "epoch": 8.980392156862745, | |
| "eval_loss": 4.056135177612305, | |
| "eval_runtime": 2.0561, | |
| "eval_samples_per_second": 4.864, | |
| "eval_steps_per_second": 0.973, | |
| "step": 229 | |
| }, | |
| { | |
| "epoch": 9.019607843137255, | |
| "grad_norm": 0.88671875, | |
| "learning_rate": 3.873830406168111e-06, | |
| "loss": 0.3294, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 9.215686274509803, | |
| "grad_norm": 0.89453125, | |
| "learning_rate": 2.1852399266194314e-06, | |
| "loss": 0.3265, | |
| "step": 235 | |
| }, | |
| { | |
| "epoch": 9.411764705882353, | |
| "grad_norm": 0.9296875, | |
| "learning_rate": 9.731931258429638e-07, | |
| "loss": 0.3268, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 9.607843137254902, | |
| "grad_norm": 0.91796875, | |
| "learning_rate": 2.4359497401758024e-07, | |
| "loss": 0.3257, | |
| "step": 245 | |
| }, | |
| { | |
| "epoch": 9.803921568627452, | |
| "grad_norm": 0.9140625, | |
| "learning_rate": 0.0, | |
| "loss": 0.3252, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 9.803921568627452, | |
| "eval_loss": 4.065803050994873, | |
| "eval_runtime": 2.0429, | |
| "eval_samples_per_second": 4.895, | |
| "eval_steps_per_second": 0.979, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 9.803921568627452, | |
| "step": 250, | |
| "total_flos": 1.9110914639160934e+17, | |
| "train_loss": 3.4213723726272582, | |
| "train_runtime": 1646.3335, | |
| "train_samples_per_second": 2.448, | |
| "train_steps_per_second": 0.152 | |
| } | |
| ], | |
| "logging_steps": 5, | |
| "max_steps": 250, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 10, | |
| "save_steps": 100, | |
| "total_flos": 1.9110914639160934e+17, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |