| { | |
| "best_metric": 0.47586047649383545, | |
| "best_model_checkpoint": "../../experiments_checkpoints/MAdAiLab/FacebookAI/roberta_base_twitter/checkpoint-50", | |
| "epoch": 3.0, | |
| "eval_steps": 50, | |
| "global_step": 408, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.07, | |
| "grad_norm": 0.8274936676025391, | |
| "learning_rate": 1.950980392156863e-05, | |
| "loss": 0.6101, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "grad_norm": 3.8833539485931396, | |
| "learning_rate": 1.9019607843137255e-05, | |
| "loss": 0.6154, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.22, | |
| "grad_norm": 6.267473220825195, | |
| "learning_rate": 1.8529411764705884e-05, | |
| "loss": 0.4968, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "grad_norm": 6.414277076721191, | |
| "learning_rate": 1.8039215686274513e-05, | |
| "loss": 0.5094, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.37, | |
| "grad_norm": 3.6427950859069824, | |
| "learning_rate": 1.7549019607843138e-05, | |
| "loss": 0.4867, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.37, | |
| "eval_accuracy": 0.7711397058823529, | |
| "eval_f1_macro": 0.7371988324345717, | |
| "eval_f1_micro": 0.7711397058823529, | |
| "eval_loss": 0.47586047649383545, | |
| "eval_runtime": 1.0164, | |
| "eval_samples_per_second": 1070.408, | |
| "eval_steps_per_second": 16.725, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.44, | |
| "grad_norm": 3.9136903285980225, | |
| "learning_rate": 1.7058823529411767e-05, | |
| "loss": 0.4718, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.51, | |
| "grad_norm": 5.659265995025635, | |
| "learning_rate": 1.6568627450980395e-05, | |
| "loss": 0.5029, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "grad_norm": 4.554523468017578, | |
| "learning_rate": 1.607843137254902e-05, | |
| "loss": 0.4713, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.66, | |
| "grad_norm": 7.2100324630737305, | |
| "learning_rate": 1.558823529411765e-05, | |
| "loss": 0.5155, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.74, | |
| "grad_norm": 3.8660807609558105, | |
| "learning_rate": 1.5098039215686276e-05, | |
| "loss": 0.4633, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.74, | |
| "eval_accuracy": 0.7711397058823529, | |
| "eval_f1_macro": 0.7285404661201644, | |
| "eval_f1_micro": 0.7711397058823529, | |
| "eval_loss": 0.4787723422050476, | |
| "eval_runtime": 1.0179, | |
| "eval_samples_per_second": 1068.897, | |
| "eval_steps_per_second": 16.702, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.81, | |
| "grad_norm": 5.878931999206543, | |
| "learning_rate": 1.4607843137254903e-05, | |
| "loss": 0.4712, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "grad_norm": 3.9491829872131348, | |
| "learning_rate": 1.4117647058823532e-05, | |
| "loss": 0.4386, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "grad_norm": 5.058815002441406, | |
| "learning_rate": 1.3627450980392158e-05, | |
| "loss": 0.4621, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 1.03, | |
| "grad_norm": 3.4084646701812744, | |
| "learning_rate": 1.3137254901960785e-05, | |
| "loss": 0.4788, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 1.1, | |
| "grad_norm": 3.3157005310058594, | |
| "learning_rate": 1.2647058823529412e-05, | |
| "loss": 0.4582, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 1.1, | |
| "eval_accuracy": 0.7738970588235294, | |
| "eval_f1_macro": 0.7356166061472933, | |
| "eval_f1_micro": 0.7738970588235294, | |
| "eval_loss": 0.48208916187286377, | |
| "eval_runtime": 1.0226, | |
| "eval_samples_per_second": 1063.923, | |
| "eval_steps_per_second": 16.624, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 1.18, | |
| "grad_norm": 3.4631340503692627, | |
| "learning_rate": 1.215686274509804e-05, | |
| "loss": 0.4219, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 1.25, | |
| "grad_norm": 3.0762877464294434, | |
| "learning_rate": 1.1666666666666668e-05, | |
| "loss": 0.4258, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 1.32, | |
| "grad_norm": 3.2170896530151367, | |
| "learning_rate": 1.1176470588235295e-05, | |
| "loss": 0.4192, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 1.4, | |
| "grad_norm": 3.9244940280914307, | |
| "learning_rate": 1.0686274509803922e-05, | |
| "loss": 0.4207, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 1.47, | |
| "grad_norm": 3.8159372806549072, | |
| "learning_rate": 1.0196078431372549e-05, | |
| "loss": 0.4642, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 1.47, | |
| "eval_accuracy": 0.7591911764705882, | |
| "eval_f1_macro": 0.7292144010639309, | |
| "eval_f1_micro": 0.7591911764705882, | |
| "eval_loss": 0.4841248691082001, | |
| "eval_runtime": 1.0248, | |
| "eval_samples_per_second": 1061.624, | |
| "eval_steps_per_second": 16.588, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 1.54, | |
| "grad_norm": 4.037954330444336, | |
| "learning_rate": 9.705882352941177e-06, | |
| "loss": 0.4269, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 1.62, | |
| "grad_norm": 5.522130012512207, | |
| "learning_rate": 9.215686274509804e-06, | |
| "loss": 0.4788, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 1.69, | |
| "grad_norm": 3.8915343284606934, | |
| "learning_rate": 8.725490196078433e-06, | |
| "loss": 0.3953, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 1.76, | |
| "grad_norm": 3.3338236808776855, | |
| "learning_rate": 8.23529411764706e-06, | |
| "loss": 0.422, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 1.84, | |
| "grad_norm": 3.794468402862549, | |
| "learning_rate": 7.745098039215687e-06, | |
| "loss": 0.458, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 1.84, | |
| "eval_accuracy": 0.7738970588235294, | |
| "eval_f1_macro": 0.7368983957219251, | |
| "eval_f1_micro": 0.7738970588235294, | |
| "eval_loss": 0.48644232749938965, | |
| "eval_runtime": 1.0281, | |
| "eval_samples_per_second": 1058.283, | |
| "eval_steps_per_second": 16.536, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 1.91, | |
| "grad_norm": 4.221856117248535, | |
| "learning_rate": 7.2549019607843145e-06, | |
| "loss": 0.4065, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 1.99, | |
| "grad_norm": 3.7018020153045654, | |
| "learning_rate": 6.764705882352942e-06, | |
| "loss": 0.3799, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 2.06, | |
| "grad_norm": 3.910475730895996, | |
| "learning_rate": 6.274509803921569e-06, | |
| "loss": 0.4244, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 2.13, | |
| "grad_norm": 6.957982063293457, | |
| "learning_rate": 5.784313725490197e-06, | |
| "loss": 0.4203, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 2.21, | |
| "grad_norm": 3.897538423538208, | |
| "learning_rate": 5.294117647058824e-06, | |
| "loss": 0.4001, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 2.21, | |
| "eval_accuracy": 0.7683823529411765, | |
| "eval_f1_macro": 0.7346341463414634, | |
| "eval_f1_micro": 0.7683823529411765, | |
| "eval_loss": 0.4866744875907898, | |
| "eval_runtime": 1.0289, | |
| "eval_samples_per_second": 1057.414, | |
| "eval_steps_per_second": 16.522, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 2.28, | |
| "grad_norm": 7.027072906494141, | |
| "learning_rate": 4.803921568627452e-06, | |
| "loss": 0.3947, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 2.35, | |
| "grad_norm": 5.167274475097656, | |
| "learning_rate": 4.313725490196079e-06, | |
| "loss": 0.3697, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 2.43, | |
| "grad_norm": 4.295684337615967, | |
| "learning_rate": 3.8235294117647055e-06, | |
| "loss": 0.4043, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 2.5, | |
| "grad_norm": 3.2442514896392822, | |
| "learning_rate": 3.3333333333333333e-06, | |
| "loss": 0.3848, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 2.57, | |
| "grad_norm": 4.748172760009766, | |
| "learning_rate": 2.843137254901961e-06, | |
| "loss": 0.443, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 2.57, | |
| "eval_accuracy": 0.7601102941176471, | |
| "eval_f1_macro": 0.7257722938932072, | |
| "eval_f1_micro": 0.7601102941176471, | |
| "eval_loss": 0.4885733127593994, | |
| "eval_runtime": 1.0305, | |
| "eval_samples_per_second": 1055.812, | |
| "eval_steps_per_second": 16.497, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 2.65, | |
| "grad_norm": 3.841963529586792, | |
| "learning_rate": 2.3529411764705885e-06, | |
| "loss": 0.4069, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 2.72, | |
| "grad_norm": 4.17108154296875, | |
| "learning_rate": 1.8627450980392158e-06, | |
| "loss": 0.3803, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 2.79, | |
| "grad_norm": 3.6617419719696045, | |
| "learning_rate": 1.3725490196078434e-06, | |
| "loss": 0.3775, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 2.87, | |
| "grad_norm": 4.859780788421631, | |
| "learning_rate": 8.823529411764707e-07, | |
| "loss": 0.4131, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 2.94, | |
| "grad_norm": 4.684369087219238, | |
| "learning_rate": 3.921568627450981e-07, | |
| "loss": 0.3461, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 2.94, | |
| "eval_accuracy": 0.765625, | |
| "eval_f1_macro": 0.7296262187723838, | |
| "eval_f1_micro": 0.765625, | |
| "eval_loss": 0.4941970109939575, | |
| "eval_runtime": 1.0298, | |
| "eval_samples_per_second": 1056.475, | |
| "eval_steps_per_second": 16.507, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "step": 408, | |
| "total_flos": 1717588929282048.0, | |
| "train_loss": 0.44208094419217575, | |
| "train_runtime": 97.8515, | |
| "train_samples_per_second": 266.731, | |
| "train_steps_per_second": 4.17 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 408, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "save_steps": 50, | |
| "total_flos": 1717588929282048.0, | |
| "train_batch_size": 32, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |