| { | |
| "best_metric": 0.8634595701125896, | |
| "best_model_checkpoint": "./outputs/finetuning/mnli_AppE/checkpoint-26000", | |
| "epoch": 5.0, | |
| "global_step": 60000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.17, | |
| "learning_rate": 1.9333333333333333e-05, | |
| "loss": 0.3562, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.17, | |
| "eval_acc": 0.8524053224155578, | |
| "eval_loss": 0.41664522886276245, | |
| "eval_runtime": 16.7668, | |
| "eval_samples_per_second": 582.7, | |
| "eval_steps_per_second": 18.25, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "learning_rate": 1.866666666666667e-05, | |
| "loss": 0.3414, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "eval_acc": 0.8475946775844422, | |
| "eval_loss": 0.43986669182777405, | |
| "eval_runtime": 16.7213, | |
| "eval_samples_per_second": 584.285, | |
| "eval_steps_per_second": 18.3, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "learning_rate": 1.8e-05, | |
| "loss": 0.3308, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "eval_acc": 0.8560900716479017, | |
| "eval_loss": 0.395113080739975, | |
| "eval_runtime": 16.7306, | |
| "eval_samples_per_second": 583.96, | |
| "eval_steps_per_second": 18.29, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "learning_rate": 1.7333333333333336e-05, | |
| "loss": 0.3306, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "eval_acc": 0.8488229273285568, | |
| "eval_loss": 0.41524460911750793, | |
| "eval_runtime": 16.7003, | |
| "eval_samples_per_second": 585.021, | |
| "eval_steps_per_second": 18.323, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 0.83, | |
| "learning_rate": 1.6666666666666667e-05, | |
| "loss": 0.3295, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 0.83, | |
| "eval_acc": 0.8575230296827021, | |
| "eval_loss": 0.39166104793548584, | |
| "eval_runtime": 16.6831, | |
| "eval_samples_per_second": 585.622, | |
| "eval_steps_per_second": 18.342, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "learning_rate": 1.6000000000000003e-05, | |
| "loss": 0.3288, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_acc": 0.8581371545547595, | |
| "eval_loss": 0.40951114892959595, | |
| "eval_runtime": 16.6934, | |
| "eval_samples_per_second": 585.263, | |
| "eval_steps_per_second": 18.331, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 1.17, | |
| "learning_rate": 1.5333333333333334e-05, | |
| "loss": 0.229, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 1.17, | |
| "eval_acc": 0.8611054247697032, | |
| "eval_loss": 0.4420296251773834, | |
| "eval_runtime": 16.7189, | |
| "eval_samples_per_second": 584.367, | |
| "eval_steps_per_second": 18.303, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 1.33, | |
| "learning_rate": 1.4666666666666666e-05, | |
| "loss": 0.2307, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 1.33, | |
| "eval_acc": 0.8560900716479017, | |
| "eval_loss": 0.4876081645488739, | |
| "eval_runtime": 16.6966, | |
| "eval_samples_per_second": 585.148, | |
| "eval_steps_per_second": 18.327, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 1.5, | |
| "learning_rate": 1.4e-05, | |
| "loss": 0.2334, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 1.5, | |
| "eval_acc": 0.857727737973388, | |
| "eval_loss": 0.449561208486557, | |
| "eval_runtime": 16.7074, | |
| "eval_samples_per_second": 584.771, | |
| "eval_steps_per_second": 18.315, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 1.67, | |
| "learning_rate": 1.3333333333333333e-05, | |
| "loss": 0.2392, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 1.67, | |
| "eval_acc": 0.8607983623336745, | |
| "eval_loss": 0.4188033640384674, | |
| "eval_runtime": 16.694, | |
| "eval_samples_per_second": 585.239, | |
| "eval_steps_per_second": 18.33, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 1.83, | |
| "learning_rate": 1.2666666666666667e-05, | |
| "loss": 0.233, | |
| "step": 22000 | |
| }, | |
| { | |
| "epoch": 1.83, | |
| "eval_acc": 0.8578300921187308, | |
| "eval_loss": 0.44934678077697754, | |
| "eval_runtime": 16.6959, | |
| "eval_samples_per_second": 585.174, | |
| "eval_steps_per_second": 18.328, | |
| "step": 22000 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "learning_rate": 1.2e-05, | |
| "loss": 0.2343, | |
| "step": 24000 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_acc": 0.8602865916069601, | |
| "eval_loss": 0.42783743143081665, | |
| "eval_runtime": 16.7147, | |
| "eval_samples_per_second": 584.517, | |
| "eval_steps_per_second": 18.307, | |
| "step": 24000 | |
| }, | |
| { | |
| "epoch": 2.17, | |
| "learning_rate": 1.1333333333333334e-05, | |
| "loss": 0.163, | |
| "step": 26000 | |
| }, | |
| { | |
| "epoch": 2.17, | |
| "eval_acc": 0.8634595701125896, | |
| "eval_loss": 0.5700035095214844, | |
| "eval_runtime": 16.7469, | |
| "eval_samples_per_second": 583.393, | |
| "eval_steps_per_second": 18.272, | |
| "step": 26000 | |
| }, | |
| { | |
| "epoch": 2.33, | |
| "learning_rate": 1.0666666666666667e-05, | |
| "loss": 0.1657, | |
| "step": 28000 | |
| }, | |
| { | |
| "epoch": 2.33, | |
| "eval_acc": 0.8560900716479017, | |
| "eval_loss": 0.5322555303573608, | |
| "eval_runtime": 16.7051, | |
| "eval_samples_per_second": 584.851, | |
| "eval_steps_per_second": 18.318, | |
| "step": 28000 | |
| }, | |
| { | |
| "epoch": 2.5, | |
| "learning_rate": 1e-05, | |
| "loss": 0.1652, | |
| "step": 30000 | |
| }, | |
| { | |
| "epoch": 2.5, | |
| "eval_acc": 0.8588536335721597, | |
| "eval_loss": 0.5046815872192383, | |
| "eval_runtime": 16.7454, | |
| "eval_samples_per_second": 583.445, | |
| "eval_steps_per_second": 18.274, | |
| "step": 30000 | |
| }, | |
| { | |
| "epoch": 2.67, | |
| "learning_rate": 9.333333333333334e-06, | |
| "loss": 0.1654, | |
| "step": 32000 | |
| }, | |
| { | |
| "epoch": 2.67, | |
| "eval_acc": 0.8589559877175026, | |
| "eval_loss": 0.5081538558006287, | |
| "eval_runtime": 16.6963, | |
| "eval_samples_per_second": 585.158, | |
| "eval_steps_per_second": 18.327, | |
| "step": 32000 | |
| }, | |
| { | |
| "epoch": 2.83, | |
| "learning_rate": 8.666666666666668e-06, | |
| "loss": 0.1659, | |
| "step": 34000 | |
| }, | |
| { | |
| "epoch": 2.83, | |
| "eval_acc": 0.8630501535312181, | |
| "eval_loss": 0.5075551271438599, | |
| "eval_runtime": 16.7114, | |
| "eval_samples_per_second": 584.632, | |
| "eval_steps_per_second": 18.311, | |
| "step": 34000 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "learning_rate": 8.000000000000001e-06, | |
| "loss": 0.1654, | |
| "step": 36000 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_acc": 0.8612077789150461, | |
| "eval_loss": 0.5035998821258545, | |
| "eval_runtime": 16.7468, | |
| "eval_samples_per_second": 583.394, | |
| "eval_steps_per_second": 18.272, | |
| "step": 36000 | |
| }, | |
| { | |
| "epoch": 3.17, | |
| "learning_rate": 7.333333333333333e-06, | |
| "loss": 0.1195, | |
| "step": 38000 | |
| }, | |
| { | |
| "epoch": 3.17, | |
| "eval_acc": 0.8599795291709315, | |
| "eval_loss": 0.6220871806144714, | |
| "eval_runtime": 16.6963, | |
| "eval_samples_per_second": 585.161, | |
| "eval_steps_per_second": 18.327, | |
| "step": 38000 | |
| }, | |
| { | |
| "epoch": 3.33, | |
| "learning_rate": 6.666666666666667e-06, | |
| "loss": 0.1194, | |
| "step": 40000 | |
| }, | |
| { | |
| "epoch": 3.33, | |
| "eval_acc": 0.8575230296827021, | |
| "eval_loss": 0.6466721892356873, | |
| "eval_runtime": 16.6726, | |
| "eval_samples_per_second": 585.991, | |
| "eval_steps_per_second": 18.353, | |
| "step": 40000 | |
| }, | |
| { | |
| "epoch": 3.5, | |
| "learning_rate": 6e-06, | |
| "loss": 0.1205, | |
| "step": 42000 | |
| }, | |
| { | |
| "epoch": 3.5, | |
| "eval_acc": 0.8578300921187308, | |
| "eval_loss": 0.6742368340492249, | |
| "eval_runtime": 16.6999, | |
| "eval_samples_per_second": 585.034, | |
| "eval_steps_per_second": 18.323, | |
| "step": 42000 | |
| }, | |
| { | |
| "epoch": 3.67, | |
| "learning_rate": 5.333333333333334e-06, | |
| "loss": 0.1209, | |
| "step": 44000 | |
| }, | |
| { | |
| "epoch": 3.67, | |
| "eval_acc": 0.8580348004094166, | |
| "eval_loss": 0.6536660194396973, | |
| "eval_runtime": 16.7907, | |
| "eval_samples_per_second": 581.87, | |
| "eval_steps_per_second": 18.224, | |
| "step": 44000 | |
| }, | |
| { | |
| "epoch": 3.83, | |
| "learning_rate": 4.666666666666667e-06, | |
| "loss": 0.1234, | |
| "step": 46000 | |
| }, | |
| { | |
| "epoch": 3.83, | |
| "eval_acc": 0.8580348004094166, | |
| "eval_loss": 0.6385037899017334, | |
| "eval_runtime": 16.6773, | |
| "eval_samples_per_second": 585.826, | |
| "eval_steps_per_second": 18.348, | |
| "step": 46000 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "learning_rate": 4.000000000000001e-06, | |
| "loss": 0.122, | |
| "step": 48000 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_acc": 0.8628454452405322, | |
| "eval_loss": 0.6103753447532654, | |
| "eval_runtime": 16.681, | |
| "eval_samples_per_second": 585.697, | |
| "eval_steps_per_second": 18.344, | |
| "step": 48000 | |
| }, | |
| { | |
| "epoch": 4.17, | |
| "learning_rate": 3.3333333333333333e-06, | |
| "loss": 0.0937, | |
| "step": 50000 | |
| }, | |
| { | |
| "epoch": 4.17, | |
| "eval_acc": 0.8602865916069601, | |
| "eval_loss": 0.7260550260543823, | |
| "eval_runtime": 16.6866, | |
| "eval_samples_per_second": 585.501, | |
| "eval_steps_per_second": 18.338, | |
| "step": 50000 | |
| }, | |
| { | |
| "epoch": 4.33, | |
| "learning_rate": 2.666666666666667e-06, | |
| "loss": 0.0941, | |
| "step": 52000 | |
| }, | |
| { | |
| "epoch": 4.33, | |
| "eval_acc": 0.8606960081883316, | |
| "eval_loss": 0.7633941769599915, | |
| "eval_runtime": 16.696, | |
| "eval_samples_per_second": 585.17, | |
| "eval_steps_per_second": 18.328, | |
| "step": 52000 | |
| }, | |
| { | |
| "epoch": 4.5, | |
| "learning_rate": 2.0000000000000003e-06, | |
| "loss": 0.0926, | |
| "step": 54000 | |
| }, | |
| { | |
| "epoch": 4.5, | |
| "eval_acc": 0.8617195496417605, | |
| "eval_loss": 0.7736791968345642, | |
| "eval_runtime": 16.6764, | |
| "eval_samples_per_second": 585.859, | |
| "eval_steps_per_second": 18.349, | |
| "step": 54000 | |
| }, | |
| { | |
| "epoch": 4.67, | |
| "learning_rate": 1.3333333333333334e-06, | |
| "loss": 0.093, | |
| "step": 56000 | |
| }, | |
| { | |
| "epoch": 4.67, | |
| "eval_acc": 0.8632548618219038, | |
| "eval_loss": 0.757666289806366, | |
| "eval_runtime": 16.6803, | |
| "eval_samples_per_second": 585.722, | |
| "eval_steps_per_second": 18.345, | |
| "step": 56000 | |
| }, | |
| { | |
| "epoch": 4.83, | |
| "learning_rate": 6.666666666666667e-07, | |
| "loss": 0.0918, | |
| "step": 58000 | |
| }, | |
| { | |
| "epoch": 4.83, | |
| "eval_acc": 0.8630501535312181, | |
| "eval_loss": 0.7534608840942383, | |
| "eval_runtime": 16.6933, | |
| "eval_samples_per_second": 585.263, | |
| "eval_steps_per_second": 18.331, | |
| "step": 58000 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "learning_rate": 0.0, | |
| "loss": 0.0935, | |
| "step": 60000 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_acc": 0.8622313203684749, | |
| "eval_loss": 0.743323802947998, | |
| "eval_runtime": 16.7307, | |
| "eval_samples_per_second": 583.956, | |
| "eval_steps_per_second": 18.29, | |
| "step": 60000 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "step": 60000, | |
| "total_flos": 1.2628556043159168e+17, | |
| "train_loss": 0.18972002029418944, | |
| "train_runtime": 11504.4253, | |
| "train_samples_per_second": 166.881, | |
| "train_steps_per_second": 5.215 | |
| } | |
| ], | |
| "max_steps": 60000, | |
| "num_train_epochs": 5, | |
| "total_flos": 1.2628556043159168e+17, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |