| { |
| "best_global_step": null, |
| "best_metric": null, |
| "best_model_checkpoint": null, |
| "epoch": 1.2784, |
| "eval_steps": 500, |
| "global_step": 400, |
| "is_hyper_param_search": false, |
| "is_local_process_zero": true, |
| "is_world_process_zero": true, |
| "log_history": [ |
| { |
| "alignment_loss": 0.957948, |
| "epoch": 0.032, |
| "grad_norm": 13216.34765625, |
| "learning_rate": 9e-06, |
| "loss": 604.8516, |
| "mean_token_accuracy": 0.6210215302184224, |
| "num_tokens": 988428.0, |
| "step": 10, |
| "teacher_ce_loss": 39.105175 |
| }, |
| { |
| "alignment_loss": 0.946621, |
| "epoch": 0.064, |
| "grad_norm": 3091.22509765625, |
| "learning_rate": 9.853896103896104e-06, |
| "loss": 171.2203, |
| "mean_token_accuracy": 0.7685627818107605, |
| "num_tokens": 1972507.0, |
| "step": 20, |
| "teacher_ce_loss": 10.471479 |
| }, |
| { |
| "alignment_loss": 0.94924, |
| "epoch": 0.096, |
| "grad_norm": 3218.343017578125, |
| "learning_rate": 9.691558441558442e-06, |
| "loss": 115.0068, |
| "mean_token_accuracy": 0.7847720004618168, |
| "num_tokens": 2964436.0, |
| "step": 30, |
| "teacher_ce_loss": 7.006755 |
| }, |
| { |
| "alignment_loss": 0.938964, |
| "epoch": 0.128, |
| "grad_norm": 2620.491455078125, |
| "learning_rate": 9.52922077922078e-06, |
| "loss": 83.4615, |
| "mean_token_accuracy": 0.8658171582967042, |
| "num_tokens": 3939005.0, |
| "step": 40, |
| "teacher_ce_loss": 5.536875 |
| }, |
| { |
| "alignment_loss": 0.9248, |
| "epoch": 0.16, |
| "grad_norm": 1182.9658203125, |
| "learning_rate": 9.366883116883117e-06, |
| "loss": 56.0003, |
| "mean_token_accuracy": 0.8695644401013851, |
| "num_tokens": 4920499.0, |
| "step": 50, |
| "teacher_ce_loss": 3.622461 |
| }, |
| { |
| "alignment_loss": 0.910064, |
| "epoch": 0.192, |
| "grad_norm": 556.953369140625, |
| "learning_rate": 9.204545454545455e-06, |
| "loss": 46.6464, |
| "mean_token_accuracy": 0.8793487269431353, |
| "num_tokens": 5904564.0, |
| "step": 60, |
| "teacher_ce_loss": 3.133064 |
| }, |
| { |
| "alignment_loss": 0.896394, |
| "epoch": 0.224, |
| "grad_norm": 573.4602661132812, |
| "learning_rate": 9.042207792207793e-06, |
| "loss": 45.5978, |
| "mean_token_accuracy": 0.8741835005581379, |
| "num_tokens": 6882920.0, |
| "step": 70, |
| "teacher_ce_loss": 2.513722 |
| }, |
| { |
| "alignment_loss": 0.885299, |
| "epoch": 0.256, |
| "grad_norm": 481.3504333496094, |
| "learning_rate": 8.879870129870131e-06, |
| "loss": 43.9065, |
| "mean_token_accuracy": 0.8817355066537857, |
| "num_tokens": 7882978.0, |
| "step": 80, |
| "teacher_ce_loss": 2.451343 |
| }, |
| { |
| "alignment_loss": 0.872261, |
| "epoch": 0.288, |
| "grad_norm": 567.1941528320312, |
| "learning_rate": 8.717532467532468e-06, |
| "loss": 42.6578, |
| "mean_token_accuracy": 0.884648934751749, |
| "num_tokens": 8868924.0, |
| "step": 90, |
| "teacher_ce_loss": 2.764072 |
| }, |
| { |
| "alignment_loss": 0.866291, |
| "epoch": 0.32, |
| "grad_norm": 437.5704345703125, |
| "learning_rate": 8.555194805194806e-06, |
| "loss": 41.3949, |
| "mean_token_accuracy": 0.8872841037809849, |
| "num_tokens": 9861471.0, |
| "step": 100, |
| "teacher_ce_loss": 2.50916 |
| }, |
| { |
| "alignment_loss": 0.849671, |
| "epoch": 0.352, |
| "grad_norm": 419.76153564453125, |
| "learning_rate": 8.392857142857144e-06, |
| "loss": 42.0742, |
| "mean_token_accuracy": 0.8809192053973675, |
| "num_tokens": 10827332.0, |
| "step": 110, |
| "teacher_ce_loss": 2.530303 |
| }, |
| { |
| "alignment_loss": 0.839253, |
| "epoch": 0.384, |
| "grad_norm": 403.6441345214844, |
| "learning_rate": 8.23051948051948e-06, |
| "loss": 42.7459, |
| "mean_token_accuracy": 0.8831386599689722, |
| "num_tokens": 11814144.0, |
| "step": 120, |
| "teacher_ce_loss": 2.503242 |
| }, |
| { |
| "alignment_loss": 0.832565, |
| "epoch": 0.416, |
| "grad_norm": 458.2369384765625, |
| "learning_rate": 8.068181818181819e-06, |
| "loss": 43.0414, |
| "mean_token_accuracy": 0.8847327880561352, |
| "num_tokens": 12808535.0, |
| "step": 130, |
| "teacher_ce_loss": 2.526565 |
| }, |
| { |
| "alignment_loss": 0.820745, |
| "epoch": 0.448, |
| "grad_norm": 6519.52978515625, |
| "learning_rate": 7.905844155844157e-06, |
| "loss": 40.8972, |
| "mean_token_accuracy": 0.8860506009310484, |
| "num_tokens": 13803034.0, |
| "step": 140, |
| "teacher_ce_loss": 2.56256 |
| }, |
| { |
| "alignment_loss": 0.811062, |
| "epoch": 0.48, |
| "grad_norm": 411.56884765625, |
| "learning_rate": 7.743506493506494e-06, |
| "loss": 43.3087, |
| "mean_token_accuracy": 0.8850693836808204, |
| "num_tokens": 14792532.0, |
| "step": 150, |
| "teacher_ce_loss": 3.044439 |
| }, |
| { |
| "alignment_loss": 0.801598, |
| "epoch": 0.512, |
| "grad_norm": 400.7724304199219, |
| "learning_rate": 7.581168831168831e-06, |
| "loss": 40.9337, |
| "mean_token_accuracy": 0.8875109016895294, |
| "num_tokens": 15762961.0, |
| "step": 160, |
| "teacher_ce_loss": 2.304987 |
| }, |
| { |
| "alignment_loss": 0.791421, |
| "epoch": 0.544, |
| "grad_norm": 413.0471496582031, |
| "learning_rate": 7.41883116883117e-06, |
| "loss": 40.8531, |
| "mean_token_accuracy": 0.887184364721179, |
| "num_tokens": 16744298.0, |
| "step": 170, |
| "teacher_ce_loss": 2.488554 |
| }, |
| { |
| "alignment_loss": 0.789574, |
| "epoch": 0.576, |
| "grad_norm": 416.1972351074219, |
| "learning_rate": 7.2564935064935075e-06, |
| "loss": 39.661, |
| "mean_token_accuracy": 0.8898494675755501, |
| "num_tokens": 17711917.0, |
| "step": 180, |
| "teacher_ce_loss": 2.359948 |
| }, |
| { |
| "alignment_loss": 0.777883, |
| "epoch": 0.608, |
| "grad_norm": 546.523193359375, |
| "learning_rate": 7.094155844155845e-06, |
| "loss": 41.7644, |
| "mean_token_accuracy": 0.885301411151886, |
| "num_tokens": 18708337.0, |
| "step": 190, |
| "teacher_ce_loss": 3.000633 |
| }, |
| { |
| "alignment_loss": 0.762139, |
| "epoch": 0.64, |
| "grad_norm": 408.8435363769531, |
| "learning_rate": 6.931818181818183e-06, |
| "loss": 38.7193, |
| "mean_token_accuracy": 0.8898769337683916, |
| "num_tokens": 19692844.0, |
| "step": 200, |
| "teacher_ce_loss": 2.393609 |
| }, |
| { |
| "alignment_loss": 0.755599, |
| "epoch": 0.672, |
| "grad_norm": 437.32293701171875, |
| "learning_rate": 6.76948051948052e-06, |
| "loss": 38.0133, |
| "mean_token_accuracy": 0.8948926456272602, |
| "num_tokens": 20680335.0, |
| "step": 210, |
| "teacher_ce_loss": 2.142819 |
| }, |
| { |
| "alignment_loss": 0.740096, |
| "epoch": 0.704, |
| "grad_norm": 428.99456787109375, |
| "learning_rate": 6.607142857142858e-06, |
| "loss": 38.9905, |
| "mean_token_accuracy": 0.8879964012652636, |
| "num_tokens": 21661088.0, |
| "step": 220, |
| "teacher_ce_loss": 2.698433 |
| }, |
| { |
| "alignment_loss": 0.736316, |
| "epoch": 0.736, |
| "grad_norm": 416.6829833984375, |
| "learning_rate": 6.444805194805195e-06, |
| "loss": 38.9761, |
| "mean_token_accuracy": 0.8850245524197817, |
| "num_tokens": 22649359.0, |
| "step": 230, |
| "teacher_ce_loss": 2.48835 |
| }, |
| { |
| "alignment_loss": 0.728679, |
| "epoch": 0.768, |
| "grad_norm": 456.9922180175781, |
| "learning_rate": 6.282467532467533e-06, |
| "loss": 38.8624, |
| "mean_token_accuracy": 0.8895205993205308, |
| "num_tokens": 23626957.0, |
| "step": 240, |
| "teacher_ce_loss": 2.382476 |
| }, |
| { |
| "alignment_loss": 0.719706, |
| "epoch": 0.8, |
| "grad_norm": 449.4363708496094, |
| "learning_rate": 6.120129870129871e-06, |
| "loss": 37.791, |
| "mean_token_accuracy": 0.8879898354411125, |
| "num_tokens": 24620619.0, |
| "step": 250, |
| "teacher_ce_loss": 1.988624 |
| }, |
| { |
| "alignment_loss": 0.726101, |
| "epoch": 0.832, |
| "grad_norm": 746.3112182617188, |
| "learning_rate": 5.957792207792208e-06, |
| "loss": 37.5596, |
| "mean_token_accuracy": 0.8911128289997577, |
| "num_tokens": 25622825.0, |
| "step": 260, |
| "teacher_ce_loss": 2.721366 |
| }, |
| { |
| "alignment_loss": 0.717496, |
| "epoch": 0.864, |
| "grad_norm": 440.4930114746094, |
| "learning_rate": 5.795454545454546e-06, |
| "loss": 36.7011, |
| "mean_token_accuracy": 0.8905739679932594, |
| "num_tokens": 26582415.0, |
| "step": 270, |
| "teacher_ce_loss": 2.252957 |
| }, |
| { |
| "alignment_loss": 0.715658, |
| "epoch": 0.896, |
| "grad_norm": 430.8921813964844, |
| "learning_rate": 5.633116883116884e-06, |
| "loss": 37.5255, |
| "mean_token_accuracy": 0.889144329354167, |
| "num_tokens": 27575630.0, |
| "step": 280, |
| "teacher_ce_loss": 2.475199 |
| }, |
| { |
| "alignment_loss": 0.710406, |
| "epoch": 0.928, |
| "grad_norm": 430.91119384765625, |
| "learning_rate": 5.470779220779221e-06, |
| "loss": 37.6495, |
| "mean_token_accuracy": 0.8862563863396644, |
| "num_tokens": 28564913.0, |
| "step": 290, |
| "teacher_ce_loss": 2.185168 |
| }, |
| { |
| "alignment_loss": 0.702479, |
| "epoch": 0.96, |
| "grad_norm": 438.2897033691406, |
| "learning_rate": 5.308441558441559e-06, |
| "loss": 33.5691, |
| "mean_token_accuracy": 0.8983726523816585, |
| "num_tokens": 29538345.0, |
| "step": 300, |
| "teacher_ce_loss": 2.326633 |
| }, |
| { |
| "alignment_loss": 0.695988, |
| "epoch": 0.992, |
| "grad_norm": 454.2060852050781, |
| "learning_rate": 5.1461038961038965e-06, |
| "loss": 38.3144, |
| "mean_token_accuracy": 0.8862502921372652, |
| "num_tokens": 30546300.0, |
| "step": 310, |
| "teacher_ce_loss": 2.276398 |
| }, |
| { |
| "alignment_loss": 0.693005, |
| "epoch": 1.0224, |
| "grad_norm": 358.0450134277344, |
| "learning_rate": 4.983766233766234e-06, |
| "loss": 32.0346, |
| "mean_token_accuracy": 0.8995122603679958, |
| "num_tokens": 31455923.0, |
| "step": 320, |
| "teacher_ce_loss": 2.014774 |
| }, |
| { |
| "alignment_loss": 0.688472, |
| "epoch": 1.0544, |
| "grad_norm": 456.7706298828125, |
| "learning_rate": 4.821428571428572e-06, |
| "loss": 31.4628, |
| "mean_token_accuracy": 0.9042000655084849, |
| "num_tokens": 32438290.0, |
| "step": 330, |
| "teacher_ce_loss": 1.883295 |
| }, |
| { |
| "alignment_loss": 0.680988, |
| "epoch": 1.0864, |
| "grad_norm": 403.8180236816406, |
| "learning_rate": 4.6590909090909095e-06, |
| "loss": 29.5765, |
| "mean_token_accuracy": 0.9086590971797704, |
| "num_tokens": 33419384.0, |
| "step": 340, |
| "teacher_ce_loss": 1.749079 |
| }, |
| { |
| "alignment_loss": 0.672583, |
| "epoch": 1.1184, |
| "grad_norm": 410.38751220703125, |
| "learning_rate": 4.496753246753247e-06, |
| "loss": 32.8291, |
| "mean_token_accuracy": 0.8951050832867622, |
| "num_tokens": 34412656.0, |
| "step": 350, |
| "teacher_ce_loss": 1.865192 |
| }, |
| { |
| "alignment_loss": 0.669429, |
| "epoch": 1.1504, |
| "grad_norm": 394.5881042480469, |
| "learning_rate": 4.334415584415585e-06, |
| "loss": 30.7807, |
| "mean_token_accuracy": 0.9080792762339115, |
| "num_tokens": 35408817.0, |
| "step": 360, |
| "teacher_ce_loss": 1.798086 |
| }, |
| { |
| "alignment_loss": 0.669555, |
| "epoch": 1.1824, |
| "grad_norm": 348.6483459472656, |
| "learning_rate": 4.172077922077922e-06, |
| "loss": 30.4721, |
| "mean_token_accuracy": 0.9052681710571051, |
| "num_tokens": 36398372.0, |
| "step": 370, |
| "teacher_ce_loss": 1.90278 |
| }, |
| { |
| "alignment_loss": 0.659917, |
| "epoch": 1.2144, |
| "grad_norm": 418.82513427734375, |
| "learning_rate": 4.009740259740261e-06, |
| "loss": 30.013, |
| "mean_token_accuracy": 0.9085521008819342, |
| "num_tokens": 37379426.0, |
| "step": 380, |
| "teacher_ce_loss": 1.628472 |
| }, |
| { |
| "alignment_loss": 0.656111, |
| "epoch": 1.2464, |
| "grad_norm": 395.5487365722656, |
| "learning_rate": 3.847402597402597e-06, |
| "loss": 28.2695, |
| "mean_token_accuracy": 0.9148196790367364, |
| "num_tokens": 38361069.0, |
| "step": 390, |
| "teacher_ce_loss": 1.799535 |
| }, |
| { |
| "alignment_loss": 0.651777, |
| "epoch": 1.2784, |
| "grad_norm": 352.61102294921875, |
| "learning_rate": 3.6850649350649353e-06, |
| "loss": 30.5087, |
| "mean_token_accuracy": 0.9080361116677522, |
| "num_tokens": 39356772.0, |
| "step": 400, |
| "teacher_ce_loss": 2.015757 |
| } |
| ], |
| "logging_steps": 10, |
| "max_steps": 626, |
| "num_input_tokens_seen": 0, |
| "num_train_epochs": 2, |
| "save_steps": 50, |
| "stateful_callbacks": { |
| "TrainerControl": { |
| "args": { |
| "should_epoch_stop": false, |
| "should_evaluate": false, |
| "should_log": false, |
| "should_save": true, |
| "should_training_stop": false |
| }, |
| "attributes": {} |
| } |
| }, |
| "total_flos": 1.829668660940112e+18, |
| "train_batch_size": 1, |
| "trial_name": null, |
| "trial_params": null |
| } |
|
|
