| { | |
| "best_metric": 0.8406790060333368, | |
| "best_model_checkpoint": "./outputs/finetuning/mnli_MULTI/checkpoint-26000", | |
| "epoch": 5.0, | |
| "global_step": 60065, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.17, | |
| "learning_rate": 1.933405477399484e-05, | |
| "loss": 0.4798, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.17, | |
| "eval_acc": 0.8211473565804275, | |
| "eval_loss": 0.4810026288032532, | |
| "eval_runtime": 16.7598, | |
| "eval_samples_per_second": 583.48, | |
| "eval_steps_per_second": 18.258, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "learning_rate": 1.866810954798968e-05, | |
| "loss": 0.4494, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "eval_acc": 0.8267716535433071, | |
| "eval_loss": 0.46056851744651794, | |
| "eval_runtime": 16.6747, | |
| "eval_samples_per_second": 586.457, | |
| "eval_steps_per_second": 18.351, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "learning_rate": 1.8002164321984518e-05, | |
| "loss": 0.431, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "eval_acc": 0.8301462317210349, | |
| "eval_loss": 0.4622686803340912, | |
| "eval_runtime": 16.6907, | |
| "eval_samples_per_second": 585.896, | |
| "eval_steps_per_second": 18.334, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "learning_rate": 1.7336219095979357e-05, | |
| "loss": 0.4371, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "eval_acc": 0.8297371919419163, | |
| "eval_loss": 0.44371065497398376, | |
| "eval_runtime": 16.7034, | |
| "eval_samples_per_second": 585.451, | |
| "eval_steps_per_second": 18.32, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 0.83, | |
| "learning_rate": 1.6670273869974196e-05, | |
| "loss": 0.4297, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 0.83, | |
| "eval_acc": 0.8312710911136107, | |
| "eval_loss": 0.45482972264289856, | |
| "eval_runtime": 16.6876, | |
| "eval_samples_per_second": 586.004, | |
| "eval_steps_per_second": 18.337, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "learning_rate": 1.6004328643969035e-05, | |
| "loss": 0.4214, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_acc": 0.8337253297883219, | |
| "eval_loss": 0.4565775692462921, | |
| "eval_runtime": 16.6965, | |
| "eval_samples_per_second": 585.692, | |
| "eval_steps_per_second": 18.327, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 1.17, | |
| "learning_rate": 1.5338383417963873e-05, | |
| "loss": 0.3123, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 1.17, | |
| "eval_acc": 0.8322936905614071, | |
| "eval_loss": 0.48932746052742004, | |
| "eval_runtime": 16.699, | |
| "eval_samples_per_second": 585.603, | |
| "eval_steps_per_second": 18.324, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 1.33, | |
| "learning_rate": 1.4672438191958714e-05, | |
| "loss": 0.3158, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 1.33, | |
| "eval_acc": 0.8342366295122201, | |
| "eval_loss": 0.4861135184764862, | |
| "eval_runtime": 16.699, | |
| "eval_samples_per_second": 585.602, | |
| "eval_steps_per_second": 18.324, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 1.5, | |
| "learning_rate": 1.4006492965953551e-05, | |
| "loss": 0.324, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 1.5, | |
| "eval_acc": 0.8307597913897127, | |
| "eval_loss": 0.4812241792678833, | |
| "eval_runtime": 16.6905, | |
| "eval_samples_per_second": 585.902, | |
| "eval_steps_per_second": 18.334, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 1.66, | |
| "learning_rate": 1.3340547739948392e-05, | |
| "loss": 0.3161, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 1.66, | |
| "eval_acc": 0.8364863482973719, | |
| "eval_loss": 0.4630277454853058, | |
| "eval_runtime": 16.702, | |
| "eval_samples_per_second": 585.498, | |
| "eval_steps_per_second": 18.321, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 1.83, | |
| "learning_rate": 1.2674602513943229e-05, | |
| "loss": 0.32, | |
| "step": 22000 | |
| }, | |
| { | |
| "epoch": 1.83, | |
| "eval_acc": 0.8365886082421515, | |
| "eval_loss": 0.46297991275787354, | |
| "eval_runtime": 16.6817, | |
| "eval_samples_per_second": 586.213, | |
| "eval_steps_per_second": 18.343, | |
| "step": 22000 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "learning_rate": 1.2008657287938067e-05, | |
| "loss": 0.3195, | |
| "step": 24000 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_acc": 0.8354637488495756, | |
| "eval_loss": 0.4681137800216675, | |
| "eval_runtime": 16.7123, | |
| "eval_samples_per_second": 585.138, | |
| "eval_steps_per_second": 18.31, | |
| "step": 24000 | |
| }, | |
| { | |
| "epoch": 2.16, | |
| "learning_rate": 1.1342712061932908e-05, | |
| "loss": 0.2274, | |
| "step": 26000 | |
| }, | |
| { | |
| "epoch": 2.16, | |
| "eval_acc": 0.8406790060333368, | |
| "eval_loss": 0.534744143486023, | |
| "eval_runtime": 16.6972, | |
| "eval_samples_per_second": 585.668, | |
| "eval_steps_per_second": 18.326, | |
| "step": 26000 | |
| }, | |
| { | |
| "epoch": 2.33, | |
| "learning_rate": 1.0676766835927745e-05, | |
| "loss": 0.2311, | |
| "step": 28000 | |
| }, | |
| { | |
| "epoch": 2.33, | |
| "eval_acc": 0.830964311279272, | |
| "eval_loss": 0.5649741291999817, | |
| "eval_runtime": 16.7014, | |
| "eval_samples_per_second": 585.518, | |
| "eval_steps_per_second": 18.322, | |
| "step": 28000 | |
| }, | |
| { | |
| "epoch": 2.5, | |
| "learning_rate": 1.0010821609922586e-05, | |
| "loss": 0.2293, | |
| "step": 30000 | |
| }, | |
| { | |
| "epoch": 2.5, | |
| "eval_acc": 0.8354637488495756, | |
| "eval_loss": 0.5407743453979492, | |
| "eval_runtime": 16.7112, | |
| "eval_samples_per_second": 585.175, | |
| "eval_steps_per_second": 18.311, | |
| "step": 30000 | |
| }, | |
| { | |
| "epoch": 2.66, | |
| "learning_rate": 9.344876383917424e-06, | |
| "loss": 0.2296, | |
| "step": 32000 | |
| }, | |
| { | |
| "epoch": 2.66, | |
| "eval_acc": 0.8374066878003886, | |
| "eval_loss": 0.5207422971725464, | |
| "eval_runtime": 16.6974, | |
| "eval_samples_per_second": 585.661, | |
| "eval_steps_per_second": 18.326, | |
| "step": 32000 | |
| }, | |
| { | |
| "epoch": 2.83, | |
| "learning_rate": 8.678931157912263e-06, | |
| "loss": 0.2274, | |
| "step": 34000 | |
| }, | |
| { | |
| "epoch": 2.83, | |
| "eval_acc": 0.8352592289600164, | |
| "eval_loss": 0.5696293115615845, | |
| "eval_runtime": 16.683, | |
| "eval_samples_per_second": 586.165, | |
| "eval_steps_per_second": 18.342, | |
| "step": 34000 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "learning_rate": 8.012985931907102e-06, | |
| "loss": 0.23, | |
| "step": 36000 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_acc": 0.8365886082421515, | |
| "eval_loss": 0.5331636071205139, | |
| "eval_runtime": 16.7244, | |
| "eval_samples_per_second": 584.714, | |
| "eval_steps_per_second": 18.297, | |
| "step": 36000 | |
| }, | |
| { | |
| "epoch": 3.16, | |
| "learning_rate": 7.34704070590194e-06, | |
| "loss": 0.1686, | |
| "step": 38000 | |
| }, | |
| { | |
| "epoch": 3.16, | |
| "eval_acc": 0.8343388894569997, | |
| "eval_loss": 0.6275357007980347, | |
| "eval_runtime": 16.6784, | |
| "eval_samples_per_second": 586.326, | |
| "eval_steps_per_second": 18.347, | |
| "step": 38000 | |
| }, | |
| { | |
| "epoch": 3.33, | |
| "learning_rate": 6.681095479896779e-06, | |
| "loss": 0.1632, | |
| "step": 40000 | |
| }, | |
| { | |
| "epoch": 3.33, | |
| "eval_acc": 0.8348501891808978, | |
| "eval_loss": 0.6457108855247498, | |
| "eval_runtime": 16.7007, | |
| "eval_samples_per_second": 585.544, | |
| "eval_steps_per_second": 18.323, | |
| "step": 40000 | |
| }, | |
| { | |
| "epoch": 3.5, | |
| "learning_rate": 6.0151502538916185e-06, | |
| "loss": 0.1686, | |
| "step": 42000 | |
| }, | |
| { | |
| "epoch": 3.5, | |
| "eval_acc": 0.8338275897331016, | |
| "eval_loss": 0.5964699983596802, | |
| "eval_runtime": 16.6934, | |
| "eval_samples_per_second": 585.8, | |
| "eval_steps_per_second": 18.331, | |
| "step": 42000 | |
| }, | |
| { | |
| "epoch": 3.66, | |
| "learning_rate": 5.349205027886457e-06, | |
| "loss": 0.1634, | |
| "step": 44000 | |
| }, | |
| { | |
| "epoch": 3.66, | |
| "eval_acc": 0.8342366295122201, | |
| "eval_loss": 0.6272006034851074, | |
| "eval_runtime": 16.6672, | |
| "eval_samples_per_second": 586.722, | |
| "eval_steps_per_second": 18.359, | |
| "step": 44000 | |
| }, | |
| { | |
| "epoch": 3.83, | |
| "learning_rate": 4.683259801881296e-06, | |
| "loss": 0.1656, | |
| "step": 46000 | |
| }, | |
| { | |
| "epoch": 3.83, | |
| "eval_acc": 0.8311688311688312, | |
| "eval_loss": 0.6541053652763367, | |
| "eval_runtime": 16.6856, | |
| "eval_samples_per_second": 586.076, | |
| "eval_steps_per_second": 18.339, | |
| "step": 46000 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "learning_rate": 4.017314575876134e-06, | |
| "loss": 0.162, | |
| "step": 48000 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_acc": 0.8316801308927293, | |
| "eval_loss": 0.6408036947250366, | |
| "eval_runtime": 16.6809, | |
| "eval_samples_per_second": 586.239, | |
| "eval_steps_per_second": 18.344, | |
| "step": 48000 | |
| }, | |
| { | |
| "epoch": 4.16, | |
| "learning_rate": 3.3513693498709734e-06, | |
| "loss": 0.1288, | |
| "step": 50000 | |
| }, | |
| { | |
| "epoch": 4.16, | |
| "eval_acc": 0.8348501891808978, | |
| "eval_loss": 0.7236860990524292, | |
| "eval_runtime": 16.7011, | |
| "eval_samples_per_second": 585.529, | |
| "eval_steps_per_second": 18.322, | |
| "step": 50000 | |
| }, | |
| { | |
| "epoch": 4.33, | |
| "learning_rate": 2.6854241238658126e-06, | |
| "loss": 0.1275, | |
| "step": 52000 | |
| }, | |
| { | |
| "epoch": 4.33, | |
| "eval_acc": 0.8295326720523571, | |
| "eval_loss": 0.7558159828186035, | |
| "eval_runtime": 16.7056, | |
| "eval_samples_per_second": 585.372, | |
| "eval_steps_per_second": 18.317, | |
| "step": 52000 | |
| }, | |
| { | |
| "epoch": 4.5, | |
| "learning_rate": 2.019478897860651e-06, | |
| "loss": 0.1291, | |
| "step": 54000 | |
| }, | |
| { | |
| "epoch": 4.5, | |
| "eval_acc": 0.8305552715001534, | |
| "eval_loss": 0.7729807496070862, | |
| "eval_runtime": 16.6766, | |
| "eval_samples_per_second": 586.391, | |
| "eval_steps_per_second": 18.349, | |
| "step": 54000 | |
| }, | |
| { | |
| "epoch": 4.66, | |
| "learning_rate": 1.35353367185549e-06, | |
| "loss": 0.1261, | |
| "step": 56000 | |
| }, | |
| { | |
| "epoch": 4.66, | |
| "eval_acc": 0.8300439717762552, | |
| "eval_loss": 0.7523751258850098, | |
| "eval_runtime": 16.688, | |
| "eval_samples_per_second": 585.99, | |
| "eval_steps_per_second": 18.337, | |
| "step": 56000 | |
| }, | |
| { | |
| "epoch": 4.83, | |
| "learning_rate": 6.875884458503289e-07, | |
| "loss": 0.1272, | |
| "step": 58000 | |
| }, | |
| { | |
| "epoch": 4.83, | |
| "eval_acc": 0.8316801308927293, | |
| "eval_loss": 0.7572413682937622, | |
| "eval_runtime": 16.7304, | |
| "eval_samples_per_second": 584.506, | |
| "eval_steps_per_second": 18.29, | |
| "step": 58000 | |
| }, | |
| { | |
| "epoch": 4.99, | |
| "learning_rate": 2.164321984516774e-08, | |
| "loss": 0.1242, | |
| "step": 60000 | |
| }, | |
| { | |
| "epoch": 4.99, | |
| "eval_acc": 0.830964311279272, | |
| "eval_loss": 0.7606698870658875, | |
| "eval_runtime": 16.6838, | |
| "eval_samples_per_second": 586.137, | |
| "eval_steps_per_second": 18.341, | |
| "step": 60000 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "step": 60065, | |
| "total_flos": 1.2642205051703808e+17, | |
| "train_loss": 0.2560187041367518, | |
| "train_runtime": 11508.9997, | |
| "train_samples_per_second": 166.995, | |
| "train_steps_per_second": 5.219 | |
| } | |
| ], | |
| "max_steps": 60065, | |
| "num_train_epochs": 5, | |
| "total_flos": 1.2642205051703808e+17, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |