| { | |
| "best_metric": 0.948394495412844, | |
| "best_model_checkpoint": "output/fine_tuned/t5-base/SST2/checkpoint-8420", | |
| "epoch": 10.0, | |
| "eval_steps": 500, | |
| "global_step": 21050, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.2375296912114014, | |
| "grad_norm": 3.5314207077026367, | |
| "learning_rate": 4.8812351543942996e-05, | |
| "loss": 0.2839, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.4750593824228028, | |
| "grad_norm": 3.3338215351104736, | |
| "learning_rate": 4.7624703087885984e-05, | |
| "loss": 0.1706, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.7125890736342043, | |
| "grad_norm": 2.111057996749878, | |
| "learning_rate": 4.643705463182898e-05, | |
| "loss": 0.1579, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.9501187648456056, | |
| "grad_norm": 5.4748430252075195, | |
| "learning_rate": 4.524940617577197e-05, | |
| "loss": 0.1443, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.9323394495412844, | |
| "eval_loss": 0.20724497735500336, | |
| "eval_runtime": 2.2627, | |
| "eval_samples_per_second": 385.376, | |
| "eval_steps_per_second": 48.172, | |
| "step": 2105 | |
| }, | |
| { | |
| "epoch": 1.187648456057007, | |
| "grad_norm": 4.547887325286865, | |
| "learning_rate": 4.406175771971497e-05, | |
| "loss": 0.1207, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 1.4251781472684084, | |
| "grad_norm": 3.1552112102508545, | |
| "learning_rate": 4.287410926365796e-05, | |
| "loss": 0.113, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 1.66270783847981, | |
| "grad_norm": 0.2997496724128723, | |
| "learning_rate": 4.168646080760095e-05, | |
| "loss": 0.1112, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 1.9002375296912115, | |
| "grad_norm": 1.9386916160583496, | |
| "learning_rate": 4.049881235154395e-05, | |
| "loss": 0.1152, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.9403669724770642, | |
| "eval_loss": 0.21269434690475464, | |
| "eval_runtime": 1.9647, | |
| "eval_samples_per_second": 443.834, | |
| "eval_steps_per_second": 55.479, | |
| "step": 4210 | |
| }, | |
| { | |
| "epoch": 2.137767220902613, | |
| "grad_norm": 0.23684050142765045, | |
| "learning_rate": 3.9311163895486934e-05, | |
| "loss": 0.0952, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 2.375296912114014, | |
| "grad_norm": 1.943021297454834, | |
| "learning_rate": 3.812351543942993e-05, | |
| "loss": 0.0881, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 2.6128266033254155, | |
| "grad_norm": 4.209224700927734, | |
| "learning_rate": 3.693586698337292e-05, | |
| "loss": 0.0913, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 2.850356294536817, | |
| "grad_norm": 4.389578819274902, | |
| "learning_rate": 3.5748218527315916e-05, | |
| "loss": 0.0849, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.9438073394495413, | |
| "eval_loss": 0.2156314253807068, | |
| "eval_runtime": 1.9686, | |
| "eval_samples_per_second": 442.964, | |
| "eval_steps_per_second": 55.371, | |
| "step": 6315 | |
| }, | |
| { | |
| "epoch": 3.0878859857482186, | |
| "grad_norm": 0.4832597076892853, | |
| "learning_rate": 3.456057007125891e-05, | |
| "loss": 0.075, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 3.32541567695962, | |
| "grad_norm": 6.154821395874023, | |
| "learning_rate": 3.3372921615201904e-05, | |
| "loss": 0.0719, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 3.5629453681710213, | |
| "grad_norm": 2.8581926822662354, | |
| "learning_rate": 3.21852731591449e-05, | |
| "loss": 0.0727, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 3.800475059382423, | |
| "grad_norm": 4.29818868637085, | |
| "learning_rate": 3.0997624703087885e-05, | |
| "loss": 0.0709, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.948394495412844, | |
| "eval_loss": 0.222471222281456, | |
| "eval_runtime": 1.9642, | |
| "eval_samples_per_second": 443.937, | |
| "eval_steps_per_second": 55.492, | |
| "step": 8420 | |
| }, | |
| { | |
| "epoch": 4.038004750593824, | |
| "grad_norm": 1.9306527376174927, | |
| "learning_rate": 2.980997624703088e-05, | |
| "loss": 0.0724, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 4.275534441805226, | |
| "grad_norm": 2.4774186611175537, | |
| "learning_rate": 2.8622327790973873e-05, | |
| "loss": 0.0541, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 4.513064133016627, | |
| "grad_norm": 2.690717935562134, | |
| "learning_rate": 2.7434679334916867e-05, | |
| "loss": 0.0568, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 4.750593824228028, | |
| "grad_norm": 1.1091784238815308, | |
| "learning_rate": 2.6247030878859858e-05, | |
| "loss": 0.0602, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 4.98812351543943, | |
| "grad_norm": 1.5014886856079102, | |
| "learning_rate": 2.5059382422802852e-05, | |
| "loss": 0.06, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.9403669724770642, | |
| "eval_loss": 0.2718953788280487, | |
| "eval_runtime": 1.9734, | |
| "eval_samples_per_second": 441.874, | |
| "eval_steps_per_second": 55.234, | |
| "step": 10525 | |
| }, | |
| { | |
| "epoch": 5.225653206650831, | |
| "grad_norm": 1.5026357173919678, | |
| "learning_rate": 2.3871733966745842e-05, | |
| "loss": 0.0428, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 5.463182897862232, | |
| "grad_norm": 6.937454700469971, | |
| "learning_rate": 2.2684085510688836e-05, | |
| "loss": 0.0428, | |
| "step": 11500 | |
| }, | |
| { | |
| "epoch": 5.7007125890736345, | |
| "grad_norm": 2.493163824081421, | |
| "learning_rate": 2.149643705463183e-05, | |
| "loss": 0.0502, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 5.938242280285036, | |
| "grad_norm": 0.8779876828193665, | |
| "learning_rate": 2.0308788598574824e-05, | |
| "loss": 0.0507, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_accuracy": 0.9403669724770642, | |
| "eval_loss": 0.29110807180404663, | |
| "eval_runtime": 1.965, | |
| "eval_samples_per_second": 443.769, | |
| "eval_steps_per_second": 55.471, | |
| "step": 12630 | |
| }, | |
| { | |
| "epoch": 6.175771971496437, | |
| "grad_norm": 1.818724513053894, | |
| "learning_rate": 1.9121140142517815e-05, | |
| "loss": 0.0409, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 6.4133016627078385, | |
| "grad_norm": 0.3184266984462738, | |
| "learning_rate": 1.793349168646081e-05, | |
| "loss": 0.0378, | |
| "step": 13500 | |
| }, | |
| { | |
| "epoch": 6.65083135391924, | |
| "grad_norm": 0.03643421456217766, | |
| "learning_rate": 1.6745843230403803e-05, | |
| "loss": 0.0397, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 6.888361045130641, | |
| "grad_norm": 0.04083951190114021, | |
| "learning_rate": 1.5558194774346793e-05, | |
| "loss": 0.0435, | |
| "step": 14500 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_accuracy": 0.9334862385321101, | |
| "eval_loss": 0.3279486894607544, | |
| "eval_runtime": 1.9653, | |
| "eval_samples_per_second": 443.688, | |
| "eval_steps_per_second": 55.461, | |
| "step": 14735 | |
| }, | |
| { | |
| "epoch": 7.1258907363420425, | |
| "grad_norm": 0.06795825809240341, | |
| "learning_rate": 1.4370546318289787e-05, | |
| "loss": 0.0371, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 7.363420427553444, | |
| "grad_norm": 3.7526097297668457, | |
| "learning_rate": 1.318289786223278e-05, | |
| "loss": 0.0299, | |
| "step": 15500 | |
| }, | |
| { | |
| "epoch": 7.600950118764846, | |
| "grad_norm": 2.537559747695923, | |
| "learning_rate": 1.1995249406175772e-05, | |
| "loss": 0.0302, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 7.838479809976247, | |
| "grad_norm": 3.0933053493499756, | |
| "learning_rate": 1.0807600950118766e-05, | |
| "loss": 0.0357, | |
| "step": 16500 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_accuracy": 0.9311926605504587, | |
| "eval_loss": 0.3565831780433655, | |
| "eval_runtime": 1.9696, | |
| "eval_samples_per_second": 442.724, | |
| "eval_steps_per_second": 55.341, | |
| "step": 16840 | |
| }, | |
| { | |
| "epoch": 8.076009501187649, | |
| "grad_norm": 2.52713680267334, | |
| "learning_rate": 9.619952494061758e-06, | |
| "loss": 0.037, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 8.31353919239905, | |
| "grad_norm": 0.04675888270139694, | |
| "learning_rate": 8.432304038004752e-06, | |
| "loss": 0.0276, | |
| "step": 17500 | |
| }, | |
| { | |
| "epoch": 8.551068883610451, | |
| "grad_norm": 5.1887593269348145, | |
| "learning_rate": 7.244655581947744e-06, | |
| "loss": 0.0294, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 8.788598574821853, | |
| "grad_norm": 0.14877742528915405, | |
| "learning_rate": 6.0570071258907366e-06, | |
| "loss": 0.0274, | |
| "step": 18500 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_accuracy": 0.9357798165137615, | |
| "eval_loss": 0.38757723569869995, | |
| "eval_runtime": 1.9636, | |
| "eval_samples_per_second": 444.076, | |
| "eval_steps_per_second": 55.509, | |
| "step": 18945 | |
| }, | |
| { | |
| "epoch": 9.026128266033254, | |
| "grad_norm": 0.024900976568460464, | |
| "learning_rate": 4.86935866983373e-06, | |
| "loss": 0.0282, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 9.263657957244655, | |
| "grad_norm": 0.1904192417860031, | |
| "learning_rate": 3.681710213776722e-06, | |
| "loss": 0.0213, | |
| "step": 19500 | |
| }, | |
| { | |
| "epoch": 9.501187648456057, | |
| "grad_norm": 5.56980562210083, | |
| "learning_rate": 2.494061757719715e-06, | |
| "loss": 0.0218, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 9.738717339667458, | |
| "grad_norm": 3.3612709045410156, | |
| "learning_rate": 1.306413301662708e-06, | |
| "loss": 0.0272, | |
| "step": 20500 | |
| }, | |
| { | |
| "epoch": 9.97624703087886, | |
| "grad_norm": 4.481935024261475, | |
| "learning_rate": 1.1876484560570073e-07, | |
| "loss": 0.0253, | |
| "step": 21000 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_accuracy": 0.9380733944954128, | |
| "eval_loss": 0.4033842086791992, | |
| "eval_runtime": 1.9746, | |
| "eval_samples_per_second": 441.601, | |
| "eval_steps_per_second": 55.2, | |
| "step": 21050 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "step": 21050, | |
| "total_flos": 1.028380359389952e+17, | |
| "train_loss": 0.06886951911477748, | |
| "train_runtime": 3499.6769, | |
| "train_samples_per_second": 192.443, | |
| "train_steps_per_second": 6.015 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 21050, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 10, | |
| "save_steps": 500, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 1.028380359389952e+17, | |
| "train_batch_size": 32, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |