| { | |
| "best_metric": 22.9493, | |
| "best_model_checkpoint": "./ko-en_mbartLarge_mid3/checkpoint-22000", | |
| "epoch": 3.4806822137138878, | |
| "eval_steps": 2000, | |
| "global_step": 30000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.06, | |
| "learning_rate": 2.5e-05, | |
| "loss": 1.9104, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.12, | |
| "learning_rate": 5e-05, | |
| "loss": 1.7744, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.17, | |
| "learning_rate": 4.99997390013085e-05, | |
| "loss": 1.6281, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.23, | |
| "learning_rate": 4.9998956010683606e-05, | |
| "loss": 1.5377, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.23, | |
| "eval_bleu": 17.2009, | |
| "eval_gen_len": 18.7106, | |
| "eval_loss": 1.6122242212295532, | |
| "eval_runtime": 1146.4034, | |
| "eval_samples_per_second": 15.036, | |
| "eval_steps_per_second": 0.94, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "learning_rate": 4.9997651044474085e-05, | |
| "loss": 1.4666, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 0.35, | |
| "learning_rate": 4.999582412992751e-05, | |
| "loss": 1.4294, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.41, | |
| "learning_rate": 4.999347530518964e-05, | |
| "loss": 1.3956, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "learning_rate": 4.999060461930371e-05, | |
| "loss": 1.3891, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "eval_bleu": 19.3345, | |
| "eval_gen_len": 18.7688, | |
| "eval_loss": 1.5059210062026978, | |
| "eval_runtime": 1145.6172, | |
| "eval_samples_per_second": 15.046, | |
| "eval_steps_per_second": 0.941, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "learning_rate": 4.998721213220933e-05, | |
| "loss": 1.3651, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 0.58, | |
| "learning_rate": 4.9983297914741276e-05, | |
| "loss": 1.3337, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "learning_rate": 4.9978862048628e-05, | |
| "loss": 1.3017, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "learning_rate": 4.997390462648993e-05, | |
| "loss": 1.2812, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "eval_bleu": 20.6032, | |
| "eval_gen_len": 18.9022, | |
| "eval_loss": 1.4347867965698242, | |
| "eval_runtime": 1146.1733, | |
| "eval_samples_per_second": 15.039, | |
| "eval_steps_per_second": 0.941, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "learning_rate": 4.996842575183751e-05, | |
| "loss": 1.2664, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 0.81, | |
| "learning_rate": 4.996242553906908e-05, | |
| "loss": 1.2599, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 0.87, | |
| "learning_rate": 4.9955904113468444e-05, | |
| "loss": 1.2513, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 0.93, | |
| "learning_rate": 4.994886161120229e-05, | |
| "loss": 1.2374, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 0.93, | |
| "eval_bleu": 21.2391, | |
| "eval_gen_len": 18.8434, | |
| "eval_loss": 1.4034597873687744, | |
| "eval_runtime": 1157.6007, | |
| "eval_samples_per_second": 14.89, | |
| "eval_steps_per_second": 0.931, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 0.99, | |
| "learning_rate": 4.994129817931734e-05, | |
| "loss": 1.2306, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 1.04, | |
| "learning_rate": 4.993321397573724e-05, | |
| "loss": 1.1716, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 1.1, | |
| "learning_rate": 4.9924609169259325e-05, | |
| "loss": 1.1718, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 1.16, | |
| "learning_rate": 4.991548393955104e-05, | |
| "loss": 1.1734, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 1.16, | |
| "eval_bleu": 21.304, | |
| "eval_gen_len": 18.9964, | |
| "eval_loss": 1.4039467573165894, | |
| "eval_runtime": 1151.81, | |
| "eval_samples_per_second": 14.965, | |
| "eval_steps_per_second": 0.936, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 1.22, | |
| "learning_rate": 4.990583847714626e-05, | |
| "loss": 1.1702, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 1.28, | |
| "learning_rate": 4.9895672983441194e-05, | |
| "loss": 1.1688, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 1.33, | |
| "learning_rate": 4.988498767069031e-05, | |
| "loss": 1.1581, | |
| "step": 11500 | |
| }, | |
| { | |
| "epoch": 1.39, | |
| "learning_rate": 4.98737827620018e-05, | |
| "loss": 1.1531, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 1.39, | |
| "eval_bleu": 21.9087, | |
| "eval_gen_len": 18.8573, | |
| "eval_loss": 1.3693655729293823, | |
| "eval_runtime": 1144.9147, | |
| "eval_samples_per_second": 15.055, | |
| "eval_steps_per_second": 0.942, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 1.45, | |
| "learning_rate": 4.9862058491333e-05, | |
| "loss": 1.1184, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 1.51, | |
| "learning_rate": 4.984981510348545e-05, | |
| "loss": 1.1171, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 1.57, | |
| "learning_rate": 4.9837052854099804e-05, | |
| "loss": 1.1149, | |
| "step": 13500 | |
| }, | |
| { | |
| "epoch": 1.62, | |
| "learning_rate": 4.9823772009650496e-05, | |
| "loss": 1.1158, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 1.62, | |
| "eval_bleu": 22.004, | |
| "eval_gen_len": 18.5485, | |
| "eval_loss": 1.357371211051941, | |
| "eval_runtime": 1129.6253, | |
| "eval_samples_per_second": 15.259, | |
| "eval_steps_per_second": 0.954, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 1.68, | |
| "learning_rate": 4.980997284744017e-05, | |
| "loss": 1.1163, | |
| "step": 14500 | |
| }, | |
| { | |
| "epoch": 1.74, | |
| "learning_rate": 4.979565565559388e-05, | |
| "loss": 1.1147, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 1.8, | |
| "learning_rate": 4.97808207330531e-05, | |
| "loss": 1.0988, | |
| "step": 15500 | |
| }, | |
| { | |
| "epoch": 1.86, | |
| "learning_rate": 4.9765468389569455e-05, | |
| "loss": 1.0941, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 1.86, | |
| "eval_bleu": 21.9785, | |
| "eval_gen_len": 18.7119, | |
| "eval_loss": 1.3456674814224243, | |
| "eval_runtime": 1149.9948, | |
| "eval_samples_per_second": 14.989, | |
| "eval_steps_per_second": 0.937, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 1.91, | |
| "learning_rate": 4.9749598945698276e-05, | |
| "loss": 1.1112, | |
| "step": 16500 | |
| }, | |
| { | |
| "epoch": 1.97, | |
| "learning_rate": 4.973321273279189e-05, | |
| "loss": 1.1, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 2.03, | |
| "learning_rate": 4.97163100929927e-05, | |
| "loss": 1.0405, | |
| "step": 17500 | |
| }, | |
| { | |
| "epoch": 2.09, | |
| "learning_rate": 4.9698891379226064e-05, | |
| "loss": 0.9809, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 2.09, | |
| "eval_bleu": 22.7983, | |
| "eval_gen_len": 18.8011, | |
| "eval_loss": 1.3494887351989746, | |
| "eval_runtime": 1141.8247, | |
| "eval_samples_per_second": 15.096, | |
| "eval_steps_per_second": 0.944, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 2.15, | |
| "learning_rate": 4.9680956955192896e-05, | |
| "loss": 0.9747, | |
| "step": 18500 | |
| }, | |
| { | |
| "epoch": 2.2, | |
| "learning_rate": 4.966250719536209e-05, | |
| "loss": 0.9933, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 2.26, | |
| "learning_rate": 4.964354248496271e-05, | |
| "loss": 0.9972, | |
| "step": 19500 | |
| }, | |
| { | |
| "epoch": 2.32, | |
| "learning_rate": 4.962406321997592e-05, | |
| "loss": 0.9834, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 2.32, | |
| "eval_bleu": 22.5654, | |
| "eval_gen_len": 18.9416, | |
| "eval_loss": 1.3429056406021118, | |
| "eval_runtime": 1150.5664, | |
| "eval_samples_per_second": 14.981, | |
| "eval_steps_per_second": 0.937, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 2.38, | |
| "learning_rate": 4.960406980712672e-05, | |
| "loss": 0.9855, | |
| "step": 20500 | |
| }, | |
| { | |
| "epoch": 2.44, | |
| "learning_rate": 4.9583562663875496e-05, | |
| "loss": 0.9939, | |
| "step": 21000 | |
| }, | |
| { | |
| "epoch": 2.49, | |
| "learning_rate": 4.956254221840925e-05, | |
| "loss": 0.985, | |
| "step": 21500 | |
| }, | |
| { | |
| "epoch": 2.55, | |
| "learning_rate": 4.9541008909632674e-05, | |
| "loss": 0.9981, | |
| "step": 22000 | |
| }, | |
| { | |
| "epoch": 2.55, | |
| "eval_bleu": 22.9493, | |
| "eval_gen_len": 18.7364, | |
| "eval_loss": 1.324600100517273, | |
| "eval_runtime": 1151.8793, | |
| "eval_samples_per_second": 14.964, | |
| "eval_steps_per_second": 0.936, | |
| "step": 22000 | |
| }, | |
| { | |
| "epoch": 2.61, | |
| "learning_rate": 4.951896318715901e-05, | |
| "loss": 0.995, | |
| "step": 22500 | |
| }, | |
| { | |
| "epoch": 2.67, | |
| "learning_rate": 4.949640551130063e-05, | |
| "loss": 1.009, | |
| "step": 23000 | |
| }, | |
| { | |
| "epoch": 2.73, | |
| "learning_rate": 4.947333635305944e-05, | |
| "loss": 1.0026, | |
| "step": 23500 | |
| }, | |
| { | |
| "epoch": 2.78, | |
| "learning_rate": 4.9449756194117056e-05, | |
| "loss": 1.0074, | |
| "step": 24000 | |
| }, | |
| { | |
| "epoch": 2.78, | |
| "eval_bleu": 22.3874, | |
| "eval_gen_len": 18.4428, | |
| "eval_loss": 1.35391366481781, | |
| "eval_runtime": 1140.2393, | |
| "eval_samples_per_second": 15.117, | |
| "eval_steps_per_second": 0.945, | |
| "step": 24000 | |
| }, | |
| { | |
| "epoch": 2.84, | |
| "learning_rate": 4.942566552682473e-05, | |
| "loss": 1.0106, | |
| "step": 24500 | |
| }, | |
| { | |
| "epoch": 2.9, | |
| "learning_rate": 4.940106485419307e-05, | |
| "loss": 1.0039, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 2.96, | |
| "learning_rate": 4.937595468988155e-05, | |
| "loss": 0.9942, | |
| "step": 25500 | |
| }, | |
| { | |
| "epoch": 3.02, | |
| "learning_rate": 4.935033555818776e-05, | |
| "loss": 0.9752, | |
| "step": 26000 | |
| }, | |
| { | |
| "epoch": 3.02, | |
| "eval_bleu": 22.1907, | |
| "eval_gen_len": 18.8139, | |
| "eval_loss": 1.3586821556091309, | |
| "eval_runtime": 1147.1144, | |
| "eval_samples_per_second": 15.026, | |
| "eval_steps_per_second": 0.94, | |
| "step": 26000 | |
| }, | |
| { | |
| "epoch": 3.07, | |
| "learning_rate": 4.932420799403651e-05, | |
| "loss": 0.8804, | |
| "step": 26500 | |
| }, | |
| { | |
| "epoch": 3.13, | |
| "learning_rate": 4.929757254296859e-05, | |
| "loss": 0.8868, | |
| "step": 27000 | |
| }, | |
| { | |
| "epoch": 3.19, | |
| "learning_rate": 4.9270429761129415e-05, | |
| "loss": 0.8998, | |
| "step": 27500 | |
| }, | |
| { | |
| "epoch": 3.25, | |
| "learning_rate": 4.924278021525745e-05, | |
| "loss": 0.8858, | |
| "step": 28000 | |
| }, | |
| { | |
| "epoch": 3.25, | |
| "eval_bleu": 22.82, | |
| "eval_gen_len": 18.8021, | |
| "eval_loss": 1.345699667930603, | |
| "eval_runtime": 1150.474, | |
| "eval_samples_per_second": 14.983, | |
| "eval_steps_per_second": 0.937, | |
| "step": 28000 | |
| }, | |
| { | |
| "epoch": 3.31, | |
| "learning_rate": 4.921462448267232e-05, | |
| "loss": 0.8772, | |
| "step": 28500 | |
| }, | |
| { | |
| "epoch": 3.36, | |
| "learning_rate": 4.918596315126276e-05, | |
| "loss": 0.8855, | |
| "step": 29000 | |
| }, | |
| { | |
| "epoch": 3.42, | |
| "learning_rate": 4.9156796819474384e-05, | |
| "loss": 0.8879, | |
| "step": 29500 | |
| }, | |
| { | |
| "epoch": 3.48, | |
| "learning_rate": 4.912712609629714e-05, | |
| "loss": 0.8895, | |
| "step": 30000 | |
| }, | |
| { | |
| "epoch": 3.48, | |
| "eval_bleu": 22.1575, | |
| "eval_gen_len": 18.5638, | |
| "eval_loss": 1.3602744340896606, | |
| "eval_runtime": 1140.37, | |
| "eval_samples_per_second": 15.115, | |
| "eval_steps_per_second": 0.945, | |
| "step": 30000 | |
| }, | |
| { | |
| "epoch": 3.48, | |
| "step": 30000, | |
| "total_flos": 1.0403263105697055e+18, | |
| "train_loss": 1.1391656209309895, | |
| "train_runtime": 39291.176, | |
| "train_samples_per_second": 140.388, | |
| "train_steps_per_second": 8.774 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 344760, | |
| "num_train_epochs": 40, | |
| "save_steps": 2000, | |
| "total_flos": 1.0403263105697055e+18, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |