| { | |
| "best_metric": 41.2618, | |
| "best_model_checkpoint": "checkpoint/kobigbirdbart_tapt_ep3_bs16_pre_noam_LB/checkpoint-34340", | |
| "epoch": 2.9999708802888674, | |
| "global_step": 51510, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.06, | |
| "learning_rate": 0.0001, | |
| "loss": 4.0417, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.12, | |
| "learning_rate": 7.072834474681765e-05, | |
| "loss": 3.6688, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.17, | |
| "learning_rate": 5.775426231077362e-05, | |
| "loss": 3.5243, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.23, | |
| "learning_rate": 5.001874180112048e-05, | |
| "loss": 3.4178, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "learning_rate": 4.473924094197383e-05, | |
| "loss": 3.367, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.35, | |
| "learning_rate": 4.084183301607125e-05, | |
| "loss": 3.3013, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 0.41, | |
| "learning_rate": 3.7812639995975144e-05, | |
| "loss": 3.2593, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 0.47, | |
| "learning_rate": 3.537080170562066e-05, | |
| "loss": 3.212, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "learning_rate": 3.3348143212252787e-05, | |
| "loss": 3.1631, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 0.58, | |
| "learning_rate": 3.16370022285456e-05, | |
| "loss": 3.1368, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "learning_rate": 3.0164835160286395e-05, | |
| "loss": 3.106, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "learning_rate": 2.8880740270466042e-05, | |
| "loss": 3.0851, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 0.76, | |
| "learning_rate": 2.774780664821333e-05, | |
| "loss": 3.0639, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 0.82, | |
| "learning_rate": 2.6738528983830193e-05, | |
| "loss": 3.0412, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 0.87, | |
| "learning_rate": 2.583193464319257e-05, | |
| "loss": 3.0153, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 0.93, | |
| "learning_rate": 2.5011715272671623e-05, | |
| "loss": 2.9905, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 0.99, | |
| "learning_rate": 2.426497258953522e-05, | |
| "loss": 2.9887, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_gen_len": 20.0, | |
| "eval_loss": 3.2666966915130615, | |
| "eval_rouge1": 43.6697, | |
| "eval_rouge2": 28.3814, | |
| "eval_rougeL": 40.2632, | |
| "eval_rougeLsum": 40.2565, | |
| "eval_runtime": 2924.633, | |
| "eval_samples_per_second": 31.314, | |
| "eval_steps_per_second": 1.957, | |
| "step": 17170 | |
| }, | |
| { | |
| "epoch": 1.05, | |
| "learning_rate": 2.3581353179274502e-05, | |
| "loss": 2.7893, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 1.11, | |
| "learning_rate": 2.2952437303925296e-05, | |
| "loss": 2.724, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 1.16, | |
| "learning_rate": 2.2371298045736806e-05, | |
| "loss": 2.722, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 1.22, | |
| "learning_rate": 2.183217738419129e-05, | |
| "loss": 2.7157, | |
| "step": 21000 | |
| }, | |
| { | |
| "epoch": 1.28, | |
| "learning_rate": 2.1330244234932196e-05, | |
| "loss": 2.7097, | |
| "step": 22000 | |
| }, | |
| { | |
| "epoch": 1.34, | |
| "learning_rate": 2.0861411017276734e-05, | |
| "loss": 2.6939, | |
| "step": 23000 | |
| }, | |
| { | |
| "epoch": 1.4, | |
| "learning_rate": 2.0422192722261335e-05, | |
| "loss": 2.7026, | |
| "step": 24000 | |
| }, | |
| { | |
| "epoch": 1.46, | |
| "learning_rate": 2.0009597313304793e-05, | |
| "loss": 2.6871, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 1.51, | |
| "learning_rate": 1.9621039546293084e-05, | |
| "loss": 2.6776, | |
| "step": 26000 | |
| }, | |
| { | |
| "epoch": 1.57, | |
| "learning_rate": 1.925427251574924e-05, | |
| "loss": 2.6799, | |
| "step": 27000 | |
| }, | |
| { | |
| "epoch": 1.63, | |
| "learning_rate": 1.8907332773261164e-05, | |
| "loss": 2.6613, | |
| "step": 28000 | |
| }, | |
| { | |
| "epoch": 1.69, | |
| "learning_rate": 1.8578495948447844e-05, | |
| "loss": 2.6756, | |
| "step": 29000 | |
| }, | |
| { | |
| "epoch": 1.75, | |
| "learning_rate": 1.8266240576954328e-05, | |
| "loss": 2.6459, | |
| "step": 30000 | |
| }, | |
| { | |
| "epoch": 1.81, | |
| "learning_rate": 1.796921840006843e-05, | |
| "loss": 2.6352, | |
| "step": 31000 | |
| }, | |
| { | |
| "epoch": 1.86, | |
| "learning_rate": 1.768622981064213e-05, | |
| "loss": 2.6379, | |
| "step": 32000 | |
| }, | |
| { | |
| "epoch": 1.92, | |
| "learning_rate": 1.7416203423613194e-05, | |
| "loss": 2.6138, | |
| "step": 33000 | |
| }, | |
| { | |
| "epoch": 1.98, | |
| "learning_rate": 1.7158178976540783e-05, | |
| "loss": 2.6193, | |
| "step": 34000 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_gen_len": 20.0, | |
| "eval_loss": 3.2624921798706055, | |
| "eval_rouge1": 44.7484, | |
| "eval_rouge2": 29.1819, | |
| "eval_rougeL": 41.2553, | |
| "eval_rougeLsum": 41.2618, | |
| "eval_runtime": 2925.4113, | |
| "eval_samples_per_second": 31.306, | |
| "eval_steps_per_second": 1.957, | |
| "step": 34340 | |
| }, | |
| { | |
| "epoch": 2.04, | |
| "learning_rate": 1.6911292937114278e-05, | |
| "loss": 2.5111, | |
| "step": 35000 | |
| }, | |
| { | |
| "epoch": 2.1, | |
| "learning_rate": 1.6674766325339214e-05, | |
| "loss": 2.4361, | |
| "step": 36000 | |
| }, | |
| { | |
| "epoch": 2.15, | |
| "learning_rate": 1.644789435859962e-05, | |
| "loss": 2.4523, | |
| "step": 37000 | |
| }, | |
| { | |
| "epoch": 2.21, | |
| "learning_rate": 1.6230037605656824e-05, | |
| "loss": 2.4485, | |
| "step": 38000 | |
| }, | |
| { | |
| "epoch": 2.27, | |
| "learning_rate": 1.6020614396421562e-05, | |
| "loss": 2.443, | |
| "step": 39000 | |
| }, | |
| { | |
| "epoch": 2.33, | |
| "learning_rate": 1.58190942821137e-05, | |
| "loss": 2.4408, | |
| "step": 40000 | |
| }, | |
| { | |
| "epoch": 2.39, | |
| "learning_rate": 1.562499237823282e-05, | |
| "loss": 2.4376, | |
| "step": 41000 | |
| }, | |
| { | |
| "epoch": 2.45, | |
| "learning_rate": 1.5437864452872526e-05, | |
| "loss": 2.4237, | |
| "step": 42000 | |
| }, | |
| { | |
| "epoch": 2.5, | |
| "learning_rate": 1.5257302647033036e-05, | |
| "loss": 2.4302, | |
| "step": 43000 | |
| }, | |
| { | |
| "epoch": 2.56, | |
| "learning_rate": 1.508293173302068e-05, | |
| "loss": 2.4272, | |
| "step": 44000 | |
| }, | |
| { | |
| "epoch": 2.62, | |
| "learning_rate": 1.4914405832764997e-05, | |
| "loss": 2.4191, | |
| "step": 45000 | |
| }, | |
| { | |
| "epoch": 2.68, | |
| "learning_rate": 1.4751405530700888e-05, | |
| "loss": 2.415, | |
| "step": 46000 | |
| }, | |
| { | |
| "epoch": 2.74, | |
| "learning_rate": 1.4593635326349197e-05, | |
| "loss": 2.4139, | |
| "step": 47000 | |
| }, | |
| { | |
| "epoch": 2.8, | |
| "learning_rate": 1.4440821380348212e-05, | |
| "loss": 2.409, | |
| "step": 48000 | |
| }, | |
| { | |
| "epoch": 2.85, | |
| "learning_rate": 1.4292709514804612e-05, | |
| "loss": 2.4043, | |
| "step": 49000 | |
| }, | |
| { | |
| "epoch": 2.91, | |
| "learning_rate": 1.4149063434732295e-05, | |
| "loss": 2.3984, | |
| "step": 50000 | |
| }, | |
| { | |
| "epoch": 2.97, | |
| "learning_rate": 1.4009663142259388e-05, | |
| "loss": 2.4011, | |
| "step": 51000 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_gen_len": 20.0, | |
| "eval_loss": 3.302915573120117, | |
| "eval_rouge1": 44.3235, | |
| "eval_rouge2": 28.8361, | |
| "eval_rougeL": 40.7694, | |
| "eval_rougeLsum": 40.7674, | |
| "eval_runtime": 2928.7847, | |
| "eval_samples_per_second": 31.27, | |
| "eval_steps_per_second": 1.954, | |
| "step": 51510 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "step": 51510, | |
| "total_flos": 3.55488261007809e+17, | |
| "train_loss": 2.7859744007345357, | |
| "train_runtime": 40305.1206, | |
| "train_samples_per_second": 20.449, | |
| "train_steps_per_second": 1.278 | |
| } | |
| ], | |
| "max_steps": 51510, | |
| "num_train_epochs": 3, | |
| "total_flos": 3.55488261007809e+17, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |