| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 10.0, | |
| "global_step": 68220, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.73, | |
| "learning_rate": 2.780123131046614e-05, | |
| "loss": 0.7441, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "eval_gen_len": 33.7, | |
| "eval_loss": 0.633007287979126, | |
| "eval_rouge1": 19.0253, | |
| "eval_rouge2": 11.6724, | |
| "eval_rougeL": 18.7632, | |
| "eval_rougeLsum": 18.8697, | |
| "eval_runtime": 1962.245, | |
| "eval_samples_per_second": 1.739, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 1.47, | |
| "learning_rate": 2.5602462620932278e-05, | |
| "loss": 0.5879, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 1.47, | |
| "eval_gen_len": 33.8, | |
| "eval_loss": 0.6442069411277771, | |
| "eval_rouge1": 18.9287, | |
| "eval_rouge2": 11.7109, | |
| "eval_rougeL": 18.6885, | |
| "eval_rougeLsum": 18.7824, | |
| "eval_runtime": 1627.9481, | |
| "eval_samples_per_second": 2.097, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 2.2, | |
| "learning_rate": 2.3403693931398417e-05, | |
| "loss": 0.5011, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 2.2, | |
| "eval_gen_len": 33.7, | |
| "eval_loss": 0.6601932048797607, | |
| "eval_rouge1": 19.157, | |
| "eval_rouge2": 11.7904, | |
| "eval_rougeL": 18.8919, | |
| "eval_rougeLsum": 18.9663, | |
| "eval_runtime": 1218.763, | |
| "eval_samples_per_second": 2.8, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 2.93, | |
| "learning_rate": 2.120492524186456e-05, | |
| "loss": 0.4365, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 2.93, | |
| "eval_gen_len": 33.8, | |
| "eval_loss": 0.6567361950874329, | |
| "eval_rouge1": 19.402, | |
| "eval_rouge2": 11.846, | |
| "eval_rougeL": 19.1085, | |
| "eval_rougeLsum": 19.1695, | |
| "eval_runtime": 1214.5476, | |
| "eval_samples_per_second": 2.81, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 3.66, | |
| "learning_rate": 1.9006156552330694e-05, | |
| "loss": 0.3731, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 3.66, | |
| "eval_gen_len": 33.5, | |
| "eval_loss": 0.7168177366256714, | |
| "eval_rouge1": 20.1185, | |
| "eval_rouge2": 12.3512, | |
| "eval_rougeL": 19.8009, | |
| "eval_rougeLsum": 19.9205, | |
| "eval_runtime": 1198.2455, | |
| "eval_samples_per_second": 2.848, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 4.4, | |
| "learning_rate": 1.6807387862796836e-05, | |
| "loss": 0.3377, | |
| "step": 30000 | |
| }, | |
| { | |
| "epoch": 4.4, | |
| "eval_gen_len": 33.4, | |
| "eval_loss": 0.7644200325012207, | |
| "eval_rouge1": 20.026, | |
| "eval_rouge2": 12.1819, | |
| "eval_rougeL": 19.7149, | |
| "eval_rougeLsum": 19.7857, | |
| "eval_runtime": 1196.6336, | |
| "eval_samples_per_second": 2.852, | |
| "step": 30000 | |
| }, | |
| { | |
| "epoch": 5.13, | |
| "learning_rate": 1.4608619173262973e-05, | |
| "loss": 0.3003, | |
| "step": 35000 | |
| }, | |
| { | |
| "epoch": 5.13, | |
| "eval_gen_len": 33.5, | |
| "eval_loss": 0.8117037415504456, | |
| "eval_rouge1": 20.0409, | |
| "eval_rouge2": 12.1197, | |
| "eval_rougeL": 19.737, | |
| "eval_rougeLsum": 19.8529, | |
| "eval_runtime": 1196.6657, | |
| "eval_samples_per_second": 2.852, | |
| "step": 35000 | |
| }, | |
| { | |
| "epoch": 5.86, | |
| "learning_rate": 1.2409850483729111e-05, | |
| "loss": 0.272, | |
| "step": 40000 | |
| }, | |
| { | |
| "epoch": 5.86, | |
| "eval_gen_len": 33.4, | |
| "eval_loss": 0.8414672613143921, | |
| "eval_rouge1": 20.3612, | |
| "eval_rouge2": 12.4433, | |
| "eval_rougeL": 20.0321, | |
| "eval_rougeLsum": 20.1456, | |
| "eval_runtime": 1188.9907, | |
| "eval_samples_per_second": 2.871, | |
| "step": 40000 | |
| }, | |
| { | |
| "epoch": 6.6, | |
| "learning_rate": 1.0211081794195251e-05, | |
| "loss": 0.2467, | |
| "step": 45000 | |
| }, | |
| { | |
| "epoch": 6.6, | |
| "eval_gen_len": 33.6, | |
| "eval_loss": 0.9067189693450928, | |
| "eval_rouge1": 19.9276, | |
| "eval_rouge2": 12.162, | |
| "eval_rougeL": 19.6218, | |
| "eval_rougeLsum": 19.7163, | |
| "eval_runtime": 1207.9436, | |
| "eval_samples_per_second": 2.825, | |
| "step": 45000 | |
| }, | |
| { | |
| "epoch": 7.33, | |
| "learning_rate": 8.01231310466139e-06, | |
| "loss": 0.2253, | |
| "step": 50000 | |
| }, | |
| { | |
| "epoch": 7.33, | |
| "eval_gen_len": 33.2, | |
| "eval_loss": 0.9487330317497253, | |
| "eval_rouge1": 20.7483, | |
| "eval_rouge2": 12.591, | |
| "eval_rougeL": 20.4197, | |
| "eval_rougeLsum": 20.5083, | |
| "eval_runtime": 1187.4731, | |
| "eval_samples_per_second": 2.874, | |
| "step": 50000 | |
| }, | |
| { | |
| "epoch": 8.06, | |
| "learning_rate": 5.813544415127529e-06, | |
| "loss": 0.2141, | |
| "step": 55000 | |
| }, | |
| { | |
| "epoch": 8.06, | |
| "eval_gen_len": 33.3, | |
| "eval_loss": 0.9719064235687256, | |
| "eval_rouge1": 20.2944, | |
| "eval_rouge2": 12.35, | |
| "eval_rougeL": 19.9563, | |
| "eval_rougeLsum": 20.0559, | |
| "eval_runtime": 1191.6755, | |
| "eval_samples_per_second": 2.864, | |
| "step": 55000 | |
| }, | |
| { | |
| "epoch": 8.8, | |
| "learning_rate": 3.6147757255936676e-06, | |
| "loss": 0.1951, | |
| "step": 60000 | |
| }, | |
| { | |
| "epoch": 8.8, | |
| "eval_gen_len": 33.5, | |
| "eval_loss": 0.9840695261955261, | |
| "eval_rouge1": 20.0284, | |
| "eval_rouge2": 12.0984, | |
| "eval_rougeL": 19.6653, | |
| "eval_rougeLsum": 19.7619, | |
| "eval_runtime": 1228.4982, | |
| "eval_samples_per_second": 2.778, | |
| "step": 60000 | |
| }, | |
| { | |
| "epoch": 9.53, | |
| "learning_rate": 1.4160070360598064e-06, | |
| "loss": 0.185, | |
| "step": 65000 | |
| }, | |
| { | |
| "epoch": 9.53, | |
| "eval_gen_len": 33.4, | |
| "eval_loss": 1.021308422088623, | |
| "eval_rouge1": 20.5124, | |
| "eval_rouge2": 12.4505, | |
| "eval_rougeL": 20.1615, | |
| "eval_rougeLsum": 20.2679, | |
| "eval_runtime": 1216.5148, | |
| "eval_samples_per_second": 2.806, | |
| "step": 65000 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "step": 68220, | |
| "total_flos": 1.7131748370415616e+17, | |
| "train_runtime": 36592.0856, | |
| "train_samples_per_second": 1.864 | |
| } | |
| ], | |
| "max_steps": 68220, | |
| "num_train_epochs": 10, | |
| "total_flos": 1.7131748370415616e+17, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |