| { | |
| "best_metric": 12.0444, | |
| "best_model_checkpoint": "./tst-translation-output5/checkpoint-6000", | |
| "epoch": 5.143591941705958, | |
| "eval_steps": 2000, | |
| "global_step": 12000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.21, | |
| "learning_rate": 1.6566666666666665e-05, | |
| "loss": 2.2544, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.43, | |
| "learning_rate": 3.323333333333333e-05, | |
| "loss": 1.5518, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "learning_rate": 4.99e-05, | |
| "loss": 1.4716, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.86, | |
| "learning_rate": 4.9729361794815946e-05, | |
| "loss": 1.3822, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.86, | |
| "eval_bleu": 8.4562, | |
| "eval_gen_len": 15.8491, | |
| "eval_loss": 1.369325041770935, | |
| "eval_runtime": 299.4199, | |
| "eval_samples_per_second": 15.58, | |
| "eval_steps_per_second": 0.975, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 1.07, | |
| "learning_rate": 4.9457089958614686e-05, | |
| "loss": 1.2265, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 1.29, | |
| "learning_rate": 4.918481812241342e-05, | |
| "loss": 1.0251, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 1.5, | |
| "learning_rate": 4.891254628621216e-05, | |
| "loss": 1.0351, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 1.71, | |
| "learning_rate": 4.864027445001089e-05, | |
| "loss": 1.0081, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 1.71, | |
| "eval_bleu": 11.3956, | |
| "eval_gen_len": 15.239, | |
| "eval_loss": 1.2806029319763184, | |
| "eval_runtime": 274.3132, | |
| "eval_samples_per_second": 17.006, | |
| "eval_steps_per_second": 1.064, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 1.93, | |
| "learning_rate": 4.836854715748203e-05, | |
| "loss": 1.0022, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 2.14, | |
| "learning_rate": 4.809627532128077e-05, | |
| "loss": 0.774, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 2.36, | |
| "learning_rate": 4.78240034850795e-05, | |
| "loss": 0.673, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 2.57, | |
| "learning_rate": 4.755227619255064e-05, | |
| "loss": 0.7007, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 2.57, | |
| "eval_bleu": 12.0444, | |
| "eval_gen_len": 15.34, | |
| "eval_loss": 1.3149274587631226, | |
| "eval_runtime": 275.621, | |
| "eval_samples_per_second": 16.925, | |
| "eval_steps_per_second": 1.059, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 2.79, | |
| "learning_rate": 4.728000435634938e-05, | |
| "loss": 0.7025, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "learning_rate": 4.7007732520148115e-05, | |
| "loss": 0.7042, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 3.21, | |
| "learning_rate": 4.6735460683946854e-05, | |
| "loss": 0.4363, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 3.43, | |
| "learning_rate": 4.6463733391417995e-05, | |
| "loss": 0.4483, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 3.43, | |
| "eval_bleu": 11.5647, | |
| "eval_gen_len": 15.2139, | |
| "eval_loss": 1.417746901512146, | |
| "eval_runtime": 266.1794, | |
| "eval_samples_per_second": 17.526, | |
| "eval_steps_per_second": 1.097, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 3.64, | |
| "learning_rate": 4.6191461555216734e-05, | |
| "loss": 0.4609, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 3.86, | |
| "learning_rate": 4.5919189719015466e-05, | |
| "loss": 0.4811, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 4.07, | |
| "learning_rate": 4.564800697015901e-05, | |
| "loss": 0.4172, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 4.29, | |
| "learning_rate": 4.537573513395775e-05, | |
| "loss": 0.2837, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 4.29, | |
| "eval_bleu": 11.1306, | |
| "eval_gen_len": 15.2984, | |
| "eval_loss": 1.5126218795776367, | |
| "eval_runtime": 263.8822, | |
| "eval_samples_per_second": 17.678, | |
| "eval_steps_per_second": 1.107, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 4.5, | |
| "learning_rate": 4.510346329775648e-05, | |
| "loss": 0.2955, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 4.71, | |
| "learning_rate": 4.483119146155522e-05, | |
| "loss": 0.3166, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 4.93, | |
| "learning_rate": 4.455891962535396e-05, | |
| "loss": 0.323, | |
| "step": 11500 | |
| }, | |
| { | |
| "epoch": 5.14, | |
| "learning_rate": 4.428664778915269e-05, | |
| "loss": 0.2276, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 5.14, | |
| "eval_bleu": 11.1467, | |
| "eval_gen_len": 15.2992, | |
| "eval_loss": 1.5938239097595215, | |
| "eval_runtime": 266.5302, | |
| "eval_samples_per_second": 17.503, | |
| "eval_steps_per_second": 1.096, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 5.14, | |
| "step": 12000, | |
| "total_flos": 4.16088915443712e+17, | |
| "train_loss": 0.8000754559834798, | |
| "train_runtime": 9676.4117, | |
| "train_samples_per_second": 154.289, | |
| "train_steps_per_second": 9.644 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 93320, | |
| "num_train_epochs": 40, | |
| "save_steps": 2000, | |
| "total_flos": 4.16088915443712e+17, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |