| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 10.0, | |
| "eval_steps": 500, | |
| "global_step": 3415, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.29, | |
| "learning_rate": 3e-05, | |
| "loss": 1.574, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "learning_rate": 3e-05, | |
| "loss": 1.474, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "learning_rate": 3e-05, | |
| "loss": 1.4104, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.4536923076923077, | |
| "eval_loss": 3.3574578762054443, | |
| "eval_runtime": 4.4299, | |
| "eval_samples_per_second": 112.869, | |
| "eval_steps_per_second": 14.221, | |
| "step": 341 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_bleu": 0.08840557310176213, | |
| "eval_exact_match": 0.0, | |
| "eval_prefix_exact_match": 0.058, | |
| "step": 341 | |
| }, | |
| { | |
| "epoch": 1.17, | |
| "learning_rate": 3e-05, | |
| "loss": 1.3889, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.46, | |
| "learning_rate": 3e-05, | |
| "loss": 1.387, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.76, | |
| "learning_rate": 3e-05, | |
| "loss": 1.389, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.4543589743589744, | |
| "eval_loss": 3.4179794788360596, | |
| "eval_runtime": 4.1997, | |
| "eval_samples_per_second": 119.055, | |
| "eval_steps_per_second": 15.001, | |
| "step": 683 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_bleu": 0.09820569565497561, | |
| "eval_exact_match": 0.0, | |
| "eval_prefix_exact_match": 0.056, | |
| "step": 683 | |
| }, | |
| { | |
| "epoch": 2.05, | |
| "learning_rate": 3e-05, | |
| "loss": 1.3725, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 2.34, | |
| "learning_rate": 3e-05, | |
| "loss": 1.3588, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 2.64, | |
| "learning_rate": 3e-05, | |
| "loss": 1.348, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 2.93, | |
| "learning_rate": 3e-05, | |
| "loss": 1.3414, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.45476923076923076, | |
| "eval_loss": 3.511923313140869, | |
| "eval_runtime": 4.1985, | |
| "eval_samples_per_second": 119.089, | |
| "eval_steps_per_second": 15.005, | |
| "step": 1024 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_bleu": 0.09889831624589397, | |
| "eval_exact_match": 0.0, | |
| "eval_prefix_exact_match": 0.056, | |
| "step": 1024 | |
| }, | |
| { | |
| "epoch": 3.22, | |
| "learning_rate": 3e-05, | |
| "loss": 1.3166, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 3.51, | |
| "learning_rate": 3e-05, | |
| "loss": 1.3035, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 3.81, | |
| "learning_rate": 3e-05, | |
| "loss": 1.3002, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.45543589743589746, | |
| "eval_loss": 3.5288071632385254, | |
| "eval_runtime": 4.4066, | |
| "eval_samples_per_second": 113.466, | |
| "eval_steps_per_second": 14.297, | |
| "step": 1366 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_bleu": 0.11209345982903633, | |
| "eval_exact_match": 0.0, | |
| "eval_prefix_exact_match": 0.054, | |
| "step": 1366 | |
| }, | |
| { | |
| "epoch": 4.1, | |
| "learning_rate": 3e-05, | |
| "loss": 1.2929, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 4.39, | |
| "learning_rate": 3e-05, | |
| "loss": 1.2693, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 4.69, | |
| "learning_rate": 3e-05, | |
| "loss": 1.2697, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 4.98, | |
| "learning_rate": 3e-05, | |
| "loss": 1.2574, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.45394871794871794, | |
| "eval_loss": 3.6893365383148193, | |
| "eval_runtime": 4.1122, | |
| "eval_samples_per_second": 121.591, | |
| "eval_steps_per_second": 15.32, | |
| "step": 1707 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_bleu": 0.12076940663377221, | |
| "eval_exact_match": 0.0, | |
| "eval_prefix_exact_match": 0.056, | |
| "step": 1707 | |
| }, | |
| { | |
| "epoch": 5.27, | |
| "learning_rate": 3e-05, | |
| "loss": 1.2097, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 5.56, | |
| "learning_rate": 3e-05, | |
| "loss": 1.2188, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 5.86, | |
| "learning_rate": 3e-05, | |
| "loss": 1.2258, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_accuracy": 0.45615384615384613, | |
| "eval_loss": 3.725893497467041, | |
| "eval_runtime": 4.0989, | |
| "eval_samples_per_second": 121.984, | |
| "eval_steps_per_second": 15.37, | |
| "step": 2049 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_bleu": 0.11884272898137646, | |
| "eval_exact_match": 0.0, | |
| "eval_prefix_exact_match": 0.058, | |
| "step": 2049 | |
| }, | |
| { | |
| "epoch": 6.15, | |
| "learning_rate": 3e-05, | |
| "loss": 1.205, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 6.44, | |
| "learning_rate": 3e-05, | |
| "loss": 1.1643, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 6.73, | |
| "learning_rate": 3e-05, | |
| "loss": 1.1844, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_accuracy": 0.45594871794871794, | |
| "eval_loss": 3.724449396133423, | |
| "eval_runtime": 4.8158, | |
| "eval_samples_per_second": 103.825, | |
| "eval_steps_per_second": 13.082, | |
| "step": 2390 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_bleu": 0.12407896336995003, | |
| "eval_exact_match": 0.0, | |
| "eval_prefix_exact_match": 0.048, | |
| "step": 2390 | |
| }, | |
| { | |
| "epoch": 7.03, | |
| "learning_rate": 3e-05, | |
| "loss": 1.1771, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 7.32, | |
| "learning_rate": 3e-05, | |
| "loss": 1.1218, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 7.61, | |
| "learning_rate": 3e-05, | |
| "loss": 1.1491, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 7.91, | |
| "learning_rate": 3e-05, | |
| "loss": 1.1363, | |
| "step": 2700 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_accuracy": 0.4543589743589744, | |
| "eval_loss": 3.8138701915740967, | |
| "eval_runtime": 4.1963, | |
| "eval_samples_per_second": 119.152, | |
| "eval_steps_per_second": 15.013, | |
| "step": 2732 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_bleu": 0.11947629327387163, | |
| "eval_exact_match": 0.0, | |
| "eval_prefix_exact_match": 0.046, | |
| "step": 2732 | |
| }, | |
| { | |
| "epoch": 8.2, | |
| "learning_rate": 3e-05, | |
| "loss": 1.0858, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 8.49, | |
| "learning_rate": 3e-05, | |
| "loss": 1.0951, | |
| "step": 2900 | |
| }, | |
| { | |
| "epoch": 8.78, | |
| "learning_rate": 3e-05, | |
| "loss": 1.0903, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_accuracy": 0.4524102564102564, | |
| "eval_loss": 3.9115548133850098, | |
| "eval_runtime": 4.4011, | |
| "eval_samples_per_second": 113.609, | |
| "eval_steps_per_second": 14.315, | |
| "step": 3073 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_bleu": 0.13426062809240774, | |
| "eval_exact_match": 0.0, | |
| "eval_prefix_exact_match": 0.034, | |
| "step": 3073 | |
| }, | |
| { | |
| "epoch": 9.08, | |
| "learning_rate": 3e-05, | |
| "loss": 1.0822, | |
| "step": 3100 | |
| }, | |
| { | |
| "epoch": 9.37, | |
| "learning_rate": 3e-05, | |
| "loss": 1.0414, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 9.66, | |
| "learning_rate": 3e-05, | |
| "loss": 1.0563, | |
| "step": 3300 | |
| }, | |
| { | |
| "epoch": 9.96, | |
| "learning_rate": 3e-05, | |
| "loss": 1.0538, | |
| "step": 3400 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_accuracy": 0.4515897435897436, | |
| "eval_loss": 3.92203426361084, | |
| "eval_runtime": 4.4105, | |
| "eval_samples_per_second": 113.366, | |
| "eval_steps_per_second": 14.284, | |
| "step": 3415 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_bleu": 0.11780869016319889, | |
| "eval_exact_match": 0.0, | |
| "eval_prefix_exact_match": 0.032, | |
| "step": 3415 | |
| } | |
| ], | |
| "logging_steps": 100, | |
| "max_steps": 17050, | |
| "num_train_epochs": 50, | |
| "save_steps": 500, | |
| "total_flos": 6.763068176298148e+17, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |