| { | |
| "best_metric": 16.6036, | |
| "best_model_checkpoint": "./zh-kr_mid/checkpoint-12000", | |
| "epoch": 11.976047904191617, | |
| "eval_steps": 1000, | |
| "global_step": 16000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.37, | |
| "learning_rate": 4.91e-05, | |
| "loss": 4.4024, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "learning_rate": 4.95362674726105e-05, | |
| "loss": 2.7248, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "eval_bleu": 3.2381, | |
| "eval_gen_len": 48.6095, | |
| "eval_loss": 1.9409534931182861, | |
| "eval_runtime": 1917.4714, | |
| "eval_samples_per_second": 1.393, | |
| "eval_steps_per_second": 0.087, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.12, | |
| "learning_rate": 4.9064979221760484e-05, | |
| "loss": 1.8483, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 1.5, | |
| "learning_rate": 4.85927465054779e-05, | |
| "loss": 1.5683, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 1.5, | |
| "eval_bleu": 10.2345, | |
| "eval_gen_len": 20.4433, | |
| "eval_loss": 1.6889078617095947, | |
| "eval_runtime": 497.5454, | |
| "eval_samples_per_second": 5.368, | |
| "eval_steps_per_second": 0.336, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 1.87, | |
| "learning_rate": 4.812051378919532e-05, | |
| "loss": 1.5205, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 2.25, | |
| "learning_rate": 4.764828107291273e-05, | |
| "loss": 1.1916, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 2.25, | |
| "eval_bleu": 13.4571, | |
| "eval_gen_len": 18.8854, | |
| "eval_loss": 1.6842907667160034, | |
| "eval_runtime": 538.3261, | |
| "eval_samples_per_second": 4.962, | |
| "eval_steps_per_second": 0.31, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 2.62, | |
| "learning_rate": 4.7176048356630146e-05, | |
| "loss": 1.0458, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 2.99, | |
| "learning_rate": 4.670381564034756e-05, | |
| "loss": 1.068, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 2.99, | |
| "eval_bleu": 15.6862, | |
| "eval_gen_len": 15.5054, | |
| "eval_loss": 1.6389988660812378, | |
| "eval_runtime": 200.5304, | |
| "eval_samples_per_second": 13.32, | |
| "eval_steps_per_second": 0.833, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 3.37, | |
| "learning_rate": 4.623158292406498e-05, | |
| "loss": 0.7097, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 3.74, | |
| "learning_rate": 4.57593502077824e-05, | |
| "loss": 0.7313, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 3.74, | |
| "eval_bleu": 15.2014, | |
| "eval_gen_len": 16.5938, | |
| "eval_loss": 1.7002748250961304, | |
| "eval_runtime": 222.4222, | |
| "eval_samples_per_second": 12.009, | |
| "eval_steps_per_second": 0.751, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 4.12, | |
| "learning_rate": 4.5287117491499814e-05, | |
| "loss": 0.6587, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 4.49, | |
| "learning_rate": 4.481488477521723e-05, | |
| "loss": 0.4832, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 4.49, | |
| "eval_bleu": 15.0381, | |
| "eval_gen_len": 16.9068, | |
| "eval_loss": 1.8982346057891846, | |
| "eval_runtime": 258.8867, | |
| "eval_samples_per_second": 10.317, | |
| "eval_steps_per_second": 0.645, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 4.87, | |
| "learning_rate": 4.434265205893465e-05, | |
| "loss": 0.5089, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 5.24, | |
| "learning_rate": 4.3870419342652066e-05, | |
| "loss": 0.3862, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 5.24, | |
| "eval_bleu": 15.5397, | |
| "eval_gen_len": 15.6451, | |
| "eval_loss": 2.1426022052764893, | |
| "eval_runtime": 170.9728, | |
| "eval_samples_per_second": 15.622, | |
| "eval_steps_per_second": 0.977, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 5.61, | |
| "learning_rate": 4.339913109180204e-05, | |
| "loss": 0.3495, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 5.99, | |
| "learning_rate": 4.292689837551946e-05, | |
| "loss": 0.3675, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 5.99, | |
| "eval_bleu": 15.8847, | |
| "eval_gen_len": 15.6926, | |
| "eval_loss": 2.1168086528778076, | |
| "eval_runtime": 169.7878, | |
| "eval_samples_per_second": 15.731, | |
| "eval_steps_per_second": 0.984, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 6.36, | |
| "learning_rate": 4.2454665659236876e-05, | |
| "loss": 0.2443, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 6.74, | |
| "learning_rate": 4.198243294295429e-05, | |
| "loss": 0.2627, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 6.74, | |
| "eval_bleu": 16.3603, | |
| "eval_gen_len": 15.9671, | |
| "eval_loss": 2.260291814804077, | |
| "eval_runtime": 176.4048, | |
| "eval_samples_per_second": 15.141, | |
| "eval_steps_per_second": 0.947, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 7.11, | |
| "learning_rate": 4.151020022667171e-05, | |
| "loss": 0.2471, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 7.49, | |
| "learning_rate": 4.103796751038912e-05, | |
| "loss": 0.1955, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 7.49, | |
| "eval_bleu": 15.7447, | |
| "eval_gen_len": 15.979, | |
| "eval_loss": 2.4113738536834717, | |
| "eval_runtime": 200.4127, | |
| "eval_samples_per_second": 13.327, | |
| "eval_steps_per_second": 0.833, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 7.86, | |
| "learning_rate": 4.056573479410654e-05, | |
| "loss": 0.2088, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 8.23, | |
| "learning_rate": 4.0093502077823954e-05, | |
| "loss": 0.171, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 8.23, | |
| "eval_bleu": 15.7852, | |
| "eval_gen_len": 15.9244, | |
| "eval_loss": 2.5141146183013916, | |
| "eval_runtime": 180.5467, | |
| "eval_samples_per_second": 14.794, | |
| "eval_steps_per_second": 0.925, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 8.61, | |
| "learning_rate": 3.962126936154137e-05, | |
| "loss": 0.157, | |
| "step": 11500 | |
| }, | |
| { | |
| "epoch": 8.98, | |
| "learning_rate": 3.9149981110691354e-05, | |
| "loss": 0.1702, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 8.98, | |
| "eval_bleu": 16.6036, | |
| "eval_gen_len": 15.4901, | |
| "eval_loss": 2.5557188987731934, | |
| "eval_runtime": 168.9011, | |
| "eval_samples_per_second": 15.814, | |
| "eval_steps_per_second": 0.989, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 9.36, | |
| "learning_rate": 3.8677748394408764e-05, | |
| "loss": 0.1206, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 9.73, | |
| "learning_rate": 3.820551567812618e-05, | |
| "loss": 0.1298, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 9.73, | |
| "eval_bleu": 16.1319, | |
| "eval_gen_len": 15.5492, | |
| "eval_loss": 2.653599739074707, | |
| "eval_runtime": 168.2024, | |
| "eval_samples_per_second": 15.88, | |
| "eval_steps_per_second": 0.993, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 10.1, | |
| "learning_rate": 3.77332829618436e-05, | |
| "loss": 0.1263, | |
| "step": 13500 | |
| }, | |
| { | |
| "epoch": 10.48, | |
| "learning_rate": 3.7261050245561016e-05, | |
| "loss": 0.1052, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 10.48, | |
| "eval_bleu": 16.1807, | |
| "eval_gen_len": 15.8884, | |
| "eval_loss": 2.7585606575012207, | |
| "eval_runtime": 171.7593, | |
| "eval_samples_per_second": 15.551, | |
| "eval_steps_per_second": 0.972, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 10.85, | |
| "learning_rate": 3.678881752927843e-05, | |
| "loss": 0.1137, | |
| "step": 14500 | |
| }, | |
| { | |
| "epoch": 11.23, | |
| "learning_rate": 3.6320362674726105e-05, | |
| "loss": 0.2268, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 11.23, | |
| "eval_bleu": 15.1752, | |
| "eval_gen_len": 15.5346, | |
| "eval_loss": 2.7257988452911377, | |
| "eval_runtime": 171.5928, | |
| "eval_samples_per_second": 15.566, | |
| "eval_steps_per_second": 0.973, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 11.6, | |
| "learning_rate": 3.584812995844352e-05, | |
| "loss": 0.1133, | |
| "step": 15500 | |
| }, | |
| { | |
| "epoch": 11.98, | |
| "learning_rate": 3.5376841707593505e-05, | |
| "loss": 0.1327, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 11.98, | |
| "eval_bleu": 15.8563, | |
| "eval_gen_len": 15.7971, | |
| "eval_loss": 2.719254493713379, | |
| "eval_runtime": 171.7089, | |
| "eval_samples_per_second": 15.555, | |
| "eval_steps_per_second": 0.973, | |
| "step": 16000 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 53440, | |
| "num_train_epochs": 40, | |
| "save_steps": 1000, | |
| "total_flos": 5.54785220591616e+17, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |