| { | |
| "best_metric": 26.3346, | |
| "best_model_checkpoint": "/local1/hfs/gs_stuff/ft-wmt14/checkpoint-100000", | |
| "epoch": 0.354862233609357, | |
| "eval_steps": 10000, | |
| "global_step": 100000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.01774311168046785, | |
| "grad_norm": 1.5128649473190308, | |
| "learning_rate": 0.000475, | |
| "loss": 2.4524, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.0354862233609357, | |
| "grad_norm": 1.242777705192566, | |
| "learning_rate": 0.00045000000000000004, | |
| "loss": 2.3103, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 0.0354862233609357, | |
| "eval_bleu": 22.058, | |
| "eval_gen_len": 27.7263, | |
| "eval_loss": 1.8454290628433228, | |
| "eval_runtime": 238.9627, | |
| "eval_samples_per_second": 12.554, | |
| "eval_steps_per_second": 1.569, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 0.05322933504140355, | |
| "grad_norm": 0.988516628742218, | |
| "learning_rate": 0.000425, | |
| "loss": 2.2594, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 0.0709724467218714, | |
| "grad_norm": 1.4387503862380981, | |
| "learning_rate": 0.0004, | |
| "loss": 2.2141, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 0.0709724467218714, | |
| "eval_bleu": 23.339, | |
| "eval_gen_len": 26.7147, | |
| "eval_loss": 1.7811188697814941, | |
| "eval_runtime": 225.1917, | |
| "eval_samples_per_second": 13.322, | |
| "eval_steps_per_second": 1.665, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 0.08871555840233925, | |
| "grad_norm": 1.2291666269302368, | |
| "learning_rate": 0.000375, | |
| "loss": 2.1898, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 0.1064586700828071, | |
| "grad_norm": 1.1732761859893799, | |
| "learning_rate": 0.00035, | |
| "loss": 2.176, | |
| "step": 30000 | |
| }, | |
| { | |
| "epoch": 0.1064586700828071, | |
| "eval_bleu": 24.3234, | |
| "eval_gen_len": 27.125, | |
| "eval_loss": 1.7360602617263794, | |
| "eval_runtime": 227.357, | |
| "eval_samples_per_second": 13.195, | |
| "eval_steps_per_second": 1.649, | |
| "step": 30000 | |
| }, | |
| { | |
| "epoch": 0.12420178176327495, | |
| "grad_norm": 1.187321662902832, | |
| "learning_rate": 0.00032500000000000004, | |
| "loss": 2.1468, | |
| "step": 35000 | |
| }, | |
| { | |
| "epoch": 0.1419448934437428, | |
| "grad_norm": 1.3599053621292114, | |
| "learning_rate": 0.0003, | |
| "loss": 2.139, | |
| "step": 40000 | |
| }, | |
| { | |
| "epoch": 0.1419448934437428, | |
| "eval_bleu": 25.0888, | |
| "eval_gen_len": 26.8213, | |
| "eval_loss": 1.7130982875823975, | |
| "eval_runtime": 221.7983, | |
| "eval_samples_per_second": 13.526, | |
| "eval_steps_per_second": 1.691, | |
| "step": 40000 | |
| }, | |
| { | |
| "epoch": 0.15968800512421066, | |
| "grad_norm": 1.4392811059951782, | |
| "learning_rate": 0.000275, | |
| "loss": 2.1151, | |
| "step": 45000 | |
| }, | |
| { | |
| "epoch": 0.1774311168046785, | |
| "grad_norm": 1.4162044525146484, | |
| "learning_rate": 0.00025, | |
| "loss": 2.1084, | |
| "step": 50000 | |
| }, | |
| { | |
| "epoch": 0.1774311168046785, | |
| "eval_bleu": 24.9992, | |
| "eval_gen_len": 26.824, | |
| "eval_loss": 1.687427043914795, | |
| "eval_runtime": 224.0057, | |
| "eval_samples_per_second": 13.393, | |
| "eval_steps_per_second": 1.674, | |
| "step": 50000 | |
| }, | |
| { | |
| "epoch": 0.19517422848514634, | |
| "grad_norm": 1.2046048641204834, | |
| "learning_rate": 0.00022500000000000002, | |
| "loss": 2.0914, | |
| "step": 55000 | |
| }, | |
| { | |
| "epoch": 0.2129173401656142, | |
| "grad_norm": 1.2651879787445068, | |
| "learning_rate": 0.0002, | |
| "loss": 2.0826, | |
| "step": 60000 | |
| }, | |
| { | |
| "epoch": 0.2129173401656142, | |
| "eval_bleu": 25.7297, | |
| "eval_gen_len": 26.62, | |
| "eval_loss": 1.6685482263565063, | |
| "eval_runtime": 221.6914, | |
| "eval_samples_per_second": 13.532, | |
| "eval_steps_per_second": 1.692, | |
| "step": 60000 | |
| }, | |
| { | |
| "epoch": 0.23066045184608205, | |
| "grad_norm": 1.212643027305603, | |
| "learning_rate": 0.000175, | |
| "loss": 2.0778, | |
| "step": 65000 | |
| }, | |
| { | |
| "epoch": 0.2484035635265499, | |
| "grad_norm": 1.2400418519973755, | |
| "learning_rate": 0.00015, | |
| "loss": 2.068, | |
| "step": 70000 | |
| }, | |
| { | |
| "epoch": 0.2484035635265499, | |
| "eval_bleu": 25.9031, | |
| "eval_gen_len": 26.685, | |
| "eval_loss": 1.648539662361145, | |
| "eval_runtime": 223.211, | |
| "eval_samples_per_second": 13.44, | |
| "eval_steps_per_second": 1.68, | |
| "step": 70000 | |
| }, | |
| { | |
| "epoch": 0.26614667520701774, | |
| "grad_norm": 1.3389995098114014, | |
| "learning_rate": 0.000125, | |
| "loss": 2.0566, | |
| "step": 75000 | |
| }, | |
| { | |
| "epoch": 0.2838897868874856, | |
| "grad_norm": 1.1512677669525146, | |
| "learning_rate": 0.0001, | |
| "loss": 2.05, | |
| "step": 80000 | |
| }, | |
| { | |
| "epoch": 0.2838897868874856, | |
| "eval_bleu": 26.143, | |
| "eval_gen_len": 26.8693, | |
| "eval_loss": 1.6370748281478882, | |
| "eval_runtime": 225.2245, | |
| "eval_samples_per_second": 13.32, | |
| "eval_steps_per_second": 1.665, | |
| "step": 80000 | |
| }, | |
| { | |
| "epoch": 0.30163289856795344, | |
| "grad_norm": 1.1607016324996948, | |
| "learning_rate": 7.5e-05, | |
| "loss": 2.0235, | |
| "step": 85000 | |
| }, | |
| { | |
| "epoch": 0.3193760102484213, | |
| "grad_norm": 1.2967106103897095, | |
| "learning_rate": 5e-05, | |
| "loss": 2.0331, | |
| "step": 90000 | |
| }, | |
| { | |
| "epoch": 0.3193760102484213, | |
| "eval_bleu": 26.3038, | |
| "eval_gen_len": 26.5183, | |
| "eval_loss": 1.6311123371124268, | |
| "eval_runtime": 219.2546, | |
| "eval_samples_per_second": 13.683, | |
| "eval_steps_per_second": 1.71, | |
| "step": 90000 | |
| }, | |
| { | |
| "epoch": 0.33711912192888915, | |
| "grad_norm": 1.2956724166870117, | |
| "learning_rate": 2.5e-05, | |
| "loss": 2.0346, | |
| "step": 95000 | |
| }, | |
| { | |
| "epoch": 0.354862233609357, | |
| "grad_norm": 1.1822398900985718, | |
| "learning_rate": 0.0, | |
| "loss": 2.0273, | |
| "step": 100000 | |
| }, | |
| { | |
| "epoch": 0.354862233609357, | |
| "eval_bleu": 26.3346, | |
| "eval_gen_len": 26.6907, | |
| "eval_loss": 1.6253596544265747, | |
| "eval_runtime": 221.5579, | |
| "eval_samples_per_second": 13.54, | |
| "eval_steps_per_second": 1.693, | |
| "step": 100000 | |
| }, | |
| { | |
| "epoch": 0.354862233609357, | |
| "step": 100000, | |
| "total_flos": 3.589813132276531e+16, | |
| "train_loss": 2.132816611328125, | |
| "train_runtime": 15981.9131, | |
| "train_samples_per_second": 100.113, | |
| "train_steps_per_second": 6.257 | |
| } | |
| ], | |
| "logging_steps": 5000, | |
| "max_steps": 100000, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 10000, | |
| "total_flos": 3.589813132276531e+16, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |