| { | |
| "best_metric": NaN, | |
| "best_model_checkpoint": "/mnt/bn/qingyi-bn-lq/llama/saved_llama-7b-hf_TQA/checkpoint-200", | |
| "epoch": 2.7962861824139815, | |
| "global_step": 800, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.07, | |
| "learning_rate": 5.1e-05, | |
| "loss": 15278261862.4, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.14, | |
| "learning_rate": 0.00010799999999999998, | |
| "loss": 155534963507.2, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "learning_rate": 0.000168, | |
| "loss": 26388629094.4, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.28, | |
| "learning_rate": 0.00022799999999999999, | |
| "loss": 1740693708.8, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.35, | |
| "learning_rate": 0.00028799999999999995, | |
| "loss": 353188315136.0, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "learning_rate": 0.0002936675461741425, | |
| "loss": 15921389568.0, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.49, | |
| "learning_rate": 0.00028575197889182057, | |
| "loss": 8279447142.4, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.56, | |
| "learning_rate": 0.00027783641160949866, | |
| "loss": 31305360998.4, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.63, | |
| "learning_rate": 0.00026992084432717674, | |
| "loss": 100943947366.4, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "learning_rate": 0.0002620052770448549, | |
| "loss": 1337943859.2, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "eval_loss": NaN, | |
| "eval_runtime": 33.6458, | |
| "eval_samples_per_second": 59.443, | |
| "eval_steps_per_second": 1.872, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.77, | |
| "learning_rate": 0.00025408970976253297, | |
| "loss": 13503238963.2, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "learning_rate": 0.00024617414248021105, | |
| "loss": 4071580057.6, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.91, | |
| "learning_rate": 0.00023825857519788916, | |
| "loss": 2072524800.0, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.98, | |
| "learning_rate": 0.00023034300791556725, | |
| "loss": 771705753.6, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 1.05, | |
| "learning_rate": 0.00022242744063324536, | |
| "loss": 12987478016.0, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 1.12, | |
| "learning_rate": 0.00021451187335092345, | |
| "loss": 7428961075.2, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 1.19, | |
| "learning_rate": 0.0002065963060686016, | |
| "loss": 20587793612.8, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 1.26, | |
| "learning_rate": 0.00019868073878627967, | |
| "loss": 4394051174.4, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 1.33, | |
| "learning_rate": 0.00019076517150395776, | |
| "loss": 83019392614.4, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 1.4, | |
| "learning_rate": 0.00018284960422163587, | |
| "loss": 15752210022.4, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.4, | |
| "eval_loss": NaN, | |
| "eval_runtime": 33.6928, | |
| "eval_samples_per_second": 59.36, | |
| "eval_steps_per_second": 1.87, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.47, | |
| "learning_rate": 0.00017493403693931398, | |
| "loss": 22903119872.0, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 1.54, | |
| "learning_rate": 0.00016701846965699207, | |
| "loss": 3805777510.4, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 1.61, | |
| "learning_rate": 0.00015910290237467018, | |
| "loss": 8790727065.6, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 1.68, | |
| "learning_rate": 0.00015118733509234827, | |
| "loss": 16184664064.0, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 1.75, | |
| "learning_rate": 0.00014327176781002638, | |
| "loss": 479622384844.8, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.82, | |
| "learning_rate": 0.00013535620052770446, | |
| "loss": 59689546547.2, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 1.89, | |
| "learning_rate": 0.00012744063324538258, | |
| "loss": 7117320192.0, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 1.96, | |
| "learning_rate": 0.00011952506596306068, | |
| "loss": 165947716403.2, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 2.03, | |
| "learning_rate": 0.00011160949868073879, | |
| "loss": 1006318694.4, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 2.1, | |
| "learning_rate": 0.00010369393139841687, | |
| "loss": 476977612.8, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 2.1, | |
| "eval_loss": NaN, | |
| "eval_runtime": 33.5099, | |
| "eval_samples_per_second": 59.684, | |
| "eval_steps_per_second": 1.88, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 2.17, | |
| "learning_rate": 9.577836411609499e-05, | |
| "loss": 13798570393.6, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 2.24, | |
| "learning_rate": 8.786279683377308e-05, | |
| "loss": 1035509760.0, | |
| "step": 640 | |
| }, | |
| { | |
| "epoch": 2.31, | |
| "learning_rate": 7.994722955145117e-05, | |
| "loss": 27405954252.8, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 2.38, | |
| "learning_rate": 7.203166226912928e-05, | |
| "loss": 26138355302.4, | |
| "step": 680 | |
| }, | |
| { | |
| "epoch": 2.45, | |
| "learning_rate": 6.411609498680738e-05, | |
| "loss": 13618489753.6, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 2.52, | |
| "learning_rate": 5.620052770448549e-05, | |
| "loss": 389702352896.0, | |
| "step": 720 | |
| }, | |
| { | |
| "epoch": 2.59, | |
| "learning_rate": 4.828496042216358e-05, | |
| "loss": 128603114700.8, | |
| "step": 740 | |
| }, | |
| { | |
| "epoch": 2.66, | |
| "learning_rate": 4.036939313984169e-05, | |
| "loss": 342587443.2, | |
| "step": 760 | |
| }, | |
| { | |
| "epoch": 2.73, | |
| "learning_rate": 3.2453825857519784e-05, | |
| "loss": 27623394508.8, | |
| "step": 780 | |
| }, | |
| { | |
| "epoch": 2.8, | |
| "learning_rate": 2.4538258575197886e-05, | |
| "loss": 15703303782.4, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 2.8, | |
| "eval_loss": NaN, | |
| "eval_runtime": 33.5559, | |
| "eval_samples_per_second": 59.602, | |
| "eval_steps_per_second": 1.877, | |
| "step": 800 | |
| } | |
| ], | |
| "max_steps": 858, | |
| "num_train_epochs": 3, | |
| "total_flos": 8.316244295052952e+18, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |