| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 10.0, | |
| "eval_steps": 100, | |
| "global_step": 450, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.22, | |
| "grad_norm": 0.2616782486438751, | |
| "learning_rate": 0.00025, | |
| "loss": 0.7467, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.44, | |
| "grad_norm": 0.4230063855648041, | |
| "learning_rate": 0.0005, | |
| "loss": 0.0676, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "grad_norm": 0.4279615581035614, | |
| "learning_rate": 0.0004993330709158879, | |
| "loss": 0.0669, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.89, | |
| "grad_norm": 0.26194414496421814, | |
| "learning_rate": 0.0004973358420187776, | |
| "loss": 0.0401, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 1.11, | |
| "grad_norm": 0.18304042518138885, | |
| "learning_rate": 0.0004940189693889818, | |
| "loss": 0.0267, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 1.33, | |
| "grad_norm": 0.15659470856189728, | |
| "learning_rate": 0.0004894001499771015, | |
| "loss": 0.0172, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 1.56, | |
| "grad_norm": 0.8856974840164185, | |
| "learning_rate": 0.00048350402718313703, | |
| "loss": 0.0141, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 1.78, | |
| "grad_norm": 0.2954822778701782, | |
| "learning_rate": 0.0004763620593732867, | |
| "loss": 0.0225, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 0.36144667863845825, | |
| "learning_rate": 0.00046801235203595195, | |
| "loss": 0.0182, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 2.22, | |
| "grad_norm": 0.20678772032260895, | |
| "learning_rate": 0.0004584994544724695, | |
| "loss": 0.013, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 2.22, | |
| "eval_loss": 0.013783249072730541, | |
| "eval_runtime": 4.8183, | |
| "eval_samples_per_second": 39.433, | |
| "eval_steps_per_second": 1.66, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 2.44, | |
| "grad_norm": 0.3987113833427429, | |
| "learning_rate": 0.00044787412210731353, | |
| "loss": 0.0078, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 2.67, | |
| "grad_norm": 0.23118920624256134, | |
| "learning_rate": 0.00043619304568594545, | |
| "loss": 0.0104, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 2.89, | |
| "grad_norm": 0.1873694211244583, | |
| "learning_rate": 0.0004235185488051585, | |
| "loss": 0.0191, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 3.11, | |
| "grad_norm": 0.03233395144343376, | |
| "learning_rate": 0.0004099182553897228, | |
| "loss": 0.0065, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 3.33, | |
| "grad_norm": 0.5932971239089966, | |
| "learning_rate": 0.00039546472888948826, | |
| "loss": 0.0034, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 3.56, | |
| "grad_norm": 0.13725461065769196, | |
| "learning_rate": 0.0003802350851219826, | |
| "loss": 0.007, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 3.78, | |
| "grad_norm": 0.22745588421821594, | |
| "learning_rate": 0.00036431058082615964, | |
| "loss": 0.0058, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.0794651210308075, | |
| "learning_rate": 0.00034777618012253895, | |
| "loss": 0.0039, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 4.22, | |
| "grad_norm": 0.06090324744582176, | |
| "learning_rate": 0.00033072010119286155, | |
| "loss": 0.0035, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 4.44, | |
| "grad_norm": 0.17350110411643982, | |
| "learning_rate": 0.0003132333455979202, | |
| "loss": 0.0034, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 4.44, | |
| "eval_loss": 0.01827273517847061, | |
| "eval_runtime": 4.8208, | |
| "eval_samples_per_second": 39.413, | |
| "eval_steps_per_second": 1.659, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 4.67, | |
| "grad_norm": 0.0679766833782196, | |
| "learning_rate": 0.0002954092127448591, | |
| "loss": 0.0043, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 4.89, | |
| "grad_norm": 0.2287779450416565, | |
| "learning_rate": 0.00027734280209446867, | |
| "loss": 0.0033, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 5.11, | |
| "grad_norm": 0.022623347118496895, | |
| "learning_rate": 0.00025913050576441477, | |
| "loss": 0.0051, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 5.33, | |
| "grad_norm": 0.023388344794511795, | |
| "learning_rate": 0.00024086949423558527, | |
| "loss": 0.0011, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 5.56, | |
| "grad_norm": 0.06873177736997604, | |
| "learning_rate": 0.00022265719790553147, | |
| "loss": 0.0032, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 5.78, | |
| "grad_norm": 0.00556823518127203, | |
| "learning_rate": 0.0002045907872551409, | |
| "loss": 0.0016, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "grad_norm": 0.00377739779651165, | |
| "learning_rate": 0.0001867666544020798, | |
| "loss": 0.0008, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 6.22, | |
| "grad_norm": 0.017272206023335457, | |
| "learning_rate": 0.00016927989880713852, | |
| "loss": 0.0008, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 6.44, | |
| "grad_norm": 0.028082353994250298, | |
| "learning_rate": 0.00015222381987746103, | |
| "loss": 0.0002, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 6.67, | |
| "grad_norm": 0.0010064858943223953, | |
| "learning_rate": 0.00013568941917384037, | |
| "loss": 0.0001, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 6.67, | |
| "eval_loss": 0.01472838968038559, | |
| "eval_runtime": 4.8243, | |
| "eval_samples_per_second": 39.384, | |
| "eval_steps_per_second": 1.658, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 6.89, | |
| "grad_norm": 0.0007610166212543845, | |
| "learning_rate": 0.00011976491487801746, | |
| "loss": 0.0008, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 7.11, | |
| "grad_norm": 0.04695134982466698, | |
| "learning_rate": 0.00010453527111051184, | |
| "loss": 0.0002, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 7.33, | |
| "grad_norm": 0.0033624598290771246, | |
| "learning_rate": 9.008174461027724e-05, | |
| "loss": 0.0, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 7.56, | |
| "grad_norm": 0.003984487149864435, | |
| "learning_rate": 7.648145119484151e-05, | |
| "loss": 0.0001, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 7.78, | |
| "grad_norm": 0.002355258446186781, | |
| "learning_rate": 6.380695431405453e-05, | |
| "loss": 0.0, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "grad_norm": 0.00041891244472935796, | |
| "learning_rate": 5.21258778926865e-05, | |
| "loss": 0.0, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 8.22, | |
| "grad_norm": 0.0004031055432278663, | |
| "learning_rate": 4.150054552753055e-05, | |
| "loss": 0.0, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 8.44, | |
| "grad_norm": 0.00040033855475485325, | |
| "learning_rate": 3.198764796404807e-05, | |
| "loss": 0.0, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 8.67, | |
| "grad_norm": 0.0015161971095949411, | |
| "learning_rate": 2.3637940626713344e-05, | |
| "loss": 0.0, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 8.89, | |
| "grad_norm": 0.0011603726306930184, | |
| "learning_rate": 1.649597281686302e-05, | |
| "loss": 0.0, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 8.89, | |
| "eval_loss": 0.014345706440508366, | |
| "eval_runtime": 4.8186, | |
| "eval_samples_per_second": 39.43, | |
| "eval_steps_per_second": 1.66, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 9.11, | |
| "grad_norm": 0.0003027151105925441, | |
| "learning_rate": 1.0599850022898538e-05, | |
| "loss": 0.0001, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 9.33, | |
| "grad_norm": 0.0009049432119354606, | |
| "learning_rate": 5.981030611018234e-06, | |
| "loss": 0.0, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 9.56, | |
| "grad_norm": 0.0018125231144949794, | |
| "learning_rate": 2.664157981222437e-06, | |
| "loss": 0.0, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 9.78, | |
| "grad_norm": 0.003249464090913534, | |
| "learning_rate": 6.66929084112089e-07, | |
| "loss": 0.0, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "grad_norm": 0.0004782461328431964, | |
| "learning_rate": 0.0, | |
| "loss": 0.0, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "step": 450, | |
| "total_flos": 4.3483361730730394e+17, | |
| "train_loss": 0.025018121934683425, | |
| "train_runtime": 909.2296, | |
| "train_samples_per_second": 11.834, | |
| "train_steps_per_second": 0.495 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 450, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 10, | |
| "save_steps": 500, | |
| "total_flos": 4.3483361730730394e+17, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |