| { | |
| "best_metric": 0.25722527503967285, | |
| "best_model_checkpoint": "whisper-medium-v3-ff4/checkpoint-20000", | |
| "epoch": 3.3243830712185143, | |
| "eval_steps": 5000, | |
| "global_step": 26000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.06393044367727913, | |
| "grad_norm": 8.506481170654297, | |
| "learning_rate": 1.965626204238921e-05, | |
| "loss": 0.9546, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.12786088735455825, | |
| "grad_norm": 9.10377025604248, | |
| "learning_rate": 1.9270905587668595e-05, | |
| "loss": 0.5883, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.19179133103183735, | |
| "grad_norm": 6.146899223327637, | |
| "learning_rate": 1.888554913294798e-05, | |
| "loss": 0.5081, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.2557217747091165, | |
| "grad_norm": 4.1534929275512695, | |
| "learning_rate": 1.8500192678227363e-05, | |
| "loss": 0.4631, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.3196522183863956, | |
| "grad_norm": 5.8231587409973145, | |
| "learning_rate": 1.8114836223506743e-05, | |
| "loss": 0.4313, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 0.3835826620636747, | |
| "grad_norm": 7.882592678070068, | |
| "learning_rate": 1.772947976878613e-05, | |
| "loss": 0.417, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.4475131057409538, | |
| "grad_norm": 5.645440101623535, | |
| "learning_rate": 1.7344123314065514e-05, | |
| "loss": 0.4162, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 0.511443549418233, | |
| "grad_norm": 4.277677536010742, | |
| "learning_rate": 1.6958766859344894e-05, | |
| "loss": 0.378, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.5753739930955121, | |
| "grad_norm": 4.594604969024658, | |
| "learning_rate": 1.6573410404624278e-05, | |
| "loss": 0.3653, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 0.6393044367727913, | |
| "grad_norm": 6.918728351593018, | |
| "learning_rate": 1.6188053949903662e-05, | |
| "loss": 0.3543, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.6393044367727913, | |
| "eval_loss": 0.3379935026168823, | |
| "eval_runtime": 1120.3014, | |
| "eval_samples_per_second": 3.096, | |
| "eval_steps_per_second": 0.387, | |
| "eval_wer": 0.2361289200878013, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.7032348804500703, | |
| "grad_norm": 5.27606725692749, | |
| "learning_rate": 1.5802697495183046e-05, | |
| "loss": 0.34, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 0.7671653241273494, | |
| "grad_norm": 8.176605224609375, | |
| "learning_rate": 1.541734104046243e-05, | |
| "loss": 0.3273, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 0.8310957678046286, | |
| "grad_norm": 4.268400192260742, | |
| "learning_rate": 1.5032755298651254e-05, | |
| "loss": 0.3305, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 0.8950262114819076, | |
| "grad_norm": 7.465890884399414, | |
| "learning_rate": 1.4647398843930638e-05, | |
| "loss": 0.318, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 0.9589566551591868, | |
| "grad_norm": 7.091923236846924, | |
| "learning_rate": 1.426204238921002e-05, | |
| "loss": 0.2965, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 1.022887098836466, | |
| "grad_norm": 5.541090965270996, | |
| "learning_rate": 1.3878227360308287e-05, | |
| "loss": 0.2865, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 1.086817542513745, | |
| "grad_norm": 5.313994407653809, | |
| "learning_rate": 1.349287090558767e-05, | |
| "loss": 0.2109, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 1.1507479861910241, | |
| "grad_norm": 4.2993621826171875, | |
| "learning_rate": 1.3107514450867053e-05, | |
| "loss": 0.2191, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 1.2146784298683033, | |
| "grad_norm": 5.328224182128906, | |
| "learning_rate": 1.2722157996146436e-05, | |
| "loss": 0.2196, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 1.2786088735455823, | |
| "grad_norm": 5.929138660430908, | |
| "learning_rate": 1.2336801541425821e-05, | |
| "loss": 0.2259, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 1.2786088735455823, | |
| "eval_loss": 0.2893017828464508, | |
| "eval_runtime": 1112.5698, | |
| "eval_samples_per_second": 3.117, | |
| "eval_steps_per_second": 0.39, | |
| "eval_wer": 0.2021819920457262, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 1.3425393172228615, | |
| "grad_norm": 4.919782638549805, | |
| "learning_rate": 1.1951445086705203e-05, | |
| "loss": 0.2224, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 1.4064697609001406, | |
| "grad_norm": 5.959251403808594, | |
| "learning_rate": 1.1566088631984587e-05, | |
| "loss": 0.2125, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 1.4704002045774198, | |
| "grad_norm": 4.974765777587891, | |
| "learning_rate": 1.118073217726397e-05, | |
| "loss": 0.2074, | |
| "step": 11500 | |
| }, | |
| { | |
| "epoch": 1.534330648254699, | |
| "grad_norm": 5.756904125213623, | |
| "learning_rate": 1.0795375722543353e-05, | |
| "loss": 0.2089, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 1.598261091931978, | |
| "grad_norm": 3.7231180667877197, | |
| "learning_rate": 1.0410019267822737e-05, | |
| "loss": 0.215, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 1.6621915356092571, | |
| "grad_norm": 4.505956172943115, | |
| "learning_rate": 1.0024662813102119e-05, | |
| "loss": 0.204, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 1.7261219792865363, | |
| "grad_norm": 2.797830820083618, | |
| "learning_rate": 9.639306358381504e-06, | |
| "loss": 0.1983, | |
| "step": 13500 | |
| }, | |
| { | |
| "epoch": 1.7900524229638153, | |
| "grad_norm": 6.62880802154541, | |
| "learning_rate": 9.254720616570328e-06, | |
| "loss": 0.2004, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 1.8539828666410945, | |
| "grad_norm": 4.8896284103393555, | |
| "learning_rate": 8.869364161849712e-06, | |
| "loss": 0.2023, | |
| "step": 14500 | |
| }, | |
| { | |
| "epoch": 1.9179133103183736, | |
| "grad_norm": 4.615833759307861, | |
| "learning_rate": 8.484007707129094e-06, | |
| "loss": 0.1938, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 1.9179133103183736, | |
| "eval_loss": 0.2596728801727295, | |
| "eval_runtime": 1104.7223, | |
| "eval_samples_per_second": 3.139, | |
| "eval_steps_per_second": 0.393, | |
| "eval_wer": 0.1790146263012627, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 1.9818437539956526, | |
| "grad_norm": 2.99739408493042, | |
| "learning_rate": 8.098651252408478e-06, | |
| "loss": 0.1871, | |
| "step": 15500 | |
| }, | |
| { | |
| "epoch": 2.045774197672932, | |
| "grad_norm": 1.4894932508468628, | |
| "learning_rate": 7.713294797687862e-06, | |
| "loss": 0.1419, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 2.109704641350211, | |
| "grad_norm": 3.1801295280456543, | |
| "learning_rate": 7.327938342967246e-06, | |
| "loss": 0.1193, | |
| "step": 16500 | |
| }, | |
| { | |
| "epoch": 2.17363508502749, | |
| "grad_norm": 3.464812994003296, | |
| "learning_rate": 6.942581888246629e-06, | |
| "loss": 0.1196, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 2.2375655287047693, | |
| "grad_norm": 2.4718642234802246, | |
| "learning_rate": 6.5572254335260125e-06, | |
| "loss": 0.1173, | |
| "step": 17500 | |
| }, | |
| { | |
| "epoch": 2.3014959723820483, | |
| "grad_norm": 4.073084354400635, | |
| "learning_rate": 6.172639691714837e-06, | |
| "loss": 0.1179, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 2.3654264160593277, | |
| "grad_norm": 2.377967119216919, | |
| "learning_rate": 5.78728323699422e-06, | |
| "loss": 0.1158, | |
| "step": 18500 | |
| }, | |
| { | |
| "epoch": 2.4293568597366066, | |
| "grad_norm": 6.478204250335693, | |
| "learning_rate": 5.401926782273603e-06, | |
| "loss": 0.1229, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 2.4932873034138856, | |
| "grad_norm": 1.8411247730255127, | |
| "learning_rate": 5.016570327552986e-06, | |
| "loss": 0.1144, | |
| "step": 19500 | |
| }, | |
| { | |
| "epoch": 2.5572177470911646, | |
| "grad_norm": 3.3542532920837402, | |
| "learning_rate": 4.631984585741811e-06, | |
| "loss": 0.1119, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 2.5572177470911646, | |
| "eval_loss": 0.25722527503967285, | |
| "eval_runtime": 1107.3303, | |
| "eval_samples_per_second": 3.132, | |
| "eval_steps_per_second": 0.392, | |
| "eval_wer": 0.16938691239432335, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 2.621148190768444, | |
| "grad_norm": 4.66868257522583, | |
| "learning_rate": 4.246628131021195e-06, | |
| "loss": 0.1126, | |
| "step": 20500 | |
| }, | |
| { | |
| "epoch": 2.685078634445723, | |
| "grad_norm": 5.47554874420166, | |
| "learning_rate": 3.861271676300578e-06, | |
| "loss": 0.1171, | |
| "step": 21000 | |
| }, | |
| { | |
| "epoch": 2.7490090781230023, | |
| "grad_norm": 3.905674695968628, | |
| "learning_rate": 3.475915221579962e-06, | |
| "loss": 0.1158, | |
| "step": 21500 | |
| }, | |
| { | |
| "epoch": 2.8129395218002813, | |
| "grad_norm": 3.0056424140930176, | |
| "learning_rate": 3.0913294797687864e-06, | |
| "loss": 0.1133, | |
| "step": 22000 | |
| }, | |
| { | |
| "epoch": 2.8768699654775602, | |
| "grad_norm": 1.467344880104065, | |
| "learning_rate": 2.70597302504817e-06, | |
| "loss": 0.1085, | |
| "step": 22500 | |
| }, | |
| { | |
| "epoch": 2.9408004091548396, | |
| "grad_norm": 1.9786620140075684, | |
| "learning_rate": 2.3213872832369944e-06, | |
| "loss": 0.1084, | |
| "step": 23000 | |
| }, | |
| { | |
| "epoch": 3.0047308528321186, | |
| "grad_norm": 3.245360851287842, | |
| "learning_rate": 1.9360308285163777e-06, | |
| "loss": 0.102, | |
| "step": 23500 | |
| }, | |
| { | |
| "epoch": 3.0686612965093976, | |
| "grad_norm": 0.7349840998649597, | |
| "learning_rate": 1.5506743737957611e-06, | |
| "loss": 0.063, | |
| "step": 24000 | |
| }, | |
| { | |
| "epoch": 3.132591740186677, | |
| "grad_norm": 1.5207897424697876, | |
| "learning_rate": 1.1653179190751445e-06, | |
| "loss": 0.0616, | |
| "step": 24500 | |
| }, | |
| { | |
| "epoch": 3.196522183863956, | |
| "grad_norm": 3.616459846496582, | |
| "learning_rate": 7.799614643545279e-07, | |
| "loss": 0.0632, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 3.196522183863956, | |
| "eval_loss": 0.2614915072917938, | |
| "eval_runtime": 1121.1739, | |
| "eval_samples_per_second": 3.093, | |
| "eval_steps_per_second": 0.387, | |
| "eval_wer": 0.16601829917631974, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 3.2604526275412353, | |
| "grad_norm": 1.1991757154464722, | |
| "learning_rate": 3.946050096339114e-07, | |
| "loss": 0.0598, | |
| "step": 25500 | |
| }, | |
| { | |
| "epoch": 3.3243830712185143, | |
| "grad_norm": 3.6829333305358887, | |
| "learning_rate": 9.248554913294799e-09, | |
| "loss": 0.0593, | |
| "step": 26000 | |
| }, | |
| { | |
| "epoch": 3.3243830712185143, | |
| "step": 26000, | |
| "total_flos": 6.002316599795712e+19, | |
| "train_loss": 0.2320793916262113, | |
| "train_runtime": 50624.8997, | |
| "train_samples_per_second": 4.109, | |
| "train_steps_per_second": 0.514 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 26000, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 4, | |
| "save_steps": 5000, | |
| "stateful_callbacks": { | |
| "EarlyStoppingCallback": { | |
| "args": { | |
| "early_stopping_patience": 3, | |
| "early_stopping_threshold": 0.0 | |
| }, | |
| "attributes": { | |
| "early_stopping_patience_counter": 1 | |
| } | |
| }, | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 6.002316599795712e+19, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |