| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 30.0, | |
| "eval_steps": 500, | |
| "global_step": 67440, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 20.024471282958984, | |
| "learning_rate": 9.66755634638197e-06, | |
| "loss": 1.2057, | |
| "step": 2248 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.450177937746048, | |
| "eval_loss": 1.4521995782852173, | |
| "eval_runtime": 46.9769, | |
| "eval_samples_per_second": 11.963, | |
| "eval_steps_per_second": 11.963, | |
| "step": 2248 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 40.918148040771484, | |
| "learning_rate": 9.334371293001186e-06, | |
| "loss": 1.3873, | |
| "step": 4496 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.6423487663269043, | |
| "eval_loss": 1.4503390789031982, | |
| "eval_runtime": 47.9141, | |
| "eval_samples_per_second": 11.729, | |
| "eval_steps_per_second": 11.729, | |
| "step": 4496 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 14.9369478225708, | |
| "learning_rate": 9.001334519572953e-06, | |
| "loss": 1.4246, | |
| "step": 6744 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.6672598123550415, | |
| "eval_loss": 1.6165008544921875, | |
| "eval_runtime": 47.7427, | |
| "eval_samples_per_second": 11.771, | |
| "eval_steps_per_second": 11.771, | |
| "step": 6744 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.12427664548158646, | |
| "learning_rate": 8.668149466192172e-06, | |
| "loss": 1.3335, | |
| "step": 8992 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.7206405401229858, | |
| "eval_loss": 1.478636384010315, | |
| "eval_runtime": 47.5592, | |
| "eval_samples_per_second": 11.817, | |
| "eval_steps_per_second": 11.817, | |
| "step": 8992 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "grad_norm": 0.1769200563430786, | |
| "learning_rate": 8.334964412811389e-06, | |
| "loss": 1.251, | |
| "step": 11240 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.6886121034622192, | |
| "eval_loss": 1.6414002180099487, | |
| "eval_runtime": 47.4426, | |
| "eval_samples_per_second": 11.846, | |
| "eval_steps_per_second": 11.846, | |
| "step": 11240 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "grad_norm": 0.14434462785720825, | |
| "learning_rate": 8.001779359430606e-06, | |
| "loss": 1.1859, | |
| "step": 13488 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_accuracy": 0.754448413848877, | |
| "eval_loss": 1.329976201057434, | |
| "eval_runtime": 47.6847, | |
| "eval_samples_per_second": 11.786, | |
| "eval_steps_per_second": 11.786, | |
| "step": 13488 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "grad_norm": 24.07697105407715, | |
| "learning_rate": 7.668446026097273e-06, | |
| "loss": 1.1132, | |
| "step": 15736 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_accuracy": 0.7508896589279175, | |
| "eval_loss": 1.3665430545806885, | |
| "eval_runtime": 47.6469, | |
| "eval_samples_per_second": 11.795, | |
| "eval_steps_per_second": 11.795, | |
| "step": 15736 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "grad_norm": 3.64510178565979, | |
| "learning_rate": 7.335112692763939e-06, | |
| "loss": 1.0189, | |
| "step": 17984 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_accuracy": 0.7153024673461914, | |
| "eval_loss": 1.6664679050445557, | |
| "eval_runtime": 47.663, | |
| "eval_samples_per_second": 11.791, | |
| "eval_steps_per_second": 11.791, | |
| "step": 17984 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "grad_norm": 0.2391408234834671, | |
| "learning_rate": 7.001927639383156e-06, | |
| "loss": 0.9807, | |
| "step": 20232 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_accuracy": 0.7793594598770142, | |
| "eval_loss": 1.1174825429916382, | |
| "eval_runtime": 47.784, | |
| "eval_samples_per_second": 11.761, | |
| "eval_steps_per_second": 11.761, | |
| "step": 20232 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "grad_norm": 0.037992771714925766, | |
| "learning_rate": 6.668890865954924e-06, | |
| "loss": 0.8786, | |
| "step": 22480 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_accuracy": 0.7882562279701233, | |
| "eval_loss": 1.1786267757415771, | |
| "eval_runtime": 47.6281, | |
| "eval_samples_per_second": 11.8, | |
| "eval_steps_per_second": 11.8, | |
| "step": 22480 | |
| }, | |
| { | |
| "epoch": 11.0, | |
| "grad_norm": 0.07834123820066452, | |
| "learning_rate": 6.3355575326215905e-06, | |
| "loss": 0.8677, | |
| "step": 24728 | |
| }, | |
| { | |
| "epoch": 11.0, | |
| "eval_accuracy": 0.7811387777328491, | |
| "eval_loss": 1.1294955015182495, | |
| "eval_runtime": 47.9525, | |
| "eval_samples_per_second": 11.72, | |
| "eval_steps_per_second": 11.72, | |
| "step": 24728 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "grad_norm": 0.0959361270070076, | |
| "learning_rate": 6.0025207591933574e-06, | |
| "loss": 0.7554, | |
| "step": 26976 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "eval_accuracy": 0.8185053467750549, | |
| "eval_loss": 1.1184855699539185, | |
| "eval_runtime": 47.4663, | |
| "eval_samples_per_second": 11.84, | |
| "eval_steps_per_second": 11.84, | |
| "step": 26976 | |
| }, | |
| { | |
| "epoch": 13.0, | |
| "grad_norm": 0.06532555818557739, | |
| "learning_rate": 5.669335705812574e-06, | |
| "loss": 0.7196, | |
| "step": 29224 | |
| }, | |
| { | |
| "epoch": 13.0, | |
| "eval_accuracy": 0.7846975326538086, | |
| "eval_loss": 1.4066756963729858, | |
| "eval_runtime": 48.1233, | |
| "eval_samples_per_second": 11.678, | |
| "eval_steps_per_second": 11.678, | |
| "step": 29224 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "grad_norm": 177.36297607421875, | |
| "learning_rate": 5.336150652431792e-06, | |
| "loss": 0.692, | |
| "step": 31472 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "eval_accuracy": 0.8202847242355347, | |
| "eval_loss": 1.1174949407577515, | |
| "eval_runtime": 47.9299, | |
| "eval_samples_per_second": 11.725, | |
| "eval_steps_per_second": 11.725, | |
| "step": 31472 | |
| }, | |
| { | |
| "epoch": 15.0, | |
| "grad_norm": 0.7878779768943787, | |
| "learning_rate": 5.002817319098459e-06, | |
| "loss": 0.6276, | |
| "step": 33720 | |
| }, | |
| { | |
| "epoch": 15.0, | |
| "eval_accuracy": 0.7882562279701233, | |
| "eval_loss": 1.4490171670913696, | |
| "eval_runtime": 47.6011, | |
| "eval_samples_per_second": 11.806, | |
| "eval_steps_per_second": 11.806, | |
| "step": 33720 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "grad_norm": 0.04724971204996109, | |
| "learning_rate": 4.6696322657176755e-06, | |
| "loss": 0.6083, | |
| "step": 35968 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "eval_accuracy": 0.8345195651054382, | |
| "eval_loss": 1.0982587337493896, | |
| "eval_runtime": 47.8014, | |
| "eval_samples_per_second": 11.757, | |
| "eval_steps_per_second": 11.757, | |
| "step": 35968 | |
| }, | |
| { | |
| "epoch": 17.0, | |
| "grad_norm": 0.3455263674259186, | |
| "learning_rate": 4.336447212336892e-06, | |
| "loss": 0.5204, | |
| "step": 38216 | |
| }, | |
| { | |
| "epoch": 17.0, | |
| "eval_accuracy": 0.8256227970123291, | |
| "eval_loss": 1.181368350982666, | |
| "eval_runtime": 47.9669, | |
| "eval_samples_per_second": 11.716, | |
| "eval_steps_per_second": 11.716, | |
| "step": 38216 | |
| }, | |
| { | |
| "epoch": 18.0, | |
| "grad_norm": 0.03767940774559975, | |
| "learning_rate": 4.003262158956109e-06, | |
| "loss": 0.5197, | |
| "step": 40464 | |
| }, | |
| { | |
| "epoch": 18.0, | |
| "eval_accuracy": 0.8167259693145752, | |
| "eval_loss": 1.2945315837860107, | |
| "eval_runtime": 47.911, | |
| "eval_samples_per_second": 11.73, | |
| "eval_steps_per_second": 11.73, | |
| "step": 40464 | |
| }, | |
| { | |
| "epoch": 19.0, | |
| "grad_norm": 365.2795104980469, | |
| "learning_rate": 3.6700771055753265e-06, | |
| "loss": 0.488, | |
| "step": 42712 | |
| }, | |
| { | |
| "epoch": 19.0, | |
| "eval_accuracy": 0.8024911284446716, | |
| "eval_loss": 1.4494409561157227, | |
| "eval_runtime": 44.9229, | |
| "eval_samples_per_second": 12.51, | |
| "eval_steps_per_second": 12.51, | |
| "step": 42712 | |
| }, | |
| { | |
| "epoch": 20.0, | |
| "grad_norm": 31.586769104003906, | |
| "learning_rate": 3.3368920521945437e-06, | |
| "loss": 0.4714, | |
| "step": 44960 | |
| }, | |
| { | |
| "epoch": 20.0, | |
| "eval_accuracy": 0.8113878965377808, | |
| "eval_loss": 1.3498995304107666, | |
| "eval_runtime": 44.7889, | |
| "eval_samples_per_second": 12.548, | |
| "eval_steps_per_second": 12.548, | |
| "step": 44960 | |
| }, | |
| { | |
| "epoch": 21.0, | |
| "grad_norm": 4.5781145095825195, | |
| "learning_rate": 3.0035587188612105e-06, | |
| "loss": 0.3641, | |
| "step": 47208 | |
| }, | |
| { | |
| "epoch": 21.0, | |
| "eval_accuracy": 0.8380783200263977, | |
| "eval_loss": 1.252496600151062, | |
| "eval_runtime": 44.8551, | |
| "eval_samples_per_second": 12.529, | |
| "eval_steps_per_second": 12.529, | |
| "step": 47208 | |
| }, | |
| { | |
| "epoch": 22.0, | |
| "grad_norm": 0.028242342174053192, | |
| "learning_rate": 2.6703736654804273e-06, | |
| "loss": 0.3877, | |
| "step": 49456 | |
| }, | |
| { | |
| "epoch": 22.0, | |
| "eval_accuracy": 0.8380783200263977, | |
| "eval_loss": 1.2610100507736206, | |
| "eval_runtime": 45.2543, | |
| "eval_samples_per_second": 12.419, | |
| "eval_steps_per_second": 12.419, | |
| "step": 49456 | |
| }, | |
| { | |
| "epoch": 23.0, | |
| "grad_norm": 0.02240828238427639, | |
| "learning_rate": 2.337188612099644e-06, | |
| "loss": 0.3253, | |
| "step": 51704 | |
| }, | |
| { | |
| "epoch": 23.0, | |
| "eval_accuracy": 0.8274021148681641, | |
| "eval_loss": 1.3913415670394897, | |
| "eval_runtime": 45.0743, | |
| "eval_samples_per_second": 12.468, | |
| "eval_steps_per_second": 12.468, | |
| "step": 51704 | |
| }, | |
| { | |
| "epoch": 24.0, | |
| "grad_norm": 776.1561279296875, | |
| "learning_rate": 2.0040035587188614e-06, | |
| "loss": 0.2978, | |
| "step": 53952 | |
| }, | |
| { | |
| "epoch": 24.0, | |
| "eval_accuracy": 0.8416370153427124, | |
| "eval_loss": 1.2989881038665771, | |
| "eval_runtime": 44.7173, | |
| "eval_samples_per_second": 12.568, | |
| "eval_steps_per_second": 12.568, | |
| "step": 53952 | |
| }, | |
| { | |
| "epoch": 25.0, | |
| "grad_norm": 118.77944946289062, | |
| "learning_rate": 1.6706702253855281e-06, | |
| "loss": 0.3238, | |
| "step": 56200 | |
| }, | |
| { | |
| "epoch": 25.0, | |
| "eval_accuracy": 0.8274021148681641, | |
| "eval_loss": 1.4328011274337769, | |
| "eval_runtime": 45.0042, | |
| "eval_samples_per_second": 12.488, | |
| "eval_steps_per_second": 12.488, | |
| "step": 56200 | |
| }, | |
| { | |
| "epoch": 26.0, | |
| "grad_norm": 0.006892771925777197, | |
| "learning_rate": 1.3373368920521945e-06, | |
| "loss": 0.2669, | |
| "step": 58448 | |
| }, | |
| { | |
| "epoch": 26.0, | |
| "eval_accuracy": 0.8327401876449585, | |
| "eval_loss": 1.3079112768173218, | |
| "eval_runtime": 45.1126, | |
| "eval_samples_per_second": 12.458, | |
| "eval_steps_per_second": 12.458, | |
| "step": 58448 | |
| }, | |
| { | |
| "epoch": 27.0, | |
| "grad_norm": 0.0030675730668008327, | |
| "learning_rate": 1.0041518386714117e-06, | |
| "loss": 0.2521, | |
| "step": 60696 | |
| }, | |
| { | |
| "epoch": 27.0, | |
| "eval_accuracy": 0.8398576378822327, | |
| "eval_loss": 1.3250336647033691, | |
| "eval_runtime": 45.3922, | |
| "eval_samples_per_second": 12.381, | |
| "eval_steps_per_second": 12.381, | |
| "step": 60696 | |
| }, | |
| { | |
| "epoch": 28.0, | |
| "grad_norm": 0.06381255388259888, | |
| "learning_rate": 6.709667852906288e-07, | |
| "loss": 0.2632, | |
| "step": 62944 | |
| }, | |
| { | |
| "epoch": 28.0, | |
| "eval_accuracy": 0.8416370153427124, | |
| "eval_loss": 1.3356966972351074, | |
| "eval_runtime": 44.8744, | |
| "eval_samples_per_second": 12.524, | |
| "eval_steps_per_second": 12.524, | |
| "step": 62944 | |
| }, | |
| { | |
| "epoch": 29.0, | |
| "grad_norm": 0.0834624245762825, | |
| "learning_rate": 3.377817319098458e-07, | |
| "loss": 0.2655, | |
| "step": 65192 | |
| }, | |
| { | |
| "epoch": 29.0, | |
| "eval_accuracy": 0.8434163928031921, | |
| "eval_loss": 1.2957476377487183, | |
| "eval_runtime": 44.9511, | |
| "eval_samples_per_second": 12.502, | |
| "eval_steps_per_second": 12.502, | |
| "step": 65192 | |
| }, | |
| { | |
| "epoch": 30.0, | |
| "grad_norm": 0.026273438706994057, | |
| "learning_rate": 4.4483985765124555e-09, | |
| "loss": 0.2379, | |
| "step": 67440 | |
| }, | |
| { | |
| "epoch": 30.0, | |
| "eval_accuracy": 0.8487544655799866, | |
| "eval_loss": 1.2928475141525269, | |
| "eval_runtime": 44.8607, | |
| "eval_samples_per_second": 12.528, | |
| "eval_steps_per_second": 12.528, | |
| "step": 67440 | |
| }, | |
| { | |
| "epoch": 30.0, | |
| "step": 67440, | |
| "total_flos": 6.148286618112e+18, | |
| "train_loss": 0.7011247130334024, | |
| "train_runtime": 11844.845, | |
| "train_samples_per_second": 5.694, | |
| "train_steps_per_second": 5.694 | |
| } | |
| ], | |
| "logging_steps": 35, | |
| "max_steps": 67440, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 30, | |
| "save_steps": 500, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": false, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 6.148286618112e+18, | |
| "train_batch_size": 1, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |