| { | |
| "best_global_step": 24815, | |
| "best_metric": 2.6661746501922607, | |
| "best_model_checkpoint": "lmsys-prompts/checkpoint-24815", | |
| "epoch": 5.0, | |
| "eval_steps": 500, | |
| "global_step": 24815, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.10074551682450131, | |
| "grad_norm": 216.5968017578125, | |
| "learning_rate": 4.832426623681913e-05, | |
| "loss": 5.5953, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.20149103364900262, | |
| "grad_norm": 90.29191589355469, | |
| "learning_rate": 4.664517428974411e-05, | |
| "loss": 4.1114, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.30223655047350395, | |
| "grad_norm": 27.48844337463379, | |
| "learning_rate": 4.496608234266908e-05, | |
| "loss": 3.9607, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.40298206729800523, | |
| "grad_norm": 29.192981719970703, | |
| "learning_rate": 4.328699039559406e-05, | |
| "loss": 3.7888, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.5037275841225065, | |
| "grad_norm": 27.68057632446289, | |
| "learning_rate": 4.160789844851904e-05, | |
| "loss": 3.6636, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 0.6044731009470079, | |
| "grad_norm": 26.73339080810547, | |
| "learning_rate": 3.992880650144402e-05, | |
| "loss": 3.5414, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.7052186177715092, | |
| "grad_norm": 13.406410217285156, | |
| "learning_rate": 3.8249714554368995e-05, | |
| "loss": 3.467, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 0.8059641345960105, | |
| "grad_norm": 17.33591079711914, | |
| "learning_rate": 3.657062260729397e-05, | |
| "loss": 3.3742, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.9067096514205117, | |
| "grad_norm": 11.469970703125, | |
| "learning_rate": 3.4891530660218955e-05, | |
| "loss": 3.3254, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.42128131925101936, | |
| "eval_loss": 3.2208595275878906, | |
| "eval_runtime": 16.4691, | |
| "eval_samples_per_second": 133.037, | |
| "eval_steps_per_second": 16.637, | |
| "step": 4963 | |
| }, | |
| { | |
| "epoch": 1.007455168245013, | |
| "grad_norm": 10.576227188110352, | |
| "learning_rate": 3.321243871314393e-05, | |
| "loss": 3.2632, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 1.1082006850695145, | |
| "grad_norm": 10.408350944519043, | |
| "learning_rate": 3.153334676606891e-05, | |
| "loss": 3.177, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 1.2089462018940158, | |
| "grad_norm": 8.68959903717041, | |
| "learning_rate": 2.9854254818993887e-05, | |
| "loss": 3.1321, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 1.309691718718517, | |
| "grad_norm": 10.844682693481445, | |
| "learning_rate": 2.817516287191887e-05, | |
| "loss": 3.0864, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 1.4104372355430184, | |
| "grad_norm": 11.559004783630371, | |
| "learning_rate": 2.6496070924843847e-05, | |
| "loss": 3.0802, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 1.5111827523675196, | |
| "grad_norm": 11.332049369812012, | |
| "learning_rate": 2.4816978977768823e-05, | |
| "loss": 3.021, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 1.611928269192021, | |
| "grad_norm": 8.480116844177246, | |
| "learning_rate": 2.3137887030693803e-05, | |
| "loss": 3.0068, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 1.7126737860165222, | |
| "grad_norm": 11.450610160827637, | |
| "learning_rate": 2.145879508361878e-05, | |
| "loss": 2.9865, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 1.8134193028410235, | |
| "grad_norm": 7.126410007476807, | |
| "learning_rate": 1.977970313654376e-05, | |
| "loss": 2.946, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 1.9141648196655248, | |
| "grad_norm": 9.249116897583008, | |
| "learning_rate": 1.8100611189468735e-05, | |
| "loss": 2.9236, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.4685653073780457, | |
| "eval_loss": 2.902540445327759, | |
| "eval_runtime": 17.401, | |
| "eval_samples_per_second": 125.912, | |
| "eval_steps_per_second": 15.746, | |
| "step": 9926 | |
| }, | |
| { | |
| "epoch": 2.014910336490026, | |
| "grad_norm": 7.414788246154785, | |
| "learning_rate": 1.6421519242393715e-05, | |
| "loss": 2.9077, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 2.1156558533145273, | |
| "grad_norm": 10.728218078613281, | |
| "learning_rate": 1.4742427295318693e-05, | |
| "loss": 2.8403, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 2.216401370139029, | |
| "grad_norm": 9.297683715820312, | |
| "learning_rate": 1.3063335348243671e-05, | |
| "loss": 2.8482, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 2.3171468869635303, | |
| "grad_norm": 6.445867538452148, | |
| "learning_rate": 1.1384243401168649e-05, | |
| "loss": 2.8066, | |
| "step": 11500 | |
| }, | |
| { | |
| "epoch": 2.4178924037880316, | |
| "grad_norm": 8.223403930664062, | |
| "learning_rate": 9.705151454093627e-06, | |
| "loss": 2.8088, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 2.518637920612533, | |
| "grad_norm": 6.288245677947998, | |
| "learning_rate": 8.026059507018605e-06, | |
| "loss": 2.7869, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 2.619383437437034, | |
| "grad_norm": 7.528985977172852, | |
| "learning_rate": 6.346967559943583e-06, | |
| "loss": 2.7774, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 2.7201289542615354, | |
| "grad_norm": 8.420631408691406, | |
| "learning_rate": 4.667875612868561e-06, | |
| "loss": 2.7582, | |
| "step": 13500 | |
| }, | |
| { | |
| "epoch": 2.8208744710860367, | |
| "grad_norm": 7.883354187011719, | |
| "learning_rate": 2.988783665793539e-06, | |
| "loss": 2.7613, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 2.921619987910538, | |
| "grad_norm": 7.071943759918213, | |
| "learning_rate": 1.309691718718517e-06, | |
| "loss": 2.7526, | |
| "step": 14500 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.48612581550848066, | |
| "eval_loss": 2.79274320602417, | |
| "eval_runtime": 16.6412, | |
| "eval_samples_per_second": 131.661, | |
| "eval_steps_per_second": 16.465, | |
| "step": 14889 | |
| }, | |
| { | |
| "epoch": 3.0223655047350393, | |
| "grad_norm": 7.703913688659668, | |
| "learning_rate": 1.97783598629861e-05, | |
| "loss": 2.7412, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 3.1231110215595406, | |
| "grad_norm": 9.797307968139648, | |
| "learning_rate": 1.8770904694741083e-05, | |
| "loss": 2.7521, | |
| "step": 15500 | |
| }, | |
| { | |
| "epoch": 3.223856538384042, | |
| "grad_norm": 11.222009658813477, | |
| "learning_rate": 1.776344952649607e-05, | |
| "loss": 2.7435, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 3.324602055208543, | |
| "grad_norm": 9.812458038330078, | |
| "learning_rate": 1.675599435825106e-05, | |
| "loss": 2.7507, | |
| "step": 16500 | |
| }, | |
| { | |
| "epoch": 3.4253475720330444, | |
| "grad_norm": 7.502840995788574, | |
| "learning_rate": 1.5748539190006045e-05, | |
| "loss": 2.7181, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 3.5260930888575457, | |
| "grad_norm": 6.958070278167725, | |
| "learning_rate": 1.4741084021761032e-05, | |
| "loss": 2.7264, | |
| "step": 17500 | |
| }, | |
| { | |
| "epoch": 3.626838605682047, | |
| "grad_norm": 7.174421787261963, | |
| "learning_rate": 1.373362885351602e-05, | |
| "loss": 2.7289, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 3.7275841225065482, | |
| "grad_norm": 5.985280990600586, | |
| "learning_rate": 1.2726173685271006e-05, | |
| "loss": 2.6752, | |
| "step": 18500 | |
| }, | |
| { | |
| "epoch": 3.8283296393310495, | |
| "grad_norm": 7.28540563583374, | |
| "learning_rate": 1.1718718517025991e-05, | |
| "loss": 2.6896, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 3.929075156155551, | |
| "grad_norm": 9.331765174865723, | |
| "learning_rate": 1.071126334878098e-05, | |
| "loss": 2.683, | |
| "step": 19500 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.4999047467356238, | |
| "eval_loss": 2.7130751609802246, | |
| "eval_runtime": 16.5004, | |
| "eval_samples_per_second": 132.785, | |
| "eval_steps_per_second": 16.606, | |
| "step": 19852 | |
| }, | |
| { | |
| "epoch": 4.029820672980052, | |
| "grad_norm": 8.58707046508789, | |
| "learning_rate": 9.703808180535967e-06, | |
| "loss": 2.6525, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 4.130566189804553, | |
| "grad_norm": 7.965948581695557, | |
| "learning_rate": 8.696353012290953e-06, | |
| "loss": 2.6099, | |
| "step": 20500 | |
| }, | |
| { | |
| "epoch": 4.231311706629055, | |
| "grad_norm": 5.951071739196777, | |
| "learning_rate": 7.68889784404594e-06, | |
| "loss": 2.633, | |
| "step": 21000 | |
| }, | |
| { | |
| "epoch": 4.332057223453556, | |
| "grad_norm": 6.427402973175049, | |
| "learning_rate": 6.681442675800926e-06, | |
| "loss": 2.6239, | |
| "step": 21500 | |
| }, | |
| { | |
| "epoch": 4.432802740278058, | |
| "grad_norm": 6.754240989685059, | |
| "learning_rate": 5.673987507555914e-06, | |
| "loss": 2.6033, | |
| "step": 22000 | |
| }, | |
| { | |
| "epoch": 4.533548257102559, | |
| "grad_norm": 6.711804389953613, | |
| "learning_rate": 4.666532339310901e-06, | |
| "loss": 2.6101, | |
| "step": 22500 | |
| }, | |
| { | |
| "epoch": 4.634293773927061, | |
| "grad_norm": 7.764159202575684, | |
| "learning_rate": 3.6590771710658874e-06, | |
| "loss": 2.6117, | |
| "step": 23000 | |
| }, | |
| { | |
| "epoch": 4.735039290751562, | |
| "grad_norm": 8.959000587463379, | |
| "learning_rate": 2.6516220028208746e-06, | |
| "loss": 2.6101, | |
| "step": 23500 | |
| }, | |
| { | |
| "epoch": 4.835784807576063, | |
| "grad_norm": 6.05490255355835, | |
| "learning_rate": 1.6441668345758613e-06, | |
| "loss": 2.5942, | |
| "step": 24000 | |
| }, | |
| { | |
| "epoch": 4.9365303244005645, | |
| "grad_norm": 6.731525897979736, | |
| "learning_rate": 6.367116663308483e-07, | |
| "loss": 2.6099, | |
| "step": 24500 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.5068343659501033, | |
| "eval_loss": 2.6661746501922607, | |
| "eval_runtime": 16.4299, | |
| "eval_samples_per_second": 133.355, | |
| "eval_steps_per_second": 16.677, | |
| "step": 24815 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "step": 24815, | |
| "total_flos": 5.448549676351488e+16, | |
| "train_loss": 1.0647943758421696, | |
| "train_runtime": 2230.6964, | |
| "train_samples_per_second": 88.995, | |
| "train_steps_per_second": 11.124 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 24815, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 5, | |
| "save_steps": 500, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 5.448549676351488e+16, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |