| { | |
| "best_metric": 2.3333284854888916, | |
| "best_model_checkpoint": "./output/training_results/C019_random_sample_Meta-Llama-3-8B_pretrain_20240726_033210/checkpoint-40700", | |
| "epoch": 4.0, | |
| "eval_steps": 4070, | |
| "global_step": 40700, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 9.828009828009828e-05, | |
| "grad_norm": 0.0, | |
| "learning_rate": 0.0, | |
| "loss": 2.7384, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "grad_norm": 2.155671963962152, | |
| "learning_rate": 1.9927939731411727e-06, | |
| "loss": 2.4775, | |
| "step": 2035 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "grad_norm": 2.0610799346731716, | |
| "learning_rate": 2.238606233181306e-06, | |
| "loss": 2.404, | |
| "step": 4070 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "eval_loss": 2.388622283935547, | |
| "eval_runtime": 386.8721, | |
| "eval_samples_per_second": 187.02, | |
| "eval_steps_per_second": 1.463, | |
| "step": 4070 | |
| }, | |
| { | |
| "epoch": 0.6, | |
| "grad_norm": 1.852185161682542, | |
| "learning_rate": 1.2183995585808034e-06, | |
| "loss": 2.3695, | |
| "step": 6105 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "grad_norm": 1.963949683462878, | |
| "learning_rate": 6.503021712782451e-07, | |
| "loss": 2.3519, | |
| "step": 8140 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "eval_loss": 2.347768545150757, | |
| "eval_runtime": 353.4879, | |
| "eval_samples_per_second": 204.683, | |
| "eval_steps_per_second": 1.601, | |
| "step": 8140 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 1.8257205248825334, | |
| "learning_rate": 3.454633956957804e-07, | |
| "loss": 2.341, | |
| "step": 10175 | |
| }, | |
| { | |
| "epoch": 1.2, | |
| "grad_norm": 1.934651438097829, | |
| "learning_rate": 1.8843776358550853e-07, | |
| "loss": 2.272, | |
| "step": 12210 | |
| }, | |
| { | |
| "epoch": 1.2, | |
| "eval_loss": 2.3399369716644287, | |
| "eval_runtime": 353.2432, | |
| "eval_samples_per_second": 204.825, | |
| "eval_steps_per_second": 1.602, | |
| "step": 12210 | |
| }, | |
| { | |
| "epoch": 1.4, | |
| "grad_norm": 1.9087851196083852, | |
| "learning_rate": 1.1134838176940004e-07, | |
| "loss": 2.2758, | |
| "step": 14245 | |
| }, | |
| { | |
| "epoch": 1.6, | |
| "grad_norm": 1.9434977796166109, | |
| "learning_rate": 7.549690625694449e-08, | |
| "loss": 2.2737, | |
| "step": 16280 | |
| }, | |
| { | |
| "epoch": 1.6, | |
| "eval_loss": 2.3373029232025146, | |
| "eval_runtime": 353.221, | |
| "eval_samples_per_second": 204.838, | |
| "eval_steps_per_second": 1.602, | |
| "step": 16280 | |
| }, | |
| { | |
| "epoch": 1.8, | |
| "grad_norm": 1.8796392127857566, | |
| "learning_rate": 5.980471884190691e-08, | |
| "loss": 2.2725, | |
| "step": 18315 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 1.9345766812761183, | |
| "learning_rate": 5.3444261786522375e-08, | |
| "loss": 2.2735, | |
| "step": 20350 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_loss": 2.3360562324523926, | |
| "eval_runtime": 353.5294, | |
| "eval_samples_per_second": 204.659, | |
| "eval_steps_per_second": 1.601, | |
| "step": 20350 | |
| }, | |
| { | |
| "epoch": 2.2, | |
| "grad_norm": 1.941719791382991, | |
| "learning_rate": 5.108383189787308e-08, | |
| "loss": 2.2617, | |
| "step": 22385 | |
| }, | |
| { | |
| "epoch": 2.4, | |
| "grad_norm": 1.9268741466862327, | |
| "learning_rate": 5.029810361474273e-08, | |
| "loss": 2.2629, | |
| "step": 24420 | |
| }, | |
| { | |
| "epoch": 2.4, | |
| "eval_loss": 2.3360707759857178, | |
| "eval_runtime": 353.1028, | |
| "eval_samples_per_second": 204.906, | |
| "eval_steps_per_second": 1.603, | |
| "step": 24420 | |
| }, | |
| { | |
| "epoch": 2.6, | |
| "grad_norm": 1.9707963592912294, | |
| "learning_rate": 5.006893878600473e-08, | |
| "loss": 2.2617, | |
| "step": 26455 | |
| }, | |
| { | |
| "epoch": 2.8, | |
| "grad_norm": 1.9597319665023907, | |
| "learning_rate": 5.001272696084112e-08, | |
| "loss": 2.2655, | |
| "step": 28490 | |
| }, | |
| { | |
| "epoch": 2.8, | |
| "eval_loss": 2.3352677822113037, | |
| "eval_runtime": 353.4276, | |
| "eval_samples_per_second": 204.718, | |
| "eval_steps_per_second": 1.601, | |
| "step": 28490 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 34.02276184361365, | |
| "learning_rate": 5.0001729586514593e-08, | |
| "loss": 2.2615, | |
| "step": 30525 | |
| }, | |
| { | |
| "epoch": 3.2, | |
| "grad_norm": 2.124382265375981, | |
| "learning_rate": 5.0000150744726626e-08, | |
| "loss": 2.2567, | |
| "step": 32560 | |
| }, | |
| { | |
| "epoch": 3.2, | |
| "eval_loss": 2.334810733795166, | |
| "eval_runtime": 353.1613, | |
| "eval_samples_per_second": 204.872, | |
| "eval_steps_per_second": 1.603, | |
| "step": 32560 | |
| }, | |
| { | |
| "epoch": 3.4, | |
| "grad_norm": 1.9867886821885041, | |
| "learning_rate": 5.000000648758777e-08, | |
| "loss": 2.2555, | |
| "step": 34595 | |
| }, | |
| { | |
| "epoch": 3.6, | |
| "grad_norm": 2.0413183498941763, | |
| "learning_rate": 5.000000007849755e-08, | |
| "loss": 2.2581, | |
| "step": 36630 | |
| }, | |
| { | |
| "epoch": 3.6, | |
| "eval_loss": 2.3342058658599854, | |
| "eval_runtime": 353.3438, | |
| "eval_samples_per_second": 204.767, | |
| "eval_steps_per_second": 1.602, | |
| "step": 36630 | |
| }, | |
| { | |
| "epoch": 3.8, | |
| "grad_norm": 1.954005554664922, | |
| "learning_rate": 5.000000000004389e-08, | |
| "loss": 2.2592, | |
| "step": 38665 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 1.9904895703350474, | |
| "learning_rate": 5e-08, | |
| "loss": 2.2607, | |
| "step": 40700 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_loss": 2.3333284854888916, | |
| "eval_runtime": 353.3086, | |
| "eval_samples_per_second": 204.787, | |
| "eval_steps_per_second": 1.602, | |
| "step": 40700 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "step": 40700, | |
| "total_flos": 4255641501696000.0, | |
| "train_loss": 2.2957492570853644, | |
| "train_runtime": 58446.3006, | |
| "train_samples_per_second": 44.566, | |
| "train_steps_per_second": 0.696 | |
| } | |
| ], | |
| "logging_steps": 2035, | |
| "max_steps": 40700, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 4, | |
| "save_steps": 4070, | |
| "total_flos": 4255641501696000.0, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |