| { | |
| "best_global_step": 4000, | |
| "best_metric": 0.8441482186317444, | |
| "best_model_checkpoint": "/home/ksokim/finetune-llm/workspace/25w17_0421/250423_llm2vec/mntp/gemma3-multidomain_sampled_2m_eval/checkpoint-4000", | |
| "epoch": 1.0, | |
| "eval_steps": 1000, | |
| "global_step": 4210, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.1187789523696401, | |
| "grad_norm": 3.54020094871521, | |
| "learning_rate": 4.407363420427554e-05, | |
| "loss": 4.86, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.2375579047392802, | |
| "grad_norm": 3.5299084186553955, | |
| "learning_rate": 3.81353919239905e-05, | |
| "loss": 3.3164, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.2375579047392802, | |
| "eval_accuracy": 0.7672592058420341, | |
| "eval_loss": 0.965872585773468, | |
| "eval_runtime": 1354.4622, | |
| "eval_samples_per_second": 73.279, | |
| "eval_steps_per_second": 0.655, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.3563368571089203, | |
| "grad_norm": 3.1422698497772217, | |
| "learning_rate": 3.2197149643705464e-05, | |
| "loss": 3.1505, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.4751158094785604, | |
| "grad_norm": 3.2540836334228516, | |
| "learning_rate": 2.6258907363420427e-05, | |
| "loss": 3.0442, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.4751158094785604, | |
| "eval_accuracy": 0.7774986952473916, | |
| "eval_loss": 0.8946214318275452, | |
| "eval_runtime": 1351.4638, | |
| "eval_samples_per_second": 73.441, | |
| "eval_steps_per_second": 0.656, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.5938947618482004, | |
| "grad_norm": 3.3036434650421143, | |
| "learning_rate": 2.0320665083135394e-05, | |
| "loss": 2.9783, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 0.7126737142178406, | |
| "grad_norm": 3.374356985092163, | |
| "learning_rate": 1.4382422802850357e-05, | |
| "loss": 2.9391, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.7126737142178406, | |
| "eval_accuracy": 0.7826880250906799, | |
| "eval_loss": 0.860706090927124, | |
| "eval_runtime": 1352.0159, | |
| "eval_samples_per_second": 73.411, | |
| "eval_steps_per_second": 0.656, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.8314526665874807, | |
| "grad_norm": 3.2680838108062744, | |
| "learning_rate": 8.444180522565321e-06, | |
| "loss": 2.9004, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 0.9502316189571208, | |
| "grad_norm": 3.7923641204833984, | |
| "learning_rate": 2.505938242280285e-06, | |
| "loss": 2.8718, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.9502316189571208, | |
| "eval_accuracy": 0.7852508588547956, | |
| "eval_loss": 0.8441482186317444, | |
| "eval_runtime": 1352.0218, | |
| "eval_samples_per_second": 73.411, | |
| "eval_steps_per_second": 0.656, | |
| "step": 4000 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 4210, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "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": 1.5106917580956238e+20, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |