| { | |
| "best_global_step": 180, | |
| "best_metric": 0.9728425145149231, | |
| "best_model_checkpoint": "j05hr3d/SFT-Qwen2.5-Coder-3B/checkpoint-180", | |
| "epoch": 3.582089552238806, | |
| "eval_steps": 20, | |
| "global_step": 240, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.29850746268656714, | |
| "grad_norm": 0.48786428570747375, | |
| "learning_rate": 9.753086419753087e-05, | |
| "loss": 1.0721, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.29850746268656714, | |
| "eval_loss": 1.1757386922836304, | |
| "eval_runtime": 10.4355, | |
| "eval_samples_per_second": 5.75, | |
| "eval_steps_per_second": 0.767, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.5970149253731343, | |
| "grad_norm": 0.688591718673706, | |
| "learning_rate": 9.135802469135802e-05, | |
| "loss": 0.8989, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.5970149253731343, | |
| "eval_loss": 1.1058913469314575, | |
| "eval_runtime": 8.5976, | |
| "eval_samples_per_second": 6.979, | |
| "eval_steps_per_second": 0.93, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.8955223880597015, | |
| "grad_norm": 0.4469507336616516, | |
| "learning_rate": 8.518518518518518e-05, | |
| "loss": 0.8293, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.8955223880597015, | |
| "eval_loss": 1.0656379461288452, | |
| "eval_runtime": 8.5963, | |
| "eval_samples_per_second": 6.98, | |
| "eval_steps_per_second": 0.931, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 1.1940298507462686, | |
| "grad_norm": 0.3167116641998291, | |
| "learning_rate": 7.901234567901235e-05, | |
| "loss": 0.787, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 1.1940298507462686, | |
| "eval_loss": 1.0364410877227783, | |
| "eval_runtime": 8.6008, | |
| "eval_samples_per_second": 6.976, | |
| "eval_steps_per_second": 0.93, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 1.4925373134328357, | |
| "grad_norm": 0.588107705116272, | |
| "learning_rate": 7.283950617283951e-05, | |
| "loss": 0.7025, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.4925373134328357, | |
| "eval_loss": 1.0205715894699097, | |
| "eval_runtime": 8.6057, | |
| "eval_samples_per_second": 6.972, | |
| "eval_steps_per_second": 0.93, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.7910447761194028, | |
| "grad_norm": 0.6192528009414673, | |
| "learning_rate": 6.666666666666667e-05, | |
| "loss": 0.7386, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 1.7910447761194028, | |
| "eval_loss": 0.9961486458778381, | |
| "eval_runtime": 8.5981, | |
| "eval_samples_per_second": 6.978, | |
| "eval_steps_per_second": 0.93, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 2.08955223880597, | |
| "grad_norm": 0.49059680104255676, | |
| "learning_rate": 6.049382716049383e-05, | |
| "loss": 0.7471, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 2.08955223880597, | |
| "eval_loss": 0.9915516972541809, | |
| "eval_runtime": 8.6047, | |
| "eval_samples_per_second": 6.973, | |
| "eval_steps_per_second": 0.93, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 2.388059701492537, | |
| "grad_norm": 0.41391539573669434, | |
| "learning_rate": 5.4320987654320986e-05, | |
| "loss": 0.624, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 2.388059701492537, | |
| "eval_loss": 0.9842909574508667, | |
| "eval_runtime": 8.6, | |
| "eval_samples_per_second": 6.977, | |
| "eval_steps_per_second": 0.93, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 2.6865671641791042, | |
| "grad_norm": 0.5299602746963501, | |
| "learning_rate": 4.814814814814815e-05, | |
| "loss": 0.6839, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 2.6865671641791042, | |
| "eval_loss": 0.9728425145149231, | |
| "eval_runtime": 8.6028, | |
| "eval_samples_per_second": 6.974, | |
| "eval_steps_per_second": 0.93, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 2.9850746268656714, | |
| "grad_norm": 0.6335314512252808, | |
| "learning_rate": 4.197530864197531e-05, | |
| "loss": 0.6561, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 2.9850746268656714, | |
| "eval_loss": 0.973730206489563, | |
| "eval_runtime": 8.5961, | |
| "eval_samples_per_second": 6.98, | |
| "eval_steps_per_second": 0.931, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.283582089552239, | |
| "grad_norm": 0.6432749032974243, | |
| "learning_rate": 3.580246913580247e-05, | |
| "loss": 0.6027, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 3.283582089552239, | |
| "eval_loss": 0.9784586429595947, | |
| "eval_runtime": 8.5988, | |
| "eval_samples_per_second": 6.978, | |
| "eval_steps_per_second": 0.93, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 3.582089552238806, | |
| "grad_norm": 0.638783872127533, | |
| "learning_rate": 2.962962962962963e-05, | |
| "loss": 0.5221, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 3.582089552238806, | |
| "eval_loss": 0.9842756986618042, | |
| "eval_runtime": 8.5989, | |
| "eval_samples_per_second": 6.978, | |
| "eval_steps_per_second": 0.93, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 3.582089552238806, | |
| "step": 240, | |
| "total_flos": 3.78148862442455e+16, | |
| "train_loss": 0.7386853178342183, | |
| "train_runtime": 1029.4211, | |
| "train_samples_per_second": 2.599, | |
| "train_steps_per_second": 0.325 | |
| }, | |
| { | |
| "epoch": 3.582089552238806, | |
| "eval_loss": 0.9728425145149231, | |
| "eval_runtime": 8.6231, | |
| "eval_samples_per_second": 6.958, | |
| "eval_steps_per_second": 0.928, | |
| "step": 240 | |
| } | |
| ], | |
| "logging_steps": 20, | |
| "max_steps": 335, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 5, | |
| "save_steps": 20, | |
| "stateful_callbacks": { | |
| "EarlyStoppingCallback": { | |
| "args": { | |
| "early_stopping_patience": 3, | |
| "early_stopping_threshold": 0.0 | |
| }, | |
| "attributes": { | |
| "early_stopping_patience_counter": 3 | |
| } | |
| }, | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 3.78148862442455e+16, | |
| "train_batch_size": 2, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |