| { | |
| "best_global_step": 200, | |
| "best_metric": 0.7061131596565247, | |
| "best_model_checkpoint": "j05hr3d/SFT-Qwen2.5-Coder-7B_v1.1/checkpoint-200", | |
| "epoch": 3.0, | |
| "eval_steps": 20, | |
| "global_step": 210, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.2857142857142857, | |
| "grad_norm": 0.24874337017536163, | |
| "learning_rate": 9.408866995073891e-05, | |
| "loss": 0.9132, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.2857142857142857, | |
| "eval_loss": 0.8498335480690002, | |
| "eval_runtime": 20.8728, | |
| "eval_samples_per_second": 2.97, | |
| "eval_steps_per_second": 0.383, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.5714285714285714, | |
| "grad_norm": 0.24814729392528534, | |
| "learning_rate": 8.423645320197044e-05, | |
| "loss": 0.9442, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.5714285714285714, | |
| "eval_loss": 0.7972759008407593, | |
| "eval_runtime": 18.6915, | |
| "eval_samples_per_second": 3.317, | |
| "eval_steps_per_second": 0.428, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.8571428571428571, | |
| "grad_norm": 0.3184037208557129, | |
| "learning_rate": 7.438423645320197e-05, | |
| "loss": 0.7474, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.8571428571428571, | |
| "eval_loss": 0.7688223123550415, | |
| "eval_runtime": 18.6846, | |
| "eval_samples_per_second": 3.318, | |
| "eval_steps_per_second": 0.428, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 1.1428571428571428, | |
| "grad_norm": 0.2819560468196869, | |
| "learning_rate": 6.45320197044335e-05, | |
| "loss": 0.755, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 1.1428571428571428, | |
| "eval_loss": 0.7504354119300842, | |
| "eval_runtime": 18.6934, | |
| "eval_samples_per_second": 3.317, | |
| "eval_steps_per_second": 0.428, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 1.4285714285714286, | |
| "grad_norm": 0.349388062953949, | |
| "learning_rate": 5.467980295566503e-05, | |
| "loss": 0.7374, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.4285714285714286, | |
| "eval_loss": 0.7384942770004272, | |
| "eval_runtime": 18.7081, | |
| "eval_samples_per_second": 3.314, | |
| "eval_steps_per_second": 0.428, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.7142857142857144, | |
| "grad_norm": 0.4439921975135803, | |
| "learning_rate": 4.482758620689655e-05, | |
| "loss": 0.6531, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 1.7142857142857144, | |
| "eval_loss": 0.7255800366401672, | |
| "eval_runtime": 18.704, | |
| "eval_samples_per_second": 3.315, | |
| "eval_steps_per_second": 0.428, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 1.2030707597732544, | |
| "learning_rate": 3.497536945812808e-05, | |
| "loss": 0.6193, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_loss": 0.7122625112533569, | |
| "eval_runtime": 18.6953, | |
| "eval_samples_per_second": 3.316, | |
| "eval_steps_per_second": 0.428, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 2.2857142857142856, | |
| "grad_norm": 0.7790128588676453, | |
| "learning_rate": 2.512315270935961e-05, | |
| "loss": 0.5379, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 2.2857142857142856, | |
| "eval_loss": 0.7121083736419678, | |
| "eval_runtime": 18.7059, | |
| "eval_samples_per_second": 3.314, | |
| "eval_steps_per_second": 0.428, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 2.571428571428571, | |
| "grad_norm": 0.5455464720726013, | |
| "learning_rate": 1.5270935960591133e-05, | |
| "loss": 0.5749, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 2.571428571428571, | |
| "eval_loss": 0.71072918176651, | |
| "eval_runtime": 18.7175, | |
| "eval_samples_per_second": 3.312, | |
| "eval_steps_per_second": 0.427, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 2.857142857142857, | |
| "grad_norm": 0.8019999861717224, | |
| "learning_rate": 5.418719211822661e-06, | |
| "loss": 0.7175, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 2.857142857142857, | |
| "eval_loss": 0.7061131596565247, | |
| "eval_runtime": 18.7094, | |
| "eval_samples_per_second": 3.314, | |
| "eval_steps_per_second": 0.428, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "step": 210, | |
| "total_flos": 7.87369399526062e+16, | |
| "train_loss": 0.7112103757404146, | |
| "train_runtime": 1206.0601, | |
| "train_samples_per_second": 1.39, | |
| "train_steps_per_second": 0.174 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_loss": 0.7061131596565247, | |
| "eval_runtime": 18.6112, | |
| "eval_samples_per_second": 3.331, | |
| "eval_steps_per_second": 0.43, | |
| "step": 210 | |
| } | |
| ], | |
| "logging_steps": 20, | |
| "max_steps": 210, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "save_steps": 20, | |
| "stateful_callbacks": { | |
| "EarlyStoppingCallback": { | |
| "args": { | |
| "early_stopping_patience": 3, | |
| "early_stopping_threshold": 0.0 | |
| }, | |
| "attributes": { | |
| "early_stopping_patience_counter": 0 | |
| } | |
| }, | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 7.87369399526062e+16, | |
| "train_batch_size": 2, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |