| { |
| "best_global_step": 40, |
| "best_metric": 2.39176344871521, |
| "best_model_checkpoint": "./outputs/astrid-llama-8b/checkpoint-40", |
| "epoch": 13.444444444444445, |
| "eval_steps": 5, |
| "global_step": 40, |
| "is_hyper_param_search": false, |
| "is_local_process_zero": true, |
| "is_world_process_zero": true, |
| "log_history": [ |
| { |
| "epoch": 0, |
| "eval_loss": 3.2546439170837402, |
| "eval_runtime": 3.0517, |
| "eval_samples_per_second": 7.209, |
| "eval_steps_per_second": 3.605, |
| "memory/device_reserved (GiB)": 12.15, |
| "memory/max_active (GiB)": 11.95, |
| "memory/max_allocated (GiB)": 11.95, |
| "step": 0 |
| }, |
| { |
| "epoch": 1.8888888888888888, |
| "grad_norm": 9.879115104675293, |
| "learning_rate": 8.000000000000001e-06, |
| "loss": 3.1725, |
| "memory/device_reserved (GiB)": 17.46, |
| "memory/max_active (GiB)": 15.28, |
| "memory/max_allocated (GiB)": 15.28, |
| "step": 5, |
| "tokens_per_second_per_gpu": 606.65 |
| }, |
| { |
| "epoch": 1.8888888888888888, |
| "eval_loss": 3.214911460876465, |
| "eval_runtime": 2.571, |
| "eval_samples_per_second": 8.557, |
| "eval_steps_per_second": 4.278, |
| "memory/device_reserved (GiB)": 17.46, |
| "memory/max_active (GiB)": 12.13, |
| "memory/max_allocated (GiB)": 12.13, |
| "step": 5 |
| }, |
| { |
| "epoch": 3.4444444444444446, |
| "grad_norm": 8.826162338256836, |
| "learning_rate": 1.8e-05, |
| "loss": 3.1171, |
| "memory/device_reserved (GiB)": 16.52, |
| "memory/max_active (GiB)": 15.28, |
| "memory/max_allocated (GiB)": 15.28, |
| "step": 10, |
| "tokens_per_second_per_gpu": 453.74 |
| }, |
| { |
| "epoch": 3.4444444444444446, |
| "eval_loss": 2.9428627490997314, |
| "eval_runtime": 2.5614, |
| "eval_samples_per_second": 8.589, |
| "eval_steps_per_second": 4.295, |
| "memory/device_reserved (GiB)": 16.52, |
| "memory/max_active (GiB)": 12.13, |
| "memory/max_allocated (GiB)": 12.13, |
| "step": 10 |
| }, |
| { |
| "epoch": 5.0, |
| "grad_norm": 9.603759765625, |
| "learning_rate": 1.913545457642601e-05, |
| "loss": 2.6355, |
| "memory/device_reserved (GiB)": 16.52, |
| "memory/max_active (GiB)": 15.28, |
| "memory/max_allocated (GiB)": 15.28, |
| "step": 15, |
| "tokens_per_second_per_gpu": 2687.33 |
| }, |
| { |
| "epoch": 5.0, |
| "eval_loss": 2.639833688735962, |
| "eval_runtime": 2.5641, |
| "eval_samples_per_second": 8.58, |
| "eval_steps_per_second": 4.29, |
| "memory/device_reserved (GiB)": 16.52, |
| "memory/max_active (GiB)": 12.13, |
| "memory/max_allocated (GiB)": 12.13, |
| "step": 15 |
| }, |
| { |
| "epoch": 6.888888888888889, |
| "grad_norm": 4.490882873535156, |
| "learning_rate": 1.5877852522924733e-05, |
| "loss": 2.3752, |
| "memory/device_reserved (GiB)": 16.52, |
| "memory/max_active (GiB)": 15.28, |
| "memory/max_allocated (GiB)": 15.28, |
| "step": 20, |
| "tokens_per_second_per_gpu": 575.3 |
| }, |
| { |
| "epoch": 6.888888888888889, |
| "eval_loss": 2.5206339359283447, |
| "eval_runtime": 2.5709, |
| "eval_samples_per_second": 8.557, |
| "eval_steps_per_second": 4.279, |
| "memory/device_reserved (GiB)": 16.52, |
| "memory/max_active (GiB)": 12.13, |
| "memory/max_allocated (GiB)": 12.13, |
| "step": 20 |
| }, |
| { |
| "epoch": 8.444444444444445, |
| "grad_norm": 3.118549346923828, |
| "learning_rate": 1.1045284632676535e-05, |
| "loss": 2.1869, |
| "memory/device_reserved (GiB)": 16.52, |
| "memory/max_active (GiB)": 15.28, |
| "memory/max_allocated (GiB)": 15.28, |
| "step": 25, |
| "tokens_per_second_per_gpu": 447.65 |
| }, |
| { |
| "epoch": 8.444444444444445, |
| "eval_loss": 2.446424961090088, |
| "eval_runtime": 2.572, |
| "eval_samples_per_second": 8.554, |
| "eval_steps_per_second": 4.277, |
| "memory/device_reserved (GiB)": 16.52, |
| "memory/max_active (GiB)": 12.13, |
| "memory/max_allocated (GiB)": 12.13, |
| "step": 25 |
| }, |
| { |
| "epoch": 10.0, |
| "grad_norm": 7.509883880615234, |
| "learning_rate": 5.932633569242e-06, |
| "loss": 2.0751, |
| "memory/device_reserved (GiB)": 16.52, |
| "memory/max_active (GiB)": 15.28, |
| "memory/max_allocated (GiB)": 15.28, |
| "step": 30, |
| "tokens_per_second_per_gpu": 2704.46 |
| }, |
| { |
| "epoch": 10.0, |
| "eval_loss": 2.41866397857666, |
| "eval_runtime": 2.5749, |
| "eval_samples_per_second": 8.544, |
| "eval_steps_per_second": 4.272, |
| "memory/device_reserved (GiB)": 16.52, |
| "memory/max_active (GiB)": 12.13, |
| "memory/max_allocated (GiB)": 12.13, |
| "step": 30 |
| }, |
| { |
| "epoch": 11.88888888888889, |
| "grad_norm": 3.0285682678222656, |
| "learning_rate": 1.9098300562505266e-06, |
| "loss": 2.0616, |
| "memory/device_reserved (GiB)": 16.52, |
| "memory/max_active (GiB)": 15.28, |
| "memory/max_allocated (GiB)": 15.28, |
| "step": 35, |
| "tokens_per_second_per_gpu": 575.33 |
| }, |
| { |
| "epoch": 11.88888888888889, |
| "eval_loss": 2.40836238861084, |
| "eval_runtime": 2.5669, |
| "eval_samples_per_second": 8.571, |
| "eval_steps_per_second": 4.285, |
| "memory/device_reserved (GiB)": 16.52, |
| "memory/max_active (GiB)": 12.13, |
| "memory/max_allocated (GiB)": 12.13, |
| "step": 35 |
| }, |
| { |
| "epoch": 13.444444444444445, |
| "grad_norm": 3.364469528198242, |
| "learning_rate": 5.4781046317267103e-08, |
| "loss": 2.0263, |
| "memory/device_reserved (GiB)": 16.52, |
| "memory/max_active (GiB)": 15.28, |
| "memory/max_allocated (GiB)": 15.28, |
| "step": 40, |
| "tokens_per_second_per_gpu": 447.5 |
| }, |
| { |
| "epoch": 13.444444444444445, |
| "eval_loss": 2.39176344871521, |
| "eval_runtime": 2.5947, |
| "eval_samples_per_second": 8.479, |
| "eval_steps_per_second": 4.239, |
| "memory/device_reserved (GiB)": 16.52, |
| "memory/max_active (GiB)": 12.13, |
| "memory/max_allocated (GiB)": 12.13, |
| "step": 40 |
| } |
| ], |
| "logging_steps": 5, |
| "max_steps": 40, |
| "num_input_tokens_seen": 0, |
| "num_train_epochs": 14, |
| "save_steps": 100, |
| "stateful_callbacks": { |
| "EarlyStoppingCallback": { |
| "args": { |
| "early_stopping_patience": 2, |
| "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": 2.301236913949901e+16, |
| "train_batch_size": 2, |
| "trial_name": null, |
| "trial_params": null |
| } |
|
|
