|
{ |
|
"best_metric": null, |
|
"best_model_checkpoint": null, |
|
"epoch": 8.0, |
|
"eval_steps": 500, |
|
"global_step": 2000, |
|
"is_hyper_param_search": false, |
|
"is_local_process_zero": true, |
|
"is_world_process_zero": true, |
|
"log_history": [ |
|
{ |
|
"epoch": 0.4, |
|
"grad_norm": 0.6251392960548401, |
|
"learning_rate": 5e-05, |
|
"loss": 2.427, |
|
"step": 100 |
|
}, |
|
{ |
|
"epoch": 0.8, |
|
"grad_norm": 0.6590011715888977, |
|
"learning_rate": 5e-05, |
|
"loss": 2.2748, |
|
"step": 200 |
|
}, |
|
{ |
|
"epoch": 1.0, |
|
"eval_accuracy": 0.5142539682539683, |
|
"eval_loss": 2.336690902709961, |
|
"eval_runtime": 5.4733, |
|
"eval_samples_per_second": 91.353, |
|
"eval_steps_per_second": 11.51, |
|
"step": 250 |
|
}, |
|
{ |
|
"epoch": 1.0, |
|
"eval_exact_match": 16.8, |
|
"eval_f1": 24.850246583481884, |
|
"step": 250 |
|
}, |
|
{ |
|
"epoch": 1.2, |
|
"grad_norm": 0.8194239139556885, |
|
"learning_rate": 5e-05, |
|
"loss": 2.2322, |
|
"step": 300 |
|
}, |
|
{ |
|
"epoch": 1.6, |
|
"grad_norm": 0.9867294430732727, |
|
"learning_rate": 5e-05, |
|
"loss": 2.1904, |
|
"step": 400 |
|
}, |
|
{ |
|
"epoch": 2.0, |
|
"grad_norm": 0.9660400152206421, |
|
"learning_rate": 5e-05, |
|
"loss": 2.1657, |
|
"step": 500 |
|
}, |
|
{ |
|
"epoch": 2.0, |
|
"eval_accuracy": 0.5168253968253969, |
|
"eval_loss": 2.323688507080078, |
|
"eval_runtime": 5.1683, |
|
"eval_samples_per_second": 96.744, |
|
"eval_steps_per_second": 12.19, |
|
"step": 500 |
|
}, |
|
{ |
|
"epoch": 2.0, |
|
"eval_exact_match": 17.2, |
|
"eval_f1": 24.620036075036083, |
|
"step": 500 |
|
}, |
|
{ |
|
"epoch": 2.4, |
|
"grad_norm": 1.281128168106079, |
|
"learning_rate": 5e-05, |
|
"loss": 2.0251, |
|
"step": 600 |
|
}, |
|
{ |
|
"epoch": 2.8, |
|
"grad_norm": 1.3365116119384766, |
|
"learning_rate": 5e-05, |
|
"loss": 2.0158, |
|
"step": 700 |
|
}, |
|
{ |
|
"epoch": 3.0, |
|
"eval_accuracy": 0.5156507936507937, |
|
"eval_loss": 2.3560824394226074, |
|
"eval_runtime": 5.4675, |
|
"eval_samples_per_second": 91.449, |
|
"eval_steps_per_second": 11.523, |
|
"step": 750 |
|
}, |
|
{ |
|
"epoch": 3.0, |
|
"eval_exact_match": 19.0, |
|
"eval_f1": 27.242790688843332, |
|
"step": 750 |
|
}, |
|
{ |
|
"epoch": 3.2, |
|
"grad_norm": 1.7840828895568848, |
|
"learning_rate": 5e-05, |
|
"loss": 1.9394, |
|
"step": 800 |
|
}, |
|
{ |
|
"epoch": 3.6, |
|
"grad_norm": 1.8030190467834473, |
|
"learning_rate": 5e-05, |
|
"loss": 1.8502, |
|
"step": 900 |
|
}, |
|
{ |
|
"epoch": 4.0, |
|
"grad_norm": 1.9682222604751587, |
|
"learning_rate": 5e-05, |
|
"loss": 1.8682, |
|
"step": 1000 |
|
}, |
|
{ |
|
"epoch": 4.0, |
|
"eval_accuracy": 0.5131746031746032, |
|
"eval_loss": 2.4071946144104004, |
|
"eval_runtime": 5.2825, |
|
"eval_samples_per_second": 94.653, |
|
"eval_steps_per_second": 11.926, |
|
"step": 1000 |
|
}, |
|
{ |
|
"epoch": 4.0, |
|
"eval_exact_match": 17.6, |
|
"eval_f1": 26.010988455988468, |
|
"step": 1000 |
|
}, |
|
{ |
|
"epoch": 4.4, |
|
"grad_norm": 2.138505220413208, |
|
"learning_rate": 5e-05, |
|
"loss": 1.6795, |
|
"step": 1100 |
|
}, |
|
{ |
|
"epoch": 4.8, |
|
"grad_norm": 2.3212926387786865, |
|
"learning_rate": 5e-05, |
|
"loss": 1.7036, |
|
"step": 1200 |
|
}, |
|
{ |
|
"epoch": 5.0, |
|
"eval_accuracy": 0.5099365079365079, |
|
"eval_loss": 2.494770050048828, |
|
"eval_runtime": 5.1799, |
|
"eval_samples_per_second": 96.528, |
|
"eval_steps_per_second": 12.162, |
|
"step": 1250 |
|
}, |
|
{ |
|
"epoch": 5.0, |
|
"eval_exact_match": 17.2, |
|
"eval_f1": 25.84752525252526, |
|
"step": 1250 |
|
}, |
|
{ |
|
"epoch": 5.2, |
|
"grad_norm": 2.4878485202789307, |
|
"learning_rate": 5e-05, |
|
"loss": 1.6305, |
|
"step": 1300 |
|
}, |
|
{ |
|
"epoch": 5.6, |
|
"grad_norm": 2.599329710006714, |
|
"learning_rate": 5e-05, |
|
"loss": 1.5477, |
|
"step": 1400 |
|
}, |
|
{ |
|
"epoch": 6.0, |
|
"grad_norm": 2.5542759895324707, |
|
"learning_rate": 5e-05, |
|
"loss": 1.5708, |
|
"step": 1500 |
|
}, |
|
{ |
|
"epoch": 6.0, |
|
"eval_accuracy": 0.5065079365079365, |
|
"eval_loss": 2.5797038078308105, |
|
"eval_runtime": 5.6827, |
|
"eval_samples_per_second": 87.986, |
|
"eval_steps_per_second": 11.086, |
|
"step": 1500 |
|
}, |
|
{ |
|
"epoch": 6.0, |
|
"eval_exact_match": 17.0, |
|
"eval_f1": 25.310858585858586, |
|
"step": 1500 |
|
}, |
|
{ |
|
"epoch": 6.4, |
|
"grad_norm": 3.1968846321105957, |
|
"learning_rate": 5e-05, |
|
"loss": 1.3896, |
|
"step": 1600 |
|
}, |
|
{ |
|
"epoch": 6.8, |
|
"grad_norm": 3.8637375831604004, |
|
"learning_rate": 5e-05, |
|
"loss": 1.421, |
|
"step": 1700 |
|
}, |
|
{ |
|
"epoch": 7.0, |
|
"eval_accuracy": 0.5038412698412699, |
|
"eval_loss": 2.6977317333221436, |
|
"eval_runtime": 5.3312, |
|
"eval_samples_per_second": 93.787, |
|
"eval_steps_per_second": 11.817, |
|
"step": 1750 |
|
}, |
|
{ |
|
"epoch": 7.0, |
|
"eval_exact_match": 16.0, |
|
"eval_f1": 23.99441558441559, |
|
"step": 1750 |
|
}, |
|
{ |
|
"epoch": 7.2, |
|
"grad_norm": 3.468459367752075, |
|
"learning_rate": 5e-05, |
|
"loss": 1.3366, |
|
"step": 1800 |
|
}, |
|
{ |
|
"epoch": 7.6, |
|
"grad_norm": 3.5727667808532715, |
|
"learning_rate": 5e-05, |
|
"loss": 1.2585, |
|
"step": 1900 |
|
}, |
|
{ |
|
"epoch": 8.0, |
|
"grad_norm": 3.508258819580078, |
|
"learning_rate": 5e-05, |
|
"loss": 1.296, |
|
"step": 2000 |
|
}, |
|
{ |
|
"epoch": 8.0, |
|
"eval_accuracy": 0.5013333333333333, |
|
"eval_loss": 2.786928653717041, |
|
"eval_runtime": 5.8398, |
|
"eval_samples_per_second": 85.62, |
|
"eval_steps_per_second": 10.788, |
|
"step": 2000 |
|
}, |
|
{ |
|
"epoch": 8.0, |
|
"eval_exact_match": 15.4, |
|
"eval_f1": 22.542871572871572, |
|
"step": 2000 |
|
} |
|
], |
|
"logging_steps": 100, |
|
"max_steps": 12500, |
|
"num_input_tokens_seen": 0, |
|
"num_train_epochs": 50, |
|
"save_steps": 500, |
|
"stateful_callbacks": { |
|
"TrainerControl": { |
|
"args": { |
|
"should_epoch_stop": false, |
|
"should_evaluate": false, |
|
"should_log": false, |
|
"should_save": true, |
|
"should_training_stop": false |
|
}, |
|
"attributes": {} |
|
} |
|
}, |
|
"total_flos": 5.174154322404966e+16, |
|
"train_batch_size": 1, |
|
"trial_name": null, |
|
"trial_params": null |
|
} |
|
|