SLM_vs_LLM_experiments
/
max_seq_length_128_experiments
/google_t5
/t5_small_twitter
/checkpoint-200
/trainer_state.json
{ | |
"best_metric": 0.49757373332977295, | |
"best_model_checkpoint": "../../experiments_checkpoints/MAdAiLab/google_t5/t5_small_twitter/checkpoint-200", | |
"epoch": 0.7352941176470589, | |
"eval_steps": 50, | |
"global_step": 200, | |
"is_hyper_param_search": false, | |
"is_local_process_zero": true, | |
"is_world_process_zero": true, | |
"log_history": [ | |
{ | |
"epoch": 0.04, | |
"grad_norm": 0.9986526370048523, | |
"learning_rate": 0.0004938725490196079, | |
"loss": 0.6031, | |
"step": 10 | |
}, | |
{ | |
"epoch": 0.07, | |
"grad_norm": 0.9707651734352112, | |
"learning_rate": 0.0004877450980392157, | |
"loss": 0.5921, | |
"step": 20 | |
}, | |
{ | |
"epoch": 0.11, | |
"grad_norm": 1.3216536045074463, | |
"learning_rate": 0.00048161764705882356, | |
"loss": 0.5945, | |
"step": 30 | |
}, | |
{ | |
"epoch": 0.15, | |
"grad_norm": 0.8769760131835938, | |
"learning_rate": 0.00047549019607843134, | |
"loss": 0.5244, | |
"step": 40 | |
}, | |
{ | |
"epoch": 0.18, | |
"grad_norm": 1.3491202592849731, | |
"learning_rate": 0.0004693627450980392, | |
"loss": 0.4779, | |
"step": 50 | |
}, | |
{ | |
"epoch": 0.18, | |
"eval_accuracy": 0.7564338235294118, | |
"eval_f1_macro": 0.7177028105371891, | |
"eval_f1_micro": 0.7564338235294118, | |
"eval_loss": 0.5150883793830872, | |
"eval_runtime": 0.9275, | |
"eval_samples_per_second": 1173.066, | |
"eval_steps_per_second": 36.658, | |
"step": 50 | |
}, | |
{ | |
"epoch": 0.22, | |
"grad_norm": 2.29738450050354, | |
"learning_rate": 0.0004632352941176471, | |
"loss": 0.4934, | |
"step": 60 | |
}, | |
{ | |
"epoch": 0.26, | |
"grad_norm": 3.7542002201080322, | |
"learning_rate": 0.0004571078431372549, | |
"loss": 0.5007, | |
"step": 70 | |
}, | |
{ | |
"epoch": 0.29, | |
"grad_norm": 1.893402099609375, | |
"learning_rate": 0.0004509803921568628, | |
"loss": 0.4738, | |
"step": 80 | |
}, | |
{ | |
"epoch": 0.33, | |
"grad_norm": 2.0344367027282715, | |
"learning_rate": 0.00044485294117647056, | |
"loss": 0.4594, | |
"step": 90 | |
}, | |
{ | |
"epoch": 0.37, | |
"grad_norm": 2.5279059410095215, | |
"learning_rate": 0.00043872549019607844, | |
"loss": 0.5123, | |
"step": 100 | |
}, | |
{ | |
"epoch": 0.37, | |
"eval_accuracy": 0.7527573529411765, | |
"eval_f1_macro": 0.6986562303405088, | |
"eval_f1_micro": 0.7527573529411765, | |
"eval_loss": 0.5059758424758911, | |
"eval_runtime": 0.8718, | |
"eval_samples_per_second": 1248.006, | |
"eval_steps_per_second": 39.0, | |
"step": 100 | |
}, | |
{ | |
"epoch": 0.4, | |
"grad_norm": 1.4078208208084106, | |
"learning_rate": 0.0004325980392156863, | |
"loss": 0.516, | |
"step": 110 | |
}, | |
{ | |
"epoch": 0.44, | |
"grad_norm": 2.224971055984497, | |
"learning_rate": 0.0004264705882352941, | |
"loss": 0.5001, | |
"step": 120 | |
}, | |
{ | |
"epoch": 0.48, | |
"grad_norm": 1.39370858669281, | |
"learning_rate": 0.000420343137254902, | |
"loss": 0.4835, | |
"step": 130 | |
}, | |
{ | |
"epoch": 0.51, | |
"grad_norm": 0.9512380361557007, | |
"learning_rate": 0.0004142156862745098, | |
"loss": 0.5218, | |
"step": 140 | |
}, | |
{ | |
"epoch": 0.55, | |
"grad_norm": 1.359387755393982, | |
"learning_rate": 0.00040808823529411766, | |
"loss": 0.4617, | |
"step": 150 | |
}, | |
{ | |
"epoch": 0.55, | |
"eval_accuracy": 0.7270220588235294, | |
"eval_f1_macro": 0.614936097953347, | |
"eval_f1_micro": 0.7270220588235294, | |
"eval_loss": 0.5286622047424316, | |
"eval_runtime": 0.9283, | |
"eval_samples_per_second": 1172.017, | |
"eval_steps_per_second": 36.626, | |
"step": 150 | |
}, | |
{ | |
"epoch": 0.59, | |
"grad_norm": 1.1093153953552246, | |
"learning_rate": 0.0004019607843137255, | |
"loss": 0.4961, | |
"step": 160 | |
}, | |
{ | |
"epoch": 0.62, | |
"grad_norm": 2.4549243450164795, | |
"learning_rate": 0.0003958333333333333, | |
"loss": 0.5351, | |
"step": 170 | |
}, | |
{ | |
"epoch": 0.66, | |
"grad_norm": 1.2358920574188232, | |
"learning_rate": 0.0003897058823529412, | |
"loss": 0.5129, | |
"step": 180 | |
}, | |
{ | |
"epoch": 0.7, | |
"grad_norm": 1.4391891956329346, | |
"learning_rate": 0.00038357843137254904, | |
"loss": 0.4945, | |
"step": 190 | |
}, | |
{ | |
"epoch": 0.74, | |
"grad_norm": 1.3604400157928467, | |
"learning_rate": 0.0003774509803921569, | |
"loss": 0.4942, | |
"step": 200 | |
}, | |
{ | |
"epoch": 0.74, | |
"eval_accuracy": 0.7444852941176471, | |
"eval_f1_macro": 0.6956024668695554, | |
"eval_f1_micro": 0.7444852941176471, | |
"eval_loss": 0.49757373332977295, | |
"eval_runtime": 0.9302, | |
"eval_samples_per_second": 1169.596, | |
"eval_steps_per_second": 36.55, | |
"step": 200 | |
} | |
], | |
"logging_steps": 10, | |
"max_steps": 816, | |
"num_input_tokens_seen": 0, | |
"num_train_epochs": 3, | |
"save_steps": 50, | |
"total_flos": 217842922291200.0, | |
"train_batch_size": 16, | |
"trial_name": null, | |
"trial_params": null | |
} | |