305 lines
9.0 KiB
JSON
305 lines
9.0 KiB
JSON
|
{
|
||
|
|
"best_global_step": 15,
|
||
|
|
"best_metric": 0.7146788990825688,
|
||
|
|
"best_model_checkpoint": "/root/BoolQ_Llama-3.2-1B/None/26t8ytsb_dazzling-shape-27/checkpoint-15",
|
||
|
|
"epoch": 5.0,
|
||
|
|
"eval_steps": 500,
|
||
|
|
"global_step": 15,
|
||
|
|
"is_hyper_param_search": false,
|
||
|
|
"is_local_process_zero": true,
|
||
|
|
"is_world_process_zero": true,
|
||
|
|
"log_history": [
|
||
|
|
{
|
||
|
|
"epoch": 0,
|
||
|
|
"eval_accumulated_loss": 2315.0376377105713,
|
||
|
|
"eval_accuracy": 0.6214067278287462,
|
||
|
|
"eval_accuracy_2822": 0.01949634443541836,
|
||
|
|
"eval_accuracy_9642": 0.9911154985192497,
|
||
|
|
"eval_correct_gen_preds": 2040.0,
|
||
|
|
"eval_correct_gen_preds_2822": 24.0,
|
||
|
|
"eval_correct_gen_preds_9642": 2007.0,
|
||
|
|
"eval_correct_preds": 2032.0,
|
||
|
|
"eval_correct_preds_2822": 24.0,
|
||
|
|
"eval_correct_preds_9642": 2008.0,
|
||
|
|
"eval_gen_accuracy": 0.6238532110091743,
|
||
|
|
"eval_gen_accuracy_2822": 0.01949634443541836,
|
||
|
|
"eval_gen_accuracy_9642": 0.9906219151036525,
|
||
|
|
"eval_loss": 0.7079626321792603,
|
||
|
|
"eval_mdl": 3339.893319396342,
|
||
|
|
"eval_model_preparation_time": 0.0059,
|
||
|
|
"eval_runtime": 17.2936,
|
||
|
|
"eval_samples_per_second": 189.088,
|
||
|
|
"eval_steps_per_second": 1.619,
|
||
|
|
"eval_total_labels_2822": 1231.0,
|
||
|
|
"eval_total_labels_9642": 2026.0,
|
||
|
|
"eval_total_preds": 3270.0,
|
||
|
|
"step": 0
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 0.3333333333333333,
|
||
|
|
"grad_norm": 19.625,
|
||
|
|
"learning_rate": 0.0,
|
||
|
|
"loss": 0.5234,
|
||
|
|
"step": 1
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 0.6666666666666666,
|
||
|
|
"grad_norm": 44.25,
|
||
|
|
"learning_rate": 6.666666666666667e-06,
|
||
|
|
"loss": 0.7261,
|
||
|
|
"step": 2
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 1.0,
|
||
|
|
"grad_norm": 19.75,
|
||
|
|
"learning_rate": 1.3333333333333333e-05,
|
||
|
|
"loss": 0.5194,
|
||
|
|
"step": 3
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 1.0,
|
||
|
|
"eval_accumulated_loss": 2543.702173233032,
|
||
|
|
"eval_accuracy": 0.5678899082568807,
|
||
|
|
"eval_accuracy_2822": 0.8911454102355808,
|
||
|
|
"eval_accuracy_9642": 0.3751233958538993,
|
||
|
|
"eval_correct_gen_preds": 1701.0,
|
||
|
|
"eval_correct_gen_preds_2822": 1022.0,
|
||
|
|
"eval_correct_gen_preds_9642": 670.0,
|
||
|
|
"eval_correct_preds": 1857.0,
|
||
|
|
"eval_correct_preds_2822": 1097.0,
|
||
|
|
"eval_correct_preds_9642": 760.0,
|
||
|
|
"eval_gen_accuracy": 0.5201834862385321,
|
||
|
|
"eval_gen_accuracy_2822": 0.8302193338748984,
|
||
|
|
"eval_gen_accuracy_9642": 0.33070088845014806,
|
||
|
|
"eval_loss": 0.7778906226158142,
|
||
|
|
"eval_mdl": 3669.7865108217748,
|
||
|
|
"eval_model_preparation_time": 0.0059,
|
||
|
|
"eval_runtime": 17.8318,
|
||
|
|
"eval_samples_per_second": 183.381,
|
||
|
|
"eval_steps_per_second": 1.57,
|
||
|
|
"eval_total_labels_2822": 1231.0,
|
||
|
|
"eval_total_labels_9642": 2026.0,
|
||
|
|
"eval_total_preds": 3270.0,
|
||
|
|
"step": 3
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 1.3333333333333333,
|
||
|
|
"grad_norm": 83.5,
|
||
|
|
"learning_rate": 2e-05,
|
||
|
|
"loss": 0.7045,
|
||
|
|
"step": 4
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 1.6666666666666665,
|
||
|
|
"grad_norm": 102.5,
|
||
|
|
"learning_rate": 1.9999440560919153e-05,
|
||
|
|
"loss": 0.6378,
|
||
|
|
"step": 5
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 2.0,
|
||
|
|
"grad_norm": 58.5,
|
||
|
|
"learning_rate": 1.999776230627102e-05,
|
||
|
|
"loss": 0.2429,
|
||
|
|
"step": 6
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 2.0,
|
||
|
|
"eval_accumulated_loss": 2881.3052768707275,
|
||
|
|
"eval_accuracy": 0.6385321100917432,
|
||
|
|
"eval_accuracy_2822": 0.20389926888708368,
|
||
|
|
"eval_accuracy_9642": 0.9067127344521224,
|
||
|
|
"eval_correct_gen_preds": 1960.0,
|
||
|
|
"eval_correct_gen_preds_2822": 228.0,
|
||
|
|
"eval_correct_gen_preds_9642": 1723.0,
|
||
|
|
"eval_correct_preds": 2088.0,
|
||
|
|
"eval_correct_preds_2822": 251.0,
|
||
|
|
"eval_correct_preds_9642": 1837.0,
|
||
|
|
"eval_gen_accuracy": 0.599388379204893,
|
||
|
|
"eval_gen_accuracy_2822": 0.1852152721364744,
|
||
|
|
"eval_gen_accuracy_9642": 0.8504442250740375,
|
||
|
|
"eval_loss": 0.8811332583427429,
|
||
|
|
"eval_mdl": 4156.8448342286,
|
||
|
|
"eval_model_preparation_time": 0.0059,
|
||
|
|
"eval_runtime": 17.8483,
|
||
|
|
"eval_samples_per_second": 183.21,
|
||
|
|
"eval_steps_per_second": 1.569,
|
||
|
|
"eval_total_labels_2822": 1231.0,
|
||
|
|
"eval_total_labels_9642": 2026.0,
|
||
|
|
"eval_total_preds": 3270.0,
|
||
|
|
"step": 6
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 2.3333333333333335,
|
||
|
|
"grad_norm": 35.0,
|
||
|
|
"learning_rate": 1.9994965423831853e-05,
|
||
|
|
"loss": 0.251,
|
||
|
|
"step": 7
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 2.6666666666666665,
|
||
|
|
"grad_norm": 18.5,
|
||
|
|
"learning_rate": 1.999105022653872e-05,
|
||
|
|
"loss": 0.1897,
|
||
|
|
"step": 8
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 3.0,
|
||
|
|
"grad_norm": 20.125,
|
||
|
|
"learning_rate": 1.9986017152454497e-05,
|
||
|
|
"loss": 0.0895,
|
||
|
|
"step": 9
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 3.0,
|
||
|
|
"eval_accumulated_loss": 2496.1307315826416,
|
||
|
|
"eval_accuracy": 0.6886850152905198,
|
||
|
|
"eval_accuracy_2822": 0.5889520714865962,
|
||
|
|
"eval_accuracy_9642": 0.7537018756169793,
|
||
|
|
"eval_correct_gen_preds": 2036.0,
|
||
|
|
"eval_correct_gen_preds_2822": 670.0,
|
||
|
|
"eval_correct_gen_preds_9642": 1357.0,
|
||
|
|
"eval_correct_preds": 2252.0,
|
||
|
|
"eval_correct_preds_2822": 725.0,
|
||
|
|
"eval_correct_preds_9642": 1527.0,
|
||
|
|
"eval_gen_accuracy": 0.6226299694189602,
|
||
|
|
"eval_gen_accuracy_2822": 0.5442729488220959,
|
||
|
|
"eval_gen_accuracy_9642": 0.6697926949654491,
|
||
|
|
"eval_loss": 0.7633427977561951,
|
||
|
|
"eval_mdl": 3601.1554278648173,
|
||
|
|
"eval_model_preparation_time": 0.0059,
|
||
|
|
"eval_runtime": 18.2462,
|
||
|
|
"eval_samples_per_second": 179.216,
|
||
|
|
"eval_steps_per_second": 1.535,
|
||
|
|
"eval_total_labels_2822": 1231.0,
|
||
|
|
"eval_total_labels_9642": 2026.0,
|
||
|
|
"eval_total_preds": 3270.0,
|
||
|
|
"step": 9
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 3.3333333333333335,
|
||
|
|
"grad_norm": 9.4375,
|
||
|
|
"learning_rate": 1.9979866764718846e-05,
|
||
|
|
"loss": 0.0289,
|
||
|
|
"step": 10
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 3.6666666666666665,
|
||
|
|
"grad_norm": 1.1484375,
|
||
|
|
"learning_rate": 1.9972599751485225e-05,
|
||
|
|
"loss": 0.0049,
|
||
|
|
"step": 11
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 4.0,
|
||
|
|
"grad_norm": 207.0,
|
||
|
|
"learning_rate": 1.9964216925843876e-05,
|
||
|
|
"loss": 0.3695,
|
||
|
|
"step": 12
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 4.0,
|
||
|
|
"eval_accumulated_loss": 6579.203971862793,
|
||
|
|
"eval_accuracy": 0.6996941896024464,
|
||
|
|
"eval_accuracy_2822": 0.3013809910641755,
|
||
|
|
"eval_accuracy_9642": 0.9461994076999013,
|
||
|
|
"eval_correct_gen_preds": 2274.0,
|
||
|
|
"eval_correct_gen_preds_2822": 360.0,
|
||
|
|
"eval_correct_gen_preds_9642": 1905.0,
|
||
|
|
"eval_correct_preds": 2288.0,
|
||
|
|
"eval_correct_preds_2822": 371.0,
|
||
|
|
"eval_correct_preds_9642": 1917.0,
|
||
|
|
"eval_gen_accuracy": 0.6954128440366972,
|
||
|
|
"eval_gen_accuracy_2822": 0.2924451665312754,
|
||
|
|
"eval_gen_accuracy_9642": 0.9402764067127345,
|
||
|
|
"eval_loss": 2.011989116668701,
|
||
|
|
"eval_mdl": 9491.784943203424,
|
||
|
|
"eval_model_preparation_time": 0.0059,
|
||
|
|
"eval_runtime": 18.2678,
|
||
|
|
"eval_samples_per_second": 179.003,
|
||
|
|
"eval_steps_per_second": 1.533,
|
||
|
|
"eval_total_labels_2822": 1231.0,
|
||
|
|
"eval_total_labels_9642": 2026.0,
|
||
|
|
"eval_total_preds": 3270.0,
|
||
|
|
"step": 12
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 4.333333333333333,
|
||
|
|
"grad_norm": 41.0,
|
||
|
|
"learning_rate": 1.9954719225730847e-05,
|
||
|
|
"loss": 0.2212,
|
||
|
|
"step": 13
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 4.666666666666667,
|
||
|
|
"grad_norm": 0.033447265625,
|
||
|
|
"learning_rate": 1.9944107713823068e-05,
|
||
|
|
"loss": 0.0001,
|
||
|
|
"step": 14
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 5.0,
|
||
|
|
"grad_norm": 0.0299072265625,
|
||
|
|
"learning_rate": 1.9932383577419432e-05,
|
||
|
|
"loss": 0.0001,
|
||
|
|
"step": 15
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 5.0,
|
||
|
|
"eval_accumulated_loss": 5369.499008178711,
|
||
|
|
"eval_accuracy": 0.7146788990825688,
|
||
|
|
"eval_accuracy_2822": 0.6344435418359058,
|
||
|
|
"eval_accuracy_9642": 0.7680157946692991,
|
||
|
|
"eval_correct_gen_preds": 2302.0,
|
||
|
|
"eval_correct_gen_preds_2822": 768.0,
|
||
|
|
"eval_correct_gen_preds_9642": 1525.0,
|
||
|
|
"eval_correct_preds": 2337.0,
|
||
|
|
"eval_correct_preds_2822": 781.0,
|
||
|
|
"eval_correct_preds_9642": 1556.0,
|
||
|
|
"eval_gen_accuracy": 0.7039755351681957,
|
||
|
|
"eval_gen_accuracy_2822": 0.6238830219333875,
|
||
|
|
"eval_gen_accuracy_9642": 0.7527147087857848,
|
||
|
|
"eval_loss": 1.6420485973358154,
|
||
|
|
"eval_mdl": 7746.5495911576345,
|
||
|
|
"eval_model_preparation_time": 0.0059,
|
||
|
|
"eval_runtime": 18.1931,
|
||
|
|
"eval_samples_per_second": 179.739,
|
||
|
|
"eval_steps_per_second": 1.539,
|
||
|
|
"eval_total_labels_2822": 1231.0,
|
||
|
|
"eval_total_labels_9642": 2026.0,
|
||
|
|
"eval_total_preds": 3270.0,
|
||
|
|
"step": 15
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"logging_steps": 1,
|
||
|
|
"max_steps": 300,
|
||
|
|
"num_input_tokens_seen": 0,
|
||
|
|
"num_train_epochs": 100,
|
||
|
|
"save_steps": 500,
|
||
|
|
"stateful_callbacks": {
|
||
|
|
"EarlyStoppingCallback": {
|
||
|
|
"args": {
|
||
|
|
"early_stopping_patience": 30,
|
||
|
|
"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": false
|
||
|
|
},
|
||
|
|
"attributes": {}
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"total_flos": 735978726752256.0,
|
||
|
|
"train_batch_size": 32,
|
||
|
|
"trial_name": null,
|
||
|
|
"trial_params": null
|
||
|
|
}
|