rakhman-llm's picture
Training in progress, step 15462, checkpoint
71eee4b verified
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 3.0,
"eval_steps": 1000,
"global_step": 15462,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.09701202949165696,
"grad_norm": 0.9203131794929504,
"learning_rate": 1.9353253136722286e-05,
"loss": 1.7024,
"step": 500
},
{
"epoch": 0.19402405898331393,
"grad_norm": 1.0445950031280518,
"learning_rate": 1.8706506273444574e-05,
"loss": 1.3896,
"step": 1000
},
{
"epoch": 0.19402405898331393,
"eval_loss": 1.1814582347869873,
"eval_runtime": 49.6691,
"eval_samples_per_second": 104.351,
"eval_steps_per_second": 3.262,
"step": 1000
},
{
"epoch": 0.2910360884749709,
"grad_norm": 0.8253631591796875,
"learning_rate": 1.8059759410166862e-05,
"loss": 1.3302,
"step": 1500
},
{
"epoch": 0.38804811796662786,
"grad_norm": 0.9224209189414978,
"learning_rate": 1.741301254688915e-05,
"loss": 1.2914,
"step": 2000
},
{
"epoch": 0.38804811796662786,
"eval_loss": 1.1394085884094238,
"eval_runtime": 49.6277,
"eval_samples_per_second": 104.438,
"eval_steps_per_second": 3.264,
"step": 2000
},
{
"epoch": 0.48506014745828485,
"grad_norm": 0.7809765934944153,
"learning_rate": 1.6766265683611435e-05,
"loss": 1.2695,
"step": 2500
},
{
"epoch": 0.5820721769499418,
"grad_norm": 0.6769764423370361,
"learning_rate": 1.6119518820333723e-05,
"loss": 1.2523,
"step": 3000
},
{
"epoch": 0.5820721769499418,
"eval_loss": 1.1180071830749512,
"eval_runtime": 49.6817,
"eval_samples_per_second": 104.324,
"eval_steps_per_second": 3.261,
"step": 3000
},
{
"epoch": 0.6790842064415987,
"grad_norm": 0.8303701877593994,
"learning_rate": 1.547277195705601e-05,
"loss": 1.233,
"step": 3500
},
{
"epoch": 0.7760962359332557,
"grad_norm": 0.8391227722167969,
"learning_rate": 1.4826025093778296e-05,
"loss": 1.2294,
"step": 4000
},
{
"epoch": 0.7760962359332557,
"eval_loss": 1.1015874147415161,
"eval_runtime": 49.6378,
"eval_samples_per_second": 104.416,
"eval_steps_per_second": 3.264,
"step": 4000
},
{
"epoch": 0.8731082654249127,
"grad_norm": 0.7971065640449524,
"learning_rate": 1.4179278230500584e-05,
"loss": 1.2111,
"step": 4500
},
{
"epoch": 0.9701202949165697,
"grad_norm": 2.1875197887420654,
"learning_rate": 1.353253136722287e-05,
"loss": 1.2006,
"step": 5000
},
{
"epoch": 0.9701202949165697,
"eval_loss": 1.091620922088623,
"eval_runtime": 49.6213,
"eval_samples_per_second": 104.451,
"eval_steps_per_second": 3.265,
"step": 5000
},
{
"epoch": 1.0671323244082267,
"grad_norm": 0.729252278804779,
"learning_rate": 1.2885784503945158e-05,
"loss": 1.1861,
"step": 5500
},
{
"epoch": 1.1641443538998835,
"grad_norm": 0.6466144323348999,
"learning_rate": 1.2239037640667444e-05,
"loss": 1.1714,
"step": 6000
},
{
"epoch": 1.1641443538998835,
"eval_loss": 1.0831941366195679,
"eval_runtime": 49.6774,
"eval_samples_per_second": 104.333,
"eval_steps_per_second": 3.261,
"step": 6000
},
{
"epoch": 1.2611563833915405,
"grad_norm": 0.7486400008201599,
"learning_rate": 1.1592290777389729e-05,
"loss": 1.1696,
"step": 6500
},
{
"epoch": 1.3581684128831975,
"grad_norm": 0.7059735059738159,
"learning_rate": 1.0945543914112017e-05,
"loss": 1.1566,
"step": 7000
},
{
"epoch": 1.3581684128831975,
"eval_loss": 1.0764732360839844,
"eval_runtime": 49.6981,
"eval_samples_per_second": 104.29,
"eval_steps_per_second": 3.26,
"step": 7000
},
{
"epoch": 1.4551804423748544,
"grad_norm": 0.7840662002563477,
"learning_rate": 1.0298797050834303e-05,
"loss": 1.154,
"step": 7500
},
{
"epoch": 1.5521924718665114,
"grad_norm": 1.1568676233291626,
"learning_rate": 9.652050187556591e-06,
"loss": 1.1596,
"step": 8000
},
{
"epoch": 1.5521924718665114,
"eval_loss": 1.07025945186615,
"eval_runtime": 49.6881,
"eval_samples_per_second": 104.311,
"eval_steps_per_second": 3.26,
"step": 8000
},
{
"epoch": 1.6492045013581684,
"grad_norm": 0.7761677503585815,
"learning_rate": 9.005303324278878e-06,
"loss": 1.1541,
"step": 8500
},
{
"epoch": 1.7462165308498254,
"grad_norm": 0.8179004788398743,
"learning_rate": 8.358556461001164e-06,
"loss": 1.1483,
"step": 9000
},
{
"epoch": 1.7462165308498254,
"eval_loss": 1.0663487911224365,
"eval_runtime": 49.6731,
"eval_samples_per_second": 104.342,
"eval_steps_per_second": 3.261,
"step": 9000
},
{
"epoch": 1.8432285603414824,
"grad_norm": 0.7527589201927185,
"learning_rate": 7.711809597723452e-06,
"loss": 1.1429,
"step": 9500
},
{
"epoch": 1.9402405898331394,
"grad_norm": 0.8005126714706421,
"learning_rate": 7.065062734445739e-06,
"loss": 1.1281,
"step": 10000
},
{
"epoch": 1.9402405898331394,
"eval_loss": 1.0637460947036743,
"eval_runtime": 49.672,
"eval_samples_per_second": 104.344,
"eval_steps_per_second": 3.261,
"step": 10000
},
{
"epoch": 2.0372526193247964,
"grad_norm": 0.6758320927619934,
"learning_rate": 6.418315871168025e-06,
"loss": 1.1251,
"step": 10500
},
{
"epoch": 2.1342646488164534,
"grad_norm": 0.7758616805076599,
"learning_rate": 5.771569007890312e-06,
"loss": 1.1252,
"step": 11000
},
{
"epoch": 2.1342646488164534,
"eval_loss": 1.0619136095046997,
"eval_runtime": 49.6813,
"eval_samples_per_second": 104.325,
"eval_steps_per_second": 3.261,
"step": 11000
},
{
"epoch": 2.2312766783081104,
"grad_norm": 0.8191428780555725,
"learning_rate": 5.124822144612599e-06,
"loss": 1.1191,
"step": 11500
},
{
"epoch": 2.328288707799767,
"grad_norm": 0.7311928272247314,
"learning_rate": 4.478075281334886e-06,
"loss": 1.1187,
"step": 12000
},
{
"epoch": 2.328288707799767,
"eval_loss": 1.0580228567123413,
"eval_runtime": 49.6253,
"eval_samples_per_second": 104.443,
"eval_steps_per_second": 3.264,
"step": 12000
},
{
"epoch": 2.4253007372914244,
"grad_norm": 0.6671691536903381,
"learning_rate": 3.831328418057173e-06,
"loss": 1.115,
"step": 12500
},
{
"epoch": 2.522312766783081,
"grad_norm": 0.7181034088134766,
"learning_rate": 3.1845815547794594e-06,
"loss": 1.1025,
"step": 13000
},
{
"epoch": 2.522312766783081,
"eval_loss": 1.0561449527740479,
"eval_runtime": 49.663,
"eval_samples_per_second": 104.363,
"eval_steps_per_second": 3.262,
"step": 13000
},
{
"epoch": 2.619324796274738,
"grad_norm": 1.3909169435501099,
"learning_rate": 2.5378346915017466e-06,
"loss": 1.1197,
"step": 13500
},
{
"epoch": 2.716336825766395,
"grad_norm": 0.7657527327537537,
"learning_rate": 1.8910878282240333e-06,
"loss": 1.1113,
"step": 14000
},
{
"epoch": 2.716336825766395,
"eval_loss": 1.0557653903961182,
"eval_runtime": 49.6516,
"eval_samples_per_second": 104.387,
"eval_steps_per_second": 3.263,
"step": 14000
},
{
"epoch": 2.813348855258052,
"grad_norm": 0.6979902386665344,
"learning_rate": 1.24434096494632e-06,
"loss": 1.1123,
"step": 14500
},
{
"epoch": 2.910360884749709,
"grad_norm": 0.7375929355621338,
"learning_rate": 5.975941016686069e-07,
"loss": 1.1155,
"step": 15000
},
{
"epoch": 2.910360884749709,
"eval_loss": 1.054747223854065,
"eval_runtime": 49.5431,
"eval_samples_per_second": 104.616,
"eval_steps_per_second": 3.27,
"step": 15000
}
],
"logging_steps": 500,
"max_steps": 15462,
"num_input_tokens_seen": 0,
"num_train_epochs": 3,
"save_steps": 50000,
"stateful_callbacks": {
"TrainerControl": {
"args": {
"should_epoch_stop": false,
"should_evaluate": false,
"should_log": false,
"should_save": true,
"should_training_stop": true
},
"attributes": {}
}
},
"total_flos": 7.532337163862016e+16,
"train_batch_size": 32,
"trial_name": null,
"trial_params": null
}