209 lines
5.4 KiB
JSON
209 lines
5.4 KiB
JSON
{
|
|
"best_global_step": null,
|
|
"best_metric": null,
|
|
"best_model_checkpoint": null,
|
|
"epoch": 7.0,
|
|
"eval_steps": 500,
|
|
"global_step": 77,
|
|
"is_hyper_param_search": false,
|
|
"is_local_process_zero": true,
|
|
"is_world_process_zero": true,
|
|
"log_history": [
|
|
{
|
|
"epoch": 0.46875,
|
|
"grad_norm": 4.618053901343934,
|
|
"learning_rate": 2e-05,
|
|
"loss": 0.9267,
|
|
"loss_nan_ranks": 0,
|
|
"loss_rank_avg": 0.2775108516216278,
|
|
"step": 5,
|
|
"valid_targets_mean": 10397.8,
|
|
"valid_targets_min": 2361
|
|
},
|
|
{
|
|
"epoch": 0.9375,
|
|
"grad_norm": 1.790584402959514,
|
|
"learning_rate": 3.9979273472934556e-05,
|
|
"loss": 0.7924,
|
|
"loss_nan_ranks": 0,
|
|
"loss_rank_avg": 0.24930989742279053,
|
|
"step": 10,
|
|
"valid_targets_mean": 11106.3,
|
|
"valid_targets_min": 2024
|
|
},
|
|
{
|
|
"epoch": 1.375,
|
|
"grad_norm": 0.7677108313035453,
|
|
"learning_rate": 3.925834574695599e-05,
|
|
"loss": 0.7306,
|
|
"loss_nan_ranks": 0,
|
|
"loss_rank_avg": 0.26121610403060913,
|
|
"step": 15,
|
|
"valid_targets_mean": 11634.1,
|
|
"valid_targets_min": 2357
|
|
},
|
|
{
|
|
"epoch": 1.84375,
|
|
"grad_norm": 0.5966576040620364,
|
|
"learning_rate": 3.754365204805189e-05,
|
|
"loss": 0.6863,
|
|
"loss_nan_ranks": 0,
|
|
"loss_rank_avg": 0.26030510663986206,
|
|
"step": 20,
|
|
"valid_targets_mean": 11225.4,
|
|
"valid_targets_min": 2461
|
|
},
|
|
{
|
|
"epoch": 2.28125,
|
|
"grad_norm": 0.5154141849393353,
|
|
"learning_rate": 3.4923673716953717e-05,
|
|
"loss": 0.6654,
|
|
"loss_nan_ranks": 0,
|
|
"loss_rank_avg": 0.22374999523162842,
|
|
"step": 25,
|
|
"valid_targets_mean": 10048.5,
|
|
"valid_targets_min": 1446
|
|
},
|
|
{
|
|
"epoch": 2.75,
|
|
"grad_norm": 0.36786823091452653,
|
|
"learning_rate": 3.153360644229735e-05,
|
|
"loss": 0.6265,
|
|
"loss_nan_ranks": 0,
|
|
"loss_rank_avg": 0.21615391969680786,
|
|
"step": 30,
|
|
"valid_targets_mean": 10494.9,
|
|
"valid_targets_min": 2380
|
|
},
|
|
{
|
|
"epoch": 3.1875,
|
|
"grad_norm": 0.2829903030397986,
|
|
"learning_rate": 2.7548383915052287e-05,
|
|
"loss": 0.6287,
|
|
"loss_nan_ranks": 0,
|
|
"loss_rank_avg": 0.19593819975852966,
|
|
"step": 35,
|
|
"valid_targets_mean": 10709.0,
|
|
"valid_targets_min": 2289
|
|
},
|
|
{
|
|
"epoch": 3.65625,
|
|
"grad_norm": 0.26049127760963475,
|
|
"learning_rate": 2.317365092912503e-05,
|
|
"loss": 0.6027,
|
|
"loss_nan_ranks": 0,
|
|
"loss_rank_avg": 0.2036541998386383,
|
|
"step": 40,
|
|
"valid_targets_mean": 10705.1,
|
|
"valid_targets_min": 3649
|
|
},
|
|
{
|
|
"epoch": 4.09375,
|
|
"grad_norm": 0.22627786472946976,
|
|
"learning_rate": 1.8635151732706586e-05,
|
|
"loss": 0.5976,
|
|
"loss_nan_ranks": 0,
|
|
"loss_rank_avg": 0.1876196265220642,
|
|
"step": 45,
|
|
"valid_targets_mean": 11062.9,
|
|
"valid_targets_min": 2003
|
|
},
|
|
{
|
|
"epoch": 4.5625,
|
|
"grad_norm": 0.26569312607529016,
|
|
"learning_rate": 1.4167081210903501e-05,
|
|
"loss": 0.5903,
|
|
"loss_nan_ranks": 0,
|
|
"loss_rank_avg": 0.20486825704574585,
|
|
"step": 50,
|
|
"valid_targets_mean": 9806.9,
|
|
"valid_targets_min": 3222
|
|
},
|
|
{
|
|
"epoch": 5.0,
|
|
"grad_norm": 0.23478162249932177,
|
|
"learning_rate": 1.0000000000000006e-05,
|
|
"loss": 0.5752,
|
|
"loss_nan_ranks": 0,
|
|
"loss_rank_avg": 0.3025435209274292,
|
|
"step": 55,
|
|
"valid_targets_mean": 12590.4,
|
|
"valid_targets_min": 2809
|
|
},
|
|
{
|
|
"epoch": 5.46875,
|
|
"grad_norm": 0.20415423264064336,
|
|
"learning_rate": 6.348937135626922e-06,
|
|
"loss": 0.5848,
|
|
"loss_nan_ranks": 0,
|
|
"loss_rank_avg": 0.19273033738136292,
|
|
"step": 60,
|
|
"valid_targets_mean": 10790.1,
|
|
"valid_targets_min": 1523
|
|
},
|
|
{
|
|
"epoch": 5.9375,
|
|
"grad_norm": 0.19045595214793523,
|
|
"learning_rate": 3.402294160019499e-06,
|
|
"loss": 0.5652,
|
|
"loss_nan_ranks": 0,
|
|
"loss_rank_avg": 0.1997273564338684,
|
|
"step": 65,
|
|
"valid_targets_mean": 11000.0,
|
|
"valid_targets_min": 1559
|
|
},
|
|
{
|
|
"epoch": 6.375,
|
|
"grad_norm": 0.18501011315226856,
|
|
"learning_rate": 1.3121232567865793e-06,
|
|
"loss": 0.5675,
|
|
"loss_nan_ranks": 0,
|
|
"loss_rank_avg": 0.19713328778743744,
|
|
"step": 70,
|
|
"valid_targets_mean": 12062.0,
|
|
"valid_targets_min": 3167
|
|
},
|
|
{
|
|
"epoch": 6.84375,
|
|
"grad_norm": 0.1910650829334838,
|
|
"learning_rate": 1.862810792733849e-07,
|
|
"loss": 0.5731,
|
|
"loss_nan_ranks": 0,
|
|
"loss_rank_avg": 0.1840532273054123,
|
|
"step": 75,
|
|
"valid_targets_mean": 9972.5,
|
|
"valid_targets_min": 3772
|
|
},
|
|
{
|
|
"epoch": 7.0,
|
|
"step": 77,
|
|
"total_flos": 2.891732460272353e+17,
|
|
"train_loss": 0.0,
|
|
"train_runtime": 0.8675,
|
|
"train_samples_per_second": 8068.811,
|
|
"train_steps_per_second": 88.757
|
|
}
|
|
],
|
|
"logging_steps": 5,
|
|
"max_steps": 77,
|
|
"num_input_tokens_seen": 0,
|
|
"num_train_epochs": 7,
|
|
"save_steps": 300,
|
|
"stateful_callbacks": {
|
|
"TrainerControl": {
|
|
"args": {
|
|
"should_epoch_stop": false,
|
|
"should_evaluate": false,
|
|
"should_log": false,
|
|
"should_save": true,
|
|
"should_training_stop": true
|
|
},
|
|
"attributes": {}
|
|
}
|
|
},
|
|
"total_flos": 2.891732460272353e+17,
|
|
"train_batch_size": 1,
|
|
"trial_name": null,
|
|
"trial_params": null
|
|
}
|