ryangensyn's picture
End of training
f105124 verified
Raw
History Blame Contribute Delete
8.36 kB
{
"best_global_step": null,
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 5.0,
"eval_steps": 500,
"global_step": 20,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"completion_length": 310.25,
"epoch": 0.5714285714285714,
"grad_norm": 3.4802045822143555,
"kl": 0.0025676235600258224,
"learning_rate": 5e-07,
"loss": 0.0,
"reward": 0.36369822337292135,
"reward_std": 0.5284897601231933,
"rewards/concensus_correctness_reward_func": 0.04618750140070915,
"rewards/consensus_reward_func": 0.0,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.0,
"rewards/question_recreation_reward_func": 0.21319822408258915,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.0,
"rewards/xmlcount_reward_func": 0.10431249998509884,
"step": 2
},
{
"completion_length": 368.8333333333333,
"epoch": 1.0,
"grad_norm": 1.3146730661392212,
"kl": 0.0012442187289707363,
"learning_rate": 4.864543104251586e-07,
"loss": 0.0,
"reward": 0.5365445253749689,
"reward_std": 1.0781507529318333,
"rewards/concensus_correctness_reward_func": 0.2206666668256124,
"rewards/consensus_reward_func": 0.08333333333333333,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.16666666666666666,
"rewards/question_recreation_reward_func": 0.20641951945920786,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.0,
"rewards/xmlcount_reward_func": -0.1405416834168136,
"step": 4
},
{
"completion_length": 286.625,
"epoch": 1.5714285714285714,
"grad_norm": 2.5264699459075928,
"kl": 0.0014120559499133378,
"learning_rate": 4.472851273490984e-07,
"loss": 0.0,
"reward": 0.3478491101413965,
"reward_std": 0.5067278523929417,
"rewards/concensus_correctness_reward_func": 0.0,
"rewards/consensus_reward_func": 0.0,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.0625,
"rewards/question_recreation_reward_func": 0.2111303503625095,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.0,
"rewards/xmlcount_reward_func": 0.07421875232830644,
"step": 6
},
{
"completion_length": 262.5416666666667,
"epoch": 2.0,
"grad_norm": 2.2129807472229004,
"kl": 0.005514397392592703,
"learning_rate": 3.867370395306068e-07,
"loss": 0.0,
"reward": 0.6012521783510844,
"reward_std": 0.7437834901114305,
"rewards/concensus_correctness_reward_func": 0.0,
"rewards/consensus_reward_func": 0.0,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.25,
"rewards/question_recreation_reward_func": 0.21700217574834824,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.0,
"rewards/xmlcount_reward_func": 0.1342500050086528,
"step": 8
},
{
"completion_length": 318.5,
"epoch": 2.571428571428571,
"grad_norm": 4.46564245223999,
"kl": 0.001195534598082304,
"learning_rate": 3.1137137178519977e-07,
"loss": 0.0,
"reward": 0.2478014836087823,
"reward_std": 0.31371984258294106,
"rewards/concensus_correctness_reward_func": 0.0,
"rewards/consensus_reward_func": 0.0,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.0625,
"rewards/question_recreation_reward_func": 0.19030148535966873,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.0,
"rewards/xmlcount_reward_func": -0.005000000004656613,
"step": 10
},
{
"completion_length": 458.8333333333333,
"epoch": 3.0,
"grad_norm": 1.14970862865448,
"kl": 0.0018530109955463558,
"learning_rate": 2.2935516363191693e-07,
"loss": 0.0,
"reward": 0.16758247216542563,
"reward_std": 0.45210413355380297,
"rewards/concensus_correctness_reward_func": 0.0,
"rewards/consensus_reward_func": 0.0,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.0,
"rewards/question_recreation_reward_func": 0.24441581157346567,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.0,
"rewards/xmlcount_reward_func": -0.07683332761128743,
"step": 12
},
{
"completion_length": 224.4375,
"epoch": 3.571428571428571,
"grad_norm": 5.530119895935059,
"kl": 0.0022977004118729383,
"learning_rate": 1.4957614383675767e-07,
"loss": 0.0,
"reward": 0.5765039762482047,
"reward_std": 0.5979496333748102,
"rewards/concensus_correctness_reward_func": 0.07175000011920929,
"rewards/consensus_reward_func": 0.0625,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.1875,
"rewards/question_recreation_reward_func": 0.23669149680063128,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.0,
"rewards/xmlcount_reward_func": 0.018062498420476913,
"step": 14
},
{
"completion_length": 391.4166666666667,
"epoch": 4.0,
"grad_norm": 1.2258388996124268,
"kl": 0.00157317992610236,
"learning_rate": 8.067960709356478e-08,
"loss": 0.0,
"reward": 0.25179099415739375,
"reward_std": 0.29438598081469536,
"rewards/concensus_correctness_reward_func": 0.0,
"rewards/consensus_reward_func": 0.0,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.0,
"rewards/question_recreation_reward_func": 0.102999330808719,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.0,
"rewards/xmlcount_reward_func": 0.14879166831572851,
"step": 16
},
{
"completion_length": 369.4375,
"epoch": 4.571428571428571,
"grad_norm": 3.678051471710205,
"kl": 0.0020257429860066622,
"learning_rate": 3.013156219837776e-08,
"loss": 0.0,
"reward": 0.9741338808089495,
"reward_std": 1.6201133914291859,
"rewards/concensus_correctness_reward_func": 0.625,
"rewards/consensus_reward_func": 0.0,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.0,
"rewards/question_recreation_reward_func": 0.26800887705758214,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.0,
"rewards/xmlcount_reward_func": 0.08112500421702862,
"step": 18
},
{
"completion_length": 252.125,
"epoch": 5.0,
"grad_norm": 1.8960169553756714,
"kl": 0.0015857915762656678,
"learning_rate": 3.4096741493194193e-09,
"loss": 0.0,
"reward": 1.625936284661293,
"reward_std": 2.337537464685738,
"rewards/concensus_correctness_reward_func": 0.9583333333333334,
"rewards/consensus_reward_func": 0.0,
"rewards/cumulative_reward_2": 0.0,
"rewards/final_correctness_reward_func": 0.25,
"rewards/question_recreation_reward_func": 0.32685289718210697,
"rewards/soft_format_reward_func": 0.0,
"rewards/strict_format_reward_func": 0.0,
"rewards/xmlcount_reward_func": 0.09074999143679936,
"step": 20
},
{
"epoch": 5.0,
"step": 20,
"total_flos": 0.0,
"train_loss": 1.8312789222818537e-06,
"train_runtime": 356.1099,
"train_samples_per_second": 0.899,
"train_steps_per_second": 0.056
}
],
"logging_steps": 2,
"max_steps": 20,
"num_input_tokens_seen": 0,
"num_train_epochs": 7,
"save_steps": 25,
"stateful_callbacks": {
"TrainerControl": {
"args": {
"should_epoch_stop": false,
"should_evaluate": false,
"should_log": false,
"should_save": true,
"should_training_stop": true
},
"attributes": {}
}
},
"total_flos": 0.0,
"train_batch_size": 4,
"trial_name": null,
"trial_params": null
}