234 lines
7.6 KiB
JSON
234 lines
7.6 KiB
JSON
|
{
|
||
|
|
"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": 112.125,
|
||
|
|
"epoch": 0.5714285714285714,
|
||
|
|
"grad_norm": 0.0,
|
||
|
|
"kl": 0.0,
|
||
|
|
"learning_rate": 5e-07,
|
||
|
|
"loss": 0.0,
|
||
|
|
"reward": 6.719051718711853,
|
||
|
|
"reward_std": 0.0,
|
||
|
|
"rewards/concensus_correctness_reward_func": 1.9263750240206718,
|
||
|
|
"rewards/consensus_reward_func": 1.875,
|
||
|
|
"rewards/cumulative_reward_2": 0.0,
|
||
|
|
"rewards/final_correctness_reward_func": 0.25,
|
||
|
|
"rewards/question_recreation_reward_func": 0.9958017989993095,
|
||
|
|
"rewards/soft_format_reward_func": 0.0,
|
||
|
|
"rewards/strict_format_reward_func": 0.46875,
|
||
|
|
"rewards/xmlcount_reward_func": 1.203125,
|
||
|
|
"step": 2
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"completion_length": 109.5,
|
||
|
|
"epoch": 1.0,
|
||
|
|
"grad_norm": 0.0,
|
||
|
|
"kl": 0.0,
|
||
|
|
"learning_rate": 4.864543104251586e-07,
|
||
|
|
"loss": 0.0,
|
||
|
|
"reward": 7.673666477203369,
|
||
|
|
"reward_std": 0.0,
|
||
|
|
"rewards/concensus_correctness_reward_func": 2.423666685819626,
|
||
|
|
"rewards/consensus_reward_func": 2.0,
|
||
|
|
"rewards/cumulative_reward_2": 0.0,
|
||
|
|
"rewards/final_correctness_reward_func": 0.5,
|
||
|
|
"rewards/question_recreation_reward_func": 1.0,
|
||
|
|
"rewards/soft_format_reward_func": 0.0,
|
||
|
|
"rewards/strict_format_reward_func": 0.5,
|
||
|
|
"rewards/xmlcount_reward_func": 1.25,
|
||
|
|
"step": 4
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"completion_length": 112.0625,
|
||
|
|
"epoch": 1.5714285714285714,
|
||
|
|
"grad_norm": 0.0,
|
||
|
|
"kl": 0.0,
|
||
|
|
"learning_rate": 4.472851273490984e-07,
|
||
|
|
"loss": 0.0,
|
||
|
|
"reward": 6.97005170583725,
|
||
|
|
"reward_std": 0.0,
|
||
|
|
"rewards/concensus_correctness_reward_func": 2.052375018596649,
|
||
|
|
"rewards/consensus_reward_func": 1.875,
|
||
|
|
"rewards/cumulative_reward_2": 0.0,
|
||
|
|
"rewards/final_correctness_reward_func": 0.375,
|
||
|
|
"rewards/question_recreation_reward_func": 0.9958017989993095,
|
||
|
|
"rewards/soft_format_reward_func": 0.0,
|
||
|
|
"rewards/strict_format_reward_func": 0.46875,
|
||
|
|
"rewards/xmlcount_reward_func": 1.203125,
|
||
|
|
"step": 6
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"completion_length": 109.375,
|
||
|
|
"epoch": 2.0,
|
||
|
|
"grad_norm": 0.0,
|
||
|
|
"kl": 0.0,
|
||
|
|
"learning_rate": 3.867370395306068e-07,
|
||
|
|
"loss": 0.0,
|
||
|
|
"reward": 7.33899982770284,
|
||
|
|
"reward_std": 0.0,
|
||
|
|
"rewards/concensus_correctness_reward_func": 2.2556666930516562,
|
||
|
|
"rewards/consensus_reward_func": 2.0,
|
||
|
|
"rewards/cumulative_reward_2": 0.0,
|
||
|
|
"rewards/final_correctness_reward_func": 0.3333333333333333,
|
||
|
|
"rewards/question_recreation_reward_func": 1.0,
|
||
|
|
"rewards/soft_format_reward_func": 0.0,
|
||
|
|
"rewards/strict_format_reward_func": 0.5,
|
||
|
|
"rewards/xmlcount_reward_func": 1.25,
|
||
|
|
"step": 8
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"completion_length": 114.59375,
|
||
|
|
"epoch": 2.571428571428571,
|
||
|
|
"grad_norm": 0.0,
|
||
|
|
"kl": 0.0,
|
||
|
|
"learning_rate": 3.1137137178519977e-07,
|
||
|
|
"loss": 0.0,
|
||
|
|
"reward": 6.970051676034927,
|
||
|
|
"reward_std": 0.0,
|
||
|
|
"rewards/concensus_correctness_reward_func": 2.0523750260472298,
|
||
|
|
"rewards/consensus_reward_func": 1.875,
|
||
|
|
"rewards/cumulative_reward_2": 0.0,
|
||
|
|
"rewards/final_correctness_reward_func": 0.375,
|
||
|
|
"rewards/question_recreation_reward_func": 0.9958017989993095,
|
||
|
|
"rewards/soft_format_reward_func": 0.0,
|
||
|
|
"rewards/strict_format_reward_func": 0.46875,
|
||
|
|
"rewards/xmlcount_reward_func": 1.203125,
|
||
|
|
"step": 10
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"completion_length": 105.91666666666667,
|
||
|
|
"epoch": 3.0,
|
||
|
|
"grad_norm": 0.0,
|
||
|
|
"kl": 0.0,
|
||
|
|
"learning_rate": 2.2935516363191693e-07,
|
||
|
|
"loss": 0.0,
|
||
|
|
"reward": 7.33899986743927,
|
||
|
|
"reward_std": 0.0,
|
||
|
|
"rewards/concensus_correctness_reward_func": 2.2556666831175485,
|
||
|
|
"rewards/consensus_reward_func": 2.0,
|
||
|
|
"rewards/cumulative_reward_2": 0.0,
|
||
|
|
"rewards/final_correctness_reward_func": 0.3333333333333333,
|
||
|
|
"rewards/question_recreation_reward_func": 1.0,
|
||
|
|
"rewards/soft_format_reward_func": 0.0,
|
||
|
|
"rewards/strict_format_reward_func": 0.5,
|
||
|
|
"rewards/xmlcount_reward_func": 1.25,
|
||
|
|
"step": 12
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"completion_length": 110.78125,
|
||
|
|
"epoch": 3.571428571428571,
|
||
|
|
"grad_norm": 0.0,
|
||
|
|
"kl": 0.0,
|
||
|
|
"learning_rate": 1.4957614383675767e-07,
|
||
|
|
"loss": 0.0,
|
||
|
|
"reward": 7.47067666053772,
|
||
|
|
"reward_std": 0.0,
|
||
|
|
"rewards/concensus_correctness_reward_func": 2.3030000180006027,
|
||
|
|
"rewards/consensus_reward_func": 1.875,
|
||
|
|
"rewards/cumulative_reward_2": 0.0,
|
||
|
|
"rewards/final_correctness_reward_func": 0.625,
|
||
|
|
"rewards/question_recreation_reward_func": 0.9958017989993095,
|
||
|
|
"rewards/soft_format_reward_func": 0.0,
|
||
|
|
"rewards/strict_format_reward_func": 0.46875,
|
||
|
|
"rewards/xmlcount_reward_func": 1.203125,
|
||
|
|
"step": 14
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"completion_length": 110.95833333333333,
|
||
|
|
"epoch": 4.0,
|
||
|
|
"grad_norm": 0.0,
|
||
|
|
"kl": 0.0,
|
||
|
|
"learning_rate": 8.067960709356478e-08,
|
||
|
|
"loss": 0.0,
|
||
|
|
"reward": 6.671499888102214,
|
||
|
|
"reward_std": 0.0,
|
||
|
|
"rewards/concensus_correctness_reward_func": 1.921500027179718,
|
||
|
|
"rewards/consensus_reward_func": 2.0,
|
||
|
|
"rewards/cumulative_reward_2": 0.0,
|
||
|
|
"rewards/final_correctness_reward_func": 0.0,
|
||
|
|
"rewards/question_recreation_reward_func": 1.0,
|
||
|
|
"rewards/soft_format_reward_func": 0.0,
|
||
|
|
"rewards/strict_format_reward_func": 0.5,
|
||
|
|
"rewards/xmlcount_reward_func": 1.25,
|
||
|
|
"step": 16
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"completion_length": 108.28125,
|
||
|
|
"epoch": 4.571428571428571,
|
||
|
|
"grad_norm": 0.0,
|
||
|
|
"kl": 0.0,
|
||
|
|
"learning_rate": 3.013156219837776e-08,
|
||
|
|
"loss": 0.0,
|
||
|
|
"reward": 7.220551639795303,
|
||
|
|
"reward_std": 0.0,
|
||
|
|
"rewards/concensus_correctness_reward_func": 2.177875019609928,
|
||
|
|
"rewards/consensus_reward_func": 1.875,
|
||
|
|
"rewards/cumulative_reward_2": 0.0,
|
||
|
|
"rewards/final_correctness_reward_func": 0.5,
|
||
|
|
"rewards/question_recreation_reward_func": 0.9958017989993095,
|
||
|
|
"rewards/soft_format_reward_func": 0.0,
|
||
|
|
"rewards/strict_format_reward_func": 0.46875,
|
||
|
|
"rewards/xmlcount_reward_func": 1.203125,
|
||
|
|
"step": 18
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"completion_length": 114.20833333333333,
|
||
|
|
"epoch": 5.0,
|
||
|
|
"grad_norm": 0.0,
|
||
|
|
"kl": 0.0,
|
||
|
|
"learning_rate": 3.4096741493194193e-09,
|
||
|
|
"loss": 0.0,
|
||
|
|
"reward": 7.004999915758769,
|
||
|
|
"reward_std": 0.0,
|
||
|
|
"rewards/concensus_correctness_reward_func": 2.0883333583672843,
|
||
|
|
"rewards/consensus_reward_func": 2.0,
|
||
|
|
"rewards/cumulative_reward_2": 0.0,
|
||
|
|
"rewards/final_correctness_reward_func": 0.16666666666666666,
|
||
|
|
"rewards/question_recreation_reward_func": 1.0,
|
||
|
|
"rewards/soft_format_reward_func": 0.0,
|
||
|
|
"rewards/strict_format_reward_func": 0.5,
|
||
|
|
"rewards/xmlcount_reward_func": 1.25,
|
||
|
|
"step": 20
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 5.0,
|
||
|
|
"step": 20,
|
||
|
|
"total_flos": 0.0,
|
||
|
|
"train_loss": 0.0,
|
||
|
|
"train_runtime": 1736.0203,
|
||
|
|
"train_samples_per_second": 0.184,
|
||
|
|
"train_steps_per_second": 0.012
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"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": 2,
|
||
|
|
"trial_name": null,
|
||
|
|
"trial_params": null
|
||
|
|
}
|