| { |
| "best_global_step": 1, |
| "best_metric": 0.42203541100025177, |
| "best_model_checkpoint": "/root/workspace/Shaer/grpo/outputs/sanity_check/sanity_20260411_123535/checkpoint-1", |
| "epoch": 0.5, |
| "eval_steps": 1, |
| "global_step": 1, |
| "is_hyper_param_search": false, |
| "is_local_process_zero": true, |
| "is_world_process_zero": true, |
| "log_history": [ |
| { |
| "clip_ratio/high_max": 0.10928299836814404, |
| "clip_ratio/high_mean": 0.10928299836814404, |
| "clip_ratio/low_mean": 0.09094957076013088, |
| "clip_ratio/low_min": 0.09094957076013088, |
| "clip_ratio/region_mean": 0.20023256912827492, |
| "completions/clipped_ratio": 0.25, |
| "completions/max_length": 640.0, |
| "completions/max_terminated_length": 237.0, |
| "completions/mean_length": 282.125, |
| "completions/mean_terminated_length": 162.83334350585938, |
| "completions/min_length": 86.0, |
| "completions/min_terminated_length": 86.0, |
| "entropy": 3.2345626056194305, |
| "epoch": 0.5, |
| "frac_reward_zero_std": 0.0, |
| "grad_norm": 4.5472893714904785, |
| "kl": 0.9608133509755135, |
| "learning_rate": 1e-05, |
| "loss": 0.0268, |
| "num_tokens": 3887.0, |
| "reward": 0.2736811637878418, |
| "reward_exact_count_bonus_mean": 0.25, |
| "reward_exact_count_bonus_std": 0.4629100561141968, |
| "reward_meter_mean": 0.22368116676807404, |
| "reward_meter_std": 0.30434706807136536, |
| "reward_std": 0.25279727578163147, |
| "reward_total_mean": 0.2736811637878418, |
| "rewards/exact_count_bonus/mean": 0.25, |
| "rewards/exact_count_bonus/std": 0.4629100561141968, |
| "rewards/meter/mean": 0.22368116676807404, |
| "rewards/meter/std": 0.30434706807136536, |
| "sampling/importance_sampling_ratio/max": 2.0, |
| "sampling/importance_sampling_ratio/mean": 0.9976902008056641, |
| "sampling/importance_sampling_ratio/min": 0.24163144826889038, |
| "sampling/sampling_logp_difference/max": 1.4203417301177979, |
| "sampling/sampling_logp_difference/mean": 0.15930314362049103, |
| "step": 1 |
| }, |
| { |
| "epoch": 0.5, |
| "eval_clip_ratio/high_max": 0.0, |
| "eval_clip_ratio/high_mean": 0.0, |
| "eval_clip_ratio/low_mean": 0.0, |
| "eval_clip_ratio/low_min": 0.0, |
| "eval_clip_ratio/region_mean": 0.0, |
| "eval_completions/clipped_ratio": 0.0, |
| "eval_completions/max_length": 156.0, |
| "eval_completions/max_terminated_length": 156.0, |
| "eval_completions/mean_length": 106.125, |
| "eval_completions/mean_terminated_length": 106.125, |
| "eval_completions/min_length": 69.25, |
| "eval_completions/min_terminated_length": 69.25, |
| "eval_entropy": 2.1008258759975433, |
| "eval_frac_reward_zero_std": 0.0, |
| "eval_kl": 1.3679395914077759, |
| "eval_loss": 0.0866735428571701, |
| "eval_num_tokens": 3887.0, |
| "eval_reward": 0.42203541100025177, |
| "eval_reward_exact_count_bonus_mean": 0.9375, |
| "eval_reward_exact_count_bonus_std": 0.125, |
| "eval_reward_meter_mean": 0.23453541472554207, |
| "eval_reward_meter_std": 0.22889509424567223, |
| "eval_reward_std": 0.2391926608979702, |
| "eval_reward_total_mean": 0.42203541100025177, |
| "eval_rewards/exact_count_bonus/mean": 0.9375, |
| "eval_rewards/exact_count_bonus/std": 0.125, |
| "eval_rewards/meter/mean": 0.23453541472554207, |
| "eval_rewards/meter/std": 0.22889509424567223, |
| "eval_runtime": 15.7215, |
| "eval_samples_per_second": 0.509, |
| "eval_sampling/importance_sampling_ratio/max": 1.2397247850894928, |
| "eval_sampling/importance_sampling_ratio/mean": 0.9995981901884079, |
| "eval_sampling/importance_sampling_ratio/min": 0.7994033098220825, |
| "eval_sampling/sampling_logp_difference/max": 0.24606388807296753, |
| "eval_sampling/sampling_logp_difference/mean": 0.034631784074008465, |
| "eval_steps_per_second": 0.127, |
| "step": 1 |
| } |
| ], |
| "logging_steps": 1, |
| "max_steps": 2, |
| "num_input_tokens_seen": 3887, |
| "num_train_epochs": 1, |
| "save_steps": 1, |
| "stateful_callbacks": { |
| "TrainerControl": { |
| "args": { |
| "should_epoch_stop": false, |
| "should_evaluate": false, |
| "should_log": false, |
| "should_save": true, |
| "should_training_stop": false |
| }, |
| "attributes": {} |
| } |
| }, |
| "total_flos": 0.0, |
| "train_batch_size": 1, |
| "trial_name": null, |
| "trial_params": null |
| } |
|
|
