| { | |
| "best_metric": 0.4733360707759857, | |
| "best_model_checkpoint": "./mixstral/05-04-24-Weni-WeniGPT-Agents-Mixstral-Instruct-2.0.1-KTO_KTO with Agents 1.2.0 dataset and Mixstral model, with tokenization zephyr chat template-3_max_steps-145_batch_16_2024-04-05_ppid_9/checkpoint-100", | |
| "epoch": 0.684931506849315, | |
| "eval_steps": 50, | |
| "global_step": 100, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.14, | |
| "grad_norm": Infinity, | |
| "kl": 16.759851455688477, | |
| "learning_rate": 0.00018, | |
| "logps/chosen": -217.2803497314453, | |
| "logps/rejected": -266.78955078125, | |
| "loss": 0.4135, | |
| "rewards/chosen": 2.8302645683288574, | |
| "rewards/margins": 1.9116979837417603, | |
| "rewards/rejected": 1.01555597782135, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.27, | |
| "grad_norm": 4.954351902008057, | |
| "kl": 0.40393954515457153, | |
| "learning_rate": 0.00015142857142857143, | |
| "logps/chosen": -323.0667419433594, | |
| "logps/rejected": -331.9349365234375, | |
| "loss": 0.427, | |
| "rewards/chosen": -5.341065883636475, | |
| "rewards/margins": 2.56813383102417, | |
| "rewards/rejected": -8.01456356048584, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.34, | |
| "eval_kl": 0.31787964701652527, | |
| "eval_logps/chosen": -276.96771240234375, | |
| "eval_logps/rejected": -285.83453369140625, | |
| "eval_loss": 0.4442897439002991, | |
| "eval_rewards/chosen": -2.62730073928833, | |
| "eval_rewards/margins": 1.0439932346343994, | |
| "eval_rewards/rejected": -3.7775051593780518, | |
| "eval_runtime": 357.2813, | |
| "eval_samples_per_second": 0.84, | |
| "eval_steps_per_second": 0.21, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.41, | |
| "grad_norm": 1.9756174879148602e-05, | |
| "kl": 0.9693483114242554, | |
| "learning_rate": 0.00012571428571428572, | |
| "logps/chosen": -549.1575927734375, | |
| "logps/rejected": -583.3649291992188, | |
| "loss": 0.4167, | |
| "rewards/chosen": -29.825927734375, | |
| "rewards/margins": 3.2253201007843018, | |
| "rewards/rejected": -32.24267578125, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.55, | |
| "grad_norm": 1.4999230870671454e-06, | |
| "kl": 0.0, | |
| "learning_rate": 9.714285714285715e-05, | |
| "logps/chosen": -1577.5938720703125, | |
| "logps/rejected": -1497.0914306640625, | |
| "loss": 0.4625, | |
| "rewards/chosen": -132.58737182617188, | |
| "rewards/margins": -8.486028671264648, | |
| "rewards/rejected": -123.95598602294922, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.68, | |
| "grad_norm": 0.0, | |
| "kl": 0.0, | |
| "learning_rate": 6.857142857142858e-05, | |
| "logps/chosen": -1709.744140625, | |
| "logps/rejected": -1653.6820068359375, | |
| "loss": 0.425, | |
| "rewards/chosen": -144.38365173339844, | |
| "rewards/margins": -5.729578971862793, | |
| "rewards/rejected": -138.68692016601562, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.68, | |
| "eval_kl": 0.0, | |
| "eval_logps/chosen": -1588.1917724609375, | |
| "eval_logps/rejected": -1473.307373046875, | |
| "eval_loss": 0.4733360707759857, | |
| "eval_rewards/chosen": -133.74969482421875, | |
| "eval_rewards/margins": -10.205015182495117, | |
| "eval_rewards/rejected": -122.52478790283203, | |
| "eval_runtime": 356.0942, | |
| "eval_samples_per_second": 0.842, | |
| "eval_steps_per_second": 0.211, | |
| "step": 100 | |
| } | |
| ], | |
| "logging_steps": 20, | |
| "max_steps": 145, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 100, | |
| "total_flos": 0.0, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |