| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.99581589958159, | |
| "eval_steps": 500, | |
| "global_step": 119, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.01, | |
| "learning_rate": 2.083333333333333e-08, | |
| "logits/chosen": -0.29348519444465637, | |
| "logits/rejected": 0.2745956778526306, | |
| "logps/chosen": -335.8572082519531, | |
| "logps/pi_response": -247.91940307617188, | |
| "logps/ref_response": -247.91940307617188, | |
| "logps/rejected": -424.2975769042969, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "learning_rate": 2.0833333333333333e-07, | |
| "logits/chosen": -0.016916176304221153, | |
| "logits/rejected": 0.3731401562690735, | |
| "logps/chosen": -313.33453369140625, | |
| "logps/pi_response": -241.23513793945312, | |
| "logps/ref_response": -240.8330078125, | |
| "logps/rejected": -450.3893127441406, | |
| "loss": 0.6904, | |
| "rewards/accuracies": 0.5208333134651184, | |
| "rewards/chosen": -0.006909824442118406, | |
| "rewards/margins": 0.008017909713089466, | |
| "rewards/rejected": -0.014927736483514309, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.17, | |
| "learning_rate": 2.4656762641186986e-07, | |
| "logits/chosen": 0.31496933102607727, | |
| "logits/rejected": 0.8171666860580444, | |
| "logps/chosen": -352.06365966796875, | |
| "logps/pi_response": -238.83694458007812, | |
| "logps/ref_response": -230.2179718017578, | |
| "logps/rejected": -454.1331481933594, | |
| "loss": 0.6634, | |
| "rewards/accuracies": 0.550000011920929, | |
| "rewards/chosen": -0.21408605575561523, | |
| "rewards/margins": 0.10318803787231445, | |
| "rewards/rejected": -0.3172740936279297, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "learning_rate": 2.3294604018447765e-07, | |
| "logits/chosen": 0.6205056309700012, | |
| "logits/rejected": 1.315418004989624, | |
| "logps/chosen": -350.58807373046875, | |
| "logps/pi_response": -252.17074584960938, | |
| "logps/ref_response": -247.3958282470703, | |
| "logps/rejected": -490.704345703125, | |
| "loss": 0.66, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": -0.2715633511543274, | |
| "rewards/margins": 0.2851446270942688, | |
| "rewards/rejected": -0.5567079782485962, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "learning_rate": 2.1008562769363286e-07, | |
| "logits/chosen": 0.4185073971748352, | |
| "logits/rejected": 1.0126984119415283, | |
| "logps/chosen": -352.34619140625, | |
| "logps/pi_response": -251.9999237060547, | |
| "logps/ref_response": -249.653564453125, | |
| "logps/rejected": -483.85797119140625, | |
| "loss": 0.6384, | |
| "rewards/accuracies": 0.5874999761581421, | |
| "rewards/chosen": -0.19181792438030243, | |
| "rewards/margins": 0.26109081506729126, | |
| "rewards/rejected": -0.4529086947441101, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "learning_rate": 1.799429533390377e-07, | |
| "logits/chosen": 0.5379453301429749, | |
| "logits/rejected": 1.0736875534057617, | |
| "logps/chosen": -371.8102722167969, | |
| "logps/pi_response": -259.549560546875, | |
| "logps/ref_response": -243.3005828857422, | |
| "logps/rejected": -455.50726318359375, | |
| "loss": 0.6237, | |
| "rewards/accuracies": 0.581250011920929, | |
| "rewards/chosen": -0.2562716603279114, | |
| "rewards/margins": 0.19361995160579681, | |
| "rewards/rejected": -0.4498916268348694, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "learning_rate": 1.4509785173993353e-07, | |
| "logits/chosen": 0.6267152428627014, | |
| "logits/rejected": 1.2144675254821777, | |
| "logps/chosen": -361.84320068359375, | |
| "logps/pi_response": -248.54367065429688, | |
| "logps/ref_response": -226.105224609375, | |
| "logps/rejected": -504.59698486328125, | |
| "loss": 0.6139, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -0.3499058783054352, | |
| "rewards/margins": 0.31102269887924194, | |
| "rewards/rejected": -0.6609286069869995, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "learning_rate": 1.0853262626989767e-07, | |
| "logits/chosen": 0.5657416582107544, | |
| "logits/rejected": 1.2980989217758179, | |
| "logps/chosen": -384.35614013671875, | |
| "logps/pi_response": -273.2232971191406, | |
| "logps/ref_response": -245.3466033935547, | |
| "logps/rejected": -527.5657958984375, | |
| "loss": 0.6125, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": -0.4277406632900238, | |
| "rewards/margins": 0.41355961561203003, | |
| "rewards/rejected": -0.841300368309021, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "learning_rate": 7.337680131745147e-08, | |
| "logits/chosen": 0.7760927081108093, | |
| "logits/rejected": 1.4221640825271606, | |
| "logps/chosen": -380.4726257324219, | |
| "logps/pi_response": -258.78680419921875, | |
| "logps/ref_response": -240.14889526367188, | |
| "logps/rejected": -459.6902770996094, | |
| "loss": 0.6073, | |
| "rewards/accuracies": 0.574999988079071, | |
| "rewards/chosen": -0.30700814723968506, | |
| "rewards/margins": 0.28355568647384644, | |
| "rewards/rejected": -0.5905638933181763, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "learning_rate": 4.263927427548612e-08, | |
| "logits/chosen": 0.7522335648536682, | |
| "logits/rejected": 1.2077925205230713, | |
| "logps/chosen": -352.4031677246094, | |
| "logps/pi_response": -252.8551788330078, | |
| "logps/ref_response": -233.1865692138672, | |
| "logps/rejected": -482.69781494140625, | |
| "loss": 0.6154, | |
| "rewards/accuracies": 0.574999988079071, | |
| "rewards/chosen": -0.29174160957336426, | |
| "rewards/margins": 0.2862595021724701, | |
| "rewards/rejected": -0.578001081943512, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "learning_rate": 1.8950791687585634e-08, | |
| "logits/chosen": 0.6551345586776733, | |
| "logits/rejected": 1.1925513744354248, | |
| "logps/chosen": -372.08685302734375, | |
| "logps/pi_response": -266.1755676269531, | |
| "logps/ref_response": -243.803466796875, | |
| "logps/rejected": -469.5684509277344, | |
| "loss": 0.6298, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": -0.31566232442855835, | |
| "rewards/margins": 0.2684438228607178, | |
| "rewards/rejected": -0.5841061472892761, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.92, | |
| "learning_rate": 4.338790361069836e-09, | |
| "logits/chosen": 0.6615439653396606, | |
| "logits/rejected": 1.032914400100708, | |
| "logps/chosen": -393.44122314453125, | |
| "logps/pi_response": -262.7541198730469, | |
| "logps/ref_response": -238.8057098388672, | |
| "logps/rejected": -429.650634765625, | |
| "loss": 0.6117, | |
| "rewards/accuracies": 0.574999988079071, | |
| "rewards/chosen": -0.35318782925605774, | |
| "rewards/margins": 0.17442646622657776, | |
| "rewards/rejected": -0.5276142358779907, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 119, | |
| "total_flos": 0.0, | |
| "train_loss": 0.6335267820278135, | |
| "train_runtime": 3410.0395, | |
| "train_samples_per_second": 4.482, | |
| "train_steps_per_second": 0.035 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 119, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 100, | |
| "total_flos": 0.0, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |