| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.9968652037617555, | |
| "eval_steps": 500, | |
| "global_step": 159, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.01, | |
| "learning_rate": 3.125e-08, | |
| "logits/chosen": -2.172743320465088, | |
| "logits/rejected": -2.1919965744018555, | |
| "logps/chosen": -359.53131103515625, | |
| "logps/pi_response": -340.22088623046875, | |
| "logps/ref_response": -340.22088623046875, | |
| "logps/rejected": -498.2034606933594, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.06, | |
| "learning_rate": 3.1249999999999997e-07, | |
| "logits/chosen": -2.2195708751678467, | |
| "logits/rejected": -2.1858723163604736, | |
| "logps/chosen": -295.3121337890625, | |
| "logps/pi_response": -236.75331115722656, | |
| "logps/ref_response": -236.98028564453125, | |
| "logps/rejected": -324.92376708984375, | |
| "loss": 0.6909, | |
| "rewards/accuracies": 0.4236111044883728, | |
| "rewards/chosen": -0.0021301452070474625, | |
| "rewards/margins": 0.0030618596356362104, | |
| "rewards/rejected": -0.005192004609853029, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "learning_rate": 4.990353313429303e-07, | |
| "logits/chosen": -2.1758713722229004, | |
| "logits/rejected": -2.157261371612549, | |
| "logps/chosen": -257.73443603515625, | |
| "logps/pi_response": -232.18289184570312, | |
| "logps/ref_response": -234.2621307373047, | |
| "logps/rejected": -355.9681701660156, | |
| "loss": 0.656, | |
| "rewards/accuracies": 0.643750011920929, | |
| "rewards/chosen": -0.05295618623495102, | |
| "rewards/margins": 0.16491419076919556, | |
| "rewards/rejected": -0.21787038445472717, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.19, | |
| "learning_rate": 4.882681251368548e-07, | |
| "logits/chosen": -2.071769952774048, | |
| "logits/rejected": -2.0303213596343994, | |
| "logps/chosen": -316.31121826171875, | |
| "logps/pi_response": -246.3899688720703, | |
| "logps/ref_response": -248.18966674804688, | |
| "logps/rejected": -371.10943603515625, | |
| "loss": 0.6629, | |
| "rewards/accuracies": 0.606249988079071, | |
| "rewards/chosen": -0.15825249254703522, | |
| "rewards/margins": 0.24746175110340118, | |
| "rewards/rejected": -0.4057142734527588, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "learning_rate": 4.6604720940421207e-07, | |
| "logits/chosen": -2.1430344581604004, | |
| "logits/rejected": -2.1164464950561523, | |
| "logps/chosen": -277.8004455566406, | |
| "logps/pi_response": -238.34353637695312, | |
| "logps/ref_response": -238.4957275390625, | |
| "logps/rejected": -371.0489196777344, | |
| "loss": 0.6686, | |
| "rewards/accuracies": 0.6000000238418579, | |
| "rewards/chosen": -0.0016612813342362642, | |
| "rewards/margins": 0.3003852069377899, | |
| "rewards/rejected": -0.3020465075969696, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.31, | |
| "learning_rate": 4.3344075855595097e-07, | |
| "logits/chosen": -2.2157797813415527, | |
| "logits/rejected": -2.1781415939331055, | |
| "logps/chosen": -331.3121337890625, | |
| "logps/pi_response": -255.37350463867188, | |
| "logps/ref_response": -250.7471923828125, | |
| "logps/rejected": -403.6841125488281, | |
| "loss": 0.6758, | |
| "rewards/accuracies": 0.6187499761581421, | |
| "rewards/chosen": -0.2203460931777954, | |
| "rewards/margins": 0.23005735874176025, | |
| "rewards/rejected": -0.45040345191955566, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.38, | |
| "learning_rate": 3.920161866827889e-07, | |
| "logits/chosen": -2.1681432723999023, | |
| "logits/rejected": -2.143216133117676, | |
| "logps/chosen": -316.63531494140625, | |
| "logps/pi_response": -254.83169555664062, | |
| "logps/ref_response": -235.57656860351562, | |
| "logps/rejected": -410.3238830566406, | |
| "loss": 0.6451, | |
| "rewards/accuracies": 0.6187499761581421, | |
| "rewards/chosen": -0.34151777625083923, | |
| "rewards/margins": 0.26650214195251465, | |
| "rewards/rejected": -0.6080199480056763, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.44, | |
| "learning_rate": 3.4376480090239047e-07, | |
| "logits/chosen": -2.193446636199951, | |
| "logits/rejected": -2.1542165279388428, | |
| "logps/chosen": -303.29180908203125, | |
| "logps/pi_response": -248.5313262939453, | |
| "logps/ref_response": -231.83877563476562, | |
| "logps/rejected": -365.41229248046875, | |
| "loss": 0.6366, | |
| "rewards/accuracies": 0.574999988079071, | |
| "rewards/chosen": -0.21329638361930847, | |
| "rewards/margins": 0.2295345515012741, | |
| "rewards/rejected": -0.4428309500217438, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "learning_rate": 2.910060778827554e-07, | |
| "logits/chosen": -2.191551923751831, | |
| "logits/rejected": -2.1738388538360596, | |
| "logps/chosen": -357.8346862792969, | |
| "logps/pi_response": -266.311767578125, | |
| "logps/ref_response": -249.003173828125, | |
| "logps/rejected": -391.28167724609375, | |
| "loss": 0.6417, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": -0.39481186866760254, | |
| "rewards/margins": 0.19340470433235168, | |
| "rewards/rejected": -0.5882166028022766, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.56, | |
| "learning_rate": 2.3627616503391812e-07, | |
| "logits/chosen": -2.1639695167541504, | |
| "logits/rejected": -2.132662773132324, | |
| "logps/chosen": -359.39361572265625, | |
| "logps/pi_response": -290.38165283203125, | |
| "logps/ref_response": -242.9669952392578, | |
| "logps/rejected": -430.5426330566406, | |
| "loss": 0.6297, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": -0.6065701842308044, | |
| "rewards/margins": 0.2715144753456116, | |
| "rewards/rejected": -0.878084659576416, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.63, | |
| "learning_rate": 1.8220596619089573e-07, | |
| "logits/chosen": -2.162834644317627, | |
| "logits/rejected": -2.1406049728393555, | |
| "logps/chosen": -338.8243103027344, | |
| "logps/pi_response": -289.0437316894531, | |
| "logps/ref_response": -250.93484497070312, | |
| "logps/rejected": -463.4561462402344, | |
| "loss": 0.6366, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.5193153619766235, | |
| "rewards/margins": 0.3801336884498596, | |
| "rewards/rejected": -0.8994490504264832, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.69, | |
| "learning_rate": 1.3139467229135998e-07, | |
| "logits/chosen": -2.139284610748291, | |
| "logits/rejected": -2.122141122817993, | |
| "logps/chosen": -355.6954040527344, | |
| "logps/pi_response": -282.2654724121094, | |
| "logps/ref_response": -240.23483276367188, | |
| "logps/rejected": -451.9892578125, | |
| "loss": 0.6546, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -0.5263327360153198, | |
| "rewards/margins": 0.34479039907455444, | |
| "rewards/rejected": -0.871123194694519, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "learning_rate": 8.628481651367875e-08, | |
| "logits/chosen": -2.162773847579956, | |
| "logits/rejected": -2.1433143615722656, | |
| "logps/chosen": -335.36541748046875, | |
| "logps/pi_response": -257.69769287109375, | |
| "logps/ref_response": -224.2637176513672, | |
| "logps/rejected": -387.4789123535156, | |
| "loss": 0.6329, | |
| "rewards/accuracies": 0.581250011920929, | |
| "rewards/chosen": -0.5079654455184937, | |
| "rewards/margins": 0.1786581575870514, | |
| "rewards/rejected": -0.6866235733032227, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.82, | |
| "learning_rate": 4.904486005914027e-08, | |
| "logits/chosen": -2.1650211811065674, | |
| "logits/rejected": -2.1384124755859375, | |
| "logps/chosen": -336.9052429199219, | |
| "logps/pi_response": -273.1711730957031, | |
| "logps/ref_response": -237.3932342529297, | |
| "logps/rejected": -422.136474609375, | |
| "loss": 0.6235, | |
| "rewards/accuracies": 0.606249988079071, | |
| "rewards/chosen": -0.44264864921569824, | |
| "rewards/margins": 0.29537224769592285, | |
| "rewards/rejected": -0.7380208969116211, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "learning_rate": 2.1464952759020856e-08, | |
| "logits/chosen": -2.167060136795044, | |
| "logits/rejected": -2.121119737625122, | |
| "logps/chosen": -322.75836181640625, | |
| "logps/pi_response": -271.6685485839844, | |
| "logps/ref_response": -237.94180297851562, | |
| "logps/rejected": -429.82281494140625, | |
| "loss": 0.6211, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -0.4370552599430084, | |
| "rewards/margins": 0.34228402376174927, | |
| "rewards/rejected": -0.7793391942977905, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.94, | |
| "learning_rate": 4.8708793644441086e-09, | |
| "logits/chosen": -2.1732852458953857, | |
| "logits/rejected": -2.1708812713623047, | |
| "logps/chosen": -351.72601318359375, | |
| "logps/pi_response": -296.03118896484375, | |
| "logps/ref_response": -258.5296936035156, | |
| "logps/rejected": -457.9854431152344, | |
| "loss": 0.6301, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": -0.4874304234981537, | |
| "rewards/margins": 0.2991498112678528, | |
| "rewards/rejected": -0.7865801453590393, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 159, | |
| "total_flos": 0.0, | |
| "train_loss": 0.6433063363129238, | |
| "train_runtime": 4130.1551, | |
| "train_samples_per_second": 4.934, | |
| "train_steps_per_second": 0.038 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 159, | |
| "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 | |
| } | |