| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 3.0, | |
| "eval_steps": 500, | |
| "global_step": 930, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0032258064516129032, | |
| "grad_norm": 115.22052507274148, | |
| "learning_rate": 5.3763440860215056e-09, | |
| "logits/chosen": -4.125, | |
| "logits/rejected": -4.0625, | |
| "logps/chosen": -185.0, | |
| "logps/rejected": -196.0, | |
| "loss": 0.6914, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.03225806451612903, | |
| "grad_norm": 94.47396535248316, | |
| "learning_rate": 5.3763440860215054e-08, | |
| "logits/chosen": -4.125, | |
| "logits/rejected": -4.15625, | |
| "logps/chosen": -186.0, | |
| "logps/rejected": -195.0, | |
| "loss": 0.6886, | |
| "rewards/accuracies": 0.2222222238779068, | |
| "rewards/chosen": 1.3589859008789062e-05, | |
| "rewards/margins": 0.009765625, | |
| "rewards/rejected": -0.009765625, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.06451612903225806, | |
| "grad_norm": 94.33148772179948, | |
| "learning_rate": 1.0752688172043011e-07, | |
| "logits/chosen": -4.15625, | |
| "logits/rejected": -4.0625, | |
| "logps/chosen": -188.0, | |
| "logps/rejected": -200.0, | |
| "loss": 0.6872, | |
| "rewards/accuracies": 0.2874999940395355, | |
| "rewards/chosen": -0.005645751953125, | |
| "rewards/margins": 0.01312255859375, | |
| "rewards/rejected": -0.018798828125, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.0967741935483871, | |
| "grad_norm": 92.49633755921806, | |
| "learning_rate": 1.6129032258064515e-07, | |
| "logits/chosen": -4.0625, | |
| "logits/rejected": -4.03125, | |
| "logps/chosen": -185.0, | |
| "logps/rejected": -191.0, | |
| "loss": 0.669, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": 0.016845703125, | |
| "rewards/margins": 0.054443359375, | |
| "rewards/rejected": -0.03759765625, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.12903225806451613, | |
| "grad_norm": 82.30619665044928, | |
| "learning_rate": 2.1505376344086022e-07, | |
| "logits/chosen": -4.03125, | |
| "logits/rejected": -4.03125, | |
| "logps/chosen": -189.0, | |
| "logps/rejected": -200.0, | |
| "loss": 0.6341, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": 0.038818359375, | |
| "rewards/margins": 0.13671875, | |
| "rewards/rejected": -0.09765625, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.16129032258064516, | |
| "grad_norm": 73.60890561659406, | |
| "learning_rate": 2.6881720430107523e-07, | |
| "logits/chosen": -4.03125, | |
| "logits/rejected": -3.890625, | |
| "logps/chosen": -189.0, | |
| "logps/rejected": -201.0, | |
| "loss": 0.5644, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.029296875, | |
| "rewards/margins": 0.328125, | |
| "rewards/rejected": -0.357421875, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.1935483870967742, | |
| "grad_norm": 83.24582572373353, | |
| "learning_rate": 3.225806451612903e-07, | |
| "logits/chosen": -3.890625, | |
| "logits/rejected": -3.828125, | |
| "logps/chosen": -191.0, | |
| "logps/rejected": -204.0, | |
| "loss": 0.5283, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.25390625, | |
| "rewards/margins": 0.484375, | |
| "rewards/rejected": -0.73828125, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.22580645161290322, | |
| "grad_norm": 70.93826710868849, | |
| "learning_rate": 3.7634408602150537e-07, | |
| "logits/chosen": -3.9375, | |
| "logits/rejected": -3.890625, | |
| "logps/chosen": -187.0, | |
| "logps/rejected": -206.0, | |
| "loss": 0.4758, | |
| "rewards/accuracies": 0.887499988079071, | |
| "rewards/chosen": -0.1494140625, | |
| "rewards/margins": 0.83984375, | |
| "rewards/rejected": -0.98828125, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.25806451612903225, | |
| "grad_norm": 67.89432749815619, | |
| "learning_rate": 4.3010752688172043e-07, | |
| "logits/chosen": -3.875, | |
| "logits/rejected": -3.84375, | |
| "logps/chosen": -188.0, | |
| "logps/rejected": -215.0, | |
| "loss": 0.4225, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -0.06494140625, | |
| "rewards/margins": 1.3359375, | |
| "rewards/rejected": -1.3984375, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.2903225806451613, | |
| "grad_norm": 56.57247809818219, | |
| "learning_rate": 4.838709677419355e-07, | |
| "logits/chosen": -3.9375, | |
| "logits/rejected": -3.921875, | |
| "logps/chosen": -195.0, | |
| "logps/rejected": -220.0, | |
| "loss": 0.3769, | |
| "rewards/accuracies": 0.862500011920929, | |
| "rewards/chosen": -0.8125, | |
| "rewards/margins": 1.390625, | |
| "rewards/rejected": -2.203125, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.3225806451612903, | |
| "grad_norm": 54.52791585727774, | |
| "learning_rate": 4.958183990442055e-07, | |
| "logits/chosen": -3.90625, | |
| "logits/rejected": -3.890625, | |
| "logps/chosen": -198.0, | |
| "logps/rejected": -225.0, | |
| "loss": 0.3733, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -1.03125, | |
| "rewards/margins": 1.6015625, | |
| "rewards/rejected": -2.625, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.3548387096774194, | |
| "grad_norm": 70.14660020921133, | |
| "learning_rate": 4.898446833930704e-07, | |
| "logits/chosen": -3.96875, | |
| "logits/rejected": -3.90625, | |
| "logps/chosen": -195.0, | |
| "logps/rejected": -225.0, | |
| "loss": 0.3651, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": -1.0390625, | |
| "rewards/margins": 1.7578125, | |
| "rewards/rejected": -2.796875, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.3870967741935484, | |
| "grad_norm": 70.71484436398696, | |
| "learning_rate": 4.838709677419355e-07, | |
| "logits/chosen": -3.9375, | |
| "logits/rejected": -3.90625, | |
| "logps/chosen": -196.0, | |
| "logps/rejected": -224.0, | |
| "loss": 0.4252, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": -0.9609375, | |
| "rewards/margins": 1.7421875, | |
| "rewards/rejected": -2.703125, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.41935483870967744, | |
| "grad_norm": 60.790615448301835, | |
| "learning_rate": 4.778972520908004e-07, | |
| "logits/chosen": -3.9375, | |
| "logits/rejected": -3.9375, | |
| "logps/chosen": -197.0, | |
| "logps/rejected": -231.0, | |
| "loss": 0.3266, | |
| "rewards/accuracies": 0.9125000238418579, | |
| "rewards/chosen": -0.95703125, | |
| "rewards/margins": 2.21875, | |
| "rewards/rejected": -3.1875, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.45161290322580644, | |
| "grad_norm": 69.89033127044969, | |
| "learning_rate": 4.7192353643966544e-07, | |
| "logits/chosen": -3.9375, | |
| "logits/rejected": -3.890625, | |
| "logps/chosen": -201.0, | |
| "logps/rejected": -230.0, | |
| "loss": 0.3487, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": -1.171875, | |
| "rewards/margins": 1.8359375, | |
| "rewards/rejected": -3.0, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.4838709677419355, | |
| "grad_norm": 70.95003315189264, | |
| "learning_rate": 4.6594982078853044e-07, | |
| "logits/chosen": -3.90625, | |
| "logits/rejected": -3.828125, | |
| "logps/chosen": -196.0, | |
| "logps/rejected": -229.0, | |
| "loss": 0.3011, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -1.140625, | |
| "rewards/margins": 2.078125, | |
| "rewards/rejected": -3.21875, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.5161290322580645, | |
| "grad_norm": 78.12294251323976, | |
| "learning_rate": 4.5997610513739544e-07, | |
| "logits/chosen": -3.90625, | |
| "logits/rejected": -3.859375, | |
| "logps/chosen": -194.0, | |
| "logps/rejected": -229.0, | |
| "loss": 0.2904, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": -0.80859375, | |
| "rewards/margins": 2.453125, | |
| "rewards/rejected": -3.265625, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.5483870967741935, | |
| "grad_norm": 58.17040697560013, | |
| "learning_rate": 4.540023894862604e-07, | |
| "logits/chosen": -3.984375, | |
| "logits/rejected": -3.890625, | |
| "logps/chosen": -202.0, | |
| "logps/rejected": -232.0, | |
| "loss": 0.286, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": -1.53125, | |
| "rewards/margins": 2.484375, | |
| "rewards/rejected": -4.0, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.5806451612903226, | |
| "grad_norm": 75.00483410106833, | |
| "learning_rate": 4.4802867383512544e-07, | |
| "logits/chosen": -3.8125, | |
| "logits/rejected": -3.765625, | |
| "logps/chosen": -200.0, | |
| "logps/rejected": -239.0, | |
| "loss": 0.3374, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -1.5625, | |
| "rewards/margins": 2.484375, | |
| "rewards/rejected": -4.03125, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.6129032258064516, | |
| "grad_norm": 72.167130272584, | |
| "learning_rate": 4.4205495818399044e-07, | |
| "logits/chosen": -3.90625, | |
| "logits/rejected": -3.875, | |
| "logps/chosen": -200.0, | |
| "logps/rejected": -239.0, | |
| "loss": 0.3004, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": -1.0703125, | |
| "rewards/margins": 2.890625, | |
| "rewards/rejected": -3.96875, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.6451612903225806, | |
| "grad_norm": 61.76892460773258, | |
| "learning_rate": 4.3608124253285543e-07, | |
| "logits/chosen": -3.90625, | |
| "logits/rejected": -3.90625, | |
| "logps/chosen": -199.0, | |
| "logps/rejected": -230.0, | |
| "loss": 0.301, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -1.375, | |
| "rewards/margins": 2.015625, | |
| "rewards/rejected": -3.390625, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.6774193548387096, | |
| "grad_norm": 52.49765267963524, | |
| "learning_rate": 4.3010752688172043e-07, | |
| "logits/chosen": -3.890625, | |
| "logits/rejected": -3.78125, | |
| "logps/chosen": -196.0, | |
| "logps/rejected": -233.0, | |
| "loss": 0.258, | |
| "rewards/accuracies": 0.887499988079071, | |
| "rewards/chosen": -0.8828125, | |
| "rewards/margins": 2.84375, | |
| "rewards/rejected": -3.734375, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.7096774193548387, | |
| "grad_norm": 58.816815163007185, | |
| "learning_rate": 4.241338112305854e-07, | |
| "logits/chosen": -3.859375, | |
| "logits/rejected": -3.84375, | |
| "logps/chosen": -194.0, | |
| "logps/rejected": -237.0, | |
| "loss": 0.278, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -1.0390625, | |
| "rewards/margins": 2.6875, | |
| "rewards/rejected": -3.734375, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.7419354838709677, | |
| "grad_norm": 77.22842288731327, | |
| "learning_rate": 4.1816009557945043e-07, | |
| "logits/chosen": -3.859375, | |
| "logits/rejected": -3.734375, | |
| "logps/chosen": -192.0, | |
| "logps/rejected": -234.0, | |
| "loss": 0.2537, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -0.7890625, | |
| "rewards/margins": 3.046875, | |
| "rewards/rejected": -3.84375, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.7741935483870968, | |
| "grad_norm": 55.73484353697734, | |
| "learning_rate": 4.121863799283154e-07, | |
| "logits/chosen": -3.875, | |
| "logits/rejected": -3.796875, | |
| "logps/chosen": -199.0, | |
| "logps/rejected": -235.0, | |
| "loss": 0.3267, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": -1.5078125, | |
| "rewards/margins": 2.8125, | |
| "rewards/rejected": -4.3125, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.8064516129032258, | |
| "grad_norm": 73.3120879364035, | |
| "learning_rate": 4.0621266427718037e-07, | |
| "logits/chosen": -3.953125, | |
| "logits/rejected": -3.890625, | |
| "logps/chosen": -194.0, | |
| "logps/rejected": -233.0, | |
| "loss": 0.2759, | |
| "rewards/accuracies": 0.887499988079071, | |
| "rewards/chosen": -0.85546875, | |
| "rewards/margins": 2.96875, | |
| "rewards/rejected": -3.8125, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.8387096774193549, | |
| "grad_norm": 53.82592061402359, | |
| "learning_rate": 4.002389486260454e-07, | |
| "logits/chosen": -3.953125, | |
| "logits/rejected": -3.890625, | |
| "logps/chosen": -195.0, | |
| "logps/rejected": -245.0, | |
| "loss": 0.2255, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -0.59765625, | |
| "rewards/margins": 3.875, | |
| "rewards/rejected": -4.46875, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.8709677419354839, | |
| "grad_norm": 73.21099669158846, | |
| "learning_rate": 3.9426523297491037e-07, | |
| "logits/chosen": -3.953125, | |
| "logits/rejected": -3.921875, | |
| "logps/chosen": -195.0, | |
| "logps/rejected": -244.0, | |
| "loss": 0.2336, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -1.0703125, | |
| "rewards/margins": 3.53125, | |
| "rewards/rejected": -4.59375, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.9032258064516129, | |
| "grad_norm": 61.8854330576029, | |
| "learning_rate": 3.8829151732377537e-07, | |
| "logits/chosen": -3.9375, | |
| "logits/rejected": -3.90625, | |
| "logps/chosen": -199.0, | |
| "logps/rejected": -232.0, | |
| "loss": 0.2834, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": -1.15625, | |
| "rewards/margins": 2.890625, | |
| "rewards/rejected": -4.03125, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.9354838709677419, | |
| "grad_norm": 58.02390241120245, | |
| "learning_rate": 3.8231780167264037e-07, | |
| "logits/chosen": -4.03125, | |
| "logits/rejected": -3.9375, | |
| "logps/chosen": -196.0, | |
| "logps/rejected": -230.0, | |
| "loss": 0.2844, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": -0.7734375, | |
| "rewards/margins": 2.9375, | |
| "rewards/rejected": -3.71875, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.967741935483871, | |
| "grad_norm": 78.57587657653356, | |
| "learning_rate": 3.7634408602150537e-07, | |
| "logits/chosen": -3.984375, | |
| "logits/rejected": -3.875, | |
| "logps/chosen": -189.0, | |
| "logps/rejected": -236.0, | |
| "loss": 0.2828, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -0.6171875, | |
| "rewards/margins": 3.234375, | |
| "rewards/rejected": -3.859375, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 67.82380071078845, | |
| "learning_rate": 3.703703703703703e-07, | |
| "logits/chosen": -3.890625, | |
| "logits/rejected": -3.859375, | |
| "logps/chosen": -196.0, | |
| "logps/rejected": -230.0, | |
| "loss": 0.271, | |
| "rewards/accuracies": 0.887499988079071, | |
| "rewards/chosen": -0.9921875, | |
| "rewards/margins": 3.0625, | |
| "rewards/rejected": -4.0625, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_logits/chosen": -3.953125, | |
| "eval_logits/rejected": -3.890625, | |
| "eval_logps/chosen": -217.0, | |
| "eval_logps/rejected": -255.0, | |
| "eval_loss": 0.20095214247703552, | |
| "eval_rewards/accuracies": 0.9285714030265808, | |
| "eval_rewards/chosen": -1.7890625, | |
| "eval_rewards/margins": 3.21875, | |
| "eval_rewards/rejected": -5.0, | |
| "eval_runtime": 13.4206, | |
| "eval_samples_per_second": 14.902, | |
| "eval_steps_per_second": 0.522, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 1.032258064516129, | |
| "grad_norm": 14.7763712358255, | |
| "learning_rate": 3.6439665471923536e-07, | |
| "logits/chosen": -4.03125, | |
| "logits/rejected": -3.90625, | |
| "logps/chosen": -197.0, | |
| "logps/rejected": -268.0, | |
| "loss": 0.032, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -1.0234375, | |
| "rewards/margins": 5.96875, | |
| "rewards/rejected": -7.0, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 1.064516129032258, | |
| "grad_norm": 6.092608827300573, | |
| "learning_rate": 3.5842293906810036e-07, | |
| "logits/chosen": -3.9375, | |
| "logits/rejected": -3.859375, | |
| "logps/chosen": -193.0, | |
| "logps/rejected": -268.0, | |
| "loss": 0.0204, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -0.349609375, | |
| "rewards/margins": 6.0, | |
| "rewards/rejected": -6.34375, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 1.096774193548387, | |
| "grad_norm": 13.413597095885587, | |
| "learning_rate": 3.524492234169653e-07, | |
| "logits/chosen": -3.859375, | |
| "logits/rejected": -3.75, | |
| "logps/chosen": -192.0, | |
| "logps/rejected": -270.0, | |
| "loss": 0.0397, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -0.123046875, | |
| "rewards/margins": 6.40625, | |
| "rewards/rejected": -6.53125, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 1.129032258064516, | |
| "grad_norm": 23.482097218429256, | |
| "learning_rate": 3.4647550776583036e-07, | |
| "logits/chosen": -3.875, | |
| "logits/rejected": -3.84375, | |
| "logps/chosen": -193.0, | |
| "logps/rejected": -272.0, | |
| "loss": 0.0236, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -0.376953125, | |
| "rewards/margins": 7.0, | |
| "rewards/rejected": -7.375, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 1.1612903225806452, | |
| "grad_norm": 23.358340694800074, | |
| "learning_rate": 3.405017921146953e-07, | |
| "logits/chosen": -3.828125, | |
| "logits/rejected": -3.8125, | |
| "logps/chosen": -183.0, | |
| "logps/rejected": -260.0, | |
| "loss": 0.0311, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": 0.154296875, | |
| "rewards/margins": 6.3125, | |
| "rewards/rejected": -6.1875, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 1.1935483870967742, | |
| "grad_norm": 3.914331549525913, | |
| "learning_rate": 3.345280764635603e-07, | |
| "logits/chosen": -3.890625, | |
| "logits/rejected": -3.8125, | |
| "logps/chosen": -193.0, | |
| "logps/rejected": -274.0, | |
| "loss": 0.0253, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.0299072265625, | |
| "rewards/margins": 7.03125, | |
| "rewards/rejected": -7.0625, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 1.2258064516129032, | |
| "grad_norm": 7.274490128378863, | |
| "learning_rate": 3.285543608124253e-07, | |
| "logits/chosen": -3.96875, | |
| "logits/rejected": -3.84375, | |
| "logps/chosen": -190.0, | |
| "logps/rejected": -270.0, | |
| "loss": 0.0083, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 0.08349609375, | |
| "rewards/margins": 7.3125, | |
| "rewards/rejected": -7.21875, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 1.2580645161290323, | |
| "grad_norm": 3.5384348394162375, | |
| "learning_rate": 3.225806451612903e-07, | |
| "logits/chosen": -3.8125, | |
| "logits/rejected": -3.828125, | |
| "logps/chosen": -189.0, | |
| "logps/rejected": -276.0, | |
| "loss": 0.0157, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -0.119140625, | |
| "rewards/margins": 7.09375, | |
| "rewards/rejected": -7.1875, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 1.2903225806451613, | |
| "grad_norm": 27.375764382001474, | |
| "learning_rate": 3.1660692951015535e-07, | |
| "logits/chosen": -3.75, | |
| "logits/rejected": -3.78125, | |
| "logps/chosen": -194.0, | |
| "logps/rejected": -274.0, | |
| "loss": 0.0192, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.42578125, | |
| "rewards/margins": 7.4375, | |
| "rewards/rejected": -7.875, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.3225806451612903, | |
| "grad_norm": 21.54322951685049, | |
| "learning_rate": 3.106332138590203e-07, | |
| "logits/chosen": -3.859375, | |
| "logits/rejected": -3.859375, | |
| "logps/chosen": -186.0, | |
| "logps/rejected": -274.0, | |
| "loss": 0.0143, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.150390625, | |
| "rewards/margins": 8.0, | |
| "rewards/rejected": -8.125, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 1.3548387096774195, | |
| "grad_norm": 25.577249022826813, | |
| "learning_rate": 3.046594982078853e-07, | |
| "logits/chosen": -3.78125, | |
| "logits/rejected": -3.78125, | |
| "logps/chosen": -193.0, | |
| "logps/rejected": -276.0, | |
| "loss": 0.0173, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.37890625, | |
| "rewards/margins": 7.28125, | |
| "rewards/rejected": -7.65625, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 1.3870967741935485, | |
| "grad_norm": 0.8297796576977869, | |
| "learning_rate": 2.986857825567503e-07, | |
| "logits/chosen": -3.828125, | |
| "logits/rejected": -3.765625, | |
| "logps/chosen": -190.0, | |
| "logps/rejected": -274.0, | |
| "loss": 0.0179, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.2333984375, | |
| "rewards/margins": 7.53125, | |
| "rewards/rejected": -7.75, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 1.4193548387096775, | |
| "grad_norm": 37.60882100528414, | |
| "learning_rate": 2.927120669056153e-07, | |
| "logits/chosen": -3.71875, | |
| "logits/rejected": -3.765625, | |
| "logps/chosen": -189.0, | |
| "logps/rejected": -278.0, | |
| "loss": 0.0169, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.3671875, | |
| "rewards/margins": 7.84375, | |
| "rewards/rejected": -8.1875, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 1.4516129032258065, | |
| "grad_norm": 9.165654729466173, | |
| "learning_rate": 2.8673835125448024e-07, | |
| "logits/chosen": -3.75, | |
| "logits/rejected": -3.6875, | |
| "logps/chosen": -196.0, | |
| "logps/rejected": -284.0, | |
| "loss": 0.0261, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -0.1904296875, | |
| "rewards/margins": 7.90625, | |
| "rewards/rejected": -8.0625, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 1.4838709677419355, | |
| "grad_norm": 2.112943941209443, | |
| "learning_rate": 2.807646356033453e-07, | |
| "logits/chosen": -3.640625, | |
| "logits/rejected": -3.65625, | |
| "logps/chosen": -190.0, | |
| "logps/rejected": -280.0, | |
| "loss": 0.0271, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -0.66796875, | |
| "rewards/margins": 7.8125, | |
| "rewards/rejected": -8.5, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 1.5161290322580645, | |
| "grad_norm": 53.66978666939312, | |
| "learning_rate": 2.747909199522103e-07, | |
| "logits/chosen": -3.6875, | |
| "logits/rejected": -3.6875, | |
| "logps/chosen": -198.0, | |
| "logps/rejected": -276.0, | |
| "loss": 0.024, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -0.890625, | |
| "rewards/margins": 7.5, | |
| "rewards/rejected": -8.375, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 1.5483870967741935, | |
| "grad_norm": 13.967124404944409, | |
| "learning_rate": 2.6881720430107523e-07, | |
| "logits/chosen": -3.671875, | |
| "logits/rejected": -3.671875, | |
| "logps/chosen": -198.0, | |
| "logps/rejected": -284.0, | |
| "loss": 0.0212, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.66796875, | |
| "rewards/margins": 7.8125, | |
| "rewards/rejected": -8.5, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 1.5806451612903225, | |
| "grad_norm": 1.505490222650231, | |
| "learning_rate": 2.628434886499403e-07, | |
| "logits/chosen": -3.59375, | |
| "logits/rejected": -3.625, | |
| "logps/chosen": -180.0, | |
| "logps/rejected": -266.0, | |
| "loss": 0.0126, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 0.1103515625, | |
| "rewards/margins": 7.96875, | |
| "rewards/rejected": -7.875, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 1.6129032258064515, | |
| "grad_norm": 4.552495049762715, | |
| "learning_rate": 2.5686977299880523e-07, | |
| "logits/chosen": -3.65625, | |
| "logits/rejected": -3.703125, | |
| "logps/chosen": -190.0, | |
| "logps/rejected": -282.0, | |
| "loss": 0.0185, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.5234375, | |
| "rewards/margins": 8.125, | |
| "rewards/rejected": -8.6875, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.6451612903225805, | |
| "grad_norm": 7.1849195987738925, | |
| "learning_rate": 2.508960573476702e-07, | |
| "logits/chosen": -3.625, | |
| "logits/rejected": -3.6875, | |
| "logps/chosen": -196.0, | |
| "logps/rejected": -280.0, | |
| "loss": 0.0084, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.63671875, | |
| "rewards/margins": 7.53125, | |
| "rewards/rejected": -8.1875, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 1.6774193548387095, | |
| "grad_norm": 1.599699862303645, | |
| "learning_rate": 2.449223416965352e-07, | |
| "logits/chosen": -3.53125, | |
| "logits/rejected": -3.578125, | |
| "logps/chosen": -188.0, | |
| "logps/rejected": -282.0, | |
| "loss": 0.0167, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -0.388671875, | |
| "rewards/margins": 8.1875, | |
| "rewards/rejected": -8.625, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 1.7096774193548387, | |
| "grad_norm": 2.623858535249452, | |
| "learning_rate": 2.389486260454002e-07, | |
| "logits/chosen": -3.734375, | |
| "logits/rejected": -3.6875, | |
| "logps/chosen": -212.0, | |
| "logps/rejected": -302.0, | |
| "loss": 0.0408, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.8046875, | |
| "rewards/margins": 8.1875, | |
| "rewards/rejected": -10.0, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 1.7419354838709677, | |
| "grad_norm": 72.63368145693447, | |
| "learning_rate": 2.3297491039426522e-07, | |
| "logits/chosen": -3.65625, | |
| "logits/rejected": -3.640625, | |
| "logps/chosen": -202.0, | |
| "logps/rejected": -296.0, | |
| "loss": 0.0512, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -1.2421875, | |
| "rewards/margins": 8.4375, | |
| "rewards/rejected": -9.6875, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 1.7741935483870968, | |
| "grad_norm": 3.1206520900533667, | |
| "learning_rate": 2.270011947431302e-07, | |
| "logits/chosen": -3.640625, | |
| "logits/rejected": -3.6875, | |
| "logps/chosen": -180.0, | |
| "logps/rejected": -266.0, | |
| "loss": 0.04, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 0.1953125, | |
| "rewards/margins": 7.46875, | |
| "rewards/rejected": -7.28125, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 1.8064516129032258, | |
| "grad_norm": 41.020674776975646, | |
| "learning_rate": 2.2102747909199522e-07, | |
| "logits/chosen": -3.6875, | |
| "logits/rejected": -3.6875, | |
| "logps/chosen": -194.0, | |
| "logps/rejected": -290.0, | |
| "loss": 0.0124, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.03125, | |
| "rewards/margins": 8.0625, | |
| "rewards/rejected": -9.125, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 1.838709677419355, | |
| "grad_norm": 4.010133968357106, | |
| "learning_rate": 2.1505376344086022e-07, | |
| "logits/chosen": -3.59375, | |
| "logits/rejected": -3.546875, | |
| "logps/chosen": -212.0, | |
| "logps/rejected": -316.0, | |
| "loss": 0.0198, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.328125, | |
| "rewards/margins": 9.25, | |
| "rewards/rejected": -11.625, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 1.870967741935484, | |
| "grad_norm": 7.069397017435533, | |
| "learning_rate": 2.0908004778972521e-07, | |
| "logits/chosen": -3.65625, | |
| "logits/rejected": -3.6875, | |
| "logps/chosen": -188.0, | |
| "logps/rejected": -282.0, | |
| "loss": 0.0133, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.1337890625, | |
| "rewards/margins": 8.375, | |
| "rewards/rejected": -8.5625, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 1.903225806451613, | |
| "grad_norm": 8.936652736646373, | |
| "learning_rate": 2.0310633213859019e-07, | |
| "logits/chosen": -3.625, | |
| "logits/rejected": -3.546875, | |
| "logps/chosen": -196.0, | |
| "logps/rejected": -300.0, | |
| "loss": 0.0118, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -1.171875, | |
| "rewards/margins": 8.875, | |
| "rewards/rejected": -10.0625, | |
| "step": 590 | |
| }, | |
| { | |
| "epoch": 1.935483870967742, | |
| "grad_norm": 7.8141795170212465, | |
| "learning_rate": 1.9713261648745518e-07, | |
| "logits/chosen": -3.546875, | |
| "logits/rejected": -3.640625, | |
| "logps/chosen": -190.0, | |
| "logps/rejected": -280.0, | |
| "loss": 0.0089, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.8515625, | |
| "rewards/margins": 8.125, | |
| "rewards/rejected": -9.0, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.967741935483871, | |
| "grad_norm": 5.395939675702438, | |
| "learning_rate": 1.9115890083632018e-07, | |
| "logits/chosen": -3.6875, | |
| "logits/rejected": -3.703125, | |
| "logps/chosen": -199.0, | |
| "logps/rejected": -292.0, | |
| "loss": 0.0136, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.86328125, | |
| "rewards/margins": 8.75, | |
| "rewards/rejected": -9.625, | |
| "step": 610 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 17.221158878981882, | |
| "learning_rate": 1.8518518518518516e-07, | |
| "logits/chosen": -3.6875, | |
| "logits/rejected": -3.6875, | |
| "logps/chosen": -197.0, | |
| "logps/rejected": -286.0, | |
| "loss": 0.0135, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.76953125, | |
| "rewards/margins": 8.3125, | |
| "rewards/rejected": -9.125, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_logits/chosen": -3.796875, | |
| "eval_logits/rejected": -3.765625, | |
| "eval_logps/chosen": -222.0, | |
| "eval_logps/rejected": -280.0, | |
| "eval_loss": 0.191436767578125, | |
| "eval_rewards/accuracies": 0.8928571343421936, | |
| "eval_rewards/chosen": -2.21875, | |
| "eval_rewards/margins": 5.40625, | |
| "eval_rewards/rejected": -7.625, | |
| "eval_runtime": 13.2451, | |
| "eval_samples_per_second": 15.1, | |
| "eval_steps_per_second": 0.528, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 2.032258064516129, | |
| "grad_norm": 0.45384363049400467, | |
| "learning_rate": 1.7921146953405018e-07, | |
| "logits/chosen": -3.5625, | |
| "logits/rejected": -3.578125, | |
| "logps/chosen": -191.0, | |
| "logps/rejected": -300.0, | |
| "loss": 0.0012, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.28125, | |
| "rewards/margins": 10.1875, | |
| "rewards/rejected": -10.4375, | |
| "step": 630 | |
| }, | |
| { | |
| "epoch": 2.064516129032258, | |
| "grad_norm": 2.0283023878654722, | |
| "learning_rate": 1.7323775388291518e-07, | |
| "logits/chosen": -3.609375, | |
| "logits/rejected": -3.640625, | |
| "logps/chosen": -191.0, | |
| "logps/rejected": -298.0, | |
| "loss": 0.0019, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.302734375, | |
| "rewards/margins": 10.25, | |
| "rewards/rejected": -10.5625, | |
| "step": 640 | |
| }, | |
| { | |
| "epoch": 2.096774193548387, | |
| "grad_norm": 0.9217664266560652, | |
| "learning_rate": 1.6726403823178015e-07, | |
| "logits/chosen": -3.6875, | |
| "logits/rejected": -3.75, | |
| "logps/chosen": -193.0, | |
| "logps/rejected": -298.0, | |
| "loss": 0.0013, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.93359375, | |
| "rewards/margins": 9.625, | |
| "rewards/rejected": -10.5625, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 2.129032258064516, | |
| "grad_norm": 0.44858819876518125, | |
| "learning_rate": 1.6129032258064515e-07, | |
| "logits/chosen": -3.6875, | |
| "logits/rejected": -3.703125, | |
| "logps/chosen": -207.0, | |
| "logps/rejected": -316.0, | |
| "loss": 0.0012, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.3046875, | |
| "rewards/margins": 10.125, | |
| "rewards/rejected": -11.4375, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 2.161290322580645, | |
| "grad_norm": 9.792046659236782, | |
| "learning_rate": 1.5531660692951015e-07, | |
| "logits/chosen": -3.578125, | |
| "logits/rejected": -3.59375, | |
| "logps/chosen": -199.0, | |
| "logps/rejected": -310.0, | |
| "loss": 0.0019, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.1875, | |
| "rewards/margins": 9.875, | |
| "rewards/rejected": -11.0625, | |
| "step": 670 | |
| }, | |
| { | |
| "epoch": 2.193548387096774, | |
| "grad_norm": 0.3983602269153926, | |
| "learning_rate": 1.4934289127837515e-07, | |
| "logits/chosen": -3.5, | |
| "logits/rejected": -3.703125, | |
| "logps/chosen": -188.0, | |
| "logps/rejected": -302.0, | |
| "loss": 0.0024, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.59375, | |
| "rewards/margins": 10.25, | |
| "rewards/rejected": -10.875, | |
| "step": 680 | |
| }, | |
| { | |
| "epoch": 2.225806451612903, | |
| "grad_norm": 1.2472683997191045, | |
| "learning_rate": 1.4336917562724012e-07, | |
| "logits/chosen": -3.59375, | |
| "logits/rejected": -3.65625, | |
| "logps/chosen": -202.0, | |
| "logps/rejected": -312.0, | |
| "loss": 0.0011, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.0546875, | |
| "rewards/margins": 10.25, | |
| "rewards/rejected": -11.3125, | |
| "step": 690 | |
| }, | |
| { | |
| "epoch": 2.258064516129032, | |
| "grad_norm": 0.9104056031581937, | |
| "learning_rate": 1.3739545997610514e-07, | |
| "logits/chosen": -3.453125, | |
| "logits/rejected": -3.53125, | |
| "logps/chosen": -192.0, | |
| "logps/rejected": -308.0, | |
| "loss": 0.0008, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.90625, | |
| "rewards/margins": 10.625, | |
| "rewards/rejected": -11.5625, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 2.2903225806451615, | |
| "grad_norm": 0.5005896742747854, | |
| "learning_rate": 1.3142174432497014e-07, | |
| "logits/chosen": -3.453125, | |
| "logits/rejected": -3.5625, | |
| "logps/chosen": -196.0, | |
| "logps/rejected": -316.0, | |
| "loss": 0.0015, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.51953125, | |
| "rewards/margins": 11.4375, | |
| "rewards/rejected": -11.9375, | |
| "step": 710 | |
| }, | |
| { | |
| "epoch": 2.3225806451612905, | |
| "grad_norm": 2.0961727669442247, | |
| "learning_rate": 1.254480286738351e-07, | |
| "logits/chosen": -3.53125, | |
| "logits/rejected": -3.546875, | |
| "logps/chosen": -206.0, | |
| "logps/rejected": -318.0, | |
| "loss": 0.0011, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.515625, | |
| "rewards/margins": 10.5, | |
| "rewards/rejected": -12.0, | |
| "step": 720 | |
| }, | |
| { | |
| "epoch": 2.3548387096774195, | |
| "grad_norm": 0.33710671811391674, | |
| "learning_rate": 1.194743130227001e-07, | |
| "logits/chosen": -3.53125, | |
| "logits/rejected": -3.578125, | |
| "logps/chosen": -201.0, | |
| "logps/rejected": -318.0, | |
| "loss": 0.0008, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.625, | |
| "rewards/margins": 10.375, | |
| "rewards/rejected": -12.0, | |
| "step": 730 | |
| }, | |
| { | |
| "epoch": 2.3870967741935485, | |
| "grad_norm": 3.7387698089394634, | |
| "learning_rate": 1.135005973715651e-07, | |
| "logits/chosen": -3.59375, | |
| "logits/rejected": -3.625, | |
| "logps/chosen": -202.0, | |
| "logps/rejected": -314.0, | |
| "loss": 0.0014, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.3671875, | |
| "rewards/margins": 10.375, | |
| "rewards/rejected": -11.75, | |
| "step": 740 | |
| }, | |
| { | |
| "epoch": 2.4193548387096775, | |
| "grad_norm": 3.5076436110564426, | |
| "learning_rate": 1.0752688172043011e-07, | |
| "logits/chosen": -3.5625, | |
| "logits/rejected": -3.6875, | |
| "logps/chosen": -192.0, | |
| "logps/rejected": -306.0, | |
| "loss": 0.0019, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.53125, | |
| "rewards/margins": 10.4375, | |
| "rewards/rejected": -11.0, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 2.4516129032258065, | |
| "grad_norm": 0.3051005963902528, | |
| "learning_rate": 1.0155316606929509e-07, | |
| "logits/chosen": -3.46875, | |
| "logits/rejected": -3.53125, | |
| "logps/chosen": -194.0, | |
| "logps/rejected": -312.0, | |
| "loss": 0.0009, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.94140625, | |
| "rewards/margins": 10.625, | |
| "rewards/rejected": -11.5625, | |
| "step": 760 | |
| }, | |
| { | |
| "epoch": 2.4838709677419355, | |
| "grad_norm": 0.4055555200925197, | |
| "learning_rate": 9.557945041816009e-08, | |
| "logits/chosen": -3.40625, | |
| "logits/rejected": -3.546875, | |
| "logps/chosen": -190.0, | |
| "logps/rejected": -314.0, | |
| "loss": 0.0014, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.76953125, | |
| "rewards/margins": 11.0, | |
| "rewards/rejected": -11.75, | |
| "step": 770 | |
| }, | |
| { | |
| "epoch": 2.5161290322580645, | |
| "grad_norm": 0.4423005388531537, | |
| "learning_rate": 8.960573476702509e-08, | |
| "logits/chosen": -3.5625, | |
| "logits/rejected": -3.5625, | |
| "logps/chosen": -211.0, | |
| "logps/rejected": -336.0, | |
| "loss": 0.0013, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.640625, | |
| "rewards/margins": 11.5, | |
| "rewards/rejected": -13.1875, | |
| "step": 780 | |
| }, | |
| { | |
| "epoch": 2.5483870967741935, | |
| "grad_norm": 0.34201712686499286, | |
| "learning_rate": 8.363201911589008e-08, | |
| "logits/chosen": -3.4375, | |
| "logits/rejected": -3.515625, | |
| "logps/chosen": -197.0, | |
| "logps/rejected": -312.0, | |
| "loss": 0.001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.81640625, | |
| "rewards/margins": 10.9375, | |
| "rewards/rejected": -11.75, | |
| "step": 790 | |
| }, | |
| { | |
| "epoch": 2.5806451612903225, | |
| "grad_norm": 0.08324339497209432, | |
| "learning_rate": 7.765830346475507e-08, | |
| "logits/chosen": -3.421875, | |
| "logits/rejected": -3.59375, | |
| "logps/chosen": -207.0, | |
| "logps/rejected": -330.0, | |
| "loss": 0.0005, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.609375, | |
| "rewards/margins": 11.5, | |
| "rewards/rejected": -13.125, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 2.6129032258064515, | |
| "grad_norm": 0.35882151597975387, | |
| "learning_rate": 7.168458781362006e-08, | |
| "logits/chosen": -3.59375, | |
| "logits/rejected": -3.6875, | |
| "logps/chosen": -202.0, | |
| "logps/rejected": -328.0, | |
| "loss": 0.0045, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.03125, | |
| "rewards/margins": 11.625, | |
| "rewards/rejected": -12.6875, | |
| "step": 810 | |
| }, | |
| { | |
| "epoch": 2.6451612903225805, | |
| "grad_norm": 0.3006723416131319, | |
| "learning_rate": 6.571087216248507e-08, | |
| "logits/chosen": -3.4375, | |
| "logits/rejected": -3.609375, | |
| "logps/chosen": -197.0, | |
| "logps/rejected": -322.0, | |
| "loss": 0.0026, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.40625, | |
| "rewards/margins": 11.0, | |
| "rewards/rejected": -12.4375, | |
| "step": 820 | |
| }, | |
| { | |
| "epoch": 2.6774193548387095, | |
| "grad_norm": 0.9135709315414355, | |
| "learning_rate": 5.973715651135006e-08, | |
| "logits/chosen": -3.40625, | |
| "logits/rejected": -3.546875, | |
| "logps/chosen": -194.0, | |
| "logps/rejected": -312.0, | |
| "loss": 0.0026, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.046875, | |
| "rewards/margins": 11.3125, | |
| "rewards/rejected": -12.3125, | |
| "step": 830 | |
| }, | |
| { | |
| "epoch": 2.709677419354839, | |
| "grad_norm": 0.11799710974470022, | |
| "learning_rate": 5.3763440860215054e-08, | |
| "logits/chosen": -3.453125, | |
| "logits/rejected": -3.609375, | |
| "logps/chosen": -193.0, | |
| "logps/rejected": -330.0, | |
| "loss": 0.001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.79296875, | |
| "rewards/margins": 12.8125, | |
| "rewards/rejected": -13.625, | |
| "step": 840 | |
| }, | |
| { | |
| "epoch": 2.741935483870968, | |
| "grad_norm": 0.24029168708895726, | |
| "learning_rate": 4.7789725209080046e-08, | |
| "logits/chosen": -3.484375, | |
| "logits/rejected": -3.5625, | |
| "logps/chosen": -200.0, | |
| "logps/rejected": -320.0, | |
| "loss": 0.0006, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.2734375, | |
| "rewards/margins": 10.875, | |
| "rewards/rejected": -12.1875, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 2.774193548387097, | |
| "grad_norm": 1.5601788534872705, | |
| "learning_rate": 4.181600955794504e-08, | |
| "logits/chosen": -3.484375, | |
| "logits/rejected": -3.625, | |
| "logps/chosen": -206.0, | |
| "logps/rejected": -326.0, | |
| "loss": 0.0009, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.265625, | |
| "rewards/margins": 11.1875, | |
| "rewards/rejected": -12.4375, | |
| "step": 860 | |
| }, | |
| { | |
| "epoch": 2.806451612903226, | |
| "grad_norm": 2.7780970090627104, | |
| "learning_rate": 3.584229390681003e-08, | |
| "logits/chosen": -3.53125, | |
| "logits/rejected": -3.65625, | |
| "logps/chosen": -201.0, | |
| "logps/rejected": -320.0, | |
| "loss": 0.0025, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.0546875, | |
| "rewards/margins": 10.8125, | |
| "rewards/rejected": -11.875, | |
| "step": 870 | |
| }, | |
| { | |
| "epoch": 2.838709677419355, | |
| "grad_norm": 0.36034720795454384, | |
| "learning_rate": 2.986857825567503e-08, | |
| "logits/chosen": -3.546875, | |
| "logits/rejected": -3.578125, | |
| "logps/chosen": -198.0, | |
| "logps/rejected": -318.0, | |
| "loss": 0.0008, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.86328125, | |
| "rewards/margins": 11.25, | |
| "rewards/rejected": -12.125, | |
| "step": 880 | |
| }, | |
| { | |
| "epoch": 2.870967741935484, | |
| "grad_norm": 0.2764304820275906, | |
| "learning_rate": 2.3894862604540023e-08, | |
| "logits/chosen": -3.59375, | |
| "logits/rejected": -3.671875, | |
| "logps/chosen": -203.0, | |
| "logps/rejected": -316.0, | |
| "loss": 0.0037, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.03125, | |
| "rewards/margins": 10.6875, | |
| "rewards/rejected": -11.75, | |
| "step": 890 | |
| }, | |
| { | |
| "epoch": 2.903225806451613, | |
| "grad_norm": 0.4293826404732342, | |
| "learning_rate": 1.7921146953405015e-08, | |
| "logits/chosen": -3.4375, | |
| "logits/rejected": -3.546875, | |
| "logps/chosen": -208.0, | |
| "logps/rejected": -322.0, | |
| "loss": 0.0011, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.859375, | |
| "rewards/margins": 10.875, | |
| "rewards/rejected": -12.6875, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 2.935483870967742, | |
| "grad_norm": 0.34603080271508346, | |
| "learning_rate": 1.1947431302270011e-08, | |
| "logits/chosen": -3.546875, | |
| "logits/rejected": -3.703125, | |
| "logps/chosen": -207.0, | |
| "logps/rejected": -320.0, | |
| "loss": 0.0007, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.265625, | |
| "rewards/margins": 11.0, | |
| "rewards/rejected": -12.25, | |
| "step": 910 | |
| }, | |
| { | |
| "epoch": 2.967741935483871, | |
| "grad_norm": 2.3868143014919885, | |
| "learning_rate": 5.973715651135006e-09, | |
| "logits/chosen": -3.4375, | |
| "logits/rejected": -3.65625, | |
| "logps/chosen": -188.0, | |
| "logps/rejected": -304.0, | |
| "loss": 0.002, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.71484375, | |
| "rewards/margins": 10.375, | |
| "rewards/rejected": -11.125, | |
| "step": 920 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 0.40772402916251804, | |
| "learning_rate": 0.0, | |
| "logits/chosen": -3.5, | |
| "logits/rejected": -3.609375, | |
| "logps/chosen": -194.0, | |
| "logps/rejected": -320.0, | |
| "loss": 0.0006, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.1796875, | |
| "rewards/margins": 10.9375, | |
| "rewards/rejected": -12.125, | |
| "step": 930 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_logits/chosen": -3.734375, | |
| "eval_logits/rejected": -3.75, | |
| "eval_logps/chosen": -226.0, | |
| "eval_logps/rejected": -294.0, | |
| "eval_loss": 0.19125854969024658, | |
| "eval_rewards/accuracies": 0.8928571343421936, | |
| "eval_rewards/chosen": -2.703125, | |
| "eval_rewards/margins": 6.3125, | |
| "eval_rewards/rejected": -9.0, | |
| "eval_runtime": 16.0795, | |
| "eval_samples_per_second": 12.438, | |
| "eval_steps_per_second": 0.435, | |
| "step": 930 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "step": 930, | |
| "total_flos": 0.0, | |
| "train_loss": 0.13316144410480735, | |
| "train_runtime": 5554.5038, | |
| "train_samples_per_second": 5.355, | |
| "train_steps_per_second": 0.167 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 930, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "save_steps": 500, | |
| "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": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |