| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 1.0, | |
| "eval_steps": 500, | |
| "global_step": 250, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.004, | |
| "grad_norm": 90.84001974484835, | |
| "learning_rate": 1.25e-07, | |
| "logits/chosen": 0.98046875, | |
| "logits/rejected": 0.96484375, | |
| "logps/chosen": -25.75, | |
| "logps/rejected": -28.625, | |
| "loss": 3.419, | |
| "nll_loss": 2.703125, | |
| "rewards/accuracies": 0.53125, | |
| "rewards/chosen": 0.056396484375, | |
| "rewards/margins": 0.06689453125, | |
| "rewards/rejected": -0.010498046875, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.008, | |
| "grad_norm": 93.63863607247485, | |
| "learning_rate": 2.5e-07, | |
| "logits/chosen": 1.1015625, | |
| "logits/rejected": 1.078125, | |
| "logps/chosen": -32.25, | |
| "logps/rejected": -30.25, | |
| "loss": 3.6316, | |
| "nll_loss": 3.359375, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": 0.0115966796875, | |
| "rewards/margins": 0.01458740234375, | |
| "rewards/rejected": -0.003021240234375, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.012, | |
| "grad_norm": 89.09751114393859, | |
| "learning_rate": 3.75e-07, | |
| "logits/chosen": 0.875, | |
| "logits/rejected": 0.8984375, | |
| "logps/chosen": -27.5, | |
| "logps/rejected": -30.5, | |
| "loss": 3.603, | |
| "nll_loss": 2.671875, | |
| "rewards/accuracies": 0.46875, | |
| "rewards/chosen": -0.024658203125, | |
| "rewards/margins": -0.03662109375, | |
| "rewards/rejected": 0.0120849609375, | |
| "step": 3 | |
| }, | |
| { | |
| "epoch": 0.016, | |
| "grad_norm": 93.60471022707792, | |
| "learning_rate": 5e-07, | |
| "logits/chosen": 0.94921875, | |
| "logits/rejected": 0.94140625, | |
| "logps/chosen": -23.5, | |
| "logps/rejected": -29.125, | |
| "loss": 3.7088, | |
| "nll_loss": 2.296875, | |
| "rewards/accuracies": 0.375, | |
| "rewards/chosen": 0.027587890625, | |
| "rewards/margins": -0.0021209716796875, | |
| "rewards/rejected": 0.0296630859375, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.02, | |
| "grad_norm": 93.20328417318827, | |
| "learning_rate": 6.249999999999999e-07, | |
| "logits/chosen": 0.82421875, | |
| "logits/rejected": 0.87109375, | |
| "logps/chosen": -24.375, | |
| "logps/rejected": -30.625, | |
| "loss": 3.4743, | |
| "nll_loss": 2.53125, | |
| "rewards/accuracies": 0.53125, | |
| "rewards/chosen": 0.051025390625, | |
| "rewards/margins": 0.0145263671875, | |
| "rewards/rejected": 0.036376953125, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.024, | |
| "grad_norm": 91.22976543634373, | |
| "learning_rate": 7.5e-07, | |
| "logits/chosen": 0.94140625, | |
| "logits/rejected": 0.97265625, | |
| "logps/chosen": -28.125, | |
| "logps/rejected": -30.375, | |
| "loss": 3.5928, | |
| "nll_loss": 2.78125, | |
| "rewards/accuracies": 0.46875, | |
| "rewards/chosen": -0.00156402587890625, | |
| "rewards/margins": -0.03955078125, | |
| "rewards/rejected": 0.0380859375, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.028, | |
| "grad_norm": 101.68787999817542, | |
| "learning_rate": 8.75e-07, | |
| "logits/chosen": 0.9453125, | |
| "logits/rejected": 0.9140625, | |
| "logps/chosen": -21.625, | |
| "logps/rejected": -28.25, | |
| "loss": 3.546, | |
| "nll_loss": 2.125, | |
| "rewards/accuracies": 0.53125, | |
| "rewards/chosen": 0.06884765625, | |
| "rewards/margins": 0.051025390625, | |
| "rewards/rejected": 0.01806640625, | |
| "step": 7 | |
| }, | |
| { | |
| "epoch": 0.032, | |
| "grad_norm": 86.70077888201412, | |
| "learning_rate": 1e-06, | |
| "logits/chosen": 0.875, | |
| "logits/rejected": 0.87109375, | |
| "logps/chosen": -24.25, | |
| "logps/rejected": -26.625, | |
| "loss": 3.4702, | |
| "nll_loss": 2.609375, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 0.11181640625, | |
| "rewards/margins": 0.07763671875, | |
| "rewards/rejected": 0.033935546875, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.036, | |
| "grad_norm": 97.99302928138263, | |
| "learning_rate": 9.999578688879084e-07, | |
| "logits/chosen": 0.98046875, | |
| "logits/rejected": 0.98828125, | |
| "logps/chosen": -25.0, | |
| "logps/rejected": -30.625, | |
| "loss": 3.7406, | |
| "nll_loss": 2.4375, | |
| "rewards/accuracies": 0.40625, | |
| "rewards/chosen": 0.10595703125, | |
| "rewards/margins": 0.004241943359375, | |
| "rewards/rejected": 0.10205078125, | |
| "step": 9 | |
| }, | |
| { | |
| "epoch": 0.04, | |
| "grad_norm": 81.98951839046792, | |
| "learning_rate": 9.998314826517563e-07, | |
| "logits/chosen": 0.9921875, | |
| "logits/rejected": 0.99609375, | |
| "logps/chosen": -22.125, | |
| "logps/rejected": -25.0, | |
| "loss": 3.3388, | |
| "nll_loss": 2.21875, | |
| "rewards/accuracies": 0.53125, | |
| "rewards/chosen": 0.201171875, | |
| "rewards/margins": 0.0306396484375, | |
| "rewards/rejected": 0.1708984375, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.044, | |
| "grad_norm": 76.29651704091002, | |
| "learning_rate": 9.99620862590714e-07, | |
| "logits/chosen": 0.94921875, | |
| "logits/rejected": 0.9375, | |
| "logps/chosen": -21.375, | |
| "logps/rejected": -25.75, | |
| "loss": 3.3012, | |
| "nll_loss": 2.1875, | |
| "rewards/accuracies": 0.46875, | |
| "rewards/chosen": 0.27734375, | |
| "rewards/margins": -0.040771484375, | |
| "rewards/rejected": 0.318359375, | |
| "step": 11 | |
| }, | |
| { | |
| "epoch": 0.048, | |
| "grad_norm": 81.7871405369469, | |
| "learning_rate": 9.993260441994114e-07, | |
| "logits/chosen": 0.83984375, | |
| "logits/rejected": 0.83984375, | |
| "logps/chosen": -21.875, | |
| "logps/rejected": -22.125, | |
| "loss": 3.2064, | |
| "nll_loss": 2.296875, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": 0.3359375, | |
| "rewards/margins": 0.0595703125, | |
| "rewards/rejected": 0.275390625, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 0.052, | |
| "grad_norm": 103.64982635961991, | |
| "learning_rate": 9.989470771619553e-07, | |
| "logits/chosen": 0.9921875, | |
| "logits/rejected": 1.015625, | |
| "logps/chosen": -26.625, | |
| "logps/rejected": -24.5, | |
| "loss": 3.3844, | |
| "nll_loss": 2.671875, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 0.451171875, | |
| "rewards/margins": 0.11865234375, | |
| "rewards/rejected": 0.333984375, | |
| "step": 13 | |
| }, | |
| { | |
| "epoch": 0.056, | |
| "grad_norm": 42.34708755561702, | |
| "learning_rate": 9.984840253435568e-07, | |
| "logits/chosen": 0.9296875, | |
| "logits/rejected": 1.0234375, | |
| "logps/chosen": -18.375, | |
| "logps/rejected": -19.0, | |
| "loss": 2.7424, | |
| "nll_loss": 1.828125, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": 0.7109375, | |
| "rewards/margins": 0.08544921875, | |
| "rewards/rejected": 0.625, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 0.06, | |
| "grad_norm": 37.96207461928269, | |
| "learning_rate": 9.979369667797674e-07, | |
| "logits/chosen": 0.98046875, | |
| "logits/rejected": 0.95703125, | |
| "logps/chosen": -16.625, | |
| "logps/rejected": -22.625, | |
| "loss": 2.7065, | |
| "nll_loss": 1.625, | |
| "rewards/accuracies": 0.4375, | |
| "rewards/chosen": 0.74609375, | |
| "rewards/margins": -0.111328125, | |
| "rewards/rejected": 0.85546875, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.064, | |
| "grad_norm": 35.96184288602604, | |
| "learning_rate": 9.973059936633306e-07, | |
| "logits/chosen": 0.9609375, | |
| "logits/rejected": 0.97265625, | |
| "logps/chosen": -15.25, | |
| "logps/rejected": -22.0, | |
| "loss": 2.7223, | |
| "nll_loss": 1.515625, | |
| "rewards/accuracies": 0.34375, | |
| "rewards/chosen": 0.6875, | |
| "rewards/margins": -0.1298828125, | |
| "rewards/rejected": 0.81640625, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 0.068, | |
| "grad_norm": 38.480584517300954, | |
| "learning_rate": 9.965912123286424e-07, | |
| "logits/chosen": 0.9921875, | |
| "logits/rejected": 1.015625, | |
| "logps/chosen": -20.625, | |
| "logps/rejected": -23.5, | |
| "loss": 2.7521, | |
| "nll_loss": 2.015625, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": 1.0, | |
| "rewards/margins": 0.3046875, | |
| "rewards/rejected": 0.69921875, | |
| "step": 17 | |
| }, | |
| { | |
| "epoch": 0.072, | |
| "grad_norm": 42.985647717347646, | |
| "learning_rate": 9.95792743233833e-07, | |
| "logits/chosen": 0.9296875, | |
| "logits/rejected": 0.96484375, | |
| "logps/chosen": -17.75, | |
| "logps/rejected": -22.25, | |
| "loss": 2.6274, | |
| "nll_loss": 1.875, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": 0.890625, | |
| "rewards/margins": 0.0089111328125, | |
| "rewards/rejected": 0.8828125, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 0.076, | |
| "grad_norm": 35.66016932258093, | |
| "learning_rate": 9.949107209404663e-07, | |
| "logits/chosen": 0.984375, | |
| "logits/rejected": 1.0078125, | |
| "logps/chosen": -17.0, | |
| "logps/rejected": -20.0, | |
| "loss": 2.5359, | |
| "nll_loss": 1.734375, | |
| "rewards/accuracies": 0.53125, | |
| "rewards/chosen": 0.9921875, | |
| "rewards/margins": 0.0654296875, | |
| "rewards/rejected": 0.92578125, | |
| "step": 19 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "grad_norm": 32.78942676966944, | |
| "learning_rate": 9.939452940908626e-07, | |
| "logits/chosen": 0.93359375, | |
| "logits/rejected": 0.9140625, | |
| "logps/chosen": -19.375, | |
| "logps/rejected": -20.75, | |
| "loss": 2.583, | |
| "nll_loss": 1.921875, | |
| "rewards/accuracies": 0.46875, | |
| "rewards/chosen": 1.0703125, | |
| "rewards/margins": 0.03466796875, | |
| "rewards/rejected": 1.03125, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.084, | |
| "grad_norm": 30.390224965176014, | |
| "learning_rate": 9.92896625383049e-07, | |
| "logits/chosen": 0.9609375, | |
| "logits/rejected": 0.97265625, | |
| "logps/chosen": -16.0, | |
| "logps/rejected": -20.0, | |
| "loss": 2.4877, | |
| "nll_loss": 1.65625, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": 1.140625, | |
| "rewards/margins": 0.0869140625, | |
| "rewards/rejected": 1.046875, | |
| "step": 21 | |
| }, | |
| { | |
| "epoch": 0.088, | |
| "grad_norm": 26.189318520687596, | |
| "learning_rate": 9.917648915433411e-07, | |
| "logits/chosen": 1.0234375, | |
| "logits/rejected": 1.015625, | |
| "logps/chosen": -14.75, | |
| "logps/rejected": -16.625, | |
| "loss": 2.3129, | |
| "nll_loss": 1.4296875, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 1.296875, | |
| "rewards/margins": 0.31640625, | |
| "rewards/rejected": 0.984375, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 0.092, | |
| "grad_norm": 22.39496682670829, | |
| "learning_rate": 9.905502832965603e-07, | |
| "logits/chosen": 1.0234375, | |
| "logits/rejected": 1.015625, | |
| "logps/chosen": -14.6875, | |
| "logps/rejected": -17.5, | |
| "loss": 2.2336, | |
| "nll_loss": 1.4609375, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": 1.2421875, | |
| "rewards/margins": 0.07958984375, | |
| "rewards/rejected": 1.1640625, | |
| "step": 23 | |
| }, | |
| { | |
| "epoch": 0.096, | |
| "grad_norm": 22.358547425083277, | |
| "learning_rate": 9.892530053338907e-07, | |
| "logits/chosen": 1.078125, | |
| "logits/rejected": 1.0625, | |
| "logps/chosen": -15.0625, | |
| "logps/rejected": -20.625, | |
| "loss": 2.2508, | |
| "nll_loss": 1.53125, | |
| "rewards/accuracies": 0.53125, | |
| "rewards/chosen": 1.1875, | |
| "rewards/margins": -0.14453125, | |
| "rewards/rejected": 1.3359375, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "grad_norm": 25.286222055847446, | |
| "learning_rate": 9.87873276278386e-07, | |
| "logits/chosen": 1.046875, | |
| "logits/rejected": 1.0234375, | |
| "logps/chosen": -13.9375, | |
| "logps/rejected": -17.125, | |
| "loss": 2.2224, | |
| "nll_loss": 1.3828125, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": 1.421875, | |
| "rewards/margins": 0.3125, | |
| "rewards/rejected": 1.109375, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.104, | |
| "grad_norm": 22.227874316855747, | |
| "learning_rate": 9.864113286481236e-07, | |
| "logits/chosen": 1.0, | |
| "logits/rejected": 1.0234375, | |
| "logps/chosen": -12.875, | |
| "logps/rejected": -18.75, | |
| "loss": 2.1859, | |
| "nll_loss": 1.3515625, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 1.4609375, | |
| "rewards/margins": 0.263671875, | |
| "rewards/rejected": 1.1953125, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 0.108, | |
| "grad_norm": 20.247096321249927, | |
| "learning_rate": 9.84867408817022e-07, | |
| "logits/chosen": 1.046875, | |
| "logits/rejected": 1.0703125, | |
| "logps/chosen": -12.5, | |
| "logps/rejected": -16.375, | |
| "loss": 2.0957, | |
| "nll_loss": 1.265625, | |
| "rewards/accuracies": 0.53125, | |
| "rewards/chosen": 1.4375, | |
| "rewards/margins": 0.26953125, | |
| "rewards/rejected": 1.1640625, | |
| "step": 27 | |
| }, | |
| { | |
| "epoch": 0.112, | |
| "grad_norm": 19.436858660866225, | |
| "learning_rate": 9.832417769733185e-07, | |
| "logits/chosen": 1.0703125, | |
| "logits/rejected": 1.0859375, | |
| "logps/chosen": -12.25, | |
| "logps/rejected": -16.625, | |
| "loss": 2.1721, | |
| "nll_loss": 1.1640625, | |
| "rewards/accuracies": 0.5625, | |
| "rewards/chosen": 1.484375, | |
| "rewards/margins": 0.1494140625, | |
| "rewards/rejected": 1.3359375, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 0.116, | |
| "grad_norm": 18.047996057799406, | |
| "learning_rate": 9.815347070757231e-07, | |
| "logits/chosen": 1.015625, | |
| "logits/rejected": 1.0, | |
| "logps/chosen": -14.3125, | |
| "logps/rejected": -16.875, | |
| "loss": 2.1233, | |
| "nll_loss": 1.3203125, | |
| "rewards/accuracies": 0.53125, | |
| "rewards/chosen": 1.8515625, | |
| "rewards/margins": 0.283203125, | |
| "rewards/rejected": 1.5703125, | |
| "step": 29 | |
| }, | |
| { | |
| "epoch": 0.12, | |
| "grad_norm": 18.07514136915611, | |
| "learning_rate": 9.797464868072486e-07, | |
| "logits/chosen": 1.0625, | |
| "logits/rejected": 1.109375, | |
| "logps/chosen": -12.5, | |
| "logps/rejected": -16.875, | |
| "loss": 2.1027, | |
| "nll_loss": 1.2421875, | |
| "rewards/accuracies": 0.5625, | |
| "rewards/chosen": 1.390625, | |
| "rewards/margins": 0.22265625, | |
| "rewards/rejected": 1.171875, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.124, | |
| "grad_norm": 16.76574433996455, | |
| "learning_rate": 9.778774175267295e-07, | |
| "logits/chosen": 1.0390625, | |
| "logits/rejected": 1.0546875, | |
| "logps/chosen": -11.125, | |
| "logps/rejected": -17.625, | |
| "loss": 1.9701, | |
| "nll_loss": 1.125, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 1.5546875, | |
| "rewards/margins": 0.296875, | |
| "rewards/rejected": 1.2578125, | |
| "step": 31 | |
| }, | |
| { | |
| "epoch": 0.128, | |
| "grad_norm": 18.753306438627558, | |
| "learning_rate": 9.759278142180347e-07, | |
| "logits/chosen": 1.0546875, | |
| "logits/rejected": 1.0703125, | |
| "logps/chosen": -10.125, | |
| "logps/rejected": -15.4375, | |
| "loss": 1.9336, | |
| "nll_loss": 1.0078125, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 1.15625, | |
| "rewards/margins": 0.150390625, | |
| "rewards/rejected": 1.0078125, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 0.132, | |
| "grad_norm": 15.50417437616818, | |
| "learning_rate": 9.73898005436987e-07, | |
| "logits/chosen": 1.03125, | |
| "logits/rejected": 1.03125, | |
| "logps/chosen": -11.0625, | |
| "logps/rejected": -14.75, | |
| "loss": 1.9432, | |
| "nll_loss": 1.0703125, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 1.4296875, | |
| "rewards/margins": 0.333984375, | |
| "rewards/rejected": 1.09375, | |
| "step": 33 | |
| }, | |
| { | |
| "epoch": 0.136, | |
| "grad_norm": 16.396192118981354, | |
| "learning_rate": 9.71788333255991e-07, | |
| "logits/chosen": 1.0390625, | |
| "logits/rejected": 1.0546875, | |
| "logps/chosen": -11.5625, | |
| "logps/rejected": -15.4375, | |
| "loss": 1.8603, | |
| "nll_loss": 1.2265625, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": 1.796875, | |
| "rewards/margins": 0.78125, | |
| "rewards/rejected": 1.015625, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 0.14, | |
| "grad_norm": 13.69362502638984, | |
| "learning_rate": 9.695991532063874e-07, | |
| "logits/chosen": 1.0390625, | |
| "logits/rejected": 1.0625, | |
| "logps/chosen": -11.125, | |
| "logps/rejected": -14.875, | |
| "loss": 1.9364, | |
| "nll_loss": 1.1015625, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 1.90625, | |
| "rewards/margins": 0.380859375, | |
| "rewards/rejected": 1.5234375, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.144, | |
| "grad_norm": 13.536022508840675, | |
| "learning_rate": 9.673308342185365e-07, | |
| "logits/chosen": 1.0703125, | |
| "logits/rejected": 1.0859375, | |
| "logps/chosen": -11.5625, | |
| "logps/rejected": -15.625, | |
| "loss": 1.8253, | |
| "nll_loss": 1.15625, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 1.78125, | |
| "rewards/margins": 0.3359375, | |
| "rewards/rejected": 1.4453125, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 0.148, | |
| "grad_norm": 15.477743522256894, | |
| "learning_rate": 9.649837585596444e-07, | |
| "logits/chosen": 1.125, | |
| "logits/rejected": 1.0859375, | |
| "logps/chosen": -9.5625, | |
| "logps/rejected": -16.375, | |
| "loss": 1.7562, | |
| "nll_loss": 0.88671875, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 1.8671875, | |
| "rewards/margins": 0.5078125, | |
| "rewards/rejected": 1.359375, | |
| "step": 37 | |
| }, | |
| { | |
| "epoch": 0.152, | |
| "grad_norm": 15.139542721455184, | |
| "learning_rate": 9.625583217693418e-07, | |
| "logits/chosen": 1.1171875, | |
| "logits/rejected": 1.125, | |
| "logps/chosen": -10.75, | |
| "logps/rejected": -14.625, | |
| "loss": 1.8019, | |
| "nll_loss": 1.0234375, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 2.265625, | |
| "rewards/margins": 0.83984375, | |
| "rewards/rejected": 1.4296875, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 0.156, | |
| "grad_norm": 17.039592344946968, | |
| "learning_rate": 9.600549325930259e-07, | |
| "logits/chosen": 1.1953125, | |
| "logits/rejected": 1.1953125, | |
| "logps/chosen": -10.6875, | |
| "logps/rejected": -16.125, | |
| "loss": 1.7975, | |
| "nll_loss": 1.078125, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 1.9921875, | |
| "rewards/margins": 0.37890625, | |
| "rewards/rejected": 1.609375, | |
| "step": 39 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "grad_norm": 11.945267822896529, | |
| "learning_rate": 9.574740129129765e-07, | |
| "logits/chosen": 1.1953125, | |
| "logits/rejected": 1.1875, | |
| "logps/chosen": -10.5625, | |
| "logps/rejected": -14.125, | |
| "loss": 1.6304, | |
| "nll_loss": 1.0078125, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 1.796875, | |
| "rewards/margins": 0.482421875, | |
| "rewards/rejected": 1.3125, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.164, | |
| "grad_norm": 13.26304188910449, | |
| "learning_rate": 9.548159976772592e-07, | |
| "logits/chosen": 1.171875, | |
| "logits/rejected": 1.1796875, | |
| "logps/chosen": -10.3125, | |
| "logps/rejected": -15.125, | |
| "loss": 1.7527, | |
| "nll_loss": 1.03125, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": 1.8359375, | |
| "rewards/margins": 0.46875, | |
| "rewards/rejected": 1.3671875, | |
| "step": 41 | |
| }, | |
| { | |
| "epoch": 0.168, | |
| "grad_norm": 13.145142193151658, | |
| "learning_rate": 9.520813348264251e-07, | |
| "logits/chosen": 1.1953125, | |
| "logits/rejected": 1.2109375, | |
| "logps/chosen": -9.25, | |
| "logps/rejected": -19.75, | |
| "loss": 1.6491, | |
| "nll_loss": 0.8828125, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 2.109375, | |
| "rewards/margins": 0.859375, | |
| "rewards/rejected": 1.25, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 0.172, | |
| "grad_norm": 14.23868046028076, | |
| "learning_rate": 9.492704852180226e-07, | |
| "logits/chosen": 1.1796875, | |
| "logits/rejected": 1.171875, | |
| "logps/chosen": -8.75, | |
| "logps/rejected": -16.0, | |
| "loss": 1.6864, | |
| "nll_loss": 0.796875, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 1.8671875, | |
| "rewards/margins": 0.7578125, | |
| "rewards/rejected": 1.109375, | |
| "step": 43 | |
| }, | |
| { | |
| "epoch": 0.176, | |
| "grad_norm": 13.971650217328486, | |
| "learning_rate": 9.463839225489318e-07, | |
| "logits/chosen": 1.2109375, | |
| "logits/rejected": 1.1796875, | |
| "logps/chosen": -10.0625, | |
| "logps/rejected": -15.125, | |
| "loss": 1.6841, | |
| "nll_loss": 1.0, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 2.046875, | |
| "rewards/margins": 0.83203125, | |
| "rewards/rejected": 1.21875, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 0.18, | |
| "grad_norm": 12.015757364604882, | |
| "learning_rate": 9.434221332755338e-07, | |
| "logits/chosen": 1.1640625, | |
| "logits/rejected": 1.15625, | |
| "logps/chosen": -9.375, | |
| "logps/rejected": -11.75, | |
| "loss": 1.6724, | |
| "nll_loss": 0.9375, | |
| "rewards/accuracies": 0.53125, | |
| "rewards/chosen": 1.921875, | |
| "rewards/margins": 0.134765625, | |
| "rewards/rejected": 1.7890625, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 0.184, | |
| "grad_norm": 11.903742021506595, | |
| "learning_rate": 9.403856165317321e-07, | |
| "logits/chosen": 1.1640625, | |
| "logits/rejected": 1.1953125, | |
| "logps/chosen": -8.75, | |
| "logps/rejected": -19.0, | |
| "loss": 1.546, | |
| "nll_loss": 0.8671875, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 1.7578125, | |
| "rewards/margins": 0.60546875, | |
| "rewards/rejected": 1.15625, | |
| "step": 46 | |
| }, | |
| { | |
| "epoch": 0.188, | |
| "grad_norm": 11.460580941237025, | |
| "learning_rate": 9.372748840448361e-07, | |
| "logits/chosen": 1.203125, | |
| "logits/rejected": 1.2421875, | |
| "logps/chosen": -6.375, | |
| "logps/rejected": -17.375, | |
| "loss": 1.5881, | |
| "nll_loss": 0.65234375, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 1.8046875, | |
| "rewards/margins": 0.46875, | |
| "rewards/rejected": 1.3359375, | |
| "step": 47 | |
| }, | |
| { | |
| "epoch": 0.192, | |
| "grad_norm": 12.66271092485468, | |
| "learning_rate": 9.340904600493219e-07, | |
| "logits/chosen": 1.1875, | |
| "logits/rejected": 1.1953125, | |
| "logps/chosen": -7.65625, | |
| "logps/rejected": -17.25, | |
| "loss": 1.583, | |
| "nll_loss": 0.828125, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 1.8984375, | |
| "rewards/margins": 0.53125, | |
| "rewards/rejected": 1.3671875, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 0.196, | |
| "grad_norm": 14.901259679926502, | |
| "learning_rate": 9.308328811984869e-07, | |
| "logits/chosen": 1.171875, | |
| "logits/rejected": 1.2265625, | |
| "logps/chosen": -7.90625, | |
| "logps/rejected": -18.375, | |
| "loss": 1.4589, | |
| "nll_loss": 0.7421875, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 1.875, | |
| "rewards/margins": 0.828125, | |
| "rewards/rejected": 1.046875, | |
| "step": 49 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "grad_norm": 11.737584663990637, | |
| "learning_rate": 9.275026964740101e-07, | |
| "logits/chosen": 1.140625, | |
| "logits/rejected": 1.1171875, | |
| "logps/chosen": -8.125, | |
| "logps/rejected": -13.1875, | |
| "loss": 1.4422, | |
| "nll_loss": 0.7578125, | |
| "rewards/accuracies": 0.5625, | |
| "rewards/chosen": 1.96875, | |
| "rewards/margins": 0.5390625, | |
| "rewards/rejected": 1.4296875, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.204, | |
| "grad_norm": 10.465507566086854, | |
| "learning_rate": 9.241004670934347e-07, | |
| "logits/chosen": 1.1328125, | |
| "logits/rejected": 1.1484375, | |
| "logps/chosen": -7.25, | |
| "logps/rejected": -17.875, | |
| "loss": 1.5697, | |
| "nll_loss": 0.671875, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 1.6875, | |
| "rewards/margins": 0.5859375, | |
| "rewards/rejected": 1.1015625, | |
| "step": 51 | |
| }, | |
| { | |
| "epoch": 0.208, | |
| "grad_norm": 11.645713255347133, | |
| "learning_rate": 9.206267664155906e-07, | |
| "logits/chosen": 1.203125, | |
| "logits/rejected": 1.1875, | |
| "logps/chosen": -9.125, | |
| "logps/rejected": -16.0, | |
| "loss": 1.5648, | |
| "nll_loss": 0.94140625, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 2.109375, | |
| "rewards/margins": 0.8046875, | |
| "rewards/rejected": 1.3046875, | |
| "step": 52 | |
| }, | |
| { | |
| "epoch": 0.212, | |
| "grad_norm": 12.122656566323922, | |
| "learning_rate": 9.170821798439683e-07, | |
| "logits/chosen": 1.15625, | |
| "logits/rejected": 1.171875, | |
| "logps/chosen": -6.90625, | |
| "logps/rejected": -17.125, | |
| "loss": 1.525, | |
| "nll_loss": 0.7109375, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 1.8828125, | |
| "rewards/margins": 0.74609375, | |
| "rewards/rejected": 1.1328125, | |
| "step": 53 | |
| }, | |
| { | |
| "epoch": 0.216, | |
| "grad_norm": 10.904740492194488, | |
| "learning_rate": 9.134673047280643e-07, | |
| "logits/chosen": 1.140625, | |
| "logits/rejected": 1.171875, | |
| "logps/chosen": -7.375, | |
| "logps/rejected": -11.5, | |
| "loss": 1.5872, | |
| "nll_loss": 0.6796875, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 1.6953125, | |
| "rewards/margins": 0.31640625, | |
| "rewards/rejected": 1.375, | |
| "step": 54 | |
| }, | |
| { | |
| "epoch": 0.22, | |
| "grad_norm": 11.94244354707238, | |
| "learning_rate": 9.097827502627136e-07, | |
| "logits/chosen": 1.1796875, | |
| "logits/rejected": 1.1953125, | |
| "logps/chosen": -6.71875, | |
| "logps/rejected": -13.625, | |
| "loss": 1.4571, | |
| "nll_loss": 0.62890625, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 2.203125, | |
| "rewards/margins": 0.326171875, | |
| "rewards/rejected": 1.875, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 0.224, | |
| "grad_norm": 10.898501851960246, | |
| "learning_rate": 9.06029137385425e-07, | |
| "logits/chosen": 1.21875, | |
| "logits/rejected": 1.171875, | |
| "logps/chosen": -8.25, | |
| "logps/rejected": -15.375, | |
| "loss": 1.4988, | |
| "nll_loss": 0.79296875, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 2.421875, | |
| "rewards/margins": 0.65234375, | |
| "rewards/rejected": 1.7734375, | |
| "step": 56 | |
| }, | |
| { | |
| "epoch": 0.228, | |
| "grad_norm": 11.292375773132523, | |
| "learning_rate": 9.022070986717378e-07, | |
| "logits/chosen": 1.203125, | |
| "logits/rejected": 1.1796875, | |
| "logps/chosen": -8.8125, | |
| "logps/rejected": -17.5, | |
| "loss": 1.5525, | |
| "nll_loss": 0.828125, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 1.9765625, | |
| "rewards/margins": 0.74609375, | |
| "rewards/rejected": 1.234375, | |
| "step": 57 | |
| }, | |
| { | |
| "epoch": 0.232, | |
| "grad_norm": 13.677184544578704, | |
| "learning_rate": 8.983172782286179e-07, | |
| "logits/chosen": 1.21875, | |
| "logits/rejected": 1.21875, | |
| "logps/chosen": -6.5, | |
| "logps/rejected": -12.9375, | |
| "loss": 1.5552, | |
| "nll_loss": 0.7109375, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": 2.515625, | |
| "rewards/margins": 0.2275390625, | |
| "rewards/rejected": 2.28125, | |
| "step": 58 | |
| }, | |
| { | |
| "epoch": 0.236, | |
| "grad_norm": 13.654723770150948, | |
| "learning_rate": 8.943603315859101e-07, | |
| "logits/chosen": 1.2421875, | |
| "logits/rejected": 1.2421875, | |
| "logps/chosen": -6.96875, | |
| "logps/rejected": -15.375, | |
| "loss": 1.4327, | |
| "nll_loss": 0.66015625, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": 1.9921875, | |
| "rewards/margins": 1.0078125, | |
| "rewards/rejected": 0.984375, | |
| "step": 59 | |
| }, | |
| { | |
| "epoch": 0.24, | |
| "grad_norm": 10.361981541893595, | |
| "learning_rate": 8.90336925585864e-07, | |
| "logits/chosen": 1.1796875, | |
| "logits/rejected": 1.2109375, | |
| "logps/chosen": -5.71875, | |
| "logps/rejected": -16.5, | |
| "loss": 1.5013, | |
| "nll_loss": 0.5859375, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": 2.296875, | |
| "rewards/margins": 0.99609375, | |
| "rewards/rejected": 1.296875, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.244, | |
| "grad_norm": 12.834755538872065, | |
| "learning_rate": 8.862477382707568e-07, | |
| "logits/chosen": 1.15625, | |
| "logits/rejected": 1.140625, | |
| "logps/chosen": -8.5625, | |
| "logps/rejected": -15.4375, | |
| "loss": 1.5555, | |
| "nll_loss": 0.74609375, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 1.9140625, | |
| "rewards/margins": 0.83984375, | |
| "rewards/rejected": 1.0703125, | |
| "step": 61 | |
| }, | |
| { | |
| "epoch": 0.248, | |
| "grad_norm": 10.425380312118477, | |
| "learning_rate": 8.820934587686245e-07, | |
| "logits/chosen": 1.1875, | |
| "logits/rejected": 1.21875, | |
| "logps/chosen": -7.8125, | |
| "logps/rejected": -15.3125, | |
| "loss": 1.5102, | |
| "nll_loss": 0.76171875, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 2.078125, | |
| "rewards/margins": 0.62890625, | |
| "rewards/rejected": 1.453125, | |
| "step": 62 | |
| }, | |
| { | |
| "epoch": 0.252, | |
| "grad_norm": 10.316061787295993, | |
| "learning_rate": 8.778747871771291e-07, | |
| "logits/chosen": 1.2265625, | |
| "logits/rejected": 1.265625, | |
| "logps/chosen": -6.15625, | |
| "logps/rejected": -16.75, | |
| "loss": 1.4838, | |
| "nll_loss": 0.59375, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 2.171875, | |
| "rewards/margins": 0.75, | |
| "rewards/rejected": 1.421875, | |
| "step": 63 | |
| }, | |
| { | |
| "epoch": 0.256, | |
| "grad_norm": 11.866868830656644, | |
| "learning_rate": 8.735924344455732e-07, | |
| "logits/chosen": 1.1875, | |
| "logits/rejected": 1.203125, | |
| "logps/chosen": -7.71875, | |
| "logps/rejected": -16.25, | |
| "loss": 1.3914, | |
| "nll_loss": 0.7578125, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 2.203125, | |
| "rewards/margins": 0.92578125, | |
| "rewards/rejected": 1.2734375, | |
| "step": 64 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "grad_norm": 13.159883371136859, | |
| "learning_rate": 8.692471222550885e-07, | |
| "logits/chosen": 1.203125, | |
| "logits/rejected": 1.1953125, | |
| "logps/chosen": -8.8125, | |
| "logps/rejected": -18.5, | |
| "loss": 1.5373, | |
| "nll_loss": 0.8359375, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 1.7578125, | |
| "rewards/margins": 0.94921875, | |
| "rewards/rejected": 0.80859375, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 0.264, | |
| "grad_norm": 10.735331657215488, | |
| "learning_rate": 8.64839582897015e-07, | |
| "logits/chosen": 1.234375, | |
| "logits/rejected": 1.25, | |
| "logps/chosen": -8.9375, | |
| "logps/rejected": -12.0, | |
| "loss": 1.4267, | |
| "nll_loss": 0.84375, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": 2.25, | |
| "rewards/margins": 0.30859375, | |
| "rewards/rejected": 1.9453125, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 0.268, | |
| "grad_norm": 11.19711245010709, | |
| "learning_rate": 8.603705591494916e-07, | |
| "logits/chosen": 1.25, | |
| "logits/rejected": 1.21875, | |
| "logps/chosen": -6.84375, | |
| "logps/rejected": -14.75, | |
| "loss": 1.357, | |
| "nll_loss": 0.72265625, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 1.6953125, | |
| "rewards/margins": 0.58984375, | |
| "rewards/rejected": 1.1015625, | |
| "step": 67 | |
| }, | |
| { | |
| "epoch": 0.272, | |
| "grad_norm": 11.97550661008315, | |
| "learning_rate": 8.5584080415228e-07, | |
| "logits/chosen": 1.2265625, | |
| "logits/rejected": 1.2265625, | |
| "logps/chosen": -6.0625, | |
| "logps/rejected": -11.4375, | |
| "loss": 1.4243, | |
| "nll_loss": 0.56640625, | |
| "rewards/accuracies": 0.53125, | |
| "rewards/chosen": 1.984375, | |
| "rewards/margins": 0.1904296875, | |
| "rewards/rejected": 1.796875, | |
| "step": 68 | |
| }, | |
| { | |
| "epoch": 0.276, | |
| "grad_norm": 12.20745761010426, | |
| "learning_rate": 8.512510812798425e-07, | |
| "logits/chosen": 1.203125, | |
| "logits/rejected": 1.2109375, | |
| "logps/chosen": -7.5, | |
| "logps/rejected": -13.8125, | |
| "loss": 1.4222, | |
| "nll_loss": 0.69921875, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 2.296875, | |
| "rewards/margins": 0.65625, | |
| "rewards/rejected": 1.640625, | |
| "step": 69 | |
| }, | |
| { | |
| "epoch": 0.28, | |
| "grad_norm": 12.64765706733198, | |
| "learning_rate": 8.466021640126945e-07, | |
| "logits/chosen": 1.203125, | |
| "logits/rejected": 1.1796875, | |
| "logps/chosen": -6.34375, | |
| "logps/rejected": -15.25, | |
| "loss": 1.4294, | |
| "nll_loss": 0.58203125, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 2.109375, | |
| "rewards/margins": 0.76953125, | |
| "rewards/rejected": 1.3359375, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.284, | |
| "grad_norm": 12.94971569659866, | |
| "learning_rate": 8.418948358070535e-07, | |
| "logits/chosen": 1.1953125, | |
| "logits/rejected": 1.171875, | |
| "logps/chosen": -8.1875, | |
| "logps/rejected": -13.8125, | |
| "loss": 1.4353, | |
| "nll_loss": 0.7734375, | |
| "rewards/accuracies": 0.5625, | |
| "rewards/chosen": 2.171875, | |
| "rewards/margins": 0.59765625, | |
| "rewards/rejected": 1.5703125, | |
| "step": 71 | |
| }, | |
| { | |
| "epoch": 0.288, | |
| "grad_norm": 12.30530684787801, | |
| "learning_rate": 8.37129889962809e-07, | |
| "logits/chosen": 1.21875, | |
| "logits/rejected": 1.2578125, | |
| "logps/chosen": -8.875, | |
| "logps/rejected": -18.875, | |
| "loss": 1.4423, | |
| "nll_loss": 0.8046875, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 2.140625, | |
| "rewards/margins": 1.078125, | |
| "rewards/rejected": 1.0625, | |
| "step": 72 | |
| }, | |
| { | |
| "epoch": 0.292, | |
| "grad_norm": 11.236244027667452, | |
| "learning_rate": 8.323081294898307e-07, | |
| "logits/chosen": 1.171875, | |
| "logits/rejected": 1.203125, | |
| "logps/chosen": -5.9375, | |
| "logps/rejected": -16.5, | |
| "loss": 1.3335, | |
| "nll_loss": 0.625, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 2.328125, | |
| "rewards/margins": 0.921875, | |
| "rewards/rejected": 1.40625, | |
| "step": 73 | |
| }, | |
| { | |
| "epoch": 0.296, | |
| "grad_norm": 10.369839744968973, | |
| "learning_rate": 8.274303669726426e-07, | |
| "logits/chosen": 1.2421875, | |
| "logits/rejected": 1.2421875, | |
| "logps/chosen": -6.5625, | |
| "logps/rejected": -15.375, | |
| "loss": 1.4023, | |
| "nll_loss": 0.58984375, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 2.078125, | |
| "rewards/margins": 0.9453125, | |
| "rewards/rejected": 1.1328125, | |
| "step": 74 | |
| }, | |
| { | |
| "epoch": 0.3, | |
| "grad_norm": 10.5653636945963, | |
| "learning_rate": 8.22497424433482e-07, | |
| "logits/chosen": 1.265625, | |
| "logits/rejected": 1.3125, | |
| "logps/chosen": -6.59375, | |
| "logps/rejected": -14.5, | |
| "loss": 1.3905, | |
| "nll_loss": 0.65625, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 2.171875, | |
| "rewards/margins": 0.8203125, | |
| "rewards/rejected": 1.359375, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.304, | |
| "grad_norm": 11.160863929220326, | |
| "learning_rate": 8.175101331937692e-07, | |
| "logits/chosen": 1.203125, | |
| "logits/rejected": 1.2109375, | |
| "logps/chosen": -6.3125, | |
| "logps/rejected": -13.6875, | |
| "loss": 1.3124, | |
| "nll_loss": 0.6796875, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 2.390625, | |
| "rewards/margins": 1.0859375, | |
| "rewards/rejected": 1.296875, | |
| "step": 76 | |
| }, | |
| { | |
| "epoch": 0.308, | |
| "grad_norm": 14.15246433516739, | |
| "learning_rate": 8.124693337340092e-07, | |
| "logits/chosen": 1.2421875, | |
| "logits/rejected": 1.2265625, | |
| "logps/chosen": -7.5625, | |
| "logps/rejected": -17.375, | |
| "loss": 1.4235, | |
| "nll_loss": 0.78515625, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 2.078125, | |
| "rewards/margins": 0.86328125, | |
| "rewards/rejected": 1.21875, | |
| "step": 77 | |
| }, | |
| { | |
| "epoch": 0.312, | |
| "grad_norm": 10.407569832039918, | |
| "learning_rate": 8.073758755521504e-07, | |
| "logits/chosen": 1.171875, | |
| "logits/rejected": 1.171875, | |
| "logps/chosen": -6.5625, | |
| "logps/rejected": -17.75, | |
| "loss": 1.3378, | |
| "nll_loss": 0.6484375, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 2.125, | |
| "rewards/margins": 1.0, | |
| "rewards/rejected": 1.125, | |
| "step": 78 | |
| }, | |
| { | |
| "epoch": 0.316, | |
| "grad_norm": 9.802442992278246, | |
| "learning_rate": 8.022306170204231e-07, | |
| "logits/chosen": 1.1328125, | |
| "logits/rejected": 1.1640625, | |
| "logps/chosen": -5.84375, | |
| "logps/rejected": -18.875, | |
| "loss": 1.4524, | |
| "nll_loss": 0.5625, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 2.3125, | |
| "rewards/margins": 1.3515625, | |
| "rewards/rejected": 0.96484375, | |
| "step": 79 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "grad_norm": 11.527882089480844, | |
| "learning_rate": 7.970344252406831e-07, | |
| "logits/chosen": 1.2109375, | |
| "logits/rejected": 1.2109375, | |
| "logps/chosen": -5.78125, | |
| "logps/rejected": -15.9375, | |
| "loss": 1.4225, | |
| "nll_loss": 0.5703125, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 2.171875, | |
| "rewards/margins": 0.74609375, | |
| "rewards/rejected": 1.421875, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.324, | |
| "grad_norm": 10.591457439226241, | |
| "learning_rate": 7.917881758982836e-07, | |
| "logits/chosen": 1.2265625, | |
| "logits/rejected": 1.2421875, | |
| "logps/chosen": -7.0625, | |
| "logps/rejected": -13.0, | |
| "loss": 1.3526, | |
| "nll_loss": 0.67578125, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 1.9375, | |
| "rewards/margins": 0.494140625, | |
| "rewards/rejected": 1.4453125, | |
| "step": 81 | |
| }, | |
| { | |
| "epoch": 0.328, | |
| "grad_norm": 12.214948597017989, | |
| "learning_rate": 7.86492753114501e-07, | |
| "logits/chosen": 1.203125, | |
| "logits/rejected": 1.21875, | |
| "logps/chosen": -6.3125, | |
| "logps/rejected": -20.375, | |
| "loss": 1.279, | |
| "nll_loss": 0.60546875, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 2.03125, | |
| "rewards/margins": 1.4375, | |
| "rewards/rejected": 0.5859375, | |
| "step": 82 | |
| }, | |
| { | |
| "epoch": 0.332, | |
| "grad_norm": 9.417779464605982, | |
| "learning_rate": 7.811490492975389e-07, | |
| "logits/chosen": 1.2421875, | |
| "logits/rejected": 1.265625, | |
| "logps/chosen": -4.96875, | |
| "logps/rejected": -17.0, | |
| "loss": 1.2994, | |
| "nll_loss": 0.5234375, | |
| "rewards/accuracies": 0.5625, | |
| "rewards/chosen": 1.9609375, | |
| "rewards/margins": 0.77734375, | |
| "rewards/rejected": 1.1796875, | |
| "step": 83 | |
| }, | |
| { | |
| "epoch": 0.336, | |
| "grad_norm": 10.535703587473606, | |
| "learning_rate": 7.757579649921353e-07, | |
| "logits/chosen": 1.171875, | |
| "logits/rejected": 1.1953125, | |
| "logps/chosen": -5.125, | |
| "logps/rejected": -19.75, | |
| "loss": 1.3724, | |
| "nll_loss": 0.51171875, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 2.203125, | |
| "rewards/margins": 0.92578125, | |
| "rewards/rejected": 1.28125, | |
| "step": 84 | |
| }, | |
| { | |
| "epoch": 0.34, | |
| "grad_norm": 9.31264272586202, | |
| "learning_rate": 7.703204087277988e-07, | |
| "logits/chosen": 1.15625, | |
| "logits/rejected": 1.15625, | |
| "logps/chosen": -6.1875, | |
| "logps/rejected": -15.625, | |
| "loss": 1.2552, | |
| "nll_loss": 0.58203125, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 1.9609375, | |
| "rewards/margins": 0.7734375, | |
| "rewards/rejected": 1.1875, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 0.344, | |
| "grad_norm": 9.834771755358236, | |
| "learning_rate": 7.648372968656993e-07, | |
| "logits/chosen": 1.2109375, | |
| "logits/rejected": 1.203125, | |
| "logps/chosen": -6.71875, | |
| "logps/rejected": -20.375, | |
| "loss": 1.3302, | |
| "nll_loss": 0.62109375, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 1.8203125, | |
| "rewards/margins": 0.96484375, | |
| "rewards/rejected": 0.85546875, | |
| "step": 86 | |
| }, | |
| { | |
| "epoch": 0.348, | |
| "grad_norm": 10.165771096354893, | |
| "learning_rate": 7.593095534442386e-07, | |
| "logits/chosen": 1.1796875, | |
| "logits/rejected": 1.203125, | |
| "logps/chosen": -5.15625, | |
| "logps/rejected": -15.25, | |
| "loss": 1.2375, | |
| "nll_loss": 0.55859375, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 2.1875, | |
| "rewards/margins": 0.8984375, | |
| "rewards/rejected": 1.2890625, | |
| "step": 87 | |
| }, | |
| { | |
| "epoch": 0.352, | |
| "grad_norm": 16.22798280791584, | |
| "learning_rate": 7.537381100233278e-07, | |
| "logits/chosen": 1.2421875, | |
| "logits/rejected": 1.2421875, | |
| "logps/chosen": -4.84375, | |
| "logps/rejected": -19.25, | |
| "loss": 1.3853, | |
| "nll_loss": 0.4609375, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 2.109375, | |
| "rewards/margins": 1.2734375, | |
| "rewards/rejected": 0.83984375, | |
| "step": 88 | |
| }, | |
| { | |
| "epoch": 0.356, | |
| "grad_norm": 10.189762721333606, | |
| "learning_rate": 7.481239055273958e-07, | |
| "logits/chosen": 1.3046875, | |
| "logits/rejected": 1.296875, | |
| "logps/chosen": -5.0, | |
| "logps/rejected": -8.75, | |
| "loss": 1.3135, | |
| "nll_loss": 0.49609375, | |
| "rewards/accuracies": 0.375, | |
| "rewards/chosen": 2.109375, | |
| "rewards/margins": 0.1474609375, | |
| "rewards/rejected": 1.953125, | |
| "step": 89 | |
| }, | |
| { | |
| "epoch": 0.36, | |
| "grad_norm": 9.700804030641727, | |
| "learning_rate": 7.424678860871584e-07, | |
| "logits/chosen": 1.203125, | |
| "logits/rejected": 1.265625, | |
| "logps/chosen": -4.875, | |
| "logps/rejected": -19.5, | |
| "loss": 1.3442, | |
| "nll_loss": 0.478515625, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 2.21875, | |
| "rewards/margins": 1.6015625, | |
| "rewards/rejected": 0.6171875, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.364, | |
| "grad_norm": 9.412262822397452, | |
| "learning_rate": 7.367710048801715e-07, | |
| "logits/chosen": 1.1875, | |
| "logits/rejected": 1.1953125, | |
| "logps/chosen": -3.96875, | |
| "logps/rejected": -13.5, | |
| "loss": 1.3565, | |
| "nll_loss": 0.39453125, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 2.125, | |
| "rewards/margins": 0.65234375, | |
| "rewards/rejected": 1.46875, | |
| "step": 91 | |
| }, | |
| { | |
| "epoch": 0.368, | |
| "grad_norm": 10.893195650456414, | |
| "learning_rate": 7.31034221970198e-07, | |
| "logits/chosen": 1.2109375, | |
| "logits/rejected": 1.21875, | |
| "logps/chosen": -5.8125, | |
| "logps/rejected": -23.0, | |
| "loss": 1.3479, | |
| "nll_loss": 0.6015625, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 1.9375, | |
| "rewards/margins": 1.2109375, | |
| "rewards/rejected": 0.7265625, | |
| "step": 92 | |
| }, | |
| { | |
| "epoch": 0.372, | |
| "grad_norm": 10.215336342693359, | |
| "learning_rate": 7.252585041454133e-07, | |
| "logits/chosen": 1.1640625, | |
| "logits/rejected": 1.1875, | |
| "logps/chosen": -5.09375, | |
| "logps/rejected": -19.25, | |
| "loss": 1.3127, | |
| "nll_loss": 0.51953125, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 1.921875, | |
| "rewards/margins": 1.078125, | |
| "rewards/rejected": 0.84375, | |
| "step": 93 | |
| }, | |
| { | |
| "epoch": 0.376, | |
| "grad_norm": 10.852698170529848, | |
| "learning_rate": 7.194448247554778e-07, | |
| "logits/chosen": 1.203125, | |
| "logits/rejected": 1.21875, | |
| "logps/chosen": -6.28125, | |
| "logps/rejected": -14.6875, | |
| "loss": 1.3314, | |
| "nll_loss": 0.671875, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 2.046875, | |
| "rewards/margins": 0.65625, | |
| "rewards/rejected": 1.390625, | |
| "step": 94 | |
| }, | |
| { | |
| "epoch": 0.38, | |
| "grad_norm": 11.160157641355154, | |
| "learning_rate": 7.135941635475036e-07, | |
| "logits/chosen": 1.203125, | |
| "logits/rejected": 1.2265625, | |
| "logps/chosen": -7.125, | |
| "logps/rejected": -22.25, | |
| "loss": 1.3336, | |
| "nll_loss": 0.63671875, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": 2.25, | |
| "rewards/margins": 1.40625, | |
| "rewards/rejected": 0.8359375, | |
| "step": 95 | |
| }, | |
| { | |
| "epoch": 0.384, | |
| "grad_norm": 11.83268475024706, | |
| "learning_rate": 7.077075065009433e-07, | |
| "logits/chosen": 1.2109375, | |
| "logits/rejected": 1.21875, | |
| "logps/chosen": -5.96875, | |
| "logps/rejected": -15.8125, | |
| "loss": 1.3246, | |
| "nll_loss": 0.58984375, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 1.9375, | |
| "rewards/margins": 0.62890625, | |
| "rewards/rejected": 1.3046875, | |
| "step": 96 | |
| }, | |
| { | |
| "epoch": 0.388, | |
| "grad_norm": 10.1751967830661, | |
| "learning_rate": 7.017858456614283e-07, | |
| "logits/chosen": 1.2578125, | |
| "logits/rejected": 1.234375, | |
| "logps/chosen": -4.96875, | |
| "logps/rejected": -19.5, | |
| "loss": 1.2867, | |
| "nll_loss": 0.490234375, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": 2.296875, | |
| "rewards/margins": 1.3203125, | |
| "rewards/rejected": 0.96875, | |
| "step": 97 | |
| }, | |
| { | |
| "epoch": 0.392, | |
| "grad_norm": 10.601998971866884, | |
| "learning_rate": 6.958301789735852e-07, | |
| "logits/chosen": 1.15625, | |
| "logits/rejected": 1.171875, | |
| "logps/chosen": -4.78125, | |
| "logps/rejected": -17.0, | |
| "loss": 1.3506, | |
| "nll_loss": 0.484375, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 1.6796875, | |
| "rewards/margins": 0.76953125, | |
| "rewards/rejected": 0.90625, | |
| "step": 98 | |
| }, | |
| { | |
| "epoch": 0.396, | |
| "grad_norm": 9.779406708019934, | |
| "learning_rate": 6.898415101128569e-07, | |
| "logits/chosen": 1.2578125, | |
| "logits/rejected": 1.28125, | |
| "logps/chosen": -6.28125, | |
| "logps/rejected": -13.875, | |
| "loss": 1.2903, | |
| "nll_loss": 0.6171875, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": 2.203125, | |
| "rewards/margins": 0.70703125, | |
| "rewards/rejected": 1.4921875, | |
| "step": 99 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "grad_norm": 11.255228657596119, | |
| "learning_rate": 6.8382084831636e-07, | |
| "logits/chosen": 1.1796875, | |
| "logits/rejected": 1.203125, | |
| "logps/chosen": -4.4375, | |
| "logps/rejected": -25.0, | |
| "loss": 1.2663, | |
| "nll_loss": 0.408203125, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 2.09375, | |
| "rewards/margins": 1.6953125, | |
| "rewards/rejected": 0.400390625, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.404, | |
| "grad_norm": 9.493234666356846, | |
| "learning_rate": 6.777692082128024e-07, | |
| "logits/chosen": 1.234375, | |
| "logits/rejected": 1.234375, | |
| "logps/chosen": -6.625, | |
| "logps/rejected": -19.0, | |
| "loss": 1.29, | |
| "nll_loss": 0.60546875, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": 2.0625, | |
| "rewards/margins": 1.1875, | |
| "rewards/rejected": 0.87890625, | |
| "step": 101 | |
| }, | |
| { | |
| "epoch": 0.408, | |
| "grad_norm": 10.02656201009763, | |
| "learning_rate": 6.716876096514943e-07, | |
| "logits/chosen": 1.171875, | |
| "logits/rejected": 1.1796875, | |
| "logps/chosen": -3.640625, | |
| "logps/rejected": -28.625, | |
| "loss": 1.2926, | |
| "nll_loss": 0.328125, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 1.765625, | |
| "rewards/margins": 1.734375, | |
| "rewards/rejected": 0.03125, | |
| "step": 102 | |
| }, | |
| { | |
| "epoch": 0.412, | |
| "grad_norm": 12.936276111177527, | |
| "learning_rate": 6.655770775304782e-07, | |
| "logits/chosen": 1.1875, | |
| "logits/rejected": 1.203125, | |
| "logps/chosen": -6.46875, | |
| "logps/rejected": -10.6875, | |
| "loss": 1.3398, | |
| "nll_loss": 0.65234375, | |
| "rewards/accuracies": 0.53125, | |
| "rewards/chosen": 2.28125, | |
| "rewards/margins": 0.48046875, | |
| "rewards/rejected": 1.8046875, | |
| "step": 103 | |
| }, | |
| { | |
| "epoch": 0.416, | |
| "grad_norm": 11.080497578488936, | |
| "learning_rate": 6.594386416238094e-07, | |
| "logits/chosen": 1.1953125, | |
| "logits/rejected": 1.1953125, | |
| "logps/chosen": -6.9375, | |
| "logps/rejected": -14.625, | |
| "loss": 1.2289, | |
| "nll_loss": 0.65234375, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": 1.796875, | |
| "rewards/margins": 0.6875, | |
| "rewards/rejected": 1.109375, | |
| "step": 104 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "grad_norm": 9.566982136498279, | |
| "learning_rate": 6.532733364080125e-07, | |
| "logits/chosen": 1.203125, | |
| "logits/rejected": 1.21875, | |
| "logps/chosen": -5.0625, | |
| "logps/rejected": -22.25, | |
| "loss": 1.2164, | |
| "nll_loss": 0.462890625, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": 2.171875, | |
| "rewards/margins": 1.2421875, | |
| "rewards/rejected": 0.9296875, | |
| "step": 105 | |
| }, | |
| { | |
| "epoch": 0.424, | |
| "grad_norm": 9.565841106095371, | |
| "learning_rate": 6.470822008877482e-07, | |
| "logits/chosen": 1.25, | |
| "logits/rejected": 1.28125, | |
| "logps/chosen": -6.28125, | |
| "logps/rejected": -14.3125, | |
| "loss": 1.2998, | |
| "nll_loss": 0.55859375, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 2.328125, | |
| "rewards/margins": 0.62890625, | |
| "rewards/rejected": 1.6953125, | |
| "step": 106 | |
| }, | |
| { | |
| "epoch": 0.428, | |
| "grad_norm": 10.706543148334308, | |
| "learning_rate": 6.408662784207149e-07, | |
| "logits/chosen": 1.1796875, | |
| "logits/rejected": 1.1796875, | |
| "logps/chosen": -5.0625, | |
| "logps/rejected": -21.75, | |
| "loss": 1.2774, | |
| "nll_loss": 0.53125, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 1.953125, | |
| "rewards/margins": 0.95703125, | |
| "rewards/rejected": 1.0, | |
| "step": 107 | |
| }, | |
| { | |
| "epoch": 0.432, | |
| "grad_norm": 10.606219490157203, | |
| "learning_rate": 6.346266165418172e-07, | |
| "logits/chosen": 1.2109375, | |
| "logits/rejected": 1.234375, | |
| "logps/chosen": -4.96875, | |
| "logps/rejected": -15.5625, | |
| "loss": 1.331, | |
| "nll_loss": 0.498046875, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 2.625, | |
| "rewards/margins": 0.9765625, | |
| "rewards/rejected": 1.6484375, | |
| "step": 108 | |
| }, | |
| { | |
| "epoch": 0.436, | |
| "grad_norm": 14.097117309386473, | |
| "learning_rate": 6.283642667866315e-07, | |
| "logits/chosen": 1.234375, | |
| "logits/rejected": 1.234375, | |
| "logps/chosen": -5.3125, | |
| "logps/rejected": -18.5, | |
| "loss": 1.2949, | |
| "nll_loss": 0.5078125, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 1.859375, | |
| "rewards/margins": 0.85546875, | |
| "rewards/rejected": 1.0, | |
| "step": 109 | |
| }, | |
| { | |
| "epoch": 0.44, | |
| "grad_norm": 9.292312467170014, | |
| "learning_rate": 6.220802845141957e-07, | |
| "logits/chosen": 1.203125, | |
| "logits/rejected": 1.2421875, | |
| "logps/chosen": -6.34375, | |
| "logps/rejected": -30.0, | |
| "loss": 1.2458, | |
| "nll_loss": 0.578125, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 2.203125, | |
| "rewards/margins": 2.0625, | |
| "rewards/rejected": 0.146484375, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.444, | |
| "grad_norm": 11.721210491862369, | |
| "learning_rate": 6.157757287291556e-07, | |
| "logits/chosen": 1.21875, | |
| "logits/rejected": 1.1796875, | |
| "logps/chosen": -3.96875, | |
| "logps/rejected": -25.125, | |
| "loss": 1.199, | |
| "nll_loss": 0.40625, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": 2.015625, | |
| "rewards/margins": 1.921875, | |
| "rewards/rejected": 0.10302734375, | |
| "step": 111 | |
| }, | |
| { | |
| "epoch": 0.448, | |
| "grad_norm": 14.194347503608423, | |
| "learning_rate": 6.094516619032975e-07, | |
| "logits/chosen": 1.1875, | |
| "logits/rejected": 1.1953125, | |
| "logps/chosen": -6.03125, | |
| "logps/rejected": -20.625, | |
| "loss": 1.2678, | |
| "nll_loss": 0.63671875, | |
| "rewards/accuracies": 0.53125, | |
| "rewards/chosen": 2.015625, | |
| "rewards/margins": 1.046875, | |
| "rewards/rejected": 0.96875, | |
| "step": 112 | |
| }, | |
| { | |
| "epoch": 0.452, | |
| "grad_norm": 11.663474005311688, | |
| "learning_rate": 6.03109149796494e-07, | |
| "logits/chosen": 1.1953125, | |
| "logits/rejected": 1.203125, | |
| "logps/chosen": -4.09375, | |
| "logps/rejected": -18.625, | |
| "loss": 1.2308, | |
| "nll_loss": 0.400390625, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 2.21875, | |
| "rewards/margins": 1.015625, | |
| "rewards/rejected": 1.1953125, | |
| "step": 113 | |
| }, | |
| { | |
| "epoch": 0.456, | |
| "grad_norm": 11.73787149228671, | |
| "learning_rate": 5.967492612770997e-07, | |
| "logits/chosen": 1.21875, | |
| "logits/rejected": 1.2578125, | |
| "logps/chosen": -7.0625, | |
| "logps/rejected": -22.875, | |
| "loss": 1.2719, | |
| "nll_loss": 0.6875, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 2.109375, | |
| "rewards/margins": 1.4921875, | |
| "rewards/rejected": 0.61328125, | |
| "step": 114 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "grad_norm": 18.765562279107698, | |
| "learning_rate": 5.90373068141819e-07, | |
| "logits/chosen": 1.2421875, | |
| "logits/rejected": 1.234375, | |
| "logps/chosen": -6.65625, | |
| "logps/rejected": -23.875, | |
| "loss": 1.2868, | |
| "nll_loss": 0.64453125, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": 2.0625, | |
| "rewards/margins": 1.4921875, | |
| "rewards/rejected": 0.56640625, | |
| "step": 115 | |
| }, | |
| { | |
| "epoch": 0.464, | |
| "grad_norm": 11.835101741684044, | |
| "learning_rate": 5.839816449350824e-07, | |
| "logits/chosen": 1.265625, | |
| "logits/rejected": 1.265625, | |
| "logps/chosen": -6.3125, | |
| "logps/rejected": -18.625, | |
| "loss": 1.2367, | |
| "nll_loss": 0.625, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 2.40625, | |
| "rewards/margins": 0.99609375, | |
| "rewards/rejected": 1.4140625, | |
| "step": 116 | |
| }, | |
| { | |
| "epoch": 0.468, | |
| "grad_norm": 9.913706129948022, | |
| "learning_rate": 5.775760687679603e-07, | |
| "logits/chosen": 1.2265625, | |
| "logits/rejected": 1.25, | |
| "logps/chosen": -5.75, | |
| "logps/rejected": -16.375, | |
| "loss": 1.1554, | |
| "nll_loss": 0.55859375, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": 2.109375, | |
| "rewards/margins": 1.234375, | |
| "rewards/rejected": 0.875, | |
| "step": 117 | |
| }, | |
| { | |
| "epoch": 0.472, | |
| "grad_norm": 10.913087492915105, | |
| "learning_rate": 5.711574191366427e-07, | |
| "logits/chosen": 1.1875, | |
| "logits/rejected": 1.21875, | |
| "logps/chosen": -3.25, | |
| "logps/rejected": -20.25, | |
| "loss": 1.3168, | |
| "nll_loss": 0.333984375, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 2.15625, | |
| "rewards/margins": 1.140625, | |
| "rewards/rejected": 1.0234375, | |
| "step": 118 | |
| }, | |
| { | |
| "epoch": 0.476, | |
| "grad_norm": 10.973356270302578, | |
| "learning_rate": 5.647267777405177e-07, | |
| "logits/chosen": 1.203125, | |
| "logits/rejected": 1.21875, | |
| "logps/chosen": -5.875, | |
| "logps/rejected": -13.4375, | |
| "loss": 1.2444, | |
| "nll_loss": 0.609375, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 1.96875, | |
| "rewards/margins": 0.57421875, | |
| "rewards/rejected": 1.3984375, | |
| "step": 119 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "grad_norm": 10.36977304647522, | |
| "learning_rate": 5.582852282998795e-07, | |
| "logits/chosen": 1.203125, | |
| "logits/rejected": 1.1875, | |
| "logps/chosen": -6.1875, | |
| "logps/rejected": -18.625, | |
| "loss": 1.2106, | |
| "nll_loss": 0.54296875, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 2.34375, | |
| "rewards/margins": 0.96484375, | |
| "rewards/rejected": 1.375, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.484, | |
| "grad_norm": 10.479478969202331, | |
| "learning_rate": 5.518338563732944e-07, | |
| "logits/chosen": 1.15625, | |
| "logits/rejected": 1.1640625, | |
| "logps/chosen": -5.21875, | |
| "logps/rejected": -18.125, | |
| "loss": 1.2196, | |
| "nll_loss": 0.55859375, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 1.90625, | |
| "rewards/margins": 0.875, | |
| "rewards/rejected": 1.03125, | |
| "step": 121 | |
| }, | |
| { | |
| "epoch": 0.488, | |
| "grad_norm": 12.292060691577667, | |
| "learning_rate": 5.453737491746572e-07, | |
| "logits/chosen": 1.1171875, | |
| "logits/rejected": 1.140625, | |
| "logps/chosen": -4.46875, | |
| "logps/rejected": -15.8125, | |
| "loss": 1.2847, | |
| "nll_loss": 0.44921875, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 2.4375, | |
| "rewards/margins": 1.1328125, | |
| "rewards/rejected": 1.3125, | |
| "step": 122 | |
| }, | |
| { | |
| "epoch": 0.492, | |
| "grad_norm": 13.647850653915034, | |
| "learning_rate": 5.389059953899698e-07, | |
| "logits/chosen": 1.109375, | |
| "logits/rejected": 1.1328125, | |
| "logps/chosen": -4.75, | |
| "logps/rejected": -23.0, | |
| "loss": 1.3749, | |
| "nll_loss": 0.47265625, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 2.421875, | |
| "rewards/margins": 1.4453125, | |
| "rewards/rejected": 0.96875, | |
| "step": 123 | |
| }, | |
| { | |
| "epoch": 0.496, | |
| "grad_norm": 13.449091878369773, | |
| "learning_rate": 5.324316849938714e-07, | |
| "logits/chosen": 1.1875, | |
| "logits/rejected": 1.21875, | |
| "logps/chosen": -6.28125, | |
| "logps/rejected": -15.8125, | |
| "loss": 1.262, | |
| "nll_loss": 0.56640625, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": 2.140625, | |
| "rewards/margins": 1.125, | |
| "rewards/rejected": 1.015625, | |
| "step": 124 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "grad_norm": 12.727260122548977, | |
| "learning_rate": 5.259519090659498e-07, | |
| "logits/chosen": 1.1328125, | |
| "logits/rejected": 1.1015625, | |
| "logps/chosen": -4.4375, | |
| "logps/rejected": -17.5, | |
| "loss": 1.2112, | |
| "nll_loss": 0.3984375, | |
| "rewards/accuracies": 0.5625, | |
| "rewards/chosen": 2.4375, | |
| "rewards/margins": 0.8359375, | |
| "rewards/rejected": 1.609375, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 0.504, | |
| "grad_norm": 12.904271672786484, | |
| "learning_rate": 5.194677596068688e-07, | |
| "logits/chosen": 1.203125, | |
| "logits/rejected": 1.1875, | |
| "logps/chosen": -6.46875, | |
| "logps/rejected": -14.9375, | |
| "loss": 1.3043, | |
| "nll_loss": 0.6171875, | |
| "rewards/accuracies": 0.5625, | |
| "rewards/chosen": 2.078125, | |
| "rewards/margins": 0.203125, | |
| "rewards/rejected": 1.875, | |
| "step": 126 | |
| }, | |
| { | |
| "epoch": 0.508, | |
| "grad_norm": 11.664774077427383, | |
| "learning_rate": 5.129803293543391e-07, | |
| "logits/chosen": 1.078125, | |
| "logits/rejected": 1.109375, | |
| "logps/chosen": -3.9375, | |
| "logps/rejected": -23.375, | |
| "loss": 1.259, | |
| "nll_loss": 0.373046875, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": 2.234375, | |
| "rewards/margins": 1.6875, | |
| "rewards/rejected": 0.5390625, | |
| "step": 127 | |
| }, | |
| { | |
| "epoch": 0.512, | |
| "grad_norm": 11.48911637332855, | |
| "learning_rate": 5.064907115989654e-07, | |
| "logits/chosen": 1.1796875, | |
| "logits/rejected": 1.171875, | |
| "logps/chosen": -6.0, | |
| "logps/rejected": -21.625, | |
| "loss": 1.2457, | |
| "nll_loss": 0.54296875, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 2.03125, | |
| "rewards/margins": 1.515625, | |
| "rewards/rejected": 0.515625, | |
| "step": 128 | |
| }, | |
| { | |
| "epoch": 0.516, | |
| "grad_norm": 11.221354732561968, | |
| "learning_rate": 5e-07, | |
| "logits/chosen": 1.1484375, | |
| "logits/rejected": 1.1328125, | |
| "logps/chosen": -5.21875, | |
| "logps/rejected": -25.375, | |
| "loss": 1.2597, | |
| "nll_loss": 0.5078125, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": 1.96875, | |
| "rewards/margins": 1.546875, | |
| "rewards/rejected": 0.419921875, | |
| "step": 129 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "grad_norm": 10.330065282564254, | |
| "learning_rate": 4.935092884010346e-07, | |
| "logits/chosen": 1.1484375, | |
| "logits/rejected": 1.125, | |
| "logps/chosen": -4.875, | |
| "logps/rejected": -18.125, | |
| "loss": 1.2075, | |
| "nll_loss": 0.48046875, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 2.171875, | |
| "rewards/margins": 0.84375, | |
| "rewards/rejected": 1.3359375, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.524, | |
| "grad_norm": 10.96765827131898, | |
| "learning_rate": 4.870196706456609e-07, | |
| "logits/chosen": 1.21875, | |
| "logits/rejected": 1.203125, | |
| "logps/chosen": -5.0, | |
| "logps/rejected": -21.5, | |
| "loss": 1.271, | |
| "nll_loss": 0.484375, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 2.015625, | |
| "rewards/margins": 1.046875, | |
| "rewards/rejected": 0.97265625, | |
| "step": 131 | |
| }, | |
| { | |
| "epoch": 0.528, | |
| "grad_norm": 11.306955835707726, | |
| "learning_rate": 4.805322403931311e-07, | |
| "logits/chosen": 1.078125, | |
| "logits/rejected": 1.125, | |
| "logps/chosen": -4.53125, | |
| "logps/rejected": -21.5, | |
| "loss": 1.2285, | |
| "nll_loss": 0.412109375, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 2.171875, | |
| "rewards/margins": 1.703125, | |
| "rewards/rejected": 0.474609375, | |
| "step": 132 | |
| }, | |
| { | |
| "epoch": 0.532, | |
| "grad_norm": 11.491623837247948, | |
| "learning_rate": 4.7404809093405017e-07, | |
| "logits/chosen": 1.1796875, | |
| "logits/rejected": 1.1640625, | |
| "logps/chosen": -5.3125, | |
| "logps/rejected": -18.0, | |
| "loss": 1.2724, | |
| "nll_loss": 0.44921875, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 2.21875, | |
| "rewards/margins": 1.078125, | |
| "rewards/rejected": 1.140625, | |
| "step": 133 | |
| }, | |
| { | |
| "epoch": 0.536, | |
| "grad_norm": 13.318991974643362, | |
| "learning_rate": 4.675683150061284e-07, | |
| "logits/chosen": 1.171875, | |
| "logits/rejected": 1.1953125, | |
| "logps/chosen": -5.25, | |
| "logps/rejected": -18.625, | |
| "loss": 1.3086, | |
| "nll_loss": 0.515625, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 2.328125, | |
| "rewards/margins": 0.87890625, | |
| "rewards/rejected": 1.453125, | |
| "step": 134 | |
| }, | |
| { | |
| "epoch": 0.54, | |
| "grad_norm": 12.79845150607262, | |
| "learning_rate": 4.6109400461003003e-07, | |
| "logits/chosen": 1.234375, | |
| "logits/rejected": 1.203125, | |
| "logps/chosen": -6.0, | |
| "logps/rejected": -16.875, | |
| "loss": 1.2625, | |
| "nll_loss": 0.515625, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 2.53125, | |
| "rewards/margins": 1.234375, | |
| "rewards/rejected": 1.2890625, | |
| "step": 135 | |
| }, | |
| { | |
| "epoch": 0.544, | |
| "grad_norm": 9.182288245306129, | |
| "learning_rate": 4.5462625082534286e-07, | |
| "logits/chosen": 1.1015625, | |
| "logits/rejected": 1.1328125, | |
| "logps/chosen": -3.890625, | |
| "logps/rejected": -17.5, | |
| "loss": 1.1632, | |
| "nll_loss": 0.39453125, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 2.34375, | |
| "rewards/margins": 1.6328125, | |
| "rewards/rejected": 0.71484375, | |
| "step": 136 | |
| }, | |
| { | |
| "epoch": 0.548, | |
| "grad_norm": 13.192726716701998, | |
| "learning_rate": 4.481661436267057e-07, | |
| "logits/chosen": 1.1953125, | |
| "logits/rejected": 1.1875, | |
| "logps/chosen": -5.46875, | |
| "logps/rejected": -20.75, | |
| "loss": 1.3656, | |
| "nll_loss": 0.498046875, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": 2.40625, | |
| "rewards/margins": 1.65625, | |
| "rewards/rejected": 0.75390625, | |
| "step": 137 | |
| }, | |
| { | |
| "epoch": 0.552, | |
| "grad_norm": 10.563546313174774, | |
| "learning_rate": 4.4171477170012043e-07, | |
| "logits/chosen": 1.2265625, | |
| "logits/rejected": 1.2421875, | |
| "logps/chosen": -4.25, | |
| "logps/rejected": -22.875, | |
| "loss": 1.2065, | |
| "nll_loss": 0.390625, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 2.484375, | |
| "rewards/margins": 1.53125, | |
| "rewards/rejected": 0.9453125, | |
| "step": 138 | |
| }, | |
| { | |
| "epoch": 0.556, | |
| "grad_norm": 10.127912435160546, | |
| "learning_rate": 4.352732222594823e-07, | |
| "logits/chosen": 1.2265625, | |
| "logits/rejected": 1.25, | |
| "logps/chosen": -5.59375, | |
| "logps/rejected": -17.625, | |
| "loss": 1.2103, | |
| "nll_loss": 0.5234375, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 2.3125, | |
| "rewards/margins": 1.1640625, | |
| "rewards/rejected": 1.140625, | |
| "step": 139 | |
| }, | |
| { | |
| "epoch": 0.56, | |
| "grad_norm": 10.137596180429622, | |
| "learning_rate": 4.2884258086335745e-07, | |
| "logits/chosen": 1.1875, | |
| "logits/rejected": 1.1875, | |
| "logps/chosen": -4.9375, | |
| "logps/rejected": -14.125, | |
| "loss": 1.3043, | |
| "nll_loss": 0.4375, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 2.421875, | |
| "rewards/margins": 0.94921875, | |
| "rewards/rejected": 1.4765625, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.564, | |
| "grad_norm": 10.569099883454175, | |
| "learning_rate": 4.224239312320398e-07, | |
| "logits/chosen": 1.234375, | |
| "logits/rejected": 1.2734375, | |
| "logps/chosen": -6.90625, | |
| "logps/rejected": -23.125, | |
| "loss": 1.3308, | |
| "nll_loss": 0.58984375, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 1.7109375, | |
| "rewards/margins": 1.1875, | |
| "rewards/rejected": 0.5234375, | |
| "step": 141 | |
| }, | |
| { | |
| "epoch": 0.568, | |
| "grad_norm": 10.854956318015526, | |
| "learning_rate": 4.160183550649175e-07, | |
| "logits/chosen": 1.1875, | |
| "logits/rejected": 1.1875, | |
| "logps/chosen": -4.46875, | |
| "logps/rejected": -19.25, | |
| "loss": 1.2546, | |
| "nll_loss": 0.455078125, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 2.40625, | |
| "rewards/margins": 1.3671875, | |
| "rewards/rejected": 1.03125, | |
| "step": 142 | |
| }, | |
| { | |
| "epoch": 0.572, | |
| "grad_norm": 10.941246000391953, | |
| "learning_rate": 4.09626931858181e-07, | |
| "logits/chosen": 1.265625, | |
| "logits/rejected": 1.265625, | |
| "logps/chosen": -4.40625, | |
| "logps/rejected": -13.25, | |
| "loss": 1.2541, | |
| "nll_loss": 0.435546875, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 2.203125, | |
| "rewards/margins": 0.5390625, | |
| "rewards/rejected": 1.671875, | |
| "step": 143 | |
| }, | |
| { | |
| "epoch": 0.576, | |
| "grad_norm": 11.708985201825207, | |
| "learning_rate": 4.0325073872290015e-07, | |
| "logits/chosen": 1.21875, | |
| "logits/rejected": 1.2421875, | |
| "logps/chosen": -6.375, | |
| "logps/rejected": -15.75, | |
| "loss": 1.2988, | |
| "nll_loss": 0.57421875, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 2.21875, | |
| "rewards/margins": 1.03125, | |
| "rewards/rejected": 1.1953125, | |
| "step": 144 | |
| }, | |
| { | |
| "epoch": 0.58, | |
| "grad_norm": 17.168400775199007, | |
| "learning_rate": 3.96890850203506e-07, | |
| "logits/chosen": 1.25, | |
| "logits/rejected": 1.265625, | |
| "logps/chosen": -6.0, | |
| "logps/rejected": -18.375, | |
| "loss": 1.1609, | |
| "nll_loss": 0.5703125, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": 2.625, | |
| "rewards/margins": 1.4140625, | |
| "rewards/rejected": 1.2109375, | |
| "step": 145 | |
| }, | |
| { | |
| "epoch": 0.584, | |
| "grad_norm": 12.001725465676962, | |
| "learning_rate": 3.9054833809670265e-07, | |
| "logits/chosen": 1.2109375, | |
| "logits/rejected": 1.2265625, | |
| "logps/chosen": -5.84375, | |
| "logps/rejected": -20.5, | |
| "loss": 1.211, | |
| "nll_loss": 0.5234375, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 2.125, | |
| "rewards/margins": 1.328125, | |
| "rewards/rejected": 0.796875, | |
| "step": 146 | |
| }, | |
| { | |
| "epoch": 0.588, | |
| "grad_norm": 11.340434870524739, | |
| "learning_rate": 3.8422427127084436e-07, | |
| "logits/chosen": 1.1875, | |
| "logits/rejected": 1.203125, | |
| "logps/chosen": -7.0, | |
| "logps/rejected": -21.75, | |
| "loss": 1.2451, | |
| "nll_loss": 0.6328125, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 2.0625, | |
| "rewards/margins": 1.1328125, | |
| "rewards/rejected": 0.9296875, | |
| "step": 147 | |
| }, | |
| { | |
| "epoch": 0.592, | |
| "grad_norm": 9.888995312706117, | |
| "learning_rate": 3.779197154858044e-07, | |
| "logits/chosen": 1.1953125, | |
| "logits/rejected": 1.1953125, | |
| "logps/chosen": -5.21875, | |
| "logps/rejected": -17.25, | |
| "loss": 1.1759, | |
| "nll_loss": 0.5078125, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": 2.53125, | |
| "rewards/margins": 1.140625, | |
| "rewards/rejected": 1.390625, | |
| "step": 148 | |
| }, | |
| { | |
| "epoch": 0.596, | |
| "grad_norm": 11.055589823055614, | |
| "learning_rate": 3.716357332133686e-07, | |
| "logits/chosen": 1.125, | |
| "logits/rejected": 1.1640625, | |
| "logps/chosen": -5.5625, | |
| "logps/rejected": -21.875, | |
| "loss": 1.2459, | |
| "nll_loss": 0.52734375, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 2.109375, | |
| "rewards/margins": 1.359375, | |
| "rewards/rejected": 0.75, | |
| "step": 149 | |
| }, | |
| { | |
| "epoch": 0.6, | |
| "grad_norm": 10.860785033516322, | |
| "learning_rate": 3.653733834581827e-07, | |
| "logits/chosen": 1.125, | |
| "logits/rejected": 1.09375, | |
| "logps/chosen": -4.40625, | |
| "logps/rejected": -13.625, | |
| "loss": 1.2623, | |
| "nll_loss": 0.396484375, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 1.9140625, | |
| "rewards/margins": 0.9453125, | |
| "rewards/rejected": 0.96484375, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.604, | |
| "grad_norm": 10.973234026852442, | |
| "learning_rate": 3.591337215792851e-07, | |
| "logits/chosen": 1.125, | |
| "logits/rejected": 1.1640625, | |
| "logps/chosen": -4.65625, | |
| "logps/rejected": -18.0, | |
| "loss": 1.1942, | |
| "nll_loss": 0.46875, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 2.203125, | |
| "rewards/margins": 1.0078125, | |
| "rewards/rejected": 1.203125, | |
| "step": 151 | |
| }, | |
| { | |
| "epoch": 0.608, | |
| "grad_norm": 10.179285058727595, | |
| "learning_rate": 3.5291779911225185e-07, | |
| "logits/chosen": 1.203125, | |
| "logits/rejected": 1.203125, | |
| "logps/chosen": -4.46875, | |
| "logps/rejected": -23.5, | |
| "loss": 1.1627, | |
| "nll_loss": 0.431640625, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 2.390625, | |
| "rewards/margins": 1.6328125, | |
| "rewards/rejected": 0.7578125, | |
| "step": 152 | |
| }, | |
| { | |
| "epoch": 0.612, | |
| "grad_norm": 10.366061661226649, | |
| "learning_rate": 3.4672666359198757e-07, | |
| "logits/chosen": 1.1796875, | |
| "logits/rejected": 1.1875, | |
| "logps/chosen": -3.796875, | |
| "logps/rejected": -17.75, | |
| "loss": 1.2135, | |
| "nll_loss": 0.412109375, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 2.828125, | |
| "rewards/margins": 1.5078125, | |
| "rewards/rejected": 1.3203125, | |
| "step": 153 | |
| }, | |
| { | |
| "epoch": 0.616, | |
| "grad_norm": 11.300580663859082, | |
| "learning_rate": 3.405613583761907e-07, | |
| "logits/chosen": 1.1875, | |
| "logits/rejected": 1.1953125, | |
| "logps/chosen": -6.4375, | |
| "logps/rejected": -13.9375, | |
| "loss": 1.2437, | |
| "nll_loss": 0.6796875, | |
| "rewards/accuracies": 0.5625, | |
| "rewards/chosen": 2.203125, | |
| "rewards/margins": 0.61328125, | |
| "rewards/rejected": 1.59375, | |
| "step": 154 | |
| }, | |
| { | |
| "epoch": 0.62, | |
| "grad_norm": 11.765886781112657, | |
| "learning_rate": 3.3442292246952186e-07, | |
| "logits/chosen": 1.1875, | |
| "logits/rejected": 1.1796875, | |
| "logps/chosen": -6.4375, | |
| "logps/rejected": -25.75, | |
| "loss": 1.2218, | |
| "nll_loss": 0.61328125, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 1.734375, | |
| "rewards/margins": 1.484375, | |
| "rewards/rejected": 0.248046875, | |
| "step": 155 | |
| }, | |
| { | |
| "epoch": 0.624, | |
| "grad_norm": 12.965990069205224, | |
| "learning_rate": 3.283123903485059e-07, | |
| "logits/chosen": 1.2109375, | |
| "logits/rejected": 1.21875, | |
| "logps/chosen": -5.84375, | |
| "logps/rejected": -21.0, | |
| "loss": 1.2737, | |
| "nll_loss": 0.5703125, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 2.34375, | |
| "rewards/margins": 1.3125, | |
| "rewards/rejected": 1.0234375, | |
| "step": 156 | |
| }, | |
| { | |
| "epoch": 0.628, | |
| "grad_norm": 21.892665126768843, | |
| "learning_rate": 3.222307917871977e-07, | |
| "logits/chosen": 1.1640625, | |
| "logits/rejected": 1.21875, | |
| "logps/chosen": -5.28125, | |
| "logps/rejected": -24.75, | |
| "loss": 1.2363, | |
| "nll_loss": 0.4765625, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": 2.0, | |
| "rewards/margins": 2.046875, | |
| "rewards/rejected": -0.05517578125, | |
| "step": 157 | |
| }, | |
| { | |
| "epoch": 0.632, | |
| "grad_norm": 11.166231452253136, | |
| "learning_rate": 3.161791516836399e-07, | |
| "logits/chosen": 1.109375, | |
| "logits/rejected": 1.1328125, | |
| "logps/chosen": -6.84375, | |
| "logps/rejected": -25.25, | |
| "loss": 1.2571, | |
| "nll_loss": 0.60546875, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 2.171875, | |
| "rewards/margins": 1.453125, | |
| "rewards/rejected": 0.7109375, | |
| "step": 158 | |
| }, | |
| { | |
| "epoch": 0.636, | |
| "grad_norm": 12.283746566871063, | |
| "learning_rate": 3.1015848988714307e-07, | |
| "logits/chosen": 1.203125, | |
| "logits/rejected": 1.21875, | |
| "logps/chosen": -6.90625, | |
| "logps/rejected": -23.5, | |
| "loss": 1.2295, | |
| "nll_loss": 0.64453125, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 2.484375, | |
| "rewards/margins": 1.78125, | |
| "rewards/rejected": 0.70703125, | |
| "step": 159 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "grad_norm": 11.997709728929365, | |
| "learning_rate": 3.0416982102641487e-07, | |
| "logits/chosen": 1.1796875, | |
| "logits/rejected": 1.1953125, | |
| "logps/chosen": -5.78125, | |
| "logps/rejected": -19.5, | |
| "loss": 1.2439, | |
| "nll_loss": 0.578125, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": 2.40625, | |
| "rewards/margins": 1.515625, | |
| "rewards/rejected": 0.8828125, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.644, | |
| "grad_norm": 12.312941284655942, | |
| "learning_rate": 2.982141543385717e-07, | |
| "logits/chosen": 1.1953125, | |
| "logits/rejected": 1.203125, | |
| "logps/chosen": -4.5625, | |
| "logps/rejected": -17.0, | |
| "loss": 1.2541, | |
| "nll_loss": 0.44921875, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": 2.203125, | |
| "rewards/margins": 0.85546875, | |
| "rewards/rejected": 1.34375, | |
| "step": 161 | |
| }, | |
| { | |
| "epoch": 0.648, | |
| "grad_norm": 10.583664987026914, | |
| "learning_rate": 2.922924934990568e-07, | |
| "logits/chosen": 1.1875, | |
| "logits/rejected": 1.2109375, | |
| "logps/chosen": -4.5, | |
| "logps/rejected": -19.625, | |
| "loss": 1.1992, | |
| "nll_loss": 0.462890625, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": 2.046875, | |
| "rewards/margins": 1.390625, | |
| "rewards/rejected": 0.65625, | |
| "step": 162 | |
| }, | |
| { | |
| "epoch": 0.652, | |
| "grad_norm": 12.369672652872328, | |
| "learning_rate": 2.864058364524964e-07, | |
| "logits/chosen": 1.2578125, | |
| "logits/rejected": 1.25, | |
| "logps/chosen": -4.59375, | |
| "logps/rejected": -17.125, | |
| "loss": 1.2223, | |
| "nll_loss": 0.478515625, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": 2.296875, | |
| "rewards/margins": 1.484375, | |
| "rewards/rejected": 0.8125, | |
| "step": 163 | |
| }, | |
| { | |
| "epoch": 0.656, | |
| "grad_norm": 10.926472150443256, | |
| "learning_rate": 2.8055517524452215e-07, | |
| "logits/chosen": 1.1875, | |
| "logits/rejected": 1.203125, | |
| "logps/chosen": -5.28125, | |
| "logps/rejected": -19.0, | |
| "loss": 1.1861, | |
| "nll_loss": 0.48046875, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": 2.140625, | |
| "rewards/margins": 1.21875, | |
| "rewards/rejected": 0.91796875, | |
| "step": 164 | |
| }, | |
| { | |
| "epoch": 0.66, | |
| "grad_norm": 19.45457445267617, | |
| "learning_rate": 2.7474149585458664e-07, | |
| "logits/chosen": 1.109375, | |
| "logits/rejected": 1.109375, | |
| "logps/chosen": -3.953125, | |
| "logps/rejected": -26.625, | |
| "loss": 1.2098, | |
| "nll_loss": 0.35546875, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 1.890625, | |
| "rewards/margins": 1.6484375, | |
| "rewards/rejected": 0.236328125, | |
| "step": 165 | |
| }, | |
| { | |
| "epoch": 0.664, | |
| "grad_norm": 13.019845340147054, | |
| "learning_rate": 2.6896577802980184e-07, | |
| "logits/chosen": 1.21875, | |
| "logits/rejected": 1.21875, | |
| "logps/chosen": -3.125, | |
| "logps/rejected": -24.0, | |
| "loss": 1.2028, | |
| "nll_loss": 0.291015625, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 2.0625, | |
| "rewards/margins": 1.6953125, | |
| "rewards/rejected": 0.375, | |
| "step": 166 | |
| }, | |
| { | |
| "epoch": 0.668, | |
| "grad_norm": 11.31879875355034, | |
| "learning_rate": 2.632289951198285e-07, | |
| "logits/chosen": 1.171875, | |
| "logits/rejected": 1.171875, | |
| "logps/chosen": -5.28125, | |
| "logps/rejected": -21.875, | |
| "loss": 1.2914, | |
| "nll_loss": 0.482421875, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 2.4375, | |
| "rewards/margins": 1.3828125, | |
| "rewards/rejected": 1.0546875, | |
| "step": 167 | |
| }, | |
| { | |
| "epoch": 0.672, | |
| "grad_norm": 14.312221004137449, | |
| "learning_rate": 2.575321139128417e-07, | |
| "logits/chosen": 1.2265625, | |
| "logits/rejected": 1.2734375, | |
| "logps/chosen": -5.1875, | |
| "logps/rejected": -20.375, | |
| "loss": 1.2442, | |
| "nll_loss": 0.484375, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 2.4375, | |
| "rewards/margins": 1.3203125, | |
| "rewards/rejected": 1.109375, | |
| "step": 168 | |
| }, | |
| { | |
| "epoch": 0.676, | |
| "grad_norm": 11.88685392821398, | |
| "learning_rate": 2.5187609447260417e-07, | |
| "logits/chosen": 1.1328125, | |
| "logits/rejected": 1.09375, | |
| "logps/chosen": -4.4375, | |
| "logps/rejected": -19.875, | |
| "loss": 1.1703, | |
| "nll_loss": 0.453125, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": 2.578125, | |
| "rewards/margins": 1.8828125, | |
| "rewards/rejected": 0.703125, | |
| "step": 169 | |
| }, | |
| { | |
| "epoch": 0.68, | |
| "grad_norm": 17.314779451265547, | |
| "learning_rate": 2.462618899766722e-07, | |
| "logits/chosen": 1.1328125, | |
| "logits/rejected": 1.1328125, | |
| "logps/chosen": -6.1875, | |
| "logps/rejected": -23.625, | |
| "loss": 1.2089, | |
| "nll_loss": 0.6015625, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 1.921875, | |
| "rewards/margins": 1.453125, | |
| "rewards/rejected": 0.470703125, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.684, | |
| "grad_norm": 11.85551915443719, | |
| "learning_rate": 2.406904465557614e-07, | |
| "logits/chosen": 1.1640625, | |
| "logits/rejected": 1.1328125, | |
| "logps/chosen": -5.15625, | |
| "logps/rejected": -23.0, | |
| "loss": 1.1893, | |
| "nll_loss": 0.5078125, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 2.171875, | |
| "rewards/margins": 1.28125, | |
| "rewards/rejected": 0.8828125, | |
| "step": 171 | |
| }, | |
| { | |
| "epoch": 0.688, | |
| "grad_norm": 12.228210769141855, | |
| "learning_rate": 2.3516270313430081e-07, | |
| "logits/chosen": 1.203125, | |
| "logits/rejected": 1.1875, | |
| "logps/chosen": -6.8125, | |
| "logps/rejected": -20.5, | |
| "loss": 1.2107, | |
| "nll_loss": 0.6171875, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 2.328125, | |
| "rewards/margins": 1.59375, | |
| "rewards/rejected": 0.734375, | |
| "step": 172 | |
| }, | |
| { | |
| "epoch": 0.692, | |
| "grad_norm": 11.027522143396848, | |
| "learning_rate": 2.2967959127220137e-07, | |
| "logits/chosen": 1.1328125, | |
| "logits/rejected": 1.109375, | |
| "logps/chosen": -3.84375, | |
| "logps/rejected": -17.75, | |
| "loss": 1.1306, | |
| "nll_loss": 0.365234375, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 2.375, | |
| "rewards/margins": 1.390625, | |
| "rewards/rejected": 0.9765625, | |
| "step": 173 | |
| }, | |
| { | |
| "epoch": 0.696, | |
| "grad_norm": 13.079512510371146, | |
| "learning_rate": 2.242420350078647e-07, | |
| "logits/chosen": 1.1796875, | |
| "logits/rejected": 1.15625, | |
| "logps/chosen": -5.625, | |
| "logps/rejected": -19.125, | |
| "loss": 1.2236, | |
| "nll_loss": 0.53125, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 2.453125, | |
| "rewards/margins": 1.2421875, | |
| "rewards/rejected": 1.2109375, | |
| "step": 174 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "grad_norm": 12.476234076652355, | |
| "learning_rate": 2.1885095070246113e-07, | |
| "logits/chosen": 1.1796875, | |
| "logits/rejected": 1.1484375, | |
| "logps/chosen": -5.3125, | |
| "logps/rejected": -22.125, | |
| "loss": 1.2699, | |
| "nll_loss": 0.5, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": 2.3125, | |
| "rewards/margins": 1.5703125, | |
| "rewards/rejected": 0.75, | |
| "step": 175 | |
| }, | |
| { | |
| "epoch": 0.704, | |
| "grad_norm": 13.41091780256611, | |
| "learning_rate": 2.1350724688549904e-07, | |
| "logits/chosen": 1.0625, | |
| "logits/rejected": 1.1015625, | |
| "logps/chosen": -4.40625, | |
| "logps/rejected": -22.5, | |
| "loss": 1.2699, | |
| "nll_loss": 0.421875, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 1.890625, | |
| "rewards/margins": 1.296875, | |
| "rewards/rejected": 0.59375, | |
| "step": 176 | |
| }, | |
| { | |
| "epoch": 0.708, | |
| "grad_norm": 11.558122765112593, | |
| "learning_rate": 2.0821182410171633e-07, | |
| "logits/chosen": 1.1328125, | |
| "logits/rejected": 1.1171875, | |
| "logps/chosen": -4.3125, | |
| "logps/rejected": -19.375, | |
| "loss": 1.3392, | |
| "nll_loss": 0.40625, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 2.015625, | |
| "rewards/margins": 1.1875, | |
| "rewards/rejected": 0.83203125, | |
| "step": 177 | |
| }, | |
| { | |
| "epoch": 0.712, | |
| "grad_norm": 13.675221151303951, | |
| "learning_rate": 2.0296557475931687e-07, | |
| "logits/chosen": 1.1015625, | |
| "logits/rejected": 1.1171875, | |
| "logps/chosen": -6.09375, | |
| "logps/rejected": -17.0, | |
| "loss": 1.286, | |
| "nll_loss": 0.58203125, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 2.40625, | |
| "rewards/margins": 1.015625, | |
| "rewards/rejected": 1.3984375, | |
| "step": 178 | |
| }, | |
| { | |
| "epoch": 0.716, | |
| "grad_norm": 11.77241428897747, | |
| "learning_rate": 1.9776938297957686e-07, | |
| "logits/chosen": 1.203125, | |
| "logits/rejected": 1.2734375, | |
| "logps/chosen": -3.75, | |
| "logps/rejected": -15.8125, | |
| "loss": 1.2497, | |
| "nll_loss": 0.400390625, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 2.515625, | |
| "rewards/margins": 1.125, | |
| "rewards/rejected": 1.3828125, | |
| "step": 179 | |
| }, | |
| { | |
| "epoch": 0.72, | |
| "grad_norm": 13.110435333261227, | |
| "learning_rate": 1.9262412444784958e-07, | |
| "logits/chosen": 1.2890625, | |
| "logits/rejected": 1.2578125, | |
| "logps/chosen": -6.53125, | |
| "logps/rejected": -12.3125, | |
| "loss": 1.2314, | |
| "nll_loss": 0.625, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 2.328125, | |
| "rewards/margins": 0.59375, | |
| "rewards/rejected": 1.7265625, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.724, | |
| "grad_norm": 11.650307902708178, | |
| "learning_rate": 1.8753066626599084e-07, | |
| "logits/chosen": 1.125, | |
| "logits/rejected": 1.125, | |
| "logps/chosen": -4.46875, | |
| "logps/rejected": -24.0, | |
| "loss": 1.2924, | |
| "nll_loss": 0.4609375, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 2.25, | |
| "rewards/margins": 1.4140625, | |
| "rewards/rejected": 0.8359375, | |
| "step": 181 | |
| }, | |
| { | |
| "epoch": 0.728, | |
| "grad_norm": 13.580946392431395, | |
| "learning_rate": 1.8248986680623074e-07, | |
| "logits/chosen": 1.2109375, | |
| "logits/rejected": 1.2421875, | |
| "logps/chosen": -5.375, | |
| "logps/rejected": -16.125, | |
| "loss": 1.2858, | |
| "nll_loss": 0.546875, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 2.03125, | |
| "rewards/margins": 0.703125, | |
| "rewards/rejected": 1.328125, | |
| "step": 182 | |
| }, | |
| { | |
| "epoch": 0.732, | |
| "grad_norm": 12.225985724922241, | |
| "learning_rate": 1.7750257556651805e-07, | |
| "logits/chosen": 1.15625, | |
| "logits/rejected": 1.171875, | |
| "logps/chosen": -4.5625, | |
| "logps/rejected": -22.125, | |
| "loss": 1.2099, | |
| "nll_loss": 0.443359375, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 2.046875, | |
| "rewards/margins": 1.625, | |
| "rewards/rejected": 0.416015625, | |
| "step": 183 | |
| }, | |
| { | |
| "epoch": 0.736, | |
| "grad_norm": 12.563946894349364, | |
| "learning_rate": 1.725696330273575e-07, | |
| "logits/chosen": 1.2265625, | |
| "logits/rejected": 1.234375, | |
| "logps/chosen": -4.40625, | |
| "logps/rejected": -18.875, | |
| "loss": 1.186, | |
| "nll_loss": 0.447265625, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 2.421875, | |
| "rewards/margins": 1.2578125, | |
| "rewards/rejected": 1.1640625, | |
| "step": 184 | |
| }, | |
| { | |
| "epoch": 0.74, | |
| "grad_norm": 13.255428758141953, | |
| "learning_rate": 1.6769187051016932e-07, | |
| "logits/chosen": 1.1796875, | |
| "logits/rejected": 1.1484375, | |
| "logps/chosen": -4.5, | |
| "logps/rejected": -17.875, | |
| "loss": 1.2881, | |
| "nll_loss": 0.470703125, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 2.21875, | |
| "rewards/margins": 1.109375, | |
| "rewards/rejected": 1.109375, | |
| "step": 185 | |
| }, | |
| { | |
| "epoch": 0.744, | |
| "grad_norm": 11.194260302105462, | |
| "learning_rate": 1.6287011003719102e-07, | |
| "logits/chosen": 1.2109375, | |
| "logits/rejected": 1.1796875, | |
| "logps/chosen": -5.09375, | |
| "logps/rejected": -18.25, | |
| "loss": 1.1767, | |
| "nll_loss": 0.5, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 2.640625, | |
| "rewards/margins": 1.5859375, | |
| "rewards/rejected": 1.0625, | |
| "step": 186 | |
| }, | |
| { | |
| "epoch": 0.748, | |
| "grad_norm": 12.148074503850838, | |
| "learning_rate": 1.581051641929465e-07, | |
| "logits/chosen": 1.1171875, | |
| "logits/rejected": 1.15625, | |
| "logps/chosen": -5.28125, | |
| "logps/rejected": -16.5, | |
| "loss": 1.2704, | |
| "nll_loss": 0.5, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 2.15625, | |
| "rewards/margins": 0.98046875, | |
| "rewards/rejected": 1.171875, | |
| "step": 187 | |
| }, | |
| { | |
| "epoch": 0.752, | |
| "grad_norm": 12.395801551212454, | |
| "learning_rate": 1.5339783598730565e-07, | |
| "logits/chosen": 1.1796875, | |
| "logits/rejected": 1.2109375, | |
| "logps/chosen": -4.71875, | |
| "logps/rejected": -18.75, | |
| "loss": 1.2776, | |
| "nll_loss": 0.458984375, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 2.078125, | |
| "rewards/margins": 1.296875, | |
| "rewards/rejected": 0.7734375, | |
| "step": 188 | |
| }, | |
| { | |
| "epoch": 0.756, | |
| "grad_norm": 10.67438569858004, | |
| "learning_rate": 1.4874891872015734e-07, | |
| "logits/chosen": 1.2578125, | |
| "logits/rejected": 1.203125, | |
| "logps/chosen": -6.125, | |
| "logps/rejected": -18.75, | |
| "loss": 1.2794, | |
| "nll_loss": 0.5546875, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 2.5, | |
| "rewards/margins": 1.40625, | |
| "rewards/rejected": 1.09375, | |
| "step": 189 | |
| }, | |
| { | |
| "epoch": 0.76, | |
| "grad_norm": 12.475216208621594, | |
| "learning_rate": 1.4415919584771995e-07, | |
| "logits/chosen": 1.2578125, | |
| "logits/rejected": 1.265625, | |
| "logps/chosen": -4.75, | |
| "logps/rejected": -18.625, | |
| "loss": 1.2209, | |
| "nll_loss": 0.451171875, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 2.421875, | |
| "rewards/margins": 1.4765625, | |
| "rewards/rejected": 0.94921875, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.764, | |
| "grad_norm": 10.90649391121936, | |
| "learning_rate": 1.3962944085050832e-07, | |
| "logits/chosen": 1.234375, | |
| "logits/rejected": 1.2109375, | |
| "logps/chosen": -6.90625, | |
| "logps/rejected": -17.625, | |
| "loss": 1.2429, | |
| "nll_loss": 0.625, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 2.125, | |
| "rewards/margins": 1.0859375, | |
| "rewards/rejected": 1.0390625, | |
| "step": 191 | |
| }, | |
| { | |
| "epoch": 0.768, | |
| "grad_norm": 12.353299907170381, | |
| "learning_rate": 1.3516041710298497e-07, | |
| "logits/chosen": 1.1875, | |
| "logits/rejected": 1.1953125, | |
| "logps/chosen": -3.75, | |
| "logps/rejected": -12.75, | |
| "loss": 1.137, | |
| "nll_loss": 0.3828125, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 2.5, | |
| "rewards/margins": 0.6796875, | |
| "rewards/rejected": 1.8203125, | |
| "step": 192 | |
| }, | |
| { | |
| "epoch": 0.772, | |
| "grad_norm": 11.398941153389819, | |
| "learning_rate": 1.3075287774491146e-07, | |
| "logits/chosen": 1.203125, | |
| "logits/rejected": 1.1796875, | |
| "logps/chosen": -3.5625, | |
| "logps/rejected": -20.25, | |
| "loss": 1.2329, | |
| "nll_loss": 0.369140625, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": 2.46875, | |
| "rewards/margins": 1.296875, | |
| "rewards/rejected": 1.1640625, | |
| "step": 193 | |
| }, | |
| { | |
| "epoch": 0.776, | |
| "grad_norm": 13.85380958569379, | |
| "learning_rate": 1.2640756555442683e-07, | |
| "logits/chosen": 1.203125, | |
| "logits/rejected": 1.2109375, | |
| "logps/chosen": -6.28125, | |
| "logps/rejected": -18.25, | |
| "loss": 1.2127, | |
| "nll_loss": 0.5703125, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 2.578125, | |
| "rewards/margins": 1.4140625, | |
| "rewards/rejected": 1.1640625, | |
| "step": 194 | |
| }, | |
| { | |
| "epoch": 0.78, | |
| "grad_norm": 13.267633587995695, | |
| "learning_rate": 1.2212521282287093e-07, | |
| "logits/chosen": 1.21875, | |
| "logits/rejected": 1.21875, | |
| "logps/chosen": -5.40625, | |
| "logps/rejected": -19.625, | |
| "loss": 1.2959, | |
| "nll_loss": 0.55078125, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 2.53125, | |
| "rewards/margins": 1.1640625, | |
| "rewards/rejected": 1.359375, | |
| "step": 195 | |
| }, | |
| { | |
| "epoch": 0.784, | |
| "grad_norm": 15.787904836631533, | |
| "learning_rate": 1.179065412313755e-07, | |
| "logits/chosen": 1.1796875, | |
| "logits/rejected": 1.15625, | |
| "logps/chosen": -7.03125, | |
| "logps/rejected": -15.25, | |
| "loss": 1.3018, | |
| "nll_loss": 0.69921875, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 1.921875, | |
| "rewards/margins": 0.61328125, | |
| "rewards/rejected": 1.3125, | |
| "step": 196 | |
| }, | |
| { | |
| "epoch": 0.788, | |
| "grad_norm": 11.642080128952893, | |
| "learning_rate": 1.137522617292434e-07, | |
| "logits/chosen": 1.2109375, | |
| "logits/rejected": 1.21875, | |
| "logps/chosen": -4.71875, | |
| "logps/rejected": -9.6875, | |
| "loss": 1.2841, | |
| "nll_loss": 0.45703125, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 2.546875, | |
| "rewards/margins": 0.7265625, | |
| "rewards/rejected": 1.828125, | |
| "step": 197 | |
| }, | |
| { | |
| "epoch": 0.792, | |
| "grad_norm": 19.23607159744212, | |
| "learning_rate": 1.0966307441413597e-07, | |
| "logits/chosen": 1.1171875, | |
| "logits/rejected": 1.140625, | |
| "logps/chosen": -6.0, | |
| "logps/rejected": -16.0, | |
| "loss": 1.2262, | |
| "nll_loss": 0.6015625, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 2.796875, | |
| "rewards/margins": 1.3359375, | |
| "rewards/rejected": 1.4609375, | |
| "step": 198 | |
| }, | |
| { | |
| "epoch": 0.796, | |
| "grad_norm": 14.441184784024456, | |
| "learning_rate": 1.0563966841408995e-07, | |
| "logits/chosen": 1.171875, | |
| "logits/rejected": 1.1875, | |
| "logps/chosen": -4.9375, | |
| "logps/rejected": -18.75, | |
| "loss": 1.1646, | |
| "nll_loss": 0.486328125, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": 2.5625, | |
| "rewards/margins": 1.3671875, | |
| "rewards/rejected": 1.203125, | |
| "step": 199 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "grad_norm": 12.165181869834347, | |
| "learning_rate": 1.0168272177138199e-07, | |
| "logits/chosen": 1.2109375, | |
| "logits/rejected": 1.2265625, | |
| "logps/chosen": -5.625, | |
| "logps/rejected": -21.25, | |
| "loss": 1.2264, | |
| "nll_loss": 0.53515625, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 2.328125, | |
| "rewards/margins": 1.2421875, | |
| "rewards/rejected": 1.0859375, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.804, | |
| "grad_norm": 12.286496963943826, | |
| "learning_rate": 9.779290132826224e-08, | |
| "logits/chosen": 1.1640625, | |
| "logits/rejected": 1.1875, | |
| "logps/chosen": -6.21875, | |
| "logps/rejected": -15.25, | |
| "loss": 1.1976, | |
| "nll_loss": 0.5625, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 2.453125, | |
| "rewards/margins": 0.83984375, | |
| "rewards/rejected": 1.609375, | |
| "step": 201 | |
| }, | |
| { | |
| "epoch": 0.808, | |
| "grad_norm": 12.358923809267944, | |
| "learning_rate": 9.39708626145751e-08, | |
| "logits/chosen": 1.203125, | |
| "logits/rejected": 1.203125, | |
| "logps/chosen": -6.15625, | |
| "logps/rejected": -15.9375, | |
| "loss": 1.3051, | |
| "nll_loss": 0.60546875, | |
| "rewards/accuracies": 0.53125, | |
| "rewards/chosen": 2.4375, | |
| "rewards/margins": 0.890625, | |
| "rewards/rejected": 1.5546875, | |
| "step": 202 | |
| }, | |
| { | |
| "epoch": 0.812, | |
| "grad_norm": 12.346183865053801, | |
| "learning_rate": 9.021724973728634e-08, | |
| "logits/chosen": 1.21875, | |
| "logits/rejected": 1.2421875, | |
| "logps/chosen": -5.21875, | |
| "logps/rejected": -22.375, | |
| "loss": 1.2555, | |
| "nll_loss": 0.498046875, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 2.234375, | |
| "rewards/margins": 1.6015625, | |
| "rewards/rejected": 0.63671875, | |
| "step": 203 | |
| }, | |
| { | |
| "epoch": 0.816, | |
| "grad_norm": 12.034662551259737, | |
| "learning_rate": 8.653269527193568e-08, | |
| "logits/chosen": 1.1953125, | |
| "logits/rejected": 1.1953125, | |
| "logps/chosen": -5.21875, | |
| "logps/rejected": -19.75, | |
| "loss": 1.2054, | |
| "nll_loss": 0.482421875, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 2.734375, | |
| "rewards/margins": 1.7421875, | |
| "rewards/rejected": 0.99609375, | |
| "step": 204 | |
| }, | |
| { | |
| "epoch": 0.82, | |
| "grad_norm": 11.034715459854144, | |
| "learning_rate": 8.291782015603176e-08, | |
| "logits/chosen": 1.1484375, | |
| "logits/rejected": 1.171875, | |
| "logps/chosen": -5.25, | |
| "logps/rejected": -13.3125, | |
| "loss": 1.2401, | |
| "nll_loss": 0.52734375, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": 2.25, | |
| "rewards/margins": 0.73046875, | |
| "rewards/rejected": 1.5078125, | |
| "step": 205 | |
| }, | |
| { | |
| "epoch": 0.824, | |
| "grad_norm": 34.96908950850571, | |
| "learning_rate": 7.937323358440934e-08, | |
| "logits/chosen": 1.140625, | |
| "logits/rejected": 1.1875, | |
| "logps/chosen": -3.765625, | |
| "logps/rejected": -19.0, | |
| "loss": 1.1788, | |
| "nll_loss": 0.37109375, | |
| "rewards/accuracies": 0.5625, | |
| "rewards/chosen": 2.59375, | |
| "rewards/margins": 0.99609375, | |
| "rewards/rejected": 1.5859375, | |
| "step": 206 | |
| }, | |
| { | |
| "epoch": 0.828, | |
| "grad_norm": 13.161837149601567, | |
| "learning_rate": 7.589953290656532e-08, | |
| "logits/chosen": 1.1171875, | |
| "logits/rejected": 1.1328125, | |
| "logps/chosen": -3.546875, | |
| "logps/rejected": -22.0, | |
| "loss": 1.2688, | |
| "nll_loss": 0.361328125, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 2.234375, | |
| "rewards/margins": 1.6953125, | |
| "rewards/rejected": 0.54296875, | |
| "step": 207 | |
| }, | |
| { | |
| "epoch": 0.832, | |
| "grad_norm": 11.065362037869805, | |
| "learning_rate": 7.249730352598998e-08, | |
| "logits/chosen": 1.1484375, | |
| "logits/rejected": 1.1875, | |
| "logps/chosen": -4.71875, | |
| "logps/rejected": -22.5, | |
| "loss": 1.2256, | |
| "nll_loss": 0.419921875, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 2.0625, | |
| "rewards/margins": 1.5625, | |
| "rewards/rejected": 0.50390625, | |
| "step": 208 | |
| }, | |
| { | |
| "epoch": 0.836, | |
| "grad_norm": 15.587435364904376, | |
| "learning_rate": 6.916711880151304e-08, | |
| "logits/chosen": 1.1171875, | |
| "logits/rejected": 1.109375, | |
| "logps/chosen": -7.25, | |
| "logps/rejected": -20.125, | |
| "loss": 1.3117, | |
| "nll_loss": 0.70703125, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 2.0625, | |
| "rewards/margins": 0.875, | |
| "rewards/rejected": 1.1875, | |
| "step": 209 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "grad_norm": 13.795300358361615, | |
| "learning_rate": 6.59095399506781e-08, | |
| "logits/chosen": 1.15625, | |
| "logits/rejected": 1.15625, | |
| "logps/chosen": -5.5, | |
| "logps/rejected": -18.875, | |
| "loss": 1.2623, | |
| "nll_loss": 0.52734375, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 2.390625, | |
| "rewards/margins": 1.1640625, | |
| "rewards/rejected": 1.2265625, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.844, | |
| "grad_norm": 14.092757534806013, | |
| "learning_rate": 6.2725115955164e-08, | |
| "logits/chosen": 1.171875, | |
| "logits/rejected": 1.1796875, | |
| "logps/chosen": -4.625, | |
| "logps/rejected": -25.125, | |
| "loss": 1.2361, | |
| "nll_loss": 0.443359375, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 2.25, | |
| "rewards/margins": 1.6484375, | |
| "rewards/rejected": 0.60546875, | |
| "step": 211 | |
| }, | |
| { | |
| "epoch": 0.848, | |
| "grad_norm": 10.562244070406441, | |
| "learning_rate": 5.961438346826792e-08, | |
| "logits/chosen": 1.1328125, | |
| "logits/rejected": 1.1796875, | |
| "logps/chosen": -4.34375, | |
| "logps/rejected": -19.5, | |
| "loss": 1.171, | |
| "nll_loss": 0.419921875, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 2.234375, | |
| "rewards/margins": 1.265625, | |
| "rewards/rejected": 0.96875, | |
| "step": 212 | |
| }, | |
| { | |
| "epoch": 0.852, | |
| "grad_norm": 11.070817604602208, | |
| "learning_rate": 5.6577866724466296e-08, | |
| "logits/chosen": 1.109375, | |
| "logits/rejected": 1.1328125, | |
| "logps/chosen": -6.125, | |
| "logps/rejected": -34.0, | |
| "loss": 1.1387, | |
| "nll_loss": 0.5625, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 1.96875, | |
| "rewards/margins": 2.578125, | |
| "rewards/rejected": -0.60546875, | |
| "step": 213 | |
| }, | |
| { | |
| "epoch": 0.856, | |
| "grad_norm": 15.61141696529417, | |
| "learning_rate": 5.361607745106816e-08, | |
| "logits/chosen": 1.21875, | |
| "logits/rejected": 1.2109375, | |
| "logps/chosen": -3.703125, | |
| "logps/rejected": -13.0625, | |
| "loss": 1.223, | |
| "nll_loss": 0.408203125, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 2.453125, | |
| "rewards/margins": 1.015625, | |
| "rewards/rejected": 1.4375, | |
| "step": 214 | |
| }, | |
| { | |
| "epoch": 0.86, | |
| "grad_norm": 12.525723751388632, | |
| "learning_rate": 5.072951478197723e-08, | |
| "logits/chosen": 1.1171875, | |
| "logits/rejected": 1.1328125, | |
| "logps/chosen": -4.90625, | |
| "logps/rejected": -21.5, | |
| "loss": 1.2809, | |
| "nll_loss": 0.498046875, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 2.21875, | |
| "rewards/margins": 1.3203125, | |
| "rewards/rejected": 0.89453125, | |
| "step": 215 | |
| }, | |
| { | |
| "epoch": 0.864, | |
| "grad_norm": 12.527213509383836, | |
| "learning_rate": 4.79186651735749e-08, | |
| "logits/chosen": 1.1640625, | |
| "logits/rejected": 1.171875, | |
| "logps/chosen": -4.1875, | |
| "logps/rejected": -14.4375, | |
| "loss": 1.1196, | |
| "nll_loss": 0.427734375, | |
| "rewards/accuracies": 0.53125, | |
| "rewards/chosen": 2.140625, | |
| "rewards/margins": 0.54296875, | |
| "rewards/rejected": 1.6015625, | |
| "step": 216 | |
| }, | |
| { | |
| "epoch": 0.868, | |
| "grad_norm": 11.692116725453017, | |
| "learning_rate": 4.518400232274078e-08, | |
| "logits/chosen": 1.09375, | |
| "logits/rejected": 1.1171875, | |
| "logps/chosen": -5.21875, | |
| "logps/rejected": -20.0, | |
| "loss": 1.2534, | |
| "nll_loss": 0.470703125, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 1.9375, | |
| "rewards/margins": 0.93359375, | |
| "rewards/rejected": 1.0, | |
| "step": 217 | |
| }, | |
| { | |
| "epoch": 0.872, | |
| "grad_norm": 11.469140530461623, | |
| "learning_rate": 4.252598708702343e-08, | |
| "logits/chosen": 1.203125, | |
| "logits/rejected": 1.1796875, | |
| "logps/chosen": -4.6875, | |
| "logps/rejected": -15.5625, | |
| "loss": 1.1908, | |
| "nll_loss": 0.47265625, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 2.359375, | |
| "rewards/margins": 1.296875, | |
| "rewards/rejected": 1.0625, | |
| "step": 218 | |
| }, | |
| { | |
| "epoch": 0.876, | |
| "grad_norm": 14.44996786417216, | |
| "learning_rate": 3.994506740697406e-08, | |
| "logits/chosen": 1.25, | |
| "logits/rejected": 1.2265625, | |
| "logps/chosen": -4.78125, | |
| "logps/rejected": -19.0, | |
| "loss": 1.1148, | |
| "nll_loss": 0.447265625, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 2.296875, | |
| "rewards/margins": 1.3125, | |
| "rewards/rejected": 0.98046875, | |
| "step": 219 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "grad_norm": 11.827189877497567, | |
| "learning_rate": 3.744167823065814e-08, | |
| "logits/chosen": 1.0703125, | |
| "logits/rejected": 1.1171875, | |
| "logps/chosen": -4.40625, | |
| "logps/rejected": -24.875, | |
| "loss": 1.2424, | |
| "nll_loss": 0.388671875, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 1.8125, | |
| "rewards/margins": 1.5390625, | |
| "rewards/rejected": 0.2734375, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.884, | |
| "grad_norm": 12.018027049330724, | |
| "learning_rate": 3.501624144035559e-08, | |
| "logits/chosen": 1.15625, | |
| "logits/rejected": 1.203125, | |
| "logps/chosen": -5.21875, | |
| "logps/rejected": -17.0, | |
| "loss": 1.2421, | |
| "nll_loss": 0.46484375, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 2.265625, | |
| "rewards/margins": 1.0625, | |
| "rewards/rejected": 1.2109375, | |
| "step": 221 | |
| }, | |
| { | |
| "epoch": 0.888, | |
| "grad_norm": 12.005772353362655, | |
| "learning_rate": 3.2669165781463395e-08, | |
| "logits/chosen": 1.171875, | |
| "logits/rejected": 1.1875, | |
| "logps/chosen": -5.03125, | |
| "logps/rejected": -19.875, | |
| "loss": 1.301, | |
| "nll_loss": 0.50390625, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 2.09375, | |
| "rewards/margins": 0.6953125, | |
| "rewards/rejected": 1.3984375, | |
| "step": 222 | |
| }, | |
| { | |
| "epoch": 0.892, | |
| "grad_norm": 10.902340209779693, | |
| "learning_rate": 3.0400846793612544e-08, | |
| "logits/chosen": 1.1796875, | |
| "logits/rejected": 1.2265625, | |
| "logps/chosen": -4.71875, | |
| "logps/rejected": -15.4375, | |
| "loss": 1.1767, | |
| "nll_loss": 0.447265625, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": 2.390625, | |
| "rewards/margins": 0.5390625, | |
| "rewards/rejected": 1.8515625, | |
| "step": 223 | |
| }, | |
| { | |
| "epoch": 0.896, | |
| "grad_norm": 12.355828877171342, | |
| "learning_rate": 2.8211666744009045e-08, | |
| "logits/chosen": 1.109375, | |
| "logits/rejected": 1.09375, | |
| "logps/chosen": -4.25, | |
| "logps/rejected": -21.625, | |
| "loss": 1.1988, | |
| "nll_loss": 0.38671875, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 2.1875, | |
| "rewards/margins": 1.1953125, | |
| "rewards/rejected": 1.0, | |
| "step": 224 | |
| }, | |
| { | |
| "epoch": 0.9, | |
| "grad_norm": 13.420061159806268, | |
| "learning_rate": 2.6101994563013098e-08, | |
| "logits/chosen": 1.140625, | |
| "logits/rejected": 1.1640625, | |
| "logps/chosen": -3.953125, | |
| "logps/rejected": -21.0, | |
| "loss": 1.2244, | |
| "nll_loss": 0.39453125, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": 2.21875, | |
| "rewards/margins": 1.375, | |
| "rewards/rejected": 0.84375, | |
| "step": 225 | |
| }, | |
| { | |
| "epoch": 0.904, | |
| "grad_norm": 11.000297680662745, | |
| "learning_rate": 2.4072185781965238e-08, | |
| "logits/chosen": 1.125, | |
| "logits/rejected": 1.125, | |
| "logps/chosen": -5.96875, | |
| "logps/rejected": -23.375, | |
| "loss": 1.2071, | |
| "nll_loss": 0.54296875, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 2.3125, | |
| "rewards/margins": 1.5859375, | |
| "rewards/rejected": 0.71875, | |
| "step": 226 | |
| }, | |
| { | |
| "epoch": 0.908, | |
| "grad_norm": 14.17266394056175, | |
| "learning_rate": 2.212258247327059e-08, | |
| "logits/chosen": 1.125, | |
| "logits/rejected": 1.109375, | |
| "logps/chosen": -3.75, | |
| "logps/rejected": -22.5, | |
| "loss": 1.171, | |
| "nll_loss": 0.3828125, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": 2.15625, | |
| "rewards/margins": 1.6484375, | |
| "rewards/rejected": 0.51171875, | |
| "step": 227 | |
| }, | |
| { | |
| "epoch": 0.912, | |
| "grad_norm": 11.33899892588413, | |
| "learning_rate": 2.025351319275137e-08, | |
| "logits/chosen": 1.125, | |
| "logits/rejected": 1.1640625, | |
| "logps/chosen": -6.84375, | |
| "logps/rejected": -28.0, | |
| "loss": 1.2811, | |
| "nll_loss": 0.58203125, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 2.171875, | |
| "rewards/margins": 2.03125, | |
| "rewards/rejected": 0.1337890625, | |
| "step": 228 | |
| }, | |
| { | |
| "epoch": 0.916, | |
| "grad_norm": 12.453357132102518, | |
| "learning_rate": 1.8465292924276844e-08, | |
| "logits/chosen": 1.0859375, | |
| "logits/rejected": 1.125, | |
| "logps/chosen": -5.9375, | |
| "logps/rejected": -32.75, | |
| "loss": 1.1113, | |
| "nll_loss": 0.5390625, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 2.4375, | |
| "rewards/margins": 2.4375, | |
| "rewards/rejected": -3.0517578125e-05, | |
| "step": 229 | |
| }, | |
| { | |
| "epoch": 0.92, | |
| "grad_norm": 11.898178262181379, | |
| "learning_rate": 1.6758223026681507e-08, | |
| "logits/chosen": 1.1328125, | |
| "logits/rejected": 1.140625, | |
| "logps/chosen": -5.8125, | |
| "logps/rejected": -19.125, | |
| "loss": 1.2299, | |
| "nll_loss": 0.54296875, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 1.8984375, | |
| "rewards/margins": 0.96875, | |
| "rewards/rejected": 0.9296875, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.924, | |
| "grad_norm": 10.263964920299092, | |
| "learning_rate": 1.5132591182978105e-08, | |
| "logits/chosen": 1.2109375, | |
| "logits/rejected": 1.234375, | |
| "logps/chosen": -7.09375, | |
| "logps/rejected": -23.875, | |
| "loss": 1.2085, | |
| "nll_loss": 0.640625, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 2.359375, | |
| "rewards/margins": 1.34375, | |
| "rewards/rejected": 1.0234375, | |
| "step": 231 | |
| }, | |
| { | |
| "epoch": 0.928, | |
| "grad_norm": 12.654190061391278, | |
| "learning_rate": 1.3588671351876357e-08, | |
| "logits/chosen": 1.2109375, | |
| "logits/rejected": 1.2265625, | |
| "logps/chosen": -4.4375, | |
| "logps/rejected": -15.8125, | |
| "loss": 1.2357, | |
| "nll_loss": 0.443359375, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 2.265625, | |
| "rewards/margins": 0.76953125, | |
| "rewards/rejected": 1.4921875, | |
| "step": 232 | |
| }, | |
| { | |
| "epoch": 0.932, | |
| "grad_norm": 14.352106481713815, | |
| "learning_rate": 1.2126723721614052e-08, | |
| "logits/chosen": 1.15625, | |
| "logits/rejected": 1.1484375, | |
| "logps/chosen": -4.59375, | |
| "logps/rejected": -20.5, | |
| "loss": 1.2552, | |
| "nll_loss": 0.458984375, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 2.0625, | |
| "rewards/margins": 1.1484375, | |
| "rewards/rejected": 0.921875, | |
| "step": 233 | |
| }, | |
| { | |
| "epoch": 0.936, | |
| "grad_norm": 11.269660437873329, | |
| "learning_rate": 1.0746994666109233e-08, | |
| "logits/chosen": 1.1875, | |
| "logits/rejected": 1.1875, | |
| "logps/chosen": -4.5625, | |
| "logps/rejected": -12.6875, | |
| "loss": 1.1757, | |
| "nll_loss": 0.416015625, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 2.546875, | |
| "rewards/margins": 0.81640625, | |
| "rewards/rejected": 1.734375, | |
| "step": 234 | |
| }, | |
| { | |
| "epoch": 0.94, | |
| "grad_norm": 12.81841594760115, | |
| "learning_rate": 9.449716703439803e-09, | |
| "logits/chosen": 1.0546875, | |
| "logits/rejected": 1.046875, | |
| "logps/chosen": -5.09375, | |
| "logps/rejected": -14.875, | |
| "loss": 1.2255, | |
| "nll_loss": 0.466796875, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 2.625, | |
| "rewards/margins": 1.2578125, | |
| "rewards/rejected": 1.3671875, | |
| "step": 235 | |
| }, | |
| { | |
| "epoch": 0.944, | |
| "grad_norm": 10.920504150516587, | |
| "learning_rate": 8.235108456658812e-09, | |
| "logits/chosen": 1.1640625, | |
| "logits/rejected": 1.15625, | |
| "logps/chosen": -4.59375, | |
| "logps/rejected": -27.75, | |
| "loss": 1.1622, | |
| "nll_loss": 0.4375, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 2.078125, | |
| "rewards/margins": 1.859375, | |
| "rewards/rejected": 0.2138671875, | |
| "step": 236 | |
| }, | |
| { | |
| "epoch": 0.948, | |
| "grad_norm": 17.488700410279908, | |
| "learning_rate": 7.103374616951041e-09, | |
| "logits/chosen": 1.15625, | |
| "logits/rejected": 1.1484375, | |
| "logps/chosen": -4.40625, | |
| "logps/rejected": -13.5625, | |
| "loss": 1.2803, | |
| "nll_loss": 0.43359375, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 2.546875, | |
| "rewards/margins": 0.8125, | |
| "rewards/rejected": 1.7421875, | |
| "step": 237 | |
| }, | |
| { | |
| "epoch": 0.952, | |
| "grad_norm": 15.482877470290028, | |
| "learning_rate": 6.054705909137425e-09, | |
| "logits/chosen": 1.1171875, | |
| "logits/rejected": 1.125, | |
| "logps/chosen": -4.75, | |
| "logps/rejected": -24.125, | |
| "loss": 1.2733, | |
| "nll_loss": 0.50390625, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": 2.125, | |
| "rewards/margins": 1.78125, | |
| "rewards/rejected": 0.349609375, | |
| "step": 238 | |
| }, | |
| { | |
| "epoch": 0.956, | |
| "grad_norm": 11.061399534591029, | |
| "learning_rate": 5.0892790595336575e-09, | |
| "logits/chosen": 1.109375, | |
| "logits/rejected": 1.1484375, | |
| "logps/chosen": -5.125, | |
| "logps/rejected": -19.375, | |
| "loss": 1.139, | |
| "nll_loss": 0.5078125, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": 2.109375, | |
| "rewards/margins": 1.4140625, | |
| "rewards/rejected": 0.6953125, | |
| "step": 239 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "grad_norm": 11.714703245702147, | |
| "learning_rate": 4.207256766166845e-09, | |
| "logits/chosen": 1.2109375, | |
| "logits/rejected": 1.1875, | |
| "logps/chosen": -5.53125, | |
| "logps/rejected": -18.375, | |
| "loss": 1.2334, | |
| "nll_loss": 0.51953125, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 2.015625, | |
| "rewards/margins": 1.0546875, | |
| "rewards/rejected": 0.953125, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.964, | |
| "grad_norm": 10.51394851397197, | |
| "learning_rate": 3.4087876713574935e-09, | |
| "logits/chosen": 1.1484375, | |
| "logits/rejected": 1.0859375, | |
| "logps/chosen": -5.8125, | |
| "logps/rejected": -21.5, | |
| "loss": 1.1431, | |
| "nll_loss": 0.59765625, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 2.40625, | |
| "rewards/margins": 1.7265625, | |
| "rewards/rejected": 0.6796875, | |
| "step": 241 | |
| }, | |
| { | |
| "epoch": 0.968, | |
| "grad_norm": 13.245501985890717, | |
| "learning_rate": 2.69400633666933e-09, | |
| "logits/chosen": 1.1328125, | |
| "logits/rejected": 1.15625, | |
| "logps/chosen": -4.40625, | |
| "logps/rejected": -14.125, | |
| "loss": 1.2378, | |
| "nll_loss": 0.40625, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 1.8984375, | |
| "rewards/margins": 0.6796875, | |
| "rewards/rejected": 1.21875, | |
| "step": 242 | |
| }, | |
| { | |
| "epoch": 0.972, | |
| "grad_norm": 11.496395504691947, | |
| "learning_rate": 2.0630332202324906e-09, | |
| "logits/chosen": 1.15625, | |
| "logits/rejected": 1.140625, | |
| "logps/chosen": -6.15625, | |
| "logps/rejected": -27.0, | |
| "loss": 1.2662, | |
| "nll_loss": 0.55078125, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": 2.28125, | |
| "rewards/margins": 1.8515625, | |
| "rewards/rejected": 0.42578125, | |
| "step": 243 | |
| }, | |
| { | |
| "epoch": 0.976, | |
| "grad_norm": 11.885785036184581, | |
| "learning_rate": 1.5159746564433196e-09, | |
| "logits/chosen": 1.1796875, | |
| "logits/rejected": 1.171875, | |
| "logps/chosen": -5.6875, | |
| "logps/rejected": -20.375, | |
| "loss": 1.2716, | |
| "nll_loss": 0.546875, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 1.984375, | |
| "rewards/margins": 1.328125, | |
| "rewards/rejected": 0.6484375, | |
| "step": 244 | |
| }, | |
| { | |
| "epoch": 0.98, | |
| "grad_norm": 14.24854333069195, | |
| "learning_rate": 1.0529228380445899e-09, | |
| "logits/chosen": 1.140625, | |
| "logits/rejected": 1.1796875, | |
| "logps/chosen": -5.34375, | |
| "logps/rejected": -28.25, | |
| "loss": 1.1959, | |
| "nll_loss": 0.49609375, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 2.328125, | |
| "rewards/margins": 2.015625, | |
| "rewards/rejected": 0.31640625, | |
| "step": 245 | |
| }, | |
| { | |
| "epoch": 0.984, | |
| "grad_norm": 12.20087123285817, | |
| "learning_rate": 6.739558005884883e-10, | |
| "logits/chosen": 1.1953125, | |
| "logits/rejected": 1.203125, | |
| "logps/chosen": -6.1875, | |
| "logps/rejected": -20.25, | |
| "loss": 1.2076, | |
| "nll_loss": 0.5546875, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 2.1875, | |
| "rewards/margins": 1.140625, | |
| "rewards/rejected": 1.046875, | |
| "step": 246 | |
| }, | |
| { | |
| "epoch": 0.988, | |
| "grad_norm": 13.016658555990439, | |
| "learning_rate": 3.791374092859678e-10, | |
| "logits/chosen": 1.0546875, | |
| "logits/rejected": 1.0390625, | |
| "logps/chosen": -7.15625, | |
| "logps/rejected": -21.5, | |
| "loss": 1.2678, | |
| "nll_loss": 0.640625, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": 1.953125, | |
| "rewards/margins": 1.265625, | |
| "rewards/rejected": 0.6875, | |
| "step": 247 | |
| }, | |
| { | |
| "epoch": 0.992, | |
| "grad_norm": 13.049688068801364, | |
| "learning_rate": 1.6851734824380182e-10, | |
| "logits/chosen": 1.203125, | |
| "logits/rejected": 1.203125, | |
| "logps/chosen": -3.828125, | |
| "logps/rejected": -27.0, | |
| "loss": 1.3064, | |
| "nll_loss": 0.3671875, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 2.421875, | |
| "rewards/margins": 1.765625, | |
| "rewards/rejected": 0.65625, | |
| "step": 248 | |
| }, | |
| { | |
| "epoch": 0.996, | |
| "grad_norm": 11.481752941680474, | |
| "learning_rate": 4.213111209155906e-11, | |
| "logits/chosen": 1.0859375, | |
| "logits/rejected": 1.1015625, | |
| "logps/chosen": -4.34375, | |
| "logps/rejected": -20.0, | |
| "loss": 1.153, | |
| "nll_loss": 0.3984375, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 2.328125, | |
| "rewards/margins": 1.4765625, | |
| "rewards/rejected": 0.85546875, | |
| "step": 249 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 12.415871966366533, | |
| "learning_rate": 0.0, | |
| "logits/chosen": 1.1484375, | |
| "logits/rejected": 1.140625, | |
| "logps/chosen": -4.96875, | |
| "logps/rejected": -21.75, | |
| "loss": 1.1798, | |
| "nll_loss": 0.482421875, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 1.921875, | |
| "rewards/margins": 1.328125, | |
| "rewards/rejected": 0.59765625, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 250, | |
| "total_flos": 1.540582387141627e+20, | |
| "train_loss": 1.5058151931762696, | |
| "train_runtime": 11811.4633, | |
| "train_samples_per_second": 5.418, | |
| "train_steps_per_second": 0.021 | |
| } | |
| ], | |
| "logging_steps": 1.0, | |
| "max_steps": 250, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 100.0, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": false, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 1.540582387141627e+20, | |
| "train_batch_size": 1, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |