| { | |
| "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": 108.45250504320283, | |
| "learning_rate": 1.25e-07, | |
| "logits/chosen": -1.1796875, | |
| "logits/rejected": -1.1796875, | |
| "logps/chosen": -32.0, | |
| "logps/rejected": -35.25, | |
| "loss": 4.0839, | |
| "nll_loss": 3.421875, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 0.04150390625, | |
| "rewards/margins": 0.099609375, | |
| "rewards/rejected": -0.058349609375, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.008, | |
| "grad_norm": 120.95987737750156, | |
| "learning_rate": 2.5e-07, | |
| "logits/chosen": -1.0703125, | |
| "logits/rejected": -1.0703125, | |
| "logps/chosen": -38.0, | |
| "logps/rejected": -35.25, | |
| "loss": 4.3121, | |
| "nll_loss": 3.984375, | |
| "rewards/accuracies": 0.375, | |
| "rewards/chosen": 0.046142578125, | |
| "rewards/margins": 0.01171875, | |
| "rewards/rejected": 0.034423828125, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.012, | |
| "grad_norm": 109.89014011391664, | |
| "learning_rate": 3.75e-07, | |
| "logits/chosen": -1.1875, | |
| "logits/rejected": -1.171875, | |
| "logps/chosen": -33.0, | |
| "logps/rejected": -37.0, | |
| "loss": 4.2, | |
| "nll_loss": 3.3125, | |
| "rewards/accuracies": 0.53125, | |
| "rewards/chosen": 0.05859375, | |
| "rewards/margins": 0.07421875, | |
| "rewards/rejected": -0.0155029296875, | |
| "step": 3 | |
| }, | |
| { | |
| "epoch": 0.016, | |
| "grad_norm": 105.87845131816027, | |
| "learning_rate": 5e-07, | |
| "logits/chosen": -1.234375, | |
| "logits/rejected": -1.203125, | |
| "logps/chosen": -31.25, | |
| "logps/rejected": -36.75, | |
| "loss": 4.2118, | |
| "nll_loss": 3.03125, | |
| "rewards/accuracies": 0.5625, | |
| "rewards/chosen": 0.0595703125, | |
| "rewards/margins": 0.0390625, | |
| "rewards/rejected": 0.020263671875, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.02, | |
| "grad_norm": 112.81877726787484, | |
| "learning_rate": 6.249999999999999e-07, | |
| "logits/chosen": -1.2265625, | |
| "logits/rejected": -1.2109375, | |
| "logps/chosen": -33.0, | |
| "logps/rejected": -34.25, | |
| "loss": 4.0945, | |
| "nll_loss": 3.453125, | |
| "rewards/accuracies": 0.46875, | |
| "rewards/chosen": 0.06298828125, | |
| "rewards/margins": -0.00994873046875, | |
| "rewards/rejected": 0.07275390625, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.024, | |
| "grad_norm": 106.29693927531781, | |
| "learning_rate": 7.5e-07, | |
| "logits/chosen": -1.1796875, | |
| "logits/rejected": -1.1875, | |
| "logps/chosen": -33.0, | |
| "logps/rejected": -36.5, | |
| "loss": 4.1827, | |
| "nll_loss": 3.3125, | |
| "rewards/accuracies": 0.46875, | |
| "rewards/chosen": 0.08642578125, | |
| "rewards/margins": 0.037109375, | |
| "rewards/rejected": 0.04931640625, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.028, | |
| "grad_norm": 114.94478205565038, | |
| "learning_rate": 8.75e-07, | |
| "logits/chosen": -1.1796875, | |
| "logits/rejected": -1.171875, | |
| "logps/chosen": -27.875, | |
| "logps/rejected": -33.5, | |
| "loss": 4.1499, | |
| "nll_loss": 2.84375, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": 0.0625, | |
| "rewards/margins": 0.01165771484375, | |
| "rewards/rejected": 0.051025390625, | |
| "step": 7 | |
| }, | |
| { | |
| "epoch": 0.032, | |
| "grad_norm": 97.37219322957407, | |
| "learning_rate": 1e-06, | |
| "logits/chosen": -1.171875, | |
| "logits/rejected": -1.171875, | |
| "logps/chosen": -32.5, | |
| "logps/rejected": -30.625, | |
| "loss": 4.0641, | |
| "nll_loss": 3.515625, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": 0.1552734375, | |
| "rewards/margins": 0.030517578125, | |
| "rewards/rejected": 0.125, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.036, | |
| "grad_norm": 108.05293741004135, | |
| "learning_rate": 9.999578688879084e-07, | |
| "logits/chosen": -1.15625, | |
| "logits/rejected": -1.1484375, | |
| "logps/chosen": -31.875, | |
| "logps/rejected": -37.25, | |
| "loss": 4.2044, | |
| "nll_loss": 3.234375, | |
| "rewards/accuracies": 0.53125, | |
| "rewards/chosen": 0.19140625, | |
| "rewards/margins": 0.055908203125, | |
| "rewards/rejected": 0.134765625, | |
| "step": 9 | |
| }, | |
| { | |
| "epoch": 0.04, | |
| "grad_norm": 89.47841349842803, | |
| "learning_rate": 9.998314826517563e-07, | |
| "logits/chosen": -1.1015625, | |
| "logits/rejected": -1.1171875, | |
| "logps/chosen": -27.75, | |
| "logps/rejected": -29.125, | |
| "loss": 3.8131, | |
| "nll_loss": 2.875, | |
| "rewards/accuracies": 0.375, | |
| "rewards/chosen": 0.318359375, | |
| "rewards/margins": -0.049072265625, | |
| "rewards/rejected": 0.3671875, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.044, | |
| "grad_norm": 82.2909196761966, | |
| "learning_rate": 9.99620862590714e-07, | |
| "logits/chosen": -1.1875, | |
| "logits/rejected": -1.1640625, | |
| "logps/chosen": -26.0, | |
| "logps/rejected": -30.625, | |
| "loss": 3.7873, | |
| "nll_loss": 2.640625, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": 0.5390625, | |
| "rewards/margins": 0.028076171875, | |
| "rewards/rejected": 0.51171875, | |
| "step": 11 | |
| }, | |
| { | |
| "epoch": 0.048, | |
| "grad_norm": 77.5417011003063, | |
| "learning_rate": 9.993260441994114e-07, | |
| "logits/chosen": -1.1796875, | |
| "logits/rejected": -1.1328125, | |
| "logps/chosen": -26.375, | |
| "logps/rejected": -25.125, | |
| "loss": 3.6977, | |
| "nll_loss": 2.8125, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": 0.515625, | |
| "rewards/margins": 0.02099609375, | |
| "rewards/rejected": 0.49609375, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 0.052, | |
| "grad_norm": 77.36031138542103, | |
| "learning_rate": 9.989470771619553e-07, | |
| "logits/chosen": -1.09375, | |
| "logits/rejected": -1.109375, | |
| "logps/chosen": -31.375, | |
| "logps/rejected": -31.75, | |
| "loss": 3.84, | |
| "nll_loss": 3.171875, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": 0.66015625, | |
| "rewards/margins": 0.0380859375, | |
| "rewards/rejected": 0.62109375, | |
| "step": 13 | |
| }, | |
| { | |
| "epoch": 0.056, | |
| "grad_norm": 38.474184849981015, | |
| "learning_rate": 9.984840253435568e-07, | |
| "logits/chosen": -1.1640625, | |
| "logits/rejected": -1.140625, | |
| "logps/chosen": -21.5, | |
| "logps/rejected": -23.125, | |
| "loss": 2.9964, | |
| "nll_loss": 2.15625, | |
| "rewards/accuracies": 0.46875, | |
| "rewards/chosen": 1.0625, | |
| "rewards/margins": 0.07958984375, | |
| "rewards/rejected": 0.984375, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 0.06, | |
| "grad_norm": 37.90903240520294, | |
| "learning_rate": 9.979369667797674e-07, | |
| "logits/chosen": -1.0703125, | |
| "logits/rejected": -1.09375, | |
| "logps/chosen": -20.5, | |
| "logps/rejected": -25.25, | |
| "loss": 3.0721, | |
| "nll_loss": 2.0, | |
| "rewards/accuracies": 0.46875, | |
| "rewards/chosen": 1.078125, | |
| "rewards/margins": -0.060546875, | |
| "rewards/rejected": 1.140625, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.064, | |
| "grad_norm": 35.90082325141723, | |
| "learning_rate": 9.973059936633306e-07, | |
| "logits/chosen": -1.078125, | |
| "logits/rejected": -1.1015625, | |
| "logps/chosen": -17.0, | |
| "logps/rejected": -25.25, | |
| "loss": 2.9806, | |
| "nll_loss": 1.71875, | |
| "rewards/accuracies": 0.4375, | |
| "rewards/chosen": 0.875, | |
| "rewards/margins": -0.0517578125, | |
| "rewards/rejected": 0.92578125, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 0.068, | |
| "grad_norm": 34.39419041472641, | |
| "learning_rate": 9.965912123286424e-07, | |
| "logits/chosen": -0.99609375, | |
| "logits/rejected": -0.98046875, | |
| "logps/chosen": -21.875, | |
| "logps/rejected": -24.375, | |
| "loss": 2.9524, | |
| "nll_loss": 2.125, | |
| "rewards/accuracies": 0.5625, | |
| "rewards/chosen": 1.109375, | |
| "rewards/margins": 0.1552734375, | |
| "rewards/rejected": 0.95703125, | |
| "step": 17 | |
| }, | |
| { | |
| "epoch": 0.072, | |
| "grad_norm": 33.10304291271997, | |
| "learning_rate": 9.95792743233833e-07, | |
| "logits/chosen": -1.03125, | |
| "logits/rejected": -1.0078125, | |
| "logps/chosen": -18.75, | |
| "logps/rejected": -24.25, | |
| "loss": 2.8946, | |
| "nll_loss": 1.9765625, | |
| "rewards/accuracies": 0.53125, | |
| "rewards/chosen": 1.2265625, | |
| "rewards/margins": 0.0380859375, | |
| "rewards/rejected": 1.1875, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 0.076, | |
| "grad_norm": 28.411495189507995, | |
| "learning_rate": 9.949107209404663e-07, | |
| "logits/chosen": -1.03125, | |
| "logits/rejected": -1.0234375, | |
| "logps/chosen": -19.5, | |
| "logps/rejected": -22.875, | |
| "loss": 2.7523, | |
| "nll_loss": 1.984375, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": 1.5703125, | |
| "rewards/margins": 0.0096435546875, | |
| "rewards/rejected": 1.5625, | |
| "step": 19 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "grad_norm": 30.85467104218471, | |
| "learning_rate": 9.939452940908626e-07, | |
| "logits/chosen": -1.0703125, | |
| "logits/rejected": -1.046875, | |
| "logps/chosen": -20.875, | |
| "logps/rejected": -23.25, | |
| "loss": 2.8015, | |
| "nll_loss": 2.0625, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 1.6171875, | |
| "rewards/margins": 0.296875, | |
| "rewards/rejected": 1.3203125, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.084, | |
| "grad_norm": 20.81104325101837, | |
| "learning_rate": 9.92896625383049e-07, | |
| "logits/chosen": -0.98828125, | |
| "logits/rejected": -0.9765625, | |
| "logps/chosen": -16.125, | |
| "logps/rejected": -21.0, | |
| "loss": 2.6257, | |
| "nll_loss": 1.640625, | |
| "rewards/accuracies": 0.46875, | |
| "rewards/chosen": 1.4609375, | |
| "rewards/margins": -0.09423828125, | |
| "rewards/rejected": 1.5546875, | |
| "step": 21 | |
| }, | |
| { | |
| "epoch": 0.088, | |
| "grad_norm": 16.555834400457428, | |
| "learning_rate": 9.917648915433411e-07, | |
| "logits/chosen": -0.95703125, | |
| "logits/rejected": -0.94140625, | |
| "logps/chosen": -15.8125, | |
| "logps/rejected": -16.875, | |
| "loss": 2.4551, | |
| "nll_loss": 1.53125, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": 1.703125, | |
| "rewards/margins": 0.16796875, | |
| "rewards/rejected": 1.53125, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 0.092, | |
| "grad_norm": 15.376428724494609, | |
| "learning_rate": 9.905502832965603e-07, | |
| "logits/chosen": -0.9296875, | |
| "logits/rejected": -0.91796875, | |
| "logps/chosen": -15.4375, | |
| "logps/rejected": -18.5, | |
| "loss": 2.4073, | |
| "nll_loss": 1.546875, | |
| "rewards/accuracies": 0.53125, | |
| "rewards/chosen": 1.6640625, | |
| "rewards/margins": -0.029052734375, | |
| "rewards/rejected": 1.6953125, | |
| "step": 23 | |
| }, | |
| { | |
| "epoch": 0.096, | |
| "grad_norm": 16.237442660961605, | |
| "learning_rate": 9.892530053338907e-07, | |
| "logits/chosen": -0.89453125, | |
| "logits/rejected": -0.8984375, | |
| "logps/chosen": -15.1875, | |
| "logps/rejected": -19.75, | |
| "loss": 2.3768, | |
| "nll_loss": 1.5546875, | |
| "rewards/accuracies": 0.46875, | |
| "rewards/chosen": 1.875, | |
| "rewards/margins": 0.08251953125, | |
| "rewards/rejected": 1.7890625, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "grad_norm": 15.758858425294349, | |
| "learning_rate": 9.87873276278386e-07, | |
| "logits/chosen": -0.8984375, | |
| "logits/rejected": -0.8828125, | |
| "logps/chosen": -14.5625, | |
| "logps/rejected": -16.875, | |
| "loss": 2.3852, | |
| "nll_loss": 1.4453125, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": 2.125, | |
| "rewards/margins": 0.05029296875, | |
| "rewards/rejected": 2.078125, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.104, | |
| "grad_norm": 16.347740539403798, | |
| "learning_rate": 9.864113286481236e-07, | |
| "logits/chosen": -0.9453125, | |
| "logits/rejected": -0.93359375, | |
| "logps/chosen": -14.125, | |
| "logps/rejected": -20.0, | |
| "loss": 2.317, | |
| "nll_loss": 1.46875, | |
| "rewards/accuracies": 0.5625, | |
| "rewards/chosen": 1.828125, | |
| "rewards/margins": -0.001220703125, | |
| "rewards/rejected": 1.828125, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 0.108, | |
| "grad_norm": 15.099343265144629, | |
| "learning_rate": 9.84867408817022e-07, | |
| "logits/chosen": -0.95703125, | |
| "logits/rejected": -0.9375, | |
| "logps/chosen": -13.125, | |
| "logps/rejected": -16.625, | |
| "loss": 2.2059, | |
| "nll_loss": 1.328125, | |
| "rewards/accuracies": 0.46875, | |
| "rewards/chosen": 2.015625, | |
| "rewards/margins": 0.1494140625, | |
| "rewards/rejected": 1.859375, | |
| "step": 27 | |
| }, | |
| { | |
| "epoch": 0.112, | |
| "grad_norm": 15.421632558925108, | |
| "learning_rate": 9.832417769733185e-07, | |
| "logits/chosen": -0.94140625, | |
| "logits/rejected": -0.9375, | |
| "logps/chosen": -13.3125, | |
| "logps/rejected": -17.0, | |
| "loss": 2.2687, | |
| "nll_loss": 1.2578125, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": 1.890625, | |
| "rewards/margins": 0.044921875, | |
| "rewards/rejected": 1.84375, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 0.116, | |
| "grad_norm": 15.06033701313281, | |
| "learning_rate": 9.815347070757231e-07, | |
| "logits/chosen": -0.97265625, | |
| "logits/rejected": -0.9609375, | |
| "logps/chosen": -14.875, | |
| "logps/rejected": -17.875, | |
| "loss": 2.2442, | |
| "nll_loss": 1.375, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": 2.109375, | |
| "rewards/margins": 0.3984375, | |
| "rewards/rejected": 1.71875, | |
| "step": 29 | |
| }, | |
| { | |
| "epoch": 0.12, | |
| "grad_norm": 13.532947619496891, | |
| "learning_rate": 9.797464868072486e-07, | |
| "logits/chosen": -0.91796875, | |
| "logits/rejected": -0.89453125, | |
| "logps/chosen": -13.0625, | |
| "logps/rejected": -16.75, | |
| "loss": 2.1401, | |
| "nll_loss": 1.3125, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 1.8203125, | |
| "rewards/margins": 0.18359375, | |
| "rewards/rejected": 1.6328125, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.124, | |
| "grad_norm": 13.807093125286677, | |
| "learning_rate": 9.778774175267295e-07, | |
| "logits/chosen": -0.9375, | |
| "logits/rejected": -0.9296875, | |
| "logps/chosen": -11.75, | |
| "logps/rejected": -17.75, | |
| "loss": 2.1347, | |
| "nll_loss": 1.1953125, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 2.109375, | |
| "rewards/margins": 0.1552734375, | |
| "rewards/rejected": 1.9453125, | |
| "step": 31 | |
| }, | |
| { | |
| "epoch": 0.128, | |
| "grad_norm": 14.709867281760857, | |
| "learning_rate": 9.759278142180347e-07, | |
| "logits/chosen": -0.9765625, | |
| "logits/rejected": -0.94921875, | |
| "logps/chosen": -9.8125, | |
| "logps/rejected": -15.3125, | |
| "loss": 2.0944, | |
| "nll_loss": 0.98828125, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": 1.96875, | |
| "rewards/margins": 0.11962890625, | |
| "rewards/rejected": 1.8515625, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 0.132, | |
| "grad_norm": 12.980725519975202, | |
| "learning_rate": 9.73898005436987e-07, | |
| "logits/chosen": -0.95703125, | |
| "logits/rejected": -0.96484375, | |
| "logps/chosen": -11.0, | |
| "logps/rejected": -14.8125, | |
| "loss": 2.032, | |
| "nll_loss": 1.078125, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 1.9921875, | |
| "rewards/margins": 0.345703125, | |
| "rewards/rejected": 1.6484375, | |
| "step": 33 | |
| }, | |
| { | |
| "epoch": 0.136, | |
| "grad_norm": 12.423187407094442, | |
| "learning_rate": 9.71788333255991e-07, | |
| "logits/chosen": -0.97265625, | |
| "logits/rejected": -0.98046875, | |
| "logps/chosen": -12.0625, | |
| "logps/rejected": -15.125, | |
| "loss": 1.9855, | |
| "nll_loss": 1.2890625, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": 2.125, | |
| "rewards/margins": 0.392578125, | |
| "rewards/rejected": 1.734375, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 0.14, | |
| "grad_norm": 10.78340956295933, | |
| "learning_rate": 9.695991532063874e-07, | |
| "logits/chosen": -1.0078125, | |
| "logits/rejected": -0.97265625, | |
| "logps/chosen": -11.5, | |
| "logps/rejected": -14.5625, | |
| "loss": 1.9101, | |
| "nll_loss": 1.1484375, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 2.53125, | |
| "rewards/margins": 0.71875, | |
| "rewards/rejected": 1.8125, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.144, | |
| "grad_norm": 10.539848066893843, | |
| "learning_rate": 9.673308342185365e-07, | |
| "logits/chosen": -0.9609375, | |
| "logits/rejected": -0.9609375, | |
| "logps/chosen": -11.6875, | |
| "logps/rejected": -16.625, | |
| "loss": 1.8997, | |
| "nll_loss": 1.1640625, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": 2.21875, | |
| "rewards/margins": 0.224609375, | |
| "rewards/rejected": 1.984375, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 0.148, | |
| "grad_norm": 10.7200500702218, | |
| "learning_rate": 9.649837585596444e-07, | |
| "logits/chosen": -0.921875, | |
| "logits/rejected": -0.9296875, | |
| "logps/chosen": -10.25, | |
| "logps/rejected": -16.125, | |
| "loss": 1.8621, | |
| "nll_loss": 0.94921875, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": 2.171875, | |
| "rewards/margins": 0.345703125, | |
| "rewards/rejected": 1.828125, | |
| "step": 37 | |
| }, | |
| { | |
| "epoch": 0.152, | |
| "grad_norm": 9.54892533874284, | |
| "learning_rate": 9.625583217693418e-07, | |
| "logits/chosen": -0.93359375, | |
| "logits/rejected": -0.9453125, | |
| "logps/chosen": -11.125, | |
| "logps/rejected": -14.375, | |
| "loss": 1.7924, | |
| "nll_loss": 1.0703125, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": 3.09375, | |
| "rewards/margins": 1.09375, | |
| "rewards/rejected": 1.9921875, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 0.156, | |
| "grad_norm": 10.315046477409297, | |
| "learning_rate": 9.600549325930259e-07, | |
| "logits/chosen": -0.921875, | |
| "logits/rejected": -0.92578125, | |
| "logps/chosen": -10.375, | |
| "logps/rejected": -15.5, | |
| "loss": 1.8186, | |
| "nll_loss": 1.0546875, | |
| "rewards/accuracies": 0.53125, | |
| "rewards/chosen": 2.5625, | |
| "rewards/margins": 0.51171875, | |
| "rewards/rejected": 2.046875, | |
| "step": 39 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "grad_norm": 8.518386491338127, | |
| "learning_rate": 9.574740129129765e-07, | |
| "logits/chosen": -0.92578125, | |
| "logits/rejected": -0.90234375, | |
| "logps/chosen": -10.8125, | |
| "logps/rejected": -14.4375, | |
| "loss": 1.7032, | |
| "nll_loss": 1.0234375, | |
| "rewards/accuracies": 0.5625, | |
| "rewards/chosen": 2.015625, | |
| "rewards/margins": 0.07958984375, | |
| "rewards/rejected": 1.9375, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.164, | |
| "grad_norm": 9.17761100715551, | |
| "learning_rate": 9.548159976772592e-07, | |
| "logits/chosen": -0.95703125, | |
| "logits/rejected": -0.93359375, | |
| "logps/chosen": -9.375, | |
| "logps/rejected": -14.375, | |
| "loss": 1.8061, | |
| "nll_loss": 0.93359375, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": 2.4375, | |
| "rewards/margins": 0.357421875, | |
| "rewards/rejected": 2.09375, | |
| "step": 41 | |
| }, | |
| { | |
| "epoch": 0.168, | |
| "grad_norm": 9.6427558562046, | |
| "learning_rate": 9.520813348264251e-07, | |
| "logits/chosen": -0.91015625, | |
| "logits/rejected": -0.90625, | |
| "logps/chosen": -8.6875, | |
| "logps/rejected": -20.75, | |
| "loss": 1.6474, | |
| "nll_loss": 0.82421875, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 2.8125, | |
| "rewards/margins": 0.98828125, | |
| "rewards/rejected": 1.8203125, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 0.172, | |
| "grad_norm": 10.106867319730664, | |
| "learning_rate": 9.492704852180226e-07, | |
| "logits/chosen": -0.88671875, | |
| "logits/rejected": -0.8984375, | |
| "logps/chosen": -8.0625, | |
| "logps/rejected": -17.625, | |
| "loss": 1.6427, | |
| "nll_loss": 0.734375, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 2.609375, | |
| "rewards/margins": 0.9296875, | |
| "rewards/rejected": 1.671875, | |
| "step": 43 | |
| }, | |
| { | |
| "epoch": 0.176, | |
| "grad_norm": 9.780716883236895, | |
| "learning_rate": 9.463839225489318e-07, | |
| "logits/chosen": -0.91015625, | |
| "logits/rejected": -0.91015625, | |
| "logps/chosen": -9.5, | |
| "logps/rejected": -15.375, | |
| "loss": 1.7142, | |
| "nll_loss": 0.93359375, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 2.625, | |
| "rewards/margins": 0.515625, | |
| "rewards/rejected": 2.109375, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 0.18, | |
| "grad_norm": 9.324567213910779, | |
| "learning_rate": 9.434221332755338e-07, | |
| "logits/chosen": -0.90234375, | |
| "logits/rejected": -0.88671875, | |
| "logps/chosen": -8.125, | |
| "logps/rejected": -12.125, | |
| "loss": 1.6251, | |
| "nll_loss": 0.81640625, | |
| "rewards/accuracies": 0.53125, | |
| "rewards/chosen": 2.515625, | |
| "rewards/margins": 0.138671875, | |
| "rewards/rejected": 2.375, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 0.184, | |
| "grad_norm": 9.236623014788846, | |
| "learning_rate": 9.403856165317321e-07, | |
| "logits/chosen": -0.91015625, | |
| "logits/rejected": -0.9140625, | |
| "logps/chosen": -8.875, | |
| "logps/rejected": -18.75, | |
| "loss": 1.542, | |
| "nll_loss": 0.87890625, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 2.515625, | |
| "rewards/margins": 0.6796875, | |
| "rewards/rejected": 1.828125, | |
| "step": 46 | |
| }, | |
| { | |
| "epoch": 0.188, | |
| "grad_norm": 9.078752548744747, | |
| "learning_rate": 9.372748840448361e-07, | |
| "logits/chosen": -0.8828125, | |
| "logits/rejected": -0.8828125, | |
| "logps/chosen": -6.3125, | |
| "logps/rejected": -19.25, | |
| "loss": 1.5166, | |
| "nll_loss": 0.62890625, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 2.375, | |
| "rewards/margins": 0.68359375, | |
| "rewards/rejected": 1.6875, | |
| "step": 47 | |
| }, | |
| { | |
| "epoch": 0.192, | |
| "grad_norm": 9.007283454701437, | |
| "learning_rate": 9.340904600493219e-07, | |
| "logits/chosen": -0.890625, | |
| "logits/rejected": -0.90625, | |
| "logps/chosen": -7.15625, | |
| "logps/rejected": -17.125, | |
| "loss": 1.5793, | |
| "nll_loss": 0.78125, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": 2.421875, | |
| "rewards/margins": 0.55078125, | |
| "rewards/rejected": 1.8671875, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 0.196, | |
| "grad_norm": 9.111837952613337, | |
| "learning_rate": 9.308328811984869e-07, | |
| "logits/chosen": -0.92578125, | |
| "logits/rejected": -0.90234375, | |
| "logps/chosen": -7.375, | |
| "logps/rejected": -19.25, | |
| "loss": 1.4635, | |
| "nll_loss": 0.6796875, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 2.4375, | |
| "rewards/margins": 0.90625, | |
| "rewards/rejected": 1.5390625, | |
| "step": 49 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "grad_norm": 8.886170629930112, | |
| "learning_rate": 9.275026964740101e-07, | |
| "logits/chosen": -0.93359375, | |
| "logits/rejected": -0.9296875, | |
| "logps/chosen": -7.25, | |
| "logps/rejected": -14.3125, | |
| "loss": 1.4143, | |
| "nll_loss": 0.6796875, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 2.65625, | |
| "rewards/margins": 0.474609375, | |
| "rewards/rejected": 2.171875, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.204, | |
| "grad_norm": 8.951054512095832, | |
| "learning_rate": 9.241004670934347e-07, | |
| "logits/chosen": -0.9140625, | |
| "logits/rejected": -0.88671875, | |
| "logps/chosen": -6.46875, | |
| "logps/rejected": -18.625, | |
| "loss": 1.4843, | |
| "nll_loss": 0.59375, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 2.5, | |
| "rewards/margins": 0.78515625, | |
| "rewards/rejected": 1.71875, | |
| "step": 51 | |
| }, | |
| { | |
| "epoch": 0.208, | |
| "grad_norm": 8.428410886373113, | |
| "learning_rate": 9.206267664155906e-07, | |
| "logits/chosen": -0.9140625, | |
| "logits/rejected": -0.8828125, | |
| "logps/chosen": -8.5625, | |
| "logps/rejected": -17.0, | |
| "loss": 1.5279, | |
| "nll_loss": 0.875, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 2.59375, | |
| "rewards/margins": 0.46875, | |
| "rewards/rejected": 2.125, | |
| "step": 52 | |
| }, | |
| { | |
| "epoch": 0.212, | |
| "grad_norm": 13.359303723068443, | |
| "learning_rate": 9.170821798439683e-07, | |
| "logits/chosen": -0.921875, | |
| "logits/rejected": -0.921875, | |
| "logps/chosen": -7.21875, | |
| "logps/rejected": -19.75, | |
| "loss": 1.4738, | |
| "nll_loss": 0.7265625, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 2.25, | |
| "rewards/margins": 0.84375, | |
| "rewards/rejected": 1.4140625, | |
| "step": 53 | |
| }, | |
| { | |
| "epoch": 0.216, | |
| "grad_norm": 8.607311027896719, | |
| "learning_rate": 9.134673047280643e-07, | |
| "logits/chosen": -0.90234375, | |
| "logits/rejected": -0.890625, | |
| "logps/chosen": -7.3125, | |
| "logps/rejected": -12.75, | |
| "loss": 1.5091, | |
| "nll_loss": 0.6875, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 2.53125, | |
| "rewards/margins": 0.66796875, | |
| "rewards/rejected": 1.859375, | |
| "step": 54 | |
| }, | |
| { | |
| "epoch": 0.22, | |
| "grad_norm": 10.84727026791628, | |
| "learning_rate": 9.097827502627136e-07, | |
| "logits/chosen": -0.890625, | |
| "logits/rejected": -0.9140625, | |
| "logps/chosen": -5.40625, | |
| "logps/rejected": -15.125, | |
| "loss": 1.3565, | |
| "nll_loss": 0.50390625, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 2.90625, | |
| "rewards/margins": 0.72265625, | |
| "rewards/rejected": 2.1875, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 0.224, | |
| "grad_norm": 10.663731973556606, | |
| "learning_rate": 9.06029137385425e-07, | |
| "logits/chosen": -0.90234375, | |
| "logits/rejected": -0.875, | |
| "logps/chosen": -8.0625, | |
| "logps/rejected": -16.875, | |
| "loss": 1.4568, | |
| "nll_loss": 0.77734375, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 2.890625, | |
| "rewards/margins": 0.625, | |
| "rewards/rejected": 2.265625, | |
| "step": 56 | |
| }, | |
| { | |
| "epoch": 0.228, | |
| "grad_norm": 9.34103541373727, | |
| "learning_rate": 9.022070986717378e-07, | |
| "logits/chosen": -0.8984375, | |
| "logits/rejected": -0.90234375, | |
| "logps/chosen": -7.6875, | |
| "logps/rejected": -20.75, | |
| "loss": 1.4427, | |
| "nll_loss": 0.71875, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 2.40625, | |
| "rewards/margins": 1.125, | |
| "rewards/rejected": 1.28125, | |
| "step": 57 | |
| }, | |
| { | |
| "epoch": 0.232, | |
| "grad_norm": 9.905364057698737, | |
| "learning_rate": 8.983172782286179e-07, | |
| "logits/chosen": -0.9140625, | |
| "logits/rejected": -0.90234375, | |
| "logps/chosen": -6.46875, | |
| "logps/rejected": -14.125, | |
| "loss": 1.4267, | |
| "nll_loss": 0.69140625, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 2.875, | |
| "rewards/margins": 0.326171875, | |
| "rewards/rejected": 2.546875, | |
| "step": 58 | |
| }, | |
| { | |
| "epoch": 0.236, | |
| "grad_norm": 9.50298929918346, | |
| "learning_rate": 8.943603315859101e-07, | |
| "logits/chosen": -0.91796875, | |
| "logits/rejected": -0.92578125, | |
| "logps/chosen": -6.90625, | |
| "logps/rejected": -16.25, | |
| "loss": 1.3527, | |
| "nll_loss": 0.64453125, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": 2.921875, | |
| "rewards/margins": 1.4375, | |
| "rewards/rejected": 1.484375, | |
| "step": 59 | |
| }, | |
| { | |
| "epoch": 0.24, | |
| "grad_norm": 9.418798977945164, | |
| "learning_rate": 8.90336925585864e-07, | |
| "logits/chosen": -0.9140625, | |
| "logits/rejected": -0.94140625, | |
| "logps/chosen": -5.25, | |
| "logps/rejected": -19.0, | |
| "loss": 1.4071, | |
| "nll_loss": 0.546875, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 2.578125, | |
| "rewards/margins": 1.03125, | |
| "rewards/rejected": 1.5546875, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.244, | |
| "grad_norm": 12.471943357949753, | |
| "learning_rate": 8.862477382707568e-07, | |
| "logits/chosen": -0.9453125, | |
| "logits/rejected": -0.9453125, | |
| "logps/chosen": -7.8125, | |
| "logps/rejected": -19.5, | |
| "loss": 1.516, | |
| "nll_loss": 0.6640625, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 2.640625, | |
| "rewards/margins": 0.98828125, | |
| "rewards/rejected": 1.6640625, | |
| "step": 61 | |
| }, | |
| { | |
| "epoch": 0.248, | |
| "grad_norm": 10.996603171043574, | |
| "learning_rate": 8.820934587686245e-07, | |
| "logits/chosen": -0.94921875, | |
| "logits/rejected": -0.91796875, | |
| "logps/chosen": -7.21875, | |
| "logps/rejected": -17.0, | |
| "loss": 1.4125, | |
| "nll_loss": 0.703125, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 2.40625, | |
| "rewards/margins": 0.515625, | |
| "rewards/rejected": 1.890625, | |
| "step": 62 | |
| }, | |
| { | |
| "epoch": 0.252, | |
| "grad_norm": 10.253437222104349, | |
| "learning_rate": 8.778747871771291e-07, | |
| "logits/chosen": -0.91015625, | |
| "logits/rejected": -0.90234375, | |
| "logps/chosen": -5.75, | |
| "logps/rejected": -18.75, | |
| "loss": 1.3891, | |
| "nll_loss": 0.55078125, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": 2.9375, | |
| "rewards/margins": 1.140625, | |
| "rewards/rejected": 1.796875, | |
| "step": 63 | |
| }, | |
| { | |
| "epoch": 0.256, | |
| "grad_norm": 9.86186104574288, | |
| "learning_rate": 8.735924344455732e-07, | |
| "logits/chosen": -0.90234375, | |
| "logits/rejected": -0.91015625, | |
| "logps/chosen": -6.875, | |
| "logps/rejected": -18.5, | |
| "loss": 1.3061, | |
| "nll_loss": 0.67578125, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 3.015625, | |
| "rewards/margins": 1.171875, | |
| "rewards/rejected": 1.84375, | |
| "step": 64 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "grad_norm": 11.007376728694611, | |
| "learning_rate": 8.692471222550885e-07, | |
| "logits/chosen": -0.9453125, | |
| "logits/rejected": -0.9375, | |
| "logps/chosen": -7.625, | |
| "logps/rejected": -23.125, | |
| "loss": 1.4576, | |
| "nll_loss": 0.71875, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 2.625, | |
| "rewards/margins": 1.4765625, | |
| "rewards/rejected": 1.1484375, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 0.264, | |
| "grad_norm": 10.331552283839107, | |
| "learning_rate": 8.64839582897015e-07, | |
| "logits/chosen": -0.9375, | |
| "logits/rejected": -0.953125, | |
| "logps/chosen": -7.3125, | |
| "logps/rejected": -13.5625, | |
| "loss": 1.3471, | |
| "nll_loss": 0.6953125, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 2.921875, | |
| "rewards/margins": 0.384765625, | |
| "rewards/rejected": 2.546875, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 0.268, | |
| "grad_norm": 9.444511843257002, | |
| "learning_rate": 8.603705591494916e-07, | |
| "logits/chosen": -0.91015625, | |
| "logits/rejected": -0.921875, | |
| "logps/chosen": -6.53125, | |
| "logps/rejected": -16.5, | |
| "loss": 1.2306, | |
| "nll_loss": 0.6796875, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 2.25, | |
| "rewards/margins": 0.80078125, | |
| "rewards/rejected": 1.4453125, | |
| "step": 67 | |
| }, | |
| { | |
| "epoch": 0.272, | |
| "grad_norm": 12.577093117314861, | |
| "learning_rate": 8.5584080415228e-07, | |
| "logits/chosen": -0.9140625, | |
| "logits/rejected": -0.91015625, | |
| "logps/chosen": -5.46875, | |
| "logps/rejected": -15.5625, | |
| "loss": 1.2947, | |
| "nll_loss": 0.5078125, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 2.65625, | |
| "rewards/margins": 0.87890625, | |
| "rewards/rejected": 1.78125, | |
| "step": 68 | |
| }, | |
| { | |
| "epoch": 0.276, | |
| "grad_norm": 10.641105295546792, | |
| "learning_rate": 8.512510812798425e-07, | |
| "logits/chosen": -0.94140625, | |
| "logits/rejected": -0.9453125, | |
| "logps/chosen": -6.3125, | |
| "logps/rejected": -16.875, | |
| "loss": 1.2781, | |
| "nll_loss": 0.578125, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 2.921875, | |
| "rewards/margins": 1.0546875, | |
| "rewards/rejected": 1.875, | |
| "step": 69 | |
| }, | |
| { | |
| "epoch": 0.28, | |
| "grad_norm": 12.570856113406187, | |
| "learning_rate": 8.466021640126945e-07, | |
| "logits/chosen": -0.953125, | |
| "logits/rejected": -0.984375, | |
| "logps/chosen": -5.34375, | |
| "logps/rejected": -18.625, | |
| "loss": 1.3463, | |
| "nll_loss": 0.490234375, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 2.9375, | |
| "rewards/margins": 1.25, | |
| "rewards/rejected": 1.6953125, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.284, | |
| "grad_norm": 12.634804813000724, | |
| "learning_rate": 8.418948358070535e-07, | |
| "logits/chosen": -0.9453125, | |
| "logits/rejected": -0.9453125, | |
| "logps/chosen": -7.125, | |
| "logps/rejected": -17.25, | |
| "loss": 1.2968, | |
| "nll_loss": 0.66015625, | |
| "rewards/accuracies": 0.5625, | |
| "rewards/chosen": 2.703125, | |
| "rewards/margins": 0.9609375, | |
| "rewards/rejected": 1.7421875, | |
| "step": 71 | |
| }, | |
| { | |
| "epoch": 0.288, | |
| "grad_norm": 11.739576042187055, | |
| "learning_rate": 8.37129889962809e-07, | |
| "logits/chosen": -1.0, | |
| "logits/rejected": -0.9765625, | |
| "logps/chosen": -7.4375, | |
| "logps/rejected": -20.875, | |
| "loss": 1.3581, | |
| "nll_loss": 0.67578125, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 2.703125, | |
| "rewards/margins": 1.3046875, | |
| "rewards/rejected": 1.40625, | |
| "step": 72 | |
| }, | |
| { | |
| "epoch": 0.292, | |
| "grad_norm": 12.097928995634907, | |
| "learning_rate": 8.323081294898307e-07, | |
| "logits/chosen": -0.953125, | |
| "logits/rejected": -0.9453125, | |
| "logps/chosen": -5.3125, | |
| "logps/rejected": -21.5, | |
| "loss": 1.199, | |
| "nll_loss": 0.56640625, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 2.75, | |
| "rewards/margins": 1.2890625, | |
| "rewards/rejected": 1.4609375, | |
| "step": 73 | |
| }, | |
| { | |
| "epoch": 0.296, | |
| "grad_norm": 10.920824389039185, | |
| "learning_rate": 8.274303669726426e-07, | |
| "logits/chosen": -0.95703125, | |
| "logits/rejected": -0.94921875, | |
| "logps/chosen": -5.75, | |
| "logps/rejected": -18.625, | |
| "loss": 1.3358, | |
| "nll_loss": 0.51953125, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 2.609375, | |
| "rewards/margins": 1.15625, | |
| "rewards/rejected": 1.4609375, | |
| "step": 74 | |
| }, | |
| { | |
| "epoch": 0.3, | |
| "grad_norm": 14.177002319797584, | |
| "learning_rate": 8.22497424433482e-07, | |
| "logits/chosen": -0.93359375, | |
| "logits/rejected": -0.9296875, | |
| "logps/chosen": -6.125, | |
| "logps/rejected": -17.25, | |
| "loss": 1.2862, | |
| "nll_loss": 0.61328125, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 2.890625, | |
| "rewards/margins": 1.3984375, | |
| "rewards/rejected": 1.5, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.304, | |
| "grad_norm": 13.6699848015144, | |
| "learning_rate": 8.175101331937692e-07, | |
| "logits/chosen": -0.94921875, | |
| "logits/rejected": -0.95703125, | |
| "logps/chosen": -6.34375, | |
| "logps/rejected": -15.625, | |
| "loss": 1.2559, | |
| "nll_loss": 0.66015625, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": 2.8125, | |
| "rewards/margins": 1.140625, | |
| "rewards/rejected": 1.671875, | |
| "step": 76 | |
| }, | |
| { | |
| "epoch": 0.308, | |
| "grad_norm": 13.010408259816026, | |
| "learning_rate": 8.124693337340092e-07, | |
| "logits/chosen": -0.99609375, | |
| "logits/rejected": -0.96484375, | |
| "logps/chosen": -6.96875, | |
| "logps/rejected": -21.625, | |
| "loss": 1.3234, | |
| "nll_loss": 0.6875, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 2.890625, | |
| "rewards/margins": 1.3125, | |
| "rewards/rejected": 1.5703125, | |
| "step": 77 | |
| }, | |
| { | |
| "epoch": 0.312, | |
| "grad_norm": 14.267160801384883, | |
| "learning_rate": 8.073758755521504e-07, | |
| "logits/chosen": -0.99609375, | |
| "logits/rejected": -0.98828125, | |
| "logps/chosen": -6.1875, | |
| "logps/rejected": -23.875, | |
| "loss": 1.2073, | |
| "nll_loss": 0.6171875, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": 2.921875, | |
| "rewards/margins": 1.8203125, | |
| "rewards/rejected": 1.09375, | |
| "step": 78 | |
| }, | |
| { | |
| "epoch": 0.316, | |
| "grad_norm": 10.978973661742327, | |
| "learning_rate": 8.022306170204231e-07, | |
| "logits/chosen": -1.0234375, | |
| "logits/rejected": -1.0, | |
| "logps/chosen": -5.8125, | |
| "logps/rejected": -23.5, | |
| "loss": 1.3188, | |
| "nll_loss": 0.5625, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 3.09375, | |
| "rewards/margins": 2.046875, | |
| "rewards/rejected": 1.0390625, | |
| "step": 79 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "grad_norm": 11.702933301921057, | |
| "learning_rate": 7.970344252406831e-07, | |
| "logits/chosen": -0.9765625, | |
| "logits/rejected": -0.94921875, | |
| "logps/chosen": -5.03125, | |
| "logps/rejected": -18.5, | |
| "loss": 1.3078, | |
| "nll_loss": 0.48828125, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 2.90625, | |
| "rewards/margins": 1.2109375, | |
| "rewards/rejected": 1.6875, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.324, | |
| "grad_norm": 12.12062750517865, | |
| "learning_rate": 7.917881758982836e-07, | |
| "logits/chosen": -0.953125, | |
| "logits/rejected": -0.9375, | |
| "logps/chosen": -6.0, | |
| "logps/rejected": -15.875, | |
| "loss": 1.2015, | |
| "nll_loss": 0.59765625, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 2.65625, | |
| "rewards/margins": 0.796875, | |
| "rewards/rejected": 1.859375, | |
| "step": 81 | |
| }, | |
| { | |
| "epoch": 0.328, | |
| "grad_norm": 14.451380371080663, | |
| "learning_rate": 7.86492753114501e-07, | |
| "logits/chosen": -1.015625, | |
| "logits/rejected": -1.015625, | |
| "logps/chosen": -6.03125, | |
| "logps/rejected": -24.875, | |
| "loss": 1.173, | |
| "nll_loss": 0.55859375, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": 2.859375, | |
| "rewards/margins": 2.078125, | |
| "rewards/rejected": 0.7734375, | |
| "step": 82 | |
| }, | |
| { | |
| "epoch": 0.332, | |
| "grad_norm": 12.677944938939618, | |
| "learning_rate": 7.811490492975389e-07, | |
| "logits/chosen": -0.9609375, | |
| "logits/rejected": -0.953125, | |
| "logps/chosen": -4.5, | |
| "logps/rejected": -20.5, | |
| "loss": 1.2099, | |
| "nll_loss": 0.478515625, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 2.703125, | |
| "rewards/margins": 1.1484375, | |
| "rewards/rejected": 1.5625, | |
| "step": 83 | |
| }, | |
| { | |
| "epoch": 0.336, | |
| "grad_norm": 12.912178140489386, | |
| "learning_rate": 7.757579649921353e-07, | |
| "logits/chosen": -1.0234375, | |
| "logits/rejected": -0.99609375, | |
| "logps/chosen": -4.96875, | |
| "logps/rejected": -26.0, | |
| "loss": 1.2007, | |
| "nll_loss": 0.48828125, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 2.796875, | |
| "rewards/margins": 1.4765625, | |
| "rewards/rejected": 1.328125, | |
| "step": 84 | |
| }, | |
| { | |
| "epoch": 0.34, | |
| "grad_norm": 11.95102680724988, | |
| "learning_rate": 7.703204087277988e-07, | |
| "logits/chosen": -1.0, | |
| "logits/rejected": -0.9921875, | |
| "logps/chosen": -5.125, | |
| "logps/rejected": -20.75, | |
| "loss": 1.1374, | |
| "nll_loss": 0.470703125, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 2.640625, | |
| "rewards/margins": 1.1953125, | |
| "rewards/rejected": 1.4453125, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 0.344, | |
| "grad_norm": 15.312482138498904, | |
| "learning_rate": 7.648372968656993e-07, | |
| "logits/chosen": -0.9921875, | |
| "logits/rejected": -0.99609375, | |
| "logps/chosen": -6.15625, | |
| "logps/rejected": -30.5, | |
| "loss": 1.1599, | |
| "nll_loss": 0.55859375, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 2.515625, | |
| "rewards/margins": 1.765625, | |
| "rewards/rejected": 0.75, | |
| "step": 86 | |
| }, | |
| { | |
| "epoch": 0.348, | |
| "grad_norm": 13.90159120685156, | |
| "learning_rate": 7.593095534442386e-07, | |
| "logits/chosen": -0.9921875, | |
| "logits/rejected": -0.984375, | |
| "logps/chosen": -4.84375, | |
| "logps/rejected": -19.375, | |
| "loss": 1.116, | |
| "nll_loss": 0.52734375, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 2.9375, | |
| "rewards/margins": 1.46875, | |
| "rewards/rejected": 1.46875, | |
| "step": 87 | |
| }, | |
| { | |
| "epoch": 0.352, | |
| "grad_norm": 14.950945255344106, | |
| "learning_rate": 7.537381100233278e-07, | |
| "logits/chosen": -1.0078125, | |
| "logits/rejected": -1.0078125, | |
| "logps/chosen": -5.0625, | |
| "logps/rejected": -23.375, | |
| "loss": 1.2363, | |
| "nll_loss": 0.482421875, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 2.640625, | |
| "rewards/margins": 1.6796875, | |
| "rewards/rejected": 0.9609375, | |
| "step": 88 | |
| }, | |
| { | |
| "epoch": 0.356, | |
| "grad_norm": 12.117659621877523, | |
| "learning_rate": 7.481239055273958e-07, | |
| "logits/chosen": -0.94921875, | |
| "logits/rejected": -0.9609375, | |
| "logps/chosen": -4.625, | |
| "logps/rejected": -12.8125, | |
| "loss": 1.169, | |
| "nll_loss": 0.462890625, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 2.828125, | |
| "rewards/margins": 0.546875, | |
| "rewards/rejected": 2.28125, | |
| "step": 89 | |
| }, | |
| { | |
| "epoch": 0.36, | |
| "grad_norm": 15.946889828294005, | |
| "learning_rate": 7.424678860871584e-07, | |
| "logits/chosen": -1.0078125, | |
| "logits/rejected": -0.984375, | |
| "logps/chosen": -6.125, | |
| "logps/rejected": -28.0, | |
| "loss": 1.185, | |
| "nll_loss": 0.58203125, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": 2.5625, | |
| "rewards/margins": 1.703125, | |
| "rewards/rejected": 0.8515625, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.364, | |
| "grad_norm": 21.734506154061158, | |
| "learning_rate": 7.367710048801715e-07, | |
| "logits/chosen": -1.015625, | |
| "logits/rejected": -1.015625, | |
| "logps/chosen": -3.390625, | |
| "logps/rejected": -19.375, | |
| "loss": 1.1948, | |
| "nll_loss": 0.337890625, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 2.703125, | |
| "rewards/margins": 1.2421875, | |
| "rewards/rejected": 1.453125, | |
| "step": 91 | |
| }, | |
| { | |
| "epoch": 0.368, | |
| "grad_norm": 12.733439669594082, | |
| "learning_rate": 7.31034221970198e-07, | |
| "logits/chosen": -1.03125, | |
| "logits/rejected": -1.015625, | |
| "logps/chosen": -5.3125, | |
| "logps/rejected": -29.625, | |
| "loss": 1.1723, | |
| "nll_loss": 0.55078125, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 2.828125, | |
| "rewards/margins": 2.015625, | |
| "rewards/rejected": 0.81640625, | |
| "step": 92 | |
| }, | |
| { | |
| "epoch": 0.372, | |
| "grad_norm": 13.39954488299055, | |
| "learning_rate": 7.252585041454133e-07, | |
| "logits/chosen": -1.015625, | |
| "logits/rejected": -1.0, | |
| "logps/chosen": -4.4375, | |
| "logps/rejected": -29.5, | |
| "loss": 1.1454, | |
| "nll_loss": 0.443359375, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": 2.6875, | |
| "rewards/margins": 2.15625, | |
| "rewards/rejected": 0.5390625, | |
| "step": 93 | |
| }, | |
| { | |
| "epoch": 0.376, | |
| "grad_norm": 15.64435560107329, | |
| "learning_rate": 7.194448247554778e-07, | |
| "logits/chosen": -1.0, | |
| "logits/rejected": -0.9921875, | |
| "logps/chosen": -5.15625, | |
| "logps/rejected": -20.0, | |
| "loss": 1.1604, | |
| "nll_loss": 0.546875, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 2.8125, | |
| "rewards/margins": 1.1015625, | |
| "rewards/rejected": 1.703125, | |
| "step": 94 | |
| }, | |
| { | |
| "epoch": 0.38, | |
| "grad_norm": 17.13096115933346, | |
| "learning_rate": 7.135941635475036e-07, | |
| "logits/chosen": -1.015625, | |
| "logits/rejected": -1.0078125, | |
| "logps/chosen": -6.8125, | |
| "logps/rejected": -28.5, | |
| "loss": 1.2684, | |
| "nll_loss": 0.6015625, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 2.953125, | |
| "rewards/margins": 1.9921875, | |
| "rewards/rejected": 0.96875, | |
| "step": 95 | |
| }, | |
| { | |
| "epoch": 0.384, | |
| "grad_norm": 16.845492514129948, | |
| "learning_rate": 7.077075065009433e-07, | |
| "logits/chosen": -1.0234375, | |
| "logits/rejected": -0.99609375, | |
| "logps/chosen": -6.0625, | |
| "logps/rejected": -22.625, | |
| "loss": 1.2163, | |
| "nll_loss": 0.5859375, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 2.3125, | |
| "rewards/margins": 1.2265625, | |
| "rewards/rejected": 1.0859375, | |
| "step": 96 | |
| }, | |
| { | |
| "epoch": 0.388, | |
| "grad_norm": 24.19829824906313, | |
| "learning_rate": 7.017858456614283e-07, | |
| "logits/chosen": -1.03125, | |
| "logits/rejected": -1.0078125, | |
| "logps/chosen": -4.8125, | |
| "logps/rejected": -24.125, | |
| "loss": 1.2011, | |
| "nll_loss": 0.46875, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 3.0, | |
| "rewards/margins": 1.75, | |
| "rewards/rejected": 1.2578125, | |
| "step": 97 | |
| }, | |
| { | |
| "epoch": 0.392, | |
| "grad_norm": 16.528677943117593, | |
| "learning_rate": 6.958301789735852e-07, | |
| "logits/chosen": -1.03125, | |
| "logits/rejected": -1.0234375, | |
| "logps/chosen": -5.46875, | |
| "logps/rejected": -26.125, | |
| "loss": 1.2671, | |
| "nll_loss": 0.5546875, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 2.5, | |
| "rewards/margins": 1.8359375, | |
| "rewards/rejected": 0.66015625, | |
| "step": 98 | |
| }, | |
| { | |
| "epoch": 0.396, | |
| "grad_norm": 15.542683127815902, | |
| "learning_rate": 6.898415101128569e-07, | |
| "logits/chosen": -1.0234375, | |
| "logits/rejected": -1.03125, | |
| "logps/chosen": -5.5625, | |
| "logps/rejected": -18.25, | |
| "loss": 1.2051, | |
| "nll_loss": 0.53125, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 2.796875, | |
| "rewards/margins": 1.109375, | |
| "rewards/rejected": 1.6796875, | |
| "step": 99 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "grad_norm": 16.703597419369395, | |
| "learning_rate": 6.8382084831636e-07, | |
| "logits/chosen": -1.0546875, | |
| "logits/rejected": -1.0234375, | |
| "logps/chosen": -4.34375, | |
| "logps/rejected": -30.875, | |
| "loss": 1.1695, | |
| "nll_loss": 0.392578125, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": 2.46875, | |
| "rewards/margins": 2.140625, | |
| "rewards/rejected": 0.333984375, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.404, | |
| "grad_norm": 20.289662822876423, | |
| "learning_rate": 6.777692082128024e-07, | |
| "logits/chosen": -1.046875, | |
| "logits/rejected": -1.015625, | |
| "logps/chosen": -6.9375, | |
| "logps/rejected": -23.125, | |
| "loss": 1.1355, | |
| "nll_loss": 0.6328125, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": 2.84375, | |
| "rewards/margins": 1.734375, | |
| "rewards/rejected": 1.109375, | |
| "step": 101 | |
| }, | |
| { | |
| "epoch": 0.408, | |
| "grad_norm": 10.57908068816297, | |
| "learning_rate": 6.716876096514943e-07, | |
| "logits/chosen": -1.0078125, | |
| "logits/rejected": -1.0234375, | |
| "logps/chosen": -3.34375, | |
| "logps/rejected": -35.25, | |
| "loss": 1.1203, | |
| "nll_loss": 0.30078125, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": 2.75, | |
| "rewards/margins": 2.953125, | |
| "rewards/rejected": -0.2001953125, | |
| "step": 102 | |
| }, | |
| { | |
| "epoch": 0.412, | |
| "grad_norm": 12.448311996517877, | |
| "learning_rate": 6.655770775304782e-07, | |
| "logits/chosen": -0.98828125, | |
| "logits/rejected": -0.98828125, | |
| "logps/chosen": -5.5, | |
| "logps/rejected": -13.8125, | |
| "loss": 1.2512, | |
| "nll_loss": 0.5625, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 2.90625, | |
| "rewards/margins": 0.66015625, | |
| "rewards/rejected": 2.25, | |
| "step": 103 | |
| }, | |
| { | |
| "epoch": 0.416, | |
| "grad_norm": 19.442937641890275, | |
| "learning_rate": 6.594386416238094e-07, | |
| "logits/chosen": -0.98046875, | |
| "logits/rejected": -0.98828125, | |
| "logps/chosen": -8.0625, | |
| "logps/rejected": -20.0, | |
| "loss": 1.1029, | |
| "nll_loss": 0.75, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 2.546875, | |
| "rewards/margins": 1.3984375, | |
| "rewards/rejected": 1.1484375, | |
| "step": 104 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "grad_norm": 17.597984347295238, | |
| "learning_rate": 6.532733364080125e-07, | |
| "logits/chosen": -0.98828125, | |
| "logits/rejected": -0.97265625, | |
| "logps/chosen": -5.53125, | |
| "logps/rejected": -26.125, | |
| "loss": 1.0465, | |
| "nll_loss": 0.5078125, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 2.78125, | |
| "rewards/margins": 1.84375, | |
| "rewards/rejected": 0.9375, | |
| "step": 105 | |
| }, | |
| { | |
| "epoch": 0.424, | |
| "grad_norm": 42.65485477575747, | |
| "learning_rate": 6.470822008877482e-07, | |
| "logits/chosen": -1.0, | |
| "logits/rejected": -0.98828125, | |
| "logps/chosen": -5.9375, | |
| "logps/rejected": -19.125, | |
| "loss": 1.1754, | |
| "nll_loss": 0.53515625, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 2.875, | |
| "rewards/margins": 1.390625, | |
| "rewards/rejected": 1.4921875, | |
| "step": 106 | |
| }, | |
| { | |
| "epoch": 0.428, | |
| "grad_norm": 14.418794180437754, | |
| "learning_rate": 6.408662784207149e-07, | |
| "logits/chosen": -1.0078125, | |
| "logits/rejected": -1.0, | |
| "logps/chosen": -4.875, | |
| "logps/rejected": -28.375, | |
| "loss": 1.1556, | |
| "nll_loss": 0.51171875, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 2.71875, | |
| "rewards/margins": 1.875, | |
| "rewards/rejected": 0.84375, | |
| "step": 107 | |
| }, | |
| { | |
| "epoch": 0.432, | |
| "grad_norm": 24.069184880501325, | |
| "learning_rate": 6.346266165418172e-07, | |
| "logits/chosen": -0.9765625, | |
| "logits/rejected": -0.95703125, | |
| "logps/chosen": -4.71875, | |
| "logps/rejected": -23.0, | |
| "loss": 1.1594, | |
| "nll_loss": 0.48046875, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": 3.15625, | |
| "rewards/margins": 1.875, | |
| "rewards/rejected": 1.2734375, | |
| "step": 108 | |
| }, | |
| { | |
| "epoch": 0.436, | |
| "grad_norm": 33.99711892618498, | |
| "learning_rate": 6.283642667866315e-07, | |
| "logits/chosen": -1.03125, | |
| "logits/rejected": -0.98046875, | |
| "logps/chosen": -6.28125, | |
| "logps/rejected": -27.375, | |
| "loss": 1.1293, | |
| "nll_loss": 0.58984375, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 2.5625, | |
| "rewards/margins": 1.625, | |
| "rewards/rejected": 0.9375, | |
| "step": 109 | |
| }, | |
| { | |
| "epoch": 0.44, | |
| "grad_norm": 35.15701910686572, | |
| "learning_rate": 6.220802845141957e-07, | |
| "logits/chosen": -1.03125, | |
| "logits/rejected": -1.03125, | |
| "logps/chosen": -5.71875, | |
| "logps/rejected": -37.25, | |
| "loss": 1.1224, | |
| "nll_loss": 0.51953125, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 2.578125, | |
| "rewards/margins": 2.609375, | |
| "rewards/rejected": -0.03564453125, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.444, | |
| "grad_norm": 16.617005010321233, | |
| "learning_rate": 6.157757287291556e-07, | |
| "logits/chosen": -1.015625, | |
| "logits/rejected": -1.015625, | |
| "logps/chosen": -3.859375, | |
| "logps/rejected": -31.75, | |
| "loss": 1.0481, | |
| "nll_loss": 0.40234375, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": 2.6875, | |
| "rewards/margins": 2.6875, | |
| "rewards/rejected": 0.0074462890625, | |
| "step": 111 | |
| }, | |
| { | |
| "epoch": 0.448, | |
| "grad_norm": 12.591607226736722, | |
| "learning_rate": 6.094516619032975e-07, | |
| "logits/chosen": -1.046875, | |
| "logits/rejected": -1.0390625, | |
| "logps/chosen": -5.125, | |
| "logps/rejected": -26.5, | |
| "loss": 1.0584, | |
| "nll_loss": 0.51171875, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 2.71875, | |
| "rewards/margins": 1.90625, | |
| "rewards/rejected": 0.82421875, | |
| "step": 112 | |
| }, | |
| { | |
| "epoch": 0.452, | |
| "grad_norm": 12.509284400507388, | |
| "learning_rate": 6.03109149796494e-07, | |
| "logits/chosen": -1.0546875, | |
| "logits/rejected": -1.046875, | |
| "logps/chosen": -3.984375, | |
| "logps/rejected": -23.0, | |
| "loss": 1.0653, | |
| "nll_loss": 0.390625, | |
| "rewards/accuracies": 0.90625, | |
| "rewards/chosen": 2.859375, | |
| "rewards/margins": 1.7890625, | |
| "rewards/rejected": 1.0625, | |
| "step": 113 | |
| }, | |
| { | |
| "epoch": 0.456, | |
| "grad_norm": 33.81401585194403, | |
| "learning_rate": 5.967492612770997e-07, | |
| "logits/chosen": -1.03125, | |
| "logits/rejected": -1.015625, | |
| "logps/chosen": -6.65625, | |
| "logps/rejected": -28.125, | |
| "loss": 1.0973, | |
| "nll_loss": 0.6640625, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 2.625, | |
| "rewards/margins": 2.015625, | |
| "rewards/rejected": 0.6015625, | |
| "step": 114 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "grad_norm": 15.815532311957057, | |
| "learning_rate": 5.90373068141819e-07, | |
| "logits/chosen": -1.0, | |
| "logits/rejected": -0.9765625, | |
| "logps/chosen": -6.875, | |
| "logps/rejected": -29.125, | |
| "loss": 1.2058, | |
| "nll_loss": 0.69140625, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 2.390625, | |
| "rewards/margins": 2.015625, | |
| "rewards/rejected": 0.375, | |
| "step": 115 | |
| }, | |
| { | |
| "epoch": 0.464, | |
| "grad_norm": 22.102704341248568, | |
| "learning_rate": 5.839816449350824e-07, | |
| "logits/chosen": -0.9921875, | |
| "logits/rejected": -0.9921875, | |
| "logps/chosen": -5.78125, | |
| "logps/rejected": -23.625, | |
| "loss": 1.1099, | |
| "nll_loss": 0.5703125, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 3.015625, | |
| "rewards/margins": 1.5390625, | |
| "rewards/rejected": 1.46875, | |
| "step": 116 | |
| }, | |
| { | |
| "epoch": 0.468, | |
| "grad_norm": 16.24899670304244, | |
| "learning_rate": 5.775760687679603e-07, | |
| "logits/chosen": -1.0078125, | |
| "logits/rejected": -0.984375, | |
| "logps/chosen": -4.28125, | |
| "logps/rejected": -23.125, | |
| "loss": 1.0259, | |
| "nll_loss": 0.41015625, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 2.78125, | |
| "rewards/margins": 1.765625, | |
| "rewards/rejected": 1.0234375, | |
| "step": 117 | |
| }, | |
| { | |
| "epoch": 0.472, | |
| "grad_norm": 14.705074940426503, | |
| "learning_rate": 5.711574191366427e-07, | |
| "logits/chosen": -1.0234375, | |
| "logits/rejected": -1.0234375, | |
| "logps/chosen": -3.203125, | |
| "logps/rejected": -25.875, | |
| "loss": 1.1279, | |
| "nll_loss": 0.330078125, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 3.0, | |
| "rewards/margins": 1.796875, | |
| "rewards/rejected": 1.1953125, | |
| "step": 118 | |
| }, | |
| { | |
| "epoch": 0.476, | |
| "grad_norm": 20.221686205664327, | |
| "learning_rate": 5.647267777405177e-07, | |
| "logits/chosen": -1.0234375, | |
| "logits/rejected": -1.03125, | |
| "logps/chosen": -4.40625, | |
| "logps/rejected": -23.375, | |
| "loss": 1.1415, | |
| "nll_loss": 0.478515625, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 2.578125, | |
| "rewards/margins": 1.3984375, | |
| "rewards/rejected": 1.1796875, | |
| "step": 119 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "grad_norm": 24.368984032622865, | |
| "learning_rate": 5.582852282998795e-07, | |
| "logits/chosen": -1.0234375, | |
| "logits/rejected": -1.03125, | |
| "logps/chosen": -5.40625, | |
| "logps/rejected": -23.125, | |
| "loss": 1.0667, | |
| "nll_loss": 0.474609375, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 2.84375, | |
| "rewards/margins": 1.546875, | |
| "rewards/rejected": 1.296875, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.484, | |
| "grad_norm": 22.432812053652427, | |
| "learning_rate": 5.518338563732944e-07, | |
| "logits/chosen": -1.046875, | |
| "logits/rejected": -1.0390625, | |
| "logps/chosen": -4.90625, | |
| "logps/rejected": -23.625, | |
| "loss": 1.0827, | |
| "nll_loss": 0.54296875, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": 2.546875, | |
| "rewards/margins": 1.6484375, | |
| "rewards/rejected": 0.90625, | |
| "step": 121 | |
| }, | |
| { | |
| "epoch": 0.488, | |
| "grad_norm": 16.17860106216787, | |
| "learning_rate": 5.453737491746572e-07, | |
| "logits/chosen": -1.0234375, | |
| "logits/rejected": -1.03125, | |
| "logps/chosen": -3.625, | |
| "logps/rejected": -22.875, | |
| "loss": 1.0798, | |
| "nll_loss": 0.375, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 3.046875, | |
| "rewards/margins": 1.609375, | |
| "rewards/rejected": 1.4296875, | |
| "step": 122 | |
| }, | |
| { | |
| "epoch": 0.492, | |
| "grad_norm": 24.70109578919281, | |
| "learning_rate": 5.389059953899698e-07, | |
| "logits/chosen": -1.09375, | |
| "logits/rejected": -1.0703125, | |
| "logps/chosen": -4.53125, | |
| "logps/rejected": -31.375, | |
| "loss": 1.2218, | |
| "nll_loss": 0.451171875, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": 3.109375, | |
| "rewards/margins": 2.34375, | |
| "rewards/rejected": 0.765625, | |
| "step": 123 | |
| }, | |
| { | |
| "epoch": 0.496, | |
| "grad_norm": 31.00791322114362, | |
| "learning_rate": 5.324316849938714e-07, | |
| "logits/chosen": -0.98046875, | |
| "logits/rejected": -0.97265625, | |
| "logps/chosen": -6.375, | |
| "logps/rejected": -21.75, | |
| "loss": 1.166, | |
| "nll_loss": 0.57421875, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 2.71875, | |
| "rewards/margins": 1.5703125, | |
| "rewards/rejected": 1.15625, | |
| "step": 124 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "grad_norm": 37.45373935605177, | |
| "learning_rate": 5.259519090659498e-07, | |
| "logits/chosen": -1.0546875, | |
| "logits/rejected": -1.046875, | |
| "logps/chosen": -4.40625, | |
| "logps/rejected": -25.5, | |
| "loss": 1.0699, | |
| "nll_loss": 0.392578125, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 2.890625, | |
| "rewards/margins": 1.703125, | |
| "rewards/rejected": 1.1953125, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 0.504, | |
| "grad_norm": 12.712491511773225, | |
| "learning_rate": 5.194677596068688e-07, | |
| "logits/chosen": -1.0078125, | |
| "logits/rejected": -1.03125, | |
| "logps/chosen": -5.84375, | |
| "logps/rejected": -21.375, | |
| "loss": 1.1111, | |
| "nll_loss": 0.55859375, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 2.65625, | |
| "rewards/margins": 1.1328125, | |
| "rewards/rejected": 1.5234375, | |
| "step": 126 | |
| }, | |
| { | |
| "epoch": 0.508, | |
| "grad_norm": 14.71397944686033, | |
| "learning_rate": 5.129803293543391e-07, | |
| "logits/chosen": -1.0546875, | |
| "logits/rejected": -1.046875, | |
| "logps/chosen": -4.375, | |
| "logps/rejected": -28.875, | |
| "loss": 1.0949, | |
| "nll_loss": 0.41015625, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": 2.984375, | |
| "rewards/margins": 2.390625, | |
| "rewards/rejected": 0.58984375, | |
| "step": 127 | |
| }, | |
| { | |
| "epoch": 0.512, | |
| "grad_norm": 20.40214155927554, | |
| "learning_rate": 5.064907115989654e-07, | |
| "logits/chosen": -1.078125, | |
| "logits/rejected": -1.09375, | |
| "logps/chosen": -5.375, | |
| "logps/rejected": -30.25, | |
| "loss": 1.0997, | |
| "nll_loss": 0.46484375, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": 2.984375, | |
| "rewards/margins": 2.84375, | |
| "rewards/rejected": 0.1357421875, | |
| "step": 128 | |
| }, | |
| { | |
| "epoch": 0.516, | |
| "grad_norm": 20.076566430152866, | |
| "learning_rate": 5e-07, | |
| "logits/chosen": -1.0546875, | |
| "logits/rejected": -1.0546875, | |
| "logps/chosen": -5.84375, | |
| "logps/rejected": -31.5, | |
| "loss": 1.1843, | |
| "nll_loss": 0.5703125, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 2.484375, | |
| "rewards/margins": 1.984375, | |
| "rewards/rejected": 0.5078125, | |
| "step": 129 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "grad_norm": 15.69512072569519, | |
| "learning_rate": 4.935092884010346e-07, | |
| "logits/chosen": -1.0234375, | |
| "logits/rejected": -1.015625, | |
| "logps/chosen": -4.6875, | |
| "logps/rejected": -24.75, | |
| "loss": 1.0591, | |
| "nll_loss": 0.4609375, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": 2.84375, | |
| "rewards/margins": 1.8828125, | |
| "rewards/rejected": 0.96484375, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.524, | |
| "grad_norm": 15.184658306188565, | |
| "learning_rate": 4.870196706456609e-07, | |
| "logits/chosen": -1.0625, | |
| "logits/rejected": -1.09375, | |
| "logps/chosen": -5.90625, | |
| "logps/rejected": -29.5, | |
| "loss": 1.1098, | |
| "nll_loss": 0.58203125, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 2.828125, | |
| "rewards/margins": 1.7109375, | |
| "rewards/rejected": 1.1171875, | |
| "step": 131 | |
| }, | |
| { | |
| "epoch": 0.528, | |
| "grad_norm": 30.109778759207604, | |
| "learning_rate": 4.805322403931311e-07, | |
| "logits/chosen": -1.109375, | |
| "logits/rejected": -1.1015625, | |
| "logps/chosen": -4.78125, | |
| "logps/rejected": -28.25, | |
| "loss": 1.0583, | |
| "nll_loss": 0.41796875, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": 3.078125, | |
| "rewards/margins": 2.765625, | |
| "rewards/rejected": 0.30859375, | |
| "step": 132 | |
| }, | |
| { | |
| "epoch": 0.532, | |
| "grad_norm": 19.180407490964914, | |
| "learning_rate": 4.7404809093405017e-07, | |
| "logits/chosen": -1.0703125, | |
| "logits/rejected": -1.0703125, | |
| "logps/chosen": -4.5625, | |
| "logps/rejected": -24.5, | |
| "loss": 1.0781, | |
| "nll_loss": 0.3984375, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 2.6875, | |
| "rewards/margins": 1.75, | |
| "rewards/rejected": 0.9375, | |
| "step": 133 | |
| }, | |
| { | |
| "epoch": 0.536, | |
| "grad_norm": 11.865585536243879, | |
| "learning_rate": 4.675683150061284e-07, | |
| "logits/chosen": -1.078125, | |
| "logits/rejected": -1.0390625, | |
| "logps/chosen": -4.65625, | |
| "logps/rejected": -27.75, | |
| "loss": 1.1264, | |
| "nll_loss": 0.470703125, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": 3.140625, | |
| "rewards/margins": 1.9375, | |
| "rewards/rejected": 1.203125, | |
| "step": 134 | |
| }, | |
| { | |
| "epoch": 0.54, | |
| "grad_norm": 14.216123496302846, | |
| "learning_rate": 4.6109400461003003e-07, | |
| "logits/chosen": -1.0234375, | |
| "logits/rejected": -1.0234375, | |
| "logps/chosen": -5.34375, | |
| "logps/rejected": -23.5, | |
| "loss": 1.0383, | |
| "nll_loss": 0.453125, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 3.3125, | |
| "rewards/margins": 2.25, | |
| "rewards/rejected": 1.0703125, | |
| "step": 135 | |
| }, | |
| { | |
| "epoch": 0.544, | |
| "grad_norm": 28.567567684765535, | |
| "learning_rate": 4.5462625082534286e-07, | |
| "logits/chosen": -1.078125, | |
| "logits/rejected": -1.0703125, | |
| "logps/chosen": -3.9375, | |
| "logps/rejected": -28.125, | |
| "loss": 0.9913, | |
| "nll_loss": 0.388671875, | |
| "rewards/accuracies": 0.90625, | |
| "rewards/chosen": 2.875, | |
| "rewards/margins": 2.53125, | |
| "rewards/rejected": 0.353515625, | |
| "step": 136 | |
| }, | |
| { | |
| "epoch": 0.548, | |
| "grad_norm": 19.382351209373788, | |
| "learning_rate": 4.481661436267057e-07, | |
| "logits/chosen": -1.0703125, | |
| "logits/rejected": -1.078125, | |
| "logps/chosen": -4.65625, | |
| "logps/rejected": -33.5, | |
| "loss": 1.1359, | |
| "nll_loss": 0.42578125, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": 2.828125, | |
| "rewards/margins": 2.328125, | |
| "rewards/rejected": 0.5078125, | |
| "step": 137 | |
| }, | |
| { | |
| "epoch": 0.552, | |
| "grad_norm": 28.26009731578296, | |
| "learning_rate": 4.4171477170012043e-07, | |
| "logits/chosen": -1.0546875, | |
| "logits/rejected": -1.046875, | |
| "logps/chosen": -4.21875, | |
| "logps/rejected": -32.0, | |
| "loss": 1.1338, | |
| "nll_loss": 0.380859375, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": 2.953125, | |
| "rewards/margins": 2.578125, | |
| "rewards/rejected": 0.37890625, | |
| "step": 138 | |
| }, | |
| { | |
| "epoch": 0.556, | |
| "grad_norm": 48.13783869940889, | |
| "learning_rate": 4.352732222594823e-07, | |
| "logits/chosen": -1.046875, | |
| "logits/rejected": -1.03125, | |
| "logps/chosen": -5.1875, | |
| "logps/rejected": -25.625, | |
| "loss": 1.1776, | |
| "nll_loss": 0.466796875, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 2.859375, | |
| "rewards/margins": 1.8515625, | |
| "rewards/rejected": 1.0078125, | |
| "step": 139 | |
| }, | |
| { | |
| "epoch": 0.56, | |
| "grad_norm": 23.304268399995618, | |
| "learning_rate": 4.2884258086335745e-07, | |
| "logits/chosen": -1.0390625, | |
| "logits/rejected": -1.0390625, | |
| "logps/chosen": -5.65625, | |
| "logps/rejected": -18.5, | |
| "loss": 1.1203, | |
| "nll_loss": 0.490234375, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 3.03125, | |
| "rewards/margins": 1.421875, | |
| "rewards/rejected": 1.6015625, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.564, | |
| "grad_norm": 19.29343324391366, | |
| "learning_rate": 4.224239312320398e-07, | |
| "logits/chosen": -1.078125, | |
| "logits/rejected": -1.0703125, | |
| "logps/chosen": -7.90625, | |
| "logps/rejected": -29.375, | |
| "loss": 1.1849, | |
| "nll_loss": 0.67578125, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 2.5, | |
| "rewards/margins": 1.9375, | |
| "rewards/rejected": 0.5625, | |
| "step": 141 | |
| }, | |
| { | |
| "epoch": 0.568, | |
| "grad_norm": 17.35581169203161, | |
| "learning_rate": 4.160183550649175e-07, | |
| "logits/chosen": -1.0546875, | |
| "logits/rejected": -1.0546875, | |
| "logps/chosen": -4.65625, | |
| "logps/rejected": -28.25, | |
| "loss": 1.0433, | |
| "nll_loss": 0.48046875, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": 2.796875, | |
| "rewards/margins": 2.03125, | |
| "rewards/rejected": 0.76953125, | |
| "step": 142 | |
| }, | |
| { | |
| "epoch": 0.572, | |
| "grad_norm": 31.43926475473609, | |
| "learning_rate": 4.09626931858181e-07, | |
| "logits/chosen": -1.0625, | |
| "logits/rejected": -1.0546875, | |
| "logps/chosen": -4.21875, | |
| "logps/rejected": -19.125, | |
| "loss": 1.112, | |
| "nll_loss": 0.41015625, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 2.96875, | |
| "rewards/margins": 1.3515625, | |
| "rewards/rejected": 1.625, | |
| "step": 143 | |
| }, | |
| { | |
| "epoch": 0.576, | |
| "grad_norm": 19.985361598611938, | |
| "learning_rate": 4.0325073872290015e-07, | |
| "logits/chosen": -1.0234375, | |
| "logits/rejected": -1.015625, | |
| "logps/chosen": -6.6875, | |
| "logps/rejected": -26.375, | |
| "loss": 1.1101, | |
| "nll_loss": 0.5859375, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 2.859375, | |
| "rewards/margins": 2.328125, | |
| "rewards/rejected": 0.52734375, | |
| "step": 144 | |
| }, | |
| { | |
| "epoch": 0.58, | |
| "grad_norm": 35.012979635813075, | |
| "learning_rate": 3.96890850203506e-07, | |
| "logits/chosen": -1.0625, | |
| "logits/rejected": -1.0625, | |
| "logps/chosen": -6.375, | |
| "logps/rejected": -26.625, | |
| "loss": 1.0092, | |
| "nll_loss": 0.59765625, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": 3.21875, | |
| "rewards/margins": 2.265625, | |
| "rewards/rejected": 0.94921875, | |
| "step": 145 | |
| }, | |
| { | |
| "epoch": 0.584, | |
| "grad_norm": 20.26642224735176, | |
| "learning_rate": 3.9054833809670265e-07, | |
| "logits/chosen": -1.0546875, | |
| "logits/rejected": -1.0546875, | |
| "logps/chosen": -4.625, | |
| "logps/rejected": -29.5, | |
| "loss": 1.0404, | |
| "nll_loss": 0.4140625, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 2.75, | |
| "rewards/margins": 2.5, | |
| "rewards/rejected": 0.2392578125, | |
| "step": 146 | |
| }, | |
| { | |
| "epoch": 0.588, | |
| "grad_norm": 13.154350173457996, | |
| "learning_rate": 3.8422427127084436e-07, | |
| "logits/chosen": -1.078125, | |
| "logits/rejected": -1.0625, | |
| "logps/chosen": -6.53125, | |
| "logps/rejected": -31.875, | |
| "loss": 0.9968, | |
| "nll_loss": 0.58203125, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": 2.921875, | |
| "rewards/margins": 2.59375, | |
| "rewards/rejected": 0.32421875, | |
| "step": 147 | |
| }, | |
| { | |
| "epoch": 0.592, | |
| "grad_norm": 23.129629439239768, | |
| "learning_rate": 3.779197154858044e-07, | |
| "logits/chosen": -1.0546875, | |
| "logits/rejected": -1.0390625, | |
| "logps/chosen": -5.21875, | |
| "logps/rejected": -23.375, | |
| "loss": 0.9492, | |
| "nll_loss": 0.4921875, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": 3.09375, | |
| "rewards/margins": 1.7265625, | |
| "rewards/rejected": 1.3671875, | |
| "step": 148 | |
| }, | |
| { | |
| "epoch": 0.596, | |
| "grad_norm": 14.081159200332113, | |
| "learning_rate": 3.716357332133686e-07, | |
| "logits/chosen": -1.1171875, | |
| "logits/rejected": -1.078125, | |
| "logps/chosen": -5.65625, | |
| "logps/rejected": -30.5, | |
| "loss": 1.0858, | |
| "nll_loss": 0.58984375, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 2.828125, | |
| "rewards/margins": 2.0625, | |
| "rewards/rejected": 0.76953125, | |
| "step": 149 | |
| }, | |
| { | |
| "epoch": 0.6, | |
| "grad_norm": 30.59966116072142, | |
| "learning_rate": 3.653733834581827e-07, | |
| "logits/chosen": -1.09375, | |
| "logits/rejected": -1.078125, | |
| "logps/chosen": -4.0, | |
| "logps/rejected": -26.125, | |
| "loss": 1.0947, | |
| "nll_loss": 0.373046875, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": 2.765625, | |
| "rewards/margins": 2.21875, | |
| "rewards/rejected": 0.546875, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.604, | |
| "grad_norm": 14.36839336730757, | |
| "learning_rate": 3.591337215792851e-07, | |
| "logits/chosen": -1.09375, | |
| "logits/rejected": -1.109375, | |
| "logps/chosen": -4.5625, | |
| "logps/rejected": -27.5, | |
| "loss": 1.085, | |
| "nll_loss": 0.45703125, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 2.90625, | |
| "rewards/margins": 1.8984375, | |
| "rewards/rejected": 1.0, | |
| "step": 151 | |
| }, | |
| { | |
| "epoch": 0.608, | |
| "grad_norm": 25.95175246611989, | |
| "learning_rate": 3.5291779911225185e-07, | |
| "logits/chosen": -1.0859375, | |
| "logits/rejected": -1.0859375, | |
| "logps/chosen": -4.5, | |
| "logps/rejected": -30.75, | |
| "loss": 1.0704, | |
| "nll_loss": 0.427734375, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": 2.984375, | |
| "rewards/margins": 2.265625, | |
| "rewards/rejected": 0.71484375, | |
| "step": 152 | |
| }, | |
| { | |
| "epoch": 0.612, | |
| "grad_norm": 30.341569831825527, | |
| "learning_rate": 3.4672666359198757e-07, | |
| "logits/chosen": -1.078125, | |
| "logits/rejected": -1.0625, | |
| "logps/chosen": -4.21875, | |
| "logps/rejected": -27.75, | |
| "loss": 1.0871, | |
| "nll_loss": 0.453125, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 3.265625, | |
| "rewards/margins": 2.359375, | |
| "rewards/rejected": 0.91015625, | |
| "step": 153 | |
| }, | |
| { | |
| "epoch": 0.616, | |
| "grad_norm": 15.724055467705009, | |
| "learning_rate": 3.405613583761907e-07, | |
| "logits/chosen": -1.0546875, | |
| "logits/rejected": -1.0390625, | |
| "logps/chosen": -5.53125, | |
| "logps/rejected": -22.25, | |
| "loss": 1.0548, | |
| "nll_loss": 0.5625, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 3.078125, | |
| "rewards/margins": 1.5625, | |
| "rewards/rejected": 1.5234375, | |
| "step": 154 | |
| }, | |
| { | |
| "epoch": 0.62, | |
| "grad_norm": 23.186267615438663, | |
| "learning_rate": 3.3442292246952186e-07, | |
| "logits/chosen": -1.109375, | |
| "logits/rejected": -1.0859375, | |
| "logps/chosen": -5.71875, | |
| "logps/rejected": -33.75, | |
| "loss": 1.0691, | |
| "nll_loss": 0.5390625, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 2.5, | |
| "rewards/margins": 2.390625, | |
| "rewards/rejected": 0.10595703125, | |
| "step": 155 | |
| }, | |
| { | |
| "epoch": 0.624, | |
| "grad_norm": 19.08999867823956, | |
| "learning_rate": 3.283123903485059e-07, | |
| "logits/chosen": -1.0546875, | |
| "logits/rejected": -1.0390625, | |
| "logps/chosen": -6.15625, | |
| "logps/rejected": -28.625, | |
| "loss": 1.1063, | |
| "nll_loss": 0.6015625, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 3.046875, | |
| "rewards/margins": 2.03125, | |
| "rewards/rejected": 1.015625, | |
| "step": 156 | |
| }, | |
| { | |
| "epoch": 0.628, | |
| "grad_norm": 24.530260893855512, | |
| "learning_rate": 3.222307917871977e-07, | |
| "logits/chosen": -1.1015625, | |
| "logits/rejected": -1.078125, | |
| "logps/chosen": -5.25, | |
| "logps/rejected": -36.0, | |
| "loss": 1.1445, | |
| "nll_loss": 0.46875, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": 2.53125, | |
| "rewards/margins": 2.90625, | |
| "rewards/rejected": -0.376953125, | |
| "step": 157 | |
| }, | |
| { | |
| "epoch": 0.632, | |
| "grad_norm": 42.24994864204084, | |
| "learning_rate": 3.161791516836399e-07, | |
| "logits/chosen": -1.078125, | |
| "logits/rejected": -1.0859375, | |
| "logps/chosen": -6.28125, | |
| "logps/rejected": -32.25, | |
| "loss": 1.0581, | |
| "nll_loss": 0.55859375, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 2.8125, | |
| "rewards/margins": 2.328125, | |
| "rewards/rejected": 0.484375, | |
| "step": 158 | |
| }, | |
| { | |
| "epoch": 0.636, | |
| "grad_norm": 14.715119673174955, | |
| "learning_rate": 3.1015848988714307e-07, | |
| "logits/chosen": -1.0546875, | |
| "logits/rejected": -1.0390625, | |
| "logps/chosen": -6.5, | |
| "logps/rejected": -33.75, | |
| "loss": 0.9939, | |
| "nll_loss": 0.609375, | |
| "rewards/accuracies": 0.90625, | |
| "rewards/chosen": 3.203125, | |
| "rewards/margins": 2.90625, | |
| "rewards/rejected": 0.296875, | |
| "step": 159 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "grad_norm": 20.14211721361461, | |
| "learning_rate": 3.0416982102641487e-07, | |
| "logits/chosen": -1.0625, | |
| "logits/rejected": -1.0625, | |
| "logps/chosen": -5.625, | |
| "logps/rejected": -27.25, | |
| "loss": 1.0876, | |
| "nll_loss": 0.5703125, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": 3.296875, | |
| "rewards/margins": 2.390625, | |
| "rewards/rejected": 0.8984375, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.644, | |
| "grad_norm": 15.676197977531485, | |
| "learning_rate": 2.982141543385717e-07, | |
| "logits/chosen": -1.0703125, | |
| "logits/rejected": -1.0625, | |
| "logps/chosen": -4.75, | |
| "logps/rejected": -25.25, | |
| "loss": 1.0416, | |
| "nll_loss": 0.490234375, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 2.953125, | |
| "rewards/margins": 1.8515625, | |
| "rewards/rejected": 1.1015625, | |
| "step": 161 | |
| }, | |
| { | |
| "epoch": 0.648, | |
| "grad_norm": 18.922807692403683, | |
| "learning_rate": 2.922924934990568e-07, | |
| "logits/chosen": -1.03125, | |
| "logits/rejected": -1.0390625, | |
| "logps/chosen": -4.21875, | |
| "logps/rejected": -28.25, | |
| "loss": 1.0214, | |
| "nll_loss": 0.427734375, | |
| "rewards/accuracies": 0.90625, | |
| "rewards/chosen": 2.578125, | |
| "rewards/margins": 2.1875, | |
| "rewards/rejected": 0.3984375, | |
| "step": 162 | |
| }, | |
| { | |
| "epoch": 0.652, | |
| "grad_norm": 15.211250498950244, | |
| "learning_rate": 2.864058364524964e-07, | |
| "logits/chosen": -1.0625, | |
| "logits/rejected": -1.03125, | |
| "logps/chosen": -3.875, | |
| "logps/rejected": -26.625, | |
| "loss": 1.0185, | |
| "nll_loss": 0.39453125, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": 2.984375, | |
| "rewards/margins": 2.453125, | |
| "rewards/rejected": 0.53515625, | |
| "step": 163 | |
| }, | |
| { | |
| "epoch": 0.656, | |
| "grad_norm": 32.56527905487464, | |
| "learning_rate": 2.8055517524452215e-07, | |
| "logits/chosen": -1.1171875, | |
| "logits/rejected": -1.09375, | |
| "logps/chosen": -4.875, | |
| "logps/rejected": -27.125, | |
| "loss": 0.9807, | |
| "nll_loss": 0.4453125, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 2.6875, | |
| "rewards/margins": 2.09375, | |
| "rewards/rejected": 0.58984375, | |
| "step": 164 | |
| }, | |
| { | |
| "epoch": 0.66, | |
| "grad_norm": 10.018587276474245, | |
| "learning_rate": 2.7474149585458664e-07, | |
| "logits/chosen": -1.09375, | |
| "logits/rejected": -1.1171875, | |
| "logps/chosen": -3.390625, | |
| "logps/rejected": -32.75, | |
| "loss": 1.0211, | |
| "nll_loss": 0.310546875, | |
| "rewards/accuracies": 0.90625, | |
| "rewards/chosen": 2.484375, | |
| "rewards/margins": 2.421875, | |
| "rewards/rejected": 0.0595703125, | |
| "step": 165 | |
| }, | |
| { | |
| "epoch": 0.664, | |
| "grad_norm": 37.10528373181994, | |
| "learning_rate": 2.6896577802980184e-07, | |
| "logits/chosen": -1.078125, | |
| "logits/rejected": -1.078125, | |
| "logps/chosen": -3.296875, | |
| "logps/rejected": -33.25, | |
| "loss": 1.0286, | |
| "nll_loss": 0.310546875, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": 2.734375, | |
| "rewards/margins": 2.71875, | |
| "rewards/rejected": 0.0106201171875, | |
| "step": 166 | |
| }, | |
| { | |
| "epoch": 0.668, | |
| "grad_norm": 17.20600574229811, | |
| "learning_rate": 2.632289951198285e-07, | |
| "logits/chosen": -1.0859375, | |
| "logits/rejected": -1.109375, | |
| "logps/chosen": -5.1875, | |
| "logps/rejected": -30.375, | |
| "loss": 1.1248, | |
| "nll_loss": 0.47265625, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 3.078125, | |
| "rewards/margins": 2.46875, | |
| "rewards/rejected": 0.609375, | |
| "step": 167 | |
| }, | |
| { | |
| "epoch": 0.672, | |
| "grad_norm": 38.3608360818198, | |
| "learning_rate": 2.575321139128417e-07, | |
| "logits/chosen": -1.1171875, | |
| "logits/rejected": -1.1015625, | |
| "logps/chosen": -5.25, | |
| "logps/rejected": -26.75, | |
| "loss": 0.9945, | |
| "nll_loss": 0.49609375, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 2.984375, | |
| "rewards/margins": 1.9765625, | |
| "rewards/rejected": 1.0, | |
| "step": 168 | |
| }, | |
| { | |
| "epoch": 0.676, | |
| "grad_norm": 17.759279183503754, | |
| "learning_rate": 2.5187609447260417e-07, | |
| "logits/chosen": -1.1015625, | |
| "logits/rejected": -1.09375, | |
| "logps/chosen": -4.375, | |
| "logps/rejected": -30.25, | |
| "loss": 1.0025, | |
| "nll_loss": 0.439453125, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": 2.96875, | |
| "rewards/margins": 2.625, | |
| "rewards/rejected": 0.3515625, | |
| "step": 169 | |
| }, | |
| { | |
| "epoch": 0.68, | |
| "grad_norm": 14.507361507718214, | |
| "learning_rate": 2.462618899766722e-07, | |
| "logits/chosen": -1.0546875, | |
| "logits/rejected": -1.078125, | |
| "logps/chosen": -6.21875, | |
| "logps/rejected": -32.75, | |
| "loss": 1.0907, | |
| "nll_loss": 0.60546875, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 2.359375, | |
| "rewards/margins": 2.078125, | |
| "rewards/rejected": 0.2734375, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.684, | |
| "grad_norm": 22.962611189789666, | |
| "learning_rate": 2.406904465557614e-07, | |
| "logits/chosen": -1.109375, | |
| "logits/rejected": -1.1171875, | |
| "logps/chosen": -4.5625, | |
| "logps/rejected": -28.25, | |
| "loss": 1.0443, | |
| "nll_loss": 0.4296875, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": 2.671875, | |
| "rewards/margins": 1.859375, | |
| "rewards/rejected": 0.8203125, | |
| "step": 171 | |
| }, | |
| { | |
| "epoch": 0.688, | |
| "grad_norm": 18.306407029166536, | |
| "learning_rate": 2.3516270313430081e-07, | |
| "logits/chosen": -1.1015625, | |
| "logits/rejected": -1.0703125, | |
| "logps/chosen": -7.21875, | |
| "logps/rejected": -33.0, | |
| "loss": 1.0611, | |
| "nll_loss": 0.6484375, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 2.9375, | |
| "rewards/margins": 2.796875, | |
| "rewards/rejected": 0.140625, | |
| "step": 172 | |
| }, | |
| { | |
| "epoch": 0.692, | |
| "grad_norm": 12.791171107272646, | |
| "learning_rate": 2.2967959127220137e-07, | |
| "logits/chosen": -1.0859375, | |
| "logits/rejected": -1.0703125, | |
| "logps/chosen": -3.765625, | |
| "logps/rejected": -26.25, | |
| "loss": 0.9565, | |
| "nll_loss": 0.3515625, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": 3.125, | |
| "rewards/margins": 2.5, | |
| "rewards/rejected": 0.6328125, | |
| "step": 173 | |
| }, | |
| { | |
| "epoch": 0.696, | |
| "grad_norm": 24.89717076883767, | |
| "learning_rate": 2.242420350078647e-07, | |
| "logits/chosen": -1.0703125, | |
| "logits/rejected": -1.0703125, | |
| "logps/chosen": -5.1875, | |
| "logps/rejected": -27.75, | |
| "loss": 1.0301, | |
| "nll_loss": 0.486328125, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 3.015625, | |
| "rewards/margins": 2.421875, | |
| "rewards/rejected": 0.59765625, | |
| "step": 174 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "grad_norm": 30.549643731351598, | |
| "learning_rate": 2.1885095070246113e-07, | |
| "logits/chosen": -1.1015625, | |
| "logits/rejected": -1.109375, | |
| "logps/chosen": -4.53125, | |
| "logps/rejected": -34.0, | |
| "loss": 1.1014, | |
| "nll_loss": 0.439453125, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 3.046875, | |
| "rewards/margins": 2.8125, | |
| "rewards/rejected": 0.2412109375, | |
| "step": 175 | |
| }, | |
| { | |
| "epoch": 0.704, | |
| "grad_norm": 42.66332171882467, | |
| "learning_rate": 2.1350724688549904e-07, | |
| "logits/chosen": -1.15625, | |
| "logits/rejected": -1.1484375, | |
| "logps/chosen": -5.46875, | |
| "logps/rejected": -31.5, | |
| "loss": 1.1445, | |
| "nll_loss": 0.5625, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": 2.390625, | |
| "rewards/margins": 2.09375, | |
| "rewards/rejected": 0.306640625, | |
| "step": 176 | |
| }, | |
| { | |
| "epoch": 0.708, | |
| "grad_norm": 19.430831101872652, | |
| "learning_rate": 2.0821182410171633e-07, | |
| "logits/chosen": -1.109375, | |
| "logits/rejected": -1.125, | |
| "logps/chosen": -3.5625, | |
| "logps/rejected": -32.75, | |
| "loss": 1.0655, | |
| "nll_loss": 0.328125, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 2.390625, | |
| "rewards/margins": 2.171875, | |
| "rewards/rejected": 0.2099609375, | |
| "step": 177 | |
| }, | |
| { | |
| "epoch": 0.712, | |
| "grad_norm": 40.306067494211455, | |
| "learning_rate": 2.0296557475931687e-07, | |
| "logits/chosen": -1.09375, | |
| "logits/rejected": -1.09375, | |
| "logps/chosen": -6.21875, | |
| "logps/rejected": -28.0, | |
| "loss": 1.1514, | |
| "nll_loss": 0.59375, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 2.84375, | |
| "rewards/margins": 1.65625, | |
| "rewards/rejected": 1.1796875, | |
| "step": 178 | |
| }, | |
| { | |
| "epoch": 0.716, | |
| "grad_norm": 21.550210732610292, | |
| "learning_rate": 1.9776938297957686e-07, | |
| "logits/chosen": -1.0625, | |
| "logits/rejected": -1.03125, | |
| "logps/chosen": -3.140625, | |
| "logps/rejected": -18.75, | |
| "loss": 1.0927, | |
| "nll_loss": 0.33203125, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 3.0, | |
| "rewards/margins": 1.4609375, | |
| "rewards/rejected": 1.546875, | |
| "step": 179 | |
| }, | |
| { | |
| "epoch": 0.72, | |
| "grad_norm": 16.05057780811457, | |
| "learning_rate": 1.9262412444784958e-07, | |
| "logits/chosen": -1.0078125, | |
| "logits/rejected": -1.0234375, | |
| "logps/chosen": -7.0, | |
| "logps/rejected": -21.25, | |
| "loss": 1.0314, | |
| "nll_loss": 0.6875, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 2.75, | |
| "rewards/margins": 1.4765625, | |
| "rewards/rejected": 1.2578125, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.724, | |
| "grad_norm": 22.311212494168778, | |
| "learning_rate": 1.8753066626599084e-07, | |
| "logits/chosen": -1.1015625, | |
| "logits/rejected": -1.09375, | |
| "logps/chosen": -3.75, | |
| "logps/rejected": -34.25, | |
| "loss": 1.1563, | |
| "nll_loss": 0.38671875, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 2.84375, | |
| "rewards/margins": 2.234375, | |
| "rewards/rejected": 0.6015625, | |
| "step": 181 | |
| }, | |
| { | |
| "epoch": 0.728, | |
| "grad_norm": 24.732581587145912, | |
| "learning_rate": 1.8248986680623074e-07, | |
| "logits/chosen": -1.0625, | |
| "logits/rejected": -1.0625, | |
| "logps/chosen": -5.0625, | |
| "logps/rejected": -23.625, | |
| "loss": 1.1566, | |
| "nll_loss": 0.5078125, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 2.59375, | |
| "rewards/margins": 1.546875, | |
| "rewards/rejected": 1.046875, | |
| "step": 182 | |
| }, | |
| { | |
| "epoch": 0.732, | |
| "grad_norm": 52.267590702636376, | |
| "learning_rate": 1.7750257556651805e-07, | |
| "logits/chosen": -1.078125, | |
| "logits/rejected": -1.078125, | |
| "logps/chosen": -4.09375, | |
| "logps/rejected": -31.0, | |
| "loss": 1.1124, | |
| "nll_loss": 0.384765625, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 2.796875, | |
| "rewards/margins": 2.421875, | |
| "rewards/rejected": 0.375, | |
| "step": 183 | |
| }, | |
| { | |
| "epoch": 0.736, | |
| "grad_norm": 29.867593385467835, | |
| "learning_rate": 1.725696330273575e-07, | |
| "logits/chosen": -1.0390625, | |
| "logits/rejected": -1.0390625, | |
| "logps/chosen": -3.9375, | |
| "logps/rejected": -27.625, | |
| "loss": 1.0221, | |
| "nll_loss": 0.41015625, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 3.09375, | |
| "rewards/margins": 2.203125, | |
| "rewards/rejected": 0.890625, | |
| "step": 184 | |
| }, | |
| { | |
| "epoch": 0.74, | |
| "grad_norm": 25.170290330540713, | |
| "learning_rate": 1.6769187051016932e-07, | |
| "logits/chosen": -1.0703125, | |
| "logits/rejected": -1.0703125, | |
| "logps/chosen": -3.90625, | |
| "logps/rejected": -24.75, | |
| "loss": 1.0845, | |
| "nll_loss": 0.400390625, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 2.96875, | |
| "rewards/margins": 2.046875, | |
| "rewards/rejected": 0.90625, | |
| "step": 185 | |
| }, | |
| { | |
| "epoch": 0.744, | |
| "grad_norm": 17.759260894082026, | |
| "learning_rate": 1.6287011003719102e-07, | |
| "logits/chosen": -1.1015625, | |
| "logits/rejected": -1.109375, | |
| "logps/chosen": -4.5, | |
| "logps/rejected": -26.75, | |
| "loss": 1.0185, | |
| "nll_loss": 0.455078125, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": 3.234375, | |
| "rewards/margins": 2.640625, | |
| "rewards/rejected": 0.6015625, | |
| "step": 186 | |
| }, | |
| { | |
| "epoch": 0.748, | |
| "grad_norm": 33.46964477076418, | |
| "learning_rate": 1.581051641929465e-07, | |
| "logits/chosen": -1.1015625, | |
| "logits/rejected": -1.0703125, | |
| "logps/chosen": -5.40625, | |
| "logps/rejected": -22.875, | |
| "loss": 1.1376, | |
| "nll_loss": 0.51953125, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 2.5625, | |
| "rewards/margins": 1.5, | |
| "rewards/rejected": 1.0625, | |
| "step": 187 | |
| }, | |
| { | |
| "epoch": 0.752, | |
| "grad_norm": 24.562745533923515, | |
| "learning_rate": 1.5339783598730565e-07, | |
| "logits/chosen": -1.046875, | |
| "logits/rejected": -1.046875, | |
| "logps/chosen": -5.5, | |
| "logps/rejected": -27.0, | |
| "loss": 1.1846, | |
| "nll_loss": 0.52734375, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 2.640625, | |
| "rewards/margins": 1.90625, | |
| "rewards/rejected": 0.73828125, | |
| "step": 188 | |
| }, | |
| { | |
| "epoch": 0.756, | |
| "grad_norm": 41.7355263163881, | |
| "learning_rate": 1.4874891872015734e-07, | |
| "logits/chosen": -1.046875, | |
| "logits/rejected": -1.046875, | |
| "logps/chosen": -5.90625, | |
| "logps/rejected": -25.25, | |
| "loss": 1.0575, | |
| "nll_loss": 0.546875, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 2.96875, | |
| "rewards/margins": 2.078125, | |
| "rewards/rejected": 0.89453125, | |
| "step": 189 | |
| }, | |
| { | |
| "epoch": 0.76, | |
| "grad_norm": 28.3791093205466, | |
| "learning_rate": 1.4415919584771995e-07, | |
| "logits/chosen": -1.0625, | |
| "logits/rejected": -1.0859375, | |
| "logps/chosen": -4.90625, | |
| "logps/rejected": -25.75, | |
| "loss": 1.1332, | |
| "nll_loss": 0.48828125, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 2.40625, | |
| "rewards/margins": 1.484375, | |
| "rewards/rejected": 0.92578125, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.764, | |
| "grad_norm": 19.154245446560328, | |
| "learning_rate": 1.3962944085050832e-07, | |
| "logits/chosen": -1.0234375, | |
| "logits/rejected": -1.0390625, | |
| "logps/chosen": -6.125, | |
| "logps/rejected": -25.125, | |
| "loss": 1.0726, | |
| "nll_loss": 0.55078125, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": 2.859375, | |
| "rewards/margins": 1.921875, | |
| "rewards/rejected": 0.93359375, | |
| "step": 191 | |
| }, | |
| { | |
| "epoch": 0.768, | |
| "grad_norm": 17.938238931793443, | |
| "learning_rate": 1.3516041710298497e-07, | |
| "logits/chosen": -1.0859375, | |
| "logits/rejected": -1.0703125, | |
| "logps/chosen": -4.125, | |
| "logps/rejected": -19.875, | |
| "loss": 0.9941, | |
| "nll_loss": 0.423828125, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 2.890625, | |
| "rewards/margins": 1.6015625, | |
| "rewards/rejected": 1.2890625, | |
| "step": 192 | |
| }, | |
| { | |
| "epoch": 0.772, | |
| "grad_norm": 20.81847675232155, | |
| "learning_rate": 1.3075287774491146e-07, | |
| "logits/chosen": -1.0703125, | |
| "logits/rejected": -1.1015625, | |
| "logps/chosen": -2.65625, | |
| "logps/rejected": -28.625, | |
| "loss": 1.0433, | |
| "nll_loss": 0.271484375, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": 3.046875, | |
| "rewards/margins": 2.203125, | |
| "rewards/rejected": 0.83203125, | |
| "step": 193 | |
| }, | |
| { | |
| "epoch": 0.776, | |
| "grad_norm": 44.286085577032274, | |
| "learning_rate": 1.2640756555442683e-07, | |
| "logits/chosen": -1.0703125, | |
| "logits/rejected": -1.0546875, | |
| "logps/chosen": -6.5625, | |
| "logps/rejected": -25.125, | |
| "loss": 1.0916, | |
| "nll_loss": 0.64453125, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 3.078125, | |
| "rewards/margins": 1.9140625, | |
| "rewards/rejected": 1.1640625, | |
| "step": 194 | |
| }, | |
| { | |
| "epoch": 0.78, | |
| "grad_norm": 22.029622182069097, | |
| "learning_rate": 1.2212521282287093e-07, | |
| "logits/chosen": -1.0625, | |
| "logits/rejected": -1.0625, | |
| "logps/chosen": -4.5625, | |
| "logps/rejected": -28.125, | |
| "loss": 1.0946, | |
| "nll_loss": 0.462890625, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": 2.9375, | |
| "rewards/margins": 2.0625, | |
| "rewards/rejected": 0.87109375, | |
| "step": 195 | |
| }, | |
| { | |
| "epoch": 0.784, | |
| "grad_norm": 11.861056714322848, | |
| "learning_rate": 1.179065412313755e-07, | |
| "logits/chosen": -1.1171875, | |
| "logits/rejected": -1.1015625, | |
| "logps/chosen": -6.5, | |
| "logps/rejected": -21.75, | |
| "loss": 1.0592, | |
| "nll_loss": 0.640625, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": 2.90625, | |
| "rewards/margins": 1.6796875, | |
| "rewards/rejected": 1.2265625, | |
| "step": 196 | |
| }, | |
| { | |
| "epoch": 0.788, | |
| "grad_norm": 14.546735669018146, | |
| "learning_rate": 1.137522617292434e-07, | |
| "logits/chosen": -1.03125, | |
| "logits/rejected": -1.0234375, | |
| "logps/chosen": -4.34375, | |
| "logps/rejected": -18.5, | |
| "loss": 1.0507, | |
| "nll_loss": 0.41796875, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 3.109375, | |
| "rewards/margins": 1.3203125, | |
| "rewards/rejected": 1.7890625, | |
| "step": 197 | |
| }, | |
| { | |
| "epoch": 0.792, | |
| "grad_norm": 55.63915697780125, | |
| "learning_rate": 1.0966307441413597e-07, | |
| "logits/chosen": -1.1015625, | |
| "logits/rejected": -1.1015625, | |
| "logps/chosen": -6.125, | |
| "logps/rejected": -23.25, | |
| "loss": 1.0632, | |
| "nll_loss": 0.60546875, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": 3.140625, | |
| "rewards/margins": 1.9921875, | |
| "rewards/rejected": 1.140625, | |
| "step": 198 | |
| }, | |
| { | |
| "epoch": 0.796, | |
| "grad_norm": 20.1761931612358, | |
| "learning_rate": 1.0563966841408995e-07, | |
| "logits/chosen": -1.0703125, | |
| "logits/rejected": -1.046875, | |
| "logps/chosen": -5.03125, | |
| "logps/rejected": -29.0, | |
| "loss": 0.9663, | |
| "nll_loss": 0.494140625, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 3.25, | |
| "rewards/margins": 2.546875, | |
| "rewards/rejected": 0.703125, | |
| "step": 199 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "grad_norm": 23.608862708071296, | |
| "learning_rate": 1.0168272177138199e-07, | |
| "logits/chosen": -1.0625, | |
| "logits/rejected": -1.0546875, | |
| "logps/chosen": -4.9375, | |
| "logps/rejected": -28.5, | |
| "loss": 1.0909, | |
| "nll_loss": 0.466796875, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": 2.90625, | |
| "rewards/margins": 2.109375, | |
| "rewards/rejected": 0.8046875, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.804, | |
| "grad_norm": 41.55024587771104, | |
| "learning_rate": 9.779290132826224e-08, | |
| "logits/chosen": -1.09375, | |
| "logits/rejected": -1.0859375, | |
| "logps/chosen": -4.84375, | |
| "logps/rejected": -24.125, | |
| "loss": 1.0829, | |
| "nll_loss": 0.44140625, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": 3.390625, | |
| "rewards/margins": 1.9453125, | |
| "rewards/rejected": 1.4453125, | |
| "step": 201 | |
| }, | |
| { | |
| "epoch": 0.808, | |
| "grad_norm": 34.54633889712777, | |
| "learning_rate": 9.39708626145751e-08, | |
| "logits/chosen": -1.0546875, | |
| "logits/rejected": -1.046875, | |
| "logps/chosen": -5.90625, | |
| "logps/rejected": -27.25, | |
| "loss": 1.0733, | |
| "nll_loss": 0.57421875, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 3.28125, | |
| "rewards/margins": 2.203125, | |
| "rewards/rejected": 1.078125, | |
| "step": 202 | |
| }, | |
| { | |
| "epoch": 0.812, | |
| "grad_norm": 32.77492433560517, | |
| "learning_rate": 9.021724973728634e-08, | |
| "logits/chosen": -1.0546875, | |
| "logits/rejected": -1.078125, | |
| "logps/chosen": -4.875, | |
| "logps/rejected": -31.75, | |
| "loss": 1.0815, | |
| "nll_loss": 0.478515625, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": 2.921875, | |
| "rewards/margins": 2.59375, | |
| "rewards/rejected": 0.318359375, | |
| "step": 203 | |
| }, | |
| { | |
| "epoch": 0.816, | |
| "grad_norm": 12.483799623280103, | |
| "learning_rate": 8.653269527193568e-08, | |
| "logits/chosen": -1.109375, | |
| "logits/rejected": -1.109375, | |
| "logps/chosen": -5.25, | |
| "logps/rejected": -25.375, | |
| "loss": 1.0666, | |
| "nll_loss": 0.486328125, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 3.0, | |
| "rewards/margins": 1.9140625, | |
| "rewards/rejected": 1.0859375, | |
| "step": 204 | |
| }, | |
| { | |
| "epoch": 0.82, | |
| "grad_norm": 52.55865885582505, | |
| "learning_rate": 8.291782015603176e-08, | |
| "logits/chosen": -1.140625, | |
| "logits/rejected": -1.09375, | |
| "logps/chosen": -6.0625, | |
| "logps/rejected": -20.375, | |
| "loss": 1.1337, | |
| "nll_loss": 0.63671875, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 2.625, | |
| "rewards/margins": 1.2578125, | |
| "rewards/rejected": 1.359375, | |
| "step": 205 | |
| }, | |
| { | |
| "epoch": 0.824, | |
| "grad_norm": 32.755044229250124, | |
| "learning_rate": 7.937323358440934e-08, | |
| "logits/chosen": -1.0703125, | |
| "logits/rejected": -1.0859375, | |
| "logps/chosen": -4.0625, | |
| "logps/rejected": -22.875, | |
| "loss": 1.0425, | |
| "nll_loss": 0.404296875, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 3.203125, | |
| "rewards/margins": 1.7109375, | |
| "rewards/rejected": 1.484375, | |
| "step": 206 | |
| }, | |
| { | |
| "epoch": 0.828, | |
| "grad_norm": 49.57046953822534, | |
| "learning_rate": 7.589953290656532e-08, | |
| "logits/chosen": -1.1171875, | |
| "logits/rejected": -1.1015625, | |
| "logps/chosen": -3.234375, | |
| "logps/rejected": -28.25, | |
| "loss": 1.1414, | |
| "nll_loss": 0.32421875, | |
| "rewards/accuracies": 0.90625, | |
| "rewards/chosen": 2.859375, | |
| "rewards/margins": 2.4375, | |
| "rewards/rejected": 0.435546875, | |
| "step": 207 | |
| }, | |
| { | |
| "epoch": 0.832, | |
| "grad_norm": 29.02033137242745, | |
| "learning_rate": 7.249730352598998e-08, | |
| "logits/chosen": -1.09375, | |
| "logits/rejected": -1.0859375, | |
| "logps/chosen": -4.8125, | |
| "logps/rejected": -28.75, | |
| "loss": 1.004, | |
| "nll_loss": 0.451171875, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": 2.671875, | |
| "rewards/margins": 2.171875, | |
| "rewards/rejected": 0.51171875, | |
| "step": 208 | |
| }, | |
| { | |
| "epoch": 0.836, | |
| "grad_norm": 17.714740130325648, | |
| "learning_rate": 6.916711880151304e-08, | |
| "logits/chosen": -1.109375, | |
| "logits/rejected": -1.1015625, | |
| "logps/chosen": -6.25, | |
| "logps/rejected": -30.125, | |
| "loss": 1.1389, | |
| "nll_loss": 0.609375, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": 2.703125, | |
| "rewards/margins": 2.109375, | |
| "rewards/rejected": 0.58984375, | |
| "step": 209 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "grad_norm": 17.72865807824904, | |
| "learning_rate": 6.59095399506781e-08, | |
| "logits/chosen": -1.1015625, | |
| "logits/rejected": -1.1015625, | |
| "logps/chosen": -6.5625, | |
| "logps/rejected": -26.5, | |
| "loss": 1.1747, | |
| "nll_loss": 0.6484375, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 2.8125, | |
| "rewards/margins": 1.875, | |
| "rewards/rejected": 0.94140625, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.844, | |
| "grad_norm": 13.99041401845889, | |
| "learning_rate": 6.2725115955164e-08, | |
| "logits/chosen": -1.09375, | |
| "logits/rejected": -1.0859375, | |
| "logps/chosen": -4.1875, | |
| "logps/rejected": -33.5, | |
| "loss": 1.0243, | |
| "nll_loss": 0.3984375, | |
| "rewards/accuracies": 0.90625, | |
| "rewards/chosen": 2.96875, | |
| "rewards/margins": 2.828125, | |
| "rewards/rejected": 0.140625, | |
| "step": 211 | |
| }, | |
| { | |
| "epoch": 0.848, | |
| "grad_norm": 15.906985339916742, | |
| "learning_rate": 5.961438346826792e-08, | |
| "logits/chosen": -1.1015625, | |
| "logits/rejected": -1.0703125, | |
| "logps/chosen": -4.625, | |
| "logps/rejected": -28.0, | |
| "loss": 1.0347, | |
| "nll_loss": 0.443359375, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 2.59375, | |
| "rewards/margins": 2.03125, | |
| "rewards/rejected": 0.55859375, | |
| "step": 212 | |
| }, | |
| { | |
| "epoch": 0.852, | |
| "grad_norm": 36.27762496758833, | |
| "learning_rate": 5.6577866724466296e-08, | |
| "logits/chosen": -1.140625, | |
| "logits/rejected": -1.1171875, | |
| "logps/chosen": -6.53125, | |
| "logps/rejected": -47.75, | |
| "loss": 1.0298, | |
| "nll_loss": 0.58203125, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": 2.328125, | |
| "rewards/margins": 3.59375, | |
| "rewards/rejected": -1.265625, | |
| "step": 213 | |
| }, | |
| { | |
| "epoch": 0.856, | |
| "grad_norm": 22.188902025969416, | |
| "learning_rate": 5.361607745106816e-08, | |
| "logits/chosen": -1.078125, | |
| "logits/rejected": -1.0703125, | |
| "logps/chosen": -3.734375, | |
| "logps/rejected": -22.5, | |
| "loss": 1.0263, | |
| "nll_loss": 0.431640625, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": 3.0, | |
| "rewards/margins": 1.9375, | |
| "rewards/rejected": 1.0625, | |
| "step": 214 | |
| }, | |
| { | |
| "epoch": 0.86, | |
| "grad_norm": 26.514396481742928, | |
| "learning_rate": 5.072951478197723e-08, | |
| "logits/chosen": -1.1015625, | |
| "logits/rejected": -1.0703125, | |
| "logps/chosen": -4.46875, | |
| "logps/rejected": -31.5, | |
| "loss": 1.0853, | |
| "nll_loss": 0.4375, | |
| "rewards/accuracies": 0.90625, | |
| "rewards/chosen": 2.8125, | |
| "rewards/margins": 2.296875, | |
| "rewards/rejected": 0.515625, | |
| "step": 215 | |
| }, | |
| { | |
| "epoch": 0.864, | |
| "grad_norm": 39.18691508367433, | |
| "learning_rate": 4.79186651735749e-08, | |
| "logits/chosen": -1.0703125, | |
| "logits/rejected": -1.078125, | |
| "logps/chosen": -4.25, | |
| "logps/rejected": -23.625, | |
| "loss": 0.9658, | |
| "nll_loss": 0.427734375, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": 2.828125, | |
| "rewards/margins": 1.625, | |
| "rewards/rejected": 1.2109375, | |
| "step": 216 | |
| }, | |
| { | |
| "epoch": 0.868, | |
| "grad_norm": 17.053221125940922, | |
| "learning_rate": 4.518400232274078e-08, | |
| "logits/chosen": -1.1171875, | |
| "logits/rejected": -1.109375, | |
| "logps/chosen": -4.90625, | |
| "logps/rejected": -27.125, | |
| "loss": 1.0356, | |
| "nll_loss": 0.4375, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 2.65625, | |
| "rewards/margins": 1.9140625, | |
| "rewards/rejected": 0.7421875, | |
| "step": 217 | |
| }, | |
| { | |
| "epoch": 0.872, | |
| "grad_norm": 30.04204392176379, | |
| "learning_rate": 4.252598708702343e-08, | |
| "logits/chosen": -1.0625, | |
| "logits/rejected": -1.0546875, | |
| "logps/chosen": -5.375, | |
| "logps/rejected": -24.25, | |
| "loss": 1.083, | |
| "nll_loss": 0.53125, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 2.765625, | |
| "rewards/margins": 1.5546875, | |
| "rewards/rejected": 1.203125, | |
| "step": 218 | |
| }, | |
| { | |
| "epoch": 0.876, | |
| "grad_norm": 12.813171720804812, | |
| "learning_rate": 3.994506740697406e-08, | |
| "logits/chosen": -1.046875, | |
| "logits/rejected": -1.0546875, | |
| "logps/chosen": -4.5625, | |
| "logps/rejected": -23.0, | |
| "loss": 0.9825, | |
| "nll_loss": 0.419921875, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 2.734375, | |
| "rewards/margins": 1.578125, | |
| "rewards/rejected": 1.15625, | |
| "step": 219 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "grad_norm": 31.189313546893203, | |
| "learning_rate": 3.744167823065814e-08, | |
| "logits/chosen": -1.0859375, | |
| "logits/rejected": -1.078125, | |
| "logps/chosen": -4.96875, | |
| "logps/rejected": -36.0, | |
| "loss": 1.0805, | |
| "nll_loss": 0.478515625, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 2.296875, | |
| "rewards/margins": 2.71875, | |
| "rewards/rejected": -0.4296875, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.884, | |
| "grad_norm": 38.49107579612584, | |
| "learning_rate": 3.501624144035559e-08, | |
| "logits/chosen": -1.1015625, | |
| "logits/rejected": -1.0703125, | |
| "logps/chosen": -5.5, | |
| "logps/rejected": -23.375, | |
| "loss": 1.097, | |
| "nll_loss": 0.53125, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 2.78125, | |
| "rewards/margins": 1.6953125, | |
| "rewards/rejected": 1.09375, | |
| "step": 221 | |
| }, | |
| { | |
| "epoch": 0.888, | |
| "grad_norm": 13.5024131062282, | |
| "learning_rate": 3.2669165781463395e-08, | |
| "logits/chosen": -1.0859375, | |
| "logits/rejected": -1.078125, | |
| "logps/chosen": -4.53125, | |
| "logps/rejected": -29.875, | |
| "loss": 1.0892, | |
| "nll_loss": 0.44921875, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 2.765625, | |
| "rewards/margins": 1.625, | |
| "rewards/rejected": 1.140625, | |
| "step": 222 | |
| }, | |
| { | |
| "epoch": 0.892, | |
| "grad_norm": 28.282704129587508, | |
| "learning_rate": 3.0400846793612544e-08, | |
| "logits/chosen": -1.0859375, | |
| "logits/rejected": -1.0546875, | |
| "logps/chosen": -4.71875, | |
| "logps/rejected": -23.25, | |
| "loss": 1.0451, | |
| "nll_loss": 0.462890625, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 2.859375, | |
| "rewards/margins": 1.28125, | |
| "rewards/rejected": 1.578125, | |
| "step": 223 | |
| }, | |
| { | |
| "epoch": 0.896, | |
| "grad_norm": 37.60742006233887, | |
| "learning_rate": 2.8211666744009045e-08, | |
| "logits/chosen": -1.1328125, | |
| "logits/rejected": -1.125, | |
| "logps/chosen": -3.796875, | |
| "logps/rejected": -29.875, | |
| "loss": 1.0323, | |
| "nll_loss": 0.333984375, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 2.671875, | |
| "rewards/margins": 2.265625, | |
| "rewards/rejected": 0.41015625, | |
| "step": 224 | |
| }, | |
| { | |
| "epoch": 0.9, | |
| "grad_norm": 29.516319250938352, | |
| "learning_rate": 2.6101994563013098e-08, | |
| "logits/chosen": -1.078125, | |
| "logits/rejected": -1.09375, | |
| "logps/chosen": -3.453125, | |
| "logps/rejected": -26.25, | |
| "loss": 1.0479, | |
| "nll_loss": 0.341796875, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": 2.59375, | |
| "rewards/margins": 1.796875, | |
| "rewards/rejected": 0.7890625, | |
| "step": 225 | |
| }, | |
| { | |
| "epoch": 0.904, | |
| "grad_norm": 14.847458095289273, | |
| "learning_rate": 2.4072185781965238e-08, | |
| "logits/chosen": -1.0859375, | |
| "logits/rejected": -1.09375, | |
| "logps/chosen": -4.8125, | |
| "logps/rejected": -29.25, | |
| "loss": 1.0546, | |
| "nll_loss": 0.43359375, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": 2.9375, | |
| "rewards/margins": 2.203125, | |
| "rewards/rejected": 0.7265625, | |
| "step": 226 | |
| }, | |
| { | |
| "epoch": 0.908, | |
| "grad_norm": 17.36079249416162, | |
| "learning_rate": 2.212258247327059e-08, | |
| "logits/chosen": -1.109375, | |
| "logits/rejected": -1.09375, | |
| "logps/chosen": -3.578125, | |
| "logps/rejected": -31.875, | |
| "loss": 1.0537, | |
| "nll_loss": 0.359375, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 2.765625, | |
| "rewards/margins": 2.34375, | |
| "rewards/rejected": 0.4375, | |
| "step": 227 | |
| }, | |
| { | |
| "epoch": 0.912, | |
| "grad_norm": 18.768525578098558, | |
| "learning_rate": 2.025351319275137e-08, | |
| "logits/chosen": -1.1015625, | |
| "logits/rejected": -1.0859375, | |
| "logps/chosen": -6.59375, | |
| "logps/rejected": -38.25, | |
| "loss": 1.106, | |
| "nll_loss": 0.56640625, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 2.875, | |
| "rewards/margins": 3.203125, | |
| "rewards/rejected": -0.31640625, | |
| "step": 228 | |
| }, | |
| { | |
| "epoch": 0.916, | |
| "grad_norm": 33.5917307298862, | |
| "learning_rate": 1.8465292924276844e-08, | |
| "logits/chosen": -1.1484375, | |
| "logits/rejected": -1.1484375, | |
| "logps/chosen": -5.09375, | |
| "logps/rejected": -42.75, | |
| "loss": 0.9739, | |
| "nll_loss": 0.46875, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 3.078125, | |
| "rewards/margins": 3.578125, | |
| "rewards/rejected": -0.50390625, | |
| "step": 229 | |
| }, | |
| { | |
| "epoch": 0.92, | |
| "grad_norm": 12.955837772485086, | |
| "learning_rate": 1.6758223026681507e-08, | |
| "logits/chosen": -1.0703125, | |
| "logits/rejected": -1.078125, | |
| "logps/chosen": -6.1875, | |
| "logps/rejected": -30.75, | |
| "loss": 1.0963, | |
| "nll_loss": 0.58203125, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 2.734375, | |
| "rewards/margins": 2.234375, | |
| "rewards/rejected": 0.5, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.924, | |
| "grad_norm": 16.2992963110193, | |
| "learning_rate": 1.5132591182978105e-08, | |
| "logits/chosen": -1.0625, | |
| "logits/rejected": -1.0546875, | |
| "logps/chosen": -6.28125, | |
| "logps/rejected": -31.875, | |
| "loss": 1.0419, | |
| "nll_loss": 0.56640625, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": 2.75, | |
| "rewards/margins": 2.265625, | |
| "rewards/rejected": 0.482421875, | |
| "step": 231 | |
| }, | |
| { | |
| "epoch": 0.928, | |
| "grad_norm": 12.293312024242535, | |
| "learning_rate": 1.3588671351876357e-08, | |
| "logits/chosen": -1.0703125, | |
| "logits/rejected": -1.0859375, | |
| "logps/chosen": -4.53125, | |
| "logps/rejected": -22.875, | |
| "loss": 1.0256, | |
| "nll_loss": 0.439453125, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 2.59375, | |
| "rewards/margins": 1.4375, | |
| "rewards/rejected": 1.15625, | |
| "step": 232 | |
| }, | |
| { | |
| "epoch": 0.932, | |
| "grad_norm": 18.415969639868365, | |
| "learning_rate": 1.2126723721614052e-08, | |
| "logits/chosen": -1.0859375, | |
| "logits/rejected": -1.109375, | |
| "logps/chosen": -4.71875, | |
| "logps/rejected": -28.5, | |
| "loss": 1.0731, | |
| "nll_loss": 0.46484375, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 2.59375, | |
| "rewards/margins": 2.203125, | |
| "rewards/rejected": 0.376953125, | |
| "step": 233 | |
| }, | |
| { | |
| "epoch": 0.936, | |
| "grad_norm": 17.63813786850423, | |
| "learning_rate": 1.0746994666109233e-08, | |
| "logits/chosen": -1.109375, | |
| "logits/rejected": -1.0859375, | |
| "logps/chosen": -4.75, | |
| "logps/rejected": -20.5, | |
| "loss": 0.9922, | |
| "nll_loss": 0.431640625, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 3.46875, | |
| "rewards/margins": 1.765625, | |
| "rewards/rejected": 1.6953125, | |
| "step": 234 | |
| }, | |
| { | |
| "epoch": 0.94, | |
| "grad_norm": 15.498735501179883, | |
| "learning_rate": 9.449716703439803e-09, | |
| "logits/chosen": -1.1171875, | |
| "logits/rejected": -1.140625, | |
| "logps/chosen": -4.6875, | |
| "logps/rejected": -23.625, | |
| "loss": 1.0851, | |
| "nll_loss": 0.404296875, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 2.9375, | |
| "rewards/margins": 2.046875, | |
| "rewards/rejected": 0.89453125, | |
| "step": 235 | |
| }, | |
| { | |
| "epoch": 0.944, | |
| "grad_norm": 11.978784225505468, | |
| "learning_rate": 8.235108456658812e-09, | |
| "logits/chosen": -1.140625, | |
| "logits/rejected": -1.1328125, | |
| "logps/chosen": -4.03125, | |
| "logps/rejected": -39.25, | |
| "loss": 1.0219, | |
| "nll_loss": 0.390625, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": 2.59375, | |
| "rewards/margins": 3.0, | |
| "rewards/rejected": -0.41015625, | |
| "step": 236 | |
| }, | |
| { | |
| "epoch": 0.948, | |
| "grad_norm": 13.770414732547914, | |
| "learning_rate": 7.103374616951041e-09, | |
| "logits/chosen": -1.0546875, | |
| "logits/rejected": -1.0546875, | |
| "logps/chosen": -3.25, | |
| "logps/rejected": -21.75, | |
| "loss": 1.0617, | |
| "nll_loss": 0.322265625, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": 3.3125, | |
| "rewards/margins": 1.7109375, | |
| "rewards/rejected": 1.6015625, | |
| "step": 237 | |
| }, | |
| { | |
| "epoch": 0.952, | |
| "grad_norm": 26.345272412615557, | |
| "learning_rate": 6.054705909137425e-09, | |
| "logits/chosen": -1.1015625, | |
| "logits/rejected": -1.125, | |
| "logps/chosen": -4.1875, | |
| "logps/rejected": -38.0, | |
| "loss": 1.078, | |
| "nll_loss": 0.443359375, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": 2.734375, | |
| "rewards/margins": 3.03125, | |
| "rewards/rejected": -0.30078125, | |
| "step": 238 | |
| }, | |
| { | |
| "epoch": 0.956, | |
| "grad_norm": 17.86401941431699, | |
| "learning_rate": 5.0892790595336575e-09, | |
| "logits/chosen": -1.125, | |
| "logits/rejected": -1.1328125, | |
| "logps/chosen": -4.4375, | |
| "logps/rejected": -27.5, | |
| "loss": 0.9794, | |
| "nll_loss": 0.451171875, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": 2.640625, | |
| "rewards/margins": 2.140625, | |
| "rewards/rejected": 0.49609375, | |
| "step": 239 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "grad_norm": 17.07158211116171, | |
| "learning_rate": 4.207256766166845e-09, | |
| "logits/chosen": -1.078125, | |
| "logits/rejected": -1.0859375, | |
| "logps/chosen": -4.90625, | |
| "logps/rejected": -26.625, | |
| "loss": 1.1286, | |
| "nll_loss": 0.443359375, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 2.453125, | |
| "rewards/margins": 1.578125, | |
| "rewards/rejected": 0.875, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.964, | |
| "grad_norm": 40.38091842339886, | |
| "learning_rate": 3.4087876713574935e-09, | |
| "logits/chosen": -1.1015625, | |
| "logits/rejected": -1.09375, | |
| "logps/chosen": -5.96875, | |
| "logps/rejected": -31.375, | |
| "loss": 0.9722, | |
| "nll_loss": 0.609375, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": 3.015625, | |
| "rewards/margins": 2.421875, | |
| "rewards/rejected": 0.58984375, | |
| "step": 241 | |
| }, | |
| { | |
| "epoch": 0.968, | |
| "grad_norm": 14.247494421593556, | |
| "learning_rate": 2.69400633666933e-09, | |
| "logits/chosen": -1.1171875, | |
| "logits/rejected": -1.109375, | |
| "logps/chosen": -4.4375, | |
| "logps/rejected": -21.375, | |
| "loss": 1.0203, | |
| "nll_loss": 0.408203125, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 2.71875, | |
| "rewards/margins": 1.640625, | |
| "rewards/rejected": 1.078125, | |
| "step": 242 | |
| }, | |
| { | |
| "epoch": 0.972, | |
| "grad_norm": 30.909416470072028, | |
| "learning_rate": 2.0630332202324906e-09, | |
| "logits/chosen": -1.0546875, | |
| "logits/rejected": -1.0859375, | |
| "logps/chosen": -5.65625, | |
| "logps/rejected": -34.25, | |
| "loss": 1.0594, | |
| "nll_loss": 0.5, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 3.03125, | |
| "rewards/margins": 2.96875, | |
| "rewards/rejected": 0.06396484375, | |
| "step": 243 | |
| }, | |
| { | |
| "epoch": 0.976, | |
| "grad_norm": 13.79697350875317, | |
| "learning_rate": 1.5159746564433196e-09, | |
| "logits/chosen": -1.09375, | |
| "logits/rejected": -1.09375, | |
| "logps/chosen": -5.875, | |
| "logps/rejected": -31.625, | |
| "loss": 1.0882, | |
| "nll_loss": 0.59375, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 2.625, | |
| "rewards/margins": 2.5, | |
| "rewards/rejected": 0.1337890625, | |
| "step": 244 | |
| }, | |
| { | |
| "epoch": 0.98, | |
| "grad_norm": 26.292855916544468, | |
| "learning_rate": 1.0529228380445899e-09, | |
| "logits/chosen": -1.125, | |
| "logits/rejected": -1.1015625, | |
| "logps/chosen": -4.9375, | |
| "logps/rejected": -36.0, | |
| "loss": 1.0991, | |
| "nll_loss": 0.4453125, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": 2.921875, | |
| "rewards/margins": 2.765625, | |
| "rewards/rejected": 0.1630859375, | |
| "step": 245 | |
| }, | |
| { | |
| "epoch": 0.984, | |
| "grad_norm": 9.993888997764955, | |
| "learning_rate": 6.739558005884883e-10, | |
| "logits/chosen": -1.0625, | |
| "logits/rejected": -1.0546875, | |
| "logps/chosen": -6.125, | |
| "logps/rejected": -25.625, | |
| "loss": 1.0262, | |
| "nll_loss": 0.52734375, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 2.78125, | |
| "rewards/margins": 1.671875, | |
| "rewards/rejected": 1.109375, | |
| "step": 246 | |
| }, | |
| { | |
| "epoch": 0.988, | |
| "grad_norm": 26.071889768691495, | |
| "learning_rate": 3.791374092859678e-10, | |
| "logits/chosen": -1.1171875, | |
| "logits/rejected": -1.1328125, | |
| "logps/chosen": -6.34375, | |
| "logps/rejected": -31.75, | |
| "loss": 1.0238, | |
| "nll_loss": 0.55078125, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 2.515625, | |
| "rewards/margins": 2.046875, | |
| "rewards/rejected": 0.46875, | |
| "step": 247 | |
| }, | |
| { | |
| "epoch": 0.992, | |
| "grad_norm": 27.845797479287505, | |
| "learning_rate": 1.6851734824380182e-10, | |
| "logits/chosen": -1.0703125, | |
| "logits/rejected": -1.078125, | |
| "logps/chosen": -3.578125, | |
| "logps/rejected": -32.75, | |
| "loss": 1.1748, | |
| "nll_loss": 0.341796875, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 2.953125, | |
| "rewards/margins": 2.265625, | |
| "rewards/rejected": 0.6875, | |
| "step": 248 | |
| }, | |
| { | |
| "epoch": 0.996, | |
| "grad_norm": 16.83674597083853, | |
| "learning_rate": 4.213111209155906e-11, | |
| "logits/chosen": -1.140625, | |
| "logits/rejected": -1.125, | |
| "logps/chosen": -4.1875, | |
| "logps/rejected": -30.125, | |
| "loss": 0.9346, | |
| "nll_loss": 0.38671875, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 2.984375, | |
| "rewards/margins": 2.625, | |
| "rewards/rejected": 0.35546875, | |
| "step": 249 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 24.975249728436633, | |
| "learning_rate": 0.0, | |
| "logits/chosen": -1.09375, | |
| "logits/rejected": -1.0859375, | |
| "logps/chosen": -5.1875, | |
| "logps/rejected": -30.875, | |
| "loss": 1.0326, | |
| "nll_loss": 0.5078125, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 2.375, | |
| "rewards/margins": 2.171875, | |
| "rewards/rejected": 0.2060546875, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 250, | |
| "total_flos": 1.517913978562874e+20, | |
| "train_loss": 1.4306324501037597, | |
| "train_runtime": 16604.645, | |
| "train_samples_per_second": 3.854, | |
| "train_steps_per_second": 0.015 | |
| } | |
| ], | |
| "logging_steps": 1.0, | |
| "max_steps": 250, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 100, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 1.517913978562874e+20, | |
| "train_batch_size": 1, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |