| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 5.0, | |
| "eval_steps": 100, | |
| "global_step": 1250, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0, | |
| "learning_rate": 4e-08, | |
| "logits/chosen": -1.8503975868225098, | |
| "logits/rejected": -1.8503975868225098, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.4075, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.04, | |
| "learning_rate": 4.0000000000000003e-07, | |
| "logits/chosen": -1.8588156700134277, | |
| "logits/rejected": -1.8588156700134277, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3636, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "learning_rate": 8.000000000000001e-07, | |
| "logits/chosen": -1.970517873764038, | |
| "logits/rejected": -1.970517873764038, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3902, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.12, | |
| "learning_rate": 1.2000000000000002e-06, | |
| "logits/chosen": -1.9209930896759033, | |
| "logits/rejected": -1.9209930896759033, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3482, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "learning_rate": 1.6000000000000001e-06, | |
| "logits/chosen": -1.883547067642212, | |
| "logits/rejected": -1.883547067642212, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3507, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "learning_rate": 2.0000000000000003e-06, | |
| "logits/chosen": -1.9128715991973877, | |
| "logits/rejected": -1.9128715991973877, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3359, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.24, | |
| "learning_rate": 2.4000000000000003e-06, | |
| "logits/chosen": -2.0107295513153076, | |
| "logits/rejected": -2.0107295513153076, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3828, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.28, | |
| "learning_rate": 2.8000000000000003e-06, | |
| "logits/chosen": -1.9920228719711304, | |
| "logits/rejected": -1.9920228719711304, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3112, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "learning_rate": 3.2000000000000003e-06, | |
| "logits/chosen": -1.8801155090332031, | |
| "logits/rejected": -1.8801155090332031, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3778, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.36, | |
| "learning_rate": 3.6000000000000003e-06, | |
| "logits/chosen": -2.050198793411255, | |
| "logits/rejected": -2.050198793411255, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3655, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "learning_rate": 4.000000000000001e-06, | |
| "logits/chosen": -1.8852717876434326, | |
| "logits/rejected": -1.8852717876434326, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3803, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "eval_logits/chosen": -1.9617642164230347, | |
| "eval_logits/rejected": -1.8066532611846924, | |
| "eval_logps/chosen": -266.6976013183594, | |
| "eval_logps/rejected": -254.9398193359375, | |
| "eval_loss": 0.053734518587589264, | |
| "eval_rewards/accuracies": 0.0, | |
| "eval_rewards/chosen": 0.0, | |
| "eval_rewards/margins": 0.0, | |
| "eval_rewards/rejected": 0.0, | |
| "eval_runtime": 701.958, | |
| "eval_samples_per_second": 2.849, | |
| "eval_steps_per_second": 1.425, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.44, | |
| "learning_rate": 4.4e-06, | |
| "logits/chosen": -1.731688141822815, | |
| "logits/rejected": -1.731688141822815, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.2717, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "learning_rate": 4.800000000000001e-06, | |
| "logits/chosen": -1.8530235290527344, | |
| "logits/rejected": -1.8530235290527344, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3482, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "learning_rate": 4.999756310023261e-06, | |
| "logits/chosen": -2.0225424766540527, | |
| "logits/rejected": -2.0225424766540527, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3507, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.56, | |
| "learning_rate": 4.997807075247147e-06, | |
| "logits/chosen": -1.8995482921600342, | |
| "logits/rejected": -1.8995482921600342, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3186, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.6, | |
| "learning_rate": 4.993910125649561e-06, | |
| "logits/chosen": -1.8709551095962524, | |
| "logits/rejected": -1.8500862121582031, | |
| "logps/chosen": -4.896004676818848, | |
| "logps/rejected": -1.6084611415863037, | |
| "loss": 0.3112, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "learning_rate": 4.988068499954578e-06, | |
| "logits/chosen": -2.044778823852539, | |
| "logits/rejected": -2.044778823852539, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3383, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.68, | |
| "learning_rate": 4.980286753286196e-06, | |
| "logits/chosen": -1.8561241626739502, | |
| "logits/rejected": -1.8561241626739502, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3531, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.72, | |
| "learning_rate": 4.970570953616383e-06, | |
| "logits/chosen": -1.956254243850708, | |
| "logits/rejected": -1.922114372253418, | |
| "logps/chosen": -12.786833763122559, | |
| "logps/rejected": -5.852889060974121, | |
| "loss": 0.3401, | |
| "rewards/accuracies": 0.02500000037252903, | |
| "rewards/chosen": -0.015111232176423073, | |
| "rewards/margins": 0.0011773437727242708, | |
| "rewards/rejected": -0.016288574784994125, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.76, | |
| "learning_rate": 4.958928677033465e-06, | |
| "logits/chosen": -1.8586145639419556, | |
| "logits/rejected": -1.8586145639419556, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3383, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "learning_rate": 4.9453690018345144e-06, | |
| "logits/chosen": -1.908278226852417, | |
| "logits/rejected": -1.9099280834197998, | |
| "logps/chosen": -5.618971824645996, | |
| "logps/rejected": -6.245630741119385, | |
| "loss": 0.2725, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": -0.00010124015534529462, | |
| "rewards/margins": -0.0011139239650219679, | |
| "rewards/rejected": 0.0010126838460564613, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "eval_logits/chosen": -1.9946638345718384, | |
| "eval_logits/rejected": -1.8361071348190308, | |
| "eval_logps/chosen": -270.886474609375, | |
| "eval_logps/rejected": -259.3279724121094, | |
| "eval_loss": 0.054894037544727325, | |
| "eval_rewards/accuracies": 0.4410000145435333, | |
| "eval_rewards/chosen": -0.02094435505568981, | |
| "eval_rewards/margins": 0.0009965725475922227, | |
| "eval_rewards/rejected": -0.021940927952528, | |
| "eval_runtime": 702.6065, | |
| "eval_samples_per_second": 2.847, | |
| "eval_steps_per_second": 1.423, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "learning_rate": 4.9299025014463665e-06, | |
| "logits/chosen": -1.8702771663665771, | |
| "logits/rejected": -1.8702771663665771, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3062, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "learning_rate": 4.912541236180779e-06, | |
| "logits/chosen": -2.0095884799957275, | |
| "logits/rejected": -2.0095884799957275, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3408, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.92, | |
| "learning_rate": 4.893298743830168e-06, | |
| "logits/chosen": -1.931212067604065, | |
| "logits/rejected": -1.930567741394043, | |
| "logps/chosen": -2.1643919944763184, | |
| "logps/rejected": -2.322953462600708, | |
| "loss": 0.3507, | |
| "rewards/accuracies": 0.02500000037252903, | |
| "rewards/chosen": -0.003725755959749222, | |
| "rewards/margins": 6.57886266708374e-06, | |
| "rewards/rejected": -0.0037323348224163055, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "learning_rate": 4.8721900291112415e-06, | |
| "logits/chosen": -2.0807600021362305, | |
| "logits/rejected": -2.0807600021362305, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3211, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "learning_rate": 4.849231551964771e-06, | |
| "logits/chosen": -1.9053341150283813, | |
| "logits/rejected": -1.9053341150283813, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3581, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 1.04, | |
| "learning_rate": 4.824441214720629e-06, | |
| "logits/chosen": -1.7558910846710205, | |
| "logits/rejected": -1.7558910846710205, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3778, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 1.08, | |
| "learning_rate": 4.7978383481380865e-06, | |
| "logits/chosen": -1.8789889812469482, | |
| "logits/rejected": -1.847283959388733, | |
| "logps/chosen": -14.500409126281738, | |
| "logps/rejected": -22.52828598022461, | |
| "loss": 0.5408, | |
| "rewards/accuracies": 0.02500000037252903, | |
| "rewards/chosen": -0.023679107427597046, | |
| "rewards/margins": 0.07598645985126495, | |
| "rewards/rejected": -0.099665567278862, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 1.12, | |
| "learning_rate": 4.769443696332272e-06, | |
| "logits/chosen": -1.9984527826309204, | |
| "logits/rejected": -1.9984527826309204, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3754, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 1.16, | |
| "learning_rate": 4.7392794005985324e-06, | |
| "logits/chosen": -1.9221277236938477, | |
| "logits/rejected": -1.9221277236938477, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3087, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 1.2, | |
| "learning_rate": 4.707368982147318e-06, | |
| "logits/chosen": -2.0028936862945557, | |
| "logits/rejected": -2.0028936862945557, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3013, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 1.2, | |
| "eval_logits/chosen": -1.990155816078186, | |
| "eval_logits/rejected": -1.8278305530548096, | |
| "eval_logps/chosen": -300.0851135253906, | |
| "eval_logps/rejected": -291.15228271484375, | |
| "eval_loss": 0.05316770076751709, | |
| "eval_rewards/accuracies": 0.4675000011920929, | |
| "eval_rewards/chosen": -0.1669376790523529, | |
| "eval_rewards/margins": 0.014124665409326553, | |
| "eval_rewards/rejected": -0.18106235563755035, | |
| "eval_runtime": 702.8753, | |
| "eval_samples_per_second": 2.845, | |
| "eval_steps_per_second": 1.423, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 1.24, | |
| "learning_rate": 4.673737323763048e-06, | |
| "logits/chosen": -1.927883505821228, | |
| "logits/rejected": -1.927883505821228, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3383, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 1.28, | |
| "learning_rate": 4.638410650401267e-06, | |
| "logits/chosen": -2.018455982208252, | |
| "logits/rejected": -2.018455982208252, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3951, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 1.32, | |
| "learning_rate": 4.601416508739211e-06, | |
| "logits/chosen": -2.0109939575195312, | |
| "logits/rejected": -2.0108282566070557, | |
| "logps/chosen": -2.7990431785583496, | |
| "logps/rejected": -4.229598045349121, | |
| "loss": 0.3279, | |
| "rewards/accuracies": 0.02500000037252903, | |
| "rewards/chosen": -0.006899011321365833, | |
| "rewards/margins": 0.006366544868797064, | |
| "rewards/rejected": -0.01326555572450161, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 1.36, | |
| "learning_rate": 4.562783745695738e-06, | |
| "logits/chosen": -1.74103581905365, | |
| "logits/rejected": -1.74103581905365, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3408, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 1.4, | |
| "learning_rate": 4.522542485937369e-06, | |
| "logits/chosen": -1.9381316900253296, | |
| "logits/rejected": -1.9381316900253296, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3754, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 1.44, | |
| "learning_rate": 4.4807241083879774e-06, | |
| "logits/chosen": -1.9683544635772705, | |
| "logits/rejected": -1.9683544635772705, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3778, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 1.48, | |
| "learning_rate": 4.437361221760449e-06, | |
| "logits/chosen": -1.944864273071289, | |
| "logits/rejected": -1.944864273071289, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3852, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 1.52, | |
| "learning_rate": 4.3924876391293915e-06, | |
| "logits/chosen": -1.887290358543396, | |
| "logits/rejected": -1.887290358543396, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3852, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 1.56, | |
| "learning_rate": 4.346138351564711e-06, | |
| "logits/chosen": -1.8793548345565796, | |
| "logits/rejected": -1.8793548345565796, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3408, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 1.6, | |
| "learning_rate": 4.2983495008466285e-06, | |
| "logits/chosen": -1.9503295421600342, | |
| "logits/rejected": -1.9503295421600342, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3433, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.6, | |
| "eval_logits/chosen": -1.9909396171569824, | |
| "eval_logits/rejected": -1.8287452459335327, | |
| "eval_logps/chosen": -301.09478759765625, | |
| "eval_logps/rejected": -292.806884765625, | |
| "eval_loss": 0.05226544290781021, | |
| "eval_rewards/accuracies": 0.4779999852180481, | |
| "eval_rewards/chosen": -0.17198601365089417, | |
| "eval_rewards/margins": 0.017349353060126305, | |
| "eval_rewards/rejected": -0.18933536112308502, | |
| "eval_runtime": 702.207, | |
| "eval_samples_per_second": 2.848, | |
| "eval_steps_per_second": 1.424, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.64, | |
| "learning_rate": 4.249158351283414e-06, | |
| "logits/chosen": -1.9229347705841064, | |
| "logits/rejected": -1.9229347705841064, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3186, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 1.68, | |
| "learning_rate": 4.198603260653792e-06, | |
| "logits/chosen": -1.6426639556884766, | |
| "logits/rejected": -1.6426639556884766, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.284, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 1.72, | |
| "learning_rate": 4.146723650296701e-06, | |
| "logits/chosen": -1.8314998149871826, | |
| "logits/rejected": -1.8095579147338867, | |
| "logps/chosen": -7.950663089752197, | |
| "logps/rejected": -6.032578468322754, | |
| "loss": 0.2585, | |
| "rewards/accuracies": 0.02500000037252903, | |
| "rewards/chosen": -0.015273292548954487, | |
| "rewards/margins": 0.006847293581813574, | |
| "rewards/rejected": -0.022120585665106773, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 1.76, | |
| "learning_rate": 4.093559974371725e-06, | |
| "logits/chosen": -2.019379138946533, | |
| "logits/rejected": -2.0201730728149414, | |
| "logps/chosen": -3.879962205886841, | |
| "logps/rejected": -14.236474990844727, | |
| "loss": 0.3847, | |
| "rewards/accuracies": 0.02500000037252903, | |
| "rewards/chosen": 0.008593808859586716, | |
| "rewards/margins": 0.04753534495830536, | |
| "rewards/rejected": -0.03894153609871864, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 1.8, | |
| "learning_rate": 4.039153688314146e-06, | |
| "logits/chosen": -1.8816545009613037, | |
| "logits/rejected": -1.8816545009613037, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3408, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 1.84, | |
| "learning_rate": 3.983547216509254e-06, | |
| "logits/chosen": -2.0326955318450928, | |
| "logits/rejected": -2.0326955318450928, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3211, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 1.88, | |
| "learning_rate": 3.92678391921108e-06, | |
| "logits/chosen": -1.758758783340454, | |
| "logits/rejected": -1.758758783340454, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3087, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 1.92, | |
| "learning_rate": 3.868908058731376e-06, | |
| "logits/chosen": -1.7623834609985352, | |
| "logits/rejected": -1.7623834609985352, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3606, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 1.96, | |
| "learning_rate": 3.8099647649251984e-06, | |
| "logits/chosen": -1.9364054203033447, | |
| "logits/rejected": -1.9364054203033447, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3062, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "learning_rate": 3.7500000000000005e-06, | |
| "logits/chosen": -1.7947114706039429, | |
| "logits/rejected": -1.7947114706039429, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3606, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_logits/chosen": -2.01118803024292, | |
| "eval_logits/rejected": -1.84914231300354, | |
| "eval_logps/chosen": -296.96820068359375, | |
| "eval_logps/rejected": -284.247802734375, | |
| "eval_loss": 0.06230342388153076, | |
| "eval_rewards/accuracies": 0.4050000011920929, | |
| "eval_rewards/chosen": -0.1513528972864151, | |
| "eval_rewards/margins": -0.004812860395759344, | |
| "eval_rewards/rejected": -0.14654004573822021, | |
| "eval_runtime": 706.3052, | |
| "eval_samples_per_second": 2.832, | |
| "eval_steps_per_second": 1.416, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 2.04, | |
| "learning_rate": 3.689060522675689e-06, | |
| "logits/chosen": -1.8099037408828735, | |
| "logits/rejected": -1.809592843055725, | |
| "logps/chosen": -1.9884506464004517, | |
| "logps/rejected": -3.5382981300354004, | |
| "loss": 0.2955, | |
| "rewards/accuracies": 0.02500000037252903, | |
| "rewards/chosen": -0.0028460482135415077, | |
| "rewards/margins": 0.006963008549064398, | |
| "rewards/rejected": -0.009809056296944618, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 2.08, | |
| "learning_rate": 3.627193851723577e-06, | |
| "logits/chosen": -1.992038369178772, | |
| "logits/rejected": -1.992038369178772, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3383, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 2.12, | |
| "learning_rate": 3.564448228912682e-06, | |
| "logits/chosen": -2.0280678272247314, | |
| "logits/rejected": -2.0280678272247314, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3408, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 2.16, | |
| "learning_rate": 3.5008725813922383e-06, | |
| "logits/chosen": -1.9298874139785767, | |
| "logits/rejected": -1.9298874139785767, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3926, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 2.2, | |
| "learning_rate": 3.436516483539781e-06, | |
| "logits/chosen": -1.9765506982803345, | |
| "logits/rejected": -1.9765506982803345, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3235, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 2.24, | |
| "learning_rate": 3.3714301183045382e-06, | |
| "logits/chosen": -1.7808916568756104, | |
| "logits/rejected": -1.7808916568756104, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3334, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 2.28, | |
| "learning_rate": 3.3056642380762783e-06, | |
| "logits/chosen": -1.847675085067749, | |
| "logits/rejected": -1.847675085067749, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3581, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 2.32, | |
| "learning_rate": 3.2392701251101172e-06, | |
| "logits/chosen": -1.9916722774505615, | |
| "logits/rejected": -1.9916722774505615, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3334, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 2.36, | |
| "learning_rate": 3.1722995515381644e-06, | |
| "logits/chosen": -2.0442426204681396, | |
| "logits/rejected": -2.0442426204681396, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3754, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 590 | |
| }, | |
| { | |
| "epoch": 2.4, | |
| "learning_rate": 3.1048047389991693e-06, | |
| "logits/chosen": -2.0616250038146973, | |
| "logits/rejected": -2.0616250038146973, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3038, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 2.4, | |
| "eval_logits/chosen": -2.0113329887390137, | |
| "eval_logits/rejected": -1.8490314483642578, | |
| "eval_logps/chosen": -298.9019775390625, | |
| "eval_logps/rejected": -286.5705261230469, | |
| "eval_loss": 0.06159820035099983, | |
| "eval_rewards/accuracies": 0.4090000092983246, | |
| "eval_rewards/chosen": -0.16102181375026703, | |
| "eval_rewards/margins": -0.002868250710889697, | |
| "eval_rewards/rejected": -0.15815356373786926, | |
| "eval_runtime": 704.1706, | |
| "eval_samples_per_second": 2.84, | |
| "eval_steps_per_second": 1.42, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 2.44, | |
| "learning_rate": 3.0368383179176584e-06, | |
| "logits/chosen": -2.056675434112549, | |
| "logits/rejected": -2.056675434112549, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3211, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 610 | |
| }, | |
| { | |
| "epoch": 2.48, | |
| "learning_rate": 2.9684532864643123e-06, | |
| "logits/chosen": -1.8173696994781494, | |
| "logits/rejected": -1.8173696994781494, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3062, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 2.52, | |
| "learning_rate": 2.8997029692295875e-06, | |
| "logits/chosen": -1.9147436618804932, | |
| "logits/rejected": -1.9147436618804932, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3457, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 630 | |
| }, | |
| { | |
| "epoch": 2.56, | |
| "learning_rate": 2.8306409756428067e-06, | |
| "logits/chosen": -1.8020957708358765, | |
| "logits/rejected": -1.8020957708358765, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3359, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 640 | |
| }, | |
| { | |
| "epoch": 2.6, | |
| "learning_rate": 2.761321158169134e-06, | |
| "logits/chosen": -1.6737282276153564, | |
| "logits/rejected": -1.6737282276153564, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.4099, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 2.64, | |
| "learning_rate": 2.6917975703170466e-06, | |
| "logits/chosen": -1.8559491634368896, | |
| "logits/rejected": -1.8559491634368896, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3902, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 2.68, | |
| "learning_rate": 2.6221244244890336e-06, | |
| "logits/chosen": -1.8110735416412354, | |
| "logits/rejected": -1.789332628250122, | |
| "logps/chosen": -5.0051374435424805, | |
| "logps/rejected": -3.856720447540283, | |
| "loss": 0.294, | |
| "rewards/accuracies": 0.02500000037252903, | |
| "rewards/chosen": -0.0005456657381728292, | |
| "rewards/margins": 0.010695630684494972, | |
| "rewards/rejected": -0.01124129630625248, | |
| "step": 670 | |
| }, | |
| { | |
| "epoch": 2.72, | |
| "learning_rate": 2.5523560497083927e-06, | |
| "logits/chosen": -1.8780758380889893, | |
| "logits/rejected": -1.8780758380889893, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3112, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 680 | |
| }, | |
| { | |
| "epoch": 2.76, | |
| "learning_rate": 2.482546849255096e-06, | |
| "logits/chosen": -1.8856204748153687, | |
| "logits/rejected": -1.8856204748153687, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3852, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 690 | |
| }, | |
| { | |
| "epoch": 2.8, | |
| "learning_rate": 2.4127512582437486e-06, | |
| "logits/chosen": -1.9257001876831055, | |
| "logits/rejected": -1.9257001876831055, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3161, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 2.8, | |
| "eval_logits/chosen": -2.0096490383148193, | |
| "eval_logits/rejected": -1.8473448753356934, | |
| "eval_logps/chosen": -299.4931640625, | |
| "eval_logps/rejected": -287.3162536621094, | |
| "eval_loss": 0.06128256395459175, | |
| "eval_rewards/accuracies": 0.4124999940395355, | |
| "eval_rewards/chosen": -0.1639777570962906, | |
| "eval_rewards/margins": -0.0020956520456820726, | |
| "eval_rewards/rejected": -0.16188210248947144, | |
| "eval_runtime": 702.8312, | |
| "eval_samples_per_second": 2.846, | |
| "eval_steps_per_second": 1.423, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 2.84, | |
| "learning_rate": 2.3430237011767166e-06, | |
| "logits/chosen": -2.005964756011963, | |
| "logits/rejected": -2.005964756011963, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3704, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 710 | |
| }, | |
| { | |
| "epoch": 2.88, | |
| "learning_rate": 2.2734185495055503e-06, | |
| "logits/chosen": -1.7200000286102295, | |
| "logits/rejected": -1.6872419118881226, | |
| "logps/chosen": -21.240198135375977, | |
| "logps/rejected": -10.208666801452637, | |
| "loss": 0.418, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": -0.057378046214580536, | |
| "rewards/margins": -0.01931058242917061, | |
| "rewards/rejected": -0.038067467510700226, | |
| "step": 720 | |
| }, | |
| { | |
| "epoch": 2.92, | |
| "learning_rate": 2.2039900792337477e-06, | |
| "logits/chosen": -1.8157488107681274, | |
| "logits/rejected": -1.8157488107681274, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3013, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 730 | |
| }, | |
| { | |
| "epoch": 2.96, | |
| "learning_rate": 2.134792428593971e-06, | |
| "logits/chosen": -1.9513260126113892, | |
| "logits/rejected": -1.9513260126113892, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3531, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 740 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "learning_rate": 2.0658795558326745e-06, | |
| "logits/chosen": -2.0181970596313477, | |
| "logits/rejected": -2.0197319984436035, | |
| "logps/chosen": -5.593586444854736, | |
| "logps/rejected": -8.769502639770508, | |
| "loss": 0.2891, | |
| "rewards/accuracies": 0.02500000037252903, | |
| "rewards/chosen": 2.568817035353277e-05, | |
| "rewards/margins": 0.011632362380623817, | |
| "rewards/rejected": -0.011606673710048199, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 3.04, | |
| "learning_rate": 1.997305197135089e-06, | |
| "logits/chosen": -1.9442050457000732, | |
| "logits/rejected": -1.9442050457000732, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3087, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 760 | |
| }, | |
| { | |
| "epoch": 3.08, | |
| "learning_rate": 1.9291228247233607e-06, | |
| "logits/chosen": -1.9681403636932373, | |
| "logits/rejected": -1.9681403636932373, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.2939, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 770 | |
| }, | |
| { | |
| "epoch": 3.12, | |
| "learning_rate": 1.8613856051605242e-06, | |
| "logits/chosen": -1.9257924556732178, | |
| "logits/rejected": -1.9257924556732178, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.4025, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 780 | |
| }, | |
| { | |
| "epoch": 3.16, | |
| "learning_rate": 1.7941463578928088e-06, | |
| "logits/chosen": -1.9661931991577148, | |
| "logits/rejected": -1.9661931991577148, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3038, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 790 | |
| }, | |
| { | |
| "epoch": 3.2, | |
| "learning_rate": 1.7274575140626318e-06, | |
| "logits/chosen": -2.049327850341797, | |
| "logits/rejected": -2.049327850341797, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3852, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 3.2, | |
| "eval_logits/chosen": -2.015657663345337, | |
| "eval_logits/rejected": -1.853660225868225, | |
| "eval_logps/chosen": -293.5318603515625, | |
| "eval_logps/rejected": -282.0105895996094, | |
| "eval_loss": 0.057407207787036896, | |
| "eval_rewards/accuracies": 0.4259999990463257, | |
| "eval_rewards/chosen": -0.1341715157032013, | |
| "eval_rewards/margins": 0.0011823932873085141, | |
| "eval_rewards/rejected": -0.1353539079427719, | |
| "eval_runtime": 703.0551, | |
| "eval_samples_per_second": 2.845, | |
| "eval_steps_per_second": 1.422, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 3.24, | |
| "learning_rate": 1.661371075624363e-06, | |
| "logits/chosen": -1.9243803024291992, | |
| "logits/rejected": -1.9243803024291992, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.2964, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 810 | |
| }, | |
| { | |
| "epoch": 3.28, | |
| "learning_rate": 1.5959385747947697e-06, | |
| "logits/chosen": -2.015688419342041, | |
| "logits/rejected": -2.015688419342041, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3433, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 820 | |
| }, | |
| { | |
| "epoch": 3.32, | |
| "learning_rate": 1.5312110338697427e-06, | |
| "logits/chosen": -1.9285329580307007, | |
| "logits/rejected": -1.9285329580307007, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3136, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 830 | |
| }, | |
| { | |
| "epoch": 3.36, | |
| "learning_rate": 1.467238925438646e-06, | |
| "logits/chosen": -1.7059694528579712, | |
| "logits/rejected": -1.6732099056243896, | |
| "logps/chosen": -13.629135131835938, | |
| "logps/rejected": -13.050885200500488, | |
| "loss": 0.3573, | |
| "rewards/accuracies": 0.02500000037252903, | |
| "rewards/chosen": -0.01932273805141449, | |
| "rewards/margins": 0.03295581787824631, | |
| "rewards/rejected": -0.0522785559296608, | |
| "step": 840 | |
| }, | |
| { | |
| "epoch": 3.4, | |
| "learning_rate": 1.4040721330273063e-06, | |
| "logits/chosen": -1.8047428131103516, | |
| "logits/rejected": -1.8059968948364258, | |
| "logps/chosen": -7.918952941894531, | |
| "logps/rejected": -12.563011169433594, | |
| "loss": 0.3272, | |
| "rewards/accuracies": 0.05000000074505806, | |
| "rewards/chosen": -0.00450493860989809, | |
| "rewards/margins": 0.018186846747994423, | |
| "rewards/rejected": -0.02269178442656994, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 3.44, | |
| "learning_rate": 1.3417599122003464e-06, | |
| "logits/chosen": -2.0053718090057373, | |
| "logits/rejected": -2.0053718090057373, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3828, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 860 | |
| }, | |
| { | |
| "epoch": 3.48, | |
| "learning_rate": 1.280350852153168e-06, | |
| "logits/chosen": -2.047586441040039, | |
| "logits/rejected": -2.047586441040039, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3211, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 870 | |
| }, | |
| { | |
| "epoch": 3.52, | |
| "learning_rate": 1.2198928378235717e-06, | |
| "logits/chosen": -1.888236403465271, | |
| "logits/rejected": -1.888236403465271, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3408, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 880 | |
| }, | |
| { | |
| "epoch": 3.56, | |
| "learning_rate": 1.160433012552508e-06, | |
| "logits/chosen": -1.9232513904571533, | |
| "logits/rejected": -1.9232513904571533, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.2791, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 890 | |
| }, | |
| { | |
| "epoch": 3.6, | |
| "learning_rate": 1.1020177413231334e-06, | |
| "logits/chosen": -1.919179916381836, | |
| "logits/rejected": -1.919179916381836, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3359, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 3.6, | |
| "eval_logits/chosen": -2.0115582942962646, | |
| "eval_logits/rejected": -1.8507084846496582, | |
| "eval_logps/chosen": -289.3095397949219, | |
| "eval_logps/rejected": -276.667236328125, | |
| "eval_loss": 0.059549059718847275, | |
| "eval_rewards/accuracies": 0.40049999952316284, | |
| "eval_rewards/chosen": -0.11305965483188629, | |
| "eval_rewards/margins": -0.004422400612384081, | |
| "eval_rewards/rejected": -0.1086372584104538, | |
| "eval_runtime": 701.696, | |
| "eval_samples_per_second": 2.85, | |
| "eval_steps_per_second": 1.425, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 3.64, | |
| "learning_rate": 1.0446925746067768e-06, | |
| "logits/chosen": -1.8642170429229736, | |
| "logits/rejected": -1.8642170429229736, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3383, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 910 | |
| }, | |
| { | |
| "epoch": 3.68, | |
| "learning_rate": 9.88502212844063e-07, | |
| "logits/chosen": -1.916985273361206, | |
| "logits/rejected": -1.916985273361206, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3433, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 920 | |
| }, | |
| { | |
| "epoch": 3.72, | |
| "learning_rate": 9.334904715888496e-07, | |
| "logits/chosen": -1.7782055139541626, | |
| "logits/rejected": -1.7782055139541626, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3606, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 930 | |
| }, | |
| { | |
| "epoch": 3.76, | |
| "learning_rate": 8.797002473421729e-07, | |
| "logits/chosen": -1.8740730285644531, | |
| "logits/rejected": -1.8740730285644531, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.2939, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 940 | |
| }, | |
| { | |
| "epoch": 3.8, | |
| "learning_rate": 8.271734841028553e-07, | |
| "logits/chosen": -1.9757732152938843, | |
| "logits/rejected": -1.9757732152938843, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3334, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 3.84, | |
| "learning_rate": 7.759511406608255e-07, | |
| "logits/chosen": -2.042243480682373, | |
| "logits/rejected": -2.042243480682373, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3556, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 960 | |
| }, | |
| { | |
| "epoch": 3.88, | |
| "learning_rate": 7.260731586586983e-07, | |
| "logits/chosen": -2.03684139251709, | |
| "logits/rejected": -2.03684139251709, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3408, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 970 | |
| }, | |
| { | |
| "epoch": 3.92, | |
| "learning_rate": 6.775784314464717e-07, | |
| "logits/chosen": -1.9843076467514038, | |
| "logits/rejected": -1.9843076467514038, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.4149, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 980 | |
| }, | |
| { | |
| "epoch": 3.96, | |
| "learning_rate": 6.305047737536707e-07, | |
| "logits/chosen": -1.831923484802246, | |
| "logits/rejected": -1.831923484802246, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3927, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 990 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "learning_rate": 5.848888922025553e-07, | |
| "logits/chosen": -1.9579887390136719, | |
| "logits/rejected": -1.9367748498916626, | |
| "logps/chosen": -4.523745536804199, | |
| "logps/rejected": -3.9445228576660156, | |
| "loss": 0.3701, | |
| "rewards/accuracies": 0.02500000037252903, | |
| "rewards/chosen": 0.00186129380017519, | |
| "rewards/margins": 0.013541601598262787, | |
| "rewards/rejected": -0.011680307798087597, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_logits/chosen": -2.0123445987701416, | |
| "eval_logits/rejected": -1.8512893915176392, | |
| "eval_logps/chosen": -289.3762512207031, | |
| "eval_logps/rejected": -276.7308654785156, | |
| "eval_loss": 0.05955318734049797, | |
| "eval_rewards/accuracies": 0.4000000059604645, | |
| "eval_rewards/chosen": -0.1133933812379837, | |
| "eval_rewards/margins": -0.004438146483153105, | |
| "eval_rewards/rejected": -0.10895523428916931, | |
| "eval_runtime": 701.4448, | |
| "eval_samples_per_second": 2.851, | |
| "eval_steps_per_second": 1.426, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 4.04, | |
| "learning_rate": 5.407663566854008e-07, | |
| "logits/chosen": -1.8984695672988892, | |
| "logits/rejected": -1.876189947128296, | |
| "logps/chosen": -15.62798023223877, | |
| "logps/rejected": -10.268167495727539, | |
| "loss": 0.3084, | |
| "rewards/accuracies": 0.02500000037252903, | |
| "rewards/chosen": -0.029316961765289307, | |
| "rewards/margins": 0.009048005566000938, | |
| "rewards/rejected": -0.038364969193935394, | |
| "step": 1010 | |
| }, | |
| { | |
| "epoch": 4.08, | |
| "learning_rate": 4.981715726281666e-07, | |
| "logits/chosen": -1.95705246925354, | |
| "logits/rejected": -1.95705246925354, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.2618, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1020 | |
| }, | |
| { | |
| "epoch": 4.12, | |
| "learning_rate": 4.5713775416217884e-07, | |
| "logits/chosen": -1.9425817728042603, | |
| "logits/rejected": -1.9425817728042603, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3655, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1030 | |
| }, | |
| { | |
| "epoch": 4.16, | |
| "learning_rate": 4.1769689822475147e-07, | |
| "logits/chosen": -2.1460211277008057, | |
| "logits/rejected": -2.1460211277008057, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3359, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1040 | |
| }, | |
| { | |
| "epoch": 4.2, | |
| "learning_rate": 3.798797596089351e-07, | |
| "logits/chosen": -2.158970594406128, | |
| "logits/rejected": -2.158970594406128, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3507, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1050 | |
| }, | |
| { | |
| "epoch": 4.24, | |
| "learning_rate": 3.4371582698185636e-07, | |
| "logits/chosen": -1.8473154306411743, | |
| "logits/rejected": -1.8473154306411743, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3235, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1060 | |
| }, | |
| { | |
| "epoch": 4.28, | |
| "learning_rate": 3.092332998903416e-07, | |
| "logits/chosen": -2.088589668273926, | |
| "logits/rejected": -2.088589668273926, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3606, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1070 | |
| }, | |
| { | |
| "epoch": 4.32, | |
| "learning_rate": 2.764590667717562e-07, | |
| "logits/chosen": -2.165351390838623, | |
| "logits/rejected": -2.165351390838623, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3902, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1080 | |
| }, | |
| { | |
| "epoch": 4.36, | |
| "learning_rate": 2.454186839872158e-07, | |
| "logits/chosen": -1.7897307872772217, | |
| "logits/rejected": -1.7897307872772217, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3803, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1090 | |
| }, | |
| { | |
| "epoch": 4.4, | |
| "learning_rate": 2.1613635589349756e-07, | |
| "logits/chosen": -2.0501227378845215, | |
| "logits/rejected": -2.0501227378845215, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.4025, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 4.4, | |
| "eval_logits/chosen": -2.012660264968872, | |
| "eval_logits/rejected": -1.8515597581863403, | |
| "eval_logps/chosen": -289.3721923828125, | |
| "eval_logps/rejected": -276.7074279785156, | |
| "eval_loss": 0.05955703556537628, | |
| "eval_rewards/accuracies": 0.40299999713897705, | |
| "eval_rewards/chosen": -0.11337299644947052, | |
| "eval_rewards/margins": -0.0045348722487688065, | |
| "eval_rewards/rejected": -0.10883813351392746, | |
| "eval_runtime": 702.6742, | |
| "eval_samples_per_second": 2.846, | |
| "eval_steps_per_second": 1.423, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 4.44, | |
| "learning_rate": 1.8863491596921745e-07, | |
| "logits/chosen": -1.87649667263031, | |
| "logits/rejected": -1.87649667263031, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3408, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1110 | |
| }, | |
| { | |
| "epoch": 4.48, | |
| "learning_rate": 1.629358090099639e-07, | |
| "logits/chosen": -1.876665472984314, | |
| "logits/rejected": -1.876665472984314, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.2692, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1120 | |
| }, | |
| { | |
| "epoch": 4.52, | |
| "learning_rate": 1.3905907440629752e-07, | |
| "logits/chosen": -2.0340044498443604, | |
| "logits/rejected": -2.0340044498443604, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3112, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1130 | |
| }, | |
| { | |
| "epoch": 4.56, | |
| "learning_rate": 1.1702333051763271e-07, | |
| "logits/chosen": -2.0787806510925293, | |
| "logits/rejected": -2.0787806510925293, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3581, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1140 | |
| }, | |
| { | |
| "epoch": 4.6, | |
| "learning_rate": 9.684576015420277e-08, | |
| "logits/chosen": -1.9570324420928955, | |
| "logits/rejected": -1.9570324420928955, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3852, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1150 | |
| }, | |
| { | |
| "epoch": 4.64, | |
| "learning_rate": 7.854209717842231e-08, | |
| "logits/chosen": -1.659864068031311, | |
| "logits/rejected": -1.659592866897583, | |
| "logps/chosen": -1.9274085760116577, | |
| "logps/rejected": -4.110763072967529, | |
| "loss": 0.3386, | |
| "rewards/accuracies": 0.02500000037252903, | |
| "rewards/chosen": -0.002540838671848178, | |
| "rewards/margins": 0.010130545124411583, | |
| "rewards/rejected": -0.012671384029090405, | |
| "step": 1160 | |
| }, | |
| { | |
| "epoch": 4.68, | |
| "learning_rate": 6.212661423609184e-08, | |
| "logits/chosen": -1.8022902011871338, | |
| "logits/rejected": -1.8022902011871338, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3655, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1170 | |
| }, | |
| { | |
| "epoch": 4.72, | |
| "learning_rate": 4.761211162702117e-08, | |
| "logits/chosen": -2.0399909019470215, | |
| "logits/rejected": -2.018575668334961, | |
| "logps/chosen": -4.519004821777344, | |
| "logps/rejected": -3.9544930458068848, | |
| "loss": 0.3454, | |
| "rewards/accuracies": 0.02500000037252903, | |
| "rewards/chosen": 0.0018849981715902686, | |
| "rewards/margins": 0.013615156523883343, | |
| "rewards/rejected": -0.01173015870153904, | |
| "step": 1180 | |
| }, | |
| { | |
| "epoch": 4.76, | |
| "learning_rate": 3.5009907323737826e-08, | |
| "logits/chosen": -1.8716793060302734, | |
| "logits/rejected": -1.8716793060302734, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3334, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1190 | |
| }, | |
| { | |
| "epoch": 4.8, | |
| "learning_rate": 2.4329828146074096e-08, | |
| "logits/chosen": -1.8407377004623413, | |
| "logits/rejected": -1.8407377004623413, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3754, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 4.8, | |
| "eval_logits/chosen": -2.012519359588623, | |
| "eval_logits/rejected": -1.8515361547470093, | |
| "eval_logps/chosen": -289.41143798828125, | |
| "eval_logps/rejected": -276.7694091796875, | |
| "eval_loss": 0.05953500419855118, | |
| "eval_rewards/accuracies": 0.4009999930858612, | |
| "eval_rewards/chosen": -0.11356925219297409, | |
| "eval_rewards/margins": -0.004421224817633629, | |
| "eval_rewards/rejected": -0.1091480404138565, | |
| "eval_runtime": 705.2081, | |
| "eval_samples_per_second": 2.836, | |
| "eval_steps_per_second": 1.418, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 4.84, | |
| "learning_rate": 1.5580202098509078e-08, | |
| "logits/chosen": -2.0209155082702637, | |
| "logits/rejected": -2.022444486618042, | |
| "logps/chosen": -5.759219169616699, | |
| "logps/rejected": -8.652463912963867, | |
| "loss": 0.3135, | |
| "rewards/accuracies": 0.02500000037252903, | |
| "rewards/chosen": -0.0008024768903851509, | |
| "rewards/margins": 0.010219003073871136, | |
| "rewards/rejected": -0.011021479964256287, | |
| "step": 1210 | |
| }, | |
| { | |
| "epoch": 4.88, | |
| "learning_rate": 8.767851876239075e-09, | |
| "logits/chosen": -1.9093472957611084, | |
| "logits/rejected": -1.9093472957611084, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3112, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1220 | |
| }, | |
| { | |
| "epoch": 4.92, | |
| "learning_rate": 3.8980895450474455e-09, | |
| "logits/chosen": -1.937591314315796, | |
| "logits/rejected": -1.937591314315796, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.284, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1230 | |
| }, | |
| { | |
| "epoch": 4.96, | |
| "learning_rate": 9.747123991141193e-10, | |
| "logits/chosen": -2.029675006866455, | |
| "logits/rejected": -2.029675006866455, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3606, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1240 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "learning_rate": 0.0, | |
| "logits/chosen": -1.9245744943618774, | |
| "logits/rejected": -1.9245744943618774, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3383, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1250 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "step": 1250, | |
| "total_flos": 0.0, | |
| "train_loss": 0.3429992333650589, | |
| "train_runtime": 12841.5285, | |
| "train_samples_per_second": 0.389, | |
| "train_steps_per_second": 0.097 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 1250, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 5, | |
| "save_steps": 100, | |
| "total_flos": 0.0, | |
| "train_batch_size": 2, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |