| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 1.0, | |
| "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": -1.9152824878692627, | |
| "logits/rejected": -1.9152824878692627, | |
| "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": 1 | |
| }, | |
| { | |
| "epoch": 0.01, | |
| "learning_rate": 4.0000000000000003e-07, | |
| "logits/chosen": -2.035064220428467, | |
| "logits/rejected": -2.035064220428467, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3197, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.02, | |
| "learning_rate": 8.000000000000001e-07, | |
| "logits/chosen": -2.0738348960876465, | |
| "logits/rejected": -2.0738348960876465, | |
| "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": 20 | |
| }, | |
| { | |
| "epoch": 0.02, | |
| "learning_rate": 1.2000000000000002e-06, | |
| "logits/chosen": -1.9596742391586304, | |
| "logits/rejected": -1.9596742391586304, | |
| "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": 30 | |
| }, | |
| { | |
| "epoch": 0.03, | |
| "learning_rate": 1.6000000000000001e-06, | |
| "logits/chosen": -1.9157626628875732, | |
| "logits/rejected": -1.9157626628875732, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.368, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.04, | |
| "learning_rate": 2.0000000000000003e-06, | |
| "logits/chosen": -1.9507601261138916, | |
| "logits/rejected": -1.9507601261138916, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3285, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.05, | |
| "learning_rate": 2.4000000000000003e-06, | |
| "logits/chosen": -1.9532804489135742, | |
| "logits/rejected": -1.9532804489135742, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.368, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.06, | |
| "learning_rate": 2.8000000000000003e-06, | |
| "logits/chosen": -1.8357080221176147, | |
| "logits/rejected": -1.8357080221176147, | |
| "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": 70 | |
| }, | |
| { | |
| "epoch": 0.06, | |
| "learning_rate": 3.2000000000000003e-06, | |
| "logits/chosen": -1.8322147130966187, | |
| "logits/rejected": -1.8322147130966187, | |
| "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": 80 | |
| }, | |
| { | |
| "epoch": 0.07, | |
| "learning_rate": 3.6000000000000003e-06, | |
| "logits/chosen": -1.8425270318984985, | |
| "logits/rejected": -1.8425270318984985, | |
| "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": 90 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "learning_rate": 4.000000000000001e-06, | |
| "logits/chosen": -1.8621313571929932, | |
| "logits/rejected": -1.8621313571929932, | |
| "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": 100 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "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": 727.871, | |
| "eval_samples_per_second": 2.748, | |
| "eval_steps_per_second": 1.374, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.09, | |
| "learning_rate": 4.4e-06, | |
| "logits/chosen": -1.909868597984314, | |
| "logits/rejected": -1.909868597984314, | |
| "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": 110 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "learning_rate": 4.800000000000001e-06, | |
| "logits/chosen": -1.7610801458358765, | |
| "logits/rejected": -1.7610801458358765, | |
| "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": 120 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "learning_rate": 4.999756310023261e-06, | |
| "logits/chosen": -2.119793653488159, | |
| "logits/rejected": -2.119793653488159, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.326, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.11, | |
| "learning_rate": 4.997807075247147e-06, | |
| "logits/chosen": -1.986210584640503, | |
| "logits/rejected": -1.986210584640503, | |
| "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": 140 | |
| }, | |
| { | |
| "epoch": 0.12, | |
| "learning_rate": 4.993910125649561e-06, | |
| "logits/chosen": -2.0813775062561035, | |
| "logits/rejected": -2.0813775062561035, | |
| "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": 150 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "learning_rate": 4.988068499954578e-06, | |
| "logits/chosen": -2.0440070629119873, | |
| "logits/rejected": -2.0440070629119873, | |
| "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": 160 | |
| }, | |
| { | |
| "epoch": 0.14, | |
| "learning_rate": 4.980286753286196e-06, | |
| "logits/chosen": -1.6871118545532227, | |
| "logits/rejected": -1.6871118545532227, | |
| "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": 170 | |
| }, | |
| { | |
| "epoch": 0.14, | |
| "learning_rate": 4.970570953616383e-06, | |
| "logits/chosen": -1.906797170639038, | |
| "logits/rejected": -1.906797170639038, | |
| "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": 180 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "learning_rate": 4.958928677033465e-06, | |
| "logits/chosen": -1.7077922821044922, | |
| "logits/rejected": -1.7077922821044922, | |
| "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": 190 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "learning_rate": 4.9453690018345144e-06, | |
| "logits/chosen": -1.7370048761367798, | |
| "logits/rejected": -1.7370048761367798, | |
| "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": 200 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "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": 730.7556, | |
| "eval_samples_per_second": 2.737, | |
| "eval_steps_per_second": 1.368, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.17, | |
| "learning_rate": 4.9299025014463665e-06, | |
| "logits/chosen": -1.7182317972183228, | |
| "logits/rejected": -1.7182317972183228, | |
| "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": 210 | |
| }, | |
| { | |
| "epoch": 0.18, | |
| "learning_rate": 4.912541236180779e-06, | |
| "logits/chosen": -1.905134916305542, | |
| "logits/rejected": -1.905134916305542, | |
| "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": 220 | |
| }, | |
| { | |
| "epoch": 0.18, | |
| "learning_rate": 4.893298743830168e-06, | |
| "logits/chosen": -2.071720600128174, | |
| "logits/rejected": -2.071720600128174, | |
| "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": 230 | |
| }, | |
| { | |
| "epoch": 0.19, | |
| "learning_rate": 4.8721900291112415e-06, | |
| "logits/chosen": -2.025371789932251, | |
| "logits/rejected": -2.025371789932251, | |
| "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": 240 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "learning_rate": 4.849231551964771e-06, | |
| "logits/chosen": -1.8689197301864624, | |
| "logits/rejected": -1.8689197301864624, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.363, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "learning_rate": 4.824441214720629e-06, | |
| "logits/chosen": -1.937535285949707, | |
| "logits/rejected": -1.937535285949707, | |
| "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": 260 | |
| }, | |
| { | |
| "epoch": 0.22, | |
| "learning_rate": 4.7978383481380865e-06, | |
| "logits/chosen": -2.058104991912842, | |
| "logits/rejected": -2.058104991912842, | |
| "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": 270 | |
| }, | |
| { | |
| "epoch": 0.22, | |
| "learning_rate": 4.769443696332272e-06, | |
| "logits/chosen": -1.8560088872909546, | |
| "logits/rejected": -1.8560088872909546, | |
| "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": 280 | |
| }, | |
| { | |
| "epoch": 0.23, | |
| "learning_rate": 4.7392794005985324e-06, | |
| "logits/chosen": -1.8305753469467163, | |
| "logits/rejected": -1.8305753469467163, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.363, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.24, | |
| "learning_rate": 4.707368982147318e-06, | |
| "logits/chosen": -1.7557750940322876, | |
| "logits/rejected": -1.7557750940322876, | |
| "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": 300 | |
| }, | |
| { | |
| "epoch": 0.24, | |
| "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": 730.4441, | |
| "eval_samples_per_second": 2.738, | |
| "eval_steps_per_second": 1.369, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "learning_rate": 4.673737323763048e-06, | |
| "logits/chosen": -2.111258029937744, | |
| "logits/rejected": -2.111258029937744, | |
| "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": 310 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "learning_rate": 4.638410650401267e-06, | |
| "logits/chosen": -1.8872848749160767, | |
| "logits/rejected": -1.8872848749160767, | |
| "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": 320 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "learning_rate": 4.601416508739211e-06, | |
| "logits/chosen": -1.9492517709732056, | |
| "logits/rejected": -1.9492517709732056, | |
| "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": 330 | |
| }, | |
| { | |
| "epoch": 0.27, | |
| "learning_rate": 4.562783745695738e-06, | |
| "logits/chosen": -1.8537830114364624, | |
| "logits/rejected": -1.8537830114364624, | |
| "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": 340 | |
| }, | |
| { | |
| "epoch": 0.28, | |
| "learning_rate": 4.522542485937369e-06, | |
| "logits/chosen": -2.1256401538848877, | |
| "logits/rejected": -2.1256401538848877, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3285, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "learning_rate": 4.4807241083879774e-06, | |
| "logits/chosen": -1.9888694286346436, | |
| "logits/rejected": -1.9888694286346436, | |
| "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": 360 | |
| }, | |
| { | |
| "epoch": 0.3, | |
| "learning_rate": 4.437361221760449e-06, | |
| "logits/chosen": -2.064541816711426, | |
| "logits/rejected": -2.064541816711426, | |
| "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": 370 | |
| }, | |
| { | |
| "epoch": 0.3, | |
| "learning_rate": 4.3924876391293915e-06, | |
| "logits/chosen": -2.1313929557800293, | |
| "logits/rejected": -2.1313929557800293, | |
| "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": 380 | |
| }, | |
| { | |
| "epoch": 0.31, | |
| "learning_rate": 4.346138351564711e-06, | |
| "logits/chosen": -2.0341124534606934, | |
| "logits/rejected": -2.0341124534606934, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.2914, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "learning_rate": 4.2983495008466285e-06, | |
| "logits/chosen": -1.7964853048324585, | |
| "logits/rejected": -1.7964853048324585, | |
| "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": 400 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "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": 725.8132, | |
| "eval_samples_per_second": 2.756, | |
| "eval_steps_per_second": 1.378, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "learning_rate": 4.249158351283414e-06, | |
| "logits/chosen": -1.9106441736221313, | |
| "logits/rejected": -1.9106441736221313, | |
| "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": 410 | |
| }, | |
| { | |
| "epoch": 0.34, | |
| "learning_rate": 4.198603260653792e-06, | |
| "logits/chosen": -1.9656860828399658, | |
| "logits/rejected": -1.9656860828399658, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3729, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.34, | |
| "learning_rate": 4.146723650296701e-06, | |
| "logits/chosen": -1.9669355154037476, | |
| "logits/rejected": -1.9669355154037476, | |
| "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": 430 | |
| }, | |
| { | |
| "epoch": 0.35, | |
| "learning_rate": 4.093559974371725e-06, | |
| "logits/chosen": -1.8838989734649658, | |
| "logits/rejected": -1.8838989734649658, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3976, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.36, | |
| "learning_rate": 4.039153688314146e-06, | |
| "logits/chosen": -1.890795111656189, | |
| "logits/rejected": -1.890795111656189, | |
| "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": 450 | |
| }, | |
| { | |
| "epoch": 0.37, | |
| "learning_rate": 3.983547216509254e-06, | |
| "logits/chosen": -1.9264202117919922, | |
| "logits/rejected": -1.9264202117919922, | |
| "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": 460 | |
| }, | |
| { | |
| "epoch": 0.38, | |
| "learning_rate": 3.92678391921108e-06, | |
| "logits/chosen": -1.8948800563812256, | |
| "logits/rejected": -1.8948800563812256, | |
| "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": 470 | |
| }, | |
| { | |
| "epoch": 0.38, | |
| "learning_rate": 3.868908058731376e-06, | |
| "logits/chosen": -1.9177709817886353, | |
| "logits/rejected": -1.8989362716674805, | |
| "logps/chosen": -10.277307510375977, | |
| "logps/rejected": -4.643392562866211, | |
| "loss": 0.3284, | |
| "rewards/accuracies": 0.02500000037252903, | |
| "rewards/chosen": 9.182358189718798e-05, | |
| "rewards/margins": 2.3801421775715426e-05, | |
| "rewards/rejected": 6.802215648349375e-05, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 0.39, | |
| "learning_rate": 3.8099647649251984e-06, | |
| "logits/chosen": -1.9777803421020508, | |
| "logits/rejected": -1.9777803421020508, | |
| "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": 490 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "learning_rate": 3.7500000000000005e-06, | |
| "logits/chosen": -1.8519232273101807, | |
| "logits/rejected": -1.8519232273101807, | |
| "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": 0.4, | |
| "eval_logits/chosen": -1.9201807975769043, | |
| "eval_logits/rejected": -1.7627112865447998, | |
| "eval_logps/chosen": -300.9842834472656, | |
| "eval_logps/rejected": -281.2868957519531, | |
| "eval_loss": 0.058566804975271225, | |
| "eval_rewards/accuracies": 0.3125, | |
| "eval_rewards/chosen": -0.034286707639694214, | |
| "eval_rewards/margins": -0.00793963111937046, | |
| "eval_rewards/rejected": -0.026347076520323753, | |
| "eval_runtime": 729.5568, | |
| "eval_samples_per_second": 2.741, | |
| "eval_steps_per_second": 1.371, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.41, | |
| "learning_rate": 3.689060522675689e-06, | |
| "logits/chosen": -1.7138468027114868, | |
| "logits/rejected": -1.7138468027114868, | |
| "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": 510 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "learning_rate": 3.627193851723577e-06, | |
| "logits/chosen": -1.802970290184021, | |
| "logits/rejected": -1.802970290184021, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.2643, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "learning_rate": 3.564448228912682e-06, | |
| "logits/chosen": -1.6650207042694092, | |
| "logits/rejected": -1.6650207042694092, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.368, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 0.43, | |
| "learning_rate": 3.5008725813922383e-06, | |
| "logits/chosen": -1.8073577880859375, | |
| "logits/rejected": -1.8073577880859375, | |
| "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": 540 | |
| }, | |
| { | |
| "epoch": 0.44, | |
| "learning_rate": 3.436516483539781e-06, | |
| "logits/chosen": -1.7098474502563477, | |
| "logits/rejected": -1.7098474502563477, | |
| "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": 550 | |
| }, | |
| { | |
| "epoch": 0.45, | |
| "learning_rate": 3.3714301183045382e-06, | |
| "logits/chosen": -1.8214161396026611, | |
| "logits/rejected": -1.8214161396026611, | |
| "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": 560 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "learning_rate": 3.3056642380762783e-06, | |
| "logits/chosen": -1.8090301752090454, | |
| "logits/rejected": -1.8090301752090454, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.4001, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "learning_rate": 3.2392701251101172e-06, | |
| "logits/chosen": -1.6492881774902344, | |
| "logits/rejected": -1.6492881774902344, | |
| "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": 580 | |
| }, | |
| { | |
| "epoch": 0.47, | |
| "learning_rate": 3.1722995515381644e-06, | |
| "logits/chosen": -1.7389822006225586, | |
| "logits/rejected": -1.7389822006225586, | |
| "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": 590 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "learning_rate": 3.1048047389991693e-06, | |
| "logits/chosen": -1.6927127838134766, | |
| "logits/rejected": -1.6927127838134766, | |
| "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": 600 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "eval_logits/chosen": -1.8917298316955566, | |
| "eval_logits/rejected": -1.7360574007034302, | |
| "eval_logps/chosen": -305.4412841796875, | |
| "eval_logps/rejected": -285.37774658203125, | |
| "eval_loss": 0.058740295469760895, | |
| "eval_rewards/accuracies": 0.31200000643730164, | |
| "eval_rewards/chosen": -0.038743674755096436, | |
| "eval_rewards/margins": -0.008305751718580723, | |
| "eval_rewards/rejected": -0.030437923967838287, | |
| "eval_runtime": 730.0666, | |
| "eval_samples_per_second": 2.739, | |
| "eval_steps_per_second": 1.37, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.49, | |
| "learning_rate": 3.0368383179176584e-06, | |
| "logits/chosen": -1.760236382484436, | |
| "logits/rejected": -1.760236382484436, | |
| "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": 610 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "learning_rate": 2.9684532864643123e-06, | |
| "logits/chosen": -1.7868738174438477, | |
| "logits/rejected": -1.7868738174438477, | |
| "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": 620 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "learning_rate": 2.8997029692295875e-06, | |
| "logits/chosen": -1.7506424188613892, | |
| "logits/rejected": -1.7506424188613892, | |
| "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": 630 | |
| }, | |
| { | |
| "epoch": 0.51, | |
| "learning_rate": 2.8306409756428067e-06, | |
| "logits/chosen": -1.6889375448226929, | |
| "logits/rejected": -1.6889375448226929, | |
| "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": 640 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "learning_rate": 2.761321158169134e-06, | |
| "logits/chosen": -1.7860199213027954, | |
| "logits/rejected": -1.7860199213027954, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.368, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 0.53, | |
| "learning_rate": 2.6917975703170466e-06, | |
| "logits/chosen": -1.8066167831420898, | |
| "logits/rejected": -1.8066167831420898, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.368, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 0.54, | |
| "learning_rate": 2.6221244244890336e-06, | |
| "logits/chosen": -1.998725175857544, | |
| "logits/rejected": -1.998725175857544, | |
| "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": 670 | |
| }, | |
| { | |
| "epoch": 0.54, | |
| "learning_rate": 2.5523560497083927e-06, | |
| "logits/chosen": -1.7736581563949585, | |
| "logits/rejected": -1.7736581563949585, | |
| "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": 680 | |
| }, | |
| { | |
| "epoch": 0.55, | |
| "learning_rate": 2.482546849255096e-06, | |
| "logits/chosen": -1.7235050201416016, | |
| "logits/rejected": -1.7235050201416016, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.2865, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 690 | |
| }, | |
| { | |
| "epoch": 0.56, | |
| "learning_rate": 2.4127512582437486e-06, | |
| "logits/chosen": -1.740098237991333, | |
| "logits/rejected": -1.740098237991333, | |
| "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": 700 | |
| }, | |
| { | |
| "epoch": 0.56, | |
| "eval_logits/chosen": -1.8920072317123413, | |
| "eval_logits/rejected": -1.7363018989562988, | |
| "eval_logps/chosen": -305.3719787597656, | |
| "eval_logps/rejected": -285.3293762207031, | |
| "eval_loss": 0.05872537940740585, | |
| "eval_rewards/accuracies": 0.3095000088214874, | |
| "eval_rewards/chosen": -0.038674384355545044, | |
| "eval_rewards/margins": -0.008284823969006538, | |
| "eval_rewards/rejected": -0.030389558523893356, | |
| "eval_runtime": 729.5557, | |
| "eval_samples_per_second": 2.741, | |
| "eval_steps_per_second": 1.371, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.57, | |
| "learning_rate": 2.3430237011767166e-06, | |
| "logits/chosen": -1.9098726511001587, | |
| "logits/rejected": -1.9098726511001587, | |
| "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": 710 | |
| }, | |
| { | |
| "epoch": 0.58, | |
| "learning_rate": 2.2734185495055503e-06, | |
| "logits/chosen": -1.7153629064559937, | |
| "logits/rejected": -1.7153629064559937, | |
| "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": 720 | |
| }, | |
| { | |
| "epoch": 0.58, | |
| "learning_rate": 2.2039900792337477e-06, | |
| "logits/chosen": -1.611203908920288, | |
| "logits/rejected": -1.611203908920288, | |
| "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": 730 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "learning_rate": 2.134792428593971e-06, | |
| "logits/chosen": -1.7782968282699585, | |
| "logits/rejected": -1.7782968282699585, | |
| "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": 740 | |
| }, | |
| { | |
| "epoch": 0.6, | |
| "learning_rate": 2.0658795558326745e-06, | |
| "logits/chosen": -1.8602126836776733, | |
| "logits/rejected": -1.8602126836776733, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3877, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 0.61, | |
| "learning_rate": 1.997305197135089e-06, | |
| "logits/chosen": -1.7727556228637695, | |
| "logits/rejected": -1.7727556228637695, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3309, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 760 | |
| }, | |
| { | |
| "epoch": 0.62, | |
| "learning_rate": 1.9291228247233607e-06, | |
| "logits/chosen": -1.8260834217071533, | |
| "logits/rejected": -1.8260834217071533, | |
| "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": 770 | |
| }, | |
| { | |
| "epoch": 0.62, | |
| "learning_rate": 1.8613856051605242e-06, | |
| "logits/chosen": -1.7438442707061768, | |
| "logits/rejected": -1.7438442707061768, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.2988, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 780 | |
| }, | |
| { | |
| "epoch": 0.63, | |
| "learning_rate": 1.7941463578928088e-06, | |
| "logits/chosen": -1.9674087762832642, | |
| "logits/rejected": -1.9656404256820679, | |
| "logps/chosen": -3.2285637855529785, | |
| "logps/rejected": -2.4141287803649902, | |
| "loss": 0.3516, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": -0.001752424635924399, | |
| "rewards/margins": -0.0005892428453080356, | |
| "rewards/rejected": -0.0011631817324087024, | |
| "step": 790 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "learning_rate": 1.7274575140626318e-06, | |
| "logits/chosen": -1.8979122638702393, | |
| "logits/rejected": -1.8979122638702393, | |
| "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": 800 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "eval_logits/chosen": -1.9784096479415894, | |
| "eval_logits/rejected": -1.8172101974487305, | |
| "eval_logps/chosen": -291.8357238769531, | |
| "eval_logps/rejected": -274.8357238769531, | |
| "eval_loss": 0.05687940865755081, | |
| "eval_rewards/accuracies": 0.3215000033378601, | |
| "eval_rewards/chosen": -0.02513808384537697, | |
| "eval_rewards/margins": -0.005242162849754095, | |
| "eval_rewards/rejected": -0.019895924255251884, | |
| "eval_runtime": 728.993, | |
| "eval_samples_per_second": 2.744, | |
| "eval_steps_per_second": 1.372, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.65, | |
| "learning_rate": 1.661371075624363e-06, | |
| "logits/chosen": -1.7325031757354736, | |
| "logits/rejected": -1.7325031757354736, | |
| "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": 810 | |
| }, | |
| { | |
| "epoch": 0.66, | |
| "learning_rate": 1.5959385747947697e-06, | |
| "logits/chosen": -1.9031715393066406, | |
| "logits/rejected": -1.9031715393066406, | |
| "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": 820 | |
| }, | |
| { | |
| "epoch": 0.66, | |
| "learning_rate": 1.5312110338697427e-06, | |
| "logits/chosen": -1.878941297531128, | |
| "logits/rejected": -1.8776565790176392, | |
| "logps/chosen": -4.952629089355469, | |
| "logps/rejected": -4.525970458984375, | |
| "loss": 0.3682, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": -0.000667310378048569, | |
| "rewards/margins": -0.0001743949978845194, | |
| "rewards/rejected": -0.0004929153947159648, | |
| "step": 830 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "learning_rate": 1.467238925438646e-06, | |
| "logits/chosen": -1.8778276443481445, | |
| "logits/rejected": -1.8778276443481445, | |
| "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": 840 | |
| }, | |
| { | |
| "epoch": 0.68, | |
| "learning_rate": 1.4040721330273063e-06, | |
| "logits/chosen": -1.8605072498321533, | |
| "logits/rejected": -1.8605072498321533, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.289, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 0.69, | |
| "learning_rate": 1.3417599122003464e-06, | |
| "logits/chosen": -1.521748423576355, | |
| "logits/rejected": -1.521748423576355, | |
| "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": 860 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "learning_rate": 1.280350852153168e-06, | |
| "logits/chosen": -1.753326416015625, | |
| "logits/rejected": -1.7436802387237549, | |
| "logps/chosen": -11.578987121582031, | |
| "logps/rejected": -9.436164855957031, | |
| "loss": 0.3037, | |
| "rewards/accuracies": 0.02500000037252903, | |
| "rewards/chosen": -0.0006309890886768699, | |
| "rewards/margins": 4.580840322887525e-05, | |
| "rewards/rejected": -0.0006767975282855332, | |
| "step": 870 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "learning_rate": 1.2198928378235717e-06, | |
| "logits/chosen": -1.9987977743148804, | |
| "logits/rejected": -1.9987977743148804, | |
| "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": 880 | |
| }, | |
| { | |
| "epoch": 0.71, | |
| "learning_rate": 1.160433012552508e-06, | |
| "logits/chosen": -1.9310076236724854, | |
| "logits/rejected": -1.9310076236724854, | |
| "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": 890 | |
| }, | |
| { | |
| "epoch": 0.72, | |
| "learning_rate": 1.1020177413231334e-06, | |
| "logits/chosen": -1.7657368183135986, | |
| "logits/rejected": -1.7657368183135986, | |
| "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": 900 | |
| }, | |
| { | |
| "epoch": 0.72, | |
| "eval_logits/chosen": -1.95650053024292, | |
| "eval_logits/rejected": -1.7981593608856201, | |
| "eval_logps/chosen": -291.2066650390625, | |
| "eval_logps/rejected": -277.3841552734375, | |
| "eval_loss": 0.05500446632504463, | |
| "eval_rewards/accuracies": 0.3840000033378601, | |
| "eval_rewards/chosen": -0.024509064853191376, | |
| "eval_rewards/margins": -0.002064692322164774, | |
| "eval_rewards/rejected": -0.02244437113404274, | |
| "eval_runtime": 726.2446, | |
| "eval_samples_per_second": 2.754, | |
| "eval_steps_per_second": 1.377, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "learning_rate": 1.0446925746067768e-06, | |
| "logits/chosen": -1.853788137435913, | |
| "logits/rejected": -1.853788137435913, | |
| "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": 910 | |
| }, | |
| { | |
| "epoch": 0.74, | |
| "learning_rate": 9.88502212844063e-07, | |
| "logits/chosen": -1.7569034099578857, | |
| "logits/rejected": -1.7553638219833374, | |
| "logps/chosen": -3.4119091033935547, | |
| "logps/rejected": -3.395718812942505, | |
| "loss": 0.335, | |
| "rewards/accuracies": 0.02500000037252903, | |
| "rewards/chosen": -0.0012354147620499134, | |
| "rewards/margins": 0.0007618310628458858, | |
| "rewards/rejected": -0.001997245941311121, | |
| "step": 920 | |
| }, | |
| { | |
| "epoch": 0.74, | |
| "learning_rate": 9.334904715888496e-07, | |
| "logits/chosen": -1.8945032358169556, | |
| "logits/rejected": -1.8945032358169556, | |
| "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": 930 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "learning_rate": 8.797002473421729e-07, | |
| "logits/chosen": -1.8704814910888672, | |
| "logits/rejected": -1.8704814910888672, | |
| "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": 940 | |
| }, | |
| { | |
| "epoch": 0.76, | |
| "learning_rate": 8.271734841028553e-07, | |
| "logits/chosen": -1.9085773229599, | |
| "logits/rejected": -1.9085773229599, | |
| "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": 950 | |
| }, | |
| { | |
| "epoch": 0.77, | |
| "learning_rate": 7.759511406608255e-07, | |
| "logits/chosen": -1.8419487476348877, | |
| "logits/rejected": -1.8419487476348877, | |
| "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": 960 | |
| }, | |
| { | |
| "epoch": 0.78, | |
| "learning_rate": 7.260731586586983e-07, | |
| "logits/chosen": -1.9136149883270264, | |
| "logits/rejected": -1.9136149883270264, | |
| "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": 970 | |
| }, | |
| { | |
| "epoch": 0.78, | |
| "learning_rate": 6.775784314464717e-07, | |
| "logits/chosen": -1.899609923362732, | |
| "logits/rejected": -1.899609923362732, | |
| "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": 980 | |
| }, | |
| { | |
| "epoch": 0.79, | |
| "learning_rate": 6.305047737536707e-07, | |
| "logits/chosen": -1.7761380672454834, | |
| "logits/rejected": -1.7761380672454834, | |
| "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": 990 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "learning_rate": 5.848888922025553e-07, | |
| "logits/chosen": -1.7907581329345703, | |
| "logits/rejected": -1.7907581329345703, | |
| "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": 1000 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "eval_logits/chosen": -1.948217511177063, | |
| "eval_logits/rejected": -1.7905058860778809, | |
| "eval_logps/chosen": -292.09368896484375, | |
| "eval_logps/rejected": -278.4593505859375, | |
| "eval_loss": 0.05491425469517708, | |
| "eval_rewards/accuracies": 0.38600000739097595, | |
| "eval_rewards/chosen": -0.025396063923835754, | |
| "eval_rewards/margins": -0.001876543858088553, | |
| "eval_rewards/rejected": -0.023519521579146385, | |
| "eval_runtime": 727.6638, | |
| "eval_samples_per_second": 2.749, | |
| "eval_steps_per_second": 1.374, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.81, | |
| "learning_rate": 5.407663566854008e-07, | |
| "logits/chosen": -1.9497054815292358, | |
| "logits/rejected": -1.9497054815292358, | |
| "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": 1010 | |
| }, | |
| { | |
| "epoch": 0.82, | |
| "learning_rate": 4.981715726281666e-07, | |
| "logits/chosen": -1.6699402332305908, | |
| "logits/rejected": -1.6699402332305908, | |
| "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": 1020 | |
| }, | |
| { | |
| "epoch": 0.82, | |
| "learning_rate": 4.5713775416217884e-07, | |
| "logits/chosen": -1.782236099243164, | |
| "logits/rejected": -1.782236099243164, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.368, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1030 | |
| }, | |
| { | |
| "epoch": 0.83, | |
| "learning_rate": 4.1769689822475147e-07, | |
| "logits/chosen": -1.8598238229751587, | |
| "logits/rejected": -1.8598238229751587, | |
| "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": 1040 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "learning_rate": 3.798797596089351e-07, | |
| "logits/chosen": -2.01692795753479, | |
| "logits/rejected": -2.01692795753479, | |
| "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": 1050 | |
| }, | |
| { | |
| "epoch": 0.85, | |
| "learning_rate": 3.4371582698185636e-07, | |
| "logits/chosen": -1.8278007507324219, | |
| "logits/rejected": -1.8278007507324219, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.326, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1060 | |
| }, | |
| { | |
| "epoch": 0.86, | |
| "learning_rate": 3.092332998903416e-07, | |
| "logits/chosen": -2.0067131519317627, | |
| "logits/rejected": -2.0067131519317627, | |
| "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": 1070 | |
| }, | |
| { | |
| "epoch": 0.86, | |
| "learning_rate": 2.764590667717562e-07, | |
| "logits/chosen": -1.8435783386230469, | |
| "logits/rejected": -1.8435783386230469, | |
| "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": 1080 | |
| }, | |
| { | |
| "epoch": 0.87, | |
| "learning_rate": 2.454186839872158e-07, | |
| "logits/chosen": -1.865335464477539, | |
| "logits/rejected": -1.865335464477539, | |
| "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": 0.88, | |
| "learning_rate": 2.1613635589349756e-07, | |
| "logits/chosen": -1.829878807067871, | |
| "logits/rejected": -1.8288015127182007, | |
| "logps/chosen": -3.383352756500244, | |
| "logps/rejected": -2.7885169982910156, | |
| "loss": 0.3682, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": -0.0015257485210895538, | |
| "rewards/margins": -0.00039975493564270437, | |
| "rewards/rejected": -0.001125993556343019, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "eval_logits/chosen": -1.9496779441833496, | |
| "eval_logits/rejected": -1.791935920715332, | |
| "eval_logps/chosen": -292.0441589355469, | |
| "eval_logps/rejected": -278.3316955566406, | |
| "eval_loss": 0.054946158081293106, | |
| "eval_rewards/accuracies": 0.38499999046325684, | |
| "eval_rewards/chosen": -0.025346575304865837, | |
| "eval_rewards/margins": -0.0019547012634575367, | |
| "eval_rewards/rejected": -0.02339187264442444, | |
| "eval_runtime": 727.7574, | |
| "eval_samples_per_second": 2.748, | |
| "eval_steps_per_second": 1.374, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 0.89, | |
| "learning_rate": 1.8863491596921745e-07, | |
| "logits/chosen": -1.9723230600357056, | |
| "logits/rejected": -1.9723230600357056, | |
| "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": 1110 | |
| }, | |
| { | |
| "epoch": 0.9, | |
| "learning_rate": 1.629358090099639e-07, | |
| "logits/chosen": -1.8381221294403076, | |
| "logits/rejected": -1.8381221294403076, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.3877, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1120 | |
| }, | |
| { | |
| "epoch": 0.9, | |
| "learning_rate": 1.3905907440629752e-07, | |
| "logits/chosen": -1.6961113214492798, | |
| "logits/rejected": -1.6961113214492798, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.363, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1130 | |
| }, | |
| { | |
| "epoch": 0.91, | |
| "learning_rate": 1.1702333051763271e-07, | |
| "logits/chosen": -1.796628713607788, | |
| "logits/rejected": -1.796628713607788, | |
| "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": 1140 | |
| }, | |
| { | |
| "epoch": 0.92, | |
| "learning_rate": 9.684576015420277e-08, | |
| "logits/chosen": -1.8420875072479248, | |
| "logits/rejected": -1.8420875072479248, | |
| "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": 1150 | |
| }, | |
| { | |
| "epoch": 0.93, | |
| "learning_rate": 7.854209717842231e-08, | |
| "logits/chosen": -1.8557026386260986, | |
| "logits/rejected": -1.8557026386260986, | |
| "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": 1160 | |
| }, | |
| { | |
| "epoch": 0.94, | |
| "learning_rate": 6.212661423609184e-08, | |
| "logits/chosen": -1.9385162591934204, | |
| "logits/rejected": -1.9385162591934204, | |
| "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": 0.94, | |
| "learning_rate": 4.761211162702117e-08, | |
| "logits/chosen": -1.9021384716033936, | |
| "logits/rejected": -1.9021384716033936, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.363, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1180 | |
| }, | |
| { | |
| "epoch": 0.95, | |
| "learning_rate": 3.5009907323737826e-08, | |
| "logits/chosen": -1.8466148376464844, | |
| "logits/rejected": -1.8466148376464844, | |
| "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": 1190 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "learning_rate": 2.4329828146074096e-08, | |
| "logits/chosen": -1.7545801401138306, | |
| "logits/rejected": -1.7545801401138306, | |
| "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": 1200 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "eval_logits/chosen": -1.949344515800476, | |
| "eval_logits/rejected": -1.7915234565734863, | |
| "eval_logps/chosen": -291.8377990722656, | |
| "eval_logps/rejected": -278.0787048339844, | |
| "eval_loss": 0.0550062395632267, | |
| "eval_rewards/accuracies": 0.39100000262260437, | |
| "eval_rewards/chosen": -0.02514021284878254, | |
| "eval_rewards/margins": -0.0020012950990349054, | |
| "eval_rewards/rejected": -0.02313891611993313, | |
| "eval_runtime": 729.194, | |
| "eval_samples_per_second": 2.743, | |
| "eval_steps_per_second": 1.371, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 0.97, | |
| "learning_rate": 1.5580202098509078e-08, | |
| "logits/chosen": -1.6881771087646484, | |
| "logits/rejected": -1.6881771087646484, | |
| "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": 1210 | |
| }, | |
| { | |
| "epoch": 0.98, | |
| "learning_rate": 8.767851876239075e-09, | |
| "logits/chosen": -1.9288533926010132, | |
| "logits/rejected": -1.9288533926010132, | |
| "logps/chosen": 0.0, | |
| "logps/rejected": 0.0, | |
| "loss": 0.368, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1220 | |
| }, | |
| { | |
| "epoch": 0.98, | |
| "learning_rate": 3.8980895450474455e-09, | |
| "logits/chosen": -1.8713302612304688, | |
| "logits/rejected": -1.8713302612304688, | |
| "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": 1230 | |
| }, | |
| { | |
| "epoch": 0.99, | |
| "learning_rate": 9.747123991141193e-10, | |
| "logits/chosen": -1.8912433385849, | |
| "logits/rejected": -1.8912433385849, | |
| "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": 1240 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "learning_rate": 0.0, | |
| "logits/chosen": -1.8090438842773438, | |
| "logits/rejected": -1.800258994102478, | |
| "logps/chosen": -5.058285713195801, | |
| "logps/rejected": -3.845151901245117, | |
| "loss": 0.3094, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": -0.002232617698609829, | |
| "rewards/margins": -0.0004155985952820629, | |
| "rewards/rejected": -0.0018170190742239356, | |
| "step": 1250 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 1250, | |
| "total_flos": 0.0, | |
| "train_loss": 0.3462253446340561, | |
| "train_runtime": 13267.124, | |
| "train_samples_per_second": 0.377, | |
| "train_steps_per_second": 0.094 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 1250, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 100, | |
| "total_flos": 0.0, | |
| "train_batch_size": 2, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |