| { | |
| "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": 700.7393, | |
| "eval_samples_per_second": 2.854, | |
| "eval_steps_per_second": 1.427, | |
| "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": -1.8702564239501953, | |
| "logits/rejected": -1.8493874073028564, | |
| "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": -2.04287052154541, | |
| "logits/rejected": -2.04287052154541, | |
| "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": -1.8564621210098267, | |
| "logits/rejected": -1.8564621210098267, | |
| "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": -1.9510726928710938, | |
| "logits/rejected": -1.9173896312713623, | |
| "logps/chosen": -12.76134967803955, | |
| "logps/rejected": -5.861204624176025, | |
| "loss": 0.3393, | |
| "rewards/accuracies": 0.02500000037252903, | |
| "rewards/chosen": -0.029967620968818665, | |
| "rewards/margins": 0.002692684531211853, | |
| "rewards/rejected": -0.03266030550003052, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.76, | |
| "learning_rate": 4.958928677033465e-06, | |
| "logits/chosen": -1.8616416454315186, | |
| "logits/rejected": -1.8616416454315186, | |
| "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": -1.911077857017517, | |
| "logits/rejected": -1.9127223491668701, | |
| "logps/chosen": -5.607743740081787, | |
| "logps/rejected": -6.2597527503967285, | |
| "loss": 0.2732, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": -9.019851859193295e-05, | |
| "rewards/margins": -0.0019743461161851883, | |
| "rewards/rejected": 0.0018841475248336792, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "eval_logits/chosen": -1.9952213764190674, | |
| "eval_logits/rejected": -1.8367009162902832, | |
| "eval_logps/chosen": -270.7552795410156, | |
| "eval_logps/rejected": -259.2743835449219, | |
| "eval_loss": 0.05847138166427612, | |
| "eval_rewards/accuracies": 0.4404999911785126, | |
| "eval_rewards/chosen": -0.040576834231615067, | |
| "eval_rewards/margins": 0.0027689056005328894, | |
| "eval_rewards/rejected": -0.04334573447704315, | |
| "eval_runtime": 702.202, | |
| "eval_samples_per_second": 2.848, | |
| "eval_steps_per_second": 1.424, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "learning_rate": 4.9299025014463665e-06, | |
| "logits/chosen": -1.871522307395935, | |
| "logits/rejected": -1.871522307395935, | |
| "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": -2.014587640762329, | |
| "logits/rejected": -2.014587640762329, | |
| "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": -1.9391746520996094, | |
| "logits/rejected": -1.9385459423065186, | |
| "logps/chosen": -2.180995464324951, | |
| "logps/rejected": -2.3552231788635254, | |
| "loss": 0.3505, | |
| "rewards/accuracies": 0.02500000037252903, | |
| "rewards/chosen": -0.007617546711117029, | |
| "rewards/margins": 0.00016982034139800817, | |
| "rewards/rejected": -0.007787366863340139, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "learning_rate": 4.8721900291112415e-06, | |
| "logits/chosen": -2.1026082038879395, | |
| "logits/rejected": -2.1026082038879395, | |
| "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": -1.923709511756897, | |
| "logits/rejected": -1.923709511756897, | |
| "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": -1.7751576900482178, | |
| "logits/rejected": -1.7751576900482178, | |
| "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": -1.8949896097183228, | |
| "logits/rejected": -1.8624740839004517, | |
| "logps/chosen": -13.423696517944336, | |
| "logps/rejected": -20.979846954345703, | |
| "loss": 1.1297, | |
| "rewards/accuracies": 0.02500000037252903, | |
| "rewards/chosen": -0.036591093987226486, | |
| "rewards/margins": 0.14725562930107117, | |
| "rewards/rejected": -0.18384674191474915, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 1.12, | |
| "learning_rate": 4.769443696332272e-06, | |
| "logits/chosen": -1.9459644556045532, | |
| "logits/rejected": -1.9459644556045532, | |
| "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": -1.8699764013290405, | |
| "logits/rejected": -1.8699764013290405, | |
| "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": -1.9606857299804688, | |
| "logits/rejected": -1.9606857299804688, | |
| "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": -1.975152850151062, | |
| "eval_logits/rejected": -1.813112497329712, | |
| "eval_logps/chosen": -299.82257080078125, | |
| "eval_logps/rejected": -291.25750732421875, | |
| "eval_loss": 0.08001529425382614, | |
| "eval_rewards/accuracies": 0.4645000100135803, | |
| "eval_rewards/chosen": -0.3312495946884155, | |
| "eval_rewards/margins": 0.031927283853292465, | |
| "eval_rewards/rejected": -0.3631769120693207, | |
| "eval_runtime": 703.8743, | |
| "eval_samples_per_second": 2.841, | |
| "eval_steps_per_second": 1.421, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 1.24, | |
| "learning_rate": 4.673737323763048e-06, | |
| "logits/chosen": -1.8876497745513916, | |
| "logits/rejected": -1.8876497745513916, | |
| "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": -1.9741500616073608, | |
| "logits/rejected": -1.9741500616073608, | |
| "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": -1.9750694036483765, | |
| "logits/rejected": -1.974765419960022, | |
| "logps/chosen": -2.6847405433654785, | |
| "logps/rejected": -4.218203067779541, | |
| "loss": 0.3257, | |
| "rewards/accuracies": 0.02500000037252903, | |
| "rewards/chosen": -0.012654995545744896, | |
| "rewards/margins": 0.01376216672360897, | |
| "rewards/rejected": -0.026417162269353867, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 1.36, | |
| "learning_rate": 4.562783745695738e-06, | |
| "logits/chosen": -1.7031259536743164, | |
| "logits/rejected": -1.7031259536743164, | |
| "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": -1.9031295776367188, | |
| "logits/rejected": -1.9031295776367188, | |
| "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": -1.9317785501480103, | |
| "logits/rejected": -1.9317785501480103, | |
| "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": -1.8960540294647217, | |
| "logits/rejected": -1.8960540294647217, | |
| "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": -1.8491512537002563, | |
| "logits/rejected": -1.8491512537002563, | |
| "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": -1.8366947174072266, | |
| "logits/rejected": -1.8366947174072266, | |
| "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": -1.9089797735214233, | |
| "logits/rejected": -1.9089797735214233, | |
| "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": -1.9721235036849976, | |
| "eval_logits/rejected": -1.8102209568023682, | |
| "eval_logps/chosen": -300.3360595703125, | |
| "eval_logps/rejected": -291.88916015625, | |
| "eval_loss": 0.08119545131921768, | |
| "eval_rewards/accuracies": 0.4675000011920929, | |
| "eval_rewards/chosen": -0.33638474345207214, | |
| "eval_rewards/margins": 0.03310885280370712, | |
| "eval_rewards/rejected": -0.3694935739040375, | |
| "eval_runtime": 703.1815, | |
| "eval_samples_per_second": 2.844, | |
| "eval_steps_per_second": 1.422, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.64, | |
| "learning_rate": 4.249158351283414e-06, | |
| "logits/chosen": -1.8887426853179932, | |
| "logits/rejected": -1.8887426853179932, | |
| "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": -1.6066404581069946, | |
| "logits/rejected": -1.6066404581069946, | |
| "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": -1.7861597537994385, | |
| "logits/rejected": -1.764651894569397, | |
| "logps/chosen": -6.981114864349365, | |
| "logps/rejected": -4.9876909255981445, | |
| "loss": 0.2566, | |
| "rewards/accuracies": 0.02500000037252903, | |
| "rewards/chosen": -0.020851103588938713, | |
| "rewards/margins": 0.012941191904246807, | |
| "rewards/rejected": -0.033792294561862946, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 1.76, | |
| "learning_rate": 4.093559974371725e-06, | |
| "logits/chosen": -1.9894205331802368, | |
| "logits/rejected": -1.9900938272476196, | |
| "logps/chosen": -4.070672988891602, | |
| "logps/rejected": -14.303924560546875, | |
| "loss": 0.6156, | |
| "rewards/accuracies": 0.02500000037252903, | |
| "rewards/chosen": 0.015280509367585182, | |
| "rewards/margins": 0.09383808076381683, | |
| "rewards/rejected": -0.0785575658082962, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 1.8, | |
| "learning_rate": 4.039153688314146e-06, | |
| "logits/chosen": -1.858513593673706, | |
| "logits/rejected": -1.858513593673706, | |
| "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": -2.0184249877929688, | |
| "logits/rejected": -2.0184249877929688, | |
| "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": -1.7378623485565186, | |
| "logits/rejected": -1.7378623485565186, | |
| "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": -1.7385492324829102, | |
| "logits/rejected": -1.7385492324829102, | |
| "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": -1.9141228199005127, | |
| "logits/rejected": -1.9141228199005127, | |
| "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": -1.7758957147598267, | |
| "logits/rejected": -1.7758957147598267, | |
| "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": -1.9969795942306519, | |
| "eval_logits/rejected": -1.8348422050476074, | |
| "eval_logps/chosen": -298.5122985839844, | |
| "eval_logps/rejected": -284.1371154785156, | |
| "eval_loss": 0.10996392369270325, | |
| "eval_rewards/accuracies": 0.3734999895095825, | |
| "eval_rewards/chosen": -0.3181473910808563, | |
| "eval_rewards/margins": -0.026174278929829597, | |
| "eval_rewards/rejected": -0.2919731140136719, | |
| "eval_runtime": 702.666, | |
| "eval_samples_per_second": 2.846, | |
| "eval_steps_per_second": 1.423, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 2.04, | |
| "learning_rate": 3.689060522675689e-06, | |
| "logits/chosen": -1.7798951864242554, | |
| "logits/rejected": -1.7795917987823486, | |
| "logps/chosen": -2.3812079429626465, | |
| "logps/rejected": -3.1179988384246826, | |
| "loss": 0.2962, | |
| "rewards/accuracies": 0.02500000037252903, | |
| "rewards/chosen": -0.009619669988751411, | |
| "rewards/margins": 0.005795452743768692, | |
| "rewards/rejected": -0.015415122732520103, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 2.08, | |
| "learning_rate": 3.627193851723577e-06, | |
| "logits/chosen": -1.9688339233398438, | |
| "logits/rejected": -1.9688339233398438, | |
| "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": -2.014033794403076, | |
| "logits/rejected": -2.014033794403076, | |
| "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": -1.9099280834197998, | |
| "logits/rejected": -1.9099280834197998, | |
| "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": -1.9664745330810547, | |
| "logits/rejected": -1.9664745330810547, | |
| "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": -1.767690658569336, | |
| "logits/rejected": -1.767690658569336, | |
| "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": -1.8395519256591797, | |
| "logits/rejected": -1.8395519256591797, | |
| "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": -1.9773021936416626, | |
| "logits/rejected": -1.9773021936416626, | |
| "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": -2.025939464569092, | |
| "logits/rejected": -2.025939464569092, | |
| "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": -2.046207904815674, | |
| "logits/rejected": -2.046207904815674, | |
| "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": -1.9935928583145142, | |
| "eval_logits/rejected": -1.831691026687622, | |
| "eval_logps/chosen": -299.02557373046875, | |
| "eval_logps/rejected": -284.72607421875, | |
| "eval_loss": 0.10917193442583084, | |
| "eval_rewards/accuracies": 0.37700000405311584, | |
| "eval_rewards/chosen": -0.3232795298099518, | |
| "eval_rewards/margins": -0.025417106226086617, | |
| "eval_rewards/rejected": -0.2978624105453491, | |
| "eval_runtime": 703.1518, | |
| "eval_samples_per_second": 2.844, | |
| "eval_steps_per_second": 1.422, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 2.44, | |
| "learning_rate": 3.0368383179176584e-06, | |
| "logits/chosen": -2.0336036682128906, | |
| "logits/rejected": -2.0336036682128906, | |
| "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": -1.7936891317367554, | |
| "logits/rejected": -1.7936891317367554, | |
| "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": -1.8989810943603516, | |
| "logits/rejected": -1.8989810943603516, | |
| "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": -1.7690341472625732, | |
| "logits/rejected": -1.7690341472625732, | |
| "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": -1.6524708271026611, | |
| "logits/rejected": -1.6524708271026611, | |
| "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": -1.835680365562439, | |
| "logits/rejected": -1.835680365562439, | |
| "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": -1.7938703298568726, | |
| "logits/rejected": -1.7727596759796143, | |
| "logps/chosen": -5.777490615844727, | |
| "logps/rejected": -3.1091294288635254, | |
| "loss": 0.296, | |
| "rewards/accuracies": 0.02500000037252903, | |
| "rewards/chosen": -0.008814861066639423, | |
| "rewards/margins": 0.006191821303218603, | |
| "rewards/rejected": -0.01500668190419674, | |
| "step": 670 | |
| }, | |
| { | |
| "epoch": 2.72, | |
| "learning_rate": 2.5523560497083927e-06, | |
| "logits/chosen": -1.8663837909698486, | |
| "logits/rejected": -1.8663837909698486, | |
| "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": -1.8712513446807861, | |
| "logits/rejected": -1.8712513446807861, | |
| "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": -1.9196503162384033, | |
| "logits/rejected": -1.9196503162384033, | |
| "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": -1.9965624809265137, | |
| "eval_logits/rejected": -1.8345471620559692, | |
| "eval_logps/chosen": -298.41583251953125, | |
| "eval_logps/rejected": -284.232177734375, | |
| "eval_loss": 0.10689055174589157, | |
| "eval_rewards/accuracies": 0.3799999952316284, | |
| "eval_rewards/chosen": -0.317182332277298, | |
| "eval_rewards/margins": -0.024258404970169067, | |
| "eval_rewards/rejected": -0.2929239571094513, | |
| "eval_runtime": 703.3493, | |
| "eval_samples_per_second": 2.844, | |
| "eval_steps_per_second": 1.422, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 2.84, | |
| "learning_rate": 2.3430237011767166e-06, | |
| "logits/chosen": -1.9961488246917725, | |
| "logits/rejected": -1.9961488246917725, | |
| "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": -1.7007853984832764, | |
| "logits/rejected": -1.668602705001831, | |
| "logps/chosen": -22.714946746826172, | |
| "logps/rejected": -10.431631088256836, | |
| "loss": 0.5289, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": -0.12950357794761658, | |
| "rewards/margins": -0.05113900825381279, | |
| "rewards/rejected": -0.07836457341909409, | |
| "step": 720 | |
| }, | |
| { | |
| "epoch": 2.92, | |
| "learning_rate": 2.2039900792337477e-06, | |
| "logits/chosen": -1.8030157089233398, | |
| "logits/rejected": -1.8030157089233398, | |
| "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": -1.9432337284088135, | |
| "logits/rejected": -1.9432337284088135, | |
| "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": -1.9992624521255493, | |
| "logits/rejected": -2.000507116317749, | |
| "logps/chosen": -6.309609889984131, | |
| "logps/rejected": -9.059834480285645, | |
| "loss": 0.2932, | |
| "rewards/accuracies": 0.02500000037252903, | |
| "rewards/chosen": -0.007108859717845917, | |
| "rewards/margins": 0.019007809460163116, | |
| "rewards/rejected": -0.026116669178009033, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 3.04, | |
| "learning_rate": 1.997305197135089e-06, | |
| "logits/chosen": -1.9350688457489014, | |
| "logits/rejected": -1.9350688457489014, | |
| "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": -1.943377137184143, | |
| "logits/rejected": -1.943377137184143, | |
| "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": -1.9116928577423096, | |
| "logits/rejected": -1.9116928577423096, | |
| "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": -1.9572219848632812, | |
| "logits/rejected": -1.9572219848632812, | |
| "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": -2.023911952972412, | |
| "logits/rejected": -2.023911952972412, | |
| "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": -2.001863479614258, | |
| "eval_logits/rejected": -1.8408547639846802, | |
| "eval_logps/chosen": -289.7388000488281, | |
| "eval_logps/rejected": -275.51025390625, | |
| "eval_loss": 0.09182017296552658, | |
| "eval_rewards/accuracies": 0.3684999942779541, | |
| "eval_rewards/chosen": -0.2304122895002365, | |
| "eval_rewards/margins": -0.02470785565674305, | |
| "eval_rewards/rejected": -0.2057044357061386, | |
| "eval_runtime": 704.5723, | |
| "eval_samples_per_second": 2.839, | |
| "eval_steps_per_second": 1.419, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 3.24, | |
| "learning_rate": 1.661371075624363e-06, | |
| "logits/chosen": -1.9029079675674438, | |
| "logits/rejected": -1.9029079675674438, | |
| "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": -1.9884204864501953, | |
| "logits/rejected": -1.9884204864501953, | |
| "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": -1.9079113006591797, | |
| "logits/rejected": -1.9079113006591797, | |
| "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": -1.6772514581680298, | |
| "logits/rejected": -1.6457149982452393, | |
| "logps/chosen": -14.745180130004883, | |
| "logps/rejected": -14.53711223602295, | |
| "loss": 0.4833, | |
| "rewards/accuracies": 0.02500000037252903, | |
| "rewards/chosen": -0.04980592057108879, | |
| "rewards/margins": 0.06961346417665482, | |
| "rewards/rejected": -0.11941938102245331, | |
| "step": 840 | |
| }, | |
| { | |
| "epoch": 3.4, | |
| "learning_rate": 1.4040721330273063e-06, | |
| "logits/chosen": -1.7772667407989502, | |
| "logits/rejected": -1.7781718969345093, | |
| "logps/chosen": -8.622848510742188, | |
| "logps/rejected": -12.08240032196045, | |
| "loss": 0.3293, | |
| "rewards/accuracies": 0.05000000074505806, | |
| "rewards/chosen": -0.01604883186519146, | |
| "rewards/margins": 0.024528637528419495, | |
| "rewards/rejected": -0.040577471256256104, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 3.44, | |
| "learning_rate": 1.3417599122003464e-06, | |
| "logits/chosen": -1.9655358791351318, | |
| "logits/rejected": -1.9655358791351318, | |
| "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": -2.0000901222229004, | |
| "logits/rejected": -2.0000901222229004, | |
| "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": -1.8563473224639893, | |
| "logits/rejected": -1.8563473224639893, | |
| "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": -1.8875033855438232, | |
| "logits/rejected": -1.8875033855438232, | |
| "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": -1.8840911388397217, | |
| "logits/rejected": -1.8840911388397217, | |
| "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": -1.983768343925476, | |
| "eval_logits/rejected": -1.823969841003418, | |
| "eval_logps/chosen": -287.3323059082031, | |
| "eval_logps/rejected": -271.89581298828125, | |
| "eval_loss": 0.09828919917345047, | |
| "eval_rewards/accuracies": 0.34299999475479126, | |
| "eval_rewards/chosen": -0.20634719729423523, | |
| "eval_rewards/margins": -0.036787137389183044, | |
| "eval_rewards/rejected": -0.16956007480621338, | |
| "eval_runtime": 702.0734, | |
| "eval_samples_per_second": 2.849, | |
| "eval_steps_per_second": 1.424, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 3.64, | |
| "learning_rate": 1.0446925746067768e-06, | |
| "logits/chosen": -1.8285129070281982, | |
| "logits/rejected": -1.8285129070281982, | |
| "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": -1.877623200416565, | |
| "logits/rejected": -1.877623200416565, | |
| "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": -1.7445094585418701, | |
| "logits/rejected": -1.7445094585418701, | |
| "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": -1.8405430316925049, | |
| "logits/rejected": -1.8405430316925049, | |
| "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": -1.934522271156311, | |
| "logits/rejected": -1.934522271156311, | |
| "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": -2.002326011657715, | |
| "logits/rejected": -2.002326011657715, | |
| "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": -1.997032880783081, | |
| "logits/rejected": -1.997032880783081, | |
| "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": -1.9426374435424805, | |
| "logits/rejected": -1.9426374435424805, | |
| "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": -1.7942962646484375, | |
| "logits/rejected": -1.7942962646484375, | |
| "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": -1.9319026470184326, | |
| "logits/rejected": -1.9107824563980103, | |
| "logps/chosen": -4.969229698181152, | |
| "logps/rejected": -3.057560682296753, | |
| "loss": 0.3701, | |
| "rewards/accuracies": 0.02500000037252903, | |
| "rewards/chosen": -0.0007322501624003053, | |
| "rewards/margins": 0.013758744113147259, | |
| "rewards/rejected": -0.01449099462479353, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_logits/chosen": -1.9837957620620728, | |
| "eval_logits/rejected": -1.8240842819213867, | |
| "eval_logps/chosen": -287.31591796875, | |
| "eval_logps/rejected": -271.8734130859375, | |
| "eval_loss": 0.09820234775543213, | |
| "eval_rewards/accuracies": 0.34549999237060547, | |
| "eval_rewards/chosen": -0.20618313550949097, | |
| "eval_rewards/margins": -0.03684700280427933, | |
| "eval_rewards/rejected": -0.16933614015579224, | |
| "eval_runtime": 705.2974, | |
| "eval_samples_per_second": 2.836, | |
| "eval_steps_per_second": 1.418, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 4.04, | |
| "learning_rate": 5.407663566854008e-07, | |
| "logits/chosen": -1.8698651790618896, | |
| "logits/rejected": -1.8484447002410889, | |
| "logps/chosen": -16.939109802246094, | |
| "logps/rejected": -9.7978515625, | |
| "loss": 0.311, | |
| "rewards/accuracies": 0.02500000037252903, | |
| "rewards/chosen": -0.07174522429704666, | |
| "rewards/margins": 0.0002815544721670449, | |
| "rewards/rejected": -0.07202677428722382, | |
| "step": 1010 | |
| }, | |
| { | |
| "epoch": 4.08, | |
| "learning_rate": 4.981715726281666e-07, | |
| "logits/chosen": -1.9195115566253662, | |
| "logits/rejected": -1.9195115566253662, | |
| "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": -1.9020694494247437, | |
| "logits/rejected": -1.9020694494247437, | |
| "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": -2.1050262451171875, | |
| "logits/rejected": -2.1050262451171875, | |
| "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": -2.110747814178467, | |
| "logits/rejected": -2.110747814178467, | |
| "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": -1.8137989044189453, | |
| "logits/rejected": -1.8137989044189453, | |
| "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": -2.048144817352295, | |
| "logits/rejected": -2.048144817352295, | |
| "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": -2.115201473236084, | |
| "logits/rejected": -2.115201473236084, | |
| "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": -1.7584993839263916, | |
| "logits/rejected": -1.7584993839263916, | |
| "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": -2.012417793273926, | |
| "logits/rejected": -2.012417793273926, | |
| "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": -1.9857844114303589, | |
| "eval_logits/rejected": -1.8259761333465576, | |
| "eval_logps/chosen": -287.1649475097656, | |
| "eval_logps/rejected": -271.81268310546875, | |
| "eval_loss": 0.09747015684843063, | |
| "eval_rewards/accuracies": 0.34549999237060547, | |
| "eval_rewards/chosen": -0.20467324554920197, | |
| "eval_rewards/margins": -0.03594454750418663, | |
| "eval_rewards/rejected": -0.16872867941856384, | |
| "eval_runtime": 705.0834, | |
| "eval_samples_per_second": 2.837, | |
| "eval_steps_per_second": 1.418, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 4.44, | |
| "learning_rate": 1.8863491596921745e-07, | |
| "logits/chosen": -1.8358827829360962, | |
| "logits/rejected": -1.8358827829360962, | |
| "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": -1.8439744710922241, | |
| "logits/rejected": -1.8439744710922241, | |
| "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": -1.9981298446655273, | |
| "logits/rejected": -1.9981298446655273, | |
| "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": -2.0399842262268066, | |
| "logits/rejected": -2.0399842262268066, | |
| "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": -1.9201631546020508, | |
| "logits/rejected": -1.9201631546020508, | |
| "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": -1.6203396320343018, | |
| "logits/rejected": -1.6199716329574585, | |
| "logps/chosen": -2.1602871417999268, | |
| "logps/rejected": -3.176687240600586, | |
| "loss": 0.3391, | |
| "rewards/accuracies": 0.02500000037252903, | |
| "rewards/chosen": -0.007410462014377117, | |
| "rewards/margins": 0.0085915457457304, | |
| "rewards/rejected": -0.016002008691430092, | |
| "step": 1160 | |
| }, | |
| { | |
| "epoch": 4.68, | |
| "learning_rate": 6.212661423609184e-08, | |
| "logits/chosen": -1.7701547145843506, | |
| "logits/rejected": -1.7701547145843506, | |
| "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": -1.9981458187103271, | |
| "logits/rejected": -1.976782202720642, | |
| "logps/chosen": -4.941376686096191, | |
| "logps/rejected": -3.0533976554870605, | |
| "loss": 0.3454, | |
| "rewards/accuracies": 0.02500000037252903, | |
| "rewards/chosen": -0.0004537239146884531, | |
| "rewards/margins": 0.013995639979839325, | |
| "rewards/rejected": -0.014449363574385643, | |
| "step": 1180 | |
| }, | |
| { | |
| "epoch": 4.76, | |
| "learning_rate": 3.5009907323737826e-08, | |
| "logits/chosen": -1.8368213176727295, | |
| "logits/rejected": -1.8368213176727295, | |
| "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": -1.806884765625, | |
| "logits/rejected": -1.806884765625, | |
| "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": -1.9853414297103882, | |
| "eval_logits/rejected": -1.825589895248413, | |
| "eval_logps/chosen": -287.1330871582031, | |
| "eval_logps/rejected": -271.7890319824219, | |
| "eval_loss": 0.09735800325870514, | |
| "eval_rewards/accuracies": 0.3440000116825104, | |
| "eval_rewards/chosen": -0.20435477793216705, | |
| "eval_rewards/margins": -0.03586255759000778, | |
| "eval_rewards/rejected": -0.16849222779273987, | |
| "eval_runtime": 702.0059, | |
| "eval_samples_per_second": 2.849, | |
| "eval_steps_per_second": 1.424, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 4.84, | |
| "learning_rate": 1.5580202098509078e-08, | |
| "logits/chosen": -1.985327124595642, | |
| "logits/rejected": -1.9865849018096924, | |
| "logps/chosen": -6.517402648925781, | |
| "logps/rejected": -8.804891586303711, | |
| "loss": 0.3148, | |
| "rewards/accuracies": 0.02500000037252903, | |
| "rewards/chosen": -0.009186786599457264, | |
| "rewards/margins": 0.014380457811057568, | |
| "rewards/rejected": -0.02356724441051483, | |
| "step": 1210 | |
| }, | |
| { | |
| "epoch": 4.88, | |
| "learning_rate": 8.767851876239075e-09, | |
| "logits/chosen": -1.8783695697784424, | |
| "logits/rejected": -1.8783695697784424, | |
| "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": -1.9094947576522827, | |
| "logits/rejected": -1.9094947576522827, | |
| "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": -1.9814590215682983, | |
| "logits/rejected": -1.9814590215682983, | |
| "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": -1.8839404582977295, | |
| "logits/rejected": -1.8839404582977295, | |
| "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": 0.3515237546205521, | |
| "train_runtime": 12848.6235, | |
| "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 | |
| } | |