| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.9986824769433466, | |
| "eval_steps": 100, | |
| "global_step": 379, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0, | |
| "grad_norm": 32.10821322863283, | |
| "learning_rate": 1.3157894736842104e-08, | |
| "logits/chosen": -2.219799041748047, | |
| "logits/rejected": -2.229109525680542, | |
| "logps/chosen": -269.856201171875, | |
| "logps/rejected": -192.3697509765625, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.01, | |
| "grad_norm": 35.006735982686784, | |
| "learning_rate": 6.578947368421052e-08, | |
| "logits/chosen": -2.276287078857422, | |
| "logits/rejected": -2.2080204486846924, | |
| "logps/chosen": -283.5397644042969, | |
| "logps/rejected": -208.59442138671875, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.40625, | |
| "rewards/chosen": 0.0013411559630185366, | |
| "rewards/margins": 0.00017823810048867017, | |
| "rewards/rejected": 0.0011629178188741207, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.03, | |
| "grad_norm": 36.439618208906154, | |
| "learning_rate": 1.3157894736842104e-07, | |
| "logits/chosen": -2.3274893760681152, | |
| "logits/rejected": -2.2322466373443604, | |
| "logps/chosen": -286.6865539550781, | |
| "logps/rejected": -206.5416717529297, | |
| "loss": 0.6926, | |
| "rewards/accuracies": 0.48750001192092896, | |
| "rewards/chosen": 0.00261278566904366, | |
| "rewards/margins": -3.606556128943339e-05, | |
| "rewards/rejected": 0.0026488511357456446, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.04, | |
| "grad_norm": 34.873528802219326, | |
| "learning_rate": 1.9736842105263157e-07, | |
| "logits/chosen": -2.3293230533599854, | |
| "logits/rejected": -2.2996506690979004, | |
| "logps/chosen": -289.71771240234375, | |
| "logps/rejected": -237.6280059814453, | |
| "loss": 0.6906, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": 0.017419874668121338, | |
| "rewards/margins": 0.004653422627598047, | |
| "rewards/rejected": 0.012766450643539429, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.05, | |
| "grad_norm": 31.898185632963774, | |
| "learning_rate": 2.631578947368421e-07, | |
| "logits/chosen": -2.3196263313293457, | |
| "logits/rejected": -2.2592759132385254, | |
| "logps/chosen": -263.24908447265625, | |
| "logps/rejected": -202.05458068847656, | |
| "loss": 0.6857, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": 0.05536944791674614, | |
| "rewards/margins": 0.01851554773747921, | |
| "rewards/rejected": 0.03685389831662178, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.07, | |
| "grad_norm": 28.076450357377052, | |
| "learning_rate": 3.2894736842105264e-07, | |
| "logits/chosen": -2.3921802043914795, | |
| "logits/rejected": -2.377410650253296, | |
| "logps/chosen": -260.9800720214844, | |
| "logps/rejected": -214.38821411132812, | |
| "loss": 0.6763, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": 0.13004128634929657, | |
| "rewards/margins": 0.03864985704421997, | |
| "rewards/rejected": 0.0913914293050766, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "grad_norm": 25.341140024963043, | |
| "learning_rate": 3.9473684210526315e-07, | |
| "logits/chosen": -2.477496385574341, | |
| "logits/rejected": -2.506354808807373, | |
| "logps/chosen": -255.5835723876953, | |
| "logps/rejected": -223.4532470703125, | |
| "loss": 0.6664, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": 0.2109464406967163, | |
| "rewards/margins": 0.04907592386007309, | |
| "rewards/rejected": 0.16187050938606262, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.09, | |
| "grad_norm": 23.40327238126855, | |
| "learning_rate": 4.6052631578947365e-07, | |
| "logits/chosen": -2.6838037967681885, | |
| "logits/rejected": -2.6064510345458984, | |
| "logps/chosen": -272.02825927734375, | |
| "logps/rejected": -207.57565307617188, | |
| "loss": 0.6576, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": 0.3294451832771301, | |
| "rewards/margins": 0.12016526609659195, | |
| "rewards/rejected": 0.20927992463111877, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.11, | |
| "grad_norm": 21.272616898577162, | |
| "learning_rate": 4.999575626062319e-07, | |
| "logits/chosen": -2.747399091720581, | |
| "logits/rejected": -2.7062249183654785, | |
| "logps/chosen": -249.28085327148438, | |
| "logps/rejected": -202.8720245361328, | |
| "loss": 0.6504, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": 0.3287069499492645, | |
| "rewards/margins": 0.0975252240896225, | |
| "rewards/rejected": 0.23118171095848083, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.12, | |
| "grad_norm": 18.3841620762551, | |
| "learning_rate": 4.994803073715569e-07, | |
| "logits/chosen": -2.788217067718506, | |
| "logits/rejected": -2.736720561981201, | |
| "logps/chosen": -260.8193359375, | |
| "logps/rejected": -202.5912322998047, | |
| "loss": 0.6444, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": 0.4205988049507141, | |
| "rewards/margins": 0.14394915103912354, | |
| "rewards/rejected": 0.2766496539115906, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "grad_norm": 18.656759383093647, | |
| "learning_rate": 4.984737660598186e-07, | |
| "logits/chosen": -2.8456408977508545, | |
| "logits/rejected": -2.7947795391082764, | |
| "logps/chosen": -223.0256805419922, | |
| "logps/rejected": -197.88247680664062, | |
| "loss": 0.6469, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": 0.3571344316005707, | |
| "rewards/margins": 0.09834496676921844, | |
| "rewards/rejected": 0.2587894797325134, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.14, | |
| "grad_norm": 20.93264303618464, | |
| "learning_rate": 4.969400741032999e-07, | |
| "logits/chosen": -2.8368711471557617, | |
| "logits/rejected": -2.8192451000213623, | |
| "logps/chosen": -234.53515625, | |
| "logps/rejected": -199.7665252685547, | |
| "loss": 0.6341, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": 0.33271390199661255, | |
| "rewards/margins": 0.14281830191612244, | |
| "rewards/rejected": 0.18989557027816772, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "grad_norm": 19.192901766300064, | |
| "learning_rate": 4.948824853131236e-07, | |
| "logits/chosen": -2.8852925300598145, | |
| "logits/rejected": -2.859867572784424, | |
| "logps/chosen": -253.8678741455078, | |
| "logps/rejected": -212.1206512451172, | |
| "loss": 0.6257, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 0.3522695302963257, | |
| "rewards/margins": 0.1734958440065384, | |
| "rewards/rejected": 0.1787737011909485, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.17, | |
| "grad_norm": 19.6177452377838, | |
| "learning_rate": 4.923053649761152e-07, | |
| "logits/chosen": -2.8169569969177246, | |
| "logits/rejected": -2.7849762439727783, | |
| "logps/chosen": -244.50808715820312, | |
| "logps/rejected": -201.7557830810547, | |
| "loss": 0.6149, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": 0.35878387093544006, | |
| "rewards/margins": 0.19476714730262756, | |
| "rewards/rejected": 0.1640167087316513, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 0.18, | |
| "grad_norm": 20.424079652700506, | |
| "learning_rate": 4.892141805936084e-07, | |
| "logits/chosen": -2.789267063140869, | |
| "logits/rejected": -2.73822283744812, | |
| "logps/chosen": -250.8653564453125, | |
| "logps/rejected": -224.6582794189453, | |
| "loss": 0.6149, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": 0.4033736288547516, | |
| "rewards/margins": 0.23386640846729279, | |
| "rewards/rejected": 0.1695072501897812, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "grad_norm": 25.663818502097516, | |
| "learning_rate": 4.856154902818431e-07, | |
| "logits/chosen": -2.8126118183135986, | |
| "logits/rejected": -2.7708253860473633, | |
| "logps/chosen": -229.9053955078125, | |
| "logps/rejected": -206.87258911132812, | |
| "loss": 0.6004, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.2887989580631256, | |
| "rewards/margins": 0.22907009720802307, | |
| "rewards/rejected": 0.05972885340452194, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "grad_norm": 24.00962454367728, | |
| "learning_rate": 4.81516928858564e-07, | |
| "logits/chosen": -2.7702012062072754, | |
| "logits/rejected": -2.7202858924865723, | |
| "logps/chosen": -271.06317138671875, | |
| "logps/rejected": -220.15475463867188, | |
| "loss": 0.5964, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 0.3578353524208069, | |
| "rewards/margins": 0.31964898109436035, | |
| "rewards/rejected": 0.03818630054593086, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.22, | |
| "grad_norm": 26.200013285180134, | |
| "learning_rate": 4.769271916453385e-07, | |
| "logits/chosen": -2.817336320877075, | |
| "logits/rejected": -2.7950100898742676, | |
| "logps/chosen": -257.9813537597656, | |
| "logps/rejected": -234.77108764648438, | |
| "loss": 0.5692, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": 0.16794440150260925, | |
| "rewards/margins": 0.3747493624687195, | |
| "rewards/rejected": -0.20680496096611023, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 0.24, | |
| "grad_norm": 31.71282371862922, | |
| "learning_rate": 4.7185601601995784e-07, | |
| "logits/chosen": -2.8321759700775146, | |
| "logits/rejected": -2.8417580127716064, | |
| "logps/chosen": -252.80313110351562, | |
| "logps/rejected": -233.3778839111328, | |
| "loss": 0.5452, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": 0.17818805575370789, | |
| "rewards/margins": 0.4816998839378357, | |
| "rewards/rejected": -0.3035118579864502, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "grad_norm": 30.203747907953968, | |
| "learning_rate": 4.6631416075805886e-07, | |
| "logits/chosen": -2.9183051586151123, | |
| "logits/rejected": -2.8942832946777344, | |
| "logps/chosen": -292.6710510253906, | |
| "logps/rejected": -278.3623962402344, | |
| "loss": 0.5293, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": 0.0014784678351134062, | |
| "rewards/margins": 0.45240503549575806, | |
| "rewards/rejected": -0.4509265422821045, | |
| "step": 95 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "grad_norm": 35.20356531841964, | |
| "learning_rate": 4.603133832077953e-07, | |
| "logits/chosen": -3.0372838973999023, | |
| "logits/rejected": -3.03391432762146, | |
| "logps/chosen": -264.54412841796875, | |
| "logps/rejected": -271.9236755371094, | |
| "loss": 0.5292, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -0.24657364189624786, | |
| "rewards/margins": 0.4146839678287506, | |
| "rewards/rejected": -0.6612575650215149, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "eval_logits/chosen": -2.981271743774414, | |
| "eval_logits/rejected": -3.007460117340088, | |
| "eval_logps/chosen": -500.8688049316406, | |
| "eval_logps/rejected": -504.0709228515625, | |
| "eval_loss": 0.6958277821540833, | |
| "eval_rewards/accuracies": 0.53125, | |
| "eval_rewards/chosen": -1.2149137258529663, | |
| "eval_rewards/margins": 0.09968078136444092, | |
| "eval_rewards/rejected": -1.3145945072174072, | |
| "eval_runtime": 97.6692, | |
| "eval_samples_per_second": 20.477, | |
| "eval_steps_per_second": 0.328, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.28, | |
| "grad_norm": 35.37658079116714, | |
| "learning_rate": 4.538664143459818e-07, | |
| "logits/chosen": -3.1525657176971436, | |
| "logits/rejected": -3.170300245285034, | |
| "logps/chosen": -301.29736328125, | |
| "logps/rejected": -306.07171630859375, | |
| "loss": 0.4875, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": -0.40715986490249634, | |
| "rewards/margins": 0.6011335849761963, | |
| "rewards/rejected": -1.0082933902740479, | |
| "step": 105 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "grad_norm": 44.5542186040465, | |
| "learning_rate": 4.4698693176863316e-07, | |
| "logits/chosen": -3.1490285396575928, | |
| "logits/rejected": -3.167726993560791, | |
| "logps/chosen": -326.18792724609375, | |
| "logps/rejected": -354.35784912109375, | |
| "loss": 0.4645, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -0.6683047413825989, | |
| "rewards/margins": 0.806756854057312, | |
| "rewards/rejected": -1.4750616550445557, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.3, | |
| "grad_norm": 46.11054779372386, | |
| "learning_rate": 4.396895306731977e-07, | |
| "logits/chosen": -3.2922375202178955, | |
| "logits/rejected": -3.3002562522888184, | |
| "logps/chosen": -316.46331787109375, | |
| "logps/rejected": -344.7227783203125, | |
| "loss": 0.478, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -0.5892685651779175, | |
| "rewards/margins": 0.8242443203926086, | |
| "rewards/rejected": -1.413512945175171, | |
| "step": 115 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "grad_norm": 39.587491646799435, | |
| "learning_rate": 4.319896928940505e-07, | |
| "logits/chosen": -3.4349207878112793, | |
| "logits/rejected": -3.4556515216827393, | |
| "logps/chosen": -430.2960510253906, | |
| "logps/rejected": -467.32061767578125, | |
| "loss": 0.4502, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -1.621132254600525, | |
| "rewards/margins": 0.7838995456695557, | |
| "rewards/rejected": -2.405031681060791, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "grad_norm": 36.84462226750656, | |
| "learning_rate": 4.2390375405693723e-07, | |
| "logits/chosen": -3.474997043609619, | |
| "logits/rejected": -3.5272536277770996, | |
| "logps/chosen": -415.80438232421875, | |
| "logps/rejected": -468.0596618652344, | |
| "loss": 0.4422, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -1.5106348991394043, | |
| "rewards/margins": 1.0294691324234009, | |
| "rewards/rejected": -2.5401039123535156, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 0.34, | |
| "grad_norm": 47.09166107339635, | |
| "learning_rate": 4.1544886892205354e-07, | |
| "logits/chosen": -3.5425872802734375, | |
| "logits/rejected": -3.5473670959472656, | |
| "logps/chosen": -377.5634765625, | |
| "logps/rejected": -425.98016357421875, | |
| "loss": 0.4589, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -0.8434383273124695, | |
| "rewards/margins": 0.9450035095214844, | |
| "rewards/rejected": -1.7884416580200195, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.36, | |
| "grad_norm": 49.77750455648269, | |
| "learning_rate": 4.0664297498928534e-07, | |
| "logits/chosen": -3.6749653816223145, | |
| "logits/rejected": -3.6941752433776855, | |
| "logps/chosen": -392.3956604003906, | |
| "logps/rejected": -454.9180603027344, | |
| "loss": 0.4261, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -1.5022857189178467, | |
| "rewards/margins": 1.0955121517181396, | |
| "rewards/rejected": -2.5977978706359863, | |
| "step": 135 | |
| }, | |
| { | |
| "epoch": 0.37, | |
| "grad_norm": 49.72958773280421, | |
| "learning_rate": 3.975047544428254e-07, | |
| "logits/chosen": -3.7381629943847656, | |
| "logits/rejected": -3.768315076828003, | |
| "logps/chosen": -459.23455810546875, | |
| "logps/rejected": -531.2666625976562, | |
| "loss": 0.4262, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -1.8736200332641602, | |
| "rewards/margins": 1.138641119003296, | |
| "rewards/rejected": -3.012261152267456, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.38, | |
| "grad_norm": 37.39808184046548, | |
| "learning_rate": 3.880535945158997e-07, | |
| "logits/chosen": -3.7747676372528076, | |
| "logits/rejected": -3.7997519969940186, | |
| "logps/chosen": -367.59576416015625, | |
| "logps/rejected": -449.857177734375, | |
| "loss": 0.4127, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -1.2651011943817139, | |
| "rewards/margins": 1.1386528015136719, | |
| "rewards/rejected": -2.403754234313965, | |
| "step": 145 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "grad_norm": 47.294132664555036, | |
| "learning_rate": 3.78309546359696e-07, | |
| "logits/chosen": -3.9040164947509766, | |
| "logits/rejected": -3.91438627243042, | |
| "logps/chosen": -410.07159423828125, | |
| "logps/rejected": -498.1920471191406, | |
| "loss": 0.4235, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -1.4813024997711182, | |
| "rewards/margins": 1.1341646909713745, | |
| "rewards/rejected": -2.6154673099517822, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.41, | |
| "grad_norm": 40.77943292899045, | |
| "learning_rate": 3.6829328250375227e-07, | |
| "logits/chosen": -3.932398557662964, | |
| "logits/rejected": -4.020986080169678, | |
| "logps/chosen": -427.6490783691406, | |
| "logps/rejected": -508.22930908203125, | |
| "loss": 0.3809, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -1.581094741821289, | |
| "rewards/margins": 1.3298404216766357, | |
| "rewards/rejected": -2.910935401916504, | |
| "step": 155 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "grad_norm": 41.2026716936047, | |
| "learning_rate": 3.580260529980584e-07, | |
| "logits/chosen": -4.01393461227417, | |
| "logits/rejected": -4.080017566680908, | |
| "logps/chosen": -387.573974609375, | |
| "logps/rejected": -487.1441345214844, | |
| "loss": 0.3859, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -1.2360130548477173, | |
| "rewards/margins": 1.409085988998413, | |
| "rewards/rejected": -2.64509916305542, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.43, | |
| "grad_norm": 40.85253910181408, | |
| "learning_rate": 3.475296403299163e-07, | |
| "logits/chosen": -4.092155456542969, | |
| "logits/rejected": -4.100491523742676, | |
| "logps/chosen": -388.4344482421875, | |
| "logps/rejected": -494.84771728515625, | |
| "loss": 0.391, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -1.4383682012557983, | |
| "rewards/margins": 1.5076179504394531, | |
| "rewards/rejected": -2.945986032485962, | |
| "step": 165 | |
| }, | |
| { | |
| "epoch": 0.45, | |
| "grad_norm": 43.630879623871635, | |
| "learning_rate": 3.36826313211205e-07, | |
| "logits/chosen": -4.23541784286499, | |
| "logits/rejected": -4.378731727600098, | |
| "logps/chosen": -423.3770446777344, | |
| "logps/rejected": -521.7530517578125, | |
| "loss": 0.3784, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -1.7406959533691406, | |
| "rewards/margins": 1.5629594326019287, | |
| "rewards/rejected": -3.3036551475524902, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "grad_norm": 40.910837008073884, | |
| "learning_rate": 3.259387793340943e-07, | |
| "logits/chosen": -4.390842437744141, | |
| "logits/rejected": -4.466560363769531, | |
| "logps/chosen": -482.41748046875, | |
| "logps/rejected": -587.3478393554688, | |
| "loss": 0.3845, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -2.141714096069336, | |
| "rewards/margins": 1.4235177040100098, | |
| "rewards/rejected": -3.5652313232421875, | |
| "step": 175 | |
| }, | |
| { | |
| "epoch": 0.47, | |
| "grad_norm": 43.846997204513016, | |
| "learning_rate": 3.14890137195437e-07, | |
| "logits/chosen": -4.389448642730713, | |
| "logits/rejected": -4.424112319946289, | |
| "logps/chosen": -481.9625549316406, | |
| "logps/rejected": -567.7194213867188, | |
| "loss": 0.3682, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -2.1195263862609863, | |
| "rewards/margins": 1.2820372581481934, | |
| "rewards/rejected": -3.4015636444091797, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.49, | |
| "grad_norm": 48.89695189111101, | |
| "learning_rate": 3.0370382709204883e-07, | |
| "logits/chosen": -4.416136264801025, | |
| "logits/rejected": -4.519248962402344, | |
| "logps/chosen": -472.10931396484375, | |
| "logps/rejected": -609.0546264648438, | |
| "loss": 0.3879, | |
| "rewards/accuracies": 0.887499988079071, | |
| "rewards/chosen": -1.9735286235809326, | |
| "rewards/margins": 1.7880016565322876, | |
| "rewards/rejected": -3.7615303993225098, | |
| "step": 185 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "grad_norm": 49.97288424600573, | |
| "learning_rate": 2.9240358139084013e-07, | |
| "logits/chosen": -4.548556804656982, | |
| "logits/rejected": -4.62185001373291, | |
| "logps/chosen": -421.85052490234375, | |
| "logps/rejected": -534.3924560546875, | |
| "loss": 0.3824, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -1.734278917312622, | |
| "rewards/margins": 1.5305709838867188, | |
| "rewards/rejected": -3.264849901199341, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.51, | |
| "grad_norm": 42.794361716855846, | |
| "learning_rate": 2.810133741793052e-07, | |
| "logits/chosen": -4.521183490753174, | |
| "logits/rejected": -4.65042781829834, | |
| "logps/chosen": -472.19500732421875, | |
| "logps/rejected": -594.0242309570312, | |
| "loss": 0.3875, | |
| "rewards/accuracies": 0.862500011920929, | |
| "rewards/chosen": -2.0321810245513916, | |
| "rewards/margins": 1.7259247303009033, | |
| "rewards/rejected": -3.758105516433716, | |
| "step": 195 | |
| }, | |
| { | |
| "epoch": 0.53, | |
| "grad_norm": 43.28984360773978, | |
| "learning_rate": 2.695573704031885e-07, | |
| "logits/chosen": -4.522828102111816, | |
| "logits/rejected": -4.58953332901001, | |
| "logps/chosen": -470.6817321777344, | |
| "logps/rejected": -592.047119140625, | |
| "loss": 0.3733, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": -1.8410589694976807, | |
| "rewards/margins": 1.6147441864013672, | |
| "rewards/rejected": -3.4558029174804688, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.53, | |
| "eval_logits/chosen": -4.407002925872803, | |
| "eval_logits/rejected": -4.5463128089904785, | |
| "eval_logps/chosen": -804.1498413085938, | |
| "eval_logps/rejected": -826.000732421875, | |
| "eval_loss": 0.9587702751159668, | |
| "eval_rewards/accuracies": 0.52734375, | |
| "eval_rewards/chosen": -4.247724533081055, | |
| "eval_rewards/margins": 0.28616809844970703, | |
| "eval_rewards/rejected": -4.53389310836792, | |
| "eval_runtime": 97.528, | |
| "eval_samples_per_second": 20.507, | |
| "eval_steps_per_second": 0.328, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.54, | |
| "grad_norm": 45.90321806356986, | |
| "learning_rate": 2.580598745992342e-07, | |
| "logits/chosen": -4.592051029205322, | |
| "logits/rejected": -4.729592800140381, | |
| "logps/chosen": -462.078857421875, | |
| "logps/rejected": -578.6028442382812, | |
| "loss": 0.3683, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": -1.6925300359725952, | |
| "rewards/margins": 1.7680349349975586, | |
| "rewards/rejected": -3.4605648517608643, | |
| "step": 205 | |
| }, | |
| { | |
| "epoch": 0.55, | |
| "grad_norm": 45.69561704557737, | |
| "learning_rate": 2.465452793317865e-07, | |
| "logits/chosen": -4.680180549621582, | |
| "logits/rejected": -4.785167694091797, | |
| "logps/chosen": -477.2386779785156, | |
| "logps/rejected": -609.8983154296875, | |
| "loss": 0.3534, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -1.8637679815292358, | |
| "rewards/margins": 1.7234246730804443, | |
| "rewards/rejected": -3.587193012237549, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.57, | |
| "grad_norm": 50.566593421325194, | |
| "learning_rate": 2.3503801344263344e-07, | |
| "logits/chosen": -4.860222339630127, | |
| "logits/rejected": -4.944591045379639, | |
| "logps/chosen": -456.22357177734375, | |
| "logps/rejected": -573.6387939453125, | |
| "loss": 0.3712, | |
| "rewards/accuracies": 0.862500011920929, | |
| "rewards/chosen": -2.1773338317871094, | |
| "rewards/margins": 1.6265901327133179, | |
| "rewards/rejected": -3.8039238452911377, | |
| "step": 215 | |
| }, | |
| { | |
| "epoch": 0.58, | |
| "grad_norm": 49.14997910435826, | |
| "learning_rate": 2.2356249022388789e-07, | |
| "logits/chosen": -4.710982322692871, | |
| "logits/rejected": -4.878017425537109, | |
| "logps/chosen": -468.77978515625, | |
| "logps/rejected": -580.1436157226562, | |
| "loss": 0.3706, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -1.8084490299224854, | |
| "rewards/margins": 1.884009599685669, | |
| "rewards/rejected": -3.6924586296081543, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "grad_norm": 48.71640334228914, | |
| "learning_rate": 2.121430556238559e-07, | |
| "logits/chosen": -4.8175554275512695, | |
| "logits/rejected": -4.996689319610596, | |
| "logps/chosen": -471.17132568359375, | |
| "logps/rejected": -604.3876342773438, | |
| "loss": 0.3386, | |
| "rewards/accuracies": 0.862500011920929, | |
| "rewards/chosen": -2.008169412612915, | |
| "rewards/margins": 1.8934684991836548, | |
| "rewards/rejected": -3.9016380310058594, | |
| "step": 225 | |
| }, | |
| { | |
| "epoch": 0.61, | |
| "grad_norm": 39.742825593785554, | |
| "learning_rate": 2.0080393659578038e-07, | |
| "logits/chosen": -4.740202903747559, | |
| "logits/rejected": -5.004001617431641, | |
| "logps/chosen": -513.0294189453125, | |
| "logps/rejected": -656.8961181640625, | |
| "loss": 0.3492, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -2.400604248046875, | |
| "rewards/margins": 2.0355873107910156, | |
| "rewards/rejected": -4.436191558837891, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.62, | |
| "grad_norm": 50.20217799621439, | |
| "learning_rate": 1.895691896990388e-07, | |
| "logits/chosen": -4.822530269622803, | |
| "logits/rejected": -4.9560699462890625, | |
| "logps/chosen": -474.0023498535156, | |
| "logps/rejected": -579.5379638671875, | |
| "loss": 0.3354, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -2.1355338096618652, | |
| "rewards/margins": 1.5200008153915405, | |
| "rewards/rejected": -3.655534267425537, | |
| "step": 235 | |
| }, | |
| { | |
| "epoch": 0.63, | |
| "grad_norm": 43.0245343621866, | |
| "learning_rate": 1.7846265006183976e-07, | |
| "logits/chosen": -4.6783013343811035, | |
| "logits/rejected": -4.8442277908325195, | |
| "logps/chosen": -524.712158203125, | |
| "logps/rejected": -632.6328125, | |
| "loss": 0.3455, | |
| "rewards/accuracies": 0.862500011920929, | |
| "rewards/chosen": -2.490675687789917, | |
| "rewards/margins": 1.6402429342269897, | |
| "rewards/rejected": -4.130918979644775, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.65, | |
| "grad_norm": 50.41005236551354, | |
| "learning_rate": 1.6750788081369948e-07, | |
| "logits/chosen": -4.760382175445557, | |
| "logits/rejected": -4.9731292724609375, | |
| "logps/chosen": -475.59765625, | |
| "logps/rejected": -638.8780517578125, | |
| "loss": 0.3533, | |
| "rewards/accuracies": 0.887499988079071, | |
| "rewards/chosen": -2.0123469829559326, | |
| "rewards/margins": 2.291203737258911, | |
| "rewards/rejected": -4.303550720214844, | |
| "step": 245 | |
| }, | |
| { | |
| "epoch": 0.66, | |
| "grad_norm": 41.2044329360216, | |
| "learning_rate": 1.5672812309497722e-07, | |
| "logits/chosen": -4.802388668060303, | |
| "logits/rejected": -4.928206443786621, | |
| "logps/chosen": -410.943115234375, | |
| "logps/rejected": -523.5797119140625, | |
| "loss": 0.3523, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": -1.6920316219329834, | |
| "rewards/margins": 1.4624649286270142, | |
| "rewards/rejected": -3.154496669769287, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "grad_norm": 44.69671736661672, | |
| "learning_rate": 1.461462467495284e-07, | |
| "logits/chosen": -4.758819580078125, | |
| "logits/rejected": -4.947306156158447, | |
| "logps/chosen": -495.4164123535156, | |
| "logps/rejected": -593.2689208984375, | |
| "loss": 0.3703, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -2.218562602996826, | |
| "rewards/margins": 1.439822793006897, | |
| "rewards/rejected": -3.6583850383758545, | |
| "step": 255 | |
| }, | |
| { | |
| "epoch": 0.69, | |
| "grad_norm": 47.815462029236265, | |
| "learning_rate": 1.357847018050843e-07, | |
| "logits/chosen": -4.747325420379639, | |
| "logits/rejected": -4.958649158477783, | |
| "logps/chosen": -592.2781372070312, | |
| "logps/rejected": -758.1558837890625, | |
| "loss": 0.3513, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": -2.605052947998047, | |
| "rewards/margins": 2.2088265419006348, | |
| "rewards/rejected": -4.81387996673584, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "grad_norm": 49.07485622430381, | |
| "learning_rate": 1.2566547084429324e-07, | |
| "logits/chosen": -4.789057731628418, | |
| "logits/rejected": -5.023972988128662, | |
| "logps/chosen": -472.7372131347656, | |
| "logps/rejected": -615.0753784179688, | |
| "loss": 0.3591, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -2.2092373371124268, | |
| "rewards/margins": 1.8947007656097412, | |
| "rewards/rejected": -4.103938579559326, | |
| "step": 265 | |
| }, | |
| { | |
| "epoch": 0.71, | |
| "grad_norm": 45.44279309151153, | |
| "learning_rate": 1.1581002236747328e-07, | |
| "logits/chosen": -4.7014241218566895, | |
| "logits/rejected": -4.898508548736572, | |
| "logps/chosen": -470.0, | |
| "logps/rejected": -624.503662109375, | |
| "loss": 0.368, | |
| "rewards/accuracies": 0.887499988079071, | |
| "rewards/chosen": -2.254678249359131, | |
| "rewards/margins": 2.110576868057251, | |
| "rewards/rejected": -4.365254878997803, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.72, | |
| "grad_norm": 53.087469451573526, | |
| "learning_rate": 1.062392652460177e-07, | |
| "logits/chosen": -4.696314811706543, | |
| "logits/rejected": -4.83748197555542, | |
| "logps/chosen": -505.9961853027344, | |
| "logps/rejected": -626.0050048828125, | |
| "loss": 0.3245, | |
| "rewards/accuracies": 0.862500011920929, | |
| "rewards/chosen": -1.9974615573883057, | |
| "rewards/margins": 1.828150987625122, | |
| "rewards/rejected": -3.8256123065948486, | |
| "step": 275 | |
| }, | |
| { | |
| "epoch": 0.74, | |
| "grad_norm": 41.5564187525106, | |
| "learning_rate": 9.697350436308427e-08, | |
| "logits/chosen": -4.737555027008057, | |
| "logits/rejected": -4.948160648345947, | |
| "logps/chosen": -507.66607666015625, | |
| "logps/rejected": -594.7119750976562, | |
| "loss": 0.3584, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -2.3149771690368652, | |
| "rewards/margins": 1.4441829919815063, | |
| "rewards/rejected": -3.759159803390503, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "grad_norm": 44.45010163032659, | |
| "learning_rate": 8.803239753567829e-08, | |
| "logits/chosen": -4.868664741516113, | |
| "logits/rejected": -4.975742340087891, | |
| "logps/chosen": -469.70263671875, | |
| "logps/rejected": -589.5235595703125, | |
| "loss": 0.3498, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": -2.1795949935913086, | |
| "rewards/margins": 1.7358201742172241, | |
| "rewards/rejected": -3.9154155254364014, | |
| "step": 285 | |
| }, | |
| { | |
| "epoch": 0.76, | |
| "grad_norm": 38.027371906598745, | |
| "learning_rate": 7.943491380952188e-08, | |
| "logits/chosen": -4.954745292663574, | |
| "logits/rejected": -5.065755844116211, | |
| "logps/chosen": -429.4737243652344, | |
| "logps/rejected": -571.6141967773438, | |
| "loss": 0.3285, | |
| "rewards/accuracies": 0.862500011920929, | |
| "rewards/chosen": -1.8640539646148682, | |
| "rewards/margins": 1.8129631280899048, | |
| "rewards/rejected": -3.6770172119140625, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.78, | |
| "grad_norm": 47.705363932547336, | |
| "learning_rate": 7.119929321518875e-08, | |
| "logits/chosen": -4.735751152038574, | |
| "logits/rejected": -5.044283866882324, | |
| "logps/chosen": -445.70367431640625, | |
| "logps/rejected": -583.5177612304688, | |
| "loss": 0.3436, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -1.9234256744384766, | |
| "rewards/margins": 1.9092767238616943, | |
| "rewards/rejected": -3.832702159881592, | |
| "step": 295 | |
| }, | |
| { | |
| "epoch": 0.79, | |
| "grad_norm": 52.02157401864257, | |
| "learning_rate": 6.334300807088508e-08, | |
| "logits/chosen": -4.835855007171631, | |
| "logits/rejected": -5.021653652191162, | |
| "logps/chosen": -518.7296142578125, | |
| "logps/rejected": -639.1578369140625, | |
| "loss": 0.3689, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -2.530418634414673, | |
| "rewards/margins": 1.8332151174545288, | |
| "rewards/rejected": -4.363633155822754, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.79, | |
| "eval_logits/chosen": -4.618500709533691, | |
| "eval_logits/rejected": -4.810915946960449, | |
| "eval_logps/chosen": -895.0606079101562, | |
| "eval_logps/rejected": -917.572265625, | |
| "eval_loss": 1.0593960285186768, | |
| "eval_rewards/accuracies": 0.5625, | |
| "eval_rewards/chosen": -5.156832218170166, | |
| "eval_rewards/margins": 0.2927757203578949, | |
| "eval_rewards/rejected": -5.449607849121094, | |
| "eval_runtime": 97.5523, | |
| "eval_samples_per_second": 20.502, | |
| "eval_steps_per_second": 0.328, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "grad_norm": 50.197907161154795, | |
| "learning_rate": 5.588272591397336e-08, | |
| "logits/chosen": -4.786068916320801, | |
| "logits/rejected": -4.9945068359375, | |
| "logps/chosen": -485.31280517578125, | |
| "logps/rejected": -626.9613647460938, | |
| "loss": 0.3457, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -2.227161169052124, | |
| "rewards/margins": 2.0316004753112793, | |
| "rewards/rejected": -4.258761405944824, | |
| "step": 305 | |
| }, | |
| { | |
| "epoch": 0.82, | |
| "grad_norm": 48.34581690217935, | |
| "learning_rate": 4.8834274139883084e-08, | |
| "logits/chosen": -4.855401039123535, | |
| "logits/rejected": -5.10861873626709, | |
| "logps/chosen": -469.53704833984375, | |
| "logps/rejected": -616.6036376953125, | |
| "loss": 0.3427, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": -2.20617413520813, | |
| "rewards/margins": 1.9504497051239014, | |
| "rewards/rejected": -4.156623840332031, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.83, | |
| "grad_norm": 43.92355028448363, | |
| "learning_rate": 4.221260642342786e-08, | |
| "logits/chosen": -4.848982810974121, | |
| "logits/rejected": -4.972110271453857, | |
| "logps/chosen": -502.3720703125, | |
| "logps/rejected": -609.923828125, | |
| "loss": 0.3357, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": -2.262599229812622, | |
| "rewards/margins": 1.6350996494293213, | |
| "rewards/rejected": -3.8976986408233643, | |
| "step": 315 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "grad_norm": 48.906535443168416, | |
| "learning_rate": 3.60317709937693e-08, | |
| "logits/chosen": -4.723662853240967, | |
| "logits/rejected": -4.999013900756836, | |
| "logps/chosen": -511.27691650390625, | |
| "logps/rejected": -656.3660888671875, | |
| "loss": 0.361, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -2.2201476097106934, | |
| "rewards/margins": 2.0478897094726562, | |
| "rewards/rejected": -4.26803731918335, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.86, | |
| "grad_norm": 38.024316440823085, | |
| "learning_rate": 3.030488083033273e-08, | |
| "logits/chosen": -4.8308892250061035, | |
| "logits/rejected": -5.002086162567139, | |
| "logps/chosen": -500.25579833984375, | |
| "logps/rejected": -703.7189331054688, | |
| "loss": 0.3434, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": -2.2706263065338135, | |
| "rewards/margins": 2.5099949836730957, | |
| "rewards/rejected": -4.780620574951172, | |
| "step": 325 | |
| }, | |
| { | |
| "epoch": 0.87, | |
| "grad_norm": 46.39394071999461, | |
| "learning_rate": 2.5044085842905683e-08, | |
| "logits/chosen": -4.880900859832764, | |
| "logits/rejected": -5.102107048034668, | |
| "logps/chosen": -522.2579345703125, | |
| "logps/rejected": -679.2675170898438, | |
| "loss": 0.375, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": -2.569988965988159, | |
| "rewards/margins": 2.0331876277923584, | |
| "rewards/rejected": -4.603176593780518, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "grad_norm": 44.94685644229088, | |
| "learning_rate": 2.0260547094942348e-08, | |
| "logits/chosen": -4.779486656188965, | |
| "logits/rejected": -4.989696502685547, | |
| "logps/chosen": -506.99920654296875, | |
| "logps/rejected": -654.0205688476562, | |
| "loss": 0.3649, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": -2.142956495285034, | |
| "rewards/margins": 2.1117589473724365, | |
| "rewards/rejected": -4.254715919494629, | |
| "step": 335 | |
| }, | |
| { | |
| "epoch": 0.9, | |
| "grad_norm": 44.46821814832556, | |
| "learning_rate": 1.5964413124758493e-08, | |
| "logits/chosen": -4.752711296081543, | |
| "logits/rejected": -4.939455986022949, | |
| "logps/chosen": -438.1668395996094, | |
| "logps/rejected": -598.4129028320312, | |
| "loss": 0.3357, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": -1.8868913650512695, | |
| "rewards/margins": 1.9845082759857178, | |
| "rewards/rejected": -3.8714001178741455, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.91, | |
| "grad_norm": 49.902887693125585, | |
| "learning_rate": 1.2164798414854071e-08, | |
| "logits/chosen": -4.861344814300537, | |
| "logits/rejected": -4.92690896987915, | |
| "logps/chosen": -514.7852783203125, | |
| "logps/rejected": -692.4171142578125, | |
| "loss": 0.3343, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -2.289367437362671, | |
| "rewards/margins": 2.1290056705474854, | |
| "rewards/rejected": -4.418373107910156, | |
| "step": 345 | |
| }, | |
| { | |
| "epoch": 0.92, | |
| "grad_norm": 46.82416143070829, | |
| "learning_rate": 8.869764055041501e-09, | |
| "logits/chosen": -4.845822811126709, | |
| "logits/rejected": -5.040514945983887, | |
| "logps/chosen": -476.73358154296875, | |
| "logps/rejected": -638.4616088867188, | |
| "loss": 0.3795, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -2.128450870513916, | |
| "rewards/margins": 1.8921457529067993, | |
| "rewards/rejected": -4.020596504211426, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.94, | |
| "grad_norm": 45.59470033908034, | |
| "learning_rate": 6.086300640404079e-09, | |
| "logits/chosen": -4.7604827880859375, | |
| "logits/rejected": -4.900928020477295, | |
| "logps/chosen": -530.9966430664062, | |
| "logps/rejected": -647.1216430664062, | |
| "loss": 0.3474, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -2.3973419666290283, | |
| "rewards/margins": 1.63766610622406, | |
| "rewards/rejected": -4.035007953643799, | |
| "step": 355 | |
| }, | |
| { | |
| "epoch": 0.95, | |
| "grad_norm": 38.33546651783018, | |
| "learning_rate": 3.82031344036729e-09, | |
| "logits/chosen": -4.72461462020874, | |
| "logits/rejected": -4.975947380065918, | |
| "logps/chosen": -490.10650634765625, | |
| "logps/rejected": -617.1689453125, | |
| "loss": 0.3241, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": -2.2612645626068115, | |
| "rewards/margins": 1.7519699335098267, | |
| "rewards/rejected": -4.013234615325928, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "grad_norm": 48.183739221855284, | |
| "learning_rate": 2.0766098703477173e-09, | |
| "logits/chosen": -4.829585552215576, | |
| "logits/rejected": -5.007233619689941, | |
| "logps/chosen": -458.9842834472656, | |
| "logps/rejected": -589.1207275390625, | |
| "loss": 0.3698, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -1.9293123483657837, | |
| "rewards/margins": 1.888193130493164, | |
| "rewards/rejected": -3.8175055980682373, | |
| "step": 365 | |
| }, | |
| { | |
| "epoch": 0.97, | |
| "grad_norm": 47.43892343620843, | |
| "learning_rate": 8.588892925590063e-10, | |
| "logits/chosen": -4.886306285858154, | |
| "logits/rejected": -5.221497535705566, | |
| "logps/chosen": -480.17462158203125, | |
| "logps/rejected": -637.3043823242188, | |
| "loss": 0.3276, | |
| "rewards/accuracies": 0.862500011920929, | |
| "rewards/chosen": -2.0167715549468994, | |
| "rewards/margins": 2.319883346557617, | |
| "rewards/rejected": -4.3366546630859375, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.99, | |
| "grad_norm": 48.471497735173386, | |
| "learning_rate": 1.6973516761317752e-10, | |
| "logits/chosen": -4.912912845611572, | |
| "logits/rejected": -5.170907020568848, | |
| "logps/chosen": -459.01531982421875, | |
| "logps/rejected": -576.1976318359375, | |
| "loss": 0.3505, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": -2.2589094638824463, | |
| "rewards/margins": 1.6588354110717773, | |
| "rewards/rejected": -3.9177448749542236, | |
| "step": 375 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 379, | |
| "total_flos": 0.0, | |
| "train_loss": 0.4399322837512537, | |
| "train_runtime": 5833.2156, | |
| "train_samples_per_second": 8.32, | |
| "train_steps_per_second": 0.065 | |
| } | |
| ], | |
| "logging_steps": 5, | |
| "max_steps": 379, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 100, | |
| "total_flos": 0.0, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |