qwen2_coder_adamw_iter5 / trainer_state.json
yiran-wang3's picture
End of training
f607847 verified
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 1.0,
"eval_steps": 100,
"global_step": 37,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"debug/policy_chosen_logits": -3.41206693649292,
"debug/policy_chosen_logps": -143.21307373046875,
"debug/policy_rejected_logits": -3.2551214694976807,
"debug/policy_rejected_logps": -150.2328338623047,
"debug/reference_chosen_logps": -143.21307373046875,
"debug/reference_rejected_logps": -150.2328338623047,
"epoch": 0.02702702702702703,
"grad_norm": 3.5140838495338116,
"learning_rate": 1e-06,
"logits/chosen": -3.41206693649292,
"logits/rejected": -3.2551214694976807,
"logps/chosen": -143.21307373046875,
"logps/rejected": -150.2328338623047,
"loss": 0.5,
"rewards/accuracies": 0.0,
"rewards/chosen": 0.0,
"rewards/margins": 0.0,
"rewards/rejected": 0.0,
"step": 1
},
{
"debug/policy_chosen_logits": -3.4049649238586426,
"debug/policy_chosen_logps": -143.7670135498047,
"debug/policy_rejected_logits": -3.4008069038391113,
"debug/policy_rejected_logps": -138.0754852294922,
"debug/reference_chosen_logps": -143.82278442382812,
"debug/reference_rejected_logps": -138.1428985595703,
"epoch": 0.05405405405405406,
"grad_norm": 3.526231895119079,
"learning_rate": 1e-06,
"logits/chosen": -3.4049649238586426,
"logits/rejected": -3.4008069038391113,
"logps/chosen": -143.7670135498047,
"logps/rejected": -138.0754852294922,
"loss": 0.4994,
"rewards/accuracies": 0.375,
"rewards/chosen": 0.0005577945848926902,
"rewards/margins": -0.00011638639261946082,
"rewards/rejected": 0.000674180977512151,
"step": 2
},
{
"debug/policy_chosen_logits": -3.2883780002593994,
"debug/policy_chosen_logps": -143.35385131835938,
"debug/policy_rejected_logits": -3.2103564739227295,
"debug/policy_rejected_logps": -139.85372924804688,
"debug/reference_chosen_logps": -143.38311767578125,
"debug/reference_rejected_logps": -139.98326110839844,
"epoch": 0.08108108108108109,
"grad_norm": 3.618148106626628,
"learning_rate": 1e-06,
"logits/chosen": -3.2883780002593994,
"logits/rejected": -3.2103564739227295,
"logps/chosen": -143.35385131835938,
"logps/rejected": -139.85372924804688,
"loss": 0.5006,
"rewards/accuracies": 0.25,
"rewards/chosen": 0.00029258732683956623,
"rewards/margins": -0.001002683537080884,
"rewards/rejected": 0.0012952708639204502,
"step": 3
},
{
"debug/policy_chosen_logits": -3.294776678085327,
"debug/policy_chosen_logps": -132.79519653320312,
"debug/policy_rejected_logits": -3.4150912761688232,
"debug/policy_rejected_logps": -138.08180236816406,
"debug/reference_chosen_logps": -132.62167358398438,
"debug/reference_rejected_logps": -137.927978515625,
"epoch": 0.10810810810810811,
"grad_norm": 3.5093502617713077,
"learning_rate": 1e-06,
"logits/chosen": -3.294776678085327,
"logits/rejected": -3.4150912761688232,
"logps/chosen": -132.79519653320312,
"logps/rejected": -138.08180236816406,
"loss": 0.4994,
"rewards/accuracies": 0.5,
"rewards/chosen": -0.0017352199647575617,
"rewards/margins": -0.0001968955621123314,
"rewards/rejected": -0.0015383241698145866,
"step": 4
},
{
"debug/policy_chosen_logits": -3.343012809753418,
"debug/policy_chosen_logps": -126.95188903808594,
"debug/policy_rejected_logits": -3.3124771118164062,
"debug/policy_rejected_logps": -129.9374237060547,
"debug/reference_chosen_logps": -126.82910919189453,
"debug/reference_rejected_logps": -129.91493225097656,
"epoch": 0.13513513513513514,
"grad_norm": 3.250245901043351,
"learning_rate": 1e-06,
"logits/chosen": -3.343012809753418,
"logits/rejected": -3.3124771118164062,
"logps/chosen": -126.95188903808594,
"logps/rejected": -129.9374237060547,
"loss": 0.5009,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.001227750675752759,
"rewards/margins": -0.001002778997644782,
"rewards/rejected": -0.00022497176541946828,
"step": 5
},
{
"debug/policy_chosen_logits": -3.2374277114868164,
"debug/policy_chosen_logps": -154.26406860351562,
"debug/policy_rejected_logits": -3.2978382110595703,
"debug/policy_rejected_logps": -143.40101623535156,
"debug/reference_chosen_logps": -153.9961395263672,
"debug/reference_rejected_logps": -143.3165283203125,
"epoch": 0.16216216216216217,
"grad_norm": 3.439711062354016,
"learning_rate": 1e-06,
"logits/chosen": -3.2374277114868164,
"logits/rejected": -3.2978382110595703,
"logps/chosen": -154.26406860351562,
"logps/rejected": -143.40101623535156,
"loss": 0.5009,
"rewards/accuracies": 0.375,
"rewards/chosen": -0.002679176162928343,
"rewards/margins": -0.0018340873066335917,
"rewards/rejected": -0.000845088972710073,
"step": 6
},
{
"debug/policy_chosen_logits": -3.2252252101898193,
"debug/policy_chosen_logps": -162.3474578857422,
"debug/policy_rejected_logits": -3.230537176132202,
"debug/policy_rejected_logps": -147.3297576904297,
"debug/reference_chosen_logps": -162.26889038085938,
"debug/reference_rejected_logps": -147.57199096679688,
"epoch": 0.1891891891891892,
"grad_norm": 3.5128248367150468,
"learning_rate": 1e-06,
"logits/chosen": -3.2252252101898193,
"logits/rejected": -3.230537176132202,
"logps/chosen": -162.3474578857422,
"logps/rejected": -147.3297576904297,
"loss": 0.5005,
"rewards/accuracies": 0.25,
"rewards/chosen": -0.0007855510921217501,
"rewards/margins": -0.003207969479262829,
"rewards/rejected": 0.0024224184453487396,
"step": 7
},
{
"debug/policy_chosen_logits": -3.3006181716918945,
"debug/policy_chosen_logps": -149.18630981445312,
"debug/policy_rejected_logits": -3.3146488666534424,
"debug/policy_rejected_logps": -147.72348022460938,
"debug/reference_chosen_logps": -149.20960998535156,
"debug/reference_rejected_logps": -147.54893493652344,
"epoch": 0.21621621621621623,
"grad_norm": 3.499360817762659,
"learning_rate": 1e-06,
"logits/chosen": -3.3006181716918945,
"logits/rejected": -3.3146488666534424,
"logps/chosen": -149.18630981445312,
"logps/rejected": -147.72348022460938,
"loss": 0.4992,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.000233030179515481,
"rewards/margins": 0.001978444866836071,
"rewards/rejected": -0.00174541468732059,
"step": 8
},
{
"debug/policy_chosen_logits": -3.3503048419952393,
"debug/policy_chosen_logps": -150.99197387695312,
"debug/policy_rejected_logits": -3.2822837829589844,
"debug/policy_rejected_logps": -145.26345825195312,
"debug/reference_chosen_logps": -151.06939697265625,
"debug/reference_rejected_logps": -145.52816772460938,
"epoch": 0.24324324324324326,
"grad_norm": 3.4264003876223215,
"learning_rate": 1e-06,
"logits/chosen": -3.3503048419952393,
"logits/rejected": -3.2822837829589844,
"logps/chosen": -150.99197387695312,
"logps/rejected": -145.26345825195312,
"loss": 0.4987,
"rewards/accuracies": 0.375,
"rewards/chosen": 0.0007742022862657905,
"rewards/margins": -0.0018729494186118245,
"rewards/rejected": 0.002647151704877615,
"step": 9
},
{
"debug/policy_chosen_logits": -3.304420232772827,
"debug/policy_chosen_logps": -133.21044921875,
"debug/policy_rejected_logits": -3.412870168685913,
"debug/policy_rejected_logps": -125.90888977050781,
"debug/reference_chosen_logps": -133.26504516601562,
"debug/reference_rejected_logps": -125.95523834228516,
"epoch": 0.2702702702702703,
"grad_norm": 3.3820605783151074,
"learning_rate": 1e-06,
"logits/chosen": -3.304420232772827,
"logits/rejected": -3.412870168685913,
"logps/chosen": -133.21044921875,
"logps/rejected": -125.90888977050781,
"loss": 0.5005,
"rewards/accuracies": 0.5,
"rewards/chosen": 0.0005457591032609344,
"rewards/margins": 8.22543297545053e-05,
"rewards/rejected": 0.0004635047516785562,
"step": 10
},
{
"debug/policy_chosen_logits": -3.3585329055786133,
"debug/policy_chosen_logps": -139.380859375,
"debug/policy_rejected_logits": -3.3599538803100586,
"debug/policy_rejected_logps": -132.3596649169922,
"debug/reference_chosen_logps": -139.71542358398438,
"debug/reference_rejected_logps": -132.61981201171875,
"epoch": 0.2972972972972973,
"grad_norm": 3.288218575224273,
"learning_rate": 1e-06,
"logits/chosen": -3.3585329055786133,
"logits/rejected": -3.3599538803100586,
"logps/chosen": -139.380859375,
"logps/rejected": -132.3596649169922,
"loss": 0.4999,
"rewards/accuracies": 0.5,
"rewards/chosen": 0.0033456420060247183,
"rewards/margins": 0.0007441901252605021,
"rewards/rejected": 0.0026014517061412334,
"step": 11
},
{
"debug/policy_chosen_logits": -3.2882096767425537,
"debug/policy_chosen_logps": -156.53334045410156,
"debug/policy_rejected_logits": -3.2960500717163086,
"debug/policy_rejected_logps": -143.13995361328125,
"debug/reference_chosen_logps": -156.6325225830078,
"debug/reference_rejected_logps": -142.83905029296875,
"epoch": 0.32432432432432434,
"grad_norm": 3.813316903837083,
"learning_rate": 1e-06,
"logits/chosen": -3.2882096767425537,
"logits/rejected": -3.2960500717163086,
"logps/chosen": -156.53334045410156,
"logps/rejected": -143.13995361328125,
"loss": 0.4989,
"rewards/accuracies": 0.625,
"rewards/chosen": 0.0009917640127241611,
"rewards/margins": 0.004000749439001083,
"rewards/rejected": -0.0030089854262769222,
"step": 12
},
{
"debug/policy_chosen_logits": -3.2771499156951904,
"debug/policy_chosen_logps": -133.85403442382812,
"debug/policy_rejected_logits": -3.3577096462249756,
"debug/policy_rejected_logps": -145.81051635742188,
"debug/reference_chosen_logps": -134.21636962890625,
"debug/reference_rejected_logps": -145.71588134765625,
"epoch": 0.35135135135135137,
"grad_norm": 3.5238547026932254,
"learning_rate": 1e-06,
"logits/chosen": -3.2771499156951904,
"logits/rejected": -3.3577096462249756,
"logps/chosen": -133.85403442382812,
"logps/rejected": -145.81051635742188,
"loss": 0.5004,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.003623266238719225,
"rewards/margins": 0.004569797310978174,
"rewards/rejected": -0.0009465311886742711,
"step": 13
},
{
"debug/policy_chosen_logits": -3.4804084300994873,
"debug/policy_chosen_logps": -165.59515380859375,
"debug/policy_rejected_logits": -3.5207834243774414,
"debug/policy_rejected_logps": -139.4119873046875,
"debug/reference_chosen_logps": -165.61618041992188,
"debug/reference_rejected_logps": -139.546142578125,
"epoch": 0.3783783783783784,
"grad_norm": 3.2715325631312377,
"learning_rate": 1e-06,
"logits/chosen": -3.4804084300994873,
"logits/rejected": -3.5207834243774414,
"logps/chosen": -165.59515380859375,
"logps/rejected": -139.4119873046875,
"loss": 0.4997,
"rewards/accuracies": 0.625,
"rewards/chosen": 0.00021032325457781553,
"rewards/margins": -0.0011312393471598625,
"rewards/rejected": 0.0013415624853223562,
"step": 14
},
{
"debug/policy_chosen_logits": -3.3723671436309814,
"debug/policy_chosen_logps": -138.40748596191406,
"debug/policy_rejected_logits": -3.2726492881774902,
"debug/policy_rejected_logps": -146.87710571289062,
"debug/reference_chosen_logps": -138.04966735839844,
"debug/reference_rejected_logps": -146.4534454345703,
"epoch": 0.40540540540540543,
"grad_norm": 3.5479335407438923,
"learning_rate": 1e-06,
"logits/chosen": -3.3723671436309814,
"logits/rejected": -3.2726492881774902,
"logps/chosen": -138.40748596191406,
"logps/rejected": -146.87710571289062,
"loss": 0.4971,
"rewards/accuracies": 0.5,
"rewards/chosen": -0.003578138304874301,
"rewards/margins": 0.0006584453512914479,
"rewards/rejected": -0.004236583597958088,
"step": 15
},
{
"debug/policy_chosen_logits": -3.280038356781006,
"debug/policy_chosen_logps": -144.99691772460938,
"debug/policy_rejected_logits": -3.265355110168457,
"debug/policy_rejected_logps": -136.1260223388672,
"debug/reference_chosen_logps": -145.38519287109375,
"debug/reference_rejected_logps": -136.1790313720703,
"epoch": 0.43243243243243246,
"grad_norm": 3.3905692790559057,
"learning_rate": 1e-06,
"logits/chosen": -3.280038356781006,
"logits/rejected": -3.265355110168457,
"logps/chosen": -144.99691772460938,
"logps/rejected": -136.1260223388672,
"loss": 0.4987,
"rewards/accuracies": 0.625,
"rewards/chosen": 0.0038827608805149794,
"rewards/margins": 0.0033527181949466467,
"rewards/rejected": 0.0005300427437759936,
"step": 16
},
{
"debug/policy_chosen_logits": -3.4201347827911377,
"debug/policy_chosen_logps": -154.7392578125,
"debug/policy_rejected_logits": -3.348952293395996,
"debug/policy_rejected_logps": -133.3173828125,
"debug/reference_chosen_logps": -154.70895385742188,
"debug/reference_rejected_logps": -133.4116668701172,
"epoch": 0.4594594594594595,
"grad_norm": 3.6801385333166086,
"learning_rate": 1e-06,
"logits/chosen": -3.4201347827911377,
"logits/rejected": -3.348952293395996,
"logps/chosen": -154.7392578125,
"logps/rejected": -133.3173828125,
"loss": 0.4999,
"rewards/accuracies": 0.375,
"rewards/chosen": -0.0003031063242815435,
"rewards/margins": -0.0012458800338208675,
"rewards/rejected": 0.0009427738841623068,
"step": 17
},
{
"debug/policy_chosen_logits": -3.2982780933380127,
"debug/policy_chosen_logps": -146.35263061523438,
"debug/policy_rejected_logits": -3.3202931880950928,
"debug/policy_rejected_logps": -156.7283172607422,
"debug/reference_chosen_logps": -146.34840393066406,
"debug/reference_rejected_logps": -156.84378051757812,
"epoch": 0.4864864864864865,
"grad_norm": 3.5196394680278327,
"learning_rate": 1e-06,
"logits/chosen": -3.2982780933380127,
"logits/rejected": -3.3202931880950928,
"logps/chosen": -146.35263061523438,
"logps/rejected": -156.7283172607422,
"loss": 0.4974,
"rewards/accuracies": 0.375,
"rewards/chosen": -4.220963455736637e-05,
"rewards/margins": -0.0011968993349000812,
"rewards/rejected": 0.0011546898167580366,
"step": 18
},
{
"debug/policy_chosen_logits": -3.3398616313934326,
"debug/policy_chosen_logps": -131.99713134765625,
"debug/policy_rejected_logits": -3.3604841232299805,
"debug/policy_rejected_logps": -148.30227661132812,
"debug/reference_chosen_logps": -131.88931274414062,
"debug/reference_rejected_logps": -148.13345336914062,
"epoch": 0.5135135135135135,
"grad_norm": 3.609790844356815,
"learning_rate": 1e-06,
"logits/chosen": -3.3398616313934326,
"logits/rejected": -3.3604841232299805,
"logps/chosen": -131.99713134765625,
"logps/rejected": -148.30227661132812,
"loss": 0.4981,
"rewards/accuracies": 0.375,
"rewards/chosen": -0.0010780429001897573,
"rewards/margins": 0.0006100271129980683,
"rewards/rejected": -0.0016880702460184693,
"step": 19
},
{
"debug/policy_chosen_logits": -3.2643423080444336,
"debug/policy_chosen_logps": -134.54617309570312,
"debug/policy_rejected_logits": -3.255173921585083,
"debug/policy_rejected_logps": -124.35269165039062,
"debug/reference_chosen_logps": -134.53549194335938,
"debug/reference_rejected_logps": -123.25413513183594,
"epoch": 0.5405405405405406,
"grad_norm": 3.4308218789715035,
"learning_rate": 1e-06,
"logits/chosen": -3.2643423080444336,
"logits/rejected": -3.255173921585083,
"logps/chosen": -134.54617309570312,
"logps/rejected": -124.35269165039062,
"loss": 0.4969,
"rewards/accuracies": 0.375,
"rewards/chosen": -0.00010671629570424557,
"rewards/margins": 0.010878859087824821,
"rewards/rejected": -0.010985574685037136,
"step": 20
},
{
"debug/policy_chosen_logits": -3.39021635055542,
"debug/policy_chosen_logps": -145.20547485351562,
"debug/policy_rejected_logits": -3.324305295944214,
"debug/policy_rejected_logps": -140.68399047851562,
"debug/reference_chosen_logps": -145.19155883789062,
"debug/reference_rejected_logps": -139.9624481201172,
"epoch": 0.5675675675675675,
"grad_norm": 3.572312158938155,
"learning_rate": 1e-06,
"logits/chosen": -3.39021635055542,
"logits/rejected": -3.324305295944214,
"logps/chosen": -145.20547485351562,
"logps/rejected": -140.68399047851562,
"loss": 0.4975,
"rewards/accuracies": 0.5,
"rewards/chosen": -0.00013918871991336346,
"rewards/margins": 0.007076186593621969,
"rewards/rejected": -0.007215375080704689,
"step": 21
},
{
"debug/policy_chosen_logits": -3.267886161804199,
"debug/policy_chosen_logps": -160.2618865966797,
"debug/policy_rejected_logits": -3.2591264247894287,
"debug/policy_rejected_logps": -141.91876220703125,
"debug/reference_chosen_logps": -159.80300903320312,
"debug/reference_rejected_logps": -142.28695678710938,
"epoch": 0.5945945945945946,
"grad_norm": 3.81196901542863,
"learning_rate": 1e-06,
"logits/chosen": -3.267886161804199,
"logits/rejected": -3.2591264247894287,
"logps/chosen": -160.2618865966797,
"logps/rejected": -141.91876220703125,
"loss": 0.4995,
"rewards/accuracies": 0.125,
"rewards/chosen": -0.004588775336742401,
"rewards/margins": -0.008270883932709694,
"rewards/rejected": 0.0036821081303060055,
"step": 22
},
{
"debug/policy_chosen_logits": -3.3076066970825195,
"debug/policy_chosen_logps": -147.66110229492188,
"debug/policy_rejected_logits": -3.3076066970825195,
"debug/policy_rejected_logps": -156.1669464111328,
"debug/reference_chosen_logps": -147.6890869140625,
"debug/reference_rejected_logps": -155.93359375,
"epoch": 0.6216216216216216,
"grad_norm": 3.584513321934443,
"learning_rate": 1e-06,
"logits/chosen": -3.3076066970825195,
"logits/rejected": -3.3076066970825195,
"logps/chosen": -147.66110229492188,
"logps/rejected": -156.1669464111328,
"loss": 0.4981,
"rewards/accuracies": 0.5,
"rewards/chosen": 0.0002799700014293194,
"rewards/margins": 0.002613373100757599,
"rewards/rejected": -0.002333402633666992,
"step": 23
},
{
"debug/policy_chosen_logits": -3.3157832622528076,
"debug/policy_chosen_logps": -145.93922424316406,
"debug/policy_rejected_logits": -3.2806382179260254,
"debug/policy_rejected_logps": -150.69491577148438,
"debug/reference_chosen_logps": -146.20852661132812,
"debug/reference_rejected_logps": -151.0212860107422,
"epoch": 0.6486486486486487,
"grad_norm": 3.4973421303279575,
"learning_rate": 1e-06,
"logits/chosen": -3.3157832622528076,
"logits/rejected": -3.2806382179260254,
"logps/chosen": -145.93922424316406,
"logps/rejected": -150.69491577148438,
"loss": 0.4979,
"rewards/accuracies": 0.625,
"rewards/chosen": 0.0026929855812340975,
"rewards/margins": -0.0005707358941435814,
"rewards/rejected": 0.0032637212425470352,
"step": 24
},
{
"debug/policy_chosen_logits": -3.340818405151367,
"debug/policy_chosen_logps": -145.6500244140625,
"debug/policy_rejected_logits": -3.2608158588409424,
"debug/policy_rejected_logps": -148.91769409179688,
"debug/reference_chosen_logps": -145.28700256347656,
"debug/reference_rejected_logps": -148.38320922851562,
"epoch": 0.6756756756756757,
"grad_norm": 3.493835228930821,
"learning_rate": 1e-06,
"logits/chosen": -3.340818405151367,
"logits/rejected": -3.2608158588409424,
"logps/chosen": -145.6500244140625,
"logps/rejected": -148.91769409179688,
"loss": 0.4997,
"rewards/accuracies": 0.375,
"rewards/chosen": -0.003630065592005849,
"rewards/margins": 0.0017146589234471321,
"rewards/rejected": -0.005344724282622337,
"step": 25
},
{
"debug/policy_chosen_logits": -3.3708367347717285,
"debug/policy_chosen_logps": -134.6257781982422,
"debug/policy_rejected_logits": -3.2685859203338623,
"debug/policy_rejected_logps": -164.87184143066406,
"debug/reference_chosen_logps": -134.9357452392578,
"debug/reference_rejected_logps": -164.06057739257812,
"epoch": 0.7027027027027027,
"grad_norm": 3.5262056978807492,
"learning_rate": 1e-06,
"logits/chosen": -3.3708367347717285,
"logits/rejected": -3.2685859203338623,
"logps/chosen": -134.6257781982422,
"logps/rejected": -164.87184143066406,
"loss": 0.4954,
"rewards/accuracies": 0.875,
"rewards/chosen": 0.0030995942652225494,
"rewards/margins": 0.011212329380214214,
"rewards/rejected": -0.008112735114991665,
"step": 26
},
{
"debug/policy_chosen_logits": -3.276146650314331,
"debug/policy_chosen_logps": -127.90383911132812,
"debug/policy_rejected_logits": -3.2199714183807373,
"debug/policy_rejected_logps": -148.08279418945312,
"debug/reference_chosen_logps": -127.79987335205078,
"debug/reference_rejected_logps": -148.602783203125,
"epoch": 0.7297297297297297,
"grad_norm": 3.5419489962048973,
"learning_rate": 1e-06,
"logits/chosen": -3.276146650314331,
"logits/rejected": -3.2199714183807373,
"logps/chosen": -127.90383911132812,
"logps/rejected": -148.08279418945312,
"loss": 0.4972,
"rewards/accuracies": 0.25,
"rewards/chosen": -0.0010396192083135247,
"rewards/margins": -0.006239423528313637,
"rewards/rejected": 0.005199803970754147,
"step": 27
},
{
"debug/policy_chosen_logits": -3.3671457767486572,
"debug/policy_chosen_logps": -142.98072814941406,
"debug/policy_rejected_logits": -3.313788652420044,
"debug/policy_rejected_logps": -159.15054321289062,
"debug/reference_chosen_logps": -143.54417419433594,
"debug/reference_rejected_logps": -159.19894409179688,
"epoch": 0.7567567567567568,
"grad_norm": 3.6334386387375948,
"learning_rate": 1e-06,
"logits/chosen": -3.3671457767486572,
"logits/rejected": -3.313788652420044,
"logps/chosen": -142.98072814941406,
"logps/rejected": -159.15054321289062,
"loss": 0.4957,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.005634431727230549,
"rewards/margins": 0.005150537472218275,
"rewards/rejected": 0.0004838943714275956,
"step": 28
},
{
"debug/policy_chosen_logits": -3.451542377471924,
"debug/policy_chosen_logps": -144.62733459472656,
"debug/policy_rejected_logits": -3.500192403793335,
"debug/policy_rejected_logps": -129.41226196289062,
"debug/reference_chosen_logps": -144.27435302734375,
"debug/reference_rejected_logps": -128.7982177734375,
"epoch": 0.7837837837837838,
"grad_norm": 3.525191396719062,
"learning_rate": 1e-06,
"logits/chosen": -3.451542377471924,
"logits/rejected": -3.500192403793335,
"logps/chosen": -144.62733459472656,
"logps/rejected": -129.41226196289062,
"loss": 0.4925,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.003529853653162718,
"rewards/margins": 0.0026106643490493298,
"rewards/rejected": -0.006140518002212048,
"step": 29
},
{
"debug/policy_chosen_logits": -3.4419167041778564,
"debug/policy_chosen_logps": -128.94613647460938,
"debug/policy_rejected_logits": -3.3551814556121826,
"debug/policy_rejected_logps": -125.87129974365234,
"debug/reference_chosen_logps": -129.1351318359375,
"debug/reference_rejected_logps": -125.61131286621094,
"epoch": 0.8108108108108109,
"grad_norm": 3.4357174715493812,
"learning_rate": 1e-06,
"logits/chosen": -3.4419167041778564,
"logits/rejected": -3.3551814556121826,
"logps/chosen": -128.94613647460938,
"logps/rejected": -125.87129974365234,
"loss": 0.4956,
"rewards/accuracies": 0.375,
"rewards/chosen": 0.001890067826025188,
"rewards/margins": 0.00448994617909193,
"rewards/rejected": -0.0025998782366514206,
"step": 30
},
{
"debug/policy_chosen_logits": -3.464578628540039,
"debug/policy_chosen_logps": -148.08114624023438,
"debug/policy_rejected_logits": -3.42940354347229,
"debug/policy_rejected_logps": -155.03309631347656,
"debug/reference_chosen_logps": -147.857666015625,
"debug/reference_rejected_logps": -153.7999267578125,
"epoch": 0.8378378378378378,
"grad_norm": 3.569444014406672,
"learning_rate": 1e-06,
"logits/chosen": -3.464578628540039,
"logits/rejected": -3.42940354347229,
"logps/chosen": -148.08114624023438,
"logps/rejected": -155.03309631347656,
"loss": 0.4994,
"rewards/accuracies": 0.5,
"rewards/chosen": -0.0022347732447087765,
"rewards/margins": 0.010096902027726173,
"rewards/rejected": -0.012331675738096237,
"step": 31
},
{
"debug/policy_chosen_logits": -3.2954604625701904,
"debug/policy_chosen_logps": -146.03802490234375,
"debug/policy_rejected_logits": -3.384243965148926,
"debug/policy_rejected_logps": -145.24456787109375,
"debug/reference_chosen_logps": -146.03688049316406,
"debug/reference_rejected_logps": -144.96478271484375,
"epoch": 0.8648648648648649,
"grad_norm": 3.794862233420595,
"learning_rate": 1e-06,
"logits/chosen": -3.2954604625701904,
"logits/rejected": -3.384243965148926,
"logps/chosen": -146.03802490234375,
"logps/rejected": -145.24456787109375,
"loss": 0.4995,
"rewards/accuracies": 0.5,
"rewards/chosen": -1.1482159607112408e-05,
"rewards/margins": 0.002786235883831978,
"rewards/rejected": -0.0027977179270237684,
"step": 32
},
{
"debug/policy_chosen_logits": -3.393167734146118,
"debug/policy_chosen_logps": -135.54547119140625,
"debug/policy_rejected_logits": -3.4107766151428223,
"debug/policy_rejected_logps": -126.21063995361328,
"debug/reference_chosen_logps": -135.9659423828125,
"debug/reference_rejected_logps": -123.26311492919922,
"epoch": 0.8918918918918919,
"grad_norm": 3.7044090203550195,
"learning_rate": 1e-06,
"logits/chosen": -3.393167734146118,
"logits/rejected": -3.4107766151428223,
"logps/chosen": -135.54547119140625,
"logps/rejected": -126.21063995361328,
"loss": 0.4985,
"rewards/accuracies": 0.5,
"rewards/chosen": 0.004204845521599054,
"rewards/margins": 0.03368005529046059,
"rewards/rejected": -0.02947521209716797,
"step": 33
},
{
"debug/policy_chosen_logits": -3.3235180377960205,
"debug/policy_chosen_logps": -159.286865234375,
"debug/policy_rejected_logits": -3.369253635406494,
"debug/policy_rejected_logps": -165.31944274902344,
"debug/reference_chosen_logps": -158.15670776367188,
"debug/reference_rejected_logps": -164.94976806640625,
"epoch": 0.918918918918919,
"grad_norm": 3.550860802017907,
"learning_rate": 1e-06,
"logits/chosen": -3.3235180377960205,
"logits/rejected": -3.369253635406494,
"logps/chosen": -159.286865234375,
"logps/rejected": -165.31944274902344,
"loss": 0.4997,
"rewards/accuracies": 0.25,
"rewards/chosen": -0.011301536113023758,
"rewards/margins": -0.007604693062603474,
"rewards/rejected": -0.0036968423519283533,
"step": 34
},
{
"debug/policy_chosen_logits": -3.367034673690796,
"debug/policy_chosen_logps": -129.437744140625,
"debug/policy_rejected_logits": -3.396902561187744,
"debug/policy_rejected_logps": -157.3302001953125,
"debug/reference_chosen_logps": -129.46514892578125,
"debug/reference_rejected_logps": -157.58755493164062,
"epoch": 0.9459459459459459,
"grad_norm": 3.6785421821803044,
"learning_rate": 1e-06,
"logits/chosen": -3.367034673690796,
"logits/rejected": -3.396902561187744,
"logps/chosen": -129.437744140625,
"logps/rejected": -157.3302001953125,
"loss": 0.498,
"rewards/accuracies": 0.5,
"rewards/chosen": 0.00027409568428993225,
"rewards/margins": -0.002299328101798892,
"rewards/rejected": 0.002573424018919468,
"step": 35
},
{
"debug/policy_chosen_logits": -3.252328634262085,
"debug/policy_chosen_logps": -139.81039428710938,
"debug/policy_rejected_logits": -3.3797786235809326,
"debug/policy_rejected_logps": -127.12653350830078,
"debug/reference_chosen_logps": -140.35733032226562,
"debug/reference_rejected_logps": -127.09129333496094,
"epoch": 0.972972972972973,
"grad_norm": 3.7364723193101446,
"learning_rate": 1e-06,
"logits/chosen": -3.252328634262085,
"logits/rejected": -3.3797786235809326,
"logps/chosen": -139.81039428710938,
"logps/rejected": -127.12653350830078,
"loss": 0.4951,
"rewards/accuracies": 0.625,
"rewards/chosen": 0.005469245836138725,
"rewards/margins": 0.0058215707540512085,
"rewards/rejected": -0.0003523253835737705,
"step": 36
},
{
"debug/policy_chosen_logits": -3.166999101638794,
"debug/policy_chosen_logps": -165.9640350341797,
"debug/policy_rejected_logits": -3.1885428428649902,
"debug/policy_rejected_logps": -135.82620239257812,
"debug/reference_chosen_logps": -164.97256469726562,
"debug/reference_rejected_logps": -135.76358032226562,
"epoch": 1.0,
"grad_norm": 3.5142818885997116,
"learning_rate": 1e-06,
"logits/chosen": -3.166999101638794,
"logits/rejected": -3.1885428428649902,
"logps/chosen": -165.9640350341797,
"logps/rejected": -135.82620239257812,
"loss": 0.481,
"rewards/accuracies": 0.375,
"rewards/chosen": -0.009914855472743511,
"rewards/margins": -0.009288596920669079,
"rewards/rejected": -0.0006262586684897542,
"step": 37
},
{
"epoch": 1.0,
"step": 37,
"total_flos": 0.0,
"train_loss": 0.4980396749200048,
"train_runtime": 139.9378,
"train_samples_per_second": 16.757,
"train_steps_per_second": 0.264
}
],
"logging_steps": 1,
"max_steps": 37,
"num_input_tokens_seen": 0,
"num_train_epochs": 1,
"save_steps": 500,
"stateful_callbacks": {
"TrainerControl": {
"args": {
"should_epoch_stop": false,
"should_evaluate": false,
"should_log": false,
"should_save": true,
"should_training_stop": true
},
"attributes": {}
}
},
"total_flos": 0.0,
"train_batch_size": 8,
"trial_name": null,
"trial_params": null
}