qwen2_coder_reflct_adamw_iter8 / trainer_state.json
yiran-wang3's picture
End of training
eeef5a8 verified
raw
history blame
27.3 kB
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 1.0,
"eval_steps": 100,
"global_step": 31,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"debug/policy_chosen_logits": -3.0837113857269287,
"debug/policy_chosen_logps": -156.1900634765625,
"debug/policy_rejected_logits": -3.0168697834014893,
"debug/policy_rejected_logps": -213.46035766601562,
"debug/reference_chosen_logps": -156.1900634765625,
"debug/reference_rejected_logps": -213.46035766601562,
"epoch": 0.03225806451612903,
"grad_norm": 33.635131555501395,
"learning_rate": 1e-06,
"logits/chosen": -3.0837113857269287,
"logits/rejected": -3.0168697834014893,
"logps/chosen": -156.1900634765625,
"logps/rejected": -213.46035766601562,
"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": -3.188256025314331,
"debug/policy_chosen_logps": -173.60809326171875,
"debug/policy_rejected_logits": -3.2439301013946533,
"debug/policy_rejected_logps": -204.31166076660156,
"debug/reference_chosen_logps": -172.73226928710938,
"debug/reference_rejected_logps": -198.9984588623047,
"epoch": 0.06451612903225806,
"grad_norm": 12.890042733243831,
"learning_rate": 1e-06,
"logits/chosen": -3.188256025314331,
"logits/rejected": -3.2439301013946533,
"logps/chosen": -173.60809326171875,
"logps/rejected": -204.31166076660156,
"loss": 0.498,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.008758220821619034,
"rewards/margins": 0.04437387362122536,
"rewards/rejected": -0.05313209444284439,
"step": 2
},
{
"debug/policy_chosen_logits": -3.208176374435425,
"debug/policy_chosen_logps": -172.39854431152344,
"debug/policy_rejected_logits": -3.1775765419006348,
"debug/policy_rejected_logps": -211.1417236328125,
"debug/reference_chosen_logps": -171.08482360839844,
"debug/reference_rejected_logps": -208.74172973632812,
"epoch": 0.0967741935483871,
"grad_norm": 19.598718664572683,
"learning_rate": 1e-06,
"logits/chosen": -3.208176374435425,
"logits/rejected": -3.1775765419006348,
"logps/chosen": -172.39854431152344,
"logps/rejected": -211.1417236328125,
"loss": 0.4905,
"rewards/accuracies": 0.5,
"rewards/chosen": -0.013137312605977058,
"rewards/margins": 0.010862598195672035,
"rewards/rejected": -0.023999910801649094,
"step": 3
},
{
"debug/policy_chosen_logits": -3.0628340244293213,
"debug/policy_chosen_logps": -170.09300231933594,
"debug/policy_rejected_logits": -2.9403834342956543,
"debug/policy_rejected_logps": -206.1832275390625,
"debug/reference_chosen_logps": -162.90957641601562,
"debug/reference_rejected_logps": -198.26235961914062,
"epoch": 0.12903225806451613,
"grad_norm": 19.89756602440321,
"learning_rate": 1e-06,
"logits/chosen": -3.0628340244293213,
"logits/rejected": -2.9403834342956543,
"logps/chosen": -170.09300231933594,
"logps/rejected": -206.1832275390625,
"loss": 0.4921,
"rewards/accuracies": 0.375,
"rewards/chosen": -0.07183413207530975,
"rewards/margins": 0.0073744915425777435,
"rewards/rejected": -0.0792086198925972,
"step": 4
},
{
"debug/policy_chosen_logits": -3.1047394275665283,
"debug/policy_chosen_logps": -169.22952270507812,
"debug/policy_rejected_logits": -3.178271532058716,
"debug/policy_rejected_logps": -193.72286987304688,
"debug/reference_chosen_logps": -168.54092407226562,
"debug/reference_rejected_logps": -186.80690002441406,
"epoch": 0.16129032258064516,
"grad_norm": 13.088334676288056,
"learning_rate": 1e-06,
"logits/chosen": -3.1047394275665283,
"logits/rejected": -3.178271532058716,
"logps/chosen": -169.22952270507812,
"logps/rejected": -193.72286987304688,
"loss": 0.4858,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.006886100862175226,
"rewards/margins": 0.062273673713207245,
"rewards/rejected": -0.06915977597236633,
"step": 5
},
{
"debug/policy_chosen_logits": -3.1648635864257812,
"debug/policy_chosen_logps": -161.87496948242188,
"debug/policy_rejected_logits": -3.1185457706451416,
"debug/policy_rejected_logps": -213.39263916015625,
"debug/reference_chosen_logps": -154.76771545410156,
"debug/reference_rejected_logps": -211.57601928710938,
"epoch": 0.1935483870967742,
"grad_norm": 67.05357176610438,
"learning_rate": 1e-06,
"logits/chosen": -3.1648635864257812,
"logits/rejected": -3.1185457706451416,
"logps/chosen": -161.87496948242188,
"logps/rejected": -213.39263916015625,
"loss": 0.5157,
"rewards/accuracies": 0.375,
"rewards/chosen": -0.07107262313365936,
"rewards/margins": -0.052906349301338196,
"rewards/rejected": -0.018166273832321167,
"step": 6
},
{
"debug/policy_chosen_logits": -3.2189788818359375,
"debug/policy_chosen_logps": -177.56878662109375,
"debug/policy_rejected_logits": -3.11997389793396,
"debug/policy_rejected_logps": -189.31292724609375,
"debug/reference_chosen_logps": -179.44760131835938,
"debug/reference_rejected_logps": -187.13697814941406,
"epoch": 0.22580645161290322,
"grad_norm": 11.843697734815683,
"learning_rate": 1e-06,
"logits/chosen": -3.2189788818359375,
"logits/rejected": -3.11997389793396,
"logps/chosen": -177.56878662109375,
"logps/rejected": -189.31292724609375,
"loss": 0.4878,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.01878824271261692,
"rewards/margins": 0.04054774343967438,
"rewards/rejected": -0.021759498864412308,
"step": 7
},
{
"debug/policy_chosen_logits": -3.2227280139923096,
"debug/policy_chosen_logps": -163.78762817382812,
"debug/policy_rejected_logits": -3.290872573852539,
"debug/policy_rejected_logps": -193.32077026367188,
"debug/reference_chosen_logps": -165.69235229492188,
"debug/reference_rejected_logps": -189.6918182373047,
"epoch": 0.25806451612903225,
"grad_norm": 15.145376881649522,
"learning_rate": 1e-06,
"logits/chosen": -3.2227280139923096,
"logits/rejected": -3.290872573852539,
"logps/chosen": -163.78762817382812,
"logps/rejected": -193.32077026367188,
"loss": 0.4912,
"rewards/accuracies": 0.625,
"rewards/chosen": 0.019047202542424202,
"rewards/margins": 0.05533674359321594,
"rewards/rejected": -0.03628953918814659,
"step": 8
},
{
"debug/policy_chosen_logits": -3.159303903579712,
"debug/policy_chosen_logps": -167.47830200195312,
"debug/policy_rejected_logits": -3.1892285346984863,
"debug/policy_rejected_logps": -166.064208984375,
"debug/reference_chosen_logps": -168.6324462890625,
"debug/reference_rejected_logps": -168.05596923828125,
"epoch": 0.2903225806451613,
"grad_norm": 14.37792159617078,
"learning_rate": 1e-06,
"logits/chosen": -3.159303903579712,
"logits/rejected": -3.1892285346984863,
"logps/chosen": -167.47830200195312,
"logps/rejected": -166.064208984375,
"loss": 0.4962,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.011541502550244331,
"rewards/margins": -0.0083759855479002,
"rewards/rejected": 0.01991748809814453,
"step": 9
},
{
"debug/policy_chosen_logits": -2.971123218536377,
"debug/policy_chosen_logps": -211.6798858642578,
"debug/policy_rejected_logits": -3.1354713439941406,
"debug/policy_rejected_logps": -175.7381134033203,
"debug/reference_chosen_logps": -216.06980895996094,
"debug/reference_rejected_logps": -179.71377563476562,
"epoch": 0.3225806451612903,
"grad_norm": 12.173478354163274,
"learning_rate": 1e-06,
"logits/chosen": -2.971123218536377,
"logits/rejected": -3.1354713439941406,
"logps/chosen": -211.6798858642578,
"logps/rejected": -175.7381134033203,
"loss": 0.4989,
"rewards/accuracies": 0.5,
"rewards/chosen": 0.04389926791191101,
"rewards/margins": 0.004142570309340954,
"rewards/rejected": 0.03975669667124748,
"step": 10
},
{
"debug/policy_chosen_logits": -3.1046411991119385,
"debug/policy_chosen_logps": -179.2141571044922,
"debug/policy_rejected_logits": -3.006396770477295,
"debug/policy_rejected_logps": -231.02593994140625,
"debug/reference_chosen_logps": -183.09750366210938,
"debug/reference_rejected_logps": -234.02491760253906,
"epoch": 0.3548387096774194,
"grad_norm": 19.69641784212433,
"learning_rate": 1e-06,
"logits/chosen": -3.1046411991119385,
"logits/rejected": -3.006396770477295,
"logps/chosen": -179.2141571044922,
"logps/rejected": -231.02593994140625,
"loss": 0.4927,
"rewards/accuracies": 0.625,
"rewards/chosen": 0.038833435624837875,
"rewards/margins": 0.008843660354614258,
"rewards/rejected": 0.029989775270223618,
"step": 11
},
{
"debug/policy_chosen_logits": -3.148892402648926,
"debug/policy_chosen_logps": -164.12550354003906,
"debug/policy_rejected_logits": -3.2069592475891113,
"debug/policy_rejected_logps": -196.435791015625,
"debug/reference_chosen_logps": -173.92369079589844,
"debug/reference_rejected_logps": -200.88906860351562,
"epoch": 0.3870967741935484,
"grad_norm": 39.94934023839017,
"learning_rate": 1e-06,
"logits/chosen": -3.148892402648926,
"logits/rejected": -3.2069592475891113,
"logps/chosen": -164.12550354003906,
"logps/rejected": -196.435791015625,
"loss": 0.5015,
"rewards/accuracies": 0.875,
"rewards/chosen": 0.09798187762498856,
"rewards/margins": 0.05344902724027634,
"rewards/rejected": 0.04453285038471222,
"step": 12
},
{
"debug/policy_chosen_logits": -3.147637367248535,
"debug/policy_chosen_logps": -142.30979919433594,
"debug/policy_rejected_logits": -3.1137590408325195,
"debug/policy_rejected_logps": -179.00892639160156,
"debug/reference_chosen_logps": -149.29544067382812,
"debug/reference_rejected_logps": -185.37518310546875,
"epoch": 0.41935483870967744,
"grad_norm": 65.71234154752543,
"learning_rate": 1e-06,
"logits/chosen": -3.147637367248535,
"logits/rejected": -3.1137590408325195,
"logps/chosen": -142.30979919433594,
"logps/rejected": -179.00892639160156,
"loss": 0.5062,
"rewards/accuracies": 0.625,
"rewards/chosen": 0.06985651701688766,
"rewards/margins": 0.006193867418915033,
"rewards/rejected": 0.06366264820098877,
"step": 13
},
{
"debug/policy_chosen_logits": -3.2451846599578857,
"debug/policy_chosen_logps": -183.0009765625,
"debug/policy_rejected_logits": -3.2852346897125244,
"debug/policy_rejected_logps": -197.824951171875,
"debug/reference_chosen_logps": -183.507568359375,
"debug/reference_rejected_logps": -201.4381103515625,
"epoch": 0.45161290322580644,
"grad_norm": 26.124571295930675,
"learning_rate": 1e-06,
"logits/chosen": -3.2451846599578857,
"logits/rejected": -3.2852346897125244,
"logps/chosen": -183.0009765625,
"logps/rejected": -197.824951171875,
"loss": 0.5122,
"rewards/accuracies": 0.375,
"rewards/chosen": 0.005065831355750561,
"rewards/margins": -0.031065817922353745,
"rewards/rejected": 0.03613165393471718,
"step": 14
},
{
"debug/policy_chosen_logits": -3.228227376937866,
"debug/policy_chosen_logps": -150.36203002929688,
"debug/policy_rejected_logits": -3.2135226726531982,
"debug/policy_rejected_logps": -236.50494384765625,
"debug/reference_chosen_logps": -158.17320251464844,
"debug/reference_rejected_logps": -239.26577758789062,
"epoch": 0.4838709677419355,
"grad_norm": 15.171711401478728,
"learning_rate": 1e-06,
"logits/chosen": -3.228227376937866,
"logits/rejected": -3.2135226726531982,
"logps/chosen": -150.36203002929688,
"logps/rejected": -236.50494384765625,
"loss": 0.5003,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.07811184972524643,
"rewards/margins": 0.05050337687134743,
"rewards/rejected": 0.027608470991253853,
"step": 15
},
{
"debug/policy_chosen_logits": -3.164198398590088,
"debug/policy_chosen_logps": -178.2091064453125,
"debug/policy_rejected_logits": -3.2197299003601074,
"debug/policy_rejected_logps": -188.035888671875,
"debug/reference_chosen_logps": -180.6068115234375,
"debug/reference_rejected_logps": -190.4276123046875,
"epoch": 0.5161290322580645,
"grad_norm": 21.853695731678606,
"learning_rate": 1e-06,
"logits/chosen": -3.164198398590088,
"logits/rejected": -3.2197299003601074,
"logps/chosen": -178.2091064453125,
"logps/rejected": -188.035888671875,
"loss": 0.5007,
"rewards/accuracies": 0.375,
"rewards/chosen": 0.023976972326636314,
"rewards/margins": 5.9641897678375244e-05,
"rewards/rejected": 0.02391733229160309,
"step": 16
},
{
"debug/policy_chosen_logits": -3.1247196197509766,
"debug/policy_chosen_logps": -138.4497528076172,
"debug/policy_rejected_logits": -3.148193359375,
"debug/policy_rejected_logps": -177.52459716796875,
"debug/reference_chosen_logps": -141.9833526611328,
"debug/reference_rejected_logps": -176.2555694580078,
"epoch": 0.5483870967741935,
"grad_norm": 16.184529966630762,
"learning_rate": 1e-06,
"logits/chosen": -3.1247196197509766,
"logits/rejected": -3.148193359375,
"logps/chosen": -138.4497528076172,
"logps/rejected": -177.52459716796875,
"loss": 0.5052,
"rewards/accuracies": 0.5,
"rewards/chosen": 0.035335883498191833,
"rewards/margins": 0.04802631214261055,
"rewards/rejected": -0.012690428644418716,
"step": 17
},
{
"debug/policy_chosen_logits": -3.064986228942871,
"debug/policy_chosen_logps": -200.385009765625,
"debug/policy_rejected_logits": -3.0621964931488037,
"debug/policy_rejected_logps": -207.35183715820312,
"debug/reference_chosen_logps": -202.05747985839844,
"debug/reference_rejected_logps": -206.5642547607422,
"epoch": 0.5806451612903226,
"grad_norm": 12.667063930754026,
"learning_rate": 1e-06,
"logits/chosen": -3.064986228942871,
"logits/rejected": -3.0621964931488037,
"logps/chosen": -200.385009765625,
"logps/rejected": -207.35183715820312,
"loss": 0.4804,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.016724586486816406,
"rewards/margins": 0.024600541219115257,
"rewards/rejected": -0.007875956594944,
"step": 18
},
{
"debug/policy_chosen_logits": -3.2486557960510254,
"debug/policy_chosen_logps": -152.92401123046875,
"debug/policy_rejected_logits": -3.239227771759033,
"debug/policy_rejected_logps": -164.85879516601562,
"debug/reference_chosen_logps": -152.27206420898438,
"debug/reference_rejected_logps": -161.17071533203125,
"epoch": 0.6129032258064516,
"grad_norm": 17.915281455929325,
"learning_rate": 1e-06,
"logits/chosen": -3.2486557960510254,
"logits/rejected": -3.239227771759033,
"logps/chosen": -152.92401123046875,
"logps/rejected": -164.85879516601562,
"loss": 0.4612,
"rewards/accuracies": 0.875,
"rewards/chosen": -0.00651948805898428,
"rewards/margins": 0.03036128729581833,
"rewards/rejected": -0.03688077628612518,
"step": 19
},
{
"debug/policy_chosen_logits": -3.223217010498047,
"debug/policy_chosen_logps": -216.85635375976562,
"debug/policy_rejected_logits": -3.2773444652557373,
"debug/policy_rejected_logps": -170.92140197753906,
"debug/reference_chosen_logps": -215.15817260742188,
"debug/reference_rejected_logps": -165.83981323242188,
"epoch": 0.6451612903225806,
"grad_norm": 27.794458161611285,
"learning_rate": 1e-06,
"logits/chosen": -3.223217010498047,
"logits/rejected": -3.2773444652557373,
"logps/chosen": -216.85635375976562,
"logps/rejected": -170.92140197753906,
"loss": 0.4863,
"rewards/accuracies": 0.375,
"rewards/chosen": -0.01698194444179535,
"rewards/margins": 0.03383391723036766,
"rewards/rejected": -0.05081585794687271,
"step": 20
},
{
"debug/policy_chosen_logits": -3.0951452255249023,
"debug/policy_chosen_logps": -184.65277099609375,
"debug/policy_rejected_logits": -3.077794313430786,
"debug/policy_rejected_logps": -177.56741333007812,
"debug/reference_chosen_logps": -183.5254669189453,
"debug/reference_rejected_logps": -175.02011108398438,
"epoch": 0.6774193548387096,
"grad_norm": 20.0849838488403,
"learning_rate": 1e-06,
"logits/chosen": -3.0951452255249023,
"logits/rejected": -3.077794313430786,
"logps/chosen": -184.65277099609375,
"logps/rejected": -177.56741333007812,
"loss": 0.4889,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.011273040436208248,
"rewards/margins": 0.014199809171259403,
"rewards/rejected": -0.02547284960746765,
"step": 21
},
{
"debug/policy_chosen_logits": -3.1871633529663086,
"debug/policy_chosen_logps": -147.81625366210938,
"debug/policy_rejected_logits": -3.2192680835723877,
"debug/policy_rejected_logps": -198.98519897460938,
"debug/reference_chosen_logps": -145.34873962402344,
"debug/reference_rejected_logps": -196.41769409179688,
"epoch": 0.7096774193548387,
"grad_norm": 22.495490487928247,
"learning_rate": 1e-06,
"logits/chosen": -3.1871633529663086,
"logits/rejected": -3.2192680835723877,
"logps/chosen": -147.81625366210938,
"logps/rejected": -198.98519897460938,
"loss": 0.4847,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.024675264954566956,
"rewards/margins": 0.0009998120367527008,
"rewards/rejected": -0.025675076991319656,
"step": 22
},
{
"debug/policy_chosen_logits": -3.077518939971924,
"debug/policy_chosen_logps": -193.34442138671875,
"debug/policy_rejected_logits": -3.218148708343506,
"debug/policy_rejected_logps": -198.03993225097656,
"debug/reference_chosen_logps": -191.85537719726562,
"debug/reference_rejected_logps": -191.48312377929688,
"epoch": 0.7419354838709677,
"grad_norm": 25.201442820556927,
"learning_rate": 1e-06,
"logits/chosen": -3.077518939971924,
"logits/rejected": -3.218148708343506,
"logps/chosen": -193.34442138671875,
"logps/rejected": -198.03993225097656,
"loss": 0.4618,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.014890595339238644,
"rewards/margins": 0.05067773163318634,
"rewards/rejected": -0.06556832045316696,
"step": 23
},
{
"debug/policy_chosen_logits": -3.2286179065704346,
"debug/policy_chosen_logps": -161.75,
"debug/policy_rejected_logits": -3.1389119625091553,
"debug/policy_rejected_logps": -189.98956298828125,
"debug/reference_chosen_logps": -161.94760131835938,
"debug/reference_rejected_logps": -190.900634765625,
"epoch": 0.7741935483870968,
"grad_norm": 12.967586243161568,
"learning_rate": 1e-06,
"logits/chosen": -3.2286179065704346,
"logits/rejected": -3.1389119625091553,
"logps/chosen": -161.75,
"logps/rejected": -189.98956298828125,
"loss": 0.4749,
"rewards/accuracies": 0.375,
"rewards/chosen": 0.0019761761650443077,
"rewards/margins": -0.007134380750358105,
"rewards/rejected": 0.009110554121434689,
"step": 24
},
{
"debug/policy_chosen_logits": -3.272780418395996,
"debug/policy_chosen_logps": -159.73602294921875,
"debug/policy_rejected_logits": -3.2736470699310303,
"debug/policy_rejected_logps": -184.5146484375,
"debug/reference_chosen_logps": -164.45745849609375,
"debug/reference_rejected_logps": -187.22787475585938,
"epoch": 0.8064516129032258,
"grad_norm": 30.149125998144832,
"learning_rate": 1e-06,
"logits/chosen": -3.272780418395996,
"logits/rejected": -3.2736470699310303,
"logps/chosen": -159.73602294921875,
"logps/rejected": -184.5146484375,
"loss": 0.4964,
"rewards/accuracies": 0.5,
"rewards/chosen": 0.0472143180668354,
"rewards/margins": 0.02008194290101528,
"rewards/rejected": 0.027132373303174973,
"step": 25
},
{
"debug/policy_chosen_logits": -3.0852928161621094,
"debug/policy_chosen_logps": -206.41693115234375,
"debug/policy_rejected_logits": -3.1496071815490723,
"debug/policy_rejected_logps": -199.11636352539062,
"debug/reference_chosen_logps": -207.51861572265625,
"debug/reference_rejected_logps": -200.44711303710938,
"epoch": 0.8387096774193549,
"grad_norm": 53.54284328779407,
"learning_rate": 1e-06,
"logits/chosen": -3.0852928161621094,
"logits/rejected": -3.1496071815490723,
"logps/chosen": -206.41693115234375,
"logps/rejected": -199.11636352539062,
"loss": 0.5242,
"rewards/accuracies": 0.5,
"rewards/chosen": 0.01101672276854515,
"rewards/margins": -0.0022906875237822533,
"rewards/rejected": 0.013307408429682255,
"step": 26
},
{
"debug/policy_chosen_logits": -3.176145315170288,
"debug/policy_chosen_logps": -136.85189819335938,
"debug/policy_rejected_logits": -3.165332555770874,
"debug/policy_rejected_logps": -172.10556030273438,
"debug/reference_chosen_logps": -137.76473999023438,
"debug/reference_rejected_logps": -177.917236328125,
"epoch": 0.8709677419354839,
"grad_norm": 48.38261399039204,
"learning_rate": 1e-06,
"logits/chosen": -3.176145315170288,
"logits/rejected": -3.165332555770874,
"logps/chosen": -136.85189819335938,
"logps/rejected": -172.10556030273438,
"loss": 0.5204,
"rewards/accuracies": 0.375,
"rewards/chosen": 0.009128322824835777,
"rewards/margins": -0.048988424241542816,
"rewards/rejected": 0.05811674892902374,
"step": 27
},
{
"debug/policy_chosen_logits": -3.070044755935669,
"debug/policy_chosen_logps": -188.4156036376953,
"debug/policy_rejected_logits": -3.1209747791290283,
"debug/policy_rejected_logps": -268.2892150878906,
"debug/reference_chosen_logps": -187.7012939453125,
"debug/reference_rejected_logps": -255.53848266601562,
"epoch": 0.9032258064516129,
"grad_norm": 12.887123147417002,
"learning_rate": 1e-06,
"logits/chosen": -3.070044755935669,
"logits/rejected": -3.1209747791290283,
"logps/chosen": -188.4156036376953,
"logps/rejected": -268.2892150878906,
"loss": 0.4722,
"rewards/accuracies": 1.0,
"rewards/chosen": -0.007143107242882252,
"rewards/margins": 0.12036430835723877,
"rewards/rejected": -0.12750741839408875,
"step": 28
},
{
"debug/policy_chosen_logits": -3.107494831085205,
"debug/policy_chosen_logps": -160.94583129882812,
"debug/policy_rejected_logits": -3.0909152030944824,
"debug/policy_rejected_logps": -195.18310546875,
"debug/reference_chosen_logps": -164.90542602539062,
"debug/reference_rejected_logps": -192.39305114746094,
"epoch": 0.9354838709677419,
"grad_norm": 43.32899603798361,
"learning_rate": 1e-06,
"logits/chosen": -3.107494831085205,
"logits/rejected": -3.0909152030944824,
"logps/chosen": -160.94583129882812,
"logps/rejected": -195.18310546875,
"loss": 0.4825,
"rewards/accuracies": 0.625,
"rewards/chosen": 0.03959597647190094,
"rewards/margins": 0.06749667972326279,
"rewards/rejected": -0.027900703251361847,
"step": 29
},
{
"debug/policy_chosen_logits": -3.106745958328247,
"debug/policy_chosen_logps": -197.70811462402344,
"debug/policy_rejected_logits": -3.103618860244751,
"debug/policy_rejected_logps": -207.11260986328125,
"debug/reference_chosen_logps": -194.18881225585938,
"debug/reference_rejected_logps": -205.00173950195312,
"epoch": 0.967741935483871,
"grad_norm": 9.647676539368454,
"learning_rate": 1e-06,
"logits/chosen": -3.106745958328247,
"logits/rejected": -3.103618860244751,
"logps/chosen": -197.70811462402344,
"logps/rejected": -207.11260986328125,
"loss": 0.4565,
"rewards/accuracies": 0.5,
"rewards/chosen": -0.035193100571632385,
"rewards/margins": -0.01408432051539421,
"rewards/rejected": -0.021108778193593025,
"step": 30
},
{
"debug/policy_chosen_logits": -3.026829957962036,
"debug/policy_chosen_logps": -198.42453002929688,
"debug/policy_rejected_logits": -2.9937751293182373,
"debug/policy_rejected_logps": -243.72132873535156,
"debug/reference_chosen_logps": -197.02398681640625,
"debug/reference_rejected_logps": -237.82403564453125,
"epoch": 1.0,
"grad_norm": 10.307557167599814,
"learning_rate": 1e-06,
"logits/chosen": -3.026829957962036,
"logits/rejected": -2.9937751293182373,
"logps/chosen": -198.42453002929688,
"logps/rejected": -243.72132873535156,
"loss": 0.461,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.014005661010742188,
"rewards/margins": 0.04496738314628601,
"rewards/rejected": -0.0589730404317379,
"step": 31
},
{
"epoch": 1.0,
"step": 31,
"total_flos": 0.0,
"train_loss": 0.4911675299367597,
"train_runtime": 120.215,
"train_samples_per_second": 16.371,
"train_steps_per_second": 0.258
}
],
"logging_steps": 1,
"max_steps": 31,
"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
}