ds_coder_adamw_iter5 / trainer_state.json
yiran-wang3's picture
End of training
c3652ca verified
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 1.0,
"eval_steps": 100,
"global_step": 33,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"debug/policy_chosen_logits": 29.093868255615234,
"debug/policy_chosen_logps": -394.1856689453125,
"debug/policy_rejected_logits": 25.88316535949707,
"debug/policy_rejected_logps": -393.7677307128906,
"debug/reference_chosen_logps": -394.1856689453125,
"debug/reference_rejected_logps": -393.7677307128906,
"epoch": 0.030303030303030304,
"grad_norm": 5.030868755913746,
"learning_rate": 1e-06,
"logits/chosen": 29.093868255615234,
"logits/rejected": 25.88316535949707,
"logps/chosen": -394.1856689453125,
"logps/rejected": -393.7677307128906,
"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": 24.404436111450195,
"debug/policy_chosen_logps": -377.15850830078125,
"debug/policy_rejected_logits": 27.354230880737305,
"debug/policy_rejected_logps": -394.011962890625,
"debug/reference_chosen_logps": -376.9351806640625,
"debug/reference_rejected_logps": -393.5160827636719,
"epoch": 0.06060606060606061,
"grad_norm": 6.25518297669566,
"learning_rate": 1e-06,
"logits/chosen": 24.404436111450195,
"logits/rejected": 27.354230880737305,
"logps/chosen": -377.15850830078125,
"logps/rejected": -394.011962890625,
"loss": 0.4994,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.0022330854553729296,
"rewards/margins": 0.0027256396133452654,
"rewards/rejected": -0.004958725534379482,
"step": 2
},
{
"debug/policy_chosen_logits": 28.0947208404541,
"debug/policy_chosen_logps": -412.77642822265625,
"debug/policy_rejected_logits": 26.817684173583984,
"debug/policy_rejected_logps": -395.5458679199219,
"debug/reference_chosen_logps": -412.11737060546875,
"debug/reference_rejected_logps": -395.4435729980469,
"epoch": 0.09090909090909091,
"grad_norm": 5.536591360832764,
"learning_rate": 1e-06,
"logits/chosen": 28.0947208404541,
"logits/rejected": 26.817684173583984,
"logps/chosen": -412.77642822265625,
"logps/rejected": -395.5458679199219,
"loss": 0.4984,
"rewards/accuracies": 0.375,
"rewards/chosen": -0.00659027136862278,
"rewards/margins": -0.005567245185375214,
"rewards/rejected": -0.001023025717586279,
"step": 3
},
{
"debug/policy_chosen_logits": 24.52150535583496,
"debug/policy_chosen_logps": -417.7979736328125,
"debug/policy_rejected_logits": 26.255943298339844,
"debug/policy_rejected_logps": -416.5947570800781,
"debug/reference_chosen_logps": -417.81341552734375,
"debug/reference_rejected_logps": -416.580322265625,
"epoch": 0.12121212121212122,
"grad_norm": 5.63081759397413,
"learning_rate": 1e-06,
"logits/chosen": 24.52150535583496,
"logits/rejected": 26.255943298339844,
"logps/chosen": -417.7979736328125,
"logps/rejected": -416.5947570800781,
"loss": 0.5028,
"rewards/accuracies": 0.625,
"rewards/chosen": 0.00015426566824316978,
"rewards/margins": 0.0002988434862345457,
"rewards/rejected": -0.0001445768866688013,
"step": 4
},
{
"debug/policy_chosen_logits": 26.004487991333008,
"debug/policy_chosen_logps": -388.8539733886719,
"debug/policy_rejected_logits": 28.043724060058594,
"debug/policy_rejected_logps": -416.6321105957031,
"debug/reference_chosen_logps": -388.64068603515625,
"debug/reference_rejected_logps": -416.25311279296875,
"epoch": 0.15151515151515152,
"grad_norm": 5.64911268621937,
"learning_rate": 1e-06,
"logits/chosen": 26.004487991333008,
"logits/rejected": 28.043724060058594,
"logps/chosen": -388.8539733886719,
"logps/rejected": -416.6321105957031,
"loss": 0.4994,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.0021327207796275616,
"rewards/margins": 0.0016573716420680285,
"rewards/rejected": -0.00379009242169559,
"step": 5
},
{
"debug/policy_chosen_logits": 30.30371856689453,
"debug/policy_chosen_logps": -390.15362548828125,
"debug/policy_rejected_logits": 27.941354751586914,
"debug/policy_rejected_logps": -398.5522155761719,
"debug/reference_chosen_logps": -390.7499694824219,
"debug/reference_rejected_logps": -398.3262023925781,
"epoch": 0.18181818181818182,
"grad_norm": 7.405074953159111,
"learning_rate": 1e-06,
"logits/chosen": 30.30371856689453,
"logits/rejected": 27.941354751586914,
"logps/chosen": -390.15362548828125,
"logps/rejected": -398.5522155761719,
"loss": 0.4994,
"rewards/accuracies": 0.625,
"rewards/chosen": 0.005963439587503672,
"rewards/margins": 0.008223343640565872,
"rewards/rejected": -0.0022599028889089823,
"step": 6
},
{
"debug/policy_chosen_logits": 26.8443603515625,
"debug/policy_chosen_logps": -393.0838623046875,
"debug/policy_rejected_logits": 25.924436569213867,
"debug/policy_rejected_logps": -411.60015869140625,
"debug/reference_chosen_logps": -393.399169921875,
"debug/reference_rejected_logps": -410.80096435546875,
"epoch": 0.21212121212121213,
"grad_norm": 5.507645597677901,
"learning_rate": 1e-06,
"logits/chosen": 26.8443603515625,
"logits/rejected": 25.924436569213867,
"logps/chosen": -393.0838623046875,
"logps/rejected": -411.60015869140625,
"loss": 0.4981,
"rewards/accuracies": 0.625,
"rewards/chosen": 0.003153343452140689,
"rewards/margins": 0.011145325377583504,
"rewards/rejected": -0.007991980761289597,
"step": 7
},
{
"debug/policy_chosen_logits": 26.57024383544922,
"debug/policy_chosen_logps": -404.73974609375,
"debug/policy_rejected_logits": 23.179155349731445,
"debug/policy_rejected_logps": -381.5411376953125,
"debug/reference_chosen_logps": -405.3680419921875,
"debug/reference_rejected_logps": -382.30010986328125,
"epoch": 0.24242424242424243,
"grad_norm": 5.23273638015015,
"learning_rate": 1e-06,
"logits/chosen": 26.57024383544922,
"logits/rejected": 23.179155349731445,
"logps/chosen": -404.73974609375,
"logps/rejected": -381.5411376953125,
"loss": 0.5002,
"rewards/accuracies": 0.375,
"rewards/chosen": 0.00628303550183773,
"rewards/margins": -0.0013064193772152066,
"rewards/rejected": 0.007589454762637615,
"step": 8
},
{
"debug/policy_chosen_logits": 27.397151947021484,
"debug/policy_chosen_logps": -389.5999755859375,
"debug/policy_rejected_logits": 26.98357391357422,
"debug/policy_rejected_logps": -410.5415954589844,
"debug/reference_chosen_logps": -388.646728515625,
"debug/reference_rejected_logps": -410.95263671875,
"epoch": 0.2727272727272727,
"grad_norm": 4.833818621369551,
"learning_rate": 1e-06,
"logits/chosen": 27.397151947021484,
"logits/rejected": 26.98357391357422,
"logps/chosen": -389.5999755859375,
"logps/rejected": -410.5415954589844,
"loss": 0.5042,
"rewards/accuracies": 0.125,
"rewards/chosen": -0.009532546624541283,
"rewards/margins": -0.013642844744026661,
"rewards/rejected": 0.004110298119485378,
"step": 9
},
{
"debug/policy_chosen_logits": 26.80759620666504,
"debug/policy_chosen_logps": -399.30169677734375,
"debug/policy_rejected_logits": 28.478506088256836,
"debug/policy_rejected_logps": -394.08697509765625,
"debug/reference_chosen_logps": -398.6517333984375,
"debug/reference_rejected_logps": -394.05316162109375,
"epoch": 0.30303030303030304,
"grad_norm": 5.107308392124932,
"learning_rate": 1e-06,
"logits/chosen": 26.80759620666504,
"logits/rejected": 28.478506088256836,
"logps/chosen": -399.30169677734375,
"logps/rejected": -394.08697509765625,
"loss": 0.5001,
"rewards/accuracies": 0.375,
"rewards/chosen": -0.006500015035271645,
"rewards/margins": -0.0061615752056241035,
"rewards/rejected": -0.00033844029530882835,
"step": 10
},
{
"debug/policy_chosen_logits": 26.44737434387207,
"debug/policy_chosen_logps": -408.57733154296875,
"debug/policy_rejected_logits": 26.65883445739746,
"debug/policy_rejected_logps": -391.9090576171875,
"debug/reference_chosen_logps": -409.13458251953125,
"debug/reference_rejected_logps": -391.6588439941406,
"epoch": 0.3333333333333333,
"grad_norm": 4.974671977298996,
"learning_rate": 1e-06,
"logits/chosen": 26.44737434387207,
"logits/rejected": 26.65883445739746,
"logps/chosen": -408.57733154296875,
"logps/rejected": -391.9090576171875,
"loss": 0.4995,
"rewards/accuracies": 0.625,
"rewards/chosen": 0.005572776310145855,
"rewards/margins": 0.008074606768786907,
"rewards/rejected": -0.002501830691471696,
"step": 11
},
{
"debug/policy_chosen_logits": 27.20985221862793,
"debug/policy_chosen_logps": -377.276611328125,
"debug/policy_rejected_logits": 28.053346633911133,
"debug/policy_rejected_logps": -383.6709289550781,
"debug/reference_chosen_logps": -376.66357421875,
"debug/reference_rejected_logps": -383.989990234375,
"epoch": 0.36363636363636365,
"grad_norm": 5.327027762788645,
"learning_rate": 1e-06,
"logits/chosen": 27.20985221862793,
"logits/rejected": 28.053346633911133,
"logps/chosen": -377.276611328125,
"logps/rejected": -383.6709289550781,
"loss": 0.4994,
"rewards/accuracies": 0.25,
"rewards/chosen": -0.00613048579543829,
"rewards/margins": -0.00932083185762167,
"rewards/rejected": 0.003190345596522093,
"step": 12
},
{
"debug/policy_chosen_logits": 25.8612117767334,
"debug/policy_chosen_logps": -392.2992858886719,
"debug/policy_rejected_logits": 24.847339630126953,
"debug/policy_rejected_logps": -402.8356628417969,
"debug/reference_chosen_logps": -391.6371765136719,
"debug/reference_rejected_logps": -402.23583984375,
"epoch": 0.3939393939393939,
"grad_norm": 5.361687372093286,
"learning_rate": 1e-06,
"logits/chosen": 25.8612117767334,
"logits/rejected": 24.847339630126953,
"logps/chosen": -392.2992858886719,
"logps/rejected": -402.8356628417969,
"loss": 0.496,
"rewards/accuracies": 0.375,
"rewards/chosen": -0.0066207884810864925,
"rewards/margins": -0.0006225210381671786,
"rewards/rejected": -0.005998267792165279,
"step": 13
},
{
"debug/policy_chosen_logits": 29.508888244628906,
"debug/policy_chosen_logps": -387.1098327636719,
"debug/policy_rejected_logits": 29.381460189819336,
"debug/policy_rejected_logps": -405.422607421875,
"debug/reference_chosen_logps": -387.8492126464844,
"debug/reference_rejected_logps": -406.4776306152344,
"epoch": 0.42424242424242425,
"grad_norm": 5.461096503844373,
"learning_rate": 1e-06,
"logits/chosen": 29.508888244628906,
"logits/rejected": 29.381460189819336,
"logps/chosen": -387.1098327636719,
"logps/rejected": -405.422607421875,
"loss": 0.4966,
"rewards/accuracies": 0.25,
"rewards/chosen": 0.007393989246338606,
"rewards/margins": -0.0031561660580337048,
"rewards/rejected": 0.010550156235694885,
"step": 14
},
{
"debug/policy_chosen_logits": 30.304807662963867,
"debug/policy_chosen_logps": -388.3260498046875,
"debug/policy_rejected_logits": 29.830333709716797,
"debug/policy_rejected_logps": -399.8182373046875,
"debug/reference_chosen_logps": -387.77960205078125,
"debug/reference_rejected_logps": -398.8104553222656,
"epoch": 0.45454545454545453,
"grad_norm": 5.250363808071301,
"learning_rate": 1e-06,
"logits/chosen": 30.304807662963867,
"logits/rejected": 29.830333709716797,
"logps/chosen": -388.3260498046875,
"logps/rejected": -399.8182373046875,
"loss": 0.5039,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.0054648201912641525,
"rewards/margins": 0.004612846300005913,
"rewards/rejected": -0.01007766742259264,
"step": 15
},
{
"debug/policy_chosen_logits": 28.09935188293457,
"debug/policy_chosen_logps": -397.7403259277344,
"debug/policy_rejected_logits": 31.767385482788086,
"debug/policy_rejected_logps": -376.7595520019531,
"debug/reference_chosen_logps": -397.562744140625,
"debug/reference_rejected_logps": -377.139892578125,
"epoch": 0.48484848484848486,
"grad_norm": 5.660459055452235,
"learning_rate": 1e-06,
"logits/chosen": 28.09935188293457,
"logits/rejected": 31.767385482788086,
"logps/chosen": -397.7403259277344,
"logps/rejected": -376.7595520019531,
"loss": 0.4983,
"rewards/accuracies": 0.125,
"rewards/chosen": -0.0017757411114871502,
"rewards/margins": -0.00557926157489419,
"rewards/rejected": 0.0038035199977457523,
"step": 16
},
{
"debug/policy_chosen_logits": 26.62754249572754,
"debug/policy_chosen_logps": -388.61407470703125,
"debug/policy_rejected_logits": 27.771495819091797,
"debug/policy_rejected_logps": -422.8292541503906,
"debug/reference_chosen_logps": -389.57568359375,
"debug/reference_rejected_logps": -422.9500732421875,
"epoch": 0.5151515151515151,
"grad_norm": 5.458787195966993,
"learning_rate": 1e-06,
"logits/chosen": 26.62754249572754,
"logits/rejected": 27.771495819091797,
"logps/chosen": -388.61407470703125,
"logps/rejected": -422.8292541503906,
"loss": 0.4997,
"rewards/accuracies": 0.625,
"rewards/chosen": 0.009616240859031677,
"rewards/margins": 0.008408088237047195,
"rewards/rejected": 0.0012081530876457691,
"step": 17
},
{
"debug/policy_chosen_logits": 26.25288200378418,
"debug/policy_chosen_logps": -395.8739318847656,
"debug/policy_rejected_logits": 26.34918975830078,
"debug/policy_rejected_logps": -418.2298889160156,
"debug/reference_chosen_logps": -396.3597412109375,
"debug/reference_rejected_logps": -419.4313049316406,
"epoch": 0.5454545454545454,
"grad_norm": 6.045438153278639,
"learning_rate": 1e-06,
"logits/chosen": 26.25288200378418,
"logits/rejected": 26.34918975830078,
"logps/chosen": -395.8739318847656,
"logps/rejected": -418.2298889160156,
"loss": 0.4971,
"rewards/accuracies": 0.375,
"rewards/chosen": 0.004857787862420082,
"rewards/margins": -0.007156638894230127,
"rewards/rejected": 0.012014427222311497,
"step": 18
},
{
"debug/policy_chosen_logits": 28.588306427001953,
"debug/policy_chosen_logps": -380.94671630859375,
"debug/policy_rejected_logits": 29.382522583007812,
"debug/policy_rejected_logps": -404.1590576171875,
"debug/reference_chosen_logps": -381.28717041015625,
"debug/reference_rejected_logps": -404.83154296875,
"epoch": 0.5757575757575758,
"grad_norm": 5.5743203305237,
"learning_rate": 1e-06,
"logits/chosen": 28.588306427001953,
"logits/rejected": 29.382522583007812,
"logps/chosen": -380.94671630859375,
"logps/rejected": -404.1590576171875,
"loss": 0.5007,
"rewards/accuracies": 0.5,
"rewards/chosen": 0.0034041597973555326,
"rewards/margins": -0.0033205407671630383,
"rewards/rejected": 0.00672469986602664,
"step": 19
},
{
"debug/policy_chosen_logits": 24.93841552734375,
"debug/policy_chosen_logps": -415.7802734375,
"debug/policy_rejected_logits": 27.447294235229492,
"debug/policy_rejected_logps": -404.31024169921875,
"debug/reference_chosen_logps": -416.76080322265625,
"debug/reference_rejected_logps": -404.63677978515625,
"epoch": 0.6060606060606061,
"grad_norm": 5.261485130685494,
"learning_rate": 1e-06,
"logits/chosen": 24.93841552734375,
"logits/rejected": 27.447294235229492,
"logps/chosen": -415.7802734375,
"logps/rejected": -404.31024169921875,
"loss": 0.4912,
"rewards/accuracies": 0.5,
"rewards/chosen": 0.009805488400161266,
"rewards/margins": 0.0065397643484175205,
"rewards/rejected": 0.003265724051743746,
"step": 20
},
{
"debug/policy_chosen_logits": 26.202421188354492,
"debug/policy_chosen_logps": -425.5368957519531,
"debug/policy_rejected_logits": 24.03887939453125,
"debug/policy_rejected_logps": -397.7702941894531,
"debug/reference_chosen_logps": -426.2981262207031,
"debug/reference_rejected_logps": -397.6723937988281,
"epoch": 0.6363636363636364,
"grad_norm": 5.590167982725568,
"learning_rate": 1e-06,
"logits/chosen": 26.202421188354492,
"logits/rejected": 24.03887939453125,
"logps/chosen": -425.5368957519531,
"logps/rejected": -397.7702941894531,
"loss": 0.4985,
"rewards/accuracies": 0.875,
"rewards/chosen": 0.007612571120262146,
"rewards/margins": 0.008591575548052788,
"rewards/rejected": -0.0009790037292987108,
"step": 21
},
{
"debug/policy_chosen_logits": 28.777624130249023,
"debug/policy_chosen_logps": -383.98968505859375,
"debug/policy_rejected_logits": 30.540597915649414,
"debug/policy_rejected_logps": -404.1961975097656,
"debug/reference_chosen_logps": -385.23529052734375,
"debug/reference_rejected_logps": -405.3203125,
"epoch": 0.6666666666666666,
"grad_norm": 6.204641362035411,
"learning_rate": 1e-06,
"logits/chosen": 28.777624130249023,
"logits/rejected": 30.540597915649414,
"logps/chosen": -383.98968505859375,
"logps/rejected": -404.1961975097656,
"loss": 0.5028,
"rewards/accuracies": 0.375,
"rewards/chosen": 0.012456131167709827,
"rewards/margins": 0.001215210184454918,
"rewards/rejected": 0.011240921914577484,
"step": 22
},
{
"debug/policy_chosen_logits": 29.740785598754883,
"debug/policy_chosen_logps": -386.64471435546875,
"debug/policy_rejected_logits": 28.254274368286133,
"debug/policy_rejected_logps": -388.5518493652344,
"debug/reference_chosen_logps": -386.5996398925781,
"debug/reference_rejected_logps": -387.55853271484375,
"epoch": 0.696969696969697,
"grad_norm": 5.534634639642078,
"learning_rate": 1e-06,
"logits/chosen": 29.740785598754883,
"logits/rejected": 28.254274368286133,
"logps/chosen": -386.64471435546875,
"logps/rejected": -388.5518493652344,
"loss": 0.4938,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.000450668390840292,
"rewards/margins": 0.00948253646492958,
"rewards/rejected": -0.00993320532143116,
"step": 23
},
{
"debug/policy_chosen_logits": 27.11961555480957,
"debug/policy_chosen_logps": -401.6017761230469,
"debug/policy_rejected_logits": 27.432086944580078,
"debug/policy_rejected_logps": -412.71929931640625,
"debug/reference_chosen_logps": -400.8846130371094,
"debug/reference_rejected_logps": -412.95587158203125,
"epoch": 0.7272727272727273,
"grad_norm": 5.536134798973258,
"learning_rate": 1e-06,
"logits/chosen": 27.11961555480957,
"logits/rejected": 27.432086944580078,
"logps/chosen": -401.6017761230469,
"logps/rejected": -412.71929931640625,
"loss": 0.492,
"rewards/accuracies": 0.25,
"rewards/chosen": -0.007171629928052425,
"rewards/margins": -0.00953700952231884,
"rewards/rejected": 0.0023653795942664146,
"step": 24
},
{
"debug/policy_chosen_logits": 24.8132266998291,
"debug/policy_chosen_logps": -396.219482421875,
"debug/policy_rejected_logits": 27.473514556884766,
"debug/policy_rejected_logps": -407.45550537109375,
"debug/reference_chosen_logps": -395.4239501953125,
"debug/reference_rejected_logps": -405.3516540527344,
"epoch": 0.7575757575757576,
"grad_norm": 5.248733249901598,
"learning_rate": 1e-06,
"logits/chosen": 24.8132266998291,
"logits/rejected": 27.473514556884766,
"logps/chosen": -396.219482421875,
"logps/rejected": -407.45550537109375,
"loss": 0.4923,
"rewards/accuracies": 0.875,
"rewards/chosen": -0.007955474779009819,
"rewards/margins": 0.013082961551845074,
"rewards/rejected": -0.021038437262177467,
"step": 25
},
{
"debug/policy_chosen_logits": 26.792572021484375,
"debug/policy_chosen_logps": -388.32012939453125,
"debug/policy_rejected_logits": 26.965463638305664,
"debug/policy_rejected_logps": -401.08880615234375,
"debug/reference_chosen_logps": -388.4382629394531,
"debug/reference_rejected_logps": -399.5977783203125,
"epoch": 0.7878787878787878,
"grad_norm": 5.024632086381175,
"learning_rate": 1e-06,
"logits/chosen": 26.792572021484375,
"logits/rejected": 26.965463638305664,
"logps/chosen": -388.32012939453125,
"logps/rejected": -401.08880615234375,
"loss": 0.4907,
"rewards/accuracies": 0.875,
"rewards/chosen": 0.0011813733726739883,
"rewards/margins": 0.016091765835881233,
"rewards/rejected": -0.014910392463207245,
"step": 26
},
{
"debug/policy_chosen_logits": 21.125619888305664,
"debug/policy_chosen_logps": -363.61517333984375,
"debug/policy_rejected_logits": 26.54195213317871,
"debug/policy_rejected_logps": -408.9364013671875,
"debug/reference_chosen_logps": -363.9629821777344,
"debug/reference_rejected_logps": -408.07269287109375,
"epoch": 0.8181818181818182,
"grad_norm": 5.432070893156278,
"learning_rate": 1e-06,
"logits/chosen": 21.125619888305664,
"logits/rejected": 26.54195213317871,
"logps/chosen": -363.61517333984375,
"logps/rejected": -408.9364013671875,
"loss": 0.4934,
"rewards/accuracies": 0.625,
"rewards/chosen": 0.0034781266003847122,
"rewards/margins": 0.012114867568016052,
"rewards/rejected": -0.008636741898953915,
"step": 27
},
{
"debug/policy_chosen_logits": 24.872333526611328,
"debug/policy_chosen_logps": -372.0343933105469,
"debug/policy_rejected_logits": 26.66975975036621,
"debug/policy_rejected_logps": -398.46728515625,
"debug/reference_chosen_logps": -371.37255859375,
"debug/reference_rejected_logps": -398.9984130859375,
"epoch": 0.8484848484848485,
"grad_norm": 5.03229338667667,
"learning_rate": 1e-06,
"logits/chosen": 24.872333526611328,
"logits/rejected": 26.66975975036621,
"logps/chosen": -372.0343933105469,
"logps/rejected": -398.46728515625,
"loss": 0.4967,
"rewards/accuracies": 0.5,
"rewards/chosen": -0.006618117913603783,
"rewards/margins": -0.011929434724152088,
"rewards/rejected": 0.0053113182075321674,
"step": 28
},
{
"debug/policy_chosen_logits": 29.60181427001953,
"debug/policy_chosen_logps": -406.96124267578125,
"debug/policy_rejected_logits": 28.26951026916504,
"debug/policy_rejected_logps": -390.26763916015625,
"debug/reference_chosen_logps": -407.25262451171875,
"debug/reference_rejected_logps": -390.38494873046875,
"epoch": 0.8787878787878788,
"grad_norm": 5.040026373746743,
"learning_rate": 1e-06,
"logits/chosen": 29.60181427001953,
"logits/rejected": 28.26951026916504,
"logps/chosen": -406.96124267578125,
"logps/rejected": -390.26763916015625,
"loss": 0.494,
"rewards/accuracies": 0.25,
"rewards/chosen": 0.0029140852857381105,
"rewards/margins": 0.0017412565648555756,
"rewards/rejected": 0.0011728284880518913,
"step": 29
},
{
"debug/policy_chosen_logits": 27.035560607910156,
"debug/policy_chosen_logps": -392.25396728515625,
"debug/policy_rejected_logits": 27.442691802978516,
"debug/policy_rejected_logps": -391.0264892578125,
"debug/reference_chosen_logps": -392.8270263671875,
"debug/reference_rejected_logps": -390.5660400390625,
"epoch": 0.9090909090909091,
"grad_norm": 5.259600833580871,
"learning_rate": 1e-06,
"logits/chosen": 27.035560607910156,
"logits/rejected": 27.442691802978516,
"logps/chosen": -392.25396728515625,
"logps/rejected": -391.0264892578125,
"loss": 0.498,
"rewards/accuracies": 0.625,
"rewards/chosen": 0.005730743054300547,
"rewards/margins": 0.01033527310937643,
"rewards/rejected": -0.0046045295894145966,
"step": 30
},
{
"debug/policy_chosen_logits": 23.07676887512207,
"debug/policy_chosen_logps": -368.69171142578125,
"debug/policy_rejected_logits": 27.108469009399414,
"debug/policy_rejected_logps": -401.52239990234375,
"debug/reference_chosen_logps": -369.28192138671875,
"debug/reference_rejected_logps": -402.61138916015625,
"epoch": 0.9393939393939394,
"grad_norm": 5.210897913069892,
"learning_rate": 1e-06,
"logits/chosen": 23.07676887512207,
"logits/rejected": 27.108469009399414,
"logps/chosen": -368.69171142578125,
"logps/rejected": -401.52239990234375,
"loss": 0.5014,
"rewards/accuracies": 0.25,
"rewards/chosen": 0.005902099423110485,
"rewards/margins": -0.004987678490579128,
"rewards/rejected": 0.010889777913689613,
"step": 31
},
{
"debug/policy_chosen_logits": 26.6132869720459,
"debug/policy_chosen_logps": -397.4622802734375,
"debug/policy_rejected_logits": 27.003686904907227,
"debug/policy_rejected_logps": -413.6773681640625,
"debug/reference_chosen_logps": -397.5897216796875,
"debug/reference_rejected_logps": -411.7838134765625,
"epoch": 0.9696969696969697,
"grad_norm": 5.060163056079328,
"learning_rate": 1e-06,
"logits/chosen": 26.6132869720459,
"logits/rejected": 27.003686904907227,
"logps/chosen": -397.4622802734375,
"logps/rejected": -413.6773681640625,
"loss": 0.498,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.0012746425345540047,
"rewards/margins": 0.020210150629281998,
"rewards/rejected": -0.018935509026050568,
"step": 32
},
{
"debug/policy_chosen_logits": 25.71558952331543,
"debug/policy_chosen_logps": -366.22833251953125,
"debug/policy_rejected_logits": 26.291645050048828,
"debug/policy_rejected_logps": -416.4896545410156,
"debug/reference_chosen_logps": -366.5025939941406,
"debug/reference_rejected_logps": -416.4383544921875,
"epoch": 1.0,
"grad_norm": 5.67721223467369,
"learning_rate": 1e-06,
"logits/chosen": 25.71558952331543,
"logits/rejected": 26.291645050048828,
"logps/chosen": -366.22833251953125,
"logps/rejected": -416.4896545410156,
"loss": 0.4695,
"rewards/accuracies": 0.625,
"rewards/chosen": 0.0027426911983639,
"rewards/margins": 0.003255615010857582,
"rewards/rejected": -0.0005129238124936819,
"step": 33
},
{
"epoch": 1.0,
"step": 33,
"total_flos": 0.0,
"train_loss": 0.4971384649926966,
"train_runtime": 409.9588,
"train_samples_per_second": 5.066,
"train_steps_per_second": 0.08
}
],
"logging_steps": 1,
"max_steps": 33,
"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
}