{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.9989528795811519,
  "eval_steps": 200,
  "global_step": 477,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0020942408376963353,
      "grad_norm": 7.1476731300354,
      "learning_rate": 0.0,
      "logits/chosen": -0.6103914976119995,
      "logits/rejected": -0.6099507808685303,
      "logps/chosen": -318.31317138671875,
      "logps/ref_chosen": -318.28057861328125,
      "logps/ref_rejected": -203.32687377929688,
      "logps/rejected": -203.17298889160156,
      "loss": 1.3866,
      "margin_dpo/margin_mean": -0.18641114234924316,
      "margin_dpo/margin_std": 0.6639037132263184,
      "step": 1
    },
    {
      "epoch": 0.010471204188481676,
      "grad_norm": 7.389297008514404,
      "learning_rate": 4.166666666666666e-08,
      "logits/chosen": -0.6768993139266968,
      "logits/rejected": -0.6286869049072266,
      "logps/chosen": -284.77685546875,
      "logps/ref_chosen": -284.8314514160156,
      "logps/ref_rejected": -286.5647888183594,
      "logps/rejected": -286.6424865722656,
      "loss": 1.3862,
      "margin_dpo/margin_mean": 0.1322835385799408,
      "margin_dpo/margin_std": 0.8204990029335022,
      "step": 5
    },
    {
      "epoch": 0.020942408376963352,
      "grad_norm": 6.636824131011963,
      "learning_rate": 9.375e-08,
      "logits/chosen": -0.6935982704162598,
      "logits/rejected": -0.6999162435531616,
      "logps/chosen": -276.5225524902344,
      "logps/ref_chosen": -276.4944152832031,
      "logps/ref_rejected": -242.5192108154297,
      "logps/rejected": -242.5146484375,
      "loss": 1.3861,
      "margin_dpo/margin_mean": -0.03269507735967636,
      "margin_dpo/margin_std": 0.6867764592170715,
      "step": 10
    },
    {
      "epoch": 0.031413612565445025,
      "grad_norm": 7.5604400634765625,
      "learning_rate": 1.4583333333333335e-07,
      "logits/chosen": -0.634182333946228,
      "logits/rejected": -0.6508210897445679,
      "logps/chosen": -307.7691345214844,
      "logps/ref_chosen": -307.81634521484375,
      "logps/ref_rejected": -287.94805908203125,
      "logps/rejected": -288.03228759765625,
      "loss": 1.386,
      "margin_dpo/margin_mean": 0.13143400847911835,
      "margin_dpo/margin_std": 0.7205628156661987,
      "step": 15
    },
    {
      "epoch": 0.041884816753926704,
      "grad_norm": 7.414900779724121,
      "learning_rate": 1.9791666666666664e-07,
      "logits/chosen": -0.6475816965103149,
      "logits/rejected": -0.6553324460983276,
      "logps/chosen": -283.70404052734375,
      "logps/ref_chosen": -283.7494201660156,
      "logps/ref_rejected": -234.01278686523438,
      "logps/rejected": -234.0884246826172,
      "loss": 1.3854,
      "margin_dpo/margin_mean": 0.1209900826215744,
      "margin_dpo/margin_std": 0.5572749376296997,
      "step": 20
    },
    {
      "epoch": 0.05235602094240838,
      "grad_norm": 6.6036601066589355,
      "learning_rate": 2.5e-07,
      "logits/chosen": -0.646403431892395,
      "logits/rejected": -0.6549252271652222,
      "logps/chosen": -234.9982147216797,
      "logps/ref_chosen": -235.056884765625,
      "logps/ref_rejected": -214.2563934326172,
      "logps/rejected": -214.32626342773438,
      "loss": 1.385,
      "margin_dpo/margin_mean": 0.12857410311698914,
      "margin_dpo/margin_std": 0.6414791941642761,
      "step": 25
    },
    {
      "epoch": 0.06282722513089005,
      "grad_norm": 7.229597568511963,
      "learning_rate": 3.020833333333333e-07,
      "logits/chosen": -0.6887374520301819,
      "logits/rejected": -0.6938886642456055,
      "logps/chosen": -322.88897705078125,
      "logps/ref_chosen": -323.2079772949219,
      "logps/ref_rejected": -253.31405639648438,
      "logps/rejected": -253.1912841796875,
      "loss": 1.3844,
      "margin_dpo/margin_mean": 0.19620926678180695,
      "margin_dpo/margin_std": 0.8317287564277649,
      "step": 30
    },
    {
      "epoch": 0.07329842931937172,
      "grad_norm": 6.7626142501831055,
      "learning_rate": 3.541666666666667e-07,
      "logits/chosen": -0.6849234700202942,
      "logits/rejected": -0.6820663809776306,
      "logps/chosen": -300.1553039550781,
      "logps/ref_chosen": -300.67559814453125,
      "logps/ref_rejected": -275.8673400878906,
      "logps/rejected": -275.5356750488281,
      "loss": 1.3823,
      "margin_dpo/margin_mean": 0.18862803280353546,
      "margin_dpo/margin_std": 0.9617260098457336,
      "step": 35
    },
    {
      "epoch": 0.08376963350785341,
      "grad_norm": 6.857061862945557,
      "learning_rate": 4.0625e-07,
      "logits/chosen": -0.5887177586555481,
      "logits/rejected": -0.6046378016471863,
      "logps/chosen": -251.55557250976562,
      "logps/ref_chosen": -252.3434600830078,
      "logps/ref_rejected": -278.35101318359375,
      "logps/rejected": -278.2633361816406,
      "loss": 1.3786,
      "margin_dpo/margin_mean": 0.7002249956130981,
      "margin_dpo/margin_std": 1.3109245300292969,
      "step": 40
    },
    {
      "epoch": 0.09424083769633508,
      "grad_norm": 7.225438594818115,
      "learning_rate": 4.5833333333333327e-07,
      "logits/chosen": -0.7441970705986023,
      "logits/rejected": -0.7299541234970093,
      "logps/chosen": -311.53662109375,
      "logps/ref_chosen": -312.97418212890625,
      "logps/ref_rejected": -304.2184753417969,
      "logps/rejected": -303.91888427734375,
      "loss": 1.3746,
      "margin_dpo/margin_mean": 1.1379705667495728,
      "margin_dpo/margin_std": 2.3105995655059814,
      "step": 45
    },
    {
      "epoch": 0.10471204188481675,
      "grad_norm": 6.961248397827148,
      "learning_rate": 4.999932966293553e-07,
      "logits/chosen": -0.6633109450340271,
      "logits/rejected": -0.6941882371902466,
      "logps/chosen": -274.2547302246094,
      "logps/ref_chosen": -276.328369140625,
      "logps/ref_rejected": -249.0668182373047,
      "logps/rejected": -248.53305053710938,
      "loss": 1.373,
      "margin_dpo/margin_mean": 1.5398980379104614,
      "margin_dpo/margin_std": 2.8042705059051514,
      "step": 50
    },
    {
      "epoch": 0.11518324607329843,
      "grad_norm": 6.207220554351807,
      "learning_rate": 4.997587164001815e-07,
      "logits/chosen": -0.6882608532905579,
      "logits/rejected": -0.6976534128189087,
      "logps/chosen": -306.1382751464844,
      "logps/ref_chosen": -308.47393798828125,
      "logps/ref_rejected": -298.5058288574219,
      "logps/rejected": -298.40655517578125,
      "loss": 1.3643,
      "margin_dpo/margin_mean": 2.236351728439331,
      "margin_dpo/margin_std": 4.03792142868042,
      "step": 55
    },
    {
      "epoch": 0.1256544502617801,
      "grad_norm": 7.07670783996582,
      "learning_rate": 4.991893270335525e-07,
      "logits/chosen": -0.6674671173095703,
      "logits/rejected": -0.6547614336013794,
      "logps/chosen": -309.98895263671875,
      "logps/ref_chosen": -312.65618896484375,
      "logps/ref_rejected": -274.500244140625,
      "logps/rejected": -274.29986572265625,
      "loss": 1.3582,
      "margin_dpo/margin_mean": 2.4668667316436768,
      "margin_dpo/margin_std": 6.715214729309082,
      "step": 60
    },
    {
      "epoch": 0.13612565445026178,
      "grad_norm": 6.950833320617676,
      "learning_rate": 4.982858918131906e-07,
      "logits/chosen": -0.6531001925468445,
      "logits/rejected": -0.6600942611694336,
      "logps/chosen": -329.8468017578125,
      "logps/ref_chosen": -334.0863952636719,
      "logps/ref_rejected": -310.85491943359375,
      "logps/rejected": -310.4468688964844,
      "loss": 1.3531,
      "margin_dpo/margin_mean": 3.8315443992614746,
      "margin_dpo/margin_std": 8.632562637329102,
      "step": 65
    },
    {
      "epoch": 0.14659685863874344,
      "grad_norm": 7.140095233917236,
      "learning_rate": 4.970496218214204e-07,
      "logits/chosen": -0.7720015645027161,
      "logits/rejected": -0.7810764908790588,
      "logps/chosen": -281.643798828125,
      "logps/ref_chosen": -286.09478759765625,
      "logps/ref_rejected": -269.44683837890625,
      "logps/rejected": -269.636962890625,
      "loss": 1.3407,
      "margin_dpo/margin_mean": 4.641018867492676,
      "margin_dpo/margin_std": 9.235776901245117,
      "step": 70
    },
    {
      "epoch": 0.15706806282722513,
      "grad_norm": 7.371065616607666,
      "learning_rate": 4.954821743156767e-07,
      "logits/chosen": -0.70263671875,
      "logits/rejected": -0.7102752923965454,
      "logps/chosen": -324.5908508300781,
      "logps/ref_chosen": -329.2369384765625,
      "logps/ref_rejected": -309.39324951171875,
      "logps/rejected": -308.8836669921875,
      "loss": 1.3387,
      "margin_dpo/margin_mean": 4.13649845123291,
      "margin_dpo/margin_std": 11.341253280639648,
      "step": 75
    },
    {
      "epoch": 0.16753926701570682,
      "grad_norm": 7.241215229034424,
      "learning_rate": 4.935856505068998e-07,
      "logits/chosen": -0.7236490845680237,
      "logits/rejected": -0.722406268119812,
      "logps/chosen": -255.44656372070312,
      "logps/ref_chosen": -257.80487060546875,
      "logps/ref_rejected": -247.0098114013672,
      "logps/rejected": -249.47775268554688,
      "loss": 1.3201,
      "margin_dpo/margin_mean": 4.8262553215026855,
      "margin_dpo/margin_std": 11.936810493469238,
      "step": 80
    },
    {
      "epoch": 0.17801047120418848,
      "grad_norm": 7.340602874755859,
      "learning_rate": 4.913625927427995e-07,
      "logits/chosen": -0.7124683260917664,
      "logits/rejected": -0.735012412071228,
      "logps/chosen": -273.90252685546875,
      "logps/ref_chosen": -277.0785827636719,
      "logps/ref_rejected": -261.55169677734375,
      "logps/rejected": -270.66497802734375,
      "loss": 1.3219,
      "margin_dpo/margin_mean": 12.289302825927734,
      "margin_dpo/margin_std": 17.269710540771484,
      "step": 85
    },
    {
      "epoch": 0.18848167539267016,
      "grad_norm": 7.4462690353393555,
      "learning_rate": 4.8881598109976e-07,
      "logits/chosen": -0.7061265707015991,
      "logits/rejected": -0.7180206775665283,
      "logps/chosen": -301.48211669921875,
      "logps/ref_chosen": -300.3891296386719,
      "logps/ref_rejected": -289.59014892578125,
      "logps/rejected": -301.53326416015625,
      "loss": 1.3065,
      "margin_dpo/margin_mean": 10.850162506103516,
      "margin_dpo/margin_std": 16.798553466796875,
      "step": 90
    },
    {
      "epoch": 0.19895287958115182,
      "grad_norm": 8.369804382324219,
      "learning_rate": 4.859492293879573e-07,
      "logits/chosen": -0.687911331653595,
      "logits/rejected": -0.7407578825950623,
      "logps/chosen": -245.9025421142578,
      "logps/ref_chosen": -243.02804565429688,
      "logps/ref_rejected": -219.5611572265625,
      "logps/rejected": -229.8905792236328,
      "loss": 1.3033,
      "margin_dpo/margin_mean": 7.454855918884277,
      "margin_dpo/margin_std": 19.36737632751465,
      "step": 95
    },
    {
      "epoch": 0.2094240837696335,
      "grad_norm": 8.873656272888184,
      "learning_rate": 4.827661805750437e-07,
      "logits/chosen": -0.7412772178649902,
      "logits/rejected": -0.7370281219482422,
      "logps/chosen": -302.4429626464844,
      "logps/ref_chosen": -297.3129577636719,
      "logps/ref_rejected": -304.67572021484375,
      "logps/rejected": -320.0408935546875,
      "loss": 1.2866,
      "margin_dpo/margin_mean": 10.235170364379883,
      "margin_dpo/margin_std": 25.688796997070312,
      "step": 100
    },
    {
      "epoch": 0.2198952879581152,
      "grad_norm": 8.666266441345215,
      "learning_rate": 4.792711016345321e-07,
      "logits/chosen": -0.7506468892097473,
      "logits/rejected": -0.7723590135574341,
      "logps/chosen": -285.6722412109375,
      "logps/ref_chosen": -279.5523376464844,
      "logps/ref_rejected": -258.6304016113281,
      "logps/rejected": -277.24481201171875,
      "loss": 1.2624,
      "margin_dpo/margin_mean": 12.494502067565918,
      "margin_dpo/margin_std": 27.29207992553711,
      "step": 105
    },
    {
      "epoch": 0.23036649214659685,
      "grad_norm": 8.676697731018066,
      "learning_rate": 4.75468677825789e-07,
      "logits/chosen": -0.7946863770484924,
      "logits/rejected": -0.7978917360305786,
      "logps/chosen": -284.75701904296875,
      "logps/ref_chosen": -278.9017639160156,
      "logps/ref_rejected": -230.1315460205078,
      "logps/rejected": -254.346923828125,
      "loss": 1.2393,
      "margin_dpo/margin_mean": 18.360124588012695,
      "margin_dpo/margin_std": 28.798229217529297,
      "step": 110
    },
    {
      "epoch": 0.24083769633507854,
      "grad_norm": 10.881580352783203,
      "learning_rate": 4.7136400641330245e-07,
      "logits/chosen": -0.7917270660400391,
      "logits/rejected": -0.7990630865097046,
      "logps/chosen": -277.28607177734375,
      "logps/ref_chosen": -262.6755676269531,
      "logps/ref_rejected": -234.7182159423828,
      "logps/rejected": -267.9288635253906,
      "loss": 1.259,
      "margin_dpo/margin_mean": 18.60015106201172,
      "margin_dpo/margin_std": 37.60885238647461,
      "step": 115
    },
    {
      "epoch": 0.2513089005235602,
      "grad_norm": 10.508712768554688,
      "learning_rate": 4.669625898336438e-07,
      "logits/chosen": -0.8892138600349426,
      "logits/rejected": -0.8844587206840515,
      "logps/chosen": -294.91656494140625,
      "logps/ref_chosen": -269.8807373046875,
      "logps/ref_rejected": -268.51702880859375,
      "logps/rejected": -313.973388671875,
      "loss": 1.241,
      "margin_dpo/margin_mean": 20.420486450195312,
      "margin_dpo/margin_std": 38.39332580566406,
      "step": 120
    },
    {
      "epoch": 0.2617801047120419,
      "grad_norm": 13.029239654541016,
      "learning_rate": 4.6227032831928483e-07,
      "logits/chosen": -0.8507975339889526,
      "logits/rejected": -0.8206876516342163,
      "logps/chosen": -322.9510192871094,
      "logps/ref_chosen": -293.70062255859375,
      "logps/ref_rejected": -286.66217041015625,
      "logps/rejected": -330.2171630859375,
      "loss": 1.2444,
      "margin_dpo/margin_mean": 14.304577827453613,
      "margin_dpo/margin_std": 44.559913635253906,
      "step": 125
    },
    {
      "epoch": 0.27225130890052357,
      "grad_norm": 12.968539237976074,
      "learning_rate": 4.5729351198915705e-07,
      "logits/chosen": -0.8192211389541626,
      "logits/rejected": -0.8181384205818176,
      "logps/chosen": -308.5555725097656,
      "logps/ref_chosen": -284.30474853515625,
      "logps/ref_rejected": -289.4891662597656,
      "logps/rejected": -341.5650329589844,
      "loss": 1.1922,
      "margin_dpo/margin_mean": 27.825063705444336,
      "margin_dpo/margin_std": 42.415462493896484,
      "step": 130
    },
    {
      "epoch": 0.28272251308900526,
      "grad_norm": 15.247875213623047,
      "learning_rate": 4.520388124165564e-07,
      "logits/chosen": -0.8112742304801941,
      "logits/rejected": -0.790899932384491,
      "logps/chosen": -308.3522033691406,
      "logps/ref_chosen": -279.0638732910156,
      "logps/ref_rejected": -271.4653015136719,
      "logps/rejected": -325.9285888671875,
      "loss": 1.2114,
      "margin_dpo/margin_mean": 25.174943923950195,
      "margin_dpo/margin_std": 43.65618896484375,
      "step": 135
    },
    {
      "epoch": 0.2931937172774869,
      "grad_norm": 15.881244659423828,
      "learning_rate": 4.4651327368569684e-07,
      "logits/chosen": -0.8326481580734253,
      "logits/rejected": -0.8528935313224792,
      "logps/chosen": -360.31134033203125,
      "logps/ref_chosen": -319.4598693847656,
      "logps/ref_rejected": -272.49420166015625,
      "logps/rejected": -338.5896911621094,
      "loss": 1.2095,
      "margin_dpo/margin_mean": 25.244089126586914,
      "margin_dpo/margin_std": 55.747283935546875,
      "step": 140
    },
    {
      "epoch": 0.3036649214659686,
      "grad_norm": 13.514267921447754,
      "learning_rate": 4.4072430294890166e-07,
      "logits/chosen": -0.8288572430610657,
      "logits/rejected": -0.8065798878669739,
      "logps/chosen": -310.3006896972656,
      "logps/ref_chosen": -282.9032287597656,
      "logps/ref_rejected": -256.1629638671875,
      "logps/rejected": -322.1570129394531,
      "loss": 1.1832,
      "margin_dpo/margin_mean": 38.596588134765625,
      "margin_dpo/margin_std": 52.25310134887695,
      "step": 145
    },
    {
      "epoch": 0.31413612565445026,
      "grad_norm": 15.282773971557617,
      "learning_rate": 4.346796604970912e-07,
      "logits/chosen": -0.8817797899246216,
      "logits/rejected": -0.8567675352096558,
      "logps/chosen": -337.6163635253906,
      "logps/ref_chosen": -294.38629150390625,
      "logps/ref_rejected": -283.258544921875,
      "logps/rejected": -359.8990783691406,
      "loss": 1.1792,
      "margin_dpo/margin_mean": 33.410484313964844,
      "margin_dpo/margin_std": 48.665000915527344,
      "step": 150
    },
    {
      "epoch": 0.32460732984293195,
      "grad_norm": 21.630462646484375,
      "learning_rate": 4.2838744935687716e-07,
      "logits/chosen": -0.8599667549133301,
      "logits/rejected": -0.8464357256889343,
      "logps/chosen": -336.33258056640625,
      "logps/ref_chosen": -282.7860107421875,
      "logps/ref_rejected": -277.7061462402344,
      "logps/rejected": -382.0819396972656,
      "loss": 1.1005,
      "margin_dpo/margin_mean": 50.82927703857422,
      "margin_dpo/margin_std": 65.71281433105469,
      "step": 155
    },
    {
      "epoch": 0.33507853403141363,
      "grad_norm": 15.092132568359375,
      "learning_rate": 4.218561044282098e-07,
      "logits/chosen": -0.8716852068901062,
      "logits/rejected": -0.8610923886299133,
      "logps/chosen": -297.11187744140625,
      "logps/ref_chosen": -245.03564453125,
      "logps/ref_rejected": -193.0067596435547,
      "logps/rejected": -286.109130859375,
      "loss": 1.1212,
      "margin_dpo/margin_mean": 41.026119232177734,
      "margin_dpo/margin_std": 55.911903381347656,
      "step": 160
    },
    {
      "epoch": 0.34554973821989526,
      "grad_norm": 25.69545555114746,
      "learning_rate": 4.1509438117713863e-07,
      "logits/chosen": -0.8722988963127136,
      "logits/rejected": -0.8764654397964478,
      "logps/chosen": -323.02362060546875,
      "logps/ref_chosen": -271.78961181640625,
      "logps/ref_rejected": -231.51443481445312,
      "logps/rejected": -323.4990539550781,
      "loss": 1.1301,
      "margin_dpo/margin_mean": 40.750648498535156,
      "margin_dpo/margin_std": 60.6348991394043,
      "step": 165
    },
    {
      "epoch": 0.35602094240837695,
      "grad_norm": 22.663711547851562,
      "learning_rate": 4.081113438988443e-07,
      "logits/chosen": -0.8557019233703613,
      "logits/rejected": -0.8207084536552429,
      "logps/chosen": -391.30767822265625,
      "logps/ref_chosen": -335.8885803222656,
      "logps/ref_rejected": -260.59735107421875,
      "logps/rejected": -372.26214599609375,
      "loss": 1.1057,
      "margin_dpo/margin_mean": 56.2457160949707,
      "margin_dpo/margin_std": 64.02092742919922,
      "step": 170
    },
    {
      "epoch": 0.36649214659685864,
      "grad_norm": 24.62014389038086,
      "learning_rate": 4.00916353566676e-07,
      "logits/chosen": -0.7955960035324097,
      "logits/rejected": -0.8043614625930786,
      "logps/chosen": -392.2655029296875,
      "logps/ref_chosen": -332.5082092285156,
      "logps/ref_rejected": -267.3958740234375,
      "logps/rejected": -373.43511962890625,
      "loss": 1.1212,
      "margin_dpo/margin_mean": 46.28200149536133,
      "margin_dpo/margin_std": 70.63272857666016,
      "step": 175
    },
    {
      "epoch": 0.3769633507853403,
      "grad_norm": 23.055593490600586,
      "learning_rate": 3.935190552834828e-07,
      "logits/chosen": -0.7882012724876404,
      "logits/rejected": -0.773046612739563,
      "logps/chosen": -312.5518493652344,
      "logps/ref_chosen": -251.58407592773438,
      "logps/ref_rejected": -225.3298797607422,
      "logps/rejected": -328.76324462890625,
      "loss": 1.1412,
      "margin_dpo/margin_mean": 42.46554946899414,
      "margin_dpo/margin_std": 68.0986328125,
      "step": 180
    },
    {
      "epoch": 0.387434554973822,
      "grad_norm": 23.7900333404541,
      "learning_rate": 3.859293653520604e-07,
      "logits/chosen": -0.8637909889221191,
      "logits/rejected": -0.856473445892334,
      "logps/chosen": -346.76739501953125,
      "logps/ref_chosen": -280.81097412109375,
      "logps/ref_rejected": -264.32025146484375,
      "logps/rejected": -396.57025146484375,
      "loss": 1.0657,
      "margin_dpo/margin_mean": 66.29359436035156,
      "margin_dpo/margin_std": 73.91728973388672,
      "step": 185
    },
    {
      "epoch": 0.39790575916230364,
      "grad_norm": 22.772708892822266,
      "learning_rate": 3.781574579820464e-07,
      "logits/chosen": -0.8728398084640503,
      "logits/rejected": -0.8765541315078735,
      "logps/chosen": -379.94354248046875,
      "logps/ref_chosen": -300.13128662109375,
      "logps/ref_rejected": -263.724609375,
      "logps/rejected": -397.33709716796875,
      "loss": 1.1047,
      "margin_dpo/margin_mean": 53.800254821777344,
      "margin_dpo/margin_std": 73.82844543457031,
      "step": 190
    },
    {
      "epoch": 0.4083769633507853,
      "grad_norm": 25.31096649169922,
      "learning_rate": 3.7021375165108377e-07,
      "logits/chosen": -0.8635498881340027,
      "logits/rejected": -0.8397191762924194,
      "logps/chosen": -384.876708984375,
      "logps/ref_chosen": -287.50128173828125,
      "logps/ref_rejected": -278.3941650390625,
      "logps/rejected": -409.2007141113281,
      "loss": 1.139,
      "margin_dpo/margin_mean": 33.431087493896484,
      "margin_dpo/margin_std": 66.89299011230469,
      "step": 195
    },
    {
      "epoch": 0.418848167539267,
      "grad_norm": 29.776451110839844,
      "learning_rate": 3.621088951385353e-07,
      "logits/chosen": -0.8681343793869019,
      "logits/rejected": -0.8792493939399719,
      "logps/chosen": -378.0604553222656,
      "logps/ref_chosen": -289.7194519042969,
      "logps/ref_rejected": -264.33685302734375,
      "logps/rejected": -412.09326171875,
      "loss": 1.0912,
      "margin_dpo/margin_mean": 59.41547393798828,
      "margin_dpo/margin_std": 78.97772216796875,
      "step": 200
    },
    {
      "epoch": 0.418848167539267,
      "eval_logits/chosen": -0.8521618843078613,
      "eval_logits/rejected": -0.8391309380531311,
      "eval_logps/chosen": -374.4195251464844,
      "eval_logps/ref_chosen": -280.7076110839844,
      "eval_logps/ref_rejected": -280.1632385253906,
      "eval_logps/rejected": -434.01593017578125,
      "eval_loss": 0.5626364350318909,
      "eval_margin_dpo/margin_mean": 60.140811920166016,
      "eval_margin_dpo/margin_std": 83.66898345947266,
      "eval_runtime": 50.9457,
      "eval_samples_per_second": 39.258,
      "eval_steps_per_second": 0.628,
      "step": 200
    },
    {
      "epoch": 0.4293193717277487,
      "grad_norm": 20.72422981262207,
      "learning_rate": 3.5385375325047163e-07,
      "logits/chosen": -0.8376694917678833,
      "logits/rejected": -0.832554817199707,
      "logps/chosen": -339.31396484375,
      "logps/ref_chosen": -241.3316650390625,
      "logps/ref_rejected": -217.2843017578125,
      "logps/rejected": -353.23785400390625,
      "loss": 1.1129,
      "margin_dpo/margin_mean": 37.97121810913086,
      "margin_dpo/margin_std": 76.61421203613281,
      "step": 205
    },
    {
      "epoch": 0.4397905759162304,
      "grad_norm": 23.13551139831543,
      "learning_rate": 3.454593922550693e-07,
      "logits/chosen": -0.8016048669815063,
      "logits/rejected": -0.7972103953361511,
      "logps/chosen": -345.01312255859375,
      "logps/ref_chosen": -269.7131652832031,
      "logps/ref_rejected": -256.3033447265625,
      "logps/rejected": -377.6368408203125,
      "loss": 1.1331,
      "margin_dpo/margin_mean": 46.03350830078125,
      "margin_dpo/margin_std": 87.02532958984375,
      "step": 210
    },
    {
      "epoch": 0.450261780104712,
      "grad_norm": 28.872802734375,
      "learning_rate": 3.3693706504794243e-07,
      "logits/chosen": -0.8741207122802734,
      "logits/rejected": -0.8464077115058899,
      "logps/chosen": -361.7008361816406,
      "logps/ref_chosen": -275.9259033203125,
      "logps/ref_rejected": -259.9688415527344,
      "logps/rejected": -391.6307678222656,
      "loss": 1.1091,
      "margin_dpo/margin_mean": 45.88698959350586,
      "margin_dpo/margin_std": 67.70562744140625,
      "step": 215
    },
    {
      "epoch": 0.4607329842931937,
      "grad_norm": 30.65153694152832,
      "learning_rate": 3.2829819606729477e-07,
      "logits/chosen": -0.8631395101547241,
      "logits/rejected": -0.8379348516464233,
      "logps/chosen": -368.6546936035156,
      "logps/ref_chosen": -272.54815673828125,
      "logps/ref_rejected": -296.7209167480469,
      "logps/rejected": -459.22528076171875,
      "loss": 1.0561,
      "margin_dpo/margin_mean": 66.39784240722656,
      "margin_dpo/margin_std": 84.69873809814453,
      "step": 220
    },
    {
      "epoch": 0.4712041884816754,
      "grad_norm": 26.21302604675293,
      "learning_rate": 3.1955436597911315e-07,
      "logits/chosen": -0.8575876355171204,
      "logits/rejected": -0.8386489748954773,
      "logps/chosen": -449.95989990234375,
      "logps/ref_chosen": -353.4187927246094,
      "logps/ref_rejected": -275.82666015625,
      "logps/rejected": -434.48590087890625,
      "loss": 1.1035,
      "margin_dpo/margin_mean": 62.11814498901367,
      "margin_dpo/margin_std": 93.59648132324219,
      "step": 225
    },
    {
      "epoch": 0.4816753926701571,
      "grad_norm": 22.068635940551758,
      "learning_rate": 3.1071729615293424e-07,
      "logits/chosen": -0.8830578923225403,
      "logits/rejected": -0.8567687273025513,
      "logps/chosen": -436.341796875,
      "logps/ref_chosen": -346.2792663574219,
      "logps/ref_rejected": -333.0526123046875,
      "logps/rejected": -487.6953125,
      "loss": 1.0555,
      "margin_dpo/margin_mean": 64.58020782470703,
      "margin_dpo/margin_std": 82.20867919921875,
      "step": 230
    },
    {
      "epoch": 0.49214659685863876,
      "grad_norm": 22.247276306152344,
      "learning_rate": 3.017988329489923e-07,
      "logits/chosen": -0.8050928115844727,
      "logits/rejected": -0.8064786791801453,
      "logps/chosen": -377.37249755859375,
      "logps/ref_chosen": -293.1265869140625,
      "logps/ref_rejected": -292.41009521484375,
      "logps/rejected": -430.84149169921875,
      "loss": 1.0875,
      "margin_dpo/margin_mean": 54.1854362487793,
      "margin_dpo/margin_std": 79.34012603759766,
      "step": 235
    },
    {
      "epoch": 0.5026178010471204,
      "grad_norm": 28.167221069335938,
      "learning_rate": 2.9281093183781403e-07,
      "logits/chosen": -0.7880058288574219,
      "logits/rejected": -0.8285211324691772,
      "logps/chosen": -289.57073974609375,
      "logps/ref_chosen": -212.98837280273438,
      "logps/ref_rejected": -238.8592987060547,
      "logps/rejected": -380.24237060546875,
      "loss": 1.0831,
      "margin_dpo/margin_mean": 64.80075073242188,
      "margin_dpo/margin_std": 87.5298843383789,
      "step": 240
    },
    {
      "epoch": 0.5130890052356021,
      "grad_norm": 21.444618225097656,
      "learning_rate": 2.837656413735479e-07,
      "logits/chosen": -0.8537348508834839,
      "logits/rejected": -0.8319905400276184,
      "logps/chosen": -341.90264892578125,
      "logps/ref_chosen": -270.50933837890625,
      "logps/ref_rejected": -283.96795654296875,
      "logps/rejected": -410.18414306640625,
      "loss": 1.0671,
      "margin_dpo/margin_mean": 54.822853088378906,
      "margin_dpo/margin_std": 83.7713394165039,
      "step": 245
    },
    {
      "epoch": 0.5235602094240838,
      "grad_norm": 24.929975509643555,
      "learning_rate": 2.7467508704251135e-07,
      "logits/chosen": -0.872177243232727,
      "logits/rejected": -0.8667652010917664,
      "logps/chosen": -363.5415344238281,
      "logps/ref_chosen": -278.5409240722656,
      "logps/ref_rejected": -289.39508056640625,
      "logps/rejected": -423.7264709472656,
      "loss": 1.1319,
      "margin_dpo/margin_mean": 49.33078384399414,
      "margin_dpo/margin_std": 79.461669921875,
      "step": 250
    },
    {
      "epoch": 0.5340314136125655,
      "grad_norm": 24.554033279418945,
      "learning_rate": 2.655514550086086e-07,
      "logits/chosen": -0.8809840083122253,
      "logits/rejected": -0.8628867864608765,
      "logps/chosen": -393.5577392578125,
      "logps/ref_chosen": -294.0670166015625,
      "logps/ref_rejected": -255.25021362304688,
      "logps/rejected": -419.11676025390625,
      "loss": 1.091,
      "margin_dpo/margin_mean": 64.37581634521484,
      "margin_dpo/margin_std": 85.99962615966797,
      "step": 255
    },
    {
      "epoch": 0.5445026178010471,
      "grad_norm": 24.0455265045166,
      "learning_rate": 2.5640697577740815e-07,
      "logits/chosen": -0.8196985125541687,
      "logits/rejected": -0.8147414326667786,
      "logps/chosen": -386.85418701171875,
      "logps/ref_chosen": -276.14508056640625,
      "logps/ref_rejected": -225.4951171875,
      "logps/rejected": -392.1053466796875,
      "loss": 1.1138,
      "margin_dpo/margin_mean": 55.9011344909668,
      "margin_dpo/margin_std": 81.55867004394531,
      "step": 260
    },
    {
      "epoch": 0.5549738219895288,
      "grad_norm": 22.498727798461914,
      "learning_rate": 2.4725390780077905e-07,
      "logits/chosen": -0.7659963965415955,
      "logits/rejected": -0.7464796900749207,
      "logps/chosen": -369.7604675292969,
      "logps/ref_chosen": -257.68145751953125,
      "logps/ref_rejected": -239.24484252929688,
      "logps/rejected": -427.9410705566406,
      "loss": 1.0651,
      "margin_dpo/margin_mean": 76.6172103881836,
      "margin_dpo/margin_std": 92.36399841308594,
      "step": 265
    },
    {
      "epoch": 0.5654450261780105,
      "grad_norm": 19.695539474487305,
      "learning_rate": 2.381045210440644e-07,
      "logits/chosen": -0.8541289567947388,
      "logits/rejected": -0.8301209211349487,
      "logps/chosen": -380.1869812011719,
      "logps/ref_chosen": -258.50482177734375,
      "logps/ref_rejected": -252.1217498779297,
      "logps/rejected": -426.6290588378906,
      "loss": 1.0646,
      "margin_dpo/margin_mean": 52.8251838684082,
      "margin_dpo/margin_std": 84.49119567871094,
      "step": 270
    },
    {
      "epoch": 0.5759162303664922,
      "grad_norm": 17.279773712158203,
      "learning_rate": 2.2897108053782e-07,
      "logits/chosen": -0.8132762908935547,
      "logits/rejected": -0.815741240978241,
      "logps/chosen": -415.64617919921875,
      "logps/ref_chosen": -320.1709289550781,
      "logps/ref_rejected": -278.55364990234375,
      "logps/rejected": -443.93499755859375,
      "loss": 1.0293,
      "margin_dpo/margin_mean": 69.90602111816406,
      "margin_dpo/margin_std": 85.13655853271484,
      "step": 275
    },
    {
      "epoch": 0.5863874345549738,
      "grad_norm": 17.42207145690918,
      "learning_rate": 2.1986582993616925e-07,
      "logits/chosen": -0.7942547798156738,
      "logits/rejected": -0.8006545901298523,
      "logps/chosen": -399.39056396484375,
      "logps/ref_chosen": -303.01043701171875,
      "logps/ref_rejected": -304.8597412109375,
      "logps/rejected": -473.74407958984375,
      "loss": 1.0884,
      "margin_dpo/margin_mean": 72.5042724609375,
      "margin_dpo/margin_std": 83.59529876708984,
      "step": 280
    },
    {
      "epoch": 0.5968586387434555,
      "grad_norm": 29.439176559448242,
      "learning_rate": 2.1080097510381294e-07,
      "logits/chosen": -0.8379713296890259,
      "logits/rejected": -0.8423024415969849,
      "logps/chosen": -410.0982360839844,
      "logps/ref_chosen": -284.4503479003906,
      "logps/ref_rejected": -295.3260803222656,
      "logps/rejected": -488.0785217285156,
      "loss": 1.1025,
      "margin_dpo/margin_mean": 67.1044692993164,
      "margin_dpo/margin_std": 85.94883728027344,
      "step": 285
    },
    {
      "epoch": 0.6073298429319371,
      "grad_norm": 26.986276626586914,
      "learning_rate": 2.0178866775369774e-07,
      "logits/chosen": -0.8594253659248352,
      "logits/rejected": -0.842856764793396,
      "logps/chosen": -406.8668212890625,
      "logps/ref_chosen": -282.28424072265625,
      "logps/ref_rejected": -245.0934295654297,
      "logps/rejected": -418.887939453125,
      "loss": 1.1327,
      "margin_dpo/margin_mean": 49.21195602416992,
      "margin_dpo/margin_std": 83.4120101928711,
      "step": 290
    },
    {
      "epoch": 0.6178010471204188,
      "grad_norm": 17.56438446044922,
      "learning_rate": 1.928409891572757e-07,
      "logits/chosen": -0.826249897480011,
      "logits/rejected": -0.8039449453353882,
      "logps/chosen": -403.50164794921875,
      "logps/ref_chosen": -280.5242004394531,
      "logps/ref_rejected": -262.0945129394531,
      "logps/rejected": -443.8013610839844,
      "loss": 1.0541,
      "margin_dpo/margin_mean": 58.729454040527344,
      "margin_dpo/margin_std": 91.69632720947266,
      "step": 295
    },
    {
      "epoch": 0.6282722513089005,
      "grad_norm": 28.401439666748047,
      "learning_rate": 1.839699339491937e-07,
      "logits/chosen": -0.8613263964653015,
      "logits/rejected": -0.8485943675041199,
      "logps/chosen": -364.01141357421875,
      "logps/ref_chosen": -255.989990234375,
      "logps/ref_rejected": -263.67828369140625,
      "logps/rejected": -434.90423583984375,
      "loss": 1.0651,
      "margin_dpo/margin_mean": 63.20460891723633,
      "margin_dpo/margin_std": 98.46973419189453,
      "step": 300
    },
    {
      "epoch": 0.6387434554973822,
      "grad_norm": 21.466861724853516,
      "learning_rate": 1.7518739404812155e-07,
      "logits/chosen": -0.8165709376335144,
      "logits/rejected": -0.813851535320282,
      "logps/chosen": -397.8428649902344,
      "logps/ref_chosen": -275.48211669921875,
      "logps/ref_rejected": -239.78201293945312,
      "logps/rejected": -420.2521057128906,
      "loss": 1.0263,
      "margin_dpo/margin_mean": 58.10936737060547,
      "margin_dpo/margin_std": 86.70993041992188,
      "step": 305
    },
    {
      "epoch": 0.6492146596858639,
      "grad_norm": 23.69497299194336,
      "learning_rate": 1.6650514271527465e-07,
      "logits/chosen": -0.8850182294845581,
      "logits/rejected": -0.8541525602340698,
      "logps/chosen": -434.433349609375,
      "logps/ref_chosen": -298.96624755859375,
      "logps/ref_rejected": -281.476806640625,
      "logps/rejected": -478.77020263671875,
      "loss": 1.061,
      "margin_dpo/margin_mean": 61.826271057128906,
      "margin_dpo/margin_std": 83.03723907470703,
      "step": 310
    },
    {
      "epoch": 0.6596858638743456,
      "grad_norm": 23.33465576171875,
      "learning_rate": 1.5793481877199943e-07,
      "logits/chosen": -0.8339045643806458,
      "logits/rejected": -0.8051811456680298,
      "logps/chosen": -461.6744079589844,
      "logps/ref_chosen": -319.5059814453125,
      "logps/ref_rejected": -295.2557067871094,
      "logps/rejected": -506.0589294433594,
      "loss": 1.0603,
      "margin_dpo/margin_mean": 68.63478088378906,
      "margin_dpo/margin_std": 97.61984252929688,
      "step": 315
    },
    {
      "epoch": 0.6701570680628273,
      "grad_norm": 23.42766761779785,
      "learning_rate": 1.4948791099758052e-07,
      "logits/chosen": -0.8916772603988647,
      "logits/rejected": -0.8678488731384277,
      "logps/chosen": -403.6375732421875,
      "logps/ref_chosen": -274.22412109375,
      "logps/ref_rejected": -253.42117309570312,
      "logps/rejected": -431.6688537597656,
      "loss": 0.9986,
      "margin_dpo/margin_mean": 48.834320068359375,
      "margin_dpo/margin_std": 76.32218933105469,
      "step": 320
    },
    {
      "epoch": 0.680628272251309,
      "grad_norm": 31.045774459838867,
      "learning_rate": 1.4117574272818386e-07,
      "logits/chosen": -0.8033218383789062,
      "logits/rejected": -0.7930227518081665,
      "logps/chosen": -395.0705871582031,
      "logps/ref_chosen": -259.73590087890625,
      "logps/ref_rejected": -271.35369873046875,
      "logps/rejected": -458.1595153808594,
      "loss": 1.0877,
      "margin_dpo/margin_mean": 51.47113037109375,
      "margin_dpo/margin_std": 90.4473648071289,
      "step": 325
    },
    {
      "epoch": 0.6910994764397905,
      "grad_norm": 24.11140251159668,
      "learning_rate": 1.3300945667758012e-07,
      "logits/chosen": -0.8636866807937622,
      "logits/rejected": -0.8337670564651489,
      "logps/chosen": -421.36749267578125,
      "logps/ref_chosen": -288.8438720703125,
      "logps/ref_rejected": -285.99853515625,
      "logps/rejected": -491.48919677734375,
      "loss": 1.0256,
      "margin_dpo/margin_mean": 72.96711730957031,
      "margin_dpo/margin_std": 98.3145523071289,
      "step": 330
    },
    {
      "epoch": 0.7015706806282722,
      "grad_norm": 30.136749267578125,
      "learning_rate": 1.2500000000000005e-07,
      "logits/chosen": -0.8628665208816528,
      "logits/rejected": -0.8348671197891235,
      "logps/chosen": -424.05352783203125,
      "logps/ref_chosen": -284.08929443359375,
      "logps/ref_rejected": -245.40060424804688,
      "logps/rejected": -441.79852294921875,
      "loss": 1.0605,
      "margin_dpo/margin_mean": 56.433692932128906,
      "margin_dpo/margin_std": 105.3170166015625,
      "step": 335
    },
    {
      "epoch": 0.7120418848167539,
      "grad_norm": 33.01960754394531,
      "learning_rate": 1.1715810961514072e-07,
      "logits/chosen": -0.9020859599113464,
      "logits/rejected": -0.8989129066467285,
      "logps/chosen": -425.54400634765625,
      "logps/ref_chosen": -281.47247314453125,
      "logps/ref_rejected": -256.0491027832031,
      "logps/rejected": -472.6197204589844,
      "loss": 1.0478,
      "margin_dpo/margin_mean": 72.49906921386719,
      "margin_dpo/margin_std": 109.55391693115234,
      "step": 340
    },
    {
      "epoch": 0.7225130890052356,
      "grad_norm": 26.777374267578125,
      "learning_rate": 1.09494297815e-07,
      "logits/chosen": -0.8701359033584595,
      "logits/rejected": -0.8566125631332397,
      "logps/chosen": -399.4504699707031,
      "logps/ref_chosen": -258.94598388671875,
      "logps/ref_rejected": -262.75830078125,
      "logps/rejected": -455.2816467285156,
      "loss": 1.1001,
      "margin_dpo/margin_mean": 52.018768310546875,
      "margin_dpo/margin_std": 89.29933166503906,
      "step": 345
    },
    {
      "epoch": 0.7329842931937173,
      "grad_norm": 26.882720947265625,
      "learning_rate": 1.0201883817182949e-07,
      "logits/chosen": -0.8565577268600464,
      "logits/rejected": -0.8601759672164917,
      "logps/chosen": -385.0584411621094,
      "logps/ref_chosen": -242.0128936767578,
      "logps/ref_rejected": -219.5886993408203,
      "logps/rejected": -417.0812072753906,
      "loss": 1.065,
      "margin_dpo/margin_mean": 54.44697952270508,
      "margin_dpo/margin_std": 90.08226013183594,
      "step": 350
    },
    {
      "epoch": 0.743455497382199,
      "grad_norm": 26.08757781982422,
      "learning_rate": 9.474175176609956e-08,
      "logits/chosen": -0.8443655967712402,
      "logits/rejected": -0.838814914226532,
      "logps/chosen": -453.3646545410156,
      "logps/ref_chosen": -309.3603820800781,
      "logps/ref_rejected": -281.742431640625,
      "logps/rejected": -493.95404052734375,
      "loss": 1.1191,
      "margin_dpo/margin_mean": 68.20734405517578,
      "margin_dpo/margin_std": 108.90913391113281,
      "step": 355
    },
    {
      "epoch": 0.7539267015706806,
      "grad_norm": 26.802879333496094,
      "learning_rate": 8.76727937529367e-08,
      "logits/chosen": -0.7590880393981934,
      "logits/rejected": -0.7674987316131592,
      "logps/chosen": -392.798828125,
      "logps/ref_chosen": -254.25436401367188,
      "logps/ref_rejected": -221.65145874023438,
      "logps/rejected": -422.1822204589844,
      "loss": 1.0568,
      "margin_dpo/margin_mean": 61.9863166809082,
      "margin_dpo/margin_std": 87.15711975097656,
      "step": 360
    },
    {
      "epoch": 0.7643979057591623,
      "grad_norm": 21.973236083984375,
      "learning_rate": 8.082144028504231e-08,
      "logits/chosen": -0.8197676539421082,
      "logits/rejected": -0.8475183248519897,
      "logps/chosen": -364.2659912109375,
      "logps/ref_chosen": -242.38778686523438,
      "logps/ref_rejected": -208.74813842773438,
      "logps/rejected": -386.58636474609375,
      "loss": 1.055,
      "margin_dpo/margin_mean": 55.96006393432617,
      "margin_dpo/margin_std": 79.6556167602539,
      "step": 365
    },
    {
      "epoch": 0.774869109947644,
      "grad_norm": 18.004220962524414,
      "learning_rate": 7.419687580962222e-08,
      "logits/chosen": -0.850188136100769,
      "logits/rejected": -0.8511862754821777,
      "logps/chosen": -414.29644775390625,
      "logps/ref_chosen": -293.2469787597656,
      "logps/ref_rejected": -280.17156982421875,
      "logps/rejected": -468.7672424316406,
      "loss": 0.9856,
      "margin_dpo/margin_mean": 67.54612731933594,
      "margin_dpo/margin_std": 89.1795425415039,
      "step": 370
    },
    {
      "epoch": 0.7853403141361257,
      "grad_norm": 22.512737274169922,
      "learning_rate": 6.780798075635675e-08,
      "logits/chosen": -0.8369812965393066,
      "logits/rejected": -0.8140292167663574,
      "logps/chosen": -450.01275634765625,
      "logps/ref_chosen": -329.85968017578125,
      "logps/ref_rejected": -237.68679809570312,
      "logps/rejected": -409.7642517089844,
      "loss": 1.0446,
      "margin_dpo/margin_mean": 51.924415588378906,
      "margin_dpo/margin_std": 90.19831848144531,
      "step": 375
    },
    {
      "epoch": 0.7958115183246073,
      "grad_norm": 29.517168045043945,
      "learning_rate": 6.166331963291519e-08,
      "logits/chosen": -0.8067518472671509,
      "logits/rejected": -0.8046929240226746,
      "logps/chosen": -407.7606506347656,
      "logps/ref_chosen": -294.3583679199219,
      "logps/ref_rejected": -266.3445739746094,
      "logps/rejected": -433.68463134765625,
      "loss": 1.1103,
      "margin_dpo/margin_mean": 53.93775177001953,
      "margin_dpo/margin_std": 93.24676513671875,
      "step": 380
    },
    {
      "epoch": 0.806282722513089,
      "grad_norm": 24.049985885620117,
      "learning_rate": 5.57711295439732e-08,
      "logits/chosen": -0.8399543762207031,
      "logits/rejected": -0.8283275365829468,
      "logps/chosen": -422.85137939453125,
      "logps/ref_chosen": -314.8267822265625,
      "logps/ref_rejected": -292.8830871582031,
      "logps/rejected": -470.80560302734375,
      "loss": 1.0282,
      "margin_dpo/margin_mean": 69.89784240722656,
      "margin_dpo/margin_std": 93.79012298583984,
      "step": 385
    },
    {
      "epoch": 0.8167539267015707,
      "grad_norm": 19.402238845825195,
      "learning_rate": 5.013930914912476e-08,
      "logits/chosen": -0.8243298530578613,
      "logits/rejected": -0.8369429707527161,
      "logps/chosen": -411.41290283203125,
      "logps/ref_chosen": -283.4703063964844,
      "logps/ref_rejected": -237.56201171875,
      "logps/rejected": -418.91900634765625,
      "loss": 1.0603,
      "margin_dpo/margin_mean": 53.41436767578125,
      "margin_dpo/margin_std": 72.03800964355469,
      "step": 390
    },
    {
      "epoch": 0.8272251308900523,
      "grad_norm": 19.115346908569336,
      "learning_rate": 4.477540807448832e-08,
      "logits/chosen": -0.8676174879074097,
      "logits/rejected": -0.8561701774597168,
      "logps/chosen": -418.7401428222656,
      "logps/ref_chosen": -307.20220947265625,
      "logps/ref_rejected": -245.660400390625,
      "logps/rejected": -418.5318908691406,
      "loss": 1.0248,
      "margin_dpo/margin_mean": 61.33369827270508,
      "margin_dpo/margin_std": 90.22969055175781,
      "step": 395
    },
    {
      "epoch": 0.837696335078534,
      "grad_norm": 28.535594940185547,
      "learning_rate": 3.968661679220467e-08,
      "logits/chosen": -0.8202043771743774,
      "logits/rejected": -0.8029823303222656,
      "logps/chosen": -394.8623046875,
      "logps/ref_chosen": -266.63018798828125,
      "logps/ref_rejected": -285.74224853515625,
      "logps/rejected": -481.6102600097656,
      "loss": 1.1088,
      "margin_dpo/margin_mean": 67.63584899902344,
      "margin_dpo/margin_std": 88.04866027832031,
      "step": 400
    },
    {
      "epoch": 0.837696335078534,
      "eval_logits/chosen": -0.8223316073417664,
      "eval_logits/rejected": -0.8121299743652344,
      "eval_logps/chosen": -403.2485046386719,
      "eval_logps/ref_chosen": -280.7076110839844,
      "eval_logps/ref_rejected": -280.1632385253906,
      "eval_logps/rejected": -474.8625183105469,
      "eval_loss": 0.5358365774154663,
      "eval_margin_dpo/margin_mean": 72.15837860107422,
      "eval_margin_dpo/margin_std": 96.15919494628906,
      "eval_runtime": 50.8127,
      "eval_samples_per_second": 39.36,
      "eval_steps_per_second": 0.63,
      "step": 400
    },
    {
      "epoch": 0.8481675392670157,
      "grad_norm": 16.877710342407227,
      "learning_rate": 3.487975698139084e-08,
      "logits/chosen": -0.8321081399917603,
      "logits/rejected": -0.8053818941116333,
      "logps/chosen": -420.28216552734375,
      "logps/ref_chosen": -293.24041748046875,
      "logps/ref_rejected": -270.2939453125,
      "logps/rejected": -457.31097412109375,
      "loss": 1.0084,
      "margin_dpo/margin_mean": 59.9753303527832,
      "margin_dpo/margin_std": 80.72087097167969,
      "step": 405
    },
    {
      "epoch": 0.8586387434554974,
      "grad_norm": 21.314437866210938,
      "learning_rate": 3.036127238347164e-08,
      "logits/chosen": -0.9076123237609863,
      "logits/rejected": -0.9141916036605835,
      "logps/chosen": -440.6188049316406,
      "logps/ref_chosen": -298.21405029296875,
      "logps/ref_rejected": -287.6912536621094,
      "logps/rejected": -481.20684814453125,
      "loss": 1.0709,
      "margin_dpo/margin_mean": 51.1108512878418,
      "margin_dpo/margin_std": 80.4267349243164,
      "step": 410
    },
    {
      "epoch": 0.8691099476439791,
      "grad_norm": 19.411727905273438,
      "learning_rate": 2.613722016414943e-08,
      "logits/chosen": -0.7660075426101685,
      "logits/rejected": -0.7644788026809692,
      "logps/chosen": -384.01080322265625,
      "logps/ref_chosen": -254.6023712158203,
      "logps/ref_rejected": -278.03485107421875,
      "logps/rejected": -464.08380126953125,
      "loss": 1.0408,
      "margin_dpo/margin_mean": 56.64048385620117,
      "margin_dpo/margin_std": 90.71663665771484,
      "step": 415
    },
    {
      "epoch": 0.8795811518324608,
      "grad_norm": 33.511924743652344,
      "learning_rate": 2.2213262793589482e-08,
      "logits/chosen": -0.8396707773208618,
      "logits/rejected": -0.7929636240005493,
      "logps/chosen": -455.02288818359375,
      "logps/ref_chosen": -328.2628479003906,
      "logps/ref_rejected": -267.16888427734375,
      "logps/rejected": -482.4369201660156,
      "loss": 1.0272,
      "margin_dpo/margin_mean": 88.50798034667969,
      "margin_dpo/margin_std": 90.31407928466797,
      "step": 420
    },
    {
      "epoch": 0.8900523560209425,
      "grad_norm": 25.034595489501953,
      "learning_rate": 1.8594660455706763e-08,
      "logits/chosen": -0.8549168705940247,
      "logits/rejected": -0.8669727444648743,
      "logps/chosen": -471.65289306640625,
      "logps/ref_chosen": -341.20465087890625,
      "logps/ref_rejected": -259.51910400390625,
      "logps/rejected": -463.3876037597656,
      "loss": 1.0265,
      "margin_dpo/margin_mean": 73.42025756835938,
      "margin_dpo/margin_std": 100.29209899902344,
      "step": 425
    },
    {
      "epoch": 0.900523560209424,
      "grad_norm": 26.290597915649414,
      "learning_rate": 1.5286263996730026e-08,
      "logits/chosen": -0.8484354019165039,
      "logits/rejected": -0.831134021282196,
      "logps/chosen": -436.2567443847656,
      "logps/ref_chosen": -300.48785400390625,
      "logps/ref_rejected": -252.441162109375,
      "logps/rejected": -447.345703125,
      "loss": 1.0004,
      "margin_dpo/margin_mean": 59.13561248779297,
      "margin_dpo/margin_std": 92.1114273071289,
      "step": 430
    },
    {
      "epoch": 0.9109947643979057,
      "grad_norm": 32.428932189941406,
      "learning_rate": 1.2292508422495157e-08,
      "logits/chosen": -0.8825214505195618,
      "logits/rejected": -0.8753899335861206,
      "logps/chosen": -389.99371337890625,
      "logps/ref_chosen": -264.25225830078125,
      "logps/ref_rejected": -237.9748992919922,
      "logps/rejected": -427.88104248046875,
      "loss": 1.0395,
      "margin_dpo/margin_mean": 64.16473388671875,
      "margin_dpo/margin_std": 94.56092834472656,
      "step": 435
    },
    {
      "epoch": 0.9214659685863874,
      "grad_norm": 24.546009063720703,
      "learning_rate": 9.617406953185136e-09,
      "logits/chosen": -0.8393553495407104,
      "logits/rejected": -0.8456501960754395,
      "logps/chosen": -449.50836181640625,
      "logps/ref_chosen": -304.78619384765625,
      "logps/ref_rejected": -334.1107482910156,
      "logps/rejected": -516.7224731445312,
      "loss": 1.1006,
      "margin_dpo/margin_mean": 37.8895378112793,
      "margin_dpo/margin_std": 85.40432739257812,
      "step": 440
    },
    {
      "epoch": 0.9319371727748691,
      "grad_norm": 23.76576805114746,
      "learning_rate": 7.2645456434869965e-09,
      "logits/chosen": -0.838398277759552,
      "logits/rejected": -0.8250478506088257,
      "logps/chosen": -448.4908142089844,
      "logps/ref_chosen": -293.74560546875,
      "logps/ref_rejected": -246.31625366210938,
      "logps/rejected": -445.4239807128906,
      "loss": 1.0524,
      "margin_dpo/margin_mean": 44.362464904785156,
      "margin_dpo/margin_std": 85.54446411132812,
      "step": 445
    },
    {
      "epoch": 0.9424083769633508,
      "grad_norm": 23.184553146362305,
      "learning_rate": 5.2370785753763356e-09,
      "logits/chosen": -0.8070716857910156,
      "logits/rejected": -0.8125391006469727,
      "logps/chosen": -381.4776306152344,
      "logps/ref_chosen": -258.7225646972656,
      "logps/ref_rejected": -240.88504028320312,
      "logps/rejected": -421.876220703125,
      "loss": 1.0018,
      "margin_dpo/margin_mean": 58.236106872558594,
      "margin_dpo/margin_std": 83.58892822265625,
      "step": 450
    },
    {
      "epoch": 0.9528795811518325,
      "grad_norm": 22.7464542388916,
      "learning_rate": 3.5377236299748147e-09,
      "logits/chosen": -0.8362730741500854,
      "logits/rejected": -0.8388243913650513,
      "logps/chosen": -474.7862854003906,
      "logps/ref_chosen": -331.77642822265625,
      "logps/ref_rejected": -305.6571960449219,
      "logps/rejected": -511.6985778808594,
      "loss": 1.0562,
      "margin_dpo/margin_mean": 63.031532287597656,
      "margin_dpo/margin_std": 97.10409545898438,
      "step": 455
    },
    {
      "epoch": 0.9633507853403142,
      "grad_norm": 24.231159210205078,
      "learning_rate": 2.168758844148272e-09,
      "logits/chosen": -0.8514804840087891,
      "logits/rejected": -0.8368584513664246,
      "logps/chosen": -404.0189514160156,
      "logps/ref_chosen": -267.02508544921875,
      "logps/ref_rejected": -267.7117919921875,
      "logps/rejected": -456.15289306640625,
      "loss": 1.0653,
      "margin_dpo/margin_mean": 51.44722366333008,
      "margin_dpo/margin_std": 92.69256591796875,
      "step": 460
    },
    {
      "epoch": 0.9738219895287958,
      "grad_norm": 19.262453079223633,
      "learning_rate": 1.1320193567288527e-09,
      "logits/chosen": -0.859015166759491,
      "logits/rejected": -0.8542447090148926,
      "logps/chosen": -407.23431396484375,
      "logps/ref_chosen": -276.2010803222656,
      "logps/ref_rejected": -229.0570068359375,
      "logps/rejected": -442.80023193359375,
      "loss": 1.0555,
      "margin_dpo/margin_mean": 82.71009826660156,
      "margin_dpo/margin_std": 78.12190246582031,
      "step": 465
    },
    {
      "epoch": 0.9842931937172775,
      "grad_norm": 20.87303352355957,
      "learning_rate": 4.288949484559934e-10,
      "logits/chosen": -0.8113743662834167,
      "logits/rejected": -0.7894054651260376,
      "logps/chosen": -377.26910400390625,
      "logps/ref_chosen": -257.05743408203125,
      "logps/ref_rejected": -235.47738647460938,
      "logps/rejected": -443.84393310546875,
      "loss": 1.0227,
      "margin_dpo/margin_mean": 88.15492248535156,
      "margin_dpo/margin_std": 88.85401153564453,
      "step": 470
    },
    {
      "epoch": 0.9947643979057592,
      "grad_norm": 22.590627670288086,
      "learning_rate": 6.032817893297793e-11,
      "logits/chosen": -0.8724172711372375,
      "logits/rejected": -0.8867457509040833,
      "logps/chosen": -366.89349365234375,
      "logps/ref_chosen": -235.49948120117188,
      "logps/ref_rejected": -264.59051513671875,
      "logps/rejected": -458.58197021484375,
      "loss": 1.0225,
      "margin_dpo/margin_mean": 62.59749221801758,
      "margin_dpo/margin_std": 81.80487060546875,
      "step": 475
    },
    {
      "epoch": 0.9989528795811519,
      "step": 477,
      "total_flos": 0.0,
      "train_loss": 1.1476524381017785,
      "train_runtime": 4456.6592,
      "train_samples_per_second": 13.718,
      "train_steps_per_second": 0.107
    }
  ],
  "logging_steps": 5,
  "max_steps": 477,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 200,
  "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
}