ds_coder_adamw_iter4 / trainer_state.json
yiran-wang3's picture
End of training
c18b83e verified
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 1.0,
"eval_steps": 100,
"global_step": 34,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"debug/policy_chosen_logits": 29.054555892944336,
"debug/policy_chosen_logps": -394.62103271484375,
"debug/policy_rejected_logits": 30.527902603149414,
"debug/policy_rejected_logps": -401.28692626953125,
"debug/reference_chosen_logps": -394.62103271484375,
"debug/reference_rejected_logps": -401.28692626953125,
"epoch": 0.029411764705882353,
"grad_norm": 5.461222893396727,
"learning_rate": 1e-06,
"logits/chosen": 29.054555892944336,
"logits/rejected": 30.527902603149414,
"logps/chosen": -394.62103271484375,
"logps/rejected": -401.28692626953125,
"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": 26.592811584472656,
"debug/policy_chosen_logps": -404.9052429199219,
"debug/policy_rejected_logits": 23.82817268371582,
"debug/policy_rejected_logps": -377.16888427734375,
"debug/reference_chosen_logps": -405.39239501953125,
"debug/reference_rejected_logps": -377.3977966308594,
"epoch": 0.058823529411764705,
"grad_norm": 5.26155447031347,
"learning_rate": 1e-06,
"logits/chosen": 26.592811584472656,
"logits/rejected": 23.82817268371582,
"logps/chosen": -404.9052429199219,
"logps/rejected": -377.16888427734375,
"loss": 0.4996,
"rewards/accuracies": 0.625,
"rewards/chosen": 0.004871406592428684,
"rewards/margins": 0.002582397311925888,
"rewards/rejected": 0.002289008814841509,
"step": 2
},
{
"debug/policy_chosen_logits": 29.616479873657227,
"debug/policy_chosen_logps": -402.9261474609375,
"debug/policy_rejected_logits": 27.229713439941406,
"debug/policy_rejected_logps": -402.63165283203125,
"debug/reference_chosen_logps": -402.37353515625,
"debug/reference_rejected_logps": -403.0528564453125,
"epoch": 0.08823529411764706,
"grad_norm": 5.686734369927699,
"learning_rate": 1e-06,
"logits/chosen": 29.616479873657227,
"logits/rejected": 27.229713439941406,
"logps/chosen": -402.9261474609375,
"logps/rejected": -402.63165283203125,
"loss": 0.5037,
"rewards/accuracies": 0.375,
"rewards/chosen": -0.005526161286979914,
"rewards/margins": -0.009738006629049778,
"rewards/rejected": 0.004211845342069864,
"step": 3
},
{
"debug/policy_chosen_logits": 28.472673416137695,
"debug/policy_chosen_logps": -406.1566467285156,
"debug/policy_rejected_logits": 28.92209815979004,
"debug/policy_rejected_logps": -408.373046875,
"debug/reference_chosen_logps": -406.66375732421875,
"debug/reference_rejected_logps": -407.8741760253906,
"epoch": 0.11764705882352941,
"grad_norm": 5.378959566289513,
"learning_rate": 1e-06,
"logits/chosen": 28.472673416137695,
"logits/rejected": 28.92209815979004,
"logps/chosen": -406.1566467285156,
"logps/rejected": -408.373046875,
"loss": 0.5029,
"rewards/accuracies": 0.5,
"rewards/chosen": 0.005071220453828573,
"rewards/margins": 0.010060157626867294,
"rewards/rejected": -0.004988937638700008,
"step": 4
},
{
"debug/policy_chosen_logits": 31.492525100708008,
"debug/policy_chosen_logps": -403.90069580078125,
"debug/policy_rejected_logits": 30.92120361328125,
"debug/policy_rejected_logps": -411.68096923828125,
"debug/reference_chosen_logps": -404.43756103515625,
"debug/reference_rejected_logps": -412.2107849121094,
"epoch": 0.14705882352941177,
"grad_norm": 5.270120936652155,
"learning_rate": 1e-06,
"logits/chosen": 31.492525100708008,
"logits/rejected": 30.92120361328125,
"logps/chosen": -403.90069580078125,
"logps/rejected": -411.68096923828125,
"loss": 0.4989,
"rewards/accuracies": 0.5,
"rewards/chosen": 0.005368461366742849,
"rewards/margins": 7.053330773487687e-05,
"rewards/rejected": 0.005297927651554346,
"step": 5
},
{
"debug/policy_chosen_logits": 31.689584732055664,
"debug/policy_chosen_logps": -426.2255554199219,
"debug/policy_rejected_logits": 31.215805053710938,
"debug/policy_rejected_logps": -414.5820007324219,
"debug/reference_chosen_logps": -426.4078369140625,
"debug/reference_rejected_logps": -414.718017578125,
"epoch": 0.17647058823529413,
"grad_norm": 5.511863687507664,
"learning_rate": 1e-06,
"logits/chosen": 31.689584732055664,
"logits/rejected": 31.215805053710938,
"logps/chosen": -426.2255554199219,
"logps/rejected": -414.5820007324219,
"loss": 0.5044,
"rewards/accuracies": 0.5,
"rewards/chosen": 0.001822624122723937,
"rewards/margins": 0.0004624177236109972,
"rewards/rejected": 0.0013602066319435835,
"step": 6
},
{
"debug/policy_chosen_logits": 27.33298683166504,
"debug/policy_chosen_logps": -410.829345703125,
"debug/policy_rejected_logits": 26.893630981445312,
"debug/policy_rejected_logps": -394.69989013671875,
"debug/reference_chosen_logps": -410.4607849121094,
"debug/reference_rejected_logps": -395.7179870605469,
"epoch": 0.20588235294117646,
"grad_norm": 5.290456967833764,
"learning_rate": 1e-06,
"logits/chosen": 27.33298683166504,
"logits/rejected": 26.893630981445312,
"logps/chosen": -410.829345703125,
"logps/rejected": -394.69989013671875,
"loss": 0.501,
"rewards/accuracies": 0.25,
"rewards/chosen": -0.003685684408992529,
"rewards/margins": -0.013866348192095757,
"rewards/rejected": 0.010180664248764515,
"step": 7
},
{
"debug/policy_chosen_logits": 26.145978927612305,
"debug/policy_chosen_logps": -409.4605712890625,
"debug/policy_rejected_logits": 27.502052307128906,
"debug/policy_rejected_logps": -401.6300048828125,
"debug/reference_chosen_logps": -409.36480712890625,
"debug/reference_rejected_logps": -402.3456726074219,
"epoch": 0.23529411764705882,
"grad_norm": 5.259206650868919,
"learning_rate": 1e-06,
"logits/chosen": 26.145978927612305,
"logits/rejected": 27.502052307128906,
"logps/chosen": -409.4605712890625,
"logps/rejected": -401.6300048828125,
"loss": 0.4965,
"rewards/accuracies": 0.375,
"rewards/chosen": -0.0009576798183843493,
"rewards/margins": -0.00811409018933773,
"rewards/rejected": 0.007156409788876772,
"step": 8
},
{
"debug/policy_chosen_logits": 29.184545516967773,
"debug/policy_chosen_logps": -393.89373779296875,
"debug/policy_rejected_logits": 28.960050582885742,
"debug/policy_rejected_logps": -382.4853820800781,
"debug/reference_chosen_logps": -393.92431640625,
"debug/reference_rejected_logps": -382.33984375,
"epoch": 0.2647058823529412,
"grad_norm": 5.662262990371031,
"learning_rate": 1e-06,
"logits/chosen": 29.184545516967773,
"logits/rejected": 28.960050582885742,
"logps/chosen": -393.89373779296875,
"logps/rejected": -382.4853820800781,
"loss": 0.5007,
"rewards/accuracies": 0.5,
"rewards/chosen": 0.0003057096619158983,
"rewards/margins": 0.001760864513926208,
"rewards/rejected": -0.0014551542699337006,
"step": 9
},
{
"debug/policy_chosen_logits": 24.076282501220703,
"debug/policy_chosen_logps": -394.2970275878906,
"debug/policy_rejected_logits": 21.97032928466797,
"debug/policy_rejected_logps": -381.5108947753906,
"debug/reference_chosen_logps": -393.9175720214844,
"debug/reference_rejected_logps": -381.532470703125,
"epoch": 0.29411764705882354,
"grad_norm": 5.055283348007549,
"learning_rate": 1e-06,
"logits/chosen": 24.076282501220703,
"logits/rejected": 21.97032928466797,
"logps/chosen": -394.2970275878906,
"logps/rejected": -381.5108947753906,
"loss": 0.4978,
"rewards/accuracies": 0.5,
"rewards/chosen": -0.0037943648640066385,
"rewards/margins": -0.004009971860796213,
"rewards/rejected": 0.0002156068803742528,
"step": 10
},
{
"debug/policy_chosen_logits": 30.054981231689453,
"debug/policy_chosen_logps": -386.2349853515625,
"debug/policy_rejected_logits": 29.39519691467285,
"debug/policy_rejected_logps": -412.88177490234375,
"debug/reference_chosen_logps": -386.55413818359375,
"debug/reference_rejected_logps": -413.00701904296875,
"epoch": 0.3235294117647059,
"grad_norm": 5.392953865442379,
"learning_rate": 1e-06,
"logits/chosen": 30.054981231689453,
"logits/rejected": 29.39519691467285,
"logps/chosen": -386.2349853515625,
"logps/rejected": -412.88177490234375,
"loss": 0.4972,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.0031916426960378885,
"rewards/margins": 0.0019394302507862449,
"rewards/rejected": 0.0012522124452516437,
"step": 11
},
{
"debug/policy_chosen_logits": 27.566011428833008,
"debug/policy_chosen_logps": -374.5032653808594,
"debug/policy_rejected_logits": 29.68182373046875,
"debug/policy_rejected_logps": -401.15179443359375,
"debug/reference_chosen_logps": -374.087646484375,
"debug/reference_rejected_logps": -402.13214111328125,
"epoch": 0.35294117647058826,
"grad_norm": 6.451284347500255,
"learning_rate": 1e-06,
"logits/chosen": 27.566011428833008,
"logits/rejected": 29.68182373046875,
"logps/chosen": -374.5032653808594,
"logps/rejected": -401.15179443359375,
"loss": 0.4998,
"rewards/accuracies": 0.25,
"rewards/chosen": -0.0041561126708984375,
"rewards/margins": -0.013959312811493874,
"rewards/rejected": 0.009803200140595436,
"step": 12
},
{
"debug/policy_chosen_logits": 28.40235137939453,
"debug/policy_chosen_logps": -399.9848327636719,
"debug/policy_rejected_logits": 29.182313919067383,
"debug/policy_rejected_logps": -415.01519775390625,
"debug/reference_chosen_logps": -400.4940185546875,
"debug/reference_rejected_logps": -414.45550537109375,
"epoch": 0.38235294117647056,
"grad_norm": 4.8489563556671005,
"learning_rate": 1e-06,
"logits/chosen": 28.40235137939453,
"logits/rejected": 29.182313919067383,
"logps/chosen": -399.9848327636719,
"logps/rejected": -415.01519775390625,
"loss": 0.4979,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.005092009902000427,
"rewards/margins": 0.010688590817153454,
"rewards/rejected": -0.005596580449491739,
"step": 13
},
{
"debug/policy_chosen_logits": 21.881561279296875,
"debug/policy_chosen_logps": -402.83416748046875,
"debug/policy_rejected_logits": 23.709701538085938,
"debug/policy_rejected_logps": -417.4763488769531,
"debug/reference_chosen_logps": -402.1183166503906,
"debug/reference_rejected_logps": -417.23797607421875,
"epoch": 0.4117647058823529,
"grad_norm": 5.498578228190857,
"learning_rate": 1e-06,
"logits/chosen": 21.881561279296875,
"logits/rejected": 23.709701538085938,
"logps/chosen": -402.83416748046875,
"logps/rejected": -417.4763488769531,
"loss": 0.5019,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.007158737629652023,
"rewards/margins": -0.004775161389261484,
"rewards/rejected": -0.0023835753090679646,
"step": 14
},
{
"debug/policy_chosen_logits": 29.381258010864258,
"debug/policy_chosen_logps": -395.54803466796875,
"debug/policy_rejected_logits": 28.470420837402344,
"debug/policy_rejected_logps": -387.3695373535156,
"debug/reference_chosen_logps": -395.508056640625,
"debug/reference_rejected_logps": -386.67974853515625,
"epoch": 0.4411764705882353,
"grad_norm": 6.635697402918325,
"learning_rate": 1e-06,
"logits/chosen": 29.381258010864258,
"logits/rejected": 28.470420837402344,
"logps/chosen": -395.54803466796875,
"logps/rejected": -387.3695373535156,
"loss": 0.4963,
"rewards/accuracies": 0.5,
"rewards/chosen": -0.0003996272571384907,
"rewards/margins": 0.006498145870864391,
"rewards/rejected": -0.006897774059325457,
"step": 15
},
{
"debug/policy_chosen_logits": 27.256446838378906,
"debug/policy_chosen_logps": -403.69305419921875,
"debug/policy_rejected_logits": 26.9769344329834,
"debug/policy_rejected_logps": -387.86083984375,
"debug/reference_chosen_logps": -404.33697509765625,
"debug/reference_rejected_logps": -388.44683837890625,
"epoch": 0.47058823529411764,
"grad_norm": 5.539928310976471,
"learning_rate": 1e-06,
"logits/chosen": 27.256446838378906,
"logits/rejected": 26.9769344329834,
"logps/chosen": -403.69305419921875,
"logps/rejected": -387.86083984375,
"loss": 0.497,
"rewards/accuracies": 0.5,
"rewards/chosen": 0.006439666263759136,
"rewards/margins": 0.0005795289762318134,
"rewards/rejected": 0.00586013775318861,
"step": 16
},
{
"debug/policy_chosen_logits": 27.54875373840332,
"debug/policy_chosen_logps": -379.23291015625,
"debug/policy_rejected_logits": 28.37594985961914,
"debug/policy_rejected_logps": -391.45654296875,
"debug/reference_chosen_logps": -380.18060302734375,
"debug/reference_rejected_logps": -390.98968505859375,
"epoch": 0.5,
"grad_norm": 5.2957197578010415,
"learning_rate": 1e-06,
"logits/chosen": 27.54875373840332,
"logits/rejected": 28.37594985961914,
"logps/chosen": -379.23291015625,
"logps/rejected": -391.45654296875,
"loss": 0.4938,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.009477118961513042,
"rewards/margins": 0.01414569839835167,
"rewards/rejected": -0.0046685789711773396,
"step": 17
},
{
"debug/policy_chosen_logits": 25.160919189453125,
"debug/policy_chosen_logps": -389.9479675292969,
"debug/policy_rejected_logits": 25.0775146484375,
"debug/policy_rejected_logps": -372.14349365234375,
"debug/reference_chosen_logps": -389.97222900390625,
"debug/reference_rejected_logps": -371.96075439453125,
"epoch": 0.5294117647058824,
"grad_norm": 4.954808294891358,
"learning_rate": 1e-06,
"logits/chosen": 25.160919189453125,
"logits/rejected": 25.0775146484375,
"logps/chosen": -389.9479675292969,
"logps/rejected": -372.14349365234375,
"loss": 0.4953,
"rewards/accuracies": 0.375,
"rewards/chosen": 0.00024272920563817024,
"rewards/margins": 0.002069931011646986,
"rewards/rejected": -0.0018272020388394594,
"step": 18
},
{
"debug/policy_chosen_logits": 25.586524963378906,
"debug/policy_chosen_logps": -374.25006103515625,
"debug/policy_rejected_logits": 28.893770217895508,
"debug/policy_rejected_logps": -412.3420104980469,
"debug/reference_chosen_logps": -374.282958984375,
"debug/reference_rejected_logps": -411.41131591796875,
"epoch": 0.5588235294117647,
"grad_norm": 4.9850918433013724,
"learning_rate": 1e-06,
"logits/chosen": 25.586524963378906,
"logits/rejected": 28.893770217895508,
"logps/chosen": -374.25006103515625,
"logps/rejected": -412.3420104980469,
"loss": 0.5002,
"rewards/accuracies": 0.375,
"rewards/chosen": 0.0003293610643595457,
"rewards/margins": 0.009636610746383667,
"rewards/rejected": -0.009307250380516052,
"step": 19
},
{
"debug/policy_chosen_logits": 24.718896865844727,
"debug/policy_chosen_logps": -363.05120849609375,
"debug/policy_rejected_logits": 24.986190795898438,
"debug/policy_rejected_logps": -388.0316162109375,
"debug/reference_chosen_logps": -362.204833984375,
"debug/reference_rejected_logps": -386.9569091796875,
"epoch": 0.5882352941176471,
"grad_norm": 5.05147280128332,
"learning_rate": 1e-06,
"logits/chosen": 24.718896865844727,
"logits/rejected": 24.986190795898438,
"logps/chosen": -363.05120849609375,
"logps/rejected": -388.0316162109375,
"loss": 0.4997,
"rewards/accuracies": 0.5,
"rewards/chosen": -0.00846382137387991,
"rewards/margins": 0.0022830581292510033,
"rewards/rejected": -0.010746879503130913,
"step": 20
},
{
"debug/policy_chosen_logits": 28.18993377685547,
"debug/policy_chosen_logps": -397.7740478515625,
"debug/policy_rejected_logits": 23.599760055541992,
"debug/policy_rejected_logps": -400.73492431640625,
"debug/reference_chosen_logps": -399.1761169433594,
"debug/reference_rejected_logps": -399.78948974609375,
"epoch": 0.6176470588235294,
"grad_norm": 6.701576850276441,
"learning_rate": 1e-06,
"logits/chosen": 28.18993377685547,
"logits/rejected": 23.599760055541992,
"logps/chosen": -397.7740478515625,
"logps/rejected": -400.73492431640625,
"loss": 0.4981,
"rewards/accuracies": 0.5,
"rewards/chosen": 0.014020653441548347,
"rewards/margins": 0.023475077003240585,
"rewards/rejected": -0.009454423561692238,
"step": 21
},
{
"debug/policy_chosen_logits": 29.8790283203125,
"debug/policy_chosen_logps": -413.7144775390625,
"debug/policy_rejected_logits": 28.82781982421875,
"debug/policy_rejected_logps": -409.18292236328125,
"debug/reference_chosen_logps": -414.69781494140625,
"debug/reference_rejected_logps": -407.171142578125,
"epoch": 0.6470588235294118,
"grad_norm": 6.323947957916722,
"learning_rate": 1e-06,
"logits/chosen": 29.8790283203125,
"logits/rejected": 28.82781982421875,
"logps/chosen": -413.7144775390625,
"logps/rejected": -409.18292236328125,
"loss": 0.4999,
"rewards/accuracies": 0.5,
"rewards/chosen": 0.009833297692239285,
"rewards/margins": 0.029951322823762894,
"rewards/rejected": -0.020118026062846184,
"step": 22
},
{
"debug/policy_chosen_logits": 27.423133850097656,
"debug/policy_chosen_logps": -377.9825439453125,
"debug/policy_rejected_logits": 29.817766189575195,
"debug/policy_rejected_logps": -400.9486389160156,
"debug/reference_chosen_logps": -378.6519775390625,
"debug/reference_rejected_logps": -400.3038024902344,
"epoch": 0.6764705882352942,
"grad_norm": 7.650698869625993,
"learning_rate": 1e-06,
"logits/chosen": 27.423133850097656,
"logits/rejected": 29.817766189575195,
"logps/chosen": -377.9825439453125,
"logps/rejected": -400.9486389160156,
"loss": 0.4944,
"rewards/accuracies": 0.875,
"rewards/chosen": 0.006694145500659943,
"rewards/margins": 0.013142585754394531,
"rewards/rejected": -0.006448440253734589,
"step": 23
},
{
"debug/policy_chosen_logits": 26.779478073120117,
"debug/policy_chosen_logps": -392.8223571777344,
"debug/policy_rejected_logits": 28.28175163269043,
"debug/policy_rejected_logps": -394.74676513671875,
"debug/reference_chosen_logps": -393.3936462402344,
"debug/reference_rejected_logps": -394.11358642578125,
"epoch": 0.7058823529411765,
"grad_norm": 5.251742483181524,
"learning_rate": 1e-06,
"logits/chosen": 26.779478073120117,
"logits/rejected": 28.28175163269043,
"logps/chosen": -392.8223571777344,
"logps/rejected": -394.74676513671875,
"loss": 0.5016,
"rewards/accuracies": 0.5,
"rewards/chosen": 0.005712966434657574,
"rewards/margins": 0.012044791132211685,
"rewards/rejected": -0.0063318246975541115,
"step": 24
},
{
"debug/policy_chosen_logits": 28.304908752441406,
"debug/policy_chosen_logps": -395.7744445800781,
"debug/policy_rejected_logits": 27.35923957824707,
"debug/policy_rejected_logps": -403.1441955566406,
"debug/reference_chosen_logps": -395.40789794921875,
"debug/reference_rejected_logps": -403.5139465332031,
"epoch": 0.7352941176470589,
"grad_norm": 4.968117146678153,
"learning_rate": 1e-06,
"logits/chosen": 28.304908752441406,
"logits/rejected": 27.35923957824707,
"logps/chosen": -395.7744445800781,
"logps/rejected": -403.1441955566406,
"loss": 0.4985,
"rewards/accuracies": 0.125,
"rewards/chosen": -0.003665314055979252,
"rewards/margins": -0.007362784817814827,
"rewards/rejected": 0.0036974712274968624,
"step": 25
},
{
"debug/policy_chosen_logits": 28.73921012878418,
"debug/policy_chosen_logps": -396.11834716796875,
"debug/policy_rejected_logits": 28.473979949951172,
"debug/policy_rejected_logps": -419.5787353515625,
"debug/reference_chosen_logps": -396.2462158203125,
"debug/reference_rejected_logps": -418.2436218261719,
"epoch": 0.7647058823529411,
"grad_norm": 5.07020088847226,
"learning_rate": 1e-06,
"logits/chosen": 28.73921012878418,
"logits/rejected": 28.473979949951172,
"logps/chosen": -396.11834716796875,
"logps/rejected": -419.5787353515625,
"loss": 0.4994,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.0012784956488758326,
"rewards/margins": 0.014629820361733437,
"rewards/rejected": -0.013351325877010822,
"step": 26
},
{
"debug/policy_chosen_logits": 30.456932067871094,
"debug/policy_chosen_logps": -410.0908203125,
"debug/policy_rejected_logits": 28.822614669799805,
"debug/policy_rejected_logps": -418.9249572753906,
"debug/reference_chosen_logps": -410.21063232421875,
"debug/reference_rejected_logps": -417.00714111328125,
"epoch": 0.7941176470588235,
"grad_norm": 6.10176209535134,
"learning_rate": 1e-06,
"logits/chosen": 30.456932067871094,
"logits/rejected": 28.822614669799805,
"logps/chosen": -410.0908203125,
"logps/rejected": -418.9249572753906,
"loss": 0.4853,
"rewards/accuracies": 0.625,
"rewards/chosen": 0.0011978147085756063,
"rewards/margins": 0.020375745370984077,
"rewards/rejected": -0.019177932292222977,
"step": 27
},
{
"debug/policy_chosen_logits": 28.178850173950195,
"debug/policy_chosen_logps": -408.07464599609375,
"debug/policy_rejected_logits": 28.366477966308594,
"debug/policy_rejected_logps": -414.43853759765625,
"debug/reference_chosen_logps": -407.51043701171875,
"debug/reference_rejected_logps": -410.44268798828125,
"epoch": 0.8235294117647058,
"grad_norm": 4.919999726189566,
"learning_rate": 1e-06,
"logits/chosen": 28.178850173950195,
"logits/rejected": 28.366477966308594,
"logps/chosen": -408.07464599609375,
"logps/rejected": -414.43853759765625,
"loss": 0.499,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.005642051808536053,
"rewards/margins": 0.03431610018014908,
"rewards/rejected": -0.039958152920007706,
"step": 28
},
{
"debug/policy_chosen_logits": 29.156843185424805,
"debug/policy_chosen_logps": -409.95208740234375,
"debug/policy_rejected_logits": 27.127243041992188,
"debug/policy_rejected_logps": -393.4117431640625,
"debug/reference_chosen_logps": -409.63983154296875,
"debug/reference_rejected_logps": -392.2528076171875,
"epoch": 0.8529411764705882,
"grad_norm": 5.038344787361152,
"learning_rate": 1e-06,
"logits/chosen": 29.156843185424805,
"logits/rejected": 27.127243041992188,
"logps/chosen": -409.95208740234375,
"logps/rejected": -393.4117431640625,
"loss": 0.4935,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.003122864291071892,
"rewards/margins": 0.008466909639537334,
"rewards/rejected": -0.0115897748619318,
"step": 29
},
{
"debug/policy_chosen_logits": 26.275718688964844,
"debug/policy_chosen_logps": -406.03277587890625,
"debug/policy_rejected_logits": 26.794349670410156,
"debug/policy_rejected_logps": -410.08282470703125,
"debug/reference_chosen_logps": -405.848876953125,
"debug/reference_rejected_logps": -410.5152587890625,
"epoch": 0.8823529411764706,
"grad_norm": 6.078237487272278,
"learning_rate": 1e-06,
"logits/chosen": 26.275718688964844,
"logits/rejected": 26.794349670410156,
"logps/chosen": -406.03277587890625,
"logps/rejected": -410.08282470703125,
"loss": 0.4941,
"rewards/accuracies": 0.5,
"rewards/chosen": -0.0018388747703284025,
"rewards/margins": -0.006163253448903561,
"rewards/rejected": 0.0043243784457445145,
"step": 30
},
{
"debug/policy_chosen_logits": 27.000301361083984,
"debug/policy_chosen_logps": -391.47613525390625,
"debug/policy_rejected_logits": 29.51080322265625,
"debug/policy_rejected_logps": -386.7709655761719,
"debug/reference_chosen_logps": -390.591552734375,
"debug/reference_rejected_logps": -387.83935546875,
"epoch": 0.9117647058823529,
"grad_norm": 5.608762984875432,
"learning_rate": 1e-06,
"logits/chosen": 27.000301361083984,
"logits/rejected": 29.51080322265625,
"logps/chosen": -391.47613525390625,
"logps/rejected": -386.7709655761719,
"loss": 0.4983,
"rewards/accuracies": 0.125,
"rewards/chosen": -0.008846206590533257,
"rewards/margins": -0.019530104473233223,
"rewards/rejected": 0.010683897882699966,
"step": 31
},
{
"debug/policy_chosen_logits": 29.469999313354492,
"debug/policy_chosen_logps": -390.48291015625,
"debug/policy_rejected_logits": 28.674427032470703,
"debug/policy_rejected_logps": -414.2558288574219,
"debug/reference_chosen_logps": -392.53704833984375,
"debug/reference_rejected_logps": -416.4234619140625,
"epoch": 0.9411764705882353,
"grad_norm": 5.952413372138852,
"learning_rate": 1e-06,
"logits/chosen": 29.469999313354492,
"logits/rejected": 28.674427032470703,
"logps/chosen": -390.48291015625,
"logps/rejected": -414.2558288574219,
"loss": 0.4928,
"rewards/accuracies": 0.625,
"rewards/chosen": 0.02054138109087944,
"rewards/margins": -0.0011348333209753036,
"rewards/rejected": 0.021676216274499893,
"step": 32
},
{
"debug/policy_chosen_logits": 30.11175537109375,
"debug/policy_chosen_logps": -393.00079345703125,
"debug/policy_rejected_logits": 31.391122817993164,
"debug/policy_rejected_logps": -419.4772644042969,
"debug/reference_chosen_logps": -393.62738037109375,
"debug/reference_rejected_logps": -416.98974609375,
"epoch": 0.9705882352941176,
"grad_norm": 5.204183724206507,
"learning_rate": 1e-06,
"logits/chosen": 30.11175537109375,
"logits/rejected": 31.391122817993164,
"logps/chosen": -393.00079345703125,
"logps/rejected": -419.4772644042969,
"loss": 0.4877,
"rewards/accuracies": 0.875,
"rewards/chosen": 0.006265563890337944,
"rewards/margins": 0.031140554696321487,
"rewards/rejected": -0.024874988943338394,
"step": 33
},
{
"debug/policy_chosen_logits": 28.005823135375977,
"debug/policy_chosen_logps": -417.3971252441406,
"debug/policy_rejected_logits": 24.94090461730957,
"debug/policy_rejected_logps": -404.98779296875,
"debug/reference_chosen_logps": -416.21746826171875,
"debug/reference_rejected_logps": -404.94940185546875,
"epoch": 1.0,
"grad_norm": 5.733599790823146,
"learning_rate": 1e-06,
"logits/chosen": 28.005823135375977,
"logits/rejected": 24.94090461730957,
"logps/chosen": -417.3971252441406,
"logps/rejected": -404.98779296875,
"loss": 0.4673,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.011796417646110058,
"rewards/margins": -0.01141277328133583,
"rewards/rejected": -0.00038364529609680176,
"step": 34
},
{
"epoch": 1.0,
"step": 34,
"total_flos": 0.0,
"train_loss": 0.4969010160249822,
"train_runtime": 407.6559,
"train_samples_per_second": 5.247,
"train_steps_per_second": 0.083
}
],
"logging_steps": 1,
"max_steps": 34,
"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
}