| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 2.9762532981530345, | |
| "eval_steps": 100, | |
| "global_step": 282, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.052770448548812667, | |
| "grad_norm": 64.76068032491155, | |
| "learning_rate": 5e-07, | |
| "logits/chosen": -2.746584892272949, | |
| "logits/rejected": -2.7205777168273926, | |
| "logps/chosen": -309.4463806152344, | |
| "logps/rejected": -245.45993041992188, | |
| "loss": 0.6897, | |
| "rewards/accuracies": 0.39375001192092896, | |
| "rewards/chosen": 0.031672243028879166, | |
| "rewards/margins": 0.014508840627968311, | |
| "rewards/rejected": 0.01716339960694313, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.10554089709762533, | |
| "grad_norm": 49.64963638540198, | |
| "learning_rate": 1e-06, | |
| "logits/chosen": -2.679844379425049, | |
| "logits/rejected": -2.671976327896118, | |
| "logps/chosen": -282.12872314453125, | |
| "logps/rejected": -240.8015594482422, | |
| "loss": 0.6204, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": 0.6456395983695984, | |
| "rewards/margins": 0.19399957358837128, | |
| "rewards/rejected": 0.4516400396823883, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.158311345646438, | |
| "grad_norm": 45.577976942543465, | |
| "learning_rate": 9.991664699681798e-07, | |
| "logits/chosen": -2.459498882293701, | |
| "logits/rejected": -2.4666247367858887, | |
| "logps/chosen": -281.7549133300781, | |
| "logps/rejected": -230.15090942382812, | |
| "loss": 0.5611, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 2.039682626724243, | |
| "rewards/margins": 0.6908785104751587, | |
| "rewards/rejected": 1.348804235458374, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.21108179419525067, | |
| "grad_norm": 42.33752347466586, | |
| "learning_rate": 9.966686589619748e-07, | |
| "logits/chosen": -2.287057399749756, | |
| "logits/rejected": -2.2771310806274414, | |
| "logps/chosen": -288.93560791015625, | |
| "logps/rejected": -228.0941925048828, | |
| "loss": 0.5667, | |
| "rewards/accuracies": 0.8187500238418579, | |
| "rewards/chosen": 2.559917449951172, | |
| "rewards/margins": 1.54579758644104, | |
| "rewards/rejected": 1.0141198635101318, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.2638522427440633, | |
| "grad_norm": 39.85680411240784, | |
| "learning_rate": 9.925148949833354e-07, | |
| "logits/chosen": -2.126183032989502, | |
| "logits/rejected": -2.1005260944366455, | |
| "logps/chosen": -274.09747314453125, | |
| "logps/rejected": -219.9466094970703, | |
| "loss": 0.565, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 2.0806262493133545, | |
| "rewards/margins": 1.2727241516113281, | |
| "rewards/rejected": 0.8079022169113159, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.316622691292876, | |
| "grad_norm": 43.37392233672497, | |
| "learning_rate": 9.867190271803463e-07, | |
| "logits/chosen": -2.031036853790283, | |
| "logits/rejected": -2.001377820968628, | |
| "logps/chosen": -270.2996520996094, | |
| "logps/rejected": -211.66836547851562, | |
| "loss": 0.5599, | |
| "rewards/accuracies": 0.793749988079071, | |
| "rewards/chosen": 2.6486361026763916, | |
| "rewards/margins": 1.5665667057037354, | |
| "rewards/rejected": 1.0820693969726562, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.36939313984168864, | |
| "grad_norm": 41.48710119002139, | |
| "learning_rate": 9.793003796725049e-07, | |
| "logits/chosen": -2.0884552001953125, | |
| "logits/rejected": -2.0888595581054688, | |
| "logps/chosen": -267.67620849609375, | |
| "logps/rejected": -226.6994171142578, | |
| "loss": 0.5654, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": 2.3934950828552246, | |
| "rewards/margins": 1.489709734916687, | |
| "rewards/rejected": 0.9037855863571167, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.42216358839050133, | |
| "grad_norm": 41.87295119287348, | |
| "learning_rate": 9.702836871217837e-07, | |
| "logits/chosen": -2.189497470855713, | |
| "logits/rejected": -2.181183338165283, | |
| "logps/chosen": -271.53692626953125, | |
| "logps/rejected": -235.14474487304688, | |
| "loss": 0.4739, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": 2.4543251991271973, | |
| "rewards/margins": 1.8336181640625, | |
| "rewards/rejected": 0.6207070350646973, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.47493403693931396, | |
| "grad_norm": 38.88914869924963, | |
| "learning_rate": 9.596990122642983e-07, | |
| "logits/chosen": -2.101018190383911, | |
| "logits/rejected": -2.0873799324035645, | |
| "logps/chosen": -267.4029541015625, | |
| "logps/rejected": -231.83963012695312, | |
| "loss": 0.5017, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": 1.969552755355835, | |
| "rewards/margins": 1.2634941339492798, | |
| "rewards/rejected": 0.7060589790344238, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 0.5277044854881267, | |
| "grad_norm": 32.96198476610442, | |
| "learning_rate": 9.475816456775312e-07, | |
| "logits/chosen": -1.9850475788116455, | |
| "logits/rejected": -1.9540290832519531, | |
| "logps/chosen": -278.4673156738281, | |
| "logps/rejected": -219.38278198242188, | |
| "loss": 0.513, | |
| "rewards/accuracies": 0.793749988079071, | |
| "rewards/chosen": 2.193819284439087, | |
| "rewards/margins": 1.707207441329956, | |
| "rewards/rejected": 0.4866116940975189, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.5804749340369393, | |
| "grad_norm": 37.97806674219141, | |
| "learning_rate": 9.339719881173092e-07, | |
| "logits/chosen": -2.1176700592041016, | |
| "logits/rejected": -2.092395305633545, | |
| "logps/chosen": -251.4981689453125, | |
| "logps/rejected": -235.84164428710938, | |
| "loss": 0.5447, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": 2.199078321456909, | |
| "rewards/margins": 1.7228357791900635, | |
| "rewards/rejected": 0.47624263167381287, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 0.633245382585752, | |
| "grad_norm": 38.132896833898045, | |
| "learning_rate": 9.18915415816829e-07, | |
| "logits/chosen": -2.226740837097168, | |
| "logits/rejected": -2.220912218093872, | |
| "logps/chosen": -274.7961730957031, | |
| "logps/rejected": -219.2300262451172, | |
| "loss": 0.5351, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": 2.3536345958709717, | |
| "rewards/margins": 1.691295862197876, | |
| "rewards/rejected": 0.6623384952545166, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.6860158311345647, | |
| "grad_norm": 39.82817078264063, | |
| "learning_rate": 9.024621291968459e-07, | |
| "logits/chosen": -2.2491986751556396, | |
| "logits/rejected": -2.2434003353118896, | |
| "logps/chosen": -275.55938720703125, | |
| "logps/rejected": -221.8572235107422, | |
| "loss": 0.5159, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 1.9517409801483154, | |
| "rewards/margins": 1.8589683771133423, | |
| "rewards/rejected": 0.09277255833148956, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 0.7387862796833773, | |
| "grad_norm": 37.683212183008145, | |
| "learning_rate": 8.846669854914395e-07, | |
| "logits/chosen": -2.192699670791626, | |
| "logits/rejected": -2.196516752243042, | |
| "logps/chosen": -264.14141845703125, | |
| "logps/rejected": -232.97012329101562, | |
| "loss": 0.5269, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": 1.64827561378479, | |
| "rewards/margins": 1.7010596990585327, | |
| "rewards/rejected": -0.052784264087677, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.7915567282321899, | |
| "grad_norm": 33.80834744359574, | |
| "learning_rate": 8.655893158474054e-07, | |
| "logits/chosen": -2.113145351409912, | |
| "logits/rejected": -2.065735340118408, | |
| "logps/chosen": -269.6071472167969, | |
| "logps/rejected": -251.72726440429688, | |
| "loss": 0.5091, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": 1.1491832733154297, | |
| "rewards/margins": 1.6937177181243896, | |
| "rewards/rejected": -0.5445345640182495, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.8443271767810027, | |
| "grad_norm": 39.647954556629706, | |
| "learning_rate": 8.452927275070857e-07, | |
| "logits/chosen": -2.097092866897583, | |
| "logits/rejected": -2.0857818126678467, | |
| "logps/chosen": -292.34619140625, | |
| "logps/rejected": -246.6460418701172, | |
| "loss": 0.5064, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": 1.3558640480041504, | |
| "rewards/margins": 1.6969764232635498, | |
| "rewards/rejected": -0.34111255407333374, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.8970976253298153, | |
| "grad_norm": 35.686723090952604, | |
| "learning_rate": 8.238448917341809e-07, | |
| "logits/chosen": -2.1606364250183105, | |
| "logits/rejected": -2.151815891265869, | |
| "logps/chosen": -278.24261474609375, | |
| "logps/rejected": -225.29812622070312, | |
| "loss": 0.4882, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": 1.774441123008728, | |
| "rewards/margins": 2.115816831588745, | |
| "rewards/rejected": -0.3413757383823395, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 0.9498680738786279, | |
| "grad_norm": 30.74062773369213, | |
| "learning_rate": 8.013173181896282e-07, | |
| "logits/chosen": -2.119166135787964, | |
| "logits/rejected": -2.0985186100006104, | |
| "logps/chosen": -291.9129333496094, | |
| "logps/rejected": -275.41473388671875, | |
| "loss": 0.4649, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": 1.9476991891860962, | |
| "rewards/margins": 2.8862874507904053, | |
| "rewards/rejected": -0.9385882616043091, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 1.0026385224274406, | |
| "grad_norm": 37.02461141617717, | |
| "learning_rate": 7.777851165098011e-07, | |
| "logits/chosen": -2.0612759590148926, | |
| "logits/rejected": -2.0387701988220215, | |
| "logps/chosen": -282.53094482421875, | |
| "logps/rejected": -248.63037109375, | |
| "loss": 0.4735, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": 1.4429569244384766, | |
| "rewards/margins": 2.1116912364959717, | |
| "rewards/rejected": -0.6687344312667847, | |
| "step": 95 | |
| }, | |
| { | |
| "epoch": 1.0554089709762533, | |
| "grad_norm": 16.66169739789075, | |
| "learning_rate": 7.533267458819597e-07, | |
| "logits/chosen": -2.0679678916931152, | |
| "logits/rejected": -2.042463779449463, | |
| "logps/chosen": -279.1990051269531, | |
| "logps/rejected": -248.1520233154297, | |
| "loss": 0.1988, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": 2.0477817058563232, | |
| "rewards/margins": 3.6042072772979736, | |
| "rewards/rejected": -1.5564252138137817, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.0554089709762533, | |
| "eval_logits/chosen": -2.049605131149292, | |
| "eval_logits/rejected": -2.027388334274292, | |
| "eval_logps/chosen": -292.00994873046875, | |
| "eval_logps/rejected": -231.3839874267578, | |
| "eval_loss": 0.45807966589927673, | |
| "eval_rewards/accuracies": 0.805232584476471, | |
| "eval_rewards/chosen": 1.2440974712371826, | |
| "eval_rewards/margins": 1.994946837425232, | |
| "eval_rewards/rejected": -0.7508493661880493, | |
| "eval_runtime": 181.9718, | |
| "eval_samples_per_second": 14.777, | |
| "eval_steps_per_second": 0.236, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.108179419525066, | |
| "grad_norm": 17.836098027593945, | |
| "learning_rate": 7.280237534518947e-07, | |
| "logits/chosen": -2.0414350032806396, | |
| "logits/rejected": -2.029930591583252, | |
| "logps/chosen": -270.83984375, | |
| "logps/rejected": -252.8948974609375, | |
| "loss": 0.1806, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": 2.0295984745025635, | |
| "rewards/margins": 3.588951826095581, | |
| "rewards/rejected": -1.5593531131744385, | |
| "step": 105 | |
| }, | |
| { | |
| "epoch": 1.1609498680738786, | |
| "grad_norm": 17.80012843476481, | |
| "learning_rate": 7.019605024359474e-07, | |
| "logits/chosen": -2.067286968231201, | |
| "logits/rejected": -2.0599799156188965, | |
| "logps/chosen": -277.5862731933594, | |
| "logps/rejected": -236.6626739501953, | |
| "loss": 0.1727, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": 2.3234915733337402, | |
| "rewards/margins": 3.590890884399414, | |
| "rewards/rejected": -1.2673994302749634, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 1.2137203166226913, | |
| "grad_norm": 22.157519816362168, | |
| "learning_rate": 6.75223890843913e-07, | |
| "logits/chosen": -2.0847175121307373, | |
| "logits/rejected": -2.0452258586883545, | |
| "logps/chosen": -256.7383117675781, | |
| "logps/rejected": -249.1232452392578, | |
| "loss": 0.1747, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": 2.8541388511657715, | |
| "rewards/margins": 4.252060413360596, | |
| "rewards/rejected": -1.3979219198226929, | |
| "step": 115 | |
| }, | |
| { | |
| "epoch": 1.266490765171504, | |
| "grad_norm": 18.868288476503547, | |
| "learning_rate": 6.479030617506353e-07, | |
| "logits/chosen": -2.1164755821228027, | |
| "logits/rejected": -2.0933194160461426, | |
| "logps/chosen": -274.2527160644531, | |
| "logps/rejected": -229.81961059570312, | |
| "loss": 0.168, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 3.0210556983947754, | |
| "rewards/margins": 4.186809062957764, | |
| "rewards/rejected": -1.1657536029815674, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 1.3192612137203166, | |
| "grad_norm": 14.436192471094481, | |
| "learning_rate": 6.200891060822883e-07, | |
| "logits/chosen": -2.2040228843688965, | |
| "logits/rejected": -2.13104248046875, | |
| "logps/chosen": -269.16143798828125, | |
| "logps/rejected": -254.4296417236328, | |
| "loss": 0.18, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": 3.200051784515381, | |
| "rewards/margins": 4.450863361358643, | |
| "rewards/rejected": -1.250811219215393, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 1.3720316622691293, | |
| "grad_norm": 23.727100289583046, | |
| "learning_rate": 5.918747589082852e-07, | |
| "logits/chosen": -2.265648365020752, | |
| "logits/rejected": -2.187023162841797, | |
| "logps/chosen": -266.95074462890625, | |
| "logps/rejected": -228.5242462158203, | |
| "loss": 0.1983, | |
| "rewards/accuracies": 0.9437500238418579, | |
| "rewards/chosen": 2.6789937019348145, | |
| "rewards/margins": 4.175280570983887, | |
| "rewards/rejected": -1.4962866306304932, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 1.424802110817942, | |
| "grad_norm": 23.502630941365233, | |
| "learning_rate": 5.633540902514169e-07, | |
| "logits/chosen": -2.209868907928467, | |
| "logits/rejected": -2.1866037845611572, | |
| "logps/chosen": -262.3369140625, | |
| "logps/rejected": -241.46603393554688, | |
| "loss": 0.2006, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 2.625196695327759, | |
| "rewards/margins": 4.23705530166626, | |
| "rewards/rejected": -1.611858606338501, | |
| "step": 135 | |
| }, | |
| { | |
| "epoch": 1.4775725593667546, | |
| "grad_norm": 23.084811806663915, | |
| "learning_rate": 5.346221914470958e-07, | |
| "logits/chosen": -2.1318600177764893, | |
| "logits/rejected": -2.097346782684326, | |
| "logps/chosen": -248.5489501953125, | |
| "logps/rejected": -255.60360717773438, | |
| "loss": 0.2217, | |
| "rewards/accuracies": 0.918749988079071, | |
| "rewards/chosen": 2.0362741947174072, | |
| "rewards/margins": 3.8567168712615967, | |
| "rewards/rejected": -1.8204425573349, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 1.5303430079155673, | |
| "grad_norm": 23.157038660609793, | |
| "learning_rate": 5.057748580974204e-07, | |
| "logits/chosen": -2.0590291023254395, | |
| "logits/rejected": -2.0120911598205566, | |
| "logps/chosen": -270.0731506347656, | |
| "logps/rejected": -235.39254760742188, | |
| "loss": 0.2198, | |
| "rewards/accuracies": 0.9125000238418579, | |
| "rewards/chosen": 2.321479320526123, | |
| "rewards/margins": 4.0736308097839355, | |
| "rewards/rejected": -1.7521512508392334, | |
| "step": 145 | |
| }, | |
| { | |
| "epoch": 1.58311345646438, | |
| "grad_norm": 20.586644732633363, | |
| "learning_rate": 4.769082706771303e-07, | |
| "logits/chosen": -2.0520288944244385, | |
| "logits/rejected": -2.0188193321228027, | |
| "logps/chosen": -268.6761474609375, | |
| "logps/rejected": -248.79989624023438, | |
| "loss": 0.2577, | |
| "rewards/accuracies": 0.9125000238418579, | |
| "rewards/chosen": 2.286835193634033, | |
| "rewards/margins": 4.1830949783325195, | |
| "rewards/rejected": -1.8962600231170654, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 1.6358839050131926, | |
| "grad_norm": 17.91805124210978, | |
| "learning_rate": 4.481186738563491e-07, | |
| "logits/chosen": -2.0534660816192627, | |
| "logits/rejected": -2.016132116317749, | |
| "logps/chosen": -250.5113067626953, | |
| "logps/rejected": -258.28509521484375, | |
| "loss": 0.212, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 2.313445806503296, | |
| "rewards/margins": 3.963188886642456, | |
| "rewards/rejected": -1.6497433185577393, | |
| "step": 155 | |
| }, | |
| { | |
| "epoch": 1.6886543535620053, | |
| "grad_norm": 18.273291701902995, | |
| "learning_rate": 4.1950205560929346e-07, | |
| "logits/chosen": -2.0628087520599365, | |
| "logits/rejected": -2.022836923599243, | |
| "logps/chosen": -289.82684326171875, | |
| "logps/rejected": -254.5777587890625, | |
| "loss": 0.2345, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": 2.853477716445923, | |
| "rewards/margins": 4.158769607543945, | |
| "rewards/rejected": -1.305292010307312, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 1.741424802110818, | |
| "grad_norm": 23.98541469015997, | |
| "learning_rate": 3.9115382717883583e-07, | |
| "logits/chosen": -2.007776975631714, | |
| "logits/rejected": -1.9897905588150024, | |
| "logps/chosen": -260.97076416015625, | |
| "logps/rejected": -222.6240997314453, | |
| "loss": 0.2437, | |
| "rewards/accuracies": 0.9312499761581421, | |
| "rewards/chosen": 2.0936646461486816, | |
| "rewards/margins": 3.4598960876464844, | |
| "rewards/rejected": -1.366231083869934, | |
| "step": 165 | |
| }, | |
| { | |
| "epoch": 1.7941952506596306, | |
| "grad_norm": 18.592544096757504, | |
| "learning_rate": 3.6316850496395855e-07, | |
| "logits/chosen": -1.9580230712890625, | |
| "logits/rejected": -1.9140596389770508, | |
| "logps/chosen": -279.15875244140625, | |
| "logps/rejected": -241.4788818359375, | |
| "loss": 0.2303, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": 2.5783209800720215, | |
| "rewards/margins": 3.7805628776550293, | |
| "rewards/rejected": -1.2022416591644287, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 1.8469656992084431, | |
| "grad_norm": 17.349768780533378, | |
| "learning_rate": 3.3563939539072707e-07, | |
| "logits/chosen": -1.9430656433105469, | |
| "logits/rejected": -1.8976306915283203, | |
| "logps/chosen": -290.4911804199219, | |
| "logps/rejected": -256.70452880859375, | |
| "loss": 0.2243, | |
| "rewards/accuracies": 0.90625, | |
| "rewards/chosen": 2.43168306350708, | |
| "rewards/margins": 4.526327133178711, | |
| "rewards/rejected": -2.094644069671631, | |
| "step": 175 | |
| }, | |
| { | |
| "epoch": 1.899736147757256, | |
| "grad_norm": 18.360513371639943, | |
| "learning_rate": 3.086582838174551e-07, | |
| "logits/chosen": -1.8910833597183228, | |
| "logits/rejected": -1.8794158697128296, | |
| "logps/chosen": -276.2688293457031, | |
| "logps/rejected": -249.4001007080078, | |
| "loss": 0.2411, | |
| "rewards/accuracies": 0.9125000238418579, | |
| "rewards/chosen": 2.026005983352661, | |
| "rewards/margins": 3.8342995643615723, | |
| "rewards/rejected": -1.8082935810089111, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 1.9525065963060686, | |
| "grad_norm": 21.66091551271331, | |
| "learning_rate": 2.823151285112959e-07, | |
| "logits/chosen": -1.9270591735839844, | |
| "logits/rejected": -1.884987473487854, | |
| "logps/chosen": -275.21209716796875, | |
| "logps/rejected": -244.12063598632812, | |
| "loss": 0.1771, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": 1.7359602451324463, | |
| "rewards/margins": 3.962980270385742, | |
| "rewards/rejected": -2.227019786834717, | |
| "step": 185 | |
| }, | |
| { | |
| "epoch": 2.005277044854881, | |
| "grad_norm": 13.35667357704236, | |
| "learning_rate": 2.566977607165719e-07, | |
| "logits/chosen": -1.991843819618225, | |
| "logits/rejected": -1.9527130126953125, | |
| "logps/chosen": -266.570068359375, | |
| "logps/rejected": -249.18722534179688, | |
| "loss": 0.2033, | |
| "rewards/accuracies": 0.956250011920929, | |
| "rewards/chosen": 1.811069130897522, | |
| "rewards/margins": 4.211352348327637, | |
| "rewards/rejected": -2.400282621383667, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 2.058047493403694, | |
| "grad_norm": 13.486567649894633, | |
| "learning_rate": 2.3189159181485511e-07, | |
| "logits/chosen": -2.0401811599731445, | |
| "logits/rejected": -1.998416543006897, | |
| "logps/chosen": -280.8411560058594, | |
| "logps/rejected": -275.76739501953125, | |
| "loss": 0.0984, | |
| "rewards/accuracies": 0.981249988079071, | |
| "rewards/chosen": 2.4434828758239746, | |
| "rewards/margins": 5.145030498504639, | |
| "rewards/rejected": -2.7015483379364014, | |
| "step": 195 | |
| }, | |
| { | |
| "epoch": 2.1108179419525066, | |
| "grad_norm": 13.112761330007803, | |
| "learning_rate": 2.079793285531618e-07, | |
| "logits/chosen": -2.068660259246826, | |
| "logits/rejected": -2.026952028274536, | |
| "logps/chosen": -294.8218688964844, | |
| "logps/rejected": -248.809326171875, | |
| "loss": 0.0898, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 2.1426644325256348, | |
| "rewards/margins": 4.40965461730957, | |
| "rewards/rejected": -2.2669899463653564, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 2.1108179419525066, | |
| "eval_logits/chosen": -2.0563266277313232, | |
| "eval_logits/rejected": -2.038892984390259, | |
| "eval_logps/chosen": -292.2633972167969, | |
| "eval_logps/rejected": -238.73812866210938, | |
| "eval_loss": 0.45805448293685913, | |
| "eval_rewards/accuracies": 0.8284883499145508, | |
| "eval_rewards/chosen": 1.218752145767212, | |
| "eval_rewards/margins": 2.705014228820801, | |
| "eval_rewards/rejected": -1.486262321472168, | |
| "eval_runtime": 180.6142, | |
| "eval_samples_per_second": 14.888, | |
| "eval_steps_per_second": 0.238, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 2.163588390501319, | |
| "grad_norm": 17.553767678675094, | |
| "learning_rate": 1.8504069728972122e-07, | |
| "logits/chosen": -2.0627760887145996, | |
| "logits/rejected": -2.0340285301208496, | |
| "logps/chosen": -261.85357666015625, | |
| "logps/rejected": -273.9400939941406, | |
| "loss": 0.1351, | |
| "rewards/accuracies": 0.9437500238418579, | |
| "rewards/chosen": 1.646632432937622, | |
| "rewards/margins": 4.253820896148682, | |
| "rewards/rejected": -2.6071884632110596, | |
| "step": 205 | |
| }, | |
| { | |
| "epoch": 2.216358839050132, | |
| "grad_norm": 17.23688737993748, | |
| "learning_rate": 1.631521781767214e-07, | |
| "logits/chosen": -2.108936309814453, | |
| "logits/rejected": -2.0527260303497314, | |
| "logps/chosen": -282.0605163574219, | |
| "logps/rejected": -248.5954132080078, | |
| "loss": 0.1067, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 2.1594605445861816, | |
| "rewards/margins": 4.530497074127197, | |
| "rewards/rejected": -2.371037244796753, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 2.2691292875989446, | |
| "grad_norm": 13.361975367104742, | |
| "learning_rate": 1.4238675016629336e-07, | |
| "logits/chosen": -2.0880515575408936, | |
| "logits/rejected": -2.045135974884033, | |
| "logps/chosen": -272.15704345703125, | |
| "logps/rejected": -254.84268188476562, | |
| "loss": 0.0998, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": 2.2964234352111816, | |
| "rewards/margins": 4.824353218078613, | |
| "rewards/rejected": -2.5279300212860107, | |
| "step": 215 | |
| }, | |
| { | |
| "epoch": 2.321899736147757, | |
| "grad_norm": 15.498546111351358, | |
| "learning_rate": 1.2281364768991804e-07, | |
| "logits/chosen": -2.091918468475342, | |
| "logits/rejected": -2.061227798461914, | |
| "logps/chosen": -256.5397033691406, | |
| "logps/rejected": -271.6669921875, | |
| "loss": 0.1052, | |
| "rewards/accuracies": 0.96875, | |
| "rewards/chosen": 2.6319310665130615, | |
| "rewards/margins": 5.266372203826904, | |
| "rewards/rejected": -2.63444185256958, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 2.37467018469657, | |
| "grad_norm": 14.379930581256865, | |
| "learning_rate": 1.0449812982251554e-07, | |
| "logits/chosen": -2.0655429363250732, | |
| "logits/rejected": -2.0514378547668457, | |
| "logps/chosen": -258.11456298828125, | |
| "logps/rejected": -248.6119842529297, | |
| "loss": 0.1248, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": 2.4950919151306152, | |
| "rewards/margins": 5.067896366119385, | |
| "rewards/rejected": -2.5728042125701904, | |
| "step": 225 | |
| }, | |
| { | |
| "epoch": 2.4274406332453826, | |
| "grad_norm": 13.318614530112333, | |
| "learning_rate": 8.75012627008489e-08, | |
| "logits/chosen": -2.0971901416778564, | |
| "logits/rejected": -2.0480291843414307, | |
| "logps/chosen": -265.1176452636719, | |
| "logps/rejected": -253.163818359375, | |
| "loss": 0.0959, | |
| "rewards/accuracies": 0.956250011920929, | |
| "rewards/chosen": 2.4935193061828613, | |
| "rewards/margins": 5.0423784255981445, | |
| "rewards/rejected": -2.548858880996704, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 2.480211081794195, | |
| "grad_norm": 17.75846671409766, | |
| "learning_rate": 7.187971592168934e-08, | |
| "logits/chosen": -2.102832317352295, | |
| "logits/rejected": -2.0520293712615967, | |
| "logps/chosen": -277.07965087890625, | |
| "logps/rejected": -245.1716766357422, | |
| "loss": 0.0977, | |
| "rewards/accuracies": 0.956250011920929, | |
| "rewards/chosen": 2.9457013607025146, | |
| "rewards/margins": 4.983917236328125, | |
| "rewards/rejected": -2.0382158756256104, | |
| "step": 235 | |
| }, | |
| { | |
| "epoch": 2.532981530343008, | |
| "grad_norm": 12.586486984541308, | |
| "learning_rate": 5.76855735985724e-08, | |
| "logits/chosen": -2.107192039489746, | |
| "logits/rejected": -2.062079668045044, | |
| "logps/chosen": -273.0597229003906, | |
| "logps/rejected": -258.69244384765625, | |
| "loss": 0.1106, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 2.69569730758667, | |
| "rewards/margins": 4.899806499481201, | |
| "rewards/rejected": -2.204108953475952, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 2.5857519788918206, | |
| "grad_norm": 13.220379474429727, | |
| "learning_rate": 4.4966160707107073e-08, | |
| "logits/chosen": -2.090113878250122, | |
| "logits/rejected": -2.0582504272460938, | |
| "logps/chosen": -262.7442932128906, | |
| "logps/rejected": -238.4239501953125, | |
| "loss": 0.1081, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": 2.032378911972046, | |
| "rewards/margins": 4.627095699310303, | |
| "rewards/rejected": -2.594717025756836, | |
| "step": 245 | |
| }, | |
| { | |
| "epoch": 2.638522427440633, | |
| "grad_norm": 8.791919342695266, | |
| "learning_rate": 3.376388529782215e-08, | |
| "logits/chosen": -2.1072940826416016, | |
| "logits/rejected": -2.0487256050109863, | |
| "logps/chosen": -260.51971435546875, | |
| "logps/rejected": -278.62286376953125, | |
| "loss": 0.0921, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": 2.3468101024627686, | |
| "rewards/margins": 5.20367431640625, | |
| "rewards/rejected": -2.8568639755249023, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 2.691292875989446, | |
| "grad_norm": 14.283825096734653, | |
| "learning_rate": 2.4116097102630906e-08, | |
| "logits/chosen": -2.0812129974365234, | |
| "logits/rejected": -2.054689645767212, | |
| "logps/chosen": -250.6894989013672, | |
| "logps/rejected": -239.05288696289062, | |
| "loss": 0.079, | |
| "rewards/accuracies": 0.96875, | |
| "rewards/chosen": 2.3491179943084717, | |
| "rewards/margins": 4.9853925704956055, | |
| "rewards/rejected": -2.636274576187134, | |
| "step": 255 | |
| }, | |
| { | |
| "epoch": 2.7440633245382586, | |
| "grad_norm": 13.777761464657969, | |
| "learning_rate": 1.605496300633874e-08, | |
| "logits/chosen": -2.073850631713867, | |
| "logits/rejected": -2.032045841217041, | |
| "logps/chosen": -260.50152587890625, | |
| "logps/rejected": -249.5687255859375, | |
| "loss": 0.0941, | |
| "rewards/accuracies": 0.96875, | |
| "rewards/chosen": 2.5012388229370117, | |
| "rewards/margins": 4.769859790802002, | |
| "rewards/rejected": -2.268620252609253, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 2.796833773087071, | |
| "grad_norm": 12.375261710883466, | |
| "learning_rate": 9.607359798384784e-09, | |
| "logits/chosen": -2.068258762359619, | |
| "logits/rejected": -2.036041736602783, | |
| "logps/chosen": -274.81884765625, | |
| "logps/rejected": -266.08685302734375, | |
| "loss": 0.0983, | |
| "rewards/accuracies": 0.981249988079071, | |
| "rewards/chosen": 2.848701000213623, | |
| "rewards/margins": 5.480710029602051, | |
| "rewards/rejected": -2.6320090293884277, | |
| "step": 265 | |
| }, | |
| { | |
| "epoch": 2.849604221635884, | |
| "grad_norm": 14.2159531121401, | |
| "learning_rate": 4.794784562397458e-09, | |
| "logits/chosen": -2.06805682182312, | |
| "logits/rejected": -2.047144651412964, | |
| "logps/chosen": -273.3573913574219, | |
| "logps/rejected": -254.7665252685547, | |
| "loss": 0.1101, | |
| "rewards/accuracies": 0.956250011920929, | |
| "rewards/chosen": 2.480971574783325, | |
| "rewards/margins": 4.715611457824707, | |
| "rewards/rejected": -2.2346396446228027, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 2.9023746701846966, | |
| "grad_norm": 16.497357456749207, | |
| "learning_rate": 1.6332830023350065e-09, | |
| "logits/chosen": -2.0658648014068604, | |
| "logits/rejected": -2.0133416652679443, | |
| "logps/chosen": -274.0085754394531, | |
| "logps/rejected": -252.18295288085938, | |
| "loss": 0.1145, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": 2.764272928237915, | |
| "rewards/margins": 4.97408390045166, | |
| "rewards/rejected": -2.209810733795166, | |
| "step": 275 | |
| }, | |
| { | |
| "epoch": 2.955145118733509, | |
| "grad_norm": 16.33614190524302, | |
| "learning_rate": 1.3339594418138035e-10, | |
| "logits/chosen": -2.0698370933532715, | |
| "logits/rejected": -2.033106803894043, | |
| "logps/chosen": -275.19708251953125, | |
| "logps/rejected": -241.0928192138672, | |
| "loss": 0.0954, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": 2.7163026332855225, | |
| "rewards/margins": 4.763990879058838, | |
| "rewards/rejected": -2.0476880073547363, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 2.9762532981530345, | |
| "step": 282, | |
| "total_flos": 3325131736743936.0, | |
| "train_loss": 0.285313792368199, | |
| "train_runtime": 10060.2707, | |
| "train_samples_per_second": 7.214, | |
| "train_steps_per_second": 0.028 | |
| } | |
| ], | |
| "logging_steps": 5, | |
| "max_steps": 282, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "save_steps": 100, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 3325131736743936.0, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |