| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.998691442030882, | |
| "eval_steps": 500, | |
| "global_step": 477, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.002093692750588851, | |
| "grad_norm": 3.652106715378975, | |
| "learning_rate": 1.0416666666666666e-08, | |
| "logits/chosen": 4904.83349609375, | |
| "logits/rejected": 4510.8515625, | |
| "logps/chosen": -227.15560913085938, | |
| "logps/rejected": -163.90182495117188, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.02093692750588851, | |
| "grad_norm": 3.4125691438258094, | |
| "learning_rate": 1.0416666666666667e-07, | |
| "logits/chosen": 5918.8134765625, | |
| "logits/rejected": 4984.322265625, | |
| "logps/chosen": -247.88807678222656, | |
| "logps/rejected": -215.54205322265625, | |
| "loss": 0.6932, | |
| "rewards/accuracies": 0.4722222089767456, | |
| "rewards/chosen": -0.0003395989187993109, | |
| "rewards/margins": -0.0001249208435183391, | |
| "rewards/rejected": -0.00021467804617714137, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.04187385501177702, | |
| "grad_norm": 3.574432275378761, | |
| "learning_rate": 2.0833333333333333e-07, | |
| "logits/chosen": 5801.2060546875, | |
| "logits/rejected": 4794.20458984375, | |
| "logps/chosen": -245.21127319335938, | |
| "logps/rejected": -208.40121459960938, | |
| "loss": 0.693, | |
| "rewards/accuracies": 0.543749988079071, | |
| "rewards/chosen": 0.0002933222276624292, | |
| "rewards/margins": 0.0006072597461752594, | |
| "rewards/rejected": -0.0003139374894089997, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.06281078251766553, | |
| "grad_norm": 3.454558671637055, | |
| "learning_rate": 3.1249999999999997e-07, | |
| "logits/chosen": 5605.74072265625, | |
| "logits/rejected": 4871.9853515625, | |
| "logps/chosen": -221.7666778564453, | |
| "logps/rejected": -195.719482421875, | |
| "loss": 0.6922, | |
| "rewards/accuracies": 0.606249988079071, | |
| "rewards/chosen": 0.0020034019835293293, | |
| "rewards/margins": 0.004107593558728695, | |
| "rewards/rejected": -0.0021041915751993656, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.08374771002355404, | |
| "grad_norm": 3.4260672878459753, | |
| "learning_rate": 4.1666666666666667e-07, | |
| "logits/chosen": 5728.78662109375, | |
| "logits/rejected": 4998.1181640625, | |
| "logps/chosen": -219.33212280273438, | |
| "logps/rejected": -207.6959991455078, | |
| "loss": 0.69, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": 0.0023276545107364655, | |
| "rewards/margins": 0.013038136065006256, | |
| "rewards/rejected": -0.010710480622947216, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.10468463752944256, | |
| "grad_norm": 3.894605994772773, | |
| "learning_rate": 4.999731868769026e-07, | |
| "logits/chosen": 6021.73388671875, | |
| "logits/rejected": 5236.958984375, | |
| "logps/chosen": -226.1421356201172, | |
| "logps/rejected": -219.1000518798828, | |
| "loss": 0.6867, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -0.019845347851514816, | |
| "rewards/margins": 0.022543836385011673, | |
| "rewards/rejected": -0.04238918051123619, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.12562156503533106, | |
| "grad_norm": 3.9844472720303865, | |
| "learning_rate": 4.990353313429303e-07, | |
| "logits/chosen": 5815.6328125, | |
| "logits/rejected": 4951.1953125, | |
| "logps/chosen": -229.83743286132812, | |
| "logps/rejected": -212.3618927001953, | |
| "loss": 0.6803, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": -0.08109215646982193, | |
| "rewards/margins": 0.061417847871780396, | |
| "rewards/rejected": -0.14251002669334412, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.14655849254121958, | |
| "grad_norm": 4.688078197523215, | |
| "learning_rate": 4.967625656594781e-07, | |
| "logits/chosen": 5626.0810546875, | |
| "logits/rejected": 5155.11083984375, | |
| "logps/chosen": -245.9986114501953, | |
| "logps/rejected": -248.27816772460938, | |
| "loss": 0.674, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -0.2200128138065338, | |
| "rewards/margins": 0.11577703803777695, | |
| "rewards/rejected": -0.33578985929489136, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.16749542004710807, | |
| "grad_norm": 4.970596658767943, | |
| "learning_rate": 4.93167072587771e-07, | |
| "logits/chosen": 5786.61962890625, | |
| "logits/rejected": 4942.87890625, | |
| "logps/chosen": -262.2926025390625, | |
| "logps/rejected": -260.3222351074219, | |
| "loss": 0.6634, | |
| "rewards/accuracies": 0.690625011920929, | |
| "rewards/chosen": -0.4204093813896179, | |
| "rewards/margins": 0.15950994193553925, | |
| "rewards/rejected": -0.5799193382263184, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.1884323475529966, | |
| "grad_norm": 5.267468034186471, | |
| "learning_rate": 4.882681251368548e-07, | |
| "logits/chosen": 5978.1640625, | |
| "logits/rejected": 5005.2685546875, | |
| "logps/chosen": -267.36871337890625, | |
| "logps/rejected": -272.29107666015625, | |
| "loss": 0.6562, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.47232165932655334, | |
| "rewards/margins": 0.23299944400787354, | |
| "rewards/rejected": -0.7053210735321045, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.2093692750588851, | |
| "grad_norm": 8.174427342804387, | |
| "learning_rate": 4.820919832540181e-07, | |
| "logits/chosen": 6043.4521484375, | |
| "logits/rejected": 5533.69580078125, | |
| "logps/chosen": -278.51739501953125, | |
| "logps/rejected": -301.57855224609375, | |
| "loss": 0.6545, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.541197657585144, | |
| "rewards/margins": 0.2974393367767334, | |
| "rewards/rejected": -0.8386369943618774, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.23030620256477363, | |
| "grad_norm": 9.05953079902565, | |
| "learning_rate": 4.7467175306295647e-07, | |
| "logits/chosen": 6159.54736328125, | |
| "logits/rejected": 4946.470703125, | |
| "logps/chosen": -292.4051513671875, | |
| "logps/rejected": -324.51922607421875, | |
| "loss": 0.6364, | |
| "rewards/accuracies": 0.7281249761581421, | |
| "rewards/chosen": -0.7568130493164062, | |
| "rewards/margins": 0.4172437787055969, | |
| "rewards/rejected": -1.1740567684173584, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.2512431300706621, | |
| "grad_norm": 6.739997326262223, | |
| "learning_rate": 4.6604720940421207e-07, | |
| "logits/chosen": 5756.60009765625, | |
| "logits/rejected": 5299.6982421875, | |
| "logps/chosen": -324.49249267578125, | |
| "logps/rejected": -363.9969177246094, | |
| "loss": 0.6485, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -1.1217855215072632, | |
| "rewards/margins": 0.3946547508239746, | |
| "rewards/rejected": -1.516440510749817, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.2721800575765506, | |
| "grad_norm": 6.621930453733046, | |
| "learning_rate": 4.5626458262912735e-07, | |
| "logits/chosen": 5827.3271484375, | |
| "logits/rejected": 5155.0693359375, | |
| "logps/chosen": -297.32525634765625, | |
| "logps/rejected": -331.3907775878906, | |
| "loss": 0.65, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.8195554614067078, | |
| "rewards/margins": 0.33257201313972473, | |
| "rewards/rejected": -1.1521273851394653, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.29311698508243916, | |
| "grad_norm": 8.184989902661265, | |
| "learning_rate": 4.453763107901675e-07, | |
| "logits/chosen": 5592.46826171875, | |
| "logits/rejected": 4921.5244140625, | |
| "logps/chosen": -312.6219177246094, | |
| "logps/rejected": -336.4300842285156, | |
| "loss": 0.6409, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.929578959941864, | |
| "rewards/margins": 0.4212140440940857, | |
| "rewards/rejected": -1.3507930040359497, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.31405391258832765, | |
| "grad_norm": 7.770526198736392, | |
| "learning_rate": 4.3344075855595097e-07, | |
| "logits/chosen": 6135.81396484375, | |
| "logits/rejected": 4984.74560546875, | |
| "logps/chosen": -348.8048095703125, | |
| "logps/rejected": -365.989013671875, | |
| "loss": 0.6372, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -1.1235905885696411, | |
| "rewards/margins": 0.5181370973587036, | |
| "rewards/rejected": -1.6417274475097656, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.33499084009421615, | |
| "grad_norm": 10.490577519744107, | |
| "learning_rate": 4.2052190435769554e-07, | |
| "logits/chosen": 5809.59912109375, | |
| "logits/rejected": 4767.5107421875, | |
| "logps/chosen": -363.4933166503906, | |
| "logps/rejected": -403.258544921875, | |
| "loss": 0.6231, | |
| "rewards/accuracies": 0.746874988079071, | |
| "rewards/chosen": -1.3820796012878418, | |
| "rewards/margins": 0.6052602529525757, | |
| "rewards/rejected": -1.987339973449707, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.3559277676001047, | |
| "grad_norm": 9.415988436151078, | |
| "learning_rate": 4.0668899744407567e-07, | |
| "logits/chosen": 6179.80029296875, | |
| "logits/rejected": 5211.10791015625, | |
| "logps/chosen": -347.8581237792969, | |
| "logps/rejected": -359.56597900390625, | |
| "loss": 0.6284, | |
| "rewards/accuracies": 0.721875011920929, | |
| "rewards/chosen": -1.145115852355957, | |
| "rewards/margins": 0.5148926377296448, | |
| "rewards/rejected": -1.660008430480957, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.3768646951059932, | |
| "grad_norm": 9.226681349355893, | |
| "learning_rate": 3.920161866827889e-07, | |
| "logits/chosen": 5761.98828125, | |
| "logits/rejected": 4862.2822265625, | |
| "logps/chosen": -358.5799560546875, | |
| "logps/rejected": -377.93389892578125, | |
| "loss": 0.6314, | |
| "rewards/accuracies": 0.715624988079071, | |
| "rewards/chosen": -1.224297285079956, | |
| "rewards/margins": 0.5621532201766968, | |
| "rewards/rejected": -1.7864506244659424, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.39780162261188173, | |
| "grad_norm": 8.252168204957982, | |
| "learning_rate": 3.765821230985757e-07, | |
| "logits/chosen": 5842.5185546875, | |
| "logits/rejected": 4813.79052734375, | |
| "logps/chosen": -352.1559143066406, | |
| "logps/rejected": -399.0413818359375, | |
| "loss": 0.622, | |
| "rewards/accuracies": 0.7718750238418579, | |
| "rewards/chosen": -1.3295438289642334, | |
| "rewards/margins": 0.6773251891136169, | |
| "rewards/rejected": -2.006869077682495, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.4187385501177702, | |
| "grad_norm": 11.315021257291264, | |
| "learning_rate": 3.604695382782159e-07, | |
| "logits/chosen": 5441.796875, | |
| "logits/rejected": 4646.61767578125, | |
| "logps/chosen": -335.6774597167969, | |
| "logps/rejected": -384.28009033203125, | |
| "loss": 0.6346, | |
| "rewards/accuracies": 0.7593749761581421, | |
| "rewards/chosen": -1.2916799783706665, | |
| "rewards/margins": 0.5930269360542297, | |
| "rewards/rejected": -1.8847068548202515, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.4396754776236587, | |
| "grad_norm": 8.501093826534886, | |
| "learning_rate": 3.4376480090239047e-07, | |
| "logits/chosen": 6127.7705078125, | |
| "logits/rejected": 5139.142578125, | |
| "logps/chosen": -358.6240234375, | |
| "logps/rejected": -389.53839111328125, | |
| "loss": 0.6294, | |
| "rewards/accuracies": 0.7406250238418579, | |
| "rewards/chosen": -1.222839117050171, | |
| "rewards/margins": 0.5366731882095337, | |
| "rewards/rejected": -1.7595125436782837, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.46061240512954726, | |
| "grad_norm": 10.392980676816252, | |
| "learning_rate": 3.265574537815398e-07, | |
| "logits/chosen": 5526.58642578125, | |
| "logits/rejected": 5086.4755859375, | |
| "logps/chosen": -333.91912841796875, | |
| "logps/rejected": -384.867919921875, | |
| "loss": 0.62, | |
| "rewards/accuracies": 0.796875, | |
| "rewards/chosen": -1.1029677391052246, | |
| "rewards/margins": 0.635188102722168, | |
| "rewards/rejected": -1.738155722618103, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.48154933263543576, | |
| "grad_norm": 11.453344580020602, | |
| "learning_rate": 3.0893973387735683e-07, | |
| "logits/chosen": 5728.69580078125, | |
| "logits/rejected": 4758.76220703125, | |
| "logps/chosen": -353.7453308105469, | |
| "logps/rejected": -412.8028259277344, | |
| "loss": 0.6164, | |
| "rewards/accuracies": 0.753125011920929, | |
| "rewards/chosen": -1.3611700534820557, | |
| "rewards/margins": 0.7657342553138733, | |
| "rewards/rejected": -2.126904010772705, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.5024862601413242, | |
| "grad_norm": 8.180627717186571, | |
| "learning_rate": 2.910060778827554e-07, | |
| "logits/chosen": 5866.203125, | |
| "logits/rejected": 4879.27001953125, | |
| "logps/chosen": -354.0733337402344, | |
| "logps/rejected": -401.80963134765625, | |
| "loss": 0.6215, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -1.3669780492782593, | |
| "rewards/margins": 0.6288621425628662, | |
| "rewards/rejected": -1.995840072631836, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.5234231876472127, | |
| "grad_norm": 10.371103059038116, | |
| "learning_rate": 2.7285261601056697e-07, | |
| "logits/chosen": 5912.509765625, | |
| "logits/rejected": 4911.0068359375, | |
| "logps/chosen": -373.91375732421875, | |
| "logps/rejected": -420.58148193359375, | |
| "loss": 0.627, | |
| "rewards/accuracies": 0.746874988079071, | |
| "rewards/chosen": -1.393939733505249, | |
| "rewards/margins": 0.6612668633460999, | |
| "rewards/rejected": -2.055206537246704, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.5443601151531012, | |
| "grad_norm": 13.008161119529388, | |
| "learning_rate": 2.5457665670441937e-07, | |
| "logits/chosen": 5701.7294921875, | |
| "logits/rejected": 5080.9052734375, | |
| "logps/chosen": -376.61328125, | |
| "logps/rejected": -419.1117248535156, | |
| "loss": 0.6204, | |
| "rewards/accuracies": 0.784375011920929, | |
| "rewards/chosen": -1.4007240533828735, | |
| "rewards/margins": 0.651884138584137, | |
| "rewards/rejected": -2.052608013153076, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.5652970426589898, | |
| "grad_norm": 11.430857626049596, | |
| "learning_rate": 2.3627616503391812e-07, | |
| "logits/chosen": 5770.25048828125, | |
| "logits/rejected": 5216.2509765625, | |
| "logps/chosen": -362.5931091308594, | |
| "logps/rejected": -424.2901306152344, | |
| "loss": 0.6203, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -1.371613621711731, | |
| "rewards/margins": 0.7380041480064392, | |
| "rewards/rejected": -2.1096177101135254, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.5862339701648783, | |
| "grad_norm": 10.238288027151336, | |
| "learning_rate": 2.1804923757009882e-07, | |
| "logits/chosen": 5784.783203125, | |
| "logits/rejected": 4712.6181640625, | |
| "logps/chosen": -361.86749267578125, | |
| "logps/rejected": -408.02581787109375, | |
| "loss": 0.6187, | |
| "rewards/accuracies": 0.746874988079071, | |
| "rewards/chosen": -1.283023476600647, | |
| "rewards/margins": 0.7839054465293884, | |
| "rewards/rejected": -2.0669288635253906, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.6071708976707668, | |
| "grad_norm": 11.836706849708237, | |
| "learning_rate": 1.9999357655598891e-07, | |
| "logits/chosen": 6021.28955078125, | |
| "logits/rejected": 5582.01708984375, | |
| "logps/chosen": -380.15704345703125, | |
| "logps/rejected": -437.7364196777344, | |
| "loss": 0.6278, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -1.3834381103515625, | |
| "rewards/margins": 0.6087765097618103, | |
| "rewards/rejected": -1.992214560508728, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.6281078251766553, | |
| "grad_norm": 11.52456328666972, | |
| "learning_rate": 1.8220596619089573e-07, | |
| "logits/chosen": 5510.0224609375, | |
| "logits/rejected": 5133.35791015625, | |
| "logps/chosen": -356.4972839355469, | |
| "logps/rejected": -416.02044677734375, | |
| "loss": 0.6191, | |
| "rewards/accuracies": 0.746874988079071, | |
| "rewards/chosen": -1.291826605796814, | |
| "rewards/margins": 0.646382749080658, | |
| "rewards/rejected": -1.9382092952728271, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.6490447526825438, | |
| "grad_norm": 11.01114898946855, | |
| "learning_rate": 1.647817538357072e-07, | |
| "logits/chosen": 6207.8447265625, | |
| "logits/rejected": 5361.03564453125, | |
| "logps/chosen": -390.5240783691406, | |
| "logps/rejected": -438.2767028808594, | |
| "loss": 0.6149, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -1.5012855529785156, | |
| "rewards/margins": 0.6348371505737305, | |
| "rewards/rejected": -2.136122703552246, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.6699816801884323, | |
| "grad_norm": 12.51384214714789, | |
| "learning_rate": 1.478143389201113e-07, | |
| "logits/chosen": 5781.5048828125, | |
| "logits/rejected": 4920.125, | |
| "logps/chosen": -377.29022216796875, | |
| "logps/rejected": -443.91180419921875, | |
| "loss": 0.6203, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -1.5008214712142944, | |
| "rewards/margins": 0.7988003492355347, | |
| "rewards/rejected": -2.299621820449829, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.6909186076943209, | |
| "grad_norm": 10.78460049890849, | |
| "learning_rate": 1.3139467229135998e-07, | |
| "logits/chosen": 5555.7392578125, | |
| "logits/rejected": 5078.744140625, | |
| "logps/chosen": -372.0219421386719, | |
| "logps/rejected": -430.1839904785156, | |
| "loss": 0.6236, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -1.3855758905410767, | |
| "rewards/margins": 0.7292943596839905, | |
| "rewards/rejected": -2.114870071411133, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.7118555352002094, | |
| "grad_norm": 11.721311586946122, | |
| "learning_rate": 1.1561076868822755e-07, | |
| "logits/chosen": 5405.2060546875, | |
| "logits/rejected": 4614.5302734375, | |
| "logps/chosen": -345.7624206542969, | |
| "logps/rejected": -404.03363037109375, | |
| "loss": 0.6217, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": -1.3347055912017822, | |
| "rewards/margins": 0.718868613243103, | |
| "rewards/rejected": -2.0535740852355957, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.7327924627060979, | |
| "grad_norm": 10.945499562820919, | |
| "learning_rate": 1.0054723495346482e-07, | |
| "logits/chosen": 6087.5595703125, | |
| "logits/rejected": 4836.3017578125, | |
| "logps/chosen": -384.67120361328125, | |
| "logps/rejected": -427.477294921875, | |
| "loss": 0.6204, | |
| "rewards/accuracies": 0.7593749761581421, | |
| "rewards/chosen": -1.3101894855499268, | |
| "rewards/margins": 0.7134831547737122, | |
| "rewards/rejected": -2.023672580718994, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.7537293902119864, | |
| "grad_norm": 11.287611167045075, | |
| "learning_rate": 8.628481651367875e-08, | |
| "logits/chosen": 5760.5537109375, | |
| "logits/rejected": 4939.3935546875, | |
| "logps/chosen": -358.7193908691406, | |
| "logps/rejected": -413.4532165527344, | |
| "loss": 0.6245, | |
| "rewards/accuracies": 0.7406250238418579, | |
| "rewards/chosen": -1.2639433145523071, | |
| "rewards/margins": 0.7279934287071228, | |
| "rewards/rejected": -1.9919369220733643, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.7746663177178749, | |
| "grad_norm": 12.141387495601492, | |
| "learning_rate": 7.289996455765748e-08, | |
| "logits/chosen": 5631.306640625, | |
| "logits/rejected": 4705.03125, | |
| "logps/chosen": -380.37921142578125, | |
| "logps/rejected": -429.5140075683594, | |
| "loss": 0.6102, | |
| "rewards/accuracies": 0.7593749761581421, | |
| "rewards/chosen": -1.317676067352295, | |
| "rewards/margins": 0.7852860689163208, | |
| "rewards/rejected": -2.1029622554779053, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.7956032452237635, | |
| "grad_norm": 12.405125307958853, | |
| "learning_rate": 6.046442623320145e-08, | |
| "logits/chosen": 5931.7939453125, | |
| "logits/rejected": 5068.625, | |
| "logps/chosen": -378.31756591796875, | |
| "logps/rejected": -409.650634765625, | |
| "loss": 0.6244, | |
| "rewards/accuracies": 0.7406250238418579, | |
| "rewards/chosen": -1.3470385074615479, | |
| "rewards/margins": 0.6615097522735596, | |
| "rewards/rejected": -2.0085482597351074, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.816540172729652, | |
| "grad_norm": 9.677482662034228, | |
| "learning_rate": 4.904486005914027e-08, | |
| "logits/chosen": 6173.0888671875, | |
| "logits/rejected": 5347.0029296875, | |
| "logps/chosen": -382.66802978515625, | |
| "logps/rejected": -447.6190490722656, | |
| "loss": 0.6134, | |
| "rewards/accuracies": 0.7593749761581421, | |
| "rewards/chosen": -1.3735501766204834, | |
| "rewards/margins": 0.7947303056716919, | |
| "rewards/rejected": -2.1682803630828857, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.8374771002355405, | |
| "grad_norm": 8.859063561472233, | |
| "learning_rate": 3.8702478614051345e-08, | |
| "logits/chosen": 5736.6416015625, | |
| "logits/rejected": 4989.16943359375, | |
| "logps/chosen": -381.01031494140625, | |
| "logps/rejected": -414.2621154785156, | |
| "loss": 0.6216, | |
| "rewards/accuracies": 0.7593749761581421, | |
| "rewards/chosen": -1.3569508790969849, | |
| "rewards/margins": 0.6679108738899231, | |
| "rewards/rejected": -2.0248618125915527, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.8584140277414289, | |
| "grad_norm": 10.739392345746163, | |
| "learning_rate": 2.9492720416985e-08, | |
| "logits/chosen": 5716.4189453125, | |
| "logits/rejected": 4977.70068359375, | |
| "logps/chosen": -373.8516540527344, | |
| "logps/rejected": -419.96539306640625, | |
| "loss": 0.6141, | |
| "rewards/accuracies": 0.715624988079071, | |
| "rewards/chosen": -1.4153944253921509, | |
| "rewards/margins": 0.6307347416877747, | |
| "rewards/rejected": -2.046128749847412, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.8793509552473174, | |
| "grad_norm": 14.1735949631008, | |
| "learning_rate": 2.1464952759020856e-08, | |
| "logits/chosen": 6084.1474609375, | |
| "logits/rejected": 5216.306640625, | |
| "logps/chosen": -386.87371826171875, | |
| "logps/rejected": -443.08349609375, | |
| "loss": 0.6137, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -1.3818824291229248, | |
| "rewards/margins": 0.7465057373046875, | |
| "rewards/rejected": -2.1283881664276123, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.9002878827532059, | |
| "grad_norm": 11.556772540890211, | |
| "learning_rate": 1.4662207078575684e-08, | |
| "logits/chosen": 6019.7197265625, | |
| "logits/rejected": 4750.1123046875, | |
| "logps/chosen": -381.2154235839844, | |
| "logps/rejected": -432.96014404296875, | |
| "loss": 0.6068, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -1.374026894569397, | |
| "rewards/margins": 0.7791653871536255, | |
| "rewards/rejected": -2.1531922817230225, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 0.9212248102590945, | |
| "grad_norm": 11.373149208327353, | |
| "learning_rate": 9.12094829893642e-09, | |
| "logits/chosen": 5833.07080078125, | |
| "logits/rejected": 5466.525390625, | |
| "logps/chosen": -373.42926025390625, | |
| "logps/rejected": -441.54510498046875, | |
| "loss": 0.6153, | |
| "rewards/accuracies": 0.746874988079071, | |
| "rewards/chosen": -1.3335464000701904, | |
| "rewards/margins": 0.7267743945121765, | |
| "rewards/rejected": -2.0603206157684326, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.942161737764983, | |
| "grad_norm": 13.629808310438557, | |
| "learning_rate": 4.8708793644441086e-09, | |
| "logits/chosen": 5978.48193359375, | |
| "logits/rejected": 5039.6708984375, | |
| "logps/chosen": -384.88446044921875, | |
| "logps/rejected": -428.22052001953125, | |
| "loss": 0.6148, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -1.3654634952545166, | |
| "rewards/margins": 0.7422479391098022, | |
| "rewards/rejected": -2.1077115535736084, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.9630986652708715, | |
| "grad_norm": 12.426315535607491, | |
| "learning_rate": 1.9347820230782295e-09, | |
| "logits/chosen": 6148.8798828125, | |
| "logits/rejected": 5086.49658203125, | |
| "logps/chosen": -411.7068786621094, | |
| "logps/rejected": -458.9427795410156, | |
| "loss": 0.6211, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -1.4349669218063354, | |
| "rewards/margins": 0.7559154033660889, | |
| "rewards/rejected": -2.1908822059631348, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 0.98403559277676, | |
| "grad_norm": 10.932899404920517, | |
| "learning_rate": 3.2839470889836627e-10, | |
| "logits/chosen": 5713.95703125, | |
| "logits/rejected": 4984.78515625, | |
| "logps/chosen": -375.77093505859375, | |
| "logps/rejected": -430.6996154785156, | |
| "loss": 0.6118, | |
| "rewards/accuracies": 0.7718750238418579, | |
| "rewards/chosen": -1.386940598487854, | |
| "rewards/margins": 0.7800599932670593, | |
| "rewards/rejected": -2.1670005321502686, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 0.998691442030882, | |
| "step": 477, | |
| "total_flos": 0.0, | |
| "train_loss": 0.634834620437782, | |
| "train_runtime": 14636.479, | |
| "train_samples_per_second": 4.177, | |
| "train_steps_per_second": 0.033 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 477, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 1000, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": false, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 0.0, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |