| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 1.0, | |
| "eval_steps": 100, | |
| "global_step": 478, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0, | |
| "learning_rate": 1.0416666666666666e-08, | |
| "logits/chosen": -2.6553587913513184, | |
| "logits/rejected": -2.6760501861572266, | |
| "logps/chosen": -153.28274536132812, | |
| "logps/rejected": -127.2953872680664, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.02, | |
| "learning_rate": 1.0416666666666667e-07, | |
| "logits/chosen": -2.793574094772339, | |
| "logits/rejected": -2.7928600311279297, | |
| "logps/chosen": -169.88270568847656, | |
| "logps/rejected": -174.67208862304688, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.4375, | |
| "rewards/chosen": 0.0004781727911904454, | |
| "rewards/margins": 0.0005411193706095219, | |
| "rewards/rejected": -6.294667400652543e-05, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.04, | |
| "learning_rate": 2.0833333333333333e-07, | |
| "logits/chosen": -2.75508975982666, | |
| "logits/rejected": -2.7587742805480957, | |
| "logps/chosen": -173.97691345214844, | |
| "logps/rejected": -172.56187438964844, | |
| "loss": 0.693, | |
| "rewards/accuracies": 0.5625, | |
| "rewards/chosen": -0.00017117620154749602, | |
| "rewards/margins": 0.00056259281700477, | |
| "rewards/rejected": -0.0007337690331041813, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.06, | |
| "learning_rate": 3.1249999999999997e-07, | |
| "logits/chosen": -2.7977731227874756, | |
| "logits/rejected": -2.7815966606140137, | |
| "logps/chosen": -149.19393920898438, | |
| "logps/rejected": -148.75160217285156, | |
| "loss": 0.6923, | |
| "rewards/accuracies": 0.5249999761581421, | |
| "rewards/chosen": -0.0018012176733464003, | |
| "rewards/margins": 0.0007020062766969204, | |
| "rewards/rejected": -0.0025032239500433207, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "learning_rate": 4.1666666666666667e-07, | |
| "logits/chosen": -2.7791695594787598, | |
| "logits/rejected": -2.765839099884033, | |
| "logps/chosen": -145.63992309570312, | |
| "logps/rejected": -153.24330139160156, | |
| "loss": 0.691, | |
| "rewards/accuracies": 0.6312500238418579, | |
| "rewards/chosen": -0.0012001590803265572, | |
| "rewards/margins": 0.004328931216150522, | |
| "rewards/rejected": -0.005529090296477079, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "learning_rate": 4.999733114418725e-07, | |
| "logits/chosen": -2.765580415725708, | |
| "logits/rejected": -2.7574915885925293, | |
| "logps/chosen": -168.78195190429688, | |
| "logps/rejected": -149.6485595703125, | |
| "loss": 0.6867, | |
| "rewards/accuracies": 0.6187499761581421, | |
| "rewards/chosen": 0.0035712323151528835, | |
| "rewards/margins": 0.012459376826882362, | |
| "rewards/rejected": -0.008888144977390766, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "learning_rate": 4.990398100856366e-07, | |
| "logits/chosen": -2.7321128845214844, | |
| "logits/rejected": -2.7279860973358154, | |
| "logps/chosen": -148.0186004638672, | |
| "logps/rejected": -141.0856475830078, | |
| "loss": 0.6803, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": 0.05182523652911186, | |
| "rewards/margins": 0.043829865753650665, | |
| "rewards/rejected": 0.00799537356942892, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "learning_rate": 4.967775735898179e-07, | |
| "logits/chosen": -2.74849271774292, | |
| "logits/rejected": -2.7429099082946777, | |
| "logps/chosen": -175.6835174560547, | |
| "logps/rejected": -169.5352783203125, | |
| "loss": 0.6726, | |
| "rewards/accuracies": 0.643750011920929, | |
| "rewards/chosen": -0.026319989934563637, | |
| "rewards/margins": 0.061176080256700516, | |
| "rewards/rejected": -0.0874960646033287, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.17, | |
| "learning_rate": 4.931986719649298e-07, | |
| "logits/chosen": -2.633291244506836, | |
| "logits/rejected": -2.629636764526367, | |
| "logps/chosen": -169.12416076660156, | |
| "logps/rejected": -163.40679931640625, | |
| "loss": 0.6604, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -0.1176857128739357, | |
| "rewards/margins": 0.0713546946644783, | |
| "rewards/rejected": -0.1890403926372528, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.19, | |
| "learning_rate": 4.883222001996351e-07, | |
| "logits/chosen": -2.708993911743164, | |
| "logits/rejected": -2.6869051456451416, | |
| "logps/chosen": -195.36166381835938, | |
| "logps/rejected": -183.77658081054688, | |
| "loss": 0.6512, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": -0.2477399855852127, | |
| "rewards/margins": 0.12695127725601196, | |
| "rewards/rejected": -0.37469127774238586, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "learning_rate": 4.821741763807186e-07, | |
| "logits/chosen": -2.5655856132507324, | |
| "logits/rejected": -2.5531303882598877, | |
| "logps/chosen": -200.37026977539062, | |
| "logps/rejected": -205.42471313476562, | |
| "loss": 0.645, | |
| "rewards/accuracies": 0.6312500238418579, | |
| "rewards/chosen": -0.4126931130886078, | |
| "rewards/margins": 0.11018091440200806, | |
| "rewards/rejected": -0.5228740572929382, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "eval_logits/chosen": -2.6051313877105713, | |
| "eval_logits/rejected": -2.5876898765563965, | |
| "eval_logps/chosen": -291.487060546875, | |
| "eval_logps/rejected": -323.70025634765625, | |
| "eval_loss": 0.6137697696685791, | |
| "eval_rewards/accuracies": 0.7109375, | |
| "eval_rewards/chosen": -0.34447404742240906, | |
| "eval_rewards/margins": 0.31899595260620117, | |
| "eval_rewards/rejected": -0.6634700298309326, | |
| "eval_runtime": 53.8393, | |
| "eval_samples_per_second": 37.148, | |
| "eval_steps_per_second": 0.594, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.23, | |
| "learning_rate": 4.747874028753375e-07, | |
| "logits/chosen": -2.5329718589782715, | |
| "logits/rejected": -2.5321857929229736, | |
| "logps/chosen": -226.0692596435547, | |
| "logps/rejected": -221.8550262451172, | |
| "loss": 0.6378, | |
| "rewards/accuracies": 0.581250011920929, | |
| "rewards/chosen": -0.476815402507782, | |
| "rewards/margins": 0.19172437489032745, | |
| "rewards/rejected": -0.6685398817062378, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "learning_rate": 4.662012913161997e-07, | |
| "logits/chosen": -2.5088794231414795, | |
| "logits/rejected": -2.5054843425750732, | |
| "logps/chosen": -209.64541625976562, | |
| "logps/rejected": -223.4107208251953, | |
| "loss": 0.6382, | |
| "rewards/accuracies": 0.6312500238418579, | |
| "rewards/chosen": -0.6152997612953186, | |
| "rewards/margins": 0.22199754416942596, | |
| "rewards/rejected": -0.837297260761261, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.27, | |
| "learning_rate": 4.5646165232345103e-07, | |
| "logits/chosen": -2.5064823627471924, | |
| "logits/rejected": -2.4858460426330566, | |
| "logps/chosen": -223.5966339111328, | |
| "logps/rejected": -242.69461059570312, | |
| "loss": 0.6146, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.559280514717102, | |
| "rewards/margins": 0.3010101616382599, | |
| "rewards/rejected": -0.8602906465530396, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "learning_rate": 4.456204510851956e-07, | |
| "logits/chosen": -2.4447319507598877, | |
| "logits/rejected": -2.4147696495056152, | |
| "logps/chosen": -247.92050170898438, | |
| "logps/rejected": -308.4876708984375, | |
| "loss": 0.6237, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -0.6928067207336426, | |
| "rewards/margins": 0.2877771854400635, | |
| "rewards/rejected": -0.9805838465690613, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.31, | |
| "learning_rate": 4.337355301007335e-07, | |
| "logits/chosen": -2.5137321949005127, | |
| "logits/rejected": -2.46433687210083, | |
| "logps/chosen": -230.3649139404297, | |
| "logps/rejected": -242.9630126953125, | |
| "loss": 0.6261, | |
| "rewards/accuracies": 0.643750011920929, | |
| "rewards/chosen": -0.818808376789093, | |
| "rewards/margins": 0.23969991505146027, | |
| "rewards/rejected": -1.0585081577301025, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "learning_rate": 4.2087030056579986e-07, | |
| "logits/chosen": -2.4309020042419434, | |
| "logits/rejected": -2.4267687797546387, | |
| "logps/chosen": -253.67276000976562, | |
| "logps/rejected": -301.7275390625, | |
| "loss": 0.62, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.9178106188774109, | |
| "rewards/margins": 0.3893894851207733, | |
| "rewards/rejected": -1.3072000741958618, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.36, | |
| "learning_rate": 4.070934040463998e-07, | |
| "logits/chosen": -2.3415331840515137, | |
| "logits/rejected": -2.289360284805298, | |
| "logps/chosen": -309.43841552734375, | |
| "logps/rejected": -346.57745361328125, | |
| "loss": 0.6056, | |
| "rewards/accuracies": 0.606249988079071, | |
| "rewards/chosen": -1.3271605968475342, | |
| "rewards/margins": 0.3612029254436493, | |
| "rewards/rejected": -1.6883633136749268, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.38, | |
| "learning_rate": 3.9247834624635404e-07, | |
| "logits/chosen": -2.3409111499786377, | |
| "logits/rejected": -2.3219058513641357, | |
| "logps/chosen": -260.40386962890625, | |
| "logps/rejected": -308.46649169921875, | |
| "loss": 0.6178, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -1.1141560077667236, | |
| "rewards/margins": 0.31085366010665894, | |
| "rewards/rejected": -1.425009846687317, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "learning_rate": 3.7710310482256523e-07, | |
| "logits/chosen": -2.3359875679016113, | |
| "logits/rejected": -2.3345441818237305, | |
| "logps/chosen": -248.61245727539062, | |
| "logps/rejected": -272.8714599609375, | |
| "loss": 0.6087, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -0.832414984703064, | |
| "rewards/margins": 0.28758493065834045, | |
| "rewards/rejected": -1.119999885559082, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "learning_rate": 3.610497133404795e-07, | |
| "logits/chosen": -2.3152999877929688, | |
| "logits/rejected": -2.2719802856445312, | |
| "logps/chosen": -283.94561767578125, | |
| "logps/rejected": -298.6034240722656, | |
| "loss": 0.6088, | |
| "rewards/accuracies": 0.606249988079071, | |
| "rewards/chosen": -1.2314116954803467, | |
| "rewards/margins": 0.2198210060596466, | |
| "rewards/rejected": -1.451232671737671, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "eval_logits/chosen": -2.296954393386841, | |
| "eval_logits/rejected": -2.2505648136138916, | |
| "eval_logps/chosen": -381.8215026855469, | |
| "eval_logps/rejected": -442.4111328125, | |
| "eval_loss": 0.5601309537887573, | |
| "eval_rewards/accuracies": 0.73828125, | |
| "eval_rewards/chosen": -1.2478187084197998, | |
| "eval_rewards/margins": 0.6027593016624451, | |
| "eval_rewards/rejected": -1.8505780696868896, | |
| "eval_runtime": 53.1318, | |
| "eval_samples_per_second": 37.642, | |
| "eval_steps_per_second": 0.602, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.44, | |
| "learning_rate": 3.4440382358952115e-07, | |
| "logits/chosen": -2.245332717895508, | |
| "logits/rejected": -2.208613634109497, | |
| "logps/chosen": -322.2597351074219, | |
| "logps/rejected": -380.9275817871094, | |
| "loss": 0.5753, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -1.3745754957199097, | |
| "rewards/margins": 0.4796102046966553, | |
| "rewards/rejected": -1.854185700416565, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "learning_rate": 3.272542485937368e-07, | |
| "logits/chosen": -2.236409902572632, | |
| "logits/rejected": -2.2042953968048096, | |
| "logps/chosen": -291.12939453125, | |
| "logps/rejected": -318.00677490234375, | |
| "loss": 0.6031, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -1.2436137199401855, | |
| "rewards/margins": 0.3306885063648224, | |
| "rewards/rejected": -1.5743024349212646, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "learning_rate": 3.096924887558854e-07, | |
| "logits/chosen": -2.22724986076355, | |
| "logits/rejected": -2.243987798690796, | |
| "logps/chosen": -267.0990295410156, | |
| "logps/rejected": -302.5461120605469, | |
| "loss": 0.6021, | |
| "rewards/accuracies": 0.6312500238418579, | |
| "rewards/chosen": -1.1229889392852783, | |
| "rewards/margins": 0.2818460762500763, | |
| "rewards/rejected": -1.4048349857330322, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "learning_rate": 2.9181224366319943e-07, | |
| "logits/chosen": -2.167080879211426, | |
| "logits/rejected": -2.164618492126465, | |
| "logps/chosen": -282.51898193359375, | |
| "logps/rejected": -297.18804931640625, | |
| "loss": 0.5853, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -1.141631007194519, | |
| "rewards/margins": 0.3452518582344055, | |
| "rewards/rejected": -1.4868830442428589, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "learning_rate": 2.7370891215954565e-07, | |
| "logits/chosen": -2.233933687210083, | |
| "logits/rejected": -2.1930792331695557, | |
| "logps/chosen": -265.99151611328125, | |
| "logps/rejected": -305.2124328613281, | |
| "loss": 0.6125, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -1.0087003707885742, | |
| "rewards/margins": 0.35329657793045044, | |
| "rewards/rejected": -1.3619968891143799, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.54, | |
| "learning_rate": 2.55479083351317e-07, | |
| "logits/chosen": -2.2159643173217773, | |
| "logits/rejected": -2.176429033279419, | |
| "logps/chosen": -258.05267333984375, | |
| "logps/rejected": -276.36248779296875, | |
| "loss": 0.6118, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -1.0803258419036865, | |
| "rewards/margins": 0.3370239734649658, | |
| "rewards/rejected": -1.4173498153686523, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.56, | |
| "learning_rate": 2.3722002126275822e-07, | |
| "logits/chosen": -2.1824676990509033, | |
| "logits/rejected": -2.1633520126342773, | |
| "logps/chosen": -275.42816162109375, | |
| "logps/rejected": -318.5284423828125, | |
| "loss": 0.5836, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -1.093287706375122, | |
| "rewards/margins": 0.48871397972106934, | |
| "rewards/rejected": -1.582001805305481, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "learning_rate": 2.19029145890313e-07, | |
| "logits/chosen": -2.166625738143921, | |
| "logits/rejected": -2.1442408561706543, | |
| "logps/chosen": -299.334228515625, | |
| "logps/rejected": -323.8576354980469, | |
| "loss": 0.5885, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -1.312145709991455, | |
| "rewards/margins": 0.33559149503707886, | |
| "rewards/rejected": -1.6477371454238892, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.61, | |
| "learning_rate": 2.0100351342479216e-07, | |
| "logits/chosen": -2.1124095916748047, | |
| "logits/rejected": -2.091318368911743, | |
| "logps/chosen": -273.45526123046875, | |
| "logps/rejected": -320.23291015625, | |
| "loss": 0.5729, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -1.2136024236679077, | |
| "rewards/margins": 0.6087918281555176, | |
| "rewards/rejected": -1.8223940134048462, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.63, | |
| "learning_rate": 1.8323929841460178e-07, | |
| "logits/chosen": -2.0944509506225586, | |
| "logits/rejected": -2.042977809906006, | |
| "logps/chosen": -318.8859558105469, | |
| "logps/rejected": -362.031005859375, | |
| "loss": 0.5713, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -1.417192816734314, | |
| "rewards/margins": 0.5179849863052368, | |
| "rewards/rejected": -1.9351780414581299, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.63, | |
| "eval_logits/chosen": -2.1609904766082764, | |
| "eval_logits/rejected": -2.1131389141082764, | |
| "eval_logps/chosen": -374.8100891113281, | |
| "eval_logps/rejected": -456.4195556640625, | |
| "eval_loss": 0.5404335856437683, | |
| "eval_rewards/accuracies": 0.765625, | |
| "eval_rewards/chosen": -1.1777048110961914, | |
| "eval_rewards/margins": 0.8129580616950989, | |
| "eval_rewards/rejected": -1.9906628131866455, | |
| "eval_runtime": 53.1243, | |
| "eval_samples_per_second": 37.648, | |
| "eval_steps_per_second": 0.602, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.65, | |
| "learning_rate": 1.6583128063291573e-07, | |
| "logits/chosen": -2.0995776653289795, | |
| "logits/rejected": -2.084737539291382, | |
| "logps/chosen": -293.4931640625, | |
| "logps/rejected": -358.046142578125, | |
| "loss": 0.5972, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -1.297886610031128, | |
| "rewards/margins": 0.4939315915107727, | |
| "rewards/rejected": -1.7918182611465454, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "learning_rate": 1.488723393865766e-07, | |
| "logits/chosen": -2.1189067363739014, | |
| "logits/rejected": -2.0870201587677, | |
| "logps/chosen": -323.58111572265625, | |
| "logps/rejected": -349.4317932128906, | |
| "loss": 0.5988, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -1.445042371749878, | |
| "rewards/margins": 0.39154213666915894, | |
| "rewards/rejected": -1.8365846872329712, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.69, | |
| "learning_rate": 1.3245295796480788e-07, | |
| "logits/chosen": -2.028104543685913, | |
| "logits/rejected": -2.0024235248565674, | |
| "logps/chosen": -300.0089111328125, | |
| "logps/rejected": -327.14935302734375, | |
| "loss": 0.5806, | |
| "rewards/accuracies": 0.643750011920929, | |
| "rewards/chosen": -1.295310378074646, | |
| "rewards/margins": 0.41160374879837036, | |
| "rewards/rejected": -1.7069141864776611, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.71, | |
| "learning_rate": 1.1666074087171627e-07, | |
| "logits/chosen": -2.106621265411377, | |
| "logits/rejected": -2.084312677383423, | |
| "logps/chosen": -303.54571533203125, | |
| "logps/rejected": -323.85528564453125, | |
| "loss": 0.5842, | |
| "rewards/accuracies": 0.6187499761581421, | |
| "rewards/chosen": -1.239495873451233, | |
| "rewards/margins": 0.33583495020866394, | |
| "rewards/rejected": -1.5753308534622192, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "learning_rate": 1.0157994641835734e-07, | |
| "logits/chosen": -2.0272693634033203, | |
| "logits/rejected": -2.014112949371338, | |
| "logps/chosen": -298.81707763671875, | |
| "logps/rejected": -342.98248291015625, | |
| "loss": 0.602, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -1.1794991493225098, | |
| "rewards/margins": 0.47226953506469727, | |
| "rewards/rejected": -1.651768684387207, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "learning_rate": 8.729103716819111e-08, | |
| "logits/chosen": -2.0633723735809326, | |
| "logits/rejected": -2.032658100128174, | |
| "logps/chosen": -288.3153991699219, | |
| "logps/rejected": -329.348876953125, | |
| "loss": 0.5949, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -1.3004556894302368, | |
| "rewards/margins": 0.3886806070804596, | |
| "rewards/rejected": -1.6891365051269531, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.77, | |
| "learning_rate": 7.387025063449081e-08, | |
| "logits/chosen": -2.0877535343170166, | |
| "logits/rejected": -2.0346691608428955, | |
| "logps/chosen": -288.3091125488281, | |
| "logps/rejected": -319.13067626953125, | |
| "loss": 0.5867, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -1.2991220951080322, | |
| "rewards/margins": 0.4315001368522644, | |
| "rewards/rejected": -1.7306222915649414, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.79, | |
| "learning_rate": 6.138919252022435e-08, | |
| "logits/chosen": -2.0052437782287598, | |
| "logits/rejected": -2.026928663253784, | |
| "logps/chosen": -279.9800720214844, | |
| "logps/rejected": -343.59759521484375, | |
| "loss": 0.5808, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -1.3590168952941895, | |
| "rewards/margins": 0.5555295348167419, | |
| "rewards/rejected": -1.9145463705062866, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.82, | |
| "learning_rate": 4.991445467064689e-08, | |
| "logits/chosen": -2.0364718437194824, | |
| "logits/rejected": -2.0138964653015137, | |
| "logps/chosen": -342.26214599609375, | |
| "logps/rejected": -371.9187316894531, | |
| "loss": 0.5832, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -1.5362588167190552, | |
| "rewards/margins": 0.39626526832580566, | |
| "rewards/rejected": -1.9325239658355713, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "learning_rate": 3.9507259776993954e-08, | |
| "logits/chosen": -2.0687859058380127, | |
| "logits/rejected": -2.024091958999634, | |
| "logps/chosen": -327.12469482421875, | |
| "logps/rejected": -350.4504699707031, | |
| "loss": 0.585, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -1.5318901538848877, | |
| "rewards/margins": 0.41996774077415466, | |
| "rewards/rejected": -1.9518581628799438, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "eval_logits/chosen": -2.0913608074188232, | |
| "eval_logits/rejected": -2.0418004989624023, | |
| "eval_logps/chosen": -387.622802734375, | |
| "eval_logps/rejected": -465.2492370605469, | |
| "eval_loss": 0.532730221748352, | |
| "eval_rewards/accuracies": 0.76953125, | |
| "eval_rewards/chosen": -1.3058316707611084, | |
| "eval_rewards/margins": 0.7731280326843262, | |
| "eval_rewards/rejected": -2.0789597034454346, | |
| "eval_runtime": 53.1693, | |
| "eval_samples_per_second": 37.616, | |
| "eval_steps_per_second": 0.602, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.86, | |
| "learning_rate": 3.022313472693447e-08, | |
| "logits/chosen": -2.051201820373535, | |
| "logits/rejected": -2.0001602172851562, | |
| "logps/chosen": -301.98968505859375, | |
| "logps/rejected": -340.74334716796875, | |
| "loss": 0.6013, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -1.455548882484436, | |
| "rewards/margins": 0.3957931697368622, | |
| "rewards/rejected": -1.8513422012329102, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "learning_rate": 2.2111614344599684e-08, | |
| "logits/chosen": -2.003211736679077, | |
| "logits/rejected": -1.9390989542007446, | |
| "logps/chosen": -301.6129455566406, | |
| "logps/rejected": -358.48712158203125, | |
| "loss": 0.588, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -1.4678657054901123, | |
| "rewards/margins": 0.4569568634033203, | |
| "rewards/rejected": -1.9248225688934326, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.9, | |
| "learning_rate": 1.521597710086439e-08, | |
| "logits/chosen": -2.049384593963623, | |
| "logits/rejected": -2.022660732269287, | |
| "logps/chosen": -276.34271240234375, | |
| "logps/rejected": -341.69525146484375, | |
| "loss": 0.577, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -1.4056159257888794, | |
| "rewards/margins": 0.5912927985191345, | |
| "rewards/rejected": -1.9969087839126587, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 0.92, | |
| "learning_rate": 9.57301420397924e-09, | |
| "logits/chosen": -2.0629231929779053, | |
| "logits/rejected": -2.0302374362945557, | |
| "logps/chosen": -324.0228576660156, | |
| "logps/rejected": -372.5835266113281, | |
| "loss": 0.5727, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -1.5117217302322388, | |
| "rewards/margins": 0.4386143088340759, | |
| "rewards/rejected": -1.9503358602523804, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.94, | |
| "learning_rate": 5.212833302556258e-09, | |
| "logits/chosen": -2.0090086460113525, | |
| "logits/rejected": -1.9827800989151, | |
| "logps/chosen": -339.3094482421875, | |
| "logps/rejected": -378.0882873535156, | |
| "loss": 0.5813, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -1.4472534656524658, | |
| "rewards/margins": 0.464362770318985, | |
| "rewards/rejected": -1.9116159677505493, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "learning_rate": 2.158697848236607e-09, | |
| "logits/chosen": -2.037766218185425, | |
| "logits/rejected": -1.9724490642547607, | |
| "logps/chosen": -298.0849609375, | |
| "logps/rejected": -351.5104675292969, | |
| "loss": 0.5765, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -1.4280312061309814, | |
| "rewards/margins": 0.4992195665836334, | |
| "rewards/rejected": -1.927250623703003, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 0.98, | |
| "learning_rate": 4.269029751107489e-10, | |
| "logits/chosen": -2.028768301010132, | |
| "logits/rejected": -1.9958856105804443, | |
| "logps/chosen": -302.0503845214844, | |
| "logps/rejected": -326.9120178222656, | |
| "loss": 0.5964, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -1.4426469802856445, | |
| "rewards/margins": 0.42627257108688354, | |
| "rewards/rejected": -1.8689197301864624, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 478, | |
| "total_flos": 0.0, | |
| "train_loss": 0.6134292899814111, | |
| "train_runtime": 4347.6461, | |
| "train_samples_per_second": 14.061, | |
| "train_steps_per_second": 0.11 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 478, | |
| "num_train_epochs": 1, | |
| "save_steps": 100, | |
| "total_flos": 0.0, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |