| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 5.0, | |
| "eval_steps": 1000, | |
| "global_step": 625, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.008, | |
| "grad_norm": 7.598883452709097, | |
| "learning_rate": 1.5873015873015874e-07, | |
| "logits/chosen": -1.7671998739242554, | |
| "logits/rejected": -2.2639822959899902, | |
| "logps/chosen": -46.430763244628906, | |
| "logps/rejected": -102.85381317138672, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "grad_norm": 9.021711112045313, | |
| "learning_rate": 1.5873015873015873e-06, | |
| "logits/chosen": -1.6467827558517456, | |
| "logits/rejected": -2.05173921585083, | |
| "logps/chosen": -50.05971145629883, | |
| "logps/rejected": -110.39069366455078, | |
| "loss": 0.6403, | |
| "rewards/accuracies": 0.7777777910232544, | |
| "rewards/chosen": 0.0030777277424931526, | |
| "rewards/margins": 0.11931005865335464, | |
| "rewards/rejected": -0.11623234301805496, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "grad_norm": 1.8190332435996819, | |
| "learning_rate": 3.1746031746031746e-06, | |
| "logits/chosen": -1.3254443407058716, | |
| "logits/rejected": -1.779170036315918, | |
| "logps/chosen": -37.68037033081055, | |
| "logps/rejected": -313.1996154785156, | |
| "loss": 0.2472, | |
| "rewards/accuracies": 0.9937499761581421, | |
| "rewards/chosen": 0.09366801381111145, | |
| "rewards/margins": 2.184483051300049, | |
| "rewards/rejected": -2.0908150672912598, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.24, | |
| "grad_norm": 8.393638030400462, | |
| "learning_rate": 4.761904761904762e-06, | |
| "logits/chosen": -0.7945032715797424, | |
| "logits/rejected": -1.1047414541244507, | |
| "logps/chosen": -55.5176887512207, | |
| "logps/rejected": -406.99609375, | |
| "loss": 0.1799, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -0.04824959114193916, | |
| "rewards/margins": 3.0284643173217773, | |
| "rewards/rejected": -3.076713800430298, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "grad_norm": 3.6170805565428488, | |
| "learning_rate": 6.349206349206349e-06, | |
| "logits/chosen": -0.3710061311721802, | |
| "logits/rejected": -0.7456581592559814, | |
| "logps/chosen": -50.16427230834961, | |
| "logps/rejected": -541.6275634765625, | |
| "loss": 0.1374, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -0.01431644894182682, | |
| "rewards/margins": 4.384586334228516, | |
| "rewards/rejected": -4.398902416229248, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "grad_norm": 3.795764554921293, | |
| "learning_rate": 7.936507936507936e-06, | |
| "logits/chosen": -0.581534206867218, | |
| "logits/rejected": -0.8648965954780579, | |
| "logps/chosen": -80.1789779663086, | |
| "logps/rejected": -672.813720703125, | |
| "loss": 0.0553, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -0.3146038353443146, | |
| "rewards/margins": 5.392758846282959, | |
| "rewards/rejected": -5.707362174987793, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "grad_norm": 1.397363269705544, | |
| "learning_rate": 9.523809523809525e-06, | |
| "logits/chosen": -0.7689735293388367, | |
| "logits/rejected": -0.9561458826065063, | |
| "logps/chosen": -145.53469848632812, | |
| "logps/rejected": -831.8748779296875, | |
| "loss": 0.0371, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.9438420534133911, | |
| "rewards/margins": 6.384782314300537, | |
| "rewards/rejected": -7.328624725341797, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.56, | |
| "grad_norm": 0.13618082387486424, | |
| "learning_rate": 9.996172565322375e-06, | |
| "logits/chosen": -1.2459557056427002, | |
| "logits/rejected": -1.4872663021087646, | |
| "logps/chosen": -132.92080688476562, | |
| "logps/rejected": -911.9078979492188, | |
| "loss": 0.0409, | |
| "rewards/accuracies": 0.9937499761581421, | |
| "rewards/chosen": -0.8237069249153137, | |
| "rewards/margins": 7.28142786026001, | |
| "rewards/rejected": -8.105135917663574, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "grad_norm": 1.3606359582288166, | |
| "learning_rate": 9.97744005136599e-06, | |
| "logits/chosen": -1.503824234008789, | |
| "logits/rejected": -1.8992702960968018, | |
| "logps/chosen": -115.5097885131836, | |
| "logps/rejected": -916.9987182617188, | |
| "loss": 0.013, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.6468590497970581, | |
| "rewards/margins": 7.513754367828369, | |
| "rewards/rejected": -8.160614013671875, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.72, | |
| "grad_norm": 8.135027509581384, | |
| "learning_rate": 9.943157907471825e-06, | |
| "logits/chosen": -1.5570601224899292, | |
| "logits/rejected": -1.9588285684585571, | |
| "logps/chosen": -122.7470703125, | |
| "logps/rejected": -1007.5989379882812, | |
| "loss": 0.0185, | |
| "rewards/accuracies": 0.9937499761581421, | |
| "rewards/chosen": -0.71613609790802, | |
| "rewards/margins": 8.348057746887207, | |
| "rewards/rejected": -9.064192771911621, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "grad_norm": 0.07850356571109195, | |
| "learning_rate": 9.893433231795864e-06, | |
| "logits/chosen": -2.478653907775879, | |
| "logits/rejected": -2.867173671722412, | |
| "logps/chosen": -142.598388671875, | |
| "logps/rejected": -989.5442504882812, | |
| "loss": 0.0079, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.9251037836074829, | |
| "rewards/margins": 7.951480865478516, | |
| "rewards/rejected": -8.876585006713867, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "grad_norm": 0.010143597010992798, | |
| "learning_rate": 9.828421365296023e-06, | |
| "logits/chosen": -1.9893125295639038, | |
| "logits/rejected": -2.527003526687622, | |
| "logps/chosen": -127.56657409667969, | |
| "logps/rejected": -1111.281005859375, | |
| "loss": 0.0025, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.7765494585037231, | |
| "rewards/margins": 9.32172679901123, | |
| "rewards/rejected": -10.098276138305664, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "grad_norm": 0.005924828079353313, | |
| "learning_rate": 9.748325406443647e-06, | |
| "logits/chosen": -1.330843448638916, | |
| "logits/rejected": -2.0016732215881348, | |
| "logps/chosen": -80.08639526367188, | |
| "logps/rejected": -1047.385009765625, | |
| "loss": 0.0004, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.33333706855773926, | |
| "rewards/margins": 9.099153518676758, | |
| "rewards/rejected": -9.43249225616455, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 1.04, | |
| "grad_norm": 0.006907045924266673, | |
| "learning_rate": 9.653395576739504e-06, | |
| "logits/chosen": -1.3165078163146973, | |
| "logits/rejected": -1.999441385269165, | |
| "logps/chosen": -142.83404541015625, | |
| "logps/rejected": -1252.3265380859375, | |
| "loss": 0.004, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.9162995219230652, | |
| "rewards/margins": 10.60672378540039, | |
| "rewards/rejected": -11.52302360534668, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 1.12, | |
| "grad_norm": 0.11065788109459221, | |
| "learning_rate": 9.543928439016445e-06, | |
| "logits/chosen": -2.337254047393799, | |
| "logits/rejected": -2.8029887676239014, | |
| "logps/chosen": -141.67652893066406, | |
| "logps/rejected": -1141.8990478515625, | |
| "loss": 0.0013, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.9172345399856567, | |
| "rewards/margins": 9.48987102508545, | |
| "rewards/rejected": -10.407105445861816, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 1.2, | |
| "grad_norm": 0.08367928566939553, | |
| "learning_rate": 9.42026597097071e-06, | |
| "logits/chosen": -2.5381298065185547, | |
| "logits/rejected": -3.0234386920928955, | |
| "logps/chosen": -162.30075073242188, | |
| "logps/rejected": -1239.561767578125, | |
| "loss": 0.0006, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.1074072122573853, | |
| "rewards/margins": 10.28199291229248, | |
| "rewards/rejected": -11.389399528503418, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 1.28, | |
| "grad_norm": 0.0036761092850006146, | |
| "learning_rate": 9.282794496816244e-06, | |
| "logits/chosen": -2.3002023696899414, | |
| "logits/rejected": -2.8048605918884277, | |
| "logps/chosen": -136.76681518554688, | |
| "logps/rejected": -1181.3677978515625, | |
| "loss": 0.0008, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.8667460680007935, | |
| "rewards/margins": 9.932957649230957, | |
| "rewards/rejected": -10.799702644348145, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 1.3599999999999999, | |
| "grad_norm": 0.0021196612287394884, | |
| "learning_rate": 9.131943480399531e-06, | |
| "logits/chosen": -2.4460995197296143, | |
| "logits/rejected": -2.9539546966552734, | |
| "logps/chosen": -154.1061553955078, | |
| "logps/rejected": -1166.7919921875, | |
| "loss": 0.0017, | |
| "rewards/accuracies": 0.9937499761581421, | |
| "rewards/chosen": -1.0605894327163696, | |
| "rewards/margins": 9.586730003356934, | |
| "rewards/rejected": -10.647318840026855, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 1.44, | |
| "grad_norm": 0.004351477647807626, | |
| "learning_rate": 8.968184183545285e-06, | |
| "logits/chosen": -1.3524295091629028, | |
| "logits/rejected": -2.148332357406616, | |
| "logps/chosen": -146.11656188964844, | |
| "logps/rejected": -1293.1431884765625, | |
| "loss": 0.0002, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.959304928779602, | |
| "rewards/margins": 10.965489387512207, | |
| "rewards/rejected": -11.92479419708252, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 1.52, | |
| "grad_norm": 0.004066031914054124, | |
| "learning_rate": 8.792028193824364e-06, | |
| "logits/chosen": -0.9521867036819458, | |
| "logits/rejected": -1.7849302291870117, | |
| "logps/chosen": -146.82675170898438, | |
| "logps/rejected": -1224.1842041015625, | |
| "loss": 0.0011, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.9709192514419556, | |
| "rewards/margins": 10.264577865600586, | |
| "rewards/rejected": -11.235495567321777, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 1.6, | |
| "grad_norm": 2.8254638303864446, | |
| "learning_rate": 8.604025826343167e-06, | |
| "logits/chosen": -1.4731212854385376, | |
| "logits/rejected": -2.2158398628234863, | |
| "logps/chosen": -166.7774200439453, | |
| "logps/rejected": -1286.593505859375, | |
| "loss": 0.0007, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.1668568849563599, | |
| "rewards/margins": 10.681219100952148, | |
| "rewards/rejected": -11.848076820373535, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 1.6800000000000002, | |
| "grad_norm": 0.027929253424062123, | |
| "learning_rate": 8.404764404547404e-06, | |
| "logits/chosen": -1.5285543203353882, | |
| "logits/rejected": -2.207693099975586, | |
| "logps/chosen": -145.4683837890625, | |
| "logps/rejected": -1240.819091796875, | |
| "loss": 0.0003, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.9586501121520996, | |
| "rewards/margins": 10.435168266296387, | |
| "rewards/rejected": -11.393818855285645, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 1.76, | |
| "grad_norm": 0.00891305422535769, | |
| "learning_rate": 8.194866425410984e-06, | |
| "logits/chosen": -1.2906019687652588, | |
| "logits/rejected": -1.9918200969696045, | |
| "logps/chosen": -109.07038879394531, | |
| "logps/rejected": -1155.131591796875, | |
| "loss": 0.0002, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.6084994077682495, | |
| "rewards/margins": 9.912607192993164, | |
| "rewards/rejected": -10.521106719970703, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 1.8399999999999999, | |
| "grad_norm": 0.0018316293380985671, | |
| "learning_rate": 7.974987614742066e-06, | |
| "logits/chosen": -0.9763646125793457, | |
| "logits/rejected": -1.7541002035140991, | |
| "logps/chosen": -96.57362365722656, | |
| "logps/rejected": -1161.369384765625, | |
| "loss": 0.0004, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.4819282591342926, | |
| "rewards/margins": 10.10207462310791, | |
| "rewards/rejected": -10.584001541137695, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 1.92, | |
| "grad_norm": 0.003080239253269156, | |
| "learning_rate": 7.745814878681516e-06, | |
| "logits/chosen": -1.5342729091644287, | |
| "logits/rejected": -2.221407175064087, | |
| "logps/chosen": -136.1558074951172, | |
| "logps/rejected": -1264.690673828125, | |
| "loss": 0.0047, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.8548258543014526, | |
| "rewards/margins": 10.780150413513184, | |
| "rewards/rejected": -11.634977340698242, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 0.029482148279890704, | |
| "learning_rate": 7.50806415779332e-06, | |
| "logits/chosen": -2.258669853210449, | |
| "logits/rejected": -2.8217015266418457, | |
| "logps/chosen": -140.51853942871094, | |
| "logps/rejected": -1197.581787109375, | |
| "loss": 0.0003, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.8967872858047485, | |
| "rewards/margins": 10.075855255126953, | |
| "rewards/rejected": -10.97264289855957, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 2.08, | |
| "grad_norm": 0.020782306025551353, | |
| "learning_rate": 7.262478190450834e-06, | |
| "logits/chosen": -2.3182389736175537, | |
| "logits/rejected": -2.92055082321167, | |
| "logps/chosen": -120.69600677490234, | |
| "logps/rejected": -1217.408447265625, | |
| "loss": 0.0003, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.7067109942436218, | |
| "rewards/margins": 10.442468643188477, | |
| "rewards/rejected": -11.14918041229248, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 2.16, | |
| "grad_norm": 0.014378242296594932, | |
| "learning_rate": 7.0098241925061215e-06, | |
| "logits/chosen": -2.2274346351623535, | |
| "logits/rejected": -2.8495638370513916, | |
| "logps/chosen": -143.10069274902344, | |
| "logps/rejected": -1256.431640625, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.919533371925354, | |
| "rewards/margins": 10.646241188049316, | |
| "rewards/rejected": -11.565774917602539, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 2.24, | |
| "grad_norm": 0.0011703826093326713, | |
| "learning_rate": 6.750891460491093e-06, | |
| "logits/chosen": -1.953466773033142, | |
| "logits/rejected": -2.550438642501831, | |
| "logps/chosen": -119.64122009277344, | |
| "logps/rejected": -1253.8546142578125, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.6891164779663086, | |
| "rewards/margins": 10.844141960144043, | |
| "rewards/rejected": -11.533258438110352, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 2.32, | |
| "grad_norm": 0.016444776166175196, | |
| "learning_rate": 6.486488905838143e-06, | |
| "logits/chosen": -1.8186432123184204, | |
| "logits/rejected": -2.431183338165283, | |
| "logps/chosen": -114.60746002197266, | |
| "logps/rejected": -1220.325927734375, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.6494568586349487, | |
| "rewards/margins": 10.547185897827148, | |
| "rewards/rejected": -11.196642875671387, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 2.4, | |
| "grad_norm": 0.0005496505009132721, | |
| "learning_rate": 6.2174425278234115e-06, | |
| "logits/chosen": -1.7246840000152588, | |
| "logits/rejected": -2.2873055934906006, | |
| "logps/chosen": -108.48016357421875, | |
| "logps/rejected": -1233.37744140625, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.5782238245010376, | |
| "rewards/margins": 10.742240905761719, | |
| "rewards/rejected": -11.320464134216309, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 2.48, | |
| "grad_norm": 0.0006580056763243573, | |
| "learning_rate": 5.944592833127253e-06, | |
| "logits/chosen": -1.8141095638275146, | |
| "logits/rejected": -2.429810047149658, | |
| "logps/chosen": -119.1697006225586, | |
| "logps/rejected": -1326.865966796875, | |
| "loss": 0.0002, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.6833556890487671, | |
| "rewards/margins": 11.57683277130127, | |
| "rewards/rejected": -12.260188102722168, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 2.56, | |
| "grad_norm": 0.0011989998803146951, | |
| "learning_rate": 5.668792210073255e-06, | |
| "logits/chosen": -1.7204673290252686, | |
| "logits/rejected": -2.3327724933624268, | |
| "logps/chosen": -103.1290512084961, | |
| "logps/rejected": -1265.806640625, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.5410266518592834, | |
| "rewards/margins": 11.101877212524414, | |
| "rewards/rejected": -11.642904281616211, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 2.64, | |
| "grad_norm": 0.01089396369261857, | |
| "learning_rate": 5.39090226574877e-06, | |
| "logits/chosen": -1.52898371219635, | |
| "logits/rejected": -2.140092134475708, | |
| "logps/chosen": -91.94349670410156, | |
| "logps/rejected": -1207.033447265625, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.43181926012039185, | |
| "rewards/margins": 10.611051559448242, | |
| "rewards/rejected": -11.042871475219727, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 2.7199999999999998, | |
| "grad_norm": 0.0015016734818494304, | |
| "learning_rate": 5.111791134325793e-06, | |
| "logits/chosen": -1.6063101291656494, | |
| "logits/rejected": -2.3003313541412354, | |
| "logps/chosen": -126.36214447021484, | |
| "logps/rejected": -1331.2056884765625, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.7603046298027039, | |
| "rewards/margins": 11.541264533996582, | |
| "rewards/rejected": -12.301568984985352, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 2.8, | |
| "grad_norm": 0.044167012422386985, | |
| "learning_rate": 4.832330764991131e-06, | |
| "logits/chosen": -1.6678760051727295, | |
| "logits/rejected": -2.332516670227051, | |
| "logps/chosen": -119.7556381225586, | |
| "logps/rejected": -1301.3248291015625, | |
| "loss": 0.0004, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.6866098642349243, | |
| "rewards/margins": 11.313838958740234, | |
| "rewards/rejected": -12.000448226928711, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 2.88, | |
| "grad_norm": 0.0005385533629396381, | |
| "learning_rate": 4.553394197958339e-06, | |
| "logits/chosen": -1.9416593313217163, | |
| "logits/rejected": -2.6904163360595703, | |
| "logps/chosen": -121.2122802734375, | |
| "logps/rejected": -1295.510498046875, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.7095136642456055, | |
| "rewards/margins": 11.224905014038086, | |
| "rewards/rejected": -11.934419631958008, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 2.96, | |
| "grad_norm": 0.0006203949257092619, | |
| "learning_rate": 4.275852837071309e-06, | |
| "logits/chosen": -1.8644075393676758, | |
| "logits/rejected": -2.7058677673339844, | |
| "logps/chosen": -114.23481750488281, | |
| "logps/rejected": -1273.7774658203125, | |
| "loss": 0.0014, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.6469079256057739, | |
| "rewards/margins": 11.084717750549316, | |
| "rewards/rejected": -11.731626510620117, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 3.04, | |
| "grad_norm": 0.0018701198703410247, | |
| "learning_rate": 4.000573727519868e-06, | |
| "logits/chosen": -1.6911453008651733, | |
| "logits/rejected": -2.5057501792907715, | |
| "logps/chosen": -115.5822982788086, | |
| "logps/rejected": -1267.122314453125, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.6596534848213196, | |
| "rewards/margins": 11.005937576293945, | |
| "rewards/rejected": -11.665590286254883, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 3.12, | |
| "grad_norm": 0.004226077469779532, | |
| "learning_rate": 3.7284168471719527e-06, | |
| "logits/chosen": -1.646296501159668, | |
| "logits/rejected": -2.4523534774780273, | |
| "logps/chosen": -110.4627914428711, | |
| "logps/rejected": -1280.544921875, | |
| "loss": 0.0003, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.6036595106124878, | |
| "rewards/margins": 11.18139934539795, | |
| "rewards/rejected": -11.785058975219727, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 3.2, | |
| "grad_norm": 0.0006904055330191614, | |
| "learning_rate": 3.4602324199842026e-06, | |
| "logits/chosen": -1.7251287698745728, | |
| "logits/rejected": -2.5812900066375732, | |
| "logps/chosen": -138.866943359375, | |
| "logps/rejected": -1301.4326171875, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.8702784776687622, | |
| "rewards/margins": 11.145977973937988, | |
| "rewards/rejected": -12.016257286071777, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 3.2800000000000002, | |
| "grad_norm": 0.0026124260205581465, | |
| "learning_rate": 3.1968582598840234e-06, | |
| "logits/chosen": -1.5730335712432861, | |
| "logits/rejected": -2.383654832839966, | |
| "logps/chosen": -109.4728012084961, | |
| "logps/rejected": -1262.404541015625, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.6025822758674622, | |
| "rewards/margins": 11.004800796508789, | |
| "rewards/rejected": -11.607383728027344, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 3.36, | |
| "grad_norm": 0.016129813594014157, | |
| "learning_rate": 2.9391171534208185e-06, | |
| "logits/chosen": -1.7119560241699219, | |
| "logits/rejected": -2.51015043258667, | |
| "logps/chosen": -122.90666198730469, | |
| "logps/rejected": -1313.09228515625, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.7348455190658569, | |
| "rewards/margins": 11.394600868225098, | |
| "rewards/rejected": -12.129446983337402, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 3.44, | |
| "grad_norm": 0.0024401290515470016, | |
| "learning_rate": 2.6878142893630904e-06, | |
| "logits/chosen": -1.6607134342193604, | |
| "logits/rejected": -2.4547019004821777, | |
| "logps/chosen": -137.77719116210938, | |
| "logps/rejected": -1349.2073974609375, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.8648022413253784, | |
| "rewards/margins": 11.622838020324707, | |
| "rewards/rejected": -12.487640380859375, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 3.52, | |
| "grad_norm": 0.001375330125575793, | |
| "learning_rate": 2.4437347432713838e-06, | |
| "logits/chosen": -1.8045127391815186, | |
| "logits/rejected": -2.586901903152466, | |
| "logps/chosen": -156.62245178222656, | |
| "logps/rejected": -1351.8687744140625, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.0343620777130127, | |
| "rewards/margins": 11.499258995056152, | |
| "rewards/rejected": -12.533620834350586, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 3.6, | |
| "grad_norm": 0.0016289824782741704, | |
| "learning_rate": 2.207641024905322e-06, | |
| "logits/chosen": -1.6454054117202759, | |
| "logits/rejected": -2.4403579235076904, | |
| "logps/chosen": -108.79969787597656, | |
| "logps/rejected": -1255.9583740234375, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.5948039293289185, | |
| "rewards/margins": 10.957615852355957, | |
| "rewards/rejected": -11.552419662475586, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 3.68, | |
| "grad_norm": 0.0003909667399708806, | |
| "learning_rate": 1.9802706961266936e-06, | |
| "logits/chosen": -1.5957154035568237, | |
| "logits/rejected": -2.4980740547180176, | |
| "logps/chosen": -124.23823547363281, | |
| "logps/rejected": -1321.4207763671875, | |
| "loss": 0.0002, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.7479596138000488, | |
| "rewards/margins": 11.454780578613281, | |
| "rewards/rejected": -12.202739715576172, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 3.76, | |
| "grad_norm": 0.0006870240069107201, | |
| "learning_rate": 1.7623340667403089e-06, | |
| "logits/chosen": -1.5271575450897217, | |
| "logits/rejected": -2.3313746452331543, | |
| "logps/chosen": -115.49822998046875, | |
| "logps/rejected": -1287.483642578125, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.6483306288719177, | |
| "rewards/margins": 11.213019371032715, | |
| "rewards/rejected": -11.861350059509277, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 3.84, | |
| "grad_norm": 0.03279067348481457, | |
| "learning_rate": 1.5545119754708682e-06, | |
| "logits/chosen": -1.719366431236267, | |
| "logits/rejected": -2.5721096992492676, | |
| "logps/chosen": -141.73440551757812, | |
| "logps/rejected": -1358.951416015625, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.9100968241691589, | |
| "rewards/margins": 11.68384838104248, | |
| "rewards/rejected": -12.59394645690918, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 3.92, | |
| "grad_norm": 0.004304506433089471, | |
| "learning_rate": 1.3574536630081208e-06, | |
| "logits/chosen": -1.5175102949142456, | |
| "logits/rejected": -2.292325496673584, | |
| "logps/chosen": -100.6703872680664, | |
| "logps/rejected": -1273.519775390625, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.5087094902992249, | |
| "rewards/margins": 11.203500747680664, | |
| "rewards/rejected": -11.71220874786377, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.0004662700999661582, | |
| "learning_rate": 1.1717747437649657e-06, | |
| "logits/chosen": -1.5407252311706543, | |
| "logits/rejected": -2.3491063117980957, | |
| "logps/chosen": -122.54182434082031, | |
| "logps/rejected": -1335.7579345703125, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.7224875092506409, | |
| "rewards/margins": 11.629507064819336, | |
| "rewards/rejected": -12.351993560791016, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 4.08, | |
| "grad_norm": 0.0017439323111767312, | |
| "learning_rate": 9.980552826847635e-07, | |
| "logits/chosen": -1.6905838251113892, | |
| "logits/rejected": -2.487473487854004, | |
| "logps/chosen": -166.93624877929688, | |
| "logps/rejected": -1409.1981201171875, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.1514984369277954, | |
| "rewards/margins": 11.959856033325195, | |
| "rewards/rejected": -13.111355781555176, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 4.16, | |
| "grad_norm": 0.0024902235420863005, | |
| "learning_rate": 8.368379831059592e-07, | |
| "logits/chosen": -1.5832383632659912, | |
| "logits/rejected": -2.3622446060180664, | |
| "logps/chosen": -130.75253295898438, | |
| "logps/rejected": -1372.863037109375, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.806990921497345, | |
| "rewards/margins": 11.918740272521973, | |
| "rewards/rejected": -12.725730895996094, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 4.24, | |
| "grad_norm": 0.0004030809278175782, | |
| "learning_rate": 6.886264913451635e-07, | |
| "logits/chosen": -1.5754905939102173, | |
| "logits/rejected": -2.314927101135254, | |
| "logps/chosen": -109.8852310180664, | |
| "logps/rejected": -1284.387451171875, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.6023492813110352, | |
| "rewards/margins": 11.232891082763672, | |
| "rewards/rejected": -11.835241317749023, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 4.32, | |
| "grad_norm": 0.014990693875413815, | |
| "learning_rate": 5.538838232952104e-07, | |
| "logits/chosen": -1.6247650384902954, | |
| "logits/rejected": -2.395631790161133, | |
| "logps/chosen": -112.5528335571289, | |
| "logps/rejected": -1308.465087890625, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.6267929673194885, | |
| "rewards/margins": 11.434622764587402, | |
| "rewards/rejected": -12.06141471862793, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 4.4, | |
| "grad_norm": 0.0008931683311901964, | |
| "learning_rate": 4.3303091795353024e-07, | |
| "logits/chosen": -1.7327635288238525, | |
| "logits/rejected": -2.521054744720459, | |
| "logps/chosen": -193.12649536132812, | |
| "logps/rejected": -1453.8939208984375, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.3928953409194946, | |
| "rewards/margins": 12.179082870483398, | |
| "rewards/rejected": -13.571980476379395, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 4.48, | |
| "grad_norm": 0.0006090223103858115, | |
| "learning_rate": 3.2644532239966444e-07, | |
| "logits/chosen": -1.540307879447937, | |
| "logits/rejected": -2.274130344390869, | |
| "logps/chosen": -116.36529541015625, | |
| "logps/rejected": -1331.5732421875, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.6516658663749695, | |
| "rewards/margins": 11.651314735412598, | |
| "rewards/rejected": -12.302980422973633, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 4.5600000000000005, | |
| "grad_norm": 0.0012600919102292293, | |
| "learning_rate": 2.3446001233004333e-07, | |
| "logits/chosen": -1.545082926750183, | |
| "logits/rejected": -2.3070833683013916, | |
| "logps/chosen": -118.43846130371094, | |
| "logps/rejected": -1310.844482421875, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.6805299520492554, | |
| "rewards/margins": 11.414255142211914, | |
| "rewards/rejected": -12.094785690307617, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 4.64, | |
| "grad_norm": 0.0033895825817679913, | |
| "learning_rate": 1.573623518347517e-07, | |
| "logits/chosen": -1.3955062627792358, | |
| "logits/rejected": -2.1139488220214844, | |
| "logps/chosen": -95.18065643310547, | |
| "logps/rejected": -1255.7984619140625, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.46435871720314026, | |
| "rewards/margins": 11.06248664855957, | |
| "rewards/rejected": -11.52684497833252, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 4.72, | |
| "grad_norm": 0.0005277953893460651, | |
| "learning_rate": 9.539319566590766e-08, | |
| "logits/chosen": -1.4713037014007568, | |
| "logits/rejected": -2.1673684120178223, | |
| "logps/chosen": -103.91761779785156, | |
| "logps/rejected": -1269.9681396484375, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.5508452653884888, | |
| "rewards/margins": 11.134601593017578, | |
| "rewards/rejected": -11.685447692871094, | |
| "step": 590 | |
| }, | |
| { | |
| "epoch": 4.8, | |
| "grad_norm": 0.0038828625096458437, | |
| "learning_rate": 4.8746136802240716e-08, | |
| "logits/chosen": -1.60639226436615, | |
| "logits/rejected": -2.2908573150634766, | |
| "logps/chosen": -111.3029556274414, | |
| "logps/rejected": -1284.128173828125, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.6181883811950684, | |
| "rewards/margins": 11.210031509399414, | |
| "rewards/rejected": -11.82822036743164, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 4.88, | |
| "grad_norm": 0.0016152355239910333, | |
| "learning_rate": 1.75669016604485e-08, | |
| "logits/chosen": -1.5895562171936035, | |
| "logits/rejected": -2.336843490600586, | |
| "logps/chosen": -118.6875991821289, | |
| "logps/rejected": -1293.158935546875, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.6965440511703491, | |
| "rewards/margins": 11.21881103515625, | |
| "rewards/rejected": -11.915353775024414, | |
| "step": 610 | |
| }, | |
| { | |
| "epoch": 4.96, | |
| "grad_norm": 0.0004233911082118195, | |
| "learning_rate": 1.952894842735531e-09, | |
| "logits/chosen": -1.5527576208114624, | |
| "logits/rejected": -2.2691588401794434, | |
| "logps/chosen": -98.73721313476562, | |
| "logps/rejected": -1254.2822265625, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.4978357255458832, | |
| "rewards/margins": 11.023382186889648, | |
| "rewards/rejected": -11.52121639251709, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "step": 625, | |
| "total_flos": 0.0, | |
| "train_loss": 0.022511796173290348, | |
| "train_runtime": 65461.2558, | |
| "train_samples_per_second": 0.611, | |
| "train_steps_per_second": 0.01 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 625, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 5, | |
| "save_steps": 1000, | |
| "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": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |