| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.9984301412872841, | |
| "eval_steps": 500, | |
| "global_step": 477, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0020931449502878076, | |
| "grad_norm": 16.340937257369013, | |
| "learning_rate": 1.0416666666666666e-08, | |
| "logits/chosen": -1.3137571811676025, | |
| "logits/rejected": -1.3155428171157837, | |
| "logps/chosen": -358.085205078125, | |
| "logps/rejected": -360.6844482421875, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.020931449502878074, | |
| "grad_norm": 15.95485055252751, | |
| "learning_rate": 1.0416666666666667e-07, | |
| "logits/chosen": -1.2563481330871582, | |
| "logits/rejected": -1.2476683855056763, | |
| "logps/chosen": -299.458251953125, | |
| "logps/rejected": -277.6595458984375, | |
| "loss": 0.6933, | |
| "rewards/accuracies": 0.5555555820465088, | |
| "rewards/chosen": 1.3408768609224353e-05, | |
| "rewards/margins": 0.00760088674724102, | |
| "rewards/rejected": -0.007587478496134281, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.04186289900575615, | |
| "grad_norm": 14.819439660572645, | |
| "learning_rate": 2.0833333333333333e-07, | |
| "logits/chosen": -1.219632863998413, | |
| "logits/rejected": -1.2203152179718018, | |
| "logps/chosen": -312.0214538574219, | |
| "logps/rejected": -280.1725769042969, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.5249999761581421, | |
| "rewards/chosen": 0.0031699701212346554, | |
| "rewards/margins": 0.00014753341383766383, | |
| "rewards/rejected": 0.003022436983883381, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.06279434850863422, | |
| "grad_norm": 15.080656700275002, | |
| "learning_rate": 3.1249999999999997e-07, | |
| "logits/chosen": -1.2369500398635864, | |
| "logits/rejected": -1.2387772798538208, | |
| "logps/chosen": -289.1150207519531, | |
| "logps/rejected": -249.2782745361328, | |
| "loss": 0.6891, | |
| "rewards/accuracies": 0.512499988079071, | |
| "rewards/chosen": 0.030798697844147682, | |
| "rewards/margins": 0.002891167066991329, | |
| "rewards/rejected": 0.027907531708478928, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.0837257980115123, | |
| "grad_norm": 14.89428955444734, | |
| "learning_rate": 4.1666666666666667e-07, | |
| "logits/chosen": -1.223825216293335, | |
| "logits/rejected": -1.2183074951171875, | |
| "logps/chosen": -257.38848876953125, | |
| "logps/rejected": -259.7073669433594, | |
| "loss": 0.6779, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": 0.10449536889791489, | |
| "rewards/margins": 0.02815863862633705, | |
| "rewards/rejected": 0.07633672654628754, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.10465724751439037, | |
| "grad_norm": 14.339684271853905, | |
| "learning_rate": 4.999731868769026e-07, | |
| "logits/chosen": -1.2123075723648071, | |
| "logits/rejected": -1.2090098857879639, | |
| "logps/chosen": -284.5002746582031, | |
| "logps/rejected": -270.3381042480469, | |
| "loss": 0.6612, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": 0.1965118646621704, | |
| "rewards/margins": 0.08109600841999054, | |
| "rewards/rejected": 0.11541587114334106, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.12558869701726844, | |
| "grad_norm": 14.404391107962436, | |
| "learning_rate": 4.990353313429303e-07, | |
| "logits/chosen": -1.2295911312103271, | |
| "logits/rejected": -1.2253326177597046, | |
| "logps/chosen": -250.2545928955078, | |
| "logps/rejected": -242.4825439453125, | |
| "loss": 0.6391, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.31995826959609985, | |
| "rewards/margins": 0.1508600413799286, | |
| "rewards/rejected": 0.16909828782081604, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.14652014652014653, | |
| "grad_norm": 13.80624198776984, | |
| "learning_rate": 4.967625656594781e-07, | |
| "logits/chosen": -1.229933500289917, | |
| "logits/rejected": -1.225915551185608, | |
| "logps/chosen": -291.14410400390625, | |
| "logps/rejected": -280.30694580078125, | |
| "loss": 0.6164, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": 0.3125011920928955, | |
| "rewards/margins": 0.19667133688926697, | |
| "rewards/rejected": 0.11582987010478973, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.1674515960230246, | |
| "grad_norm": 16.377957325997038, | |
| "learning_rate": 4.93167072587771e-07, | |
| "logits/chosen": -1.2333672046661377, | |
| "logits/rejected": -1.2266961336135864, | |
| "logps/chosen": -322.6647644042969, | |
| "logps/rejected": -255.389404296875, | |
| "loss": 0.5995, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": 0.28232404589653015, | |
| "rewards/margins": 0.330217182636261, | |
| "rewards/rejected": -0.047893110662698746, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.18838304552590268, | |
| "grad_norm": 14.043489999345669, | |
| "learning_rate": 4.882681251368548e-07, | |
| "logits/chosen": -1.2742656469345093, | |
| "logits/rejected": -1.2482376098632812, | |
| "logps/chosen": -253.7241973876953, | |
| "logps/rejected": -265.04833984375, | |
| "loss": 0.5696, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 0.06319811195135117, | |
| "rewards/margins": 0.394509494304657, | |
| "rewards/rejected": -0.3313114047050476, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.20931449502878074, | |
| "grad_norm": 12.760977388847909, | |
| "learning_rate": 4.820919832540181e-07, | |
| "logits/chosen": -1.2820088863372803, | |
| "logits/rejected": -1.2887402772903442, | |
| "logps/chosen": -302.23724365234375, | |
| "logps/rejected": -294.84930419921875, | |
| "loss": 0.569, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -0.11593605577945709, | |
| "rewards/margins": 0.5848953723907471, | |
| "rewards/rejected": -0.7008314728736877, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.2302459445316588, | |
| "grad_norm": 13.965126785422115, | |
| "learning_rate": 4.7467175306295647e-07, | |
| "logits/chosen": -1.2517648935317993, | |
| "logits/rejected": -1.2490425109863281, | |
| "logps/chosen": -291.33343505859375, | |
| "logps/rejected": -285.234130859375, | |
| "loss": 0.5798, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -0.13307788968086243, | |
| "rewards/margins": 0.5294880270957947, | |
| "rewards/rejected": -0.6625660061836243, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.25117739403453687, | |
| "grad_norm": 14.338557887990715, | |
| "learning_rate": 4.6604720940421207e-07, | |
| "logits/chosen": -1.1928507089614868, | |
| "logits/rejected": -1.2148034572601318, | |
| "logps/chosen": -269.4814758300781, | |
| "logps/rejected": -275.809814453125, | |
| "loss": 0.5552, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -0.0706900805234909, | |
| "rewards/margins": 0.551015317440033, | |
| "rewards/rejected": -0.6217054128646851, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.272108843537415, | |
| "grad_norm": 13.380124741562339, | |
| "learning_rate": 4.5626458262912735e-07, | |
| "logits/chosen": -1.2292916774749756, | |
| "logits/rejected": -1.2278550863265991, | |
| "logps/chosen": -279.93707275390625, | |
| "logps/rejected": -283.0473937988281, | |
| "loss": 0.5332, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 0.029192794114351273, | |
| "rewards/margins": 0.5927339792251587, | |
| "rewards/rejected": -0.5635411739349365, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.29304029304029305, | |
| "grad_norm": 19.13310763479197, | |
| "learning_rate": 4.453763107901675e-07, | |
| "logits/chosen": -1.2881428003311157, | |
| "logits/rejected": -1.2795203924179077, | |
| "logps/chosen": -317.7713928222656, | |
| "logps/rejected": -281.768310546875, | |
| "loss": 0.5385, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": 0.0925106629729271, | |
| "rewards/margins": 0.792411744594574, | |
| "rewards/rejected": -0.6999011635780334, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.3139717425431711, | |
| "grad_norm": 17.819771074479416, | |
| "learning_rate": 4.3344075855595097e-07, | |
| "logits/chosen": -1.2540961503982544, | |
| "logits/rejected": -1.256716251373291, | |
| "logps/chosen": -283.3349609375, | |
| "logps/rejected": -264.36187744140625, | |
| "loss": 0.5513, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -0.17489957809448242, | |
| "rewards/margins": 0.6731396913528442, | |
| "rewards/rejected": -0.8480392694473267, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.3349031920460492, | |
| "grad_norm": 14.045308941417, | |
| "learning_rate": 4.2052190435769554e-07, | |
| "logits/chosen": -1.2732188701629639, | |
| "logits/rejected": -1.2924416065216064, | |
| "logps/chosen": -279.59515380859375, | |
| "logps/rejected": -270.7580261230469, | |
| "loss": 0.5395, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -0.07183124125003815, | |
| "rewards/margins": 0.7724566459655762, | |
| "rewards/rejected": -0.8442879915237427, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.35583464154892724, | |
| "grad_norm": 14.389219742552266, | |
| "learning_rate": 4.0668899744407567e-07, | |
| "logits/chosen": -1.2052724361419678, | |
| "logits/rejected": -1.2073228359222412, | |
| "logps/chosen": -262.7742004394531, | |
| "logps/rejected": -249.1964111328125, | |
| "loss": 0.546, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -0.12547028064727783, | |
| "rewards/margins": 0.5785378217697144, | |
| "rewards/rejected": -0.7040081024169922, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.37676609105180536, | |
| "grad_norm": 14.416540854634023, | |
| "learning_rate": 3.920161866827889e-07, | |
| "logits/chosen": -1.261725664138794, | |
| "logits/rejected": -1.2505862712860107, | |
| "logps/chosen": -275.26007080078125, | |
| "logps/rejected": -264.3808288574219, | |
| "loss": 0.5191, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -0.180962473154068, | |
| "rewards/margins": 0.6575189828872681, | |
| "rewards/rejected": -0.8384815454483032, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.3976975405546834, | |
| "grad_norm": 16.121736171233948, | |
| "learning_rate": 3.765821230985757e-07, | |
| "logits/chosen": -1.2564064264297485, | |
| "logits/rejected": -1.2468352317810059, | |
| "logps/chosen": -269.68450927734375, | |
| "logps/rejected": -278.04510498046875, | |
| "loss": 0.53, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -0.22583866119384766, | |
| "rewards/margins": 0.6085832715034485, | |
| "rewards/rejected": -0.8344219326972961, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.4186289900575615, | |
| "grad_norm": 72.96841658583594, | |
| "learning_rate": 3.604695382782159e-07, | |
| "logits/chosen": -1.2191921472549438, | |
| "logits/rejected": -1.235925316810608, | |
| "logps/chosen": -258.74053955078125, | |
| "logps/rejected": -288.37237548828125, | |
| "loss": 0.5526, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -0.37098079919815063, | |
| "rewards/margins": 0.4539188742637634, | |
| "rewards/rejected": -0.8248996734619141, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.43956043956043955, | |
| "grad_norm": 17.453963247637727, | |
| "learning_rate": 3.4376480090239047e-07, | |
| "logits/chosen": -1.2523726224899292, | |
| "logits/rejected": -1.2435147762298584, | |
| "logps/chosen": -314.97119140625, | |
| "logps/rejected": -277.51654052734375, | |
| "loss": 0.5439, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.18608003854751587, | |
| "rewards/margins": 0.7605607509613037, | |
| "rewards/rejected": -0.94664067029953, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.4604918890633176, | |
| "grad_norm": 15.908490762668235, | |
| "learning_rate": 3.265574537815398e-07, | |
| "logits/chosen": -1.2073228359222412, | |
| "logits/rejected": -1.2404286861419678, | |
| "logps/chosen": -237.31582641601562, | |
| "logps/rejected": -261.70037841796875, | |
| "loss": 0.527, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -0.21793127059936523, | |
| "rewards/margins": 0.7825945019721985, | |
| "rewards/rejected": -1.0005258321762085, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.48142333856619574, | |
| "grad_norm": 16.31169658586331, | |
| "learning_rate": 3.0893973387735683e-07, | |
| "logits/chosen": -1.2508623600006104, | |
| "logits/rejected": -1.247154712677002, | |
| "logps/chosen": -260.1632080078125, | |
| "logps/rejected": -257.7186279296875, | |
| "loss": 0.527, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": -0.09356329590082169, | |
| "rewards/margins": 0.9371424913406372, | |
| "rewards/rejected": -1.0307056903839111, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.5023547880690737, | |
| "grad_norm": 14.990862329437586, | |
| "learning_rate": 2.910060778827554e-07, | |
| "logits/chosen": -1.2494900226593018, | |
| "logits/rejected": -1.2233803272247314, | |
| "logps/chosen": -288.9205322265625, | |
| "logps/rejected": -287.9413757324219, | |
| "loss": 0.4961, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.2458188533782959, | |
| "rewards/margins": 0.8366943597793579, | |
| "rewards/rejected": -1.0825130939483643, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.5232862375719518, | |
| "grad_norm": 17.284585165133276, | |
| "learning_rate": 2.7285261601056697e-07, | |
| "logits/chosen": -1.2617541551589966, | |
| "logits/rejected": -1.233609914779663, | |
| "logps/chosen": -287.88665771484375, | |
| "logps/rejected": -269.63671875, | |
| "loss": 0.5159, | |
| "rewards/accuracies": 0.793749988079071, | |
| "rewards/chosen": -0.4091765880584717, | |
| "rewards/margins": 1.0085241794586182, | |
| "rewards/rejected": -1.4177007675170898, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.54421768707483, | |
| "grad_norm": 16.30189212914588, | |
| "learning_rate": 2.5457665670441937e-07, | |
| "logits/chosen": -1.3039329051971436, | |
| "logits/rejected": -1.3189932107925415, | |
| "logps/chosen": -273.03094482421875, | |
| "logps/rejected": -272.4352111816406, | |
| "loss": 0.5116, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.6695433855056763, | |
| "rewards/margins": 0.7468977570533752, | |
| "rewards/rejected": -1.4164413213729858, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.565149136577708, | |
| "grad_norm": 16.278994247151147, | |
| "learning_rate": 2.3627616503391812e-07, | |
| "logits/chosen": -1.251950979232788, | |
| "logits/rejected": -1.2527947425842285, | |
| "logps/chosen": -309.84271240234375, | |
| "logps/rejected": -300.3086853027344, | |
| "loss": 0.5241, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": -0.5385168790817261, | |
| "rewards/margins": 0.9867380261421204, | |
| "rewards/rejected": -1.5252548456192017, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.5860805860805861, | |
| "grad_norm": 18.46433312907332, | |
| "learning_rate": 2.1804923757009882e-07, | |
| "logits/chosen": -1.201102614402771, | |
| "logits/rejected": -1.220724105834961, | |
| "logps/chosen": -264.5877685546875, | |
| "logps/rejected": -256.20599365234375, | |
| "loss": 0.5223, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.3720601201057434, | |
| "rewards/margins": 0.8117911219596863, | |
| "rewards/rejected": -1.1838513612747192, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.6070120355834642, | |
| "grad_norm": 15.847830198038507, | |
| "learning_rate": 1.9999357655598891e-07, | |
| "logits/chosen": -1.2599542140960693, | |
| "logits/rejected": -1.233518362045288, | |
| "logps/chosen": -257.46136474609375, | |
| "logps/rejected": -277.2739562988281, | |
| "loss": 0.5181, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -0.46861186623573303, | |
| "rewards/margins": 0.8282958269119263, | |
| "rewards/rejected": -1.296907663345337, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.6279434850863422, | |
| "grad_norm": 14.949024890685736, | |
| "learning_rate": 1.8220596619089573e-07, | |
| "logits/chosen": -1.2312617301940918, | |
| "logits/rejected": -1.2143806219100952, | |
| "logps/chosen": -336.4346008300781, | |
| "logps/rejected": -305.5640563964844, | |
| "loss": 0.4966, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -0.49016547203063965, | |
| "rewards/margins": 0.7988624572753906, | |
| "rewards/rejected": -1.2890279293060303, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.6488749345892203, | |
| "grad_norm": 16.813422745977434, | |
| "learning_rate": 1.647817538357072e-07, | |
| "logits/chosen": -1.2181973457336426, | |
| "logits/rejected": -1.198678970336914, | |
| "logps/chosen": -315.1355895996094, | |
| "logps/rejected": -281.01611328125, | |
| "loss": 0.4986, | |
| "rewards/accuracies": 0.793749988079071, | |
| "rewards/chosen": -0.46934041380882263, | |
| "rewards/margins": 0.9643914103507996, | |
| "rewards/rejected": -1.4337317943572998, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.6698063840920984, | |
| "grad_norm": 17.925744659126426, | |
| "learning_rate": 1.478143389201113e-07, | |
| "logits/chosen": -1.2826982736587524, | |
| "logits/rejected": -1.260918378829956, | |
| "logps/chosen": -264.7817077636719, | |
| "logps/rejected": -253.742431640625, | |
| "loss": 0.4854, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -0.37943515181541443, | |
| "rewards/margins": 0.9557247161865234, | |
| "rewards/rejected": -1.3351598978042603, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.6907378335949764, | |
| "grad_norm": 16.065419444341575, | |
| "learning_rate": 1.3139467229135998e-07, | |
| "logits/chosen": -1.2470223903656006, | |
| "logits/rejected": -1.2457056045532227, | |
| "logps/chosen": -281.67779541015625, | |
| "logps/rejected": -292.11993408203125, | |
| "loss": 0.5031, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -0.27068477869033813, | |
| "rewards/margins": 0.8681548833847046, | |
| "rewards/rejected": -1.1388397216796875, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.7116692830978545, | |
| "grad_norm": 19.080459880375884, | |
| "learning_rate": 1.1561076868822755e-07, | |
| "logits/chosen": -1.177343726158142, | |
| "logits/rejected": -1.1864651441574097, | |
| "logps/chosen": -304.8360595703125, | |
| "logps/rejected": -294.51251220703125, | |
| "loss": 0.5087, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -0.46762943267822266, | |
| "rewards/margins": 0.8512303233146667, | |
| "rewards/rejected": -1.3188598155975342, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.7326007326007326, | |
| "grad_norm": 18.441162480749096, | |
| "learning_rate": 1.0054723495346482e-07, | |
| "logits/chosen": -1.2660634517669678, | |
| "logits/rejected": -1.2635161876678467, | |
| "logps/chosen": -271.19140625, | |
| "logps/rejected": -263.1456298828125, | |
| "loss": 0.4867, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -0.43542027473449707, | |
| "rewards/margins": 1.0579180717468262, | |
| "rewards/rejected": -1.4933383464813232, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.7535321821036107, | |
| "grad_norm": 15.426287175254291, | |
| "learning_rate": 8.628481651367875e-08, | |
| "logits/chosen": -1.2429046630859375, | |
| "logits/rejected": -1.2402927875518799, | |
| "logps/chosen": -322.0977478027344, | |
| "logps/rejected": -301.01617431640625, | |
| "loss": 0.5264, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.47941383719444275, | |
| "rewards/margins": 0.9736785888671875, | |
| "rewards/rejected": -1.453092336654663, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.7744636316064888, | |
| "grad_norm": 16.948130871679638, | |
| "learning_rate": 7.289996455765748e-08, | |
| "logits/chosen": -1.291482925415039, | |
| "logits/rejected": -1.283089280128479, | |
| "logps/chosen": -270.592529296875, | |
| "logps/rejected": -257.29815673828125, | |
| "loss": 0.514, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.3671807646751404, | |
| "rewards/margins": 0.8553284406661987, | |
| "rewards/rejected": -1.2225091457366943, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.7953950811093669, | |
| "grad_norm": 14.870629372413775, | |
| "learning_rate": 6.046442623320145e-08, | |
| "logits/chosen": -1.2129778861999512, | |
| "logits/rejected": -1.2233318090438843, | |
| "logps/chosen": -257.82354736328125, | |
| "logps/rejected": -295.4395751953125, | |
| "loss": 0.505, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -0.6078884601593018, | |
| "rewards/margins": 1.0272793769836426, | |
| "rewards/rejected": -1.6351678371429443, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.8163265306122449, | |
| "grad_norm": 21.173303854564995, | |
| "learning_rate": 4.904486005914027e-08, | |
| "logits/chosen": -1.2637814283370972, | |
| "logits/rejected": -1.227350115776062, | |
| "logps/chosen": -343.4383544921875, | |
| "logps/rejected": -330.2360534667969, | |
| "loss": 0.488, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -0.3783820867538452, | |
| "rewards/margins": 0.9135732650756836, | |
| "rewards/rejected": -1.2919553518295288, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.837257980115123, | |
| "grad_norm": 17.112121473067976, | |
| "learning_rate": 3.8702478614051345e-08, | |
| "logits/chosen": -1.2346088886260986, | |
| "logits/rejected": -1.2416532039642334, | |
| "logps/chosen": -271.0476989746094, | |
| "logps/rejected": -268.0299377441406, | |
| "loss": 0.5027, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -0.40478605031967163, | |
| "rewards/margins": 0.9699586033821106, | |
| "rewards/rejected": -1.3747446537017822, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.858189429618001, | |
| "grad_norm": 17.74156653235628, | |
| "learning_rate": 2.9492720416985e-08, | |
| "logits/chosen": -1.2814921140670776, | |
| "logits/rejected": -1.2781927585601807, | |
| "logps/chosen": -317.0015869140625, | |
| "logps/rejected": -294.0724182128906, | |
| "loss": 0.5116, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -0.42354822158813477, | |
| "rewards/margins": 1.0186049938201904, | |
| "rewards/rejected": -1.4421532154083252, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.8791208791208791, | |
| "grad_norm": 16.217670115768026, | |
| "learning_rate": 2.1464952759020856e-08, | |
| "logits/chosen": -1.146841287612915, | |
| "logits/rejected": -1.1821863651275635, | |
| "logps/chosen": -274.8821105957031, | |
| "logps/rejected": -300.4873046875, | |
| "loss": 0.4967, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -0.5592753291130066, | |
| "rewards/margins": 0.9687057733535767, | |
| "rewards/rejected": -1.527981162071228, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.9000523286237572, | |
| "grad_norm": 15.24751289616232, | |
| "learning_rate": 1.4662207078575684e-08, | |
| "logits/chosen": -1.2148033380508423, | |
| "logits/rejected": -1.1892964839935303, | |
| "logps/chosen": -300.93768310546875, | |
| "logps/rejected": -297.40887451171875, | |
| "loss": 0.4729, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -0.42483001947402954, | |
| "rewards/margins": 0.965556800365448, | |
| "rewards/rejected": -1.3903868198394775, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 0.9209837781266352, | |
| "grad_norm": 16.431523627178482, | |
| "learning_rate": 9.12094829893642e-09, | |
| "logits/chosen": -1.2636325359344482, | |
| "logits/rejected": -1.2488772869110107, | |
| "logps/chosen": -273.4022216796875, | |
| "logps/rejected": -249.5026397705078, | |
| "loss": 0.5065, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -0.38708069920539856, | |
| "rewards/margins": 0.8260801434516907, | |
| "rewards/rejected": -1.213160753250122, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.9419152276295133, | |
| "grad_norm": 14.749633727793965, | |
| "learning_rate": 4.8708793644441086e-09, | |
| "logits/chosen": -1.1481744050979614, | |
| "logits/rejected": -1.1602979898452759, | |
| "logps/chosen": -281.922119140625, | |
| "logps/rejected": -289.523193359375, | |
| "loss": 0.4946, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.5180338025093079, | |
| "rewards/margins": 0.9671915173530579, | |
| "rewards/rejected": -1.4852253198623657, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.9628466771323915, | |
| "grad_norm": 16.752693571204084, | |
| "learning_rate": 1.9347820230782295e-09, | |
| "logits/chosen": -1.2104110717773438, | |
| "logits/rejected": -1.1974222660064697, | |
| "logps/chosen": -281.140380859375, | |
| "logps/rejected": -257.756103515625, | |
| "loss": 0.5119, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.385597288608551, | |
| "rewards/margins": 0.9535913467407227, | |
| "rewards/rejected": -1.339188575744629, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 0.9837781266352695, | |
| "grad_norm": 16.6487775309576, | |
| "learning_rate": 3.2839470889836627e-10, | |
| "logits/chosen": -1.2300002574920654, | |
| "logits/rejected": -1.2165695428848267, | |
| "logps/chosen": -303.2815246582031, | |
| "logps/rejected": -305.7578430175781, | |
| "loss": 0.4778, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.34391146898269653, | |
| "rewards/margins": 0.9113001823425293, | |
| "rewards/rejected": -1.255211591720581, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 0.9984301412872841, | |
| "eval_logits/chosen": -1.2443331480026245, | |
| "eval_logits/rejected": -1.2345507144927979, | |
| "eval_logps/chosen": -279.912109375, | |
| "eval_logps/rejected": -303.2041931152344, | |
| "eval_loss": 0.5042367577552795, | |
| "eval_rewards/accuracies": 0.765625, | |
| "eval_rewards/chosen": -0.3376635015010834, | |
| "eval_rewards/margins": 0.9887116551399231, | |
| "eval_rewards/rejected": -1.3263752460479736, | |
| "eval_runtime": 230.6194, | |
| "eval_samples_per_second": 8.672, | |
| "eval_steps_per_second": 0.139, | |
| "step": 477 | |
| }, | |
| { | |
| "epoch": 0.9984301412872841, | |
| "step": 477, | |
| "total_flos": 0.0, | |
| "train_loss": 0.5425664483876098, | |
| "train_runtime": 17053.3304, | |
| "train_samples_per_second": 3.585, | |
| "train_steps_per_second": 0.028 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 477, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "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": 0.0, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |