| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 2.0, | |
| "eval_steps": 100, | |
| "global_step": 658, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0, | |
| "learning_rate": 7.575757575757576e-08, | |
| "logits/chosen": -3.0401854515075684, | |
| "logits/rejected": -2.9935803413391113, | |
| "logps/chosen": -33.6235466003418, | |
| "logps/rejected": -31.46235466003418, | |
| "loss": 0.0011, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.03, | |
| "learning_rate": 7.575757575757576e-07, | |
| "logits/chosen": -3.0845224857330322, | |
| "logits/rejected": -2.939974784851074, | |
| "logps/chosen": -187.66502380371094, | |
| "logps/rejected": -172.21742248535156, | |
| "loss": 0.0011, | |
| "rewards/accuracies": 0.4444444477558136, | |
| "rewards/chosen": 4.157695730100386e-05, | |
| "rewards/margins": -0.0002048778405878693, | |
| "rewards/rejected": 0.0002464548160787672, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.06, | |
| "learning_rate": 1.5151515151515152e-06, | |
| "logits/chosen": -2.975489616394043, | |
| "logits/rejected": -2.8098347187042236, | |
| "logps/chosen": -101.97224426269531, | |
| "logps/rejected": -79.41976928710938, | |
| "loss": 0.001, | |
| "rewards/accuracies": 0.6000000238418579, | |
| "rewards/chosen": 0.0006187028484418988, | |
| "rewards/margins": 0.0002881823165807873, | |
| "rewards/rejected": 0.0003305206191726029, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.09, | |
| "learning_rate": 2.2727272727272728e-06, | |
| "logits/chosen": -2.9616599082946777, | |
| "logits/rejected": -2.853991746902466, | |
| "logps/chosen": -93.24244689941406, | |
| "logps/rejected": -93.37409973144531, | |
| "loss": 0.0011, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": 0.0018816631054505706, | |
| "rewards/margins": 0.000262443587416783, | |
| "rewards/rejected": 0.001619219547137618, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.12, | |
| "learning_rate": 3.0303030303030305e-06, | |
| "logits/chosen": -2.957383871078491, | |
| "logits/rejected": -2.8873894214630127, | |
| "logps/chosen": -98.40808868408203, | |
| "logps/rejected": -90.6903305053711, | |
| "loss": 0.0011, | |
| "rewards/accuracies": 0.6000000238418579, | |
| "rewards/chosen": 0.0038822232745587826, | |
| "rewards/margins": 0.0006505203782580793, | |
| "rewards/rejected": 0.0032317028380930424, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "learning_rate": 3.7878787878787882e-06, | |
| "logits/chosen": -2.793536424636841, | |
| "logits/rejected": -2.701014995574951, | |
| "logps/chosen": -76.89764404296875, | |
| "logps/rejected": -76.88601684570312, | |
| "loss": 0.0009, | |
| "rewards/accuracies": 0.44999998807907104, | |
| "rewards/chosen": 0.0003440352447796613, | |
| "rewards/margins": 0.003660330083221197, | |
| "rewards/rejected": -0.003316295798867941, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.18, | |
| "learning_rate": 4.5454545454545455e-06, | |
| "logits/chosen": -2.8517611026763916, | |
| "logits/rejected": -2.701003313064575, | |
| "logps/chosen": -111.06632232666016, | |
| "logps/rejected": -97.74620056152344, | |
| "loss": 0.001, | |
| "rewards/accuracies": 0.4000000059604645, | |
| "rewards/chosen": 0.0011594895040616393, | |
| "rewards/margins": 0.00348883168771863, | |
| "rewards/rejected": -0.0023293420672416687, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "learning_rate": 4.999436790436924e-06, | |
| "logits/chosen": -3.074284791946411, | |
| "logits/rejected": -2.859107494354248, | |
| "logps/chosen": -175.5392608642578, | |
| "logps/rejected": -165.30654907226562, | |
| "loss": 0.0011, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": 0.0042682006023824215, | |
| "rewards/margins": 0.004156609531491995, | |
| "rewards/rejected": 0.00011159134737681597, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.24, | |
| "learning_rate": 4.993103596812269e-06, | |
| "logits/chosen": -2.7597286701202393, | |
| "logits/rejected": -2.6338610649108887, | |
| "logps/chosen": -117.37264251708984, | |
| "logps/rejected": -96.65496826171875, | |
| "loss": 0.0014, | |
| "rewards/accuracies": 0.30000001192092896, | |
| "rewards/chosen": -0.004695770796388388, | |
| "rewards/margins": -0.0023340280167758465, | |
| "rewards/rejected": -0.002361743012443185, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.27, | |
| "learning_rate": 4.979751088147192e-06, | |
| "logits/chosen": -2.7671074867248535, | |
| "logits/rejected": -2.5734329223632812, | |
| "logps/chosen": -124.0124740600586, | |
| "logps/rejected": -117.49992370605469, | |
| "loss": 0.0013, | |
| "rewards/accuracies": 0.550000011920929, | |
| "rewards/chosen": -0.0014447696739807725, | |
| "rewards/margins": 0.006112386472523212, | |
| "rewards/rejected": -0.0075571550987660885, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.3, | |
| "learning_rate": 4.95941685833271e-06, | |
| "logits/chosen": -2.841676712036133, | |
| "logits/rejected": -2.6840147972106934, | |
| "logps/chosen": -136.9166259765625, | |
| "logps/rejected": -142.25997924804688, | |
| "loss": 0.0012, | |
| "rewards/accuracies": 0.550000011920929, | |
| "rewards/chosen": -0.005834830459207296, | |
| "rewards/margins": 0.0021175735164433718, | |
| "rewards/rejected": -0.007952402345836163, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.3, | |
| "eval_logits/chosen": -2.3582723140716553, | |
| "eval_logits/rejected": -2.1644442081451416, | |
| "eval_logps/chosen": -281.1500244140625, | |
| "eval_logps/rejected": -259.7554626464844, | |
| "eval_loss": 0.001627120072953403, | |
| "eval_rewards/accuracies": 0.5034999847412109, | |
| "eval_rewards/chosen": -0.016425278037786484, | |
| "eval_rewards/margins": -0.00047451091813854873, | |
| "eval_rewards/rejected": -0.0159507654607296, | |
| "eval_runtime": 1420.6531, | |
| "eval_samples_per_second": 1.408, | |
| "eval_steps_per_second": 0.704, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "learning_rate": 4.9321581582449365e-06, | |
| "logits/chosen": -2.8993427753448486, | |
| "logits/rejected": -2.786620616912842, | |
| "logps/chosen": -118.65157318115234, | |
| "logps/rejected": -127.34912109375, | |
| "loss": 0.0011, | |
| "rewards/accuracies": 0.44999998807907104, | |
| "rewards/chosen": -0.0042579686269164085, | |
| "rewards/margins": 0.004396714735776186, | |
| "rewards/rejected": -0.008654682897031307, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.36, | |
| "learning_rate": 4.898051734555676e-06, | |
| "logits/chosen": -3.0458617210388184, | |
| "logits/rejected": -2.99739670753479, | |
| "logps/chosen": -95.90953063964844, | |
| "logps/rejected": -102.84895324707031, | |
| "loss": 0.0013, | |
| "rewards/accuracies": 0.550000011920929, | |
| "rewards/chosen": -0.004254769999533892, | |
| "rewards/margins": -0.0005313962465152144, | |
| "rewards/rejected": -0.0037233736366033554, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "learning_rate": 4.857193613652711e-06, | |
| "logits/chosen": -3.0125811100006104, | |
| "logits/rejected": -2.8767762184143066, | |
| "logps/chosen": -66.88941955566406, | |
| "logps/rejected": -71.67733001708984, | |
| "loss": 0.0012, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": 0.004949026275426149, | |
| "rewards/margins": 0.0032058332581073046, | |
| "rewards/rejected": 0.001743193482980132, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.43, | |
| "learning_rate": 4.809698831278217e-06, | |
| "logits/chosen": -2.9552953243255615, | |
| "logits/rejected": -2.7150585651397705, | |
| "logps/chosen": -75.46265411376953, | |
| "logps/rejected": -72.64187622070312, | |
| "loss": 0.001, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": 0.00027392804622650146, | |
| "rewards/margins": 0.016122380271553993, | |
| "rewards/rejected": -0.01584845408797264, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "learning_rate": 4.7557011086464634e-06, | |
| "logits/chosen": -2.9132320880889893, | |
| "logits/rejected": -2.64630126953125, | |
| "logps/chosen": -76.73291778564453, | |
| "logps/rejected": -97.9271011352539, | |
| "loss": 0.0013, | |
| "rewards/accuracies": 0.6000000238418579, | |
| "rewards/chosen": -0.00335231376811862, | |
| "rewards/margins": 0.01173357479274273, | |
| "rewards/rejected": -0.015085889026522636, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.49, | |
| "learning_rate": 4.695352475952706e-06, | |
| "logits/chosen": -3.013190746307373, | |
| "logits/rejected": -2.926192045211792, | |
| "logps/chosen": -73.78190612792969, | |
| "logps/rejected": -84.00215148925781, | |
| "loss": 0.0016, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.006495082285255194, | |
| "rewards/margins": 0.01742926612496376, | |
| "rewards/rejected": -0.023924347013235092, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "learning_rate": 4.6288228443332786e-06, | |
| "logits/chosen": -2.892408847808838, | |
| "logits/rejected": -2.785109758377075, | |
| "logps/chosen": -88.50574493408203, | |
| "logps/rejected": -63.979164123535156, | |
| "loss": 0.0014, | |
| "rewards/accuracies": 0.44999998807907104, | |
| "rewards/chosen": -0.015971587970852852, | |
| "rewards/margins": 0.0037319730035960674, | |
| "rewards/rejected": -0.019703561440110207, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.55, | |
| "learning_rate": 4.556299527482029e-06, | |
| "logits/chosen": -3.024970293045044, | |
| "logits/rejected": -2.9614624977111816, | |
| "logps/chosen": -95.4378433227539, | |
| "logps/rejected": -86.34746551513672, | |
| "loss": 0.0009, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.0011760194320231676, | |
| "rewards/margins": 0.010892638936638832, | |
| "rewards/rejected": -0.009716618806123734, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.58, | |
| "learning_rate": 4.4779867142699715e-06, | |
| "logits/chosen": -3.0479438304901123, | |
| "logits/rejected": -2.903231620788574, | |
| "logps/chosen": -138.85531616210938, | |
| "logps/rejected": -146.4200439453125, | |
| "loss": 0.0017, | |
| "rewards/accuracies": 0.44999998807907104, | |
| "rewards/chosen": 0.013675776310265064, | |
| "rewards/margins": 0.003341741394251585, | |
| "rewards/rejected": 0.010334033519029617, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.61, | |
| "learning_rate": 4.394104893853007e-06, | |
| "logits/chosen": -2.9416115283966064, | |
| "logits/rejected": -2.8143601417541504, | |
| "logps/chosen": -85.66619110107422, | |
| "logps/rejected": -84.1985092163086, | |
| "loss": 0.0011, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": 0.003594033420085907, | |
| "rewards/margins": 0.005169789306819439, | |
| "rewards/rejected": -0.0015757561195641756, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.61, | |
| "eval_logits/chosen": -2.378091812133789, | |
| "eval_logits/rejected": -2.183748960494995, | |
| "eval_logps/chosen": -280.3858337402344, | |
| "eval_logps/rejected": -258.9316711425781, | |
| "eval_loss": 0.0018177788006141782, | |
| "eval_rewards/accuracies": 0.4814999997615814, | |
| "eval_rewards/chosen": -0.008783474564552307, | |
| "eval_rewards/margins": -0.0010705279419198632, | |
| "eval_rewards/rejected": -0.007712946273386478, | |
| "eval_runtime": 1420.2588, | |
| "eval_samples_per_second": 1.408, | |
| "eval_steps_per_second": 0.704, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "learning_rate": 4.3048902348863116e-06, | |
| "logits/chosen": -2.9989707469940186, | |
| "logits/rejected": -2.8650078773498535, | |
| "logps/chosen": -102.37027740478516, | |
| "logps/rejected": -118.4636001586914, | |
| "loss": 0.0012, | |
| "rewards/accuracies": 0.6000000238418579, | |
| "rewards/chosen": 0.001182993990369141, | |
| "rewards/margins": 0.004763666074723005, | |
| "rewards/rejected": -0.003580672200769186, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "learning_rate": 4.210593920593201e-06, | |
| "logits/chosen": -2.9398374557495117, | |
| "logits/rejected": -2.8116965293884277, | |
| "logps/chosen": -70.69538879394531, | |
| "logps/rejected": -75.53007507324219, | |
| "loss": 0.0013, | |
| "rewards/accuracies": 0.6000000238418579, | |
| "rewards/chosen": 0.005323179066181183, | |
| "rewards/margins": 0.004396263975650072, | |
| "rewards/rejected": 0.0009269150905311108, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "learning_rate": 4.111481441560598e-06, | |
| "logits/chosen": -2.972592830657959, | |
| "logits/rejected": -2.852461099624634, | |
| "logps/chosen": -160.5238037109375, | |
| "logps/rejected": -129.62954711914062, | |
| "loss": 0.0013, | |
| "rewards/accuracies": 0.44999998807907104, | |
| "rewards/chosen": 0.00035965832648798823, | |
| "rewards/margins": 0.005134374834597111, | |
| "rewards/rejected": -0.004774716217070818, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "learning_rate": 4.007831848252212e-06, | |
| "logits/chosen": -3.1746320724487305, | |
| "logits/rejected": -2.9887349605560303, | |
| "logps/chosen": -124.45001220703125, | |
| "logps/rejected": -101.69696807861328, | |
| "loss": 0.0009, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": 0.019443057477474213, | |
| "rewards/margins": 0.012666161172091961, | |
| "rewards/rejected": 0.0067768944427371025, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.76, | |
| "learning_rate": 3.899936965343989e-06, | |
| "logits/chosen": -3.0935516357421875, | |
| "logits/rejected": -3.028534412384033, | |
| "logps/chosen": -131.24337768554688, | |
| "logps/rejected": -116.88172912597656, | |
| "loss": 0.0016, | |
| "rewards/accuracies": 0.44999998807907104, | |
| "rewards/chosen": 0.005977076943963766, | |
| "rewards/margins": 0.005966751836240292, | |
| "rewards/rejected": 1.0324455615773331e-05, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.79, | |
| "learning_rate": 3.7881005700938635e-06, | |
| "logits/chosen": -2.7336764335632324, | |
| "logits/rejected": -2.7675654888153076, | |
| "logps/chosen": -95.82691955566406, | |
| "logps/rejected": -99.57430267333984, | |
| "loss": 0.0008, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": 0.012143732979893684, | |
| "rewards/margins": 0.008911015465855598, | |
| "rewards/rejected": 0.003232717514038086, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.82, | |
| "learning_rate": 3.6726375370590927e-06, | |
| "logits/chosen": -2.959117889404297, | |
| "logits/rejected": -2.841716766357422, | |
| "logps/chosen": -132.79884338378906, | |
| "logps/rejected": -122.66941833496094, | |
| "loss": 0.0015, | |
| "rewards/accuracies": 0.550000011920929, | |
| "rewards/chosen": 0.003037696471437812, | |
| "rewards/margins": 0.00202806293964386, | |
| "rewards/rejected": 0.0010096338810399175, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.85, | |
| "learning_rate": 3.553872951569236e-06, | |
| "logits/chosen": -3.0903332233428955, | |
| "logits/rejected": -3.009970188140869, | |
| "logps/chosen": -90.32659912109375, | |
| "logps/rejected": -92.47020721435547, | |
| "loss": 0.0012, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": 0.005868277512490749, | |
| "rewards/margins": 0.011209800839424133, | |
| "rewards/rejected": -0.005341522395610809, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "learning_rate": 3.432141194450772e-06, | |
| "logits/chosen": -2.920701503753662, | |
| "logits/rejected": -2.721304416656494, | |
| "logps/chosen": -111.44612121582031, | |
| "logps/rejected": -111.91583251953125, | |
| "loss": 0.0007, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.004392694681882858, | |
| "rewards/margins": 0.016653168946504593, | |
| "rewards/rejected": -0.02104586362838745, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.91, | |
| "learning_rate": 3.307785000580313e-06, | |
| "logits/chosen": -2.757887363433838, | |
| "logits/rejected": -2.5684075355529785, | |
| "logps/chosen": -128.5594024658203, | |
| "logps/rejected": -115.97843933105469, | |
| "loss": 0.0015, | |
| "rewards/accuracies": 0.4000000059604645, | |
| "rewards/chosen": -0.006182619370520115, | |
| "rewards/margins": 0.0031226237770169973, | |
| "rewards/rejected": -0.009305243380367756, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.91, | |
| "eval_logits/chosen": -2.374006748199463, | |
| "eval_logits/rejected": -2.1796298027038574, | |
| "eval_logps/chosen": -281.17401123046875, | |
| "eval_logps/rejected": -259.6520690917969, | |
| "eval_loss": 0.0019075373420491815, | |
| "eval_rewards/accuracies": 0.4805000126361847, | |
| "eval_rewards/chosen": -0.016665350645780563, | |
| "eval_rewards/margins": -0.0017483622068539262, | |
| "eval_rewards/rejected": -0.014916987158358097, | |
| "eval_runtime": 1421.116, | |
| "eval_samples_per_second": 1.407, | |
| "eval_steps_per_second": 0.704, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.94, | |
| "learning_rate": 3.1811544939170573e-06, | |
| "logits/chosen": -2.7975335121154785, | |
| "logits/rejected": -2.667598009109497, | |
| "logps/chosen": -161.1660919189453, | |
| "logps/rejected": -133.93838500976562, | |
| "loss": 0.0012, | |
| "rewards/accuracies": 0.44999998807907104, | |
| "rewards/chosen": -0.010928490199148655, | |
| "rewards/margins": 0.00381992245092988, | |
| "rewards/rejected": -0.014748412184417248, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.97, | |
| "learning_rate": 3.052606201731325e-06, | |
| "logits/chosen": -2.93320894241333, | |
| "logits/rejected": -2.7140159606933594, | |
| "logps/chosen": -79.52749633789062, | |
| "logps/rejected": -84.45579528808594, | |
| "loss": 0.0008, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.0026647502090781927, | |
| "rewards/margins": 0.009594769217073917, | |
| "rewards/rejected": -0.01225951872766018, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "learning_rate": 2.9225020508046233e-06, | |
| "logits/chosen": -2.901916742324829, | |
| "logits/rejected": -2.736692190170288, | |
| "logps/chosen": -115.4826431274414, | |
| "logps/rejected": -130.37356567382812, | |
| "loss": 0.0313, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": 0.017747070640325546, | |
| "rewards/margins": 0.05266844481229782, | |
| "rewards/rejected": -0.034921370446681976, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 1.03, | |
| "learning_rate": 2.7912083484274266e-06, | |
| "logits/chosen": -2.7787222862243652, | |
| "logits/rejected": -2.63339900970459, | |
| "logps/chosen": -137.5430450439453, | |
| "logps/rejected": -127.07392883300781, | |
| "loss": 0.0436, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": 0.0036686602979898453, | |
| "rewards/margins": 0.05118006467819214, | |
| "rewards/rejected": -0.04751140996813774, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 1.06, | |
| "learning_rate": 2.659094751063666e-06, | |
| "logits/chosen": -2.9705729484558105, | |
| "logits/rejected": -2.832822322845459, | |
| "logps/chosen": -143.50698852539062, | |
| "logps/rejected": -140.02352905273438, | |
| "loss": 0.0133, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.008256057277321815, | |
| "rewards/margins": 0.05859958380460739, | |
| "rewards/rejected": -0.05034352466464043, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 1.09, | |
| "learning_rate": 2.526533223585641e-06, | |
| "logits/chosen": -2.8163113594055176, | |
| "logits/rejected": -2.697537660598755, | |
| "logps/chosen": -85.76821899414062, | |
| "logps/rejected": -81.71627807617188, | |
| "loss": 0.0073, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": 0.010457667522132397, | |
| "rewards/margins": 0.053769804537296295, | |
| "rewards/rejected": -0.043312136083841324, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 1.12, | |
| "learning_rate": 2.39389699200963e-06, | |
| "logits/chosen": -3.000455141067505, | |
| "logits/rejected": -2.8073432445526123, | |
| "logps/chosen": -107.18000793457031, | |
| "logps/rejected": -118.2176284790039, | |
| "loss": 0.019, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": 0.008137053810060024, | |
| "rewards/margins": 0.07659684866666794, | |
| "rewards/rejected": -0.06845979392528534, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 1.16, | |
| "learning_rate": 2.2615594926807554e-06, | |
| "logits/chosen": -2.9433727264404297, | |
| "logits/rejected": -2.817605495452881, | |
| "logps/chosen": -84.96995544433594, | |
| "logps/rejected": -95.87684631347656, | |
| "loss": 0.0277, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": 0.01651051640510559, | |
| "rewards/margins": 0.06000148132443428, | |
| "rewards/rejected": -0.04349096864461899, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 1.19, | |
| "learning_rate": 2.129893320865672e-06, | |
| "logits/chosen": -3.0510168075561523, | |
| "logits/rejected": -2.938931465148926, | |
| "logps/chosen": -76.3388442993164, | |
| "logps/rejected": -62.002540588378906, | |
| "loss": 0.0302, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": 0.04483187943696976, | |
| "rewards/margins": 0.10971853882074356, | |
| "rewards/rejected": -0.0648866593837738, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 1.22, | |
| "learning_rate": 1.9992691817133025e-06, | |
| "logits/chosen": -3.05735445022583, | |
| "logits/rejected": -2.815657138824463, | |
| "logps/chosen": -251.65219116210938, | |
| "logps/rejected": -255.02658081054688, | |
| "loss": 0.0397, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.03865816444158554, | |
| "rewards/margins": 0.04983866214752197, | |
| "rewards/rejected": -0.08849682658910751, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.22, | |
| "eval_logits/chosen": -2.356828451156616, | |
| "eval_logits/rejected": -2.1631951332092285, | |
| "eval_logps/chosen": -287.29345703125, | |
| "eval_logps/rejected": -264.43231201171875, | |
| "eval_loss": 0.007440177723765373, | |
| "eval_rewards/accuracies": 0.41600000858306885, | |
| "eval_rewards/chosen": -0.0778598040342331, | |
| "eval_rewards/margins": -0.015140472911298275, | |
| "eval_rewards/rejected": -0.06271932274103165, | |
| "eval_runtime": 1421.1528, | |
| "eval_samples_per_second": 1.407, | |
| "eval_steps_per_second": 0.704, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.25, | |
| "learning_rate": 1.8700548465371877e-06, | |
| "logits/chosen": -3.085784435272217, | |
| "logits/rejected": -2.8534493446350098, | |
| "logps/chosen": -150.0517120361328, | |
| "logps/rejected": -130.98390197753906, | |
| "loss": 0.0304, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.026314150542020798, | |
| "rewards/margins": 0.05200232192873955, | |
| "rewards/rejected": -0.07831647992134094, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 1.28, | |
| "learning_rate": 1.742614117358029e-06, | |
| "logits/chosen": -2.8179147243499756, | |
| "logits/rejected": -2.72475528717041, | |
| "logps/chosen": -85.7872543334961, | |
| "logps/rejected": -98.0990219116211, | |
| "loss": 0.0839, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.052498847246170044, | |
| "rewards/margins": 0.009759088046848774, | |
| "rewards/rejected": -0.06225793436169624, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 1.31, | |
| "learning_rate": 1.617305802621748e-06, | |
| "logits/chosen": -2.978959083557129, | |
| "logits/rejected": -2.9247641563415527, | |
| "logps/chosen": -94.34339904785156, | |
| "logps/rejected": -103.36552429199219, | |
| "loss": 0.0221, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": 0.004420995246618986, | |
| "rewards/margins": 0.09026098251342773, | |
| "rewards/rejected": -0.08583998680114746, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 1.34, | |
| "learning_rate": 1.4944827069769125e-06, | |
| "logits/chosen": -2.94343900680542, | |
| "logits/rejected": -2.8201680183410645, | |
| "logps/chosen": -164.69044494628906, | |
| "logps/rejected": -141.93528747558594, | |
| "loss": 0.004, | |
| "rewards/accuracies": 0.550000011920929, | |
| "rewards/chosen": -0.01349672395735979, | |
| "rewards/margins": 0.0265080276876688, | |
| "rewards/rejected": -0.040004752576351166, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 1.37, | |
| "learning_rate": 1.3744906379558165e-06, | |
| "logits/chosen": -2.9248745441436768, | |
| "logits/rejected": -2.9322562217712402, | |
| "logps/chosen": -103.63645935058594, | |
| "logps/rejected": -123.60746765136719, | |
| "loss": 0.0383, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": 0.014714512042701244, | |
| "rewards/margins": 0.11696286499500275, | |
| "rewards/rejected": -0.10224835574626923, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 1.4, | |
| "learning_rate": 1.257667432355893e-06, | |
| "logits/chosen": -2.974544048309326, | |
| "logits/rejected": -2.780351161956787, | |
| "logps/chosen": -76.25318145751953, | |
| "logps/rejected": -95.62451934814453, | |
| "loss": 0.0157, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": 0.0021516154520213604, | |
| "rewards/margins": 0.10200606286525726, | |
| "rewards/rejected": -0.09985444694757462, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 1.43, | |
| "learning_rate": 1.1443420050626624e-06, | |
| "logits/chosen": -3.061354875564575, | |
| "logits/rejected": -2.9399447441101074, | |
| "logps/chosen": -71.61457824707031, | |
| "logps/rejected": -85.53517150878906, | |
| "loss": 0.047, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -0.00521849375218153, | |
| "rewards/margins": 0.06701686233282089, | |
| "rewards/rejected": -0.0722353607416153, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 1.46, | |
| "learning_rate": 1.0348334229922677e-06, | |
| "logits/chosen": -3.0284695625305176, | |
| "logits/rejected": -2.7861952781677246, | |
| "logps/chosen": -84.50325012207031, | |
| "logps/rejected": -80.56702423095703, | |
| "loss": 0.0331, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.010606551542878151, | |
| "rewards/margins": 0.08576520532369614, | |
| "rewards/rejected": -0.07515865564346313, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 1.49, | |
| "learning_rate": 9.294500067608941e-07, | |
| "logits/chosen": -2.929229974746704, | |
| "logits/rejected": -2.727160692214966, | |
| "logps/chosen": -106.28855895996094, | |
| "logps/rejected": -96.00597381591797, | |
| "loss": 0.0183, | |
| "rewards/accuracies": 0.6000000238418579, | |
| "rewards/chosen": -0.032337434589862823, | |
| "rewards/margins": 0.024760348722338676, | |
| "rewards/rejected": -0.05709778517484665, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 1.52, | |
| "learning_rate": 8.284884626103165e-07, | |
| "logits/chosen": -2.9482462406158447, | |
| "logits/rejected": -2.7884809970855713, | |
| "logps/chosen": -77.4437255859375, | |
| "logps/rejected": -83.18536376953125, | |
| "loss": 0.0305, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": 0.025916963815689087, | |
| "rewards/margins": 0.11217900365591049, | |
| "rewards/rejected": -0.086262047290802, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.52, | |
| "eval_logits/chosen": -2.3874802589416504, | |
| "eval_logits/rejected": -2.1901655197143555, | |
| "eval_logps/chosen": -288.4842224121094, | |
| "eval_logps/rejected": -264.83880615234375, | |
| "eval_loss": 0.011713836342096329, | |
| "eval_rewards/accuracies": 0.3944999873638153, | |
| "eval_rewards/chosen": -0.08976726979017258, | |
| "eval_rewards/margins": -0.022982925176620483, | |
| "eval_rewards/rejected": -0.06678435206413269, | |
| "eval_runtime": 1421.6286, | |
| "eval_samples_per_second": 1.407, | |
| "eval_steps_per_second": 0.703, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.55, | |
| "learning_rate": 7.322330470336314e-07, | |
| "logits/chosen": -3.0825541019439697, | |
| "logits/rejected": -2.9925718307495117, | |
| "logps/chosen": -95.14485931396484, | |
| "logps/rejected": -104.95149993896484, | |
| "loss": 0.0318, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": 0.004621284082531929, | |
| "rewards/margins": 0.12566979229450226, | |
| "rewards/rejected": -0.12104851007461548, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 1.58, | |
| "learning_rate": 6.409547664531734e-07, | |
| "logits/chosen": -2.9336235523223877, | |
| "logits/rejected": -2.8529655933380127, | |
| "logps/chosen": -130.2211151123047, | |
| "logps/rejected": -122.72911071777344, | |
| "loss": 0.0063, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": 0.002978086471557617, | |
| "rewards/margins": 0.05790657550096512, | |
| "rewards/rejected": -0.0549284890294075, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 1.61, | |
| "learning_rate": 5.549106142039018e-07, | |
| "logits/chosen": -3.001368522644043, | |
| "logits/rejected": -2.9931020736694336, | |
| "logps/chosen": -122.59478759765625, | |
| "logps/rejected": -139.2528839111328, | |
| "loss": 0.0343, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.03542733192443848, | |
| "rewards/margins": 0.028016597032546997, | |
| "rewards/rejected": -0.06344393640756607, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 1.64, | |
| "learning_rate": 4.743428469705336e-07, | |
| "logits/chosen": -2.895446300506592, | |
| "logits/rejected": -2.805732011795044, | |
| "logps/chosen": -77.88665771484375, | |
| "logps/rejected": -86.38732147216797, | |
| "loss": 0.0152, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": 0.012998471967875957, | |
| "rewards/margins": 0.09006098657846451, | |
| "rewards/rejected": -0.07706250250339508, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 1.67, | |
| "learning_rate": 3.994783027156143e-07, | |
| "logits/chosen": -2.8439111709594727, | |
| "logits/rejected": -2.765376567840576, | |
| "logps/chosen": -124.60064697265625, | |
| "logps/rejected": -115.8946533203125, | |
| "loss": 0.0116, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -0.0016067728865891695, | |
| "rewards/margins": 0.08273078501224518, | |
| "rewards/rejected": -0.08433755487203598, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 1.7, | |
| "learning_rate": 3.3052776201888266e-07, | |
| "logits/chosen": -2.8251261711120605, | |
| "logits/rejected": -2.69478178024292, | |
| "logps/chosen": -85.72721862792969, | |
| "logps/rejected": -84.4536361694336, | |
| "loss": 0.0146, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.007551318965852261, | |
| "rewards/margins": 0.08319269120693207, | |
| "rewards/rejected": -0.07564137130975723, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 1.73, | |
| "learning_rate": 2.676853546260791e-07, | |
| "logits/chosen": -2.8717586994171143, | |
| "logits/rejected": -2.805624485015869, | |
| "logps/chosen": -102.53792572021484, | |
| "logps/rejected": -89.75617980957031, | |
| "loss": 0.0086, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": -0.008890565484762192, | |
| "rewards/margins": 0.04497462511062622, | |
| "rewards/rejected": -0.05386519432067871, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 1.76, | |
| "learning_rate": 2.111280128780638e-07, | |
| "logits/chosen": -3.0598039627075195, | |
| "logits/rejected": -2.9559364318847656, | |
| "logps/chosen": -84.91885375976562, | |
| "logps/rejected": -75.99688720703125, | |
| "loss": 0.0162, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.025450268760323524, | |
| "rewards/margins": 0.03124385140836239, | |
| "rewards/rejected": -0.056694112718105316, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 1.79, | |
| "learning_rate": 1.610149735590949e-07, | |
| "logits/chosen": -2.929579496383667, | |
| "logits/rejected": -2.8060431480407715, | |
| "logps/chosen": -60.950294494628906, | |
| "logps/rejected": -59.30121994018555, | |
| "loss": 0.0299, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": -0.03121623769402504, | |
| "rewards/margins": 0.015263142995536327, | |
| "rewards/rejected": -0.04647938907146454, | |
| "step": 590 | |
| }, | |
| { | |
| "epoch": 1.82, | |
| "learning_rate": 1.1748732956682023e-07, | |
| "logits/chosen": -3.107717275619507, | |
| "logits/rejected": -2.9001567363739014, | |
| "logps/chosen": -90.45118713378906, | |
| "logps/rejected": -84.7816390991211, | |
| "loss": 0.0366, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.021300766617059708, | |
| "rewards/margins": 0.0555335208773613, | |
| "rewards/rejected": -0.0768342837691307, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.82, | |
| "eval_logits/chosen": -2.3873283863067627, | |
| "eval_logits/rejected": -2.1899969577789307, | |
| "eval_logps/chosen": -288.27227783203125, | |
| "eval_logps/rejected": -264.630126953125, | |
| "eval_loss": 0.011489564552903175, | |
| "eval_rewards/accuracies": 0.4000000059604645, | |
| "eval_rewards/chosen": -0.08764798194169998, | |
| "eval_rewards/margins": -0.02295052260160446, | |
| "eval_rewards/rejected": -0.06469745188951492, | |
| "eval_runtime": 1422.4891, | |
| "eval_samples_per_second": 1.406, | |
| "eval_steps_per_second": 0.703, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.85, | |
| "learning_rate": 8.066763266625283e-08, | |
| "logits/chosen": -2.9506607055664062, | |
| "logits/rejected": -2.825108051300049, | |
| "logps/chosen": -100.47017669677734, | |
| "logps/rejected": -127.78892517089844, | |
| "loss": 0.0349, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": 0.004581383429467678, | |
| "rewards/margins": 0.1253204494714737, | |
| "rewards/rejected": -0.12073905766010284, | |
| "step": 610 | |
| }, | |
| { | |
| "epoch": 1.88, | |
| "learning_rate": 5.065954844616722e-08, | |
| "logits/chosen": -2.9829344749450684, | |
| "logits/rejected": -2.7227559089660645, | |
| "logps/chosen": -101.5233154296875, | |
| "logps/rejected": -99.42388916015625, | |
| "loss": 0.027, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.0389946773648262, | |
| "rewards/margins": 0.016298165544867516, | |
| "rewards/rejected": -0.05529283359646797, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 1.91, | |
| "learning_rate": 2.7547564449386666e-08, | |
| "logits/chosen": -2.7666523456573486, | |
| "logits/rejected": -2.719158887863159, | |
| "logps/chosen": -215.42599487304688, | |
| "logps/rejected": -217.2696075439453, | |
| "loss": 0.0308, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.015156986191868782, | |
| "rewards/margins": 0.07782644778490067, | |
| "rewards/rejected": -0.0929834321141243, | |
| "step": 630 | |
| }, | |
| { | |
| "epoch": 1.95, | |
| "learning_rate": 1.1396752298723501e-08, | |
| "logits/chosen": -2.9985458850860596, | |
| "logits/rejected": -2.927014112472534, | |
| "logps/chosen": -109.7343521118164, | |
| "logps/rejected": -111.255859375, | |
| "loss": 0.0344, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": 0.020465310662984848, | |
| "rewards/margins": 0.10245206207036972, | |
| "rewards/rejected": -0.08198676258325577, | |
| "step": 640 | |
| }, | |
| { | |
| "epoch": 1.98, | |
| "learning_rate": 2.252584488296461e-09, | |
| "logits/chosen": -2.9206175804138184, | |
| "logits/rejected": -2.8202061653137207, | |
| "logps/chosen": -157.34161376953125, | |
| "logps/rejected": -171.29029846191406, | |
| "loss": 0.0403, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": -0.005368991754949093, | |
| "rewards/margins": 0.11630947887897491, | |
| "rewards/rejected": -0.12167844921350479, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "step": 658, | |
| "total_flos": 0.0, | |
| "train_loss": 0.014390302914878392, | |
| "train_runtime": 11750.1681, | |
| "train_samples_per_second": 0.112, | |
| "train_steps_per_second": 0.056 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 658, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 2, | |
| "save_steps": 100, | |
| "total_flos": 0.0, | |
| "train_batch_size": 1, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |