| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.9989071038251366, | |
| "eval_steps": 400, | |
| "global_step": 457, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.01092896174863388, | |
| "grad_norm": 48.58164102385424, | |
| "learning_rate": 1.0869565217391303e-07, | |
| "logits/chosen": -1.0108345746994019, | |
| "logits/rejected": -1.005958914756775, | |
| "logps/chosen": -0.2804548144340515, | |
| "logps/rejected": -0.2860378921031952, | |
| "loss": 3.1505, | |
| "rewards/accuracies": 0.53125, | |
| "rewards/chosen": -2.8045480251312256, | |
| "rewards/margins": 0.055831074714660645, | |
| "rewards/rejected": -2.8603789806365967, | |
| "semantic_entropy": 0.7518873810768127, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.02185792349726776, | |
| "grad_norm": 60.09792641017642, | |
| "learning_rate": 2.1739130434782607e-07, | |
| "logits/chosen": -1.0501697063446045, | |
| "logits/rejected": -0.9994386434555054, | |
| "logps/chosen": -0.25685304403305054, | |
| "logps/rejected": -0.27076202630996704, | |
| "loss": 3.1218, | |
| "rewards/accuracies": 0.5375000238418579, | |
| "rewards/chosen": -2.568530559539795, | |
| "rewards/margins": 0.1390899419784546, | |
| "rewards/rejected": -2.707620143890381, | |
| "semantic_entropy": 0.7095814347267151, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.03278688524590164, | |
| "grad_norm": 53.218153749761065, | |
| "learning_rate": 3.260869565217391e-07, | |
| "logits/chosen": -1.010613203048706, | |
| "logits/rejected": -0.964097797870636, | |
| "logps/chosen": -0.2674953043460846, | |
| "logps/rejected": -0.2733650207519531, | |
| "loss": 3.1131, | |
| "rewards/accuracies": 0.512499988079071, | |
| "rewards/chosen": -2.674952983856201, | |
| "rewards/margins": 0.05869739130139351, | |
| "rewards/rejected": -2.7336502075195312, | |
| "semantic_entropy": 0.7275451421737671, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.04371584699453552, | |
| "grad_norm": 67.45687349058657, | |
| "learning_rate": 4.3478260869565214e-07, | |
| "logits/chosen": -0.9450489282608032, | |
| "logits/rejected": -0.8954045176506042, | |
| "logps/chosen": -0.27203652262687683, | |
| "logps/rejected": -0.2849101722240448, | |
| "loss": 3.1483, | |
| "rewards/accuracies": 0.5687500238418579, | |
| "rewards/chosen": -2.720365047454834, | |
| "rewards/margins": 0.1287367343902588, | |
| "rewards/rejected": -2.8491017818450928, | |
| "semantic_entropy": 0.7445512413978577, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.0546448087431694, | |
| "grad_norm": 31.667955833823346, | |
| "learning_rate": 5.434782608695652e-07, | |
| "logits/chosen": -0.9527080655097961, | |
| "logits/rejected": -0.8783634901046753, | |
| "logps/chosen": -0.27410227060317993, | |
| "logps/rejected": -0.29486891627311707, | |
| "loss": 3.1117, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": -2.741022825241089, | |
| "rewards/margins": 0.2076665163040161, | |
| "rewards/rejected": -2.9486892223358154, | |
| "semantic_entropy": 0.754062831401825, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.06557377049180328, | |
| "grad_norm": 49.885531132124974, | |
| "learning_rate": 6.521739130434782e-07, | |
| "logits/chosen": -1.05032479763031, | |
| "logits/rejected": -0.9858748316764832, | |
| "logps/chosen": -0.2675837576389313, | |
| "logps/rejected": -0.28654658794403076, | |
| "loss": 3.1024, | |
| "rewards/accuracies": 0.5375000238418579, | |
| "rewards/chosen": -2.675837278366089, | |
| "rewards/margins": 0.1896287053823471, | |
| "rewards/rejected": -2.8654661178588867, | |
| "semantic_entropy": 0.726620078086853, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.07650273224043716, | |
| "grad_norm": 55.4615534844222, | |
| "learning_rate": 7.608695652173913e-07, | |
| "logits/chosen": -1.0076689720153809, | |
| "logits/rejected": -0.9416168332099915, | |
| "logps/chosen": -0.25796476006507874, | |
| "logps/rejected": -0.2931690514087677, | |
| "loss": 3.058, | |
| "rewards/accuracies": 0.512499988079071, | |
| "rewards/chosen": -2.5796475410461426, | |
| "rewards/margins": 0.3520428538322449, | |
| "rewards/rejected": -2.931690216064453, | |
| "semantic_entropy": 0.7216796875, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.08743169398907104, | |
| "grad_norm": 61.12682886329354, | |
| "learning_rate": 8.695652173913043e-07, | |
| "logits/chosen": -0.9594043493270874, | |
| "logits/rejected": -0.8999654650688171, | |
| "logps/chosen": -0.28498369455337524, | |
| "logps/rejected": -0.31468039751052856, | |
| "loss": 3.1476, | |
| "rewards/accuracies": 0.512499988079071, | |
| "rewards/chosen": -2.849836826324463, | |
| "rewards/margins": 0.296967089176178, | |
| "rewards/rejected": -3.146804094314575, | |
| "semantic_entropy": 0.7657366394996643, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.09836065573770492, | |
| "grad_norm": 30.63183736560673, | |
| "learning_rate": 9.782608695652173e-07, | |
| "logits/chosen": -1.0224157571792603, | |
| "logits/rejected": -0.9429336786270142, | |
| "logps/chosen": -0.29369351267814636, | |
| "logps/rejected": -0.3366171717643738, | |
| "loss": 3.0689, | |
| "rewards/accuracies": 0.518750011920929, | |
| "rewards/chosen": -2.9369354248046875, | |
| "rewards/margins": 0.42923641204833984, | |
| "rewards/rejected": -3.3661720752716064, | |
| "semantic_entropy": 0.7680700421333313, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 0.1092896174863388, | |
| "grad_norm": 66.51868162378878, | |
| "learning_rate": 9.997663088532014e-07, | |
| "logits/chosen": -0.9610111117362976, | |
| "logits/rejected": -0.8836190104484558, | |
| "logps/chosen": -0.3035297095775604, | |
| "logps/rejected": -0.3295886814594269, | |
| "loss": 3.0733, | |
| "rewards/accuracies": 0.4937500059604645, | |
| "rewards/chosen": -3.03529691696167, | |
| "rewards/margins": 0.26059025526046753, | |
| "rewards/rejected": -3.295886993408203, | |
| "semantic_entropy": 0.7657495141029358, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.12021857923497267, | |
| "grad_norm": 36.53428450530809, | |
| "learning_rate": 9.98817312944725e-07, | |
| "logits/chosen": -1.002439260482788, | |
| "logits/rejected": -0.8919259905815125, | |
| "logps/chosen": -0.2904648780822754, | |
| "logps/rejected": -0.3420776426792145, | |
| "loss": 2.9863, | |
| "rewards/accuracies": 0.581250011920929, | |
| "rewards/chosen": -2.904649019241333, | |
| "rewards/margins": 0.5161274671554565, | |
| "rewards/rejected": -3.4207763671875, | |
| "semantic_entropy": 0.7696514129638672, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 0.13114754098360656, | |
| "grad_norm": 31.199928628063066, | |
| "learning_rate": 9.971397915250336e-07, | |
| "logits/chosen": -1.0388410091400146, | |
| "logits/rejected": -0.9951351284980774, | |
| "logps/chosen": -0.27618950605392456, | |
| "logps/rejected": -0.3456563353538513, | |
| "loss": 2.9429, | |
| "rewards/accuracies": 0.6187499761581421, | |
| "rewards/chosen": -2.761894941329956, | |
| "rewards/margins": 0.694668173789978, | |
| "rewards/rejected": -3.4565632343292236, | |
| "semantic_entropy": 0.7672961354255676, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.14207650273224043, | |
| "grad_norm": 43.04608890956751, | |
| "learning_rate": 9.94736194623663e-07, | |
| "logits/chosen": -1.0304734706878662, | |
| "logits/rejected": -0.9656420946121216, | |
| "logps/chosen": -0.3235929608345032, | |
| "logps/rejected": -0.3579636216163635, | |
| "loss": 3.063, | |
| "rewards/accuracies": 0.53125, | |
| "rewards/chosen": -3.2359299659729004, | |
| "rewards/margins": 0.34370657801628113, | |
| "rewards/rejected": -3.579636335372925, | |
| "semantic_entropy": 0.8141781091690063, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 0.15300546448087432, | |
| "grad_norm": 210.46235590585465, | |
| "learning_rate": 9.916100327075037e-07, | |
| "logits/chosen": -0.9860013723373413, | |
| "logits/rejected": -0.9686506986618042, | |
| "logps/chosen": -0.3108309209346771, | |
| "logps/rejected": -0.36498162150382996, | |
| "loss": 2.9623, | |
| "rewards/accuracies": 0.581250011920929, | |
| "rewards/chosen": -3.108308792114258, | |
| "rewards/margins": 0.5415072441101074, | |
| "rewards/rejected": -3.6498160362243652, | |
| "semantic_entropy": 0.7746008038520813, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.16393442622950818, | |
| "grad_norm": 35.07537533113234, | |
| "learning_rate": 9.877658715537428e-07, | |
| "logits/chosen": -0.97590172290802, | |
| "logits/rejected": -0.9608650207519531, | |
| "logps/chosen": -0.32075995206832886, | |
| "logps/rejected": -0.39185312390327454, | |
| "loss": 3.0014, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -3.2076001167297363, | |
| "rewards/margins": 0.7109313011169434, | |
| "rewards/rejected": -3.9185314178466797, | |
| "semantic_entropy": 0.806628406047821, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.17486338797814208, | |
| "grad_norm": 52.85437823609445, | |
| "learning_rate": 9.832093255815216e-07, | |
| "logits/chosen": -0.9724711179733276, | |
| "logits/rejected": -0.9133187532424927, | |
| "logps/chosen": -0.33075493574142456, | |
| "logps/rejected": -0.3698170483112335, | |
| "loss": 2.9391, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -3.307549238204956, | |
| "rewards/margins": 0.39062100648880005, | |
| "rewards/rejected": -3.6981704235076904, | |
| "semantic_entropy": 0.7876138091087341, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.18579234972677597, | |
| "grad_norm": 68.01590434786175, | |
| "learning_rate": 9.779470496520441e-07, | |
| "logits/chosen": -0.973779022693634, | |
| "logits/rejected": -0.9283053278923035, | |
| "logps/chosen": -0.3297511339187622, | |
| "logps/rejected": -0.41490381956100464, | |
| "loss": 2.9392, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -3.297511339187622, | |
| "rewards/margins": 0.8515273928642273, | |
| "rewards/rejected": -4.149038791656494, | |
| "semantic_entropy": 0.8133514523506165, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 0.19672131147540983, | |
| "grad_norm": 43.8207236035398, | |
| "learning_rate": 9.719867293491144e-07, | |
| "logits/chosen": -1.0416749715805054, | |
| "logits/rejected": -0.9649559259414673, | |
| "logps/chosen": -0.3701372444629669, | |
| "logps/rejected": -0.42724722623825073, | |
| "loss": 2.9596, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": -3.7013721466064453, | |
| "rewards/margins": 0.5711001753807068, | |
| "rewards/rejected": -4.272472381591797, | |
| "semantic_entropy": 0.8610862493515015, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.20765027322404372, | |
| "grad_norm": 37.490640629213566, | |
| "learning_rate": 9.653370697542987e-07, | |
| "logits/chosen": -0.9690491557121277, | |
| "logits/rejected": -0.9741093516349792, | |
| "logps/chosen": -0.358903706073761, | |
| "logps/rejected": -0.41135644912719727, | |
| "loss": 2.8808, | |
| "rewards/accuracies": 0.6000000238418579, | |
| "rewards/chosen": -3.5890374183654785, | |
| "rewards/margins": 0.5245276689529419, | |
| "rewards/rejected": -4.113564968109131, | |
| "semantic_entropy": 0.8484892845153809, | |
| "step": 95 | |
| }, | |
| { | |
| "epoch": 0.2185792349726776, | |
| "grad_norm": 72.4012883538114, | |
| "learning_rate": 9.580077827331037e-07, | |
| "logits/chosen": -0.9927606582641602, | |
| "logits/rejected": -0.9488978385925293, | |
| "logps/chosen": -0.4221861958503723, | |
| "logps/rejected": -0.5099042654037476, | |
| "loss": 2.8815, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -4.221861839294434, | |
| "rewards/margins": 0.8771804571151733, | |
| "rewards/rejected": -5.0990424156188965, | |
| "semantic_entropy": 0.891185462474823, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.22950819672131148, | |
| "grad_norm": 53.762702275255066, | |
| "learning_rate": 9.500095727507419e-07, | |
| "logits/chosen": -1.0526697635650635, | |
| "logits/rejected": -1.025564193725586, | |
| "logps/chosen": -0.37911754846572876, | |
| "logps/rejected": -0.46409544348716736, | |
| "loss": 2.8288, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -3.7911758422851562, | |
| "rewards/margins": 0.8497791290283203, | |
| "rewards/rejected": -4.640954971313477, | |
| "semantic_entropy": 0.8762981295585632, | |
| "step": 105 | |
| }, | |
| { | |
| "epoch": 0.24043715846994534, | |
| "grad_norm": 45.04185647458636, | |
| "learning_rate": 9.413541212382004e-07, | |
| "logits/chosen": -1.0478137731552124, | |
| "logits/rejected": -1.028585433959961, | |
| "logps/chosen": -0.4319530427455902, | |
| "logps/rejected": -0.5620400905609131, | |
| "loss": 2.8058, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -4.319530487060547, | |
| "rewards/margins": 1.3008701801300049, | |
| "rewards/rejected": -5.620400428771973, | |
| "semantic_entropy": 0.922514796257019, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.25136612021857924, | |
| "grad_norm": 56.505413378989054, | |
| "learning_rate": 9.320540695314438e-07, | |
| "logits/chosen": -1.06332266330719, | |
| "logits/rejected": -1.0288236141204834, | |
| "logps/chosen": -0.43162402510643005, | |
| "logps/rejected": -0.5995403528213501, | |
| "loss": 2.7772, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -4.3162407875061035, | |
| "rewards/margins": 1.6791629791259766, | |
| "rewards/rejected": -5.995403289794922, | |
| "semantic_entropy": 0.913605809211731, | |
| "step": 115 | |
| }, | |
| { | |
| "epoch": 0.26229508196721313, | |
| "grad_norm": 44.06382434081467, | |
| "learning_rate": 9.221230004086721e-07, | |
| "logits/chosen": -1.0795866250991821, | |
| "logits/rejected": -1.0165742635726929, | |
| "logps/chosen": -0.441800594329834, | |
| "logps/rejected": -0.5526998043060303, | |
| "loss": 2.736, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -4.41800594329834, | |
| "rewards/margins": 1.1089919805526733, | |
| "rewards/rejected": -5.526998519897461, | |
| "semantic_entropy": 0.9433540105819702, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.273224043715847, | |
| "grad_norm": 115.05665224956637, | |
| "learning_rate": 9.11575418252596e-07, | |
| "logits/chosen": -0.9929405450820923, | |
| "logits/rejected": -0.9629238247871399, | |
| "logps/chosen": -0.4976869225502014, | |
| "logps/rejected": -0.6475359201431274, | |
| "loss": 2.7093, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -4.976869106292725, | |
| "rewards/margins": 1.4984896183013916, | |
| "rewards/rejected": -6.4753594398498535, | |
| "semantic_entropy": 0.9429339170455933, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 0.28415300546448086, | |
| "grad_norm": 45.61755920747003, | |
| "learning_rate": 9.004267278667031e-07, | |
| "logits/chosen": -1.0040867328643799, | |
| "logits/rejected": -1.0052454471588135, | |
| "logps/chosen": -0.5251004099845886, | |
| "logps/rejected": -0.7621506452560425, | |
| "loss": 2.7046, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -5.251004219055176, | |
| "rewards/margins": 2.3705029487609863, | |
| "rewards/rejected": -7.621507167816162, | |
| "semantic_entropy": 0.9312794804573059, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.29508196721311475, | |
| "grad_norm": 42.07329597975219, | |
| "learning_rate": 8.886932119764565e-07, | |
| "logits/chosen": -1.0416096448898315, | |
| "logits/rejected": -0.9662833213806152, | |
| "logps/chosen": -0.5473520755767822, | |
| "logps/rejected": -0.7608135938644409, | |
| "loss": 2.5919, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -5.473520755767822, | |
| "rewards/margins": 2.134615898132324, | |
| "rewards/rejected": -7.608136177062988, | |
| "semantic_entropy": 0.9771392941474915, | |
| "step": 135 | |
| }, | |
| { | |
| "epoch": 0.30601092896174864, | |
| "grad_norm": 45.36734996455639, | |
| "learning_rate": 8.763920074482809e-07, | |
| "logits/chosen": -1.0434763431549072, | |
| "logits/rejected": -0.9957412481307983, | |
| "logps/chosen": -0.594224750995636, | |
| "logps/rejected": -0.8605974316596985, | |
| "loss": 2.3803, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -5.94224739074707, | |
| "rewards/margins": 2.663726806640625, | |
| "rewards/rejected": -8.605974197387695, | |
| "semantic_entropy": 0.9883469343185425, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.31693989071038253, | |
| "grad_norm": 57.35860265502167, | |
| "learning_rate": 8.635410802610723e-07, | |
| "logits/chosen": -1.0125925540924072, | |
| "logits/rejected": -1.0003665685653687, | |
| "logps/chosen": -0.6211115121841431, | |
| "logps/rejected": -0.7806649804115295, | |
| "loss": 2.5474, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -6.211114406585693, | |
| "rewards/margins": 1.5955346822738647, | |
| "rewards/rejected": -7.806649208068848, | |
| "semantic_entropy": 0.9819900393486023, | |
| "step": 145 | |
| }, | |
| { | |
| "epoch": 0.32786885245901637, | |
| "grad_norm": 57.36843422854158, | |
| "learning_rate": 8.501591992667849e-07, | |
| "logits/chosen": -1.0663528442382812, | |
| "logits/rejected": -1.0401932001113892, | |
| "logps/chosen": -0.705702543258667, | |
| "logps/rejected": -1.0317673683166504, | |
| "loss": 2.4773, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -7.057024955749512, | |
| "rewards/margins": 3.2606492042541504, | |
| "rewards/rejected": -10.317673683166504, | |
| "semantic_entropy": 0.9794198870658875, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.33879781420765026, | |
| "grad_norm": 55.65247575131584, | |
| "learning_rate": 8.362659087784152e-07, | |
| "logits/chosen": -1.0362324714660645, | |
| "logits/rejected": -0.9899821281433105, | |
| "logps/chosen": -0.7101436257362366, | |
| "logps/rejected": -0.9879515767097473, | |
| "loss": 2.4284, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -7.101436614990234, | |
| "rewards/margins": 2.7780795097351074, | |
| "rewards/rejected": -9.879515647888184, | |
| "semantic_entropy": 0.9602919816970825, | |
| "step": 155 | |
| }, | |
| { | |
| "epoch": 0.34972677595628415, | |
| "grad_norm": 57.58698714096881, | |
| "learning_rate": 8.218815000254231e-07, | |
| "logits/chosen": -1.0922863483428955, | |
| "logits/rejected": -1.0440222024917603, | |
| "logps/chosen": -0.8645855188369751, | |
| "logps/rejected": -1.1068294048309326, | |
| "loss": 2.4302, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -8.645854949951172, | |
| "rewards/margins": 2.4224390983581543, | |
| "rewards/rejected": -11.068293571472168, | |
| "semantic_entropy": 0.9356050491333008, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.36065573770491804, | |
| "grad_norm": 98.49758980833779, | |
| "learning_rate": 8.07026981518276e-07, | |
| "logits/chosen": -1.0802171230316162, | |
| "logits/rejected": -1.0290331840515137, | |
| "logps/chosen": -0.8299033045768738, | |
| "logps/rejected": -1.0872141122817993, | |
| "loss": 2.4163, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -8.299032211303711, | |
| "rewards/margins": 2.573108196258545, | |
| "rewards/rejected": -10.872140884399414, | |
| "semantic_entropy": 0.9756767153739929, | |
| "step": 165 | |
| }, | |
| { | |
| "epoch": 0.37158469945355194, | |
| "grad_norm": 51.034857184899195, | |
| "learning_rate": 7.917240483654e-07, | |
| "logits/chosen": -1.0868977308273315, | |
| "logits/rejected": -1.0327848196029663, | |
| "logps/chosen": -0.8292908668518066, | |
| "logps/rejected": -1.0890681743621826, | |
| "loss": 2.4543, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -8.29290771484375, | |
| "rewards/margins": 2.597774028778076, | |
| "rewards/rejected": -10.890682220458984, | |
| "semantic_entropy": 0.949859619140625, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.3825136612021858, | |
| "grad_norm": 70.18286425215469, | |
| "learning_rate": 7.759950505873521e-07, | |
| "logits/chosen": -1.1532905101776123, | |
| "logits/rejected": -1.1286249160766602, | |
| "logps/chosen": -0.8882781863212585, | |
| "logps/rejected": -1.0833543539047241, | |
| "loss": 2.363, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -8.882781982421875, | |
| "rewards/margins": 1.9507627487182617, | |
| "rewards/rejected": -10.833544731140137, | |
| "semantic_entropy": 0.9320682287216187, | |
| "step": 175 | |
| }, | |
| { | |
| "epoch": 0.39344262295081966, | |
| "grad_norm": 57.51037018197599, | |
| "learning_rate": 7.598629604744872e-07, | |
| "logits/chosen": -1.1318219900131226, | |
| "logits/rejected": -1.1271886825561523, | |
| "logps/chosen": -0.8509464263916016, | |
| "logps/rejected": -1.2260388135910034, | |
| "loss": 2.2789, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -8.509464263916016, | |
| "rewards/margins": 3.7509243488311768, | |
| "rewards/rejected": -12.260388374328613, | |
| "semantic_entropy": 0.9382311105728149, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.40437158469945356, | |
| "grad_norm": 52.68844433987736, | |
| "learning_rate": 7.433513390357989e-07, | |
| "logits/chosen": -1.147756814956665, | |
| "logits/rejected": -1.167338490486145, | |
| "logps/chosen": -0.8701988458633423, | |
| "logps/rejected": -1.2299163341522217, | |
| "loss": 2.2886, | |
| "rewards/accuracies": 0.831250011920929, | |
| "rewards/chosen": -8.701990127563477, | |
| "rewards/margins": 3.5971744060516357, | |
| "rewards/rejected": -12.299162864685059, | |
| "semantic_entropy": 0.9483189582824707, | |
| "step": 185 | |
| }, | |
| { | |
| "epoch": 0.41530054644808745, | |
| "grad_norm": 67.57716327384249, | |
| "learning_rate": 7.264843015879321e-07, | |
| "logits/chosen": -1.156224012374878, | |
| "logits/rejected": -1.115622878074646, | |
| "logps/chosen": -0.8300007581710815, | |
| "logps/rejected": -1.1879067420959473, | |
| "loss": 2.3281, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -8.300007820129395, | |
| "rewards/margins": 3.579059600830078, | |
| "rewards/rejected": -11.879068374633789, | |
| "semantic_entropy": 0.9479382634162903, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.4262295081967213, | |
| "grad_norm": 49.46269476072645, | |
| "learning_rate": 7.092864825346266e-07, | |
| "logits/chosen": -1.1936676502227783, | |
| "logits/rejected": -1.173877477645874, | |
| "logps/chosen": -0.9850988388061523, | |
| "logps/rejected": -1.3840601444244385, | |
| "loss": 2.3467, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -9.850988388061523, | |
| "rewards/margins": 3.9896130561828613, | |
| "rewards/rejected": -13.840600967407227, | |
| "semantic_entropy": 0.9018501043319702, | |
| "step": 195 | |
| }, | |
| { | |
| "epoch": 0.4371584699453552, | |
| "grad_norm": 52.77170794873292, | |
| "learning_rate": 6.917829993880302e-07, | |
| "logits/chosen": -1.1570117473602295, | |
| "logits/rejected": -1.0735323429107666, | |
| "logps/chosen": -0.884504497051239, | |
| "logps/rejected": -1.2432372570037842, | |
| "loss": 2.2946, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": -8.84504508972168, | |
| "rewards/margins": 3.5873265266418457, | |
| "rewards/rejected": -12.432373046875, | |
| "semantic_entropy": 0.9433660507202148, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.44808743169398907, | |
| "grad_norm": 44.7279934671613, | |
| "learning_rate": 6.739994160844309e-07, | |
| "logits/chosen": -1.11527419090271, | |
| "logits/rejected": -1.1322429180145264, | |
| "logps/chosen": -0.8004055023193359, | |
| "logps/rejected": -1.1100412607192993, | |
| "loss": 2.2536, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -8.004055976867676, | |
| "rewards/margins": 3.0963568687438965, | |
| "rewards/rejected": -11.100412368774414, | |
| "semantic_entropy": 0.9715192914009094, | |
| "step": 205 | |
| }, | |
| { | |
| "epoch": 0.45901639344262296, | |
| "grad_norm": 74.86910632781094, | |
| "learning_rate": 6.559617056479827e-07, | |
| "logits/chosen": -1.1598584651947021, | |
| "logits/rejected": -1.1630977392196655, | |
| "logps/chosen": -0.8840648531913757, | |
| "logps/rejected": -1.3052994012832642, | |
| "loss": 2.1909, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -8.840649604797363, | |
| "rewards/margins": 4.212344169616699, | |
| "rewards/rejected": -13.052993774414062, | |
| "semantic_entropy": 0.9134693145751953, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.46994535519125685, | |
| "grad_norm": 75.49182823367542, | |
| "learning_rate": 6.376962122569567e-07, | |
| "logits/chosen": -1.1508004665374756, | |
| "logits/rejected": -1.0994528532028198, | |
| "logps/chosen": -0.8758481740951538, | |
| "logps/rejected": -1.3154475688934326, | |
| "loss": 2.2578, | |
| "rewards/accuracies": 0.831250011920929, | |
| "rewards/chosen": -8.758482933044434, | |
| "rewards/margins": 4.395995140075684, | |
| "rewards/rejected": -13.154478073120117, | |
| "semantic_entropy": 0.9499451518058777, | |
| "step": 215 | |
| }, | |
| { | |
| "epoch": 0.4808743169398907, | |
| "grad_norm": 48.75460798689588, | |
| "learning_rate": 6.192296127679192e-07, | |
| "logits/chosen": -1.216799020767212, | |
| "logits/rejected": -1.1946138143539429, | |
| "logps/chosen": -0.9245316386222839, | |
| "logps/rejected": -1.2553398609161377, | |
| "loss": 2.1888, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -9.245316505432129, | |
| "rewards/margins": 3.3080811500549316, | |
| "rewards/rejected": -12.553396224975586, | |
| "semantic_entropy": 0.920149028301239, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.4918032786885246, | |
| "grad_norm": 51.690790536277476, | |
| "learning_rate": 6.005888777540319e-07, | |
| "logits/chosen": -1.2422759532928467, | |
| "logits/rejected": -1.2019939422607422, | |
| "logps/chosen": -0.8988568186759949, | |
| "logps/rejected": -1.281859278678894, | |
| "loss": 2.1988, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -8.988569259643555, | |
| "rewards/margins": 3.8300251960754395, | |
| "rewards/rejected": -12.81859302520752, | |
| "semantic_entropy": 0.9307478070259094, | |
| "step": 225 | |
| }, | |
| { | |
| "epoch": 0.5027322404371585, | |
| "grad_norm": 54.336949187520304, | |
| "learning_rate": 5.818012321143773e-07, | |
| "logits/chosen": -1.1455602645874023, | |
| "logits/rejected": -1.1488093137741089, | |
| "logps/chosen": -0.8657233119010925, | |
| "logps/rejected": -1.2924095392227173, | |
| "loss": 2.2216, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -8.657234191894531, | |
| "rewards/margins": 4.266862392425537, | |
| "rewards/rejected": -12.924097061157227, | |
| "semantic_entropy": 0.9354921579360962, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.5136612021857924, | |
| "grad_norm": 65.64434177548276, | |
| "learning_rate": 5.628941153118388e-07, | |
| "logits/chosen": -1.158027172088623, | |
| "logits/rejected": -1.1202760934829712, | |
| "logps/chosen": -0.8916556239128113, | |
| "logps/rejected": -1.255821704864502, | |
| "loss": 2.147, | |
| "rewards/accuracies": 0.831250011920929, | |
| "rewards/chosen": -8.916555404663086, | |
| "rewards/margins": 3.641660690307617, | |
| "rewards/rejected": -12.55821704864502, | |
| "semantic_entropy": 0.9296435117721558, | |
| "step": 235 | |
| }, | |
| { | |
| "epoch": 0.5245901639344263, | |
| "grad_norm": 50.4169580491656, | |
| "learning_rate": 5.438951412976098e-07, | |
| "logits/chosen": -1.230419397354126, | |
| "logits/rejected": -1.24024498462677, | |
| "logps/chosen": -0.8406194448471069, | |
| "logps/rejected": -1.2802150249481201, | |
| "loss": 2.048, | |
| "rewards/accuracies": 0.831250011920929, | |
| "rewards/chosen": -8.406194686889648, | |
| "rewards/margins": 4.395954608917236, | |
| "rewards/rejected": -12.802148818969727, | |
| "semantic_entropy": 0.9510146379470825, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.5355191256830601, | |
| "grad_norm": 49.4313854721959, | |
| "learning_rate": 5.248320581808619e-07, | |
| "logits/chosen": -1.1573108434677124, | |
| "logits/rejected": -1.1132813692092896, | |
| "logps/chosen": -0.8796091079711914, | |
| "logps/rejected": -1.3321001529693604, | |
| "loss": 2.1276, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -8.796091079711914, | |
| "rewards/margins": 4.524909973144531, | |
| "rewards/rejected": -13.321001052856445, | |
| "semantic_entropy": 0.9308155179023743, | |
| "step": 245 | |
| }, | |
| { | |
| "epoch": 0.546448087431694, | |
| "grad_norm": 48.57429116569769, | |
| "learning_rate": 5.057327077024744e-07, | |
| "logits/chosen": -1.2085868120193481, | |
| "logits/rejected": -1.1734802722930908, | |
| "logps/chosen": -0.8411601185798645, | |
| "logps/rejected": -1.2031333446502686, | |
| "loss": 2.1896, | |
| "rewards/accuracies": 0.793749988079071, | |
| "rewards/chosen": -8.411601066589355, | |
| "rewards/margins": 3.619731903076172, | |
| "rewards/rejected": -12.031333923339844, | |
| "semantic_entropy": 0.9265835881233215, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.5573770491803278, | |
| "grad_norm": 66.94044733232491, | |
| "learning_rate": 4.866249845720132e-07, | |
| "logits/chosen": -1.1839255094528198, | |
| "logits/rejected": -1.1519018411636353, | |
| "logps/chosen": -0.9079425930976868, | |
| "logps/rejected": -1.3595575094223022, | |
| "loss": 2.1027, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -9.079425811767578, | |
| "rewards/margins": 4.516149044036865, | |
| "rewards/rejected": -13.595575332641602, | |
| "semantic_entropy": 0.9337202310562134, | |
| "step": 255 | |
| }, | |
| { | |
| "epoch": 0.5683060109289617, | |
| "grad_norm": 52.69841931787451, | |
| "learning_rate": 4.675367957273505e-07, | |
| "logits/chosen": -1.1522729396820068, | |
| "logits/rejected": -1.1428637504577637, | |
| "logps/chosen": -0.8813174366950989, | |
| "logps/rejected": -1.3057996034622192, | |
| "loss": 2.093, | |
| "rewards/accuracies": 0.8062499761581421, | |
| "rewards/chosen": -8.813173294067383, | |
| "rewards/margins": 4.2448225021362305, | |
| "rewards/rejected": -13.057995796203613, | |
| "semantic_entropy": 0.9142274856567383, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.5792349726775956, | |
| "grad_norm": 43.72494095619232, | |
| "learning_rate": 4.4849601957642285e-07, | |
| "logits/chosen": -1.1581261157989502, | |
| "logits/rejected": -1.1219886541366577, | |
| "logps/chosen": -0.851395308971405, | |
| "logps/rejected": -1.285305380821228, | |
| "loss": 2.0682, | |
| "rewards/accuracies": 0.8062499761581421, | |
| "rewards/chosen": -8.51395320892334, | |
| "rewards/margins": 4.339099407196045, | |
| "rewards/rejected": -12.853053092956543, | |
| "semantic_entropy": 0.9402335286140442, | |
| "step": 265 | |
| }, | |
| { | |
| "epoch": 0.5901639344262295, | |
| "grad_norm": 74.3004033684429, | |
| "learning_rate": 4.295304652806592e-07, | |
| "logits/chosen": -1.1579556465148926, | |
| "logits/rejected": -1.1343352794647217, | |
| "logps/chosen": -0.8544819951057434, | |
| "logps/rejected": -1.3365771770477295, | |
| "loss": 2.03, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -8.544819831848145, | |
| "rewards/margins": 4.820953369140625, | |
| "rewards/rejected": -13.36577320098877, | |
| "semantic_entropy": 0.9244714975357056, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.6010928961748634, | |
| "grad_norm": 47.25525711691707, | |
| "learning_rate": 4.106678321395433e-07, | |
| "logits/chosen": -1.1198965311050415, | |
| "logits/rejected": -1.0580785274505615, | |
| "logps/chosen": -0.8487468957901001, | |
| "logps/rejected": -1.1780356168746948, | |
| "loss": 2.0436, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -8.487468719482422, | |
| "rewards/margins": 3.2928879261016846, | |
| "rewards/rejected": -11.780357360839844, | |
| "semantic_entropy": 0.9419177174568176, | |
| "step": 275 | |
| }, | |
| { | |
| "epoch": 0.6120218579234973, | |
| "grad_norm": 50.516107327592174, | |
| "learning_rate": 3.9193566913562915e-07, | |
| "logits/chosen": -1.0839662551879883, | |
| "logits/rejected": -1.088769555091858, | |
| "logps/chosen": -0.8575767278671265, | |
| "logps/rejected": -1.3727636337280273, | |
| "loss": 2.0447, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -8.57576847076416, | |
| "rewards/margins": 5.151867866516113, | |
| "rewards/rejected": -13.727636337280273, | |
| "semantic_entropy": 0.9023324251174927, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.6229508196721312, | |
| "grad_norm": 53.0473998547049, | |
| "learning_rate": 3.7336133469909623e-07, | |
| "logits/chosen": -1.2294294834136963, | |
| "logits/rejected": -1.1984007358551025, | |
| "logps/chosen": -0.8300548791885376, | |
| "logps/rejected": -1.328940987586975, | |
| "loss": 1.9766, | |
| "rewards/accuracies": 0.8812500238418579, | |
| "rewards/chosen": -8.300548553466797, | |
| "rewards/margins": 4.988862037658691, | |
| "rewards/rejected": -13.289410591125488, | |
| "semantic_entropy": 0.929417610168457, | |
| "step": 285 | |
| }, | |
| { | |
| "epoch": 0.6338797814207651, | |
| "grad_norm": 90.39553328719175, | |
| "learning_rate": 3.549719567506076e-07, | |
| "logits/chosen": -1.153247594833374, | |
| "logits/rejected": -1.1141259670257568, | |
| "logps/chosen": -0.865193247795105, | |
| "logps/rejected": -1.3289250135421753, | |
| "loss": 2.0316, | |
| "rewards/accuracies": 0.8187500238418579, | |
| "rewards/chosen": -8.651933670043945, | |
| "rewards/margins": 4.6373186111450195, | |
| "rewards/rejected": -13.289251327514648, | |
| "semantic_entropy": 0.9263349771499634, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.644808743169399, | |
| "grad_norm": 55.4386947782287, | |
| "learning_rate": 3.3679439308082774e-07, | |
| "logits/chosen": -1.1450589895248413, | |
| "logits/rejected": -1.1411672830581665, | |
| "logps/chosen": -0.8747571706771851, | |
| "logps/rejected": -1.3667054176330566, | |
| "loss": 1.9546, | |
| "rewards/accuracies": 0.856249988079071, | |
| "rewards/chosen": -8.747570991516113, | |
| "rewards/margins": 4.919483661651611, | |
| "rewards/rejected": -13.66705322265625, | |
| "semantic_entropy": 0.9194013476371765, | |
| "step": 295 | |
| }, | |
| { | |
| "epoch": 0.6557377049180327, | |
| "grad_norm": 45.252881799709705, | |
| "learning_rate": 3.1885519212446716e-07, | |
| "logits/chosen": -1.1887695789337158, | |
| "logits/rejected": -1.1733474731445312, | |
| "logps/chosen": -0.9111166000366211, | |
| "logps/rejected": -1.3687089681625366, | |
| "loss": 1.9751, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -9.111166000366211, | |
| "rewards/margins": 4.575922966003418, | |
| "rewards/rejected": -13.687089920043945, | |
| "semantic_entropy": 0.8905431032180786, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.6666666666666666, | |
| "grad_norm": 63.443681351494455, | |
| "learning_rate": 3.0118055418614295e-07, | |
| "logits/chosen": -1.1734164953231812, | |
| "logits/rejected": -1.1228505373001099, | |
| "logps/chosen": -0.8473879098892212, | |
| "logps/rejected": -1.3491770029067993, | |
| "loss": 2.0108, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": -8.473878860473633, | |
| "rewards/margins": 5.017890453338623, | |
| "rewards/rejected": -13.491769790649414, | |
| "semantic_entropy": 0.9206064343452454, | |
| "step": 305 | |
| }, | |
| { | |
| "epoch": 0.6775956284153005, | |
| "grad_norm": 48.352495630764054, | |
| "learning_rate": 2.83796293174686e-07, | |
| "logits/chosen": -1.117762565612793, | |
| "logits/rejected": -1.1260699033737183, | |
| "logps/chosen": -0.901767909526825, | |
| "logps/rejected": -1.443933129310608, | |
| "loss": 2.0817, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -9.017679214477539, | |
| "rewards/margins": 5.421651840209961, | |
| "rewards/rejected": -14.4393310546875, | |
| "semantic_entropy": 0.9092252850532532, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.6885245901639344, | |
| "grad_norm": 48.159593233413936, | |
| "learning_rate": 2.6672779890178046e-07, | |
| "logits/chosen": -1.1639653444290161, | |
| "logits/rejected": -1.1631653308868408, | |
| "logps/chosen": -0.9492608904838562, | |
| "logps/rejected": -1.3062164783477783, | |
| "loss": 2.0243, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -9.492609977722168, | |
| "rewards/margins": 3.5695548057556152, | |
| "rewards/rejected": -13.062166213989258, | |
| "semantic_entropy": 0.9068315625190735, | |
| "step": 315 | |
| }, | |
| { | |
| "epoch": 0.6994535519125683, | |
| "grad_norm": 44.49675726254248, | |
| "learning_rate": 2.500000000000001e-07, | |
| "logits/chosen": -1.2249935865402222, | |
| "logits/rejected": -1.1779038906097412, | |
| "logps/chosen": -0.9137300252914429, | |
| "logps/rejected": -1.3746672868728638, | |
| "loss": 1.9926, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": -9.137300491333008, | |
| "rewards/margins": 4.609372138977051, | |
| "rewards/rejected": -13.746671676635742, | |
| "semantic_entropy": 0.9218708872795105, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.7103825136612022, | |
| "grad_norm": 62.06584454249009, | |
| "learning_rate": 2.3363732751439923e-07, | |
| "logits/chosen": -1.1833363771438599, | |
| "logits/rejected": -1.1670992374420166, | |
| "logps/chosen": -0.8321934938430786, | |
| "logps/rejected": -1.2726361751556396, | |
| "loss": 2.0087, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": -8.32193374633789, | |
| "rewards/margins": 4.404426097869873, | |
| "rewards/rejected": -12.726360321044922, | |
| "semantic_entropy": 0.9296501278877258, | |
| "step": 325 | |
| }, | |
| { | |
| "epoch": 0.7213114754098361, | |
| "grad_norm": 47.582405569826236, | |
| "learning_rate": 2.1766367922083283e-07, | |
| "logits/chosen": -1.1103484630584717, | |
| "logits/rejected": -1.0834376811981201, | |
| "logps/chosen": -0.7902881503105164, | |
| "logps/rejected": -1.379204511642456, | |
| "loss": 1.9477, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -7.9028825759887695, | |
| "rewards/margins": 5.889164447784424, | |
| "rewards/rejected": -13.792045593261719, | |
| "semantic_entropy": 0.9163787961006165, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.73224043715847, | |
| "grad_norm": 51.055602602949044, | |
| "learning_rate": 2.021023847231202e-07, | |
| "logits/chosen": -1.0874385833740234, | |
| "logits/rejected": -1.0472285747528076, | |
| "logps/chosen": -0.9132916331291199, | |
| "logps/rejected": -1.3749035596847534, | |
| "loss": 1.9593, | |
| "rewards/accuracies": 0.856249988079071, | |
| "rewards/chosen": -9.132916450500488, | |
| "rewards/margins": 4.616118907928467, | |
| "rewards/rejected": -13.749035835266113, | |
| "semantic_entropy": 0.8924514651298523, | |
| "step": 335 | |
| }, | |
| { | |
| "epoch": 0.7431693989071039, | |
| "grad_norm": 53.38188724969566, | |
| "learning_rate": 1.869761713800254e-07, | |
| "logits/chosen": -1.1097468137741089, | |
| "logits/rejected": -1.064247965812683, | |
| "logps/chosen": -0.8977781534194946, | |
| "logps/rejected": -1.4049198627471924, | |
| "loss": 2.0313, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -8.977781295776367, | |
| "rewards/margins": 5.071417808532715, | |
| "rewards/rejected": -14.049200057983398, | |
| "semantic_entropy": 0.896017849445343, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.7540983606557377, | |
| "grad_norm": 57.62309009254558, | |
| "learning_rate": 1.7230713111182164e-07, | |
| "logits/chosen": -1.1679933071136475, | |
| "logits/rejected": -1.167457103729248, | |
| "logps/chosen": -0.919908344745636, | |
| "logps/rejected": -1.4295835494995117, | |
| "loss": 2.0598, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -9.19908332824707, | |
| "rewards/margins": 5.096750736236572, | |
| "rewards/rejected": -14.2958345413208, | |
| "semantic_entropy": 0.8877577781677246, | |
| "step": 345 | |
| }, | |
| { | |
| "epoch": 0.7650273224043715, | |
| "grad_norm": 56.46416368551912, | |
| "learning_rate": 1.5811668813491696e-07, | |
| "logits/chosen": -1.164217233657837, | |
| "logits/rejected": -1.140291452407837, | |
| "logps/chosen": -0.8424995541572571, | |
| "logps/rejected": -1.2386906147003174, | |
| "loss": 1.9725, | |
| "rewards/accuracies": 0.831250011920929, | |
| "rewards/chosen": -8.424995422363281, | |
| "rewards/margins": 3.96191143989563, | |
| "rewards/rejected": -12.386906623840332, | |
| "semantic_entropy": 0.9234525561332703, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.7759562841530054, | |
| "grad_norm": 47.45687983835232, | |
| "learning_rate": 1.4442556767166369e-07, | |
| "logits/chosen": -1.1149598360061646, | |
| "logits/rejected": -1.0830504894256592, | |
| "logps/chosen": -0.8345939517021179, | |
| "logps/rejected": -1.2633658647537231, | |
| "loss": 1.9792, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": -8.345940589904785, | |
| "rewards/margins": 4.287718296051025, | |
| "rewards/rejected": -12.633659362792969, | |
| "semantic_entropy": 0.9200286865234375, | |
| "step": 355 | |
| }, | |
| { | |
| "epoch": 0.7868852459016393, | |
| "grad_norm": 69.3977290180621, | |
| "learning_rate": 1.312537656810549e-07, | |
| "logits/chosen": -1.0737590789794922, | |
| "logits/rejected": -1.0747511386871338, | |
| "logps/chosen": -0.9100669622421265, | |
| "logps/rejected": -1.410347580909729, | |
| "loss": 1.9621, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -9.100671768188477, | |
| "rewards/margins": 5.002806186676025, | |
| "rewards/rejected": -14.103477478027344, | |
| "semantic_entropy": 0.8941181898117065, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.7978142076502732, | |
| "grad_norm": 62.629527673002094, | |
| "learning_rate": 1.1862051965451214e-07, | |
| "logits/chosen": -1.1590697765350342, | |
| "logits/rejected": -1.1595137119293213, | |
| "logps/chosen": -0.8836385011672974, | |
| "logps/rejected": -1.3890182971954346, | |
| "loss": 1.9788, | |
| "rewards/accuracies": 0.793749988079071, | |
| "rewards/chosen": -8.836385726928711, | |
| "rewards/margins": 5.053797245025635, | |
| "rewards/rejected": -13.890182495117188, | |
| "semantic_entropy": 0.9085249900817871, | |
| "step": 365 | |
| }, | |
| { | |
| "epoch": 0.8087431693989071, | |
| "grad_norm": 65.41767650747809, | |
| "learning_rate": 1.0654428051942138e-07, | |
| "logits/chosen": -1.1633613109588623, | |
| "logits/rejected": -1.1319725513458252, | |
| "logps/chosen": -0.9073840975761414, | |
| "logps/rejected": -1.4780160188674927, | |
| "loss": 2.0047, | |
| "rewards/accuracies": 0.8062499761581421, | |
| "rewards/chosen": -9.07384204864502, | |
| "rewards/margins": 5.7063188552856445, | |
| "rewards/rejected": -14.780160903930664, | |
| "semantic_entropy": 0.8961697816848755, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.819672131147541, | |
| "grad_norm": 45.906819053741515, | |
| "learning_rate": 9.504268569144763e-08, | |
| "logits/chosen": -1.1833055019378662, | |
| "logits/rejected": -1.1207568645477295, | |
| "logps/chosen": -0.8244643211364746, | |
| "logps/rejected": -1.3258384466171265, | |
| "loss": 1.9881, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": -8.244643211364746, | |
| "rewards/margins": 5.013741970062256, | |
| "rewards/rejected": -13.258384704589844, | |
| "semantic_entropy": 0.9239455461502075, | |
| "step": 375 | |
| }, | |
| { | |
| "epoch": 0.8306010928961749, | |
| "grad_norm": 41.61552479544314, | |
| "learning_rate": 8.413253331499049e-08, | |
| "logits/chosen": -1.067228078842163, | |
| "logits/rejected": -1.083414077758789, | |
| "logps/chosen": -0.8691232800483704, | |
| "logps/rejected": -1.315148949623108, | |
| "loss": 1.9039, | |
| "rewards/accuracies": 0.8187500238418579, | |
| "rewards/chosen": -8.69123363494873, | |
| "rewards/margins": 4.4602556228637695, | |
| "rewards/rejected": -13.151487350463867, | |
| "semantic_entropy": 0.9291045069694519, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.8415300546448088, | |
| "grad_norm": 53.25699744382843, | |
| "learning_rate": 7.382975772939865e-08, | |
| "logits/chosen": -1.1655666828155518, | |
| "logits/rejected": -1.1497318744659424, | |
| "logps/chosen": -0.9494159817695618, | |
| "logps/rejected": -1.448509931564331, | |
| "loss": 2.061, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": -9.494159698486328, | |
| "rewards/margins": 4.990939617156982, | |
| "rewards/rejected": -14.485099792480469, | |
| "semantic_entropy": 0.9152476191520691, | |
| "step": 385 | |
| }, | |
| { | |
| "epoch": 0.8524590163934426, | |
| "grad_norm": 70.12618215621923, | |
| "learning_rate": 6.414940619677734e-08, | |
| "logits/chosen": -1.157553791999817, | |
| "logits/rejected": -1.1328723430633545, | |
| "logps/chosen": -0.8517138361930847, | |
| "logps/rejected": -1.4659974575042725, | |
| "loss": 1.9241, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": -8.51713752746582, | |
| "rewards/margins": 6.14283561706543, | |
| "rewards/rejected": -14.659975051879883, | |
| "semantic_entropy": 0.9240752458572388, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.8633879781420765, | |
| "grad_norm": 55.71292222513987, | |
| "learning_rate": 5.5105616925376296e-08, | |
| "logits/chosen": -1.134119987487793, | |
| "logits/rejected": -1.1163135766983032, | |
| "logps/chosen": -0.9063531160354614, | |
| "logps/rejected": -1.3211853504180908, | |
| "loss": 1.9119, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -9.063530921936035, | |
| "rewards/margins": 4.148322105407715, | |
| "rewards/rejected": -13.21185302734375, | |
| "semantic_entropy": 0.9249491691589355, | |
| "step": 395 | |
| }, | |
| { | |
| "epoch": 0.8743169398907104, | |
| "grad_norm": 55.00548916953222, | |
| "learning_rate": 4.6711598420656976e-08, | |
| "logits/chosen": -1.0777003765106201, | |
| "logits/rejected": -1.0427886247634888, | |
| "logps/chosen": -0.8970823287963867, | |
| "logps/rejected": -1.4476853609085083, | |
| "loss": 1.9268, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": -8.970823287963867, | |
| "rewards/margins": 5.506030082702637, | |
| "rewards/rejected": -14.476852416992188, | |
| "semantic_entropy": 0.8907070159912109, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.8743169398907104, | |
| "eval_logits/chosen": -1.3707057237625122, | |
| "eval_logits/rejected": -1.3261345624923706, | |
| "eval_logps/chosen": -0.8602119088172913, | |
| "eval_logps/rejected": -1.391231656074524, | |
| "eval_loss": 1.9562536478042603, | |
| "eval_rewards/accuracies": 0.8433734774589539, | |
| "eval_rewards/chosen": -8.602119445800781, | |
| "eval_rewards/margins": 5.31019926071167, | |
| "eval_rewards/rejected": -13.912318229675293, | |
| "eval_runtime": 46.5338, | |
| "eval_samples_per_second": 28.323, | |
| "eval_semantic_entropy": 0.9091164469718933, | |
| "eval_steps_per_second": 1.784, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.8852459016393442, | |
| "grad_norm": 56.49464800000407, | |
| "learning_rate": 3.897961019419516e-08, | |
| "logits/chosen": -1.1149052381515503, | |
| "logits/rejected": -1.0388867855072021, | |
| "logps/chosen": -0.8156052827835083, | |
| "logps/rejected": -1.2854751348495483, | |
| "loss": 1.9256, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -8.156051635742188, | |
| "rewards/margins": 4.698698997497559, | |
| "rewards/rejected": -12.854751586914062, | |
| "semantic_entropy": 0.9291986227035522, | |
| "step": 405 | |
| }, | |
| { | |
| "epoch": 0.8961748633879781, | |
| "grad_norm": 50.962161952757214, | |
| "learning_rate": 3.192094485859526e-08, | |
| "logits/chosen": -1.117832064628601, | |
| "logits/rejected": -1.1457985639572144, | |
| "logps/chosen": -0.8877307772636414, | |
| "logps/rejected": -1.4623702764511108, | |
| "loss": 1.9062, | |
| "rewards/accuracies": 0.856249988079071, | |
| "rewards/chosen": -8.877307891845703, | |
| "rewards/margins": 5.746395587921143, | |
| "rewards/rejected": -14.62370491027832, | |
| "semantic_entropy": 0.8888769149780273, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.907103825136612, | |
| "grad_norm": 54.776769647326134, | |
| "learning_rate": 2.5545911634565265e-08, | |
| "logits/chosen": -1.1549928188323975, | |
| "logits/rejected": -1.1519380807876587, | |
| "logps/chosen": -0.8584114909172058, | |
| "logps/rejected": -1.409808874130249, | |
| "loss": 1.9933, | |
| "rewards/accuracies": 0.856249988079071, | |
| "rewards/chosen": -8.584114074707031, | |
| "rewards/margins": 5.513972282409668, | |
| "rewards/rejected": -14.0980863571167, | |
| "semantic_entropy": 0.9221046566963196, | |
| "step": 415 | |
| }, | |
| { | |
| "epoch": 0.9180327868852459, | |
| "grad_norm": 57.98778672833022, | |
| "learning_rate": 1.9863821294241522e-08, | |
| "logits/chosen": -1.149745225906372, | |
| "logits/rejected": -1.125451683998108, | |
| "logps/chosen": -0.8198378682136536, | |
| "logps/rejected": -1.4263782501220703, | |
| "loss": 1.8964, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": -8.198378562927246, | |
| "rewards/margins": 6.065403938293457, | |
| "rewards/rejected": -14.26378345489502, | |
| "semantic_entropy": 0.9195442199707031, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.9289617486338798, | |
| "grad_norm": 52.73787429786629, | |
| "learning_rate": 1.4882972562753615e-08, | |
| "logits/chosen": -1.1456595659255981, | |
| "logits/rejected": -1.1313962936401367, | |
| "logps/chosen": -0.9214321374893188, | |
| "logps/rejected": -1.5123398303985596, | |
| "loss": 1.99, | |
| "rewards/accuracies": 0.8187500238418579, | |
| "rewards/chosen": -9.21432113647461, | |
| "rewards/margins": 5.90907621383667, | |
| "rewards/rejected": -15.123395919799805, | |
| "semantic_entropy": 0.8927605748176575, | |
| "step": 425 | |
| }, | |
| { | |
| "epoch": 0.9398907103825137, | |
| "grad_norm": 44.1912015437529, | |
| "learning_rate": 1.0610639997888915e-08, | |
| "logits/chosen": -1.081146001815796, | |
| "logits/rejected": -1.075714111328125, | |
| "logps/chosen": -0.7971862554550171, | |
| "logps/rejected": -1.3901129961013794, | |
| "loss": 1.7868, | |
| "rewards/accuracies": 0.9125000238418579, | |
| "rewards/chosen": -7.971863746643066, | |
| "rewards/margins": 5.92926549911499, | |
| "rewards/rejected": -13.901128768920898, | |
| "semantic_entropy": 0.9434274435043335, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 0.9508196721311475, | |
| "grad_norm": 61.224577711409715, | |
| "learning_rate": 7.053063365559997e-09, | |
| "logits/chosen": -1.1544030904769897, | |
| "logits/rejected": -1.1826975345611572, | |
| "logps/chosen": -0.8376800417900085, | |
| "logps/rejected": -1.3729816675186157, | |
| "loss": 1.9375, | |
| "rewards/accuracies": 0.862500011920929, | |
| "rewards/chosen": -8.376801490783691, | |
| "rewards/margins": 5.353014945983887, | |
| "rewards/rejected": -13.729815483093262, | |
| "semantic_entropy": 0.9147855639457703, | |
| "step": 435 | |
| }, | |
| { | |
| "epoch": 0.9617486338797814, | |
| "grad_norm": 53.9815806254924, | |
| "learning_rate": 4.215438526591064e-09, | |
| "logits/chosen": -1.110844612121582, | |
| "logits/rejected": -1.0655571222305298, | |
| "logps/chosen": -0.9438391923904419, | |
| "logps/rejected": -1.4074232578277588, | |
| "loss": 1.8879, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -9.438390731811523, | |
| "rewards/margins": 4.635839939117432, | |
| "rewards/rejected": -14.074231147766113, | |
| "semantic_entropy": 0.879165530204773, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.9726775956284153, | |
| "grad_norm": 57.824670175057065, | |
| "learning_rate": 2.1019098481337426e-09, | |
| "logits/chosen": -1.1616451740264893, | |
| "logits/rejected": -1.1365407705307007, | |
| "logps/chosen": -0.8202412724494934, | |
| "logps/rejected": -1.3376966714859009, | |
| "loss": 1.9265, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -8.202413558959961, | |
| "rewards/margins": 5.174552917480469, | |
| "rewards/rejected": -13.376965522766113, | |
| "semantic_entropy": 0.9334556460380554, | |
| "step": 445 | |
| }, | |
| { | |
| "epoch": 0.9836065573770492, | |
| "grad_norm": 55.05572574163813, | |
| "learning_rate": 7.155641507955445e-10, | |
| "logits/chosen": -1.076629877090454, | |
| "logits/rejected": -1.0707811117172241, | |
| "logps/chosen": -0.882774829864502, | |
| "logps/rejected": -1.3867313861846924, | |
| "loss": 1.9798, | |
| "rewards/accuracies": 0.8187500238418579, | |
| "rewards/chosen": -8.82774829864502, | |
| "rewards/margins": 5.039565563201904, | |
| "rewards/rejected": -13.86731243133545, | |
| "semantic_entropy": 0.910732626914978, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.994535519125683, | |
| "grad_norm": 64.69250426850014, | |
| "learning_rate": 5.842620032053824e-11, | |
| "logits/chosen": -1.0871715545654297, | |
| "logits/rejected": -1.0820033550262451, | |
| "logps/chosen": -0.9286049604415894, | |
| "logps/rejected": -1.3075941801071167, | |
| "loss": 2.053, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": -9.286049842834473, | |
| "rewards/margins": 3.7898917198181152, | |
| "rewards/rejected": -13.07594108581543, | |
| "semantic_entropy": 0.9091150164604187, | |
| "step": 455 | |
| }, | |
| { | |
| "epoch": 0.9989071038251366, | |
| "step": 457, | |
| "total_flos": 0.0, | |
| "train_loss": 2.3507381581112385, | |
| "train_runtime": 7991.7933, | |
| "train_samples_per_second": 7.327, | |
| "train_steps_per_second": 0.057 | |
| } | |
| ], | |
| "logging_steps": 5, | |
| "max_steps": 457, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 1000000, | |
| "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": 2, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |