| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 1.0, | |
| "eval_steps": 500, | |
| "global_step": 1274, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0007849293563579278, | |
| "grad_norm": 15.217127460584273, | |
| "learning_rate": 3.90625e-09, | |
| "logits/chosen": 5881.4375, | |
| "logits/rejected": 2834.66162109375, | |
| "logps/chosen": -257.5969543457031, | |
| "logps/rejected": -120.09489440917969, | |
| "loss": 1.0, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.007849293563579277, | |
| "grad_norm": 16.776894624882384, | |
| "learning_rate": 3.9062499999999997e-08, | |
| "logits/chosen": 4945.97705078125, | |
| "logits/rejected": 4332.49072265625, | |
| "logps/chosen": -248.80799865722656, | |
| "logps/rejected": -214.25498962402344, | |
| "loss": 1.0, | |
| "rewards/accuracies": 0.46296295523643494, | |
| "rewards/chosen": 0.008178697898983955, | |
| "rewards/margins": -0.02892056480050087, | |
| "rewards/rejected": 0.037099260836839676, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.015698587127158554, | |
| "grad_norm": 13.99592115814451, | |
| "learning_rate": 7.812499999999999e-08, | |
| "logits/chosen": 6023.15771484375, | |
| "logits/rejected": 4842.72216796875, | |
| "logps/chosen": -266.2255859375, | |
| "logps/rejected": -232.41552734375, | |
| "loss": 1.0001, | |
| "rewards/accuracies": 0.5416666269302368, | |
| "rewards/chosen": 0.08771798014640808, | |
| "rewards/margins": 0.051151834428310394, | |
| "rewards/rejected": 0.03656615689396858, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.023547880690737835, | |
| "grad_norm": 13.81414691081162, | |
| "learning_rate": 1.1718749999999999e-07, | |
| "logits/chosen": 6032.67578125, | |
| "logits/rejected": 5038.6357421875, | |
| "logps/chosen": -310.4226989746094, | |
| "logps/rejected": -262.90899658203125, | |
| "loss": 0.9997, | |
| "rewards/accuracies": 0.5000000596046448, | |
| "rewards/chosen": 0.18663737177848816, | |
| "rewards/margins": 0.0590377077460289, | |
| "rewards/rejected": 0.12759965658187866, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.03139717425431711, | |
| "grad_norm": 14.509114343975037, | |
| "learning_rate": 1.5624999999999999e-07, | |
| "logits/chosen": 5242.150390625, | |
| "logits/rejected": 4343.94189453125, | |
| "logps/chosen": -263.48541259765625, | |
| "logps/rejected": -224.81298828125, | |
| "loss": 0.9982, | |
| "rewards/accuracies": 0.6583333611488342, | |
| "rewards/chosen": 0.5348619222640991, | |
| "rewards/margins": 0.2838565707206726, | |
| "rewards/rejected": 0.2510053813457489, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.03924646781789639, | |
| "grad_norm": 15.528115845986795, | |
| "learning_rate": 1.9531249999999998e-07, | |
| "logits/chosen": 6364.38134765625, | |
| "logits/rejected": 4975.55712890625, | |
| "logps/chosen": -324.6205749511719, | |
| "logps/rejected": -258.08489990234375, | |
| "loss": 0.9964, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": 1.489025354385376, | |
| "rewards/margins": 1.024138331413269, | |
| "rewards/rejected": 0.46488696336746216, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.04709576138147567, | |
| "grad_norm": 15.841524244504972, | |
| "learning_rate": 2.3437499999999998e-07, | |
| "logits/chosen": 5383.0732421875, | |
| "logits/rejected": 4504.74755859375, | |
| "logps/chosen": -268.93963623046875, | |
| "logps/rejected": -262.5002136230469, | |
| "loss": 0.9928, | |
| "rewards/accuracies": 0.6583333611488342, | |
| "rewards/chosen": 0.9874664545059204, | |
| "rewards/margins": 1.2968491315841675, | |
| "rewards/rejected": -0.30938273668289185, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.054945054945054944, | |
| "grad_norm": 12.412319893385746, | |
| "learning_rate": 2.734375e-07, | |
| "logits/chosen": 5106.55615234375, | |
| "logits/rejected": 4840.5986328125, | |
| "logps/chosen": -232.88815307617188, | |
| "logps/rejected": -234.45919799804688, | |
| "loss": 0.9892, | |
| "rewards/accuracies": 0.6583333611488342, | |
| "rewards/chosen": 0.3698170781135559, | |
| "rewards/margins": 2.176013708114624, | |
| "rewards/rejected": -1.8061968088150024, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.06279434850863422, | |
| "grad_norm": 12.80231150764336, | |
| "learning_rate": 3.1249999999999997e-07, | |
| "logits/chosen": 5693.59423828125, | |
| "logits/rejected": 5163.55810546875, | |
| "logps/chosen": -255.7566375732422, | |
| "logps/rejected": -245.6920166015625, | |
| "loss": 0.9821, | |
| "rewards/accuracies": 0.6583333611488342, | |
| "rewards/chosen": 0.22178272902965546, | |
| "rewards/margins": 3.601436138153076, | |
| "rewards/rejected": -3.379653215408325, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.0706436420722135, | |
| "grad_norm": 14.016716726890994, | |
| "learning_rate": 3.5156249999999997e-07, | |
| "logits/chosen": 5872.6611328125, | |
| "logits/rejected": 5053.22900390625, | |
| "logps/chosen": -282.4061584472656, | |
| "logps/rejected": -248.04946899414062, | |
| "loss": 0.9734, | |
| "rewards/accuracies": 0.7416666746139526, | |
| "rewards/chosen": 2.7849347591400146, | |
| "rewards/margins": 7.367936611175537, | |
| "rewards/rejected": -4.583001136779785, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.07849293563579278, | |
| "grad_norm": 14.456274236246015, | |
| "learning_rate": 3.9062499999999997e-07, | |
| "logits/chosen": 5763.5947265625, | |
| "logits/rejected": 5547.8154296875, | |
| "logps/chosen": -277.15655517578125, | |
| "logps/rejected": -276.28717041015625, | |
| "loss": 0.9719, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": 1.1185591220855713, | |
| "rewards/margins": 9.217432022094727, | |
| "rewards/rejected": -8.098873138427734, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.08634222919937205, | |
| "grad_norm": 15.965712184122198, | |
| "learning_rate": 4.2968749999999996e-07, | |
| "logits/chosen": 6198.5576171875, | |
| "logits/rejected": 5092.955078125, | |
| "logps/chosen": -256.78546142578125, | |
| "logps/rejected": -254.78933715820312, | |
| "loss": 0.9584, | |
| "rewards/accuracies": 0.7166667580604553, | |
| "rewards/chosen": 0.352399080991745, | |
| "rewards/margins": 11.528807640075684, | |
| "rewards/rejected": -11.176408767700195, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.09419152276295134, | |
| "grad_norm": 20.370249209275958, | |
| "learning_rate": 4.6874999999999996e-07, | |
| "logits/chosen": 6045.43408203125, | |
| "logits/rejected": 5060.17822265625, | |
| "logps/chosen": -277.0526428222656, | |
| "logps/rejected": -283.88458251953125, | |
| "loss": 0.9454, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 2.5342214107513428, | |
| "rewards/margins": 16.816547393798828, | |
| "rewards/rejected": -14.282328605651855, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.10204081632653061, | |
| "grad_norm": 14.630467472384158, | |
| "learning_rate": 4.999962424962166e-07, | |
| "logits/chosen": 6011.037109375, | |
| "logits/rejected": 5550.65185546875, | |
| "logps/chosen": -289.7455749511719, | |
| "logps/rejected": -285.1080627441406, | |
| "loss": 0.9568, | |
| "rewards/accuracies": 0.658333420753479, | |
| "rewards/chosen": 0.3616622984409332, | |
| "rewards/margins": 12.038993835449219, | |
| "rewards/rejected": -11.677331924438477, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.10989010989010989, | |
| "grad_norm": 17.081001015208503, | |
| "learning_rate": 4.998647417232375e-07, | |
| "logits/chosen": 5865.60693359375, | |
| "logits/rejected": 5142.06005859375, | |
| "logps/chosen": -261.2616882324219, | |
| "logps/rejected": -265.9956359863281, | |
| "loss": 0.9585, | |
| "rewards/accuracies": 0.6583333015441895, | |
| "rewards/chosen": -1.0194756984710693, | |
| "rewards/margins": 14.62031364440918, | |
| "rewards/rejected": -15.639788627624512, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.11773940345368916, | |
| "grad_norm": 28.23881440069633, | |
| "learning_rate": 4.995454786965036e-07, | |
| "logits/chosen": 5954.14501953125, | |
| "logits/rejected": 4962.21630859375, | |
| "logps/chosen": -278.65301513671875, | |
| "logps/rejected": -264.0430908203125, | |
| "loss": 0.9532, | |
| "rewards/accuracies": 0.6833333969116211, | |
| "rewards/chosen": -0.7810039520263672, | |
| "rewards/margins": 21.372682571411133, | |
| "rewards/rejected": -22.153682708740234, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.12558869701726844, | |
| "grad_norm": 17.275009395044158, | |
| "learning_rate": 4.990386933279972e-07, | |
| "logits/chosen": 5890.1767578125, | |
| "logits/rejected": 5239.57373046875, | |
| "logps/chosen": -276.5395812988281, | |
| "logps/rejected": -283.18988037109375, | |
| "loss": 0.9486, | |
| "rewards/accuracies": 0.7083333730697632, | |
| "rewards/chosen": -2.9753592014312744, | |
| "rewards/margins": 14.08137321472168, | |
| "rewards/rejected": -17.056734085083008, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.13343799058084774, | |
| "grad_norm": 19.7776157223108, | |
| "learning_rate": 4.983447664444096e-07, | |
| "logits/chosen": 6131.412109375, | |
| "logits/rejected": 5464.201171875, | |
| "logps/chosen": -292.5121765136719, | |
| "logps/rejected": -288.33782958984375, | |
| "loss": 0.947, | |
| "rewards/accuracies": 0.658333420753479, | |
| "rewards/chosen": -2.714707136154175, | |
| "rewards/margins": 16.617053985595703, | |
| "rewards/rejected": -19.33176040649414, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.141287284144427, | |
| "grad_norm": 16.27120763569239, | |
| "learning_rate": 4.97464219500968e-07, | |
| "logits/chosen": 5405.88671875, | |
| "logits/rejected": 4714.19677734375, | |
| "logps/chosen": -262.1574401855469, | |
| "logps/rejected": -261.7540283203125, | |
| "loss": 0.9482, | |
| "rewards/accuracies": 0.6416666507720947, | |
| "rewards/chosen": 1.2868962287902832, | |
| "rewards/margins": 15.773916244506836, | |
| "rewards/rejected": -14.487022399902344, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.14913657770800628, | |
| "grad_norm": 19.668914928097266, | |
| "learning_rate": 4.963977141895843e-07, | |
| "logits/chosen": 5469.93505859375, | |
| "logits/rejected": 4702.2255859375, | |
| "logps/chosen": -289.6636047363281, | |
| "logps/rejected": -291.77496337890625, | |
| "loss": 0.9296, | |
| "rewards/accuracies": 0.7499999403953552, | |
| "rewards/chosen": -2.378950595855713, | |
| "rewards/margins": 27.93460464477539, | |
| "rewards/rejected": -30.313556671142578, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.15698587127158556, | |
| "grad_norm": 20.842473477340224, | |
| "learning_rate": 4.951460519416227e-07, | |
| "logits/chosen": 5406.935546875, | |
| "logits/rejected": 4991.7724609375, | |
| "logps/chosen": -249.13339233398438, | |
| "logps/rejected": -282.2489013671875, | |
| "loss": 0.9367, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": 2.116595983505249, | |
| "rewards/margins": 20.891857147216797, | |
| "rewards/rejected": -18.775264739990234, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.16483516483516483, | |
| "grad_norm": 33.15523072801943, | |
| "learning_rate": 4.937101733256606e-07, | |
| "logits/chosen": 4901.9677734375, | |
| "logits/rejected": 4344.95947265625, | |
| "logps/chosen": -224.97476196289062, | |
| "logps/rejected": -243.6951446533203, | |
| "loss": 0.9347, | |
| "rewards/accuracies": 0.6916666030883789, | |
| "rewards/chosen": -0.2618327736854553, | |
| "rewards/margins": 20.972537994384766, | |
| "rewards/rejected": -21.2343692779541, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.1726844583987441, | |
| "grad_norm": 21.17945225425017, | |
| "learning_rate": 4.920911573406924e-07, | |
| "logits/chosen": 5983.14404296875, | |
| "logits/rejected": 5075.22900390625, | |
| "logps/chosen": -287.1895751953125, | |
| "logps/rejected": -258.6339416503906, | |
| "loss": 0.939, | |
| "rewards/accuracies": 0.7000000476837158, | |
| "rewards/chosen": -4.556641101837158, | |
| "rewards/margins": 20.536806106567383, | |
| "rewards/rejected": -25.09344482421875, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.18053375196232338, | |
| "grad_norm": 15.05328515456915, | |
| "learning_rate": 4.902902206053098e-07, | |
| "logits/chosen": 5476.01611328125, | |
| "logits/rejected": 4954.21630859375, | |
| "logps/chosen": -268.1181945800781, | |
| "logps/rejected": -276.8870544433594, | |
| "loss": 0.9401, | |
| "rewards/accuracies": 0.6583333611488342, | |
| "rewards/chosen": 2.1590466499328613, | |
| "rewards/margins": 21.023910522460938, | |
| "rewards/rejected": -18.864864349365234, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.18838304552590268, | |
| "grad_norm": 18.01659982755688, | |
| "learning_rate": 4.883087164434672e-07, | |
| "logits/chosen": 4976.4541015625, | |
| "logits/rejected": 3982.62255859375, | |
| "logps/chosen": -239.6964111328125, | |
| "logps/rejected": -237.77529907226562, | |
| "loss": 0.9329, | |
| "rewards/accuracies": 0.7166666984558105, | |
| "rewards/chosen": -0.002855014754459262, | |
| "rewards/margins": 24.846181869506836, | |
| "rewards/rejected": -24.849035263061523, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.19623233908948196, | |
| "grad_norm": 19.701799187597075, | |
| "learning_rate": 4.861481338675183e-07, | |
| "logits/chosen": 5823.4638671875, | |
| "logits/rejected": 5177.93896484375, | |
| "logps/chosen": -241.93338012695312, | |
| "logps/rejected": -285.5468444824219, | |
| "loss": 0.9424, | |
| "rewards/accuracies": 0.6833333373069763, | |
| "rewards/chosen": -6.285933017730713, | |
| "rewards/margins": 24.868818283081055, | |
| "rewards/rejected": -31.15475082397461, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.20408163265306123, | |
| "grad_norm": 715.9229177986646, | |
| "learning_rate": 4.838100964592904e-07, | |
| "logits/chosen": 5995.43701171875, | |
| "logits/rejected": 4819.33056640625, | |
| "logps/chosen": -283.37677001953125, | |
| "logps/rejected": -258.4510192871094, | |
| "loss": 0.991, | |
| "rewards/accuracies": 0.7416667342185974, | |
| "rewards/chosen": 1.375902771949768, | |
| "rewards/margins": 21.302745819091797, | |
| "rewards/rejected": -19.92684555053711, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.2119309262166405, | |
| "grad_norm": 30.214986071208994, | |
| "learning_rate": 4.812963611500339e-07, | |
| "logits/chosen": 5903.2216796875, | |
| "logits/rejected": 5689.96435546875, | |
| "logps/chosen": -275.84600830078125, | |
| "logps/rejected": -287.48284912109375, | |
| "loss": 0.9176, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -1.337425708770752, | |
| "rewards/margins": 19.31707763671875, | |
| "rewards/rejected": -20.654502868652344, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.21978021978021978, | |
| "grad_norm": 21.071420758528312, | |
| "learning_rate": 4.786088169001671e-07, | |
| "logits/chosen": 5051.4794921875, | |
| "logits/rejected": 4381.0185546875, | |
| "logps/chosen": -228.6282196044922, | |
| "logps/rejected": -264.83636474609375, | |
| "loss": 0.9491, | |
| "rewards/accuracies": 0.7583333253860474, | |
| "rewards/chosen": 0.08709268271923065, | |
| "rewards/margins": 30.51346778869629, | |
| "rewards/rejected": -30.426376342773438, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.22762951334379905, | |
| "grad_norm": 20.756738551757568, | |
| "learning_rate": 4.7574948327980567e-07, | |
| "logits/chosen": 6978.69384765625, | |
| "logits/rejected": 5147.8818359375, | |
| "logps/chosen": -318.50439453125, | |
| "logps/rejected": -284.42498779296875, | |
| "loss": 0.9255, | |
| "rewards/accuracies": 0.7666666507720947, | |
| "rewards/chosen": 6.286134243011475, | |
| "rewards/margins": 32.94474411010742, | |
| "rewards/rejected": -26.65860939025879, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.23547880690737832, | |
| "grad_norm": 24.77524846092873, | |
| "learning_rate": 4.727205089511466e-07, | |
| "logits/chosen": 5091.29833984375, | |
| "logits/rejected": 5012.20361328125, | |
| "logps/chosen": -240.96395874023438, | |
| "logps/rejected": -272.696533203125, | |
| "loss": 0.9404, | |
| "rewards/accuracies": 0.6583333611488342, | |
| "rewards/chosen": -4.223406791687012, | |
| "rewards/margins": 24.546117782592773, | |
| "rewards/rejected": -28.769527435302734, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.24332810047095763, | |
| "grad_norm": 17.46534329970633, | |
| "learning_rate": 4.6952417005384247e-07, | |
| "logits/chosen": 5712.3994140625, | |
| "logits/rejected": 5113.927734375, | |
| "logps/chosen": -251.79891967773438, | |
| "logps/rejected": -256.73321533203125, | |
| "loss": 0.9389, | |
| "rewards/accuracies": 0.6666666865348816, | |
| "rewards/chosen": 3.8696587085723877, | |
| "rewards/margins": 18.93717384338379, | |
| "rewards/rejected": -15.06751537322998, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.25117739403453687, | |
| "grad_norm": 23.489605989743538, | |
| "learning_rate": 4.661628684945851e-07, | |
| "logits/chosen": 5749.4189453125, | |
| "logits/rejected": 4976.5224609375, | |
| "logps/chosen": -272.5289001464844, | |
| "logps/rejected": -300.0706481933594, | |
| "loss": 0.9306, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": 4.758330345153809, | |
| "rewards/margins": 32.90387725830078, | |
| "rewards/rejected": -28.145549774169922, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.25902668759811615, | |
| "grad_norm": 27.939618074872772, | |
| "learning_rate": 4.626391301421782e-07, | |
| "logits/chosen": 5560.0283203125, | |
| "logits/rejected": 5054.7841796875, | |
| "logps/chosen": -270.5875549316406, | |
| "logps/rejected": -273.18206787109375, | |
| "loss": 0.9383, | |
| "rewards/accuracies": 0.6833333969116211, | |
| "rewards/chosen": -0.4894745945930481, | |
| "rewards/margins": 23.300487518310547, | |
| "rewards/rejected": -23.789962768554688, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.2668759811616955, | |
| "grad_norm": 23.064661533697098, | |
| "learning_rate": 4.5895560292945996e-07, | |
| "logits/chosen": 5838.73681640625, | |
| "logits/rejected": 5922.8486328125, | |
| "logps/chosen": -264.6275329589844, | |
| "logps/rejected": -323.1180725097656, | |
| "loss": 0.9292, | |
| "rewards/accuracies": 0.7083333730697632, | |
| "rewards/chosen": 3.4555885791778564, | |
| "rewards/margins": 24.69220542907715, | |
| "rewards/rejected": -21.236614227294922, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.27472527472527475, | |
| "grad_norm": 26.7902596796826, | |
| "learning_rate": 4.5511505486349865e-07, | |
| "logits/chosen": 6143.82421875, | |
| "logits/rejected": 5557.6533203125, | |
| "logps/chosen": -267.94775390625, | |
| "logps/rejected": -314.7910461425781, | |
| "loss": 0.9275, | |
| "rewards/accuracies": 0.7666667103767395, | |
| "rewards/chosen": 4.959996223449707, | |
| "rewards/margins": 36.25249481201172, | |
| "rewards/rejected": -31.29250144958496, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.282574568288854, | |
| "grad_norm": 22.401607581910273, | |
| "learning_rate": 4.5112037194555876e-07, | |
| "logits/chosen": 5670.1220703125, | |
| "logits/rejected": 5525.7197265625, | |
| "logps/chosen": -243.54159545898438, | |
| "logps/rejected": -309.3414001464844, | |
| "loss": 0.9155, | |
| "rewards/accuracies": 0.7500000596046448, | |
| "rewards/chosen": 0.14337244629859924, | |
| "rewards/margins": 35.511756896972656, | |
| "rewards/rejected": -35.36838150024414, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.2904238618524333, | |
| "grad_norm": 25.673874629204995, | |
| "learning_rate": 4.4697455600239863e-07, | |
| "logits/chosen": 5165.1650390625, | |
| "logits/rejected": 4833.2763671875, | |
| "logps/chosen": -255.64859008789062, | |
| "logps/rejected": -257.18560791015625, | |
| "loss": 0.942, | |
| "rewards/accuracies": 0.7166666388511658, | |
| "rewards/chosen": 3.253772735595703, | |
| "rewards/margins": 22.0217342376709, | |
| "rewards/rejected": -18.767959594726562, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.29827315541601257, | |
| "grad_norm": 16.59204251418475, | |
| "learning_rate": 4.426807224305315e-07, | |
| "logits/chosen": 6208.30322265625, | |
| "logits/rejected": 5111.9501953125, | |
| "logps/chosen": -301.24530029296875, | |
| "logps/rejected": -280.9383850097656, | |
| "loss": 0.9236, | |
| "rewards/accuracies": 0.7666666507720947, | |
| "rewards/chosen": 6.731601715087891, | |
| "rewards/margins": 31.761554718017578, | |
| "rewards/rejected": -25.029956817626953, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.30612244897959184, | |
| "grad_norm": 18.775548887189014, | |
| "learning_rate": 4.3824209785514326e-07, | |
| "logits/chosen": 6327.4404296875, | |
| "logits/rejected": 4821.05029296875, | |
| "logps/chosen": -285.92803955078125, | |
| "logps/rejected": -288.6047058105469, | |
| "loss": 0.9076, | |
| "rewards/accuracies": 0.7916667461395264, | |
| "rewards/chosen": 5.598370552062988, | |
| "rewards/margins": 48.711387634277344, | |
| "rewards/rejected": -43.113014221191406, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.3139717425431711, | |
| "grad_norm": 18.55961719941118, | |
| "learning_rate": 4.3366201770542687e-07, | |
| "logits/chosen": 5379.22021484375, | |
| "logits/rejected": 5216.08642578125, | |
| "logps/chosen": -281.50048828125, | |
| "logps/rejected": -329.05059814453125, | |
| "loss": 0.9521, | |
| "rewards/accuracies": 0.7666666507720947, | |
| "rewards/chosen": -6.883020877838135, | |
| "rewards/margins": 35.55707931518555, | |
| "rewards/rejected": -42.440101623535156, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.3218210361067504, | |
| "grad_norm": 27.884929639357004, | |
| "learning_rate": 4.2894392370815567e-07, | |
| "logits/chosen": 5830.39111328125, | |
| "logits/rejected": 5225.48681640625, | |
| "logps/chosen": -288.3466491699219, | |
| "logps/rejected": -321.122314453125, | |
| "loss": 0.8869, | |
| "rewards/accuracies": 0.7833333611488342, | |
| "rewards/chosen": 3.522181749343872, | |
| "rewards/margins": 42.019920349121094, | |
| "rewards/rejected": -38.49774169921875, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.32967032967032966, | |
| "grad_norm": 20.26058977803537, | |
| "learning_rate": 4.2409136130137845e-07, | |
| "logits/chosen": 5500.486328125, | |
| "logits/rejected": 4993.22412109375, | |
| "logps/chosen": -286.73968505859375, | |
| "logps/rejected": -286.5626525878906, | |
| "loss": 0.9178, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.8464605212211609, | |
| "rewards/margins": 32.63377380371094, | |
| "rewards/rejected": -33.4802360534668, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.33751962323390894, | |
| "grad_norm": 23.259074401678372, | |
| "learning_rate": 4.1910797697018017e-07, | |
| "logits/chosen": 5319.9033203125, | |
| "logits/rejected": 4466.1845703125, | |
| "logps/chosen": -253.0988006591797, | |
| "logps/rejected": -259.62548828125, | |
| "loss": 0.9059, | |
| "rewards/accuracies": 0.7750000953674316, | |
| "rewards/chosen": 3.696092128753662, | |
| "rewards/margins": 38.31728744506836, | |
| "rewards/rejected": -34.621192932128906, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 0.3453689167974882, | |
| "grad_norm": 23.635617682728697, | |
| "learning_rate": 4.1399751550651084e-07, | |
| "logits/chosen": 5703.9736328125, | |
| "logits/rejected": 5649.138671875, | |
| "logps/chosen": -251.727783203125, | |
| "logps/rejected": -292.34222412109375, | |
| "loss": 0.9191, | |
| "rewards/accuracies": 0.7083333730697632, | |
| "rewards/chosen": -0.9597722887992859, | |
| "rewards/margins": 31.166919708251953, | |
| "rewards/rejected": -32.12669372558594, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.3532182103610675, | |
| "grad_norm": 19.26341973190091, | |
| "learning_rate": 4.087638171951401e-07, | |
| "logits/chosen": 6566.53125, | |
| "logits/rejected": 4723.1767578125, | |
| "logps/chosen": -286.6239318847656, | |
| "logps/rejected": -269.2875671386719, | |
| "loss": 0.9225, | |
| "rewards/accuracies": 0.7500000596046448, | |
| "rewards/chosen": 0.43297329545021057, | |
| "rewards/margins": 43.72136306762695, | |
| "rewards/rejected": -43.28838348388672, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.36106750392464676, | |
| "grad_norm": 24.15018784175595, | |
| "learning_rate": 4.034108149278543e-07, | |
| "logits/chosen": 6809.2626953125, | |
| "logits/rejected": 5260.38818359375, | |
| "logps/chosen": -328.55633544921875, | |
| "logps/rejected": -295.267578125, | |
| "loss": 0.9139, | |
| "rewards/accuracies": 0.8166667222976685, | |
| "rewards/chosen": 4.154780864715576, | |
| "rewards/margins": 34.77088928222656, | |
| "rewards/rejected": -30.616104125976562, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 0.36891679748822603, | |
| "grad_norm": 19.930362113206062, | |
| "learning_rate": 3.979425312480629e-07, | |
| "logits/chosen": 5824.78857421875, | |
| "logits/rejected": 5113.4462890625, | |
| "logps/chosen": -287.6418762207031, | |
| "logps/rejected": -312.6920471191406, | |
| "loss": 0.916, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": 0.1438978612422943, | |
| "rewards/margins": 34.73716735839844, | |
| "rewards/rejected": -34.5932731628418, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 0.37676609105180536, | |
| "grad_norm": 17.674222833195227, | |
| "learning_rate": 3.923630753280357e-07, | |
| "logits/chosen": 6285.4208984375, | |
| "logits/rejected": 5395.0048828125, | |
| "logps/chosen": -281.23046875, | |
| "logps/rejected": -289.922119140625, | |
| "loss": 0.91, | |
| "rewards/accuracies": 0.7666666507720947, | |
| "rewards/chosen": -1.4238802194595337, | |
| "rewards/margins": 41.704959869384766, | |
| "rewards/rejected": -43.128841400146484, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 0.38461538461538464, | |
| "grad_norm": 19.792313593032937, | |
| "learning_rate": 3.866766398810424e-07, | |
| "logits/chosen": 5899.3759765625, | |
| "logits/rejected": 5611.10400390625, | |
| "logps/chosen": -232.70321655273438, | |
| "logps/rejected": -309.1105651855469, | |
| "loss": 0.8918, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": 8.237863540649414, | |
| "rewards/margins": 44.7490348815918, | |
| "rewards/rejected": -36.511173248291016, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 0.3924646781789639, | |
| "grad_norm": 19.90809886153986, | |
| "learning_rate": 3.8088749801071496e-07, | |
| "logits/chosen": 6366.142578125, | |
| "logits/rejected": 4952.5068359375, | |
| "logps/chosen": -303.6125793457031, | |
| "logps/rejected": -313.29315185546875, | |
| "loss": 0.912, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -5.123102188110352, | |
| "rewards/margins": 43.737327575683594, | |
| "rewards/rejected": -48.86042785644531, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.4003139717425432, | |
| "grad_norm": 15.360327222439452, | |
| "learning_rate": 3.75e-07, | |
| "logits/chosen": 5147.7861328125, | |
| "logits/rejected": 4536.16455078125, | |
| "logps/chosen": -252.94473266601562, | |
| "logps/rejected": -273.7191162109375, | |
| "loss": 0.911, | |
| "rewards/accuracies": 0.7500001192092896, | |
| "rewards/chosen": 3.915085554122925, | |
| "rewards/margins": 48.51652526855469, | |
| "rewards/rejected": -44.601436614990234, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 0.40816326530612246, | |
| "grad_norm": 22.726045624861126, | |
| "learning_rate": 3.6901857004211443e-07, | |
| "logits/chosen": 5440.7919921875, | |
| "logits/rejected": 5028.6923828125, | |
| "logps/chosen": -270.27691650390625, | |
| "logps/rejected": -308.1585693359375, | |
| "loss": 0.9365, | |
| "rewards/accuracies": 0.6833332777023315, | |
| "rewards/chosen": -4.1040358543396, | |
| "rewards/margins": 34.928199768066406, | |
| "rewards/rejected": -39.03223419189453, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 0.41601255886970173, | |
| "grad_norm": 22.1271600593506, | |
| "learning_rate": 3.6294770291596076e-07, | |
| "logits/chosen": 6135.3037109375, | |
| "logits/rejected": 5045.927734375, | |
| "logps/chosen": -278.6271057128906, | |
| "logps/rejected": -283.05340576171875, | |
| "loss": 0.9054, | |
| "rewards/accuracies": 0.7666667699813843, | |
| "rewards/chosen": 5.665005683898926, | |
| "rewards/margins": 24.647602081298828, | |
| "rewards/rejected": -18.98259925842285, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 0.423861852433281, | |
| "grad_norm": 41.55555763430692, | |
| "learning_rate": 3.5679196060850034e-07, | |
| "logits/chosen": 5851.7666015625, | |
| "logits/rejected": 5219.24609375, | |
| "logps/chosen": -277.853271484375, | |
| "logps/rejected": -295.34912109375, | |
| "loss": 0.9161, | |
| "rewards/accuracies": 0.7333333492279053, | |
| "rewards/chosen": -0.46962302923202515, | |
| "rewards/margins": 39.48213577270508, | |
| "rewards/rejected": -39.951759338378906, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 0.4317111459968603, | |
| "grad_norm": 20.618281611990163, | |
| "learning_rate": 3.505559688866229e-07, | |
| "logits/chosen": 5663.544921875, | |
| "logits/rejected": 5269.53955078125, | |
| "logps/chosen": -283.8811340332031, | |
| "logps/rejected": -334.2888488769531, | |
| "loss": 0.9078, | |
| "rewards/accuracies": 0.7083333730697632, | |
| "rewards/chosen": -2.544511556625366, | |
| "rewards/margins": 40.56104278564453, | |
| "rewards/rejected": -43.10555648803711, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 0.43956043956043955, | |
| "grad_norm": 19.391598225192343, | |
| "learning_rate": 3.4424441382108826e-07, | |
| "logits/chosen": 5656.9873046875, | |
| "logits/rejected": 5293.77392578125, | |
| "logps/chosen": -279.07427978515625, | |
| "logps/rejected": -302.99371337890625, | |
| "loss": 0.935, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -2.946714162826538, | |
| "rewards/margins": 33.268821716308594, | |
| "rewards/rejected": -36.215538024902344, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 0.4474097331240188, | |
| "grad_norm": 27.492150331098347, | |
| "learning_rate": 3.378620382651523e-07, | |
| "logits/chosen": 6007.7392578125, | |
| "logits/rejected": 5555.06689453125, | |
| "logps/chosen": -315.08062744140625, | |
| "logps/rejected": -331.8717956542969, | |
| "loss": 0.9104, | |
| "rewards/accuracies": 0.7333333492279053, | |
| "rewards/chosen": -0.5156325101852417, | |
| "rewards/margins": 39.41523742675781, | |
| "rewards/rejected": -39.930870056152344, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 0.4552590266875981, | |
| "grad_norm": 24.81865808929877, | |
| "learning_rate": 3.314136382905234e-07, | |
| "logits/chosen": 5973.07958984375, | |
| "logits/rejected": 5394.1357421875, | |
| "logps/chosen": -284.0478515625, | |
| "logps/rejected": -333.41473388671875, | |
| "loss": 0.9205, | |
| "rewards/accuracies": 0.7916666865348816, | |
| "rewards/chosen": 3.1989758014678955, | |
| "rewards/margins": 51.7633056640625, | |
| "rewards/rejected": -48.564327239990234, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 0.4631083202511774, | |
| "grad_norm": 25.774824989224687, | |
| "learning_rate": 3.249040595833274e-07, | |
| "logits/chosen": 6467.61328125, | |
| "logits/rejected": 5461.263671875, | |
| "logps/chosen": -303.5272216796875, | |
| "logps/rejected": -285.9855041503906, | |
| "loss": 0.9112, | |
| "rewards/accuracies": 0.7666667103767395, | |
| "rewards/chosen": 2.6262047290802, | |
| "rewards/margins": 42.771522521972656, | |
| "rewards/rejected": -40.14531707763672, | |
| "step": 590 | |
| }, | |
| { | |
| "epoch": 0.47095761381475665, | |
| "grad_norm": 27.67394815870794, | |
| "learning_rate": 3.1833819380279023e-07, | |
| "logits/chosen": 6110.16455078125, | |
| "logits/rejected": 5210.61865234375, | |
| "logps/chosen": -247.8615264892578, | |
| "logps/rejected": -286.87371826171875, | |
| "loss": 0.9008, | |
| "rewards/accuracies": 0.8083332777023315, | |
| "rewards/chosen": 4.629504203796387, | |
| "rewards/margins": 36.539493560791016, | |
| "rewards/rejected": -31.909988403320312, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.478806907378336, | |
| "grad_norm": 20.718470452802894, | |
| "learning_rate": 3.11720974905373e-07, | |
| "logits/chosen": 5875.37109375, | |
| "logits/rejected": 5125.77099609375, | |
| "logps/chosen": -269.7516784667969, | |
| "logps/rejected": -293.9992370605469, | |
| "loss": 0.8922, | |
| "rewards/accuracies": 0.7666666507720947, | |
| "rewards/chosen": 3.0898051261901855, | |
| "rewards/margins": 46.59734344482422, | |
| "rewards/rejected": -43.50753402709961, | |
| "step": 610 | |
| }, | |
| { | |
| "epoch": 0.48665620094191525, | |
| "grad_norm": 20.359862853723154, | |
| "learning_rate": 3.0505737543712275e-07, | |
| "logits/chosen": 4997.51416015625, | |
| "logits/rejected": 4120.1025390625, | |
| "logps/chosen": -248.70364379882812, | |
| "logps/rejected": -253.5068817138672, | |
| "loss": 0.9011, | |
| "rewards/accuracies": 0.7833333015441895, | |
| "rewards/chosen": -0.8777009844779968, | |
| "rewards/margins": 39.05192947387695, | |
| "rewards/rejected": -39.929630279541016, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 0.4945054945054945, | |
| "grad_norm": 24.410845424528254, | |
| "learning_rate": 2.9835240279702513e-07, | |
| "logits/chosen": 6503.3544921875, | |
| "logits/rejected": 5521.38134765625, | |
| "logps/chosen": -311.4924621582031, | |
| "logps/rejected": -310.57275390625, | |
| "loss": 0.8996, | |
| "rewards/accuracies": 0.7583333849906921, | |
| "rewards/chosen": 5.219082832336426, | |
| "rewards/margins": 49.75696563720703, | |
| "rewards/rejected": -44.53789138793945, | |
| "step": 630 | |
| }, | |
| { | |
| "epoch": 0.5023547880690737, | |
| "grad_norm": 20.26453021100001, | |
| "learning_rate": 2.9161109547416667e-07, | |
| "logits/chosen": 6199.2998046875, | |
| "logits/rejected": 5321.87744140625, | |
| "logps/chosen": -277.9691162109375, | |
| "logps/rejected": -303.0850524902344, | |
| "loss": 0.9023, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": 0.7213143110275269, | |
| "rewards/margins": 25.9655704498291, | |
| "rewards/rejected": -25.24425506591797, | |
| "step": 640 | |
| }, | |
| { | |
| "epoch": 0.5102040816326531, | |
| "grad_norm": 25.607967393691553, | |
| "learning_rate": 2.848385192615339e-07, | |
| "logits/chosen": 5318.2783203125, | |
| "logits/rejected": 4340.18603515625, | |
| "logps/chosen": -256.1814270019531, | |
| "logps/rejected": -262.76361083984375, | |
| "loss": 0.9022, | |
| "rewards/accuracies": 0.7666667103767395, | |
| "rewards/chosen": 1.3570995330810547, | |
| "rewards/margins": 37.01885223388672, | |
| "rewards/rejected": -35.66175842285156, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 0.5180533751962323, | |
| "grad_norm": 28.57773283940273, | |
| "learning_rate": 2.780397634492949e-07, | |
| "logits/chosen": 5933.3408203125, | |
| "logits/rejected": 4759.22509765625, | |
| "logps/chosen": -289.2817077636719, | |
| "logps/rejected": -310.09234619140625, | |
| "loss": 0.9164, | |
| "rewards/accuracies": 0.8166667222976685, | |
| "rewards/chosen": 1.0365993976593018, | |
| "rewards/margins": 58.85808181762695, | |
| "rewards/rejected": -57.82147979736328, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 0.5259026687598116, | |
| "grad_norm": 23.773604822565275, | |
| "learning_rate": 2.71219937000424e-07, | |
| "logits/chosen": 5936.36865234375, | |
| "logits/rejected": 4900.83056640625, | |
| "logps/chosen": -274.3346862792969, | |
| "logps/rejected": -285.41607666015625, | |
| "loss": 0.9161, | |
| "rewards/accuracies": 0.7166666388511658, | |
| "rewards/chosen": 1.8758052587509155, | |
| "rewards/margins": 36.20291519165039, | |
| "rewards/rejected": -34.327110290527344, | |
| "step": 670 | |
| }, | |
| { | |
| "epoch": 0.533751962323391, | |
| "grad_norm": 23.006395031938606, | |
| "learning_rate": 2.6438416471154273e-07, | |
| "logits/chosen": 5816.591796875, | |
| "logits/rejected": 4883.8837890625, | |
| "logps/chosen": -274.1582946777344, | |
| "logps/rejected": -274.8831481933594, | |
| "loss": 0.9049, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": 3.6695594787597656, | |
| "rewards/margins": 43.54711151123047, | |
| "rewards/rejected": -39.87754821777344, | |
| "step": 680 | |
| }, | |
| { | |
| "epoch": 0.5416012558869702, | |
| "grad_norm": 26.525872068105123, | |
| "learning_rate": 2.5753758336186326e-07, | |
| "logits/chosen": 5734.0380859375, | |
| "logits/rejected": 5293.50341796875, | |
| "logps/chosen": -276.3390197753906, | |
| "logps/rejected": -324.45611572265625, | |
| "loss": 0.8992, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -1.2499253749847412, | |
| "rewards/margins": 45.74732971191406, | |
| "rewards/rejected": -46.997257232666016, | |
| "step": 690 | |
| }, | |
| { | |
| "epoch": 0.5494505494505495, | |
| "grad_norm": 26.541363934518053, | |
| "learning_rate": 2.5068533785312666e-07, | |
| "logits/chosen": 5495.10595703125, | |
| "logits/rejected": 5287.18896484375, | |
| "logps/chosen": -251.01681518554688, | |
| "logps/rejected": -289.4714660644531, | |
| "loss": 0.9318, | |
| "rewards/accuracies": 0.7166666388511658, | |
| "rewards/chosen": -0.9005786776542664, | |
| "rewards/margins": 43.11927032470703, | |
| "rewards/rejected": -44.01984405517578, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.5572998430141287, | |
| "grad_norm": 24.46002490835382, | |
| "learning_rate": 2.4383257734343794e-07, | |
| "logits/chosen": 5464.05712890625, | |
| "logits/rejected": 5455.001953125, | |
| "logps/chosen": -250.55587768554688, | |
| "logps/rejected": -300.8583679199219, | |
| "loss": 0.9064, | |
| "rewards/accuracies": 0.6916667222976685, | |
| "rewards/chosen": -1.3100314140319824, | |
| "rewards/margins": 35.45912551879883, | |
| "rewards/rejected": -36.76915740966797, | |
| "step": 710 | |
| }, | |
| { | |
| "epoch": 0.565149136577708, | |
| "grad_norm": 21.147359750230834, | |
| "learning_rate": 2.3698445137790258e-07, | |
| "logits/chosen": 5864.6064453125, | |
| "logits/rejected": 5073.22607421875, | |
| "logps/chosen": -277.67425537109375, | |
| "logps/rejected": -293.50439453125, | |
| "loss": 0.9079, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": 7.0904951095581055, | |
| "rewards/margins": 42.4764289855957, | |
| "rewards/rejected": -35.385929107666016, | |
| "step": 720 | |
| }, | |
| { | |
| "epoch": 0.5729984301412873, | |
| "grad_norm": 22.688465462752504, | |
| "learning_rate": 2.3014610601897157e-07, | |
| "logits/chosen": 6338.6025390625, | |
| "logits/rejected": 4882.97021484375, | |
| "logps/chosen": -295.96942138671875, | |
| "logps/rejected": -287.09844970703125, | |
| "loss": 0.8906, | |
| "rewards/accuracies": 0.7416667342185974, | |
| "rewards/chosen": -0.9103342294692993, | |
| "rewards/margins": 45.903358459472656, | |
| "rewards/rejected": -46.81369400024414, | |
| "step": 730 | |
| }, | |
| { | |
| "epoch": 0.5808477237048666, | |
| "grad_norm": 24.937006724806338, | |
| "learning_rate": 2.2332267997940513e-07, | |
| "logits/chosen": 5288.89111328125, | |
| "logits/rejected": 4503.5400390625, | |
| "logps/chosen": -243.6168975830078, | |
| "logps/rejected": -258.78863525390625, | |
| "loss": 0.8925, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": 4.391143321990967, | |
| "rewards/margins": 45.62685775756836, | |
| "rewards/rejected": -41.2357177734375, | |
| "step": 740 | |
| }, | |
| { | |
| "epoch": 0.5886970172684458, | |
| "grad_norm": 25.087199062301373, | |
| "learning_rate": 2.1651930076075723e-07, | |
| "logits/chosen": 5745.19140625, | |
| "logits/rejected": 5214.74365234375, | |
| "logps/chosen": -238.671875, | |
| "logps/rejected": -253.86264038085938, | |
| "loss": 0.9096, | |
| "rewards/accuracies": 0.7083333730697632, | |
| "rewards/chosen": 1.5903170108795166, | |
| "rewards/margins": 32.90300750732422, | |
| "rewards/rejected": -31.31268882751465, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 0.5965463108320251, | |
| "grad_norm": 22.6004674707504, | |
| "learning_rate": 2.0974108080028692e-07, | |
| "logits/chosen": 6041.4853515625, | |
| "logits/rejected": 4772.58740234375, | |
| "logps/chosen": -259.30780029296875, | |
| "logps/rejected": -272.39886474609375, | |
| "loss": 0.9056, | |
| "rewards/accuracies": 0.7166666984558105, | |
| "rewards/chosen": 4.05280065536499, | |
| "rewards/margins": 38.531856536865234, | |
| "rewards/rejected": -34.47905731201172, | |
| "step": 760 | |
| }, | |
| { | |
| "epoch": 0.6043956043956044, | |
| "grad_norm": 32.07154020812047, | |
| "learning_rate": 2.0299311362918773e-07, | |
| "logits/chosen": 6262.0263671875, | |
| "logits/rejected": 5383.724609375, | |
| "logps/chosen": -292.2076110839844, | |
| "logps/rejected": -327.34906005859375, | |
| "loss": 0.9143, | |
| "rewards/accuracies": 0.7083333730697632, | |
| "rewards/chosen": 2.43733549118042, | |
| "rewards/margins": 44.98815155029297, | |
| "rewards/rejected": -42.550819396972656, | |
| "step": 770 | |
| }, | |
| { | |
| "epoch": 0.6122448979591837, | |
| "grad_norm": 23.504239493179618, | |
| "learning_rate": 1.962804700450265e-07, | |
| "logits/chosen": 6072.6123046875, | |
| "logits/rejected": 5769.93896484375, | |
| "logps/chosen": -277.5558776855469, | |
| "logps/rejected": -344.66473388671875, | |
| "loss": 0.9245, | |
| "rewards/accuracies": 0.7583333849906921, | |
| "rewards/chosen": 5.5767011642456055, | |
| "rewards/margins": 35.678287506103516, | |
| "rewards/rejected": -30.10158348083496, | |
| "step": 780 | |
| }, | |
| { | |
| "epoch": 0.6200941915227629, | |
| "grad_norm": 23.391365758751537, | |
| "learning_rate": 1.8960819430126334e-07, | |
| "logits/chosen": 5679.32666015625, | |
| "logits/rejected": 5026.5771484375, | |
| "logps/chosen": -254.663330078125, | |
| "logps/rejected": -304.33038330078125, | |
| "loss": 0.8986, | |
| "rewards/accuracies": 0.7666666507720947, | |
| "rewards/chosen": -0.2119230329990387, | |
| "rewards/margins": 48.772735595703125, | |
| "rewards/rejected": -48.98466110229492, | |
| "step": 790 | |
| }, | |
| { | |
| "epoch": 0.6279434850863422, | |
| "grad_norm": 24.446664309226747, | |
| "learning_rate": 1.8298130031671972e-07, | |
| "logits/chosen": 5691.87255859375, | |
| "logits/rejected": 4992.9775390625, | |
| "logps/chosen": -277.36175537109375, | |
| "logps/rejected": -307.2684020996094, | |
| "loss": 0.9219, | |
| "rewards/accuracies": 0.7583333253860474, | |
| "rewards/chosen": -3.6252803802490234, | |
| "rewards/margins": 37.125953674316406, | |
| "rewards/rejected": -40.75123596191406, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.6357927786499215, | |
| "grad_norm": 24.78783701861456, | |
| "learning_rate": 1.7640476790784075e-07, | |
| "logits/chosen": 5191.9833984375, | |
| "logits/rejected": 4733.341796875, | |
| "logps/chosen": -273.46575927734375, | |
| "logps/rejected": -316.88641357421875, | |
| "loss": 0.9032, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": 0.3641335070133209, | |
| "rewards/margins": 43.47520446777344, | |
| "rewards/rejected": -43.11106872558594, | |
| "step": 810 | |
| }, | |
| { | |
| "epoch": 0.6436420722135008, | |
| "grad_norm": 22.35233438036302, | |
| "learning_rate": 1.6988353904658492e-07, | |
| "logits/chosen": 5694.0908203125, | |
| "logits/rejected": 4441.0947265625, | |
| "logps/chosen": -278.25262451171875, | |
| "logps/rejected": -253.1114044189453, | |
| "loss": 0.8889, | |
| "rewards/accuracies": 0.8083332777023315, | |
| "rewards/chosen": 4.820112228393555, | |
| "rewards/margins": 42.1064453125, | |
| "rewards/rejected": -37.28633499145508, | |
| "step": 820 | |
| }, | |
| { | |
| "epoch": 0.6514913657770801, | |
| "grad_norm": 33.971535716829386, | |
| "learning_rate": 1.634225141467513e-07, | |
| "logits/chosen": 5630.9619140625, | |
| "logits/rejected": 5059.73388671875, | |
| "logps/chosen": -262.0240173339844, | |
| "logps/rejected": -293.10284423828125, | |
| "loss": 0.9036, | |
| "rewards/accuracies": 0.7916666865348816, | |
| "rewards/chosen": 1.4304834604263306, | |
| "rewards/margins": 46.748069763183594, | |
| "rewards/rejected": -45.317588806152344, | |
| "step": 830 | |
| }, | |
| { | |
| "epoch": 0.6593406593406593, | |
| "grad_norm": 22.058744826701787, | |
| "learning_rate": 1.570265483815364e-07, | |
| "logits/chosen": 6187.8076171875, | |
| "logits/rejected": 5085.29052734375, | |
| "logps/chosen": -284.2825927734375, | |
| "logps/rejected": -312.46990966796875, | |
| "loss": 0.8997, | |
| "rewards/accuracies": 0.7166666984558105, | |
| "rewards/chosen": 1.250636339187622, | |
| "rewards/margins": 39.52533721923828, | |
| "rewards/rejected": -38.27470016479492, | |
| "step": 840 | |
| }, | |
| { | |
| "epoch": 0.6671899529042387, | |
| "grad_norm": 24.57030833624564, | |
| "learning_rate": 1.5070044803508691e-07, | |
| "logits/chosen": 5731.578125, | |
| "logits/rejected": 5200.16650390625, | |
| "logps/chosen": -273.11700439453125, | |
| "logps/rejected": -295.6126403808594, | |
| "loss": 0.8764, | |
| "rewards/accuracies": 0.7333332896232605, | |
| "rewards/chosen": -0.5748471021652222, | |
| "rewards/margins": 49.07981872558594, | |
| "rewards/rejected": -49.65467071533203, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 0.6750392464678179, | |
| "grad_norm": 29.643861873247765, | |
| "learning_rate": 1.444489668907914e-07, | |
| "logits/chosen": 6211.44091796875, | |
| "logits/rejected": 5296.041015625, | |
| "logps/chosen": -299.6576232910156, | |
| "logps/rejected": -300.2407531738281, | |
| "loss": 0.9124, | |
| "rewards/accuracies": 0.7666666507720947, | |
| "rewards/chosen": 3.2107672691345215, | |
| "rewards/margins": 49.146392822265625, | |
| "rewards/rejected": -45.93562698364258, | |
| "step": 860 | |
| }, | |
| { | |
| "epoch": 0.6828885400313972, | |
| "grad_norm": 19.18443294400448, | |
| "learning_rate": 1.3827680265902232e-07, | |
| "logits/chosen": 6146.9755859375, | |
| "logits/rejected": 5104.4638671875, | |
| "logps/chosen": -284.4742736816406, | |
| "logps/rejected": -296.0386962890625, | |
| "loss": 0.8972, | |
| "rewards/accuracies": 0.7416666746139526, | |
| "rewards/chosen": 4.938233375549316, | |
| "rewards/margins": 46.767799377441406, | |
| "rewards/rejected": -41.829566955566406, | |
| "step": 870 | |
| }, | |
| { | |
| "epoch": 0.6907378335949764, | |
| "grad_norm": 48.40221110161389, | |
| "learning_rate": 1.3218859344701632e-07, | |
| "logits/chosen": 5446.45849609375, | |
| "logits/rejected": 5177.0341796875, | |
| "logps/chosen": -252.7174835205078, | |
| "logps/rejected": -326.0343322753906, | |
| "loss": 0.9097, | |
| "rewards/accuracies": 0.8083333969116211, | |
| "rewards/chosen": 6.035229206085205, | |
| "rewards/margins": 43.76191711425781, | |
| "rewards/rejected": -37.726688385009766, | |
| "step": 880 | |
| }, | |
| { | |
| "epoch": 0.6985871271585558, | |
| "grad_norm": 27.2644482399342, | |
| "learning_rate": 1.2618891427354172e-07, | |
| "logits/chosen": 6387.37939453125, | |
| "logits/rejected": 5229.87353515625, | |
| "logps/chosen": -305.6217346191406, | |
| "logps/rejected": -304.56756591796875, | |
| "loss": 0.9068, | |
| "rewards/accuracies": 0.7833333015441895, | |
| "rewards/chosen": 6.906777858734131, | |
| "rewards/margins": 49.44459915161133, | |
| "rewards/rejected": -42.537818908691406, | |
| "step": 890 | |
| }, | |
| { | |
| "epoch": 0.706436420722135, | |
| "grad_norm": 22.953604786244295, | |
| "learning_rate": 1.202822736309758e-07, | |
| "logits/chosen": 5421.3154296875, | |
| "logits/rejected": 5001.1025390625, | |
| "logps/chosen": -253.27780151367188, | |
| "logps/rejected": -305.64788818359375, | |
| "loss": 0.91, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": 4.763638019561768, | |
| "rewards/margins": 45.9899787902832, | |
| "rewards/rejected": -41.226341247558594, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.7142857142857143, | |
| "grad_norm": 27.91344694605715, | |
| "learning_rate": 1.1447311009737299e-07, | |
| "logits/chosen": 5335.6162109375, | |
| "logits/rejected": 5072.3359375, | |
| "logps/chosen": -258.45941162109375, | |
| "logps/rejected": -303.97113037109375, | |
| "loss": 0.9052, | |
| "rewards/accuracies": 0.7083333730697632, | |
| "rewards/chosen": -0.8920276761054993, | |
| "rewards/margins": 45.10370635986328, | |
| "rewards/rejected": -45.9957389831543, | |
| "step": 910 | |
| }, | |
| { | |
| "epoch": 0.7221350078492935, | |
| "grad_norm": 23.26853190795506, | |
| "learning_rate": 1.0876578900107053e-07, | |
| "logits/chosen": 5871.88720703125, | |
| "logits/rejected": 4929.97900390625, | |
| "logps/chosen": -283.12799072265625, | |
| "logps/rejected": -281.83013916015625, | |
| "loss": 0.9041, | |
| "rewards/accuracies": 0.7916666865348816, | |
| "rewards/chosen": 3.5937552452087402, | |
| "rewards/margins": 40.470191955566406, | |
| "rewards/rejected": -36.876434326171875, | |
| "step": 920 | |
| }, | |
| { | |
| "epoch": 0.7299843014128728, | |
| "grad_norm": 19.12047364944922, | |
| "learning_rate": 1.0316459914033793e-07, | |
| "logits/chosen": 5825.18115234375, | |
| "logits/rejected": 4393.8466796875, | |
| "logps/chosen": -289.2558898925781, | |
| "logps/rejected": -272.1807861328125, | |
| "loss": 0.9068, | |
| "rewards/accuracies": 0.7500000596046448, | |
| "rewards/chosen": -1.3023111820220947, | |
| "rewards/margins": 37.931583404541016, | |
| "rewards/rejected": -39.233890533447266, | |
| "step": 930 | |
| }, | |
| { | |
| "epoch": 0.7378335949764521, | |
| "grad_norm": 20.651437164552735, | |
| "learning_rate": 9.767374956053584e-08, | |
| "logits/chosen": 5656.6083984375, | |
| "logits/rejected": 4979.91943359375, | |
| "logps/chosen": -261.6556396484375, | |
| "logps/rejected": -294.62158203125, | |
| "loss": 0.9014, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": 3.210376024246216, | |
| "rewards/margins": 50.671321868896484, | |
| "rewards/rejected": -47.46094512939453, | |
| "step": 940 | |
| }, | |
| { | |
| "epoch": 0.7456828885400314, | |
| "grad_norm": 21.44755274256472, | |
| "learning_rate": 9.229736639120561e-08, | |
| "logits/chosen": 5831.8720703125, | |
| "logits/rejected": 5376.9619140625, | |
| "logps/chosen": -261.1865539550781, | |
| "logps/rejected": -289.8360900878906, | |
| "loss": 0.9157, | |
| "rewards/accuracies": 0.7083333730697632, | |
| "rewards/chosen": 2.373589277267456, | |
| "rewards/margins": 29.416553497314453, | |
| "rewards/rejected": -27.042964935302734, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 0.7535321821036107, | |
| "grad_norm": 34.123365725576235, | |
| "learning_rate": 8.70394897454659e-08, | |
| "logits/chosen": 5702.736328125, | |
| "logits/rejected": 5105.47802734375, | |
| "logps/chosen": -254.7130126953125, | |
| "logps/rejected": -283.2959289550781, | |
| "loss": 0.8878, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": 9.300559997558594, | |
| "rewards/margins": 46.2738151550293, | |
| "rewards/rejected": -36.97325897216797, | |
| "step": 960 | |
| }, | |
| { | |
| "epoch": 0.7613814756671899, | |
| "grad_norm": 30.293238410586564, | |
| "learning_rate": 8.19040706840472e-08, | |
| "logits/chosen": 5806.1142578125, | |
| "logits/rejected": 4896.1630859375, | |
| "logps/chosen": -283.5286560058594, | |
| "logps/rejected": -292.33721923828125, | |
| "loss": 0.885, | |
| "rewards/accuracies": 0.7500000596046448, | |
| "rewards/chosen": 5.602731227874756, | |
| "rewards/margins": 51.605499267578125, | |
| "rewards/rejected": -46.002769470214844, | |
| "step": 970 | |
| }, | |
| { | |
| "epoch": 0.7692307692307693, | |
| "grad_norm": 23.491592765262904, | |
| "learning_rate": 7.689496824624525e-08, | |
| "logits/chosen": 5516.03369140625, | |
| "logits/rejected": 4427.40185546875, | |
| "logps/chosen": -270.09564208984375, | |
| "logps/rejected": -297.0606689453125, | |
| "loss": 0.8801, | |
| "rewards/accuracies": 0.7916666269302368, | |
| "rewards/chosen": 6.5463409423828125, | |
| "rewards/margins": 64.91987609863281, | |
| "rewards/rejected": -58.37353515625, | |
| "step": 980 | |
| }, | |
| { | |
| "epoch": 0.7770800627943485, | |
| "grad_norm": 36.05397286248155, | |
| "learning_rate": 7.201594655002458e-08, | |
| "logits/chosen": 5858.2021484375, | |
| "logits/rejected": 4895.6455078125, | |
| "logps/chosen": -267.0971984863281, | |
| "logps/rejected": -288.00775146484375, | |
| "loss": 0.8929, | |
| "rewards/accuracies": 0.7666667103767395, | |
| "rewards/chosen": 1.010213851928711, | |
| "rewards/margins": 52.7333869934082, | |
| "rewards/rejected": -51.723175048828125, | |
| "step": 990 | |
| }, | |
| { | |
| "epoch": 0.7849293563579278, | |
| "grad_norm": 35.53234033842205, | |
| "learning_rate": 6.727067196345099e-08, | |
| "logits/chosen": 5515.2802734375, | |
| "logits/rejected": 4669.0283203125, | |
| "logps/chosen": -259.7113952636719, | |
| "logps/rejected": -264.55084228515625, | |
| "loss": 0.9068, | |
| "rewards/accuracies": 0.73333340883255, | |
| "rewards/chosen": 0.6101600527763367, | |
| "rewards/margins": 39.48523712158203, | |
| "rewards/rejected": -38.875083923339844, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.792778649921507, | |
| "grad_norm": 25.616049738852727, | |
| "learning_rate": 6.26627103495786e-08, | |
| "logits/chosen": 5715.3486328125, | |
| "logits/rejected": 4768.5458984375, | |
| "logps/chosen": -254.9511260986328, | |
| "logps/rejected": -275.54779052734375, | |
| "loss": 0.9084, | |
| "rewards/accuracies": 0.7166666388511658, | |
| "rewards/chosen": 4.464234828948975, | |
| "rewards/margins": 44.496917724609375, | |
| "rewards/rejected": -40.032676696777344, | |
| "step": 1010 | |
| }, | |
| { | |
| "epoch": 0.8006279434850864, | |
| "grad_norm": 22.502131941962762, | |
| "learning_rate": 5.8195524386862374e-08, | |
| "logits/chosen": 5797.1455078125, | |
| "logits/rejected": 5151.40771484375, | |
| "logps/chosen": -287.3981018066406, | |
| "logps/rejected": -309.9508361816406, | |
| "loss": 0.902, | |
| "rewards/accuracies": 0.7916666269302368, | |
| "rewards/chosen": 10.498518943786621, | |
| "rewards/margins": 46.848793029785156, | |
| "rewards/rejected": -36.35027313232422, | |
| "step": 1020 | |
| }, | |
| { | |
| "epoch": 0.8084772370486656, | |
| "grad_norm": 20.045575816512876, | |
| "learning_rate": 5.38724709671092e-08, | |
| "logits/chosen": 6191.720703125, | |
| "logits/rejected": 5863.8681640625, | |
| "logps/chosen": -272.2727966308594, | |
| "logps/rejected": -315.62384033203125, | |
| "loss": 0.8982, | |
| "rewards/accuracies": 0.7500000596046448, | |
| "rewards/chosen": 5.527390480041504, | |
| "rewards/margins": 46.04096603393555, | |
| "rewards/rejected": -40.513572692871094, | |
| "step": 1030 | |
| }, | |
| { | |
| "epoch": 0.8163265306122449, | |
| "grad_norm": 25.302444772287263, | |
| "learning_rate": 4.969679867292276e-08, | |
| "logits/chosen": 5513.0185546875, | |
| "logits/rejected": 5018.1357421875, | |
| "logps/chosen": -264.01837158203125, | |
| "logps/rejected": -301.9166564941406, | |
| "loss": 0.8943, | |
| "rewards/accuracies": 0.7166666984558105, | |
| "rewards/chosen": 1.22486412525177, | |
| "rewards/margins": 49.567771911621094, | |
| "rewards/rejected": -48.34290313720703, | |
| "step": 1040 | |
| }, | |
| { | |
| "epoch": 0.8241758241758241, | |
| "grad_norm": 26.971621245905506, | |
| "learning_rate": 4.5671645336537416e-08, | |
| "logits/chosen": 5564.62939453125, | |
| "logits/rejected": 5052.43994140625, | |
| "logps/chosen": -279.4435119628906, | |
| "logps/rejected": -315.63885498046875, | |
| "loss": 0.8897, | |
| "rewards/accuracies": 0.7333333492279053, | |
| "rewards/chosen": 2.089627504348755, | |
| "rewards/margins": 52.004425048828125, | |
| "rewards/rejected": -49.914798736572266, | |
| "step": 1050 | |
| }, | |
| { | |
| "epoch": 0.8320251177394035, | |
| "grad_norm": 33.760753801014786, | |
| "learning_rate": 4.180003568187776e-08, | |
| "logits/chosen": 6866.97265625, | |
| "logits/rejected": 5416.5302734375, | |
| "logps/chosen": -308.72735595703125, | |
| "logps/rejected": -301.6792907714844, | |
| "loss": 0.9172, | |
| "rewards/accuracies": 0.6833333373069763, | |
| "rewards/chosen": -0.8046684265136719, | |
| "rewards/margins": 36.33965301513672, | |
| "rewards/rejected": -37.14432144165039, | |
| "step": 1060 | |
| }, | |
| { | |
| "epoch": 0.8398744113029827, | |
| "grad_norm": 42.68076165248159, | |
| "learning_rate": 3.8084879051612144e-08, | |
| "logits/chosen": 5713.95068359375, | |
| "logits/rejected": 5230.85009765625, | |
| "logps/chosen": -266.6412048339844, | |
| "logps/rejected": -279.85150146484375, | |
| "loss": 0.8971, | |
| "rewards/accuracies": 0.73333340883255, | |
| "rewards/chosen": 6.2511186599731445, | |
| "rewards/margins": 48.867332458496094, | |
| "rewards/rejected": -42.6162109375, | |
| "step": 1070 | |
| }, | |
| { | |
| "epoch": 0.847723704866562, | |
| "grad_norm": 26.287897922478123, | |
| "learning_rate": 3.452896722091128e-08, | |
| "logits/chosen": 6276.97119140625, | |
| "logits/rejected": 4863.9853515625, | |
| "logps/chosen": -304.9378967285156, | |
| "logps/rejected": -293.0082092285156, | |
| "loss": 0.8843, | |
| "rewards/accuracies": 0.8250001072883606, | |
| "rewards/chosen": 7.549924373626709, | |
| "rewards/margins": 59.31702423095703, | |
| "rewards/rejected": -51.76709747314453, | |
| "step": 1080 | |
| }, | |
| { | |
| "epoch": 0.8555729984301413, | |
| "grad_norm": 28.443568850519753, | |
| "learning_rate": 3.11349722995527e-08, | |
| "logits/chosen": 6319.14599609375, | |
| "logits/rejected": 4778.30322265625, | |
| "logps/chosen": -271.9593505859375, | |
| "logps/rejected": -292.22821044921875, | |
| "loss": 0.9029, | |
| "rewards/accuracies": 0.7166666984558105, | |
| "rewards/chosen": 1.8383331298828125, | |
| "rewards/margins": 44.397727966308594, | |
| "rewards/rejected": -42.55939865112305, | |
| "step": 1090 | |
| }, | |
| { | |
| "epoch": 0.8634222919937206, | |
| "grad_norm": 32.28685337112104, | |
| "learning_rate": 2.7905444723949762e-08, | |
| "logits/chosen": 6111.0732421875, | |
| "logits/rejected": 5072.91796875, | |
| "logps/chosen": -282.01324462890625, | |
| "logps/rejected": -285.5963134765625, | |
| "loss": 0.908, | |
| "rewards/accuracies": 0.7833333611488342, | |
| "rewards/chosen": 1.209522008895874, | |
| "rewards/margins": 54.696800231933594, | |
| "rewards/rejected": -53.487281799316406, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 0.8712715855572999, | |
| "grad_norm": 25.46638694280487, | |
| "learning_rate": 2.484281134061142e-08, | |
| "logits/chosen": 6478.18701171875, | |
| "logits/rejected": 5250.2666015625, | |
| "logps/chosen": -308.4444580078125, | |
| "logps/rejected": -311.63800048828125, | |
| "loss": 0.9019, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": 0.8299369812011719, | |
| "rewards/margins": 49.747154235839844, | |
| "rewards/rejected": -48.917213439941406, | |
| "step": 1110 | |
| }, | |
| { | |
| "epoch": 0.8791208791208791, | |
| "grad_norm": 33.894472996102174, | |
| "learning_rate": 2.194937358247506e-08, | |
| "logits/chosen": 6344.24462890625, | |
| "logits/rejected": 5178.0888671875, | |
| "logps/chosen": -291.0325622558594, | |
| "logps/rejected": -310.59222412109375, | |
| "loss": 0.89, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": 0.3055509626865387, | |
| "rewards/margins": 51.41431427001953, | |
| "rewards/rejected": -51.108768463134766, | |
| "step": 1120 | |
| }, | |
| { | |
| "epoch": 0.8869701726844584, | |
| "grad_norm": 35.09428216666554, | |
| "learning_rate": 1.9227305739481612e-08, | |
| "logits/chosen": 5788.7275390625, | |
| "logits/rejected": 4550.9697265625, | |
| "logps/chosen": -270.097412109375, | |
| "logps/rejected": -266.88616943359375, | |
| "loss": 0.8938, | |
| "rewards/accuracies": 0.7333332896232605, | |
| "rewards/chosen": 8.472616195678711, | |
| "rewards/margins": 50.75660705566406, | |
| "rewards/rejected": -42.28398895263672, | |
| "step": 1130 | |
| }, | |
| { | |
| "epoch": 0.8948194662480377, | |
| "grad_norm": 24.4757646274536, | |
| "learning_rate": 1.6678653324693787e-08, | |
| "logits/chosen": 6356.0703125, | |
| "logits/rejected": 5166.41259765625, | |
| "logps/chosen": -299.87261962890625, | |
| "logps/rejected": -303.05859375, | |
| "loss": 0.8891, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": 5.566819667816162, | |
| "rewards/margins": 47.01850128173828, | |
| "rewards/rejected": -41.45167922973633, | |
| "step": 1140 | |
| }, | |
| { | |
| "epoch": 0.902668759811617, | |
| "grad_norm": 24.046707084849317, | |
| "learning_rate": 1.4305331537183384e-08, | |
| "logits/chosen": 5655.9697265625, | |
| "logits/rejected": 5174.4541015625, | |
| "logps/chosen": -265.05810546875, | |
| "logps/rejected": -304.6141357421875, | |
| "loss": 0.8896, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": 2.137265682220459, | |
| "rewards/margins": 43.28087615966797, | |
| "rewards/rejected": -41.14360809326172, | |
| "step": 1150 | |
| }, | |
| { | |
| "epoch": 0.9105180533751962, | |
| "grad_norm": 23.734757870906822, | |
| "learning_rate": 1.2109123822844653e-08, | |
| "logits/chosen": 5815.59423828125, | |
| "logits/rejected": 4604.97900390625, | |
| "logps/chosen": -269.80438232421875, | |
| "logps/rejected": -275.7255859375, | |
| "loss": 0.8999, | |
| "rewards/accuracies": 0.7500000596046448, | |
| "rewards/chosen": 0.8734811544418335, | |
| "rewards/margins": 42.580562591552734, | |
| "rewards/rejected": -41.70708465576172, | |
| "step": 1160 | |
| }, | |
| { | |
| "epoch": 0.9183673469387755, | |
| "grad_norm": 31.83145273716174, | |
| "learning_rate": 1.0091680534213387e-08, | |
| "logits/chosen": 6320.2197265625, | |
| "logits/rejected": 6094.9072265625, | |
| "logps/chosen": -290.4259033203125, | |
| "logps/rejected": -343.50604248046875, | |
| "loss": 0.9088, | |
| "rewards/accuracies": 0.6916667222976685, | |
| "rewards/chosen": 1.7011464834213257, | |
| "rewards/margins": 41.39912796020508, | |
| "rewards/rejected": -39.6979866027832, | |
| "step": 1170 | |
| }, | |
| { | |
| "epoch": 0.9262166405023547, | |
| "grad_norm": 31.116419015526656, | |
| "learning_rate": 8.254517690300944e-09, | |
| "logits/chosen": 5599.1767578125, | |
| "logits/rejected": 5051.66455078125, | |
| "logps/chosen": -274.82489013671875, | |
| "logps/rejected": -302.0051574707031, | |
| "loss": 0.9037, | |
| "rewards/accuracies": 0.7666667699813843, | |
| "rewards/chosen": 4.126657009124756, | |
| "rewards/margins": 47.318702697753906, | |
| "rewards/rejected": -43.19204330444336, | |
| "step": 1180 | |
| }, | |
| { | |
| "epoch": 0.9340659340659341, | |
| "grad_norm": 25.672415270441174, | |
| "learning_rate": 6.599015837372907e-09, | |
| "logits/chosen": 6093.4912109375, | |
| "logits/rejected": 5290.40185546875, | |
| "logps/chosen": -295.9696044921875, | |
| "logps/rejected": -312.8633728027344, | |
| "loss": 0.8923, | |
| "rewards/accuracies": 0.7166666984558105, | |
| "rewards/chosen": -3.6509156227111816, | |
| "rewards/margins": 45.37594985961914, | |
| "rewards/rejected": -49.02686309814453, | |
| "step": 1190 | |
| }, | |
| { | |
| "epoch": 0.9419152276295133, | |
| "grad_norm": 32.1293848999246, | |
| "learning_rate": 5.126419011529992e-09, | |
| "logits/chosen": 6286.7607421875, | |
| "logits/rejected": 5351.490234375, | |
| "logps/chosen": -299.1442565917969, | |
| "logps/rejected": -305.995361328125, | |
| "loss": 0.8891, | |
| "rewards/accuracies": 0.8499999046325684, | |
| "rewards/chosen": 5.603835582733154, | |
| "rewards/margins": 53.45023727416992, | |
| "rewards/rejected": -47.84640121459961, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 0.9497645211930926, | |
| "grad_norm": 37.72170489863353, | |
| "learning_rate": 3.837833803870177e-09, | |
| "logits/chosen": 5903.0517578125, | |
| "logits/rejected": 5179.5322265625, | |
| "logps/chosen": -275.3850402832031, | |
| "logps/rejected": -299.48687744140625, | |
| "loss": 0.895, | |
| "rewards/accuracies": 0.783333420753479, | |
| "rewards/chosen": 4.417906284332275, | |
| "rewards/margins": 51.5880126953125, | |
| "rewards/rejected": -47.17010498046875, | |
| "step": 1210 | |
| }, | |
| { | |
| "epoch": 0.957613814756672, | |
| "grad_norm": 23.85369498972877, | |
| "learning_rate": 2.734228528934679e-09, | |
| "logits/chosen": 7313.7734375, | |
| "logits/rejected": 5392.05517578125, | |
| "logps/chosen": -346.1097106933594, | |
| "logps/rejected": -333.2154541015625, | |
| "loss": 0.9148, | |
| "rewards/accuracies": 0.7583333849906921, | |
| "rewards/chosen": 0.30585241317749023, | |
| "rewards/margins": 48.22237014770508, | |
| "rewards/rejected": -47.9165153503418, | |
| "step": 1220 | |
| }, | |
| { | |
| "epoch": 0.9654631083202512, | |
| "grad_norm": 26.200635812141893, | |
| "learning_rate": 1.8164324970625645e-09, | |
| "logits/chosen": 6493.0498046875, | |
| "logits/rejected": 5102.583984375, | |
| "logps/chosen": -301.46197509765625, | |
| "logps/rejected": -308.8186950683594, | |
| "loss": 0.9018, | |
| "rewards/accuracies": 0.7083333730697632, | |
| "rewards/chosen": 6.893446445465088, | |
| "rewards/margins": 50.65166091918945, | |
| "rewards/rejected": -43.758216857910156, | |
| "step": 1230 | |
| }, | |
| { | |
| "epoch": 0.9733124018838305, | |
| "grad_norm": 22.483038203329606, | |
| "learning_rate": 1.0851353912008642e-09, | |
| "logits/chosen": 5596.9306640625, | |
| "logits/rejected": 5155.1064453125, | |
| "logps/chosen": -273.6884460449219, | |
| "logps/rejected": -323.09979248046875, | |
| "loss": 0.8851, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.7349063754081726, | |
| "rewards/margins": 46.14875030517578, | |
| "rewards/rejected": -46.88364791870117, | |
| "step": 1240 | |
| }, | |
| { | |
| "epoch": 0.9811616954474097, | |
| "grad_norm": 20.587079205487154, | |
| "learning_rate": 5.408867486384471e-10, | |
| "logits/chosen": 5712.6357421875, | |
| "logits/rejected": 4821.349609375, | |
| "logps/chosen": -263.42022705078125, | |
| "logps/rejected": -261.34967041015625, | |
| "loss": 0.8966, | |
| "rewards/accuracies": 0.7833333015441895, | |
| "rewards/chosen": 7.018378257751465, | |
| "rewards/margins": 45.015357971191406, | |
| "rewards/rejected": -37.996986389160156, | |
| "step": 1250 | |
| }, | |
| { | |
| "epoch": 0.989010989010989, | |
| "grad_norm": 24.191259232021025, | |
| "learning_rate": 1.840955480532924e-10, | |
| "logits/chosen": 5454.1728515625, | |
| "logits/rejected": 5137.4345703125, | |
| "logps/chosen": -264.6805725097656, | |
| "logps/rejected": -294.5889587402344, | |
| "loss": 0.8915, | |
| "rewards/accuracies": 0.7833333611488342, | |
| "rewards/chosen": 1.9722245931625366, | |
| "rewards/margins": 38.85475540161133, | |
| "rewards/rejected": -36.882530212402344, | |
| "step": 1260 | |
| }, | |
| { | |
| "epoch": 0.9968602825745683, | |
| "grad_norm": 27.833416743925994, | |
| "learning_rate": 1.502990218302247e-11, | |
| "logits/chosen": 5687.4814453125, | |
| "logits/rejected": 4611.4326171875, | |
| "logps/chosen": -260.2027282714844, | |
| "logps/rejected": -269.2962951660156, | |
| "loss": 0.8907, | |
| "rewards/accuracies": 0.7666667103767395, | |
| "rewards/chosen": 1.270787000656128, | |
| "rewards/margins": 42.4088020324707, | |
| "rewards/rejected": -41.13801193237305, | |
| "step": 1270 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 1274, | |
| "total_flos": 0.0, | |
| "train_loss": 0.9191367263135206, | |
| "train_runtime": 14436.6556, | |
| "train_samples_per_second": 4.235, | |
| "train_steps_per_second": 0.088 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 1274, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 2000, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": false, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 0.0, | |
| "train_batch_size": 6, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |