| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 1.6632016632016633, | |
| "eval_steps": 50, | |
| "global_step": 200, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.04158004158004158, | |
| "grad_norm": 57.63412014296559, | |
| "learning_rate": 5e-07, | |
| "logits/chosen": -2.7145180702209473, | |
| "logits/rejected": -2.7382097244262695, | |
| "logps/chosen": -265.24481201171875, | |
| "logps/rejected": -201.1150665283203, | |
| "loss": 0.6915, | |
| "rewards/accuracies": 0.375, | |
| "rewards/chosen": 0.01524554193019867, | |
| "rewards/margins": 0.008573563769459724, | |
| "rewards/rejected": 0.006671978626400232, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.08316008316008316, | |
| "grad_norm": 52.91443405505758, | |
| "learning_rate": 1e-06, | |
| "logits/chosen": -2.690777063369751, | |
| "logits/rejected": -2.689495325088501, | |
| "logps/chosen": -259.32598876953125, | |
| "logps/rejected": -236.3219757080078, | |
| "loss": 0.6616, | |
| "rewards/accuracies": 0.6312500238418579, | |
| "rewards/chosen": 0.4881957471370697, | |
| "rewards/margins": 0.10135326534509659, | |
| "rewards/rejected": 0.3868425786495209, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.12474012474012475, | |
| "grad_norm": 53.28329333963272, | |
| "learning_rate": 9.994965332706572e-07, | |
| "logits/chosen": -2.5831141471862793, | |
| "logits/rejected": -2.580876350402832, | |
| "logps/chosen": -237.8153839111328, | |
| "logps/rejected": -192.7242889404297, | |
| "loss": 0.6359, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 1.1336857080459595, | |
| "rewards/margins": 0.47578057646751404, | |
| "rewards/rejected": 0.6579049825668335, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.16632016632016633, | |
| "grad_norm": 59.51771261732688, | |
| "learning_rate": 9.979871469976195e-07, | |
| "logits/chosen": -2.5097970962524414, | |
| "logits/rejected": -2.491774797439575, | |
| "logps/chosen": -241.24349975585938, | |
| "logps/rejected": -206.7909698486328, | |
| "loss": 0.6385, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 1.5591249465942383, | |
| "rewards/margins": 0.980699896812439, | |
| "rewards/rejected": 0.5784249305725098, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.2079002079002079, | |
| "grad_norm": 43.932979300169734, | |
| "learning_rate": 9.954748808839674e-07, | |
| "logits/chosen": -2.428896427154541, | |
| "logits/rejected": -2.426640510559082, | |
| "logps/chosen": -260.3008117675781, | |
| "logps/rejected": -223.4193878173828, | |
| "loss": 0.6287, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": 1.4255732297897339, | |
| "rewards/margins": 0.9219636917114258, | |
| "rewards/rejected": 0.5036097764968872, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.2494802494802495, | |
| "grad_norm": 51.9520224365334, | |
| "learning_rate": 9.919647942993147e-07, | |
| "logits/chosen": -2.270231008529663, | |
| "logits/rejected": -2.2373533248901367, | |
| "logps/chosen": -239.1925811767578, | |
| "logps/rejected": -199.03244018554688, | |
| "loss": 0.6013, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": 1.0962295532226562, | |
| "rewards/margins": 0.8090855479240417, | |
| "rewards/rejected": 0.28714415431022644, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.2910602910602911, | |
| "grad_norm": 42.07074967218912, | |
| "learning_rate": 9.874639560909118e-07, | |
| "logits/chosen": -2.2837703227996826, | |
| "logits/rejected": -2.2773120403289795, | |
| "logps/chosen": -245.8435516357422, | |
| "logps/rejected": -218.3318328857422, | |
| "loss": 0.5782, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": 0.842710018157959, | |
| "rewards/margins": 0.7606425285339355, | |
| "rewards/rejected": 0.08206750452518463, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.33264033264033266, | |
| "grad_norm": 40.52902304303911, | |
| "learning_rate": 9.819814303479267e-07, | |
| "logits/chosen": -2.3677680492401123, | |
| "logits/rejected": -2.3257031440734863, | |
| "logps/chosen": -255.2162322998047, | |
| "logps/rejected": -206.1927490234375, | |
| "loss": 0.5732, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": 0.774177610874176, | |
| "rewards/margins": 0.9143842458724976, | |
| "rewards/rejected": -0.14020657539367676, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.37422037422037424, | |
| "grad_norm": 42.34370895122762, | |
| "learning_rate": 9.755282581475767e-07, | |
| "logits/chosen": -2.39888596534729, | |
| "logits/rejected": -2.3836829662323, | |
| "logps/chosen": -237.30508422851562, | |
| "logps/rejected": -216.06588745117188, | |
| "loss": 0.5958, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.7707290649414062, | |
| "rewards/margins": 1.1094779968261719, | |
| "rewards/rejected": -0.3387489914894104, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 0.4158004158004158, | |
| "grad_norm": 46.42422960165206, | |
| "learning_rate": 9.681174353198686e-07, | |
| "logits/chosen": -2.3984861373901367, | |
| "logits/rejected": -2.3900020122528076, | |
| "logps/chosen": -233.5660400390625, | |
| "logps/rejected": -216.6641082763672, | |
| "loss": 0.635, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": 0.9606113433837891, | |
| "rewards/margins": 1.124942660331726, | |
| "rewards/rejected": -0.16433146595954895, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.4158004158004158, | |
| "eval_logits/chosen": -2.3899078369140625, | |
| "eval_logits/rejected": -2.3907740116119385, | |
| "eval_logps/chosen": -237.2779541015625, | |
| "eval_logps/rejected": -211.79153442382812, | |
| "eval_loss": 0.598716676235199, | |
| "eval_rewards/accuracies": 0.7013888955116272, | |
| "eval_rewards/chosen": 0.8451206684112549, | |
| "eval_rewards/margins": 0.8839507102966309, | |
| "eval_rewards/rejected": -0.038830023258924484, | |
| "eval_runtime": 229.0134, | |
| "eval_samples_per_second": 14.934, | |
| "eval_steps_per_second": 0.236, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.4573804573804574, | |
| "grad_norm": 38.39211344385862, | |
| "learning_rate": 9.597638862757253e-07, | |
| "logits/chosen": -2.456921339035034, | |
| "logits/rejected": -2.4394099712371826, | |
| "logps/chosen": -237.6243133544922, | |
| "logps/rejected": -233.6071014404297, | |
| "loss": 0.6009, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": 0.8705752491950989, | |
| "rewards/margins": 1.126349687576294, | |
| "rewards/rejected": -0.2557744085788727, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 0.498960498960499, | |
| "grad_norm": 37.3024427630972, | |
| "learning_rate": 9.504844339512094e-07, | |
| "logits/chosen": -2.521387815475464, | |
| "logits/rejected": -2.4951186180114746, | |
| "logps/chosen": -266.1732177734375, | |
| "logps/rejected": -242.85897827148438, | |
| "loss": 0.554, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": 0.9916666150093079, | |
| "rewards/margins": 1.3694941997528076, | |
| "rewards/rejected": -0.3778277039527893, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.5405405405405406, | |
| "grad_norm": 34.5258394053866, | |
| "learning_rate": 9.402977659283689e-07, | |
| "logits/chosen": -2.4310925006866455, | |
| "logits/rejected": -2.4042844772338867, | |
| "logps/chosen": -223.9085693359375, | |
| "logps/rejected": -203.2917022705078, | |
| "loss": 0.5364, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": 0.5567647814750671, | |
| "rewards/margins": 0.8208805918693542, | |
| "rewards/rejected": -0.26411566138267517, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 0.5821205821205822, | |
| "grad_norm": 37.746071231755344, | |
| "learning_rate": 9.29224396800933e-07, | |
| "logits/chosen": -2.3321871757507324, | |
| "logits/rejected": -2.2954535484313965, | |
| "logps/chosen": -243.26156616210938, | |
| "logps/rejected": -226.8199005126953, | |
| "loss": 0.5423, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": 0.38314875960350037, | |
| "rewards/margins": 1.1013362407684326, | |
| "rewards/rejected": -0.7181875109672546, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.6237006237006237, | |
| "grad_norm": 57.07804647920337, | |
| "learning_rate": 9.172866268606513e-07, | |
| "logits/chosen": -2.26948881149292, | |
| "logits/rejected": -2.2403252124786377, | |
| "logps/chosen": -251.2090606689453, | |
| "logps/rejected": -240.20169067382812, | |
| "loss": 0.5844, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": 0.9941379427909851, | |
| "rewards/margins": 1.1448538303375244, | |
| "rewards/rejected": -0.1507158726453781, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.6652806652806653, | |
| "grad_norm": 36.48822907311496, | |
| "learning_rate": 9.045084971874737e-07, | |
| "logits/chosen": -2.212411403656006, | |
| "logits/rejected": -2.1935043334960938, | |
| "logps/chosen": -272.8021545410156, | |
| "logps/rejected": -241.2564239501953, | |
| "loss": 0.5453, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": 1.1616096496582031, | |
| "rewards/margins": 1.850441336631775, | |
| "rewards/rejected": -0.6888317465782166, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.7068607068607069, | |
| "grad_norm": 38.5389032336389, | |
| "learning_rate": 8.909157412340149e-07, | |
| "logits/chosen": -2.180285930633545, | |
| "logits/rejected": -2.126051902770996, | |
| "logps/chosen": -253.0643310546875, | |
| "logps/rejected": -237.400390625, | |
| "loss": 0.5269, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.9046308398246765, | |
| "rewards/margins": 1.5041005611419678, | |
| "rewards/rejected": -0.599469780921936, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 0.7484407484407485, | |
| "grad_norm": 31.08445554774815, | |
| "learning_rate": 8.765357330018055e-07, | |
| "logits/chosen": -2.146207332611084, | |
| "logits/rejected": -2.1198434829711914, | |
| "logps/chosen": -246.0504608154297, | |
| "logps/rejected": -242.79714965820312, | |
| "loss": 0.5741, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": 0.4694552421569824, | |
| "rewards/margins": 1.537427544593811, | |
| "rewards/rejected": -1.0679724216461182, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.7900207900207901, | |
| "grad_norm": 32.25150255167185, | |
| "learning_rate": 8.613974319136957e-07, | |
| "logits/chosen": -2.114973306655884, | |
| "logits/rejected": -2.0754525661468506, | |
| "logps/chosen": -249.64132690429688, | |
| "logps/rejected": -233.89352416992188, | |
| "loss": 0.4787, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -0.09324260801076889, | |
| "rewards/margins": 1.446690320968628, | |
| "rewards/rejected": -1.5399329662322998, | |
| "step": 95 | |
| }, | |
| { | |
| "epoch": 0.8316008316008316, | |
| "grad_norm": 34.22313791470385, | |
| "learning_rate": 8.455313244934324e-07, | |
| "logits/chosen": -2.027057409286499, | |
| "logits/rejected": -1.9819841384887695, | |
| "logps/chosen": -250.86947631835938, | |
| "logps/rejected": -241.2573699951172, | |
| "loss": 0.4933, | |
| "rewards/accuracies": 0.8187500238418579, | |
| "rewards/chosen": 0.10764148086309433, | |
| "rewards/margins": 1.9850435256958008, | |
| "rewards/rejected": -1.8774023056030273, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.8316008316008316, | |
| "eval_logits/chosen": -1.9530320167541504, | |
| "eval_logits/rejected": -1.9127507209777832, | |
| "eval_logps/chosen": -247.99232482910156, | |
| "eval_logps/rejected": -229.55447387695312, | |
| "eval_loss": 0.5284922122955322, | |
| "eval_rewards/accuracies": 0.7523148059844971, | |
| "eval_rewards/chosen": -0.22631923854351044, | |
| "eval_rewards/margins": 1.5888041257858276, | |
| "eval_rewards/rejected": -1.815123438835144, | |
| "eval_runtime": 227.9451, | |
| "eval_samples_per_second": 15.004, | |
| "eval_steps_per_second": 0.237, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.8731808731808732, | |
| "grad_norm": 43.67585736596051, | |
| "learning_rate": 8.289693629698563e-07, | |
| "logits/chosen": -1.952876329421997, | |
| "logits/rejected": -1.9095561504364014, | |
| "logps/chosen": -250.580322265625, | |
| "logps/rejected": -255.4891357421875, | |
| "loss": 0.5324, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -0.0729295164346695, | |
| "rewards/margins": 1.8105415105819702, | |
| "rewards/rejected": -1.8834712505340576, | |
| "step": 105 | |
| }, | |
| { | |
| "epoch": 0.9147609147609148, | |
| "grad_norm": 39.1719273308311, | |
| "learning_rate": 8.117449009293668e-07, | |
| "logits/chosen": -1.8559868335723877, | |
| "logits/rejected": -1.799195647239685, | |
| "logps/chosen": -248.75961303710938, | |
| "logps/rejected": -232.480712890625, | |
| "loss": 0.4943, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.30641165375709534, | |
| "rewards/margins": 1.7232319116592407, | |
| "rewards/rejected": -1.4168202877044678, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.9563409563409564, | |
| "grad_norm": 30.704674411877335, | |
| "learning_rate": 7.938926261462365e-07, | |
| "logits/chosen": -1.7355577945709229, | |
| "logits/rejected": -1.6728878021240234, | |
| "logps/chosen": -255.2344970703125, | |
| "logps/rejected": -221.1670684814453, | |
| "loss": 0.4842, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": 0.09199291467666626, | |
| "rewards/margins": 1.9209611415863037, | |
| "rewards/rejected": -1.8289684057235718, | |
| "step": 115 | |
| }, | |
| { | |
| "epoch": 0.997920997920998, | |
| "grad_norm": 41.39067500024576, | |
| "learning_rate": 7.754484907260512e-07, | |
| "logits/chosen": -1.7632324695587158, | |
| "logits/rejected": -1.6800066232681274, | |
| "logps/chosen": -236.336669921875, | |
| "logps/rejected": -248.4853515625, | |
| "loss": 0.4959, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.09794483333826065, | |
| "rewards/margins": 2.035454273223877, | |
| "rewards/rejected": -1.937509536743164, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 1.0395010395010396, | |
| "grad_norm": 18.479902808810362, | |
| "learning_rate": 7.564496387029531e-07, | |
| "logits/chosen": -1.835211157798767, | |
| "logits/rejected": -1.744879126548767, | |
| "logps/chosen": -246.5315704345703, | |
| "logps/rejected": -238.4007568359375, | |
| "loss": 0.2765, | |
| "rewards/accuracies": 0.9312499761581421, | |
| "rewards/chosen": 1.2002815008163452, | |
| "rewards/margins": 3.028635025024414, | |
| "rewards/rejected": -1.8283536434173584, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 1.0810810810810811, | |
| "grad_norm": 20.239636952023396, | |
| "learning_rate": 7.369343312364993e-07, | |
| "logits/chosen": -1.8902709484100342, | |
| "logits/rejected": -1.841406226158142, | |
| "logps/chosen": -249.97555541992188, | |
| "logps/rejected": -225.74893188476562, | |
| "loss": 0.2499, | |
| "rewards/accuracies": 0.918749988079071, | |
| "rewards/chosen": 1.3934506177902222, | |
| "rewards/margins": 3.1044392585754395, | |
| "rewards/rejected": -1.7109886407852173, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 1.1226611226611227, | |
| "grad_norm": 27.020533908406442, | |
| "learning_rate": 7.16941869558779e-07, | |
| "logits/chosen": -2.057455062866211, | |
| "logits/rejected": -2.003194570541382, | |
| "logps/chosen": -232.67001342773438, | |
| "logps/rejected": -234.6235809326172, | |
| "loss": 0.2301, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": 1.6992406845092773, | |
| "rewards/margins": 3.422689914703369, | |
| "rewards/rejected": -1.7234487533569336, | |
| "step": 135 | |
| }, | |
| { | |
| "epoch": 1.1642411642411643, | |
| "grad_norm": 22.833931438432327, | |
| "learning_rate": 6.965125158269618e-07, | |
| "logits/chosen": -2.0912320613861084, | |
| "logits/rejected": -2.0458855628967285, | |
| "logps/chosen": -244.6804656982422, | |
| "logps/rejected": -225.2838897705078, | |
| "loss": 0.228, | |
| "rewards/accuracies": 0.90625, | |
| "rewards/chosen": 2.007575511932373, | |
| "rewards/margins": 3.5606026649475098, | |
| "rewards/rejected": -1.5530271530151367, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 1.2058212058212059, | |
| "grad_norm": 19.05547442329841, | |
| "learning_rate": 6.756874120406714e-07, | |
| "logits/chosen": -2.0981993675231934, | |
| "logits/rejected": -2.0623602867126465, | |
| "logps/chosen": -231.7192840576172, | |
| "logps/rejected": -239.6023712158203, | |
| "loss": 0.2467, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": 1.801042914390564, | |
| "rewards/margins": 3.514739513397217, | |
| "rewards/rejected": -1.713696837425232, | |
| "step": 145 | |
| }, | |
| { | |
| "epoch": 1.2474012474012475, | |
| "grad_norm": 17.479166649922572, | |
| "learning_rate": 6.545084971874736e-07, | |
| "logits/chosen": -2.1294615268707275, | |
| "logits/rejected": -2.0636110305786133, | |
| "logps/chosen": -238.5365447998047, | |
| "logps/rejected": -244.5021514892578, | |
| "loss": 0.2495, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": 1.3247355222702026, | |
| "rewards/margins": 3.6321330070495605, | |
| "rewards/rejected": -2.3073973655700684, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 1.2474012474012475, | |
| "eval_logits/chosen": -2.1232447624206543, | |
| "eval_logits/rejected": -2.0982720851898193, | |
| "eval_logps/chosen": -240.15695190429688, | |
| "eval_logps/rejected": -225.6040802001953, | |
| "eval_loss": 0.5427103042602539, | |
| "eval_rewards/accuracies": 0.7592592835426331, | |
| "eval_rewards/chosen": 0.5572175979614258, | |
| "eval_rewards/margins": 1.9773017168045044, | |
| "eval_rewards/rejected": -1.4200841188430786, | |
| "eval_runtime": 227.6445, | |
| "eval_samples_per_second": 15.023, | |
| "eval_steps_per_second": 0.237, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 1.288981288981289, | |
| "grad_norm": 23.820329780522155, | |
| "learning_rate": 6.330184227833375e-07, | |
| "logits/chosen": -2.1582248210906982, | |
| "logits/rejected": -2.1403584480285645, | |
| "logps/chosen": -229.8828582763672, | |
| "logps/rejected": -232.87924194335938, | |
| "loss": 0.2962, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": 1.5254548788070679, | |
| "rewards/margins": 3.3456084728240967, | |
| "rewards/rejected": -1.8201534748077393, | |
| "step": 155 | |
| }, | |
| { | |
| "epoch": 1.3305613305613306, | |
| "grad_norm": 17.670853805471324, | |
| "learning_rate": 6.112604669781572e-07, | |
| "logits/chosen": -2.1896843910217285, | |
| "logits/rejected": -2.151392698287964, | |
| "logps/chosen": -246.44552612304688, | |
| "logps/rejected": -224.73641967773438, | |
| "loss": 0.2666, | |
| "rewards/accuracies": 0.9125000238418579, | |
| "rewards/chosen": 1.2775523662567139, | |
| "rewards/margins": 3.355015516281128, | |
| "rewards/rejected": -2.077463150024414, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 1.3721413721413722, | |
| "grad_norm": 19.884309338358804, | |
| "learning_rate": 5.892784473993183e-07, | |
| "logits/chosen": -2.1978659629821777, | |
| "logits/rejected": -2.139770984649658, | |
| "logps/chosen": -244.56283569335938, | |
| "logps/rejected": -235.612060546875, | |
| "loss": 0.2784, | |
| "rewards/accuracies": 0.956250011920929, | |
| "rewards/chosen": 1.9508459568023682, | |
| "rewards/margins": 3.806994676589966, | |
| "rewards/rejected": -1.8561484813690186, | |
| "step": 165 | |
| }, | |
| { | |
| "epoch": 1.4137214137214138, | |
| "grad_norm": 21.536221614158325, | |
| "learning_rate": 5.671166329088277e-07, | |
| "logits/chosen": -2.159837245941162, | |
| "logits/rejected": -2.1259591579437256, | |
| "logps/chosen": -235.0163116455078, | |
| "logps/rejected": -240.001953125, | |
| "loss": 0.2998, | |
| "rewards/accuracies": 0.893750011920929, | |
| "rewards/chosen": 1.634421944618225, | |
| "rewards/margins": 3.504880428314209, | |
| "rewards/rejected": -1.870458960533142, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 1.4553014553014554, | |
| "grad_norm": 19.683343698848628, | |
| "learning_rate": 5.448196544517167e-07, | |
| "logits/chosen": -2.1057746410369873, | |
| "logits/rejected": -2.0693142414093018, | |
| "logps/chosen": -242.95492553710938, | |
| "logps/rejected": -222.35372924804688, | |
| "loss": 0.2881, | |
| "rewards/accuracies": 0.9312499761581421, | |
| "rewards/chosen": 1.408005952835083, | |
| "rewards/margins": 3.477801561355591, | |
| "rewards/rejected": -2.069795846939087, | |
| "step": 175 | |
| }, | |
| { | |
| "epoch": 1.496881496881497, | |
| "grad_norm": 23.26853368506061, | |
| "learning_rate": 5.224324151752575e-07, | |
| "logits/chosen": -2.0776679515838623, | |
| "logits/rejected": -2.0356287956237793, | |
| "logps/chosen": -252.8184356689453, | |
| "logps/rejected": -263.45013427734375, | |
| "loss": 0.2876, | |
| "rewards/accuracies": 0.887499988079071, | |
| "rewards/chosen": 1.3347493410110474, | |
| "rewards/margins": 3.841991901397705, | |
| "rewards/rejected": -2.5072426795959473, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 1.5384615384615383, | |
| "grad_norm": 19.309579616953823, | |
| "learning_rate": 5e-07, | |
| "logits/chosen": -2.0446431636810303, | |
| "logits/rejected": -1.985939621925354, | |
| "logps/chosen": -235.72598266601562, | |
| "logps/rejected": -258.97344970703125, | |
| "loss": 0.2758, | |
| "rewards/accuracies": 0.887499988079071, | |
| "rewards/chosen": 1.3530393838882446, | |
| "rewards/margins": 3.8046765327453613, | |
| "rewards/rejected": -2.451636791229248, | |
| "step": 185 | |
| }, | |
| { | |
| "epoch": 1.5800415800415801, | |
| "grad_norm": 18.625040552548324, | |
| "learning_rate": 4.775675848247427e-07, | |
| "logits/chosen": -2.0644099712371826, | |
| "logits/rejected": -2.0216641426086426, | |
| "logps/chosen": -241.5558624267578, | |
| "logps/rejected": -263.6393737792969, | |
| "loss": 0.292, | |
| "rewards/accuracies": 0.887499988079071, | |
| "rewards/chosen": 1.0546023845672607, | |
| "rewards/margins": 3.451650619506836, | |
| "rewards/rejected": -2.397047996520996, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 1.6216216216216215, | |
| "grad_norm": 25.624591590300348, | |
| "learning_rate": 4.5518034554828327e-07, | |
| "logits/chosen": -2.0048115253448486, | |
| "logits/rejected": -1.9765331745147705, | |
| "logps/chosen": -227.64968872070312, | |
| "logps/rejected": -238.43795776367188, | |
| "loss": 0.2831, | |
| "rewards/accuracies": 0.887499988079071, | |
| "rewards/chosen": 1.2037407159805298, | |
| "rewards/margins": 3.231950044631958, | |
| "rewards/rejected": -2.0282092094421387, | |
| "step": 195 | |
| }, | |
| { | |
| "epoch": 1.6632016632016633, | |
| "grad_norm": 18.88665153638521, | |
| "learning_rate": 4.328833670911724e-07, | |
| "logits/chosen": -2.0011658668518066, | |
| "logits/rejected": -1.9661411046981812, | |
| "logps/chosen": -240.28115844726562, | |
| "logps/rejected": -233.8397674560547, | |
| "loss": 0.2753, | |
| "rewards/accuracies": 0.918749988079071, | |
| "rewards/chosen": 1.5664188861846924, | |
| "rewards/margins": 3.7791213989257812, | |
| "rewards/rejected": -2.212702512741089, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 1.6632016632016633, | |
| "eval_logits/chosen": -2.0067989826202393, | |
| "eval_logits/rejected": -1.9752123355865479, | |
| "eval_logps/chosen": -239.95291137695312, | |
| "eval_logps/rejected": -228.13818359375, | |
| "eval_loss": 0.5260343551635742, | |
| "eval_rewards/accuracies": 0.7870370149612427, | |
| "eval_rewards/chosen": 0.5776218175888062, | |
| "eval_rewards/margins": 2.2511181831359863, | |
| "eval_rewards/rejected": -1.6734964847564697, | |
| "eval_runtime": 227.8787, | |
| "eval_samples_per_second": 15.008, | |
| "eval_steps_per_second": 0.237, | |
| "step": 200 | |
| } | |
| ], | |
| "logging_steps": 5, | |
| "max_steps": 360, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "save_steps": 100, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 2358113407598592.0, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |