| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.998691442030882, | |
| "eval_steps": 100, | |
| "global_step": 477, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0, | |
| "learning_rate": 1.0416666666666667e-06, | |
| "logits/chosen": -2.9089105129241943, | |
| "logits/rejected": -2.8982176780700684, | |
| "logps/chosen": -328.48699951171875, | |
| "logps/rejected": -294.0901794433594, | |
| "loss": 0.6931, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 0.02, | |
| "learning_rate": 1.0416666666666668e-05, | |
| "logits/chosen": -2.802924633026123, | |
| "logits/rejected": -2.845370292663574, | |
| "logps/chosen": -274.2305908203125, | |
| "logps/rejected": -257.4792785644531, | |
| "loss": 0.6894, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.4965277910232544, | |
| "rewards/chosen": 0.0021643945947289467, | |
| "rewards/margins": 0.006878577638417482, | |
| "rewards/rejected": -0.004714183043688536, | |
| "step": 10, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 0.04, | |
| "learning_rate": 2.0833333333333336e-05, | |
| "logits/chosen": -2.8522541522979736, | |
| "logits/rejected": -2.842360019683838, | |
| "logps/chosen": -279.8498840332031, | |
| "logps/rejected": -264.8490905761719, | |
| "loss": 0.6557, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 0.02565639279782772, | |
| "rewards/margins": 0.10996748507022858, | |
| "rewards/rejected": -0.0843110978603363, | |
| "step": 20, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 0.06, | |
| "learning_rate": 3.125e-05, | |
| "logits/chosen": -2.8415942192077637, | |
| "logits/rejected": -2.829390525817871, | |
| "logps/chosen": -283.43377685546875, | |
| "logps/rejected": -275.16021728515625, | |
| "loss": 0.6152, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -0.017193807289004326, | |
| "rewards/margins": 0.2977373003959656, | |
| "rewards/rejected": -0.31493109464645386, | |
| "step": 30, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "learning_rate": 4.166666666666667e-05, | |
| "logits/chosen": -2.808344841003418, | |
| "logits/rejected": -2.8009369373321533, | |
| "logps/chosen": -299.697998046875, | |
| "logps/rejected": -281.28240966796875, | |
| "loss": 0.5509, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.734375, | |
| "rewards/chosen": 0.011143045499920845, | |
| "rewards/margins": 0.5404817461967468, | |
| "rewards/rejected": -0.5293387174606323, | |
| "step": 40, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "learning_rate": 4.976689976689977e-05, | |
| "logits/chosen": -2.735532283782959, | |
| "logits/rejected": -2.762310743331909, | |
| "logps/chosen": -284.2992248535156, | |
| "logps/rejected": -285.3406982421875, | |
| "loss": 0.5757, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": 0.06582433730363846, | |
| "rewards/margins": 0.6093484163284302, | |
| "rewards/rejected": -0.5435240864753723, | |
| "step": 50, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "learning_rate": 4.86013986013986e-05, | |
| "logits/chosen": -2.7604079246520996, | |
| "logits/rejected": -2.749760389328003, | |
| "logps/chosen": -271.2710876464844, | |
| "logps/rejected": -253.1558074951172, | |
| "loss": 0.5778, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.690625011920929, | |
| "rewards/chosen": 0.08226754516363144, | |
| "rewards/margins": 0.5549899935722351, | |
| "rewards/rejected": -0.47272244095802307, | |
| "step": 60, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "learning_rate": 4.7435897435897435e-05, | |
| "logits/chosen": -2.825266122817993, | |
| "logits/rejected": -2.8063371181488037, | |
| "logps/chosen": -282.8874816894531, | |
| "logps/rejected": -260.37786865234375, | |
| "loss": 0.5591, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.6968749761581421, | |
| "rewards/chosen": -0.020037399604916573, | |
| "rewards/margins": 0.611237108707428, | |
| "rewards/rejected": -0.6312744617462158, | |
| "step": 70, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 0.17, | |
| "learning_rate": 4.6270396270396274e-05, | |
| "logits/chosen": -2.777280807495117, | |
| "logits/rejected": -2.760341167449951, | |
| "logps/chosen": -278.3568420410156, | |
| "logps/rejected": -281.4912109375, | |
| "loss": 0.5484, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": 0.02771860361099243, | |
| "rewards/margins": 0.6373428702354431, | |
| "rewards/rejected": -0.6096242666244507, | |
| "step": 80, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 0.19, | |
| "learning_rate": 4.5104895104895105e-05, | |
| "logits/chosen": -2.7895703315734863, | |
| "logits/rejected": -2.7826099395751953, | |
| "logps/chosen": -274.6077575683594, | |
| "logps/rejected": -265.6002197265625, | |
| "loss": 0.5476, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 0.15948975086212158, | |
| "rewards/margins": 0.703137218952179, | |
| "rewards/rejected": -0.5436475276947021, | |
| "step": 90, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "learning_rate": 4.3939393939393944e-05, | |
| "logits/chosen": -2.842555046081543, | |
| "logits/rejected": -2.8219830989837646, | |
| "logps/chosen": -282.922607421875, | |
| "logps/rejected": -272.0762939453125, | |
| "loss": 0.5218, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.753125011920929, | |
| "rewards/chosen": 0.1369757205247879, | |
| "rewards/margins": 0.8804744482040405, | |
| "rewards/rejected": -0.7434987425804138, | |
| "step": 100, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 0.23, | |
| "learning_rate": 4.2773892773892776e-05, | |
| "logits/chosen": -2.7973036766052246, | |
| "logits/rejected": -2.8159468173980713, | |
| "logps/chosen": -287.54254150390625, | |
| "logps/rejected": -269.5066223144531, | |
| "loss": 0.5455, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.734375, | |
| "rewards/chosen": 0.057723164558410645, | |
| "rewards/margins": 0.7660588622093201, | |
| "rewards/rejected": -0.7083355784416199, | |
| "step": 110, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "learning_rate": 4.1608391608391614e-05, | |
| "logits/chosen": -2.7736809253692627, | |
| "logits/rejected": -2.7753469944000244, | |
| "logps/chosen": -292.09246826171875, | |
| "logps/rejected": -252.3135986328125, | |
| "loss": 0.5353, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.07172087579965591, | |
| "rewards/margins": 0.8031927347183228, | |
| "rewards/rejected": -0.8749135732650757, | |
| "step": 120, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 0.27, | |
| "learning_rate": 4.0442890442890446e-05, | |
| "logits/chosen": -2.7675487995147705, | |
| "logits/rejected": -2.761944532394409, | |
| "logps/chosen": -271.2415771484375, | |
| "logps/rejected": -269.3936767578125, | |
| "loss": 0.5413, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.703125, | |
| "rewards/chosen": -0.16168196499347687, | |
| "rewards/margins": 0.7922881841659546, | |
| "rewards/rejected": -0.9539702534675598, | |
| "step": 130, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "learning_rate": 3.9277389277389285e-05, | |
| "logits/chosen": -2.821808338165283, | |
| "logits/rejected": -2.828434467315674, | |
| "logps/chosen": -293.92376708984375, | |
| "logps/rejected": -278.75469970703125, | |
| "loss": 0.5249, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.7593749761581421, | |
| "rewards/chosen": -0.10061223804950714, | |
| "rewards/margins": 0.9257994890213013, | |
| "rewards/rejected": -1.026411771774292, | |
| "step": 140, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 0.31, | |
| "learning_rate": 3.811188811188811e-05, | |
| "logits/chosen": -2.8263511657714844, | |
| "logits/rejected": -2.803342819213867, | |
| "logps/chosen": -280.3300476074219, | |
| "logps/rejected": -244.6310577392578, | |
| "loss": 0.5312, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.721875011920929, | |
| "rewards/chosen": -0.2029426395893097, | |
| "rewards/margins": 0.7571284174919128, | |
| "rewards/rejected": -0.9600710868835449, | |
| "step": 150, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "learning_rate": 3.694638694638695e-05, | |
| "logits/chosen": -2.7946958541870117, | |
| "logits/rejected": -2.748753309249878, | |
| "logps/chosen": -282.52154541015625, | |
| "logps/rejected": -296.4395446777344, | |
| "loss": 0.5058, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -0.10657407343387604, | |
| "rewards/margins": 1.0698082447052002, | |
| "rewards/rejected": -1.176382303237915, | |
| "step": 160, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 0.36, | |
| "learning_rate": 3.578088578088578e-05, | |
| "logits/chosen": -2.7990992069244385, | |
| "logits/rejected": -2.7839698791503906, | |
| "logps/chosen": -280.283447265625, | |
| "logps/rejected": -245.1365966796875, | |
| "loss": 0.5741, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.0219185221940279, | |
| "rewards/margins": 0.7351241707801819, | |
| "rewards/rejected": -0.7570425868034363, | |
| "step": 170, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 0.38, | |
| "learning_rate": 3.461538461538462e-05, | |
| "logits/chosen": -2.8004555702209473, | |
| "logits/rejected": -2.80672025680542, | |
| "logps/chosen": -286.1470031738281, | |
| "logps/rejected": -280.5009765625, | |
| "loss": 0.5354, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -0.1152573823928833, | |
| "rewards/margins": 0.79168301820755, | |
| "rewards/rejected": -0.9069403409957886, | |
| "step": 180, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "learning_rate": 3.344988344988345e-05, | |
| "logits/chosen": -2.820003032684326, | |
| "logits/rejected": -2.789689302444458, | |
| "logps/chosen": -281.3421936035156, | |
| "logps/rejected": -278.36883544921875, | |
| "loss": 0.5092, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -0.06256841123104095, | |
| "rewards/margins": 0.9592651128768921, | |
| "rewards/rejected": -1.0218336582183838, | |
| "step": 190, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "learning_rate": 3.228438228438229e-05, | |
| "logits/chosen": -2.7884361743927, | |
| "logits/rejected": -2.782519578933716, | |
| "logps/chosen": -294.6650390625, | |
| "logps/rejected": -261.40435791015625, | |
| "loss": 0.5052, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.734375, | |
| "rewards/chosen": 0.02564082108438015, | |
| "rewards/margins": 0.9850121736526489, | |
| "rewards/rejected": -0.9593712687492371, | |
| "step": 200, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 0.44, | |
| "learning_rate": 3.111888111888112e-05, | |
| "logits/chosen": -2.797853946685791, | |
| "logits/rejected": -2.7864465713500977, | |
| "logps/chosen": -269.8060607910156, | |
| "logps/rejected": -262.01422119140625, | |
| "loss": 0.5258, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -0.046703118830919266, | |
| "rewards/margins": 0.7715562582015991, | |
| "rewards/rejected": -0.8182594180107117, | |
| "step": 210, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "learning_rate": 2.9953379953379956e-05, | |
| "logits/chosen": -2.8102259635925293, | |
| "logits/rejected": -2.813842296600342, | |
| "logps/chosen": -280.2008972167969, | |
| "logps/rejected": -262.47161865234375, | |
| "loss": 0.5326, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.7093750238418579, | |
| "rewards/chosen": 0.038443028926849365, | |
| "rewards/margins": 0.6599202156066895, | |
| "rewards/rejected": -0.6214772462844849, | |
| "step": 220, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "learning_rate": 2.878787878787879e-05, | |
| "logits/chosen": -2.7520573139190674, | |
| "logits/rejected": -2.744856357574463, | |
| "logps/chosen": -265.61956787109375, | |
| "logps/rejected": -256.87115478515625, | |
| "loss": 0.5328, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.690625011920929, | |
| "rewards/chosen": 0.014087711460888386, | |
| "rewards/margins": 0.7686548233032227, | |
| "rewards/rejected": -0.7545671463012695, | |
| "step": 230, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "learning_rate": 2.762237762237762e-05, | |
| "logits/chosen": -2.7545018196105957, | |
| "logits/rejected": -2.730861186981201, | |
| "logps/chosen": -256.40753173828125, | |
| "logps/rejected": -260.3459167480469, | |
| "loss": 0.5315, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.7093750238418579, | |
| "rewards/chosen": -0.0070175291039049625, | |
| "rewards/margins": 0.7594768404960632, | |
| "rewards/rejected": -0.7664943933486938, | |
| "step": 240, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "learning_rate": 2.6456876456876455e-05, | |
| "logits/chosen": -2.8026585578918457, | |
| "logits/rejected": -2.778594493865967, | |
| "logps/chosen": -277.6792907714844, | |
| "logps/rejected": -256.2415771484375, | |
| "loss": 0.5158, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.7406250238418579, | |
| "rewards/chosen": 0.056344062089920044, | |
| "rewards/margins": 0.8976815938949585, | |
| "rewards/rejected": -0.8413375616073608, | |
| "step": 250, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 0.54, | |
| "learning_rate": 2.529137529137529e-05, | |
| "logits/chosen": -2.8391366004943848, | |
| "logits/rejected": -2.8148343563079834, | |
| "logps/chosen": -284.32781982421875, | |
| "logps/rejected": -262.647705078125, | |
| "loss": 0.4916, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.7406250238418579, | |
| "rewards/chosen": -0.06821730732917786, | |
| "rewards/margins": 1.0360552072525024, | |
| "rewards/rejected": -1.104272484779358, | |
| "step": 260, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 0.57, | |
| "learning_rate": 2.4125874125874125e-05, | |
| "logits/chosen": -2.841184139251709, | |
| "logits/rejected": -2.813793659210205, | |
| "logps/chosen": -295.7508544921875, | |
| "logps/rejected": -291.97467041015625, | |
| "loss": 0.5071, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.746874988079071, | |
| "rewards/chosen": -0.13246159255504608, | |
| "rewards/margins": 1.0678224563598633, | |
| "rewards/rejected": -1.2002841234207153, | |
| "step": 270, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "learning_rate": 2.296037296037296e-05, | |
| "logits/chosen": -2.8141586780548096, | |
| "logits/rejected": -2.803541898727417, | |
| "logps/chosen": -283.2509765625, | |
| "logps/rejected": -250.5514373779297, | |
| "loss": 0.5358, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.715624988079071, | |
| "rewards/chosen": -0.14472146332263947, | |
| "rewards/margins": 0.9390050768852234, | |
| "rewards/rejected": -1.0837266445159912, | |
| "step": 280, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 0.61, | |
| "learning_rate": 2.1794871794871795e-05, | |
| "logits/chosen": -2.821655035018921, | |
| "logits/rejected": -2.8060765266418457, | |
| "logps/chosen": -274.44903564453125, | |
| "logps/rejected": -272.57745361328125, | |
| "loss": 0.5206, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -0.06504921615123749, | |
| "rewards/margins": 0.8676969408988953, | |
| "rewards/rejected": -0.9327462315559387, | |
| "step": 290, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 0.63, | |
| "learning_rate": 2.062937062937063e-05, | |
| "logits/chosen": -2.8019015789031982, | |
| "logits/rejected": -2.801147222518921, | |
| "logps/chosen": -292.081787109375, | |
| "logps/rejected": -301.0537414550781, | |
| "loss": 0.5177, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.734375, | |
| "rewards/chosen": 0.03986026719212532, | |
| "rewards/margins": 0.968209445476532, | |
| "rewards/rejected": -0.9283491969108582, | |
| "step": 300, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 0.65, | |
| "learning_rate": 1.9463869463869462e-05, | |
| "logits/chosen": -2.809706449508667, | |
| "logits/rejected": -2.781670570373535, | |
| "logps/chosen": -293.81451416015625, | |
| "logps/rejected": -256.2749938964844, | |
| "loss": 0.5066, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.7281249761581421, | |
| "rewards/chosen": -0.08690959960222244, | |
| "rewards/margins": 0.949812114238739, | |
| "rewards/rejected": -1.0367217063903809, | |
| "step": 310, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "learning_rate": 1.8298368298368298e-05, | |
| "logits/chosen": -2.797341823577881, | |
| "logits/rejected": -2.787775754928589, | |
| "logps/chosen": -269.12139892578125, | |
| "logps/rejected": -266.01080322265625, | |
| "loss": 0.4996, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": 0.0034091435372829437, | |
| "rewards/margins": 1.1171700954437256, | |
| "rewards/rejected": -1.1137609481811523, | |
| "step": 320, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 0.69, | |
| "learning_rate": 1.7132867132867133e-05, | |
| "logits/chosen": -2.8247230052948, | |
| "logits/rejected": -2.80410099029541, | |
| "logps/chosen": -299.09442138671875, | |
| "logps/rejected": -273.0837707519531, | |
| "loss": 0.5236, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -0.04382320120930672, | |
| "rewards/margins": 0.871612548828125, | |
| "rewards/rejected": -0.915435791015625, | |
| "step": 330, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 0.71, | |
| "learning_rate": 1.5967365967365968e-05, | |
| "logits/chosen": -2.806962251663208, | |
| "logits/rejected": -2.79986572265625, | |
| "logps/chosen": -284.2406005859375, | |
| "logps/rejected": -258.4219665527344, | |
| "loss": 0.5258, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.7093750238418579, | |
| "rewards/chosen": -0.023638445883989334, | |
| "rewards/margins": 0.8371860384941101, | |
| "rewards/rejected": -0.8608245849609375, | |
| "step": 340, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "learning_rate": 1.4801864801864803e-05, | |
| "logits/chosen": -2.803683280944824, | |
| "logits/rejected": -2.7878894805908203, | |
| "logps/chosen": -263.1024169921875, | |
| "logps/rejected": -239.3209686279297, | |
| "loss": 0.4963, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.765625, | |
| "rewards/chosen": -0.0065319957211613655, | |
| "rewards/margins": 1.0474125146865845, | |
| "rewards/rejected": -1.05394446849823, | |
| "step": 350, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "learning_rate": 1.3636363636363637e-05, | |
| "logits/chosen": -2.8253402709960938, | |
| "logits/rejected": -2.817469596862793, | |
| "logps/chosen": -280.67022705078125, | |
| "logps/rejected": -258.00579833984375, | |
| "loss": 0.4967, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.734375, | |
| "rewards/chosen": -0.015683341771364212, | |
| "rewards/margins": 1.0513224601745605, | |
| "rewards/rejected": -1.0670057535171509, | |
| "step": 360, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 0.77, | |
| "learning_rate": 1.2470862470862472e-05, | |
| "logits/chosen": -2.793775796890259, | |
| "logits/rejected": -2.796365261077881, | |
| "logps/chosen": -302.306884765625, | |
| "logps/rejected": -273.70635986328125, | |
| "loss": 0.5115, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": 0.010395345278084278, | |
| "rewards/margins": 0.973157525062561, | |
| "rewards/rejected": -0.9627620577812195, | |
| "step": 370, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "learning_rate": 1.1305361305361307e-05, | |
| "logits/chosen": -2.8007330894470215, | |
| "logits/rejected": -2.809499740600586, | |
| "logps/chosen": -277.95599365234375, | |
| "logps/rejected": -271.8043212890625, | |
| "loss": 0.5023, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": 0.0019541799556463957, | |
| "rewards/margins": 1.0503606796264648, | |
| "rewards/rejected": -1.0484063625335693, | |
| "step": 380, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 0.82, | |
| "learning_rate": 1.013986013986014e-05, | |
| "logits/chosen": -2.838869571685791, | |
| "logits/rejected": -2.8424456119537354, | |
| "logps/chosen": -291.6543884277344, | |
| "logps/rejected": -275.7422790527344, | |
| "loss": 0.5223, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.7406250238418579, | |
| "rewards/chosen": 0.009039236232638359, | |
| "rewards/margins": 0.990314781665802, | |
| "rewards/rejected": -0.9812755584716797, | |
| "step": 390, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "learning_rate": 8.974358974358976e-06, | |
| "logits/chosen": -2.7799577713012695, | |
| "logits/rejected": -2.7700963020324707, | |
| "logps/chosen": -292.00604248046875, | |
| "logps/rejected": -278.8124694824219, | |
| "loss": 0.5059, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.7281249761581421, | |
| "rewards/chosen": -0.08654220402240753, | |
| "rewards/margins": 0.9550994634628296, | |
| "rewards/rejected": -1.0416417121887207, | |
| "step": 400, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 0.86, | |
| "learning_rate": 7.808857808857809e-06, | |
| "logits/chosen": -2.7790863513946533, | |
| "logits/rejected": -2.7839720249176025, | |
| "logps/chosen": -288.0146484375, | |
| "logps/rejected": -246.18386840820312, | |
| "loss": 0.5193, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.7593749761581421, | |
| "rewards/chosen": -0.0843573808670044, | |
| "rewards/margins": 1.01529860496521, | |
| "rewards/rejected": -1.0996559858322144, | |
| "step": 410, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "learning_rate": 6.643356643356643e-06, | |
| "logits/chosen": -2.8042919635772705, | |
| "logits/rejected": -2.814549446105957, | |
| "logps/chosen": -267.73272705078125, | |
| "logps/rejected": -282.6224060058594, | |
| "loss": 0.5131, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.23174908757209778, | |
| "rewards/margins": 0.7961767315864563, | |
| "rewards/rejected": -1.027925729751587, | |
| "step": 420, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 0.9, | |
| "learning_rate": 5.477855477855478e-06, | |
| "logits/chosen": -2.8008358478546143, | |
| "logits/rejected": -2.827676773071289, | |
| "logps/chosen": -291.685791015625, | |
| "logps/rejected": -269.49896240234375, | |
| "loss": 0.5135, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.7593749761581421, | |
| "rewards/chosen": -0.1730644553899765, | |
| "rewards/margins": 0.9842265248298645, | |
| "rewards/rejected": -1.157291054725647, | |
| "step": 430, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 0.92, | |
| "learning_rate": 4.312354312354312e-06, | |
| "logits/chosen": -2.8276031017303467, | |
| "logits/rejected": -2.81488299369812, | |
| "logps/chosen": -265.67242431640625, | |
| "logps/rejected": -261.6026611328125, | |
| "loss": 0.51, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.734375, | |
| "rewards/chosen": -0.08840381354093552, | |
| "rewards/margins": 0.9768926501274109, | |
| "rewards/rejected": -1.0652964115142822, | |
| "step": 440, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 0.94, | |
| "learning_rate": 3.1468531468531472e-06, | |
| "logits/chosen": -2.794529676437378, | |
| "logits/rejected": -2.7928192615509033, | |
| "logps/chosen": -280.88995361328125, | |
| "logps/rejected": -275.9464416503906, | |
| "loss": 0.5128, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.765625, | |
| "rewards/chosen": -0.06786171346902847, | |
| "rewards/margins": 0.9994968175888062, | |
| "rewards/rejected": -1.0673584938049316, | |
| "step": 450, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "learning_rate": 1.981351981351981e-06, | |
| "logits/chosen": -2.7951102256774902, | |
| "logits/rejected": -2.7973313331604004, | |
| "logps/chosen": -297.4739990234375, | |
| "logps/rejected": -272.06024169921875, | |
| "loss": 0.4871, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.08221320062875748, | |
| "rewards/margins": 0.9878571629524231, | |
| "rewards/rejected": -1.0700703859329224, | |
| "step": 460, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 0.98, | |
| "learning_rate": 8.158508158508159e-07, | |
| "logits/chosen": -2.8256263732910156, | |
| "logits/rejected": -2.7975211143493652, | |
| "logps/chosen": -278.83154296875, | |
| "logps/rejected": -266.3514099121094, | |
| "loss": 0.4951, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.7593749761581421, | |
| "rewards/chosen": -0.10287537425756454, | |
| "rewards/margins": 1.0982601642608643, | |
| "rewards/rejected": -1.2011353969573975, | |
| "step": 470, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_logits/chosen": -2.822110891342163, | |
| "eval_logits/rejected": -2.8178136348724365, | |
| "eval_logps/chosen": -285.4951171875, | |
| "eval_logps/rejected": -270.55743408203125, | |
| "eval_loss": 0.5198934078216553, | |
| "eval_pred_label": 0.0, | |
| "eval_rewards/accuracies": 0.7300000190734863, | |
| "eval_rewards/chosen": -0.12377375364303589, | |
| "eval_rewards/margins": 1.001997470855713, | |
| "eval_rewards/rejected": -1.1257712841033936, | |
| "eval_runtime": 453.8631, | |
| "eval_samples_per_second": 4.407, | |
| "eval_steps_per_second": 0.275, | |
| "eval_use_label": 0.0, | |
| "step": 477 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 477, | |
| "total_flos": 0.0, | |
| "train_loss": 0.5321606655040877, | |
| "train_runtime": 24451.2028, | |
| "train_samples_per_second": 2.5, | |
| "train_steps_per_second": 0.02 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 477, | |
| "num_train_epochs": 1, | |
| "save_steps": 50, | |
| "total_flos": 0.0, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |