| { |
| "best_global_step": null, |
| "best_metric": null, |
| "best_model_checkpoint": null, |
| "epoch": 0.4287628405538187, |
| "eval_steps": 300, |
| "global_step": 150, |
| "is_hyper_param_search": false, |
| "is_local_process_zero": true, |
| "is_world_process_zero": true, |
| "log_history": [ |
| { |
| "epoch": 0.002858418937025458, |
| "grad_norm": 10.349462509155273, |
| "learning_rate": 3.7735849056603773e-06, |
| "logits/chosen": -0.6422490477561951, |
| "logits/rejected": -0.6447486877441406, |
| "logps/chosen": -6.999429225921631, |
| "logps/rejected": -13.433603286743164, |
| "loss": 1.3778549432754517, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.6847077012062073, |
| "rewards/accuracies": 0.0, |
| "rewards/chosen": 0.0, |
| "rewards/margins": 0.0, |
| "rewards/rejected": 0.0, |
| "step": 1, |
| "train_speed(iter/s)": 0.005273 |
| }, |
| { |
| "epoch": 0.005716837874050916, |
| "grad_norm": 11.263213157653809, |
| "learning_rate": 7.547169811320755e-06, |
| "logits/chosen": -0.6429960131645203, |
| "logits/rejected": -0.6454926133155823, |
| "logps/chosen": -9.604641914367676, |
| "logps/rejected": -13.53369140625, |
| "loss": 1.4975696802139282, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.8044224381446838, |
| "rewards/accuracies": 0.0, |
| "rewards/chosen": 0.0, |
| "rewards/margins": 0.0, |
| "rewards/rejected": 0.0, |
| "step": 2, |
| "train_speed(iter/s)": 0.005366 |
| }, |
| { |
| "epoch": 0.008575256811076373, |
| "grad_norm": 9.09404182434082, |
| "learning_rate": 1.1320754716981132e-05, |
| "logits/chosen": -0.6486532092094421, |
| "logits/rejected": -0.6493248343467712, |
| "logps/chosen": -7.69912576675415, |
| "logps/rejected": -17.816326141357422, |
| "loss": 1.2593896389007568, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.5679630637168884, |
| "rewards/accuracies": 0.625, |
| "rewards/chosen": 0.0031128099653869867, |
| "rewards/margins": 0.0035441224463284016, |
| "rewards/rejected": -0.00043131227721460164, |
| "step": 3, |
| "train_speed(iter/s)": 0.005419 |
| }, |
| { |
| "epoch": 0.011433675748101831, |
| "grad_norm": 13.314420700073242, |
| "learning_rate": 1.509433962264151e-05, |
| "logits/chosen": -0.6668453812599182, |
| "logits/rejected": -0.6718658804893494, |
| "logps/chosen": -6.061755180358887, |
| "logps/rejected": -13.643918991088867, |
| "loss": 1.4406771659851074, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.7477570176124573, |
| "rewards/accuracies": 0.53125, |
| "rewards/chosen": 0.008774133399128914, |
| "rewards/margins": 0.0006008772179484367, |
| "rewards/rejected": 0.008173256181180477, |
| "step": 4, |
| "train_speed(iter/s)": 0.005446 |
| }, |
| { |
| "epoch": 0.014292094685127288, |
| "grad_norm": 12.340804100036621, |
| "learning_rate": 1.8867924528301888e-05, |
| "logits/chosen": -0.6321276426315308, |
| "logits/rejected": -0.6321861743927002, |
| "logps/chosen": -11.234216690063477, |
| "logps/rejected": -11.82645034790039, |
| "loss": 1.3690662384033203, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.6870464086532593, |
| "rewards/accuracies": 0.6875, |
| "rewards/chosen": 0.032103873789310455, |
| "rewards/margins": 0.022960105910897255, |
| "rewards/rejected": 0.0091437678784132, |
| "step": 5, |
| "train_speed(iter/s)": 0.005444 |
| }, |
| { |
| "epoch": 0.017150513622152745, |
| "grad_norm": 13.995262145996094, |
| "learning_rate": 2.2641509433962265e-05, |
| "logits/chosen": -0.6530164480209351, |
| "logits/rejected": -0.6550691723823547, |
| "logps/chosen": -8.742744445800781, |
| "logps/rejected": -12.622247695922852, |
| "loss": 1.3672001361846924, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.702350378036499, |
| "rewards/accuracies": 0.65625, |
| "rewards/chosen": 0.0800265222787857, |
| "rewards/margins": 0.06047351285815239, |
| "rewards/rejected": 0.019553007557988167, |
| "step": 6, |
| "train_speed(iter/s)": 0.005427 |
| }, |
| { |
| "epoch": 0.020008932559178204, |
| "grad_norm": 10.019857406616211, |
| "learning_rate": 2.641509433962264e-05, |
| "logits/chosen": -0.6211694478988647, |
| "logits/rejected": -0.6214967370033264, |
| "logps/chosen": -6.935382843017578, |
| "logps/rejected": -13.613279342651367, |
| "loss": 1.2412631511688232, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.5651571154594421, |
| "rewards/accuracies": 0.59375, |
| "rewards/chosen": 0.13799861073493958, |
| "rewards/margins": 0.053141627460718155, |
| "rewards/rejected": 0.08485697209835052, |
| "step": 7, |
| "train_speed(iter/s)": 0.005441 |
| }, |
| { |
| "epoch": 0.022867351496203663, |
| "grad_norm": 6.375326633453369, |
| "learning_rate": 3.018867924528302e-05, |
| "logits/chosen": -0.6666778326034546, |
| "logits/rejected": -0.6678147315979004, |
| "logps/chosen": -8.114872932434082, |
| "logps/rejected": -13.392560958862305, |
| "loss": 1.1162137985229492, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.4446101784706116, |
| "rewards/accuracies": 0.59375, |
| "rewards/chosen": 0.20063960552215576, |
| "rewards/margins": 0.09723334014415741, |
| "rewards/rejected": 0.10340625792741776, |
| "step": 8, |
| "train_speed(iter/s)": 0.005426 |
| }, |
| { |
| "epoch": 0.02572577043322912, |
| "grad_norm": 4.046472549438477, |
| "learning_rate": 3.39622641509434e-05, |
| "logits/chosen": -0.5842044949531555, |
| "logits/rejected": -0.5892153382301331, |
| "logps/chosen": -13.877073287963867, |
| "logps/rejected": -14.116576194763184, |
| "loss": 1.259205937385559, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.5605449676513672, |
| "rewards/accuracies": 0.53125, |
| "rewards/chosen": 0.1392103135585785, |
| "rewards/margins": 0.09627523273229599, |
| "rewards/rejected": 0.0429350808262825, |
| "step": 9, |
| "train_speed(iter/s)": 0.005422 |
| }, |
| { |
| "epoch": 0.028584189370254576, |
| "grad_norm": 7.625533103942871, |
| "learning_rate": 3.7735849056603776e-05, |
| "logits/chosen": -0.6210353374481201, |
| "logits/rejected": -0.6224341988563538, |
| "logps/chosen": -8.32575798034668, |
| "logps/rejected": -11.715011596679688, |
| "loss": 1.2772942781448364, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.48597079515457153, |
| "rewards/accuracies": 0.4375, |
| "rewards/chosen": -0.043191827833652496, |
| "rewards/margins": 0.029505327343940735, |
| "rewards/rejected": -0.07269717007875443, |
| "step": 10, |
| "train_speed(iter/s)": 0.005438 |
| }, |
| { |
| "epoch": 0.03144260830728004, |
| "grad_norm": 11.961934089660645, |
| "learning_rate": 4.150943396226415e-05, |
| "logits/chosen": -0.6265634298324585, |
| "logits/rejected": -0.6289808750152588, |
| "logps/chosen": -7.129470348358154, |
| "logps/rejected": -11.932583808898926, |
| "loss": 1.447466254234314, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.6552329659461975, |
| "rewards/accuracies": 0.46875, |
| "rewards/chosen": -0.00290237320587039, |
| "rewards/margins": 0.06162097677588463, |
| "rewards/rejected": -0.0645233541727066, |
| "step": 11, |
| "train_speed(iter/s)": 0.005473 |
| }, |
| { |
| "epoch": 0.03430102724430549, |
| "grad_norm": 10.154479026794434, |
| "learning_rate": 4.528301886792453e-05, |
| "logits/chosen": -0.6388624906539917, |
| "logits/rejected": -0.639370858669281, |
| "logps/chosen": -9.490407943725586, |
| "logps/rejected": -12.88676643371582, |
| "loss": 1.4875458478927612, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.6468713283538818, |
| "rewards/accuracies": 0.4375, |
| "rewards/chosen": -0.10463112592697144, |
| "rewards/margins": -0.059477299451828, |
| "rewards/rejected": -0.04515381157398224, |
| "step": 12, |
| "train_speed(iter/s)": 0.005465 |
| }, |
| { |
| "epoch": 0.03715944618133095, |
| "grad_norm": 4.091860771179199, |
| "learning_rate": 4.9056603773584906e-05, |
| "logits/chosen": -0.6680133938789368, |
| "logits/rejected": -0.6711105108261108, |
| "logps/chosen": -5.56148099899292, |
| "logps/rejected": -13.865943908691406, |
| "loss": 1.0154482126235962, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.35444653034210205, |
| "rewards/accuracies": 0.65625, |
| "rewards/chosen": 0.12190410494804382, |
| "rewards/margins": 0.3293726444244385, |
| "rewards/rejected": -0.20746850967407227, |
| "step": 13, |
| "train_speed(iter/s)": 0.00546 |
| }, |
| { |
| "epoch": 0.04001786511835641, |
| "grad_norm": 3.039910316467285, |
| "learning_rate": 5.283018867924528e-05, |
| "logits/chosen": -0.6245446801185608, |
| "logits/rejected": -0.6280595660209656, |
| "logps/chosen": -7.042474269866943, |
| "logps/rejected": -14.88280963897705, |
| "loss": 0.9986085295677185, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.32772910594940186, |
| "rewards/accuracies": 0.59375, |
| "rewards/chosen": 0.08946945518255234, |
| "rewards/margins": 0.2548743188381195, |
| "rewards/rejected": -0.16540484130382538, |
| "step": 14, |
| "train_speed(iter/s)": 0.005463 |
| }, |
| { |
| "epoch": 0.042876284055381866, |
| "grad_norm": 2.8388617038726807, |
| "learning_rate": 5.660377358490566e-05, |
| "logits/chosen": -0.613956868648529, |
| "logits/rejected": -0.6150951385498047, |
| "logps/chosen": -5.398593902587891, |
| "logps/rejected": -10.984697341918945, |
| "loss": 0.9217166900634766, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.21436992287635803, |
| "rewards/accuracies": 0.65625, |
| "rewards/chosen": 0.2490297257900238, |
| "rewards/margins": 0.12685683369636536, |
| "rewards/rejected": 0.12217291444540024, |
| "step": 15, |
| "train_speed(iter/s)": 0.005464 |
| }, |
| { |
| "epoch": 0.045734702992407325, |
| "grad_norm": 3.623399257659912, |
| "learning_rate": 6.037735849056604e-05, |
| "logits/chosen": -0.5890456438064575, |
| "logits/rejected": -0.5907201766967773, |
| "logps/chosen": -7.65773868560791, |
| "logps/rejected": -15.24480152130127, |
| "loss": 1.130638837814331, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.33376482129096985, |
| "rewards/accuracies": 0.46875, |
| "rewards/chosen": 0.14319562911987305, |
| "rewards/margins": 0.01443251222372055, |
| "rewards/rejected": 0.1287631392478943, |
| "step": 16, |
| "train_speed(iter/s)": 0.00546 |
| }, |
| { |
| "epoch": 0.048593121929432784, |
| "grad_norm": 2.9850075244903564, |
| "learning_rate": 6.415094339622641e-05, |
| "logits/chosen": -0.5228931903839111, |
| "logits/rejected": -0.524716854095459, |
| "logps/chosen": -10.135262489318848, |
| "logps/rejected": -14.072588920593262, |
| "loss": 1.2770482301712036, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.48515522480010986, |
| "rewards/accuracies": 0.40625, |
| "rewards/chosen": 0.13618236780166626, |
| "rewards/margins": -0.09462341666221619, |
| "rewards/rejected": 0.23080575466156006, |
| "step": 17, |
| "train_speed(iter/s)": 0.005473 |
| }, |
| { |
| "epoch": 0.05145154086645824, |
| "grad_norm": 2.2746193408966064, |
| "learning_rate": 6.79245283018868e-05, |
| "logits/chosen": -0.5764032006263733, |
| "logits/rejected": -0.5762047171592712, |
| "logps/chosen": -11.675098419189453, |
| "logps/rejected": -10.994297981262207, |
| "loss": 1.2022521495819092, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.4671342372894287, |
| "rewards/accuracies": 0.46875, |
| "rewards/chosen": 0.1593710482120514, |
| "rewards/margins": -0.00844080001115799, |
| "rewards/rejected": 0.1678118258714676, |
| "step": 18, |
| "train_speed(iter/s)": 0.005468 |
| }, |
| { |
| "epoch": 0.0543099598034837, |
| "grad_norm": 2.324984312057495, |
| "learning_rate": 7.169811320754717e-05, |
| "logits/chosen": -0.567690908908844, |
| "logits/rejected": -0.569066047668457, |
| "logps/chosen": -9.0728120803833, |
| "logps/rejected": -12.268661499023438, |
| "loss": 1.1060843467712402, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.41300466656684875, |
| "rewards/accuracies": 0.53125, |
| "rewards/chosen": 0.23081830143928528, |
| "rewards/margins": 0.0710553377866745, |
| "rewards/rejected": 0.15976294875144958, |
| "step": 19, |
| "train_speed(iter/s)": 0.005463 |
| }, |
| { |
| "epoch": 0.05716837874050915, |
| "grad_norm": 2.0991458892822266, |
| "learning_rate": 7.547169811320755e-05, |
| "logits/chosen": -0.6091416478157043, |
| "logits/rejected": -0.6100034117698669, |
| "logps/chosen": -4.26762056350708, |
| "logps/rejected": -11.32187557220459, |
| "loss": 1.0021699666976929, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.2933175265789032, |
| "rewards/accuracies": 0.4375, |
| "rewards/chosen": 0.15655732154846191, |
| "rewards/margins": 0.008277412503957748, |
| "rewards/rejected": 0.14827993512153625, |
| "step": 20, |
| "train_speed(iter/s)": 0.005467 |
| }, |
| { |
| "epoch": 0.06002679767753461, |
| "grad_norm": 3.4958455562591553, |
| "learning_rate": 7.924528301886794e-05, |
| "logits/chosen": -0.571921706199646, |
| "logits/rejected": -0.572878360748291, |
| "logps/chosen": -9.612298965454102, |
| "logps/rejected": -12.484085083007812, |
| "loss": 1.1795949935913086, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.46760252118110657, |
| "rewards/accuracies": 0.53125, |
| "rewards/chosen": 0.14643912017345428, |
| "rewards/margins": 0.0200117826461792, |
| "rewards/rejected": 0.12642733752727509, |
| "step": 21, |
| "train_speed(iter/s)": 0.005467 |
| }, |
| { |
| "epoch": 0.06288521661456008, |
| "grad_norm": 3.175546169281006, |
| "learning_rate": 8.30188679245283e-05, |
| "logits/chosen": -0.6108006834983826, |
| "logits/rejected": -0.6111243367195129, |
| "logps/chosen": -6.207424163818359, |
| "logps/rejected": -7.818144798278809, |
| "loss": 1.1549935340881348, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.4683927893638611, |
| "rewards/accuracies": 0.5625, |
| "rewards/chosen": 0.23285990953445435, |
| "rewards/margins": 0.06634702533483505, |
| "rewards/rejected": 0.16651292145252228, |
| "step": 22, |
| "train_speed(iter/s)": 0.00547 |
| }, |
| { |
| "epoch": 0.06574363555158554, |
| "grad_norm": 2.4531869888305664, |
| "learning_rate": 8.679245283018869e-05, |
| "logits/chosen": -0.6012008190155029, |
| "logits/rejected": -0.6021928191184998, |
| "logps/chosen": -5.598664283752441, |
| "logps/rejected": -13.262667655944824, |
| "loss": 0.9260164499282837, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.33532270789146423, |
| "rewards/accuracies": 0.625, |
| "rewards/chosen": 0.3285560607910156, |
| "rewards/margins": 0.3157665431499481, |
| "rewards/rejected": 0.012789532542228699, |
| "step": 23, |
| "train_speed(iter/s)": 0.005467 |
| }, |
| { |
| "epoch": 0.06860205448861098, |
| "grad_norm": 2.328031301498413, |
| "learning_rate": 9.056603773584906e-05, |
| "logits/chosen": -0.567432165145874, |
| "logits/rejected": -0.5704118609428406, |
| "logps/chosen": -4.026985168457031, |
| "logps/rejected": -14.904813766479492, |
| "loss": 0.8984246253967285, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.26223480701446533, |
| "rewards/accuracies": 0.59375, |
| "rewards/chosen": 0.26638248562812805, |
| "rewards/margins": 0.22430934011936188, |
| "rewards/rejected": 0.042073119431734085, |
| "step": 24, |
| "train_speed(iter/s)": 0.005466 |
| }, |
| { |
| "epoch": 0.07146047342563644, |
| "grad_norm": 3.5119378566741943, |
| "learning_rate": 9.433962264150944e-05, |
| "logits/chosen": -0.5709348320960999, |
| "logits/rejected": -0.5688645243644714, |
| "logps/chosen": -7.716845989227295, |
| "logps/rejected": -7.868927955627441, |
| "loss": 1.113297462463379, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.3660878837108612, |
| "rewards/accuracies": 0.5, |
| "rewards/chosen": 0.2273256480693817, |
| "rewards/margins": -0.0022776294499635696, |
| "rewards/rejected": 0.22960326075553894, |
| "step": 25, |
| "train_speed(iter/s)": 0.005475 |
| }, |
| { |
| "epoch": 0.0743188923626619, |
| "grad_norm": 3.2336673736572266, |
| "learning_rate": 9.811320754716981e-05, |
| "logits/chosen": -0.589956521987915, |
| "logits/rejected": -0.587726354598999, |
| "logps/chosen": -10.072855949401855, |
| "logps/rejected": -12.857810974121094, |
| "loss": 1.0991942882537842, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.40110355615615845, |
| "rewards/accuracies": 0.5, |
| "rewards/chosen": 0.17576897144317627, |
| "rewards/margins": 0.11343260109424591, |
| "rewards/rejected": 0.062336355447769165, |
| "step": 26, |
| "train_speed(iter/s)": 0.005466 |
| }, |
| { |
| "epoch": 0.07717731129968736, |
| "grad_norm": 2.6813321113586426, |
| "learning_rate": 0.0001018867924528302, |
| "logits/chosen": -0.5832359194755554, |
| "logits/rejected": -0.5860565304756165, |
| "logps/chosen": -8.197118759155273, |
| "logps/rejected": -17.45244598388672, |
| "loss": 0.8366767764091492, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.2886262536048889, |
| "rewards/accuracies": 0.71875, |
| "rewards/chosen": 0.3297504186630249, |
| "rewards/margins": 0.48295390605926514, |
| "rewards/rejected": -0.15320347249507904, |
| "step": 27, |
| "train_speed(iter/s)": 0.00547 |
| }, |
| { |
| "epoch": 0.08003573023671282, |
| "grad_norm": 3.3800203800201416, |
| "learning_rate": 0.00010566037735849057, |
| "logits/chosen": -0.6005962491035461, |
| "logits/rejected": -0.6035608053207397, |
| "logps/chosen": -9.117904663085938, |
| "logps/rejected": -18.400161743164062, |
| "loss": 1.0143071413040161, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.4276316463947296, |
| "rewards/accuracies": 0.65625, |
| "rewards/chosen": 0.271901398897171, |
| "rewards/margins": 0.4319708049297333, |
| "rewards/rejected": -0.16006940603256226, |
| "step": 28, |
| "train_speed(iter/s)": 0.005468 |
| }, |
| { |
| "epoch": 0.08289414917373827, |
| "grad_norm": 5.461134433746338, |
| "learning_rate": 0.00010943396226415095, |
| "logits/chosen": -0.5596082210540771, |
| "logits/rejected": -0.5635167360305786, |
| "logps/chosen": -7.607837200164795, |
| "logps/rejected": -23.174358367919922, |
| "loss": 0.803045928478241, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.32165294885635376, |
| "rewards/accuracies": 0.78125, |
| "rewards/chosen": 0.3834715783596039, |
| "rewards/margins": 0.7214637994766235, |
| "rewards/rejected": -0.33799219131469727, |
| "step": 29, |
| "train_speed(iter/s)": 0.005464 |
| }, |
| { |
| "epoch": 0.08575256811076373, |
| "grad_norm": 3.3377506732940674, |
| "learning_rate": 0.00011320754716981132, |
| "logits/chosen": -0.6003558039665222, |
| "logits/rejected": -0.6016712188720703, |
| "logps/chosen": -7.941157817840576, |
| "logps/rejected": -10.564404487609863, |
| "loss": 1.0737202167510986, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.3314566910266876, |
| "rewards/accuracies": 0.59375, |
| "rewards/chosen": 0.1379927098751068, |
| "rewards/margins": 0.08670535683631897, |
| "rewards/rejected": 0.05128733813762665, |
| "step": 30, |
| "train_speed(iter/s)": 0.005461 |
| }, |
| { |
| "epoch": 0.08861098704778919, |
| "grad_norm": 3.4214563369750977, |
| "learning_rate": 0.0001169811320754717, |
| "logits/chosen": -0.5186684727668762, |
| "logits/rejected": -0.5189406275749207, |
| "logps/chosen": -6.143683910369873, |
| "logps/rejected": -25.88001251220703, |
| "loss": 0.7792238593101501, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.2514691948890686, |
| "rewards/accuracies": 0.71875, |
| "rewards/chosen": 0.2757991552352905, |
| "rewards/margins": 0.5156725645065308, |
| "rewards/rejected": -0.23987337946891785, |
| "step": 31, |
| "train_speed(iter/s)": 0.005459 |
| }, |
| { |
| "epoch": 0.09146940598481465, |
| "grad_norm": 4.645272731781006, |
| "learning_rate": 0.00012075471698113207, |
| "logits/chosen": -0.5988867282867432, |
| "logits/rejected": -0.6016397476196289, |
| "logps/chosen": -5.776233673095703, |
| "logps/rejected": -22.404544830322266, |
| "loss": 0.8892383575439453, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.4011368155479431, |
| "rewards/accuracies": 0.75, |
| "rewards/chosen": 0.22715790569782257, |
| "rewards/margins": 0.7724226713180542, |
| "rewards/rejected": -0.545264720916748, |
| "step": 32, |
| "train_speed(iter/s)": 0.005454 |
| }, |
| { |
| "epoch": 0.09432782492184011, |
| "grad_norm": 4.586422443389893, |
| "learning_rate": 0.00012452830188679244, |
| "logits/chosen": -0.5782912969589233, |
| "logits/rejected": -0.5768243670463562, |
| "logps/chosen": -10.939208984375, |
| "logps/rejected": -13.486640930175781, |
| "loss": 1.1607584953308105, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.5877886414527893, |
| "rewards/accuracies": 0.65625, |
| "rewards/chosen": 0.21317529678344727, |
| "rewards/margins": 0.4141784906387329, |
| "rewards/rejected": -0.20100319385528564, |
| "step": 33, |
| "train_speed(iter/s)": 0.005457 |
| }, |
| { |
| "epoch": 0.09718624385886557, |
| "grad_norm": 4.306042671203613, |
| "learning_rate": 0.00012830188679245283, |
| "logits/chosen": -0.6019578576087952, |
| "logits/rejected": -0.6082974076271057, |
| "logps/chosen": -7.5170578956604, |
| "logps/rejected": -24.889406204223633, |
| "loss": 0.9509016275405884, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.4247076213359833, |
| "rewards/accuracies": 0.75, |
| "rewards/chosen": 0.10510760545730591, |
| "rewards/margins": 0.8037927150726318, |
| "rewards/rejected": -0.6986850500106812, |
| "step": 34, |
| "train_speed(iter/s)": 0.005456 |
| }, |
| { |
| "epoch": 0.10004466279589103, |
| "grad_norm": 3.002274751663208, |
| "learning_rate": 0.0001320754716981132, |
| "logits/chosen": -0.6133092641830444, |
| "logits/rejected": -0.6127456426620483, |
| "logps/chosen": -8.859593391418457, |
| "logps/rejected": -17.02096939086914, |
| "loss": 0.8848217129707336, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.3480006456375122, |
| "rewards/accuracies": 0.71875, |
| "rewards/chosen": 0.25863486528396606, |
| "rewards/margins": 0.5611037015914917, |
| "rewards/rejected": -0.302468866109848, |
| "step": 35, |
| "train_speed(iter/s)": 0.005457 |
| }, |
| { |
| "epoch": 0.10290308173291648, |
| "grad_norm": 4.299287796020508, |
| "learning_rate": 0.0001358490566037736, |
| "logits/chosen": -0.5725032091140747, |
| "logits/rejected": -0.5725142359733582, |
| "logps/chosen": -8.371686935424805, |
| "logps/rejected": -17.657852172851562, |
| "loss": 0.8825864791870117, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.3603641390800476, |
| "rewards/accuracies": 0.78125, |
| "rewards/chosen": 0.23154468834400177, |
| "rewards/margins": 0.6240193843841553, |
| "rewards/rejected": -0.3924746811389923, |
| "step": 36, |
| "train_speed(iter/s)": 0.005457 |
| }, |
| { |
| "epoch": 0.10576150066994194, |
| "grad_norm": 4.354315757751465, |
| "learning_rate": 0.00013962264150943395, |
| "logits/chosen": -0.5943324565887451, |
| "logits/rejected": -0.5940991044044495, |
| "logps/chosen": -7.657552719116211, |
| "logps/rejected": -16.183917999267578, |
| "loss": 0.8160735964775085, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.3588046133518219, |
| "rewards/accuracies": 0.84375, |
| "rewards/chosen": 0.27544379234313965, |
| "rewards/margins": 0.7471473217010498, |
| "rewards/rejected": -0.4717034101486206, |
| "step": 37, |
| "train_speed(iter/s)": 0.005452 |
| }, |
| { |
| "epoch": 0.1086199196069674, |
| "grad_norm": 7.916697978973389, |
| "learning_rate": 0.00014339622641509434, |
| "logits/chosen": -0.5676888227462769, |
| "logits/rejected": -0.5714080929756165, |
| "logps/chosen": -6.698662757873535, |
| "logps/rejected": -17.954235076904297, |
| "loss": 1.052018642425537, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.49374452233314514, |
| "rewards/accuracies": 0.78125, |
| "rewards/chosen": 0.16338011622428894, |
| "rewards/margins": 0.42942774295806885, |
| "rewards/rejected": -0.2660475969314575, |
| "step": 38, |
| "train_speed(iter/s)": 0.005452 |
| }, |
| { |
| "epoch": 0.11147833854399285, |
| "grad_norm": 3.7271363735198975, |
| "learning_rate": 0.00014716981132075472, |
| "logits/chosen": -0.6097766160964966, |
| "logits/rejected": -0.6135396957397461, |
| "logps/chosen": -7.30006217956543, |
| "logps/rejected": -22.168495178222656, |
| "loss": 0.7729511857032776, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.3125499188899994, |
| "rewards/accuracies": 0.78125, |
| "rewards/chosen": 0.28066012263298035, |
| "rewards/margins": 0.8108583092689514, |
| "rewards/rejected": -0.5301981568336487, |
| "step": 39, |
| "train_speed(iter/s)": 0.00545 |
| }, |
| { |
| "epoch": 0.1143367574810183, |
| "grad_norm": 2.960145950317383, |
| "learning_rate": 0.0001509433962264151, |
| "logits/chosen": -0.6132208108901978, |
| "logits/rejected": -0.6163071990013123, |
| "logps/chosen": -7.024638652801514, |
| "logps/rejected": -21.298852920532227, |
| "loss": 0.8106734752655029, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.37238579988479614, |
| "rewards/accuracies": 0.78125, |
| "rewards/chosen": 0.39458420872688293, |
| "rewards/margins": 0.8476477265357971, |
| "rewards/rejected": -0.45306357741355896, |
| "step": 40, |
| "train_speed(iter/s)": 0.005445 |
| }, |
| { |
| "epoch": 0.11719517641804376, |
| "grad_norm": 4.3203630447387695, |
| "learning_rate": 0.0001547169811320755, |
| "logits/chosen": -0.5976248979568481, |
| "logits/rejected": -0.6002449989318848, |
| "logps/chosen": -5.656615257263184, |
| "logps/rejected": -16.032827377319336, |
| "loss": 0.9440045952796936, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.33253738284111023, |
| "rewards/accuracies": 0.59375, |
| "rewards/chosen": 0.14414586126804352, |
| "rewards/margins": 0.5006409287452698, |
| "rewards/rejected": -0.35649505257606506, |
| "step": 41, |
| "train_speed(iter/s)": 0.005445 |
| }, |
| { |
| "epoch": 0.12005359535506922, |
| "grad_norm": 4.646491527557373, |
| "learning_rate": 0.00015849056603773587, |
| "logits/chosen": -0.5956608653068542, |
| "logits/rejected": -0.5956603288650513, |
| "logps/chosen": -11.719463348388672, |
| "logps/rejected": -27.070579528808594, |
| "loss": 1.0076483488082886, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.5070124268531799, |
| "rewards/accuracies": 0.71875, |
| "rewards/chosen": 0.15291725099086761, |
| "rewards/margins": 0.7710469961166382, |
| "rewards/rejected": -0.6181297898292542, |
| "step": 42, |
| "train_speed(iter/s)": 0.005446 |
| }, |
| { |
| "epoch": 0.12291201429209468, |
| "grad_norm": 4.248917102813721, |
| "learning_rate": 0.00016226415094339625, |
| "logits/chosen": -0.6100150942802429, |
| "logits/rejected": -0.6142972111701965, |
| "logps/chosen": -6.201865196228027, |
| "logps/rejected": -19.82131004333496, |
| "loss": 0.7969534993171692, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.31225427985191345, |
| "rewards/accuracies": 0.6875, |
| "rewards/chosen": 0.20218679308891296, |
| "rewards/margins": 0.705444872379303, |
| "rewards/rejected": -0.5032580494880676, |
| "step": 43, |
| "train_speed(iter/s)": 0.005448 |
| }, |
| { |
| "epoch": 0.12577043322912015, |
| "grad_norm": 3.1864066123962402, |
| "learning_rate": 0.0001660377358490566, |
| "logits/chosen": -0.6078452467918396, |
| "logits/rejected": -0.6100336909294128, |
| "logps/chosen": -8.217609405517578, |
| "logps/rejected": -18.432086944580078, |
| "loss": 0.945781409740448, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.4447736442089081, |
| "rewards/accuracies": 0.8125, |
| "rewards/chosen": 0.16747131943702698, |
| "rewards/margins": 0.6352076530456543, |
| "rewards/rejected": -0.46773630380630493, |
| "step": 44, |
| "train_speed(iter/s)": 0.005451 |
| }, |
| { |
| "epoch": 0.1286288521661456, |
| "grad_norm": 2.8247904777526855, |
| "learning_rate": 0.000169811320754717, |
| "logits/chosen": -0.6424505710601807, |
| "logits/rejected": -0.6455634236335754, |
| "logps/chosen": -4.626930236816406, |
| "logps/rejected": -18.966289520263672, |
| "loss": 0.6525614261627197, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.20949943363666534, |
| "rewards/accuracies": 0.90625, |
| "rewards/chosen": 0.1935967355966568, |
| "rewards/margins": 0.7181073427200317, |
| "rewards/rejected": -0.5245105624198914, |
| "step": 45, |
| "train_speed(iter/s)": 0.005449 |
| }, |
| { |
| "epoch": 0.13148727110317107, |
| "grad_norm": 5.167090892791748, |
| "learning_rate": 0.00017358490566037738, |
| "logits/chosen": -0.6390554308891296, |
| "logits/rejected": -0.6455981731414795, |
| "logps/chosen": -7.457905292510986, |
| "logps/rejected": -22.238759994506836, |
| "loss": 0.7052859663963318, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.3145278990268707, |
| "rewards/accuracies": 0.90625, |
| "rewards/chosen": 0.3406547009944916, |
| "rewards/margins": 0.9184964895248413, |
| "rewards/rejected": -0.5778417587280273, |
| "step": 46, |
| "train_speed(iter/s)": 0.005446 |
| }, |
| { |
| "epoch": 0.1343456900401965, |
| "grad_norm": 3.110797166824341, |
| "learning_rate": 0.00017735849056603776, |
| "logits/chosen": -0.6957284212112427, |
| "logits/rejected": -0.6984925866127014, |
| "logps/chosen": -6.335500717163086, |
| "logps/rejected": -26.946069717407227, |
| "loss": 0.6684616804122925, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.27696001529693604, |
| "rewards/accuracies": 0.78125, |
| "rewards/chosen": 0.36519789695739746, |
| "rewards/margins": 1.261197805404663, |
| "rewards/rejected": -0.8960000276565552, |
| "step": 47, |
| "train_speed(iter/s)": 0.005445 |
| }, |
| { |
| "epoch": 0.13720410897722196, |
| "grad_norm": 5.592209815979004, |
| "learning_rate": 0.00018113207547169812, |
| "logits/chosen": -0.6557337641716003, |
| "logits/rejected": -0.6595371961593628, |
| "logps/chosen": -6.262232780456543, |
| "logps/rejected": -22.550495147705078, |
| "loss": 0.844598650932312, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.32042089104652405, |
| "rewards/accuracies": 0.71875, |
| "rewards/chosen": 0.098446786403656, |
| "rewards/margins": 1.1042643785476685, |
| "rewards/rejected": -1.0058175325393677, |
| "step": 48, |
| "train_speed(iter/s)": 0.005444 |
| }, |
| { |
| "epoch": 0.14006252791424742, |
| "grad_norm": 5.518631458282471, |
| "learning_rate": 0.0001849056603773585, |
| "logits/chosen": -0.7261903285980225, |
| "logits/rejected": -0.7333111763000488, |
| "logps/chosen": -6.155282020568848, |
| "logps/rejected": -28.635177612304688, |
| "loss": 0.7839219570159912, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.3167538046836853, |
| "rewards/accuracies": 0.75, |
| "rewards/chosen": 0.18123595416545868, |
| "rewards/margins": 1.343546986579895, |
| "rewards/rejected": -1.1623109579086304, |
| "step": 49, |
| "train_speed(iter/s)": 0.00544 |
| }, |
| { |
| "epoch": 0.14292094685127288, |
| "grad_norm": 5.368433475494385, |
| "learning_rate": 0.00018867924528301889, |
| "logits/chosen": -0.7460846304893494, |
| "logits/rejected": -0.7433331608772278, |
| "logps/chosen": -8.476225852966309, |
| "logps/rejected": -20.479339599609375, |
| "loss": 1.041689395904541, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.5162184834480286, |
| "rewards/accuracies": 0.6875, |
| "rewards/chosen": 0.07575411349534988, |
| "rewards/margins": 0.8544278144836426, |
| "rewards/rejected": -0.7786736488342285, |
| "step": 50, |
| "train_speed(iter/s)": 0.005439 |
| }, |
| { |
| "epoch": 0.14577936578829834, |
| "grad_norm": 12.666760444641113, |
| "learning_rate": 0.00019245283018867927, |
| "logits/chosen": -0.7423916459083557, |
| "logits/rejected": -0.7536158561706543, |
| "logps/chosen": -4.518327713012695, |
| "logps/rejected": -25.71565818786621, |
| "loss": 0.8285016417503357, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.38353338837623596, |
| "rewards/accuracies": 0.6875, |
| "rewards/chosen": 0.25002920627593994, |
| "rewards/margins": 1.0534642934799194, |
| "rewards/rejected": -0.8034350872039795, |
| "step": 51, |
| "train_speed(iter/s)": 0.005432 |
| }, |
| { |
| "epoch": 0.1486377847253238, |
| "grad_norm": 5.63993501663208, |
| "learning_rate": 0.00019622641509433963, |
| "logits/chosen": -0.7661195993423462, |
| "logits/rejected": -0.7684649229049683, |
| "logps/chosen": -5.303972244262695, |
| "logps/rejected": -20.424297332763672, |
| "loss": 0.7760183811187744, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.30550119280815125, |
| "rewards/accuracies": 0.84375, |
| "rewards/chosen": 0.17865437269210815, |
| "rewards/margins": 0.9522275328636169, |
| "rewards/rejected": -0.7735730409622192, |
| "step": 52, |
| "train_speed(iter/s)": 0.005436 |
| }, |
| { |
| "epoch": 0.15149620366234925, |
| "grad_norm": 3.208770275115967, |
| "learning_rate": 0.0002, |
| "logits/chosen": -0.7615375518798828, |
| "logits/rejected": -0.7639837861061096, |
| "logps/chosen": -7.793338775634766, |
| "logps/rejected": -22.267044067382812, |
| "loss": 0.6889612078666687, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.2784683108329773, |
| "rewards/accuracies": 0.78125, |
| "rewards/chosen": 0.21318240463733673, |
| "rewards/margins": 1.1274638175964355, |
| "rewards/rejected": -0.9142814874649048, |
| "step": 53, |
| "train_speed(iter/s)": 0.005436 |
| }, |
| { |
| "epoch": 0.1543546225993747, |
| "grad_norm": 4.603616714477539, |
| "learning_rate": 0.00019999950054470762, |
| "logits/chosen": -0.7416525483131409, |
| "logits/rejected": -0.7468435168266296, |
| "logps/chosen": -11.753125190734863, |
| "logps/rejected": -23.24603271484375, |
| "loss": 0.9534075856208801, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.42802760004997253, |
| "rewards/accuracies": 0.65625, |
| "rewards/chosen": 0.2138495296239853, |
| "rewards/margins": 0.8852367997169495, |
| "rewards/rejected": -0.6713871955871582, |
| "step": 54, |
| "train_speed(iter/s)": 0.005437 |
| }, |
| { |
| "epoch": 0.15721304153640017, |
| "grad_norm": 3.226914405822754, |
| "learning_rate": 0.00019999800218381956, |
| "logits/chosen": -0.782919704914093, |
| "logits/rejected": -0.7837254405021667, |
| "logps/chosen": -10.437933921813965, |
| "logps/rejected": -21.950359344482422, |
| "loss": 0.7319959998130798, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.28649967908859253, |
| "rewards/accuracies": 0.78125, |
| "rewards/chosen": 0.17146320641040802, |
| "rewards/margins": 0.9873074889183044, |
| "rewards/rejected": -0.8158442974090576, |
| "step": 55, |
| "train_speed(iter/s)": 0.005435 |
| }, |
| { |
| "epoch": 0.16007146047342563, |
| "grad_norm": 5.957598686218262, |
| "learning_rate": 0.00019999550493230315, |
| "logits/chosen": -0.7622566819190979, |
| "logits/rejected": -0.7720569968223572, |
| "logps/chosen": -9.299139022827148, |
| "logps/rejected": -25.902114868164062, |
| "loss": 1.0733510255813599, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.5738880634307861, |
| "rewards/accuracies": 0.75, |
| "rewards/chosen": 0.1327313929796219, |
| "rewards/margins": 1.0297049283981323, |
| "rewards/rejected": -0.8969735503196716, |
| "step": 56, |
| "train_speed(iter/s)": 0.005434 |
| }, |
| { |
| "epoch": 0.1629298794104511, |
| "grad_norm": 4.4198760986328125, |
| "learning_rate": 0.00019999200881510367, |
| "logits/chosen": -0.7777268886566162, |
| "logits/rejected": -0.7817560434341431, |
| "logps/chosen": -8.051362037658691, |
| "logps/rejected": -16.21352767944336, |
| "loss": 0.9994087815284729, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.5157659649848938, |
| "rewards/accuracies": 0.71875, |
| "rewards/chosen": 0.3178602159023285, |
| "rewards/margins": 0.7738386392593384, |
| "rewards/rejected": -0.4559784233570099, |
| "step": 57, |
| "train_speed(iter/s)": 0.005437 |
| }, |
| { |
| "epoch": 0.16578829834747655, |
| "grad_norm": 10.918800354003906, |
| "learning_rate": 0.00019998751386714417, |
| "logits/chosen": -0.7735524773597717, |
| "logits/rejected": -0.7762777209281921, |
| "logps/chosen": -6.026726245880127, |
| "logps/rejected": -20.445030212402344, |
| "loss": 0.8424608707427979, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.4571954309940338, |
| "rewards/accuracies": 0.875, |
| "rewards/chosen": 0.3362487852573395, |
| "rewards/margins": 1.0223908424377441, |
| "rewards/rejected": -0.6861420273780823, |
| "step": 58, |
| "train_speed(iter/s)": 0.005438 |
| }, |
| { |
| "epoch": 0.168646717284502, |
| "grad_norm": 6.003457546234131, |
| "learning_rate": 0.00019998202013332526, |
| "logits/chosen": -0.7671990394592285, |
| "logits/rejected": -0.7679510712623596, |
| "logps/chosen": -7.208020210266113, |
| "logps/rejected": -16.45644187927246, |
| "loss": 0.8400717377662659, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.3789638876914978, |
| "rewards/accuracies": 0.8125, |
| "rewards/chosen": 0.19236475229263306, |
| "rewards/margins": 0.8348343372344971, |
| "rewards/rejected": -0.6424695253372192, |
| "step": 59, |
| "train_speed(iter/s)": 0.005439 |
| }, |
| { |
| "epoch": 0.17150513622152747, |
| "grad_norm": 2.5898213386535645, |
| "learning_rate": 0.00019997552766852432, |
| "logits/chosen": -0.800130307674408, |
| "logits/rejected": -0.8043981790542603, |
| "logps/chosen": -6.655481815338135, |
| "logps/rejected": -24.4234619140625, |
| "loss": 0.6945982575416565, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.3340861201286316, |
| "rewards/accuracies": 0.875, |
| "rewards/chosen": 0.3088517487049103, |
| "rewards/margins": 1.2354176044464111, |
| "rewards/rejected": -0.9265658855438232, |
| "step": 60, |
| "train_speed(iter/s)": 0.005436 |
| }, |
| { |
| "epoch": 0.17436355515855292, |
| "grad_norm": 3.626788377761841, |
| "learning_rate": 0.00019996803653759532, |
| "logits/chosen": -0.7841392159461975, |
| "logits/rejected": -0.7904000282287598, |
| "logps/chosen": -4.378767013549805, |
| "logps/rejected": -26.130855560302734, |
| "loss": 0.5890128016471863, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.2802823781967163, |
| "rewards/accuracies": 0.9375, |
| "rewards/chosen": 0.2776252329349518, |
| "rewards/margins": 1.491828203201294, |
| "rewards/rejected": -1.2142030000686646, |
| "step": 61, |
| "train_speed(iter/s)": 0.005438 |
| }, |
| { |
| "epoch": 0.17722197409557838, |
| "grad_norm": 6.747097492218018, |
| "learning_rate": 0.00019995954681536798, |
| "logits/chosen": -0.7469812035560608, |
| "logits/rejected": -0.7426640391349792, |
| "logps/chosen": -11.407801628112793, |
| "logps/rejected": -19.764530181884766, |
| "loss": 1.1509681940078735, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.4816472828388214, |
| "rewards/accuracies": 0.71875, |
| "rewards/chosen": -0.023781299591064453, |
| "rewards/margins": 0.7125381231307983, |
| "rewards/rejected": -0.7363194227218628, |
| "step": 62, |
| "train_speed(iter/s)": 0.005439 |
| }, |
| { |
| "epoch": 0.18008039303260384, |
| "grad_norm": 6.9256978034973145, |
| "learning_rate": 0.00019995005858664696, |
| "logits/chosen": -0.819698691368103, |
| "logits/rejected": -0.8221041560173035, |
| "logps/chosen": -7.0722761154174805, |
| "logps/rejected": -30.66091537475586, |
| "loss": 0.9850404262542725, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.42992380261421204, |
| "rewards/accuracies": 0.75, |
| "rewards/chosen": 0.07954245805740356, |
| "rewards/margins": 1.697967767715454, |
| "rewards/rejected": -1.6184254884719849, |
| "step": 63, |
| "train_speed(iter/s)": 0.005439 |
| }, |
| { |
| "epoch": 0.1829388119696293, |
| "grad_norm": 5.373042106628418, |
| "learning_rate": 0.00019993957194621128, |
| "logits/chosen": -0.8020437955856323, |
| "logits/rejected": -0.8022991418838501, |
| "logps/chosen": -12.086634635925293, |
| "logps/rejected": -21.337501525878906, |
| "loss": 0.9992907643318176, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.46714556217193604, |
| "rewards/accuracies": 0.78125, |
| "rewards/chosen": 0.11581222712993622, |
| "rewards/margins": 1.0597875118255615, |
| "rewards/rejected": -0.9439753293991089, |
| "step": 64, |
| "train_speed(iter/s)": 0.005441 |
| }, |
| { |
| "epoch": 0.18579723090665476, |
| "grad_norm": 3.77294921875, |
| "learning_rate": 0.00019992808699881303, |
| "logits/chosen": -0.7432379722595215, |
| "logits/rejected": -0.7440240979194641, |
| "logps/chosen": -6.54131555557251, |
| "logps/rejected": -25.4437255859375, |
| "loss": 0.8025262951850891, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.4345034658908844, |
| "rewards/accuracies": 0.90625, |
| "rewards/chosen": 0.1653081476688385, |
| "rewards/margins": 1.261300802230835, |
| "rewards/rejected": -1.0959926843643188, |
| "step": 65, |
| "train_speed(iter/s)": 0.005442 |
| }, |
| { |
| "epoch": 0.18865564984368022, |
| "grad_norm": 3.9820404052734375, |
| "learning_rate": 0.0001999156038591766, |
| "logits/chosen": -0.754156231880188, |
| "logits/rejected": -0.7619881629943848, |
| "logps/chosen": -6.165823459625244, |
| "logps/rejected": -23.689210891723633, |
| "loss": 0.7558284997940063, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.35462892055511475, |
| "rewards/accuracies": 0.84375, |
| "rewards/chosen": 0.18967051804065704, |
| "rewards/margins": 0.9685844779014587, |
| "rewards/rejected": -0.7789139151573181, |
| "step": 66, |
| "train_speed(iter/s)": 0.005441 |
| }, |
| { |
| "epoch": 0.19151406878070568, |
| "grad_norm": 4.361135482788086, |
| "learning_rate": 0.00019990212265199738, |
| "logits/chosen": -0.7575560212135315, |
| "logits/rejected": -0.7635327577590942, |
| "logps/chosen": -4.563057899475098, |
| "logps/rejected": -24.856014251708984, |
| "loss": 0.7012404203414917, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.3292723000049591, |
| "rewards/accuracies": 0.84375, |
| "rewards/chosen": 0.360738068819046, |
| "rewards/margins": 1.0550967454910278, |
| "rewards/rejected": -0.6943586468696594, |
| "step": 67, |
| "train_speed(iter/s)": 0.005441 |
| }, |
| { |
| "epoch": 0.19437248771773113, |
| "grad_norm": 8.01318359375, |
| "learning_rate": 0.00019988764351194056, |
| "logits/chosen": -0.7201259136199951, |
| "logits/rejected": -0.7213173508644104, |
| "logps/chosen": -14.019023895263672, |
| "logps/rejected": -23.463457107543945, |
| "loss": 1.155617594718933, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.6643218994140625, |
| "rewards/accuracies": 0.78125, |
| "rewards/chosen": 0.16900399327278137, |
| "rewards/margins": 0.8244603276252747, |
| "rewards/rejected": -0.6554563641548157, |
| "step": 68, |
| "train_speed(iter/s)": 0.005442 |
| }, |
| { |
| "epoch": 0.1972309066547566, |
| "grad_norm": 4.880486488342285, |
| "learning_rate": 0.00019987216658363984, |
| "logits/chosen": -0.7603176832199097, |
| "logits/rejected": -0.7611086368560791, |
| "logps/chosen": -7.871481418609619, |
| "logps/rejected": -18.355331420898438, |
| "loss": 0.9050899744033813, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.4671905040740967, |
| "rewards/accuracies": 0.90625, |
| "rewards/chosen": 0.14834776520729065, |
| "rewards/margins": 0.9282243251800537, |
| "rewards/rejected": -0.7798765897750854, |
| "step": 69, |
| "train_speed(iter/s)": 0.005441 |
| }, |
| { |
| "epoch": 0.20008932559178205, |
| "grad_norm": 36.09653854370117, |
| "learning_rate": 0.00019985569202169584, |
| "logits/chosen": -0.826109766960144, |
| "logits/rejected": -0.8319188952445984, |
| "logps/chosen": -7.292363166809082, |
| "logps/rejected": -19.279315948486328, |
| "loss": 0.906982421875, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.4692971408367157, |
| "rewards/accuracies": 0.8125, |
| "rewards/chosen": 0.17047010362148285, |
| "rewards/margins": 0.9899492859840393, |
| "rewards/rejected": -0.8194791674613953, |
| "step": 70, |
| "train_speed(iter/s)": 0.005443 |
| }, |
| { |
| "epoch": 0.2029477445288075, |
| "grad_norm": 5.051456451416016, |
| "learning_rate": 0.00019983821999067475, |
| "logits/chosen": -0.7659405469894409, |
| "logits/rejected": -0.773281991481781, |
| "logps/chosen": -7.996345520019531, |
| "logps/rejected": -16.987951278686523, |
| "loss": 0.9166132211685181, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.43807780742645264, |
| "rewards/accuracies": 0.71875, |
| "rewards/chosen": 0.2428373396396637, |
| "rewards/margins": 0.7172515988349915, |
| "rewards/rejected": -0.47441428899765015, |
| "step": 71, |
| "train_speed(iter/s)": 0.005444 |
| }, |
| { |
| "epoch": 0.20580616346583297, |
| "grad_norm": 3.3546969890594482, |
| "learning_rate": 0.00019981975066510655, |
| "logits/chosen": -0.7702159285545349, |
| "logits/rejected": -0.7756153345108032, |
| "logps/chosen": -3.096975088119507, |
| "logps/rejected": -24.7138671875, |
| "loss": 0.6350318193435669, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.20447328686714172, |
| "rewards/accuracies": 0.8125, |
| "rewards/chosen": 0.23610374331474304, |
| "rewards/margins": 1.1568939685821533, |
| "rewards/rejected": -0.9207903146743774, |
| "step": 72, |
| "train_speed(iter/s)": 0.005441 |
| }, |
| { |
| "epoch": 0.20866458240285843, |
| "grad_norm": 3.3056812286376953, |
| "learning_rate": 0.00019980028422948322, |
| "logits/chosen": -0.7785294651985168, |
| "logits/rejected": -0.7796329259872437, |
| "logps/chosen": -6.459807395935059, |
| "logps/rejected": -18.2646541595459, |
| "loss": 0.7826670408248901, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.299204021692276, |
| "rewards/accuracies": 0.71875, |
| "rewards/chosen": 0.17048794031143188, |
| "rewards/margins": 0.8538780212402344, |
| "rewards/rejected": -0.6833901405334473, |
| "step": 73, |
| "train_speed(iter/s)": 0.005441 |
| }, |
| { |
| "epoch": 0.2115230013398839, |
| "grad_norm": 2.804248094558716, |
| "learning_rate": 0.00019977982087825713, |
| "logits/chosen": -0.7178574204444885, |
| "logits/rejected": -0.7222320437431335, |
| "logps/chosen": -5.382746696472168, |
| "logps/rejected": -21.648345947265625, |
| "loss": 0.6176108717918396, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.2642488479614258, |
| "rewards/accuracies": 0.90625, |
| "rewards/chosen": 0.3241485357284546, |
| "rewards/margins": 1.1783626079559326, |
| "rewards/rejected": -0.854214072227478, |
| "step": 74, |
| "train_speed(iter/s)": 0.005441 |
| }, |
| { |
| "epoch": 0.21438142027690935, |
| "grad_norm": 2.3271028995513916, |
| "learning_rate": 0.0001997583608158388, |
| "logits/chosen": -0.7665284872055054, |
| "logits/rejected": -0.7763013243675232, |
| "logps/chosen": -3.597620725631714, |
| "logps/rejected": -28.286640167236328, |
| "loss": 0.4727986454963684, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.1368505358695984, |
| "rewards/accuracies": 0.90625, |
| "rewards/chosen": 0.36978626251220703, |
| "rewards/margins": 1.460551142692566, |
| "rewards/rejected": -1.0907649993896484, |
| "step": 75, |
| "train_speed(iter/s)": 0.005442 |
| }, |
| { |
| "epoch": 0.2172398392139348, |
| "grad_norm": 5.8310160636901855, |
| "learning_rate": 0.00019973590425659507, |
| "logits/chosen": -0.7936246991157532, |
| "logits/rejected": -0.7950795888900757, |
| "logps/chosen": -8.558622360229492, |
| "logps/rejected": -22.39399528503418, |
| "loss": 0.7566131949424744, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.4084041714668274, |
| "rewards/accuracies": 0.875, |
| "rewards/chosen": 0.26708710193634033, |
| "rewards/margins": 1.3221920728683472, |
| "rewards/rejected": -1.0551048517227173, |
| "step": 76, |
| "train_speed(iter/s)": 0.00544 |
| }, |
| { |
| "epoch": 0.22009825815096026, |
| "grad_norm": 4.89293909072876, |
| "learning_rate": 0.0001997124514248469, |
| "logits/chosen": -0.8093288540840149, |
| "logits/rejected": -0.8106516003608704, |
| "logps/chosen": -12.884543418884277, |
| "logps/rejected": -22.5157413482666, |
| "loss": 1.130676507949829, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.7112829685211182, |
| "rewards/accuracies": 0.8125, |
| "rewards/chosen": 0.10567857325077057, |
| "rewards/margins": 1.2312006950378418, |
| "rewards/rejected": -1.1255221366882324, |
| "step": 77, |
| "train_speed(iter/s)": 0.005441 |
| }, |
| { |
| "epoch": 0.2229566770879857, |
| "grad_norm": 7.881620407104492, |
| "learning_rate": 0.00019968800255486713, |
| "logits/chosen": -0.8169571161270142, |
| "logits/rejected": -0.8179275393486023, |
| "logps/chosen": -6.7093825340271, |
| "logps/rejected": -29.32290267944336, |
| "loss": 0.8154858350753784, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.3944161534309387, |
| "rewards/accuracies": 0.8125, |
| "rewards/chosen": 0.16114036738872528, |
| "rewards/margins": 1.4641916751861572, |
| "rewards/rejected": -1.3030513525009155, |
| "step": 78, |
| "train_speed(iter/s)": 0.005441 |
| }, |
| { |
| "epoch": 0.22581509602501115, |
| "grad_norm": 2.513998031616211, |
| "learning_rate": 0.00019966255789087808, |
| "logits/chosen": -0.8547464609146118, |
| "logits/rejected": -0.8608615398406982, |
| "logps/chosen": -7.297125339508057, |
| "logps/rejected": -27.81803321838379, |
| "loss": 0.6003709435462952, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.29905110597610474, |
| "rewards/accuracies": 0.9375, |
| "rewards/chosen": 0.13682395219802856, |
| "rewards/margins": 1.487520456314087, |
| "rewards/rejected": -1.3506965637207031, |
| "step": 79, |
| "train_speed(iter/s)": 0.005437 |
| }, |
| { |
| "epoch": 0.2286735149620366, |
| "grad_norm": 22.424015045166016, |
| "learning_rate": 0.0001996361176870492, |
| "logits/chosen": -0.8499366044998169, |
| "logits/rejected": -0.863770604133606, |
| "logps/chosen": -8.103764533996582, |
| "logps/rejected": -35.18703842163086, |
| "loss": 0.7495014071464539, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.5106410980224609, |
| "rewards/accuracies": 0.9375, |
| "rewards/chosen": 0.2272297590970993, |
| "rewards/margins": 1.9457027912139893, |
| "rewards/rejected": -1.7184730768203735, |
| "step": 80, |
| "train_speed(iter/s)": 0.005436 |
| }, |
| { |
| "epoch": 0.23153193389906207, |
| "grad_norm": 4.0974273681640625, |
| "learning_rate": 0.00019960868220749448, |
| "logits/chosen": -0.8633654117584229, |
| "logits/rejected": -0.8727495670318604, |
| "logps/chosen": -5.683876991271973, |
| "logps/rejected": -34.967037200927734, |
| "loss": 0.6187952756881714, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.32927608489990234, |
| "rewards/accuracies": 0.90625, |
| "rewards/chosen": 0.2094997614622116, |
| "rewards/margins": 2.0794010162353516, |
| "rewards/rejected": -1.8699012994766235, |
| "step": 81, |
| "train_speed(iter/s)": 0.005436 |
| }, |
| { |
| "epoch": 0.23439035283608753, |
| "grad_norm": 3.275705575942993, |
| "learning_rate": 0.00019958025172626986, |
| "logits/chosen": -0.8937872648239136, |
| "logits/rejected": -0.9034286737442017, |
| "logps/chosen": -11.346511840820312, |
| "logps/rejected": -35.471927642822266, |
| "loss": 0.7466462850570679, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.4821043014526367, |
| "rewards/accuracies": 0.9375, |
| "rewards/chosen": 0.2012697458267212, |
| "rewards/margins": 1.9956668615341187, |
| "rewards/rejected": -1.7943971157073975, |
| "step": 82, |
| "train_speed(iter/s)": 0.005436 |
| }, |
| { |
| "epoch": 0.237248771773113, |
| "grad_norm": 8.550602912902832, |
| "learning_rate": 0.0001995508265273704, |
| "logits/chosen": -0.9247998595237732, |
| "logits/rejected": -0.934203028678894, |
| "logps/chosen": -7.0234270095825195, |
| "logps/rejected": -40.43362808227539, |
| "loss": 0.6714224219322205, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.36930039525032043, |
| "rewards/accuracies": 0.84375, |
| "rewards/chosen": -0.015845157206058502, |
| "rewards/margins": 2.3306310176849365, |
| "rewards/rejected": -2.3464760780334473, |
| "step": 83, |
| "train_speed(iter/s)": 0.005436 |
| }, |
| { |
| "epoch": 0.24010719071013845, |
| "grad_norm": 10.380730628967285, |
| "learning_rate": 0.00019952040690472748, |
| "logits/chosen": -0.89071124792099, |
| "logits/rejected": -0.896521806716919, |
| "logps/chosen": -9.274986267089844, |
| "logps/rejected": -31.167112350463867, |
| "loss": 0.9556330442428589, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.4726044833660126, |
| "rewards/accuracies": 0.78125, |
| "rewards/chosen": 0.055123914033174515, |
| "rewards/margins": 1.99386465549469, |
| "rewards/rejected": -1.9387409687042236, |
| "step": 84, |
| "train_speed(iter/s)": 0.005435 |
| }, |
| { |
| "epoch": 0.2429656096471639, |
| "grad_norm": 59.343536376953125, |
| "learning_rate": 0.00019948899316220604, |
| "logits/chosen": -0.86629319190979, |
| "logits/rejected": -0.8767050504684448, |
| "logps/chosen": -6.04016637802124, |
| "logps/rejected": -28.20957374572754, |
| "loss": 0.725101888179779, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.40447694063186646, |
| "rewards/accuracies": 0.8125, |
| "rewards/chosen": 0.29112857580184937, |
| "rewards/margins": 1.845672369003296, |
| "rewards/rejected": -1.5545437335968018, |
| "step": 85, |
| "train_speed(iter/s)": 0.005436 |
| }, |
| { |
| "epoch": 0.24582402858418936, |
| "grad_norm": 6.300822734832764, |
| "learning_rate": 0.0001994565856136012, |
| "logits/chosen": -0.8398523330688477, |
| "logits/rejected": -0.8439276814460754, |
| "logps/chosen": -7.2902045249938965, |
| "logps/rejected": -32.82334899902344, |
| "loss": 0.5805257558822632, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.4190253019332886, |
| "rewards/accuracies": 0.9375, |
| "rewards/chosen": 0.3955605626106262, |
| "rewards/margins": 2.423994302749634, |
| "rewards/rejected": -2.0284337997436523, |
| "step": 86, |
| "train_speed(iter/s)": 0.005436 |
| }, |
| { |
| "epoch": 0.24868244752121482, |
| "grad_norm": 43.768619537353516, |
| "learning_rate": 0.0001994231845826354, |
| "logits/chosen": -0.8710224628448486, |
| "logits/rejected": -0.8721168637275696, |
| "logps/chosen": -7.397428035736084, |
| "logps/rejected": -29.24224853515625, |
| "loss": 1.1686397790908813, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.7285480499267578, |
| "rewards/accuracies": 0.8125, |
| "rewards/chosen": -0.11831004917621613, |
| "rewards/margins": 1.2224053144454956, |
| "rewards/rejected": -1.3407154083251953, |
| "step": 87, |
| "train_speed(iter/s)": 0.005436 |
| }, |
| { |
| "epoch": 0.2515408664582403, |
| "grad_norm": 39.749202728271484, |
| "learning_rate": 0.00019938879040295508, |
| "logits/chosen": -0.8551483750343323, |
| "logits/rejected": -0.8542779684066772, |
| "logps/chosen": -10.534215927124023, |
| "logps/rejected": -26.982810974121094, |
| "loss": 1.2515333890914917, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.7181237936019897, |
| "rewards/accuracies": 0.71875, |
| "rewards/chosen": -0.03403427451848984, |
| "rewards/margins": 1.2902536392211914, |
| "rewards/rejected": -1.32428777217865, |
| "step": 88, |
| "train_speed(iter/s)": 0.005437 |
| }, |
| { |
| "epoch": 0.25439928539526574, |
| "grad_norm": 10.098331451416016, |
| "learning_rate": 0.00019935340341812737, |
| "logits/chosen": -0.8135808706283569, |
| "logits/rejected": -0.8137494325637817, |
| "logps/chosen": -9.00113582611084, |
| "logps/rejected": -22.338396072387695, |
| "loss": 0.9040060639381409, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.5397258400917053, |
| "rewards/accuracies": 0.84375, |
| "rewards/chosen": 0.04357346147298813, |
| "rewards/margins": 1.2031430006027222, |
| "rewards/rejected": -1.159569501876831, |
| "step": 89, |
| "train_speed(iter/s)": 0.005435 |
| }, |
| { |
| "epoch": 0.2572577043322912, |
| "grad_norm": 8.489130973815918, |
| "learning_rate": 0.00019931702398163657, |
| "logits/chosen": -0.8590521216392517, |
| "logits/rejected": -0.8651595711708069, |
| "logps/chosen": -8.849437713623047, |
| "logps/rejected": -27.54432487487793, |
| "loss": 0.8707167506217957, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.48353832960128784, |
| "rewards/accuracies": 0.8125, |
| "rewards/chosen": 0.06561100482940674, |
| "rewards/margins": 1.1353689432144165, |
| "rewards/rejected": -1.0697578191757202, |
| "step": 90, |
| "train_speed(iter/s)": 0.005437 |
| }, |
| { |
| "epoch": 0.26011612326931666, |
| "grad_norm": 5.018461227416992, |
| "learning_rate": 0.00019927965245688072, |
| "logits/chosen": -0.8591816425323486, |
| "logits/rejected": -0.8660668730735779, |
| "logps/chosen": -11.013117790222168, |
| "logps/rejected": -27.660642623901367, |
| "loss": 0.9352516531944275, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.541546106338501, |
| "rewards/accuracies": 0.78125, |
| "rewards/chosen": 0.2031691074371338, |
| "rewards/margins": 1.437905192375183, |
| "rewards/rejected": -1.2347362041473389, |
| "step": 91, |
| "train_speed(iter/s)": 0.005437 |
| }, |
| { |
| "epoch": 0.26297454220634214, |
| "grad_norm": 6.94728946685791, |
| "learning_rate": 0.00019924128921716797, |
| "logits/chosen": -0.8400283455848694, |
| "logits/rejected": -0.8450547456741333, |
| "logps/chosen": -9.49151611328125, |
| "logps/rejected": -28.902881622314453, |
| "loss": 0.806583821773529, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.3471754193305969, |
| "rewards/accuracies": 0.78125, |
| "rewards/chosen": 0.21784308552742004, |
| "rewards/margins": 1.679865837097168, |
| "rewards/rejected": -1.4620226621627808, |
| "step": 92, |
| "train_speed(iter/s)": 0.005436 |
| }, |
| { |
| "epoch": 0.2658329611433676, |
| "grad_norm": 2.8959178924560547, |
| "learning_rate": 0.00019920193464571275, |
| "logits/chosen": -0.839047372341156, |
| "logits/rejected": -0.8450105786323547, |
| "logps/chosen": -10.59231948852539, |
| "logps/rejected": -29.6372013092041, |
| "loss": 0.7759805917739868, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.5248351097106934, |
| "rewards/accuracies": 0.90625, |
| "rewards/chosen": 0.3599660098552704, |
| "rewards/margins": 1.950143575668335, |
| "rewards/rejected": -1.5901774168014526, |
| "step": 93, |
| "train_speed(iter/s)": 0.005437 |
| }, |
| { |
| "epoch": 0.268691380080393, |
| "grad_norm": 5.318148612976074, |
| "learning_rate": 0.00019916158913563207, |
| "logits/chosen": -0.8162392377853394, |
| "logits/rejected": -0.8246886730194092, |
| "logps/chosen": -5.6260833740234375, |
| "logps/rejected": -29.538055419921875, |
| "loss": 0.6956499218940735, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.38102442026138306, |
| "rewards/accuracies": 0.84375, |
| "rewards/chosen": 0.2292197048664093, |
| "rewards/margins": 1.7469617128372192, |
| "rewards/rejected": -1.5177420377731323, |
| "step": 94, |
| "train_speed(iter/s)": 0.005437 |
| }, |
| { |
| "epoch": 0.2715497990174185, |
| "grad_norm": 4.5264787673950195, |
| "learning_rate": 0.00019912025308994148, |
| "logits/chosen": -0.8437933325767517, |
| "logits/rejected": -0.8463299870491028, |
| "logps/chosen": -8.831324577331543, |
| "logps/rejected": -26.610977172851562, |
| "loss": 0.7661141157150269, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.4225902557373047, |
| "rewards/accuracies": 0.875, |
| "rewards/chosen": 0.12178793549537659, |
| "rewards/margins": 1.467708945274353, |
| "rewards/rejected": -1.3459210395812988, |
| "step": 95, |
| "train_speed(iter/s)": 0.005436 |
| }, |
| { |
| "epoch": 0.2744082179544439, |
| "grad_norm": 5.978212356567383, |
| "learning_rate": 0.00019907792692155113, |
| "logits/chosen": -0.7815445065498352, |
| "logits/rejected": -0.78977370262146, |
| "logps/chosen": -6.673672199249268, |
| "logps/rejected": -33.78847122192383, |
| "loss": 0.7695144414901733, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.5254613757133484, |
| "rewards/accuracies": 0.9375, |
| "rewards/chosen": 0.2026345133781433, |
| "rewards/margins": 1.8294543027877808, |
| "rewards/rejected": -1.6268197298049927, |
| "step": 96, |
| "train_speed(iter/s)": 0.005435 |
| }, |
| { |
| "epoch": 0.2772666368914694, |
| "grad_norm": 9.170755386352539, |
| "learning_rate": 0.00019903461105326154, |
| "logits/chosen": -0.8260455131530762, |
| "logits/rejected": -0.830398678779602, |
| "logps/chosen": -6.689513206481934, |
| "logps/rejected": -22.671979904174805, |
| "loss": 0.7065067291259766, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.32698267698287964, |
| "rewards/accuracies": 0.8125, |
| "rewards/chosen": 0.2753945589065552, |
| "rewards/margins": 1.3633196353912354, |
| "rewards/rejected": -1.0879249572753906, |
| "step": 97, |
| "train_speed(iter/s)": 0.005435 |
| }, |
| { |
| "epoch": 0.28012505582849484, |
| "grad_norm": 3.521716594696045, |
| "learning_rate": 0.00019899030591775957, |
| "logits/chosen": -0.7963160872459412, |
| "logits/rejected": -0.8014420866966248, |
| "logps/chosen": -7.337916851043701, |
| "logps/rejected": -28.08383560180664, |
| "loss": 0.61917644739151, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.3110499680042267, |
| "rewards/accuracies": 0.84375, |
| "rewards/chosen": 0.23168900609016418, |
| "rewards/margins": 1.7576744556427002, |
| "rewards/rejected": -1.525985598564148, |
| "step": 98, |
| "train_speed(iter/s)": 0.005436 |
| }, |
| { |
| "epoch": 0.2829834747655203, |
| "grad_norm": 5.793614864349365, |
| "learning_rate": 0.00019894501195761392, |
| "logits/chosen": -0.8137191534042358, |
| "logits/rejected": -0.8227260708808899, |
| "logps/chosen": -6.954023838043213, |
| "logps/rejected": -35.75700378417969, |
| "loss": 0.6259573698043823, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.33839166164398193, |
| "rewards/accuracies": 0.875, |
| "rewards/chosen": 0.26017990708351135, |
| "rewards/margins": 2.326916217803955, |
| "rewards/rejected": -2.0667362213134766, |
| "step": 99, |
| "train_speed(iter/s)": 0.005437 |
| }, |
| { |
| "epoch": 0.28584189370254576, |
| "grad_norm": 5.138144493103027, |
| "learning_rate": 0.00019889872962527072, |
| "logits/chosen": -0.8491114377975464, |
| "logits/rejected": -0.8503983020782471, |
| "logps/chosen": -5.965326309204102, |
| "logps/rejected": -24.436853408813477, |
| "loss": 0.9054074287414551, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.3898293972015381, |
| "rewards/accuracies": 0.65625, |
| "rewards/chosen": 0.10434619337320328, |
| "rewards/margins": 1.369442105293274, |
| "rewards/rejected": -1.265095829963684, |
| "step": 100, |
| "train_speed(iter/s)": 0.005437 |
| }, |
| { |
| "epoch": 0.28870031263957124, |
| "grad_norm": 3.6206700801849365, |
| "learning_rate": 0.00019885145938304907, |
| "logits/chosen": -0.8245471119880676, |
| "logits/rejected": -0.8329594135284424, |
| "logps/chosen": -12.563156127929688, |
| "logps/rejected": -31.796875, |
| "loss": 0.6779420971870422, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.36515456438064575, |
| "rewards/accuracies": 0.84375, |
| "rewards/chosen": 0.2728801667690277, |
| "rewards/margins": 2.0701961517333984, |
| "rewards/rejected": -1.797316074371338, |
| "step": 101, |
| "train_speed(iter/s)": 0.005432 |
| }, |
| { |
| "epoch": 0.2915587315765967, |
| "grad_norm": 4.129011154174805, |
| "learning_rate": 0.0001988032017031364, |
| "logits/chosen": -0.7823147773742676, |
| "logits/rejected": -0.7900741696357727, |
| "logps/chosen": -5.765602111816406, |
| "logps/rejected": -31.836132049560547, |
| "loss": 0.533344030380249, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.3198188543319702, |
| "rewards/accuracies": 0.9375, |
| "rewards/chosen": 0.3069157004356384, |
| "rewards/margins": 2.2244224548339844, |
| "rewards/rejected": -1.9175068140029907, |
| "step": 102, |
| "train_speed(iter/s)": 0.005432 |
| }, |
| { |
| "epoch": 0.29441715051362216, |
| "grad_norm": 7.285937786102295, |
| "learning_rate": 0.00019875395706758388, |
| "logits/chosen": -0.8199520707130432, |
| "logits/rejected": -0.8266770839691162, |
| "logps/chosen": -6.686117172241211, |
| "logps/rejected": -35.52680969238281, |
| "loss": 0.6140994429588318, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.33540287613868713, |
| "rewards/accuracies": 0.90625, |
| "rewards/chosen": 0.3744865357875824, |
| "rewards/margins": 2.1863038539886475, |
| "rewards/rejected": -1.8118171691894531, |
| "step": 103, |
| "train_speed(iter/s)": 0.005432 |
| }, |
| { |
| "epoch": 0.2972755694506476, |
| "grad_norm": 4.08017110824585, |
| "learning_rate": 0.0001987037259683013, |
| "logits/chosen": -0.7858948707580566, |
| "logits/rejected": -0.7929791808128357, |
| "logps/chosen": -6.60341215133667, |
| "logps/rejected": -32.91640853881836, |
| "loss": 0.49151238799095154, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.26134875416755676, |
| "rewards/accuracies": 0.9375, |
| "rewards/chosen": 0.3017127513885498, |
| "rewards/margins": 2.313183546066284, |
| "rewards/rejected": -2.0114707946777344, |
| "step": 104, |
| "train_speed(iter/s)": 0.005432 |
| }, |
| { |
| "epoch": 0.3001339883876731, |
| "grad_norm": 6.110865116119385, |
| "learning_rate": 0.00019865250890705247, |
| "logits/chosen": -0.8439383506774902, |
| "logits/rejected": -0.8605507612228394, |
| "logps/chosen": -5.249790668487549, |
| "logps/rejected": -45.18266677856445, |
| "loss": 0.665439248085022, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.4175432622432709, |
| "rewards/accuracies": 0.90625, |
| "rewards/chosen": 0.19819766283035278, |
| "rewards/margins": 2.974789619445801, |
| "rewards/rejected": -2.776592254638672, |
| "step": 105, |
| "train_speed(iter/s)": 0.005432 |
| }, |
| { |
| "epoch": 0.3029924073246985, |
| "grad_norm": 8.704270362854004, |
| "learning_rate": 0.00019860030639545003, |
| "logits/chosen": -0.8581717610359192, |
| "logits/rejected": -0.8701122403144836, |
| "logps/chosen": -8.200727462768555, |
| "logps/rejected": -39.06196975708008, |
| "loss": 0.9722508788108826, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.6783410310745239, |
| "rewards/accuracies": 0.875, |
| "rewards/chosen": 0.09280453622341156, |
| "rewards/margins": 2.520831823348999, |
| "rewards/rejected": -2.4280271530151367, |
| "step": 106, |
| "train_speed(iter/s)": 0.005432 |
| }, |
| { |
| "epoch": 0.305850826261724, |
| "grad_norm": 6.919908046722412, |
| "learning_rate": 0.00019854711895495036, |
| "logits/chosen": -0.8861989974975586, |
| "logits/rejected": -0.8968295454978943, |
| "logps/chosen": -4.962037563323975, |
| "logps/rejected": -40.77632522583008, |
| "loss": 0.5569091439247131, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.37660181522369385, |
| "rewards/accuracies": 0.96875, |
| "rewards/chosen": 0.30795052647590637, |
| "rewards/margins": 3.064462184906006, |
| "rewards/rejected": -2.756511688232422, |
| "step": 107, |
| "train_speed(iter/s)": 0.005431 |
| }, |
| { |
| "epoch": 0.3087092451987494, |
| "grad_norm": 4.996776580810547, |
| "learning_rate": 0.00019849294711684845, |
| "logits/chosen": -0.9625644683837891, |
| "logits/rejected": -0.972130298614502, |
| "logps/chosen": -8.820281028747559, |
| "logps/rejected": -50.21597671508789, |
| "loss": 0.7489598989486694, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.5024483799934387, |
| "rewards/accuracies": 0.90625, |
| "rewards/chosen": 0.26294729113578796, |
| "rewards/margins": 3.5292611122131348, |
| "rewards/rejected": -3.2663135528564453, |
| "step": 108, |
| "train_speed(iter/s)": 0.005431 |
| }, |
| { |
| "epoch": 0.3115676641357749, |
| "grad_norm": 6.157798767089844, |
| "learning_rate": 0.00019843779142227256, |
| "logits/chosen": -0.8762018084526062, |
| "logits/rejected": -0.882946789264679, |
| "logps/chosen": -7.924221992492676, |
| "logps/rejected": -43.705806732177734, |
| "loss": 0.6497112512588501, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.3160628080368042, |
| "rewards/accuracies": 0.75, |
| "rewards/chosen": 0.10452399402856827, |
| "rewards/margins": 3.319636106491089, |
| "rewards/rejected": -3.2151122093200684, |
| "step": 109, |
| "train_speed(iter/s)": 0.005433 |
| }, |
| { |
| "epoch": 0.31442608307280034, |
| "grad_norm": 15.488557815551758, |
| "learning_rate": 0.00019838165242217875, |
| "logits/chosen": -0.863703727722168, |
| "logits/rejected": -0.8713690042495728, |
| "logps/chosen": -6.335309028625488, |
| "logps/rejected": -46.15715026855469, |
| "loss": 0.871519923210144, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.5997005701065063, |
| "rewards/accuracies": 0.875, |
| "rewards/chosen": 0.15678179264068604, |
| "rewards/margins": 3.3330183029174805, |
| "rewards/rejected": -3.176236629486084, |
| "step": 110, |
| "train_speed(iter/s)": 0.005432 |
| }, |
| { |
| "epoch": 0.31728450200982583, |
| "grad_norm": 4.969946384429932, |
| "learning_rate": 0.0001983245306773454, |
| "logits/chosen": -0.8029311895370483, |
| "logits/rejected": -0.8069216012954712, |
| "logps/chosen": -5.91574764251709, |
| "logps/rejected": -26.575307846069336, |
| "loss": 0.6512956619262695, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.2922568917274475, |
| "rewards/accuracies": 0.90625, |
| "rewards/chosen": 0.20303387939929962, |
| "rewards/margins": 1.892715573310852, |
| "rewards/rejected": -1.6896816492080688, |
| "step": 111, |
| "train_speed(iter/s)": 0.005433 |
| }, |
| { |
| "epoch": 0.32014292094685126, |
| "grad_norm": 6.895030498504639, |
| "learning_rate": 0.0001982664267583677, |
| "logits/chosen": -0.8168566226959229, |
| "logits/rejected": -0.8292732238769531, |
| "logps/chosen": -5.602714538574219, |
| "logps/rejected": -31.50632667541504, |
| "loss": 0.8574928641319275, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.5075439810752869, |
| "rewards/accuracies": 0.8125, |
| "rewards/chosen": 0.010208956897258759, |
| "rewards/margins": 1.8891350030899048, |
| "rewards/rejected": -1.8789262771606445, |
| "step": 112, |
| "train_speed(iter/s)": 0.005432 |
| }, |
| { |
| "epoch": 0.32300133988387675, |
| "grad_norm": 4.541954040527344, |
| "learning_rate": 0.00019820734124565182, |
| "logits/chosen": -0.7733545899391174, |
| "logits/rejected": -0.7771452069282532, |
| "logps/chosen": -6.299251079559326, |
| "logps/rejected": -27.075977325439453, |
| "loss": 0.7081068754196167, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.3878493905067444, |
| "rewards/accuracies": 0.84375, |
| "rewards/chosen": 0.13715356588363647, |
| "rewards/margins": 1.6064281463623047, |
| "rewards/rejected": -1.4692747592926025, |
| "step": 113, |
| "train_speed(iter/s)": 0.005431 |
| }, |
| { |
| "epoch": 0.3258597588209022, |
| "grad_norm": 3.1780240535736084, |
| "learning_rate": 0.00019814727472940917, |
| "logits/chosen": -0.7537604570388794, |
| "logits/rejected": -0.7559649348258972, |
| "logps/chosen": -11.733123779296875, |
| "logps/rejected": -27.673603057861328, |
| "loss": 0.71772301197052, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.42100292444229126, |
| "rewards/accuracies": 0.96875, |
| "rewards/chosen": 0.21125301718711853, |
| "rewards/margins": 1.6915286779403687, |
| "rewards/rejected": -1.4802757501602173, |
| "step": 114, |
| "train_speed(iter/s)": 0.005429 |
| }, |
| { |
| "epoch": 0.32871817775792767, |
| "grad_norm": 3.8751778602600098, |
| "learning_rate": 0.00019808622780965066, |
| "logits/chosen": -0.7520322799682617, |
| "logits/rejected": -0.7543449401855469, |
| "logps/chosen": -7.260441303253174, |
| "logps/rejected": -29.127504348754883, |
| "loss": 0.6945881843566895, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.3524607717990875, |
| "rewards/accuracies": 0.84375, |
| "rewards/chosen": 0.26789340376853943, |
| "rewards/margins": 1.691644549369812, |
| "rewards/rejected": -1.4237512350082397, |
| "step": 115, |
| "train_speed(iter/s)": 0.005429 |
| }, |
| { |
| "epoch": 0.3315765966949531, |
| "grad_norm": 3.0139174461364746, |
| "learning_rate": 0.0001980242010961803, |
| "logits/chosen": -0.7901269793510437, |
| "logits/rejected": -0.7972939014434814, |
| "logps/chosen": -7.289261341094971, |
| "logps/rejected": -47.759063720703125, |
| "loss": 0.5331582427024841, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.38446128368377686, |
| "rewards/accuracies": 1.0, |
| "rewards/chosen": 0.3029555678367615, |
| "rewards/margins": 2.9576172828674316, |
| "rewards/rejected": -2.6546616554260254, |
| "step": 116, |
| "train_speed(iter/s)": 0.005429 |
| }, |
| { |
| "epoch": 0.3344350156319786, |
| "grad_norm": 3.233109712600708, |
| "learning_rate": 0.00019796119520858955, |
| "logits/chosen": -0.7568274140357971, |
| "logits/rejected": -0.7630717158317566, |
| "logps/chosen": -10.064506530761719, |
| "logps/rejected": -32.50635528564453, |
| "loss": 0.789489209651947, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.4839721918106079, |
| "rewards/accuracies": 0.84375, |
| "rewards/chosen": 0.3997339606285095, |
| "rewards/margins": 2.3508408069610596, |
| "rewards/rejected": -1.9511069059371948, |
| "step": 117, |
| "train_speed(iter/s)": 0.005429 |
| }, |
| { |
| "epoch": 0.337293434569004, |
| "grad_norm": 6.901320934295654, |
| "learning_rate": 0.0001978972107762509, |
| "logits/chosen": -0.7462923526763916, |
| "logits/rejected": -0.7441357970237732, |
| "logps/chosen": -10.976001739501953, |
| "logps/rejected": -31.10492706298828, |
| "loss": 0.7512857913970947, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.43077993392944336, |
| "rewards/accuracies": 0.875, |
| "rewards/chosen": 0.1536616086959839, |
| "rewards/margins": 2.106844186782837, |
| "rewards/rejected": -1.953182578086853, |
| "step": 118, |
| "train_speed(iter/s)": 0.00543 |
| }, |
| { |
| "epoch": 0.3401518535060295, |
| "grad_norm": 2.286742925643921, |
| "learning_rate": 0.00019783224843831162, |
| "logits/chosen": -0.7183930277824402, |
| "logits/rejected": -0.7182736396789551, |
| "logps/chosen": -8.731864929199219, |
| "logps/rejected": -33.98211669921875, |
| "loss": 0.6520930528640747, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.3684110939502716, |
| "rewards/accuracies": 0.84375, |
| "rewards/chosen": 0.2964346706867218, |
| "rewards/margins": 2.4515674114227295, |
| "rewards/rejected": -2.15513277053833, |
| "step": 119, |
| "train_speed(iter/s)": 0.005432 |
| }, |
| { |
| "epoch": 0.34301027244305493, |
| "grad_norm": 4.936792850494385, |
| "learning_rate": 0.00019776630884368738, |
| "logits/chosen": -0.7660694122314453, |
| "logits/rejected": -0.776424765586853, |
| "logps/chosen": -4.871070861816406, |
| "logps/rejected": -47.67111587524414, |
| "loss": 0.4515693187713623, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.30822524428367615, |
| "rewards/accuracies": 0.9375, |
| "rewards/chosen": 0.38517045974731445, |
| "rewards/margins": 3.7489230632781982, |
| "rewards/rejected": -3.363752841949463, |
| "step": 120, |
| "train_speed(iter/s)": 0.005431 |
| }, |
| { |
| "epoch": 0.3458686913800804, |
| "grad_norm": 9.613821983337402, |
| "learning_rate": 0.00019769939265105573, |
| "logits/chosen": -0.7732415795326233, |
| "logits/rejected": -0.7863773703575134, |
| "logps/chosen": -4.910383701324463, |
| "logps/rejected": -44.046382904052734, |
| "loss": 0.6085942983627319, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.3416149616241455, |
| "rewards/accuracies": 0.84375, |
| "rewards/chosen": 0.3152579069137573, |
| "rewards/margins": 3.380157470703125, |
| "rewards/rejected": -3.0648996829986572, |
| "step": 121, |
| "train_speed(iter/s)": 0.005432 |
| }, |
| { |
| "epoch": 0.34872711031710585, |
| "grad_norm": 5.3062744140625, |
| "learning_rate": 0.00019763150052884966, |
| "logits/chosen": -0.7940014004707336, |
| "logits/rejected": -0.8029736280441284, |
| "logps/chosen": -3.551081895828247, |
| "logps/rejected": -38.06512451171875, |
| "loss": 0.42784249782562256, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.25387704372406006, |
| "rewards/accuracies": 0.96875, |
| "rewards/chosen": 0.37009117007255554, |
| "rewards/margins": 2.841196298599243, |
| "rewards/rejected": -2.471104860305786, |
| "step": 122, |
| "train_speed(iter/s)": 0.005431 |
| }, |
| { |
| "epoch": 0.35158552925413133, |
| "grad_norm": 1.533804178237915, |
| "learning_rate": 0.0001975626331552507, |
| "logits/chosen": -0.8097442984580994, |
| "logits/rejected": -0.8176737427711487, |
| "logps/chosen": -3.7899668216705322, |
| "logps/rejected": -51.7594108581543, |
| "loss": 0.3041819632053375, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.2387474924325943, |
| "rewards/accuracies": 1.0, |
| "rewards/chosen": 0.35475367307662964, |
| "rewards/margins": 4.306023597717285, |
| "rewards/rejected": -3.9512698650360107, |
| "step": 123, |
| "train_speed(iter/s)": 0.00543 |
| }, |
| { |
| "epoch": 0.35444394819115677, |
| "grad_norm": 4.373802661895752, |
| "learning_rate": 0.00019749279121818235, |
| "logits/chosen": -0.7605909705162048, |
| "logits/rejected": -0.770943284034729, |
| "logps/chosen": -8.83084774017334, |
| "logps/rejected": -59.24419021606445, |
| "loss": 0.5374789834022522, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.39307644963264465, |
| "rewards/accuracies": 0.96875, |
| "rewards/chosen": 0.14716514945030212, |
| "rewards/margins": 4.3210930824279785, |
| "rewards/rejected": -4.173928260803223, |
| "step": 124, |
| "train_speed(iter/s)": 0.00543 |
| }, |
| { |
| "epoch": 0.3573023671281822, |
| "grad_norm": 6.3563103675842285, |
| "learning_rate": 0.0001974219754153032, |
| "logits/chosen": -0.8258897066116333, |
| "logits/rejected": -0.8272631168365479, |
| "logps/chosen": -7.77381706237793, |
| "logps/rejected": -41.343414306640625, |
| "loss": 0.6204044818878174, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.33825787901878357, |
| "rewards/accuracies": 0.84375, |
| "rewards/chosen": 0.12217142432928085, |
| "rewards/margins": 3.0349438190460205, |
| "rewards/rejected": -2.9127724170684814, |
| "step": 125, |
| "train_speed(iter/s)": 0.005431 |
| }, |
| { |
| "epoch": 0.3601607860652077, |
| "grad_norm": 9.85932731628418, |
| "learning_rate": 0.00019735018645399967, |
| "logits/chosen": -0.8187968730926514, |
| "logits/rejected": -0.8283690214157104, |
| "logps/chosen": -4.5190534591674805, |
| "logps/rejected": -45.119789123535156, |
| "loss": 0.5073397755622864, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.33468902111053467, |
| "rewards/accuracies": 0.875, |
| "rewards/chosen": 0.34614282846450806, |
| "rewards/margins": 3.5951905250549316, |
| "rewards/rejected": -3.2490477561950684, |
| "step": 126, |
| "train_speed(iter/s)": 0.005432 |
| }, |
| { |
| "epoch": 0.3630192050022331, |
| "grad_norm": 3.943042039871216, |
| "learning_rate": 0.00019727742505137936, |
| "logits/chosen": -0.7813420295715332, |
| "logits/rejected": -0.7975005507469177, |
| "logps/chosen": -7.530571937561035, |
| "logps/rejected": -53.51607894897461, |
| "loss": 0.46567243337631226, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.2742045819759369, |
| "rewards/accuracies": 0.90625, |
| "rewards/chosen": 0.23301076889038086, |
| "rewards/margins": 3.9482827186584473, |
| "rewards/rejected": -3.7152719497680664, |
| "step": 127, |
| "train_speed(iter/s)": 0.005432 |
| }, |
| { |
| "epoch": 0.3658776239392586, |
| "grad_norm": 1.9621978998184204, |
| "learning_rate": 0.00019720369193426365, |
| "logits/chosen": -0.7805514335632324, |
| "logits/rejected": -0.7947923541069031, |
| "logps/chosen": -4.593271255493164, |
| "logps/rejected": -51.02714157104492, |
| "loss": 0.3543168008327484, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.2695479393005371, |
| "rewards/accuracies": 0.96875, |
| "rewards/chosen": 0.3205646574497223, |
| "rewards/margins": 3.938343048095703, |
| "rewards/rejected": -3.6177780628204346, |
| "step": 128, |
| "train_speed(iter/s)": 0.005433 |
| }, |
| { |
| "epoch": 0.36873604287628403, |
| "grad_norm": 6.375858783721924, |
| "learning_rate": 0.00019712898783918038, |
| "logits/chosen": -0.7720431089401245, |
| "logits/rejected": -0.7774226069450378, |
| "logps/chosen": -3.9679479598999023, |
| "logps/rejected": -39.059242248535156, |
| "loss": 0.42859622836112976, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.2718777060508728, |
| "rewards/accuracies": 0.96875, |
| "rewards/chosen": 0.335873007774353, |
| "rewards/margins": 3.0524649620056152, |
| "rewards/rejected": -2.7165920734405518, |
| "step": 129, |
| "train_speed(iter/s)": 0.005434 |
| }, |
| { |
| "epoch": 0.3715944618133095, |
| "grad_norm": 20.90467643737793, |
| "learning_rate": 0.00019705331351235674, |
| "logits/chosen": -0.7668591737747192, |
| "logits/rejected": -0.7685776948928833, |
| "logps/chosen": -9.782769203186035, |
| "logps/rejected": -33.2745361328125, |
| "loss": 1.310682773590088, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.8834658861160278, |
| "rewards/accuracies": 0.78125, |
| "rewards/chosen": -0.039194077253341675, |
| "rewards/margins": 2.1344308853149414, |
| "rewards/rejected": -2.1736247539520264, |
| "step": 130, |
| "train_speed(iter/s)": 0.005434 |
| }, |
| { |
| "epoch": 0.37445288075033495, |
| "grad_norm": 2.3716464042663574, |
| "learning_rate": 0.00019697666970971153, |
| "logits/chosen": -0.7852880954742432, |
| "logits/rejected": -0.7899134755134583, |
| "logps/chosen": -6.234109878540039, |
| "logps/rejected": -42.3105354309082, |
| "loss": 0.4260920286178589, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.25257518887519836, |
| "rewards/accuracies": 0.9375, |
| "rewards/chosen": 0.3239039182662964, |
| "rewards/margins": 3.0808892250061035, |
| "rewards/rejected": -2.7569851875305176, |
| "step": 131, |
| "train_speed(iter/s)": 0.005433 |
| }, |
| { |
| "epoch": 0.37731129968736044, |
| "grad_norm": 17.186786651611328, |
| "learning_rate": 0.00019689905719684782, |
| "logits/chosen": -0.7246553301811218, |
| "logits/rejected": -0.739700198173523, |
| "logps/chosen": -4.994725227355957, |
| "logps/rejected": -37.61749267578125, |
| "loss": 0.7901899814605713, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.529437243938446, |
| "rewards/accuracies": 0.96875, |
| "rewards/chosen": 0.3914901614189148, |
| "rewards/margins": 2.7029402256011963, |
| "rewards/rejected": -2.311450242996216, |
| "step": 132, |
| "train_speed(iter/s)": 0.005432 |
| }, |
| { |
| "epoch": 0.38016971862438587, |
| "grad_norm": 19.44753646850586, |
| "learning_rate": 0.00019682047674904525, |
| "logits/chosen": -0.7204717993736267, |
| "logits/rejected": -0.7237159013748169, |
| "logps/chosen": -10.331472396850586, |
| "logps/rejected": -44.90771484375, |
| "loss": 0.9498412609100342, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.6981635093688965, |
| "rewards/accuracies": 0.90625, |
| "rewards/chosen": 0.13750669360160828, |
| "rewards/margins": 3.0414295196533203, |
| "rewards/rejected": -2.9039225578308105, |
| "step": 133, |
| "train_speed(iter/s)": 0.005433 |
| }, |
| { |
| "epoch": 0.38302813756141135, |
| "grad_norm": 2.630051612854004, |
| "learning_rate": 0.00019674092915125218, |
| "logits/chosen": -0.7359256148338318, |
| "logits/rejected": -0.7523387670516968, |
| "logps/chosen": -5.2792582511901855, |
| "logps/rejected": -53.244354248046875, |
| "loss": 0.37174469232559204, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.19531044363975525, |
| "rewards/accuracies": 0.9375, |
| "rewards/chosen": 0.35951948165893555, |
| "rewards/margins": 3.9146037101745605, |
| "rewards/rejected": -3.555084228515625, |
| "step": 134, |
| "train_speed(iter/s)": 0.005432 |
| }, |
| { |
| "epoch": 0.3858865564984368, |
| "grad_norm": 8.38878059387207, |
| "learning_rate": 0.00019666041519807802, |
| "logits/chosen": -0.7512345910072327, |
| "logits/rejected": -0.7590532302856445, |
| "logps/chosen": -7.108774662017822, |
| "logps/rejected": -49.24845886230469, |
| "loss": 0.515214741230011, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.35548365116119385, |
| "rewards/accuracies": 0.9375, |
| "rewards/chosen": 0.11580058932304382, |
| "rewards/margins": 3.6866679191589355, |
| "rewards/rejected": -3.5708670616149902, |
| "step": 135, |
| "train_speed(iter/s)": 0.005431 |
| }, |
| { |
| "epoch": 0.38874497543546227, |
| "grad_norm": 7.971051216125488, |
| "learning_rate": 0.0001965789356937852, |
| "logits/chosen": -0.7898239493370056, |
| "logits/rejected": -0.7973160743713379, |
| "logps/chosen": -9.885854721069336, |
| "logps/rejected": -55.670047760009766, |
| "loss": 0.7229015827178955, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.43035101890563965, |
| "rewards/accuracies": 0.875, |
| "rewards/chosen": 0.13693107664585114, |
| "rewards/margins": 3.9661924839019775, |
| "rewards/rejected": -3.829261541366577, |
| "step": 136, |
| "train_speed(iter/s)": 0.00543 |
| }, |
| { |
| "epoch": 0.3916033943724877, |
| "grad_norm": 6.154712677001953, |
| "learning_rate": 0.00019649649145228102, |
| "logits/chosen": -0.7895013689994812, |
| "logits/rejected": -0.7943228483200073, |
| "logps/chosen": -5.192661285400391, |
| "logps/rejected": -39.91571044921875, |
| "loss": 0.5592948794364929, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.22298313677310944, |
| "rewards/accuracies": 0.8125, |
| "rewards/chosen": 0.19678261876106262, |
| "rewards/margins": 2.95813250541687, |
| "rewards/rejected": -2.761349678039551, |
| "step": 137, |
| "train_speed(iter/s)": 0.005431 |
| }, |
| { |
| "epoch": 0.3944618133095132, |
| "grad_norm": 2.1893486976623535, |
| "learning_rate": 0.0001964130832971098, |
| "logits/chosen": -0.7978731393814087, |
| "logits/rejected": -0.808353066444397, |
| "logps/chosen": -7.741363048553467, |
| "logps/rejected": -49.584014892578125, |
| "loss": 0.4424561560153961, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.3187883198261261, |
| "rewards/accuracies": 0.96875, |
| "rewards/chosen": 0.3768758773803711, |
| "rewards/margins": 3.4152517318725586, |
| "rewards/rejected": -3.0383756160736084, |
| "step": 138, |
| "train_speed(iter/s)": 0.005431 |
| }, |
| { |
| "epoch": 0.3973202322465386, |
| "grad_norm": 3.585315465927124, |
| "learning_rate": 0.0001963287120614444, |
| "logits/chosen": -0.8294897079467773, |
| "logits/rejected": -0.8371224403381348, |
| "logps/chosen": -4.645942211151123, |
| "logps/rejected": -41.52901840209961, |
| "loss": 0.5509465336799622, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.3694899380207062, |
| "rewards/accuracies": 0.9375, |
| "rewards/chosen": 0.15256014466285706, |
| "rewards/margins": 3.042027473449707, |
| "rewards/rejected": -2.8894670009613037, |
| "step": 139, |
| "train_speed(iter/s)": 0.005432 |
| }, |
| { |
| "epoch": 0.4001786511835641, |
| "grad_norm": 3.1583364009857178, |
| "learning_rate": 0.00019624337858807807, |
| "logits/chosen": -0.8058665990829468, |
| "logits/rejected": -0.802970290184021, |
| "logps/chosen": -12.615983009338379, |
| "logps/rejected": -42.576560974121094, |
| "loss": 0.47237128019332886, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.28210973739624023, |
| "rewards/accuracies": 0.875, |
| "rewards/chosen": -0.08335088193416595, |
| "rewards/margins": 2.9341583251953125, |
| "rewards/rejected": -3.0175094604492188, |
| "step": 140, |
| "train_speed(iter/s)": 0.005432 |
| }, |
| { |
| "epoch": 0.40303707012058954, |
| "grad_norm": 1.9583823680877686, |
| "learning_rate": 0.00019615708372941587, |
| "logits/chosen": -0.7773740887641907, |
| "logits/rejected": -0.7813900113105774, |
| "logps/chosen": -7.43025016784668, |
| "logps/rejected": -48.81037521362305, |
| "loss": 0.45620426535606384, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.2858808636665344, |
| "rewards/accuracies": 0.90625, |
| "rewards/chosen": 0.14979784190654755, |
| "rewards/margins": 3.327134847640991, |
| "rewards/rejected": -3.177337169647217, |
| "step": 141, |
| "train_speed(iter/s)": 0.005434 |
| }, |
| { |
| "epoch": 0.405895489057615, |
| "grad_norm": 17.376474380493164, |
| "learning_rate": 0.00019606982834746627, |
| "logits/chosen": -0.8055461645126343, |
| "logits/rejected": -0.8150217533111572, |
| "logps/chosen": -9.388490676879883, |
| "logps/rejected": -43.86671447753906, |
| "loss": 1.0983664989471436, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.7692541480064392, |
| "rewards/accuracies": 0.90625, |
| "rewards/chosen": 0.1326138973236084, |
| "rewards/margins": 2.8655741214752197, |
| "rewards/rejected": -2.7329602241516113, |
| "step": 142, |
| "train_speed(iter/s)": 0.005433 |
| }, |
| { |
| "epoch": 0.40875390799464045, |
| "grad_norm": 35.17894744873047, |
| "learning_rate": 0.00019598161331383257, |
| "logits/chosen": -0.8625121116638184, |
| "logits/rejected": -0.871791422367096, |
| "logps/chosen": -9.16450023651123, |
| "logps/rejected": -47.45623016357422, |
| "loss": 0.7930290102958679, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.543784499168396, |
| "rewards/accuracies": 0.875, |
| "rewards/chosen": 0.08059316873550415, |
| "rewards/margins": 3.273069381713867, |
| "rewards/rejected": -3.192476272583008, |
| "step": 143, |
| "train_speed(iter/s)": 0.005433 |
| }, |
| { |
| "epoch": 0.41161232693166594, |
| "grad_norm": 4.112196445465088, |
| "learning_rate": 0.00019589243950970402, |
| "logits/chosen": -0.8877297043800354, |
| "logits/rejected": -0.8911035060882568, |
| "logps/chosen": -6.215106964111328, |
| "logps/rejected": -41.89563751220703, |
| "loss": 0.3530213236808777, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.20008514821529388, |
| "rewards/accuracies": 0.9375, |
| "rewards/chosen": 0.3117275834083557, |
| "rewards/margins": 3.073056221008301, |
| "rewards/rejected": -2.76132869720459, |
| "step": 144, |
| "train_speed(iter/s)": 0.005434 |
| }, |
| { |
| "epoch": 0.41447074586869137, |
| "grad_norm": 12.758796691894531, |
| "learning_rate": 0.00019580230782584722, |
| "logits/chosen": -0.8928901553153992, |
| "logits/rejected": -0.8998609781265259, |
| "logps/chosen": -5.406792640686035, |
| "logps/rejected": -40.006412506103516, |
| "loss": 0.5003563165664673, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.30671095848083496, |
| "rewards/accuracies": 0.90625, |
| "rewards/chosen": 0.23732616007328033, |
| "rewards/margins": 2.932386636734009, |
| "rewards/rejected": -2.6950607299804688, |
| "step": 145, |
| "train_speed(iter/s)": 0.005434 |
| }, |
| { |
| "epoch": 0.41732916480571686, |
| "grad_norm": 4.014595985412598, |
| "learning_rate": 0.00019571121916259706, |
| "logits/chosen": -0.887598991394043, |
| "logits/rejected": -0.8942646384239197, |
| "logps/chosen": -4.3415961265563965, |
| "logps/rejected": -46.59621810913086, |
| "loss": 0.36891594529151917, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.2651711106300354, |
| "rewards/accuracies": 0.96875, |
| "rewards/chosen": 0.36169373989105225, |
| "rewards/margins": 3.6561460494995117, |
| "rewards/rejected": -3.294452428817749, |
| "step": 146, |
| "train_speed(iter/s)": 0.005434 |
| }, |
| { |
| "epoch": 0.4201875837427423, |
| "grad_norm": 10.37033462524414, |
| "learning_rate": 0.00019561917442984788, |
| "logits/chosen": -0.8872988224029541, |
| "logits/rejected": -0.8909754753112793, |
| "logps/chosen": -7.036233901977539, |
| "logps/rejected": -44.856361389160156, |
| "loss": 0.5071041584014893, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.4129847586154938, |
| "rewards/accuracies": 0.96875, |
| "rewards/chosen": 0.29683101177215576, |
| "rewards/margins": 3.6587531566619873, |
| "rewards/rejected": -3.361922025680542, |
| "step": 147, |
| "train_speed(iter/s)": 0.005434 |
| }, |
| { |
| "epoch": 0.4230460026797678, |
| "grad_norm": 12.54053783416748, |
| "learning_rate": 0.00019552617454704428, |
| "logits/chosen": -0.909309983253479, |
| "logits/rejected": -0.9173017144203186, |
| "logps/chosen": -5.930859088897705, |
| "logps/rejected": -46.784332275390625, |
| "loss": 0.6537680625915527, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.521246075630188, |
| "rewards/accuracies": 0.9375, |
| "rewards/chosen": 0.25368478894233704, |
| "rewards/margins": 3.396470308303833, |
| "rewards/rejected": -3.1427857875823975, |
| "step": 148, |
| "train_speed(iter/s)": 0.005433 |
| }, |
| { |
| "epoch": 0.4259044216167932, |
| "grad_norm": 2.5870954990386963, |
| "learning_rate": 0.00019543222044317188, |
| "logits/chosen": -0.8764325976371765, |
| "logits/rejected": -0.881737232208252, |
| "logps/chosen": -5.638083457946777, |
| "logps/rejected": -47.27805709838867, |
| "loss": 0.3462856113910675, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.22526228427886963, |
| "rewards/accuracies": 0.96875, |
| "rewards/chosen": 0.3174706697463989, |
| "rewards/margins": 3.5086669921875, |
| "rewards/rejected": -3.1911962032318115, |
| "step": 149, |
| "train_speed(iter/s)": 0.005433 |
| }, |
| { |
| "epoch": 0.4287628405538187, |
| "grad_norm": 9.161517143249512, |
| "learning_rate": 0.00019533731305674818, |
| "logits/chosen": -0.8732975721359253, |
| "logits/rejected": -0.8769663572311401, |
| "logps/chosen": -9.518259048461914, |
| "logps/rejected": -42.666114807128906, |
| "loss": 0.9442892670631409, |
| "memory(GiB)": 46.73, |
| "nll_loss": 0.6460040211677551, |
| "rewards/accuracies": 0.875, |
| "rewards/chosen": 0.07534892857074738, |
| "rewards/margins": 3.1069083213806152, |
| "rewards/rejected": -3.0315592288970947, |
| "step": 150, |
| "train_speed(iter/s)": 0.005433 |
| } |
| ], |
| "logging_steps": 1, |
| "max_steps": 1047, |
| "num_input_tokens_seen": 0, |
| "num_train_epochs": 3, |
| "save_steps": 50, |
| "stateful_callbacks": { |
| "TrainerControl": { |
| "args": { |
| "should_epoch_stop": false, |
| "should_evaluate": false, |
| "should_log": false, |
| "should_save": true, |
| "should_training_stop": false |
| }, |
| "attributes": {} |
| } |
| }, |
| "total_flos": 1.0479874209029816e+18, |
| "train_batch_size": 1, |
| "trial_name": null, |
| "trial_params": null |
| } |
|
|
