| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.16327085955304602, | |
| "eval_steps": 50, | |
| "global_step": 150, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.005442361985101534, | |
| "grad_norm": 223.0, | |
| "learning_rate": 1.4999011862064836e-06, | |
| "logits/chosen": -0.3115585446357727, | |
| "logits/rejected": -0.4312504231929779, | |
| "logps/chosen": -0.8023494482040405, | |
| "logps/rejected": -0.9114950299263, | |
| "loss": 1.7206, | |
| "nll_loss": 0.8566828966140747, | |
| "rewards/accuracies": 0.671875, | |
| "rewards/chosen": -2.005873680114746, | |
| "rewards/margins": 0.27286386489868164, | |
| "rewards/rejected": -2.2787375450134277, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.010884723970203068, | |
| "grad_norm": 213.0, | |
| "learning_rate": 1.4996047737567963e-06, | |
| "logits/chosen": -0.2838870584964752, | |
| "logits/rejected": -0.4562837481498718, | |
| "logps/chosen": -0.756219744682312, | |
| "logps/rejected": -0.893709659576416, | |
| "loss": 1.6545, | |
| "nll_loss": 0.8323467373847961, | |
| "rewards/accuracies": 0.746874988079071, | |
| "rewards/chosen": -1.8905493021011353, | |
| "rewards/margins": 0.34372463822364807, | |
| "rewards/rejected": -2.23427414894104, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.0163270859553046, | |
| "grad_norm": 207.0, | |
| "learning_rate": 1.4991108494350523e-06, | |
| "logits/chosen": -0.2802577018737793, | |
| "logits/rejected": -0.445369154214859, | |
| "logps/chosen": -0.7702305912971497, | |
| "logps/rejected": -0.9028533101081848, | |
| "loss": 1.67, | |
| "nll_loss": 0.8351463079452515, | |
| "rewards/accuracies": 0.721875011920929, | |
| "rewards/chosen": -1.9255762100219727, | |
| "rewards/margins": 0.331557035446167, | |
| "rewards/rejected": -2.2571334838867188, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.021769447940406136, | |
| "grad_norm": 193.0, | |
| "learning_rate": 1.4984195578532098e-06, | |
| "logits/chosen": -0.27219459414482117, | |
| "logits/rejected": -0.43887200951576233, | |
| "logps/chosen": -0.7428117990493774, | |
| "logps/rejected": -0.8896303176879883, | |
| "loss": 1.6298, | |
| "nll_loss": 0.8110933303833008, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -1.8570295572280884, | |
| "rewards/margins": 0.36704641580581665, | |
| "rewards/rejected": -2.2240757942199707, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.02721180992550767, | |
| "grad_norm": 196.0, | |
| "learning_rate": 1.4975311014087314e-06, | |
| "logits/chosen": -0.28368493914604187, | |
| "logits/rejected": -0.45796823501586914, | |
| "logps/chosen": -0.7394925951957703, | |
| "logps/rejected": -0.8978835344314575, | |
| "loss": 1.6038, | |
| "nll_loss": 0.810085117816925, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -1.8487316370010376, | |
| "rewards/margins": 0.3959771394729614, | |
| "rewards/rejected": -2.244708776473999, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.0326541719106092, | |
| "grad_norm": 180.0, | |
| "learning_rate": 1.4964457402253263e-06, | |
| "logits/chosen": -0.26782557368278503, | |
| "logits/rejected": -0.4345061779022217, | |
| "logps/chosen": -0.7221552133560181, | |
| "logps/rejected": -0.8976587057113647, | |
| "loss": 1.5676, | |
| "nll_loss": 0.8051786422729492, | |
| "rewards/accuracies": 0.801562488079071, | |
| "rewards/chosen": -1.80538809299469, | |
| "rewards/margins": 0.4387587606906891, | |
| "rewards/rejected": -2.2441468238830566, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.03809653389571074, | |
| "grad_norm": 184.0, | |
| "learning_rate": 1.495163792076789e-06, | |
| "logits/chosen": -0.2734525799751282, | |
| "logits/rejected": -0.4188796579837799, | |
| "logps/chosen": -0.7307429313659668, | |
| "logps/rejected": -0.9264034032821655, | |
| "loss": 1.5643, | |
| "nll_loss": 0.8087199926376343, | |
| "rewards/accuracies": 0.809374988079071, | |
| "rewards/chosen": -1.826857328414917, | |
| "rewards/margins": 0.4891512393951416, | |
| "rewards/rejected": -2.3160085678100586, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.04353889588081227, | |
| "grad_norm": 167.0, | |
| "learning_rate": 1.493685632293963e-06, | |
| "logits/chosen": -0.2776848375797272, | |
| "logits/rejected": -0.45066213607788086, | |
| "logps/chosen": -0.6951563954353333, | |
| "logps/rejected": -0.9053373336791992, | |
| "loss": 1.5161, | |
| "nll_loss": 0.7817299962043762, | |
| "rewards/accuracies": 0.8296874761581421, | |
| "rewards/chosen": -1.7378908395767212, | |
| "rewards/margins": 0.525452733039856, | |
| "rewards/rejected": -2.263343572616577, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.048981257865913806, | |
| "grad_norm": 166.0, | |
| "learning_rate": 1.49201169365485e-06, | |
| "logits/chosen": -0.2875303626060486, | |
| "logits/rejected": -0.4160293936729431, | |
| "logps/chosen": -0.7146323919296265, | |
| "logps/rejected": -0.9353097677230835, | |
| "loss": 1.509, | |
| "nll_loss": 0.789823055267334, | |
| "rewards/accuracies": 0.854687511920929, | |
| "rewards/chosen": -1.786581039428711, | |
| "rewards/margins": 0.551693320274353, | |
| "rewards/rejected": -2.3382744789123535, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 0.05442361985101534, | |
| "grad_norm": 167.0, | |
| "learning_rate": 1.490142466257901e-06, | |
| "logits/chosen": -0.2606073021888733, | |
| "logits/rejected": -0.42186981439590454, | |
| "logps/chosen": -0.6743353605270386, | |
| "logps/rejected": -0.9106225967407227, | |
| "loss": 1.472, | |
| "nll_loss": 0.7610515356063843, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": -1.6858384609222412, | |
| "rewards/margins": 0.590718150138855, | |
| "rewards/rejected": -2.2765564918518066, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.05442361985101534, | |
| "eval_logits/chosen": -0.4242976903915405, | |
| "eval_logits/rejected": -0.5250905156135559, | |
| "eval_logps/chosen": -0.6774176955223083, | |
| "eval_logps/rejected": -0.9184495806694031, | |
| "eval_loss": 1.466933012008667, | |
| "eval_nll_loss": 0.7765384316444397, | |
| "eval_rewards/accuracies": 0.8569999933242798, | |
| "eval_rewards/chosen": -1.6935441493988037, | |
| "eval_rewards/margins": 0.6025797128677368, | |
| "eval_rewards/rejected": -2.29612398147583, | |
| "eval_runtime": 119.2017, | |
| "eval_samples_per_second": 8.389, | |
| "eval_steps_per_second": 2.097, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.059865981836116876, | |
| "grad_norm": 176.0, | |
| "learning_rate": 1.4880784973785227e-06, | |
| "logits/chosen": -0.2771835923194885, | |
| "logits/rejected": -0.43040376901626587, | |
| "logps/chosen": -0.6886002421379089, | |
| "logps/rejected": -0.9180153608322144, | |
| "loss": 1.4611, | |
| "nll_loss": 0.7568584084510803, | |
| "rewards/accuracies": 0.8687499761581421, | |
| "rewards/chosen": -1.7215007543563843, | |
| "rewards/margins": 0.5735376477241516, | |
| "rewards/rejected": -2.2950384616851807, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 0.0653083438212184, | |
| "grad_norm": 173.0, | |
| "learning_rate": 1.4858203913088481e-06, | |
| "logits/chosen": -0.28170254826545715, | |
| "logits/rejected": -0.43440356850624084, | |
| "logps/chosen": -0.6944113969802856, | |
| "logps/rejected": -0.958784282207489, | |
| "loss": 1.456, | |
| "nll_loss": 0.7766550183296204, | |
| "rewards/accuracies": 0.8531249761581421, | |
| "rewards/chosen": -1.7360286712646484, | |
| "rewards/margins": 0.6609319448471069, | |
| "rewards/rejected": -2.396960496902466, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.07075070580631994, | |
| "grad_norm": 175.0, | |
| "learning_rate": 1.4833688091808084e-06, | |
| "logits/chosen": -0.29959458112716675, | |
| "logits/rejected": -0.4557119905948639, | |
| "logps/chosen": -0.6825212836265564, | |
| "logps/rejected": -0.9580196142196655, | |
| "loss": 1.4192, | |
| "nll_loss": 0.7619790434837341, | |
| "rewards/accuracies": 0.895312488079071, | |
| "rewards/chosen": -1.7063030004501343, | |
| "rewards/margins": 0.6887460947036743, | |
| "rewards/rejected": -2.3950493335723877, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 0.07619306779142147, | |
| "grad_norm": 189.0, | |
| "learning_rate": 1.4807244687725672e-06, | |
| "logits/chosen": -0.2963514029979706, | |
| "logits/rejected": -0.4536547064781189, | |
| "logps/chosen": -0.6995417475700378, | |
| "logps/rejected": -0.9971321821212769, | |
| "loss": 1.4071, | |
| "nll_loss": 0.7711254358291626, | |
| "rewards/accuracies": 0.903124988079071, | |
| "rewards/chosen": -1.748854398727417, | |
| "rewards/margins": 0.7439761757850647, | |
| "rewards/rejected": -2.492830753326416, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.08163542977652301, | |
| "grad_norm": 192.0, | |
| "learning_rate": 1.477888144298368e-06, | |
| "logits/chosen": -0.3087882399559021, | |
| "logits/rejected": -0.4929865002632141, | |
| "logps/chosen": -0.6821542978286743, | |
| "logps/rejected": -1.0003395080566406, | |
| "loss": 1.3817, | |
| "nll_loss": 0.7631580233573914, | |
| "rewards/accuracies": 0.8984375, | |
| "rewards/chosen": -1.7053858041763306, | |
| "rewards/margins": 0.7954627871513367, | |
| "rewards/rejected": -2.5008487701416016, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.08707779176162454, | |
| "grad_norm": 250.0, | |
| "learning_rate": 1.4748606661818576e-06, | |
| "logits/chosen": -0.30626195669174194, | |
| "logits/rejected": -0.5067037343978882, | |
| "logps/chosen": -0.6984843015670776, | |
| "logps/rejected": -1.0486465692520142, | |
| "loss": 1.3604, | |
| "nll_loss": 0.777851939201355, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -1.7462108135223389, | |
| "rewards/margins": 0.8754053115844727, | |
| "rewards/rejected": -2.6216163635253906, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.09252015374672608, | |
| "grad_norm": 229.0, | |
| "learning_rate": 1.4716429208129552e-06, | |
| "logits/chosen": -0.3446356952190399, | |
| "logits/rejected": -0.558761477470398, | |
| "logps/chosen": -0.6777626276016235, | |
| "logps/rejected": -1.0675888061523438, | |
| "loss": 1.3122, | |
| "nll_loss": 0.7588330507278442, | |
| "rewards/accuracies": 0.9359375238418579, | |
| "rewards/chosen": -1.694406509399414, | |
| "rewards/margins": 0.9745653867721558, | |
| "rewards/rejected": -2.6689720153808594, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 0.09796251573182761, | |
| "grad_norm": 229.0, | |
| "learning_rate": 1.4682358502883309e-06, | |
| "logits/chosen": -0.36884021759033203, | |
| "logits/rejected": -0.6074358820915222, | |
| "logps/chosen": -0.6712280511856079, | |
| "logps/rejected": -1.1392602920532227, | |
| "loss": 1.2571, | |
| "nll_loss": 0.7604485154151917, | |
| "rewards/accuracies": 0.940625011920929, | |
| "rewards/chosen": -1.678070068359375, | |
| "rewards/margins": 1.1700807809829712, | |
| "rewards/rejected": -2.8481509685516357, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.10340487771692915, | |
| "grad_norm": 248.0, | |
| "learning_rate": 1.4646404521355798e-06, | |
| "logits/chosen": -0.3828192353248596, | |
| "logits/rejected": -0.6576655507087708, | |
| "logps/chosen": -0.695937991142273, | |
| "logps/rejected": -1.2227665185928345, | |
| "loss": 1.2367, | |
| "nll_loss": 0.7761465311050415, | |
| "rewards/accuracies": 0.9468749761581421, | |
| "rewards/chosen": -1.7398450374603271, | |
| "rewards/margins": 1.3170711994171143, | |
| "rewards/rejected": -3.0569162368774414, | |
| "step": 95 | |
| }, | |
| { | |
| "epoch": 0.10884723970203068, | |
| "grad_norm": 233.0, | |
| "learning_rate": 1.4608577790211639e-06, | |
| "logits/chosen": -0.408522367477417, | |
| "logits/rejected": -0.6887016296386719, | |
| "logps/chosen": -0.6912533640861511, | |
| "logps/rejected": -1.2720332145690918, | |
| "loss": 1.1964, | |
| "nll_loss": 0.7770565748214722, | |
| "rewards/accuracies": 0.956250011920929, | |
| "rewards/chosen": -1.7281334400177002, | |
| "rewards/margins": 1.4519497156143188, | |
| "rewards/rejected": -3.1800830364227295, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.10884723970203068, | |
| "eval_logits/chosen": -0.6553338766098022, | |
| "eval_logits/rejected": -0.960581362247467, | |
| "eval_logps/chosen": -0.6551228165626526, | |
| "eval_logps/rejected": -1.286936640739441, | |
| "eval_loss": 1.1423654556274414, | |
| "eval_nll_loss": 0.7609782814979553, | |
| "eval_rewards/accuracies": 0.9639999866485596, | |
| "eval_rewards/chosen": -1.637807011604309, | |
| "eval_rewards/margins": 1.579534649848938, | |
| "eval_rewards/rejected": -3.217341899871826, | |
| "eval_runtime": 122.2807, | |
| "eval_samples_per_second": 8.178, | |
| "eval_steps_per_second": 2.044, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.11428960168713222, | |
| "grad_norm": 244.0, | |
| "learning_rate": 1.4568889384422085e-06, | |
| "logits/chosen": -0.44453367590904236, | |
| "logits/rejected": -0.8137799501419067, | |
| "logps/chosen": -0.6734666228294373, | |
| "logps/rejected": -1.3720922470092773, | |
| "loss": 1.1097, | |
| "nll_loss": 0.7629668116569519, | |
| "rewards/accuracies": 0.9703124761581421, | |
| "rewards/chosen": -1.6836665868759155, | |
| "rewards/margins": 1.7465636730194092, | |
| "rewards/rejected": -3.4302303791046143, | |
| "step": 105 | |
| }, | |
| { | |
| "epoch": 0.11973196367223375, | |
| "grad_norm": 232.0, | |
| "learning_rate": 1.4527350924022508e-06, | |
| "logits/chosen": -0.4658277928829193, | |
| "logits/rejected": -0.8643436431884766, | |
| "logps/chosen": -0.6887673735618591, | |
| "logps/rejected": -1.4408514499664307, | |
| "loss": 1.0866, | |
| "nll_loss": 0.7645546793937683, | |
| "rewards/accuracies": 0.9765625, | |
| "rewards/chosen": -1.7219183444976807, | |
| "rewards/margins": 1.8802101612091064, | |
| "rewards/rejected": -3.602128505706787, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.12517432565733527, | |
| "grad_norm": 227.0, | |
| "learning_rate": 1.4483974570710226e-06, | |
| "logits/chosen": -0.4871746897697449, | |
| "logits/rejected": -1.0338547229766846, | |
| "logps/chosen": -0.6287084817886353, | |
| "logps/rejected": -1.4971318244934082, | |
| "loss": 0.9829, | |
| "nll_loss": 0.7324297428131104, | |
| "rewards/accuracies": 0.984375, | |
| "rewards/chosen": -1.571771264076233, | |
| "rewards/margins": 2.171058177947998, | |
| "rewards/rejected": -3.7428295612335205, | |
| "step": 115 | |
| }, | |
| { | |
| "epoch": 0.1306166876424368, | |
| "grad_norm": 186.0, | |
| "learning_rate": 1.4438773024283826e-06, | |
| "logits/chosen": -0.5202233195304871, | |
| "logits/rejected": -1.1327874660491943, | |
| "logps/chosen": -0.6288429498672485, | |
| "logps/rejected": -1.5928165912628174, | |
| "loss": 0.959, | |
| "nll_loss": 0.7312449216842651, | |
| "rewards/accuracies": 0.9859374761581421, | |
| "rewards/chosen": -1.572107195854187, | |
| "rewards/margins": 2.4099345207214355, | |
| "rewards/rejected": -3.982041120529175, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.13605904962753834, | |
| "grad_norm": 177.0, | |
| "learning_rate": 1.4391759518924858e-06, | |
| "logits/chosen": -0.5006698369979858, | |
| "logits/rejected": -1.2095844745635986, | |
| "logps/chosen": -0.6492566466331482, | |
| "logps/rejected": -1.7050262689590454, | |
| "loss": 0.9298, | |
| "nll_loss": 0.742263674736023, | |
| "rewards/accuracies": 0.9906250238418579, | |
| "rewards/chosen": -1.623141884803772, | |
| "rewards/margins": 2.6394238471984863, | |
| "rewards/rejected": -4.262566089630127, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 0.14150141161263988, | |
| "grad_norm": 155.0, | |
| "learning_rate": 1.4342947819323133e-06, | |
| "logits/chosen": -0.49277132749557495, | |
| "logits/rejected": -1.2336362600326538, | |
| "logps/chosen": -0.6462730169296265, | |
| "logps/rejected": -1.7665297985076904, | |
| "loss": 0.9073, | |
| "nll_loss": 0.7340320348739624, | |
| "rewards/accuracies": 0.9906250238418579, | |
| "rewards/chosen": -1.615682601928711, | |
| "rewards/margins": 2.8006420135498047, | |
| "rewards/rejected": -4.416324615478516, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.1469437735977414, | |
| "grad_norm": 157.0, | |
| "learning_rate": 1.4292352216646672e-06, | |
| "logits/chosen": -0.4820989966392517, | |
| "logits/rejected": -1.2759294509887695, | |
| "logps/chosen": -0.6405208706855774, | |
| "logps/rejected": -1.8488733768463135, | |
| "loss": 0.8986, | |
| "nll_loss": 0.7417997121810913, | |
| "rewards/accuracies": 0.989062488079071, | |
| "rewards/chosen": -1.601301908493042, | |
| "rewards/margins": 3.020881175994873, | |
| "rewards/rejected": -4.622182846069336, | |
| "step": 135 | |
| }, | |
| { | |
| "epoch": 0.15238613558284295, | |
| "grad_norm": 117.0, | |
| "learning_rate": 1.4239987524357501e-06, | |
| "logits/chosen": -0.48703551292419434, | |
| "logits/rejected": -1.3474723100662231, | |
| "logps/chosen": -0.6453564167022705, | |
| "logps/rejected": -1.947719931602478, | |
| "loss": 0.8587, | |
| "nll_loss": 0.7325531840324402, | |
| "rewards/accuracies": 0.9937499761581421, | |
| "rewards/chosen": -1.6133911609649658, | |
| "rewards/margins": 3.2559094429016113, | |
| "rewards/rejected": -4.86929988861084, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.15782849756794448, | |
| "grad_norm": 114.0, | |
| "learning_rate": 1.4185869073874565e-06, | |
| "logits/chosen": -0.4773550033569336, | |
| "logits/rejected": -1.3999682664871216, | |
| "logps/chosen": -0.6471393704414368, | |
| "logps/rejected": -2.04455304145813, | |
| "loss": 0.8458, | |
| "nll_loss": 0.7375032305717468, | |
| "rewards/accuracies": 0.996874988079071, | |
| "rewards/chosen": -1.6178483963012695, | |
| "rewards/margins": 3.493534803390503, | |
| "rewards/rejected": -5.111382961273193, | |
| "step": 145 | |
| }, | |
| { | |
| "epoch": 0.16327085955304602, | |
| "grad_norm": 102.5, | |
| "learning_rate": 1.413001271008494e-06, | |
| "logits/chosen": -0.5018254518508911, | |
| "logits/rejected": -1.4777194261550903, | |
| "logps/chosen": -0.6435462832450867, | |
| "logps/rejected": -2.1130857467651367, | |
| "loss": 0.8377, | |
| "nll_loss": 0.7388636469841003, | |
| "rewards/accuracies": 0.9937499761581421, | |
| "rewards/chosen": -1.608865737915039, | |
| "rewards/margins": 3.673849105834961, | |
| "rewards/rejected": -5.28271484375, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.16327085955304602, | |
| "eval_logits/chosen": -0.6831561326980591, | |
| "eval_logits/rejected": -1.8252586126327515, | |
| "eval_logps/chosen": -0.6115793585777283, | |
| "eval_logps/rejected": -2.0589704513549805, | |
| "eval_loss": 0.8232702612876892, | |
| "eval_nll_loss": 0.7224369645118713, | |
| "eval_rewards/accuracies": 0.9929999709129333, | |
| "eval_rewards/chosen": -1.528948187828064, | |
| "eval_rewards/margins": 3.6184778213500977, | |
| "eval_rewards/rejected": -5.147425651550293, | |
| "eval_runtime": 112.2438, | |
| "eval_samples_per_second": 8.909, | |
| "eval_steps_per_second": 2.227, | |
| "step": 150 | |
| } | |
| ], | |
| "logging_steps": 5, | |
| "max_steps": 918, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "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": 0.0, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |