| { |
| "best_global_step": null, |
| "best_metric": null, |
| "best_model_checkpoint": null, |
| "epoch": 0.12982797792924375, |
| "eval_steps": 300, |
| "global_step": 50, |
| "is_hyper_param_search": false, |
| "is_local_process_zero": true, |
| "is_world_process_zero": true, |
| "log_history": [ |
| { |
| "epoch": 0.002596559558584875, |
| "grad_norm": 4.769598007202148, |
| "learning_rate": 5.128205128205128e-06, |
| "logits/chosen": -0.5944205522537231, |
| "logits/rejected": -0.5917393565177917, |
| "logps/chosen": -11.821638107299805, |
| "logps/rejected": -10.731678009033203, |
| "loss": 1.193467140197754, |
| "memory(GiB)": 46.82, |
| "nll_loss": 0.50031977891922, |
| "rewards/accuracies": 0.0, |
| "rewards/chosen": 0.0, |
| "rewards/margins": 0.0, |
| "rewards/rejected": 0.0, |
| "step": 1, |
| "train_speed(iter/s)": 0.004987 |
| }, |
| { |
| "epoch": 0.00519311911716975, |
| "grad_norm": 7.638933181762695, |
| "learning_rate": 1.0256410256410256e-05, |
| "logits/chosen": -0.6539207696914673, |
| "logits/rejected": -0.6607442498207092, |
| "logps/chosen": -9.628549575805664, |
| "logps/rejected": -15.572840690612793, |
| "loss": 1.1913397312164307, |
| "memory(GiB)": 46.82, |
| "nll_loss": 0.4981924891471863, |
| "rewards/accuracies": 0.0, |
| "rewards/chosen": 0.0, |
| "rewards/margins": 0.0, |
| "rewards/rejected": 0.0, |
| "step": 2, |
| "train_speed(iter/s)": 0.005274 |
| }, |
| { |
| "epoch": 0.007789678675754625, |
| "grad_norm": 8.172294616699219, |
| "learning_rate": 1.5384615384615387e-05, |
| "logits/chosen": -0.6215255260467529, |
| "logits/rejected": -0.6243188381195068, |
| "logps/chosen": -10.225044250488281, |
| "logps/rejected": -12.433060646057129, |
| "loss": 1.3010810613632202, |
| "memory(GiB)": 46.82, |
| "nll_loss": 0.6032217741012573, |
| "rewards/accuracies": 0.40625, |
| "rewards/chosen": -0.002520320238545537, |
| "rewards/margins": -0.00917358510196209, |
| "rewards/rejected": 0.006653264630585909, |
| "step": 3, |
| "train_speed(iter/s)": 0.005242 |
| }, |
| { |
| "epoch": 0.0103862382343395, |
| "grad_norm": 7.314579010009766, |
| "learning_rate": 2.0512820512820512e-05, |
| "logits/chosen": -0.6062605381011963, |
| "logits/rejected": -0.6072395443916321, |
| "logps/chosen": -9.956047058105469, |
| "logps/rejected": -11.640246391296387, |
| "loss": 1.2542610168457031, |
| "memory(GiB)": 46.82, |
| "nll_loss": 0.5589770674705505, |
| "rewards/accuracies": 0.46875, |
| "rewards/chosen": 0.007420660927891731, |
| "rewards/margins": -0.004003261215984821, |
| "rewards/rejected": 0.011423922143876553, |
| "step": 4, |
| "train_speed(iter/s)": 0.005268 |
| }, |
| { |
| "epoch": 0.012982797792924375, |
| "grad_norm": 3.7953975200653076, |
| "learning_rate": 2.564102564102564e-05, |
| "logits/chosen": -0.5976298451423645, |
| "logits/rejected": -0.5966740846633911, |
| "logps/chosen": -13.070164680480957, |
| "logps/rejected": -9.661067008972168, |
| "loss": 1.215171456336975, |
| "memory(GiB)": 46.82, |
| "nll_loss": 0.5106430649757385, |
| "rewards/accuracies": 0.375, |
| "rewards/chosen": 0.00016488437540829182, |
| "rewards/margins": -0.020555444061756134, |
| "rewards/rejected": 0.02072032354772091, |
| "step": 5, |
| "train_speed(iter/s)": 0.00529 |
| }, |
| { |
| "epoch": 0.01557935735150925, |
| "grad_norm": 7.202587604522705, |
| "learning_rate": 3.0769230769230774e-05, |
| "logits/chosen": -0.6029255390167236, |
| "logits/rejected": -0.6082680225372314, |
| "logps/chosen": -6.68419075012207, |
| "logps/rejected": -14.836040496826172, |
| "loss": 1.1022435426712036, |
| "memory(GiB)": 46.82, |
| "nll_loss": 0.4213921129703522, |
| "rewards/accuracies": 0.59375, |
| "rewards/chosen": 0.05949082225561142, |
| "rewards/margins": 0.03669985011219978, |
| "rewards/rejected": 0.02279096469283104, |
| "step": 6, |
| "train_speed(iter/s)": 0.00532 |
| }, |
| { |
| "epoch": 0.018175916910094125, |
| "grad_norm": 4.17504358291626, |
| "learning_rate": 3.58974358974359e-05, |
| "logits/chosen": -0.5737150311470032, |
| "logits/rejected": -0.5751169919967651, |
| "logps/chosen": -10.05550479888916, |
| "logps/rejected": -10.229061126708984, |
| "loss": 1.1968086957931519, |
| "memory(GiB)": 46.82, |
| "nll_loss": 0.451814740896225, |
| "rewards/accuracies": 0.40625, |
| "rewards/chosen": 0.018731240183115005, |
| "rewards/margins": -0.06073049083352089, |
| "rewards/rejected": 0.0794617310166359, |
| "step": 7, |
| "train_speed(iter/s)": 0.005319 |
| }, |
| { |
| "epoch": 0.020772476468679, |
| "grad_norm": 4.966729164123535, |
| "learning_rate": 4.1025641025641023e-05, |
| "logits/chosen": -0.6149957180023193, |
| "logits/rejected": -0.619285523891449, |
| "logps/chosen": -10.129886627197266, |
| "logps/rejected": -14.14559268951416, |
| "loss": 1.1496262550354004, |
| "memory(GiB)": 46.82, |
| "nll_loss": 0.45597678422927856, |
| "rewards/accuracies": 0.5, |
| "rewards/chosen": -0.048677604645490646, |
| "rewards/margins": 0.07307372987270355, |
| "rewards/rejected": -0.1217513307929039, |
| "step": 8, |
| "train_speed(iter/s)": 0.005322 |
| }, |
| { |
| "epoch": 0.023369036027263874, |
| "grad_norm": 4.258759021759033, |
| "learning_rate": 4.615384615384616e-05, |
| "logits/chosen": -0.6464206576347351, |
| "logits/rejected": -0.649814248085022, |
| "logps/chosen": -6.2765960693359375, |
| "logps/rejected": -14.259992599487305, |
| "loss": 0.9683928489685059, |
| "memory(GiB)": 46.82, |
| "nll_loss": 0.27603328227996826, |
| "rewards/accuracies": 0.59375, |
| "rewards/chosen": -0.028385277837514877, |
| "rewards/margins": 0.1275366246700287, |
| "rewards/rejected": -0.15592190623283386, |
| "step": 9, |
| "train_speed(iter/s)": 0.005346 |
| }, |
| { |
| "epoch": 0.02596559558584875, |
| "grad_norm": 3.5765655040740967, |
| "learning_rate": 5.128205128205128e-05, |
| "logits/chosen": -0.5905835628509521, |
| "logits/rejected": -0.5969584584236145, |
| "logps/chosen": -14.795049667358398, |
| "logps/rejected": -15.656811714172363, |
| "loss": 1.238275408744812, |
| "memory(GiB)": 46.82, |
| "nll_loss": 0.5923190116882324, |
| "rewards/accuracies": 0.5625, |
| "rewards/chosen": -0.012541369535028934, |
| "rewards/margins": 0.1883796602487564, |
| "rewards/rejected": -0.20092105865478516, |
| "step": 10, |
| "train_speed(iter/s)": 0.005347 |
| }, |
| { |
| "epoch": 0.028562155144433627, |
| "grad_norm": 5.3241682052612305, |
| "learning_rate": 5.6410256410256414e-05, |
| "logits/chosen": -0.5469571948051453, |
| "logits/rejected": -0.5478999614715576, |
| "logps/chosen": -11.43381118774414, |
| "logps/rejected": -13.521588325500488, |
| "loss": 1.2181414365768433, |
| "memory(GiB)": 46.82, |
| "nll_loss": 0.4799095690250397, |
| "rewards/accuracies": 0.5, |
| "rewards/chosen": -0.11080781370401382, |
| "rewards/margins": -0.000970199704170227, |
| "rewards/rejected": -0.109837606549263, |
| "step": 11, |
| "train_speed(iter/s)": 0.005357 |
| }, |
| { |
| "epoch": 0.0311587147030185, |
| "grad_norm": 3.739349126815796, |
| "learning_rate": 6.153846153846155e-05, |
| "logits/chosen": -0.5663439631462097, |
| "logits/rejected": -0.5685043931007385, |
| "logps/chosen": -9.113786697387695, |
| "logps/rejected": -12.898387908935547, |
| "loss": 1.1865863800048828, |
| "memory(GiB)": 46.82, |
| "nll_loss": 0.5080631971359253, |
| "rewards/accuracies": 0.59375, |
| "rewards/chosen": 0.07351399958133698, |
| "rewards/margins": 0.09553620219230652, |
| "rewards/rejected": -0.022022202610969543, |
| "step": 12, |
| "train_speed(iter/s)": 0.00536 |
| }, |
| { |
| "epoch": 0.03375527426160337, |
| "grad_norm": 2.732536792755127, |
| "learning_rate": 6.666666666666667e-05, |
| "logits/chosen": -0.6258959770202637, |
| "logits/rejected": -0.628243088722229, |
| "logps/chosen": -11.908061027526855, |
| "logps/rejected": -14.742002487182617, |
| "loss": 1.0720946788787842, |
| "memory(GiB)": 46.82, |
| "nll_loss": 0.44474831223487854, |
| "rewards/accuracies": 0.65625, |
| "rewards/chosen": 0.0872192531824112, |
| "rewards/margins": 0.16859112679958344, |
| "rewards/rejected": -0.08137187361717224, |
| "step": 13, |
| "train_speed(iter/s)": 0.005365 |
| }, |
| { |
| "epoch": 0.03635183382018825, |
| "grad_norm": 2.286402702331543, |
| "learning_rate": 7.17948717948718e-05, |
| "logits/chosen": -0.5518996119499207, |
| "logits/rejected": -0.5536460280418396, |
| "logps/chosen": -7.985256195068359, |
| "logps/rejected": -9.997514724731445, |
| "loss": 1.0465917587280273, |
| "memory(GiB)": 46.82, |
| "nll_loss": 0.37517642974853516, |
| "rewards/accuracies": 0.5625, |
| "rewards/chosen": 0.09782104939222336, |
| "rewards/margins": 0.08964069932699203, |
| "rewards/rejected": 0.008180351927876472, |
| "step": 14, |
| "train_speed(iter/s)": 0.005362 |
| }, |
| { |
| "epoch": 0.03894839337877313, |
| "grad_norm": 2.8793153762817383, |
| "learning_rate": 7.692307692307693e-05, |
| "logits/chosen": -0.5976672172546387, |
| "logits/rejected": -0.5990516543388367, |
| "logps/chosen": -11.681562423706055, |
| "logps/rejected": -15.269111633300781, |
| "loss": 1.0413833856582642, |
| "memory(GiB)": 46.82, |
| "nll_loss": 0.44159939885139465, |
| "rewards/accuracies": 0.6875, |
| "rewards/chosen": 0.18112002313137054, |
| "rewards/margins": 0.246211975812912, |
| "rewards/rejected": -0.06509192287921906, |
| "step": 15, |
| "train_speed(iter/s)": 0.005369 |
| }, |
| { |
| "epoch": 0.041544952937358, |
| "grad_norm": 3.362415313720703, |
| "learning_rate": 8.205128205128205e-05, |
| "logits/chosen": -0.6091838479042053, |
| "logits/rejected": -0.6097983717918396, |
| "logps/chosen": -6.9782915115356445, |
| "logps/rejected": -9.985608100891113, |
| "loss": 1.0336495637893677, |
| "memory(GiB)": 46.82, |
| "nll_loss": 0.3420756161212921, |
| "rewards/accuracies": 0.5, |
| "rewards/chosen": 0.05590391904115677, |
| "rewards/margins": 0.03598181903362274, |
| "rewards/rejected": 0.019922103732824326, |
| "step": 16, |
| "train_speed(iter/s)": 0.005368 |
| }, |
| { |
| "epoch": 0.04414151249594288, |
| "grad_norm": 2.314833641052246, |
| "learning_rate": 8.717948717948718e-05, |
| "logits/chosen": -0.6001038551330566, |
| "logits/rejected": -0.6009984612464905, |
| "logps/chosen": -8.711943626403809, |
| "logps/rejected": -8.347241401672363, |
| "loss": 1.0018643140792847, |
| "memory(GiB)": 46.82, |
| "nll_loss": 0.3586280345916748, |
| "rewards/accuracies": 0.53125, |
| "rewards/chosen": 0.17141370475292206, |
| "rewards/margins": 0.15122823417186737, |
| "rewards/rejected": 0.020185478031635284, |
| "step": 17, |
| "train_speed(iter/s)": 0.005365 |
| }, |
| { |
| "epoch": 0.04673807205452775, |
| "grad_norm": 2.6080470085144043, |
| "learning_rate": 9.230769230769232e-05, |
| "logits/chosen": -0.6079075932502747, |
| "logits/rejected": -0.6091002225875854, |
| "logps/chosen": -9.866250991821289, |
| "logps/rejected": -9.885502815246582, |
| "loss": 1.0599387884140015, |
| "memory(GiB)": 46.82, |
| "nll_loss": 0.3953135013580322, |
| "rewards/accuracies": 0.625, |
| "rewards/chosen": 0.1653863787651062, |
| "rewards/margins": 0.13114988803863525, |
| "rewards/rejected": 0.03423647955060005, |
| "step": 18, |
| "train_speed(iter/s)": 0.005363 |
| }, |
| { |
| "epoch": 0.04933463161311263, |
| "grad_norm": 2.5750620365142822, |
| "learning_rate": 9.743589743589744e-05, |
| "logits/chosen": -0.5754107236862183, |
| "logits/rejected": -0.5755229592323303, |
| "logps/chosen": -9.663117408752441, |
| "logps/rejected": -14.233282089233398, |
| "loss": 0.9537867307662964, |
| "memory(GiB)": 46.82, |
| "nll_loss": 0.35284289717674255, |
| "rewards/accuracies": 0.6875, |
| "rewards/chosen": 0.12623246014118195, |
| "rewards/margins": 0.24378177523612976, |
| "rewards/rejected": -0.11754930019378662, |
| "step": 19, |
| "train_speed(iter/s)": 0.00537 |
| }, |
| { |
| "epoch": 0.0519311911716975, |
| "grad_norm": 4.289522171020508, |
| "learning_rate": 0.00010256410256410256, |
| "logits/chosen": -0.5981144905090332, |
| "logits/rejected": -0.6002673506736755, |
| "logps/chosen": -8.273541450500488, |
| "logps/rejected": -12.811166763305664, |
| "loss": 1.422603964805603, |
| "memory(GiB)": 46.82, |
| "nll_loss": 0.8183298707008362, |
| "rewards/accuracies": 0.65625, |
| "rewards/chosen": 0.21172407269477844, |
| "rewards/margins": 0.2575765550136566, |
| "rewards/rejected": -0.04585248976945877, |
| "step": 20, |
| "train_speed(iter/s)": 0.005365 |
| }, |
| { |
| "epoch": 0.054527750730282376, |
| "grad_norm": 2.396873712539673, |
| "learning_rate": 0.0001076923076923077, |
| "logits/chosen": -0.5655781030654907, |
| "logits/rejected": -0.5672769546508789, |
| "logps/chosen": -7.642134189605713, |
| "logps/rejected": -16.37198257446289, |
| "loss": 0.8944156765937805, |
| "memory(GiB)": 46.82, |
| "nll_loss": 0.28976452350616455, |
| "rewards/accuracies": 0.6875, |
| "rewards/chosen": 0.2191261202096939, |
| "rewards/margins": 0.2641218602657318, |
| "rewards/rejected": -0.04499572888016701, |
| "step": 21, |
| "train_speed(iter/s)": 0.005364 |
| }, |
| { |
| "epoch": 0.057124310288867254, |
| "grad_norm": 3.0281121730804443, |
| "learning_rate": 0.00011282051282051283, |
| "logits/chosen": -0.5596702694892883, |
| "logits/rejected": -0.5609368681907654, |
| "logps/chosen": -7.446313381195068, |
| "logps/rejected": -12.651912689208984, |
| "loss": 0.9526722431182861, |
| "memory(GiB)": 46.82, |
| "nll_loss": 0.33721715211868286, |
| "rewards/accuracies": 0.75, |
| "rewards/chosen": 0.15709233283996582, |
| "rewards/margins": 0.2245473712682724, |
| "rewards/rejected": -0.06745504587888718, |
| "step": 22, |
| "train_speed(iter/s)": 0.005362 |
| }, |
| { |
| "epoch": 0.059720869847452125, |
| "grad_norm": 2.6955020427703857, |
| "learning_rate": 0.00011794871794871796, |
| "logits/chosen": -0.5400959849357605, |
| "logits/rejected": -0.5422949194908142, |
| "logps/chosen": -6.002541542053223, |
| "logps/rejected": -12.28770637512207, |
| "loss": 0.9044801592826843, |
| "memory(GiB)": 46.82, |
| "nll_loss": 0.2814594805240631, |
| "rewards/accuracies": 0.6875, |
| "rewards/chosen": 0.11246027797460556, |
| "rewards/margins": 0.21622154116630554, |
| "rewards/rejected": -0.10376127809286118, |
| "step": 23, |
| "train_speed(iter/s)": 0.005363 |
| }, |
| { |
| "epoch": 0.062317429406037, |
| "grad_norm": 3.1954290866851807, |
| "learning_rate": 0.0001230769230769231, |
| "logits/chosen": -0.5763551592826843, |
| "logits/rejected": -0.5829795598983765, |
| "logps/chosen": -8.22636890411377, |
| "logps/rejected": -12.94092845916748, |
| "loss": 0.9596253037452698, |
| "memory(GiB)": 46.82, |
| "nll_loss": 0.3679075837135315, |
| "rewards/accuracies": 0.75, |
| "rewards/chosen": 0.17844173312187195, |
| "rewards/margins": 0.2646990418434143, |
| "rewards/rejected": -0.08625732362270355, |
| "step": 24, |
| "train_speed(iter/s)": 0.005367 |
| }, |
| { |
| "epoch": 0.06491398896462187, |
| "grad_norm": 3.440917730331421, |
| "learning_rate": 0.00012820512820512823, |
| "logits/chosen": -0.5423401594161987, |
| "logits/rejected": -0.5454680919647217, |
| "logps/chosen": -6.751742362976074, |
| "logps/rejected": -14.677633285522461, |
| "loss": 0.8222842216491699, |
| "memory(GiB)": 46.82, |
| "nll_loss": 0.2530384063720703, |
| "rewards/accuracies": 0.6875, |
| "rewards/chosen": 0.20278047025203705, |
| "rewards/margins": 0.37300848960876465, |
| "rewards/rejected": -0.1702280342578888, |
| "step": 25, |
| "train_speed(iter/s)": 0.005363 |
| }, |
| { |
| "epoch": 0.06751054852320675, |
| "grad_norm": 7.046718120574951, |
| "learning_rate": 0.00013333333333333334, |
| "logits/chosen": -0.531994640827179, |
| "logits/rejected": -0.5324733257293701, |
| "logps/chosen": -7.691443920135498, |
| "logps/rejected": -12.912195205688477, |
| "loss": 0.8851491212844849, |
| "memory(GiB)": 46.82, |
| "nll_loss": 0.2668984532356262, |
| "rewards/accuracies": 0.65625, |
| "rewards/chosen": 0.03996456041932106, |
| "rewards/margins": 0.30043962597846985, |
| "rewards/rejected": -0.2604750692844391, |
| "step": 26, |
| "train_speed(iter/s)": 0.005354 |
| }, |
| { |
| "epoch": 0.07010710808179163, |
| "grad_norm": 3.7837774753570557, |
| "learning_rate": 0.00013846153846153847, |
| "logits/chosen": -0.5356835722923279, |
| "logits/rejected": -0.535167932510376, |
| "logps/chosen": -7.836236476898193, |
| "logps/rejected": -12.5257568359375, |
| "loss": 0.9785585999488831, |
| "memory(GiB)": 46.82, |
| "nll_loss": 0.39958691596984863, |
| "rewards/accuracies": 0.625, |
| "rewards/chosen": 0.11726226657629013, |
| "rewards/margins": 0.30467987060546875, |
| "rewards/rejected": -0.18741759657859802, |
| "step": 27, |
| "train_speed(iter/s)": 0.005353 |
| }, |
| { |
| "epoch": 0.0727036676403765, |
| "grad_norm": 2.986128568649292, |
| "learning_rate": 0.0001435897435897436, |
| "logits/chosen": -0.5374022126197815, |
| "logits/rejected": -0.5330510139465332, |
| "logps/chosen": -10.774028778076172, |
| "logps/rejected": -10.21718692779541, |
| "loss": 1.0210673809051514, |
| "memory(GiB)": 46.82, |
| "nll_loss": 0.44256913661956787, |
| "rewards/accuracies": 0.78125, |
| "rewards/chosen": 0.08123424649238586, |
| "rewards/margins": 0.32150229811668396, |
| "rewards/rejected": -0.2402680516242981, |
| "step": 28, |
| "train_speed(iter/s)": 0.00536 |
| }, |
| { |
| "epoch": 0.07530022719896137, |
| "grad_norm": 8.227151870727539, |
| "learning_rate": 0.00014871794871794872, |
| "logits/chosen": -0.555951714515686, |
| "logits/rejected": -0.5587239265441895, |
| "logps/chosen": -7.656264781951904, |
| "logps/rejected": -16.2135066986084, |
| "loss": 1.024498701095581, |
| "memory(GiB)": 46.82, |
| "nll_loss": 0.467337429523468, |
| "rewards/accuracies": 0.71875, |
| "rewards/chosen": 0.1679726243019104, |
| "rewards/margins": 0.43920189142227173, |
| "rewards/rejected": -0.2712292969226837, |
| "step": 29, |
| "train_speed(iter/s)": 0.005367 |
| }, |
| { |
| "epoch": 0.07789678675754626, |
| "grad_norm": 3.866678476333618, |
| "learning_rate": 0.00015384615384615385, |
| "logits/chosen": -0.5329037308692932, |
| "logits/rejected": -0.5308660864830017, |
| "logps/chosen": -13.327455520629883, |
| "logps/rejected": -12.309192657470703, |
| "loss": 1.1753352880477905, |
| "memory(GiB)": 46.82, |
| "nll_loss": 0.5743139386177063, |
| "rewards/accuracies": 0.65625, |
| "rewards/chosen": 0.10232796519994736, |
| "rewards/margins": 0.27346518635749817, |
| "rewards/rejected": -0.1711372286081314, |
| "step": 30, |
| "train_speed(iter/s)": 0.00537 |
| }, |
| { |
| "epoch": 0.08049334631613113, |
| "grad_norm": 2.7835240364074707, |
| "learning_rate": 0.00015897435897435896, |
| "logits/chosen": -0.5286478996276855, |
| "logits/rejected": -0.5297562479972839, |
| "logps/chosen": -8.105016708374023, |
| "logps/rejected": -18.577655792236328, |
| "loss": 0.8083048462867737, |
| "memory(GiB)": 46.82, |
| "nll_loss": 0.3598409593105316, |
| "rewards/accuracies": 0.875, |
| "rewards/chosen": 0.3429361879825592, |
| "rewards/margins": 0.6762320399284363, |
| "rewards/rejected": -0.3332958221435547, |
| "step": 31, |
| "train_speed(iter/s)": 0.005368 |
| }, |
| { |
| "epoch": 0.083089905874716, |
| "grad_norm": 2.621320962905884, |
| "learning_rate": 0.0001641025641025641, |
| "logits/chosen": -0.5535257458686829, |
| "logits/rejected": -0.5577558279037476, |
| "logps/chosen": -7.381773948669434, |
| "logps/rejected": -20.8293514251709, |
| "loss": 0.7857053279876709, |
| "memory(GiB)": 46.82, |
| "nll_loss": 0.3121376037597656, |
| "rewards/accuracies": 0.84375, |
| "rewards/chosen": 0.11771110445261002, |
| "rewards/margins": 0.6395547389984131, |
| "rewards/rejected": -0.5218436121940613, |
| "step": 32, |
| "train_speed(iter/s)": 0.005373 |
| }, |
| { |
| "epoch": 0.08568646543330087, |
| "grad_norm": 5.273679256439209, |
| "learning_rate": 0.00016923076923076923, |
| "logits/chosen": -0.5348193645477295, |
| "logits/rejected": -0.5335426330566406, |
| "logps/chosen": -16.012008666992188, |
| "logps/rejected": -16.090946197509766, |
| "loss": 1.150058388710022, |
| "memory(GiB)": 46.82, |
| "nll_loss": 0.5311561822891235, |
| "rewards/accuracies": 0.6875, |
| "rewards/chosen": -0.09080329537391663, |
| "rewards/margins": 0.3698381185531616, |
| "rewards/rejected": -0.46064135432243347, |
| "step": 33, |
| "train_speed(iter/s)": 0.005375 |
| }, |
| { |
| "epoch": 0.08828302499188576, |
| "grad_norm": 3.7380588054656982, |
| "learning_rate": 0.00017435897435897436, |
| "logits/chosen": -0.526316225528717, |
| "logits/rejected": -0.5241590738296509, |
| "logps/chosen": -13.188718795776367, |
| "logps/rejected": -18.877933502197266, |
| "loss": 0.8901183009147644, |
| "memory(GiB)": 46.82, |
| "nll_loss": 0.4010193347930908, |
| "rewards/accuracies": 0.78125, |
| "rewards/chosen": 0.13551224768161774, |
| "rewards/margins": 0.7849555015563965, |
| "rewards/rejected": -0.6494433283805847, |
| "step": 34, |
| "train_speed(iter/s)": 0.005376 |
| }, |
| { |
| "epoch": 0.09087958455047063, |
| "grad_norm": 3.2903590202331543, |
| "learning_rate": 0.0001794871794871795, |
| "logits/chosen": -0.5694751739501953, |
| "logits/rejected": -0.5714199542999268, |
| "logps/chosen": -13.836735725402832, |
| "logps/rejected": -17.397571563720703, |
| "loss": 0.9489783644676208, |
| "memory(GiB)": 46.82, |
| "nll_loss": 0.4817684292793274, |
| "rewards/accuracies": 0.84375, |
| "rewards/chosen": 0.17933052778244019, |
| "rewards/margins": 0.7199034690856934, |
| "rewards/rejected": -0.5405729413032532, |
| "step": 35, |
| "train_speed(iter/s)": 0.005374 |
| }, |
| { |
| "epoch": 0.0934761441090555, |
| "grad_norm": 7.443658351898193, |
| "learning_rate": 0.00018461538461538463, |
| "logits/chosen": -0.5579850077629089, |
| "logits/rejected": -0.5598161220550537, |
| "logps/chosen": -9.724953651428223, |
| "logps/rejected": -19.001192092895508, |
| "loss": 1.025741696357727, |
| "memory(GiB)": 46.82, |
| "nll_loss": 0.4933169484138489, |
| "rewards/accuracies": 0.8125, |
| "rewards/chosen": 0.09439410269260406, |
| "rewards/margins": 0.531223714351654, |
| "rewards/rejected": -0.4368295967578888, |
| "step": 36, |
| "train_speed(iter/s)": 0.005367 |
| }, |
| { |
| "epoch": 0.09607270366764038, |
| "grad_norm": 4.221776962280273, |
| "learning_rate": 0.00018974358974358974, |
| "logits/chosen": -0.5673465728759766, |
| "logits/rejected": -0.568681001663208, |
| "logps/chosen": -7.996758937835693, |
| "logps/rejected": -16.828737258911133, |
| "loss": 0.768932044506073, |
| "memory(GiB)": 46.82, |
| "nll_loss": 0.3632465600967407, |
| "rewards/accuracies": 0.90625, |
| "rewards/chosen": 0.20153376460075378, |
| "rewards/margins": 0.8081123232841492, |
| "rewards/rejected": -0.6065785884857178, |
| "step": 37, |
| "train_speed(iter/s)": 0.00536 |
| }, |
| { |
| "epoch": 0.09866926322622525, |
| "grad_norm": 4.438755989074707, |
| "learning_rate": 0.00019487179487179487, |
| "logits/chosen": -0.5432108640670776, |
| "logits/rejected": -0.5440645217895508, |
| "logps/chosen": -11.19113826751709, |
| "logps/rejected": -16.801233291625977, |
| "loss": 0.9643179178237915, |
| "memory(GiB)": 46.82, |
| "nll_loss": 0.5330875515937805, |
| "rewards/accuracies": 0.875, |
| "rewards/chosen": 0.212477445602417, |
| "rewards/margins": 0.7622401714324951, |
| "rewards/rejected": -0.5497627258300781, |
| "step": 38, |
| "train_speed(iter/s)": 0.005362 |
| }, |
| { |
| "epoch": 0.10126582278481013, |
| "grad_norm": 3.844766855239868, |
| "learning_rate": 0.0002, |
| "logits/chosen": -0.5104550719261169, |
| "logits/rejected": -0.5119606852531433, |
| "logps/chosen": -11.040297508239746, |
| "logps/rejected": -18.91219711303711, |
| "loss": 0.9704711437225342, |
| "memory(GiB)": 46.82, |
| "nll_loss": 0.5230136513710022, |
| "rewards/accuracies": 0.84375, |
| "rewards/chosen": 0.14157143235206604, |
| "rewards/margins": 0.7000695466995239, |
| "rewards/rejected": -0.5584981441497803, |
| "step": 39, |
| "train_speed(iter/s)": 0.005358 |
| }, |
| { |
| "epoch": 0.103862382343395, |
| "grad_norm": 5.2911272048950195, |
| "learning_rate": 0.00019999907650547008, |
| "logits/chosen": -0.563255250453949, |
| "logits/rejected": -0.5627263784408569, |
| "logps/chosen": -8.962218284606934, |
| "logps/rejected": -20.214658737182617, |
| "loss": 0.9730861783027649, |
| "memory(GiB)": 46.82, |
| "nll_loss": 0.4494069516658783, |
| "rewards/accuracies": 0.71875, |
| "rewards/chosen": 0.03735905885696411, |
| "rewards/margins": 0.7039426565170288, |
| "rewards/rejected": -0.6665836572647095, |
| "step": 40, |
| "train_speed(iter/s)": 0.00536 |
| }, |
| { |
| "epoch": 0.10645894190197988, |
| "grad_norm": 3.2932822704315186, |
| "learning_rate": 0.0001999963060389371, |
| "logits/chosen": -0.49407708644866943, |
| "logits/rejected": -0.4951040744781494, |
| "logps/chosen": -4.336781978607178, |
| "logps/rejected": -23.352622985839844, |
| "loss": 0.5769761800765991, |
| "memory(GiB)": 46.82, |
| "nll_loss": 0.21373578906059265, |
| "rewards/accuracies": 0.84375, |
| "rewards/chosen": 0.2602457106113434, |
| "rewards/margins": 1.0540037155151367, |
| "rewards/rejected": -0.7937580347061157, |
| "step": 41, |
| "train_speed(iter/s)": 0.005359 |
| }, |
| { |
| "epoch": 0.10905550146056475, |
| "grad_norm": 3.550082206726074, |
| "learning_rate": 0.00019999168865157137, |
| "logits/chosen": -0.5316025018692017, |
| "logits/rejected": -0.5340486764907837, |
| "logps/chosen": -8.441543579101562, |
| "logps/rejected": -18.124588012695312, |
| "loss": 0.9663518071174622, |
| "memory(GiB)": 46.82, |
| "nll_loss": 0.47912853956222534, |
| "rewards/accuracies": 0.78125, |
| "rewards/chosen": 0.14834752678871155, |
| "rewards/margins": 0.7632903456687927, |
| "rewards/rejected": -0.6149427890777588, |
| "step": 42, |
| "train_speed(iter/s)": 0.005356 |
| }, |
| { |
| "epoch": 0.11165206101914962, |
| "grad_norm": 3.8798763751983643, |
| "learning_rate": 0.0001999852244286554, |
| "logits/chosen": -0.5125827193260193, |
| "logits/rejected": -0.5118388533592224, |
| "logps/chosen": -8.259347915649414, |
| "logps/rejected": -18.246912002563477, |
| "loss": 0.8706857562065125, |
| "memory(GiB)": 46.82, |
| "nll_loss": 0.3688449263572693, |
| "rewards/accuracies": 0.78125, |
| "rewards/chosen": 0.1189112439751625, |
| "rewards/margins": 0.5402826070785522, |
| "rewards/rejected": -0.4213714003562927, |
| "step": 43, |
| "train_speed(iter/s)": 0.005355 |
| }, |
| { |
| "epoch": 0.11424862057773451, |
| "grad_norm": 5.702610015869141, |
| "learning_rate": 0.0001999769134895828, |
| "logits/chosen": -0.5676945447921753, |
| "logits/rejected": -0.5712423324584961, |
| "logps/chosen": -7.507511138916016, |
| "logps/rejected": -17.306344985961914, |
| "loss": 0.9447178840637207, |
| "memory(GiB)": 46.82, |
| "nll_loss": 0.5068771839141846, |
| "rewards/accuracies": 0.84375, |
| "rewards/chosen": 0.1572760045528412, |
| "rewards/margins": 0.8183934688568115, |
| "rewards/rejected": -0.6611174941062927, |
| "step": 44, |
| "train_speed(iter/s)": 0.005357 |
| }, |
| { |
| "epoch": 0.11684518013631938, |
| "grad_norm": 5.846091270446777, |
| "learning_rate": 0.0001999667559878556, |
| "logits/chosen": -0.5530126094818115, |
| "logits/rejected": -0.5544271469116211, |
| "logps/chosen": -9.34422779083252, |
| "logps/rejected": -21.99274444580078, |
| "loss": 1.0033745765686035, |
| "memory(GiB)": 46.82, |
| "nll_loss": 0.5510709285736084, |
| "rewards/accuracies": 0.75, |
| "rewards/chosen": 0.06781984120607376, |
| "rewards/margins": 0.8660807609558105, |
| "rewards/rejected": -0.7982609868049622, |
| "step": 45, |
| "train_speed(iter/s)": 0.005358 |
| }, |
| { |
| "epoch": 0.11944173969490425, |
| "grad_norm": 3.4377248287200928, |
| "learning_rate": 0.00019995475211108185, |
| "logits/chosen": -0.5516988039016724, |
| "logits/rejected": -0.5513719916343689, |
| "logps/chosen": -9.321004867553711, |
| "logps/rejected": -18.94799041748047, |
| "loss": 0.838591992855072, |
| "memory(GiB)": 46.82, |
| "nll_loss": 0.44976377487182617, |
| "rewards/accuracies": 0.875, |
| "rewards/chosen": 0.15304125845432281, |
| "rewards/margins": 0.9354310631752014, |
| "rewards/rejected": -0.782389760017395, |
| "step": 46, |
| "train_speed(iter/s)": 0.00536 |
| }, |
| { |
| "epoch": 0.12203829925348912, |
| "grad_norm": 3.021584987640381, |
| "learning_rate": 0.00019994090208097176, |
| "logits/chosen": -0.5453197956085205, |
| "logits/rejected": -0.5457130670547485, |
| "logps/chosen": -7.513545036315918, |
| "logps/rejected": -19.929990768432617, |
| "loss": 0.7878033518791199, |
| "memory(GiB)": 46.82, |
| "nll_loss": 0.2906043827533722, |
| "rewards/accuracies": 0.71875, |
| "rewards/chosen": 0.04635176807641983, |
| "rewards/margins": 0.8055768609046936, |
| "rewards/rejected": -0.759225070476532, |
| "step": 47, |
| "train_speed(iter/s)": 0.005359 |
| }, |
| { |
| "epoch": 0.124634858812074, |
| "grad_norm": 2.4632484912872314, |
| "learning_rate": 0.00019992520615333393, |
| "logits/chosen": -0.5288453102111816, |
| "logits/rejected": -0.5294773578643799, |
| "logps/chosen": -4.316083908081055, |
| "logps/rejected": -16.75088119506836, |
| "loss": 0.6369743347167969, |
| "memory(GiB)": 46.82, |
| "nll_loss": 0.19174346327781677, |
| "rewards/accuracies": 0.8125, |
| "rewards/chosen": 0.1844159960746765, |
| "rewards/margins": 0.7728991508483887, |
| "rewards/rejected": -0.5884830951690674, |
| "step": 48, |
| "train_speed(iter/s)": 0.005357 |
| }, |
| { |
| "epoch": 0.12723141837065888, |
| "grad_norm": 2.225041627883911, |
| "learning_rate": 0.00019990766461807039, |
| "logits/chosen": -0.5786827802658081, |
| "logits/rejected": -0.5775803923606873, |
| "logps/chosen": -7.1691179275512695, |
| "logps/rejected": -19.747949600219727, |
| "loss": 0.611931562423706, |
| "memory(GiB)": 46.82, |
| "nll_loss": 0.26027703285217285, |
| "rewards/accuracies": 0.9375, |
| "rewards/chosen": 0.32533061504364014, |
| "rewards/margins": 1.0796949863433838, |
| "rewards/rejected": -0.7543643712997437, |
| "step": 49, |
| "train_speed(iter/s)": 0.005358 |
| }, |
| { |
| "epoch": 0.12982797792924375, |
| "grad_norm": 3.631110429763794, |
| "learning_rate": 0.00019988827779917137, |
| "logits/chosen": -0.5347975492477417, |
| "logits/rejected": -0.5404946804046631, |
| "logps/chosen": -8.115504264831543, |
| "logps/rejected": -22.913742065429688, |
| "loss": 0.8714032769203186, |
| "memory(GiB)": 46.82, |
| "nll_loss": 0.4530017375946045, |
| "rewards/accuracies": 0.875, |
| "rewards/chosen": 0.22861376404762268, |
| "rewards/margins": 0.9870915412902832, |
| "rewards/rejected": -0.7584777474403381, |
| "step": 50, |
| "train_speed(iter/s)": 0.00536 |
| } |
| ], |
| "logging_steps": 1, |
| "max_steps": 770, |
| "num_input_tokens_seen": 0, |
| "num_train_epochs": 2, |
| "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": 3.5719756566337946e+17, |
| "train_batch_size": 1, |
| "trial_name": null, |
| "trial_params": null |
| } |
|
|
