| { |
| "best_metric": null, |
| "best_model_checkpoint": null, |
| "epoch": 0.060836501901140684, |
| "eval_steps": 50, |
| "global_step": 20, |
| "is_hyper_param_search": false, |
| "is_local_process_zero": true, |
| "is_world_process_zero": true, |
| "log_history": [ |
| { |
| "epoch": 0.003041825095057034, |
| "grad_norm": 0.2263541966676712, |
| "learning_rate": 9.090909090909091e-06, |
| "log_odds_chosen": 0.0022868067026138306, |
| "log_odds_ratio": -0.6926867961883545, |
| "logits/chosen": -2.233539581298828, |
| "logits/rejected": -2.2098608016967773, |
| "logps/chosen": -0.7675836682319641, |
| "logps/rejected": -0.7697547078132629, |
| "loss": 0.85230005, |
| "memory(GiB)": 243.33, |
| "nll_loss": 0.7675836682319641, |
| "rewards/accuracies": 0.375, |
| "rewards/chosen": -0.07675836980342865, |
| "rewards/margins": 0.00021709827706217766, |
| "rewards/rejected": -0.07697547227144241, |
| "step": 1, |
| "train_speed(iter/s)": 0.010532 |
| }, |
| { |
| "epoch": 0.006083650190114068, |
| "grad_norm": 0.21064028143882751, |
| "learning_rate": 1.8181818181818182e-05, |
| "log_odds_chosen": 0.411478191614151, |
| "log_odds_ratio": -0.5463609099388123, |
| "logits/chosen": -2.2685160636901855, |
| "logits/rejected": -2.260843515396118, |
| "logps/chosen": -0.9355340003967285, |
| "logps/rejected": -1.236234188079834, |
| "loss": 0.87318134, |
| "memory(GiB)": 243.33, |
| "nll_loss": 0.9355340003967285, |
| "rewards/accuracies": 0.875, |
| "rewards/chosen": -0.09355340152978897, |
| "rewards/margins": 0.030070019885897636, |
| "rewards/rejected": -0.12362341582775116, |
| "step": 2, |
| "train_speed(iter/s)": 0.009734 |
| }, |
| { |
| "epoch": 0.009125475285171103, |
| "grad_norm": 0.19250068068504333, |
| "learning_rate": 2.727272727272727e-05, |
| "log_odds_chosen": 0.3938485383987427, |
| "log_odds_ratio": -0.5777455568313599, |
| "logits/chosen": -1.993510365486145, |
| "logits/rejected": -1.9485821723937988, |
| "logps/chosen": -0.6979267001152039, |
| "logps/rejected": -0.9728450775146484, |
| "loss": 0.75437951, |
| "memory(GiB)": 243.33, |
| "nll_loss": 0.6979267597198486, |
| "rewards/accuracies": 0.375, |
| "rewards/chosen": -0.06979266554117203, |
| "rewards/margins": 0.027491841465234756, |
| "rewards/rejected": -0.09728450328111649, |
| "step": 3, |
| "train_speed(iter/s)": 0.01152 |
| }, |
| { |
| "epoch": 0.012167300380228136, |
| "grad_norm": 0.20081621408462524, |
| "learning_rate": 3.6363636363636364e-05, |
| "log_odds_chosen": 0.0849333107471466, |
| "log_odds_ratio": -0.6586424112319946, |
| "logits/chosen": -2.3017446994781494, |
| "logits/rejected": -2.300307512283325, |
| "logps/chosen": -0.7535954713821411, |
| "logps/rejected": -0.8145345449447632, |
| "loss": 0.85012603, |
| "memory(GiB)": 315.51, |
| "nll_loss": 0.7535954713821411, |
| "rewards/accuracies": 0.5, |
| "rewards/chosen": -0.0753595381975174, |
| "rewards/margins": 0.006093906704336405, |
| "rewards/rejected": -0.08145345002412796, |
| "step": 4, |
| "train_speed(iter/s)": 0.011285 |
| }, |
| { |
| "epoch": 0.015209125475285171, |
| "grad_norm": 0.20968054234981537, |
| "learning_rate": 4.545454545454545e-05, |
| "log_odds_chosen": 0.2265796959400177, |
| "log_odds_ratio": -0.605457067489624, |
| "logits/chosen": -2.132286548614502, |
| "logits/rejected": -2.0954604148864746, |
| "logps/chosen": -0.7475967407226562, |
| "logps/rejected": -0.9037913680076599, |
| "loss": 0.83076566, |
| "memory(GiB)": 349.51, |
| "nll_loss": 0.747596800327301, |
| "rewards/accuracies": 0.75, |
| "rewards/chosen": -0.07475967705249786, |
| "rewards/margins": 0.015619462355971336, |
| "rewards/rejected": -0.09037913382053375, |
| "step": 5, |
| "train_speed(iter/s)": 0.01019 |
| }, |
| { |
| "epoch": 0.018250950570342206, |
| "grad_norm": 0.28389012813568115, |
| "learning_rate": 5.454545454545454e-05, |
| "log_odds_chosen": 0.07245179265737534, |
| "log_odds_ratio": -0.6642654538154602, |
| "logits/chosen": -2.207275867462158, |
| "logits/rejected": -2.194154977798462, |
| "logps/chosen": -0.7644542455673218, |
| "logps/rejected": -0.811606228351593, |
| "loss": 0.80996728, |
| "memory(GiB)": 304.24, |
| "nll_loss": 0.7644542455673218, |
| "rewards/accuracies": 0.625, |
| "rewards/chosen": -0.07644543051719666, |
| "rewards/margins": 0.004715194460004568, |
| "rewards/rejected": -0.08116061985492706, |
| "step": 6, |
| "train_speed(iter/s)": 0.010519 |
| }, |
| { |
| "epoch": 0.02129277566539924, |
| "grad_norm": 0.2204827517271042, |
| "learning_rate": 6.363636363636363e-05, |
| "log_odds_chosen": 0.0014549940824508667, |
| "log_odds_ratio": -0.6934571266174316, |
| "logits/chosen": -2.2422585487365723, |
| "logits/rejected": -2.2558586597442627, |
| "logps/chosen": -0.7055240869522095, |
| "logps/rejected": -0.7079155445098877, |
| "loss": 0.80136919, |
| "memory(GiB)": 376.14, |
| "nll_loss": 0.7055240869522095, |
| "rewards/accuracies": 0.5, |
| "rewards/chosen": -0.07055240869522095, |
| "rewards/margins": 0.0002391412854194641, |
| "rewards/rejected": -0.07079154998064041, |
| "step": 7, |
| "train_speed(iter/s)": 0.010453 |
| }, |
| { |
| "epoch": 0.024334600760456272, |
| "grad_norm": 0.18603529036045074, |
| "learning_rate": 7.272727272727273e-05, |
| "log_odds_chosen": 0.2205890268087387, |
| "log_odds_ratio": -0.6130101084709167, |
| "logits/chosen": -2.1775014400482178, |
| "logits/rejected": -2.1533212661743164, |
| "logps/chosen": -0.6938110589981079, |
| "logps/rejected": -0.8382289409637451, |
| "loss": 0.73867655, |
| "memory(GiB)": 337.96, |
| "nll_loss": 0.6938109993934631, |
| "rewards/accuracies": 0.875, |
| "rewards/chosen": -0.06938110291957855, |
| "rewards/margins": 0.01444179005920887, |
| "rewards/rejected": -0.08382289111614227, |
| "step": 8, |
| "train_speed(iter/s)": 0.010026 |
| }, |
| { |
| "epoch": 0.02737642585551331, |
| "grad_norm": 0.2093774527311325, |
| "learning_rate": 8.18181818181818e-05, |
| "log_odds_chosen": 0.28880590200424194, |
| "log_odds_ratio": -0.5970734357833862, |
| "logits/chosen": -2.231719970703125, |
| "logits/rejected": -2.1949002742767334, |
| "logps/chosen": -0.6275652647018433, |
| "logps/rejected": -0.8133708834648132, |
| "loss": 0.73195064, |
| "memory(GiB)": 337.96, |
| "nll_loss": 0.627565324306488, |
| "rewards/accuracies": 0.75, |
| "rewards/chosen": -0.06275652348995209, |
| "rewards/margins": 0.018580568954348564, |
| "rewards/rejected": -0.0813370943069458, |
| "step": 9, |
| "train_speed(iter/s)": 0.010471 |
| }, |
| { |
| "epoch": 0.030418250950570342, |
| "grad_norm": 0.17292404174804688, |
| "learning_rate": 9.09090909090909e-05, |
| "log_odds_chosen": 0.1367553025484085, |
| "log_odds_ratio": -0.6314333081245422, |
| "logits/chosen": -2.193378448486328, |
| "logits/rejected": -2.2005226612091064, |
| "logps/chosen": -0.712138831615448, |
| "logps/rejected": -0.7979795932769775, |
| "loss": 0.72619164, |
| "memory(GiB)": 337.96, |
| "nll_loss": 0.7121387720108032, |
| "rewards/accuracies": 0.75, |
| "rewards/chosen": -0.07121387869119644, |
| "rewards/margins": 0.00858408585190773, |
| "rewards/rejected": -0.07979796826839447, |
| "step": 10, |
| "train_speed(iter/s)": 0.010603 |
| }, |
| { |
| "epoch": 0.033460076045627375, |
| "grad_norm": 0.18145819008350372, |
| "learning_rate": 9.999999999999999e-05, |
| "log_odds_chosen": 0.39647987484931946, |
| "log_odds_ratio": -0.5509629249572754, |
| "logits/chosen": -1.9035614728927612, |
| "logits/rejected": -1.8070275783538818, |
| "logps/chosen": -0.6602268218994141, |
| "logps/rejected": -0.9264889359474182, |
| "loss": 0.63924813, |
| "memory(GiB)": 337.96, |
| "nll_loss": 0.6602268218994141, |
| "rewards/accuracies": 0.75, |
| "rewards/chosen": -0.06602267920970917, |
| "rewards/margins": 0.026626210659742355, |
| "rewards/rejected": -0.09264889359474182, |
| "step": 11, |
| "train_speed(iter/s)": 0.011075 |
| }, |
| { |
| "epoch": 0.03650190114068441, |
| "grad_norm": 0.1996176391839981, |
| "learning_rate": 0.00010909090909090908, |
| "log_odds_chosen": -0.008832626044750214, |
| "log_odds_ratio": -0.6989511251449585, |
| "logits/chosen": -2.1484761238098145, |
| "logits/rejected": -2.1583542823791504, |
| "logps/chosen": -0.5783950686454773, |
| "logps/rejected": -0.5806748867034912, |
| "loss": 0.66453189, |
| "memory(GiB)": 337.96, |
| "nll_loss": 0.5783950686454773, |
| "rewards/accuracies": 0.375, |
| "rewards/chosen": -0.05783950537443161, |
| "rewards/margins": 0.0002279849722981453, |
| "rewards/rejected": -0.05806749314069748, |
| "step": 12, |
| "train_speed(iter/s)": 0.011198 |
| }, |
| { |
| "epoch": 0.03954372623574144, |
| "grad_norm": 0.18517741560935974, |
| "learning_rate": 0.00011818181818181817, |
| "log_odds_chosen": 0.1502607762813568, |
| "log_odds_ratio": -0.6237398386001587, |
| "logits/chosen": -2.078707695007324, |
| "logits/rejected": -2.0867443084716797, |
| "logps/chosen": -0.6652951240539551, |
| "logps/rejected": -0.7532068490982056, |
| "loss": 0.71045852, |
| "memory(GiB)": 337.96, |
| "nll_loss": 0.6652951240539551, |
| "rewards/accuracies": 0.875, |
| "rewards/chosen": -0.0665295198559761, |
| "rewards/margins": 0.008791167289018631, |
| "rewards/rejected": -0.07532068341970444, |
| "step": 13, |
| "train_speed(iter/s)": 0.011144 |
| }, |
| { |
| "epoch": 0.04258555133079848, |
| "grad_norm": 0.460842102766037, |
| "learning_rate": 0.00012727272727272725, |
| "log_odds_chosen": 0.08848065137863159, |
| "log_odds_ratio": -0.6523964405059814, |
| "logits/chosen": -2.0903429985046387, |
| "logits/rejected": -2.071244716644287, |
| "logps/chosen": -0.5602254867553711, |
| "logps/rejected": -0.6098269820213318, |
| "loss": 0.62673962, |
| "memory(GiB)": 337.96, |
| "nll_loss": 0.5602255463600159, |
| "rewards/accuracies": 0.75, |
| "rewards/chosen": -0.05602255091071129, |
| "rewards/margins": 0.004960143007338047, |
| "rewards/rejected": -0.06098269298672676, |
| "step": 14, |
| "train_speed(iter/s)": 0.011223 |
| }, |
| { |
| "epoch": 0.045627376425855515, |
| "grad_norm": 0.201682448387146, |
| "learning_rate": 0.00013636363636363634, |
| "log_odds_chosen": 0.0586659274995327, |
| "log_odds_ratio": -0.6654743552207947, |
| "logits/chosen": -2.460188388824463, |
| "logits/rejected": -2.4563236236572266, |
| "logps/chosen": -0.3543914556503296, |
| "logps/rejected": -0.3674345910549164, |
| "loss": 0.49967062, |
| "memory(GiB)": 337.96, |
| "nll_loss": 0.3543914258480072, |
| "rewards/accuracies": 0.875, |
| "rewards/chosen": -0.0354391448199749, |
| "rewards/margins": 0.0013043158687651157, |
| "rewards/rejected": -0.03674346208572388, |
| "step": 15, |
| "train_speed(iter/s)": 0.011535 |
| }, |
| { |
| "epoch": 0.048669201520912544, |
| "grad_norm": 0.20425710082054138, |
| "learning_rate": 0.00014545454545454546, |
| "log_odds_chosen": 0.06445352733135223, |
| "log_odds_ratio": -0.6619336009025574, |
| "logits/chosen": -1.9317834377288818, |
| "logits/rejected": -1.9417508840560913, |
| "logps/chosen": -0.5264174938201904, |
| "logps/rejected": -0.5525864958763123, |
| "loss": 0.62982225, |
| "memory(GiB)": 337.96, |
| "nll_loss": 0.5264174938201904, |
| "rewards/accuracies": 0.75, |
| "rewards/chosen": -0.05264175310730934, |
| "rewards/margins": 0.0026168967597186565, |
| "rewards/rejected": -0.055258650332689285, |
| "step": 16, |
| "train_speed(iter/s)": 0.011487 |
| }, |
| { |
| "epoch": 0.05171102661596958, |
| "grad_norm": 0.19479066133499146, |
| "learning_rate": 0.00015454545454545452, |
| "log_odds_chosen": 0.05428645759820938, |
| "log_odds_ratio": -0.6680891513824463, |
| "logits/chosen": -2.0513648986816406, |
| "logits/rejected": -2.0547256469726562, |
| "logps/chosen": -0.5574195384979248, |
| "logps/rejected": -0.5769228935241699, |
| "loss": 0.65339994, |
| "memory(GiB)": 337.96, |
| "nll_loss": 0.5574195384979248, |
| "rewards/accuracies": 0.75, |
| "rewards/chosen": -0.05574195832014084, |
| "rewards/margins": 0.001950339414179325, |
| "rewards/rejected": -0.05769229307770729, |
| "step": 17, |
| "train_speed(iter/s)": 0.011193 |
| }, |
| { |
| "epoch": 0.05475285171102662, |
| "grad_norm": 0.20096692442893982, |
| "learning_rate": 0.0001636363636363636, |
| "log_odds_chosen": 0.11649945378303528, |
| "log_odds_ratio": -0.6375795006752014, |
| "logits/chosen": -1.979109764099121, |
| "logits/rejected": -1.9783812761306763, |
| "logps/chosen": -0.6062220335006714, |
| "logps/rejected": -0.6605215072631836, |
| "loss": 0.61762023, |
| "memory(GiB)": 337.96, |
| "nll_loss": 0.6062220931053162, |
| "rewards/accuracies": 0.875, |
| "rewards/chosen": -0.0606222040951252, |
| "rewards/margins": 0.005429944489151239, |
| "rewards/rejected": -0.06605214625597, |
| "step": 18, |
| "train_speed(iter/s)": 0.010973 |
| }, |
| { |
| "epoch": 0.05779467680608365, |
| "grad_norm": 0.22299402952194214, |
| "learning_rate": 0.00017272727272727272, |
| "log_odds_chosen": 0.275879442691803, |
| "log_odds_ratio": -0.5872041583061218, |
| "logits/chosen": -1.7563273906707764, |
| "logits/rejected": -1.7962664365768433, |
| "logps/chosen": -0.5380970239639282, |
| "logps/rejected": -0.7168144583702087, |
| "loss": 0.62556648, |
| "memory(GiB)": 337.96, |
| "nll_loss": 0.5380970239639282, |
| "rewards/accuracies": 0.75, |
| "rewards/chosen": -0.05380970239639282, |
| "rewards/margins": 0.017871743068099022, |
| "rewards/rejected": -0.071681447327137, |
| "step": 19, |
| "train_speed(iter/s)": 0.010995 |
| }, |
| { |
| "epoch": 0.060836501901140684, |
| "grad_norm": 0.19399508833885193, |
| "learning_rate": 0.0001818181818181818, |
| "log_odds_chosen": -0.016166899353265762, |
| "log_odds_ratio": -0.702415943145752, |
| "logits/chosen": -2.1521270275115967, |
| "logits/rejected": -2.151902914047241, |
| "logps/chosen": -0.41145703196525574, |
| "logps/rejected": -0.4066007733345032, |
| "loss": 0.49458012, |
| "memory(GiB)": 337.96, |
| "nll_loss": 0.41145703196525574, |
| "rewards/accuracies": 0.375, |
| "rewards/chosen": -0.041145700961351395, |
| "rewards/margins": -0.000485624885186553, |
| "rewards/rejected": -0.040660079568624496, |
| "step": 20, |
| "train_speed(iter/s)": 0.011065 |
| } |
| ], |
| "logging_steps": 1, |
| "max_steps": 328, |
| "num_input_tokens_seen": 0, |
| "num_train_epochs": 1, |
| "save_steps": 10, |
| "stateful_callbacks": { |
| "TrainerControl": { |
| "args": { |
| "should_epoch_stop": false, |
| "should_evaluate": false, |
| "should_log": false, |
| "should_save": true, |
| "should_training_stop": false |
| }, |
| "attributes": {} |
| } |
| }, |
| "total_flos": 9.570806552906957e+17, |
| "train_batch_size": 1, |
| "trial_name": null, |
| "trial_params": null |
| } |
|
|
