| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 1.0, | |
| "eval_steps": 500, | |
| "global_step": 313, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.003194888178913738, | |
| "grad_norm": 1.7594991820231036, | |
| "learning_rate": 5.3191489361702125e-09, | |
| "logits/chosen": -0.84765625, | |
| "logits/rejected": -0.89453125, | |
| "logps/chosen": -15.1875, | |
| "logps/rejected": -16.25, | |
| "loss": 0.6914, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.03194888178913738, | |
| "grad_norm": 3.2874000198076727, | |
| "learning_rate": 5.3191489361702123e-08, | |
| "logits/chosen": -0.890625, | |
| "logits/rejected": -0.875, | |
| "logps/chosen": -16.125, | |
| "logps/rejected": -16.5, | |
| "loss": 0.6917, | |
| "rewards/accuracies": 0.0972222238779068, | |
| "rewards/chosen": 0.0001735687255859375, | |
| "rewards/margins": 0.00130462646484375, | |
| "rewards/rejected": -0.001129150390625, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.06389776357827476, | |
| "grad_norm": 6.20375008201732, | |
| "learning_rate": 1.0638297872340425e-07, | |
| "logits/chosen": -0.88671875, | |
| "logits/rejected": -0.8984375, | |
| "logps/chosen": -16.25, | |
| "logps/rejected": -16.625, | |
| "loss": 0.692, | |
| "rewards/accuracies": 0.03750000149011612, | |
| "rewards/chosen": -7.62939453125e-05, | |
| "rewards/margins": -0.00023746490478515625, | |
| "rewards/rejected": 0.0001583099365234375, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.09584664536741214, | |
| "grad_norm": 4.664908763761001, | |
| "learning_rate": 1.5957446808510638e-07, | |
| "logits/chosen": -0.859375, | |
| "logits/rejected": -0.875, | |
| "logps/chosen": -16.125, | |
| "logps/rejected": -16.25, | |
| "loss": 0.6913, | |
| "rewards/accuracies": 0.05000000074505806, | |
| "rewards/chosen": 0.0020751953125, | |
| "rewards/margins": 0.00019550323486328125, | |
| "rewards/rejected": 0.0018768310546875, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.12779552715654952, | |
| "grad_norm": 2.4116741194375373, | |
| "learning_rate": 2.127659574468085e-07, | |
| "logits/chosen": -0.87890625, | |
| "logits/rejected": -0.8671875, | |
| "logps/chosen": -16.375, | |
| "logps/rejected": -16.5, | |
| "loss": 0.6908, | |
| "rewards/accuracies": 0.03750000149011612, | |
| "rewards/chosen": 0.00811767578125, | |
| "rewards/margins": 0.0003108978271484375, | |
| "rewards/rejected": 0.0078125, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.1597444089456869, | |
| "grad_norm": 3.308574154629096, | |
| "learning_rate": 2.659574468085106e-07, | |
| "logits/chosen": -0.87109375, | |
| "logits/rejected": -0.84765625, | |
| "logps/chosen": -14.9375, | |
| "logps/rejected": -15.6875, | |
| "loss": 0.691, | |
| "rewards/accuracies": 0.0625, | |
| "rewards/chosen": 0.0166015625, | |
| "rewards/margins": 0.0029754638671875, | |
| "rewards/rejected": 0.01361083984375, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.19169329073482427, | |
| "grad_norm": 1.6245800907371344, | |
| "learning_rate": 3.1914893617021275e-07, | |
| "logits/chosen": -0.859375, | |
| "logits/rejected": -0.85546875, | |
| "logps/chosen": -16.125, | |
| "logps/rejected": -15.0, | |
| "loss": 0.6908, | |
| "rewards/accuracies": 0.02500000037252903, | |
| "rewards/chosen": 0.0198974609375, | |
| "rewards/margins": -0.0006256103515625, | |
| "rewards/rejected": 0.0205078125, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.22364217252396165, | |
| "grad_norm": 2.6711590089749127, | |
| "learning_rate": 3.7234042553191484e-07, | |
| "logits/chosen": -0.8828125, | |
| "logits/rejected": -0.8828125, | |
| "logps/chosen": -16.5, | |
| "logps/rejected": -16.125, | |
| "loss": 0.6904, | |
| "rewards/accuracies": 0.10000000149011612, | |
| "rewards/chosen": 0.0224609375, | |
| "rewards/margins": 0.00579833984375, | |
| "rewards/rejected": 0.0167236328125, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.25559105431309903, | |
| "grad_norm": 2.3762135272130664, | |
| "learning_rate": 4.25531914893617e-07, | |
| "logits/chosen": -0.8515625, | |
| "logits/rejected": -0.84375, | |
| "logps/chosen": -14.9375, | |
| "logps/rejected": -15.25, | |
| "loss": 0.6856, | |
| "rewards/accuracies": 0.05000000074505806, | |
| "rewards/chosen": 0.057373046875, | |
| "rewards/margins": 0.007568359375, | |
| "rewards/rejected": 0.0498046875, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.28753993610223644, | |
| "grad_norm": 3.4884770405641, | |
| "learning_rate": 4.787234042553192e-07, | |
| "logits/chosen": -0.86328125, | |
| "logits/rejected": -0.8671875, | |
| "logps/chosen": -14.125, | |
| "logps/rejected": -14.0, | |
| "loss": 0.6812, | |
| "rewards/accuracies": 0.10000000149011612, | |
| "rewards/chosen": 0.142578125, | |
| "rewards/margins": 0.02001953125, | |
| "rewards/rejected": 0.12255859375, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.3194888178913738, | |
| "grad_norm": 4.958656955255395, | |
| "learning_rate": 4.964497041420119e-07, | |
| "logits/chosen": -0.859375, | |
| "logits/rejected": -0.89453125, | |
| "logps/chosen": -13.5625, | |
| "logps/rejected": -14.5, | |
| "loss": 0.6847, | |
| "rewards/accuracies": 0.0625, | |
| "rewards/chosen": 0.1826171875, | |
| "rewards/margins": 0.0277099609375, | |
| "rewards/rejected": 0.1552734375, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.3514376996805112, | |
| "grad_norm": 4.274608448473679, | |
| "learning_rate": 4.905325443786982e-07, | |
| "logits/chosen": -0.8515625, | |
| "logits/rejected": -0.8359375, | |
| "logps/chosen": -13.3125, | |
| "logps/rejected": -13.4375, | |
| "loss": 0.6855, | |
| "rewards/accuracies": 0.05000000074505806, | |
| "rewards/chosen": 0.2392578125, | |
| "rewards/margins": 0.0169677734375, | |
| "rewards/rejected": 0.22265625, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.38338658146964855, | |
| "grad_norm": 6.050646800243223, | |
| "learning_rate": 4.846153846153846e-07, | |
| "logits/chosen": -0.85546875, | |
| "logits/rejected": -0.84375, | |
| "logps/chosen": -12.3125, | |
| "logps/rejected": -12.6875, | |
| "loss": 0.6831, | |
| "rewards/accuracies": 0.07500000298023224, | |
| "rewards/chosen": 0.3046875, | |
| "rewards/margins": 0.0400390625, | |
| "rewards/rejected": 0.263671875, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.41533546325878595, | |
| "grad_norm": 0.584658593263476, | |
| "learning_rate": 4.78698224852071e-07, | |
| "logits/chosen": -0.828125, | |
| "logits/rejected": -0.875, | |
| "logps/chosen": -15.125, | |
| "logps/rejected": -17.0, | |
| "loss": 0.6783, | |
| "rewards/accuracies": 0.13750000298023224, | |
| "rewards/chosen": 0.19140625, | |
| "rewards/margins": 0.0830078125, | |
| "rewards/rejected": 0.10888671875, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.4472843450479233, | |
| "grad_norm": 2.1563062155083426, | |
| "learning_rate": 4.727810650887574e-07, | |
| "logits/chosen": -0.8671875, | |
| "logits/rejected": -0.8515625, | |
| "logps/chosen": -14.125, | |
| "logps/rejected": -15.1875, | |
| "loss": 0.6758, | |
| "rewards/accuracies": 0.11249999701976776, | |
| "rewards/chosen": 0.2275390625, | |
| "rewards/margins": 0.06640625, | |
| "rewards/rejected": 0.1611328125, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.4792332268370607, | |
| "grad_norm": 2.43580270301018, | |
| "learning_rate": 4.668639053254438e-07, | |
| "logits/chosen": -0.83984375, | |
| "logits/rejected": -0.84375, | |
| "logps/chosen": -13.6875, | |
| "logps/rejected": -14.5, | |
| "loss": 0.6744, | |
| "rewards/accuracies": 0.10000000149011612, | |
| "rewards/chosen": 0.232421875, | |
| "rewards/margins": 0.041259765625, | |
| "rewards/rejected": 0.19140625, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.5111821086261981, | |
| "grad_norm": 4.371745508926852, | |
| "learning_rate": 4.6094674556213014e-07, | |
| "logits/chosen": -0.828125, | |
| "logits/rejected": -0.8046875, | |
| "logps/chosen": -14.75, | |
| "logps/rejected": -14.5625, | |
| "loss": 0.6823, | |
| "rewards/accuracies": 0.0625, | |
| "rewards/chosen": 0.18359375, | |
| "rewards/margins": 0.0028533935546875, | |
| "rewards/rejected": 0.1806640625, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.5431309904153354, | |
| "grad_norm": 5.276282548018104, | |
| "learning_rate": 4.5502958579881655e-07, | |
| "logits/chosen": -0.796875, | |
| "logits/rejected": -0.79296875, | |
| "logps/chosen": -12.875, | |
| "logps/rejected": -13.1875, | |
| "loss": 0.6852, | |
| "rewards/accuracies": 0.10000000149011612, | |
| "rewards/chosen": 0.234375, | |
| "rewards/margins": 0.03857421875, | |
| "rewards/rejected": 0.1953125, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.5750798722044729, | |
| "grad_norm": 3.1471054069237403, | |
| "learning_rate": 4.491124260355029e-07, | |
| "logits/chosen": -0.82421875, | |
| "logits/rejected": -0.80078125, | |
| "logps/chosen": -15.125, | |
| "logps/rejected": -15.5, | |
| "loss": 0.668, | |
| "rewards/accuracies": 0.125, | |
| "rewards/chosen": 0.1435546875, | |
| "rewards/margins": 0.0712890625, | |
| "rewards/rejected": 0.072265625, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.6070287539936102, | |
| "grad_norm": 5.526336807387032, | |
| "learning_rate": 4.4319526627218936e-07, | |
| "logits/chosen": -0.77734375, | |
| "logits/rejected": -0.7890625, | |
| "logps/chosen": -14.375, | |
| "logps/rejected": -15.625, | |
| "loss": 0.6698, | |
| "rewards/accuracies": 0.08749999850988388, | |
| "rewards/chosen": 0.13671875, | |
| "rewards/margins": 0.060546875, | |
| "rewards/rejected": 0.076171875, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.6389776357827476, | |
| "grad_norm": 6.08437673538068, | |
| "learning_rate": 4.372781065088757e-07, | |
| "logits/chosen": -0.80078125, | |
| "logits/rejected": -0.8046875, | |
| "logps/chosen": -13.625, | |
| "logps/rejected": -14.375, | |
| "loss": 0.6795, | |
| "rewards/accuracies": 0.07500000298023224, | |
| "rewards/chosen": 0.20703125, | |
| "rewards/margins": 0.053466796875, | |
| "rewards/rejected": 0.1533203125, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.670926517571885, | |
| "grad_norm": 1.6033107403006597, | |
| "learning_rate": 4.313609467455621e-07, | |
| "logits/chosen": -0.80078125, | |
| "logits/rejected": -0.7890625, | |
| "logps/chosen": -12.875, | |
| "logps/rejected": -14.0625, | |
| "loss": 0.6702, | |
| "rewards/accuracies": 0.10000000149011612, | |
| "rewards/chosen": 0.291015625, | |
| "rewards/margins": 0.09326171875, | |
| "rewards/rejected": 0.197265625, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.7028753993610224, | |
| "grad_norm": 1.7290270560712724, | |
| "learning_rate": 4.2544378698224847e-07, | |
| "logits/chosen": -0.79296875, | |
| "logits/rejected": -0.78515625, | |
| "logps/chosen": -13.25, | |
| "logps/rejected": -13.5625, | |
| "loss": 0.6655, | |
| "rewards/accuracies": 0.0625, | |
| "rewards/chosen": 0.2236328125, | |
| "rewards/margins": 0.04052734375, | |
| "rewards/rejected": 0.1826171875, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.7348242811501597, | |
| "grad_norm": 3.4471878821701267, | |
| "learning_rate": 4.195266272189349e-07, | |
| "logits/chosen": -0.7890625, | |
| "logits/rejected": -0.796875, | |
| "logps/chosen": -15.625, | |
| "logps/rejected": -16.5, | |
| "loss": 0.6706, | |
| "rewards/accuracies": 0.125, | |
| "rewards/chosen": 0.138671875, | |
| "rewards/margins": 0.11083984375, | |
| "rewards/rejected": 0.02783203125, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.7667731629392971, | |
| "grad_norm": 4.299918979944633, | |
| "learning_rate": 4.1360946745562133e-07, | |
| "logits/chosen": -0.76171875, | |
| "logits/rejected": -0.765625, | |
| "logps/chosen": -14.875, | |
| "logps/rejected": -15.8125, | |
| "loss": 0.6774, | |
| "rewards/accuracies": 0.0625, | |
| "rewards/chosen": 0.0732421875, | |
| "rewards/margins": 0.08251953125, | |
| "rewards/rejected": -0.0093994140625, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.7987220447284346, | |
| "grad_norm": 4.20769528530577, | |
| "learning_rate": 4.076923076923077e-07, | |
| "logits/chosen": -0.8046875, | |
| "logits/rejected": -0.7890625, | |
| "logps/chosen": -14.5625, | |
| "logps/rejected": -14.375, | |
| "loss": 0.6723, | |
| "rewards/accuracies": 0.08749999850988388, | |
| "rewards/chosen": 0.1806640625, | |
| "rewards/margins": 0.022705078125, | |
| "rewards/rejected": 0.158203125, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.8306709265175719, | |
| "grad_norm": 4.95860147750649, | |
| "learning_rate": 4.017751479289941e-07, | |
| "logits/chosen": -0.7578125, | |
| "logits/rejected": -0.7421875, | |
| "logps/chosen": -15.9375, | |
| "logps/rejected": -17.0, | |
| "loss": 0.6511, | |
| "rewards/accuracies": 0.10000000149011612, | |
| "rewards/chosen": 0.111328125, | |
| "rewards/margins": 0.1474609375, | |
| "rewards/rejected": -0.0361328125, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.8626198083067093, | |
| "grad_norm": 1.5354295895251373, | |
| "learning_rate": 3.9585798816568044e-07, | |
| "logits/chosen": -0.76171875, | |
| "logits/rejected": -0.7890625, | |
| "logps/chosen": -12.8125, | |
| "logps/rejected": -15.0625, | |
| "loss": 0.6686, | |
| "rewards/accuracies": 0.10000000149011612, | |
| "rewards/chosen": 0.21875, | |
| "rewards/margins": 0.10888671875, | |
| "rewards/rejected": 0.1103515625, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.8945686900958466, | |
| "grad_norm": 1.9206310698954268, | |
| "learning_rate": 3.8994082840236685e-07, | |
| "logits/chosen": -0.75390625, | |
| "logits/rejected": -0.76953125, | |
| "logps/chosen": -12.75, | |
| "logps/rejected": -14.5, | |
| "loss": 0.6707, | |
| "rewards/accuracies": 0.0625, | |
| "rewards/chosen": 0.2314453125, | |
| "rewards/margins": 0.1171875, | |
| "rewards/rejected": 0.1142578125, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.9265175718849841, | |
| "grad_norm": 2.1151877094244007, | |
| "learning_rate": 3.840236686390532e-07, | |
| "logits/chosen": -0.78515625, | |
| "logits/rejected": -0.8046875, | |
| "logps/chosen": -13.5, | |
| "logps/rejected": -15.4375, | |
| "loss": 0.6668, | |
| "rewards/accuracies": 0.07500000298023224, | |
| "rewards/chosen": 0.166015625, | |
| "rewards/margins": 0.0947265625, | |
| "rewards/rejected": 0.0712890625, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.9584664536741214, | |
| "grad_norm": 1.721440878382303, | |
| "learning_rate": 3.7810650887573966e-07, | |
| "logits/chosen": -0.796875, | |
| "logits/rejected": -0.8125, | |
| "logps/chosen": -14.5, | |
| "logps/rejected": -16.125, | |
| "loss": 0.6753, | |
| "rewards/accuracies": 0.08749999850988388, | |
| "rewards/chosen": 0.1171875, | |
| "rewards/margins": 0.1220703125, | |
| "rewards/rejected": -0.0047607421875, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.9904153354632588, | |
| "grad_norm": 0.9947102905281956, | |
| "learning_rate": 3.72189349112426e-07, | |
| "logits/chosen": -0.76171875, | |
| "logits/rejected": -0.78125, | |
| "logps/chosen": -16.5, | |
| "logps/rejected": -16.625, | |
| "loss": 0.6691, | |
| "rewards/accuracies": 0.08749999850988388, | |
| "rewards/chosen": 0.1416015625, | |
| "rewards/margins": 0.10986328125, | |
| "rewards/rejected": 0.0322265625, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_logits/chosen": -0.765625, | |
| "eval_logits/rejected": -0.7734375, | |
| "eval_logps/chosen": -14.5625, | |
| "eval_logps/rejected": -15.6875, | |
| "eval_loss": 0.6690624952316284, | |
| "eval_rewards/accuracies": 0.0892857164144516, | |
| "eval_rewards/chosen": 0.0439453125, | |
| "eval_rewards/margins": 0.06787109375, | |
| "eval_rewards/rejected": -0.0240478515625, | |
| "eval_runtime": 12.3648, | |
| "eval_samples_per_second": 16.175, | |
| "eval_steps_per_second": 0.566, | |
| "step": 313 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 939, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "save_steps": 500, | |
| "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": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |