| { | |
| "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": 26.70895145343874, | |
| "learning_rate": 5.3191489361702125e-09, | |
| "logits/chosen": -1.4921875, | |
| "logits/rejected": -1.4921875, | |
| "logps/chosen": -247.0, | |
| "logps/rejected": -252.0, | |
| "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": 27.524419104861234, | |
| "learning_rate": 5.3191489361702123e-08, | |
| "logits/chosen": -1.5390625, | |
| "logits/rejected": -1.59375, | |
| "logps/chosen": -252.0, | |
| "logps/rejected": -290.0, | |
| "loss": 0.6921, | |
| "rewards/accuracies": 0.1805555522441864, | |
| "rewards/chosen": 0.004180908203125, | |
| "rewards/margins": 0.0069580078125, | |
| "rewards/rejected": -0.0027618408203125, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.06389776357827476, | |
| "grad_norm": 25.84743681377418, | |
| "learning_rate": 1.0638297872340425e-07, | |
| "logits/chosen": -1.5546875, | |
| "logits/rejected": -1.6015625, | |
| "logps/chosen": -244.0, | |
| "logps/rejected": -284.0, | |
| "loss": 0.6878, | |
| "rewards/accuracies": 0.16249999403953552, | |
| "rewards/chosen": -0.00750732421875, | |
| "rewards/margins": -0.00750732421875, | |
| "rewards/rejected": -6.109476089477539e-06, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.09584664536741214, | |
| "grad_norm": 26.132304612422136, | |
| "learning_rate": 1.5957446808510638e-07, | |
| "logits/chosen": -1.5625, | |
| "logits/rejected": -1.5625, | |
| "logps/chosen": -245.0, | |
| "logps/rejected": -282.0, | |
| "loss": 0.6743, | |
| "rewards/accuracies": 0.4124999940395355, | |
| "rewards/chosen": 0.038818359375, | |
| "rewards/margins": 0.0244140625, | |
| "rewards/rejected": 0.014404296875, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.12779552715654952, | |
| "grad_norm": 24.407986537974015, | |
| "learning_rate": 2.127659574468085e-07, | |
| "logits/chosen": -1.515625, | |
| "logits/rejected": -1.5859375, | |
| "logps/chosen": -237.0, | |
| "logps/rejected": -308.0, | |
| "loss": 0.6435, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": 0.1591796875, | |
| "rewards/margins": 0.1240234375, | |
| "rewards/rejected": 0.03515625, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.1597444089456869, | |
| "grad_norm": 21.9426563067355, | |
| "learning_rate": 2.659574468085106e-07, | |
| "logits/chosen": -1.5625, | |
| "logits/rejected": -1.640625, | |
| "logps/chosen": -239.0, | |
| "logps/rejected": -286.0, | |
| "loss": 0.6137, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": 0.310546875, | |
| "rewards/margins": 0.2275390625, | |
| "rewards/rejected": 0.08251953125, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.19169329073482427, | |
| "grad_norm": 20.96670132168061, | |
| "learning_rate": 3.1914893617021275e-07, | |
| "logits/chosen": -1.5859375, | |
| "logits/rejected": -1.6171875, | |
| "logps/chosen": -242.0, | |
| "logps/rejected": -292.0, | |
| "loss": 0.5732, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": 0.5, | |
| "rewards/margins": 0.376953125, | |
| "rewards/rejected": 0.12109375, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.22364217252396165, | |
| "grad_norm": 20.027837530897173, | |
| "learning_rate": 3.7234042553191484e-07, | |
| "logits/chosen": -1.515625, | |
| "logits/rejected": -1.625, | |
| "logps/chosen": -235.0, | |
| "logps/rejected": -296.0, | |
| "loss": 0.5562, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": 0.6953125, | |
| "rewards/margins": 0.5390625, | |
| "rewards/rejected": 0.1552734375, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.25559105431309903, | |
| "grad_norm": 20.104841517252364, | |
| "learning_rate": 4.25531914893617e-07, | |
| "logits/chosen": -1.5859375, | |
| "logits/rejected": -1.6328125, | |
| "logps/chosen": -245.0, | |
| "logps/rejected": -274.0, | |
| "loss": 0.5644, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": 0.6875, | |
| "rewards/margins": 0.40234375, | |
| "rewards/rejected": 0.28515625, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.28753993610223644, | |
| "grad_norm": 18.79087036985414, | |
| "learning_rate": 4.787234042553192e-07, | |
| "logits/chosen": -1.515625, | |
| "logits/rejected": -1.6015625, | |
| "logps/chosen": -236.0, | |
| "logps/rejected": -284.0, | |
| "loss": 0.5118, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.87109375, | |
| "rewards/margins": 0.5859375, | |
| "rewards/rejected": 0.287109375, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.3194888178913738, | |
| "grad_norm": 22.066743357299675, | |
| "learning_rate": 4.964497041420119e-07, | |
| "logits/chosen": -1.578125, | |
| "logits/rejected": -1.6328125, | |
| "logps/chosen": -231.0, | |
| "logps/rejected": -288.0, | |
| "loss": 0.5409, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": 1.0078125, | |
| "rewards/margins": 0.890625, | |
| "rewards/rejected": 0.12060546875, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.3514376996805112, | |
| "grad_norm": 19.3322147725973, | |
| "learning_rate": 4.905325443786982e-07, | |
| "logits/chosen": -1.546875, | |
| "logits/rejected": -1.6328125, | |
| "logps/chosen": -248.0, | |
| "logps/rejected": -280.0, | |
| "loss": 0.5656, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 0.69921875, | |
| "rewards/margins": 0.458984375, | |
| "rewards/rejected": 0.2392578125, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.38338658146964855, | |
| "grad_norm": 18.958252526138903, | |
| "learning_rate": 4.846153846153846e-07, | |
| "logits/chosen": -1.625, | |
| "logits/rejected": -1.703125, | |
| "logps/chosen": -240.0, | |
| "logps/rejected": -292.0, | |
| "loss": 0.5077, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": 0.8671875, | |
| "rewards/margins": 0.75390625, | |
| "rewards/rejected": 0.11376953125, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.41533546325878595, | |
| "grad_norm": 24.459723181791798, | |
| "learning_rate": 4.78698224852071e-07, | |
| "logits/chosen": -1.5625, | |
| "logits/rejected": -1.625, | |
| "logps/chosen": -235.0, | |
| "logps/rejected": -294.0, | |
| "loss": 0.4762, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": 0.86328125, | |
| "rewards/margins": 0.81640625, | |
| "rewards/rejected": 0.04638671875, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.4472843450479233, | |
| "grad_norm": 24.8748418291271, | |
| "learning_rate": 4.727810650887574e-07, | |
| "logits/chosen": -1.578125, | |
| "logits/rejected": -1.6484375, | |
| "logps/chosen": -242.0, | |
| "logps/rejected": -282.0, | |
| "loss": 0.5377, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": 0.7890625, | |
| "rewards/margins": 0.640625, | |
| "rewards/rejected": 0.1484375, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.4792332268370607, | |
| "grad_norm": 21.92108677224862, | |
| "learning_rate": 4.668639053254438e-07, | |
| "logits/chosen": -1.578125, | |
| "logits/rejected": -1.671875, | |
| "logps/chosen": -236.0, | |
| "logps/rejected": -302.0, | |
| "loss": 0.5036, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": 0.82421875, | |
| "rewards/margins": 0.79296875, | |
| "rewards/rejected": 0.029541015625, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.5111821086261981, | |
| "grad_norm": 23.160240290542404, | |
| "learning_rate": 4.6094674556213014e-07, | |
| "logits/chosen": -1.6328125, | |
| "logits/rejected": -1.6640625, | |
| "logps/chosen": -236.0, | |
| "logps/rejected": -290.0, | |
| "loss": 0.5215, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": 0.84765625, | |
| "rewards/margins": 0.984375, | |
| "rewards/rejected": -0.1337890625, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.5431309904153354, | |
| "grad_norm": 15.682473129649416, | |
| "learning_rate": 4.5502958579881655e-07, | |
| "logits/chosen": -1.6171875, | |
| "logits/rejected": -1.6875, | |
| "logps/chosen": -230.0, | |
| "logps/rejected": -292.0, | |
| "loss": 0.4507, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": 0.9296875, | |
| "rewards/margins": 0.9609375, | |
| "rewards/rejected": -0.02880859375, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.5750798722044729, | |
| "grad_norm": 18.270260505484856, | |
| "learning_rate": 4.491124260355029e-07, | |
| "logits/chosen": -1.5859375, | |
| "logits/rejected": -1.609375, | |
| "logps/chosen": -229.0, | |
| "logps/rejected": -298.0, | |
| "loss": 0.4635, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": 1.0078125, | |
| "rewards/margins": 1.0234375, | |
| "rewards/rejected": -0.01531982421875, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.6070287539936102, | |
| "grad_norm": 20.98631167244082, | |
| "learning_rate": 4.4319526627218936e-07, | |
| "logits/chosen": -1.6171875, | |
| "logits/rejected": -1.65625, | |
| "logps/chosen": -230.0, | |
| "logps/rejected": -294.0, | |
| "loss": 0.4752, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": 0.9296875, | |
| "rewards/margins": 0.953125, | |
| "rewards/rejected": -0.021484375, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.6389776357827476, | |
| "grad_norm": 22.61128808819513, | |
| "learning_rate": 4.372781065088757e-07, | |
| "logits/chosen": -1.59375, | |
| "logits/rejected": -1.6953125, | |
| "logps/chosen": -233.0, | |
| "logps/rejected": -294.0, | |
| "loss": 0.4676, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 0.90234375, | |
| "rewards/margins": 1.0, | |
| "rewards/rejected": -0.1005859375, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.670926517571885, | |
| "grad_norm": 20.472996932111148, | |
| "learning_rate": 4.313609467455621e-07, | |
| "logits/chosen": -1.5625, | |
| "logits/rejected": -1.65625, | |
| "logps/chosen": -236.0, | |
| "logps/rejected": -292.0, | |
| "loss": 0.4827, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.890625, | |
| "rewards/margins": 0.98828125, | |
| "rewards/rejected": -0.1015625, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.7028753993610224, | |
| "grad_norm": 25.186112119569938, | |
| "learning_rate": 4.2544378698224847e-07, | |
| "logits/chosen": -1.5390625, | |
| "logits/rejected": -1.5859375, | |
| "logps/chosen": -238.0, | |
| "logps/rejected": -286.0, | |
| "loss": 0.5209, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": 0.76171875, | |
| "rewards/margins": 0.92578125, | |
| "rewards/rejected": -0.1640625, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.7348242811501597, | |
| "grad_norm": 15.532080715300815, | |
| "learning_rate": 4.195266272189349e-07, | |
| "logits/chosen": -1.578125, | |
| "logits/rejected": -1.578125, | |
| "logps/chosen": -236.0, | |
| "logps/rejected": -290.0, | |
| "loss": 0.4544, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": 0.7578125, | |
| "rewards/margins": 0.8515625, | |
| "rewards/rejected": -0.095703125, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.7667731629392971, | |
| "grad_norm": 21.562659615610453, | |
| "learning_rate": 4.1360946745562133e-07, | |
| "logits/chosen": -1.625, | |
| "logits/rejected": -1.640625, | |
| "logps/chosen": -238.0, | |
| "logps/rejected": -292.0, | |
| "loss": 0.4312, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": 0.82421875, | |
| "rewards/margins": 0.96484375, | |
| "rewards/rejected": -0.1416015625, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.7987220447284346, | |
| "grad_norm": 19.697962308480218, | |
| "learning_rate": 4.076923076923077e-07, | |
| "logits/chosen": -1.578125, | |
| "logits/rejected": -1.671875, | |
| "logps/chosen": -241.0, | |
| "logps/rejected": -286.0, | |
| "loss": 0.4683, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": 0.7421875, | |
| "rewards/margins": 0.97265625, | |
| "rewards/rejected": -0.23046875, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.8306709265175719, | |
| "grad_norm": 19.922270403408632, | |
| "learning_rate": 4.017751479289941e-07, | |
| "logits/chosen": -1.5859375, | |
| "logits/rejected": -1.6875, | |
| "logps/chosen": -235.0, | |
| "logps/rejected": -298.0, | |
| "loss": 0.4589, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 0.84765625, | |
| "rewards/margins": 1.1796875, | |
| "rewards/rejected": -0.33203125, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.8626198083067093, | |
| "grad_norm": 23.03342264376932, | |
| "learning_rate": 3.9585798816568044e-07, | |
| "logits/chosen": -1.5703125, | |
| "logits/rejected": -1.6484375, | |
| "logps/chosen": -246.0, | |
| "logps/rejected": -292.0, | |
| "loss": 0.4974, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": 0.609375, | |
| "rewards/margins": 0.8125, | |
| "rewards/rejected": -0.203125, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.8945686900958466, | |
| "grad_norm": 18.556977289134693, | |
| "learning_rate": 3.8994082840236685e-07, | |
| "logits/chosen": -1.625, | |
| "logits/rejected": -1.6640625, | |
| "logps/chosen": -248.0, | |
| "logps/rejected": -280.0, | |
| "loss": 0.4835, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": 0.66796875, | |
| "rewards/margins": 0.859375, | |
| "rewards/rejected": -0.1904296875, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.9265175718849841, | |
| "grad_norm": 20.159977656950176, | |
| "learning_rate": 3.840236686390532e-07, | |
| "logits/chosen": -1.578125, | |
| "logits/rejected": -1.6484375, | |
| "logps/chosen": -227.0, | |
| "logps/rejected": -298.0, | |
| "loss": 0.4328, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": 0.79296875, | |
| "rewards/margins": 1.046875, | |
| "rewards/rejected": -0.251953125, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.9584664536741214, | |
| "grad_norm": 18.20653927593905, | |
| "learning_rate": 3.7810650887573966e-07, | |
| "logits/chosen": -1.515625, | |
| "logits/rejected": -1.59375, | |
| "logps/chosen": -239.0, | |
| "logps/rejected": -296.0, | |
| "loss": 0.4848, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": 0.68359375, | |
| "rewards/margins": 1.09375, | |
| "rewards/rejected": -0.412109375, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.9904153354632588, | |
| "grad_norm": 22.32128048324979, | |
| "learning_rate": 3.72189349112426e-07, | |
| "logits/chosen": -1.578125, | |
| "logits/rejected": -1.59375, | |
| "logps/chosen": -241.0, | |
| "logps/rejected": -290.0, | |
| "loss": 0.5057, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.60546875, | |
| "rewards/margins": 0.86328125, | |
| "rewards/rejected": -0.255859375, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_logits/chosen": -1.609375, | |
| "eval_logits/rejected": -1.6171875, | |
| "eval_logps/chosen": -237.0, | |
| "eval_logps/rejected": -296.0, | |
| "eval_loss": 0.43648436665534973, | |
| "eval_rewards/accuracies": 0.7857142686843872, | |
| "eval_rewards/chosen": 0.58203125, | |
| "eval_rewards/margins": 0.9375, | |
| "eval_rewards/rejected": -0.35546875, | |
| "eval_runtime": 12.3846, | |
| "eval_samples_per_second": 16.149, | |
| "eval_steps_per_second": 0.565, | |
| "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 | |
| } | |