| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.9965156794425087, | |
| "eval_steps": 50, | |
| "global_step": 143, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.06968641114982578, | |
| "grad_norm": 7.942402444587964, | |
| "learning_rate": 3.333333333333333e-07, | |
| "logits/chosen": -2.747418165206909, | |
| "logits/rejected": -2.7369940280914307, | |
| "logps/chosen": -291.8316650390625, | |
| "logps/rejected": -281.5037536621094, | |
| "loss": 0.693, | |
| "rewards/accuracies": 0.45625001192092896, | |
| "rewards/chosen": 0.0003662299714051187, | |
| "rewards/margins": 0.0004625393485184759, | |
| "rewards/rejected": -9.630931162973866e-05, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.13937282229965156, | |
| "grad_norm": 8.199051257817597, | |
| "learning_rate": 4.981198836496775e-07, | |
| "logits/chosen": -2.7628424167633057, | |
| "logits/rejected": -2.7486491203308105, | |
| "logps/chosen": -298.9639587402344, | |
| "logps/rejected": -320.4031066894531, | |
| "loss": 0.6901, | |
| "rewards/accuracies": 0.550000011920929, | |
| "rewards/chosen": 0.008746958337724209, | |
| "rewards/margins": 0.006991321686655283, | |
| "rewards/rejected": 0.0017556389793753624, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.20905923344947736, | |
| "grad_norm": 8.009396280118287, | |
| "learning_rate": 4.832481997086846e-07, | |
| "logits/chosen": -2.713383436203003, | |
| "logits/rejected": -2.714205503463745, | |
| "logps/chosen": -296.2840881347656, | |
| "logps/rejected": -289.38421630859375, | |
| "loss": 0.681, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": 0.04019603878259659, | |
| "rewards/margins": 0.04010898619890213, | |
| "rewards/rejected": 8.704829815542325e-05, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.2787456445993031, | |
| "grad_norm": 8.916455655557288, | |
| "learning_rate": 4.543962032878959e-07, | |
| "logits/chosen": -2.679121732711792, | |
| "logits/rejected": -2.674286365509033, | |
| "logps/chosen": -298.8094177246094, | |
| "logps/rejected": -318.2736511230469, | |
| "loss": 0.6637, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -0.003670867532491684, | |
| "rewards/margins": 0.06940315663814545, | |
| "rewards/rejected": -0.07307400554418564, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.34843205574912894, | |
| "grad_norm": 11.007582392023577, | |
| "learning_rate": 4.1329321073844413e-07, | |
| "logits/chosen": -2.5453474521636963, | |
| "logits/rejected": -2.553844451904297, | |
| "logps/chosen": -272.39227294921875, | |
| "logps/rejected": -269.97125244140625, | |
| "loss": 0.6597, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": -0.1337047815322876, | |
| "rewards/margins": 0.06374003738164902, | |
| "rewards/rejected": -0.1974448263645172, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.34843205574912894, | |
| "eval_logits/chosen": -2.5355470180511475, | |
| "eval_logits/rejected": -2.496910572052002, | |
| "eval_logps/chosen": -273.103515625, | |
| "eval_logps/rejected": -285.6536865234375, | |
| "eval_loss": 0.6512949466705322, | |
| "eval_rewards/accuracies": 0.6640625, | |
| "eval_rewards/chosen": -0.1047358363866806, | |
| "eval_rewards/margins": 0.1251804083585739, | |
| "eval_rewards/rejected": -0.22991624474525452, | |
| "eval_runtime": 105.224, | |
| "eval_samples_per_second": 19.007, | |
| "eval_steps_per_second": 0.304, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.4181184668989547, | |
| "grad_norm": 11.130762557977343, | |
| "learning_rate": 3.624028324136517e-07, | |
| "logits/chosen": -2.5859858989715576, | |
| "logits/rejected": -2.590031147003174, | |
| "logps/chosen": -312.9039611816406, | |
| "logps/rejected": -316.5061950683594, | |
| "loss": 0.6515, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.10166505724191666, | |
| "rewards/margins": 0.13827314972877502, | |
| "rewards/rejected": -0.23993822932243347, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.4878048780487805, | |
| "grad_norm": 10.367706676784401, | |
| "learning_rate": 3.047753100392174e-07, | |
| "logits/chosen": -2.515353202819824, | |
| "logits/rejected": -2.5070269107818604, | |
| "logps/chosen": -303.077392578125, | |
| "logps/rejected": -301.082763671875, | |
| "loss": 0.6453, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.2464326173067093, | |
| "rewards/margins": 0.14911559224128723, | |
| "rewards/rejected": -0.3955482840538025, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.5574912891986062, | |
| "grad_norm": 10.328785377431698, | |
| "learning_rate": 2.4386469286927194e-07, | |
| "logits/chosen": -2.534677028656006, | |
| "logits/rejected": -2.515333652496338, | |
| "logps/chosen": -305.86944580078125, | |
| "logps/rejected": -315.22369384765625, | |
| "loss": 0.6525, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": -0.3514646887779236, | |
| "rewards/margins": 0.10919757932424545, | |
| "rewards/rejected": -0.460662305355072, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.627177700348432, | |
| "grad_norm": 10.256439047308396, | |
| "learning_rate": 1.8332181063127542e-07, | |
| "logits/chosen": -2.4058992862701416, | |
| "logits/rejected": -2.4308979511260986, | |
| "logps/chosen": -272.29986572265625, | |
| "logps/rejected": -304.98779296875, | |
| "loss": 0.6315, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": -0.19782552123069763, | |
| "rewards/margins": 0.18659016489982605, | |
| "rewards/rejected": -0.38441571593284607, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.6968641114982579, | |
| "grad_norm": 11.456138890836337, | |
| "learning_rate": 1.26775451942554e-07, | |
| "logits/chosen": -2.4451889991760254, | |
| "logits/rejected": -2.447258472442627, | |
| "logps/chosen": -332.0220031738281, | |
| "logps/rejected": -343.72979736328125, | |
| "loss": 0.6383, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.2864390015602112, | |
| "rewards/margins": 0.22521762549877167, | |
| "rewards/rejected": -0.511656641960144, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.6968641114982579, | |
| "eval_logits/chosen": -2.393218517303467, | |
| "eval_logits/rejected": -2.3483967781066895, | |
| "eval_logps/chosen": -295.345947265625, | |
| "eval_logps/rejected": -318.13665771484375, | |
| "eval_loss": 0.6311340928077698, | |
| "eval_rewards/accuracies": 0.6875, | |
| "eval_rewards/chosen": -0.3271602392196655, | |
| "eval_rewards/margins": 0.22758512198925018, | |
| "eval_rewards/rejected": -0.5547453165054321, | |
| "eval_runtime": 103.8391, | |
| "eval_samples_per_second": 19.261, | |
| "eval_steps_per_second": 0.308, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.7665505226480837, | |
| "grad_norm": 15.049705548114918, | |
| "learning_rate": 7.761486381573326e-08, | |
| "logits/chosen": -2.472832679748535, | |
| "logits/rejected": -2.4680256843566895, | |
| "logps/chosen": -347.2950439453125, | |
| "logps/rejected": -383.56903076171875, | |
| "loss": 0.6223, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -0.36344587802886963, | |
| "rewards/margins": 0.232518270611763, | |
| "rewards/rejected": -0.5959641933441162, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.8362369337979094, | |
| "grad_norm": 11.65420047853757, | |
| "learning_rate": 3.878660868757322e-08, | |
| "logits/chosen": -2.3938543796539307, | |
| "logits/rejected": -2.3905398845672607, | |
| "logps/chosen": -329.33978271484375, | |
| "logps/rejected": -346.9275817871094, | |
| "loss": 0.6335, | |
| "rewards/accuracies": 0.6187499761581421, | |
| "rewards/chosen": -0.34853675961494446, | |
| "rewards/margins": 0.21830621361732483, | |
| "rewards/rejected": -0.5668429136276245, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.9059233449477352, | |
| "grad_norm": 11.07371443957527, | |
| "learning_rate": 1.261795485174083e-08, | |
| "logits/chosen": -2.4500081539154053, | |
| "logits/rejected": -2.451570510864258, | |
| "logps/chosen": -330.7276306152344, | |
| "logps/rejected": -342.2732849121094, | |
| "loss": 0.626, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.3020505905151367, | |
| "rewards/margins": 0.23814749717712402, | |
| "rewards/rejected": -0.540198028087616, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.975609756097561, | |
| "grad_norm": 11.70940252198527, | |
| "learning_rate": 6.773858303274482e-10, | |
| "logits/chosen": -2.421135425567627, | |
| "logits/rejected": -2.4089722633361816, | |
| "logps/chosen": -300.1723937988281, | |
| "logps/rejected": -315.6776123046875, | |
| "loss": 0.6255, | |
| "rewards/accuracies": 0.5874999761581421, | |
| "rewards/chosen": -0.36181899905204773, | |
| "rewards/margins": 0.13425084948539734, | |
| "rewards/rejected": -0.49606984853744507, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.9965156794425087, | |
| "step": 143, | |
| "total_flos": 0.0, | |
| "train_loss": 0.6505647914392965, | |
| "train_runtime": 3852.398, | |
| "train_samples_per_second": 4.761, | |
| "train_steps_per_second": 0.037 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 143, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 100, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 0.0, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |