| { |
| "best_global_step": null, |
| "best_metric": null, |
| "best_model_checkpoint": null, |
| "epoch": 1.5714285714285714, |
| "eval_steps": 500, |
| "global_step": 20, |
| "is_hyper_param_search": false, |
| "is_local_process_zero": true, |
| "is_world_process_zero": true, |
| "log_history": [ |
| { |
| "epoch": 0.08163265306122448, |
| "grad_norm": 1.025234580039978, |
| "learning_rate": 0.0, |
| "logits/chosen": -2.375152349472046, |
| "logits/rejected": -2.394904375076294, |
| "logps/chosen": -84.37759399414062, |
| "logps/rejected": -96.84375, |
| "loss": 0.6931, |
| "rewards/accuracies": 0.0, |
| "rewards/chosen": 0.0, |
| "rewards/margins": 0.0, |
| "rewards/rejected": 0.0, |
| "step": 1 |
| }, |
| { |
| "epoch": 0.16326530612244897, |
| "grad_norm": 1.1001588106155396, |
| "learning_rate": 2.5e-06, |
| "logits/chosen": -2.3984360694885254, |
| "logits/rejected": -2.4267055988311768, |
| "logps/chosen": -89.42276000976562, |
| "logps/rejected": -96.12063598632812, |
| "loss": 0.6931, |
| "rewards/accuracies": 0.0, |
| "rewards/chosen": 0.0, |
| "rewards/margins": 0.0, |
| "rewards/rejected": 0.0, |
| "step": 2 |
| }, |
| { |
| "epoch": 0.24489795918367346, |
| "grad_norm": 1.0922303199768066, |
| "learning_rate": 5e-06, |
| "logits/chosen": -2.426487922668457, |
| "logits/rejected": -2.437016487121582, |
| "logps/chosen": -90.65304565429688, |
| "logps/rejected": -98.02507019042969, |
| "loss": 0.6942, |
| "rewards/accuracies": 0.48681640625, |
| "rewards/chosen": -0.004277409985661507, |
| "rewards/margins": -0.0012883469462394714, |
| "rewards/rejected": -0.0029890630394220352, |
| "step": 3 |
| }, |
| { |
| "epoch": 0.32653061224489793, |
| "grad_norm": 1.0690197944641113, |
| "learning_rate": 7.500000000000001e-06, |
| "logits/chosen": -2.4016366004943848, |
| "logits/rejected": -2.404703378677368, |
| "logps/chosen": -88.64875793457031, |
| "logps/rejected": -100.15958404541016, |
| "loss": 0.6925, |
| "rewards/accuracies": 0.48828125, |
| "rewards/chosen": 0.0017670390661805868, |
| "rewards/margins": 0.0021676155738532543, |
| "rewards/rejected": -0.0004005763039458543, |
| "step": 4 |
| }, |
| { |
| "epoch": 0.40816326530612246, |
| "grad_norm": 1.0020982027053833, |
| "learning_rate": 1e-05, |
| "logits/chosen": -2.405047655105591, |
| "logits/rejected": -2.4170143604278564, |
| "logps/chosen": -86.40444946289062, |
| "logps/rejected": -99.32535552978516, |
| "loss": 0.692, |
| "rewards/accuracies": 0.52099609375, |
| "rewards/chosen": 0.002238229615613818, |
| "rewards/margins": 0.003263898193836212, |
| "rewards/rejected": -0.001025668578222394, |
| "step": 5 |
| }, |
| { |
| "epoch": 0.4897959183673469, |
| "grad_norm": 0.9774695038795471, |
| "learning_rate": 9.979871469976197e-06, |
| "logits/chosen": -2.387697696685791, |
| "logits/rejected": -2.40444278717041, |
| "logps/chosen": -88.27230072021484, |
| "logps/rejected": -96.35655212402344, |
| "loss": 0.6868, |
| "rewards/accuracies": 0.56005859375, |
| "rewards/chosen": 0.026464153081178665, |
| "rewards/margins": 0.013977548107504845, |
| "rewards/rejected": 0.01248660497367382, |
| "step": 6 |
| }, |
| { |
| "epoch": 0.5714285714285714, |
| "grad_norm": 0.9563448429107666, |
| "learning_rate": 9.91964794299315e-06, |
| "logits/chosen": -2.3706889152526855, |
| "logits/rejected": -2.395195722579956, |
| "logps/chosen": -85.99250793457031, |
| "logps/rejected": -96.60533905029297, |
| "loss": 0.6829, |
| "rewards/accuracies": 0.59912109375, |
| "rewards/chosen": 0.03764727711677551, |
| "rewards/margins": 0.02200627699494362, |
| "rewards/rejected": 0.015641000121831894, |
| "step": 7 |
| }, |
| { |
| "epoch": 0.6530612244897959, |
| "grad_norm": 1.0272670984268188, |
| "learning_rate": 9.819814303479268e-06, |
| "logits/chosen": -2.3650197982788086, |
| "logits/rejected": -2.365727424621582, |
| "logps/chosen": -89.17411804199219, |
| "logps/rejected": -104.6484146118164, |
| "loss": 0.6763, |
| "rewards/accuracies": 0.65966796875, |
| "rewards/chosen": 0.053641848266124725, |
| "rewards/margins": 0.03600878268480301, |
| "rewards/rejected": 0.017633063718676567, |
| "step": 8 |
| }, |
| { |
| "epoch": 0.7346938775510204, |
| "grad_norm": 0.9614192247390747, |
| "learning_rate": 9.681174353198687e-06, |
| "logits/chosen": -2.3682937622070312, |
| "logits/rejected": -2.3996546268463135, |
| "logps/chosen": -84.27474975585938, |
| "logps/rejected": -93.97203063964844, |
| "loss": 0.6699, |
| "rewards/accuracies": 0.6728515625, |
| "rewards/chosen": 0.07653943449258804, |
| "rewards/margins": 0.05041676014661789, |
| "rewards/rejected": 0.026122674345970154, |
| "step": 9 |
| }, |
| { |
| "epoch": 0.8163265306122449, |
| "grad_norm": 0.9354036450386047, |
| "learning_rate": 9.504844339512096e-06, |
| "logits/chosen": -2.3933653831481934, |
| "logits/rejected": -2.3978521823883057, |
| "logps/chosen": -84.89434051513672, |
| "logps/rejected": -97.98092651367188, |
| "loss": 0.6657, |
| "rewards/accuracies": 0.68896484375, |
| "rewards/chosen": 0.10172732174396515, |
| "rewards/margins": 0.060384318232536316, |
| "rewards/rejected": 0.04134300351142883, |
| "step": 10 |
| }, |
| { |
| "epoch": 0.8979591836734694, |
| "grad_norm": 0.9544646143913269, |
| "learning_rate": 9.292243968009332e-06, |
| "logits/chosen": -2.388801336288452, |
| "logits/rejected": -2.4009108543395996, |
| "logps/chosen": -87.04521179199219, |
| "logps/rejected": -99.57186889648438, |
| "loss": 0.6578, |
| "rewards/accuracies": 0.71240234375, |
| "rewards/chosen": 0.12685298919677734, |
| "rewards/margins": 0.07852301746606827, |
| "rewards/rejected": 0.048329971730709076, |
| "step": 11 |
| }, |
| { |
| "epoch": 0.9795918367346939, |
| "grad_norm": 0.945184051990509, |
| "learning_rate": 9.045084971874738e-06, |
| "logits/chosen": -2.4180715084075928, |
| "logits/rejected": -2.4203011989593506, |
| "logps/chosen": -87.21138763427734, |
| "logps/rejected": -101.43778991699219, |
| "loss": 0.648, |
| "rewards/accuracies": 0.7294921875, |
| "rewards/chosen": 0.15163584053516388, |
| "rewards/margins": 0.10153679549694061, |
| "rewards/rejected": 0.05009904503822327, |
| "step": 12 |
| }, |
| { |
| "epoch": 1.0, |
| "grad_norm": 0.9313170909881592, |
| "learning_rate": 8.765357330018056e-06, |
| "logits/chosen": -2.4016616344451904, |
| "logits/rejected": -2.461811065673828, |
| "logps/chosen": -91.43336486816406, |
| "logps/rejected": -93.41850280761719, |
| "loss": 0.6428, |
| "rewards/accuracies": 0.71875, |
| "rewards/chosen": 0.1737888753414154, |
| "rewards/margins": 0.11637693643569946, |
| "rewards/rejected": 0.05741193890571594, |
| "step": 13 |
| }, |
| { |
| "epoch": 1.0816326530612246, |
| "grad_norm": 0.879965603351593, |
| "learning_rate": 8.455313244934324e-06, |
| "logits/chosen": -2.3858630657196045, |
| "logits/rejected": -2.3951520919799805, |
| "logps/chosen": -84.2239761352539, |
| "logps/rejected": -97.90269470214844, |
| "loss": 0.6378, |
| "rewards/accuracies": 0.7109375, |
| "rewards/chosen": 0.19609522819519043, |
| "rewards/margins": 0.13054177165031433, |
| "rewards/rejected": 0.0655534565448761, |
| "step": 14 |
| }, |
| { |
| "epoch": 1.163265306122449, |
| "grad_norm": 0.8709496855735779, |
| "learning_rate": 8.117449009293668e-06, |
| "logits/chosen": -2.3995306491851807, |
| "logits/rejected": -2.4127840995788574, |
| "logps/chosen": -85.27073669433594, |
| "logps/rejected": -98.00736236572266, |
| "loss": 0.6306, |
| "rewards/accuracies": 0.7041015625, |
| "rewards/chosen": 0.21614021062850952, |
| "rewards/margins": 0.15087103843688965, |
| "rewards/rejected": 0.06526917964220047, |
| "step": 15 |
| }, |
| { |
| "epoch": 1.2448979591836735, |
| "grad_norm": 0.9011364579200745, |
| "learning_rate": 7.754484907260513e-06, |
| "logits/chosen": -2.3790643215179443, |
| "logits/rejected": -2.3908865451812744, |
| "logps/chosen": -93.4906234741211, |
| "logps/rejected": -104.75448608398438, |
| "loss": 0.6132, |
| "rewards/accuracies": 0.7509765625, |
| "rewards/chosen": 0.2497057020664215, |
| "rewards/margins": 0.19358570873737335, |
| "rewards/rejected": 0.05611998587846756, |
| "step": 16 |
| }, |
| { |
| "epoch": 1.3265306122448979, |
| "grad_norm": 0.8256746530532837, |
| "learning_rate": 7.369343312364994e-06, |
| "logits/chosen": -2.4009318351745605, |
| "logits/rejected": -2.4279661178588867, |
| "logps/chosen": -86.41546630859375, |
| "logps/rejected": -96.96250915527344, |
| "loss": 0.6066, |
| "rewards/accuracies": 0.75, |
| "rewards/chosen": 0.27642762660980225, |
| "rewards/margins": 0.2144363969564438, |
| "rewards/rejected": 0.06199120730161667, |
| "step": 17 |
| }, |
| { |
| "epoch": 1.4081632653061225, |
| "grad_norm": 0.8295791745185852, |
| "learning_rate": 6.965125158269619e-06, |
| "logits/chosen": -2.3977432250976562, |
| "logits/rejected": -2.4250879287719727, |
| "logps/chosen": -84.85093688964844, |
| "logps/rejected": -95.09344482421875, |
| "loss": 0.6024, |
| "rewards/accuracies": 0.76025390625, |
| "rewards/chosen": 0.27905529737472534, |
| "rewards/margins": 0.2258065789937973, |
| "rewards/rejected": 0.05324871465563774, |
| "step": 18 |
| }, |
| { |
| "epoch": 1.489795918367347, |
| "grad_norm": 0.7997666001319885, |
| "learning_rate": 6.545084971874738e-06, |
| "logits/chosen": -2.403979778289795, |
| "logits/rejected": -2.417026996612549, |
| "logps/chosen": -87.35746002197266, |
| "logps/rejected": -98.11820220947266, |
| "loss": 0.5982, |
| "rewards/accuracies": 0.74658203125, |
| "rewards/chosen": 0.2821013629436493, |
| "rewards/margins": 0.24028602242469788, |
| "rewards/rejected": 0.04181533679366112, |
| "step": 19 |
| }, |
| { |
| "epoch": 1.5714285714285714, |
| "grad_norm": 0.8002594113349915, |
| "learning_rate": 6.112604669781572e-06, |
| "logits/chosen": -2.4007248878479004, |
| "logits/rejected": -2.4198200702667236, |
| "logps/chosen": -84.9500503540039, |
| "logps/rejected": -97.81690979003906, |
| "loss": 0.595, |
| "rewards/accuracies": 0.76416015625, |
| "rewards/chosen": 0.27880722284317017, |
| "rewards/margins": 0.24656042456626892, |
| "rewards/rejected": 0.03224681317806244, |
| "step": 20 |
| } |
| ], |
| "logging_steps": 1, |
| "max_steps": 39, |
| "num_input_tokens_seen": 0, |
| "num_train_epochs": 3, |
| "save_steps": 20, |
| "stateful_callbacks": { |
| "TrainerControl": { |
| "args": { |
| "should_epoch_stop": false, |
| "should_evaluate": false, |
| "should_log": false, |
| "should_save": true, |
| "should_training_stop": false |
| }, |
| "attributes": {} |
| } |
| }, |
| "total_flos": 3.832705609518547e+18, |
| "train_batch_size": 4, |
| "trial_name": null, |
| "trial_params": null |
| } |
|
|
