| { | |
| "best_metric": 0.1280793398618698, | |
| "best_model_checkpoint": "deepfake_vs_real_image_detection/checkpoint-35008", | |
| "epoch": 8.0, | |
| "eval_steps": 500, | |
| "global_step": 35008, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.11425959780621572, | |
| "grad_norm": 0.9279178977012634, | |
| "learning_rate": 9.87127410034899e-07, | |
| "loss": 0.6754, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.22851919561243145, | |
| "grad_norm": 1.445857286453247, | |
| "learning_rate": 9.72824532295898e-07, | |
| "loss": 0.5645, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.34277879341864714, | |
| "grad_norm": 2.294506788253784, | |
| "learning_rate": 9.585216545568968e-07, | |
| "loss": 0.3891, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.4570383912248629, | |
| "grad_norm": 2.4269814491271973, | |
| "learning_rate": 9.442187768178958e-07, | |
| "loss": 0.2769, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.5712979890310786, | |
| "grad_norm": 8.800315856933594, | |
| "learning_rate": 9.299158990788947e-07, | |
| "loss": 0.225, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 0.6855575868372943, | |
| "grad_norm": 2.500453233718872, | |
| "learning_rate": 9.156130213398935e-07, | |
| "loss": 0.1974, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.79981718464351, | |
| "grad_norm": 2.538663864135742, | |
| "learning_rate": 9.013101436008924e-07, | |
| "loss": 0.1816, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 0.9140767824497258, | |
| "grad_norm": 3.2626090049743652, | |
| "learning_rate": 8.870072658618913e-07, | |
| "loss": 0.1687, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.9183575124277163, | |
| "eval_loss": 0.26723140478134155, | |
| "eval_runtime": 613.6678, | |
| "eval_samples_per_second": 64.25, | |
| "eval_steps_per_second": 8.032, | |
| "step": 4376 | |
| }, | |
| { | |
| "epoch": 1.0283363802559415, | |
| "grad_norm": 7.749739170074463, | |
| "learning_rate": 8.727043881228902e-07, | |
| "loss": 0.161, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 1.1425959780621573, | |
| "grad_norm": 3.7298998832702637, | |
| "learning_rate": 8.584015103838892e-07, | |
| "loss": 0.148, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 1.2568555758683728, | |
| "grad_norm": 1.4705345630645752, | |
| "learning_rate": 8.440986326448881e-07, | |
| "loss": 0.1456, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 1.3711151736745886, | |
| "grad_norm": 3.6541574001312256, | |
| "learning_rate": 8.297957549058871e-07, | |
| "loss": 0.1427, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 1.4853747714808043, | |
| "grad_norm": 2.981785535812378, | |
| "learning_rate": 8.15492877166886e-07, | |
| "loss": 0.1376, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 1.59963436928702, | |
| "grad_norm": 5.706754684448242, | |
| "learning_rate": 8.011899994278849e-07, | |
| "loss": 0.1257, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 1.7138939670932358, | |
| "grad_norm": 4.674542427062988, | |
| "learning_rate": 7.868871216888838e-07, | |
| "loss": 0.1247, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 1.8281535648994516, | |
| "grad_norm": 10.191073417663574, | |
| "learning_rate": 7.725842439498827e-07, | |
| "loss": 0.124, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 1.9424131627056673, | |
| "grad_norm": 8.716536521911621, | |
| "learning_rate": 7.582813662108815e-07, | |
| "loss": 0.1196, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.9366440093334686, | |
| "eval_loss": 0.19802308082580566, | |
| "eval_runtime": 414.2816, | |
| "eval_samples_per_second": 95.172, | |
| "eval_steps_per_second": 11.898, | |
| "step": 8752 | |
| }, | |
| { | |
| "epoch": 2.056672760511883, | |
| "grad_norm": 0.8615439534187317, | |
| "learning_rate": 7.439784884718805e-07, | |
| "loss": 0.1161, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 2.170932358318099, | |
| "grad_norm": 4.869253158569336, | |
| "learning_rate": 7.296756107328794e-07, | |
| "loss": 0.1169, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 2.2851919561243146, | |
| "grad_norm": 4.5460333824157715, | |
| "learning_rate": 7.153727329938783e-07, | |
| "loss": 0.1095, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 2.3994515539305303, | |
| "grad_norm": 0.5965484976768494, | |
| "learning_rate": 7.010698552548773e-07, | |
| "loss": 0.1131, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 2.5137111517367456, | |
| "grad_norm": 2.817739725112915, | |
| "learning_rate": 6.867669775158762e-07, | |
| "loss": 0.1056, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 2.6279707495429614, | |
| "grad_norm": 2.231886625289917, | |
| "learning_rate": 6.724640997768751e-07, | |
| "loss": 0.1086, | |
| "step": 11500 | |
| }, | |
| { | |
| "epoch": 2.742230347349177, | |
| "grad_norm": 7.3367600440979, | |
| "learning_rate": 6.58161222037874e-07, | |
| "loss": 0.0982, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 2.856489945155393, | |
| "grad_norm": 1.7964383363723755, | |
| "learning_rate": 6.438583442988729e-07, | |
| "loss": 0.1018, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 2.9707495429616086, | |
| "grad_norm": 6.008702754974365, | |
| "learning_rate": 6.295554665598718e-07, | |
| "loss": 0.1044, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.9433904839200569, | |
| "eval_loss": 0.17182812094688416, | |
| "eval_runtime": 413.3537, | |
| "eval_samples_per_second": 95.386, | |
| "eval_steps_per_second": 11.924, | |
| "step": 13128 | |
| }, | |
| { | |
| "epoch": 3.0850091407678244, | |
| "grad_norm": 8.11001205444336, | |
| "learning_rate": 6.152525888208708e-07, | |
| "loss": 0.0962, | |
| "step": 13500 | |
| }, | |
| { | |
| "epoch": 3.19926873857404, | |
| "grad_norm": 3.732985019683838, | |
| "learning_rate": 6.009497110818696e-07, | |
| "loss": 0.0903, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 3.313528336380256, | |
| "grad_norm": 8.172978401184082, | |
| "learning_rate": 5.866468333428685e-07, | |
| "loss": 0.1025, | |
| "step": 14500 | |
| }, | |
| { | |
| "epoch": 3.4277879341864717, | |
| "grad_norm": 7.130813121795654, | |
| "learning_rate": 5.723439556038675e-07, | |
| "loss": 0.0934, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 3.5420475319926874, | |
| "grad_norm": 10.737488746643066, | |
| "learning_rate": 5.580410778648664e-07, | |
| "loss": 0.0937, | |
| "step": 15500 | |
| }, | |
| { | |
| "epoch": 3.656307129798903, | |
| "grad_norm": 5.984869480133057, | |
| "learning_rate": 5.437382001258653e-07, | |
| "loss": 0.0947, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 3.770566727605119, | |
| "grad_norm": 1.0398708581924438, | |
| "learning_rate": 5.294353223868642e-07, | |
| "loss": 0.0944, | |
| "step": 16500 | |
| }, | |
| { | |
| "epoch": 3.8848263254113347, | |
| "grad_norm": 4.8392510414123535, | |
| "learning_rate": 5.151324446478631e-07, | |
| "loss": 0.0908, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 3.9990859232175504, | |
| "grad_norm": 6.028433322906494, | |
| "learning_rate": 5.00829566908862e-07, | |
| "loss": 0.0897, | |
| "step": 17500 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.9511514659632748, | |
| "eval_loss": 0.1482020914554596, | |
| "eval_runtime": 408.5773, | |
| "eval_samples_per_second": 96.501, | |
| "eval_steps_per_second": 12.064, | |
| "step": 17504 | |
| }, | |
| { | |
| "epoch": 4.113345521023766, | |
| "grad_norm": 1.668658971786499, | |
| "learning_rate": 4.86526689169861e-07, | |
| "loss": 0.0867, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 4.227605118829982, | |
| "grad_norm": 3.3406217098236084, | |
| "learning_rate": 4.7222381143085986e-07, | |
| "loss": 0.0933, | |
| "step": 18500 | |
| }, | |
| { | |
| "epoch": 4.341864716636198, | |
| "grad_norm": 6.337904930114746, | |
| "learning_rate": 4.579209336918588e-07, | |
| "loss": 0.0876, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 4.456124314442413, | |
| "grad_norm": 4.625348091125488, | |
| "learning_rate": 4.436180559528577e-07, | |
| "loss": 0.0857, | |
| "step": 19500 | |
| }, | |
| { | |
| "epoch": 4.570383912248629, | |
| "grad_norm": 0.618135929107666, | |
| "learning_rate": 4.293151782138566e-07, | |
| "loss": 0.0875, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 4.684643510054845, | |
| "grad_norm": 17.393951416015625, | |
| "learning_rate": 4.1501230047485555e-07, | |
| "loss": 0.0866, | |
| "step": 20500 | |
| }, | |
| { | |
| "epoch": 4.798903107861061, | |
| "grad_norm": 0.8322166800498962, | |
| "learning_rate": 4.0070942273585443e-07, | |
| "loss": 0.088, | |
| "step": 21000 | |
| }, | |
| { | |
| "epoch": 4.913162705667276, | |
| "grad_norm": 0.31342649459838867, | |
| "learning_rate": 3.864065449968533e-07, | |
| "loss": 0.0884, | |
| "step": 21500 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.9540681749010855, | |
| "eval_loss": 0.13915510475635529, | |
| "eval_runtime": 410.8966, | |
| "eval_samples_per_second": 95.956, | |
| "eval_steps_per_second": 11.996, | |
| "step": 21880 | |
| }, | |
| { | |
| "epoch": 5.027422303473492, | |
| "grad_norm": 15.877787590026855, | |
| "learning_rate": 3.7210366725785225e-07, | |
| "loss": 0.0907, | |
| "step": 22000 | |
| }, | |
| { | |
| "epoch": 5.141681901279708, | |
| "grad_norm": 3.825509548187256, | |
| "learning_rate": 3.578007895188512e-07, | |
| "loss": 0.0784, | |
| "step": 22500 | |
| }, | |
| { | |
| "epoch": 5.255941499085923, | |
| "grad_norm": 8.778756141662598, | |
| "learning_rate": 3.434979117798501e-07, | |
| "loss": 0.0839, | |
| "step": 23000 | |
| }, | |
| { | |
| "epoch": 5.3702010968921385, | |
| "grad_norm": 3.6557865142822266, | |
| "learning_rate": 3.29195034040849e-07, | |
| "loss": 0.0834, | |
| "step": 23500 | |
| }, | |
| { | |
| "epoch": 5.484460694698354, | |
| "grad_norm": 4.403536319732666, | |
| "learning_rate": 3.148921563018479e-07, | |
| "loss": 0.0846, | |
| "step": 24000 | |
| }, | |
| { | |
| "epoch": 5.59872029250457, | |
| "grad_norm": 1.8743895292282104, | |
| "learning_rate": 3.005892785628468e-07, | |
| "loss": 0.0818, | |
| "step": 24500 | |
| }, | |
| { | |
| "epoch": 5.712979890310786, | |
| "grad_norm": 5.720378875732422, | |
| "learning_rate": 2.8628640082384576e-07, | |
| "loss": 0.0791, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 5.8272394881170015, | |
| "grad_norm": 9.525908470153809, | |
| "learning_rate": 2.719835230848447e-07, | |
| "loss": 0.0887, | |
| "step": 25500 | |
| }, | |
| { | |
| "epoch": 5.941499085923217, | |
| "grad_norm": 13.019225120544434, | |
| "learning_rate": 2.576806453458435e-07, | |
| "loss": 0.0785, | |
| "step": 26000 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_accuracy": 0.9561479151871767, | |
| "eval_loss": 0.13129685819149017, | |
| "eval_runtime": 380.134, | |
| "eval_samples_per_second": 103.721, | |
| "eval_steps_per_second": 12.966, | |
| "step": 26256 | |
| }, | |
| { | |
| "epoch": 6.055758683729433, | |
| "grad_norm": 8.515515327453613, | |
| "learning_rate": 2.4337776760684246e-07, | |
| "loss": 0.0781, | |
| "step": 26500 | |
| }, | |
| { | |
| "epoch": 6.170018281535649, | |
| "grad_norm": 2.822601318359375, | |
| "learning_rate": 2.290748898678414e-07, | |
| "loss": 0.0792, | |
| "step": 27000 | |
| }, | |
| { | |
| "epoch": 6.2842778793418645, | |
| "grad_norm": 0.4864025115966797, | |
| "learning_rate": 2.147720121288403e-07, | |
| "loss": 0.0789, | |
| "step": 27500 | |
| }, | |
| { | |
| "epoch": 6.39853747714808, | |
| "grad_norm": 1.266471266746521, | |
| "learning_rate": 2.0046913438983924e-07, | |
| "loss": 0.085, | |
| "step": 28000 | |
| }, | |
| { | |
| "epoch": 6.512797074954296, | |
| "grad_norm": 0.17314019799232483, | |
| "learning_rate": 1.8616625665083812e-07, | |
| "loss": 0.0771, | |
| "step": 28500 | |
| }, | |
| { | |
| "epoch": 6.627056672760512, | |
| "grad_norm": 0.11477702111005783, | |
| "learning_rate": 1.7186337891183706e-07, | |
| "loss": 0.083, | |
| "step": 29000 | |
| }, | |
| { | |
| "epoch": 6.7413162705667276, | |
| "grad_norm": 3.228081703186035, | |
| "learning_rate": 1.5756050117283597e-07, | |
| "loss": 0.0815, | |
| "step": 29500 | |
| }, | |
| { | |
| "epoch": 6.855575868372943, | |
| "grad_norm": 8.033062934875488, | |
| "learning_rate": 1.4325762343383488e-07, | |
| "loss": 0.0769, | |
| "step": 30000 | |
| }, | |
| { | |
| "epoch": 6.969835466179159, | |
| "grad_norm": 3.3470427989959717, | |
| "learning_rate": 1.289547456948338e-07, | |
| "loss": 0.0827, | |
| "step": 30500 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_accuracy": 0.9570609718981434, | |
| "eval_loss": 0.13065218925476074, | |
| "eval_runtime": 375.1187, | |
| "eval_samples_per_second": 105.108, | |
| "eval_steps_per_second": 13.14, | |
| "step": 30632 | |
| }, | |
| { | |
| "epoch": 7.084095063985375, | |
| "grad_norm": 3.7121076583862305, | |
| "learning_rate": 1.1465186795583271e-07, | |
| "loss": 0.0788, | |
| "step": 31000 | |
| }, | |
| { | |
| "epoch": 7.198354661791591, | |
| "grad_norm": 1.4164518117904663, | |
| "learning_rate": 1.0034899021683162e-07, | |
| "loss": 0.0844, | |
| "step": 31500 | |
| }, | |
| { | |
| "epoch": 7.312614259597806, | |
| "grad_norm": 0.26970207691192627, | |
| "learning_rate": 8.604611247783053e-08, | |
| "loss": 0.0795, | |
| "step": 32000 | |
| }, | |
| { | |
| "epoch": 7.426873857404022, | |
| "grad_norm": 2.8127174377441406, | |
| "learning_rate": 7.174323473882945e-08, | |
| "loss": 0.0846, | |
| "step": 32500 | |
| }, | |
| { | |
| "epoch": 7.541133455210238, | |
| "grad_norm": 11.107344627380371, | |
| "learning_rate": 5.744035699982836e-08, | |
| "loss": 0.0797, | |
| "step": 33000 | |
| }, | |
| { | |
| "epoch": 7.655393053016454, | |
| "grad_norm": 10.199967384338379, | |
| "learning_rate": 4.3137479260827274e-08, | |
| "loss": 0.0776, | |
| "step": 33500 | |
| }, | |
| { | |
| "epoch": 7.769652650822669, | |
| "grad_norm": 0.1155412346124649, | |
| "learning_rate": 2.883460152182619e-08, | |
| "loss": 0.0758, | |
| "step": 34000 | |
| }, | |
| { | |
| "epoch": 7.883912248628885, | |
| "grad_norm": 0.0839085578918457, | |
| "learning_rate": 1.4531723782825104e-08, | |
| "loss": 0.0762, | |
| "step": 34500 | |
| }, | |
| { | |
| "epoch": 7.998171846435101, | |
| "grad_norm": 7.536131858825684, | |
| "learning_rate": 2.2884604382401738e-10, | |
| "loss": 0.0767, | |
| "step": 35000 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_accuracy": 0.9575682256264584, | |
| "eval_loss": 0.1280793398618698, | |
| "eval_runtime": 395.9176, | |
| "eval_samples_per_second": 99.586, | |
| "eval_steps_per_second": 12.45, | |
| "step": 35008 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 35008, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 8, | |
| "save_steps": 500, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 8.679226824017687e+19, | |
| "train_batch_size": 32, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |