{ "train_runtime_s": 17595.0, "train_runtime_h": 4.89, "train_loss": 0.19934185723245923, "total_optimizer_steps": 266, "log_history": [ { "loss": 4.096357727050782, "grad_norm": 8.247946739196777, "learning_rate": 0.00011428571428571428, "epoch": 0.03748828491096532, "step": 5 }, { "loss": 0.5162647247314454, "grad_norm": 0.7619041800498962, "learning_rate": 0.00019987764162142613, "epoch": 0.07497656982193064, "step": 10 }, { "loss": 0.12201160192489624, "grad_norm": 0.48526251316070557, "learning_rate": 0.00019850454633172631, "epoch": 0.11246485473289597, "step": 15 }, { "loss": 0.08256307244300842, "grad_norm": 0.9356642961502075, "learning_rate": 0.00019562645670201276, "epoch": 0.14995313964386128, "step": 20 }, { "loss": 0.05312276482582092, "grad_norm": 0.35064712166786194, "learning_rate": 0.00019128734540932495, "epoch": 0.18744142455482662, "step": 25 }, { "loss": 0.027615243196487428, "grad_norm": 0.40482228994369507, "learning_rate": 0.00018555350724085162, "epoch": 0.22492970946579194, "step": 30 }, { "loss": 0.04488930106163025, "grad_norm": 0.1967865228652954, "learning_rate": 0.0001785125462139855, "epoch": 0.2624179943767573, "step": 35 }, { "loss": 0.026395457983016967, "grad_norm": 0.3612152636051178, "learning_rate": 0.000170272037124899, "epoch": 0.29990627928772257, "step": 40 }, { "loss": 0.028966447710990904, "grad_norm": 0.12086012214422226, "learning_rate": 0.00016095788197507081, "epoch": 0.3373945641986879, "step": 45 }, { "loss": 0.05698441863059998, "grad_norm": 0.17247724533081055, "learning_rate": 0.00015071238638697732, "epoch": 0.37488284910965325, "step": 50 }, { "eval_loss": 0.47494766116142273, "eval_runtime": 1033.3512, "eval_samples_per_second": 0.444, "eval_steps_per_second": 0.223, "epoch": 0.37488284910965325, "step": 50 }, { "loss": 0.04874803423881531, "grad_norm": 0.4485783874988556, "learning_rate": 0.00013969208539828872, "epoch": 0.41237113402061853, "step": 55 }, { "loss": 0.03726997673511505, "grad_norm": 0.4255518913269043, "learning_rate": 0.0001280653518530093, "epoch": 0.4498594189315839, "step": 60 }, { "loss": 0.032462659478187564, "grad_norm": 0.3632126450538635, "learning_rate": 0.00011600982392957978, "epoch": 0.4873477038425492, "step": 65 }, { "loss": 0.010317540168762207, "grad_norm": 0.08012416958808899, "learning_rate": 0.00010370969110926052, "epoch": 0.5248359887535146, "step": 70 }, { "loss": 0.006553307920694351, "grad_norm": 0.025625359266996384, "learning_rate": 9.135288005092546e-05, "epoch": 0.5623242736644799, "step": 75 }, { "loss": 0.0011172231286764144, "grad_norm": 0.005225732922554016, "learning_rate": 7.91281833676665e-05, "epoch": 0.5998125585754451, "step": 80 }, { "loss": 0.03017619550228119, "grad_norm": 0.0397188626229763, "learning_rate": 6.722237517298232e-05, "epoch": 0.6373008434864105, "step": 85 }, { "loss": 0.031118598580360413, "grad_norm": 0.03184065595269203, "learning_rate": 5.5817357466461336e-05, "epoch": 0.6747891283973758, "step": 90 }, { "loss": 0.025105661153793334, "grad_norm": 0.46336865425109863, "learning_rate": 4.5087380957692784e-05, "epoch": 0.7122774133083412, "step": 95 }, { "loss": 0.01962131857872009, "grad_norm": 0.36116623878479004, "learning_rate": 3.519638278983948e-05, "epoch": 0.7497656982193065, "step": 100 }, { "eval_loss": 0.3397081196308136, "eval_runtime": 1034.6204, "eval_samples_per_second": 0.444, "eval_steps_per_second": 0.222, "epoch": 0.7497656982193065, "step": 100 }, { "loss": 0.016365388035774232, "grad_norm": 0.0068334853276610374, "learning_rate": 2.6295481838263626e-05, "epoch": 0.7872539831302718, "step": 105 }, { "loss": 0.0065630853176116945, "grad_norm": 0.04686582460999489, "learning_rate": 1.8520669852097573e-05, "epoch": 0.8247422680412371, "step": 110 }, { "loss": 0.00702376440167427, "grad_norm": 0.06303781270980835, "learning_rate": 1.1990733714481184e-05, "epoch": 0.8622305529522024, "step": 115 }, { "loss": 0.0008668770082294941, "grad_norm": 0.010310538113117218, "learning_rate": 6.805440566056553e-06, "epoch": 0.8997188378631678, "step": 120 }, { "loss": 0.005535626038908959, "grad_norm": 0.011637466959655285, "learning_rate": 3.0440135201753374e-06, "epoch": 0.9372071227741331, "step": 125 }, { "loss": 0.0059164069592952725, "grad_norm": 0.012257271446287632, "learning_rate": 7.639212584897082e-07, "epoch": 0.9746954076850984, "step": 130 }, { "eval_loss": 0.33489006757736206, "eval_runtime": 1033.9345, "eval_samples_per_second": 0.444, "eval_steps_per_second": 0.222, "epoch": 1.0, "step": 134 }, { "train_runtime": 17594.0863, "train_samples_per_second": 0.121, "train_steps_per_second": 0.008, "total_flos": 1.5104213611029504e+17, "train_loss": 0.19934185723245923, "epoch": 1.0, "step": 134 }, { "eval_loss": 0.33489006757736206, "eval_runtime": 1032.8903, "eval_samples_per_second": 0.444, "eval_steps_per_second": 0.223, "epoch": 1.0, "step": 134 } ], "final_eval": { "eval_loss": 0.33489006757736206, "eval_runtime": 1032.8903, "eval_samples_per_second": 0.444, "eval_steps_per_second": 0.223, "epoch": 1.0 } }