| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 2.906413485758574, | |
| "global_step": 30000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.05, | |
| "learning_rate": 4.9192662920622624e-05, | |
| "loss": 3.6238, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "learning_rate": 4.838532584124524e-05, | |
| "loss": 2.7692, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "learning_rate": 4.757798876186786e-05, | |
| "loss": 2.5108, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.19, | |
| "learning_rate": 4.677065168249048e-05, | |
| "loss": 2.3409, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.24, | |
| "learning_rate": 4.596331460311309e-05, | |
| "loss": 2.2195, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "learning_rate": 4.515597752373571e-05, | |
| "loss": 2.1371, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.34, | |
| "learning_rate": 4.434864044435833e-05, | |
| "loss": 2.057, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 0.39, | |
| "learning_rate": 4.354130336498095e-05, | |
| "loss": 1.9966, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.44, | |
| "learning_rate": 4.273396628560357e-05, | |
| "loss": 1.9476, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "learning_rate": 4.192662920622619e-05, | |
| "loss": 1.8893, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.53, | |
| "learning_rate": 4.1119292126848804e-05, | |
| "loss": 1.8574, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 0.58, | |
| "learning_rate": 4.031195504747142e-05, | |
| "loss": 1.8144, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 0.63, | |
| "learning_rate": 3.950461796809404e-05, | |
| "loss": 1.7866, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 0.68, | |
| "learning_rate": 3.8697280888716657e-05, | |
| "loss": 1.7497, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "learning_rate": 3.788994380933928e-05, | |
| "loss": 1.7287, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 0.78, | |
| "learning_rate": 3.7082606729961894e-05, | |
| "loss": 1.6999, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 0.82, | |
| "learning_rate": 3.6275269650584516e-05, | |
| "loss": 1.6749, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 0.87, | |
| "learning_rate": 3.546793257120713e-05, | |
| "loss": 1.6554, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 0.92, | |
| "learning_rate": 3.466059549182975e-05, | |
| "loss": 1.6307, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 0.97, | |
| "learning_rate": 3.385325841245237e-05, | |
| "loss": 1.6148, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 1.02, | |
| "learning_rate": 3.3045921333074984e-05, | |
| "loss": 1.5938, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 1.07, | |
| "learning_rate": 3.2238584253697606e-05, | |
| "loss": 1.5652, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 1.11, | |
| "learning_rate": 3.143124717432022e-05, | |
| "loss": 1.5597, | |
| "step": 11500 | |
| }, | |
| { | |
| "epoch": 1.16, | |
| "learning_rate": 3.062391009494284e-05, | |
| "loss": 1.5457, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 1.21, | |
| "learning_rate": 2.9816573015565458e-05, | |
| "loss": 1.5278, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 1.26, | |
| "learning_rate": 2.9009235936188077e-05, | |
| "loss": 1.5221, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 1.31, | |
| "learning_rate": 2.82018988568107e-05, | |
| "loss": 1.4991, | |
| "step": 13500 | |
| }, | |
| { | |
| "epoch": 1.36, | |
| "learning_rate": 2.7394561777433314e-05, | |
| "loss": 1.4876, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 1.4, | |
| "learning_rate": 2.6587224698055936e-05, | |
| "loss": 1.4747, | |
| "step": 14500 | |
| }, | |
| { | |
| "epoch": 1.45, | |
| "learning_rate": 2.577988761867855e-05, | |
| "loss": 1.4647, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 1.5, | |
| "learning_rate": 2.497255053930117e-05, | |
| "loss": 1.4566, | |
| "step": 15500 | |
| }, | |
| { | |
| "epoch": 1.55, | |
| "learning_rate": 2.416521345992379e-05, | |
| "loss": 1.4414, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 1.6, | |
| "learning_rate": 2.3357876380546407e-05, | |
| "loss": 1.4315, | |
| "step": 16500 | |
| }, | |
| { | |
| "epoch": 1.65, | |
| "learning_rate": 2.2550539301169023e-05, | |
| "loss": 1.4296, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 1.7, | |
| "learning_rate": 2.1743202221791645e-05, | |
| "loss": 1.4214, | |
| "step": 17500 | |
| }, | |
| { | |
| "epoch": 1.74, | |
| "learning_rate": 2.0935865142414263e-05, | |
| "loss": 1.4073, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 1.79, | |
| "learning_rate": 2.0128528063036882e-05, | |
| "loss": 1.4001, | |
| "step": 18500 | |
| }, | |
| { | |
| "epoch": 1.84, | |
| "learning_rate": 1.9321190983659497e-05, | |
| "loss": 1.4014, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 1.89, | |
| "learning_rate": 1.8513853904282116e-05, | |
| "loss": 1.3846, | |
| "step": 19500 | |
| }, | |
| { | |
| "epoch": 1.94, | |
| "learning_rate": 1.7706516824904734e-05, | |
| "loss": 1.3777, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 1.99, | |
| "learning_rate": 1.6899179745527353e-05, | |
| "loss": 1.3754, | |
| "step": 20500 | |
| }, | |
| { | |
| "epoch": 2.03, | |
| "learning_rate": 1.609184266614997e-05, | |
| "loss": 1.3593, | |
| "step": 21000 | |
| }, | |
| { | |
| "epoch": 2.08, | |
| "learning_rate": 1.528450558677259e-05, | |
| "loss": 1.356, | |
| "step": 21500 | |
| }, | |
| { | |
| "epoch": 2.13, | |
| "learning_rate": 1.4477168507395209e-05, | |
| "loss": 1.3527, | |
| "step": 22000 | |
| }, | |
| { | |
| "epoch": 2.18, | |
| "learning_rate": 1.3669831428017826e-05, | |
| "loss": 1.3366, | |
| "step": 22500 | |
| }, | |
| { | |
| "epoch": 2.23, | |
| "learning_rate": 1.2862494348640444e-05, | |
| "loss": 1.3392, | |
| "step": 23000 | |
| }, | |
| { | |
| "epoch": 2.28, | |
| "learning_rate": 1.2055157269263063e-05, | |
| "loss": 1.3282, | |
| "step": 23500 | |
| }, | |
| { | |
| "epoch": 2.33, | |
| "learning_rate": 1.1247820189885682e-05, | |
| "loss": 1.3309, | |
| "step": 24000 | |
| }, | |
| { | |
| "epoch": 2.37, | |
| "learning_rate": 1.0440483110508299e-05, | |
| "loss": 1.3235, | |
| "step": 24500 | |
| }, | |
| { | |
| "epoch": 2.42, | |
| "learning_rate": 9.633146031130919e-06, | |
| "loss": 1.3232, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 2.47, | |
| "learning_rate": 8.825808951753536e-06, | |
| "loss": 1.3155, | |
| "step": 25500 | |
| }, | |
| { | |
| "epoch": 2.52, | |
| "learning_rate": 8.018471872376155e-06, | |
| "loss": 1.309, | |
| "step": 26000 | |
| }, | |
| { | |
| "epoch": 2.57, | |
| "learning_rate": 7.211134792998773e-06, | |
| "loss": 1.3101, | |
| "step": 26500 | |
| }, | |
| { | |
| "epoch": 2.62, | |
| "learning_rate": 6.403797713621391e-06, | |
| "loss": 1.304, | |
| "step": 27000 | |
| }, | |
| { | |
| "epoch": 2.66, | |
| "learning_rate": 5.59646063424401e-06, | |
| "loss": 1.3035, | |
| "step": 27500 | |
| }, | |
| { | |
| "epoch": 2.71, | |
| "learning_rate": 4.7891235548666275e-06, | |
| "loss": 1.2952, | |
| "step": 28000 | |
| }, | |
| { | |
| "epoch": 2.76, | |
| "learning_rate": 3.981786475489246e-06, | |
| "loss": 1.2988, | |
| "step": 28500 | |
| }, | |
| { | |
| "epoch": 2.81, | |
| "learning_rate": 3.174449396111865e-06, | |
| "loss": 1.2876, | |
| "step": 29000 | |
| }, | |
| { | |
| "epoch": 2.86, | |
| "learning_rate": 2.3671123167344834e-06, | |
| "loss": 1.2922, | |
| "step": 29500 | |
| }, | |
| { | |
| "epoch": 2.91, | |
| "learning_rate": 1.5597752373571014e-06, | |
| "loss": 1.2881, | |
| "step": 30000 | |
| } | |
| ], | |
| "max_steps": 30966, | |
| "num_train_epochs": 3, | |
| "total_flos": 1.5832637769977856e+17, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |