| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 20.0, | |
| "eval_steps": 500, | |
| "global_step": 10580, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.19, | |
| "learning_rate": 0.0001, | |
| "loss": 0.3787, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.38, | |
| "learning_rate": 0.0001, | |
| "loss": 0.315, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.57, | |
| "learning_rate": 0.0001, | |
| "loss": 0.3191, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.76, | |
| "learning_rate": 0.0001, | |
| "loss": 0.3082, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.95, | |
| "learning_rate": 0.0001, | |
| "loss": 0.2952, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.7924862745098039, | |
| "eval_loss": 0.37208321690559387, | |
| "eval_runtime": 7.4879, | |
| "eval_samples_per_second": 66.774, | |
| "eval_steps_per_second": 8.414, | |
| "step": 529 | |
| }, | |
| { | |
| "epoch": 1.13, | |
| "learning_rate": 0.0001, | |
| "loss": 0.2733, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.32, | |
| "learning_rate": 0.0001, | |
| "loss": 0.2549, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 1.51, | |
| "learning_rate": 0.0001, | |
| "loss": 0.2483, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 1.7, | |
| "learning_rate": 0.0001, | |
| "loss": 0.2473, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 1.89, | |
| "learning_rate": 0.0001, | |
| "loss": 0.2481, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.8003137254901961, | |
| "eval_loss": 0.3231143653392792, | |
| "eval_runtime": 7.0792, | |
| "eval_samples_per_second": 70.629, | |
| "eval_steps_per_second": 8.899, | |
| "step": 1058 | |
| }, | |
| { | |
| "epoch": 2.08, | |
| "learning_rate": 0.0001, | |
| "loss": 0.2244, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 2.27, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1981, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 2.46, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1946, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 2.65, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1978, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 2.84, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1935, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.8044156862745098, | |
| "eval_loss": 0.29632872343063354, | |
| "eval_runtime": 7.6622, | |
| "eval_samples_per_second": 65.255, | |
| "eval_steps_per_second": 8.222, | |
| "step": 1587 | |
| }, | |
| { | |
| "epoch": 3.02, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1865, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 3.21, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1586, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 3.4, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1598, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 3.59, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1596, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 3.78, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1603, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 3.97, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1593, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.806235294117647, | |
| "eval_loss": 0.28722652792930603, | |
| "eval_runtime": 6.677, | |
| "eval_samples_per_second": 74.884, | |
| "eval_steps_per_second": 9.435, | |
| "step": 2116 | |
| }, | |
| { | |
| "epoch": 4.16, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1369, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 4.35, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1371, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 4.54, | |
| "learning_rate": 0.0001, | |
| "loss": 0.136, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 4.73, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1379, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 4.91, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1405, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.8067450980392157, | |
| "eval_loss": 0.29079169034957886, | |
| "eval_runtime": 7.7882, | |
| "eval_samples_per_second": 64.2, | |
| "eval_steps_per_second": 8.089, | |
| "step": 2645 | |
| }, | |
| { | |
| "epoch": 5.1, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1278, | |
| "step": 2700 | |
| }, | |
| { | |
| "epoch": 5.29, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1208, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 5.48, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1205, | |
| "step": 2900 | |
| }, | |
| { | |
| "epoch": 5.67, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1239, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 5.86, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1235, | |
| "step": 3100 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_accuracy": 0.8071529411764706, | |
| "eval_loss": 0.29287615418434143, | |
| "eval_runtime": 7.1498, | |
| "eval_samples_per_second": 69.932, | |
| "eval_steps_per_second": 8.811, | |
| "step": 3174 | |
| }, | |
| { | |
| "epoch": 6.05, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1205, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 6.24, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1096, | |
| "step": 3300 | |
| }, | |
| { | |
| "epoch": 6.43, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1103, | |
| "step": 3400 | |
| }, | |
| { | |
| "epoch": 6.62, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1127, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 6.81, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1136, | |
| "step": 3600 | |
| }, | |
| { | |
| "epoch": 6.99, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1154, | |
| "step": 3700 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_accuracy": 0.8070666666666667, | |
| "eval_loss": 0.31088685989379883, | |
| "eval_runtime": 7.7231, | |
| "eval_samples_per_second": 64.741, | |
| "eval_steps_per_second": 8.157, | |
| "step": 3703 | |
| }, | |
| { | |
| "epoch": 7.18, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1004, | |
| "step": 3800 | |
| }, | |
| { | |
| "epoch": 7.37, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1024, | |
| "step": 3900 | |
| }, | |
| { | |
| "epoch": 7.56, | |
| "learning_rate": 0.0001, | |
| "loss": 0.104, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 7.75, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1046, | |
| "step": 4100 | |
| }, | |
| { | |
| "epoch": 7.94, | |
| "learning_rate": 0.0001, | |
| "loss": 0.106, | |
| "step": 4200 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_accuracy": 0.8069019607843138, | |
| "eval_loss": 0.3178957402706146, | |
| "eval_runtime": 7.5236, | |
| "eval_samples_per_second": 66.458, | |
| "eval_steps_per_second": 8.374, | |
| "step": 4232 | |
| }, | |
| { | |
| "epoch": 8.13, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0979, | |
| "step": 4300 | |
| }, | |
| { | |
| "epoch": 8.32, | |
| "learning_rate": 0.0001, | |
| "loss": 0.095, | |
| "step": 4400 | |
| }, | |
| { | |
| "epoch": 8.51, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0974, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 8.7, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0979, | |
| "step": 4600 | |
| }, | |
| { | |
| "epoch": 8.88, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0997, | |
| "step": 4700 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_accuracy": 0.8070901960784314, | |
| "eval_loss": 0.33391040563583374, | |
| "eval_runtime": 7.6758, | |
| "eval_samples_per_second": 65.14, | |
| "eval_steps_per_second": 8.208, | |
| "step": 4761 | |
| }, | |
| { | |
| "epoch": 9.07, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0968, | |
| "step": 4800 | |
| }, | |
| { | |
| "epoch": 9.26, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0901, | |
| "step": 4900 | |
| }, | |
| { | |
| "epoch": 9.45, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0917, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 9.64, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0936, | |
| "step": 5100 | |
| }, | |
| { | |
| "epoch": 9.83, | |
| "learning_rate": 0.0001, | |
| "loss": 0.095, | |
| "step": 5200 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_accuracy": 0.8067294117647059, | |
| "eval_loss": 0.3424333930015564, | |
| "eval_runtime": 6.6593, | |
| "eval_samples_per_second": 75.083, | |
| "eval_steps_per_second": 9.46, | |
| "step": 5290 | |
| }, | |
| { | |
| "epoch": 10.02, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0952, | |
| "step": 5300 | |
| }, | |
| { | |
| "epoch": 10.21, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0867, | |
| "step": 5400 | |
| }, | |
| { | |
| "epoch": 10.4, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0902, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 10.59, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0915, | |
| "step": 5600 | |
| }, | |
| { | |
| "epoch": 10.78, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0919, | |
| "step": 5700 | |
| }, | |
| { | |
| "epoch": 10.96, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0922, | |
| "step": 5800 | |
| }, | |
| { | |
| "epoch": 11.0, | |
| "eval_accuracy": 0.8065882352941176, | |
| "eval_loss": 0.35157716274261475, | |
| "eval_runtime": 7.0864, | |
| "eval_samples_per_second": 70.558, | |
| "eval_steps_per_second": 8.89, | |
| "step": 5819 | |
| }, | |
| { | |
| "epoch": 11.15, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0855, | |
| "step": 5900 | |
| }, | |
| { | |
| "epoch": 11.34, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0863, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 11.53, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0865, | |
| "step": 6100 | |
| }, | |
| { | |
| "epoch": 11.72, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0876, | |
| "step": 6200 | |
| }, | |
| { | |
| "epoch": 11.91, | |
| "learning_rate": 0.0001, | |
| "loss": 0.089, | |
| "step": 6300 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "eval_accuracy": 0.806321568627451, | |
| "eval_loss": 0.37196654081344604, | |
| "eval_runtime": 7.1828, | |
| "eval_samples_per_second": 69.611, | |
| "eval_steps_per_second": 8.771, | |
| "step": 6348 | |
| }, | |
| { | |
| "epoch": 12.1, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0857, | |
| "step": 6400 | |
| }, | |
| { | |
| "epoch": 12.29, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0833, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 12.48, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0851, | |
| "step": 6600 | |
| }, | |
| { | |
| "epoch": 12.67, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0857, | |
| "step": 6700 | |
| }, | |
| { | |
| "epoch": 12.85, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0862, | |
| "step": 6800 | |
| }, | |
| { | |
| "epoch": 13.0, | |
| "eval_accuracy": 0.8065254901960784, | |
| "eval_loss": 0.37400877475738525, | |
| "eval_runtime": 7.7823, | |
| "eval_samples_per_second": 64.248, | |
| "eval_steps_per_second": 8.095, | |
| "step": 6877 | |
| }, | |
| { | |
| "epoch": 13.04, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0864, | |
| "step": 6900 | |
| }, | |
| { | |
| "epoch": 13.23, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0819, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 13.42, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0835, | |
| "step": 7100 | |
| }, | |
| { | |
| "epoch": 13.61, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0837, | |
| "step": 7200 | |
| }, | |
| { | |
| "epoch": 13.8, | |
| "learning_rate": 0.0001, | |
| "loss": 0.085, | |
| "step": 7300 | |
| }, | |
| { | |
| "epoch": 13.99, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0862, | |
| "step": 7400 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "eval_accuracy": 0.8070117647058823, | |
| "eval_loss": 0.3680531084537506, | |
| "eval_runtime": 7.8108, | |
| "eval_samples_per_second": 64.014, | |
| "eval_steps_per_second": 8.066, | |
| "step": 7406 | |
| }, | |
| { | |
| "epoch": 14.18, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0796, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 14.37, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0816, | |
| "step": 7600 | |
| }, | |
| { | |
| "epoch": 14.56, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0824, | |
| "step": 7700 | |
| }, | |
| { | |
| "epoch": 14.74, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0838, | |
| "step": 7800 | |
| }, | |
| { | |
| "epoch": 14.93, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0852, | |
| "step": 7900 | |
| }, | |
| { | |
| "epoch": 15.0, | |
| "eval_accuracy": 0.8066588235294118, | |
| "eval_loss": 0.37707823514938354, | |
| "eval_runtime": 7.5371, | |
| "eval_samples_per_second": 66.339, | |
| "eval_steps_per_second": 8.359, | |
| "step": 7935 | |
| }, | |
| { | |
| "epoch": 15.12, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0817, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 15.31, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0804, | |
| "step": 8100 | |
| }, | |
| { | |
| "epoch": 15.5, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0823, | |
| "step": 8200 | |
| }, | |
| { | |
| "epoch": 15.69, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0828, | |
| "step": 8300 | |
| }, | |
| { | |
| "epoch": 15.88, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0849, | |
| "step": 8400 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "eval_accuracy": 0.8066274509803921, | |
| "eval_loss": 0.3813852071762085, | |
| "eval_runtime": 7.4973, | |
| "eval_samples_per_second": 66.691, | |
| "eval_steps_per_second": 8.403, | |
| "step": 8464 | |
| }, | |
| { | |
| "epoch": 16.07, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0813, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 16.26, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0794, | |
| "step": 8600 | |
| }, | |
| { | |
| "epoch": 16.45, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0811, | |
| "step": 8700 | |
| }, | |
| { | |
| "epoch": 16.64, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0825, | |
| "step": 8800 | |
| }, | |
| { | |
| "epoch": 16.82, | |
| "learning_rate": 0.0001, | |
| "loss": 0.083, | |
| "step": 8900 | |
| }, | |
| { | |
| "epoch": 17.0, | |
| "eval_accuracy": 0.806478431372549, | |
| "eval_loss": 0.37993746995925903, | |
| "eval_runtime": 6.6523, | |
| "eval_samples_per_second": 75.162, | |
| "eval_steps_per_second": 9.47, | |
| "step": 8993 | |
| }, | |
| { | |
| "epoch": 17.01, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0842, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 17.2, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0797, | |
| "step": 9100 | |
| }, | |
| { | |
| "epoch": 17.39, | |
| "learning_rate": 0.0001, | |
| "loss": 0.082, | |
| "step": 9200 | |
| }, | |
| { | |
| "epoch": 17.58, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0821, | |
| "step": 9300 | |
| }, | |
| { | |
| "epoch": 17.77, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0827, | |
| "step": 9400 | |
| }, | |
| { | |
| "epoch": 17.96, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0838, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 18.0, | |
| "eval_accuracy": 0.8068156862745098, | |
| "eval_loss": 0.3886854946613312, | |
| "eval_runtime": 7.5079, | |
| "eval_samples_per_second": 66.596, | |
| "eval_steps_per_second": 8.391, | |
| "step": 9522 | |
| }, | |
| { | |
| "epoch": 18.15, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0795, | |
| "step": 9600 | |
| }, | |
| { | |
| "epoch": 18.34, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0794, | |
| "step": 9700 | |
| }, | |
| { | |
| "epoch": 18.53, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0828, | |
| "step": 9800 | |
| }, | |
| { | |
| "epoch": 18.71, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0824, | |
| "step": 9900 | |
| }, | |
| { | |
| "epoch": 18.9, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0834, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 19.0, | |
| "eval_accuracy": 0.8067294117647059, | |
| "eval_loss": 0.39091432094573975, | |
| "eval_runtime": 7.0941, | |
| "eval_samples_per_second": 70.481, | |
| "eval_steps_per_second": 8.881, | |
| "step": 10051 | |
| }, | |
| { | |
| "epoch": 19.09, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0809, | |
| "step": 10100 | |
| }, | |
| { | |
| "epoch": 19.28, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0797, | |
| "step": 10200 | |
| }, | |
| { | |
| "epoch": 19.47, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0797, | |
| "step": 10300 | |
| }, | |
| { | |
| "epoch": 19.66, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0814, | |
| "step": 10400 | |
| }, | |
| { | |
| "epoch": 19.85, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0818, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 20.0, | |
| "eval_accuracy": 0.8067607843137254, | |
| "eval_loss": 0.39837226271629333, | |
| "eval_runtime": 7.7346, | |
| "eval_samples_per_second": 64.645, | |
| "eval_steps_per_second": 8.145, | |
| "step": 10580 | |
| }, | |
| { | |
| "epoch": 20.0, | |
| "step": 10580, | |
| "total_flos": 2.2427623579965194e+18, | |
| "train_loss": 0.12152552902135146, | |
| "train_runtime": 7955.4081, | |
| "train_samples_per_second": 42.55, | |
| "train_steps_per_second": 1.33 | |
| } | |
| ], | |
| "logging_steps": 100, | |
| "max_steps": 10580, | |
| "num_train_epochs": 20, | |
| "save_steps": 500, | |
| "total_flos": 2.2427623579965194e+18, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |