| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 1.0, | |
| "eval_steps": 35, | |
| "global_step": 69, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.01, | |
| "grad_norm": 43.0, | |
| "learning_rate": 2.0000000000000002e-07, | |
| "loss": 0.7175, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.01, | |
| "eval_loss": 0.7699161767959595, | |
| "eval_runtime": 19.565, | |
| "eval_samples_per_second": 200.818, | |
| "eval_steps_per_second": 200.818, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.03, | |
| "grad_norm": 42.5, | |
| "learning_rate": 4.0000000000000003e-07, | |
| "loss": 0.7271, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.04, | |
| "grad_norm": 42.5, | |
| "learning_rate": 6.000000000000001e-07, | |
| "loss": 0.7161, | |
| "step": 3 | |
| }, | |
| { | |
| "epoch": 0.06, | |
| "grad_norm": 41.25, | |
| "learning_rate": 8.000000000000001e-07, | |
| "loss": 0.7088, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.07, | |
| "grad_norm": 42.5, | |
| "learning_rate": 1.0000000000000002e-06, | |
| "loss": 0.7022, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.09, | |
| "grad_norm": 41.0, | |
| "learning_rate": 1.2000000000000002e-06, | |
| "loss": 0.6846, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "grad_norm": 38.25, | |
| "learning_rate": 1.4000000000000001e-06, | |
| "loss": 0.629, | |
| "step": 7 | |
| }, | |
| { | |
| "epoch": 0.12, | |
| "grad_norm": 32.5, | |
| "learning_rate": 1.6000000000000001e-06, | |
| "loss": 0.5753, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "grad_norm": 28.125, | |
| "learning_rate": 1.8000000000000001e-06, | |
| "loss": 0.5368, | |
| "step": 9 | |
| }, | |
| { | |
| "epoch": 0.14, | |
| "grad_norm": 22.125, | |
| "learning_rate": 2.0000000000000003e-06, | |
| "loss": 0.4946, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "grad_norm": 19.125, | |
| "learning_rate": 2.2e-06, | |
| "loss": 0.4351, | |
| "step": 11 | |
| }, | |
| { | |
| "epoch": 0.17, | |
| "grad_norm": 15.3125, | |
| "learning_rate": 2.4000000000000003e-06, | |
| "loss": 0.3873, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 0.19, | |
| "grad_norm": 12.25, | |
| "learning_rate": 2.6e-06, | |
| "loss": 0.3382, | |
| "step": 13 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "grad_norm": 10.3125, | |
| "learning_rate": 2.8000000000000003e-06, | |
| "loss": 0.2777, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 0.22, | |
| "grad_norm": 8.4375, | |
| "learning_rate": 3e-06, | |
| "loss": 0.241, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.23, | |
| "grad_norm": 7.46875, | |
| "learning_rate": 3.2000000000000003e-06, | |
| "loss": 0.2278, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "grad_norm": 5.65625, | |
| "learning_rate": 3.4000000000000005e-06, | |
| "loss": 0.1924, | |
| "step": 17 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "grad_norm": 4.65625, | |
| "learning_rate": 3.6000000000000003e-06, | |
| "loss": 0.1635, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 0.28, | |
| "grad_norm": 3.546875, | |
| "learning_rate": 3.8000000000000005e-06, | |
| "loss": 0.1549, | |
| "step": 19 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "grad_norm": 2.5, | |
| "learning_rate": 4.000000000000001e-06, | |
| "loss": 0.1374, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.3, | |
| "grad_norm": 2.875, | |
| "learning_rate": 4.2000000000000004e-06, | |
| "loss": 0.12, | |
| "step": 21 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "grad_norm": 2.59375, | |
| "learning_rate": 4.4e-06, | |
| "loss": 0.118, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "grad_norm": 2.453125, | |
| "learning_rate": 4.600000000000001e-06, | |
| "loss": 0.1095, | |
| "step": 23 | |
| }, | |
| { | |
| "epoch": 0.35, | |
| "grad_norm": 2.171875, | |
| "learning_rate": 4.800000000000001e-06, | |
| "loss": 0.0964, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 0.36, | |
| "grad_norm": 1.7578125, | |
| "learning_rate": 5e-06, | |
| "loss": 0.0887, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.38, | |
| "grad_norm": 1.703125, | |
| "learning_rate": 5.2e-06, | |
| "loss": 0.077, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 0.39, | |
| "grad_norm": 1.625, | |
| "learning_rate": 5.400000000000001e-06, | |
| "loss": 0.074, | |
| "step": 27 | |
| }, | |
| { | |
| "epoch": 0.41, | |
| "grad_norm": 1.4609375, | |
| "learning_rate": 5.600000000000001e-06, | |
| "loss": 0.0659, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "grad_norm": 1.140625, | |
| "learning_rate": 5.8e-06, | |
| "loss": 0.0586, | |
| "step": 29 | |
| }, | |
| { | |
| "epoch": 0.43, | |
| "grad_norm": 2.265625, | |
| "learning_rate": 6e-06, | |
| "loss": 0.0518, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.45, | |
| "grad_norm": 1.609375, | |
| "learning_rate": 6.200000000000001e-06, | |
| "loss": 0.0566, | |
| "step": 31 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "grad_norm": 1.453125, | |
| "learning_rate": 6.4000000000000006e-06, | |
| "loss": 0.0475, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "grad_norm": 1.296875, | |
| "learning_rate": 6.600000000000001e-06, | |
| "loss": 0.0495, | |
| "step": 33 | |
| }, | |
| { | |
| "epoch": 0.49, | |
| "grad_norm": 1.234375, | |
| "learning_rate": 6.800000000000001e-06, | |
| "loss": 0.0493, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 0.51, | |
| "grad_norm": 1.03125, | |
| "learning_rate": 7e-06, | |
| "loss": 0.055, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.51, | |
| "eval_loss": 0.03942573070526123, | |
| "eval_runtime": 19.6691, | |
| "eval_samples_per_second": 199.755, | |
| "eval_steps_per_second": 199.755, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "grad_norm": 0.94921875, | |
| "learning_rate": 7.2000000000000005e-06, | |
| "loss": 0.0457, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 0.54, | |
| "grad_norm": 1.09375, | |
| "learning_rate": 7.4e-06, | |
| "loss": 0.0452, | |
| "step": 37 | |
| }, | |
| { | |
| "epoch": 0.55, | |
| "grad_norm": 0.83203125, | |
| "learning_rate": 7.600000000000001e-06, | |
| "loss": 0.0373, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 0.57, | |
| "grad_norm": 0.9296875, | |
| "learning_rate": 7.800000000000002e-06, | |
| "loss": 0.0379, | |
| "step": 39 | |
| }, | |
| { | |
| "epoch": 0.58, | |
| "grad_norm": 1.0625, | |
| "learning_rate": 8.000000000000001e-06, | |
| "loss": 0.0403, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "grad_norm": 0.91015625, | |
| "learning_rate": 8.2e-06, | |
| "loss": 0.0445, | |
| "step": 41 | |
| }, | |
| { | |
| "epoch": 0.61, | |
| "grad_norm": 0.96484375, | |
| "learning_rate": 8.400000000000001e-06, | |
| "loss": 0.0377, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 0.62, | |
| "grad_norm": 0.73046875, | |
| "learning_rate": 8.6e-06, | |
| "loss": 0.0367, | |
| "step": 43 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "grad_norm": 0.88671875, | |
| "learning_rate": 8.8e-06, | |
| "loss": 0.035, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 0.65, | |
| "grad_norm": 0.94921875, | |
| "learning_rate": 9e-06, | |
| "loss": 0.0417, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "grad_norm": 0.81640625, | |
| "learning_rate": 9.200000000000002e-06, | |
| "loss": 0.0279, | |
| "step": 46 | |
| }, | |
| { | |
| "epoch": 0.68, | |
| "grad_norm": 0.91796875, | |
| "learning_rate": 9.4e-06, | |
| "loss": 0.0292, | |
| "step": 47 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "grad_norm": 1.5625, | |
| "learning_rate": 9.600000000000001e-06, | |
| "loss": 0.0345, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 0.71, | |
| "grad_norm": 0.8515625, | |
| "learning_rate": 9.800000000000001e-06, | |
| "loss": 0.0348, | |
| "step": 49 | |
| }, | |
| { | |
| "epoch": 0.72, | |
| "grad_norm": 0.98046875, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0337, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.74, | |
| "grad_norm": 0.75, | |
| "learning_rate": 1.02e-05, | |
| "loss": 0.0361, | |
| "step": 51 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "grad_norm": 0.71875, | |
| "learning_rate": 1.04e-05, | |
| "loss": 0.0363, | |
| "step": 52 | |
| }, | |
| { | |
| "epoch": 0.77, | |
| "grad_norm": 0.7265625, | |
| "learning_rate": 1.0600000000000002e-05, | |
| "loss": 0.0329, | |
| "step": 53 | |
| }, | |
| { | |
| "epoch": 0.78, | |
| "grad_norm": 0.68359375, | |
| "learning_rate": 1.0800000000000002e-05, | |
| "loss": 0.032, | |
| "step": 54 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "grad_norm": 0.46875, | |
| "learning_rate": 1.1000000000000001e-05, | |
| "loss": 0.0275, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 0.81, | |
| "grad_norm": 0.62890625, | |
| "learning_rate": 1.1200000000000001e-05, | |
| "loss": 0.0304, | |
| "step": 56 | |
| }, | |
| { | |
| "epoch": 0.83, | |
| "grad_norm": 0.7421875, | |
| "learning_rate": 1.14e-05, | |
| "loss": 0.0325, | |
| "step": 57 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "grad_norm": 0.51953125, | |
| "learning_rate": 1.16e-05, | |
| "loss": 0.027, | |
| "step": 58 | |
| }, | |
| { | |
| "epoch": 0.86, | |
| "grad_norm": 1.015625, | |
| "learning_rate": 1.18e-05, | |
| "loss": 0.0283, | |
| "step": 59 | |
| }, | |
| { | |
| "epoch": 0.87, | |
| "grad_norm": 0.62890625, | |
| "learning_rate": 1.2e-05, | |
| "loss": 0.0301, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "grad_norm": 0.734375, | |
| "learning_rate": 1.22e-05, | |
| "loss": 0.0334, | |
| "step": 61 | |
| }, | |
| { | |
| "epoch": 0.9, | |
| "grad_norm": 0.6875, | |
| "learning_rate": 1.2400000000000002e-05, | |
| "loss": 0.0312, | |
| "step": 62 | |
| }, | |
| { | |
| "epoch": 0.91, | |
| "grad_norm": 0.58203125, | |
| "learning_rate": 1.2600000000000001e-05, | |
| "loss": 0.0282, | |
| "step": 63 | |
| }, | |
| { | |
| "epoch": 0.93, | |
| "grad_norm": 0.65625, | |
| "learning_rate": 1.2800000000000001e-05, | |
| "loss": 0.0315, | |
| "step": 64 | |
| }, | |
| { | |
| "epoch": 0.94, | |
| "grad_norm": 0.66015625, | |
| "learning_rate": 1.3000000000000001e-05, | |
| "loss": 0.0316, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "grad_norm": 0.55078125, | |
| "learning_rate": 1.3200000000000002e-05, | |
| "loss": 0.0292, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 0.97, | |
| "grad_norm": 0.84375, | |
| "learning_rate": 1.3400000000000002e-05, | |
| "loss": 0.0304, | |
| "step": 67 | |
| }, | |
| { | |
| "epoch": 0.99, | |
| "grad_norm": 0.498046875, | |
| "learning_rate": 1.3600000000000002e-05, | |
| "loss": 0.0235, | |
| "step": 68 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 0.6875, | |
| "learning_rate": 1.38e-05, | |
| "loss": 0.0262, | |
| "step": 69 | |
| } | |
| ], | |
| "logging_steps": 1, | |
| "max_steps": 207, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "save_steps": 69, | |
| "total_flos": 2.036228957279355e+17, | |
| "train_batch_size": 1, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |