| { |
| "best_metric": null, |
| "best_model_checkpoint": null, |
| "epoch": 5.0, |
| "eval_steps": 500, |
| "global_step": 360, |
| "is_hyper_param_search": false, |
| "is_local_process_zero": true, |
| "is_world_process_zero": true, |
| "log_history": [ |
| { |
| "epoch": 0.013888888888888888, |
| "grad_norm": 95.33702850341797, |
| "learning_rate": 5.0000000000000004e-08, |
| "loss": 5.3695, |
| "step": 1 |
| }, |
| { |
| "epoch": 0.027777777777777776, |
| "grad_norm": 96.05199432373047, |
| "learning_rate": 1.0000000000000001e-07, |
| "loss": 5.4178, |
| "step": 2 |
| }, |
| { |
| "epoch": 0.041666666666666664, |
| "grad_norm": 92.71851348876953, |
| "learning_rate": 1.5000000000000002e-07, |
| "loss": 5.4472, |
| "step": 3 |
| }, |
| { |
| "epoch": 0.05555555555555555, |
| "grad_norm": 95.78789520263672, |
| "learning_rate": 2.0000000000000002e-07, |
| "loss": 5.4114, |
| "step": 4 |
| }, |
| { |
| "epoch": 0.06944444444444445, |
| "grad_norm": 95.62769317626953, |
| "learning_rate": 2.5000000000000004e-07, |
| "loss": 5.3612, |
| "step": 5 |
| }, |
| { |
| "epoch": 0.08333333333333333, |
| "grad_norm": 94.85267639160156, |
| "learning_rate": 3.0000000000000004e-07, |
| "loss": 5.4065, |
| "step": 6 |
| }, |
| { |
| "epoch": 0.09722222222222222, |
| "grad_norm": 93.81153869628906, |
| "learning_rate": 3.5000000000000004e-07, |
| "loss": 5.364, |
| "step": 7 |
| }, |
| { |
| "epoch": 0.1111111111111111, |
| "grad_norm": 95.33504486083984, |
| "learning_rate": 4.0000000000000003e-07, |
| "loss": 5.4484, |
| "step": 8 |
| }, |
| { |
| "epoch": 0.125, |
| "grad_norm": 94.11618041992188, |
| "learning_rate": 4.5000000000000003e-07, |
| "loss": 5.3494, |
| "step": 9 |
| }, |
| { |
| "epoch": 0.1388888888888889, |
| "grad_norm": 93.71621704101562, |
| "learning_rate": 5.000000000000001e-07, |
| "loss": 5.2325, |
| "step": 10 |
| }, |
| { |
| "epoch": 0.1527777777777778, |
| "grad_norm": 92.6498031616211, |
| "learning_rate": 5.5e-07, |
| "loss": 5.1224, |
| "step": 11 |
| }, |
| { |
| "epoch": 0.16666666666666666, |
| "grad_norm": 89.05821228027344, |
| "learning_rate": 6.000000000000001e-07, |
| "loss": 5.0484, |
| "step": 12 |
| }, |
| { |
| "epoch": 0.18055555555555555, |
| "grad_norm": 87.52056884765625, |
| "learning_rate": 6.5e-07, |
| "loss": 4.8831, |
| "step": 13 |
| }, |
| { |
| "epoch": 0.19444444444444445, |
| "grad_norm": 87.93170928955078, |
| "learning_rate": 7.000000000000001e-07, |
| "loss": 4.7723, |
| "step": 14 |
| }, |
| { |
| "epoch": 0.20833333333333334, |
| "grad_norm": 84.07933044433594, |
| "learning_rate": 7.5e-07, |
| "loss": 4.5843, |
| "step": 15 |
| }, |
| { |
| "epoch": 0.2222222222222222, |
| "grad_norm": 82.16165161132812, |
| "learning_rate": 8.000000000000001e-07, |
| "loss": 4.4273, |
| "step": 16 |
| }, |
| { |
| "epoch": 0.2361111111111111, |
| "grad_norm": 78.94780731201172, |
| "learning_rate": 8.500000000000001e-07, |
| "loss": 4.155, |
| "step": 17 |
| }, |
| { |
| "epoch": 0.25, |
| "grad_norm": 75.85335540771484, |
| "learning_rate": 9.000000000000001e-07, |
| "loss": 3.9005, |
| "step": 18 |
| }, |
| { |
| "epoch": 0.2638888888888889, |
| "grad_norm": 71.74459075927734, |
| "learning_rate": 9.500000000000001e-07, |
| "loss": 3.5885, |
| "step": 19 |
| }, |
| { |
| "epoch": 0.2777777777777778, |
| "grad_norm": 67.66385650634766, |
| "learning_rate": 1.0000000000000002e-06, |
| "loss": 3.308, |
| "step": 20 |
| }, |
| { |
| "epoch": 0.2916666666666667, |
| "grad_norm": 64.22036743164062, |
| "learning_rate": 1.0500000000000001e-06, |
| "loss": 3.057, |
| "step": 21 |
| }, |
| { |
| "epoch": 0.3055555555555556, |
| "grad_norm": 60.24971008300781, |
| "learning_rate": 1.1e-06, |
| "loss": 2.7939, |
| "step": 22 |
| }, |
| { |
| "epoch": 0.3194444444444444, |
| "grad_norm": 57.68950271606445, |
| "learning_rate": 1.1500000000000002e-06, |
| "loss": 2.5539, |
| "step": 23 |
| }, |
| { |
| "epoch": 0.3333333333333333, |
| "grad_norm": 56.10525894165039, |
| "learning_rate": 1.2000000000000002e-06, |
| "loss": 2.2688, |
| "step": 24 |
| }, |
| { |
| "epoch": 0.3472222222222222, |
| "grad_norm": 54.84128189086914, |
| "learning_rate": 1.25e-06, |
| "loss": 2.009, |
| "step": 25 |
| }, |
| { |
| "epoch": 0.3611111111111111, |
| "grad_norm": 53.96566390991211, |
| "learning_rate": 1.3e-06, |
| "loss": 1.6682, |
| "step": 26 |
| }, |
| { |
| "epoch": 0.375, |
| "grad_norm": 53.859439849853516, |
| "learning_rate": 1.3500000000000002e-06, |
| "loss": 1.3807, |
| "step": 27 |
| }, |
| { |
| "epoch": 0.3888888888888889, |
| "grad_norm": 53.22216796875, |
| "learning_rate": 1.4000000000000001e-06, |
| "loss": 1.1229, |
| "step": 28 |
| }, |
| { |
| "epoch": 0.4027777777777778, |
| "grad_norm": 50.3905143737793, |
| "learning_rate": 1.45e-06, |
| "loss": 0.8454, |
| "step": 29 |
| }, |
| { |
| "epoch": 0.4166666666666667, |
| "grad_norm": 40.14346694946289, |
| "learning_rate": 1.5e-06, |
| "loss": 0.5817, |
| "step": 30 |
| }, |
| { |
| "epoch": 0.4305555555555556, |
| "grad_norm": 34.043235778808594, |
| "learning_rate": 1.5500000000000002e-06, |
| "loss": 0.4234, |
| "step": 31 |
| }, |
| { |
| "epoch": 0.4444444444444444, |
| "grad_norm": 28.703292846679688, |
| "learning_rate": 1.6000000000000001e-06, |
| "loss": 0.3074, |
| "step": 32 |
| }, |
| { |
| "epoch": 0.4583333333333333, |
| "grad_norm": 22.456626892089844, |
| "learning_rate": 1.6500000000000003e-06, |
| "loss": 0.2146, |
| "step": 33 |
| }, |
| { |
| "epoch": 0.4722222222222222, |
| "grad_norm": 15.350156784057617, |
| "learning_rate": 1.7000000000000002e-06, |
| "loss": 0.1103, |
| "step": 34 |
| }, |
| { |
| "epoch": 0.4861111111111111, |
| "grad_norm": 6.5175580978393555, |
| "learning_rate": 1.75e-06, |
| "loss": 0.0574, |
| "step": 35 |
| }, |
| { |
| "epoch": 0.5, |
| "grad_norm": 6.813859939575195, |
| "learning_rate": 1.8000000000000001e-06, |
| "loss": 0.0497, |
| "step": 36 |
| }, |
| { |
| "epoch": 0.5138888888888888, |
| "grad_norm": 3.5350818634033203, |
| "learning_rate": 1.85e-06, |
| "loss": 0.0199, |
| "step": 37 |
| }, |
| { |
| "epoch": 0.5277777777777778, |
| "grad_norm": 4.9657135009765625, |
| "learning_rate": 1.9000000000000002e-06, |
| "loss": 0.0245, |
| "step": 38 |
| }, |
| { |
| "epoch": 0.5416666666666666, |
| "grad_norm": 1.2266570329666138, |
| "learning_rate": 1.9500000000000004e-06, |
| "loss": 0.0087, |
| "step": 39 |
| }, |
| { |
| "epoch": 0.5555555555555556, |
| "grad_norm": 4.982794761657715, |
| "learning_rate": 2.0000000000000003e-06, |
| "loss": 0.048, |
| "step": 40 |
| }, |
| { |
| "epoch": 0.5694444444444444, |
| "grad_norm": 7.6967387199401855, |
| "learning_rate": 2.05e-06, |
| "loss": 0.0449, |
| "step": 41 |
| }, |
| { |
| "epoch": 0.5833333333333334, |
| "grad_norm": 2.9466323852539062, |
| "learning_rate": 2.1000000000000002e-06, |
| "loss": 0.0098, |
| "step": 42 |
| }, |
| { |
| "epoch": 0.5972222222222222, |
| "grad_norm": 2.191035747528076, |
| "learning_rate": 2.15e-06, |
| "loss": 0.0087, |
| "step": 43 |
| }, |
| { |
| "epoch": 0.6111111111111112, |
| "grad_norm": 7.3228840827941895, |
| "learning_rate": 2.2e-06, |
| "loss": 0.0339, |
| "step": 44 |
| }, |
| { |
| "epoch": 0.625, |
| "grad_norm": 4.737412929534912, |
| "learning_rate": 2.25e-06, |
| "loss": 0.0149, |
| "step": 45 |
| }, |
| { |
| "epoch": 0.6388888888888888, |
| "grad_norm": 1.7612887620925903, |
| "learning_rate": 2.3000000000000004e-06, |
| "loss": 0.0071, |
| "step": 46 |
| }, |
| { |
| "epoch": 0.6527777777777778, |
| "grad_norm": 3.344369649887085, |
| "learning_rate": 2.35e-06, |
| "loss": 0.0196, |
| "step": 47 |
| }, |
| { |
| "epoch": 0.6666666666666666, |
| "grad_norm": 2.2957704067230225, |
| "learning_rate": 2.4000000000000003e-06, |
| "loss": 0.0216, |
| "step": 48 |
| }, |
| { |
| "epoch": 0.6805555555555556, |
| "grad_norm": 1.83970308303833, |
| "learning_rate": 2.4500000000000003e-06, |
| "loss": 0.0221, |
| "step": 49 |
| }, |
| { |
| "epoch": 0.6944444444444444, |
| "grad_norm": 1.547287106513977, |
| "learning_rate": 2.5e-06, |
| "loss": 0.0061, |
| "step": 50 |
| }, |
| { |
| "epoch": 0.7083333333333334, |
| "grad_norm": 1.6580262184143066, |
| "learning_rate": 2.55e-06, |
| "loss": 0.0127, |
| "step": 51 |
| }, |
| { |
| "epoch": 0.7222222222222222, |
| "grad_norm": 1.8863849639892578, |
| "learning_rate": 2.6e-06, |
| "loss": 0.0124, |
| "step": 52 |
| }, |
| { |
| "epoch": 0.7361111111111112, |
| "grad_norm": 1.2086963653564453, |
| "learning_rate": 2.6500000000000005e-06, |
| "loss": 0.0085, |
| "step": 53 |
| }, |
| { |
| "epoch": 0.75, |
| "grad_norm": 2.4265427589416504, |
| "learning_rate": 2.7000000000000004e-06, |
| "loss": 0.0255, |
| "step": 54 |
| }, |
| { |
| "epoch": 0.7638888888888888, |
| "grad_norm": 2.0333855152130127, |
| "learning_rate": 2.7500000000000004e-06, |
| "loss": 0.0145, |
| "step": 55 |
| }, |
| { |
| "epoch": 0.7777777777777778, |
| "grad_norm": 2.9129791259765625, |
| "learning_rate": 2.8000000000000003e-06, |
| "loss": 0.0228, |
| "step": 56 |
| }, |
| { |
| "epoch": 0.7916666666666666, |
| "grad_norm": 0.8359552621841431, |
| "learning_rate": 2.85e-06, |
| "loss": 0.0109, |
| "step": 57 |
| }, |
| { |
| "epoch": 0.8055555555555556, |
| "grad_norm": 1.677287220954895, |
| "learning_rate": 2.9e-06, |
| "loss": 0.0246, |
| "step": 58 |
| }, |
| { |
| "epoch": 0.8194444444444444, |
| "grad_norm": 1.0761189460754395, |
| "learning_rate": 2.95e-06, |
| "loss": 0.0058, |
| "step": 59 |
| }, |
| { |
| "epoch": 0.8333333333333334, |
| "grad_norm": 1.1098495721817017, |
| "learning_rate": 3e-06, |
| "loss": 0.0167, |
| "step": 60 |
| }, |
| { |
| "epoch": 0.8472222222222222, |
| "grad_norm": 1.0377463102340698, |
| "learning_rate": 3.05e-06, |
| "loss": 0.007, |
| "step": 61 |
| }, |
| { |
| "epoch": 0.8611111111111112, |
| "grad_norm": 0.6554276347160339, |
| "learning_rate": 3.1000000000000004e-06, |
| "loss": 0.0039, |
| "step": 62 |
| }, |
| { |
| "epoch": 0.875, |
| "grad_norm": 1.279077410697937, |
| "learning_rate": 3.1500000000000003e-06, |
| "loss": 0.0199, |
| "step": 63 |
| }, |
| { |
| "epoch": 0.8888888888888888, |
| "grad_norm": 0.7709833979606628, |
| "learning_rate": 3.2000000000000003e-06, |
| "loss": 0.0043, |
| "step": 64 |
| }, |
| { |
| "epoch": 0.9027777777777778, |
| "grad_norm": 1.3619827032089233, |
| "learning_rate": 3.2500000000000002e-06, |
| "loss": 0.0067, |
| "step": 65 |
| }, |
| { |
| "epoch": 0.9166666666666666, |
| "grad_norm": 1.3849765062332153, |
| "learning_rate": 3.3000000000000006e-06, |
| "loss": 0.0091, |
| "step": 66 |
| }, |
| { |
| "epoch": 0.9305555555555556, |
| "grad_norm": 1.1380351781845093, |
| "learning_rate": 3.3500000000000005e-06, |
| "loss": 0.0143, |
| "step": 67 |
| }, |
| { |
| "epoch": 0.9444444444444444, |
| "grad_norm": 1.9759573936462402, |
| "learning_rate": 3.4000000000000005e-06, |
| "loss": 0.0231, |
| "step": 68 |
| }, |
| { |
| "epoch": 0.9583333333333334, |
| "grad_norm": 0.5920725464820862, |
| "learning_rate": 3.45e-06, |
| "loss": 0.0063, |
| "step": 69 |
| }, |
| { |
| "epoch": 0.9722222222222222, |
| "grad_norm": 1.2489631175994873, |
| "learning_rate": 3.5e-06, |
| "loss": 0.0149, |
| "step": 70 |
| }, |
| { |
| "epoch": 0.9861111111111112, |
| "grad_norm": 1.5553513765335083, |
| "learning_rate": 3.5500000000000003e-06, |
| "loss": 0.0085, |
| "step": 71 |
| }, |
| { |
| "epoch": 1.0, |
| "grad_norm": 1.0278698205947876, |
| "learning_rate": 3.6000000000000003e-06, |
| "loss": 0.0061, |
| "step": 72 |
| }, |
| { |
| "epoch": 1.0138888888888888, |
| "grad_norm": 1.9279776811599731, |
| "learning_rate": 3.65e-06, |
| "loss": 0.0118, |
| "step": 73 |
| }, |
| { |
| "epoch": 1.0277777777777777, |
| "grad_norm": 2.592510223388672, |
| "learning_rate": 3.7e-06, |
| "loss": 0.0309, |
| "step": 74 |
| }, |
| { |
| "epoch": 1.0416666666666667, |
| "grad_norm": 1.6530096530914307, |
| "learning_rate": 3.7500000000000005e-06, |
| "loss": 0.0099, |
| "step": 75 |
| }, |
| { |
| "epoch": 1.0555555555555556, |
| "grad_norm": 0.4624968469142914, |
| "learning_rate": 3.8000000000000005e-06, |
| "loss": 0.0053, |
| "step": 76 |
| }, |
| { |
| "epoch": 1.0694444444444444, |
| "grad_norm": 0.629607617855072, |
| "learning_rate": 3.85e-06, |
| "loss": 0.0087, |
| "step": 77 |
| }, |
| { |
| "epoch": 1.0833333333333333, |
| "grad_norm": 0.14678587019443512, |
| "learning_rate": 3.900000000000001e-06, |
| "loss": 0.0007, |
| "step": 78 |
| }, |
| { |
| "epoch": 1.0972222222222223, |
| "grad_norm": 1.0052955150604248, |
| "learning_rate": 3.95e-06, |
| "loss": 0.0089, |
| "step": 79 |
| }, |
| { |
| "epoch": 1.1111111111111112, |
| "grad_norm": 0.9059052467346191, |
| "learning_rate": 4.000000000000001e-06, |
| "loss": 0.0049, |
| "step": 80 |
| }, |
| { |
| "epoch": 1.125, |
| "grad_norm": 0.871515691280365, |
| "learning_rate": 4.05e-06, |
| "loss": 0.0092, |
| "step": 81 |
| }, |
| { |
| "epoch": 1.1388888888888888, |
| "grad_norm": 0.07520482689142227, |
| "learning_rate": 4.1e-06, |
| "loss": 0.0004, |
| "step": 82 |
| }, |
| { |
| "epoch": 1.1527777777777777, |
| "grad_norm": 0.8039830923080444, |
| "learning_rate": 4.15e-06, |
| "loss": 0.0031, |
| "step": 83 |
| }, |
| { |
| "epoch": 1.1666666666666667, |
| "grad_norm": 0.05209676921367645, |
| "learning_rate": 4.2000000000000004e-06, |
| "loss": 0.0003, |
| "step": 84 |
| }, |
| { |
| "epoch": 1.1805555555555556, |
| "grad_norm": 0.04410432651638985, |
| "learning_rate": 4.25e-06, |
| "loss": 0.0002, |
| "step": 85 |
| }, |
| { |
| "epoch": 1.1944444444444444, |
| "grad_norm": 0.06021099537611008, |
| "learning_rate": 4.3e-06, |
| "loss": 0.0002, |
| "step": 86 |
| }, |
| { |
| "epoch": 1.2083333333333333, |
| "grad_norm": 0.283663272857666, |
| "learning_rate": 4.350000000000001e-06, |
| "loss": 0.001, |
| "step": 87 |
| }, |
| { |
| "epoch": 1.2222222222222223, |
| "grad_norm": 0.6798669695854187, |
| "learning_rate": 4.4e-06, |
| "loss": 0.0033, |
| "step": 88 |
| }, |
| { |
| "epoch": 1.2361111111111112, |
| "grad_norm": 0.06857122480869293, |
| "learning_rate": 4.450000000000001e-06, |
| "loss": 0.0002, |
| "step": 89 |
| }, |
| { |
| "epoch": 1.25, |
| "grad_norm": 1.9897412061691284, |
| "learning_rate": 4.5e-06, |
| "loss": 0.0126, |
| "step": 90 |
| }, |
| { |
| "epoch": 1.2638888888888888, |
| "grad_norm": 1.2174546718597412, |
| "learning_rate": 4.5500000000000005e-06, |
| "loss": 0.0023, |
| "step": 91 |
| }, |
| { |
| "epoch": 1.2777777777777777, |
| "grad_norm": 2.2762115001678467, |
| "learning_rate": 4.600000000000001e-06, |
| "loss": 0.0095, |
| "step": 92 |
| }, |
| { |
| "epoch": 1.2916666666666667, |
| "grad_norm": 3.728581428527832, |
| "learning_rate": 4.65e-06, |
| "loss": 0.0046, |
| "step": 93 |
| }, |
| { |
| "epoch": 1.3055555555555556, |
| "grad_norm": 2.942180871963501, |
| "learning_rate": 4.7e-06, |
| "loss": 0.0261, |
| "step": 94 |
| }, |
| { |
| "epoch": 1.3194444444444444, |
| "grad_norm": 0.3911007344722748, |
| "learning_rate": 4.75e-06, |
| "loss": 0.0005, |
| "step": 95 |
| }, |
| { |
| "epoch": 1.3333333333333333, |
| "grad_norm": 0.03832005336880684, |
| "learning_rate": 4.800000000000001e-06, |
| "loss": 0.0001, |
| "step": 96 |
| }, |
| { |
| "epoch": 1.3472222222222223, |
| "grad_norm": 0.7516991496086121, |
| "learning_rate": 4.85e-06, |
| "loss": 0.0017, |
| "step": 97 |
| }, |
| { |
| "epoch": 1.3611111111111112, |
| "grad_norm": 0.4136435091495514, |
| "learning_rate": 4.9000000000000005e-06, |
| "loss": 0.0014, |
| "step": 98 |
| }, |
| { |
| "epoch": 1.375, |
| "grad_norm": 0.044784627854824066, |
| "learning_rate": 4.95e-06, |
| "loss": 0.0001, |
| "step": 99 |
| }, |
| { |
| "epoch": 1.3888888888888888, |
| "grad_norm": 0.7667476534843445, |
| "learning_rate": 5e-06, |
| "loss": 0.0022, |
| "step": 100 |
| }, |
| { |
| "epoch": 1.4027777777777777, |
| "grad_norm": 0.9187825322151184, |
| "learning_rate": 4.999888074163108e-06, |
| "loss": 0.0035, |
| "step": 101 |
| }, |
| { |
| "epoch": 1.4166666666666667, |
| "grad_norm": 1.2859818935394287, |
| "learning_rate": 4.999552306674345e-06, |
| "loss": 0.0046, |
| "step": 102 |
| }, |
| { |
| "epoch": 1.4305555555555556, |
| "grad_norm": 0.5899283289909363, |
| "learning_rate": 4.998992727598557e-06, |
| "loss": 0.0027, |
| "step": 103 |
| }, |
| { |
| "epoch": 1.4444444444444444, |
| "grad_norm": 0.027552006766200066, |
| "learning_rate": 4.998209387040829e-06, |
| "loss": 0.0001, |
| "step": 104 |
| }, |
| { |
| "epoch": 1.4583333333333333, |
| "grad_norm": 0.14271241426467896, |
| "learning_rate": 4.9972023551419995e-06, |
| "loss": 0.0002, |
| "step": 105 |
| }, |
| { |
| "epoch": 1.4722222222222223, |
| "grad_norm": 0.36476054787635803, |
| "learning_rate": 4.995971722072379e-06, |
| "loss": 0.001, |
| "step": 106 |
| }, |
| { |
| "epoch": 1.4861111111111112, |
| "grad_norm": 1.7742875814437866, |
| "learning_rate": 4.9945175980236745e-06, |
| "loss": 0.0025, |
| "step": 107 |
| }, |
| { |
| "epoch": 1.5, |
| "grad_norm": 0.5356531739234924, |
| "learning_rate": 4.992840113199131e-06, |
| "loss": 0.0011, |
| "step": 108 |
| }, |
| { |
| "epoch": 1.5138888888888888, |
| "grad_norm": 0.9400438070297241, |
| "learning_rate": 4.990939417801859e-06, |
| "loss": 0.0021, |
| "step": 109 |
| }, |
| { |
| "epoch": 1.5277777777777777, |
| "grad_norm": 0.028793074190616608, |
| "learning_rate": 4.988815682021398e-06, |
| "loss": 0.0001, |
| "step": 110 |
| }, |
| { |
| "epoch": 1.5416666666666665, |
| "grad_norm": 0.016167419031262398, |
| "learning_rate": 4.986469096018472e-06, |
| "loss": 0.0001, |
| "step": 111 |
| }, |
| { |
| "epoch": 1.5555555555555556, |
| "grad_norm": 0.04863646626472473, |
| "learning_rate": 4.983899869907963e-06, |
| "loss": 0.0001, |
| "step": 112 |
| }, |
| { |
| "epoch": 1.5694444444444444, |
| "grad_norm": 0.11991345882415771, |
| "learning_rate": 4.981108233740096e-06, |
| "loss": 0.0003, |
| "step": 113 |
| }, |
| { |
| "epoch": 1.5833333333333335, |
| "grad_norm": 1.1984789371490479, |
| "learning_rate": 4.978094437479843e-06, |
| "loss": 0.0205, |
| "step": 114 |
| }, |
| { |
| "epoch": 1.5972222222222223, |
| "grad_norm": 0.2907632887363434, |
| "learning_rate": 4.97485875098454e-06, |
| "loss": 0.0005, |
| "step": 115 |
| }, |
| { |
| "epoch": 1.6111111111111112, |
| "grad_norm": 2.186234474182129, |
| "learning_rate": 4.971401463979722e-06, |
| "loss": 0.0068, |
| "step": 116 |
| }, |
| { |
| "epoch": 1.625, |
| "grad_norm": 0.17205508053302765, |
| "learning_rate": 4.967722886033181e-06, |
| "loss": 0.0004, |
| "step": 117 |
| }, |
| { |
| "epoch": 1.6388888888888888, |
| "grad_norm": 1.9445418119430542, |
| "learning_rate": 4.963823346527249e-06, |
| "loss": 0.0052, |
| "step": 118 |
| }, |
| { |
| "epoch": 1.6527777777777777, |
| "grad_norm": 0.8023346662521362, |
| "learning_rate": 4.959703194629304e-06, |
| "loss": 0.0087, |
| "step": 119 |
| }, |
| { |
| "epoch": 1.6666666666666665, |
| "grad_norm": 0.4967271685600281, |
| "learning_rate": 4.955362799260507e-06, |
| "loss": 0.0011, |
| "step": 120 |
| }, |
| { |
| "epoch": 1.6805555555555556, |
| "grad_norm": 0.18211083114147186, |
| "learning_rate": 4.950802549062764e-06, |
| "loss": 0.0007, |
| "step": 121 |
| }, |
| { |
| "epoch": 1.6944444444444444, |
| "grad_norm": 0.2594281733036041, |
| "learning_rate": 4.946022852363932e-06, |
| "loss": 0.0011, |
| "step": 122 |
| }, |
| { |
| "epoch": 1.7083333333333335, |
| "grad_norm": 0.8132149577140808, |
| "learning_rate": 4.9410241371412525e-06, |
| "loss": 0.0023, |
| "step": 123 |
| }, |
| { |
| "epoch": 1.7222222222222223, |
| "grad_norm": 0.27094146609306335, |
| "learning_rate": 4.935806850983034e-06, |
| "loss": 0.001, |
| "step": 124 |
| }, |
| { |
| "epoch": 1.7361111111111112, |
| "grad_norm": 0.10404836386442184, |
| "learning_rate": 4.9303714610485705e-06, |
| "loss": 0.0005, |
| "step": 125 |
| }, |
| { |
| "epoch": 1.75, |
| "grad_norm": 0.9027087688446045, |
| "learning_rate": 4.924718454026318e-06, |
| "loss": 0.005, |
| "step": 126 |
| }, |
| { |
| "epoch": 1.7638888888888888, |
| "grad_norm": 0.3427490293979645, |
| "learning_rate": 4.918848336090309e-06, |
| "loss": 0.0013, |
| "step": 127 |
| }, |
| { |
| "epoch": 1.7777777777777777, |
| "grad_norm": 0.41760626435279846, |
| "learning_rate": 4.912761632854834e-06, |
| "loss": 0.0032, |
| "step": 128 |
| }, |
| { |
| "epoch": 1.7916666666666665, |
| "grad_norm": 0.09792769700288773, |
| "learning_rate": 4.906458889327375e-06, |
| "loss": 0.0004, |
| "step": 129 |
| }, |
| { |
| "epoch": 1.8055555555555556, |
| "grad_norm": 0.5523985624313354, |
| "learning_rate": 4.899940669859807e-06, |
| "loss": 0.0077, |
| "step": 130 |
| }, |
| { |
| "epoch": 1.8194444444444444, |
| "grad_norm": 0.15261663496494293, |
| "learning_rate": 4.893207558097867e-06, |
| "loss": 0.0007, |
| "step": 131 |
| }, |
| { |
| "epoch": 1.8333333333333335, |
| "grad_norm": 0.12152901291847229, |
| "learning_rate": 4.8862601569288885e-06, |
| "loss": 0.0005, |
| "step": 132 |
| }, |
| { |
| "epoch": 1.8472222222222223, |
| "grad_norm": 0.1660558432340622, |
| "learning_rate": 4.879099088427824e-06, |
| "loss": 0.0007, |
| "step": 133 |
| }, |
| { |
| "epoch": 1.8611111111111112, |
| "grad_norm": 0.45732977986335754, |
| "learning_rate": 4.871724993801541e-06, |
| "loss": 0.0008, |
| "step": 134 |
| }, |
| { |
| "epoch": 1.875, |
| "grad_norm": 0.1277359277009964, |
| "learning_rate": 4.864138533331411e-06, |
| "loss": 0.0004, |
| "step": 135 |
| }, |
| { |
| "epoch": 1.8888888888888888, |
| "grad_norm": 0.08218172937631607, |
| "learning_rate": 4.8563403863141825e-06, |
| "loss": 0.0003, |
| "step": 136 |
| }, |
| { |
| "epoch": 1.9027777777777777, |
| "grad_norm": 0.35776910185813904, |
| "learning_rate": 4.84833125100116e-06, |
| "loss": 0.0012, |
| "step": 137 |
| }, |
| { |
| "epoch": 1.9166666666666665, |
| "grad_norm": 0.11485318094491959, |
| "learning_rate": 4.840111844535682e-06, |
| "loss": 0.0004, |
| "step": 138 |
| }, |
| { |
| "epoch": 1.9305555555555556, |
| "grad_norm": 0.03145573288202286, |
| "learning_rate": 4.8316829028889076e-06, |
| "loss": 0.0001, |
| "step": 139 |
| }, |
| { |
| "epoch": 1.9444444444444444, |
| "grad_norm": 0.03419964388012886, |
| "learning_rate": 4.823045180793914e-06, |
| "loss": 0.0001, |
| "step": 140 |
| }, |
| { |
| "epoch": 1.9583333333333335, |
| "grad_norm": 0.008005158044397831, |
| "learning_rate": 4.8141994516781196e-06, |
| "loss": 0.0, |
| "step": 141 |
| }, |
| { |
| "epoch": 1.9722222222222223, |
| "grad_norm": 0.30687257647514343, |
| "learning_rate": 4.805146507594034e-06, |
| "loss": 0.0006, |
| "step": 142 |
| }, |
| { |
| "epoch": 1.9861111111111112, |
| "grad_norm": 0.0030141561292111874, |
| "learning_rate": 4.7958871591483305e-06, |
| "loss": 0.0, |
| "step": 143 |
| }, |
| { |
| "epoch": 2.0, |
| "grad_norm": 0.0026682906318455935, |
| "learning_rate": 4.786422235429269e-06, |
| "loss": 0.0, |
| "step": 144 |
| }, |
| { |
| "epoch": 2.013888888888889, |
| "grad_norm": 0.008881614543497562, |
| "learning_rate": 4.776752583932455e-06, |
| "loss": 0.0, |
| "step": 145 |
| }, |
| { |
| "epoch": 2.0277777777777777, |
| "grad_norm": 0.001991939265280962, |
| "learning_rate": 4.766879070484957e-06, |
| "loss": 0.0, |
| "step": 146 |
| }, |
| { |
| "epoch": 2.0416666666666665, |
| "grad_norm": 0.006773048546165228, |
| "learning_rate": 4.756802579167781e-06, |
| "loss": 0.0, |
| "step": 147 |
| }, |
| { |
| "epoch": 2.0555555555555554, |
| "grad_norm": 0.00205055414699018, |
| "learning_rate": 4.746524012236706e-06, |
| "loss": 0.0, |
| "step": 148 |
| }, |
| { |
| "epoch": 2.0694444444444446, |
| "grad_norm": 0.0041306703351438046, |
| "learning_rate": 4.736044290041496e-06, |
| "loss": 0.0, |
| "step": 149 |
| }, |
| { |
| "epoch": 2.0833333333333335, |
| "grad_norm": 0.001996932551264763, |
| "learning_rate": 4.725364350943492e-06, |
| "loss": 0.0, |
| "step": 150 |
| }, |
| { |
| "epoch": 2.0972222222222223, |
| "grad_norm": 0.003478364087641239, |
| "learning_rate": 4.714485151231593e-06, |
| "loss": 0.0, |
| "step": 151 |
| }, |
| { |
| "epoch": 2.111111111111111, |
| "grad_norm": 0.01027267798781395, |
| "learning_rate": 4.703407665036622e-06, |
| "loss": 0.0, |
| "step": 152 |
| }, |
| { |
| "epoch": 2.125, |
| "grad_norm": 0.0020006722770631313, |
| "learning_rate": 4.692132884244113e-06, |
| "loss": 0.0, |
| "step": 153 |
| }, |
| { |
| "epoch": 2.138888888888889, |
| "grad_norm": 0.006151544861495495, |
| "learning_rate": 4.680661818405485e-06, |
| "loss": 0.0, |
| "step": 154 |
| }, |
| { |
| "epoch": 2.1527777777777777, |
| "grad_norm": 0.0020251101814210415, |
| "learning_rate": 4.668995494647653e-06, |
| "loss": 0.0, |
| "step": 155 |
| }, |
| { |
| "epoch": 2.1666666666666665, |
| "grad_norm": 0.0021893251687288284, |
| "learning_rate": 4.657134957581057e-06, |
| "loss": 0.0, |
| "step": 156 |
| }, |
| { |
| "epoch": 2.1805555555555554, |
| "grad_norm": 0.0018878229893743992, |
| "learning_rate": 4.645081269206128e-06, |
| "loss": 0.0, |
| "step": 157 |
| }, |
| { |
| "epoch": 2.1944444444444446, |
| "grad_norm": 0.759505569934845, |
| "learning_rate": 4.632835508818192e-06, |
| "loss": 0.0002, |
| "step": 158 |
| }, |
| { |
| "epoch": 2.2083333333333335, |
| "grad_norm": 0.0018357271328568459, |
| "learning_rate": 4.620398772910833e-06, |
| "loss": 0.0, |
| "step": 159 |
| }, |
| { |
| "epoch": 2.2222222222222223, |
| "grad_norm": 0.001739851082675159, |
| "learning_rate": 4.607772175077712e-06, |
| "loss": 0.0, |
| "step": 160 |
| }, |
| { |
| "epoch": 2.236111111111111, |
| "grad_norm": 0.0031391726806759834, |
| "learning_rate": 4.59495684591285e-06, |
| "loss": 0.0, |
| "step": 161 |
| }, |
| { |
| "epoch": 2.25, |
| "grad_norm": 0.001889107865281403, |
| "learning_rate": 4.581953932909403e-06, |
| "loss": 0.0, |
| "step": 162 |
| }, |
| { |
| "epoch": 2.263888888888889, |
| "grad_norm": 0.0014751956332474947, |
| "learning_rate": 4.5687646003569055e-06, |
| "loss": 0.0, |
| "step": 163 |
| }, |
| { |
| "epoch": 2.2777777777777777, |
| "grad_norm": 0.0014328192919492722, |
| "learning_rate": 4.555390029237026e-06, |
| "loss": 0.0, |
| "step": 164 |
| }, |
| { |
| "epoch": 2.2916666666666665, |
| "grad_norm": 0.0014799810014665127, |
| "learning_rate": 4.541831417117815e-06, |
| "loss": 0.0, |
| "step": 165 |
| }, |
| { |
| "epoch": 2.3055555555555554, |
| "grad_norm": 0.0013504689559340477, |
| "learning_rate": 4.528089978046481e-06, |
| "loss": 0.0, |
| "step": 166 |
| }, |
| { |
| "epoch": 2.3194444444444446, |
| "grad_norm": 0.001344231073744595, |
| "learning_rate": 4.514166942440679e-06, |
| "loss": 0.0, |
| "step": 167 |
| }, |
| { |
| "epoch": 2.3333333333333335, |
| "grad_norm": 0.001366922864690423, |
| "learning_rate": 4.5000635569783365e-06, |
| "loss": 0.0, |
| "step": 168 |
| }, |
| { |
| "epoch": 2.3472222222222223, |
| "grad_norm": 0.0013009505346417427, |
| "learning_rate": 4.4857810844860325e-06, |
| "loss": 0.0, |
| "step": 169 |
| }, |
| { |
| "epoch": 2.361111111111111, |
| "grad_norm": 0.001247942098416388, |
| "learning_rate": 4.471320803825915e-06, |
| "loss": 0.0, |
| "step": 170 |
| }, |
| { |
| "epoch": 2.375, |
| "grad_norm": 0.0011512853670865297, |
| "learning_rate": 4.4566840097811956e-06, |
| "loss": 0.0, |
| "step": 171 |
| }, |
| { |
| "epoch": 2.388888888888889, |
| "grad_norm": 0.001167183741927147, |
| "learning_rate": 4.4418720129402145e-06, |
| "loss": 0.0, |
| "step": 172 |
| }, |
| { |
| "epoch": 2.4027777777777777, |
| "grad_norm": 0.0011255062418058515, |
| "learning_rate": 4.426886139579083e-06, |
| "loss": 0.0, |
| "step": 173 |
| }, |
| { |
| "epoch": 2.4166666666666665, |
| "grad_norm": 0.0010559650836512446, |
| "learning_rate": 4.411727731542937e-06, |
| "loss": 0.0, |
| "step": 174 |
| }, |
| { |
| "epoch": 2.4305555555555554, |
| "grad_norm": 0.0010468937689438462, |
| "learning_rate": 4.39639814612578e-06, |
| "loss": 0.0, |
| "step": 175 |
| }, |
| { |
| "epoch": 2.4444444444444446, |
| "grad_norm": 0.0009965520584955812, |
| "learning_rate": 4.3808987559489536e-06, |
| "loss": 0.0, |
| "step": 176 |
| }, |
| { |
| "epoch": 2.4583333333333335, |
| "grad_norm": 0.000950520858168602, |
| "learning_rate": 4.365230948838232e-06, |
| "loss": 0.0, |
| "step": 177 |
| }, |
| { |
| "epoch": 2.4722222222222223, |
| "grad_norm": 0.0009662396623753011, |
| "learning_rate": 4.349396127699552e-06, |
| "loss": 0.0, |
| "step": 178 |
| }, |
| { |
| "epoch": 2.486111111111111, |
| "grad_norm": 0.0009316857904195786, |
| "learning_rate": 4.3333957103934025e-06, |
| "loss": 0.0, |
| "step": 179 |
| }, |
| { |
| "epoch": 2.5, |
| "grad_norm": 0.0009209261042997241, |
| "learning_rate": 4.317231129607859e-06, |
| "loss": 0.0, |
| "step": 180 |
| }, |
| { |
| "epoch": 2.513888888888889, |
| "grad_norm": 0.0008661423926241696, |
| "learning_rate": 4.30090383273031e-06, |
| "loss": 0.0, |
| "step": 181 |
| }, |
| { |
| "epoch": 2.5277777777777777, |
| "grad_norm": 0.0008176359115168452, |
| "learning_rate": 4.2844152817178476e-06, |
| "loss": 0.0, |
| "step": 182 |
| }, |
| { |
| "epoch": 2.5416666666666665, |
| "grad_norm": 0.0008025923161767423, |
| "learning_rate": 4.267766952966369e-06, |
| "loss": 0.0, |
| "step": 183 |
| }, |
| { |
| "epoch": 2.5555555555555554, |
| "grad_norm": 0.0008243928314186633, |
| "learning_rate": 4.2509603371783776e-06, |
| "loss": 0.0, |
| "step": 184 |
| }, |
| { |
| "epoch": 2.5694444444444446, |
| "grad_norm": 0.0013688765466213226, |
| "learning_rate": 4.233996939229502e-06, |
| "loss": 0.0, |
| "step": 185 |
| }, |
| { |
| "epoch": 2.5833333333333335, |
| "grad_norm": 0.0007681425777263939, |
| "learning_rate": 4.216878278033753e-06, |
| "loss": 0.0, |
| "step": 186 |
| }, |
| { |
| "epoch": 2.5972222222222223, |
| "grad_norm": 0.0007665084558539093, |
| "learning_rate": 4.199605886407515e-06, |
| "loss": 0.0, |
| "step": 187 |
| }, |
| { |
| "epoch": 2.611111111111111, |
| "grad_norm": 0.000745179713703692, |
| "learning_rate": 4.1821813109322975e-06, |
| "loss": 0.0, |
| "step": 188 |
| }, |
| { |
| "epoch": 2.625, |
| "grad_norm": 0.0007080566138029099, |
| "learning_rate": 4.164606111816256e-06, |
| "loss": 0.0, |
| "step": 189 |
| }, |
| { |
| "epoch": 2.638888888888889, |
| "grad_norm": 0.0007035013986751437, |
| "learning_rate": 4.146881862754485e-06, |
| "loss": 0.0, |
| "step": 190 |
| }, |
| { |
| "epoch": 2.6527777777777777, |
| "grad_norm": 0.001035734312608838, |
| "learning_rate": 4.129010150788112e-06, |
| "loss": 0.0, |
| "step": 191 |
| }, |
| { |
| "epoch": 2.6666666666666665, |
| "grad_norm": 0.000686554005369544, |
| "learning_rate": 4.110992576162193e-06, |
| "loss": 0.0, |
| "step": 192 |
| }, |
| { |
| "epoch": 2.6805555555555554, |
| "grad_norm": 0.0006567820673808455, |
| "learning_rate": 4.092830752182423e-06, |
| "loss": 0.0, |
| "step": 193 |
| }, |
| { |
| "epoch": 2.6944444444444446, |
| "grad_norm": 0.0006689206347800791, |
| "learning_rate": 4.074526305070679e-06, |
| "loss": 0.0, |
| "step": 194 |
| }, |
| { |
| "epoch": 2.7083333333333335, |
| "grad_norm": 0.000665506231598556, |
| "learning_rate": 4.056080873819412e-06, |
| "loss": 0.0, |
| "step": 195 |
| }, |
| { |
| "epoch": 2.7222222222222223, |
| "grad_norm": 0.0006280328962020576, |
| "learning_rate": 4.037496110044885e-06, |
| "loss": 0.0, |
| "step": 196 |
| }, |
| { |
| "epoch": 2.736111111111111, |
| "grad_norm": 0.0005998269189149141, |
| "learning_rate": 4.018773677839289e-06, |
| "loss": 0.0, |
| "step": 197 |
| }, |
| { |
| "epoch": 2.75, |
| "grad_norm": 0.0007591253379359841, |
| "learning_rate": 3.999915253621739e-06, |
| "loss": 0.0, |
| "step": 198 |
| }, |
| { |
| "epoch": 2.763888888888889, |
| "grad_norm": 0.0005739334737882018, |
| "learning_rate": 3.980922525988167e-06, |
| "loss": 0.0, |
| "step": 199 |
| }, |
| { |
| "epoch": 2.7777777777777777, |
| "grad_norm": 0.0005984510062262416, |
| "learning_rate": 3.961797195560118e-06, |
| "loss": 0.0, |
| "step": 200 |
| }, |
| { |
| "epoch": 2.7916666666666665, |
| "grad_norm": 0.0007093317108228803, |
| "learning_rate": 3.942540974832486e-06, |
| "loss": 0.0, |
| "step": 201 |
| }, |
| { |
| "epoch": 2.8055555555555554, |
| "grad_norm": 0.0005686501390300691, |
| "learning_rate": 3.9231555880201655e-06, |
| "loss": 0.0, |
| "step": 202 |
| }, |
| { |
| "epoch": 2.8194444444444446, |
| "grad_norm": 0.0005396155174821615, |
| "learning_rate": 3.903642770903671e-06, |
| "loss": 0.0, |
| "step": 203 |
| }, |
| { |
| "epoch": 2.8333333333333335, |
| "grad_norm": 0.000549685733858496, |
| "learning_rate": 3.884004270673711e-06, |
| "loss": 0.0, |
| "step": 204 |
| }, |
| { |
| "epoch": 2.8472222222222223, |
| "grad_norm": 0.0005211409879848361, |
| "learning_rate": 3.864241845774746e-06, |
| "loss": 0.0, |
| "step": 205 |
| }, |
| { |
| "epoch": 2.861111111111111, |
| "grad_norm": 0.0005807566340081394, |
| "learning_rate": 3.844357265747531e-06, |
| "loss": 0.0, |
| "step": 206 |
| }, |
| { |
| "epoch": 2.875, |
| "grad_norm": 0.0005310736014507711, |
| "learning_rate": 3.8243523110706736e-06, |
| "loss": 0.0, |
| "step": 207 |
| }, |
| { |
| "epoch": 2.888888888888889, |
| "grad_norm": 0.001211474766023457, |
| "learning_rate": 3.8042287730012117e-06, |
| "loss": 0.0, |
| "step": 208 |
| }, |
| { |
| "epoch": 2.9027777777777777, |
| "grad_norm": 0.0005012262845411897, |
| "learning_rate": 3.7839884534142157e-06, |
| "loss": 0.0, |
| "step": 209 |
| }, |
| { |
| "epoch": 2.9166666666666665, |
| "grad_norm": 0.0005107534816488624, |
| "learning_rate": 3.7636331646414524e-06, |
| "loss": 0.0, |
| "step": 210 |
| }, |
| { |
| "epoch": 2.9305555555555554, |
| "grad_norm": 0.0005153871607035398, |
| "learning_rate": 3.7431647293091076e-06, |
| "loss": 0.0, |
| "step": 211 |
| }, |
| { |
| "epoch": 2.9444444444444446, |
| "grad_norm": 0.0005016613868065178, |
| "learning_rate": 3.7225849801745835e-06, |
| "loss": 0.0, |
| "step": 212 |
| }, |
| { |
| "epoch": 2.9583333333333335, |
| "grad_norm": 0.00048161958693526685, |
| "learning_rate": 3.701895759962397e-06, |
| "loss": 0.0, |
| "step": 213 |
| }, |
| { |
| "epoch": 2.9722222222222223, |
| "grad_norm": 0.0004826453805435449, |
| "learning_rate": 3.6810989211991777e-06, |
| "loss": 0.0, |
| "step": 214 |
| }, |
| { |
| "epoch": 2.986111111111111, |
| "grad_norm": 0.0004991457099094987, |
| "learning_rate": 3.6601963260477923e-06, |
| "loss": 0.0, |
| "step": 215 |
| }, |
| { |
| "epoch": 3.0, |
| "grad_norm": 0.0004765830235555768, |
| "learning_rate": 3.6391898461406045e-06, |
| "loss": 0.0, |
| "step": 216 |
| }, |
| { |
| "epoch": 3.013888888888889, |
| "grad_norm": 0.0004497554909903556, |
| "learning_rate": 3.6180813624118898e-06, |
| "loss": 0.0, |
| "step": 217 |
| }, |
| { |
| "epoch": 3.0277777777777777, |
| "grad_norm": 0.001278965501114726, |
| "learning_rate": 3.5968727649294134e-06, |
| "loss": 0.0, |
| "step": 218 |
| }, |
| { |
| "epoch": 3.0416666666666665, |
| "grad_norm": 0.00046711499453522265, |
| "learning_rate": 3.575565952725193e-06, |
| "loss": 0.0, |
| "step": 219 |
| }, |
| { |
| "epoch": 3.0555555555555554, |
| "grad_norm": 0.0004580412060022354, |
| "learning_rate": 3.55416283362546e-06, |
| "loss": 0.0, |
| "step": 220 |
| }, |
| { |
| "epoch": 3.0694444444444446, |
| "grad_norm": 0.0004473453445825726, |
| "learning_rate": 3.5326653240798283e-06, |
| "loss": 0.0, |
| "step": 221 |
| }, |
| { |
| "epoch": 3.0833333333333335, |
| "grad_norm": 0.00044585420982912183, |
| "learning_rate": 3.5110753489896924e-06, |
| "loss": 0.0, |
| "step": 222 |
| }, |
| { |
| "epoch": 3.0972222222222223, |
| "grad_norm": 0.00044625229202210903, |
| "learning_rate": 3.4893948415358803e-06, |
| "loss": 0.0, |
| "step": 223 |
| }, |
| { |
| "epoch": 3.111111111111111, |
| "grad_norm": 0.00042979326099157333, |
| "learning_rate": 3.4676257430055438e-06, |
| "loss": 0.0, |
| "step": 224 |
| }, |
| { |
| "epoch": 3.125, |
| "grad_norm": 0.0005280164768919349, |
| "learning_rate": 3.4457700026183378e-06, |
| "loss": 0.0, |
| "step": 225 |
| }, |
| { |
| "epoch": 3.138888888888889, |
| "grad_norm": 0.0004327085625845939, |
| "learning_rate": 3.4238295773518924e-06, |
| "loss": 0.0, |
| "step": 226 |
| }, |
| { |
| "epoch": 3.1527777777777777, |
| "grad_norm": 0.0004218154354020953, |
| "learning_rate": 3.4018064317665745e-06, |
| "loss": 0.0, |
| "step": 227 |
| }, |
| { |
| "epoch": 3.1666666666666665, |
| "grad_norm": 0.0004269857599865645, |
| "learning_rate": 3.3797025378295826e-06, |
| "loss": 0.0, |
| "step": 228 |
| }, |
| { |
| "epoch": 3.1805555555555554, |
| "grad_norm": 0.0004188821476418525, |
| "learning_rate": 3.357519874738382e-06, |
| "loss": 0.0, |
| "step": 229 |
| }, |
| { |
| "epoch": 3.1944444444444446, |
| "grad_norm": 0.00043003325117751956, |
| "learning_rate": 3.3352604287434752e-06, |
| "loss": 0.0, |
| "step": 230 |
| }, |
| { |
| "epoch": 3.2083333333333335, |
| "grad_norm": 0.00041776004945859313, |
| "learning_rate": 3.31292619297056e-06, |
| "loss": 0.0, |
| "step": 231 |
| }, |
| { |
| "epoch": 3.2222222222222223, |
| "grad_norm": 0.0004131250607315451, |
| "learning_rate": 3.29051916724206e-06, |
| "loss": 0.0, |
| "step": 232 |
| }, |
| { |
| "epoch": 3.236111111111111, |
| "grad_norm": 0.0004490792634896934, |
| "learning_rate": 3.2680413578980623e-06, |
| "loss": 0.0, |
| "step": 233 |
| }, |
| { |
| "epoch": 3.25, |
| "grad_norm": 0.00040265225106850266, |
| "learning_rate": 3.245494777616664e-06, |
| "loss": 0.0, |
| "step": 234 |
| }, |
| { |
| "epoch": 3.263888888888889, |
| "grad_norm": 0.0003899024159181863, |
| "learning_rate": 3.2228814452337587e-06, |
| "loss": 0.0, |
| "step": 235 |
| }, |
| { |
| "epoch": 3.2777777777777777, |
| "grad_norm": 0.00039753629243932664, |
| "learning_rate": 3.2002033855622683e-06, |
| "loss": 0.0, |
| "step": 236 |
| }, |
| { |
| "epoch": 3.2916666666666665, |
| "grad_norm": 0.00041745221824385226, |
| "learning_rate": 3.177462629210838e-06, |
| "loss": 0.0, |
| "step": 237 |
| }, |
| { |
| "epoch": 3.3055555555555554, |
| "grad_norm": 0.00037628141581080854, |
| "learning_rate": 3.154661212402017e-06, |
| "loss": 0.0, |
| "step": 238 |
| }, |
| { |
| "epoch": 3.3194444444444446, |
| "grad_norm": 0.0004110137524548918, |
| "learning_rate": 3.131801176789934e-06, |
| "loss": 0.0, |
| "step": 239 |
| }, |
| { |
| "epoch": 3.3333333333333335, |
| "grad_norm": 0.0003878215211443603, |
| "learning_rate": 3.1088845692774798e-06, |
| "loss": 0.0, |
| "step": 240 |
| }, |
| { |
| "epoch": 3.3472222222222223, |
| "grad_norm": 0.0004746608901768923, |
| "learning_rate": 3.0859134418330373e-06, |
| "loss": 0.0, |
| "step": 241 |
| }, |
| { |
| "epoch": 3.361111111111111, |
| "grad_norm": 0.0017198650166392326, |
| "learning_rate": 3.0628898513067357e-06, |
| "loss": 0.0, |
| "step": 242 |
| }, |
| { |
| "epoch": 3.375, |
| "grad_norm": 0.00038812440470792353, |
| "learning_rate": 3.0398158592462847e-06, |
| "loss": 0.0, |
| "step": 243 |
| }, |
| { |
| "epoch": 3.388888888888889, |
| "grad_norm": 0.0003960718458984047, |
| "learning_rate": 3.0166935317123824e-06, |
| "loss": 0.0, |
| "step": 244 |
| }, |
| { |
| "epoch": 3.4027777777777777, |
| "grad_norm": 0.0003934325650334358, |
| "learning_rate": 2.9935249390937184e-06, |
| "loss": 0.0, |
| "step": 245 |
| }, |
| { |
| "epoch": 3.4166666666666665, |
| "grad_norm": 0.0003809865447692573, |
| "learning_rate": 2.970312155921585e-06, |
| "loss": 0.0, |
| "step": 246 |
| }, |
| { |
| "epoch": 3.4305555555555554, |
| "grad_norm": 0.0003733540652319789, |
| "learning_rate": 2.9470572606841295e-06, |
| "loss": 0.0, |
| "step": 247 |
| }, |
| { |
| "epoch": 3.4444444444444446, |
| "grad_norm": 0.0007575592026114464, |
| "learning_rate": 2.9237623356402423e-06, |
| "loss": 0.0, |
| "step": 248 |
| }, |
| { |
| "epoch": 3.4583333333333335, |
| "grad_norm": 0.0003538963501341641, |
| "learning_rate": 2.900429466633107e-06, |
| "loss": 0.0, |
| "step": 249 |
| }, |
| { |
| "epoch": 3.4722222222222223, |
| "grad_norm": 0.00036024293513037264, |
| "learning_rate": 2.8770607429034352e-06, |
| "loss": 0.0, |
| "step": 250 |
| }, |
| { |
| "epoch": 3.486111111111111, |
| "grad_norm": 0.0003511797694955021, |
| "learning_rate": 2.8536582569023964e-06, |
| "loss": 0.0, |
| "step": 251 |
| }, |
| { |
| "epoch": 3.5, |
| "grad_norm": 0.00034679597592912614, |
| "learning_rate": 2.8302241041042564e-06, |
| "loss": 0.0, |
| "step": 252 |
| }, |
| { |
| "epoch": 3.513888888888889, |
| "grad_norm": 0.00037080422043800354, |
| "learning_rate": 2.8067603828187446e-06, |
| "loss": 0.0, |
| "step": 253 |
| }, |
| { |
| "epoch": 3.5277777777777777, |
| "grad_norm": 0.0003417868574615568, |
| "learning_rate": 2.7832691940031755e-06, |
| "loss": 0.0, |
| "step": 254 |
| }, |
| { |
| "epoch": 3.5416666666666665, |
| "grad_norm": 0.00034637749195098877, |
| "learning_rate": 2.759752641074322e-06, |
| "loss": 0.0, |
| "step": 255 |
| }, |
| { |
| "epoch": 3.5555555555555554, |
| "grad_norm": 0.00035772964474745095, |
| "learning_rate": 2.7362128297200784e-06, |
| "loss": 0.0, |
| "step": 256 |
| }, |
| { |
| "epoch": 3.5694444444444446, |
| "grad_norm": 0.00034110111300833523, |
| "learning_rate": 2.712651867710914e-06, |
| "loss": 0.0, |
| "step": 257 |
| }, |
| { |
| "epoch": 3.5833333333333335, |
| "grad_norm": 0.00033183529740199447, |
| "learning_rate": 2.6890718647111424e-06, |
| "loss": 0.0, |
| "step": 258 |
| }, |
| { |
| "epoch": 3.5972222222222223, |
| "grad_norm": 0.00034375995164737105, |
| "learning_rate": 2.665474932090017e-06, |
| "loss": 0.0, |
| "step": 259 |
| }, |
| { |
| "epoch": 3.611111111111111, |
| "grad_norm": 0.0003534927382133901, |
| "learning_rate": 2.6418631827326857e-06, |
| "loss": 0.0, |
| "step": 260 |
| }, |
| { |
| "epoch": 3.625, |
| "grad_norm": 0.00034480984322726727, |
| "learning_rate": 2.6182387308509927e-06, |
| "loss": 0.0, |
| "step": 261 |
| }, |
| { |
| "epoch": 3.638888888888889, |
| "grad_norm": 0.00034357619006186724, |
| "learning_rate": 2.5946036917941765e-06, |
| "loss": 0.0, |
| "step": 262 |
| }, |
| { |
| "epoch": 3.6527777777777777, |
| "grad_norm": 0.0003348125028423965, |
| "learning_rate": 2.570960181859458e-06, |
| "loss": 0.0, |
| "step": 263 |
| }, |
| { |
| "epoch": 3.6666666666666665, |
| "grad_norm": 0.0003396561951376498, |
| "learning_rate": 2.547310318102548e-06, |
| "loss": 0.0, |
| "step": 264 |
| }, |
| { |
| "epoch": 3.6805555555555554, |
| "grad_norm": 0.000327035115333274, |
| "learning_rate": 2.5236562181480794e-06, |
| "loss": 0.0, |
| "step": 265 |
| }, |
| { |
| "epoch": 3.6944444444444446, |
| "grad_norm": 0.0012853932566940784, |
| "learning_rate": 2.5e-06, |
| "loss": 0.0, |
| "step": 266 |
| }, |
| { |
| "epoch": 3.7083333333333335, |
| "grad_norm": 0.00033309016725979745, |
| "learning_rate": 2.4763437818519205e-06, |
| "loss": 0.0, |
| "step": 267 |
| }, |
| { |
| "epoch": 3.7222222222222223, |
| "grad_norm": 0.0003192981530446559, |
| "learning_rate": 2.4526896818974534e-06, |
| "loss": 0.0, |
| "step": 268 |
| }, |
| { |
| "epoch": 3.736111111111111, |
| "grad_norm": 0.00032640976132825017, |
| "learning_rate": 2.429039818140543e-06, |
| "loss": 0.0, |
| "step": 269 |
| }, |
| { |
| "epoch": 3.75, |
| "grad_norm": 0.00033392352634109557, |
| "learning_rate": 2.405396308205825e-06, |
| "loss": 0.0, |
| "step": 270 |
| }, |
| { |
| "epoch": 3.763888888888889, |
| "grad_norm": 0.0003594683366827667, |
| "learning_rate": 2.381761269149009e-06, |
| "loss": 0.0, |
| "step": 271 |
| }, |
| { |
| "epoch": 3.7777777777777777, |
| "grad_norm": 0.010392317548394203, |
| "learning_rate": 2.358136817267315e-06, |
| "loss": 0.0, |
| "step": 272 |
| }, |
| { |
| "epoch": 3.7916666666666665, |
| "grad_norm": 0.00031630549347028136, |
| "learning_rate": 2.334525067909983e-06, |
| "loss": 0.0, |
| "step": 273 |
| }, |
| { |
| "epoch": 3.8055555555555554, |
| "grad_norm": 0.0003196642210241407, |
| "learning_rate": 2.3109281352888593e-06, |
| "loss": 0.0, |
| "step": 274 |
| }, |
| { |
| "epoch": 3.8194444444444446, |
| "grad_norm": 0.00031541771022602916, |
| "learning_rate": 2.2873481322890866e-06, |
| "loss": 0.0, |
| "step": 275 |
| }, |
| { |
| "epoch": 3.8333333333333335, |
| "grad_norm": 0.0003151163400616497, |
| "learning_rate": 2.263787170279922e-06, |
| "loss": 0.0, |
| "step": 276 |
| }, |
| { |
| "epoch": 3.8472222222222223, |
| "grad_norm": 0.0003220181097276509, |
| "learning_rate": 2.2402473589256793e-06, |
| "loss": 0.0, |
| "step": 277 |
| }, |
| { |
| "epoch": 3.861111111111111, |
| "grad_norm": 0.00031665421556681395, |
| "learning_rate": 2.2167308059968258e-06, |
| "loss": 0.0, |
| "step": 278 |
| }, |
| { |
| "epoch": 3.875, |
| "grad_norm": 0.0003038264694623649, |
| "learning_rate": 2.193239617181256e-06, |
| "loss": 0.0, |
| "step": 279 |
| }, |
| { |
| "epoch": 3.888888888888889, |
| "grad_norm": 0.00031601826776750386, |
| "learning_rate": 2.169775895895745e-06, |
| "loss": 0.0, |
| "step": 280 |
| }, |
| { |
| "epoch": 3.9027777777777777, |
| "grad_norm": 0.0003112203849013895, |
| "learning_rate": 2.146341743097604e-06, |
| "loss": 0.0, |
| "step": 281 |
| }, |
| { |
| "epoch": 3.9166666666666665, |
| "grad_norm": 0.00030921402503736317, |
| "learning_rate": 2.1229392570965656e-06, |
| "loss": 0.0, |
| "step": 282 |
| }, |
| { |
| "epoch": 3.9305555555555554, |
| "grad_norm": 0.00031801871955394745, |
| "learning_rate": 2.0995705333668948e-06, |
| "loss": 0.0, |
| "step": 283 |
| }, |
| { |
| "epoch": 3.9444444444444446, |
| "grad_norm": 0.0003068145888391882, |
| "learning_rate": 2.0762376643597586e-06, |
| "loss": 0.0, |
| "step": 284 |
| }, |
| { |
| "epoch": 3.9583333333333335, |
| "grad_norm": 0.000303666980471462, |
| "learning_rate": 2.0529427393158704e-06, |
| "loss": 0.0, |
| "step": 285 |
| }, |
| { |
| "epoch": 3.9722222222222223, |
| "grad_norm": 0.00037687006988562644, |
| "learning_rate": 2.0296878440784164e-06, |
| "loss": 0.0, |
| "step": 286 |
| }, |
| { |
| "epoch": 3.986111111111111, |
| "grad_norm": 0.0004179875541012734, |
| "learning_rate": 2.006475060906283e-06, |
| "loss": 0.0, |
| "step": 287 |
| }, |
| { |
| "epoch": 4.0, |
| "grad_norm": 0.00030332262394949794, |
| "learning_rate": 1.9833064682876175e-06, |
| "loss": 0.0, |
| "step": 288 |
| }, |
| { |
| "epoch": 4.013888888888889, |
| "grad_norm": 0.000312824355205521, |
| "learning_rate": 1.9601841407537157e-06, |
| "loss": 0.0, |
| "step": 289 |
| }, |
| { |
| "epoch": 4.027777777777778, |
| "grad_norm": 0.0008900280809029937, |
| "learning_rate": 1.937110148693265e-06, |
| "loss": 0.0, |
| "step": 290 |
| }, |
| { |
| "epoch": 4.041666666666667, |
| "grad_norm": 0.00030201984918676317, |
| "learning_rate": 1.9140865581669627e-06, |
| "loss": 0.0, |
| "step": 291 |
| }, |
| { |
| "epoch": 4.055555555555555, |
| "grad_norm": 0.00037866891943849623, |
| "learning_rate": 1.8911154307225204e-06, |
| "loss": 0.0, |
| "step": 292 |
| }, |
| { |
| "epoch": 4.069444444444445, |
| "grad_norm": 0.0003057627473026514, |
| "learning_rate": 1.8681988232100674e-06, |
| "loss": 0.0, |
| "step": 293 |
| }, |
| { |
| "epoch": 4.083333333333333, |
| "grad_norm": 0.0003006605838891119, |
| "learning_rate": 1.8453387875979834e-06, |
| "loss": 0.0, |
| "step": 294 |
| }, |
| { |
| "epoch": 4.097222222222222, |
| "grad_norm": 0.0003039598523173481, |
| "learning_rate": 1.822537370789163e-06, |
| "loss": 0.0, |
| "step": 295 |
| }, |
| { |
| "epoch": 4.111111111111111, |
| "grad_norm": 0.00030075316317379475, |
| "learning_rate": 1.7997966144377328e-06, |
| "loss": 0.0, |
| "step": 296 |
| }, |
| { |
| "epoch": 4.125, |
| "grad_norm": 0.0005559928831644356, |
| "learning_rate": 1.7771185547662417e-06, |
| "loss": 0.0, |
| "step": 297 |
| }, |
| { |
| "epoch": 4.138888888888889, |
| "grad_norm": 0.0003315582871437073, |
| "learning_rate": 1.754505222383337e-06, |
| "loss": 0.0, |
| "step": 298 |
| }, |
| { |
| "epoch": 4.152777777777778, |
| "grad_norm": 0.00028943977667950094, |
| "learning_rate": 1.7319586421019383e-06, |
| "loss": 0.0, |
| "step": 299 |
| }, |
| { |
| "epoch": 4.166666666666667, |
| "grad_norm": 0.0003514060808811337, |
| "learning_rate": 1.7094808327579401e-06, |
| "loss": 0.0, |
| "step": 300 |
| }, |
| { |
| "epoch": 4.180555555555555, |
| "grad_norm": 0.00029012124286964536, |
| "learning_rate": 1.6870738070294412e-06, |
| "loss": 0.0, |
| "step": 301 |
| }, |
| { |
| "epoch": 4.194444444444445, |
| "grad_norm": 0.000285898830043152, |
| "learning_rate": 1.6647395712565256e-06, |
| "loss": 0.0, |
| "step": 302 |
| }, |
| { |
| "epoch": 4.208333333333333, |
| "grad_norm": 0.0002975537208840251, |
| "learning_rate": 1.6424801252616186e-06, |
| "loss": 0.0, |
| "step": 303 |
| }, |
| { |
| "epoch": 4.222222222222222, |
| "grad_norm": 0.00029446210828609765, |
| "learning_rate": 1.6202974621704176e-06, |
| "loss": 0.0, |
| "step": 304 |
| }, |
| { |
| "epoch": 4.236111111111111, |
| "grad_norm": 0.0002837205829564482, |
| "learning_rate": 1.5981935682334266e-06, |
| "loss": 0.0, |
| "step": 305 |
| }, |
| { |
| "epoch": 4.25, |
| "grad_norm": 0.0002935337251983583, |
| "learning_rate": 1.5761704226481078e-06, |
| "loss": 0.0, |
| "step": 306 |
| }, |
| { |
| "epoch": 4.263888888888889, |
| "grad_norm": 0.00029275196720846, |
| "learning_rate": 1.5542299973816626e-06, |
| "loss": 0.0, |
| "step": 307 |
| }, |
| { |
| "epoch": 4.277777777777778, |
| "grad_norm": 0.00029929872835054994, |
| "learning_rate": 1.5323742569944573e-06, |
| "loss": 0.0, |
| "step": 308 |
| }, |
| { |
| "epoch": 4.291666666666667, |
| "grad_norm": 0.00028991978615522385, |
| "learning_rate": 1.5106051584641208e-06, |
| "loss": 0.0, |
| "step": 309 |
| }, |
| { |
| "epoch": 4.305555555555555, |
| "grad_norm": 0.0002811246959026903, |
| "learning_rate": 1.4889246510103078e-06, |
| "loss": 0.0, |
| "step": 310 |
| }, |
| { |
| "epoch": 4.319444444444445, |
| "grad_norm": 0.00028082230710424483, |
| "learning_rate": 1.4673346759201728e-06, |
| "loss": 0.0, |
| "step": 311 |
| }, |
| { |
| "epoch": 4.333333333333333, |
| "grad_norm": 0.0002810833975672722, |
| "learning_rate": 1.44583716637454e-06, |
| "loss": 0.0, |
| "step": 312 |
| }, |
| { |
| "epoch": 4.347222222222222, |
| "grad_norm": 0.00030640949262306094, |
| "learning_rate": 1.4244340472748076e-06, |
| "loss": 0.0, |
| "step": 313 |
| }, |
| { |
| "epoch": 4.361111111111111, |
| "grad_norm": 0.00027884720475412905, |
| "learning_rate": 1.403127235070587e-06, |
| "loss": 0.0, |
| "step": 314 |
| }, |
| { |
| "epoch": 4.375, |
| "grad_norm": 0.00028082833159714937, |
| "learning_rate": 1.381918637588112e-06, |
| "loss": 0.0, |
| "step": 315 |
| }, |
| { |
| "epoch": 4.388888888888889, |
| "grad_norm": 0.00028363679302856326, |
| "learning_rate": 1.3608101538593965e-06, |
| "loss": 0.0, |
| "step": 316 |
| }, |
| { |
| "epoch": 4.402777777777778, |
| "grad_norm": 0.00027668694383464754, |
| "learning_rate": 1.3398036739522088e-06, |
| "loss": 0.0, |
| "step": 317 |
| }, |
| { |
| "epoch": 4.416666666666667, |
| "grad_norm": 0.00028155132895335555, |
| "learning_rate": 1.3189010788008234e-06, |
| "loss": 0.0, |
| "step": 318 |
| }, |
| { |
| "epoch": 4.430555555555555, |
| "grad_norm": 0.0007724284077994525, |
| "learning_rate": 1.2981042400376032e-06, |
| "loss": 0.0, |
| "step": 319 |
| }, |
| { |
| "epoch": 4.444444444444445, |
| "grad_norm": 0.00036324880784377456, |
| "learning_rate": 1.277415019825417e-06, |
| "loss": 0.0, |
| "step": 320 |
| }, |
| { |
| "epoch": 4.458333333333333, |
| "grad_norm": 0.00027311878511682153, |
| "learning_rate": 1.2568352706908937e-06, |
| "loss": 0.0, |
| "step": 321 |
| }, |
| { |
| "epoch": 4.472222222222222, |
| "grad_norm": 0.0002935140219051391, |
| "learning_rate": 1.2363668353585486e-06, |
| "loss": 0.0, |
| "step": 322 |
| }, |
| { |
| "epoch": 4.486111111111111, |
| "grad_norm": 0.0002762887452263385, |
| "learning_rate": 1.216011546585785e-06, |
| "loss": 0.0, |
| "step": 323 |
| }, |
| { |
| "epoch": 4.5, |
| "grad_norm": 0.00029656346305273473, |
| "learning_rate": 1.195771226998789e-06, |
| "loss": 0.0, |
| "step": 324 |
| }, |
| { |
| "epoch": 4.513888888888889, |
| "grad_norm": 0.0002801907539833337, |
| "learning_rate": 1.1756476889293269e-06, |
| "loss": 0.0, |
| "step": 325 |
| }, |
| { |
| "epoch": 4.527777777777778, |
| "grad_norm": 0.0002702914352994412, |
| "learning_rate": 1.1556427342524698e-06, |
| "loss": 0.0, |
| "step": 326 |
| }, |
| { |
| "epoch": 4.541666666666667, |
| "grad_norm": 0.0002727097016759217, |
| "learning_rate": 1.1357581542252555e-06, |
| "loss": 0.0, |
| "step": 327 |
| }, |
| { |
| "epoch": 4.555555555555555, |
| "grad_norm": 0.0002765726239886135, |
| "learning_rate": 1.1159957293262888e-06, |
| "loss": 0.0, |
| "step": 328 |
| }, |
| { |
| "epoch": 4.569444444444445, |
| "grad_norm": 0.0002687490778043866, |
| "learning_rate": 1.0963572290963298e-06, |
| "loss": 0.0, |
| "step": 329 |
| }, |
| { |
| "epoch": 4.583333333333333, |
| "grad_norm": 0.0002718923205975443, |
| "learning_rate": 1.0768444119798357e-06, |
| "loss": 0.0, |
| "step": 330 |
| }, |
| { |
| "epoch": 4.597222222222222, |
| "grad_norm": 0.0003026482881978154, |
| "learning_rate": 1.0574590251675145e-06, |
| "loss": 0.0, |
| "step": 331 |
| }, |
| { |
| "epoch": 4.611111111111111, |
| "grad_norm": 0.00027299264911562204, |
| "learning_rate": 1.0382028044398823e-06, |
| "loss": 0.0, |
| "step": 332 |
| }, |
| { |
| "epoch": 4.625, |
| "grad_norm": 0.00026989809703081846, |
| "learning_rate": 1.0190774740118343e-06, |
| "loss": 0.0, |
| "step": 333 |
| }, |
| { |
| "epoch": 4.638888888888889, |
| "grad_norm": 0.00028061502962373197, |
| "learning_rate": 1.0000847463782615e-06, |
| "loss": 0.0, |
| "step": 334 |
| }, |
| { |
| "epoch": 4.652777777777778, |
| "grad_norm": 0.0002712835557758808, |
| "learning_rate": 9.812263221607114e-07, |
| "loss": 0.0, |
| "step": 335 |
| }, |
| { |
| "epoch": 4.666666666666667, |
| "grad_norm": 0.0002652051334735006, |
| "learning_rate": 9.625038899551162e-07, |
| "loss": 0.0, |
| "step": 336 |
| }, |
| { |
| "epoch": 4.680555555555555, |
| "grad_norm": 0.00044494643225334585, |
| "learning_rate": 9.439191261805894e-07, |
| "loss": 0.0, |
| "step": 337 |
| }, |
| { |
| "epoch": 4.694444444444445, |
| "grad_norm": 0.00026834887103177607, |
| "learning_rate": 9.254736949293216e-07, |
| "loss": 0.0, |
| "step": 338 |
| }, |
| { |
| "epoch": 4.708333333333333, |
| "grad_norm": 0.00030554624390788376, |
| "learning_rate": 9.07169247817579e-07, |
| "loss": 0.0, |
| "step": 339 |
| }, |
| { |
| "epoch": 4.722222222222222, |
| "grad_norm": 0.00027393290656618774, |
| "learning_rate": 8.890074238378074e-07, |
| "loss": 0.0, |
| "step": 340 |
| }, |
| { |
| "epoch": 4.736111111111111, |
| "grad_norm": 0.0012748402077704668, |
| "learning_rate": 8.709898492118885e-07, |
| "loss": 0.0, |
| "step": 341 |
| }, |
| { |
| "epoch": 4.75, |
| "grad_norm": 0.0002683771017473191, |
| "learning_rate": 8.531181372455161e-07, |
| "loss": 0.0, |
| "step": 342 |
| }, |
| { |
| "epoch": 4.763888888888889, |
| "grad_norm": 0.0002686232910491526, |
| "learning_rate": 8.353938881837445e-07, |
| "loss": 0.0, |
| "step": 343 |
| }, |
| { |
| "epoch": 4.777777777777778, |
| "grad_norm": 0.0002642960462253541, |
| "learning_rate": 8.178186890677029e-07, |
| "loss": 0.0, |
| "step": 344 |
| }, |
| { |
| "epoch": 4.791666666666667, |
| "grad_norm": 0.00026234518736600876, |
| "learning_rate": 8.003941135924859e-07, |
| "loss": 0.0, |
| "step": 345 |
| }, |
| { |
| "epoch": 4.805555555555555, |
| "grad_norm": 0.0002638711594045162, |
| "learning_rate": 7.83121721966248e-07, |
| "loss": 0.0, |
| "step": 346 |
| }, |
| { |
| "epoch": 4.819444444444445, |
| "grad_norm": 0.00038043910171836615, |
| "learning_rate": 7.66003060770498e-07, |
| "loss": 0.0, |
| "step": 347 |
| }, |
| { |
| "epoch": 4.833333333333333, |
| "grad_norm": 0.00027849708567373455, |
| "learning_rate": 7.490396628216237e-07, |
| "loss": 0.0, |
| "step": 348 |
| }, |
| { |
| "epoch": 4.847222222222222, |
| "grad_norm": 0.00026442529633641243, |
| "learning_rate": 7.322330470336314e-07, |
| "loss": 0.0, |
| "step": 349 |
| }, |
| { |
| "epoch": 4.861111111111111, |
| "grad_norm": 0.0002640223829075694, |
| "learning_rate": 7.155847182821524e-07, |
| "loss": 0.0, |
| "step": 350 |
| }, |
| { |
| "epoch": 4.875, |
| "grad_norm": 0.0002679000608623028, |
| "learning_rate": 6.990961672696908e-07, |
| "loss": 0.0, |
| "step": 351 |
| }, |
| { |
| "epoch": 4.888888888888889, |
| "grad_norm": 0.0003060480812564492, |
| "learning_rate": 6.827688703921407e-07, |
| "loss": 0.0, |
| "step": 352 |
| }, |
| { |
| "epoch": 4.902777777777778, |
| "grad_norm": 0.0002725216909311712, |
| "learning_rate": 6.666042896065983e-07, |
| "loss": 0.0, |
| "step": 353 |
| }, |
| { |
| "epoch": 4.916666666666667, |
| "grad_norm": 0.0007128149154596031, |
| "learning_rate": 6.506038723004484e-07, |
| "loss": 0.0, |
| "step": 354 |
| }, |
| { |
| "epoch": 4.930555555555555, |
| "grad_norm": 0.0003282964462414384, |
| "learning_rate": 6.347690511617693e-07, |
| "loss": 0.0, |
| "step": 355 |
| }, |
| { |
| "epoch": 4.944444444444445, |
| "grad_norm": 0.00026580950361676514, |
| "learning_rate": 6.191012440510469e-07, |
| "loss": 0.0, |
| "step": 356 |
| }, |
| { |
| "epoch": 4.958333333333333, |
| "grad_norm": 0.0002644002088345587, |
| "learning_rate": 6.036018538742208e-07, |
| "loss": 0.0, |
| "step": 357 |
| }, |
| { |
| "epoch": 4.972222222222222, |
| "grad_norm": 0.0002604973560664803, |
| "learning_rate": 5.882722684570638e-07, |
| "loss": 0.0, |
| "step": 358 |
| }, |
| { |
| "epoch": 4.986111111111111, |
| "grad_norm": 0.0002634973789099604, |
| "learning_rate": 5.731138604209169e-07, |
| "loss": 0.0, |
| "step": 359 |
| }, |
| { |
| "epoch": 5.0, |
| "grad_norm": 0.00026600516866892576, |
| "learning_rate": 5.581279870597866e-07, |
| "loss": 0.0, |
| "step": 360 |
| } |
| ], |
| "logging_steps": 1, |
| "max_steps": 432, |
| "num_input_tokens_seen": 0, |
| "num_train_epochs": 6, |
| "save_steps": 72, |
| "stateful_callbacks": { |
| "TrainerControl": { |
| "args": { |
| "should_epoch_stop": false, |
| "should_evaluate": false, |
| "should_log": false, |
| "should_save": true, |
| "should_training_stop": false |
| }, |
| "attributes": {} |
| } |
| }, |
| "total_flos": 9.064261680208282e+17, |
| "train_batch_size": 4, |
| "trial_name": null, |
| "trial_params": null |
| } |
|
|
