| { | |
| "best_metric": 0.4546084403991699, | |
| "best_model_checkpoint": "miner_id_24/checkpoint-150", | |
| "epoch": 0.9493670886075949, | |
| "eval_steps": 50, | |
| "global_step": 150, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.006329113924050633, | |
| "grad_norm": 4.637372970581055, | |
| "learning_rate": 1e-05, | |
| "loss": 6.7527, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.006329113924050633, | |
| "eval_loss": 4.256635665893555, | |
| "eval_runtime": 18.7359, | |
| "eval_samples_per_second": 14.197, | |
| "eval_steps_per_second": 3.576, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.012658227848101266, | |
| "grad_norm": 5.933300495147705, | |
| "learning_rate": 2e-05, | |
| "loss": 7.104, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.0189873417721519, | |
| "grad_norm": 7.110599040985107, | |
| "learning_rate": 3e-05, | |
| "loss": 6.8969, | |
| "step": 3 | |
| }, | |
| { | |
| "epoch": 0.02531645569620253, | |
| "grad_norm": 6.989790439605713, | |
| "learning_rate": 4e-05, | |
| "loss": 7.0875, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.03164556962025317, | |
| "grad_norm": 6.984792232513428, | |
| "learning_rate": 5e-05, | |
| "loss": 6.447, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.0379746835443038, | |
| "grad_norm": 8.912130355834961, | |
| "learning_rate": 6e-05, | |
| "loss": 6.4936, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.04430379746835443, | |
| "grad_norm": 7.798403739929199, | |
| "learning_rate": 7e-05, | |
| "loss": 6.7064, | |
| "step": 7 | |
| }, | |
| { | |
| "epoch": 0.05063291139240506, | |
| "grad_norm": 13.288124084472656, | |
| "learning_rate": 8e-05, | |
| "loss": 6.162, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.056962025316455694, | |
| "grad_norm": 8.237154960632324, | |
| "learning_rate": 9e-05, | |
| "loss": 6.1715, | |
| "step": 9 | |
| }, | |
| { | |
| "epoch": 0.06329113924050633, | |
| "grad_norm": 13.2232084274292, | |
| "learning_rate": 0.0001, | |
| "loss": 5.6398, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.06962025316455696, | |
| "grad_norm": 15.054797172546387, | |
| "learning_rate": 9.999316524962345e-05, | |
| "loss": 4.7921, | |
| "step": 11 | |
| }, | |
| { | |
| "epoch": 0.0759493670886076, | |
| "grad_norm": 12.283551216125488, | |
| "learning_rate": 9.997266286704631e-05, | |
| "loss": 4.852, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 0.08227848101265822, | |
| "grad_norm": 13.324081420898438, | |
| "learning_rate": 9.993849845741524e-05, | |
| "loss": 5.4802, | |
| "step": 13 | |
| }, | |
| { | |
| "epoch": 0.08860759493670886, | |
| "grad_norm": 26.37136459350586, | |
| "learning_rate": 9.989068136093873e-05, | |
| "loss": 6.1261, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 0.0949367088607595, | |
| "grad_norm": 56.204917907714844, | |
| "learning_rate": 9.98292246503335e-05, | |
| "loss": 6.2994, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.10126582278481013, | |
| "grad_norm": 36.3787956237793, | |
| "learning_rate": 9.975414512725057e-05, | |
| "loss": 3.9331, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 0.10759493670886076, | |
| "grad_norm": 15.701330184936523, | |
| "learning_rate": 9.966546331768191e-05, | |
| "loss": 1.8937, | |
| "step": 17 | |
| }, | |
| { | |
| "epoch": 0.11392405063291139, | |
| "grad_norm": 50.6202278137207, | |
| "learning_rate": 9.956320346634876e-05, | |
| "loss": 1.8509, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 0.12025316455696203, | |
| "grad_norm": 16.660131454467773, | |
| "learning_rate": 9.944739353007344e-05, | |
| "loss": 0.5337, | |
| "step": 19 | |
| }, | |
| { | |
| "epoch": 0.12658227848101267, | |
| "grad_norm": 2.1026699542999268, | |
| "learning_rate": 9.931806517013612e-05, | |
| "loss": 0.0281, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.13291139240506328, | |
| "grad_norm": 0.2913537919521332, | |
| "learning_rate": 9.917525374361912e-05, | |
| "loss": 0.0026, | |
| "step": 21 | |
| }, | |
| { | |
| "epoch": 0.13924050632911392, | |
| "grad_norm": 6.9465227127075195, | |
| "learning_rate": 9.901899829374047e-05, | |
| "loss": 0.096, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 0.14556962025316456, | |
| "grad_norm": 0.027604248374700546, | |
| "learning_rate": 9.884934153917997e-05, | |
| "loss": 0.0002, | |
| "step": 23 | |
| }, | |
| { | |
| "epoch": 0.1518987341772152, | |
| "grad_norm": 0.011274458840489388, | |
| "learning_rate": 9.86663298624003e-05, | |
| "loss": 0.0001, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 0.15822784810126583, | |
| "grad_norm": 0.020102426409721375, | |
| "learning_rate": 9.847001329696653e-05, | |
| "loss": 0.0001, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.16455696202531644, | |
| "grad_norm": 0.1387815624475479, | |
| "learning_rate": 9.826044551386744e-05, | |
| "loss": 0.0003, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 0.17088607594936708, | |
| "grad_norm": 0.005320771597325802, | |
| "learning_rate": 9.803768380684242e-05, | |
| "loss": 0.0, | |
| "step": 27 | |
| }, | |
| { | |
| "epoch": 0.17721518987341772, | |
| "grad_norm": 2.64678692817688, | |
| "learning_rate": 9.780178907671789e-05, | |
| "loss": 0.0339, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 0.18354430379746836, | |
| "grad_norm": 0.0024521031882613897, | |
| "learning_rate": 9.755282581475769e-05, | |
| "loss": 0.0, | |
| "step": 29 | |
| }, | |
| { | |
| "epoch": 0.189873417721519, | |
| "grad_norm": 0.0020075368229299784, | |
| "learning_rate": 9.729086208503174e-05, | |
| "loss": 0.0, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.1962025316455696, | |
| "grad_norm": 0.038034241646528244, | |
| "learning_rate": 9.701596950580806e-05, | |
| "loss": 0.0, | |
| "step": 31 | |
| }, | |
| { | |
| "epoch": 0.20253164556962025, | |
| "grad_norm": 0.0014749664114788175, | |
| "learning_rate": 9.672822322997305e-05, | |
| "loss": 0.0, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 0.2088607594936709, | |
| "grad_norm": 0.0013801049208268523, | |
| "learning_rate": 9.642770192448536e-05, | |
| "loss": 0.0, | |
| "step": 33 | |
| }, | |
| { | |
| "epoch": 0.21518987341772153, | |
| "grad_norm": 0.0012297042412683368, | |
| "learning_rate": 9.611448774886924e-05, | |
| "loss": 0.0, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 0.22151898734177214, | |
| "grad_norm": 0.0012208245461806655, | |
| "learning_rate": 9.578866633275288e-05, | |
| "loss": 0.0, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.22784810126582278, | |
| "grad_norm": 2.8858587741851807, | |
| "learning_rate": 9.545032675245813e-05, | |
| "loss": 0.002, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 0.23417721518987342, | |
| "grad_norm": 0.01600363664329052, | |
| "learning_rate": 9.509956150664796e-05, | |
| "loss": 0.0, | |
| "step": 37 | |
| }, | |
| { | |
| "epoch": 0.24050632911392406, | |
| "grad_norm": 6.135977745056152, | |
| "learning_rate": 9.473646649103818e-05, | |
| "loss": 0.0043, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 0.2468354430379747, | |
| "grad_norm": 0.026988012716174126, | |
| "learning_rate": 9.43611409721806e-05, | |
| "loss": 0.0001, | |
| "step": 39 | |
| }, | |
| { | |
| "epoch": 0.25316455696202533, | |
| "grad_norm": 415.4832763671875, | |
| "learning_rate": 9.397368756032445e-05, | |
| "loss": 31.5935, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.25949367088607594, | |
| "grad_norm": 137.92080688476562, | |
| "learning_rate": 9.357421218136386e-05, | |
| "loss": 13.8607, | |
| "step": 41 | |
| }, | |
| { | |
| "epoch": 0.26582278481012656, | |
| "grad_norm": 24.936744689941406, | |
| "learning_rate": 9.316282404787871e-05, | |
| "loss": 7.9948, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 0.2721518987341772, | |
| "grad_norm": 12.365960121154785, | |
| "learning_rate": 9.273963562927695e-05, | |
| "loss": 6.1583, | |
| "step": 43 | |
| }, | |
| { | |
| "epoch": 0.27848101265822783, | |
| "grad_norm": 9.34447193145752, | |
| "learning_rate": 9.230476262104677e-05, | |
| "loss": 6.7741, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 0.2848101265822785, | |
| "grad_norm": 21.50917625427246, | |
| "learning_rate": 9.185832391312644e-05, | |
| "loss": 7.0601, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 0.2911392405063291, | |
| "grad_norm": 7.162813663482666, | |
| "learning_rate": 9.140044155740101e-05, | |
| "loss": 6.2468, | |
| "step": 46 | |
| }, | |
| { | |
| "epoch": 0.2974683544303797, | |
| "grad_norm": 7.348287105560303, | |
| "learning_rate": 9.093124073433463e-05, | |
| "loss": 6.0966, | |
| "step": 47 | |
| }, | |
| { | |
| "epoch": 0.3037974683544304, | |
| "grad_norm": 9.03297233581543, | |
| "learning_rate": 9.045084971874738e-05, | |
| "loss": 4.3763, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 0.310126582278481, | |
| "grad_norm": 6.385989189147949, | |
| "learning_rate": 8.995939984474624e-05, | |
| "loss": 4.5362, | |
| "step": 49 | |
| }, | |
| { | |
| "epoch": 0.31645569620253167, | |
| "grad_norm": 6.851682186126709, | |
| "learning_rate": 8.945702546981969e-05, | |
| "loss": 3.3954, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.31645569620253167, | |
| "eval_loss": 1.0085351467132568, | |
| "eval_runtime": 19.1811, | |
| "eval_samples_per_second": 13.868, | |
| "eval_steps_per_second": 3.493, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.3227848101265823, | |
| "grad_norm": 7.560333251953125, | |
| "learning_rate": 8.894386393810563e-05, | |
| "loss": 4.8001, | |
| "step": 51 | |
| }, | |
| { | |
| "epoch": 0.3291139240506329, | |
| "grad_norm": 9.081886291503906, | |
| "learning_rate": 8.842005554284296e-05, | |
| "loss": 4.9381, | |
| "step": 52 | |
| }, | |
| { | |
| "epoch": 0.33544303797468356, | |
| "grad_norm": 12.259449005126953, | |
| "learning_rate": 8.788574348801675e-05, | |
| "loss": 5.0075, | |
| "step": 53 | |
| }, | |
| { | |
| "epoch": 0.34177215189873417, | |
| "grad_norm": 19.61177635192871, | |
| "learning_rate": 8.73410738492077e-05, | |
| "loss": 3.1041, | |
| "step": 54 | |
| }, | |
| { | |
| "epoch": 0.34810126582278483, | |
| "grad_norm": 15.831315994262695, | |
| "learning_rate": 8.678619553365659e-05, | |
| "loss": 2.9224, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 0.35443037974683544, | |
| "grad_norm": 18.207788467407227, | |
| "learning_rate": 8.622126023955446e-05, | |
| "loss": 0.983, | |
| "step": 56 | |
| }, | |
| { | |
| "epoch": 0.36075949367088606, | |
| "grad_norm": 11.649006843566895, | |
| "learning_rate": 8.564642241456986e-05, | |
| "loss": 1.6834, | |
| "step": 57 | |
| }, | |
| { | |
| "epoch": 0.3670886075949367, | |
| "grad_norm": 4.370729923248291, | |
| "learning_rate": 8.506183921362443e-05, | |
| "loss": 0.2919, | |
| "step": 58 | |
| }, | |
| { | |
| "epoch": 0.37341772151898733, | |
| "grad_norm": 0.40553605556488037, | |
| "learning_rate": 8.44676704559283e-05, | |
| "loss": 0.0066, | |
| "step": 59 | |
| }, | |
| { | |
| "epoch": 0.379746835443038, | |
| "grad_norm": 3.1576056480407715, | |
| "learning_rate": 8.386407858128706e-05, | |
| "loss": 0.0229, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.3860759493670886, | |
| "grad_norm": 0.02823661081492901, | |
| "learning_rate": 8.32512286056924e-05, | |
| "loss": 0.0003, | |
| "step": 61 | |
| }, | |
| { | |
| "epoch": 0.3924050632911392, | |
| "grad_norm": 0.023632438853383064, | |
| "learning_rate": 8.262928807620843e-05, | |
| "loss": 0.0002, | |
| "step": 62 | |
| }, | |
| { | |
| "epoch": 0.3987341772151899, | |
| "grad_norm": 0.004458832088857889, | |
| "learning_rate": 8.199842702516583e-05, | |
| "loss": 0.0, | |
| "step": 63 | |
| }, | |
| { | |
| "epoch": 0.4050632911392405, | |
| "grad_norm": 0.0025968514382839203, | |
| "learning_rate": 8.135881792367686e-05, | |
| "loss": 0.0, | |
| "step": 64 | |
| }, | |
| { | |
| "epoch": 0.41139240506329117, | |
| "grad_norm": 0.002332580741494894, | |
| "learning_rate": 8.07106356344834e-05, | |
| "loss": 0.0, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 0.4177215189873418, | |
| "grad_norm": 68.49555969238281, | |
| "learning_rate": 8.005405736415126e-05, | |
| "loss": 0.0764, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 0.4240506329113924, | |
| "grad_norm": 0.007944638840854168, | |
| "learning_rate": 7.938926261462366e-05, | |
| "loss": 0.0001, | |
| "step": 67 | |
| }, | |
| { | |
| "epoch": 0.43037974683544306, | |
| "grad_norm": 0.01654389128088951, | |
| "learning_rate": 7.871643313414718e-05, | |
| "loss": 0.0001, | |
| "step": 68 | |
| }, | |
| { | |
| "epoch": 0.43670886075949367, | |
| "grad_norm": 0.016359442844986916, | |
| "learning_rate": 7.803575286758364e-05, | |
| "loss": 0.0001, | |
| "step": 69 | |
| }, | |
| { | |
| "epoch": 0.4430379746835443, | |
| "grad_norm": 0.009918966330587864, | |
| "learning_rate": 7.734740790612136e-05, | |
| "loss": 0.0001, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.44936708860759494, | |
| "grad_norm": 0.006231650244444609, | |
| "learning_rate": 7.66515864363997e-05, | |
| "loss": 0.0, | |
| "step": 71 | |
| }, | |
| { | |
| "epoch": 0.45569620253164556, | |
| "grad_norm": 0.08754861354827881, | |
| "learning_rate": 7.594847868906076e-05, | |
| "loss": 0.0001, | |
| "step": 72 | |
| }, | |
| { | |
| "epoch": 0.4620253164556962, | |
| "grad_norm": 0.005205010063946247, | |
| "learning_rate": 7.52382768867422e-05, | |
| "loss": 0.0, | |
| "step": 73 | |
| }, | |
| { | |
| "epoch": 0.46835443037974683, | |
| "grad_norm": 0.0037575392052531242, | |
| "learning_rate": 7.452117519152542e-05, | |
| "loss": 0.0, | |
| "step": 74 | |
| }, | |
| { | |
| "epoch": 0.47468354430379744, | |
| "grad_norm": 0.0029411595314741135, | |
| "learning_rate": 7.379736965185368e-05, | |
| "loss": 0.0, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.4810126582278481, | |
| "grad_norm": 0.0023681563325226307, | |
| "learning_rate": 7.30670581489344e-05, | |
| "loss": 0.0, | |
| "step": 76 | |
| }, | |
| { | |
| "epoch": 0.4873417721518987, | |
| "grad_norm": 0.0018834106158465147, | |
| "learning_rate": 7.233044034264034e-05, | |
| "loss": 0.0, | |
| "step": 77 | |
| }, | |
| { | |
| "epoch": 0.4936708860759494, | |
| "grad_norm": 0.0017286180518567562, | |
| "learning_rate": 7.158771761692464e-05, | |
| "loss": 0.0, | |
| "step": 78 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "grad_norm": 247.8935546875, | |
| "learning_rate": 7.083909302476453e-05, | |
| "loss": 21.2202, | |
| "step": 79 | |
| }, | |
| { | |
| "epoch": 0.5063291139240507, | |
| "grad_norm": 99.23827362060547, | |
| "learning_rate": 7.008477123264848e-05, | |
| "loss": 13.1565, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.5126582278481012, | |
| "grad_norm": 22.70919418334961, | |
| "learning_rate": 6.932495846462261e-05, | |
| "loss": 8.877, | |
| "step": 81 | |
| }, | |
| { | |
| "epoch": 0.5189873417721519, | |
| "grad_norm": 12.581419944763184, | |
| "learning_rate": 6.855986244591104e-05, | |
| "loss": 5.779, | |
| "step": 82 | |
| }, | |
| { | |
| "epoch": 0.5253164556962026, | |
| "grad_norm": 6.799219131469727, | |
| "learning_rate": 6.778969234612584e-05, | |
| "loss": 6.6994, | |
| "step": 83 | |
| }, | |
| { | |
| "epoch": 0.5316455696202531, | |
| "grad_norm": 5.687674522399902, | |
| "learning_rate": 6.701465872208216e-05, | |
| "loss": 6.7271, | |
| "step": 84 | |
| }, | |
| { | |
| "epoch": 0.5379746835443038, | |
| "grad_norm": 7.574744701385498, | |
| "learning_rate": 6.623497346023418e-05, | |
| "loss": 5.9832, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 0.5443037974683544, | |
| "grad_norm": 6.770690441131592, | |
| "learning_rate": 6.545084971874738e-05, | |
| "loss": 5.3349, | |
| "step": 86 | |
| }, | |
| { | |
| "epoch": 0.5506329113924051, | |
| "grad_norm": 6.103719234466553, | |
| "learning_rate": 6.466250186922325e-05, | |
| "loss": 3.7193, | |
| "step": 87 | |
| }, | |
| { | |
| "epoch": 0.5569620253164557, | |
| "grad_norm": 6.168741226196289, | |
| "learning_rate": 6.387014543809223e-05, | |
| "loss": 4.9696, | |
| "step": 88 | |
| }, | |
| { | |
| "epoch": 0.5632911392405063, | |
| "grad_norm": 5.017352104187012, | |
| "learning_rate": 6.307399704769099e-05, | |
| "loss": 2.3407, | |
| "step": 89 | |
| }, | |
| { | |
| "epoch": 0.569620253164557, | |
| "grad_norm": 6.256300449371338, | |
| "learning_rate": 6.227427435703997e-05, | |
| "loss": 4.3512, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.5759493670886076, | |
| "grad_norm": 7.3730974197387695, | |
| "learning_rate": 6.147119600233758e-05, | |
| "loss": 3.9607, | |
| "step": 91 | |
| }, | |
| { | |
| "epoch": 0.5822784810126582, | |
| "grad_norm": 9.467050552368164, | |
| "learning_rate": 6.066498153718735e-05, | |
| "loss": 4.4029, | |
| "step": 92 | |
| }, | |
| { | |
| "epoch": 0.5886075949367089, | |
| "grad_norm": 9.84329605102539, | |
| "learning_rate": 5.985585137257401e-05, | |
| "loss": 1.7972, | |
| "step": 93 | |
| }, | |
| { | |
| "epoch": 0.5949367088607594, | |
| "grad_norm": 22.693742752075195, | |
| "learning_rate": 5.90440267166055e-05, | |
| "loss": 1.5681, | |
| "step": 94 | |
| }, | |
| { | |
| "epoch": 0.6012658227848101, | |
| "grad_norm": 13.892134666442871, | |
| "learning_rate": 5.8229729514036705e-05, | |
| "loss": 1.5012, | |
| "step": 95 | |
| }, | |
| { | |
| "epoch": 0.6075949367088608, | |
| "grad_norm": 11.440923690795898, | |
| "learning_rate": 5.74131823855921e-05, | |
| "loss": 0.4601, | |
| "step": 96 | |
| }, | |
| { | |
| "epoch": 0.6139240506329114, | |
| "grad_norm": 4.429217338562012, | |
| "learning_rate": 5.6594608567103456e-05, | |
| "loss": 0.4429, | |
| "step": 97 | |
| }, | |
| { | |
| "epoch": 0.620253164556962, | |
| "grad_norm": 8.719744682312012, | |
| "learning_rate": 5.577423184847932e-05, | |
| "loss": 1.5709, | |
| "step": 98 | |
| }, | |
| { | |
| "epoch": 0.6265822784810127, | |
| "grad_norm": 2.6932637691497803, | |
| "learning_rate": 5.495227651252315e-05, | |
| "loss": 0.0642, | |
| "step": 99 | |
| }, | |
| { | |
| "epoch": 0.6329113924050633, | |
| "grad_norm": 0.5408775806427002, | |
| "learning_rate": 5.4128967273616625e-05, | |
| "loss": 0.005, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.6329113924050633, | |
| "eval_loss": 0.48234593868255615, | |
| "eval_runtime": 19.1677, | |
| "eval_samples_per_second": 13.878, | |
| "eval_steps_per_second": 3.495, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.6392405063291139, | |
| "grad_norm": 7.4709858894348145, | |
| "learning_rate": 5.330452921628497e-05, | |
| "loss": 0.0346, | |
| "step": 101 | |
| }, | |
| { | |
| "epoch": 0.6455696202531646, | |
| "grad_norm": 0.05150521174073219, | |
| "learning_rate": 5.247918773366112e-05, | |
| "loss": 0.0008, | |
| "step": 102 | |
| }, | |
| { | |
| "epoch": 0.6518987341772152, | |
| "grad_norm": 0.018166132271289825, | |
| "learning_rate": 5.165316846586541e-05, | |
| "loss": 0.0003, | |
| "step": 103 | |
| }, | |
| { | |
| "epoch": 0.6582278481012658, | |
| "grad_norm": 0.00859126914292574, | |
| "learning_rate": 5.0826697238317935e-05, | |
| "loss": 0.0001, | |
| "step": 104 | |
| }, | |
| { | |
| "epoch": 0.6645569620253164, | |
| "grad_norm": 0.004140320234000683, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0001, | |
| "step": 105 | |
| }, | |
| { | |
| "epoch": 0.6708860759493671, | |
| "grad_norm": 0.007025420665740967, | |
| "learning_rate": 4.917330276168208e-05, | |
| "loss": 0.0001, | |
| "step": 106 | |
| }, | |
| { | |
| "epoch": 0.6772151898734177, | |
| "grad_norm": 0.0018030598293989897, | |
| "learning_rate": 4.834683153413459e-05, | |
| "loss": 0.0, | |
| "step": 107 | |
| }, | |
| { | |
| "epoch": 0.6835443037974683, | |
| "grad_norm": 0.0015602059429511428, | |
| "learning_rate": 4.7520812266338885e-05, | |
| "loss": 0.0, | |
| "step": 108 | |
| }, | |
| { | |
| "epoch": 0.689873417721519, | |
| "grad_norm": 0.0018193755531683564, | |
| "learning_rate": 4.669547078371504e-05, | |
| "loss": 0.0, | |
| "step": 109 | |
| }, | |
| { | |
| "epoch": 0.6962025316455697, | |
| "grad_norm": 0.001228783163242042, | |
| "learning_rate": 4.5871032726383386e-05, | |
| "loss": 0.0, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.7025316455696202, | |
| "grad_norm": 0.05244214087724686, | |
| "learning_rate": 4.504772348747687e-05, | |
| "loss": 0.0002, | |
| "step": 111 | |
| }, | |
| { | |
| "epoch": 0.7088607594936709, | |
| "grad_norm": 0.001256680698134005, | |
| "learning_rate": 4.4225768151520694e-05, | |
| "loss": 0.0, | |
| "step": 112 | |
| }, | |
| { | |
| "epoch": 0.7151898734177216, | |
| "grad_norm": 0.0010616511572152376, | |
| "learning_rate": 4.3405391432896555e-05, | |
| "loss": 0.0, | |
| "step": 113 | |
| }, | |
| { | |
| "epoch": 0.7215189873417721, | |
| "grad_norm": 2.2404611110687256, | |
| "learning_rate": 4.2586817614407895e-05, | |
| "loss": 0.0087, | |
| "step": 114 | |
| }, | |
| { | |
| "epoch": 0.7278481012658228, | |
| "grad_norm": 0.0010911829303950071, | |
| "learning_rate": 4.17702704859633e-05, | |
| "loss": 0.0, | |
| "step": 115 | |
| }, | |
| { | |
| "epoch": 0.7341772151898734, | |
| "grad_norm": 0.0011822691885754466, | |
| "learning_rate": 4.095597328339452e-05, | |
| "loss": 0.0, | |
| "step": 116 | |
| }, | |
| { | |
| "epoch": 0.740506329113924, | |
| "grad_norm": 1.555167555809021, | |
| "learning_rate": 4.0144148627425993e-05, | |
| "loss": 0.0048, | |
| "step": 117 | |
| }, | |
| { | |
| "epoch": 0.7468354430379747, | |
| "grad_norm": 55.501468658447266, | |
| "learning_rate": 3.933501846281267e-05, | |
| "loss": 8.1764, | |
| "step": 118 | |
| }, | |
| { | |
| "epoch": 0.7531645569620253, | |
| "grad_norm": 13.718461036682129, | |
| "learning_rate": 3.852880399766243e-05, | |
| "loss": 8.2124, | |
| "step": 119 | |
| }, | |
| { | |
| "epoch": 0.759493670886076, | |
| "grad_norm": 12.956205368041992, | |
| "learning_rate": 3.772572564296005e-05, | |
| "loss": 7.4217, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.7658227848101266, | |
| "grad_norm": 8.396835327148438, | |
| "learning_rate": 3.6926002952309016e-05, | |
| "loss": 6.8848, | |
| "step": 121 | |
| }, | |
| { | |
| "epoch": 0.7721518987341772, | |
| "grad_norm": 8.19432544708252, | |
| "learning_rate": 3.612985456190778e-05, | |
| "loss": 6.1293, | |
| "step": 122 | |
| }, | |
| { | |
| "epoch": 0.7784810126582279, | |
| "grad_norm": 7.240010738372803, | |
| "learning_rate": 3.533749813077677e-05, | |
| "loss": 6.3102, | |
| "step": 123 | |
| }, | |
| { | |
| "epoch": 0.7848101265822784, | |
| "grad_norm": 6.158444881439209, | |
| "learning_rate": 3.4549150281252636e-05, | |
| "loss": 5.9369, | |
| "step": 124 | |
| }, | |
| { | |
| "epoch": 0.7911392405063291, | |
| "grad_norm": 6.622616291046143, | |
| "learning_rate": 3.3765026539765834e-05, | |
| "loss": 4.0281, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 0.7974683544303798, | |
| "grad_norm": 7.383076190948486, | |
| "learning_rate": 3.298534127791785e-05, | |
| "loss": 3.9944, | |
| "step": 126 | |
| }, | |
| { | |
| "epoch": 0.8037974683544303, | |
| "grad_norm": 7.300906181335449, | |
| "learning_rate": 3.221030765387417e-05, | |
| "loss": 3.5384, | |
| "step": 127 | |
| }, | |
| { | |
| "epoch": 0.810126582278481, | |
| "grad_norm": 7.340206623077393, | |
| "learning_rate": 3.144013755408895e-05, | |
| "loss": 4.7221, | |
| "step": 128 | |
| }, | |
| { | |
| "epoch": 0.8164556962025317, | |
| "grad_norm": 8.666287422180176, | |
| "learning_rate": 3.0675041535377405e-05, | |
| "loss": 4.6958, | |
| "step": 129 | |
| }, | |
| { | |
| "epoch": 0.8227848101265823, | |
| "grad_norm": 10.738088607788086, | |
| "learning_rate": 2.991522876735154e-05, | |
| "loss": 4.1066, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.8291139240506329, | |
| "grad_norm": 9.52358341217041, | |
| "learning_rate": 2.916090697523549e-05, | |
| "loss": 1.5994, | |
| "step": 131 | |
| }, | |
| { | |
| "epoch": 0.8354430379746836, | |
| "grad_norm": 14.979015350341797, | |
| "learning_rate": 2.8412282383075363e-05, | |
| "loss": 1.3601, | |
| "step": 132 | |
| }, | |
| { | |
| "epoch": 0.8417721518987342, | |
| "grad_norm": 9.825060844421387, | |
| "learning_rate": 2.766955965735968e-05, | |
| "loss": 0.4732, | |
| "step": 133 | |
| }, | |
| { | |
| "epoch": 0.8481012658227848, | |
| "grad_norm": 6.621682167053223, | |
| "learning_rate": 2.693294185106562e-05, | |
| "loss": 0.1295, | |
| "step": 134 | |
| }, | |
| { | |
| "epoch": 0.8544303797468354, | |
| "grad_norm": 4.410978317260742, | |
| "learning_rate": 2.6202630348146324e-05, | |
| "loss": 0.3385, | |
| "step": 135 | |
| }, | |
| { | |
| "epoch": 0.8607594936708861, | |
| "grad_norm": 0.5501710772514343, | |
| "learning_rate": 2.547882480847461e-05, | |
| "loss": 0.0079, | |
| "step": 136 | |
| }, | |
| { | |
| "epoch": 0.8670886075949367, | |
| "grad_norm": 0.6479581594467163, | |
| "learning_rate": 2.476172311325783e-05, | |
| "loss": 0.0063, | |
| "step": 137 | |
| }, | |
| { | |
| "epoch": 0.8734177215189873, | |
| "grad_norm": 0.07755237072706223, | |
| "learning_rate": 2.405152131093926e-05, | |
| "loss": 0.0011, | |
| "step": 138 | |
| }, | |
| { | |
| "epoch": 0.879746835443038, | |
| "grad_norm": 0.22078044712543488, | |
| "learning_rate": 2.3348413563600325e-05, | |
| "loss": 0.0018, | |
| "step": 139 | |
| }, | |
| { | |
| "epoch": 0.8860759493670886, | |
| "grad_norm": 2.223050117492676, | |
| "learning_rate": 2.2652592093878666e-05, | |
| "loss": 0.0105, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.8924050632911392, | |
| "grad_norm": 0.03765476495027542, | |
| "learning_rate": 2.196424713241637e-05, | |
| "loss": 0.0004, | |
| "step": 141 | |
| }, | |
| { | |
| "epoch": 0.8987341772151899, | |
| "grad_norm": 0.029489785432815552, | |
| "learning_rate": 2.128356686585282e-05, | |
| "loss": 0.0003, | |
| "step": 142 | |
| }, | |
| { | |
| "epoch": 0.9050632911392406, | |
| "grad_norm": 4.499859809875488, | |
| "learning_rate": 2.061073738537635e-05, | |
| "loss": 0.6094, | |
| "step": 143 | |
| }, | |
| { | |
| "epoch": 0.9113924050632911, | |
| "grad_norm": 0.031161252409219742, | |
| "learning_rate": 1.9945942635848748e-05, | |
| "loss": 0.0003, | |
| "step": 144 | |
| }, | |
| { | |
| "epoch": 0.9177215189873418, | |
| "grad_norm": 0.031003016978502274, | |
| "learning_rate": 1.928936436551661e-05, | |
| "loss": 0.0003, | |
| "step": 145 | |
| }, | |
| { | |
| "epoch": 0.9240506329113924, | |
| "grad_norm": 3.472163200378418, | |
| "learning_rate": 1.8641182076323148e-05, | |
| "loss": 0.0947, | |
| "step": 146 | |
| }, | |
| { | |
| "epoch": 0.930379746835443, | |
| "grad_norm": 3.906115770339966, | |
| "learning_rate": 1.800157297483417e-05, | |
| "loss": 0.0425, | |
| "step": 147 | |
| }, | |
| { | |
| "epoch": 0.9367088607594937, | |
| "grad_norm": 0.037830282002687454, | |
| "learning_rate": 1.7370711923791567e-05, | |
| "loss": 0.0003, | |
| "step": 148 | |
| }, | |
| { | |
| "epoch": 0.9430379746835443, | |
| "grad_norm": 2.2370946407318115, | |
| "learning_rate": 1.6748771394307585e-05, | |
| "loss": 0.0052, | |
| "step": 149 | |
| }, | |
| { | |
| "epoch": 0.9493670886075949, | |
| "grad_norm": 0.03683227300643921, | |
| "learning_rate": 1.6135921418712956e-05, | |
| "loss": 0.0003, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.9493670886075949, | |
| "eval_loss": 0.4546084403991699, | |
| "eval_runtime": 19.1762, | |
| "eval_samples_per_second": 13.871, | |
| "eval_steps_per_second": 3.494, | |
| "step": 150 | |
| } | |
| ], | |
| "logging_steps": 1, | |
| "max_steps": 200, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 2, | |
| "save_steps": 50, | |
| "stateful_callbacks": { | |
| "EarlyStoppingCallback": { | |
| "args": { | |
| "early_stopping_patience": 5, | |
| "early_stopping_threshold": 0.0 | |
| }, | |
| "attributes": { | |
| "early_stopping_patience_counter": 0 | |
| } | |
| }, | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 2.1608069935713485e+17, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |