| { | |
| "best_metric": 0.7523689866065979, | |
| "best_model_checkpoint": "miner_id_24/checkpoint-150", | |
| "epoch": 0.06839945280437756, | |
| "eval_steps": 50, | |
| "global_step": 150, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.00045599635202918376, | |
| "grad_norm": 35.74802780151367, | |
| "learning_rate": 1e-05, | |
| "loss": 5.441, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.00045599635202918376, | |
| "eval_loss": 1.8368788957595825, | |
| "eval_runtime": 347.1098, | |
| "eval_samples_per_second": 10.642, | |
| "eval_steps_per_second": 2.662, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.0009119927040583675, | |
| "grad_norm": 32.76350784301758, | |
| "learning_rate": 2e-05, | |
| "loss": 5.4115, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.0013679890560875513, | |
| "grad_norm": 25.53909683227539, | |
| "learning_rate": 3e-05, | |
| "loss": 4.0681, | |
| "step": 3 | |
| }, | |
| { | |
| "epoch": 0.001823985408116735, | |
| "grad_norm": 25.518535614013672, | |
| "learning_rate": 4e-05, | |
| "loss": 4.3628, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.002279981760145919, | |
| "grad_norm": 21.8685302734375, | |
| "learning_rate": 5e-05, | |
| "loss": 4.634, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.0027359781121751026, | |
| "grad_norm": 22.06576156616211, | |
| "learning_rate": 6e-05, | |
| "loss": 4.6123, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.0031919744642042863, | |
| "grad_norm": 23.909852981567383, | |
| "learning_rate": 7e-05, | |
| "loss": 4.5423, | |
| "step": 7 | |
| }, | |
| { | |
| "epoch": 0.00364797081623347, | |
| "grad_norm": 19.35548973083496, | |
| "learning_rate": 8e-05, | |
| "loss": 3.6147, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.004103967168262654, | |
| "grad_norm": 23.741609573364258, | |
| "learning_rate": 9e-05, | |
| "loss": 4.2225, | |
| "step": 9 | |
| }, | |
| { | |
| "epoch": 0.004559963520291838, | |
| "grad_norm": 19.457712173461914, | |
| "learning_rate": 0.0001, | |
| "loss": 4.4656, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.005015959872321021, | |
| "grad_norm": 18.08391571044922, | |
| "learning_rate": 9.999316524962345e-05, | |
| "loss": 3.5505, | |
| "step": 11 | |
| }, | |
| { | |
| "epoch": 0.005471956224350205, | |
| "grad_norm": 19.420406341552734, | |
| "learning_rate": 9.997266286704631e-05, | |
| "loss": 4.1672, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 0.005927952576379389, | |
| "grad_norm": 17.068937301635742, | |
| "learning_rate": 9.993849845741524e-05, | |
| "loss": 3.8395, | |
| "step": 13 | |
| }, | |
| { | |
| "epoch": 0.006383948928408573, | |
| "grad_norm": 16.270374298095703, | |
| "learning_rate": 9.989068136093873e-05, | |
| "loss": 3.8238, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 0.006839945280437756, | |
| "grad_norm": 15.42528247833252, | |
| "learning_rate": 9.98292246503335e-05, | |
| "loss": 3.545, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.00729594163246694, | |
| "grad_norm": 17.857736587524414, | |
| "learning_rate": 9.975414512725057e-05, | |
| "loss": 4.4884, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 0.007751937984496124, | |
| "grad_norm": 13.447070121765137, | |
| "learning_rate": 9.966546331768191e-05, | |
| "loss": 3.215, | |
| "step": 17 | |
| }, | |
| { | |
| "epoch": 0.008207934336525308, | |
| "grad_norm": 15.726587295532227, | |
| "learning_rate": 9.956320346634876e-05, | |
| "loss": 3.7731, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 0.008663930688554491, | |
| "grad_norm": 13.71452808380127, | |
| "learning_rate": 9.944739353007344e-05, | |
| "loss": 3.583, | |
| "step": 19 | |
| }, | |
| { | |
| "epoch": 0.009119927040583675, | |
| "grad_norm": 16.153724670410156, | |
| "learning_rate": 9.931806517013612e-05, | |
| "loss": 3.7108, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.009575923392612859, | |
| "grad_norm": 16.07076072692871, | |
| "learning_rate": 9.917525374361912e-05, | |
| "loss": 3.871, | |
| "step": 21 | |
| }, | |
| { | |
| "epoch": 0.010031919744642043, | |
| "grad_norm": 15.137386322021484, | |
| "learning_rate": 9.901899829374047e-05, | |
| "loss": 4.0925, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 0.010487916096671226, | |
| "grad_norm": 13.691039085388184, | |
| "learning_rate": 9.884934153917997e-05, | |
| "loss": 3.8998, | |
| "step": 23 | |
| }, | |
| { | |
| "epoch": 0.01094391244870041, | |
| "grad_norm": 19.00994300842285, | |
| "learning_rate": 9.86663298624003e-05, | |
| "loss": 4.4558, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 0.011399908800729594, | |
| "grad_norm": 15.00587272644043, | |
| "learning_rate": 9.847001329696653e-05, | |
| "loss": 3.7661, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.011855905152758778, | |
| "grad_norm": 14.253700256347656, | |
| "learning_rate": 9.826044551386744e-05, | |
| "loss": 3.981, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 0.012311901504787962, | |
| "grad_norm": 14.619214057922363, | |
| "learning_rate": 9.803768380684242e-05, | |
| "loss": 4.2931, | |
| "step": 27 | |
| }, | |
| { | |
| "epoch": 0.012767897856817145, | |
| "grad_norm": 13.213558197021484, | |
| "learning_rate": 9.780178907671789e-05, | |
| "loss": 4.111, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 0.013223894208846329, | |
| "grad_norm": 13.017401695251465, | |
| "learning_rate": 9.755282581475769e-05, | |
| "loss": 4.2164, | |
| "step": 29 | |
| }, | |
| { | |
| "epoch": 0.013679890560875513, | |
| "grad_norm": 11.902800559997559, | |
| "learning_rate": 9.729086208503174e-05, | |
| "loss": 3.7145, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.014135886912904697, | |
| "grad_norm": 13.412627220153809, | |
| "learning_rate": 9.701596950580806e-05, | |
| "loss": 3.7238, | |
| "step": 31 | |
| }, | |
| { | |
| "epoch": 0.01459188326493388, | |
| "grad_norm": 13.361820220947266, | |
| "learning_rate": 9.672822322997305e-05, | |
| "loss": 4.3335, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 0.015047879616963064, | |
| "grad_norm": 13.85506534576416, | |
| "learning_rate": 9.642770192448536e-05, | |
| "loss": 4.166, | |
| "step": 33 | |
| }, | |
| { | |
| "epoch": 0.015503875968992248, | |
| "grad_norm": 17.75687026977539, | |
| "learning_rate": 9.611448774886924e-05, | |
| "loss": 3.6981, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 0.01595987232102143, | |
| "grad_norm": 13.289236068725586, | |
| "learning_rate": 9.578866633275288e-05, | |
| "loss": 3.5732, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.016415868673050615, | |
| "grad_norm": 12.643074035644531, | |
| "learning_rate": 9.545032675245813e-05, | |
| "loss": 2.7541, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 0.0168718650250798, | |
| "grad_norm": 16.69640350341797, | |
| "learning_rate": 9.509956150664796e-05, | |
| "loss": 4.423, | |
| "step": 37 | |
| }, | |
| { | |
| "epoch": 0.017327861377108983, | |
| "grad_norm": 17.78840446472168, | |
| "learning_rate": 9.473646649103818e-05, | |
| "loss": 4.4712, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 0.017783857729138167, | |
| "grad_norm": 16.33861541748047, | |
| "learning_rate": 9.43611409721806e-05, | |
| "loss": 4.8851, | |
| "step": 39 | |
| }, | |
| { | |
| "epoch": 0.01823985408116735, | |
| "grad_norm": 15.124662399291992, | |
| "learning_rate": 9.397368756032445e-05, | |
| "loss": 3.9998, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.018695850433196534, | |
| "grad_norm": 12.763321876525879, | |
| "learning_rate": 9.357421218136386e-05, | |
| "loss": 3.5617, | |
| "step": 41 | |
| }, | |
| { | |
| "epoch": 0.019151846785225718, | |
| "grad_norm": 16.94106101989746, | |
| "learning_rate": 9.316282404787871e-05, | |
| "loss": 4.9259, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 0.0196078431372549, | |
| "grad_norm": 15.05451488494873, | |
| "learning_rate": 9.273963562927695e-05, | |
| "loss": 4.3584, | |
| "step": 43 | |
| }, | |
| { | |
| "epoch": 0.020063839489284085, | |
| "grad_norm": 14.099905014038086, | |
| "learning_rate": 9.230476262104677e-05, | |
| "loss": 4.5573, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 0.02051983584131327, | |
| "grad_norm": 15.21258544921875, | |
| "learning_rate": 9.185832391312644e-05, | |
| "loss": 3.675, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 0.020975832193342453, | |
| "grad_norm": 13.130019187927246, | |
| "learning_rate": 9.140044155740101e-05, | |
| "loss": 4.0426, | |
| "step": 46 | |
| }, | |
| { | |
| "epoch": 0.021431828545371637, | |
| "grad_norm": 14.917093276977539, | |
| "learning_rate": 9.093124073433463e-05, | |
| "loss": 4.6251, | |
| "step": 47 | |
| }, | |
| { | |
| "epoch": 0.02188782489740082, | |
| "grad_norm": 14.078794479370117, | |
| "learning_rate": 9.045084971874738e-05, | |
| "loss": 3.8073, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 0.022343821249430004, | |
| "grad_norm": 16.755746841430664, | |
| "learning_rate": 8.995939984474624e-05, | |
| "loss": 4.967, | |
| "step": 49 | |
| }, | |
| { | |
| "epoch": 0.022799817601459188, | |
| "grad_norm": 19.542896270751953, | |
| "learning_rate": 8.945702546981969e-05, | |
| "loss": 6.3588, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.022799817601459188, | |
| "eval_loss": 0.9113385081291199, | |
| "eval_runtime": 349.7073, | |
| "eval_samples_per_second": 10.563, | |
| "eval_steps_per_second": 2.642, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.023255813953488372, | |
| "grad_norm": 16.446622848510742, | |
| "learning_rate": 8.894386393810563e-05, | |
| "loss": 3.0671, | |
| "step": 51 | |
| }, | |
| { | |
| "epoch": 0.023711810305517556, | |
| "grad_norm": 10.394733428955078, | |
| "learning_rate": 8.842005554284296e-05, | |
| "loss": 3.1367, | |
| "step": 52 | |
| }, | |
| { | |
| "epoch": 0.02416780665754674, | |
| "grad_norm": 11.116695404052734, | |
| "learning_rate": 8.788574348801675e-05, | |
| "loss": 2.7149, | |
| "step": 53 | |
| }, | |
| { | |
| "epoch": 0.024623803009575923, | |
| "grad_norm": 9.479421615600586, | |
| "learning_rate": 8.73410738492077e-05, | |
| "loss": 2.1254, | |
| "step": 54 | |
| }, | |
| { | |
| "epoch": 0.025079799361605107, | |
| "grad_norm": 11.42734146118164, | |
| "learning_rate": 8.678619553365659e-05, | |
| "loss": 3.393, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 0.02553579571363429, | |
| "grad_norm": 9.619385719299316, | |
| "learning_rate": 8.622126023955446e-05, | |
| "loss": 2.6756, | |
| "step": 56 | |
| }, | |
| { | |
| "epoch": 0.025991792065663474, | |
| "grad_norm": 10.887063980102539, | |
| "learning_rate": 8.564642241456986e-05, | |
| "loss": 3.0928, | |
| "step": 57 | |
| }, | |
| { | |
| "epoch": 0.026447788417692658, | |
| "grad_norm": 9.697304725646973, | |
| "learning_rate": 8.506183921362443e-05, | |
| "loss": 2.6691, | |
| "step": 58 | |
| }, | |
| { | |
| "epoch": 0.026903784769721842, | |
| "grad_norm": 10.840964317321777, | |
| "learning_rate": 8.44676704559283e-05, | |
| "loss": 2.8811, | |
| "step": 59 | |
| }, | |
| { | |
| "epoch": 0.027359781121751026, | |
| "grad_norm": 12.989250183105469, | |
| "learning_rate": 8.386407858128706e-05, | |
| "loss": 3.8215, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.02781577747378021, | |
| "grad_norm": 12.489930152893066, | |
| "learning_rate": 8.32512286056924e-05, | |
| "loss": 3.967, | |
| "step": 61 | |
| }, | |
| { | |
| "epoch": 0.028271773825809393, | |
| "grad_norm": 11.065751075744629, | |
| "learning_rate": 8.262928807620843e-05, | |
| "loss": 2.8307, | |
| "step": 62 | |
| }, | |
| { | |
| "epoch": 0.028727770177838577, | |
| "grad_norm": 10.985672950744629, | |
| "learning_rate": 8.199842702516583e-05, | |
| "loss": 3.1239, | |
| "step": 63 | |
| }, | |
| { | |
| "epoch": 0.02918376652986776, | |
| "grad_norm": 9.590174674987793, | |
| "learning_rate": 8.135881792367686e-05, | |
| "loss": 2.7839, | |
| "step": 64 | |
| }, | |
| { | |
| "epoch": 0.029639762881896944, | |
| "grad_norm": 12.237695693969727, | |
| "learning_rate": 8.07106356344834e-05, | |
| "loss": 3.5155, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 0.030095759233926128, | |
| "grad_norm": 15.1380033493042, | |
| "learning_rate": 8.005405736415126e-05, | |
| "loss": 3.4588, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 0.030551755585955312, | |
| "grad_norm": 162.89707946777344, | |
| "learning_rate": 7.938926261462366e-05, | |
| "loss": 3.0448, | |
| "step": 67 | |
| }, | |
| { | |
| "epoch": 0.031007751937984496, | |
| "grad_norm": 13.830486297607422, | |
| "learning_rate": 7.871643313414718e-05, | |
| "loss": 3.0779, | |
| "step": 68 | |
| }, | |
| { | |
| "epoch": 0.03146374829001368, | |
| "grad_norm": 11.294425010681152, | |
| "learning_rate": 7.803575286758364e-05, | |
| "loss": 2.9532, | |
| "step": 69 | |
| }, | |
| { | |
| "epoch": 0.03191974464204286, | |
| "grad_norm": 10.481371879577637, | |
| "learning_rate": 7.734740790612136e-05, | |
| "loss": 3.4687, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.03237574099407205, | |
| "grad_norm": 12.793365478515625, | |
| "learning_rate": 7.66515864363997e-05, | |
| "loss": 3.5947, | |
| "step": 71 | |
| }, | |
| { | |
| "epoch": 0.03283173734610123, | |
| "grad_norm": 12.153921127319336, | |
| "learning_rate": 7.594847868906076e-05, | |
| "loss": 3.8798, | |
| "step": 72 | |
| }, | |
| { | |
| "epoch": 0.033287733698130414, | |
| "grad_norm": 9.509449005126953, | |
| "learning_rate": 7.52382768867422e-05, | |
| "loss": 2.9595, | |
| "step": 73 | |
| }, | |
| { | |
| "epoch": 0.0337437300501596, | |
| "grad_norm": 10.440327644348145, | |
| "learning_rate": 7.452117519152542e-05, | |
| "loss": 3.7834, | |
| "step": 74 | |
| }, | |
| { | |
| "epoch": 0.03419972640218878, | |
| "grad_norm": 10.908466339111328, | |
| "learning_rate": 7.379736965185368e-05, | |
| "loss": 2.6557, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.034655722754217966, | |
| "grad_norm": 9.98528003692627, | |
| "learning_rate": 7.30670581489344e-05, | |
| "loss": 2.4296, | |
| "step": 76 | |
| }, | |
| { | |
| "epoch": 0.03511171910624715, | |
| "grad_norm": 15.787983894348145, | |
| "learning_rate": 7.233044034264034e-05, | |
| "loss": 4.7181, | |
| "step": 77 | |
| }, | |
| { | |
| "epoch": 0.03556771545827633, | |
| "grad_norm": 12.403399467468262, | |
| "learning_rate": 7.158771761692464e-05, | |
| "loss": 3.1346, | |
| "step": 78 | |
| }, | |
| { | |
| "epoch": 0.03602371181030552, | |
| "grad_norm": 12.2731351852417, | |
| "learning_rate": 7.083909302476453e-05, | |
| "loss": 2.8631, | |
| "step": 79 | |
| }, | |
| { | |
| "epoch": 0.0364797081623347, | |
| "grad_norm": 14.62734317779541, | |
| "learning_rate": 7.008477123264848e-05, | |
| "loss": 3.9177, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.036935704514363885, | |
| "grad_norm": 12.990211486816406, | |
| "learning_rate": 6.932495846462261e-05, | |
| "loss": 3.5621, | |
| "step": 81 | |
| }, | |
| { | |
| "epoch": 0.03739170086639307, | |
| "grad_norm": 14.018969535827637, | |
| "learning_rate": 6.855986244591104e-05, | |
| "loss": 3.4686, | |
| "step": 82 | |
| }, | |
| { | |
| "epoch": 0.03784769721842225, | |
| "grad_norm": 14.162925720214844, | |
| "learning_rate": 6.778969234612584e-05, | |
| "loss": 3.3132, | |
| "step": 83 | |
| }, | |
| { | |
| "epoch": 0.038303693570451436, | |
| "grad_norm": 14.519307136535645, | |
| "learning_rate": 6.701465872208216e-05, | |
| "loss": 4.188, | |
| "step": 84 | |
| }, | |
| { | |
| "epoch": 0.03875968992248062, | |
| "grad_norm": 12.667447090148926, | |
| "learning_rate": 6.623497346023418e-05, | |
| "loss": 3.8917, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 0.0392156862745098, | |
| "grad_norm": 12.733670234680176, | |
| "learning_rate": 6.545084971874738e-05, | |
| "loss": 3.1904, | |
| "step": 86 | |
| }, | |
| { | |
| "epoch": 0.03967168262653899, | |
| "grad_norm": 11.787132263183594, | |
| "learning_rate": 6.466250186922325e-05, | |
| "loss": 3.0343, | |
| "step": 87 | |
| }, | |
| { | |
| "epoch": 0.04012767897856817, | |
| "grad_norm": 10.446242332458496, | |
| "learning_rate": 6.387014543809223e-05, | |
| "loss": 3.3589, | |
| "step": 88 | |
| }, | |
| { | |
| "epoch": 0.040583675330597355, | |
| "grad_norm": 12.251482009887695, | |
| "learning_rate": 6.307399704769099e-05, | |
| "loss": 3.7822, | |
| "step": 89 | |
| }, | |
| { | |
| "epoch": 0.04103967168262654, | |
| "grad_norm": 13.403604507446289, | |
| "learning_rate": 6.227427435703997e-05, | |
| "loss": 3.5298, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.04149566803465572, | |
| "grad_norm": 9.842473030090332, | |
| "learning_rate": 6.147119600233758e-05, | |
| "loss": 2.8747, | |
| "step": 91 | |
| }, | |
| { | |
| "epoch": 0.041951664386684906, | |
| "grad_norm": 12.207097053527832, | |
| "learning_rate": 6.066498153718735e-05, | |
| "loss": 3.7642, | |
| "step": 92 | |
| }, | |
| { | |
| "epoch": 0.04240766073871409, | |
| "grad_norm": 11.435685157775879, | |
| "learning_rate": 5.985585137257401e-05, | |
| "loss": 3.9007, | |
| "step": 93 | |
| }, | |
| { | |
| "epoch": 0.04286365709074327, | |
| "grad_norm": 10.784540176391602, | |
| "learning_rate": 5.90440267166055e-05, | |
| "loss": 3.7834, | |
| "step": 94 | |
| }, | |
| { | |
| "epoch": 0.04331965344277246, | |
| "grad_norm": 12.038006782531738, | |
| "learning_rate": 5.8229729514036705e-05, | |
| "loss": 3.9517, | |
| "step": 95 | |
| }, | |
| { | |
| "epoch": 0.04377564979480164, | |
| "grad_norm": 11.988702774047852, | |
| "learning_rate": 5.74131823855921e-05, | |
| "loss": 3.4095, | |
| "step": 96 | |
| }, | |
| { | |
| "epoch": 0.044231646146830825, | |
| "grad_norm": 10.996843338012695, | |
| "learning_rate": 5.6594608567103456e-05, | |
| "loss": 3.1867, | |
| "step": 97 | |
| }, | |
| { | |
| "epoch": 0.04468764249886001, | |
| "grad_norm": 14.686224937438965, | |
| "learning_rate": 5.577423184847932e-05, | |
| "loss": 4.0562, | |
| "step": 98 | |
| }, | |
| { | |
| "epoch": 0.04514363885088919, | |
| "grad_norm": 14.393037796020508, | |
| "learning_rate": 5.495227651252315e-05, | |
| "loss": 4.388, | |
| "step": 99 | |
| }, | |
| { | |
| "epoch": 0.045599635202918376, | |
| "grad_norm": 17.71428871154785, | |
| "learning_rate": 5.4128967273616625e-05, | |
| "loss": 4.8689, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.045599635202918376, | |
| "eval_loss": 0.8197675347328186, | |
| "eval_runtime": 349.5978, | |
| "eval_samples_per_second": 10.566, | |
| "eval_steps_per_second": 2.643, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.04605563155494756, | |
| "grad_norm": 10.61023235321045, | |
| "learning_rate": 5.330452921628497e-05, | |
| "loss": 2.7583, | |
| "step": 101 | |
| }, | |
| { | |
| "epoch": 0.046511627906976744, | |
| "grad_norm": 9.402891159057617, | |
| "learning_rate": 5.247918773366112e-05, | |
| "loss": 2.6451, | |
| "step": 102 | |
| }, | |
| { | |
| "epoch": 0.04696762425900593, | |
| "grad_norm": 9.066802978515625, | |
| "learning_rate": 5.165316846586541e-05, | |
| "loss": 2.2557, | |
| "step": 103 | |
| }, | |
| { | |
| "epoch": 0.04742362061103511, | |
| "grad_norm": 8.31011962890625, | |
| "learning_rate": 5.0826697238317935e-05, | |
| "loss": 1.9061, | |
| "step": 104 | |
| }, | |
| { | |
| "epoch": 0.047879616963064295, | |
| "grad_norm": 9.681857109069824, | |
| "learning_rate": 5e-05, | |
| "loss": 3.8238, | |
| "step": 105 | |
| }, | |
| { | |
| "epoch": 0.04833561331509348, | |
| "grad_norm": 10.533971786499023, | |
| "learning_rate": 4.917330276168208e-05, | |
| "loss": 3.2304, | |
| "step": 106 | |
| }, | |
| { | |
| "epoch": 0.04879160966712266, | |
| "grad_norm": 9.498639106750488, | |
| "learning_rate": 4.834683153413459e-05, | |
| "loss": 2.0589, | |
| "step": 107 | |
| }, | |
| { | |
| "epoch": 0.049247606019151846, | |
| "grad_norm": 10.653234481811523, | |
| "learning_rate": 4.7520812266338885e-05, | |
| "loss": 2.8517, | |
| "step": 108 | |
| }, | |
| { | |
| "epoch": 0.04970360237118103, | |
| "grad_norm": 9.575819969177246, | |
| "learning_rate": 4.669547078371504e-05, | |
| "loss": 2.7365, | |
| "step": 109 | |
| }, | |
| { | |
| "epoch": 0.050159598723210214, | |
| "grad_norm": 9.632899284362793, | |
| "learning_rate": 4.5871032726383386e-05, | |
| "loss": 2.5004, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.0506155950752394, | |
| "grad_norm": 9.333025932312012, | |
| "learning_rate": 4.504772348747687e-05, | |
| "loss": 2.746, | |
| "step": 111 | |
| }, | |
| { | |
| "epoch": 0.05107159142726858, | |
| "grad_norm": 8.68212604522705, | |
| "learning_rate": 4.4225768151520694e-05, | |
| "loss": 2.1002, | |
| "step": 112 | |
| }, | |
| { | |
| "epoch": 0.051527587779297765, | |
| "grad_norm": 9.370121955871582, | |
| "learning_rate": 4.3405391432896555e-05, | |
| "loss": 2.6227, | |
| "step": 113 | |
| }, | |
| { | |
| "epoch": 0.05198358413132695, | |
| "grad_norm": 8.905264854431152, | |
| "learning_rate": 4.2586817614407895e-05, | |
| "loss": 2.4046, | |
| "step": 114 | |
| }, | |
| { | |
| "epoch": 0.05243958048335613, | |
| "grad_norm": 10.212873458862305, | |
| "learning_rate": 4.17702704859633e-05, | |
| "loss": 2.824, | |
| "step": 115 | |
| }, | |
| { | |
| "epoch": 0.052895576835385316, | |
| "grad_norm": 9.993491172790527, | |
| "learning_rate": 4.095597328339452e-05, | |
| "loss": 2.3992, | |
| "step": 116 | |
| }, | |
| { | |
| "epoch": 0.0533515731874145, | |
| "grad_norm": 11.604905128479004, | |
| "learning_rate": 4.0144148627425993e-05, | |
| "loss": 2.7306, | |
| "step": 117 | |
| }, | |
| { | |
| "epoch": 0.053807569539443684, | |
| "grad_norm": 10.31896686553955, | |
| "learning_rate": 3.933501846281267e-05, | |
| "loss": 3.0943, | |
| "step": 118 | |
| }, | |
| { | |
| "epoch": 0.05426356589147287, | |
| "grad_norm": 10.747237205505371, | |
| "learning_rate": 3.852880399766243e-05, | |
| "loss": 3.0425, | |
| "step": 119 | |
| }, | |
| { | |
| "epoch": 0.05471956224350205, | |
| "grad_norm": 9.785868644714355, | |
| "learning_rate": 3.772572564296005e-05, | |
| "loss": 2.4842, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.055175558595531235, | |
| "grad_norm": 10.849400520324707, | |
| "learning_rate": 3.6926002952309016e-05, | |
| "loss": 3.777, | |
| "step": 121 | |
| }, | |
| { | |
| "epoch": 0.05563155494756042, | |
| "grad_norm": 11.813304901123047, | |
| "learning_rate": 3.612985456190778e-05, | |
| "loss": 3.0709, | |
| "step": 122 | |
| }, | |
| { | |
| "epoch": 0.0560875512995896, | |
| "grad_norm": 10.825654029846191, | |
| "learning_rate": 3.533749813077677e-05, | |
| "loss": 3.3899, | |
| "step": 123 | |
| }, | |
| { | |
| "epoch": 0.056543547651618786, | |
| "grad_norm": 10.39290714263916, | |
| "learning_rate": 3.4549150281252636e-05, | |
| "loss": 2.7794, | |
| "step": 124 | |
| }, | |
| { | |
| "epoch": 0.05699954400364797, | |
| "grad_norm": 9.933770179748535, | |
| "learning_rate": 3.3765026539765834e-05, | |
| "loss": 2.875, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 0.057455540355677154, | |
| "grad_norm": 10.807685852050781, | |
| "learning_rate": 3.298534127791785e-05, | |
| "loss": 3.2124, | |
| "step": 126 | |
| }, | |
| { | |
| "epoch": 0.05791153670770634, | |
| "grad_norm": 10.649255752563477, | |
| "learning_rate": 3.221030765387417e-05, | |
| "loss": 3.7709, | |
| "step": 127 | |
| }, | |
| { | |
| "epoch": 0.05836753305973552, | |
| "grad_norm": 12.467437744140625, | |
| "learning_rate": 3.144013755408895e-05, | |
| "loss": 3.4853, | |
| "step": 128 | |
| }, | |
| { | |
| "epoch": 0.058823529411764705, | |
| "grad_norm": 12.292319297790527, | |
| "learning_rate": 3.0675041535377405e-05, | |
| "loss": 3.7258, | |
| "step": 129 | |
| }, | |
| { | |
| "epoch": 0.05927952576379389, | |
| "grad_norm": 9.66211986541748, | |
| "learning_rate": 2.991522876735154e-05, | |
| "loss": 3.0776, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.05973552211582307, | |
| "grad_norm": 10.147554397583008, | |
| "learning_rate": 2.916090697523549e-05, | |
| "loss": 3.1498, | |
| "step": 131 | |
| }, | |
| { | |
| "epoch": 0.060191518467852256, | |
| "grad_norm": 11.330928802490234, | |
| "learning_rate": 2.8412282383075363e-05, | |
| "loss": 3.2106, | |
| "step": 132 | |
| }, | |
| { | |
| "epoch": 0.06064751481988144, | |
| "grad_norm": 11.31750202178955, | |
| "learning_rate": 2.766955965735968e-05, | |
| "loss": 3.2155, | |
| "step": 133 | |
| }, | |
| { | |
| "epoch": 0.061103511171910624, | |
| "grad_norm": 10.476314544677734, | |
| "learning_rate": 2.693294185106562e-05, | |
| "loss": 3.3905, | |
| "step": 134 | |
| }, | |
| { | |
| "epoch": 0.06155950752393981, | |
| "grad_norm": 11.83663558959961, | |
| "learning_rate": 2.6202630348146324e-05, | |
| "loss": 3.4465, | |
| "step": 135 | |
| }, | |
| { | |
| "epoch": 0.06201550387596899, | |
| "grad_norm": 10.024785041809082, | |
| "learning_rate": 2.547882480847461e-05, | |
| "loss": 2.7385, | |
| "step": 136 | |
| }, | |
| { | |
| "epoch": 0.062471500227998175, | |
| "grad_norm": 11.038595199584961, | |
| "learning_rate": 2.476172311325783e-05, | |
| "loss": 2.9536, | |
| "step": 137 | |
| }, | |
| { | |
| "epoch": 0.06292749658002736, | |
| "grad_norm": 10.529680252075195, | |
| "learning_rate": 2.405152131093926e-05, | |
| "loss": 3.1882, | |
| "step": 138 | |
| }, | |
| { | |
| "epoch": 0.06338349293205654, | |
| "grad_norm": 10.84036636352539, | |
| "learning_rate": 2.3348413563600325e-05, | |
| "loss": 3.0197, | |
| "step": 139 | |
| }, | |
| { | |
| "epoch": 0.06383948928408573, | |
| "grad_norm": 11.032572746276855, | |
| "learning_rate": 2.2652592093878666e-05, | |
| "loss": 3.8491, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.06429548563611491, | |
| "grad_norm": 9.125802040100098, | |
| "learning_rate": 2.196424713241637e-05, | |
| "loss": 2.5575, | |
| "step": 141 | |
| }, | |
| { | |
| "epoch": 0.0647514819881441, | |
| "grad_norm": 13.208250999450684, | |
| "learning_rate": 2.128356686585282e-05, | |
| "loss": 3.8827, | |
| "step": 142 | |
| }, | |
| { | |
| "epoch": 0.06520747834017328, | |
| "grad_norm": 10.89819622039795, | |
| "learning_rate": 2.061073738537635e-05, | |
| "loss": 3.476, | |
| "step": 143 | |
| }, | |
| { | |
| "epoch": 0.06566347469220246, | |
| "grad_norm": 12.219125747680664, | |
| "learning_rate": 1.9945942635848748e-05, | |
| "loss": 3.6378, | |
| "step": 144 | |
| }, | |
| { | |
| "epoch": 0.06611947104423165, | |
| "grad_norm": 11.675637245178223, | |
| "learning_rate": 1.928936436551661e-05, | |
| "loss": 3.2438, | |
| "step": 145 | |
| }, | |
| { | |
| "epoch": 0.06657546739626083, | |
| "grad_norm": 9.782285690307617, | |
| "learning_rate": 1.8641182076323148e-05, | |
| "loss": 2.9421, | |
| "step": 146 | |
| }, | |
| { | |
| "epoch": 0.06703146374829001, | |
| "grad_norm": 10.580471992492676, | |
| "learning_rate": 1.800157297483417e-05, | |
| "loss": 2.9575, | |
| "step": 147 | |
| }, | |
| { | |
| "epoch": 0.0674874601003192, | |
| "grad_norm": 12.524446487426758, | |
| "learning_rate": 1.7370711923791567e-05, | |
| "loss": 3.7983, | |
| "step": 148 | |
| }, | |
| { | |
| "epoch": 0.06794345645234838, | |
| "grad_norm": 13.980012893676758, | |
| "learning_rate": 1.6748771394307585e-05, | |
| "loss": 4.4534, | |
| "step": 149 | |
| }, | |
| { | |
| "epoch": 0.06839945280437756, | |
| "grad_norm": 16.460304260253906, | |
| "learning_rate": 1.6135921418712956e-05, | |
| "loss": 5.1192, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.06839945280437756, | |
| "eval_loss": 0.7523689866065979, | |
| "eval_runtime": 349.5702, | |
| "eval_samples_per_second": 10.567, | |
| "eval_steps_per_second": 2.643, | |
| "step": 150 | |
| } | |
| ], | |
| "logging_steps": 1, | |
| "max_steps": 200, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "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": 2.150074508503941e+17, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |