| { | |
| "best_metric": 2.1568007469177246, | |
| "best_model_checkpoint": "miner_id_24/checkpoint-150", | |
| "epoch": 0.17311021350259664, | |
| "eval_steps": 50, | |
| "global_step": 150, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.001154068090017311, | |
| "grad_norm": 0.7504671812057495, | |
| "learning_rate": 1e-05, | |
| "loss": 1.6459, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.001154068090017311, | |
| "eval_loss": 2.573826551437378, | |
| "eval_runtime": 108.773, | |
| "eval_samples_per_second": 13.422, | |
| "eval_steps_per_second": 3.356, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.002308136180034622, | |
| "grad_norm": 0.8413335084915161, | |
| "learning_rate": 2e-05, | |
| "loss": 2.1504, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.003462204270051933, | |
| "grad_norm": 0.9515752792358398, | |
| "learning_rate": 3e-05, | |
| "loss": 2.2621, | |
| "step": 3 | |
| }, | |
| { | |
| "epoch": 0.004616272360069244, | |
| "grad_norm": 0.9131438136100769, | |
| "learning_rate": 4e-05, | |
| "loss": 2.0219, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.005770340450086555, | |
| "grad_norm": 0.8736193776130676, | |
| "learning_rate": 5e-05, | |
| "loss": 2.1878, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.006924408540103866, | |
| "grad_norm": 0.954706072807312, | |
| "learning_rate": 6e-05, | |
| "loss": 2.2973, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.008078476630121177, | |
| "grad_norm": 1.0279555320739746, | |
| "learning_rate": 7e-05, | |
| "loss": 2.0644, | |
| "step": 7 | |
| }, | |
| { | |
| "epoch": 0.009232544720138488, | |
| "grad_norm": 0.9788753986358643, | |
| "learning_rate": 8e-05, | |
| "loss": 2.0584, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.0103866128101558, | |
| "grad_norm": 1.0885084867477417, | |
| "learning_rate": 9e-05, | |
| "loss": 2.0937, | |
| "step": 9 | |
| }, | |
| { | |
| "epoch": 0.01154068090017311, | |
| "grad_norm": 1.1294866800308228, | |
| "learning_rate": 0.0001, | |
| "loss": 2.1254, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.012694748990190421, | |
| "grad_norm": 0.8876044750213623, | |
| "learning_rate": 9.999316524962345e-05, | |
| "loss": 2.0217, | |
| "step": 11 | |
| }, | |
| { | |
| "epoch": 0.013848817080207732, | |
| "grad_norm": 0.9484460949897766, | |
| "learning_rate": 9.997266286704631e-05, | |
| "loss": 2.0849, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 0.015002885170225043, | |
| "grad_norm": 0.9615907669067383, | |
| "learning_rate": 9.993849845741524e-05, | |
| "loss": 1.9606, | |
| "step": 13 | |
| }, | |
| { | |
| "epoch": 0.016156953260242354, | |
| "grad_norm": 0.8483062386512756, | |
| "learning_rate": 9.989068136093873e-05, | |
| "loss": 2.0657, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 0.017311021350259664, | |
| "grad_norm": 0.9245672821998596, | |
| "learning_rate": 9.98292246503335e-05, | |
| "loss": 2.1575, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.018465089440276975, | |
| "grad_norm": 0.9007608890533447, | |
| "learning_rate": 9.975414512725057e-05, | |
| "loss": 2.1824, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 0.019619157530294286, | |
| "grad_norm": 0.8688671588897705, | |
| "learning_rate": 9.966546331768191e-05, | |
| "loss": 2.1835, | |
| "step": 17 | |
| }, | |
| { | |
| "epoch": 0.0207732256203116, | |
| "grad_norm": 0.957897961139679, | |
| "learning_rate": 9.956320346634876e-05, | |
| "loss": 2.1329, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 0.02192729371032891, | |
| "grad_norm": 0.9246083498001099, | |
| "learning_rate": 9.944739353007344e-05, | |
| "loss": 2.0605, | |
| "step": 19 | |
| }, | |
| { | |
| "epoch": 0.02308136180034622, | |
| "grad_norm": 0.8987853527069092, | |
| "learning_rate": 9.931806517013612e-05, | |
| "loss": 2.0784, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.024235429890363532, | |
| "grad_norm": 0.9510769248008728, | |
| "learning_rate": 9.917525374361912e-05, | |
| "loss": 2.0552, | |
| "step": 21 | |
| }, | |
| { | |
| "epoch": 0.025389497980380843, | |
| "grad_norm": 0.9135078191757202, | |
| "learning_rate": 9.901899829374047e-05, | |
| "loss": 2.1988, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 0.026543566070398154, | |
| "grad_norm": 0.9757664799690247, | |
| "learning_rate": 9.884934153917997e-05, | |
| "loss": 2.1735, | |
| "step": 23 | |
| }, | |
| { | |
| "epoch": 0.027697634160415464, | |
| "grad_norm": 0.9541769027709961, | |
| "learning_rate": 9.86663298624003e-05, | |
| "loss": 2.1115, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 0.028851702250432775, | |
| "grad_norm": 0.9693971872329712, | |
| "learning_rate": 9.847001329696653e-05, | |
| "loss": 2.0816, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.030005770340450086, | |
| "grad_norm": 0.9962961673736572, | |
| "learning_rate": 9.826044551386744e-05, | |
| "loss": 2.1396, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 0.031159838430467397, | |
| "grad_norm": 0.9708167910575867, | |
| "learning_rate": 9.803768380684242e-05, | |
| "loss": 2.0916, | |
| "step": 27 | |
| }, | |
| { | |
| "epoch": 0.03231390652048471, | |
| "grad_norm": 0.9625340700149536, | |
| "learning_rate": 9.780178907671789e-05, | |
| "loss": 2.0674, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 0.03346797461050202, | |
| "grad_norm": 0.9793923497200012, | |
| "learning_rate": 9.755282581475769e-05, | |
| "loss": 2.1272, | |
| "step": 29 | |
| }, | |
| { | |
| "epoch": 0.03462204270051933, | |
| "grad_norm": 0.9763445854187012, | |
| "learning_rate": 9.729086208503174e-05, | |
| "loss": 2.2651, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.03577611079053664, | |
| "grad_norm": 0.9952139258384705, | |
| "learning_rate": 9.701596950580806e-05, | |
| "loss": 2.0592, | |
| "step": 31 | |
| }, | |
| { | |
| "epoch": 0.03693017888055395, | |
| "grad_norm": 1.035845398902893, | |
| "learning_rate": 9.672822322997305e-05, | |
| "loss": 2.1378, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 0.038084246970571264, | |
| "grad_norm": 1.1929069757461548, | |
| "learning_rate": 9.642770192448536e-05, | |
| "loss": 2.2375, | |
| "step": 33 | |
| }, | |
| { | |
| "epoch": 0.03923831506058857, | |
| "grad_norm": 1.004943609237671, | |
| "learning_rate": 9.611448774886924e-05, | |
| "loss": 2.2335, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 0.040392383150605886, | |
| "grad_norm": 1.1178886890411377, | |
| "learning_rate": 9.578866633275288e-05, | |
| "loss": 2.1909, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.0415464512406232, | |
| "grad_norm": 1.0786731243133545, | |
| "learning_rate": 9.545032675245813e-05, | |
| "loss": 2.1066, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 0.04270051933064051, | |
| "grad_norm": 1.1023880243301392, | |
| "learning_rate": 9.509956150664796e-05, | |
| "loss": 2.1938, | |
| "step": 37 | |
| }, | |
| { | |
| "epoch": 0.04385458742065782, | |
| "grad_norm": 1.1077494621276855, | |
| "learning_rate": 9.473646649103818e-05, | |
| "loss": 2.1935, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 0.04500865551067513, | |
| "grad_norm": 1.3662453889846802, | |
| "learning_rate": 9.43611409721806e-05, | |
| "loss": 2.2462, | |
| "step": 39 | |
| }, | |
| { | |
| "epoch": 0.04616272360069244, | |
| "grad_norm": 1.2383288145065308, | |
| "learning_rate": 9.397368756032445e-05, | |
| "loss": 2.2246, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.04731679169070975, | |
| "grad_norm": 1.7511357069015503, | |
| "learning_rate": 9.357421218136386e-05, | |
| "loss": 2.344, | |
| "step": 41 | |
| }, | |
| { | |
| "epoch": 0.048470859780727064, | |
| "grad_norm": 1.3569585084915161, | |
| "learning_rate": 9.316282404787871e-05, | |
| "loss": 2.1847, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 0.04962492787074437, | |
| "grad_norm": 1.3819233179092407, | |
| "learning_rate": 9.273963562927695e-05, | |
| "loss": 2.1878, | |
| "step": 43 | |
| }, | |
| { | |
| "epoch": 0.050778995960761686, | |
| "grad_norm": 1.267579436302185, | |
| "learning_rate": 9.230476262104677e-05, | |
| "loss": 2.2589, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 0.05193306405077899, | |
| "grad_norm": 1.4915302991867065, | |
| "learning_rate": 9.185832391312644e-05, | |
| "loss": 2.2349, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 0.05308713214079631, | |
| "grad_norm": 1.391606330871582, | |
| "learning_rate": 9.140044155740101e-05, | |
| "loss": 2.2085, | |
| "step": 46 | |
| }, | |
| { | |
| "epoch": 0.054241200230813615, | |
| "grad_norm": 1.5685354471206665, | |
| "learning_rate": 9.093124073433463e-05, | |
| "loss": 2.2521, | |
| "step": 47 | |
| }, | |
| { | |
| "epoch": 0.05539526832083093, | |
| "grad_norm": 1.7326840162277222, | |
| "learning_rate": 9.045084971874738e-05, | |
| "loss": 2.2518, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 0.05654933641084824, | |
| "grad_norm": 2.054521083831787, | |
| "learning_rate": 8.995939984474624e-05, | |
| "loss": 2.4485, | |
| "step": 49 | |
| }, | |
| { | |
| "epoch": 0.05770340450086555, | |
| "grad_norm": 5.3004021644592285, | |
| "learning_rate": 8.945702546981969e-05, | |
| "loss": 2.8414, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.05770340450086555, | |
| "eval_loss": 2.477184772491455, | |
| "eval_runtime": 109.846, | |
| "eval_samples_per_second": 13.291, | |
| "eval_steps_per_second": 3.323, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.058857472590882864, | |
| "grad_norm": 1.7681031227111816, | |
| "learning_rate": 8.894386393810563e-05, | |
| "loss": 1.8931, | |
| "step": 51 | |
| }, | |
| { | |
| "epoch": 0.06001154068090017, | |
| "grad_norm": 1.580583095550537, | |
| "learning_rate": 8.842005554284296e-05, | |
| "loss": 1.9345, | |
| "step": 52 | |
| }, | |
| { | |
| "epoch": 0.061165608770917486, | |
| "grad_norm": 1.0677828788757324, | |
| "learning_rate": 8.788574348801675e-05, | |
| "loss": 2.0534, | |
| "step": 53 | |
| }, | |
| { | |
| "epoch": 0.06231967686093479, | |
| "grad_norm": 1.0560821294784546, | |
| "learning_rate": 8.73410738492077e-05, | |
| "loss": 2.1798, | |
| "step": 54 | |
| }, | |
| { | |
| "epoch": 0.06347374495095211, | |
| "grad_norm": 0.8886532783508301, | |
| "learning_rate": 8.678619553365659e-05, | |
| "loss": 1.9364, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 0.06462781304096941, | |
| "grad_norm": 0.7665138840675354, | |
| "learning_rate": 8.622126023955446e-05, | |
| "loss": 2.0227, | |
| "step": 56 | |
| }, | |
| { | |
| "epoch": 0.06578188113098672, | |
| "grad_norm": 0.7530053853988647, | |
| "learning_rate": 8.564642241456986e-05, | |
| "loss": 2.0664, | |
| "step": 57 | |
| }, | |
| { | |
| "epoch": 0.06693594922100404, | |
| "grad_norm": 0.7734319567680359, | |
| "learning_rate": 8.506183921362443e-05, | |
| "loss": 2.0919, | |
| "step": 58 | |
| }, | |
| { | |
| "epoch": 0.06809001731102135, | |
| "grad_norm": 0.786206066608429, | |
| "learning_rate": 8.44676704559283e-05, | |
| "loss": 2.0941, | |
| "step": 59 | |
| }, | |
| { | |
| "epoch": 0.06924408540103866, | |
| "grad_norm": 0.8636292219161987, | |
| "learning_rate": 8.386407858128706e-05, | |
| "loss": 1.8746, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.07039815349105598, | |
| "grad_norm": 0.758708119392395, | |
| "learning_rate": 8.32512286056924e-05, | |
| "loss": 2.0833, | |
| "step": 61 | |
| }, | |
| { | |
| "epoch": 0.07155222158107329, | |
| "grad_norm": 0.7842296957969666, | |
| "learning_rate": 8.262928807620843e-05, | |
| "loss": 2.003, | |
| "step": 62 | |
| }, | |
| { | |
| "epoch": 0.0727062896710906, | |
| "grad_norm": 0.8108060956001282, | |
| "learning_rate": 8.199842702516583e-05, | |
| "loss": 2.0854, | |
| "step": 63 | |
| }, | |
| { | |
| "epoch": 0.0738603577611079, | |
| "grad_norm": 0.8185780644416809, | |
| "learning_rate": 8.135881792367686e-05, | |
| "loss": 2.0941, | |
| "step": 64 | |
| }, | |
| { | |
| "epoch": 0.07501442585112522, | |
| "grad_norm": 0.7696530818939209, | |
| "learning_rate": 8.07106356344834e-05, | |
| "loss": 1.9962, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 0.07616849394114253, | |
| "grad_norm": 0.8107040524482727, | |
| "learning_rate": 8.005405736415126e-05, | |
| "loss": 2.104, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 0.07732256203115984, | |
| "grad_norm": 0.8197031021118164, | |
| "learning_rate": 7.938926261462366e-05, | |
| "loss": 2.1132, | |
| "step": 67 | |
| }, | |
| { | |
| "epoch": 0.07847663012117714, | |
| "grad_norm": 0.8103302717208862, | |
| "learning_rate": 7.871643313414718e-05, | |
| "loss": 2.2128, | |
| "step": 68 | |
| }, | |
| { | |
| "epoch": 0.07963069821119446, | |
| "grad_norm": 0.8698779940605164, | |
| "learning_rate": 7.803575286758364e-05, | |
| "loss": 2.1603, | |
| "step": 69 | |
| }, | |
| { | |
| "epoch": 0.08078476630121177, | |
| "grad_norm": 0.7898233532905579, | |
| "learning_rate": 7.734740790612136e-05, | |
| "loss": 2.1137, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.08193883439122908, | |
| "grad_norm": 0.9188711643218994, | |
| "learning_rate": 7.66515864363997e-05, | |
| "loss": 2.1372, | |
| "step": 71 | |
| }, | |
| { | |
| "epoch": 0.0830929024812464, | |
| "grad_norm": 0.8325892090797424, | |
| "learning_rate": 7.594847868906076e-05, | |
| "loss": 2.091, | |
| "step": 72 | |
| }, | |
| { | |
| "epoch": 0.08424697057126371, | |
| "grad_norm": 0.857109010219574, | |
| "learning_rate": 7.52382768867422e-05, | |
| "loss": 2.1636, | |
| "step": 73 | |
| }, | |
| { | |
| "epoch": 0.08540103866128101, | |
| "grad_norm": 0.8908112049102783, | |
| "learning_rate": 7.452117519152542e-05, | |
| "loss": 2.1437, | |
| "step": 74 | |
| }, | |
| { | |
| "epoch": 0.08655510675129832, | |
| "grad_norm": 0.899205207824707, | |
| "learning_rate": 7.379736965185368e-05, | |
| "loss": 2.2169, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.08770917484131564, | |
| "grad_norm": 0.9425099492073059, | |
| "learning_rate": 7.30670581489344e-05, | |
| "loss": 2.2109, | |
| "step": 76 | |
| }, | |
| { | |
| "epoch": 0.08886324293133295, | |
| "grad_norm": 0.876467764377594, | |
| "learning_rate": 7.233044034264034e-05, | |
| "loss": 2.1799, | |
| "step": 77 | |
| }, | |
| { | |
| "epoch": 0.09001731102135026, | |
| "grad_norm": 0.9324733018875122, | |
| "learning_rate": 7.158771761692464e-05, | |
| "loss": 2.1933, | |
| "step": 78 | |
| }, | |
| { | |
| "epoch": 0.09117137911136756, | |
| "grad_norm": 0.9686921238899231, | |
| "learning_rate": 7.083909302476453e-05, | |
| "loss": 2.272, | |
| "step": 79 | |
| }, | |
| { | |
| "epoch": 0.09232544720138489, | |
| "grad_norm": 0.9561650156974792, | |
| "learning_rate": 7.008477123264848e-05, | |
| "loss": 2.1502, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.0934795152914022, | |
| "grad_norm": 0.9259841442108154, | |
| "learning_rate": 6.932495846462261e-05, | |
| "loss": 2.2034, | |
| "step": 81 | |
| }, | |
| { | |
| "epoch": 0.0946335833814195, | |
| "grad_norm": 1.0370311737060547, | |
| "learning_rate": 6.855986244591104e-05, | |
| "loss": 2.1821, | |
| "step": 82 | |
| }, | |
| { | |
| "epoch": 0.09578765147143682, | |
| "grad_norm": 0.994090735912323, | |
| "learning_rate": 6.778969234612584e-05, | |
| "loss": 2.1535, | |
| "step": 83 | |
| }, | |
| { | |
| "epoch": 0.09694171956145413, | |
| "grad_norm": 1.0461393594741821, | |
| "learning_rate": 6.701465872208216e-05, | |
| "loss": 2.1661, | |
| "step": 84 | |
| }, | |
| { | |
| "epoch": 0.09809578765147144, | |
| "grad_norm": 1.0954227447509766, | |
| "learning_rate": 6.623497346023418e-05, | |
| "loss": 2.1766, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 0.09924985574148874, | |
| "grad_norm": 1.007415533065796, | |
| "learning_rate": 6.545084971874738e-05, | |
| "loss": 2.1727, | |
| "step": 86 | |
| }, | |
| { | |
| "epoch": 0.10040392383150606, | |
| "grad_norm": 1.0590732097625732, | |
| "learning_rate": 6.466250186922325e-05, | |
| "loss": 2.2675, | |
| "step": 87 | |
| }, | |
| { | |
| "epoch": 0.10155799192152337, | |
| "grad_norm": 1.1013295650482178, | |
| "learning_rate": 6.387014543809223e-05, | |
| "loss": 2.2104, | |
| "step": 88 | |
| }, | |
| { | |
| "epoch": 0.10271206001154068, | |
| "grad_norm": 1.1277257204055786, | |
| "learning_rate": 6.307399704769099e-05, | |
| "loss": 2.1344, | |
| "step": 89 | |
| }, | |
| { | |
| "epoch": 0.10386612810155799, | |
| "grad_norm": 1.1298824548721313, | |
| "learning_rate": 6.227427435703997e-05, | |
| "loss": 2.1562, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.10502019619157531, | |
| "grad_norm": 1.2133063077926636, | |
| "learning_rate": 6.147119600233758e-05, | |
| "loss": 2.1781, | |
| "step": 91 | |
| }, | |
| { | |
| "epoch": 0.10617426428159261, | |
| "grad_norm": 1.1791496276855469, | |
| "learning_rate": 6.066498153718735e-05, | |
| "loss": 2.2387, | |
| "step": 92 | |
| }, | |
| { | |
| "epoch": 0.10732833237160992, | |
| "grad_norm": 1.2252390384674072, | |
| "learning_rate": 5.985585137257401e-05, | |
| "loss": 2.1686, | |
| "step": 93 | |
| }, | |
| { | |
| "epoch": 0.10848240046162723, | |
| "grad_norm": 1.2614771127700806, | |
| "learning_rate": 5.90440267166055e-05, | |
| "loss": 2.2137, | |
| "step": 94 | |
| }, | |
| { | |
| "epoch": 0.10963646855164455, | |
| "grad_norm": 1.2798051834106445, | |
| "learning_rate": 5.8229729514036705e-05, | |
| "loss": 2.2652, | |
| "step": 95 | |
| }, | |
| { | |
| "epoch": 0.11079053664166186, | |
| "grad_norm": 1.4063845872879028, | |
| "learning_rate": 5.74131823855921e-05, | |
| "loss": 2.2172, | |
| "step": 96 | |
| }, | |
| { | |
| "epoch": 0.11194460473167916, | |
| "grad_norm": 1.587537407875061, | |
| "learning_rate": 5.6594608567103456e-05, | |
| "loss": 2.1679, | |
| "step": 97 | |
| }, | |
| { | |
| "epoch": 0.11309867282169649, | |
| "grad_norm": 1.5684443712234497, | |
| "learning_rate": 5.577423184847932e-05, | |
| "loss": 2.3385, | |
| "step": 98 | |
| }, | |
| { | |
| "epoch": 0.1142527409117138, | |
| "grad_norm": 2.7121994495391846, | |
| "learning_rate": 5.495227651252315e-05, | |
| "loss": 2.2971, | |
| "step": 99 | |
| }, | |
| { | |
| "epoch": 0.1154068090017311, | |
| "grad_norm": 2.7163429260253906, | |
| "learning_rate": 5.4128967273616625e-05, | |
| "loss": 2.5509, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.1154068090017311, | |
| "eval_loss": 2.359736919403076, | |
| "eval_runtime": 109.8627, | |
| "eval_samples_per_second": 13.289, | |
| "eval_steps_per_second": 3.322, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.11656087709174841, | |
| "grad_norm": 1.1242700815200806, | |
| "learning_rate": 5.330452921628497e-05, | |
| "loss": 1.8886, | |
| "step": 101 | |
| }, | |
| { | |
| "epoch": 0.11771494518176573, | |
| "grad_norm": 1.2791775465011597, | |
| "learning_rate": 5.247918773366112e-05, | |
| "loss": 2.2275, | |
| "step": 102 | |
| }, | |
| { | |
| "epoch": 0.11886901327178304, | |
| "grad_norm": 1.175013542175293, | |
| "learning_rate": 5.165316846586541e-05, | |
| "loss": 2.2796, | |
| "step": 103 | |
| }, | |
| { | |
| "epoch": 0.12002308136180034, | |
| "grad_norm": 1.0523555278778076, | |
| "learning_rate": 5.0826697238317935e-05, | |
| "loss": 2.0782, | |
| "step": 104 | |
| }, | |
| { | |
| "epoch": 0.12117714945181765, | |
| "grad_norm": 0.8484135866165161, | |
| "learning_rate": 5e-05, | |
| "loss": 1.9032, | |
| "step": 105 | |
| }, | |
| { | |
| "epoch": 0.12233121754183497, | |
| "grad_norm": 0.7368355989456177, | |
| "learning_rate": 4.917330276168208e-05, | |
| "loss": 2.1893, | |
| "step": 106 | |
| }, | |
| { | |
| "epoch": 0.12348528563185228, | |
| "grad_norm": 0.7255375385284424, | |
| "learning_rate": 4.834683153413459e-05, | |
| "loss": 2.1238, | |
| "step": 107 | |
| }, | |
| { | |
| "epoch": 0.12463935372186959, | |
| "grad_norm": 0.7863511443138123, | |
| "learning_rate": 4.7520812266338885e-05, | |
| "loss": 1.8973, | |
| "step": 108 | |
| }, | |
| { | |
| "epoch": 0.1257934218118869, | |
| "grad_norm": 0.760050892829895, | |
| "learning_rate": 4.669547078371504e-05, | |
| "loss": 2.1914, | |
| "step": 109 | |
| }, | |
| { | |
| "epoch": 0.12694748990190421, | |
| "grad_norm": 0.7366106510162354, | |
| "learning_rate": 4.5871032726383386e-05, | |
| "loss": 2.0323, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.12810155799192152, | |
| "grad_norm": 0.9344103932380676, | |
| "learning_rate": 4.504772348747687e-05, | |
| "loss": 2.1078, | |
| "step": 111 | |
| }, | |
| { | |
| "epoch": 0.12925562608193883, | |
| "grad_norm": 0.7587749361991882, | |
| "learning_rate": 4.4225768151520694e-05, | |
| "loss": 2.0855, | |
| "step": 112 | |
| }, | |
| { | |
| "epoch": 0.13040969417195614, | |
| "grad_norm": 0.8047494292259216, | |
| "learning_rate": 4.3405391432896555e-05, | |
| "loss": 2.1731, | |
| "step": 113 | |
| }, | |
| { | |
| "epoch": 0.13156376226197344, | |
| "grad_norm": 0.8073248863220215, | |
| "learning_rate": 4.2586817614407895e-05, | |
| "loss": 2.0467, | |
| "step": 114 | |
| }, | |
| { | |
| "epoch": 0.13271783035199078, | |
| "grad_norm": 0.7996666431427002, | |
| "learning_rate": 4.17702704859633e-05, | |
| "loss": 2.0532, | |
| "step": 115 | |
| }, | |
| { | |
| "epoch": 0.13387189844200809, | |
| "grad_norm": 0.8204259276390076, | |
| "learning_rate": 4.095597328339452e-05, | |
| "loss": 2.0689, | |
| "step": 116 | |
| }, | |
| { | |
| "epoch": 0.1350259665320254, | |
| "grad_norm": 0.7988185286521912, | |
| "learning_rate": 4.0144148627425993e-05, | |
| "loss": 2.0925, | |
| "step": 117 | |
| }, | |
| { | |
| "epoch": 0.1361800346220427, | |
| "grad_norm": 0.8346726298332214, | |
| "learning_rate": 3.933501846281267e-05, | |
| "loss": 2.1164, | |
| "step": 118 | |
| }, | |
| { | |
| "epoch": 0.13733410271206, | |
| "grad_norm": 0.8538673520088196, | |
| "learning_rate": 3.852880399766243e-05, | |
| "loss": 2.1109, | |
| "step": 119 | |
| }, | |
| { | |
| "epoch": 0.13848817080207732, | |
| "grad_norm": 0.8685901165008545, | |
| "learning_rate": 3.772572564296005e-05, | |
| "loss": 2.2344, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.13964223889209462, | |
| "grad_norm": 0.8483296036720276, | |
| "learning_rate": 3.6926002952309016e-05, | |
| "loss": 2.1122, | |
| "step": 121 | |
| }, | |
| { | |
| "epoch": 0.14079630698211196, | |
| "grad_norm": 0.8561789393424988, | |
| "learning_rate": 3.612985456190778e-05, | |
| "loss": 2.1018, | |
| "step": 122 | |
| }, | |
| { | |
| "epoch": 0.14195037507212926, | |
| "grad_norm": 0.8276510238647461, | |
| "learning_rate": 3.533749813077677e-05, | |
| "loss": 2.0867, | |
| "step": 123 | |
| }, | |
| { | |
| "epoch": 0.14310444316214657, | |
| "grad_norm": 0.8368616700172424, | |
| "learning_rate": 3.4549150281252636e-05, | |
| "loss": 1.9989, | |
| "step": 124 | |
| }, | |
| { | |
| "epoch": 0.14425851125216388, | |
| "grad_norm": 0.934032678604126, | |
| "learning_rate": 3.3765026539765834e-05, | |
| "loss": 2.2721, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 0.1454125793421812, | |
| "grad_norm": 0.9174278378486633, | |
| "learning_rate": 3.298534127791785e-05, | |
| "loss": 2.1676, | |
| "step": 126 | |
| }, | |
| { | |
| "epoch": 0.1465666474321985, | |
| "grad_norm": 0.8445408940315247, | |
| "learning_rate": 3.221030765387417e-05, | |
| "loss": 2.1227, | |
| "step": 127 | |
| }, | |
| { | |
| "epoch": 0.1477207155222158, | |
| "grad_norm": 0.8779769539833069, | |
| "learning_rate": 3.144013755408895e-05, | |
| "loss": 2.049, | |
| "step": 128 | |
| }, | |
| { | |
| "epoch": 0.1488747836122331, | |
| "grad_norm": 0.9407861828804016, | |
| "learning_rate": 3.0675041535377405e-05, | |
| "loss": 2.0843, | |
| "step": 129 | |
| }, | |
| { | |
| "epoch": 0.15002885170225044, | |
| "grad_norm": 0.8824257850646973, | |
| "learning_rate": 2.991522876735154e-05, | |
| "loss": 2.1531, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.15118291979226775, | |
| "grad_norm": 0.9266088008880615, | |
| "learning_rate": 2.916090697523549e-05, | |
| "loss": 2.0011, | |
| "step": 131 | |
| }, | |
| { | |
| "epoch": 0.15233698788228506, | |
| "grad_norm": 1.0334736108779907, | |
| "learning_rate": 2.8412282383075363e-05, | |
| "loss": 2.252, | |
| "step": 132 | |
| }, | |
| { | |
| "epoch": 0.15349105597230236, | |
| "grad_norm": 0.9971547722816467, | |
| "learning_rate": 2.766955965735968e-05, | |
| "loss": 2.094, | |
| "step": 133 | |
| }, | |
| { | |
| "epoch": 0.15464512406231967, | |
| "grad_norm": 0.9868890643119812, | |
| "learning_rate": 2.693294185106562e-05, | |
| "loss": 2.2888, | |
| "step": 134 | |
| }, | |
| { | |
| "epoch": 0.15579919215233698, | |
| "grad_norm": 1.0133064985275269, | |
| "learning_rate": 2.6202630348146324e-05, | |
| "loss": 2.2104, | |
| "step": 135 | |
| }, | |
| { | |
| "epoch": 0.1569532602423543, | |
| "grad_norm": 1.077724575996399, | |
| "learning_rate": 2.547882480847461e-05, | |
| "loss": 2.2726, | |
| "step": 136 | |
| }, | |
| { | |
| "epoch": 0.15810732833237162, | |
| "grad_norm": 0.997728705406189, | |
| "learning_rate": 2.476172311325783e-05, | |
| "loss": 2.09, | |
| "step": 137 | |
| }, | |
| { | |
| "epoch": 0.15926139642238893, | |
| "grad_norm": 1.0651929378509521, | |
| "learning_rate": 2.405152131093926e-05, | |
| "loss": 1.9946, | |
| "step": 138 | |
| }, | |
| { | |
| "epoch": 0.16041546451240624, | |
| "grad_norm": 1.062555193901062, | |
| "learning_rate": 2.3348413563600325e-05, | |
| "loss": 2.2394, | |
| "step": 139 | |
| }, | |
| { | |
| "epoch": 0.16156953260242354, | |
| "grad_norm": 1.2411359548568726, | |
| "learning_rate": 2.2652592093878666e-05, | |
| "loss": 2.1167, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.16272360069244085, | |
| "grad_norm": 1.2138582468032837, | |
| "learning_rate": 2.196424713241637e-05, | |
| "loss": 1.9873, | |
| "step": 141 | |
| }, | |
| { | |
| "epoch": 0.16387766878245816, | |
| "grad_norm": 1.2360583543777466, | |
| "learning_rate": 2.128356686585282e-05, | |
| "loss": 2.28, | |
| "step": 142 | |
| }, | |
| { | |
| "epoch": 0.16503173687247547, | |
| "grad_norm": 1.1820520162582397, | |
| "learning_rate": 2.061073738537635e-05, | |
| "loss": 2.158, | |
| "step": 143 | |
| }, | |
| { | |
| "epoch": 0.1661858049624928, | |
| "grad_norm": 1.2930243015289307, | |
| "learning_rate": 1.9945942635848748e-05, | |
| "loss": 2.2475, | |
| "step": 144 | |
| }, | |
| { | |
| "epoch": 0.1673398730525101, | |
| "grad_norm": 1.3382759094238281, | |
| "learning_rate": 1.928936436551661e-05, | |
| "loss": 2.3192, | |
| "step": 145 | |
| }, | |
| { | |
| "epoch": 0.16849394114252741, | |
| "grad_norm": 1.3789187669754028, | |
| "learning_rate": 1.8641182076323148e-05, | |
| "loss": 2.1857, | |
| "step": 146 | |
| }, | |
| { | |
| "epoch": 0.16964800923254472, | |
| "grad_norm": 1.3139618635177612, | |
| "learning_rate": 1.800157297483417e-05, | |
| "loss": 2.2008, | |
| "step": 147 | |
| }, | |
| { | |
| "epoch": 0.17080207732256203, | |
| "grad_norm": 1.5240283012390137, | |
| "learning_rate": 1.7370711923791567e-05, | |
| "loss": 2.3255, | |
| "step": 148 | |
| }, | |
| { | |
| "epoch": 0.17195614541257934, | |
| "grad_norm": 1.5187519788742065, | |
| "learning_rate": 1.6748771394307585e-05, | |
| "loss": 2.1988, | |
| "step": 149 | |
| }, | |
| { | |
| "epoch": 0.17311021350259664, | |
| "grad_norm": 2.440321922302246, | |
| "learning_rate": 1.6135921418712956e-05, | |
| "loss": 2.8243, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.17311021350259664, | |
| "eval_loss": 2.1568007469177246, | |
| "eval_runtime": 109.9329, | |
| "eval_samples_per_second": 13.281, | |
| "eval_steps_per_second": 3.32, | |
| "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": 2.262770368118784e+17, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |