| { | |
| "best_metric": 0.566459059715271, | |
| "best_model_checkpoint": "miner_id_24/checkpoint-150", | |
| "epoch": 0.5876591576885406, | |
| "eval_steps": 50, | |
| "global_step": 150, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0039177277179236044, | |
| "grad_norm": 0.2880200743675232, | |
| "learning_rate": 1e-05, | |
| "loss": 0.8274, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.0039177277179236044, | |
| "eval_loss": 1.168825626373291, | |
| "eval_runtime": 7.2476, | |
| "eval_samples_per_second": 59.33, | |
| "eval_steps_per_second": 14.901, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.007835455435847209, | |
| "grad_norm": 0.3449517786502838, | |
| "learning_rate": 2e-05, | |
| "loss": 0.8578, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.011753183153770812, | |
| "grad_norm": 0.3617977797985077, | |
| "learning_rate": 3e-05, | |
| "loss": 0.8814, | |
| "step": 3 | |
| }, | |
| { | |
| "epoch": 0.015670910871694418, | |
| "grad_norm": 0.38251858949661255, | |
| "learning_rate": 4e-05, | |
| "loss": 0.9613, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.019588638589618023, | |
| "grad_norm": 0.3866579532623291, | |
| "learning_rate": 5e-05, | |
| "loss": 0.9417, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.023506366307541625, | |
| "grad_norm": 0.3757135272026062, | |
| "learning_rate": 6e-05, | |
| "loss": 0.8947, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.02742409402546523, | |
| "grad_norm": 0.37219953536987305, | |
| "learning_rate": 7e-05, | |
| "loss": 0.956, | |
| "step": 7 | |
| }, | |
| { | |
| "epoch": 0.031341821743388835, | |
| "grad_norm": 0.3264732360839844, | |
| "learning_rate": 8e-05, | |
| "loss": 0.9629, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.03525954946131244, | |
| "grad_norm": 0.2868359684944153, | |
| "learning_rate": 9e-05, | |
| "loss": 0.9351, | |
| "step": 9 | |
| }, | |
| { | |
| "epoch": 0.039177277179236046, | |
| "grad_norm": 0.27399173378944397, | |
| "learning_rate": 0.0001, | |
| "loss": 0.8791, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.043095004897159644, | |
| "grad_norm": 0.25648829340934753, | |
| "learning_rate": 9.999316524962345e-05, | |
| "loss": 0.904, | |
| "step": 11 | |
| }, | |
| { | |
| "epoch": 0.04701273261508325, | |
| "grad_norm": 0.2375715970993042, | |
| "learning_rate": 9.997266286704631e-05, | |
| "loss": 0.7775, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 0.050930460333006855, | |
| "grad_norm": 0.23313941061496735, | |
| "learning_rate": 9.993849845741524e-05, | |
| "loss": 0.8446, | |
| "step": 13 | |
| }, | |
| { | |
| "epoch": 0.05484818805093046, | |
| "grad_norm": 0.21852728724479675, | |
| "learning_rate": 9.989068136093873e-05, | |
| "loss": 0.775, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 0.058765915768854066, | |
| "grad_norm": 0.19950824975967407, | |
| "learning_rate": 9.98292246503335e-05, | |
| "loss": 0.7156, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.06268364348677767, | |
| "grad_norm": 0.20977352559566498, | |
| "learning_rate": 9.975414512725057e-05, | |
| "loss": 0.7233, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 0.06660137120470128, | |
| "grad_norm": 0.2021847516298294, | |
| "learning_rate": 9.966546331768191e-05, | |
| "loss": 0.8003, | |
| "step": 17 | |
| }, | |
| { | |
| "epoch": 0.07051909892262488, | |
| "grad_norm": 0.22458495199680328, | |
| "learning_rate": 9.956320346634876e-05, | |
| "loss": 0.7392, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 0.07443682664054849, | |
| "grad_norm": 0.22804933786392212, | |
| "learning_rate": 9.944739353007344e-05, | |
| "loss": 0.7945, | |
| "step": 19 | |
| }, | |
| { | |
| "epoch": 0.07835455435847209, | |
| "grad_norm": 0.2243671864271164, | |
| "learning_rate": 9.931806517013612e-05, | |
| "loss": 0.7368, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.08227228207639568, | |
| "grad_norm": 0.2231484204530716, | |
| "learning_rate": 9.917525374361912e-05, | |
| "loss": 0.6725, | |
| "step": 21 | |
| }, | |
| { | |
| "epoch": 0.08619000979431929, | |
| "grad_norm": 0.217988058924675, | |
| "learning_rate": 9.901899829374047e-05, | |
| "loss": 0.6689, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 0.0901077375122429, | |
| "grad_norm": 0.2150304615497589, | |
| "learning_rate": 9.884934153917997e-05, | |
| "loss": 0.7145, | |
| "step": 23 | |
| }, | |
| { | |
| "epoch": 0.0940254652301665, | |
| "grad_norm": 0.1946570724248886, | |
| "learning_rate": 9.86663298624003e-05, | |
| "loss": 0.5993, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 0.0979431929480901, | |
| "grad_norm": 0.21674545109272003, | |
| "learning_rate": 9.847001329696653e-05, | |
| "loss": 0.6743, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.10186092066601371, | |
| "grad_norm": 0.2053242027759552, | |
| "learning_rate": 9.826044551386744e-05, | |
| "loss": 0.6807, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 0.10577864838393732, | |
| "grad_norm": 0.20809248089790344, | |
| "learning_rate": 9.803768380684242e-05, | |
| "loss": 0.6412, | |
| "step": 27 | |
| }, | |
| { | |
| "epoch": 0.10969637610186092, | |
| "grad_norm": 0.18792220950126648, | |
| "learning_rate": 9.780178907671789e-05, | |
| "loss": 0.6139, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 0.11361410381978453, | |
| "grad_norm": 0.21103435754776, | |
| "learning_rate": 9.755282581475769e-05, | |
| "loss": 0.6735, | |
| "step": 29 | |
| }, | |
| { | |
| "epoch": 0.11753183153770813, | |
| "grad_norm": 0.22956202924251556, | |
| "learning_rate": 9.729086208503174e-05, | |
| "loss": 0.7389, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.12144955925563174, | |
| "grad_norm": 0.2394503802061081, | |
| "learning_rate": 9.701596950580806e-05, | |
| "loss": 0.6543, | |
| "step": 31 | |
| }, | |
| { | |
| "epoch": 0.12536728697355534, | |
| "grad_norm": 0.20829223096370697, | |
| "learning_rate": 9.672822322997305e-05, | |
| "loss": 0.622, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 0.12928501469147893, | |
| "grad_norm": 0.20771794021129608, | |
| "learning_rate": 9.642770192448536e-05, | |
| "loss": 0.6015, | |
| "step": 33 | |
| }, | |
| { | |
| "epoch": 0.13320274240940255, | |
| "grad_norm": 0.17461727559566498, | |
| "learning_rate": 9.611448774886924e-05, | |
| "loss": 0.5575, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 0.13712047012732614, | |
| "grad_norm": 0.1794537901878357, | |
| "learning_rate": 9.578866633275288e-05, | |
| "loss": 0.6257, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.14103819784524976, | |
| "grad_norm": 0.18663041293621063, | |
| "learning_rate": 9.545032675245813e-05, | |
| "loss": 0.5959, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 0.14495592556317335, | |
| "grad_norm": 0.18721981346607208, | |
| "learning_rate": 9.509956150664796e-05, | |
| "loss": 0.6126, | |
| "step": 37 | |
| }, | |
| { | |
| "epoch": 0.14887365328109697, | |
| "grad_norm": 0.1981094628572464, | |
| "learning_rate": 9.473646649103818e-05, | |
| "loss": 0.6054, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 0.15279138099902057, | |
| "grad_norm": 0.19702383875846863, | |
| "learning_rate": 9.43611409721806e-05, | |
| "loss": 0.5315, | |
| "step": 39 | |
| }, | |
| { | |
| "epoch": 0.15670910871694418, | |
| "grad_norm": 0.2137853056192398, | |
| "learning_rate": 9.397368756032445e-05, | |
| "loss": 0.6874, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.16062683643486778, | |
| "grad_norm": 0.21116779744625092, | |
| "learning_rate": 9.357421218136386e-05, | |
| "loss": 0.6414, | |
| "step": 41 | |
| }, | |
| { | |
| "epoch": 0.16454456415279137, | |
| "grad_norm": 0.20851093530654907, | |
| "learning_rate": 9.316282404787871e-05, | |
| "loss": 0.6257, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 0.168462291870715, | |
| "grad_norm": 0.21066616475582123, | |
| "learning_rate": 9.273963562927695e-05, | |
| "loss": 0.5486, | |
| "step": 43 | |
| }, | |
| { | |
| "epoch": 0.17238001958863858, | |
| "grad_norm": 0.2145363986492157, | |
| "learning_rate": 9.230476262104677e-05, | |
| "loss": 0.5826, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 0.1762977473065622, | |
| "grad_norm": 0.24160481989383698, | |
| "learning_rate": 9.185832391312644e-05, | |
| "loss": 0.638, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 0.1802154750244858, | |
| "grad_norm": 0.23180629312992096, | |
| "learning_rate": 9.140044155740101e-05, | |
| "loss": 0.5407, | |
| "step": 46 | |
| }, | |
| { | |
| "epoch": 0.1841332027424094, | |
| "grad_norm": 0.2189411222934723, | |
| "learning_rate": 9.093124073433463e-05, | |
| "loss": 0.5233, | |
| "step": 47 | |
| }, | |
| { | |
| "epoch": 0.188050930460333, | |
| "grad_norm": 0.2373291701078415, | |
| "learning_rate": 9.045084971874738e-05, | |
| "loss": 0.5369, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 0.19196865817825662, | |
| "grad_norm": 0.23794688284397125, | |
| "learning_rate": 8.995939984474624e-05, | |
| "loss": 0.5349, | |
| "step": 49 | |
| }, | |
| { | |
| "epoch": 0.1958863858961802, | |
| "grad_norm": 0.2608213424682617, | |
| "learning_rate": 8.945702546981969e-05, | |
| "loss": 0.5191, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.1958863858961802, | |
| "eval_loss": 0.6147740483283997, | |
| "eval_runtime": 7.2375, | |
| "eval_samples_per_second": 59.413, | |
| "eval_steps_per_second": 14.922, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.19980411361410383, | |
| "grad_norm": 0.277538925409317, | |
| "learning_rate": 8.894386393810563e-05, | |
| "loss": 0.6745, | |
| "step": 51 | |
| }, | |
| { | |
| "epoch": 0.20372184133202742, | |
| "grad_norm": 0.26731616258621216, | |
| "learning_rate": 8.842005554284296e-05, | |
| "loss": 0.6651, | |
| "step": 52 | |
| }, | |
| { | |
| "epoch": 0.20763956904995104, | |
| "grad_norm": 0.2555927336215973, | |
| "learning_rate": 8.788574348801675e-05, | |
| "loss": 0.611, | |
| "step": 53 | |
| }, | |
| { | |
| "epoch": 0.21155729676787463, | |
| "grad_norm": 0.23091807961463928, | |
| "learning_rate": 8.73410738492077e-05, | |
| "loss": 0.6357, | |
| "step": 54 | |
| }, | |
| { | |
| "epoch": 0.21547502448579825, | |
| "grad_norm": 0.20790523290634155, | |
| "learning_rate": 8.678619553365659e-05, | |
| "loss": 0.6747, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 0.21939275220372184, | |
| "grad_norm": 0.14857007563114166, | |
| "learning_rate": 8.622126023955446e-05, | |
| "loss": 0.6119, | |
| "step": 56 | |
| }, | |
| { | |
| "epoch": 0.22331047992164543, | |
| "grad_norm": 0.14265301823616028, | |
| "learning_rate": 8.564642241456986e-05, | |
| "loss": 0.6194, | |
| "step": 57 | |
| }, | |
| { | |
| "epoch": 0.22722820763956905, | |
| "grad_norm": 0.15514330565929413, | |
| "learning_rate": 8.506183921362443e-05, | |
| "loss": 0.674, | |
| "step": 58 | |
| }, | |
| { | |
| "epoch": 0.23114593535749264, | |
| "grad_norm": 0.1494818925857544, | |
| "learning_rate": 8.44676704559283e-05, | |
| "loss": 0.5855, | |
| "step": 59 | |
| }, | |
| { | |
| "epoch": 0.23506366307541626, | |
| "grad_norm": 0.14865058660507202, | |
| "learning_rate": 8.386407858128706e-05, | |
| "loss": 0.5688, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.23898139079333985, | |
| "grad_norm": 0.17518377304077148, | |
| "learning_rate": 8.32512286056924e-05, | |
| "loss": 0.6678, | |
| "step": 61 | |
| }, | |
| { | |
| "epoch": 0.24289911851126347, | |
| "grad_norm": 0.1810987889766693, | |
| "learning_rate": 8.262928807620843e-05, | |
| "loss": 0.6014, | |
| "step": 62 | |
| }, | |
| { | |
| "epoch": 0.24681684622918706, | |
| "grad_norm": 0.17731715738773346, | |
| "learning_rate": 8.199842702516583e-05, | |
| "loss": 0.6388, | |
| "step": 63 | |
| }, | |
| { | |
| "epoch": 0.2507345739471107, | |
| "grad_norm": 0.1768607646226883, | |
| "learning_rate": 8.135881792367686e-05, | |
| "loss": 0.5846, | |
| "step": 64 | |
| }, | |
| { | |
| "epoch": 0.2546523016650343, | |
| "grad_norm": 0.15677671134471893, | |
| "learning_rate": 8.07106356344834e-05, | |
| "loss": 0.5649, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 0.25857002938295787, | |
| "grad_norm": 0.15399052202701569, | |
| "learning_rate": 8.005405736415126e-05, | |
| "loss": 0.6758, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 0.2624877571008815, | |
| "grad_norm": 0.16651038825511932, | |
| "learning_rate": 7.938926261462366e-05, | |
| "loss": 0.5923, | |
| "step": 67 | |
| }, | |
| { | |
| "epoch": 0.2664054848188051, | |
| "grad_norm": 0.16036267578601837, | |
| "learning_rate": 7.871643313414718e-05, | |
| "loss": 0.6119, | |
| "step": 68 | |
| }, | |
| { | |
| "epoch": 0.2703232125367287, | |
| "grad_norm": 0.16696076095104218, | |
| "learning_rate": 7.803575286758364e-05, | |
| "loss": 0.5891, | |
| "step": 69 | |
| }, | |
| { | |
| "epoch": 0.2742409402546523, | |
| "grad_norm": 0.1432018131017685, | |
| "learning_rate": 7.734740790612136e-05, | |
| "loss": 0.5993, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.2781586679725759, | |
| "grad_norm": 0.15339428186416626, | |
| "learning_rate": 7.66515864363997e-05, | |
| "loss": 0.6332, | |
| "step": 71 | |
| }, | |
| { | |
| "epoch": 0.2820763956904995, | |
| "grad_norm": 0.149267315864563, | |
| "learning_rate": 7.594847868906076e-05, | |
| "loss": 0.6104, | |
| "step": 72 | |
| }, | |
| { | |
| "epoch": 0.2859941234084231, | |
| "grad_norm": 0.15613606572151184, | |
| "learning_rate": 7.52382768867422e-05, | |
| "loss": 0.5618, | |
| "step": 73 | |
| }, | |
| { | |
| "epoch": 0.2899118511263467, | |
| "grad_norm": 0.15151549875736237, | |
| "learning_rate": 7.452117519152542e-05, | |
| "loss": 0.6099, | |
| "step": 74 | |
| }, | |
| { | |
| "epoch": 0.2938295788442703, | |
| "grad_norm": 0.15164095163345337, | |
| "learning_rate": 7.379736965185368e-05, | |
| "loss": 0.6257, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.29774730656219395, | |
| "grad_norm": 0.1576971560716629, | |
| "learning_rate": 7.30670581489344e-05, | |
| "loss": 0.674, | |
| "step": 76 | |
| }, | |
| { | |
| "epoch": 0.30166503428011754, | |
| "grad_norm": 0.18924656510353088, | |
| "learning_rate": 7.233044034264034e-05, | |
| "loss": 0.6037, | |
| "step": 77 | |
| }, | |
| { | |
| "epoch": 0.30558276199804113, | |
| "grad_norm": 0.1524207592010498, | |
| "learning_rate": 7.158771761692464e-05, | |
| "loss": 0.5804, | |
| "step": 78 | |
| }, | |
| { | |
| "epoch": 0.3095004897159647, | |
| "grad_norm": 0.15630418062210083, | |
| "learning_rate": 7.083909302476453e-05, | |
| "loss": 0.5844, | |
| "step": 79 | |
| }, | |
| { | |
| "epoch": 0.31341821743388837, | |
| "grad_norm": 0.17638805508613586, | |
| "learning_rate": 7.008477123264848e-05, | |
| "loss": 0.653, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.31733594515181196, | |
| "grad_norm": 0.16743972897529602, | |
| "learning_rate": 6.932495846462261e-05, | |
| "loss": 0.5317, | |
| "step": 81 | |
| }, | |
| { | |
| "epoch": 0.32125367286973555, | |
| "grad_norm": 0.14951741695404053, | |
| "learning_rate": 6.855986244591104e-05, | |
| "loss": 0.5181, | |
| "step": 82 | |
| }, | |
| { | |
| "epoch": 0.32517140058765914, | |
| "grad_norm": 0.1458686888217926, | |
| "learning_rate": 6.778969234612584e-05, | |
| "loss": 0.5814, | |
| "step": 83 | |
| }, | |
| { | |
| "epoch": 0.32908912830558273, | |
| "grad_norm": 0.152086541056633, | |
| "learning_rate": 6.701465872208216e-05, | |
| "loss": 0.5299, | |
| "step": 84 | |
| }, | |
| { | |
| "epoch": 0.3330068560235064, | |
| "grad_norm": 0.1878979653120041, | |
| "learning_rate": 6.623497346023418e-05, | |
| "loss": 0.5678, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 0.33692458374143, | |
| "grad_norm": 0.15952187776565552, | |
| "learning_rate": 6.545084971874738e-05, | |
| "loss": 0.4566, | |
| "step": 86 | |
| }, | |
| { | |
| "epoch": 0.34084231145935356, | |
| "grad_norm": 0.1914183795452118, | |
| "learning_rate": 6.466250186922325e-05, | |
| "loss": 0.5925, | |
| "step": 87 | |
| }, | |
| { | |
| "epoch": 0.34476003917727716, | |
| "grad_norm": 0.18178632855415344, | |
| "learning_rate": 6.387014543809223e-05, | |
| "loss": 0.5653, | |
| "step": 88 | |
| }, | |
| { | |
| "epoch": 0.3486777668952008, | |
| "grad_norm": 0.17960163950920105, | |
| "learning_rate": 6.307399704769099e-05, | |
| "loss": 0.5495, | |
| "step": 89 | |
| }, | |
| { | |
| "epoch": 0.3525954946131244, | |
| "grad_norm": 0.17011374235153198, | |
| "learning_rate": 6.227427435703997e-05, | |
| "loss": 0.5173, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.356513222331048, | |
| "grad_norm": 0.1940322369337082, | |
| "learning_rate": 6.147119600233758e-05, | |
| "loss": 0.6115, | |
| "step": 91 | |
| }, | |
| { | |
| "epoch": 0.3604309500489716, | |
| "grad_norm": 0.17651766538619995, | |
| "learning_rate": 6.066498153718735e-05, | |
| "loss": 0.4876, | |
| "step": 92 | |
| }, | |
| { | |
| "epoch": 0.3643486777668952, | |
| "grad_norm": 0.19913941621780396, | |
| "learning_rate": 5.985585137257401e-05, | |
| "loss": 0.5596, | |
| "step": 93 | |
| }, | |
| { | |
| "epoch": 0.3682664054848188, | |
| "grad_norm": 0.19726422429084778, | |
| "learning_rate": 5.90440267166055e-05, | |
| "loss": 0.5489, | |
| "step": 94 | |
| }, | |
| { | |
| "epoch": 0.3721841332027424, | |
| "grad_norm": 0.21293579041957855, | |
| "learning_rate": 5.8229729514036705e-05, | |
| "loss": 0.595, | |
| "step": 95 | |
| }, | |
| { | |
| "epoch": 0.376101860920666, | |
| "grad_norm": 0.20236584544181824, | |
| "learning_rate": 5.74131823855921e-05, | |
| "loss": 0.584, | |
| "step": 96 | |
| }, | |
| { | |
| "epoch": 0.38001958863858964, | |
| "grad_norm": 0.21212337911128998, | |
| "learning_rate": 5.6594608567103456e-05, | |
| "loss": 0.5082, | |
| "step": 97 | |
| }, | |
| { | |
| "epoch": 0.38393731635651324, | |
| "grad_norm": 0.19349698722362518, | |
| "learning_rate": 5.577423184847932e-05, | |
| "loss": 0.4305, | |
| "step": 98 | |
| }, | |
| { | |
| "epoch": 0.3878550440744368, | |
| "grad_norm": 0.21214741468429565, | |
| "learning_rate": 5.495227651252315e-05, | |
| "loss": 0.4532, | |
| "step": 99 | |
| }, | |
| { | |
| "epoch": 0.3917727717923604, | |
| "grad_norm": 0.3013063073158264, | |
| "learning_rate": 5.4128967273616625e-05, | |
| "loss": 0.5204, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.3917727717923604, | |
| "eval_loss": 0.5840380191802979, | |
| "eval_runtime": 7.2272, | |
| "eval_samples_per_second": 59.497, | |
| "eval_steps_per_second": 14.943, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.395690499510284, | |
| "grad_norm": 0.24198639392852783, | |
| "learning_rate": 5.330452921628497e-05, | |
| "loss": 0.7106, | |
| "step": 101 | |
| }, | |
| { | |
| "epoch": 0.39960822722820766, | |
| "grad_norm": 0.2336902767419815, | |
| "learning_rate": 5.247918773366112e-05, | |
| "loss": 0.6235, | |
| "step": 102 | |
| }, | |
| { | |
| "epoch": 0.40352595494613125, | |
| "grad_norm": 0.2363867163658142, | |
| "learning_rate": 5.165316846586541e-05, | |
| "loss": 0.678, | |
| "step": 103 | |
| }, | |
| { | |
| "epoch": 0.40744368266405484, | |
| "grad_norm": 0.22106525301933289, | |
| "learning_rate": 5.0826697238317935e-05, | |
| "loss": 0.6102, | |
| "step": 104 | |
| }, | |
| { | |
| "epoch": 0.41136141038197843, | |
| "grad_norm": 0.2046935111284256, | |
| "learning_rate": 5e-05, | |
| "loss": 0.6045, | |
| "step": 105 | |
| }, | |
| { | |
| "epoch": 0.4152791380999021, | |
| "grad_norm": 0.1781383901834488, | |
| "learning_rate": 4.917330276168208e-05, | |
| "loss": 0.6228, | |
| "step": 106 | |
| }, | |
| { | |
| "epoch": 0.41919686581782567, | |
| "grad_norm": 0.15508726239204407, | |
| "learning_rate": 4.834683153413459e-05, | |
| "loss": 0.6151, | |
| "step": 107 | |
| }, | |
| { | |
| "epoch": 0.42311459353574926, | |
| "grad_norm": 0.14032620191574097, | |
| "learning_rate": 4.7520812266338885e-05, | |
| "loss": 0.6497, | |
| "step": 108 | |
| }, | |
| { | |
| "epoch": 0.42703232125367285, | |
| "grad_norm": 0.15394900739192963, | |
| "learning_rate": 4.669547078371504e-05, | |
| "loss": 0.61, | |
| "step": 109 | |
| }, | |
| { | |
| "epoch": 0.4309500489715965, | |
| "grad_norm": 0.14519239962100983, | |
| "learning_rate": 4.5871032726383386e-05, | |
| "loss": 0.6356, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.4348677766895201, | |
| "grad_norm": 0.13110753893852234, | |
| "learning_rate": 4.504772348747687e-05, | |
| "loss": 0.5897, | |
| "step": 111 | |
| }, | |
| { | |
| "epoch": 0.4387855044074437, | |
| "grad_norm": 0.15517979860305786, | |
| "learning_rate": 4.4225768151520694e-05, | |
| "loss": 0.5716, | |
| "step": 112 | |
| }, | |
| { | |
| "epoch": 0.4427032321253673, | |
| "grad_norm": 0.1528444141149521, | |
| "learning_rate": 4.3405391432896555e-05, | |
| "loss": 0.598, | |
| "step": 113 | |
| }, | |
| { | |
| "epoch": 0.44662095984329087, | |
| "grad_norm": 0.1469126045703888, | |
| "learning_rate": 4.2586817614407895e-05, | |
| "loss": 0.6091, | |
| "step": 114 | |
| }, | |
| { | |
| "epoch": 0.4505386875612145, | |
| "grad_norm": 0.15315893292427063, | |
| "learning_rate": 4.17702704859633e-05, | |
| "loss": 0.6851, | |
| "step": 115 | |
| }, | |
| { | |
| "epoch": 0.4544564152791381, | |
| "grad_norm": 0.1631251573562622, | |
| "learning_rate": 4.095597328339452e-05, | |
| "loss": 0.6026, | |
| "step": 116 | |
| }, | |
| { | |
| "epoch": 0.4583741429970617, | |
| "grad_norm": 0.18032890558242798, | |
| "learning_rate": 4.0144148627425993e-05, | |
| "loss": 0.6629, | |
| "step": 117 | |
| }, | |
| { | |
| "epoch": 0.4622918707149853, | |
| "grad_norm": 0.16424377262592316, | |
| "learning_rate": 3.933501846281267e-05, | |
| "loss": 0.5814, | |
| "step": 118 | |
| }, | |
| { | |
| "epoch": 0.46620959843290893, | |
| "grad_norm": 0.1458495855331421, | |
| "learning_rate": 3.852880399766243e-05, | |
| "loss": 0.5606, | |
| "step": 119 | |
| }, | |
| { | |
| "epoch": 0.4701273261508325, | |
| "grad_norm": 0.17860834300518036, | |
| "learning_rate": 3.772572564296005e-05, | |
| "loss": 0.5829, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.4740450538687561, | |
| "grad_norm": 0.17693659663200378, | |
| "learning_rate": 3.6926002952309016e-05, | |
| "loss": 0.5929, | |
| "step": 121 | |
| }, | |
| { | |
| "epoch": 0.4779627815866797, | |
| "grad_norm": 0.15140914916992188, | |
| "learning_rate": 3.612985456190778e-05, | |
| "loss": 0.5438, | |
| "step": 122 | |
| }, | |
| { | |
| "epoch": 0.48188050930460335, | |
| "grad_norm": 0.15471051633358002, | |
| "learning_rate": 3.533749813077677e-05, | |
| "loss": 0.5496, | |
| "step": 123 | |
| }, | |
| { | |
| "epoch": 0.48579823702252695, | |
| "grad_norm": 0.17251479625701904, | |
| "learning_rate": 3.4549150281252636e-05, | |
| "loss": 0.6438, | |
| "step": 124 | |
| }, | |
| { | |
| "epoch": 0.48971596474045054, | |
| "grad_norm": 0.1837255358695984, | |
| "learning_rate": 3.3765026539765834e-05, | |
| "loss": 0.5606, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 0.49363369245837413, | |
| "grad_norm": 0.1563147008419037, | |
| "learning_rate": 3.298534127791785e-05, | |
| "loss": 0.5836, | |
| "step": 126 | |
| }, | |
| { | |
| "epoch": 0.4975514201762977, | |
| "grad_norm": 0.15731315314769745, | |
| "learning_rate": 3.221030765387417e-05, | |
| "loss": 0.5528, | |
| "step": 127 | |
| }, | |
| { | |
| "epoch": 0.5014691478942214, | |
| "grad_norm": 0.14928919076919556, | |
| "learning_rate": 3.144013755408895e-05, | |
| "loss": 0.5684, | |
| "step": 128 | |
| }, | |
| { | |
| "epoch": 0.5053868756121449, | |
| "grad_norm": 0.15740834176540375, | |
| "learning_rate": 3.0675041535377405e-05, | |
| "loss": 0.5198, | |
| "step": 129 | |
| }, | |
| { | |
| "epoch": 0.5093046033300686, | |
| "grad_norm": 0.15248724818229675, | |
| "learning_rate": 2.991522876735154e-05, | |
| "loss": 0.536, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.5132223310479922, | |
| "grad_norm": 0.18244415521621704, | |
| "learning_rate": 2.916090697523549e-05, | |
| "loss": 0.6128, | |
| "step": 131 | |
| }, | |
| { | |
| "epoch": 0.5171400587659157, | |
| "grad_norm": 0.16289204359054565, | |
| "learning_rate": 2.8412282383075363e-05, | |
| "loss": 0.5504, | |
| "step": 132 | |
| }, | |
| { | |
| "epoch": 0.5210577864838394, | |
| "grad_norm": 0.15426087379455566, | |
| "learning_rate": 2.766955965735968e-05, | |
| "loss": 0.5075, | |
| "step": 133 | |
| }, | |
| { | |
| "epoch": 0.524975514201763, | |
| "grad_norm": 0.1748393326997757, | |
| "learning_rate": 2.693294185106562e-05, | |
| "loss": 0.5745, | |
| "step": 134 | |
| }, | |
| { | |
| "epoch": 0.5288932419196866, | |
| "grad_norm": 0.19530609250068665, | |
| "learning_rate": 2.6202630348146324e-05, | |
| "loss": 0.6085, | |
| "step": 135 | |
| }, | |
| { | |
| "epoch": 0.5328109696376102, | |
| "grad_norm": 0.17334362864494324, | |
| "learning_rate": 2.547882480847461e-05, | |
| "loss": 0.5229, | |
| "step": 136 | |
| }, | |
| { | |
| "epoch": 0.5367286973555337, | |
| "grad_norm": 0.17231786251068115, | |
| "learning_rate": 2.476172311325783e-05, | |
| "loss": 0.5617, | |
| "step": 137 | |
| }, | |
| { | |
| "epoch": 0.5406464250734574, | |
| "grad_norm": 0.20302419364452362, | |
| "learning_rate": 2.405152131093926e-05, | |
| "loss": 0.5756, | |
| "step": 138 | |
| }, | |
| { | |
| "epoch": 0.544564152791381, | |
| "grad_norm": 0.18278081715106964, | |
| "learning_rate": 2.3348413563600325e-05, | |
| "loss": 0.5399, | |
| "step": 139 | |
| }, | |
| { | |
| "epoch": 0.5484818805093046, | |
| "grad_norm": 0.21221497654914856, | |
| "learning_rate": 2.2652592093878666e-05, | |
| "loss": 0.5631, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.5523996082272282, | |
| "grad_norm": 0.18830539286136627, | |
| "learning_rate": 2.196424713241637e-05, | |
| "loss": 0.6303, | |
| "step": 141 | |
| }, | |
| { | |
| "epoch": 0.5563173359451518, | |
| "grad_norm": 0.18480455875396729, | |
| "learning_rate": 2.128356686585282e-05, | |
| "loss": 0.5266, | |
| "step": 142 | |
| }, | |
| { | |
| "epoch": 0.5602350636630754, | |
| "grad_norm": 0.17065569758415222, | |
| "learning_rate": 2.061073738537635e-05, | |
| "loss": 0.4521, | |
| "step": 143 | |
| }, | |
| { | |
| "epoch": 0.564152791380999, | |
| "grad_norm": 0.20241571962833405, | |
| "learning_rate": 1.9945942635848748e-05, | |
| "loss": 0.5559, | |
| "step": 144 | |
| }, | |
| { | |
| "epoch": 0.5680705190989226, | |
| "grad_norm": 0.20598803460597992, | |
| "learning_rate": 1.928936436551661e-05, | |
| "loss": 0.5682, | |
| "step": 145 | |
| }, | |
| { | |
| "epoch": 0.5719882468168462, | |
| "grad_norm": 0.2017568200826645, | |
| "learning_rate": 1.8641182076323148e-05, | |
| "loss": 0.4562, | |
| "step": 146 | |
| }, | |
| { | |
| "epoch": 0.5759059745347699, | |
| "grad_norm": 0.2124551236629486, | |
| "learning_rate": 1.800157297483417e-05, | |
| "loss": 0.4943, | |
| "step": 147 | |
| }, | |
| { | |
| "epoch": 0.5798237022526934, | |
| "grad_norm": 0.22031304240226746, | |
| "learning_rate": 1.7370711923791567e-05, | |
| "loss": 0.4337, | |
| "step": 148 | |
| }, | |
| { | |
| "epoch": 0.5837414299706171, | |
| "grad_norm": 0.2700269818305969, | |
| "learning_rate": 1.6748771394307585e-05, | |
| "loss": 0.4131, | |
| "step": 149 | |
| }, | |
| { | |
| "epoch": 0.5876591576885406, | |
| "grad_norm": 0.28507000207901, | |
| "learning_rate": 1.6135921418712956e-05, | |
| "loss": 0.4754, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.5876591576885406, | |
| "eval_loss": 0.566459059715271, | |
| "eval_runtime": 7.2467, | |
| "eval_samples_per_second": 59.337, | |
| "eval_steps_per_second": 14.903, | |
| "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": 1.0303319310336e+16, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |