| { | |
| "best_metric": 1.0405155420303345, | |
| "best_model_checkpoint": "miner_id_24/checkpoint-150", | |
| "epoch": 0.1470227885322225, | |
| "eval_steps": 50, | |
| "global_step": 150, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.00098015192354815, | |
| "grad_norm": 5.174369812011719, | |
| "learning_rate": 1e-05, | |
| "loss": 1.7379, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.00098015192354815, | |
| "eval_loss": 1.4601154327392578, | |
| "eval_runtime": 157.5868, | |
| "eval_samples_per_second": 10.908, | |
| "eval_steps_per_second": 2.729, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.0019603038470963, | |
| "grad_norm": 5.3734211921691895, | |
| "learning_rate": 2e-05, | |
| "loss": 1.6634, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.00294045577064445, | |
| "grad_norm": 6.260550022125244, | |
| "learning_rate": 3e-05, | |
| "loss": 1.7559, | |
| "step": 3 | |
| }, | |
| { | |
| "epoch": 0.0039206076941926, | |
| "grad_norm": 6.638162136077881, | |
| "learning_rate": 4e-05, | |
| "loss": 1.6658, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.00490075961774075, | |
| "grad_norm": 6.897322177886963, | |
| "learning_rate": 5e-05, | |
| "loss": 1.595, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.0058809115412889, | |
| "grad_norm": 11.182191848754883, | |
| "learning_rate": 6e-05, | |
| "loss": 1.7225, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.00686106346483705, | |
| "grad_norm": 6.777177333831787, | |
| "learning_rate": 7e-05, | |
| "loss": 1.8576, | |
| "step": 7 | |
| }, | |
| { | |
| "epoch": 0.0078412153883852, | |
| "grad_norm": 12.603493690490723, | |
| "learning_rate": 8e-05, | |
| "loss": 1.6802, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.008821367311933349, | |
| "grad_norm": 12.849211692810059, | |
| "learning_rate": 9e-05, | |
| "loss": 1.7667, | |
| "step": 9 | |
| }, | |
| { | |
| "epoch": 0.0098015192354815, | |
| "grad_norm": 11.883842468261719, | |
| "learning_rate": 0.0001, | |
| "loss": 1.7196, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.01078167115902965, | |
| "grad_norm": 10.308212280273438, | |
| "learning_rate": 9.999316524962345e-05, | |
| "loss": 1.6776, | |
| "step": 11 | |
| }, | |
| { | |
| "epoch": 0.0117618230825778, | |
| "grad_norm": 10.808917999267578, | |
| "learning_rate": 9.997266286704631e-05, | |
| "loss": 1.6303, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 0.01274197500612595, | |
| "grad_norm": 12.036140441894531, | |
| "learning_rate": 9.993849845741524e-05, | |
| "loss": 1.5666, | |
| "step": 13 | |
| }, | |
| { | |
| "epoch": 0.0137221269296741, | |
| "grad_norm": 12.335501670837402, | |
| "learning_rate": 9.989068136093873e-05, | |
| "loss": 1.2813, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 0.014702278853222249, | |
| "grad_norm": 12.53278923034668, | |
| "learning_rate": 9.98292246503335e-05, | |
| "loss": 1.2934, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.0156824307767704, | |
| "grad_norm": 23.434144973754883, | |
| "learning_rate": 9.975414512725057e-05, | |
| "loss": 1.3193, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 0.01666258270031855, | |
| "grad_norm": 14.406500816345215, | |
| "learning_rate": 9.966546331768191e-05, | |
| "loss": 1.308, | |
| "step": 17 | |
| }, | |
| { | |
| "epoch": 0.017642734623866698, | |
| "grad_norm": 17.926889419555664, | |
| "learning_rate": 9.956320346634876e-05, | |
| "loss": 1.3832, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 0.018622886547414848, | |
| "grad_norm": 19.3801326751709, | |
| "learning_rate": 9.944739353007344e-05, | |
| "loss": 1.1263, | |
| "step": 19 | |
| }, | |
| { | |
| "epoch": 0.019603038470963, | |
| "grad_norm": 19.59815216064453, | |
| "learning_rate": 9.931806517013612e-05, | |
| "loss": 1.0577, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.02058319039451115, | |
| "grad_norm": 16.549386978149414, | |
| "learning_rate": 9.917525374361912e-05, | |
| "loss": 1.2888, | |
| "step": 21 | |
| }, | |
| { | |
| "epoch": 0.0215633423180593, | |
| "grad_norm": 14.710787773132324, | |
| "learning_rate": 9.901899829374047e-05, | |
| "loss": 1.2167, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 0.02254349424160745, | |
| "grad_norm": 16.740129470825195, | |
| "learning_rate": 9.884934153917997e-05, | |
| "loss": 1.0853, | |
| "step": 23 | |
| }, | |
| { | |
| "epoch": 0.0235236461651556, | |
| "grad_norm": 17.446712493896484, | |
| "learning_rate": 9.86663298624003e-05, | |
| "loss": 1.2447, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 0.02450379808870375, | |
| "grad_norm": 15.809967041015625, | |
| "learning_rate": 9.847001329696653e-05, | |
| "loss": 1.2065, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.0254839500122519, | |
| "grad_norm": 23.9841251373291, | |
| "learning_rate": 9.826044551386744e-05, | |
| "loss": 1.2515, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 0.02646410193580005, | |
| "grad_norm": 20.032608032226562, | |
| "learning_rate": 9.803768380684242e-05, | |
| "loss": 1.3494, | |
| "step": 27 | |
| }, | |
| { | |
| "epoch": 0.0274442538593482, | |
| "grad_norm": 13.666933059692383, | |
| "learning_rate": 9.780178907671789e-05, | |
| "loss": 1.1337, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 0.028424405782896348, | |
| "grad_norm": 19.81943702697754, | |
| "learning_rate": 9.755282581475769e-05, | |
| "loss": 1.2332, | |
| "step": 29 | |
| }, | |
| { | |
| "epoch": 0.029404557706444498, | |
| "grad_norm": 19.173477172851562, | |
| "learning_rate": 9.729086208503174e-05, | |
| "loss": 1.2541, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.030384709629992648, | |
| "grad_norm": 20.487085342407227, | |
| "learning_rate": 9.701596950580806e-05, | |
| "loss": 1.2556, | |
| "step": 31 | |
| }, | |
| { | |
| "epoch": 0.0313648615535408, | |
| "grad_norm": 30.30592155456543, | |
| "learning_rate": 9.672822322997305e-05, | |
| "loss": 1.3247, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 0.03234501347708895, | |
| "grad_norm": 19.03827476501465, | |
| "learning_rate": 9.642770192448536e-05, | |
| "loss": 1.1559, | |
| "step": 33 | |
| }, | |
| { | |
| "epoch": 0.0333251654006371, | |
| "grad_norm": 18.113021850585938, | |
| "learning_rate": 9.611448774886924e-05, | |
| "loss": 1.3122, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 0.03430531732418525, | |
| "grad_norm": 18.305837631225586, | |
| "learning_rate": 9.578866633275288e-05, | |
| "loss": 1.1731, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.035285469247733396, | |
| "grad_norm": 19.005558013916016, | |
| "learning_rate": 9.545032675245813e-05, | |
| "loss": 1.147, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 0.03626562117128155, | |
| "grad_norm": 22.000139236450195, | |
| "learning_rate": 9.509956150664796e-05, | |
| "loss": 1.3035, | |
| "step": 37 | |
| }, | |
| { | |
| "epoch": 0.037245773094829696, | |
| "grad_norm": 14.228368759155273, | |
| "learning_rate": 9.473646649103818e-05, | |
| "loss": 1.1039, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 0.03822592501837785, | |
| "grad_norm": 18.42213249206543, | |
| "learning_rate": 9.43611409721806e-05, | |
| "loss": 1.3188, | |
| "step": 39 | |
| }, | |
| { | |
| "epoch": 0.039206076941926, | |
| "grad_norm": 13.70470142364502, | |
| "learning_rate": 9.397368756032445e-05, | |
| "loss": 1.2816, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.04018622886547415, | |
| "grad_norm": 20.462026596069336, | |
| "learning_rate": 9.357421218136386e-05, | |
| "loss": 1.1737, | |
| "step": 41 | |
| }, | |
| { | |
| "epoch": 0.0411663807890223, | |
| "grad_norm": 13.914624214172363, | |
| "learning_rate": 9.316282404787871e-05, | |
| "loss": 1.0558, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 0.04214653271257045, | |
| "grad_norm": 15.328235626220703, | |
| "learning_rate": 9.273963562927695e-05, | |
| "loss": 1.2211, | |
| "step": 43 | |
| }, | |
| { | |
| "epoch": 0.0431266846361186, | |
| "grad_norm": 11.389966011047363, | |
| "learning_rate": 9.230476262104677e-05, | |
| "loss": 1.1727, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 0.04410683655966675, | |
| "grad_norm": 16.29636001586914, | |
| "learning_rate": 9.185832391312644e-05, | |
| "loss": 1.2227, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 0.0450869884832149, | |
| "grad_norm": 12.993414878845215, | |
| "learning_rate": 9.140044155740101e-05, | |
| "loss": 1.1584, | |
| "step": 46 | |
| }, | |
| { | |
| "epoch": 0.046067140406763046, | |
| "grad_norm": 12.752553939819336, | |
| "learning_rate": 9.093124073433463e-05, | |
| "loss": 1.1872, | |
| "step": 47 | |
| }, | |
| { | |
| "epoch": 0.0470472923303112, | |
| "grad_norm": 13.57178020477295, | |
| "learning_rate": 9.045084971874738e-05, | |
| "loss": 1.278, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 0.048027444253859346, | |
| "grad_norm": 11.074052810668945, | |
| "learning_rate": 8.995939984474624e-05, | |
| "loss": 1.003, | |
| "step": 49 | |
| }, | |
| { | |
| "epoch": 0.0490075961774075, | |
| "grad_norm": 12.612610816955566, | |
| "learning_rate": 8.945702546981969e-05, | |
| "loss": 1.1835, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.0490075961774075, | |
| "eval_loss": 1.1867510080337524, | |
| "eval_runtime": 157.5965, | |
| "eval_samples_per_second": 10.908, | |
| "eval_steps_per_second": 2.728, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.049987748100955645, | |
| "grad_norm": 10.249678611755371, | |
| "learning_rate": 8.894386393810563e-05, | |
| "loss": 1.577, | |
| "step": 51 | |
| }, | |
| { | |
| "epoch": 0.0509679000245038, | |
| "grad_norm": 11.168072700500488, | |
| "learning_rate": 8.842005554284296e-05, | |
| "loss": 1.4897, | |
| "step": 52 | |
| }, | |
| { | |
| "epoch": 0.05194805194805195, | |
| "grad_norm": 16.988407135009766, | |
| "learning_rate": 8.788574348801675e-05, | |
| "loss": 1.337, | |
| "step": 53 | |
| }, | |
| { | |
| "epoch": 0.0529282038716001, | |
| "grad_norm": 13.803077697753906, | |
| "learning_rate": 8.73410738492077e-05, | |
| "loss": 1.4961, | |
| "step": 54 | |
| }, | |
| { | |
| "epoch": 0.05390835579514825, | |
| "grad_norm": 13.419051170349121, | |
| "learning_rate": 8.678619553365659e-05, | |
| "loss": 1.4239, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 0.0548885077186964, | |
| "grad_norm": 12.883716583251953, | |
| "learning_rate": 8.622126023955446e-05, | |
| "loss": 1.4591, | |
| "step": 56 | |
| }, | |
| { | |
| "epoch": 0.05586865964224455, | |
| "grad_norm": 21.91691780090332, | |
| "learning_rate": 8.564642241456986e-05, | |
| "loss": 1.4506, | |
| "step": 57 | |
| }, | |
| { | |
| "epoch": 0.056848811565792697, | |
| "grad_norm": 8.715700149536133, | |
| "learning_rate": 8.506183921362443e-05, | |
| "loss": 1.4636, | |
| "step": 58 | |
| }, | |
| { | |
| "epoch": 0.05782896348934085, | |
| "grad_norm": 9.098658561706543, | |
| "learning_rate": 8.44676704559283e-05, | |
| "loss": 1.4697, | |
| "step": 59 | |
| }, | |
| { | |
| "epoch": 0.058809115412888996, | |
| "grad_norm": 8.151817321777344, | |
| "learning_rate": 8.386407858128706e-05, | |
| "loss": 1.3484, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.05978926733643715, | |
| "grad_norm": 9.347545623779297, | |
| "learning_rate": 8.32512286056924e-05, | |
| "loss": 1.3751, | |
| "step": 61 | |
| }, | |
| { | |
| "epoch": 0.060769419259985295, | |
| "grad_norm": 6.9391021728515625, | |
| "learning_rate": 8.262928807620843e-05, | |
| "loss": 1.4251, | |
| "step": 62 | |
| }, | |
| { | |
| "epoch": 0.06174957118353345, | |
| "grad_norm": 7.633073806762695, | |
| "learning_rate": 8.199842702516583e-05, | |
| "loss": 1.2741, | |
| "step": 63 | |
| }, | |
| { | |
| "epoch": 0.0627297231070816, | |
| "grad_norm": 11.888652801513672, | |
| "learning_rate": 8.135881792367686e-05, | |
| "loss": 1.1978, | |
| "step": 64 | |
| }, | |
| { | |
| "epoch": 0.06370987503062975, | |
| "grad_norm": 7.112800598144531, | |
| "learning_rate": 8.07106356344834e-05, | |
| "loss": 1.2832, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 0.0646900269541779, | |
| "grad_norm": 6.586153984069824, | |
| "learning_rate": 8.005405736415126e-05, | |
| "loss": 1.1862, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 0.06567017887772605, | |
| "grad_norm": 4.565048694610596, | |
| "learning_rate": 7.938926261462366e-05, | |
| "loss": 1.0571, | |
| "step": 67 | |
| }, | |
| { | |
| "epoch": 0.0666503308012742, | |
| "grad_norm": 5.686751842498779, | |
| "learning_rate": 7.871643313414718e-05, | |
| "loss": 1.0384, | |
| "step": 68 | |
| }, | |
| { | |
| "epoch": 0.06763048272482235, | |
| "grad_norm": 4.812401294708252, | |
| "learning_rate": 7.803575286758364e-05, | |
| "loss": 1.0416, | |
| "step": 69 | |
| }, | |
| { | |
| "epoch": 0.0686106346483705, | |
| "grad_norm": 5.9184136390686035, | |
| "learning_rate": 7.734740790612136e-05, | |
| "loss": 1.0038, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.06959078657191865, | |
| "grad_norm": 4.310882568359375, | |
| "learning_rate": 7.66515864363997e-05, | |
| "loss": 0.9793, | |
| "step": 71 | |
| }, | |
| { | |
| "epoch": 0.07057093849546679, | |
| "grad_norm": 6.855759143829346, | |
| "learning_rate": 7.594847868906076e-05, | |
| "loss": 1.1017, | |
| "step": 72 | |
| }, | |
| { | |
| "epoch": 0.07155109041901495, | |
| "grad_norm": 4.950701713562012, | |
| "learning_rate": 7.52382768867422e-05, | |
| "loss": 1.0, | |
| "step": 73 | |
| }, | |
| { | |
| "epoch": 0.0725312423425631, | |
| "grad_norm": 4.762784481048584, | |
| "learning_rate": 7.452117519152542e-05, | |
| "loss": 1.1238, | |
| "step": 74 | |
| }, | |
| { | |
| "epoch": 0.07351139426611125, | |
| "grad_norm": 4.812014102935791, | |
| "learning_rate": 7.379736965185368e-05, | |
| "loss": 1.0404, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.07449154618965939, | |
| "grad_norm": 4.174313545227051, | |
| "learning_rate": 7.30670581489344e-05, | |
| "loss": 1.1405, | |
| "step": 76 | |
| }, | |
| { | |
| "epoch": 0.07547169811320754, | |
| "grad_norm": 4.601934909820557, | |
| "learning_rate": 7.233044034264034e-05, | |
| "loss": 0.9509, | |
| "step": 77 | |
| }, | |
| { | |
| "epoch": 0.0764518500367557, | |
| "grad_norm": 4.325131893157959, | |
| "learning_rate": 7.158771761692464e-05, | |
| "loss": 1.0177, | |
| "step": 78 | |
| }, | |
| { | |
| "epoch": 0.07743200196030385, | |
| "grad_norm": 4.170076370239258, | |
| "learning_rate": 7.083909302476453e-05, | |
| "loss": 1.0479, | |
| "step": 79 | |
| }, | |
| { | |
| "epoch": 0.078412153883852, | |
| "grad_norm": 4.9406633377075195, | |
| "learning_rate": 7.008477123264848e-05, | |
| "loss": 1.037, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.07939230580740014, | |
| "grad_norm": 4.136759281158447, | |
| "learning_rate": 6.932495846462261e-05, | |
| "loss": 1.0884, | |
| "step": 81 | |
| }, | |
| { | |
| "epoch": 0.0803724577309483, | |
| "grad_norm": 4.374018669128418, | |
| "learning_rate": 6.855986244591104e-05, | |
| "loss": 0.9557, | |
| "step": 82 | |
| }, | |
| { | |
| "epoch": 0.08135260965449645, | |
| "grad_norm": 3.8246161937713623, | |
| "learning_rate": 6.778969234612584e-05, | |
| "loss": 1.0854, | |
| "step": 83 | |
| }, | |
| { | |
| "epoch": 0.0823327615780446, | |
| "grad_norm": 3.964921712875366, | |
| "learning_rate": 6.701465872208216e-05, | |
| "loss": 1.0668, | |
| "step": 84 | |
| }, | |
| { | |
| "epoch": 0.08331291350159274, | |
| "grad_norm": 4.285429000854492, | |
| "learning_rate": 6.623497346023418e-05, | |
| "loss": 1.0743, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 0.0842930654251409, | |
| "grad_norm": 4.323607444763184, | |
| "learning_rate": 6.545084971874738e-05, | |
| "loss": 1.0444, | |
| "step": 86 | |
| }, | |
| { | |
| "epoch": 0.08527321734868905, | |
| "grad_norm": 4.902374744415283, | |
| "learning_rate": 6.466250186922325e-05, | |
| "loss": 1.1086, | |
| "step": 87 | |
| }, | |
| { | |
| "epoch": 0.0862533692722372, | |
| "grad_norm": 3.8147130012512207, | |
| "learning_rate": 6.387014543809223e-05, | |
| "loss": 0.9245, | |
| "step": 88 | |
| }, | |
| { | |
| "epoch": 0.08723352119578534, | |
| "grad_norm": 4.760003566741943, | |
| "learning_rate": 6.307399704769099e-05, | |
| "loss": 1.0365, | |
| "step": 89 | |
| }, | |
| { | |
| "epoch": 0.0882136731193335, | |
| "grad_norm": 4.650182723999023, | |
| "learning_rate": 6.227427435703997e-05, | |
| "loss": 1.1253, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.08919382504288165, | |
| "grad_norm": 3.8675501346588135, | |
| "learning_rate": 6.147119600233758e-05, | |
| "loss": 1.0713, | |
| "step": 91 | |
| }, | |
| { | |
| "epoch": 0.0901739769664298, | |
| "grad_norm": 3.941523551940918, | |
| "learning_rate": 6.066498153718735e-05, | |
| "loss": 0.9993, | |
| "step": 92 | |
| }, | |
| { | |
| "epoch": 0.09115412888997795, | |
| "grad_norm": 4.807412147521973, | |
| "learning_rate": 5.985585137257401e-05, | |
| "loss": 1.0744, | |
| "step": 93 | |
| }, | |
| { | |
| "epoch": 0.09213428081352609, | |
| "grad_norm": 4.303157329559326, | |
| "learning_rate": 5.90440267166055e-05, | |
| "loss": 1.0588, | |
| "step": 94 | |
| }, | |
| { | |
| "epoch": 0.09311443273707425, | |
| "grad_norm": 6.031392574310303, | |
| "learning_rate": 5.8229729514036705e-05, | |
| "loss": 0.9678, | |
| "step": 95 | |
| }, | |
| { | |
| "epoch": 0.0940945846606224, | |
| "grad_norm": 4.7057905197143555, | |
| "learning_rate": 5.74131823855921e-05, | |
| "loss": 0.9343, | |
| "step": 96 | |
| }, | |
| { | |
| "epoch": 0.09507473658417055, | |
| "grad_norm": 4.901125907897949, | |
| "learning_rate": 5.6594608567103456e-05, | |
| "loss": 0.966, | |
| "step": 97 | |
| }, | |
| { | |
| "epoch": 0.09605488850771869, | |
| "grad_norm": 4.58914041519165, | |
| "learning_rate": 5.577423184847932e-05, | |
| "loss": 1.002, | |
| "step": 98 | |
| }, | |
| { | |
| "epoch": 0.09703504043126684, | |
| "grad_norm": 5.568334579467773, | |
| "learning_rate": 5.495227651252315e-05, | |
| "loss": 0.8543, | |
| "step": 99 | |
| }, | |
| { | |
| "epoch": 0.098015192354815, | |
| "grad_norm": 5.970452308654785, | |
| "learning_rate": 5.4128967273616625e-05, | |
| "loss": 0.9819, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.098015192354815, | |
| "eval_loss": 1.0481526851654053, | |
| "eval_runtime": 158.0306, | |
| "eval_samples_per_second": 10.878, | |
| "eval_steps_per_second": 2.721, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.09899534427836315, | |
| "grad_norm": 5.198331356048584, | |
| "learning_rate": 5.330452921628497e-05, | |
| "loss": 1.5161, | |
| "step": 101 | |
| }, | |
| { | |
| "epoch": 0.09997549620191129, | |
| "grad_norm": 5.3907952308654785, | |
| "learning_rate": 5.247918773366112e-05, | |
| "loss": 1.2511, | |
| "step": 102 | |
| }, | |
| { | |
| "epoch": 0.10095564812545944, | |
| "grad_norm": 6.977193832397461, | |
| "learning_rate": 5.165316846586541e-05, | |
| "loss": 1.3599, | |
| "step": 103 | |
| }, | |
| { | |
| "epoch": 0.1019358000490076, | |
| "grad_norm": 5.693919658660889, | |
| "learning_rate": 5.0826697238317935e-05, | |
| "loss": 1.2146, | |
| "step": 104 | |
| }, | |
| { | |
| "epoch": 0.10291595197255575, | |
| "grad_norm": 5.969348907470703, | |
| "learning_rate": 5e-05, | |
| "loss": 1.2312, | |
| "step": 105 | |
| }, | |
| { | |
| "epoch": 0.1038961038961039, | |
| "grad_norm": 6.7192230224609375, | |
| "learning_rate": 4.917330276168208e-05, | |
| "loss": 1.1402, | |
| "step": 106 | |
| }, | |
| { | |
| "epoch": 0.10487625581965204, | |
| "grad_norm": 7.596022605895996, | |
| "learning_rate": 4.834683153413459e-05, | |
| "loss": 1.2494, | |
| "step": 107 | |
| }, | |
| { | |
| "epoch": 0.1058564077432002, | |
| "grad_norm": 8.891712188720703, | |
| "learning_rate": 4.7520812266338885e-05, | |
| "loss": 1.1324, | |
| "step": 108 | |
| }, | |
| { | |
| "epoch": 0.10683655966674835, | |
| "grad_norm": 11.980855941772461, | |
| "learning_rate": 4.669547078371504e-05, | |
| "loss": 1.3268, | |
| "step": 109 | |
| }, | |
| { | |
| "epoch": 0.1078167115902965, | |
| "grad_norm": 9.47231674194336, | |
| "learning_rate": 4.5871032726383386e-05, | |
| "loss": 1.2507, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.10879686351384464, | |
| "grad_norm": 7.6924567222595215, | |
| "learning_rate": 4.504772348747687e-05, | |
| "loss": 1.217, | |
| "step": 111 | |
| }, | |
| { | |
| "epoch": 0.1097770154373928, | |
| "grad_norm": 9.18214225769043, | |
| "learning_rate": 4.4225768151520694e-05, | |
| "loss": 1.3148, | |
| "step": 112 | |
| }, | |
| { | |
| "epoch": 0.11075716736094095, | |
| "grad_norm": 6.987793922424316, | |
| "learning_rate": 4.3405391432896555e-05, | |
| "loss": 1.0323, | |
| "step": 113 | |
| }, | |
| { | |
| "epoch": 0.1117373192844891, | |
| "grad_norm": 8.551517486572266, | |
| "learning_rate": 4.2586817614407895e-05, | |
| "loss": 1.3381, | |
| "step": 114 | |
| }, | |
| { | |
| "epoch": 0.11271747120803724, | |
| "grad_norm": 8.323104858398438, | |
| "learning_rate": 4.17702704859633e-05, | |
| "loss": 1.1494, | |
| "step": 115 | |
| }, | |
| { | |
| "epoch": 0.11369762313158539, | |
| "grad_norm": 10.007205963134766, | |
| "learning_rate": 4.095597328339452e-05, | |
| "loss": 1.0202, | |
| "step": 116 | |
| }, | |
| { | |
| "epoch": 0.11467777505513355, | |
| "grad_norm": 7.001098155975342, | |
| "learning_rate": 4.0144148627425993e-05, | |
| "loss": 1.099, | |
| "step": 117 | |
| }, | |
| { | |
| "epoch": 0.1156579269786817, | |
| "grad_norm": 9.060266494750977, | |
| "learning_rate": 3.933501846281267e-05, | |
| "loss": 0.9041, | |
| "step": 118 | |
| }, | |
| { | |
| "epoch": 0.11663807890222985, | |
| "grad_norm": 6.638906478881836, | |
| "learning_rate": 3.852880399766243e-05, | |
| "loss": 0.9451, | |
| "step": 119 | |
| }, | |
| { | |
| "epoch": 0.11761823082577799, | |
| "grad_norm": 10.675006866455078, | |
| "learning_rate": 3.772572564296005e-05, | |
| "loss": 1.0778, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.11859838274932614, | |
| "grad_norm": 7.329932689666748, | |
| "learning_rate": 3.6926002952309016e-05, | |
| "loss": 1.0241, | |
| "step": 121 | |
| }, | |
| { | |
| "epoch": 0.1195785346728743, | |
| "grad_norm": 9.862916946411133, | |
| "learning_rate": 3.612985456190778e-05, | |
| "loss": 1.0688, | |
| "step": 122 | |
| }, | |
| { | |
| "epoch": 0.12055868659642245, | |
| "grad_norm": 13.709855079650879, | |
| "learning_rate": 3.533749813077677e-05, | |
| "loss": 0.9626, | |
| "step": 123 | |
| }, | |
| { | |
| "epoch": 0.12153883851997059, | |
| "grad_norm": 8.847636222839355, | |
| "learning_rate": 3.4549150281252636e-05, | |
| "loss": 0.9327, | |
| "step": 124 | |
| }, | |
| { | |
| "epoch": 0.12251899044351874, | |
| "grad_norm": 9.130758285522461, | |
| "learning_rate": 3.3765026539765834e-05, | |
| "loss": 0.9667, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 0.1234991423670669, | |
| "grad_norm": 11.84672737121582, | |
| "learning_rate": 3.298534127791785e-05, | |
| "loss": 1.0622, | |
| "step": 126 | |
| }, | |
| { | |
| "epoch": 0.12447929429061505, | |
| "grad_norm": 9.412070274353027, | |
| "learning_rate": 3.221030765387417e-05, | |
| "loss": 0.9975, | |
| "step": 127 | |
| }, | |
| { | |
| "epoch": 0.1254594462141632, | |
| "grad_norm": 6.956162452697754, | |
| "learning_rate": 3.144013755408895e-05, | |
| "loss": 1.0391, | |
| "step": 128 | |
| }, | |
| { | |
| "epoch": 0.12643959813771136, | |
| "grad_norm": 10.878119468688965, | |
| "learning_rate": 3.0675041535377405e-05, | |
| "loss": 0.9248, | |
| "step": 129 | |
| }, | |
| { | |
| "epoch": 0.1274197500612595, | |
| "grad_norm": 11.166563034057617, | |
| "learning_rate": 2.991522876735154e-05, | |
| "loss": 1.0058, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.12839990198480764, | |
| "grad_norm": 8.980353355407715, | |
| "learning_rate": 2.916090697523549e-05, | |
| "loss": 1.0036, | |
| "step": 131 | |
| }, | |
| { | |
| "epoch": 0.1293800539083558, | |
| "grad_norm": 11.120211601257324, | |
| "learning_rate": 2.8412282383075363e-05, | |
| "loss": 1.1327, | |
| "step": 132 | |
| }, | |
| { | |
| "epoch": 0.13036020583190394, | |
| "grad_norm": 10.85069751739502, | |
| "learning_rate": 2.766955965735968e-05, | |
| "loss": 0.9424, | |
| "step": 133 | |
| }, | |
| { | |
| "epoch": 0.1313403577554521, | |
| "grad_norm": 9.08211898803711, | |
| "learning_rate": 2.693294185106562e-05, | |
| "loss": 0.9746, | |
| "step": 134 | |
| }, | |
| { | |
| "epoch": 0.13232050967900025, | |
| "grad_norm": 8.331099510192871, | |
| "learning_rate": 2.6202630348146324e-05, | |
| "loss": 1.079, | |
| "step": 135 | |
| }, | |
| { | |
| "epoch": 0.1333006616025484, | |
| "grad_norm": 9.91579818725586, | |
| "learning_rate": 2.547882480847461e-05, | |
| "loss": 1.0347, | |
| "step": 136 | |
| }, | |
| { | |
| "epoch": 0.13428081352609655, | |
| "grad_norm": 11.309249877929688, | |
| "learning_rate": 2.476172311325783e-05, | |
| "loss": 1.0018, | |
| "step": 137 | |
| }, | |
| { | |
| "epoch": 0.1352609654496447, | |
| "grad_norm": 8.515524864196777, | |
| "learning_rate": 2.405152131093926e-05, | |
| "loss": 0.9815, | |
| "step": 138 | |
| }, | |
| { | |
| "epoch": 0.13624111737319283, | |
| "grad_norm": 7.5276618003845215, | |
| "learning_rate": 2.3348413563600325e-05, | |
| "loss": 0.9255, | |
| "step": 139 | |
| }, | |
| { | |
| "epoch": 0.137221269296741, | |
| "grad_norm": 10.231163024902344, | |
| "learning_rate": 2.2652592093878666e-05, | |
| "loss": 0.9124, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.13820142122028914, | |
| "grad_norm": 9.131063461303711, | |
| "learning_rate": 2.196424713241637e-05, | |
| "loss": 1.0122, | |
| "step": 141 | |
| }, | |
| { | |
| "epoch": 0.1391815731438373, | |
| "grad_norm": 8.375762939453125, | |
| "learning_rate": 2.128356686585282e-05, | |
| "loss": 1.0161, | |
| "step": 142 | |
| }, | |
| { | |
| "epoch": 0.14016172506738545, | |
| "grad_norm": 10.814382553100586, | |
| "learning_rate": 2.061073738537635e-05, | |
| "loss": 1.0631, | |
| "step": 143 | |
| }, | |
| { | |
| "epoch": 0.14114187699093358, | |
| "grad_norm": 8.42130184173584, | |
| "learning_rate": 1.9945942635848748e-05, | |
| "loss": 0.979, | |
| "step": 144 | |
| }, | |
| { | |
| "epoch": 0.14212202891448175, | |
| "grad_norm": 13.615252494812012, | |
| "learning_rate": 1.928936436551661e-05, | |
| "loss": 1.0586, | |
| "step": 145 | |
| }, | |
| { | |
| "epoch": 0.1431021808380299, | |
| "grad_norm": 9.27658748626709, | |
| "learning_rate": 1.8641182076323148e-05, | |
| "loss": 0.9857, | |
| "step": 146 | |
| }, | |
| { | |
| "epoch": 0.14408233276157806, | |
| "grad_norm": 8.280218124389648, | |
| "learning_rate": 1.800157297483417e-05, | |
| "loss": 1.0255, | |
| "step": 147 | |
| }, | |
| { | |
| "epoch": 0.1450624846851262, | |
| "grad_norm": 7.738588809967041, | |
| "learning_rate": 1.7370711923791567e-05, | |
| "loss": 1.0579, | |
| "step": 148 | |
| }, | |
| { | |
| "epoch": 0.14604263660867434, | |
| "grad_norm": 7.782927989959717, | |
| "learning_rate": 1.6748771394307585e-05, | |
| "loss": 0.9789, | |
| "step": 149 | |
| }, | |
| { | |
| "epoch": 0.1470227885322225, | |
| "grad_norm": 9.94774055480957, | |
| "learning_rate": 1.6135921418712956e-05, | |
| "loss": 0.9902, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.1470227885322225, | |
| "eval_loss": 1.0405155420303345, | |
| "eval_runtime": 158.2427, | |
| "eval_samples_per_second": 10.863, | |
| "eval_steps_per_second": 2.717, | |
| "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": 1.1095943914979328e+17, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |