| { | |
| "best_metric": 1.4672940969467163, | |
| "best_model_checkpoint": "miner_id_24/checkpoint-150", | |
| "epoch": 0.2828854314002829, | |
| "eval_steps": 50, | |
| "global_step": 150, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.001885902876001886, | |
| "grad_norm": 1.6719341278076172, | |
| "learning_rate": 1e-05, | |
| "loss": 1.4085, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.001885902876001886, | |
| "eval_loss": 1.798437476158142, | |
| "eval_runtime": 15.88, | |
| "eval_samples_per_second": 56.234, | |
| "eval_steps_per_second": 14.106, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.003771805752003772, | |
| "grad_norm": 1.6418969631195068, | |
| "learning_rate": 2e-05, | |
| "loss": 1.4482, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.005657708628005658, | |
| "grad_norm": 1.6673892736434937, | |
| "learning_rate": 3e-05, | |
| "loss": 1.5452, | |
| "step": 3 | |
| }, | |
| { | |
| "epoch": 0.007543611504007544, | |
| "grad_norm": 1.5872431993484497, | |
| "learning_rate": 4e-05, | |
| "loss": 1.5037, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.00942951438000943, | |
| "grad_norm": 1.4873931407928467, | |
| "learning_rate": 5e-05, | |
| "loss": 1.362, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.011315417256011316, | |
| "grad_norm": 1.3264328241348267, | |
| "learning_rate": 6e-05, | |
| "loss": 1.385, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.013201320132013201, | |
| "grad_norm": 1.127667784690857, | |
| "learning_rate": 7e-05, | |
| "loss": 1.3468, | |
| "step": 7 | |
| }, | |
| { | |
| "epoch": 0.015087223008015087, | |
| "grad_norm": 1.2904081344604492, | |
| "learning_rate": 8e-05, | |
| "loss": 1.3893, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.016973125884016973, | |
| "grad_norm": 1.4509556293487549, | |
| "learning_rate": 9e-05, | |
| "loss": 1.5524, | |
| "step": 9 | |
| }, | |
| { | |
| "epoch": 0.01885902876001886, | |
| "grad_norm": 1.1893857717514038, | |
| "learning_rate": 0.0001, | |
| "loss": 1.5371, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.020744931636020744, | |
| "grad_norm": 1.0708231925964355, | |
| "learning_rate": 9.999316524962345e-05, | |
| "loss": 1.2872, | |
| "step": 11 | |
| }, | |
| { | |
| "epoch": 0.02263083451202263, | |
| "grad_norm": 1.1538314819335938, | |
| "learning_rate": 9.997266286704631e-05, | |
| "loss": 1.291, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 0.024516737388024516, | |
| "grad_norm": 1.117232084274292, | |
| "learning_rate": 9.993849845741524e-05, | |
| "loss": 1.4771, | |
| "step": 13 | |
| }, | |
| { | |
| "epoch": 0.026402640264026403, | |
| "grad_norm": 1.038404941558838, | |
| "learning_rate": 9.989068136093873e-05, | |
| "loss": 1.4339, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 0.028288543140028287, | |
| "grad_norm": 1.0554722547531128, | |
| "learning_rate": 9.98292246503335e-05, | |
| "loss": 1.2713, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.030174446016030174, | |
| "grad_norm": 0.9968124628067017, | |
| "learning_rate": 9.975414512725057e-05, | |
| "loss": 1.4743, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 0.03206034889203206, | |
| "grad_norm": 1.0715779066085815, | |
| "learning_rate": 9.966546331768191e-05, | |
| "loss": 1.5222, | |
| "step": 17 | |
| }, | |
| { | |
| "epoch": 0.033946251768033946, | |
| "grad_norm": 1.1179016828536987, | |
| "learning_rate": 9.956320346634876e-05, | |
| "loss": 1.4405, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 0.03583215464403583, | |
| "grad_norm": 0.9977557063102722, | |
| "learning_rate": 9.944739353007344e-05, | |
| "loss": 1.3601, | |
| "step": 19 | |
| }, | |
| { | |
| "epoch": 0.03771805752003772, | |
| "grad_norm": 1.0840213298797607, | |
| "learning_rate": 9.931806517013612e-05, | |
| "loss": 1.7028, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.039603960396039604, | |
| "grad_norm": 1.0020934343338013, | |
| "learning_rate": 9.917525374361912e-05, | |
| "loss": 1.5547, | |
| "step": 21 | |
| }, | |
| { | |
| "epoch": 0.04148986327204149, | |
| "grad_norm": 1.0338664054870605, | |
| "learning_rate": 9.901899829374047e-05, | |
| "loss": 1.4547, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 0.04337576614804337, | |
| "grad_norm": 1.02623450756073, | |
| "learning_rate": 9.884934153917997e-05, | |
| "loss": 1.5793, | |
| "step": 23 | |
| }, | |
| { | |
| "epoch": 0.04526166902404526, | |
| "grad_norm": 0.9653357863426208, | |
| "learning_rate": 9.86663298624003e-05, | |
| "loss": 1.4715, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 0.04714757190004715, | |
| "grad_norm": 1.058281421661377, | |
| "learning_rate": 9.847001329696653e-05, | |
| "loss": 1.5606, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.04903347477604903, | |
| "grad_norm": 1.0581485033035278, | |
| "learning_rate": 9.826044551386744e-05, | |
| "loss": 1.3818, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 0.05091937765205092, | |
| "grad_norm": 1.0404139757156372, | |
| "learning_rate": 9.803768380684242e-05, | |
| "loss": 1.4896, | |
| "step": 27 | |
| }, | |
| { | |
| "epoch": 0.052805280528052806, | |
| "grad_norm": 0.9869558215141296, | |
| "learning_rate": 9.780178907671789e-05, | |
| "loss": 1.6065, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 0.05469118340405469, | |
| "grad_norm": 1.03351628780365, | |
| "learning_rate": 9.755282581475769e-05, | |
| "loss": 1.4879, | |
| "step": 29 | |
| }, | |
| { | |
| "epoch": 0.056577086280056574, | |
| "grad_norm": 1.1283892393112183, | |
| "learning_rate": 9.729086208503174e-05, | |
| "loss": 1.6899, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.058462989156058465, | |
| "grad_norm": 1.1303669214248657, | |
| "learning_rate": 9.701596950580806e-05, | |
| "loss": 1.7194, | |
| "step": 31 | |
| }, | |
| { | |
| "epoch": 0.06034889203206035, | |
| "grad_norm": 1.136953592300415, | |
| "learning_rate": 9.672822322997305e-05, | |
| "loss": 1.5471, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 0.06223479490806223, | |
| "grad_norm": 1.0339196920394897, | |
| "learning_rate": 9.642770192448536e-05, | |
| "loss": 1.6283, | |
| "step": 33 | |
| }, | |
| { | |
| "epoch": 0.06412069778406412, | |
| "grad_norm": 1.0422121286392212, | |
| "learning_rate": 9.611448774886924e-05, | |
| "loss": 1.6243, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 0.066006600660066, | |
| "grad_norm": 1.0634968280792236, | |
| "learning_rate": 9.578866633275288e-05, | |
| "loss": 1.6214, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.06789250353606789, | |
| "grad_norm": 1.0582144260406494, | |
| "learning_rate": 9.545032675245813e-05, | |
| "loss": 1.4956, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 0.06977840641206978, | |
| "grad_norm": 1.1187388896942139, | |
| "learning_rate": 9.509956150664796e-05, | |
| "loss": 1.5546, | |
| "step": 37 | |
| }, | |
| { | |
| "epoch": 0.07166430928807166, | |
| "grad_norm": 1.0846741199493408, | |
| "learning_rate": 9.473646649103818e-05, | |
| "loss": 1.6097, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 0.07355021216407355, | |
| "grad_norm": 1.1303223371505737, | |
| "learning_rate": 9.43611409721806e-05, | |
| "loss": 1.6785, | |
| "step": 39 | |
| }, | |
| { | |
| "epoch": 0.07543611504007544, | |
| "grad_norm": 1.1018422842025757, | |
| "learning_rate": 9.397368756032445e-05, | |
| "loss": 1.4387, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.07732201791607732, | |
| "grad_norm": 1.0702595710754395, | |
| "learning_rate": 9.357421218136386e-05, | |
| "loss": 1.5708, | |
| "step": 41 | |
| }, | |
| { | |
| "epoch": 0.07920792079207921, | |
| "grad_norm": 1.133355975151062, | |
| "learning_rate": 9.316282404787871e-05, | |
| "loss": 1.5197, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 0.0810938236680811, | |
| "grad_norm": 1.1223176717758179, | |
| "learning_rate": 9.273963562927695e-05, | |
| "loss": 1.5539, | |
| "step": 43 | |
| }, | |
| { | |
| "epoch": 0.08297972654408298, | |
| "grad_norm": 1.097373127937317, | |
| "learning_rate": 9.230476262104677e-05, | |
| "loss": 1.4639, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 0.08486562942008487, | |
| "grad_norm": 1.259489893913269, | |
| "learning_rate": 9.185832391312644e-05, | |
| "loss": 1.6197, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 0.08675153229608674, | |
| "grad_norm": 1.1635847091674805, | |
| "learning_rate": 9.140044155740101e-05, | |
| "loss": 1.412, | |
| "step": 46 | |
| }, | |
| { | |
| "epoch": 0.08863743517208864, | |
| "grad_norm": 1.3465157747268677, | |
| "learning_rate": 9.093124073433463e-05, | |
| "loss": 1.6082, | |
| "step": 47 | |
| }, | |
| { | |
| "epoch": 0.09052333804809053, | |
| "grad_norm": 1.2739461660385132, | |
| "learning_rate": 9.045084971874738e-05, | |
| "loss": 1.5323, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 0.0924092409240924, | |
| "grad_norm": 1.239130973815918, | |
| "learning_rate": 8.995939984474624e-05, | |
| "loss": 1.4232, | |
| "step": 49 | |
| }, | |
| { | |
| "epoch": 0.0942951438000943, | |
| "grad_norm": 1.4111778736114502, | |
| "learning_rate": 8.945702546981969e-05, | |
| "loss": 1.4941, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.0942951438000943, | |
| "eval_loss": 1.5392982959747314, | |
| "eval_runtime": 15.8606, | |
| "eval_samples_per_second": 56.303, | |
| "eval_steps_per_second": 14.123, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.09618104667609619, | |
| "grad_norm": 1.2728794813156128, | |
| "learning_rate": 8.894386393810563e-05, | |
| "loss": 1.3532, | |
| "step": 51 | |
| }, | |
| { | |
| "epoch": 0.09806694955209806, | |
| "grad_norm": 1.2738091945648193, | |
| "learning_rate": 8.842005554284296e-05, | |
| "loss": 1.3604, | |
| "step": 52 | |
| }, | |
| { | |
| "epoch": 0.09995285242809995, | |
| "grad_norm": 1.0199331045150757, | |
| "learning_rate": 8.788574348801675e-05, | |
| "loss": 1.3546, | |
| "step": 53 | |
| }, | |
| { | |
| "epoch": 0.10183875530410184, | |
| "grad_norm": 0.8882190585136414, | |
| "learning_rate": 8.73410738492077e-05, | |
| "loss": 1.1719, | |
| "step": 54 | |
| }, | |
| { | |
| "epoch": 0.10372465818010372, | |
| "grad_norm": 0.8215413093566895, | |
| "learning_rate": 8.678619553365659e-05, | |
| "loss": 1.3678, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 0.10561056105610561, | |
| "grad_norm": 0.7815274000167847, | |
| "learning_rate": 8.622126023955446e-05, | |
| "loss": 1.3791, | |
| "step": 56 | |
| }, | |
| { | |
| "epoch": 0.1074964639321075, | |
| "grad_norm": 0.7754129767417908, | |
| "learning_rate": 8.564642241456986e-05, | |
| "loss": 1.3071, | |
| "step": 57 | |
| }, | |
| { | |
| "epoch": 0.10938236680810938, | |
| "grad_norm": 0.8610193729400635, | |
| "learning_rate": 8.506183921362443e-05, | |
| "loss": 1.4257, | |
| "step": 58 | |
| }, | |
| { | |
| "epoch": 0.11126826968411127, | |
| "grad_norm": 0.7398249506950378, | |
| "learning_rate": 8.44676704559283e-05, | |
| "loss": 1.3307, | |
| "step": 59 | |
| }, | |
| { | |
| "epoch": 0.11315417256011315, | |
| "grad_norm": 0.8163387179374695, | |
| "learning_rate": 8.386407858128706e-05, | |
| "loss": 1.2897, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.11504007543611504, | |
| "grad_norm": 0.8740779161453247, | |
| "learning_rate": 8.32512286056924e-05, | |
| "loss": 1.2508, | |
| "step": 61 | |
| }, | |
| { | |
| "epoch": 0.11692597831211693, | |
| "grad_norm": 0.818240225315094, | |
| "learning_rate": 8.262928807620843e-05, | |
| "loss": 1.2878, | |
| "step": 62 | |
| }, | |
| { | |
| "epoch": 0.1188118811881188, | |
| "grad_norm": 0.8404731154441833, | |
| "learning_rate": 8.199842702516583e-05, | |
| "loss": 1.3849, | |
| "step": 63 | |
| }, | |
| { | |
| "epoch": 0.1206977840641207, | |
| "grad_norm": 0.896458625793457, | |
| "learning_rate": 8.135881792367686e-05, | |
| "loss": 1.4535, | |
| "step": 64 | |
| }, | |
| { | |
| "epoch": 0.12258368694012259, | |
| "grad_norm": 0.8432134389877319, | |
| "learning_rate": 8.07106356344834e-05, | |
| "loss": 1.4789, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 0.12446958981612447, | |
| "grad_norm": 0.8519056439399719, | |
| "learning_rate": 8.005405736415126e-05, | |
| "loss": 1.5264, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 0.12635549269212634, | |
| "grad_norm": 0.7978764772415161, | |
| "learning_rate": 7.938926261462366e-05, | |
| "loss": 1.3318, | |
| "step": 67 | |
| }, | |
| { | |
| "epoch": 0.12824139556812825, | |
| "grad_norm": 0.8735756278038025, | |
| "learning_rate": 7.871643313414718e-05, | |
| "loss": 1.355, | |
| "step": 68 | |
| }, | |
| { | |
| "epoch": 0.13012729844413012, | |
| "grad_norm": 0.8317325711250305, | |
| "learning_rate": 7.803575286758364e-05, | |
| "loss": 1.3436, | |
| "step": 69 | |
| }, | |
| { | |
| "epoch": 0.132013201320132, | |
| "grad_norm": 0.9063963294029236, | |
| "learning_rate": 7.734740790612136e-05, | |
| "loss": 1.5033, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.1338991041961339, | |
| "grad_norm": 0.8298577070236206, | |
| "learning_rate": 7.66515864363997e-05, | |
| "loss": 1.4334, | |
| "step": 71 | |
| }, | |
| { | |
| "epoch": 0.13578500707213578, | |
| "grad_norm": 0.821303129196167, | |
| "learning_rate": 7.594847868906076e-05, | |
| "loss": 1.4554, | |
| "step": 72 | |
| }, | |
| { | |
| "epoch": 0.13767090994813766, | |
| "grad_norm": 0.8228156566619873, | |
| "learning_rate": 7.52382768867422e-05, | |
| "loss": 1.3866, | |
| "step": 73 | |
| }, | |
| { | |
| "epoch": 0.13955681282413956, | |
| "grad_norm": 0.8506703972816467, | |
| "learning_rate": 7.452117519152542e-05, | |
| "loss": 1.473, | |
| "step": 74 | |
| }, | |
| { | |
| "epoch": 0.14144271570014144, | |
| "grad_norm": 0.8994743824005127, | |
| "learning_rate": 7.379736965185368e-05, | |
| "loss": 1.6536, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.14332861857614332, | |
| "grad_norm": 0.9050391912460327, | |
| "learning_rate": 7.30670581489344e-05, | |
| "loss": 1.3966, | |
| "step": 76 | |
| }, | |
| { | |
| "epoch": 0.14521452145214522, | |
| "grad_norm": 0.9391245245933533, | |
| "learning_rate": 7.233044034264034e-05, | |
| "loss": 1.7276, | |
| "step": 77 | |
| }, | |
| { | |
| "epoch": 0.1471004243281471, | |
| "grad_norm": 0.8455343842506409, | |
| "learning_rate": 7.158771761692464e-05, | |
| "loss": 1.4665, | |
| "step": 78 | |
| }, | |
| { | |
| "epoch": 0.14898632720414898, | |
| "grad_norm": 0.9894546866416931, | |
| "learning_rate": 7.083909302476453e-05, | |
| "loss": 1.5861, | |
| "step": 79 | |
| }, | |
| { | |
| "epoch": 0.15087223008015088, | |
| "grad_norm": 0.883921205997467, | |
| "learning_rate": 7.008477123264848e-05, | |
| "loss": 1.4461, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.15275813295615276, | |
| "grad_norm": 0.9232316613197327, | |
| "learning_rate": 6.932495846462261e-05, | |
| "loss": 1.7158, | |
| "step": 81 | |
| }, | |
| { | |
| "epoch": 0.15464403583215464, | |
| "grad_norm": 0.9596819877624512, | |
| "learning_rate": 6.855986244591104e-05, | |
| "loss": 1.7722, | |
| "step": 82 | |
| }, | |
| { | |
| "epoch": 0.15652993870815654, | |
| "grad_norm": 0.920516848564148, | |
| "learning_rate": 6.778969234612584e-05, | |
| "loss": 1.4213, | |
| "step": 83 | |
| }, | |
| { | |
| "epoch": 0.15841584158415842, | |
| "grad_norm": 0.9526838660240173, | |
| "learning_rate": 6.701465872208216e-05, | |
| "loss": 1.6793, | |
| "step": 84 | |
| }, | |
| { | |
| "epoch": 0.1603017444601603, | |
| "grad_norm": 0.9615785479545593, | |
| "learning_rate": 6.623497346023418e-05, | |
| "loss": 1.5577, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 0.1621876473361622, | |
| "grad_norm": 0.891538679599762, | |
| "learning_rate": 6.545084971874738e-05, | |
| "loss": 1.4552, | |
| "step": 86 | |
| }, | |
| { | |
| "epoch": 0.16407355021216408, | |
| "grad_norm": 0.9908190369606018, | |
| "learning_rate": 6.466250186922325e-05, | |
| "loss": 1.4512, | |
| "step": 87 | |
| }, | |
| { | |
| "epoch": 0.16595945308816595, | |
| "grad_norm": 0.9302291870117188, | |
| "learning_rate": 6.387014543809223e-05, | |
| "loss": 1.6369, | |
| "step": 88 | |
| }, | |
| { | |
| "epoch": 0.16784535596416786, | |
| "grad_norm": 1.0175998210906982, | |
| "learning_rate": 6.307399704769099e-05, | |
| "loss": 1.5077, | |
| "step": 89 | |
| }, | |
| { | |
| "epoch": 0.16973125884016974, | |
| "grad_norm": 0.9874126315116882, | |
| "learning_rate": 6.227427435703997e-05, | |
| "loss": 1.733, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.1716171617161716, | |
| "grad_norm": 0.9775768518447876, | |
| "learning_rate": 6.147119600233758e-05, | |
| "loss": 1.5483, | |
| "step": 91 | |
| }, | |
| { | |
| "epoch": 0.1735030645921735, | |
| "grad_norm": 1.0082566738128662, | |
| "learning_rate": 6.066498153718735e-05, | |
| "loss": 1.5691, | |
| "step": 92 | |
| }, | |
| { | |
| "epoch": 0.1753889674681754, | |
| "grad_norm": 1.0059901475906372, | |
| "learning_rate": 5.985585137257401e-05, | |
| "loss": 1.4483, | |
| "step": 93 | |
| }, | |
| { | |
| "epoch": 0.17727487034417727, | |
| "grad_norm": 0.9890210628509521, | |
| "learning_rate": 5.90440267166055e-05, | |
| "loss": 1.5623, | |
| "step": 94 | |
| }, | |
| { | |
| "epoch": 0.17916077322017915, | |
| "grad_norm": 1.0527801513671875, | |
| "learning_rate": 5.8229729514036705e-05, | |
| "loss": 1.6086, | |
| "step": 95 | |
| }, | |
| { | |
| "epoch": 0.18104667609618105, | |
| "grad_norm": 1.0493412017822266, | |
| "learning_rate": 5.74131823855921e-05, | |
| "loss": 1.6354, | |
| "step": 96 | |
| }, | |
| { | |
| "epoch": 0.18293257897218293, | |
| "grad_norm": 1.129049301147461, | |
| "learning_rate": 5.6594608567103456e-05, | |
| "loss": 1.4702, | |
| "step": 97 | |
| }, | |
| { | |
| "epoch": 0.1848184818481848, | |
| "grad_norm": 1.0613797903060913, | |
| "learning_rate": 5.577423184847932e-05, | |
| "loss": 1.5875, | |
| "step": 98 | |
| }, | |
| { | |
| "epoch": 0.1867043847241867, | |
| "grad_norm": 1.1796647310256958, | |
| "learning_rate": 5.495227651252315e-05, | |
| "loss": 1.6639, | |
| "step": 99 | |
| }, | |
| { | |
| "epoch": 0.1885902876001886, | |
| "grad_norm": 1.1907509565353394, | |
| "learning_rate": 5.4128967273616625e-05, | |
| "loss": 1.5667, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.1885902876001886, | |
| "eval_loss": 1.4839268922805786, | |
| "eval_runtime": 15.8643, | |
| "eval_samples_per_second": 56.29, | |
| "eval_steps_per_second": 14.12, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.19047619047619047, | |
| "grad_norm": 0.7392789125442505, | |
| "learning_rate": 5.330452921628497e-05, | |
| "loss": 1.3007, | |
| "step": 101 | |
| }, | |
| { | |
| "epoch": 0.19236209335219237, | |
| "grad_norm": 0.8285383582115173, | |
| "learning_rate": 5.247918773366112e-05, | |
| "loss": 1.3803, | |
| "step": 102 | |
| }, | |
| { | |
| "epoch": 0.19424799622819425, | |
| "grad_norm": 0.7633514404296875, | |
| "learning_rate": 5.165316846586541e-05, | |
| "loss": 1.3377, | |
| "step": 103 | |
| }, | |
| { | |
| "epoch": 0.19613389910419612, | |
| "grad_norm": 0.7272994518280029, | |
| "learning_rate": 5.0826697238317935e-05, | |
| "loss": 1.2073, | |
| "step": 104 | |
| }, | |
| { | |
| "epoch": 0.19801980198019803, | |
| "grad_norm": 0.7535402774810791, | |
| "learning_rate": 5e-05, | |
| "loss": 1.2368, | |
| "step": 105 | |
| }, | |
| { | |
| "epoch": 0.1999057048561999, | |
| "grad_norm": 0.7271507382392883, | |
| "learning_rate": 4.917330276168208e-05, | |
| "loss": 1.426, | |
| "step": 106 | |
| }, | |
| { | |
| "epoch": 0.20179160773220178, | |
| "grad_norm": 0.7888609170913696, | |
| "learning_rate": 4.834683153413459e-05, | |
| "loss": 1.298, | |
| "step": 107 | |
| }, | |
| { | |
| "epoch": 0.2036775106082037, | |
| "grad_norm": 0.7619734406471252, | |
| "learning_rate": 4.7520812266338885e-05, | |
| "loss": 1.529, | |
| "step": 108 | |
| }, | |
| { | |
| "epoch": 0.20556341348420556, | |
| "grad_norm": 0.7262706756591797, | |
| "learning_rate": 4.669547078371504e-05, | |
| "loss": 1.3107, | |
| "step": 109 | |
| }, | |
| { | |
| "epoch": 0.20744931636020744, | |
| "grad_norm": 0.7511423826217651, | |
| "learning_rate": 4.5871032726383386e-05, | |
| "loss": 1.2902, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.20933521923620935, | |
| "grad_norm": 0.7127615809440613, | |
| "learning_rate": 4.504772348747687e-05, | |
| "loss": 1.2128, | |
| "step": 111 | |
| }, | |
| { | |
| "epoch": 0.21122112211221122, | |
| "grad_norm": 0.7513578534126282, | |
| "learning_rate": 4.4225768151520694e-05, | |
| "loss": 1.504, | |
| "step": 112 | |
| }, | |
| { | |
| "epoch": 0.2131070249882131, | |
| "grad_norm": 0.8162602186203003, | |
| "learning_rate": 4.3405391432896555e-05, | |
| "loss": 1.4341, | |
| "step": 113 | |
| }, | |
| { | |
| "epoch": 0.214992927864215, | |
| "grad_norm": 0.8222413659095764, | |
| "learning_rate": 4.2586817614407895e-05, | |
| "loss": 1.4536, | |
| "step": 114 | |
| }, | |
| { | |
| "epoch": 0.21687883074021688, | |
| "grad_norm": 0.8294031620025635, | |
| "learning_rate": 4.17702704859633e-05, | |
| "loss": 1.3297, | |
| "step": 115 | |
| }, | |
| { | |
| "epoch": 0.21876473361621876, | |
| "grad_norm": 0.7801927924156189, | |
| "learning_rate": 4.095597328339452e-05, | |
| "loss": 1.456, | |
| "step": 116 | |
| }, | |
| { | |
| "epoch": 0.22065063649222066, | |
| "grad_norm": 0.8259176015853882, | |
| "learning_rate": 4.0144148627425993e-05, | |
| "loss": 1.4551, | |
| "step": 117 | |
| }, | |
| { | |
| "epoch": 0.22253653936822254, | |
| "grad_norm": 0.7965975999832153, | |
| "learning_rate": 3.933501846281267e-05, | |
| "loss": 1.5074, | |
| "step": 118 | |
| }, | |
| { | |
| "epoch": 0.22442244224422442, | |
| "grad_norm": 0.8104461431503296, | |
| "learning_rate": 3.852880399766243e-05, | |
| "loss": 1.3188, | |
| "step": 119 | |
| }, | |
| { | |
| "epoch": 0.2263083451202263, | |
| "grad_norm": 0.8252735137939453, | |
| "learning_rate": 3.772572564296005e-05, | |
| "loss": 1.3283, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.2281942479962282, | |
| "grad_norm": 0.8110776543617249, | |
| "learning_rate": 3.6926002952309016e-05, | |
| "loss": 1.4883, | |
| "step": 121 | |
| }, | |
| { | |
| "epoch": 0.23008015087223008, | |
| "grad_norm": 0.7731505632400513, | |
| "learning_rate": 3.612985456190778e-05, | |
| "loss": 1.461, | |
| "step": 122 | |
| }, | |
| { | |
| "epoch": 0.23196605374823195, | |
| "grad_norm": 0.7795805335044861, | |
| "learning_rate": 3.533749813077677e-05, | |
| "loss": 1.3828, | |
| "step": 123 | |
| }, | |
| { | |
| "epoch": 0.23385195662423386, | |
| "grad_norm": 0.8774617910385132, | |
| "learning_rate": 3.4549150281252636e-05, | |
| "loss": 1.605, | |
| "step": 124 | |
| }, | |
| { | |
| "epoch": 0.23573785950023574, | |
| "grad_norm": 0.8436756730079651, | |
| "learning_rate": 3.3765026539765834e-05, | |
| "loss": 1.4813, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 0.2376237623762376, | |
| "grad_norm": 0.8913021683692932, | |
| "learning_rate": 3.298534127791785e-05, | |
| "loss": 1.4368, | |
| "step": 126 | |
| }, | |
| { | |
| "epoch": 0.23950966525223952, | |
| "grad_norm": 0.8260645866394043, | |
| "learning_rate": 3.221030765387417e-05, | |
| "loss": 1.5402, | |
| "step": 127 | |
| }, | |
| { | |
| "epoch": 0.2413955681282414, | |
| "grad_norm": 0.8445008993148804, | |
| "learning_rate": 3.144013755408895e-05, | |
| "loss": 1.4945, | |
| "step": 128 | |
| }, | |
| { | |
| "epoch": 0.24328147100424327, | |
| "grad_norm": 0.8457810878753662, | |
| "learning_rate": 3.0675041535377405e-05, | |
| "loss": 1.4877, | |
| "step": 129 | |
| }, | |
| { | |
| "epoch": 0.24516737388024518, | |
| "grad_norm": 0.8343524932861328, | |
| "learning_rate": 2.991522876735154e-05, | |
| "loss": 1.4436, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.24705327675624705, | |
| "grad_norm": 0.8800844550132751, | |
| "learning_rate": 2.916090697523549e-05, | |
| "loss": 1.4672, | |
| "step": 131 | |
| }, | |
| { | |
| "epoch": 0.24893917963224893, | |
| "grad_norm": 0.8727882504463196, | |
| "learning_rate": 2.8412282383075363e-05, | |
| "loss": 1.5406, | |
| "step": 132 | |
| }, | |
| { | |
| "epoch": 0.2508250825082508, | |
| "grad_norm": 0.8764476776123047, | |
| "learning_rate": 2.766955965735968e-05, | |
| "loss": 1.5357, | |
| "step": 133 | |
| }, | |
| { | |
| "epoch": 0.2527109853842527, | |
| "grad_norm": 1.0033875703811646, | |
| "learning_rate": 2.693294185106562e-05, | |
| "loss": 1.548, | |
| "step": 134 | |
| }, | |
| { | |
| "epoch": 0.2545968882602546, | |
| "grad_norm": 0.8889184594154358, | |
| "learning_rate": 2.6202630348146324e-05, | |
| "loss": 1.5711, | |
| "step": 135 | |
| }, | |
| { | |
| "epoch": 0.2564827911362565, | |
| "grad_norm": 0.906014621257782, | |
| "learning_rate": 2.547882480847461e-05, | |
| "loss": 1.4872, | |
| "step": 136 | |
| }, | |
| { | |
| "epoch": 0.25836869401225837, | |
| "grad_norm": 0.9370135068893433, | |
| "learning_rate": 2.476172311325783e-05, | |
| "loss": 1.5383, | |
| "step": 137 | |
| }, | |
| { | |
| "epoch": 0.26025459688826025, | |
| "grad_norm": 0.9925730228424072, | |
| "learning_rate": 2.405152131093926e-05, | |
| "loss": 1.4615, | |
| "step": 138 | |
| }, | |
| { | |
| "epoch": 0.2621404997642621, | |
| "grad_norm": 0.9600664973258972, | |
| "learning_rate": 2.3348413563600325e-05, | |
| "loss": 1.5633, | |
| "step": 139 | |
| }, | |
| { | |
| "epoch": 0.264026402640264, | |
| "grad_norm": 0.9619641900062561, | |
| "learning_rate": 2.2652592093878666e-05, | |
| "loss": 1.4668, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.26591230551626593, | |
| "grad_norm": 0.9631545543670654, | |
| "learning_rate": 2.196424713241637e-05, | |
| "loss": 1.5177, | |
| "step": 141 | |
| }, | |
| { | |
| "epoch": 0.2677982083922678, | |
| "grad_norm": 0.9739064574241638, | |
| "learning_rate": 2.128356686585282e-05, | |
| "loss": 1.5393, | |
| "step": 142 | |
| }, | |
| { | |
| "epoch": 0.2696841112682697, | |
| "grad_norm": 1.0108882188796997, | |
| "learning_rate": 2.061073738537635e-05, | |
| "loss": 1.6015, | |
| "step": 143 | |
| }, | |
| { | |
| "epoch": 0.27157001414427157, | |
| "grad_norm": 1.037048578262329, | |
| "learning_rate": 1.9945942635848748e-05, | |
| "loss": 1.5262, | |
| "step": 144 | |
| }, | |
| { | |
| "epoch": 0.27345591702027344, | |
| "grad_norm": 1.0594589710235596, | |
| "learning_rate": 1.928936436551661e-05, | |
| "loss": 1.5892, | |
| "step": 145 | |
| }, | |
| { | |
| "epoch": 0.2753418198962753, | |
| "grad_norm": 1.0016132593154907, | |
| "learning_rate": 1.8641182076323148e-05, | |
| "loss": 1.5237, | |
| "step": 146 | |
| }, | |
| { | |
| "epoch": 0.27722772277227725, | |
| "grad_norm": 1.119092345237732, | |
| "learning_rate": 1.800157297483417e-05, | |
| "loss": 1.5041, | |
| "step": 147 | |
| }, | |
| { | |
| "epoch": 0.27911362564827913, | |
| "grad_norm": 1.0873557329177856, | |
| "learning_rate": 1.7370711923791567e-05, | |
| "loss": 1.7046, | |
| "step": 148 | |
| }, | |
| { | |
| "epoch": 0.280999528524281, | |
| "grad_norm": 1.125348448753357, | |
| "learning_rate": 1.6748771394307585e-05, | |
| "loss": 1.415, | |
| "step": 149 | |
| }, | |
| { | |
| "epoch": 0.2828854314002829, | |
| "grad_norm": 1.322414517402649, | |
| "learning_rate": 1.6135921418712956e-05, | |
| "loss": 1.4796, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.2828854314002829, | |
| "eval_loss": 1.4672940969467163, | |
| "eval_runtime": 15.8748, | |
| "eval_samples_per_second": 56.253, | |
| "eval_steps_per_second": 14.11, | |
| "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": 1.1650687265931264e+16, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |