| { | |
| "best_metric": 0.7003475427627563, | |
| "best_model_checkpoint": "miner_id_24/checkpoint-150", | |
| "epoch": 0.019829466587348802, | |
| "eval_steps": 50, | |
| "global_step": 150, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.00013219644391565866, | |
| "grad_norm": 0.19538387656211853, | |
| "learning_rate": 1e-05, | |
| "loss": 0.7056, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.00013219644391565866, | |
| "eval_loss": 1.0835129022598267, | |
| "eval_runtime": 1275.6627, | |
| "eval_samples_per_second": 9.987, | |
| "eval_steps_per_second": 2.497, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.0002643928878313173, | |
| "grad_norm": 0.24580149352550507, | |
| "learning_rate": 2e-05, | |
| "loss": 0.782, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.000396589331746976, | |
| "grad_norm": 0.266649454832077, | |
| "learning_rate": 3e-05, | |
| "loss": 0.8056, | |
| "step": 3 | |
| }, | |
| { | |
| "epoch": 0.0005287857756626346, | |
| "grad_norm": 0.32284072041511536, | |
| "learning_rate": 4e-05, | |
| "loss": 0.8834, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.0006609822195782933, | |
| "grad_norm": 0.24443857371807098, | |
| "learning_rate": 5e-05, | |
| "loss": 0.8021, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.000793178663493952, | |
| "grad_norm": 0.24312937259674072, | |
| "learning_rate": 6e-05, | |
| "loss": 0.8884, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.0009253751074096107, | |
| "grad_norm": 0.2570464611053467, | |
| "learning_rate": 7e-05, | |
| "loss": 0.8745, | |
| "step": 7 | |
| }, | |
| { | |
| "epoch": 0.0010575715513252693, | |
| "grad_norm": 0.2607725262641907, | |
| "learning_rate": 8e-05, | |
| "loss": 0.9049, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.001189767995240928, | |
| "grad_norm": 0.27473318576812744, | |
| "learning_rate": 9e-05, | |
| "loss": 0.9292, | |
| "step": 9 | |
| }, | |
| { | |
| "epoch": 0.0013219644391565867, | |
| "grad_norm": 0.28504708409309387, | |
| "learning_rate": 0.0001, | |
| "loss": 0.8395, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.0014541608830722455, | |
| "grad_norm": 0.3271239101886749, | |
| "learning_rate": 9.999316524962345e-05, | |
| "loss": 0.7652, | |
| "step": 11 | |
| }, | |
| { | |
| "epoch": 0.001586357326987904, | |
| "grad_norm": 0.3310953378677368, | |
| "learning_rate": 9.997266286704631e-05, | |
| "loss": 0.9367, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 0.0017185537709035626, | |
| "grad_norm": 0.3015696704387665, | |
| "learning_rate": 9.993849845741524e-05, | |
| "loss": 0.8164, | |
| "step": 13 | |
| }, | |
| { | |
| "epoch": 0.0018507502148192214, | |
| "grad_norm": 0.32852622866630554, | |
| "learning_rate": 9.989068136093873e-05, | |
| "loss": 0.9113, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 0.00198294665873488, | |
| "grad_norm": 0.2933748960494995, | |
| "learning_rate": 9.98292246503335e-05, | |
| "loss": 0.9147, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.0021151431026505386, | |
| "grad_norm": 0.2997913062572479, | |
| "learning_rate": 9.975414512725057e-05, | |
| "loss": 0.7964, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 0.0022473395465661974, | |
| "grad_norm": 0.2818261384963989, | |
| "learning_rate": 9.966546331768191e-05, | |
| "loss": 0.8466, | |
| "step": 17 | |
| }, | |
| { | |
| "epoch": 0.002379535990481856, | |
| "grad_norm": 0.31236743927001953, | |
| "learning_rate": 9.956320346634876e-05, | |
| "loss": 0.8633, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 0.0025117324343975145, | |
| "grad_norm": 0.267844021320343, | |
| "learning_rate": 9.944739353007344e-05, | |
| "loss": 0.8381, | |
| "step": 19 | |
| }, | |
| { | |
| "epoch": 0.0026439288783131733, | |
| "grad_norm": 0.25400617718696594, | |
| "learning_rate": 9.931806517013612e-05, | |
| "loss": 0.8444, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.002776125322228832, | |
| "grad_norm": 0.2786080241203308, | |
| "learning_rate": 9.917525374361912e-05, | |
| "loss": 0.8191, | |
| "step": 21 | |
| }, | |
| { | |
| "epoch": 0.002908321766144491, | |
| "grad_norm": 0.2764554023742676, | |
| "learning_rate": 9.901899829374047e-05, | |
| "loss": 0.8549, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 0.0030405182100601493, | |
| "grad_norm": 0.28269195556640625, | |
| "learning_rate": 9.884934153917997e-05, | |
| "loss": 0.809, | |
| "step": 23 | |
| }, | |
| { | |
| "epoch": 0.003172714653975808, | |
| "grad_norm": 0.27460262179374695, | |
| "learning_rate": 9.86663298624003e-05, | |
| "loss": 0.7984, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 0.003304911097891467, | |
| "grad_norm": 0.3024798035621643, | |
| "learning_rate": 9.847001329696653e-05, | |
| "loss": 0.9123, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.003437107541807125, | |
| "grad_norm": 0.3165448307991028, | |
| "learning_rate": 9.826044551386744e-05, | |
| "loss": 0.9137, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 0.003569303985722784, | |
| "grad_norm": 0.30335527658462524, | |
| "learning_rate": 9.803768380684242e-05, | |
| "loss": 0.8245, | |
| "step": 27 | |
| }, | |
| { | |
| "epoch": 0.003701500429638443, | |
| "grad_norm": 0.30884429812431335, | |
| "learning_rate": 9.780178907671789e-05, | |
| "loss": 0.9312, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 0.0038336968735541016, | |
| "grad_norm": 0.303963303565979, | |
| "learning_rate": 9.755282581475769e-05, | |
| "loss": 0.9228, | |
| "step": 29 | |
| }, | |
| { | |
| "epoch": 0.00396589331746976, | |
| "grad_norm": 0.30520156025886536, | |
| "learning_rate": 9.729086208503174e-05, | |
| "loss": 0.8577, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.004098089761385418, | |
| "grad_norm": 0.3221169412136078, | |
| "learning_rate": 9.701596950580806e-05, | |
| "loss": 0.8526, | |
| "step": 31 | |
| }, | |
| { | |
| "epoch": 0.004230286205301077, | |
| "grad_norm": 0.3038342297077179, | |
| "learning_rate": 9.672822322997305e-05, | |
| "loss": 0.9267, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 0.004362482649216736, | |
| "grad_norm": 0.3137074410915375, | |
| "learning_rate": 9.642770192448536e-05, | |
| "loss": 0.809, | |
| "step": 33 | |
| }, | |
| { | |
| "epoch": 0.004494679093132395, | |
| "grad_norm": 0.3341916799545288, | |
| "learning_rate": 9.611448774886924e-05, | |
| "loss": 0.8703, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 0.0046268755370480535, | |
| "grad_norm": 0.34784042835235596, | |
| "learning_rate": 9.578866633275288e-05, | |
| "loss": 0.8523, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.004759071980963712, | |
| "grad_norm": 0.3425697982311249, | |
| "learning_rate": 9.545032675245813e-05, | |
| "loss": 0.8096, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 0.004891268424879371, | |
| "grad_norm": 0.3550679683685303, | |
| "learning_rate": 9.509956150664796e-05, | |
| "loss": 0.8501, | |
| "step": 37 | |
| }, | |
| { | |
| "epoch": 0.005023464868795029, | |
| "grad_norm": 0.35561543703079224, | |
| "learning_rate": 9.473646649103818e-05, | |
| "loss": 0.8739, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 0.005155661312710688, | |
| "grad_norm": 0.3632775843143463, | |
| "learning_rate": 9.43611409721806e-05, | |
| "loss": 0.8532, | |
| "step": 39 | |
| }, | |
| { | |
| "epoch": 0.005287857756626347, | |
| "grad_norm": 0.37438029050827026, | |
| "learning_rate": 9.397368756032445e-05, | |
| "loss": 0.7684, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.005420054200542005, | |
| "grad_norm": 0.3539464771747589, | |
| "learning_rate": 9.357421218136386e-05, | |
| "loss": 0.809, | |
| "step": 41 | |
| }, | |
| { | |
| "epoch": 0.005552250644457664, | |
| "grad_norm": 0.3721257448196411, | |
| "learning_rate": 9.316282404787871e-05, | |
| "loss": 0.8727, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 0.005684447088373323, | |
| "grad_norm": 0.3766726553440094, | |
| "learning_rate": 9.273963562927695e-05, | |
| "loss": 0.7634, | |
| "step": 43 | |
| }, | |
| { | |
| "epoch": 0.005816643532288982, | |
| "grad_norm": 0.33808189630508423, | |
| "learning_rate": 9.230476262104677e-05, | |
| "loss": 0.7711, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 0.00594883997620464, | |
| "grad_norm": 0.37606221437454224, | |
| "learning_rate": 9.185832391312644e-05, | |
| "loss": 0.7699, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 0.0060810364201202985, | |
| "grad_norm": 0.3631663918495178, | |
| "learning_rate": 9.140044155740101e-05, | |
| "loss": 0.7655, | |
| "step": 46 | |
| }, | |
| { | |
| "epoch": 0.006213232864035957, | |
| "grad_norm": 0.39946281909942627, | |
| "learning_rate": 9.093124073433463e-05, | |
| "loss": 0.7527, | |
| "step": 47 | |
| }, | |
| { | |
| "epoch": 0.006345429307951616, | |
| "grad_norm": 0.41046342253685, | |
| "learning_rate": 9.045084971874738e-05, | |
| "loss": 0.7463, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 0.006477625751867275, | |
| "grad_norm": 0.4188183844089508, | |
| "learning_rate": 8.995939984474624e-05, | |
| "loss": 0.7044, | |
| "step": 49 | |
| }, | |
| { | |
| "epoch": 0.006609822195782934, | |
| "grad_norm": 0.5937662720680237, | |
| "learning_rate": 8.945702546981969e-05, | |
| "loss": 0.6353, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.006609822195782934, | |
| "eval_loss": 0.7970728278160095, | |
| "eval_runtime": 1284.9623, | |
| "eval_samples_per_second": 9.915, | |
| "eval_steps_per_second": 2.479, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.0067420186396985925, | |
| "grad_norm": 0.2786080241203308, | |
| "learning_rate": 8.894386393810563e-05, | |
| "loss": 0.6512, | |
| "step": 51 | |
| }, | |
| { | |
| "epoch": 0.00687421508361425, | |
| "grad_norm": 0.27424800395965576, | |
| "learning_rate": 8.842005554284296e-05, | |
| "loss": 0.6597, | |
| "step": 52 | |
| }, | |
| { | |
| "epoch": 0.007006411527529909, | |
| "grad_norm": 0.2192063182592392, | |
| "learning_rate": 8.788574348801675e-05, | |
| "loss": 0.6719, | |
| "step": 53 | |
| }, | |
| { | |
| "epoch": 0.007138607971445568, | |
| "grad_norm": 0.2032398134469986, | |
| "learning_rate": 8.73410738492077e-05, | |
| "loss": 0.7025, | |
| "step": 54 | |
| }, | |
| { | |
| "epoch": 0.007270804415361227, | |
| "grad_norm": 0.19889365136623383, | |
| "learning_rate": 8.678619553365659e-05, | |
| "loss": 0.7491, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 0.007403000859276886, | |
| "grad_norm": 0.20415586233139038, | |
| "learning_rate": 8.622126023955446e-05, | |
| "loss": 0.6961, | |
| "step": 56 | |
| }, | |
| { | |
| "epoch": 0.007535197303192544, | |
| "grad_norm": 0.22474883496761322, | |
| "learning_rate": 8.564642241456986e-05, | |
| "loss": 0.7191, | |
| "step": 57 | |
| }, | |
| { | |
| "epoch": 0.007667393747108203, | |
| "grad_norm": 0.1937185525894165, | |
| "learning_rate": 8.506183921362443e-05, | |
| "loss": 0.7348, | |
| "step": 58 | |
| }, | |
| { | |
| "epoch": 0.007799590191023861, | |
| "grad_norm": 0.22836388647556305, | |
| "learning_rate": 8.44676704559283e-05, | |
| "loss": 0.7534, | |
| "step": 59 | |
| }, | |
| { | |
| "epoch": 0.00793178663493952, | |
| "grad_norm": 0.22243137657642365, | |
| "learning_rate": 8.386407858128706e-05, | |
| "loss": 0.7747, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.00806398307885518, | |
| "grad_norm": 0.2246285378932953, | |
| "learning_rate": 8.32512286056924e-05, | |
| "loss": 0.752, | |
| "step": 61 | |
| }, | |
| { | |
| "epoch": 0.008196179522770837, | |
| "grad_norm": 0.22273488342761993, | |
| "learning_rate": 8.262928807620843e-05, | |
| "loss": 0.8714, | |
| "step": 62 | |
| }, | |
| { | |
| "epoch": 0.008328375966686495, | |
| "grad_norm": 0.24067454040050507, | |
| "learning_rate": 8.199842702516583e-05, | |
| "loss": 0.8074, | |
| "step": 63 | |
| }, | |
| { | |
| "epoch": 0.008460572410602154, | |
| "grad_norm": 0.25970250368118286, | |
| "learning_rate": 8.135881792367686e-05, | |
| "loss": 0.6916, | |
| "step": 64 | |
| }, | |
| { | |
| "epoch": 0.008592768854517813, | |
| "grad_norm": 0.26505008339881897, | |
| "learning_rate": 8.07106356344834e-05, | |
| "loss": 0.8349, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 0.008724965298433472, | |
| "grad_norm": 0.292170912027359, | |
| "learning_rate": 8.005405736415126e-05, | |
| "loss": 0.8387, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 0.00885716174234913, | |
| "grad_norm": 0.2884085774421692, | |
| "learning_rate": 7.938926261462366e-05, | |
| "loss": 0.8228, | |
| "step": 67 | |
| }, | |
| { | |
| "epoch": 0.00898935818626479, | |
| "grad_norm": 0.2931410074234009, | |
| "learning_rate": 7.871643313414718e-05, | |
| "loss": 0.9008, | |
| "step": 68 | |
| }, | |
| { | |
| "epoch": 0.009121554630180448, | |
| "grad_norm": 0.2584120035171509, | |
| "learning_rate": 7.803575286758364e-05, | |
| "loss": 0.7052, | |
| "step": 69 | |
| }, | |
| { | |
| "epoch": 0.009253751074096107, | |
| "grad_norm": 0.2519213557243347, | |
| "learning_rate": 7.734740790612136e-05, | |
| "loss": 0.7115, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.009385947518011766, | |
| "grad_norm": 0.24533411860466003, | |
| "learning_rate": 7.66515864363997e-05, | |
| "loss": 0.7453, | |
| "step": 71 | |
| }, | |
| { | |
| "epoch": 0.009518143961927425, | |
| "grad_norm": 0.2644747197628021, | |
| "learning_rate": 7.594847868906076e-05, | |
| "loss": 0.8302, | |
| "step": 72 | |
| }, | |
| { | |
| "epoch": 0.009650340405843083, | |
| "grad_norm": 0.283643513917923, | |
| "learning_rate": 7.52382768867422e-05, | |
| "loss": 0.7019, | |
| "step": 73 | |
| }, | |
| { | |
| "epoch": 0.009782536849758742, | |
| "grad_norm": 0.26060959696769714, | |
| "learning_rate": 7.452117519152542e-05, | |
| "loss": 0.8799, | |
| "step": 74 | |
| }, | |
| { | |
| "epoch": 0.009914733293674401, | |
| "grad_norm": 0.2720307409763336, | |
| "learning_rate": 7.379736965185368e-05, | |
| "loss": 0.865, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.010046929737590058, | |
| "grad_norm": 0.26768940687179565, | |
| "learning_rate": 7.30670581489344e-05, | |
| "loss": 0.7436, | |
| "step": 76 | |
| }, | |
| { | |
| "epoch": 0.010179126181505717, | |
| "grad_norm": 0.27865222096443176, | |
| "learning_rate": 7.233044034264034e-05, | |
| "loss": 0.8181, | |
| "step": 77 | |
| }, | |
| { | |
| "epoch": 0.010311322625421376, | |
| "grad_norm": 0.2616931200027466, | |
| "learning_rate": 7.158771761692464e-05, | |
| "loss": 0.8217, | |
| "step": 78 | |
| }, | |
| { | |
| "epoch": 0.010443519069337034, | |
| "grad_norm": 0.2953363060951233, | |
| "learning_rate": 7.083909302476453e-05, | |
| "loss": 0.8209, | |
| "step": 79 | |
| }, | |
| { | |
| "epoch": 0.010575715513252693, | |
| "grad_norm": 0.295382022857666, | |
| "learning_rate": 7.008477123264848e-05, | |
| "loss": 0.7624, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.010707911957168352, | |
| "grad_norm": 0.30855387449264526, | |
| "learning_rate": 6.932495846462261e-05, | |
| "loss": 0.8067, | |
| "step": 81 | |
| }, | |
| { | |
| "epoch": 0.01084010840108401, | |
| "grad_norm": 0.307273268699646, | |
| "learning_rate": 6.855986244591104e-05, | |
| "loss": 0.8527, | |
| "step": 82 | |
| }, | |
| { | |
| "epoch": 0.01097230484499967, | |
| "grad_norm": 0.3054741322994232, | |
| "learning_rate": 6.778969234612584e-05, | |
| "loss": 0.6751, | |
| "step": 83 | |
| }, | |
| { | |
| "epoch": 0.011104501288915328, | |
| "grad_norm": 0.300057053565979, | |
| "learning_rate": 6.701465872208216e-05, | |
| "loss": 0.7919, | |
| "step": 84 | |
| }, | |
| { | |
| "epoch": 0.011236697732830987, | |
| "grad_norm": 0.3191482126712799, | |
| "learning_rate": 6.623497346023418e-05, | |
| "loss": 0.8327, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 0.011368894176746646, | |
| "grad_norm": 0.3450177311897278, | |
| "learning_rate": 6.545084971874738e-05, | |
| "loss": 0.7622, | |
| "step": 86 | |
| }, | |
| { | |
| "epoch": 0.011501090620662305, | |
| "grad_norm": 0.3140789270401001, | |
| "learning_rate": 6.466250186922325e-05, | |
| "loss": 0.6407, | |
| "step": 87 | |
| }, | |
| { | |
| "epoch": 0.011633287064577964, | |
| "grad_norm": 0.327387273311615, | |
| "learning_rate": 6.387014543809223e-05, | |
| "loss": 0.8469, | |
| "step": 88 | |
| }, | |
| { | |
| "epoch": 0.01176548350849362, | |
| "grad_norm": 0.33975428342819214, | |
| "learning_rate": 6.307399704769099e-05, | |
| "loss": 0.7539, | |
| "step": 89 | |
| }, | |
| { | |
| "epoch": 0.01189767995240928, | |
| "grad_norm": 0.3623250722885132, | |
| "learning_rate": 6.227427435703997e-05, | |
| "loss": 0.7197, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.012029876396324938, | |
| "grad_norm": 0.3607431948184967, | |
| "learning_rate": 6.147119600233758e-05, | |
| "loss": 0.693, | |
| "step": 91 | |
| }, | |
| { | |
| "epoch": 0.012162072840240597, | |
| "grad_norm": 0.34808632731437683, | |
| "learning_rate": 6.066498153718735e-05, | |
| "loss": 0.7834, | |
| "step": 92 | |
| }, | |
| { | |
| "epoch": 0.012294269284156256, | |
| "grad_norm": 0.36442074179649353, | |
| "learning_rate": 5.985585137257401e-05, | |
| "loss": 0.7211, | |
| "step": 93 | |
| }, | |
| { | |
| "epoch": 0.012426465728071915, | |
| "grad_norm": 0.42843687534332275, | |
| "learning_rate": 5.90440267166055e-05, | |
| "loss": 0.7012, | |
| "step": 94 | |
| }, | |
| { | |
| "epoch": 0.012558662171987573, | |
| "grad_norm": 0.419827401638031, | |
| "learning_rate": 5.8229729514036705e-05, | |
| "loss": 0.7325, | |
| "step": 95 | |
| }, | |
| { | |
| "epoch": 0.012690858615903232, | |
| "grad_norm": 0.4016064703464508, | |
| "learning_rate": 5.74131823855921e-05, | |
| "loss": 0.7227, | |
| "step": 96 | |
| }, | |
| { | |
| "epoch": 0.012823055059818891, | |
| "grad_norm": 0.5434070229530334, | |
| "learning_rate": 5.6594608567103456e-05, | |
| "loss": 0.6988, | |
| "step": 97 | |
| }, | |
| { | |
| "epoch": 0.01295525150373455, | |
| "grad_norm": 0.42109209299087524, | |
| "learning_rate": 5.577423184847932e-05, | |
| "loss": 0.6488, | |
| "step": 98 | |
| }, | |
| { | |
| "epoch": 0.013087447947650209, | |
| "grad_norm": 0.4489916265010834, | |
| "learning_rate": 5.495227651252315e-05, | |
| "loss": 0.6232, | |
| "step": 99 | |
| }, | |
| { | |
| "epoch": 0.013219644391565867, | |
| "grad_norm": 0.5862768888473511, | |
| "learning_rate": 5.4128967273616625e-05, | |
| "loss": 0.6915, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.013219644391565867, | |
| "eval_loss": 0.7334732413291931, | |
| "eval_runtime": 1283.9042, | |
| "eval_samples_per_second": 9.923, | |
| "eval_steps_per_second": 2.481, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.013351840835481526, | |
| "grad_norm": 0.24181707203388214, | |
| "learning_rate": 5.330452921628497e-05, | |
| "loss": 0.596, | |
| "step": 101 | |
| }, | |
| { | |
| "epoch": 0.013484037279397185, | |
| "grad_norm": 0.23782025277614594, | |
| "learning_rate": 5.247918773366112e-05, | |
| "loss": 0.5794, | |
| "step": 102 | |
| }, | |
| { | |
| "epoch": 0.013616233723312842, | |
| "grad_norm": 0.2528301775455475, | |
| "learning_rate": 5.165316846586541e-05, | |
| "loss": 0.5885, | |
| "step": 103 | |
| }, | |
| { | |
| "epoch": 0.0137484301672285, | |
| "grad_norm": 0.21014751493930817, | |
| "learning_rate": 5.0826697238317935e-05, | |
| "loss": 0.5951, | |
| "step": 104 | |
| }, | |
| { | |
| "epoch": 0.01388062661114416, | |
| "grad_norm": 0.22236773371696472, | |
| "learning_rate": 5e-05, | |
| "loss": 0.7092, | |
| "step": 105 | |
| }, | |
| { | |
| "epoch": 0.014012823055059818, | |
| "grad_norm": 0.21512827277183533, | |
| "learning_rate": 4.917330276168208e-05, | |
| "loss": 0.6623, | |
| "step": 106 | |
| }, | |
| { | |
| "epoch": 0.014145019498975477, | |
| "grad_norm": 0.21951237320899963, | |
| "learning_rate": 4.834683153413459e-05, | |
| "loss": 0.6856, | |
| "step": 107 | |
| }, | |
| { | |
| "epoch": 0.014277215942891136, | |
| "grad_norm": 0.21105462312698364, | |
| "learning_rate": 4.7520812266338885e-05, | |
| "loss": 0.5825, | |
| "step": 108 | |
| }, | |
| { | |
| "epoch": 0.014409412386806795, | |
| "grad_norm": 0.27910059690475464, | |
| "learning_rate": 4.669547078371504e-05, | |
| "loss": 0.8379, | |
| "step": 109 | |
| }, | |
| { | |
| "epoch": 0.014541608830722454, | |
| "grad_norm": 0.2623700499534607, | |
| "learning_rate": 4.5871032726383386e-05, | |
| "loss": 0.7664, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.014673805274638112, | |
| "grad_norm": 0.24401693046092987, | |
| "learning_rate": 4.504772348747687e-05, | |
| "loss": 0.6521, | |
| "step": 111 | |
| }, | |
| { | |
| "epoch": 0.014806001718553771, | |
| "grad_norm": 0.2733868658542633, | |
| "learning_rate": 4.4225768151520694e-05, | |
| "loss": 0.6896, | |
| "step": 112 | |
| }, | |
| { | |
| "epoch": 0.01493819816246943, | |
| "grad_norm": 0.29973694682121277, | |
| "learning_rate": 4.3405391432896555e-05, | |
| "loss": 0.8034, | |
| "step": 113 | |
| }, | |
| { | |
| "epoch": 0.015070394606385089, | |
| "grad_norm": 0.2729474604129791, | |
| "learning_rate": 4.2586817614407895e-05, | |
| "loss": 0.721, | |
| "step": 114 | |
| }, | |
| { | |
| "epoch": 0.015202591050300748, | |
| "grad_norm": 0.3053736686706543, | |
| "learning_rate": 4.17702704859633e-05, | |
| "loss": 0.7217, | |
| "step": 115 | |
| }, | |
| { | |
| "epoch": 0.015334787494216406, | |
| "grad_norm": 0.30556952953338623, | |
| "learning_rate": 4.095597328339452e-05, | |
| "loss": 0.736, | |
| "step": 116 | |
| }, | |
| { | |
| "epoch": 0.015466983938132063, | |
| "grad_norm": 0.356710284948349, | |
| "learning_rate": 4.0144148627425993e-05, | |
| "loss": 0.8814, | |
| "step": 117 | |
| }, | |
| { | |
| "epoch": 0.015599180382047722, | |
| "grad_norm": 0.28044766187667847, | |
| "learning_rate": 3.933501846281267e-05, | |
| "loss": 0.7973, | |
| "step": 118 | |
| }, | |
| { | |
| "epoch": 0.015731376825963383, | |
| "grad_norm": 0.26613718271255493, | |
| "learning_rate": 3.852880399766243e-05, | |
| "loss": 0.7435, | |
| "step": 119 | |
| }, | |
| { | |
| "epoch": 0.01586357326987904, | |
| "grad_norm": 0.2693482041358948, | |
| "learning_rate": 3.772572564296005e-05, | |
| "loss": 0.7187, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.0159957697137947, | |
| "grad_norm": 0.278363436460495, | |
| "learning_rate": 3.6926002952309016e-05, | |
| "loss": 0.6341, | |
| "step": 121 | |
| }, | |
| { | |
| "epoch": 0.01612796615771036, | |
| "grad_norm": 0.2664760947227478, | |
| "learning_rate": 3.612985456190778e-05, | |
| "loss": 0.6746, | |
| "step": 122 | |
| }, | |
| { | |
| "epoch": 0.016260162601626018, | |
| "grad_norm": 0.2626112103462219, | |
| "learning_rate": 3.533749813077677e-05, | |
| "loss": 0.6684, | |
| "step": 123 | |
| }, | |
| { | |
| "epoch": 0.016392359045541673, | |
| "grad_norm": 0.2720891237258911, | |
| "learning_rate": 3.4549150281252636e-05, | |
| "loss": 0.7468, | |
| "step": 124 | |
| }, | |
| { | |
| "epoch": 0.016524555489457332, | |
| "grad_norm": 0.32452595233917236, | |
| "learning_rate": 3.3765026539765834e-05, | |
| "loss": 0.8739, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 0.01665675193337299, | |
| "grad_norm": 0.2681269943714142, | |
| "learning_rate": 3.298534127791785e-05, | |
| "loss": 0.6395, | |
| "step": 126 | |
| }, | |
| { | |
| "epoch": 0.01678894837728865, | |
| "grad_norm": 0.2877364158630371, | |
| "learning_rate": 3.221030765387417e-05, | |
| "loss": 0.7308, | |
| "step": 127 | |
| }, | |
| { | |
| "epoch": 0.01692114482120431, | |
| "grad_norm": 0.28337937593460083, | |
| "learning_rate": 3.144013755408895e-05, | |
| "loss": 0.6387, | |
| "step": 128 | |
| }, | |
| { | |
| "epoch": 0.017053341265119967, | |
| "grad_norm": 0.3185059428215027, | |
| "learning_rate": 3.0675041535377405e-05, | |
| "loss": 0.8061, | |
| "step": 129 | |
| }, | |
| { | |
| "epoch": 0.017185537709035626, | |
| "grad_norm": 0.2694532573223114, | |
| "learning_rate": 2.991522876735154e-05, | |
| "loss": 0.6599, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.017317734152951285, | |
| "grad_norm": 0.3253261148929596, | |
| "learning_rate": 2.916090697523549e-05, | |
| "loss": 0.7728, | |
| "step": 131 | |
| }, | |
| { | |
| "epoch": 0.017449930596866944, | |
| "grad_norm": 0.3177144229412079, | |
| "learning_rate": 2.8412282383075363e-05, | |
| "loss": 0.7772, | |
| "step": 132 | |
| }, | |
| { | |
| "epoch": 0.017582127040782602, | |
| "grad_norm": 0.2779819071292877, | |
| "learning_rate": 2.766955965735968e-05, | |
| "loss": 0.6469, | |
| "step": 133 | |
| }, | |
| { | |
| "epoch": 0.01771432348469826, | |
| "grad_norm": 0.29959800839424133, | |
| "learning_rate": 2.693294185106562e-05, | |
| "loss": 0.6401, | |
| "step": 134 | |
| }, | |
| { | |
| "epoch": 0.01784651992861392, | |
| "grad_norm": 0.3584449887275696, | |
| "learning_rate": 2.6202630348146324e-05, | |
| "loss": 0.8156, | |
| "step": 135 | |
| }, | |
| { | |
| "epoch": 0.01797871637252958, | |
| "grad_norm": 0.3743143677711487, | |
| "learning_rate": 2.547882480847461e-05, | |
| "loss": 0.7496, | |
| "step": 136 | |
| }, | |
| { | |
| "epoch": 0.018110912816445238, | |
| "grad_norm": 0.37764647603034973, | |
| "learning_rate": 2.476172311325783e-05, | |
| "loss": 0.8135, | |
| "step": 137 | |
| }, | |
| { | |
| "epoch": 0.018243109260360896, | |
| "grad_norm": 0.34589067101478577, | |
| "learning_rate": 2.405152131093926e-05, | |
| "loss": 0.7126, | |
| "step": 138 | |
| }, | |
| { | |
| "epoch": 0.018375305704276555, | |
| "grad_norm": 0.3663211762905121, | |
| "learning_rate": 2.3348413563600325e-05, | |
| "loss": 0.7474, | |
| "step": 139 | |
| }, | |
| { | |
| "epoch": 0.018507502148192214, | |
| "grad_norm": 0.3387898802757263, | |
| "learning_rate": 2.2652592093878666e-05, | |
| "loss": 0.6598, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.018639698592107873, | |
| "grad_norm": 0.44442814588546753, | |
| "learning_rate": 2.196424713241637e-05, | |
| "loss": 0.7162, | |
| "step": 141 | |
| }, | |
| { | |
| "epoch": 0.01877189503602353, | |
| "grad_norm": 0.38971763849258423, | |
| "learning_rate": 2.128356686585282e-05, | |
| "loss": 0.7364, | |
| "step": 142 | |
| }, | |
| { | |
| "epoch": 0.01890409147993919, | |
| "grad_norm": 0.3823245167732239, | |
| "learning_rate": 2.061073738537635e-05, | |
| "loss": 0.7215, | |
| "step": 143 | |
| }, | |
| { | |
| "epoch": 0.01903628792385485, | |
| "grad_norm": 0.40516719222068787, | |
| "learning_rate": 1.9945942635848748e-05, | |
| "loss": 0.7439, | |
| "step": 144 | |
| }, | |
| { | |
| "epoch": 0.019168484367770508, | |
| "grad_norm": 0.3880475163459778, | |
| "learning_rate": 1.928936436551661e-05, | |
| "loss": 0.7098, | |
| "step": 145 | |
| }, | |
| { | |
| "epoch": 0.019300680811686167, | |
| "grad_norm": 0.39626526832580566, | |
| "learning_rate": 1.8641182076323148e-05, | |
| "loss": 0.6337, | |
| "step": 146 | |
| }, | |
| { | |
| "epoch": 0.019432877255601826, | |
| "grad_norm": 0.40710315108299255, | |
| "learning_rate": 1.800157297483417e-05, | |
| "loss": 0.6118, | |
| "step": 147 | |
| }, | |
| { | |
| "epoch": 0.019565073699517484, | |
| "grad_norm": 0.40455126762390137, | |
| "learning_rate": 1.7370711923791567e-05, | |
| "loss": 0.5703, | |
| "step": 148 | |
| }, | |
| { | |
| "epoch": 0.019697270143433143, | |
| "grad_norm": 0.4407866299152374, | |
| "learning_rate": 1.6748771394307585e-05, | |
| "loss": 0.5417, | |
| "step": 149 | |
| }, | |
| { | |
| "epoch": 0.019829466587348802, | |
| "grad_norm": 0.6249489188194275, | |
| "learning_rate": 1.6135921418712956e-05, | |
| "loss": 0.5839, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.019829466587348802, | |
| "eval_loss": 0.7003475427627563, | |
| "eval_runtime": 1285.1101, | |
| "eval_samples_per_second": 9.914, | |
| "eval_steps_per_second": 2.478, | |
| "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": 2.8073460836912333e+17, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |