| { | |
| "best_metric": 0.043948203325271606, | |
| "best_model_checkpoint": "miner_id_24/checkpoint-200", | |
| "epoch": 0.11273957158962795, | |
| "eval_steps": 50, | |
| "global_step": 200, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0005636978579481398, | |
| "grad_norm": 2.950288772583008, | |
| "learning_rate": 1e-05, | |
| "loss": 0.4327, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.0005636978579481398, | |
| "eval_loss": 0.8793883919715881, | |
| "eval_runtime": 73.5587, | |
| "eval_samples_per_second": 40.621, | |
| "eval_steps_per_second": 10.155, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.0011273957158962795, | |
| "grad_norm": 3.1511483192443848, | |
| "learning_rate": 2e-05, | |
| "loss": 0.4627, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.0016910935738444193, | |
| "grad_norm": 4.022647380828857, | |
| "learning_rate": 3e-05, | |
| "loss": 0.4798, | |
| "step": 3 | |
| }, | |
| { | |
| "epoch": 0.002254791431792559, | |
| "grad_norm": 42.159236907958984, | |
| "learning_rate": 4e-05, | |
| "loss": 0.5633, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.002818489289740699, | |
| "grad_norm": 7.003749847412109, | |
| "learning_rate": 5e-05, | |
| "loss": 0.4195, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.0033821871476888386, | |
| "grad_norm": 2.4085493087768555, | |
| "learning_rate": 6e-05, | |
| "loss": 0.3117, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.003945885005636978, | |
| "grad_norm": 3.7020814418792725, | |
| "learning_rate": 7e-05, | |
| "loss": 0.2846, | |
| "step": 7 | |
| }, | |
| { | |
| "epoch": 0.004509582863585118, | |
| "grad_norm": 8.4557523727417, | |
| "learning_rate": 8e-05, | |
| "loss": 0.2238, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.005073280721533258, | |
| "grad_norm": 11.59519100189209, | |
| "learning_rate": 9e-05, | |
| "loss": 0.1024, | |
| "step": 9 | |
| }, | |
| { | |
| "epoch": 0.005636978579481398, | |
| "grad_norm": 2.2653753757476807, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0583, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.0062006764374295375, | |
| "grad_norm": 1.2324670553207397, | |
| "learning_rate": 9.999316524962345e-05, | |
| "loss": 0.0264, | |
| "step": 11 | |
| }, | |
| { | |
| "epoch": 0.006764374295377677, | |
| "grad_norm": 2.2194106578826904, | |
| "learning_rate": 9.997266286704631e-05, | |
| "loss": 0.0159, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 0.007328072153325817, | |
| "grad_norm": 1.2577799558639526, | |
| "learning_rate": 9.993849845741524e-05, | |
| "loss": 0.0077, | |
| "step": 13 | |
| }, | |
| { | |
| "epoch": 0.007891770011273957, | |
| "grad_norm": 4.907573223114014, | |
| "learning_rate": 9.989068136093873e-05, | |
| "loss": 0.0238, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 0.008455467869222097, | |
| "grad_norm": 0.7712904810905457, | |
| "learning_rate": 9.98292246503335e-05, | |
| "loss": 0.0067, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.009019165727170236, | |
| "grad_norm": 0.0640689879655838, | |
| "learning_rate": 9.975414512725057e-05, | |
| "loss": 0.0022, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 0.009582863585118376, | |
| "grad_norm": 0.07617411762475967, | |
| "learning_rate": 9.966546331768191e-05, | |
| "loss": 0.0022, | |
| "step": 17 | |
| }, | |
| { | |
| "epoch": 0.010146561443066516, | |
| "grad_norm": 0.0818643644452095, | |
| "learning_rate": 9.956320346634876e-05, | |
| "loss": 0.0041, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 0.010710259301014656, | |
| "grad_norm": 0.09318102896213531, | |
| "learning_rate": 9.944739353007344e-05, | |
| "loss": 0.0017, | |
| "step": 19 | |
| }, | |
| { | |
| "epoch": 0.011273957158962795, | |
| "grad_norm": 0.18323451280593872, | |
| "learning_rate": 9.931806517013612e-05, | |
| "loss": 0.0024, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.011837655016910935, | |
| "grad_norm": 0.1340917944908142, | |
| "learning_rate": 9.917525374361912e-05, | |
| "loss": 0.0114, | |
| "step": 21 | |
| }, | |
| { | |
| "epoch": 0.012401352874859075, | |
| "grad_norm": 0.13049083948135376, | |
| "learning_rate": 9.901899829374047e-05, | |
| "loss": 0.0019, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 0.012965050732807215, | |
| "grad_norm": 0.17436546087265015, | |
| "learning_rate": 9.884934153917997e-05, | |
| "loss": 0.0077, | |
| "step": 23 | |
| }, | |
| { | |
| "epoch": 0.013528748590755355, | |
| "grad_norm": 0.12390507757663727, | |
| "learning_rate": 9.86663298624003e-05, | |
| "loss": 0.0021, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 0.014092446448703494, | |
| "grad_norm": 0.0859021544456482, | |
| "learning_rate": 9.847001329696653e-05, | |
| "loss": 0.0071, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.014656144306651634, | |
| "grad_norm": 0.032982051372528076, | |
| "learning_rate": 9.826044551386744e-05, | |
| "loss": 0.0007, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 0.015219842164599774, | |
| "grad_norm": 0.16294588148593903, | |
| "learning_rate": 9.803768380684242e-05, | |
| "loss": 0.0041, | |
| "step": 27 | |
| }, | |
| { | |
| "epoch": 0.015783540022547914, | |
| "grad_norm": 0.10073961317539215, | |
| "learning_rate": 9.780178907671789e-05, | |
| "loss": 0.0039, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 0.016347237880496055, | |
| "grad_norm": 0.13056406378746033, | |
| "learning_rate": 9.755282581475769e-05, | |
| "loss": 0.0031, | |
| "step": 29 | |
| }, | |
| { | |
| "epoch": 0.016910935738444193, | |
| "grad_norm": 0.04367830604314804, | |
| "learning_rate": 9.729086208503174e-05, | |
| "loss": 0.0007, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.017474633596392335, | |
| "grad_norm": 0.5413928031921387, | |
| "learning_rate": 9.701596950580806e-05, | |
| "loss": 0.0173, | |
| "step": 31 | |
| }, | |
| { | |
| "epoch": 0.018038331454340473, | |
| "grad_norm": 0.3649708926677704, | |
| "learning_rate": 9.672822322997305e-05, | |
| "loss": 0.0507, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 0.018602029312288614, | |
| "grad_norm": 1.2466174364089966, | |
| "learning_rate": 9.642770192448536e-05, | |
| "loss": 0.1868, | |
| "step": 33 | |
| }, | |
| { | |
| "epoch": 0.019165727170236752, | |
| "grad_norm": 1.1820372343063354, | |
| "learning_rate": 9.611448774886924e-05, | |
| "loss": 0.2321, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 0.019729425028184894, | |
| "grad_norm": 0.8594322800636292, | |
| "learning_rate": 9.578866633275288e-05, | |
| "loss": 0.1586, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.020293122886133032, | |
| "grad_norm": 0.6499341726303101, | |
| "learning_rate": 9.545032675245813e-05, | |
| "loss": 0.1365, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 0.020856820744081173, | |
| "grad_norm": 0.8684273958206177, | |
| "learning_rate": 9.509956150664796e-05, | |
| "loss": 0.1355, | |
| "step": 37 | |
| }, | |
| { | |
| "epoch": 0.02142051860202931, | |
| "grad_norm": 0.663493812084198, | |
| "learning_rate": 9.473646649103818e-05, | |
| "loss": 0.1322, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 0.021984216459977453, | |
| "grad_norm": 0.5041710138320923, | |
| "learning_rate": 9.43611409721806e-05, | |
| "loss": 0.1279, | |
| "step": 39 | |
| }, | |
| { | |
| "epoch": 0.02254791431792559, | |
| "grad_norm": 0.6379398107528687, | |
| "learning_rate": 9.397368756032445e-05, | |
| "loss": 0.103, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.023111612175873732, | |
| "grad_norm": 0.421265184879303, | |
| "learning_rate": 9.357421218136386e-05, | |
| "loss": 0.0868, | |
| "step": 41 | |
| }, | |
| { | |
| "epoch": 0.02367531003382187, | |
| "grad_norm": 0.5050786137580872, | |
| "learning_rate": 9.316282404787871e-05, | |
| "loss": 0.1548, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 0.024239007891770012, | |
| "grad_norm": 0.42436856031417847, | |
| "learning_rate": 9.273963562927695e-05, | |
| "loss": 0.1258, | |
| "step": 43 | |
| }, | |
| { | |
| "epoch": 0.02480270574971815, | |
| "grad_norm": 0.3885938227176666, | |
| "learning_rate": 9.230476262104677e-05, | |
| "loss": 0.0869, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 0.02536640360766629, | |
| "grad_norm": 0.3994680941104889, | |
| "learning_rate": 9.185832391312644e-05, | |
| "loss": 0.0937, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 0.02593010146561443, | |
| "grad_norm": 0.50786292552948, | |
| "learning_rate": 9.140044155740101e-05, | |
| "loss": 0.143, | |
| "step": 46 | |
| }, | |
| { | |
| "epoch": 0.02649379932356257, | |
| "grad_norm": 0.6507979035377502, | |
| "learning_rate": 9.093124073433463e-05, | |
| "loss": 0.1599, | |
| "step": 47 | |
| }, | |
| { | |
| "epoch": 0.02705749718151071, | |
| "grad_norm": 0.7517385482788086, | |
| "learning_rate": 9.045084971874738e-05, | |
| "loss": 0.2804, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 0.02762119503945885, | |
| "grad_norm": 0.5586478114128113, | |
| "learning_rate": 8.995939984474624e-05, | |
| "loss": 0.1852, | |
| "step": 49 | |
| }, | |
| { | |
| "epoch": 0.02818489289740699, | |
| "grad_norm": 1.1680113077163696, | |
| "learning_rate": 8.945702546981969e-05, | |
| "loss": 0.2137, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.02818489289740699, | |
| "eval_loss": 0.13373692333698273, | |
| "eval_runtime": 73.9286, | |
| "eval_samples_per_second": 40.417, | |
| "eval_steps_per_second": 10.104, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.02874859075535513, | |
| "grad_norm": 1.163334608078003, | |
| "learning_rate": 8.894386393810563e-05, | |
| "loss": 0.1193, | |
| "step": 51 | |
| }, | |
| { | |
| "epoch": 0.029312288613303268, | |
| "grad_norm": 0.8139949440956116, | |
| "learning_rate": 8.842005554284296e-05, | |
| "loss": 0.068, | |
| "step": 52 | |
| }, | |
| { | |
| "epoch": 0.02987598647125141, | |
| "grad_norm": 0.3827410638332367, | |
| "learning_rate": 8.788574348801675e-05, | |
| "loss": 0.0261, | |
| "step": 53 | |
| }, | |
| { | |
| "epoch": 0.030439684329199548, | |
| "grad_norm": 0.24657344818115234, | |
| "learning_rate": 8.73410738492077e-05, | |
| "loss": 0.0169, | |
| "step": 54 | |
| }, | |
| { | |
| "epoch": 0.03100338218714769, | |
| "grad_norm": 0.3972528874874115, | |
| "learning_rate": 8.678619553365659e-05, | |
| "loss": 0.0123, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 0.03156708004509583, | |
| "grad_norm": 0.5841863751411438, | |
| "learning_rate": 8.622126023955446e-05, | |
| "loss": 0.0126, | |
| "step": 56 | |
| }, | |
| { | |
| "epoch": 0.032130777903043965, | |
| "grad_norm": 0.2513751685619354, | |
| "learning_rate": 8.564642241456986e-05, | |
| "loss": 0.0066, | |
| "step": 57 | |
| }, | |
| { | |
| "epoch": 0.03269447576099211, | |
| "grad_norm": 0.12783212959766388, | |
| "learning_rate": 8.506183921362443e-05, | |
| "loss": 0.0037, | |
| "step": 58 | |
| }, | |
| { | |
| "epoch": 0.03325817361894025, | |
| "grad_norm": 0.16285184025764465, | |
| "learning_rate": 8.44676704559283e-05, | |
| "loss": 0.0034, | |
| "step": 59 | |
| }, | |
| { | |
| "epoch": 0.033821871476888386, | |
| "grad_norm": 0.08129656314849854, | |
| "learning_rate": 8.386407858128706e-05, | |
| "loss": 0.002, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.034385569334836524, | |
| "grad_norm": 0.05978225916624069, | |
| "learning_rate": 8.32512286056924e-05, | |
| "loss": 0.0039, | |
| "step": 61 | |
| }, | |
| { | |
| "epoch": 0.03494926719278467, | |
| "grad_norm": 0.16640277206897736, | |
| "learning_rate": 8.262928807620843e-05, | |
| "loss": 0.0113, | |
| "step": 62 | |
| }, | |
| { | |
| "epoch": 0.03551296505073281, | |
| "grad_norm": 0.12526294589042664, | |
| "learning_rate": 8.199842702516583e-05, | |
| "loss": 0.0144, | |
| "step": 63 | |
| }, | |
| { | |
| "epoch": 0.036076662908680945, | |
| "grad_norm": 0.030233556404709816, | |
| "learning_rate": 8.135881792367686e-05, | |
| "loss": 0.0011, | |
| "step": 64 | |
| }, | |
| { | |
| "epoch": 0.03664036076662908, | |
| "grad_norm": 0.03970768302679062, | |
| "learning_rate": 8.07106356344834e-05, | |
| "loss": 0.0014, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 0.03720405862457723, | |
| "grad_norm": 0.07165244221687317, | |
| "learning_rate": 8.005405736415126e-05, | |
| "loss": 0.0011, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 0.037767756482525366, | |
| "grad_norm": 0.07285843044519424, | |
| "learning_rate": 7.938926261462366e-05, | |
| "loss": 0.0049, | |
| "step": 67 | |
| }, | |
| { | |
| "epoch": 0.038331454340473504, | |
| "grad_norm": 0.10785462707281113, | |
| "learning_rate": 7.871643313414718e-05, | |
| "loss": 0.002, | |
| "step": 68 | |
| }, | |
| { | |
| "epoch": 0.03889515219842165, | |
| "grad_norm": 0.058972012251615524, | |
| "learning_rate": 7.803575286758364e-05, | |
| "loss": 0.004, | |
| "step": 69 | |
| }, | |
| { | |
| "epoch": 0.03945885005636979, | |
| "grad_norm": 0.043987058103084564, | |
| "learning_rate": 7.734740790612136e-05, | |
| "loss": 0.002, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.040022547914317926, | |
| "grad_norm": 0.027181290090084076, | |
| "learning_rate": 7.66515864363997e-05, | |
| "loss": 0.0007, | |
| "step": 71 | |
| }, | |
| { | |
| "epoch": 0.040586245772266064, | |
| "grad_norm": 0.055068809539079666, | |
| "learning_rate": 7.594847868906076e-05, | |
| "loss": 0.0058, | |
| "step": 72 | |
| }, | |
| { | |
| "epoch": 0.04114994363021421, | |
| "grad_norm": 0.04421454295516014, | |
| "learning_rate": 7.52382768867422e-05, | |
| "loss": 0.0013, | |
| "step": 73 | |
| }, | |
| { | |
| "epoch": 0.04171364148816235, | |
| "grad_norm": 0.04847081005573273, | |
| "learning_rate": 7.452117519152542e-05, | |
| "loss": 0.0024, | |
| "step": 74 | |
| }, | |
| { | |
| "epoch": 0.042277339346110485, | |
| "grad_norm": 0.1457040160894394, | |
| "learning_rate": 7.379736965185368e-05, | |
| "loss": 0.0025, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.04284103720405862, | |
| "grad_norm": 0.10540561378002167, | |
| "learning_rate": 7.30670581489344e-05, | |
| "loss": 0.0145, | |
| "step": 76 | |
| }, | |
| { | |
| "epoch": 0.04340473506200677, | |
| "grad_norm": 0.06257438659667969, | |
| "learning_rate": 7.233044034264034e-05, | |
| "loss": 0.0037, | |
| "step": 77 | |
| }, | |
| { | |
| "epoch": 0.043968432919954906, | |
| "grad_norm": 0.2079205960035324, | |
| "learning_rate": 7.158771761692464e-05, | |
| "loss": 0.0104, | |
| "step": 78 | |
| }, | |
| { | |
| "epoch": 0.044532130777903044, | |
| "grad_norm": 0.054456714540719986, | |
| "learning_rate": 7.083909302476453e-05, | |
| "loss": 0.0006, | |
| "step": 79 | |
| }, | |
| { | |
| "epoch": 0.04509582863585118, | |
| "grad_norm": 0.1439305990934372, | |
| "learning_rate": 7.008477123264848e-05, | |
| "loss": 0.0081, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.04565952649379933, | |
| "grad_norm": 0.12464113533496857, | |
| "learning_rate": 6.932495846462261e-05, | |
| "loss": 0.0055, | |
| "step": 81 | |
| }, | |
| { | |
| "epoch": 0.046223224351747465, | |
| "grad_norm": 0.1310434490442276, | |
| "learning_rate": 6.855986244591104e-05, | |
| "loss": 0.0126, | |
| "step": 82 | |
| }, | |
| { | |
| "epoch": 0.0467869222096956, | |
| "grad_norm": 1.0021061897277832, | |
| "learning_rate": 6.778969234612584e-05, | |
| "loss": 0.0838, | |
| "step": 83 | |
| }, | |
| { | |
| "epoch": 0.04735062006764374, | |
| "grad_norm": 1.1204408407211304, | |
| "learning_rate": 6.701465872208216e-05, | |
| "loss": 0.1707, | |
| "step": 84 | |
| }, | |
| { | |
| "epoch": 0.047914317925591886, | |
| "grad_norm": 1.0282710790634155, | |
| "learning_rate": 6.623497346023418e-05, | |
| "loss": 0.1412, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 0.048478015783540024, | |
| "grad_norm": 0.5508891940116882, | |
| "learning_rate": 6.545084971874738e-05, | |
| "loss": 0.1321, | |
| "step": 86 | |
| }, | |
| { | |
| "epoch": 0.04904171364148816, | |
| "grad_norm": 0.39298781752586365, | |
| "learning_rate": 6.466250186922325e-05, | |
| "loss": 0.1122, | |
| "step": 87 | |
| }, | |
| { | |
| "epoch": 0.0496054114994363, | |
| "grad_norm": 0.28032705187797546, | |
| "learning_rate": 6.387014543809223e-05, | |
| "loss": 0.0979, | |
| "step": 88 | |
| }, | |
| { | |
| "epoch": 0.050169109357384445, | |
| "grad_norm": 0.3497227430343628, | |
| "learning_rate": 6.307399704769099e-05, | |
| "loss": 0.114, | |
| "step": 89 | |
| }, | |
| { | |
| "epoch": 0.05073280721533258, | |
| "grad_norm": 0.24456681311130524, | |
| "learning_rate": 6.227427435703997e-05, | |
| "loss": 0.0908, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.05129650507328072, | |
| "grad_norm": 0.19479884207248688, | |
| "learning_rate": 6.147119600233758e-05, | |
| "loss": 0.0769, | |
| "step": 91 | |
| }, | |
| { | |
| "epoch": 0.05186020293122886, | |
| "grad_norm": 0.2992049753665924, | |
| "learning_rate": 6.066498153718735e-05, | |
| "loss": 0.1043, | |
| "step": 92 | |
| }, | |
| { | |
| "epoch": 0.052423900789177004, | |
| "grad_norm": 0.1763327270746231, | |
| "learning_rate": 5.985585137257401e-05, | |
| "loss": 0.0648, | |
| "step": 93 | |
| }, | |
| { | |
| "epoch": 0.05298759864712514, | |
| "grad_norm": 0.2888164818286896, | |
| "learning_rate": 5.90440267166055e-05, | |
| "loss": 0.1217, | |
| "step": 94 | |
| }, | |
| { | |
| "epoch": 0.05355129650507328, | |
| "grad_norm": 0.2486395686864853, | |
| "learning_rate": 5.8229729514036705e-05, | |
| "loss": 0.0691, | |
| "step": 95 | |
| }, | |
| { | |
| "epoch": 0.05411499436302142, | |
| "grad_norm": 0.634675145149231, | |
| "learning_rate": 5.74131823855921e-05, | |
| "loss": 0.0948, | |
| "step": 96 | |
| }, | |
| { | |
| "epoch": 0.05467869222096956, | |
| "grad_norm": 0.41564327478408813, | |
| "learning_rate": 5.6594608567103456e-05, | |
| "loss": 0.1033, | |
| "step": 97 | |
| }, | |
| { | |
| "epoch": 0.0552423900789177, | |
| "grad_norm": 0.3561808466911316, | |
| "learning_rate": 5.577423184847932e-05, | |
| "loss": 0.1098, | |
| "step": 98 | |
| }, | |
| { | |
| "epoch": 0.05580608793686584, | |
| "grad_norm": 0.5383174419403076, | |
| "learning_rate": 5.495227651252315e-05, | |
| "loss": 0.197, | |
| "step": 99 | |
| }, | |
| { | |
| "epoch": 0.05636978579481398, | |
| "grad_norm": 0.6031817197799683, | |
| "learning_rate": 5.4128967273616625e-05, | |
| "loss": 0.2217, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.05636978579481398, | |
| "eval_loss": 0.07134784758090973, | |
| "eval_runtime": 73.8963, | |
| "eval_samples_per_second": 40.435, | |
| "eval_steps_per_second": 10.109, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.05693348365276212, | |
| "grad_norm": 0.6925469636917114, | |
| "learning_rate": 5.330452921628497e-05, | |
| "loss": 0.0474, | |
| "step": 101 | |
| }, | |
| { | |
| "epoch": 0.05749718151071026, | |
| "grad_norm": 0.4460764527320862, | |
| "learning_rate": 5.247918773366112e-05, | |
| "loss": 0.0281, | |
| "step": 102 | |
| }, | |
| { | |
| "epoch": 0.0580608793686584, | |
| "grad_norm": 0.27097219228744507, | |
| "learning_rate": 5.165316846586541e-05, | |
| "loss": 0.0129, | |
| "step": 103 | |
| }, | |
| { | |
| "epoch": 0.058624577226606536, | |
| "grad_norm": 0.20055565237998962, | |
| "learning_rate": 5.0826697238317935e-05, | |
| "loss": 0.008, | |
| "step": 104 | |
| }, | |
| { | |
| "epoch": 0.05918827508455468, | |
| "grad_norm": 0.08980648964643478, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0037, | |
| "step": 105 | |
| }, | |
| { | |
| "epoch": 0.05975197294250282, | |
| "grad_norm": 0.05727319046854973, | |
| "learning_rate": 4.917330276168208e-05, | |
| "loss": 0.0067, | |
| "step": 106 | |
| }, | |
| { | |
| "epoch": 0.06031567080045096, | |
| "grad_norm": 0.027897421270608902, | |
| "learning_rate": 4.834683153413459e-05, | |
| "loss": 0.0016, | |
| "step": 107 | |
| }, | |
| { | |
| "epoch": 0.060879368658399095, | |
| "grad_norm": 0.02825513668358326, | |
| "learning_rate": 4.7520812266338885e-05, | |
| "loss": 0.0014, | |
| "step": 108 | |
| }, | |
| { | |
| "epoch": 0.06144306651634724, | |
| "grad_norm": 0.06746522337198257, | |
| "learning_rate": 4.669547078371504e-05, | |
| "loss": 0.0107, | |
| "step": 109 | |
| }, | |
| { | |
| "epoch": 0.06200676437429538, | |
| "grad_norm": 0.25768861174583435, | |
| "learning_rate": 4.5871032726383386e-05, | |
| "loss": 0.0013, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.06257046223224352, | |
| "grad_norm": 0.05677591636776924, | |
| "learning_rate": 4.504772348747687e-05, | |
| "loss": 0.001, | |
| "step": 111 | |
| }, | |
| { | |
| "epoch": 0.06313416009019165, | |
| "grad_norm": 0.042653512209653854, | |
| "learning_rate": 4.4225768151520694e-05, | |
| "loss": 0.0048, | |
| "step": 112 | |
| }, | |
| { | |
| "epoch": 0.0636978579481398, | |
| "grad_norm": 0.06286019831895828, | |
| "learning_rate": 4.3405391432896555e-05, | |
| "loss": 0.0009, | |
| "step": 113 | |
| }, | |
| { | |
| "epoch": 0.06426155580608793, | |
| "grad_norm": 0.1722741276025772, | |
| "learning_rate": 4.2586817614407895e-05, | |
| "loss": 0.0024, | |
| "step": 114 | |
| }, | |
| { | |
| "epoch": 0.06482525366403608, | |
| "grad_norm": 0.06494489312171936, | |
| "learning_rate": 4.17702704859633e-05, | |
| "loss": 0.0012, | |
| "step": 115 | |
| }, | |
| { | |
| "epoch": 0.06538895152198422, | |
| "grad_norm": 0.06689569354057312, | |
| "learning_rate": 4.095597328339452e-05, | |
| "loss": 0.007, | |
| "step": 116 | |
| }, | |
| { | |
| "epoch": 0.06595264937993235, | |
| "grad_norm": 0.071621835231781, | |
| "learning_rate": 4.0144148627425993e-05, | |
| "loss": 0.0038, | |
| "step": 117 | |
| }, | |
| { | |
| "epoch": 0.0665163472378805, | |
| "grad_norm": 0.024242199957370758, | |
| "learning_rate": 3.933501846281267e-05, | |
| "loss": 0.0009, | |
| "step": 118 | |
| }, | |
| { | |
| "epoch": 0.06708004509582864, | |
| "grad_norm": 0.06986844539642334, | |
| "learning_rate": 3.852880399766243e-05, | |
| "loss": 0.0083, | |
| "step": 119 | |
| }, | |
| { | |
| "epoch": 0.06764374295377677, | |
| "grad_norm": 0.019101617857813835, | |
| "learning_rate": 3.772572564296005e-05, | |
| "loss": 0.0009, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.06820744081172492, | |
| "grad_norm": 0.018207982182502747, | |
| "learning_rate": 3.6926002952309016e-05, | |
| "loss": 0.0007, | |
| "step": 121 | |
| }, | |
| { | |
| "epoch": 0.06877113866967305, | |
| "grad_norm": 0.05150972306728363, | |
| "learning_rate": 3.612985456190778e-05, | |
| "loss": 0.0056, | |
| "step": 122 | |
| }, | |
| { | |
| "epoch": 0.0693348365276212, | |
| "grad_norm": 0.029793381690979004, | |
| "learning_rate": 3.533749813077677e-05, | |
| "loss": 0.0013, | |
| "step": 123 | |
| }, | |
| { | |
| "epoch": 0.06989853438556934, | |
| "grad_norm": 0.10702550411224365, | |
| "learning_rate": 3.4549150281252636e-05, | |
| "loss": 0.0074, | |
| "step": 124 | |
| }, | |
| { | |
| "epoch": 0.07046223224351747, | |
| "grad_norm": 0.0184192918241024, | |
| "learning_rate": 3.3765026539765834e-05, | |
| "loss": 0.0007, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 0.07102593010146561, | |
| "grad_norm": 0.03218894079327583, | |
| "learning_rate": 3.298534127791785e-05, | |
| "loss": 0.0006, | |
| "step": 126 | |
| }, | |
| { | |
| "epoch": 0.07158962795941376, | |
| "grad_norm": 0.09245433658361435, | |
| "learning_rate": 3.221030765387417e-05, | |
| "loss": 0.0008, | |
| "step": 127 | |
| }, | |
| { | |
| "epoch": 0.07215332581736189, | |
| "grad_norm": 0.02014056034386158, | |
| "learning_rate": 3.144013755408895e-05, | |
| "loss": 0.0007, | |
| "step": 128 | |
| }, | |
| { | |
| "epoch": 0.07271702367531004, | |
| "grad_norm": 0.024824917316436768, | |
| "learning_rate": 3.0675041535377405e-05, | |
| "loss": 0.0007, | |
| "step": 129 | |
| }, | |
| { | |
| "epoch": 0.07328072153325817, | |
| "grad_norm": 0.04269944503903389, | |
| "learning_rate": 2.991522876735154e-05, | |
| "loss": 0.0008, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.07384441939120631, | |
| "grad_norm": 0.06853633373975754, | |
| "learning_rate": 2.916090697523549e-05, | |
| "loss": 0.0075, | |
| "step": 131 | |
| }, | |
| { | |
| "epoch": 0.07440811724915446, | |
| "grad_norm": 0.5998584628105164, | |
| "learning_rate": 2.8412282383075363e-05, | |
| "loss": 0.0308, | |
| "step": 132 | |
| }, | |
| { | |
| "epoch": 0.07497181510710259, | |
| "grad_norm": 0.5549770593643188, | |
| "learning_rate": 2.766955965735968e-05, | |
| "loss": 0.0783, | |
| "step": 133 | |
| }, | |
| { | |
| "epoch": 0.07553551296505073, | |
| "grad_norm": 0.5614036917686462, | |
| "learning_rate": 2.693294185106562e-05, | |
| "loss": 0.1189, | |
| "step": 134 | |
| }, | |
| { | |
| "epoch": 0.07609921082299888, | |
| "grad_norm": 0.48949959874153137, | |
| "learning_rate": 2.6202630348146324e-05, | |
| "loss": 0.0795, | |
| "step": 135 | |
| }, | |
| { | |
| "epoch": 0.07666290868094701, | |
| "grad_norm": 0.46451514959335327, | |
| "learning_rate": 2.547882480847461e-05, | |
| "loss": 0.0878, | |
| "step": 136 | |
| }, | |
| { | |
| "epoch": 0.07722660653889515, | |
| "grad_norm": 0.4128587543964386, | |
| "learning_rate": 2.476172311325783e-05, | |
| "loss": 0.1536, | |
| "step": 137 | |
| }, | |
| { | |
| "epoch": 0.0777903043968433, | |
| "grad_norm": 0.21919631958007812, | |
| "learning_rate": 2.405152131093926e-05, | |
| "loss": 0.0712, | |
| "step": 138 | |
| }, | |
| { | |
| "epoch": 0.07835400225479143, | |
| "grad_norm": 0.3420136570930481, | |
| "learning_rate": 2.3348413563600325e-05, | |
| "loss": 0.0998, | |
| "step": 139 | |
| }, | |
| { | |
| "epoch": 0.07891770011273957, | |
| "grad_norm": 0.27021074295043945, | |
| "learning_rate": 2.2652592093878666e-05, | |
| "loss": 0.0928, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.0794813979706877, | |
| "grad_norm": 0.2487686276435852, | |
| "learning_rate": 2.196424713241637e-05, | |
| "loss": 0.0973, | |
| "step": 141 | |
| }, | |
| { | |
| "epoch": 0.08004509582863585, | |
| "grad_norm": 0.3087523579597473, | |
| "learning_rate": 2.128356686585282e-05, | |
| "loss": 0.0928, | |
| "step": 142 | |
| }, | |
| { | |
| "epoch": 0.080608793686584, | |
| "grad_norm": 0.38431546092033386, | |
| "learning_rate": 2.061073738537635e-05, | |
| "loss": 0.1188, | |
| "step": 143 | |
| }, | |
| { | |
| "epoch": 0.08117249154453213, | |
| "grad_norm": 0.37703490257263184, | |
| "learning_rate": 1.9945942635848748e-05, | |
| "loss": 0.1107, | |
| "step": 144 | |
| }, | |
| { | |
| "epoch": 0.08173618940248027, | |
| "grad_norm": 0.31309252977371216, | |
| "learning_rate": 1.928936436551661e-05, | |
| "loss": 0.1222, | |
| "step": 145 | |
| }, | |
| { | |
| "epoch": 0.08229988726042842, | |
| "grad_norm": 0.35660338401794434, | |
| "learning_rate": 1.8641182076323148e-05, | |
| "loss": 0.1273, | |
| "step": 146 | |
| }, | |
| { | |
| "epoch": 0.08286358511837655, | |
| "grad_norm": 0.3349156081676483, | |
| "learning_rate": 1.800157297483417e-05, | |
| "loss": 0.1206, | |
| "step": 147 | |
| }, | |
| { | |
| "epoch": 0.0834272829763247, | |
| "grad_norm": 0.31433185935020447, | |
| "learning_rate": 1.7370711923791567e-05, | |
| "loss": 0.0965, | |
| "step": 148 | |
| }, | |
| { | |
| "epoch": 0.08399098083427282, | |
| "grad_norm": 0.6915076375007629, | |
| "learning_rate": 1.6748771394307585e-05, | |
| "loss": 0.2222, | |
| "step": 149 | |
| }, | |
| { | |
| "epoch": 0.08455467869222097, | |
| "grad_norm": 0.6612845063209534, | |
| "learning_rate": 1.6135921418712956e-05, | |
| "loss": 0.2635, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.08455467869222097, | |
| "eval_loss": 0.049091629683971405, | |
| "eval_runtime": 73.8615, | |
| "eval_samples_per_second": 40.454, | |
| "eval_steps_per_second": 10.114, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.08511837655016911, | |
| "grad_norm": 0.23719260096549988, | |
| "learning_rate": 1.553232954407171e-05, | |
| "loss": 0.0205, | |
| "step": 151 | |
| }, | |
| { | |
| "epoch": 0.08568207440811725, | |
| "grad_norm": 0.20940031111240387, | |
| "learning_rate": 1.4938160786375572e-05, | |
| "loss": 0.0042, | |
| "step": 152 | |
| }, | |
| { | |
| "epoch": 0.08624577226606539, | |
| "grad_norm": 0.16817377507686615, | |
| "learning_rate": 1.435357758543015e-05, | |
| "loss": 0.0051, | |
| "step": 153 | |
| }, | |
| { | |
| "epoch": 0.08680947012401354, | |
| "grad_norm": 0.11934295296669006, | |
| "learning_rate": 1.3778739760445552e-05, | |
| "loss": 0.0025, | |
| "step": 154 | |
| }, | |
| { | |
| "epoch": 0.08737316798196167, | |
| "grad_norm": 0.11391474306583405, | |
| "learning_rate": 1.3213804466343421e-05, | |
| "loss": 0.0023, | |
| "step": 155 | |
| }, | |
| { | |
| "epoch": 0.08793686583990981, | |
| "grad_norm": 0.09877330809831619, | |
| "learning_rate": 1.2658926150792322e-05, | |
| "loss": 0.0023, | |
| "step": 156 | |
| }, | |
| { | |
| "epoch": 0.08850056369785794, | |
| "grad_norm": 0.053761038929224014, | |
| "learning_rate": 1.2114256511983274e-05, | |
| "loss": 0.0015, | |
| "step": 157 | |
| }, | |
| { | |
| "epoch": 0.08906426155580609, | |
| "grad_norm": 0.077743299305439, | |
| "learning_rate": 1.157994445715706e-05, | |
| "loss": 0.0053, | |
| "step": 158 | |
| }, | |
| { | |
| "epoch": 0.08962795941375423, | |
| "grad_norm": 0.038567692041397095, | |
| "learning_rate": 1.1056136061894384e-05, | |
| "loss": 0.0013, | |
| "step": 159 | |
| }, | |
| { | |
| "epoch": 0.09019165727170236, | |
| "grad_norm": 0.03246939554810524, | |
| "learning_rate": 1.0542974530180327e-05, | |
| "loss": 0.0012, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.09075535512965051, | |
| "grad_norm": 0.04957727715373039, | |
| "learning_rate": 1.0040600155253765e-05, | |
| "loss": 0.0027, | |
| "step": 161 | |
| }, | |
| { | |
| "epoch": 0.09131905298759865, | |
| "grad_norm": 0.05332395061850548, | |
| "learning_rate": 9.549150281252633e-06, | |
| "loss": 0.0054, | |
| "step": 162 | |
| }, | |
| { | |
| "epoch": 0.09188275084554678, | |
| "grad_norm": 0.020324481651186943, | |
| "learning_rate": 9.068759265665384e-06, | |
| "loss": 0.0008, | |
| "step": 163 | |
| }, | |
| { | |
| "epoch": 0.09244644870349493, | |
| "grad_norm": 0.02156563103199005, | |
| "learning_rate": 8.599558442598998e-06, | |
| "loss": 0.0009, | |
| "step": 164 | |
| }, | |
| { | |
| "epoch": 0.09301014656144306, | |
| "grad_norm": 0.09189619868993759, | |
| "learning_rate": 8.141676086873572e-06, | |
| "loss": 0.0018, | |
| "step": 165 | |
| }, | |
| { | |
| "epoch": 0.0935738444193912, | |
| "grad_norm": 0.022914238274097443, | |
| "learning_rate": 7.695237378953223e-06, | |
| "loss": 0.0009, | |
| "step": 166 | |
| }, | |
| { | |
| "epoch": 0.09413754227733935, | |
| "grad_norm": 0.02189827896654606, | |
| "learning_rate": 7.260364370723044e-06, | |
| "loss": 0.0008, | |
| "step": 167 | |
| }, | |
| { | |
| "epoch": 0.09470124013528748, | |
| "grad_norm": 0.020868653431534767, | |
| "learning_rate": 6.837175952121306e-06, | |
| "loss": 0.0008, | |
| "step": 168 | |
| }, | |
| { | |
| "epoch": 0.09526493799323563, | |
| "grad_norm": 0.03487245365977287, | |
| "learning_rate": 6.425787818636131e-06, | |
| "loss": 0.0008, | |
| "step": 169 | |
| }, | |
| { | |
| "epoch": 0.09582863585118377, | |
| "grad_norm": 0.017460862174630165, | |
| "learning_rate": 6.026312439675552e-06, | |
| "loss": 0.0008, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.0963923337091319, | |
| "grad_norm": 0.019800834357738495, | |
| "learning_rate": 5.6388590278194096e-06, | |
| "loss": 0.0008, | |
| "step": 171 | |
| }, | |
| { | |
| "epoch": 0.09695603156708005, | |
| "grad_norm": 0.023113662376999855, | |
| "learning_rate": 5.263533508961827e-06, | |
| "loss": 0.001, | |
| "step": 172 | |
| }, | |
| { | |
| "epoch": 0.09751972942502818, | |
| "grad_norm": 0.04317307099699974, | |
| "learning_rate": 4.900438493352055e-06, | |
| "loss": 0.0011, | |
| "step": 173 | |
| }, | |
| { | |
| "epoch": 0.09808342728297632, | |
| "grad_norm": 0.05742291361093521, | |
| "learning_rate": 4.549673247541875e-06, | |
| "loss": 0.0012, | |
| "step": 174 | |
| }, | |
| { | |
| "epoch": 0.09864712514092447, | |
| "grad_norm": 0.026399783790111542, | |
| "learning_rate": 4.2113336672471245e-06, | |
| "loss": 0.001, | |
| "step": 175 | |
| }, | |
| { | |
| "epoch": 0.0992108229988726, | |
| "grad_norm": 0.10123752802610397, | |
| "learning_rate": 3.885512251130763e-06, | |
| "loss": 0.0084, | |
| "step": 176 | |
| }, | |
| { | |
| "epoch": 0.09977452085682074, | |
| "grad_norm": 0.0772150307893753, | |
| "learning_rate": 3.5722980755146517e-06, | |
| "loss": 0.0012, | |
| "step": 177 | |
| }, | |
| { | |
| "epoch": 0.10033821871476889, | |
| "grad_norm": 0.08190146833658218, | |
| "learning_rate": 3.271776770026963e-06, | |
| "loss": 0.0084, | |
| "step": 178 | |
| }, | |
| { | |
| "epoch": 0.10090191657271702, | |
| "grad_norm": 0.04198770225048065, | |
| "learning_rate": 2.9840304941919415e-06, | |
| "loss": 0.0013, | |
| "step": 179 | |
| }, | |
| { | |
| "epoch": 0.10146561443066517, | |
| "grad_norm": 0.06625619530677795, | |
| "learning_rate": 2.7091379149682685e-06, | |
| "loss": 0.0027, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.1020293122886133, | |
| "grad_norm": 0.07382658123970032, | |
| "learning_rate": 2.4471741852423237e-06, | |
| "loss": 0.0019, | |
| "step": 181 | |
| }, | |
| { | |
| "epoch": 0.10259301014656144, | |
| "grad_norm": 0.13528990745544434, | |
| "learning_rate": 2.1982109232821178e-06, | |
| "loss": 0.0223, | |
| "step": 182 | |
| }, | |
| { | |
| "epoch": 0.10315670800450959, | |
| "grad_norm": 0.31559503078460693, | |
| "learning_rate": 1.962316193157593e-06, | |
| "loss": 0.1272, | |
| "step": 183 | |
| }, | |
| { | |
| "epoch": 0.10372040586245772, | |
| "grad_norm": 0.23979054391384125, | |
| "learning_rate": 1.7395544861325718e-06, | |
| "loss": 0.0835, | |
| "step": 184 | |
| }, | |
| { | |
| "epoch": 0.10428410372040586, | |
| "grad_norm": 0.2505544126033783, | |
| "learning_rate": 1.5299867030334814e-06, | |
| "loss": 0.0606, | |
| "step": 185 | |
| }, | |
| { | |
| "epoch": 0.10484780157835401, | |
| "grad_norm": 0.27896395325660706, | |
| "learning_rate": 1.333670137599713e-06, | |
| "loss": 0.1012, | |
| "step": 186 | |
| }, | |
| { | |
| "epoch": 0.10541149943630214, | |
| "grad_norm": 0.1912788599729538, | |
| "learning_rate": 1.1506584608200367e-06, | |
| "loss": 0.0623, | |
| "step": 187 | |
| }, | |
| { | |
| "epoch": 0.10597519729425028, | |
| "grad_norm": 0.2681010663509369, | |
| "learning_rate": 9.810017062595322e-07, | |
| "loss": 0.0804, | |
| "step": 188 | |
| }, | |
| { | |
| "epoch": 0.10653889515219842, | |
| "grad_norm": 0.3429611325263977, | |
| "learning_rate": 8.247462563808817e-07, | |
| "loss": 0.1137, | |
| "step": 189 | |
| }, | |
| { | |
| "epoch": 0.10710259301014656, | |
| "grad_norm": 0.40250933170318604, | |
| "learning_rate": 6.819348298638839e-07, | |
| "loss": 0.1082, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.1076662908680947, | |
| "grad_norm": 0.31438711285591125, | |
| "learning_rate": 5.526064699265753e-07, | |
| "loss": 0.1129, | |
| "step": 191 | |
| }, | |
| { | |
| "epoch": 0.10822998872604284, | |
| "grad_norm": 0.3913666605949402, | |
| "learning_rate": 4.367965336512403e-07, | |
| "loss": 0.1284, | |
| "step": 192 | |
| }, | |
| { | |
| "epoch": 0.10879368658399098, | |
| "grad_norm": 0.251301646232605, | |
| "learning_rate": 3.3453668231809286e-07, | |
| "loss": 0.0878, | |
| "step": 193 | |
| }, | |
| { | |
| "epoch": 0.10935738444193913, | |
| "grad_norm": 0.35125431418418884, | |
| "learning_rate": 2.458548727494292e-07, | |
| "loss": 0.1154, | |
| "step": 194 | |
| }, | |
| { | |
| "epoch": 0.10992108229988726, | |
| "grad_norm": 0.3338298499584198, | |
| "learning_rate": 1.7077534966650766e-07, | |
| "loss": 0.116, | |
| "step": 195 | |
| }, | |
| { | |
| "epoch": 0.1104847801578354, | |
| "grad_norm": 0.373939573764801, | |
| "learning_rate": 1.0931863906127327e-07, | |
| "loss": 0.1167, | |
| "step": 196 | |
| }, | |
| { | |
| "epoch": 0.11104847801578353, | |
| "grad_norm": 0.21560490131378174, | |
| "learning_rate": 6.150154258476315e-08, | |
| "loss": 0.0628, | |
| "step": 197 | |
| }, | |
| { | |
| "epoch": 0.11161217587373168, | |
| "grad_norm": 0.4734821021556854, | |
| "learning_rate": 2.7337132953697554e-08, | |
| "loss": 0.161, | |
| "step": 198 | |
| }, | |
| { | |
| "epoch": 0.11217587373167982, | |
| "grad_norm": 0.5038131475448608, | |
| "learning_rate": 6.834750376549792e-09, | |
| "loss": 0.1652, | |
| "step": 199 | |
| }, | |
| { | |
| "epoch": 0.11273957158962795, | |
| "grad_norm": 0.7520269751548767, | |
| "learning_rate": 0.0, | |
| "loss": 0.2721, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.11273957158962795, | |
| "eval_loss": 0.043948203325271606, | |
| "eval_runtime": 73.8627, | |
| "eval_samples_per_second": 40.453, | |
| "eval_steps_per_second": 10.113, | |
| "step": 200 | |
| } | |
| ], | |
| "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": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 5.204880760032461e+16, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |