| { | |
| "best_metric": 0.7937383055686951, | |
| "best_model_checkpoint": "miner_id_24/checkpoint-200", | |
| "epoch": 0.9900990099009901, | |
| "eval_steps": 50, | |
| "global_step": 200, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0049504950495049506, | |
| "grad_norm": 1.3482602834701538, | |
| "learning_rate": 1e-05, | |
| "loss": 2.0919, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.0049504950495049506, | |
| "eval_loss": 2.174682140350342, | |
| "eval_runtime": 41.0873, | |
| "eval_samples_per_second": 8.275, | |
| "eval_steps_per_second": 2.069, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.009900990099009901, | |
| "grad_norm": 1.1544772386550903, | |
| "learning_rate": 2e-05, | |
| "loss": 2.1907, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.01485148514851485, | |
| "grad_norm": 1.0450679063796997, | |
| "learning_rate": 3e-05, | |
| "loss": 2.0306, | |
| "step": 3 | |
| }, | |
| { | |
| "epoch": 0.019801980198019802, | |
| "grad_norm": 1.0945274829864502, | |
| "learning_rate": 4e-05, | |
| "loss": 2.1605, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.024752475247524754, | |
| "grad_norm": 0.9220031499862671, | |
| "learning_rate": 5e-05, | |
| "loss": 2.0776, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.0297029702970297, | |
| "grad_norm": 0.6884141564369202, | |
| "learning_rate": 6e-05, | |
| "loss": 1.9284, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.034653465346534656, | |
| "grad_norm": 0.7314196228981018, | |
| "learning_rate": 7e-05, | |
| "loss": 1.9397, | |
| "step": 7 | |
| }, | |
| { | |
| "epoch": 0.039603960396039604, | |
| "grad_norm": 0.7707923650741577, | |
| "learning_rate": 8e-05, | |
| "loss": 1.8289, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.04455445544554455, | |
| "grad_norm": 0.7816151976585388, | |
| "learning_rate": 9e-05, | |
| "loss": 1.6586, | |
| "step": 9 | |
| }, | |
| { | |
| "epoch": 0.04950495049504951, | |
| "grad_norm": 0.733718991279602, | |
| "learning_rate": 0.0001, | |
| "loss": 1.6422, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.054455445544554455, | |
| "grad_norm": 0.6857354640960693, | |
| "learning_rate": 9.999316524962345e-05, | |
| "loss": 1.5693, | |
| "step": 11 | |
| }, | |
| { | |
| "epoch": 0.0594059405940594, | |
| "grad_norm": 0.5385900735855103, | |
| "learning_rate": 9.997266286704631e-05, | |
| "loss": 1.4269, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 0.06435643564356436, | |
| "grad_norm": 0.5141613483428955, | |
| "learning_rate": 9.993849845741524e-05, | |
| "loss": 1.3645, | |
| "step": 13 | |
| }, | |
| { | |
| "epoch": 0.06930693069306931, | |
| "grad_norm": 0.5355311036109924, | |
| "learning_rate": 9.989068136093873e-05, | |
| "loss": 1.3309, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 0.07425742574257425, | |
| "grad_norm": 0.607257068157196, | |
| "learning_rate": 9.98292246503335e-05, | |
| "loss": 1.3879, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.07920792079207921, | |
| "grad_norm": 0.6731489300727844, | |
| "learning_rate": 9.975414512725057e-05, | |
| "loss": 1.3898, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 0.08415841584158416, | |
| "grad_norm": 0.5959958434104919, | |
| "learning_rate": 9.966546331768191e-05, | |
| "loss": 1.3101, | |
| "step": 17 | |
| }, | |
| { | |
| "epoch": 0.0891089108910891, | |
| "grad_norm": 0.4856277108192444, | |
| "learning_rate": 9.956320346634876e-05, | |
| "loss": 1.2354, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 0.09405940594059406, | |
| "grad_norm": 0.4400613307952881, | |
| "learning_rate": 9.944739353007344e-05, | |
| "loss": 1.1536, | |
| "step": 19 | |
| }, | |
| { | |
| "epoch": 0.09900990099009901, | |
| "grad_norm": 0.4226345121860504, | |
| "learning_rate": 9.931806517013612e-05, | |
| "loss": 1.1221, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.10396039603960396, | |
| "grad_norm": 0.4286828637123108, | |
| "learning_rate": 9.917525374361912e-05, | |
| "loss": 1.0921, | |
| "step": 21 | |
| }, | |
| { | |
| "epoch": 0.10891089108910891, | |
| "grad_norm": 0.49397462606430054, | |
| "learning_rate": 9.901899829374047e-05, | |
| "loss": 1.1479, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 0.11386138613861387, | |
| "grad_norm": 0.43899601697921753, | |
| "learning_rate": 9.884934153917997e-05, | |
| "loss": 1.1172, | |
| "step": 23 | |
| }, | |
| { | |
| "epoch": 0.1188118811881188, | |
| "grad_norm": 0.4055542051792145, | |
| "learning_rate": 9.86663298624003e-05, | |
| "loss": 1.0598, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 0.12376237623762376, | |
| "grad_norm": 0.4211791157722473, | |
| "learning_rate": 9.847001329696653e-05, | |
| "loss": 1.0439, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.12871287128712872, | |
| "grad_norm": 0.3768998086452484, | |
| "learning_rate": 9.826044551386744e-05, | |
| "loss": 1.0379, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 0.13366336633663367, | |
| "grad_norm": 0.39160606265068054, | |
| "learning_rate": 9.803768380684242e-05, | |
| "loss": 1.0422, | |
| "step": 27 | |
| }, | |
| { | |
| "epoch": 0.13861386138613863, | |
| "grad_norm": 0.3590935170650482, | |
| "learning_rate": 9.780178907671789e-05, | |
| "loss": 1.0572, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 0.14356435643564355, | |
| "grad_norm": 0.32565054297447205, | |
| "learning_rate": 9.755282581475769e-05, | |
| "loss": 0.9498, | |
| "step": 29 | |
| }, | |
| { | |
| "epoch": 0.1485148514851485, | |
| "grad_norm": 0.34032031893730164, | |
| "learning_rate": 9.729086208503174e-05, | |
| "loss": 1.0356, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.15346534653465346, | |
| "grad_norm": 0.356973797082901, | |
| "learning_rate": 9.701596950580806e-05, | |
| "loss": 0.9838, | |
| "step": 31 | |
| }, | |
| { | |
| "epoch": 0.15841584158415842, | |
| "grad_norm": 0.36714276671409607, | |
| "learning_rate": 9.672822322997305e-05, | |
| "loss": 1.0421, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 0.16336633663366337, | |
| "grad_norm": 0.34872764348983765, | |
| "learning_rate": 9.642770192448536e-05, | |
| "loss": 0.9503, | |
| "step": 33 | |
| }, | |
| { | |
| "epoch": 0.16831683168316833, | |
| "grad_norm": 0.3344508111476898, | |
| "learning_rate": 9.611448774886924e-05, | |
| "loss": 0.948, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 0.17326732673267325, | |
| "grad_norm": 0.3760474622249603, | |
| "learning_rate": 9.578866633275288e-05, | |
| "loss": 1.0017, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.1782178217821782, | |
| "grad_norm": 0.36311283707618713, | |
| "learning_rate": 9.545032675245813e-05, | |
| "loss": 0.9798, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 0.18316831683168316, | |
| "grad_norm": 0.3524226248264313, | |
| "learning_rate": 9.509956150664796e-05, | |
| "loss": 0.8974, | |
| "step": 37 | |
| }, | |
| { | |
| "epoch": 0.18811881188118812, | |
| "grad_norm": 0.3375001847743988, | |
| "learning_rate": 9.473646649103818e-05, | |
| "loss": 0.9696, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 0.19306930693069307, | |
| "grad_norm": 0.3268686532974243, | |
| "learning_rate": 9.43611409721806e-05, | |
| "loss": 0.9576, | |
| "step": 39 | |
| }, | |
| { | |
| "epoch": 0.19801980198019803, | |
| "grad_norm": 0.3490821123123169, | |
| "learning_rate": 9.397368756032445e-05, | |
| "loss": 0.9445, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.20297029702970298, | |
| "grad_norm": 0.37281060218811035, | |
| "learning_rate": 9.357421218136386e-05, | |
| "loss": 1.0131, | |
| "step": 41 | |
| }, | |
| { | |
| "epoch": 0.2079207920792079, | |
| "grad_norm": 0.3628464937210083, | |
| "learning_rate": 9.316282404787871e-05, | |
| "loss": 0.9171, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 0.21287128712871287, | |
| "grad_norm": 0.31578153371810913, | |
| "learning_rate": 9.273963562927695e-05, | |
| "loss": 0.9223, | |
| "step": 43 | |
| }, | |
| { | |
| "epoch": 0.21782178217821782, | |
| "grad_norm": 0.3093617558479309, | |
| "learning_rate": 9.230476262104677e-05, | |
| "loss": 0.9156, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 0.22277227722772278, | |
| "grad_norm": 0.33633360266685486, | |
| "learning_rate": 9.185832391312644e-05, | |
| "loss": 0.9754, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 0.22772277227722773, | |
| "grad_norm": 0.3421073257923126, | |
| "learning_rate": 9.140044155740101e-05, | |
| "loss": 0.916, | |
| "step": 46 | |
| }, | |
| { | |
| "epoch": 0.23267326732673269, | |
| "grad_norm": 0.35317444801330566, | |
| "learning_rate": 9.093124073433463e-05, | |
| "loss": 0.8953, | |
| "step": 47 | |
| }, | |
| { | |
| "epoch": 0.2376237623762376, | |
| "grad_norm": 0.3569601774215698, | |
| "learning_rate": 9.045084971874738e-05, | |
| "loss": 1.0049, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 0.24257425742574257, | |
| "grad_norm": 0.35890108346939087, | |
| "learning_rate": 8.995939984474624e-05, | |
| "loss": 0.884, | |
| "step": 49 | |
| }, | |
| { | |
| "epoch": 0.24752475247524752, | |
| "grad_norm": 0.3905589282512665, | |
| "learning_rate": 8.945702546981969e-05, | |
| "loss": 0.9269, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.24752475247524752, | |
| "eval_loss": 0.9261605739593506, | |
| "eval_runtime": 41.9229, | |
| "eval_samples_per_second": 8.11, | |
| "eval_steps_per_second": 2.028, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.2524752475247525, | |
| "grad_norm": 0.37633201479911804, | |
| "learning_rate": 8.894386393810563e-05, | |
| "loss": 0.9273, | |
| "step": 51 | |
| }, | |
| { | |
| "epoch": 0.25742574257425743, | |
| "grad_norm": 0.38889002799987793, | |
| "learning_rate": 8.842005554284296e-05, | |
| "loss": 0.8986, | |
| "step": 52 | |
| }, | |
| { | |
| "epoch": 0.2623762376237624, | |
| "grad_norm": 0.35975396633148193, | |
| "learning_rate": 8.788574348801675e-05, | |
| "loss": 0.9082, | |
| "step": 53 | |
| }, | |
| { | |
| "epoch": 0.26732673267326734, | |
| "grad_norm": 0.3318845331668854, | |
| "learning_rate": 8.73410738492077e-05, | |
| "loss": 0.8967, | |
| "step": 54 | |
| }, | |
| { | |
| "epoch": 0.2722772277227723, | |
| "grad_norm": 0.32496652007102966, | |
| "learning_rate": 8.678619553365659e-05, | |
| "loss": 0.9013, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 0.27722772277227725, | |
| "grad_norm": 0.33157628774642944, | |
| "learning_rate": 8.622126023955446e-05, | |
| "loss": 0.8748, | |
| "step": 56 | |
| }, | |
| { | |
| "epoch": 0.28217821782178215, | |
| "grad_norm": 0.3596450090408325, | |
| "learning_rate": 8.564642241456986e-05, | |
| "loss": 0.917, | |
| "step": 57 | |
| }, | |
| { | |
| "epoch": 0.2871287128712871, | |
| "grad_norm": 0.38354939222335815, | |
| "learning_rate": 8.506183921362443e-05, | |
| "loss": 0.9135, | |
| "step": 58 | |
| }, | |
| { | |
| "epoch": 0.29207920792079206, | |
| "grad_norm": 0.33887022733688354, | |
| "learning_rate": 8.44676704559283e-05, | |
| "loss": 0.9312, | |
| "step": 59 | |
| }, | |
| { | |
| "epoch": 0.297029702970297, | |
| "grad_norm": 0.35705676674842834, | |
| "learning_rate": 8.386407858128706e-05, | |
| "loss": 0.9299, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.30198019801980197, | |
| "grad_norm": 0.3498329818248749, | |
| "learning_rate": 8.32512286056924e-05, | |
| "loss": 0.9354, | |
| "step": 61 | |
| }, | |
| { | |
| "epoch": 0.3069306930693069, | |
| "grad_norm": 0.33207622170448303, | |
| "learning_rate": 8.262928807620843e-05, | |
| "loss": 0.9581, | |
| "step": 62 | |
| }, | |
| { | |
| "epoch": 0.3118811881188119, | |
| "grad_norm": 0.3359692394733429, | |
| "learning_rate": 8.199842702516583e-05, | |
| "loss": 0.9013, | |
| "step": 63 | |
| }, | |
| { | |
| "epoch": 0.31683168316831684, | |
| "grad_norm": 0.39063677191734314, | |
| "learning_rate": 8.135881792367686e-05, | |
| "loss": 0.8932, | |
| "step": 64 | |
| }, | |
| { | |
| "epoch": 0.3217821782178218, | |
| "grad_norm": 0.38509872555732727, | |
| "learning_rate": 8.07106356344834e-05, | |
| "loss": 0.9267, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 0.32673267326732675, | |
| "grad_norm": 0.33846649527549744, | |
| "learning_rate": 8.005405736415126e-05, | |
| "loss": 0.9079, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 0.3316831683168317, | |
| "grad_norm": 0.3691401779651642, | |
| "learning_rate": 7.938926261462366e-05, | |
| "loss": 0.9205, | |
| "step": 67 | |
| }, | |
| { | |
| "epoch": 0.33663366336633666, | |
| "grad_norm": 0.35738834738731384, | |
| "learning_rate": 7.871643313414718e-05, | |
| "loss": 0.8643, | |
| "step": 68 | |
| }, | |
| { | |
| "epoch": 0.3415841584158416, | |
| "grad_norm": 0.37603896856307983, | |
| "learning_rate": 7.803575286758364e-05, | |
| "loss": 0.8947, | |
| "step": 69 | |
| }, | |
| { | |
| "epoch": 0.3465346534653465, | |
| "grad_norm": 0.35490500926971436, | |
| "learning_rate": 7.734740790612136e-05, | |
| "loss": 0.8817, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.35148514851485146, | |
| "grad_norm": 0.36376264691352844, | |
| "learning_rate": 7.66515864363997e-05, | |
| "loss": 0.9648, | |
| "step": 71 | |
| }, | |
| { | |
| "epoch": 0.3564356435643564, | |
| "grad_norm": 0.39898744225502014, | |
| "learning_rate": 7.594847868906076e-05, | |
| "loss": 0.8762, | |
| "step": 72 | |
| }, | |
| { | |
| "epoch": 0.3613861386138614, | |
| "grad_norm": 0.3505392074584961, | |
| "learning_rate": 7.52382768867422e-05, | |
| "loss": 0.8604, | |
| "step": 73 | |
| }, | |
| { | |
| "epoch": 0.36633663366336633, | |
| "grad_norm": 0.37836551666259766, | |
| "learning_rate": 7.452117519152542e-05, | |
| "loss": 0.8567, | |
| "step": 74 | |
| }, | |
| { | |
| "epoch": 0.3712871287128713, | |
| "grad_norm": 0.3366348147392273, | |
| "learning_rate": 7.379736965185368e-05, | |
| "loss": 0.8721, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.37623762376237624, | |
| "grad_norm": 0.3340618312358856, | |
| "learning_rate": 7.30670581489344e-05, | |
| "loss": 0.8489, | |
| "step": 76 | |
| }, | |
| { | |
| "epoch": 0.3811881188118812, | |
| "grad_norm": 0.33875200152397156, | |
| "learning_rate": 7.233044034264034e-05, | |
| "loss": 0.8472, | |
| "step": 77 | |
| }, | |
| { | |
| "epoch": 0.38613861386138615, | |
| "grad_norm": 0.33830931782722473, | |
| "learning_rate": 7.158771761692464e-05, | |
| "loss": 0.8037, | |
| "step": 78 | |
| }, | |
| { | |
| "epoch": 0.3910891089108911, | |
| "grad_norm": 0.3420986831188202, | |
| "learning_rate": 7.083909302476453e-05, | |
| "loss": 0.866, | |
| "step": 79 | |
| }, | |
| { | |
| "epoch": 0.39603960396039606, | |
| "grad_norm": 0.33411622047424316, | |
| "learning_rate": 7.008477123264848e-05, | |
| "loss": 0.8384, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.400990099009901, | |
| "grad_norm": 0.3657791018486023, | |
| "learning_rate": 6.932495846462261e-05, | |
| "loss": 0.826, | |
| "step": 81 | |
| }, | |
| { | |
| "epoch": 0.40594059405940597, | |
| "grad_norm": 0.35291025042533875, | |
| "learning_rate": 6.855986244591104e-05, | |
| "loss": 0.8474, | |
| "step": 82 | |
| }, | |
| { | |
| "epoch": 0.41089108910891087, | |
| "grad_norm": 0.3487788438796997, | |
| "learning_rate": 6.778969234612584e-05, | |
| "loss": 0.8319, | |
| "step": 83 | |
| }, | |
| { | |
| "epoch": 0.4158415841584158, | |
| "grad_norm": 0.3611389398574829, | |
| "learning_rate": 6.701465872208216e-05, | |
| "loss": 0.8758, | |
| "step": 84 | |
| }, | |
| { | |
| "epoch": 0.4207920792079208, | |
| "grad_norm": 0.33499643206596375, | |
| "learning_rate": 6.623497346023418e-05, | |
| "loss": 0.8105, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 0.42574257425742573, | |
| "grad_norm": 0.34754011034965515, | |
| "learning_rate": 6.545084971874738e-05, | |
| "loss": 0.8252, | |
| "step": 86 | |
| }, | |
| { | |
| "epoch": 0.4306930693069307, | |
| "grad_norm": 0.3610055148601532, | |
| "learning_rate": 6.466250186922325e-05, | |
| "loss": 0.816, | |
| "step": 87 | |
| }, | |
| { | |
| "epoch": 0.43564356435643564, | |
| "grad_norm": 0.37026718258857727, | |
| "learning_rate": 6.387014543809223e-05, | |
| "loss": 0.829, | |
| "step": 88 | |
| }, | |
| { | |
| "epoch": 0.4405940594059406, | |
| "grad_norm": 0.3457464873790741, | |
| "learning_rate": 6.307399704769099e-05, | |
| "loss": 0.7991, | |
| "step": 89 | |
| }, | |
| { | |
| "epoch": 0.44554455445544555, | |
| "grad_norm": 0.3476656675338745, | |
| "learning_rate": 6.227427435703997e-05, | |
| "loss": 0.7879, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.4504950495049505, | |
| "grad_norm": 0.3573887348175049, | |
| "learning_rate": 6.147119600233758e-05, | |
| "loss": 0.8041, | |
| "step": 91 | |
| }, | |
| { | |
| "epoch": 0.45544554455445546, | |
| "grad_norm": 0.3491923213005066, | |
| "learning_rate": 6.066498153718735e-05, | |
| "loss": 0.8222, | |
| "step": 92 | |
| }, | |
| { | |
| "epoch": 0.4603960396039604, | |
| "grad_norm": 0.35722678899765015, | |
| "learning_rate": 5.985585137257401e-05, | |
| "loss": 0.9178, | |
| "step": 93 | |
| }, | |
| { | |
| "epoch": 0.46534653465346537, | |
| "grad_norm": 0.3609083592891693, | |
| "learning_rate": 5.90440267166055e-05, | |
| "loss": 0.8064, | |
| "step": 94 | |
| }, | |
| { | |
| "epoch": 0.47029702970297027, | |
| "grad_norm": 0.3368019461631775, | |
| "learning_rate": 5.8229729514036705e-05, | |
| "loss": 0.8459, | |
| "step": 95 | |
| }, | |
| { | |
| "epoch": 0.4752475247524752, | |
| "grad_norm": 0.3801240622997284, | |
| "learning_rate": 5.74131823855921e-05, | |
| "loss": 0.8681, | |
| "step": 96 | |
| }, | |
| { | |
| "epoch": 0.4801980198019802, | |
| "grad_norm": 0.3618883788585663, | |
| "learning_rate": 5.6594608567103456e-05, | |
| "loss": 0.8196, | |
| "step": 97 | |
| }, | |
| { | |
| "epoch": 0.48514851485148514, | |
| "grad_norm": 0.3592188060283661, | |
| "learning_rate": 5.577423184847932e-05, | |
| "loss": 0.8456, | |
| "step": 98 | |
| }, | |
| { | |
| "epoch": 0.4900990099009901, | |
| "grad_norm": 0.35489270091056824, | |
| "learning_rate": 5.495227651252315e-05, | |
| "loss": 0.8333, | |
| "step": 99 | |
| }, | |
| { | |
| "epoch": 0.49504950495049505, | |
| "grad_norm": 0.3952294886112213, | |
| "learning_rate": 5.4128967273616625e-05, | |
| "loss": 0.8508, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.49504950495049505, | |
| "eval_loss": 0.8315855860710144, | |
| "eval_runtime": 42.2313, | |
| "eval_samples_per_second": 8.051, | |
| "eval_steps_per_second": 2.013, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "grad_norm": 0.37665534019470215, | |
| "learning_rate": 5.330452921628497e-05, | |
| "loss": 0.761, | |
| "step": 101 | |
| }, | |
| { | |
| "epoch": 0.504950495049505, | |
| "grad_norm": 0.3574286699295044, | |
| "learning_rate": 5.247918773366112e-05, | |
| "loss": 0.8327, | |
| "step": 102 | |
| }, | |
| { | |
| "epoch": 0.5099009900990099, | |
| "grad_norm": 0.34863030910491943, | |
| "learning_rate": 5.165316846586541e-05, | |
| "loss": 0.8729, | |
| "step": 103 | |
| }, | |
| { | |
| "epoch": 0.5148514851485149, | |
| "grad_norm": 0.382557213306427, | |
| "learning_rate": 5.0826697238317935e-05, | |
| "loss": 0.8704, | |
| "step": 104 | |
| }, | |
| { | |
| "epoch": 0.5198019801980198, | |
| "grad_norm": 0.3428853154182434, | |
| "learning_rate": 5e-05, | |
| "loss": 0.7896, | |
| "step": 105 | |
| }, | |
| { | |
| "epoch": 0.5247524752475248, | |
| "grad_norm": 0.3719635009765625, | |
| "learning_rate": 4.917330276168208e-05, | |
| "loss": 0.8265, | |
| "step": 106 | |
| }, | |
| { | |
| "epoch": 0.5297029702970297, | |
| "grad_norm": 0.359148234128952, | |
| "learning_rate": 4.834683153413459e-05, | |
| "loss": 0.8036, | |
| "step": 107 | |
| }, | |
| { | |
| "epoch": 0.5346534653465347, | |
| "grad_norm": 0.3508284091949463, | |
| "learning_rate": 4.7520812266338885e-05, | |
| "loss": 0.7843, | |
| "step": 108 | |
| }, | |
| { | |
| "epoch": 0.5396039603960396, | |
| "grad_norm": 0.356524258852005, | |
| "learning_rate": 4.669547078371504e-05, | |
| "loss": 0.8615, | |
| "step": 109 | |
| }, | |
| { | |
| "epoch": 0.5445544554455446, | |
| "grad_norm": 0.373365581035614, | |
| "learning_rate": 4.5871032726383386e-05, | |
| "loss": 0.8091, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.5495049504950495, | |
| "grad_norm": 0.3544781804084778, | |
| "learning_rate": 4.504772348747687e-05, | |
| "loss": 0.8245, | |
| "step": 111 | |
| }, | |
| { | |
| "epoch": 0.5544554455445545, | |
| "grad_norm": 0.3394767642021179, | |
| "learning_rate": 4.4225768151520694e-05, | |
| "loss": 0.8169, | |
| "step": 112 | |
| }, | |
| { | |
| "epoch": 0.5594059405940595, | |
| "grad_norm": 0.39101168513298035, | |
| "learning_rate": 4.3405391432896555e-05, | |
| "loss": 0.7819, | |
| "step": 113 | |
| }, | |
| { | |
| "epoch": 0.5643564356435643, | |
| "grad_norm": 0.37920862436294556, | |
| "learning_rate": 4.2586817614407895e-05, | |
| "loss": 0.8312, | |
| "step": 114 | |
| }, | |
| { | |
| "epoch": 0.5693069306930693, | |
| "grad_norm": 0.38053539395332336, | |
| "learning_rate": 4.17702704859633e-05, | |
| "loss": 0.851, | |
| "step": 115 | |
| }, | |
| { | |
| "epoch": 0.5742574257425742, | |
| "grad_norm": 0.3560599684715271, | |
| "learning_rate": 4.095597328339452e-05, | |
| "loss": 0.8321, | |
| "step": 116 | |
| }, | |
| { | |
| "epoch": 0.5792079207920792, | |
| "grad_norm": 0.3801229000091553, | |
| "learning_rate": 4.0144148627425993e-05, | |
| "loss": 0.8111, | |
| "step": 117 | |
| }, | |
| { | |
| "epoch": 0.5841584158415841, | |
| "grad_norm": 0.4061015546321869, | |
| "learning_rate": 3.933501846281267e-05, | |
| "loss": 0.846, | |
| "step": 118 | |
| }, | |
| { | |
| "epoch": 0.5891089108910891, | |
| "grad_norm": 0.3850902318954468, | |
| "learning_rate": 3.852880399766243e-05, | |
| "loss": 0.8355, | |
| "step": 119 | |
| }, | |
| { | |
| "epoch": 0.594059405940594, | |
| "grad_norm": 0.3792211711406708, | |
| "learning_rate": 3.772572564296005e-05, | |
| "loss": 0.8182, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.599009900990099, | |
| "grad_norm": 0.38233286142349243, | |
| "learning_rate": 3.6926002952309016e-05, | |
| "loss": 0.8355, | |
| "step": 121 | |
| }, | |
| { | |
| "epoch": 0.6039603960396039, | |
| "grad_norm": 0.3782663345336914, | |
| "learning_rate": 3.612985456190778e-05, | |
| "loss": 0.8763, | |
| "step": 122 | |
| }, | |
| { | |
| "epoch": 0.6089108910891089, | |
| "grad_norm": 0.3521483540534973, | |
| "learning_rate": 3.533749813077677e-05, | |
| "loss": 0.8117, | |
| "step": 123 | |
| }, | |
| { | |
| "epoch": 0.6138613861386139, | |
| "grad_norm": 0.34629184007644653, | |
| "learning_rate": 3.4549150281252636e-05, | |
| "loss": 0.779, | |
| "step": 124 | |
| }, | |
| { | |
| "epoch": 0.6188118811881188, | |
| "grad_norm": 0.361945778131485, | |
| "learning_rate": 3.3765026539765834e-05, | |
| "loss": 0.7388, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 0.6237623762376238, | |
| "grad_norm": 0.3560042083263397, | |
| "learning_rate": 3.298534127791785e-05, | |
| "loss": 0.8218, | |
| "step": 126 | |
| }, | |
| { | |
| "epoch": 0.6287128712871287, | |
| "grad_norm": 0.3332178294658661, | |
| "learning_rate": 3.221030765387417e-05, | |
| "loss": 0.7668, | |
| "step": 127 | |
| }, | |
| { | |
| "epoch": 0.6336633663366337, | |
| "grad_norm": 0.3437012732028961, | |
| "learning_rate": 3.144013755408895e-05, | |
| "loss": 0.7539, | |
| "step": 128 | |
| }, | |
| { | |
| "epoch": 0.6386138613861386, | |
| "grad_norm": 0.3601408898830414, | |
| "learning_rate": 3.0675041535377405e-05, | |
| "loss": 0.8467, | |
| "step": 129 | |
| }, | |
| { | |
| "epoch": 0.6435643564356436, | |
| "grad_norm": 0.337134450674057, | |
| "learning_rate": 2.991522876735154e-05, | |
| "loss": 0.7823, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.6485148514851485, | |
| "grad_norm": 0.3651469349861145, | |
| "learning_rate": 2.916090697523549e-05, | |
| "loss": 0.8077, | |
| "step": 131 | |
| }, | |
| { | |
| "epoch": 0.6534653465346535, | |
| "grad_norm": 0.3688299357891083, | |
| "learning_rate": 2.8412282383075363e-05, | |
| "loss": 0.8335, | |
| "step": 132 | |
| }, | |
| { | |
| "epoch": 0.6584158415841584, | |
| "grad_norm": 0.3688223958015442, | |
| "learning_rate": 2.766955965735968e-05, | |
| "loss": 0.8309, | |
| "step": 133 | |
| }, | |
| { | |
| "epoch": 0.6633663366336634, | |
| "grad_norm": 0.36464688181877136, | |
| "learning_rate": 2.693294185106562e-05, | |
| "loss": 0.7993, | |
| "step": 134 | |
| }, | |
| { | |
| "epoch": 0.6683168316831684, | |
| "grad_norm": 0.3646702766418457, | |
| "learning_rate": 2.6202630348146324e-05, | |
| "loss": 0.7761, | |
| "step": 135 | |
| }, | |
| { | |
| "epoch": 0.6732673267326733, | |
| "grad_norm": 0.3442142903804779, | |
| "learning_rate": 2.547882480847461e-05, | |
| "loss": 0.7561, | |
| "step": 136 | |
| }, | |
| { | |
| "epoch": 0.6782178217821783, | |
| "grad_norm": 0.3502415120601654, | |
| "learning_rate": 2.476172311325783e-05, | |
| "loss": 0.7802, | |
| "step": 137 | |
| }, | |
| { | |
| "epoch": 0.6831683168316832, | |
| "grad_norm": 0.3568361699581146, | |
| "learning_rate": 2.405152131093926e-05, | |
| "loss": 0.7574, | |
| "step": 138 | |
| }, | |
| { | |
| "epoch": 0.6881188118811881, | |
| "grad_norm": 0.3423681855201721, | |
| "learning_rate": 2.3348413563600325e-05, | |
| "loss": 0.7331, | |
| "step": 139 | |
| }, | |
| { | |
| "epoch": 0.693069306930693, | |
| "grad_norm": 0.3651301860809326, | |
| "learning_rate": 2.2652592093878666e-05, | |
| "loss": 0.7712, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.698019801980198, | |
| "grad_norm": 0.36788809299468994, | |
| "learning_rate": 2.196424713241637e-05, | |
| "loss": 0.8131, | |
| "step": 141 | |
| }, | |
| { | |
| "epoch": 0.7029702970297029, | |
| "grad_norm": 0.3558289408683777, | |
| "learning_rate": 2.128356686585282e-05, | |
| "loss": 0.8181, | |
| "step": 142 | |
| }, | |
| { | |
| "epoch": 0.7079207920792079, | |
| "grad_norm": 0.3655586540699005, | |
| "learning_rate": 2.061073738537635e-05, | |
| "loss": 0.8288, | |
| "step": 143 | |
| }, | |
| { | |
| "epoch": 0.7128712871287128, | |
| "grad_norm": 0.38642531633377075, | |
| "learning_rate": 1.9945942635848748e-05, | |
| "loss": 0.8351, | |
| "step": 144 | |
| }, | |
| { | |
| "epoch": 0.7178217821782178, | |
| "grad_norm": 0.3551604747772217, | |
| "learning_rate": 1.928936436551661e-05, | |
| "loss": 0.7949, | |
| "step": 145 | |
| }, | |
| { | |
| "epoch": 0.7227722772277227, | |
| "grad_norm": 0.3890242278575897, | |
| "learning_rate": 1.8641182076323148e-05, | |
| "loss": 0.8221, | |
| "step": 146 | |
| }, | |
| { | |
| "epoch": 0.7277227722772277, | |
| "grad_norm": 0.3794883191585541, | |
| "learning_rate": 1.800157297483417e-05, | |
| "loss": 0.8685, | |
| "step": 147 | |
| }, | |
| { | |
| "epoch": 0.7326732673267327, | |
| "grad_norm": 0.39160802960395813, | |
| "learning_rate": 1.7370711923791567e-05, | |
| "loss": 0.8189, | |
| "step": 148 | |
| }, | |
| { | |
| "epoch": 0.7376237623762376, | |
| "grad_norm": 0.37227171659469604, | |
| "learning_rate": 1.6748771394307585e-05, | |
| "loss": 0.8325, | |
| "step": 149 | |
| }, | |
| { | |
| "epoch": 0.7425742574257426, | |
| "grad_norm": 0.4169851839542389, | |
| "learning_rate": 1.6135921418712956e-05, | |
| "loss": 0.8412, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.7425742574257426, | |
| "eval_loss": 0.798706591129303, | |
| "eval_runtime": 41.9853, | |
| "eval_samples_per_second": 8.098, | |
| "eval_steps_per_second": 2.025, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.7475247524752475, | |
| "grad_norm": 0.3418070673942566, | |
| "learning_rate": 1.553232954407171e-05, | |
| "loss": 0.7155, | |
| "step": 151 | |
| }, | |
| { | |
| "epoch": 0.7524752475247525, | |
| "grad_norm": 0.3621049225330353, | |
| "learning_rate": 1.4938160786375572e-05, | |
| "loss": 0.7654, | |
| "step": 152 | |
| }, | |
| { | |
| "epoch": 0.7574257425742574, | |
| "grad_norm": 0.3533613085746765, | |
| "learning_rate": 1.435357758543015e-05, | |
| "loss": 0.7326, | |
| "step": 153 | |
| }, | |
| { | |
| "epoch": 0.7623762376237624, | |
| "grad_norm": 0.3653974235057831, | |
| "learning_rate": 1.3778739760445552e-05, | |
| "loss": 0.7865, | |
| "step": 154 | |
| }, | |
| { | |
| "epoch": 0.7673267326732673, | |
| "grad_norm": 0.3443511724472046, | |
| "learning_rate": 1.3213804466343421e-05, | |
| "loss": 0.859, | |
| "step": 155 | |
| }, | |
| { | |
| "epoch": 0.7722772277227723, | |
| "grad_norm": 0.35138487815856934, | |
| "learning_rate": 1.2658926150792322e-05, | |
| "loss": 0.7857, | |
| "step": 156 | |
| }, | |
| { | |
| "epoch": 0.7772277227722773, | |
| "grad_norm": 0.34442415833473206, | |
| "learning_rate": 1.2114256511983274e-05, | |
| "loss": 0.8182, | |
| "step": 157 | |
| }, | |
| { | |
| "epoch": 0.7821782178217822, | |
| "grad_norm": 0.36188143491744995, | |
| "learning_rate": 1.157994445715706e-05, | |
| "loss": 0.7763, | |
| "step": 158 | |
| }, | |
| { | |
| "epoch": 0.7871287128712872, | |
| "grad_norm": 0.34919753670692444, | |
| "learning_rate": 1.1056136061894384e-05, | |
| "loss": 0.817, | |
| "step": 159 | |
| }, | |
| { | |
| "epoch": 0.7920792079207921, | |
| "grad_norm": 0.36045747995376587, | |
| "learning_rate": 1.0542974530180327e-05, | |
| "loss": 0.7894, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.7970297029702971, | |
| "grad_norm": 0.3586371839046478, | |
| "learning_rate": 1.0040600155253765e-05, | |
| "loss": 0.7454, | |
| "step": 161 | |
| }, | |
| { | |
| "epoch": 0.801980198019802, | |
| "grad_norm": 0.3578149974346161, | |
| "learning_rate": 9.549150281252633e-06, | |
| "loss": 0.7828, | |
| "step": 162 | |
| }, | |
| { | |
| "epoch": 0.806930693069307, | |
| "grad_norm": 0.3628920316696167, | |
| "learning_rate": 9.068759265665384e-06, | |
| "loss": 0.8311, | |
| "step": 163 | |
| }, | |
| { | |
| "epoch": 0.8118811881188119, | |
| "grad_norm": 0.3866752088069916, | |
| "learning_rate": 8.599558442598998e-06, | |
| "loss": 0.8611, | |
| "step": 164 | |
| }, | |
| { | |
| "epoch": 0.8168316831683168, | |
| "grad_norm": 0.38015756011009216, | |
| "learning_rate": 8.141676086873572e-06, | |
| "loss": 0.8029, | |
| "step": 165 | |
| }, | |
| { | |
| "epoch": 0.8217821782178217, | |
| "grad_norm": 0.4192558228969574, | |
| "learning_rate": 7.695237378953223e-06, | |
| "loss": 0.8796, | |
| "step": 166 | |
| }, | |
| { | |
| "epoch": 0.8267326732673267, | |
| "grad_norm": 0.3582792282104492, | |
| "learning_rate": 7.260364370723044e-06, | |
| "loss": 0.7878, | |
| "step": 167 | |
| }, | |
| { | |
| "epoch": 0.8316831683168316, | |
| "grad_norm": 0.37938931584358215, | |
| "learning_rate": 6.837175952121306e-06, | |
| "loss": 0.8585, | |
| "step": 168 | |
| }, | |
| { | |
| "epoch": 0.8366336633663366, | |
| "grad_norm": 0.36483219265937805, | |
| "learning_rate": 6.425787818636131e-06, | |
| "loss": 0.7667, | |
| "step": 169 | |
| }, | |
| { | |
| "epoch": 0.8415841584158416, | |
| "grad_norm": 0.41834333539009094, | |
| "learning_rate": 6.026312439675552e-06, | |
| "loss": 0.8407, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.8465346534653465, | |
| "grad_norm": 0.3925648331642151, | |
| "learning_rate": 5.6388590278194096e-06, | |
| "loss": 0.8238, | |
| "step": 171 | |
| }, | |
| { | |
| "epoch": 0.8514851485148515, | |
| "grad_norm": 0.37753769755363464, | |
| "learning_rate": 5.263533508961827e-06, | |
| "loss": 0.8399, | |
| "step": 172 | |
| }, | |
| { | |
| "epoch": 0.8564356435643564, | |
| "grad_norm": 0.3624138832092285, | |
| "learning_rate": 4.900438493352055e-06, | |
| "loss": 0.7573, | |
| "step": 173 | |
| }, | |
| { | |
| "epoch": 0.8613861386138614, | |
| "grad_norm": 0.3399837613105774, | |
| "learning_rate": 4.549673247541875e-06, | |
| "loss": 0.772, | |
| "step": 174 | |
| }, | |
| { | |
| "epoch": 0.8663366336633663, | |
| "grad_norm": 0.3716399669647217, | |
| "learning_rate": 4.2113336672471245e-06, | |
| "loss": 0.7299, | |
| "step": 175 | |
| }, | |
| { | |
| "epoch": 0.8712871287128713, | |
| "grad_norm": 0.34926846623420715, | |
| "learning_rate": 3.885512251130763e-06, | |
| "loss": 0.7431, | |
| "step": 176 | |
| }, | |
| { | |
| "epoch": 0.8762376237623762, | |
| "grad_norm": 0.35869038105010986, | |
| "learning_rate": 3.5722980755146517e-06, | |
| "loss": 0.7522, | |
| "step": 177 | |
| }, | |
| { | |
| "epoch": 0.8811881188118812, | |
| "grad_norm": 0.35853707790374756, | |
| "learning_rate": 3.271776770026963e-06, | |
| "loss": 0.7753, | |
| "step": 178 | |
| }, | |
| { | |
| "epoch": 0.8861386138613861, | |
| "grad_norm": 0.34297704696655273, | |
| "learning_rate": 2.9840304941919415e-06, | |
| "loss": 0.7465, | |
| "step": 179 | |
| }, | |
| { | |
| "epoch": 0.8910891089108911, | |
| "grad_norm": 0.35985806584358215, | |
| "learning_rate": 2.7091379149682685e-06, | |
| "loss": 0.766, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.8960396039603961, | |
| "grad_norm": 0.3661913573741913, | |
| "learning_rate": 2.4471741852423237e-06, | |
| "loss": 0.845, | |
| "step": 181 | |
| }, | |
| { | |
| "epoch": 0.900990099009901, | |
| "grad_norm": 0.34959831833839417, | |
| "learning_rate": 2.1982109232821178e-06, | |
| "loss": 0.8137, | |
| "step": 182 | |
| }, | |
| { | |
| "epoch": 0.905940594059406, | |
| "grad_norm": 0.3623139262199402, | |
| "learning_rate": 1.962316193157593e-06, | |
| "loss": 0.828, | |
| "step": 183 | |
| }, | |
| { | |
| "epoch": 0.9108910891089109, | |
| "grad_norm": 0.357437402009964, | |
| "learning_rate": 1.7395544861325718e-06, | |
| "loss": 0.7631, | |
| "step": 184 | |
| }, | |
| { | |
| "epoch": 0.9158415841584159, | |
| "grad_norm": 0.35356783866882324, | |
| "learning_rate": 1.5299867030334814e-06, | |
| "loss": 0.7941, | |
| "step": 185 | |
| }, | |
| { | |
| "epoch": 0.9207920792079208, | |
| "grad_norm": 0.3628612756729126, | |
| "learning_rate": 1.333670137599713e-06, | |
| "loss": 0.7827, | |
| "step": 186 | |
| }, | |
| { | |
| "epoch": 0.9257425742574258, | |
| "grad_norm": 0.34733375906944275, | |
| "learning_rate": 1.1506584608200367e-06, | |
| "loss": 0.7717, | |
| "step": 187 | |
| }, | |
| { | |
| "epoch": 0.9306930693069307, | |
| "grad_norm": 0.36748260259628296, | |
| "learning_rate": 9.810017062595322e-07, | |
| "loss": 0.7957, | |
| "step": 188 | |
| }, | |
| { | |
| "epoch": 0.9356435643564357, | |
| "grad_norm": 0.3549409806728363, | |
| "learning_rate": 8.247462563808817e-07, | |
| "loss": 0.7806, | |
| "step": 189 | |
| }, | |
| { | |
| "epoch": 0.9405940594059405, | |
| "grad_norm": 0.33620116114616394, | |
| "learning_rate": 6.819348298638839e-07, | |
| "loss": 0.7382, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.9455445544554455, | |
| "grad_norm": 0.34486040472984314, | |
| "learning_rate": 5.526064699265753e-07, | |
| "loss": 0.7609, | |
| "step": 191 | |
| }, | |
| { | |
| "epoch": 0.9504950495049505, | |
| "grad_norm": 0.3615681529045105, | |
| "learning_rate": 4.367965336512403e-07, | |
| "loss": 0.7995, | |
| "step": 192 | |
| }, | |
| { | |
| "epoch": 0.9554455445544554, | |
| "grad_norm": 0.35651570558547974, | |
| "learning_rate": 3.3453668231809286e-07, | |
| "loss": 0.8186, | |
| "step": 193 | |
| }, | |
| { | |
| "epoch": 0.9603960396039604, | |
| "grad_norm": 0.38201525807380676, | |
| "learning_rate": 2.458548727494292e-07, | |
| "loss": 0.8277, | |
| "step": 194 | |
| }, | |
| { | |
| "epoch": 0.9653465346534653, | |
| "grad_norm": 0.39812982082366943, | |
| "learning_rate": 1.7077534966650766e-07, | |
| "loss": 0.8521, | |
| "step": 195 | |
| }, | |
| { | |
| "epoch": 0.9702970297029703, | |
| "grad_norm": 0.39788690209388733, | |
| "learning_rate": 1.0931863906127327e-07, | |
| "loss": 0.8287, | |
| "step": 196 | |
| }, | |
| { | |
| "epoch": 0.9752475247524752, | |
| "grad_norm": 0.3834620416164398, | |
| "learning_rate": 6.150154258476315e-08, | |
| "loss": 0.7986, | |
| "step": 197 | |
| }, | |
| { | |
| "epoch": 0.9801980198019802, | |
| "grad_norm": 0.3714093863964081, | |
| "learning_rate": 2.7337132953697554e-08, | |
| "loss": 0.8404, | |
| "step": 198 | |
| }, | |
| { | |
| "epoch": 0.9851485148514851, | |
| "grad_norm": 0.3981218636035919, | |
| "learning_rate": 6.834750376549792e-09, | |
| "loss": 0.8326, | |
| "step": 199 | |
| }, | |
| { | |
| "epoch": 0.9900990099009901, | |
| "grad_norm": 0.4067413806915283, | |
| "learning_rate": 0.0, | |
| "loss": 0.8359, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.9900990099009901, | |
| "eval_loss": 0.7937383055686951, | |
| "eval_runtime": 42.1655, | |
| "eval_samples_per_second": 8.063, | |
| "eval_steps_per_second": 2.016, | |
| "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": 4.873523469012173e+17, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |