| { | |
| "best_metric": 1.484871745109558, | |
| "best_model_checkpoint": "miner_id_24/checkpoint-150", | |
| "epoch": 0.02405002405002405, | |
| "eval_steps": 50, | |
| "global_step": 150, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.000160333493666827, | |
| "grad_norm": 4.413506507873535, | |
| "learning_rate": 1e-05, | |
| "loss": 1.1508, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.000160333493666827, | |
| "eval_loss": 2.065471649169922, | |
| "eval_runtime": 181.1479, | |
| "eval_samples_per_second": 57.991, | |
| "eval_steps_per_second": 14.502, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.000320666987333654, | |
| "grad_norm": 3.4024477005004883, | |
| "learning_rate": 2e-05, | |
| "loss": 1.5453, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.000481000481000481, | |
| "grad_norm": 3.406569719314575, | |
| "learning_rate": 3e-05, | |
| "loss": 1.4252, | |
| "step": 3 | |
| }, | |
| { | |
| "epoch": 0.000641333974667308, | |
| "grad_norm": 2.1866605281829834, | |
| "learning_rate": 4e-05, | |
| "loss": 1.5935, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.000801667468334135, | |
| "grad_norm": 2.536449909210205, | |
| "learning_rate": 5e-05, | |
| "loss": 1.7653, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.000962000962000962, | |
| "grad_norm": 2.7087090015411377, | |
| "learning_rate": 6e-05, | |
| "loss": 2.0449, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.001122334455667789, | |
| "grad_norm": 2.613861322402954, | |
| "learning_rate": 7e-05, | |
| "loss": 1.7385, | |
| "step": 7 | |
| }, | |
| { | |
| "epoch": 0.001282667949334616, | |
| "grad_norm": 2.6648755073547363, | |
| "learning_rate": 8e-05, | |
| "loss": 1.8911, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.001443001443001443, | |
| "grad_norm": 3.0424225330352783, | |
| "learning_rate": 9e-05, | |
| "loss": 1.9142, | |
| "step": 9 | |
| }, | |
| { | |
| "epoch": 0.00160333493666827, | |
| "grad_norm": 2.2900004386901855, | |
| "learning_rate": 0.0001, | |
| "loss": 1.9279, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.001763668430335097, | |
| "grad_norm": 1.9080548286437988, | |
| "learning_rate": 9.999316524962345e-05, | |
| "loss": 1.9104, | |
| "step": 11 | |
| }, | |
| { | |
| "epoch": 0.001924001924001924, | |
| "grad_norm": 2.1431448459625244, | |
| "learning_rate": 9.997266286704631e-05, | |
| "loss": 1.9847, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 0.002084335417668751, | |
| "grad_norm": 1.7175735235214233, | |
| "learning_rate": 9.993849845741524e-05, | |
| "loss": 1.7866, | |
| "step": 13 | |
| }, | |
| { | |
| "epoch": 0.002244668911335578, | |
| "grad_norm": 3.7480881214141846, | |
| "learning_rate": 9.989068136093873e-05, | |
| "loss": 1.8202, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 0.002405002405002405, | |
| "grad_norm": 1.6957277059555054, | |
| "learning_rate": 9.98292246503335e-05, | |
| "loss": 1.7817, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.002565335898669232, | |
| "grad_norm": 1.7331048250198364, | |
| "learning_rate": 9.975414512725057e-05, | |
| "loss": 1.6535, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 0.002725669392336059, | |
| "grad_norm": 1.8792626857757568, | |
| "learning_rate": 9.966546331768191e-05, | |
| "loss": 1.99, | |
| "step": 17 | |
| }, | |
| { | |
| "epoch": 0.002886002886002886, | |
| "grad_norm": 1.7258965969085693, | |
| "learning_rate": 9.956320346634876e-05, | |
| "loss": 2.0252, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 0.003046336379669713, | |
| "grad_norm": 2.050128936767578, | |
| "learning_rate": 9.944739353007344e-05, | |
| "loss": 1.7259, | |
| "step": 19 | |
| }, | |
| { | |
| "epoch": 0.00320666987333654, | |
| "grad_norm": 1.8676942586898804, | |
| "learning_rate": 9.931806517013612e-05, | |
| "loss": 1.6296, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.003367003367003367, | |
| "grad_norm": 1.8743293285369873, | |
| "learning_rate": 9.917525374361912e-05, | |
| "loss": 1.7827, | |
| "step": 21 | |
| }, | |
| { | |
| "epoch": 0.003527336860670194, | |
| "grad_norm": 1.8751113414764404, | |
| "learning_rate": 9.901899829374047e-05, | |
| "loss": 1.5662, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 0.003687670354337021, | |
| "grad_norm": 1.7388873100280762, | |
| "learning_rate": 9.884934153917997e-05, | |
| "loss": 1.6508, | |
| "step": 23 | |
| }, | |
| { | |
| "epoch": 0.003848003848003848, | |
| "grad_norm": 1.7969709634780884, | |
| "learning_rate": 9.86663298624003e-05, | |
| "loss": 1.9835, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 0.004008337341670675, | |
| "grad_norm": 1.921236515045166, | |
| "learning_rate": 9.847001329696653e-05, | |
| "loss": 1.7869, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.004168670835337502, | |
| "grad_norm": 1.705811619758606, | |
| "learning_rate": 9.826044551386744e-05, | |
| "loss": 1.5369, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 0.004329004329004329, | |
| "grad_norm": 1.798777461051941, | |
| "learning_rate": 9.803768380684242e-05, | |
| "loss": 1.5279, | |
| "step": 27 | |
| }, | |
| { | |
| "epoch": 0.004489337822671156, | |
| "grad_norm": 1.7530027627944946, | |
| "learning_rate": 9.780178907671789e-05, | |
| "loss": 1.735, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 0.004649671316337983, | |
| "grad_norm": 1.7224781513214111, | |
| "learning_rate": 9.755282581475769e-05, | |
| "loss": 1.3065, | |
| "step": 29 | |
| }, | |
| { | |
| "epoch": 0.00481000481000481, | |
| "grad_norm": 1.673201560974121, | |
| "learning_rate": 9.729086208503174e-05, | |
| "loss": 1.3976, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.004970338303671637, | |
| "grad_norm": 1.8575778007507324, | |
| "learning_rate": 9.701596950580806e-05, | |
| "loss": 1.6034, | |
| "step": 31 | |
| }, | |
| { | |
| "epoch": 0.005130671797338464, | |
| "grad_norm": 2.107029914855957, | |
| "learning_rate": 9.672822322997305e-05, | |
| "loss": 1.4381, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 0.005291005291005291, | |
| "grad_norm": 1.966361165046692, | |
| "learning_rate": 9.642770192448536e-05, | |
| "loss": 1.2308, | |
| "step": 33 | |
| }, | |
| { | |
| "epoch": 0.005451338784672118, | |
| "grad_norm": 1.9392153024673462, | |
| "learning_rate": 9.611448774886924e-05, | |
| "loss": 1.4583, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 0.005611672278338945, | |
| "grad_norm": 2.2352333068847656, | |
| "learning_rate": 9.578866633275288e-05, | |
| "loss": 1.291, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.005772005772005772, | |
| "grad_norm": 1.8804792165756226, | |
| "learning_rate": 9.545032675245813e-05, | |
| "loss": 1.1175, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 0.005932339265672599, | |
| "grad_norm": 2.296677589416504, | |
| "learning_rate": 9.509956150664796e-05, | |
| "loss": 1.3237, | |
| "step": 37 | |
| }, | |
| { | |
| "epoch": 0.006092672759339426, | |
| "grad_norm": 1.9979928731918335, | |
| "learning_rate": 9.473646649103818e-05, | |
| "loss": 1.2588, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 0.006253006253006253, | |
| "grad_norm": 2.2553892135620117, | |
| "learning_rate": 9.43611409721806e-05, | |
| "loss": 1.2834, | |
| "step": 39 | |
| }, | |
| { | |
| "epoch": 0.00641333974667308, | |
| "grad_norm": 2.2020952701568604, | |
| "learning_rate": 9.397368756032445e-05, | |
| "loss": 1.3339, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.006573673240339907, | |
| "grad_norm": 2.2363979816436768, | |
| "learning_rate": 9.357421218136386e-05, | |
| "loss": 1.2251, | |
| "step": 41 | |
| }, | |
| { | |
| "epoch": 0.006734006734006734, | |
| "grad_norm": 2.245211362838745, | |
| "learning_rate": 9.316282404787871e-05, | |
| "loss": 1.2634, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 0.006894340227673561, | |
| "grad_norm": 2.6141109466552734, | |
| "learning_rate": 9.273963562927695e-05, | |
| "loss": 1.1247, | |
| "step": 43 | |
| }, | |
| { | |
| "epoch": 0.007054673721340388, | |
| "grad_norm": 3.073989152908325, | |
| "learning_rate": 9.230476262104677e-05, | |
| "loss": 1.0591, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 0.007215007215007215, | |
| "grad_norm": 2.8839738368988037, | |
| "learning_rate": 9.185832391312644e-05, | |
| "loss": 1.3222, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 0.007375340708674042, | |
| "grad_norm": 2.779451370239258, | |
| "learning_rate": 9.140044155740101e-05, | |
| "loss": 1.1081, | |
| "step": 46 | |
| }, | |
| { | |
| "epoch": 0.007535674202340869, | |
| "grad_norm": 3.0295441150665283, | |
| "learning_rate": 9.093124073433463e-05, | |
| "loss": 1.2263, | |
| "step": 47 | |
| }, | |
| { | |
| "epoch": 0.007696007696007696, | |
| "grad_norm": 2.8162171840667725, | |
| "learning_rate": 9.045084971874738e-05, | |
| "loss": 1.0807, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 0.007856341189674524, | |
| "grad_norm": 4.821334362030029, | |
| "learning_rate": 8.995939984474624e-05, | |
| "loss": 1.099, | |
| "step": 49 | |
| }, | |
| { | |
| "epoch": 0.00801667468334135, | |
| "grad_norm": 4.665739059448242, | |
| "learning_rate": 8.945702546981969e-05, | |
| "loss": 1.0781, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.00801667468334135, | |
| "eval_loss": 1.7481515407562256, | |
| "eval_runtime": 181.0644, | |
| "eval_samples_per_second": 58.018, | |
| "eval_steps_per_second": 14.509, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.008177008177008177, | |
| "grad_norm": 4.216922760009766, | |
| "learning_rate": 8.894386393810563e-05, | |
| "loss": 0.9413, | |
| "step": 51 | |
| }, | |
| { | |
| "epoch": 0.008337341670675003, | |
| "grad_norm": 4.562152862548828, | |
| "learning_rate": 8.842005554284296e-05, | |
| "loss": 1.3296, | |
| "step": 52 | |
| }, | |
| { | |
| "epoch": 0.008497675164341831, | |
| "grad_norm": 3.430103302001953, | |
| "learning_rate": 8.788574348801675e-05, | |
| "loss": 1.4943, | |
| "step": 53 | |
| }, | |
| { | |
| "epoch": 0.008658008658008658, | |
| "grad_norm": 3.834789276123047, | |
| "learning_rate": 8.73410738492077e-05, | |
| "loss": 1.6902, | |
| "step": 54 | |
| }, | |
| { | |
| "epoch": 0.008818342151675485, | |
| "grad_norm": 2.8888132572174072, | |
| "learning_rate": 8.678619553365659e-05, | |
| "loss": 1.8122, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 0.008978675645342313, | |
| "grad_norm": 1.801705002784729, | |
| "learning_rate": 8.622126023955446e-05, | |
| "loss": 1.9578, | |
| "step": 56 | |
| }, | |
| { | |
| "epoch": 0.00913900913900914, | |
| "grad_norm": 2.326319932937622, | |
| "learning_rate": 8.564642241456986e-05, | |
| "loss": 1.9303, | |
| "step": 57 | |
| }, | |
| { | |
| "epoch": 0.009299342632675966, | |
| "grad_norm": 2.414191722869873, | |
| "learning_rate": 8.506183921362443e-05, | |
| "loss": 2.0898, | |
| "step": 58 | |
| }, | |
| { | |
| "epoch": 0.009459676126342792, | |
| "grad_norm": 2.116661787033081, | |
| "learning_rate": 8.44676704559283e-05, | |
| "loss": 2.2721, | |
| "step": 59 | |
| }, | |
| { | |
| "epoch": 0.00962000962000962, | |
| "grad_norm": 1.7814158201217651, | |
| "learning_rate": 8.386407858128706e-05, | |
| "loss": 1.7448, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.009780343113676447, | |
| "grad_norm": 1.5620167255401611, | |
| "learning_rate": 8.32512286056924e-05, | |
| "loss": 1.8467, | |
| "step": 61 | |
| }, | |
| { | |
| "epoch": 0.009940676607343274, | |
| "grad_norm": 1.5871108770370483, | |
| "learning_rate": 8.262928807620843e-05, | |
| "loss": 2.0595, | |
| "step": 62 | |
| }, | |
| { | |
| "epoch": 0.010101010101010102, | |
| "grad_norm": 1.481527328491211, | |
| "learning_rate": 8.199842702516583e-05, | |
| "loss": 1.6665, | |
| "step": 63 | |
| }, | |
| { | |
| "epoch": 0.010261343594676928, | |
| "grad_norm": 1.3492636680603027, | |
| "learning_rate": 8.135881792367686e-05, | |
| "loss": 1.7218, | |
| "step": 64 | |
| }, | |
| { | |
| "epoch": 0.010421677088343755, | |
| "grad_norm": 1.5018565654754639, | |
| "learning_rate": 8.07106356344834e-05, | |
| "loss": 1.9853, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 0.010582010582010581, | |
| "grad_norm": 1.436688780784607, | |
| "learning_rate": 8.005405736415126e-05, | |
| "loss": 2.061, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 0.01074234407567741, | |
| "grad_norm": 1.468376874923706, | |
| "learning_rate": 7.938926261462366e-05, | |
| "loss": 1.6864, | |
| "step": 67 | |
| }, | |
| { | |
| "epoch": 0.010902677569344236, | |
| "grad_norm": 1.5045324563980103, | |
| "learning_rate": 7.871643313414718e-05, | |
| "loss": 1.899, | |
| "step": 68 | |
| }, | |
| { | |
| "epoch": 0.011063011063011063, | |
| "grad_norm": 1.4142143726348877, | |
| "learning_rate": 7.803575286758364e-05, | |
| "loss": 2.0851, | |
| "step": 69 | |
| }, | |
| { | |
| "epoch": 0.01122334455667789, | |
| "grad_norm": 1.4256950616836548, | |
| "learning_rate": 7.734740790612136e-05, | |
| "loss": 1.9298, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.011383678050344717, | |
| "grad_norm": 1.4729069471359253, | |
| "learning_rate": 7.66515864363997e-05, | |
| "loss": 1.6926, | |
| "step": 71 | |
| }, | |
| { | |
| "epoch": 0.011544011544011544, | |
| "grad_norm": 1.4505417346954346, | |
| "learning_rate": 7.594847868906076e-05, | |
| "loss": 1.6073, | |
| "step": 72 | |
| }, | |
| { | |
| "epoch": 0.01170434503767837, | |
| "grad_norm": 1.528064489364624, | |
| "learning_rate": 7.52382768867422e-05, | |
| "loss": 1.7171, | |
| "step": 73 | |
| }, | |
| { | |
| "epoch": 0.011864678531345199, | |
| "grad_norm": 1.5659871101379395, | |
| "learning_rate": 7.452117519152542e-05, | |
| "loss": 1.9315, | |
| "step": 74 | |
| }, | |
| { | |
| "epoch": 0.012025012025012025, | |
| "grad_norm": 1.6519936323165894, | |
| "learning_rate": 7.379736965185368e-05, | |
| "loss": 1.9801, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.012185345518678852, | |
| "grad_norm": 1.474597454071045, | |
| "learning_rate": 7.30670581489344e-05, | |
| "loss": 1.5052, | |
| "step": 76 | |
| }, | |
| { | |
| "epoch": 0.012345679012345678, | |
| "grad_norm": 1.4802509546279907, | |
| "learning_rate": 7.233044034264034e-05, | |
| "loss": 1.4324, | |
| "step": 77 | |
| }, | |
| { | |
| "epoch": 0.012506012506012507, | |
| "grad_norm": 1.6556718349456787, | |
| "learning_rate": 7.158771761692464e-05, | |
| "loss": 1.6473, | |
| "step": 78 | |
| }, | |
| { | |
| "epoch": 0.012666345999679333, | |
| "grad_norm": 1.578564167022705, | |
| "learning_rate": 7.083909302476453e-05, | |
| "loss": 1.6985, | |
| "step": 79 | |
| }, | |
| { | |
| "epoch": 0.01282667949334616, | |
| "grad_norm": 1.5099948644638062, | |
| "learning_rate": 7.008477123264848e-05, | |
| "loss": 1.5524, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.012987012987012988, | |
| "grad_norm": 1.6673612594604492, | |
| "learning_rate": 6.932495846462261e-05, | |
| "loss": 1.5226, | |
| "step": 81 | |
| }, | |
| { | |
| "epoch": 0.013147346480679814, | |
| "grad_norm": 1.7997468709945679, | |
| "learning_rate": 6.855986244591104e-05, | |
| "loss": 1.5043, | |
| "step": 82 | |
| }, | |
| { | |
| "epoch": 0.013307679974346641, | |
| "grad_norm": 1.7401671409606934, | |
| "learning_rate": 6.778969234612584e-05, | |
| "loss": 1.1924, | |
| "step": 83 | |
| }, | |
| { | |
| "epoch": 0.013468013468013467, | |
| "grad_norm": 1.8019427061080933, | |
| "learning_rate": 6.701465872208216e-05, | |
| "loss": 1.1919, | |
| "step": 84 | |
| }, | |
| { | |
| "epoch": 0.013628346961680296, | |
| "grad_norm": 1.8377891778945923, | |
| "learning_rate": 6.623497346023418e-05, | |
| "loss": 1.1958, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 0.013788680455347122, | |
| "grad_norm": 1.9431484937667847, | |
| "learning_rate": 6.545084971874738e-05, | |
| "loss": 1.2923, | |
| "step": 86 | |
| }, | |
| { | |
| "epoch": 0.013949013949013949, | |
| "grad_norm": 1.7748252153396606, | |
| "learning_rate": 6.466250186922325e-05, | |
| "loss": 1.0951, | |
| "step": 87 | |
| }, | |
| { | |
| "epoch": 0.014109347442680775, | |
| "grad_norm": 1.8332815170288086, | |
| "learning_rate": 6.387014543809223e-05, | |
| "loss": 1.1729, | |
| "step": 88 | |
| }, | |
| { | |
| "epoch": 0.014269680936347603, | |
| "grad_norm": 2.066511869430542, | |
| "learning_rate": 6.307399704769099e-05, | |
| "loss": 1.2436, | |
| "step": 89 | |
| }, | |
| { | |
| "epoch": 0.01443001443001443, | |
| "grad_norm": 1.8436999320983887, | |
| "learning_rate": 6.227427435703997e-05, | |
| "loss": 1.2209, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.014590347923681257, | |
| "grad_norm": 1.9777435064315796, | |
| "learning_rate": 6.147119600233758e-05, | |
| "loss": 1.1729, | |
| "step": 91 | |
| }, | |
| { | |
| "epoch": 0.014750681417348085, | |
| "grad_norm": 1.988520622253418, | |
| "learning_rate": 6.066498153718735e-05, | |
| "loss": 1.2026, | |
| "step": 92 | |
| }, | |
| { | |
| "epoch": 0.014911014911014911, | |
| "grad_norm": 2.101755380630493, | |
| "learning_rate": 5.985585137257401e-05, | |
| "loss": 1.229, | |
| "step": 93 | |
| }, | |
| { | |
| "epoch": 0.015071348404681738, | |
| "grad_norm": 2.1809160709381104, | |
| "learning_rate": 5.90440267166055e-05, | |
| "loss": 1.0617, | |
| "step": 94 | |
| }, | |
| { | |
| "epoch": 0.015231681898348564, | |
| "grad_norm": 2.251002073287964, | |
| "learning_rate": 5.8229729514036705e-05, | |
| "loss": 1.326, | |
| "step": 95 | |
| }, | |
| { | |
| "epoch": 0.015392015392015393, | |
| "grad_norm": 2.272563934326172, | |
| "learning_rate": 5.74131823855921e-05, | |
| "loss": 1.0713, | |
| "step": 96 | |
| }, | |
| { | |
| "epoch": 0.015552348885682219, | |
| "grad_norm": 3.0095102787017822, | |
| "learning_rate": 5.6594608567103456e-05, | |
| "loss": 1.1834, | |
| "step": 97 | |
| }, | |
| { | |
| "epoch": 0.015712682379349047, | |
| "grad_norm": 2.3724913597106934, | |
| "learning_rate": 5.577423184847932e-05, | |
| "loss": 1.0536, | |
| "step": 98 | |
| }, | |
| { | |
| "epoch": 0.015873015873015872, | |
| "grad_norm": 3.3596227169036865, | |
| "learning_rate": 5.495227651252315e-05, | |
| "loss": 0.9999, | |
| "step": 99 | |
| }, | |
| { | |
| "epoch": 0.0160333493666827, | |
| "grad_norm": 4.8523383140563965, | |
| "learning_rate": 5.4128967273616625e-05, | |
| "loss": 0.9932, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.0160333493666827, | |
| "eval_loss": 1.577565312385559, | |
| "eval_runtime": 181.1273, | |
| "eval_samples_per_second": 57.998, | |
| "eval_steps_per_second": 14.504, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.01619368286034953, | |
| "grad_norm": 1.0778112411499023, | |
| "learning_rate": 5.330452921628497e-05, | |
| "loss": 0.7987, | |
| "step": 101 | |
| }, | |
| { | |
| "epoch": 0.016354016354016353, | |
| "grad_norm": 1.4329731464385986, | |
| "learning_rate": 5.247918773366112e-05, | |
| "loss": 1.2528, | |
| "step": 102 | |
| }, | |
| { | |
| "epoch": 0.01651434984768318, | |
| "grad_norm": 2.3502819538116455, | |
| "learning_rate": 5.165316846586541e-05, | |
| "loss": 1.3122, | |
| "step": 103 | |
| }, | |
| { | |
| "epoch": 0.016674683341350006, | |
| "grad_norm": 3.054696798324585, | |
| "learning_rate": 5.0826697238317935e-05, | |
| "loss": 1.8111, | |
| "step": 104 | |
| }, | |
| { | |
| "epoch": 0.016835016835016835, | |
| "grad_norm": 2.7006399631500244, | |
| "learning_rate": 5e-05, | |
| "loss": 2.0079, | |
| "step": 105 | |
| }, | |
| { | |
| "epoch": 0.016995350328683663, | |
| "grad_norm": 3.1934356689453125, | |
| "learning_rate": 4.917330276168208e-05, | |
| "loss": 1.9011, | |
| "step": 106 | |
| }, | |
| { | |
| "epoch": 0.017155683822350488, | |
| "grad_norm": 2.7194902896881104, | |
| "learning_rate": 4.834683153413459e-05, | |
| "loss": 2.2386, | |
| "step": 107 | |
| }, | |
| { | |
| "epoch": 0.017316017316017316, | |
| "grad_norm": 1.9565664529800415, | |
| "learning_rate": 4.7520812266338885e-05, | |
| "loss": 1.662, | |
| "step": 108 | |
| }, | |
| { | |
| "epoch": 0.017476350809684144, | |
| "grad_norm": 1.5166082382202148, | |
| "learning_rate": 4.669547078371504e-05, | |
| "loss": 2.062, | |
| "step": 109 | |
| }, | |
| { | |
| "epoch": 0.01763668430335097, | |
| "grad_norm": 1.7787351608276367, | |
| "learning_rate": 4.5871032726383386e-05, | |
| "loss": 2.043, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.017797017797017797, | |
| "grad_norm": 1.6311932802200317, | |
| "learning_rate": 4.504772348747687e-05, | |
| "loss": 1.6041, | |
| "step": 111 | |
| }, | |
| { | |
| "epoch": 0.017957351290684626, | |
| "grad_norm": 1.3286525011062622, | |
| "learning_rate": 4.4225768151520694e-05, | |
| "loss": 1.6954, | |
| "step": 112 | |
| }, | |
| { | |
| "epoch": 0.01811768478435145, | |
| "grad_norm": 1.4849814176559448, | |
| "learning_rate": 4.3405391432896555e-05, | |
| "loss": 1.8252, | |
| "step": 113 | |
| }, | |
| { | |
| "epoch": 0.01827801827801828, | |
| "grad_norm": 1.8410348892211914, | |
| "learning_rate": 4.2586817614407895e-05, | |
| "loss": 1.8354, | |
| "step": 114 | |
| }, | |
| { | |
| "epoch": 0.018438351771685103, | |
| "grad_norm": 1.9564002752304077, | |
| "learning_rate": 4.17702704859633e-05, | |
| "loss": 1.8874, | |
| "step": 115 | |
| }, | |
| { | |
| "epoch": 0.01859868526535193, | |
| "grad_norm": 1.6850190162658691, | |
| "learning_rate": 4.095597328339452e-05, | |
| "loss": 1.5473, | |
| "step": 116 | |
| }, | |
| { | |
| "epoch": 0.01875901875901876, | |
| "grad_norm": 1.796574354171753, | |
| "learning_rate": 4.0144148627425993e-05, | |
| "loss": 1.7481, | |
| "step": 117 | |
| }, | |
| { | |
| "epoch": 0.018919352252685585, | |
| "grad_norm": 1.8272076845169067, | |
| "learning_rate": 3.933501846281267e-05, | |
| "loss": 1.8834, | |
| "step": 118 | |
| }, | |
| { | |
| "epoch": 0.019079685746352413, | |
| "grad_norm": 1.3704986572265625, | |
| "learning_rate": 3.852880399766243e-05, | |
| "loss": 1.3574, | |
| "step": 119 | |
| }, | |
| { | |
| "epoch": 0.01924001924001924, | |
| "grad_norm": 1.4744528532028198, | |
| "learning_rate": 3.772572564296005e-05, | |
| "loss": 1.6148, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.019400352733686066, | |
| "grad_norm": 1.5670959949493408, | |
| "learning_rate": 3.6926002952309016e-05, | |
| "loss": 1.7067, | |
| "step": 121 | |
| }, | |
| { | |
| "epoch": 0.019560686227352894, | |
| "grad_norm": 1.4213238954544067, | |
| "learning_rate": 3.612985456190778e-05, | |
| "loss": 1.2831, | |
| "step": 122 | |
| }, | |
| { | |
| "epoch": 0.019721019721019722, | |
| "grad_norm": 1.4642093181610107, | |
| "learning_rate": 3.533749813077677e-05, | |
| "loss": 1.5041, | |
| "step": 123 | |
| }, | |
| { | |
| "epoch": 0.019881353214686547, | |
| "grad_norm": 1.4591537714004517, | |
| "learning_rate": 3.4549150281252636e-05, | |
| "loss": 1.4938, | |
| "step": 124 | |
| }, | |
| { | |
| "epoch": 0.020041686708353375, | |
| "grad_norm": 1.5452147722244263, | |
| "learning_rate": 3.3765026539765834e-05, | |
| "loss": 1.574, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 0.020202020202020204, | |
| "grad_norm": 1.4701132774353027, | |
| "learning_rate": 3.298534127791785e-05, | |
| "loss": 1.6096, | |
| "step": 126 | |
| }, | |
| { | |
| "epoch": 0.02036235369568703, | |
| "grad_norm": 1.6247003078460693, | |
| "learning_rate": 3.221030765387417e-05, | |
| "loss": 1.5613, | |
| "step": 127 | |
| }, | |
| { | |
| "epoch": 0.020522687189353857, | |
| "grad_norm": 1.6164422035217285, | |
| "learning_rate": 3.144013755408895e-05, | |
| "loss": 1.5354, | |
| "step": 128 | |
| }, | |
| { | |
| "epoch": 0.02068302068302068, | |
| "grad_norm": 1.541204810142517, | |
| "learning_rate": 3.0675041535377405e-05, | |
| "loss": 1.6355, | |
| "step": 129 | |
| }, | |
| { | |
| "epoch": 0.02084335417668751, | |
| "grad_norm": 1.5679917335510254, | |
| "learning_rate": 2.991522876735154e-05, | |
| "loss": 1.256, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.021003687670354338, | |
| "grad_norm": 1.5460208654403687, | |
| "learning_rate": 2.916090697523549e-05, | |
| "loss": 1.4764, | |
| "step": 131 | |
| }, | |
| { | |
| "epoch": 0.021164021164021163, | |
| "grad_norm": 1.5560331344604492, | |
| "learning_rate": 2.8412282383075363e-05, | |
| "loss": 1.3285, | |
| "step": 132 | |
| }, | |
| { | |
| "epoch": 0.02132435465768799, | |
| "grad_norm": 1.600095272064209, | |
| "learning_rate": 2.766955965735968e-05, | |
| "loss": 1.1649, | |
| "step": 133 | |
| }, | |
| { | |
| "epoch": 0.02148468815135482, | |
| "grad_norm": 1.6068081855773926, | |
| "learning_rate": 2.693294185106562e-05, | |
| "loss": 1.189, | |
| "step": 134 | |
| }, | |
| { | |
| "epoch": 0.021645021645021644, | |
| "grad_norm": 1.7338483333587646, | |
| "learning_rate": 2.6202630348146324e-05, | |
| "loss": 1.2842, | |
| "step": 135 | |
| }, | |
| { | |
| "epoch": 0.021805355138688472, | |
| "grad_norm": 1.8223528861999512, | |
| "learning_rate": 2.547882480847461e-05, | |
| "loss": 1.3976, | |
| "step": 136 | |
| }, | |
| { | |
| "epoch": 0.0219656886323553, | |
| "grad_norm": 1.7216356992721558, | |
| "learning_rate": 2.476172311325783e-05, | |
| "loss": 1.0166, | |
| "step": 137 | |
| }, | |
| { | |
| "epoch": 0.022126022126022125, | |
| "grad_norm": 1.7349539995193481, | |
| "learning_rate": 2.405152131093926e-05, | |
| "loss": 1.2565, | |
| "step": 138 | |
| }, | |
| { | |
| "epoch": 0.022286355619688954, | |
| "grad_norm": 1.7386369705200195, | |
| "learning_rate": 2.3348413563600325e-05, | |
| "loss": 1.0638, | |
| "step": 139 | |
| }, | |
| { | |
| "epoch": 0.02244668911335578, | |
| "grad_norm": 1.9614675045013428, | |
| "learning_rate": 2.2652592093878666e-05, | |
| "loss": 1.2197, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.022607022607022607, | |
| "grad_norm": 1.8332651853561401, | |
| "learning_rate": 2.196424713241637e-05, | |
| "loss": 1.2153, | |
| "step": 141 | |
| }, | |
| { | |
| "epoch": 0.022767356100689435, | |
| "grad_norm": 2.135434150695801, | |
| "learning_rate": 2.128356686585282e-05, | |
| "loss": 1.1492, | |
| "step": 142 | |
| }, | |
| { | |
| "epoch": 0.02292768959435626, | |
| "grad_norm": 2.1699466705322266, | |
| "learning_rate": 2.061073738537635e-05, | |
| "loss": 1.267, | |
| "step": 143 | |
| }, | |
| { | |
| "epoch": 0.023088023088023088, | |
| "grad_norm": 2.128176689147949, | |
| "learning_rate": 1.9945942635848748e-05, | |
| "loss": 1.0789, | |
| "step": 144 | |
| }, | |
| { | |
| "epoch": 0.023248356581689916, | |
| "grad_norm": 2.194796323776245, | |
| "learning_rate": 1.928936436551661e-05, | |
| "loss": 1.2463, | |
| "step": 145 | |
| }, | |
| { | |
| "epoch": 0.02340869007535674, | |
| "grad_norm": 2.179041624069214, | |
| "learning_rate": 1.8641182076323148e-05, | |
| "loss": 1.0174, | |
| "step": 146 | |
| }, | |
| { | |
| "epoch": 0.02356902356902357, | |
| "grad_norm": 2.251429796218872, | |
| "learning_rate": 1.800157297483417e-05, | |
| "loss": 1.0932, | |
| "step": 147 | |
| }, | |
| { | |
| "epoch": 0.023729357062690398, | |
| "grad_norm": 2.3761861324310303, | |
| "learning_rate": 1.7370711923791567e-05, | |
| "loss": 1.0235, | |
| "step": 148 | |
| }, | |
| { | |
| "epoch": 0.023889690556357222, | |
| "grad_norm": 2.7878284454345703, | |
| "learning_rate": 1.6748771394307585e-05, | |
| "loss": 1.1111, | |
| "step": 149 | |
| }, | |
| { | |
| "epoch": 0.02405002405002405, | |
| "grad_norm": 3.860334634780884, | |
| "learning_rate": 1.6135921418712956e-05, | |
| "loss": 1.1455, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.02405002405002405, | |
| "eval_loss": 1.484871745109558, | |
| "eval_runtime": 181.1474, | |
| "eval_samples_per_second": 57.991, | |
| "eval_steps_per_second": 14.502, | |
| "step": 150 | |
| } | |
| ], | |
| "logging_steps": 1, | |
| "max_steps": 200, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 50, | |
| "stateful_callbacks": { | |
| "EarlyStoppingCallback": { | |
| "args": { | |
| "early_stopping_patience": 5, | |
| "early_stopping_threshold": 0.0 | |
| }, | |
| "attributes": { | |
| "early_stopping_patience_counter": 0 | |
| } | |
| }, | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 1.15927236476928e+16, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |