| { | |
| "best_metric": 1.796425700187683, | |
| "best_model_checkpoint": "miner_id_24/checkpoint-150", | |
| "epoch": 0.05062436719541006, | |
| "eval_steps": 50, | |
| "global_step": 150, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0003374957813027337, | |
| "grad_norm": 0.659518837928772, | |
| "learning_rate": 1e-05, | |
| "loss": 1.7311, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.0003374957813027337, | |
| "eval_loss": 2.9957451820373535, | |
| "eval_runtime": 115.497, | |
| "eval_samples_per_second": 43.205, | |
| "eval_steps_per_second": 10.805, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.0006749915626054674, | |
| "grad_norm": 0.7038252353668213, | |
| "learning_rate": 2e-05, | |
| "loss": 1.9467, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.0010124873439082012, | |
| "grad_norm": 0.7937076091766357, | |
| "learning_rate": 3e-05, | |
| "loss": 2.1449, | |
| "step": 3 | |
| }, | |
| { | |
| "epoch": 0.0013499831252109348, | |
| "grad_norm": 0.7090626955032349, | |
| "learning_rate": 4e-05, | |
| "loss": 2.052, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.0016874789065136687, | |
| "grad_norm": 0.7988624572753906, | |
| "learning_rate": 5e-05, | |
| "loss": 2.3878, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.0020249746878164025, | |
| "grad_norm": 0.9716485738754272, | |
| "learning_rate": 6e-05, | |
| "loss": 2.1794, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.002362470469119136, | |
| "grad_norm": 0.8470109105110168, | |
| "learning_rate": 7e-05, | |
| "loss": 2.2386, | |
| "step": 7 | |
| }, | |
| { | |
| "epoch": 0.0026999662504218697, | |
| "grad_norm": 0.7678182125091553, | |
| "learning_rate": 8e-05, | |
| "loss": 2.2977, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.0030374620317246033, | |
| "grad_norm": 0.7406213283538818, | |
| "learning_rate": 9e-05, | |
| "loss": 2.2716, | |
| "step": 9 | |
| }, | |
| { | |
| "epoch": 0.0033749578130273373, | |
| "grad_norm": 0.6096116304397583, | |
| "learning_rate": 0.0001, | |
| "loss": 2.4181, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.003712453594330071, | |
| "grad_norm": 0.6520525217056274, | |
| "learning_rate": 9.999316524962345e-05, | |
| "loss": 2.0193, | |
| "step": 11 | |
| }, | |
| { | |
| "epoch": 0.004049949375632805, | |
| "grad_norm": 0.5647959113121033, | |
| "learning_rate": 9.997266286704631e-05, | |
| "loss": 2.1552, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 0.004387445156935539, | |
| "grad_norm": 0.7149708271026611, | |
| "learning_rate": 9.993849845741524e-05, | |
| "loss": 2.3396, | |
| "step": 13 | |
| }, | |
| { | |
| "epoch": 0.004724940938238272, | |
| "grad_norm": 0.6472476720809937, | |
| "learning_rate": 9.989068136093873e-05, | |
| "loss": 2.1707, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 0.005062436719541006, | |
| "grad_norm": 0.7469307780265808, | |
| "learning_rate": 9.98292246503335e-05, | |
| "loss": 2.3344, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.005399932500843739, | |
| "grad_norm": 0.5477833151817322, | |
| "learning_rate": 9.975414512725057e-05, | |
| "loss": 2.1046, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 0.005737428282146473, | |
| "grad_norm": 0.6585941314697266, | |
| "learning_rate": 9.966546331768191e-05, | |
| "loss": 2.26, | |
| "step": 17 | |
| }, | |
| { | |
| "epoch": 0.006074924063449207, | |
| "grad_norm": 0.6190670728683472, | |
| "learning_rate": 9.956320346634876e-05, | |
| "loss": 2.1344, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 0.00641241984475194, | |
| "grad_norm": 0.5457659363746643, | |
| "learning_rate": 9.944739353007344e-05, | |
| "loss": 2.1121, | |
| "step": 19 | |
| }, | |
| { | |
| "epoch": 0.006749915626054675, | |
| "grad_norm": 0.48350393772125244, | |
| "learning_rate": 9.931806517013612e-05, | |
| "loss": 1.9028, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.007087411407357408, | |
| "grad_norm": 0.6461214423179626, | |
| "learning_rate": 9.917525374361912e-05, | |
| "loss": 2.166, | |
| "step": 21 | |
| }, | |
| { | |
| "epoch": 0.007424907188660142, | |
| "grad_norm": 0.6612595319747925, | |
| "learning_rate": 9.901899829374047e-05, | |
| "loss": 2.2152, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 0.0077624029699628755, | |
| "grad_norm": 0.5887338519096375, | |
| "learning_rate": 9.884934153917997e-05, | |
| "loss": 2.0533, | |
| "step": 23 | |
| }, | |
| { | |
| "epoch": 0.00809989875126561, | |
| "grad_norm": 0.5120691657066345, | |
| "learning_rate": 9.86663298624003e-05, | |
| "loss": 2.1555, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 0.008437394532568343, | |
| "grad_norm": 0.5528091788291931, | |
| "learning_rate": 9.847001329696653e-05, | |
| "loss": 2.2499, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.008774890313871077, | |
| "grad_norm": 0.6201184391975403, | |
| "learning_rate": 9.826044551386744e-05, | |
| "loss": 2.1444, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 0.00911238609517381, | |
| "grad_norm": 0.6279026865959167, | |
| "learning_rate": 9.803768380684242e-05, | |
| "loss": 2.2847, | |
| "step": 27 | |
| }, | |
| { | |
| "epoch": 0.009449881876476544, | |
| "grad_norm": 0.6332525014877319, | |
| "learning_rate": 9.780178907671789e-05, | |
| "loss": 1.958, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 0.009787377657779277, | |
| "grad_norm": 0.569353461265564, | |
| "learning_rate": 9.755282581475769e-05, | |
| "loss": 2.1134, | |
| "step": 29 | |
| }, | |
| { | |
| "epoch": 0.010124873439082012, | |
| "grad_norm": 0.5256232619285583, | |
| "learning_rate": 9.729086208503174e-05, | |
| "loss": 1.9178, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.010462369220384746, | |
| "grad_norm": 0.5967089533805847, | |
| "learning_rate": 9.701596950580806e-05, | |
| "loss": 2.3453, | |
| "step": 31 | |
| }, | |
| { | |
| "epoch": 0.010799865001687479, | |
| "grad_norm": 0.6425303816795349, | |
| "learning_rate": 9.672822322997305e-05, | |
| "loss": 2.1351, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 0.011137360782990213, | |
| "grad_norm": 0.5986398458480835, | |
| "learning_rate": 9.642770192448536e-05, | |
| "loss": 2.3019, | |
| "step": 33 | |
| }, | |
| { | |
| "epoch": 0.011474856564292946, | |
| "grad_norm": 0.5946639180183411, | |
| "learning_rate": 9.611448774886924e-05, | |
| "loss": 2.2574, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 0.01181235234559568, | |
| "grad_norm": 0.6006338000297546, | |
| "learning_rate": 9.578866633275288e-05, | |
| "loss": 2.3178, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.012149848126898413, | |
| "grad_norm": 0.5436823964118958, | |
| "learning_rate": 9.545032675245813e-05, | |
| "loss": 2.2969, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 0.012487343908201148, | |
| "grad_norm": 0.589877724647522, | |
| "learning_rate": 9.509956150664796e-05, | |
| "loss": 2.4032, | |
| "step": 37 | |
| }, | |
| { | |
| "epoch": 0.01282483968950388, | |
| "grad_norm": 0.5993760228157043, | |
| "learning_rate": 9.473646649103818e-05, | |
| "loss": 2.3421, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 0.013162335470806615, | |
| "grad_norm": 0.5923803448677063, | |
| "learning_rate": 9.43611409721806e-05, | |
| "loss": 2.3048, | |
| "step": 39 | |
| }, | |
| { | |
| "epoch": 0.01349983125210935, | |
| "grad_norm": 0.6783635020256042, | |
| "learning_rate": 9.397368756032445e-05, | |
| "loss": 2.3467, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.013837327033412082, | |
| "grad_norm": 0.5711936354637146, | |
| "learning_rate": 9.357421218136386e-05, | |
| "loss": 2.3039, | |
| "step": 41 | |
| }, | |
| { | |
| "epoch": 0.014174822814714817, | |
| "grad_norm": 0.6027695536613464, | |
| "learning_rate": 9.316282404787871e-05, | |
| "loss": 2.2872, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 0.01451231859601755, | |
| "grad_norm": 0.6241564750671387, | |
| "learning_rate": 9.273963562927695e-05, | |
| "loss": 2.313, | |
| "step": 43 | |
| }, | |
| { | |
| "epoch": 0.014849814377320284, | |
| "grad_norm": 0.6565014123916626, | |
| "learning_rate": 9.230476262104677e-05, | |
| "loss": 2.3402, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 0.015187310158623017, | |
| "grad_norm": 0.6258305311203003, | |
| "learning_rate": 9.185832391312644e-05, | |
| "loss": 2.2641, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 0.015524805939925751, | |
| "grad_norm": 0.6962496042251587, | |
| "learning_rate": 9.140044155740101e-05, | |
| "loss": 2.5017, | |
| "step": 46 | |
| }, | |
| { | |
| "epoch": 0.015862301721228485, | |
| "grad_norm": 0.643883466720581, | |
| "learning_rate": 9.093124073433463e-05, | |
| "loss": 2.4286, | |
| "step": 47 | |
| }, | |
| { | |
| "epoch": 0.01619979750253122, | |
| "grad_norm": 0.6366576552391052, | |
| "learning_rate": 9.045084971874738e-05, | |
| "loss": 2.3306, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 0.01653729328383395, | |
| "grad_norm": 0.723232090473175, | |
| "learning_rate": 8.995939984474624e-05, | |
| "loss": 2.4223, | |
| "step": 49 | |
| }, | |
| { | |
| "epoch": 0.016874789065136685, | |
| "grad_norm": 0.8257110118865967, | |
| "learning_rate": 8.945702546981969e-05, | |
| "loss": 2.3262, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.016874789065136685, | |
| "eval_loss": 1.9447404146194458, | |
| "eval_runtime": 114.5205, | |
| "eval_samples_per_second": 43.573, | |
| "eval_steps_per_second": 10.898, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.01721228484643942, | |
| "grad_norm": 0.47083166241645813, | |
| "learning_rate": 8.894386393810563e-05, | |
| "loss": 1.2425, | |
| "step": 51 | |
| }, | |
| { | |
| "epoch": 0.017549780627742154, | |
| "grad_norm": 0.5380712747573853, | |
| "learning_rate": 8.842005554284296e-05, | |
| "loss": 1.3438, | |
| "step": 52 | |
| }, | |
| { | |
| "epoch": 0.017887276409044885, | |
| "grad_norm": 0.4980904757976532, | |
| "learning_rate": 8.788574348801675e-05, | |
| "loss": 1.3243, | |
| "step": 53 | |
| }, | |
| { | |
| "epoch": 0.01822477219034762, | |
| "grad_norm": 0.5424904823303223, | |
| "learning_rate": 8.73410738492077e-05, | |
| "loss": 1.3158, | |
| "step": 54 | |
| }, | |
| { | |
| "epoch": 0.018562267971650354, | |
| "grad_norm": 0.5894560217857361, | |
| "learning_rate": 8.678619553365659e-05, | |
| "loss": 1.6077, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 0.01889976375295309, | |
| "grad_norm": 0.601193368434906, | |
| "learning_rate": 8.622126023955446e-05, | |
| "loss": 1.6657, | |
| "step": 56 | |
| }, | |
| { | |
| "epoch": 0.019237259534255823, | |
| "grad_norm": 0.5369265675544739, | |
| "learning_rate": 8.564642241456986e-05, | |
| "loss": 1.7473, | |
| "step": 57 | |
| }, | |
| { | |
| "epoch": 0.019574755315558554, | |
| "grad_norm": 0.5330817103385925, | |
| "learning_rate": 8.506183921362443e-05, | |
| "loss": 1.8135, | |
| "step": 58 | |
| }, | |
| { | |
| "epoch": 0.01991225109686129, | |
| "grad_norm": 0.554940402507782, | |
| "learning_rate": 8.44676704559283e-05, | |
| "loss": 1.6549, | |
| "step": 59 | |
| }, | |
| { | |
| "epoch": 0.020249746878164023, | |
| "grad_norm": 0.500688910484314, | |
| "learning_rate": 8.386407858128706e-05, | |
| "loss": 2.022, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.020587242659466758, | |
| "grad_norm": 0.44095098972320557, | |
| "learning_rate": 8.32512286056924e-05, | |
| "loss": 1.5022, | |
| "step": 61 | |
| }, | |
| { | |
| "epoch": 0.020924738440769492, | |
| "grad_norm": 0.48454850912094116, | |
| "learning_rate": 8.262928807620843e-05, | |
| "loss": 1.8577, | |
| "step": 62 | |
| }, | |
| { | |
| "epoch": 0.021262234222072223, | |
| "grad_norm": 0.48471105098724365, | |
| "learning_rate": 8.199842702516583e-05, | |
| "loss": 1.8668, | |
| "step": 63 | |
| }, | |
| { | |
| "epoch": 0.021599730003374958, | |
| "grad_norm": 0.4809412360191345, | |
| "learning_rate": 8.135881792367686e-05, | |
| "loss": 1.7112, | |
| "step": 64 | |
| }, | |
| { | |
| "epoch": 0.021937225784677692, | |
| "grad_norm": 0.5266463160514832, | |
| "learning_rate": 8.07106356344834e-05, | |
| "loss": 1.5309, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 0.022274721565980427, | |
| "grad_norm": 0.4586962163448334, | |
| "learning_rate": 8.005405736415126e-05, | |
| "loss": 1.9233, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 0.022612217347283158, | |
| "grad_norm": 0.46809718012809753, | |
| "learning_rate": 7.938926261462366e-05, | |
| "loss": 1.8757, | |
| "step": 67 | |
| }, | |
| { | |
| "epoch": 0.022949713128585892, | |
| "grad_norm": 0.5924503207206726, | |
| "learning_rate": 7.871643313414718e-05, | |
| "loss": 1.726, | |
| "step": 68 | |
| }, | |
| { | |
| "epoch": 0.023287208909888626, | |
| "grad_norm": 0.5335246920585632, | |
| "learning_rate": 7.803575286758364e-05, | |
| "loss": 1.9682, | |
| "step": 69 | |
| }, | |
| { | |
| "epoch": 0.02362470469119136, | |
| "grad_norm": 0.5980493426322937, | |
| "learning_rate": 7.734740790612136e-05, | |
| "loss": 1.93, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.023962200472494095, | |
| "grad_norm": 0.5206842422485352, | |
| "learning_rate": 7.66515864363997e-05, | |
| "loss": 1.9585, | |
| "step": 71 | |
| }, | |
| { | |
| "epoch": 0.024299696253796826, | |
| "grad_norm": 0.541953444480896, | |
| "learning_rate": 7.594847868906076e-05, | |
| "loss": 1.8451, | |
| "step": 72 | |
| }, | |
| { | |
| "epoch": 0.02463719203509956, | |
| "grad_norm": 0.4906903803348541, | |
| "learning_rate": 7.52382768867422e-05, | |
| "loss": 2.1014, | |
| "step": 73 | |
| }, | |
| { | |
| "epoch": 0.024974687816402295, | |
| "grad_norm": 0.49488648772239685, | |
| "learning_rate": 7.452117519152542e-05, | |
| "loss": 1.8397, | |
| "step": 74 | |
| }, | |
| { | |
| "epoch": 0.02531218359770503, | |
| "grad_norm": 0.49567532539367676, | |
| "learning_rate": 7.379736965185368e-05, | |
| "loss": 1.7313, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.02564967937900776, | |
| "grad_norm": 0.5103468298912048, | |
| "learning_rate": 7.30670581489344e-05, | |
| "loss": 1.8343, | |
| "step": 76 | |
| }, | |
| { | |
| "epoch": 0.025987175160310495, | |
| "grad_norm": 0.4970777630805969, | |
| "learning_rate": 7.233044034264034e-05, | |
| "loss": 1.6868, | |
| "step": 77 | |
| }, | |
| { | |
| "epoch": 0.02632467094161323, | |
| "grad_norm": 0.5564061403274536, | |
| "learning_rate": 7.158771761692464e-05, | |
| "loss": 2.0077, | |
| "step": 78 | |
| }, | |
| { | |
| "epoch": 0.026662166722915964, | |
| "grad_norm": 0.5906986594200134, | |
| "learning_rate": 7.083909302476453e-05, | |
| "loss": 1.9044, | |
| "step": 79 | |
| }, | |
| { | |
| "epoch": 0.0269996625042187, | |
| "grad_norm": 0.5161340236663818, | |
| "learning_rate": 7.008477123264848e-05, | |
| "loss": 1.8798, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.02733715828552143, | |
| "grad_norm": 0.5620460510253906, | |
| "learning_rate": 6.932495846462261e-05, | |
| "loss": 1.8767, | |
| "step": 81 | |
| }, | |
| { | |
| "epoch": 0.027674654066824164, | |
| "grad_norm": 0.5983588695526123, | |
| "learning_rate": 6.855986244591104e-05, | |
| "loss": 2.0385, | |
| "step": 82 | |
| }, | |
| { | |
| "epoch": 0.0280121498481269, | |
| "grad_norm": 0.5406123995780945, | |
| "learning_rate": 6.778969234612584e-05, | |
| "loss": 1.9357, | |
| "step": 83 | |
| }, | |
| { | |
| "epoch": 0.028349645629429633, | |
| "grad_norm": 0.5965004563331604, | |
| "learning_rate": 6.701465872208216e-05, | |
| "loss": 1.846, | |
| "step": 84 | |
| }, | |
| { | |
| "epoch": 0.028687141410732364, | |
| "grad_norm": 0.5949670076370239, | |
| "learning_rate": 6.623497346023418e-05, | |
| "loss": 1.9988, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 0.0290246371920351, | |
| "grad_norm": 0.646133303642273, | |
| "learning_rate": 6.545084971874738e-05, | |
| "loss": 1.8797, | |
| "step": 86 | |
| }, | |
| { | |
| "epoch": 0.029362132973337833, | |
| "grad_norm": 0.5879825949668884, | |
| "learning_rate": 6.466250186922325e-05, | |
| "loss": 2.0766, | |
| "step": 87 | |
| }, | |
| { | |
| "epoch": 0.029699628754640568, | |
| "grad_norm": 0.7212510108947754, | |
| "learning_rate": 6.387014543809223e-05, | |
| "loss": 1.9622, | |
| "step": 88 | |
| }, | |
| { | |
| "epoch": 0.030037124535943302, | |
| "grad_norm": 0.5758910775184631, | |
| "learning_rate": 6.307399704769099e-05, | |
| "loss": 2.0455, | |
| "step": 89 | |
| }, | |
| { | |
| "epoch": 0.030374620317246033, | |
| "grad_norm": 0.648313581943512, | |
| "learning_rate": 6.227427435703997e-05, | |
| "loss": 2.0323, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.030712116098548767, | |
| "grad_norm": 0.6412401795387268, | |
| "learning_rate": 6.147119600233758e-05, | |
| "loss": 2.2864, | |
| "step": 91 | |
| }, | |
| { | |
| "epoch": 0.031049611879851502, | |
| "grad_norm": 0.6075218915939331, | |
| "learning_rate": 6.066498153718735e-05, | |
| "loss": 2.0373, | |
| "step": 92 | |
| }, | |
| { | |
| "epoch": 0.03138710766115423, | |
| "grad_norm": 0.614220142364502, | |
| "learning_rate": 5.985585137257401e-05, | |
| "loss": 2.1688, | |
| "step": 93 | |
| }, | |
| { | |
| "epoch": 0.03172460344245697, | |
| "grad_norm": 0.6434118151664734, | |
| "learning_rate": 5.90440267166055e-05, | |
| "loss": 2.1976, | |
| "step": 94 | |
| }, | |
| { | |
| "epoch": 0.0320620992237597, | |
| "grad_norm": 0.6182180643081665, | |
| "learning_rate": 5.8229729514036705e-05, | |
| "loss": 2.2718, | |
| "step": 95 | |
| }, | |
| { | |
| "epoch": 0.03239959500506244, | |
| "grad_norm": 0.6811787486076355, | |
| "learning_rate": 5.74131823855921e-05, | |
| "loss": 2.2379, | |
| "step": 96 | |
| }, | |
| { | |
| "epoch": 0.03273709078636517, | |
| "grad_norm": 0.6939035654067993, | |
| "learning_rate": 5.6594608567103456e-05, | |
| "loss": 2.2805, | |
| "step": 97 | |
| }, | |
| { | |
| "epoch": 0.0330745865676679, | |
| "grad_norm": 0.7259014248847961, | |
| "learning_rate": 5.577423184847932e-05, | |
| "loss": 2.2849, | |
| "step": 98 | |
| }, | |
| { | |
| "epoch": 0.03341208234897064, | |
| "grad_norm": 0.6974567174911499, | |
| "learning_rate": 5.495227651252315e-05, | |
| "loss": 2.3273, | |
| "step": 99 | |
| }, | |
| { | |
| "epoch": 0.03374957813027337, | |
| "grad_norm": 0.8715429902076721, | |
| "learning_rate": 5.4128967273616625e-05, | |
| "loss": 2.2911, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.03374957813027337, | |
| "eval_loss": 1.8426469564437866, | |
| "eval_runtime": 115.1302, | |
| "eval_samples_per_second": 43.342, | |
| "eval_steps_per_second": 10.84, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.03408707391157611, | |
| "grad_norm": 0.5058120489120483, | |
| "learning_rate": 5.330452921628497e-05, | |
| "loss": 0.9619, | |
| "step": 101 | |
| }, | |
| { | |
| "epoch": 0.03442456969287884, | |
| "grad_norm": 0.4607161283493042, | |
| "learning_rate": 5.247918773366112e-05, | |
| "loss": 1.2164, | |
| "step": 102 | |
| }, | |
| { | |
| "epoch": 0.03476206547418157, | |
| "grad_norm": 0.525092601776123, | |
| "learning_rate": 5.165316846586541e-05, | |
| "loss": 1.2863, | |
| "step": 103 | |
| }, | |
| { | |
| "epoch": 0.03509956125548431, | |
| "grad_norm": 0.482377290725708, | |
| "learning_rate": 5.0826697238317935e-05, | |
| "loss": 1.3539, | |
| "step": 104 | |
| }, | |
| { | |
| "epoch": 0.03543705703678704, | |
| "grad_norm": 0.5290610790252686, | |
| "learning_rate": 5e-05, | |
| "loss": 1.226, | |
| "step": 105 | |
| }, | |
| { | |
| "epoch": 0.03577455281808977, | |
| "grad_norm": 0.4935402572154999, | |
| "learning_rate": 4.917330276168208e-05, | |
| "loss": 1.4144, | |
| "step": 106 | |
| }, | |
| { | |
| "epoch": 0.03611204859939251, | |
| "grad_norm": 0.4793010354042053, | |
| "learning_rate": 4.834683153413459e-05, | |
| "loss": 1.5215, | |
| "step": 107 | |
| }, | |
| { | |
| "epoch": 0.03644954438069524, | |
| "grad_norm": 0.583838939666748, | |
| "learning_rate": 4.7520812266338885e-05, | |
| "loss": 1.5567, | |
| "step": 108 | |
| }, | |
| { | |
| "epoch": 0.03678704016199798, | |
| "grad_norm": 0.5468663573265076, | |
| "learning_rate": 4.669547078371504e-05, | |
| "loss": 1.7712, | |
| "step": 109 | |
| }, | |
| { | |
| "epoch": 0.03712453594330071, | |
| "grad_norm": 0.4651382267475128, | |
| "learning_rate": 4.5871032726383386e-05, | |
| "loss": 1.6748, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.03746203172460344, | |
| "grad_norm": 0.4551399350166321, | |
| "learning_rate": 4.504772348747687e-05, | |
| "loss": 1.469, | |
| "step": 111 | |
| }, | |
| { | |
| "epoch": 0.03779952750590618, | |
| "grad_norm": 0.5389413833618164, | |
| "learning_rate": 4.4225768151520694e-05, | |
| "loss": 1.6551, | |
| "step": 112 | |
| }, | |
| { | |
| "epoch": 0.03813702328720891, | |
| "grad_norm": 0.5110684633255005, | |
| "learning_rate": 4.3405391432896555e-05, | |
| "loss": 1.7447, | |
| "step": 113 | |
| }, | |
| { | |
| "epoch": 0.038474519068511646, | |
| "grad_norm": 0.4776013493537903, | |
| "learning_rate": 4.2586817614407895e-05, | |
| "loss": 1.6281, | |
| "step": 114 | |
| }, | |
| { | |
| "epoch": 0.03881201484981438, | |
| "grad_norm": 0.5724086165428162, | |
| "learning_rate": 4.17702704859633e-05, | |
| "loss": 1.7601, | |
| "step": 115 | |
| }, | |
| { | |
| "epoch": 0.03914951063111711, | |
| "grad_norm": 0.5197787284851074, | |
| "learning_rate": 4.095597328339452e-05, | |
| "loss": 2.0058, | |
| "step": 116 | |
| }, | |
| { | |
| "epoch": 0.039487006412419846, | |
| "grad_norm": 0.48396414518356323, | |
| "learning_rate": 4.0144148627425993e-05, | |
| "loss": 1.8034, | |
| "step": 117 | |
| }, | |
| { | |
| "epoch": 0.03982450219372258, | |
| "grad_norm": 0.577552318572998, | |
| "learning_rate": 3.933501846281267e-05, | |
| "loss": 1.9499, | |
| "step": 118 | |
| }, | |
| { | |
| "epoch": 0.040161997975025315, | |
| "grad_norm": 0.5266735553741455, | |
| "learning_rate": 3.852880399766243e-05, | |
| "loss": 1.6601, | |
| "step": 119 | |
| }, | |
| { | |
| "epoch": 0.040499493756328046, | |
| "grad_norm": 0.5228984951972961, | |
| "learning_rate": 3.772572564296005e-05, | |
| "loss": 1.843, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.04083698953763078, | |
| "grad_norm": 0.5022814273834229, | |
| "learning_rate": 3.6926002952309016e-05, | |
| "loss": 1.606, | |
| "step": 121 | |
| }, | |
| { | |
| "epoch": 0.041174485318933515, | |
| "grad_norm": 0.5120666027069092, | |
| "learning_rate": 3.612985456190778e-05, | |
| "loss": 1.6744, | |
| "step": 122 | |
| }, | |
| { | |
| "epoch": 0.041511981100236246, | |
| "grad_norm": 0.5393658876419067, | |
| "learning_rate": 3.533749813077677e-05, | |
| "loss": 1.847, | |
| "step": 123 | |
| }, | |
| { | |
| "epoch": 0.041849476881538984, | |
| "grad_norm": 0.578925609588623, | |
| "learning_rate": 3.4549150281252636e-05, | |
| "loss": 1.8629, | |
| "step": 124 | |
| }, | |
| { | |
| "epoch": 0.042186972662841715, | |
| "grad_norm": 0.5105193257331848, | |
| "learning_rate": 3.3765026539765834e-05, | |
| "loss": 1.8517, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 0.042524468444144446, | |
| "grad_norm": 0.5774948000907898, | |
| "learning_rate": 3.298534127791785e-05, | |
| "loss": 1.8201, | |
| "step": 126 | |
| }, | |
| { | |
| "epoch": 0.042861964225447184, | |
| "grad_norm": 0.5090852975845337, | |
| "learning_rate": 3.221030765387417e-05, | |
| "loss": 1.8044, | |
| "step": 127 | |
| }, | |
| { | |
| "epoch": 0.043199460006749915, | |
| "grad_norm": 0.5966777801513672, | |
| "learning_rate": 3.144013755408895e-05, | |
| "loss": 1.9381, | |
| "step": 128 | |
| }, | |
| { | |
| "epoch": 0.043536955788052646, | |
| "grad_norm": 0.5561245083808899, | |
| "learning_rate": 3.0675041535377405e-05, | |
| "loss": 1.8774, | |
| "step": 129 | |
| }, | |
| { | |
| "epoch": 0.043874451569355384, | |
| "grad_norm": 0.5605421662330627, | |
| "learning_rate": 2.991522876735154e-05, | |
| "loss": 1.9417, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.044211947350658115, | |
| "grad_norm": 0.5726792812347412, | |
| "learning_rate": 2.916090697523549e-05, | |
| "loss": 1.865, | |
| "step": 131 | |
| }, | |
| { | |
| "epoch": 0.04454944313196085, | |
| "grad_norm": 0.4975649416446686, | |
| "learning_rate": 2.8412282383075363e-05, | |
| "loss": 1.7276, | |
| "step": 132 | |
| }, | |
| { | |
| "epoch": 0.044886938913263584, | |
| "grad_norm": 0.5293015241622925, | |
| "learning_rate": 2.766955965735968e-05, | |
| "loss": 1.9916, | |
| "step": 133 | |
| }, | |
| { | |
| "epoch": 0.045224434694566315, | |
| "grad_norm": 0.5684001445770264, | |
| "learning_rate": 2.693294185106562e-05, | |
| "loss": 2.0413, | |
| "step": 134 | |
| }, | |
| { | |
| "epoch": 0.04556193047586905, | |
| "grad_norm": 0.6660090684890747, | |
| "learning_rate": 2.6202630348146324e-05, | |
| "loss": 2.2261, | |
| "step": 135 | |
| }, | |
| { | |
| "epoch": 0.045899426257171784, | |
| "grad_norm": 0.6161995530128479, | |
| "learning_rate": 2.547882480847461e-05, | |
| "loss": 2.0809, | |
| "step": 136 | |
| }, | |
| { | |
| "epoch": 0.04623692203847452, | |
| "grad_norm": 0.5865457653999329, | |
| "learning_rate": 2.476172311325783e-05, | |
| "loss": 2.1305, | |
| "step": 137 | |
| }, | |
| { | |
| "epoch": 0.04657441781977725, | |
| "grad_norm": 0.6083557605743408, | |
| "learning_rate": 2.405152131093926e-05, | |
| "loss": 2.1383, | |
| "step": 138 | |
| }, | |
| { | |
| "epoch": 0.046911913601079984, | |
| "grad_norm": 0.6430215239524841, | |
| "learning_rate": 2.3348413563600325e-05, | |
| "loss": 2.0086, | |
| "step": 139 | |
| }, | |
| { | |
| "epoch": 0.04724940938238272, | |
| "grad_norm": 0.6370031237602234, | |
| "learning_rate": 2.2652592093878666e-05, | |
| "loss": 2.0208, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.04758690516368545, | |
| "grad_norm": 0.6458551287651062, | |
| "learning_rate": 2.196424713241637e-05, | |
| "loss": 2.066, | |
| "step": 141 | |
| }, | |
| { | |
| "epoch": 0.04792440094498819, | |
| "grad_norm": 0.5669720768928528, | |
| "learning_rate": 2.128356686585282e-05, | |
| "loss": 1.9751, | |
| "step": 142 | |
| }, | |
| { | |
| "epoch": 0.04826189672629092, | |
| "grad_norm": 0.6706758737564087, | |
| "learning_rate": 2.061073738537635e-05, | |
| "loss": 2.1718, | |
| "step": 143 | |
| }, | |
| { | |
| "epoch": 0.04859939250759365, | |
| "grad_norm": 0.6760194897651672, | |
| "learning_rate": 1.9945942635848748e-05, | |
| "loss": 2.1928, | |
| "step": 144 | |
| }, | |
| { | |
| "epoch": 0.04893688828889639, | |
| "grad_norm": 0.6566415429115295, | |
| "learning_rate": 1.928936436551661e-05, | |
| "loss": 1.973, | |
| "step": 145 | |
| }, | |
| { | |
| "epoch": 0.04927438407019912, | |
| "grad_norm": 0.7219246029853821, | |
| "learning_rate": 1.8641182076323148e-05, | |
| "loss": 2.1041, | |
| "step": 146 | |
| }, | |
| { | |
| "epoch": 0.04961187985150186, | |
| "grad_norm": 0.7868941426277161, | |
| "learning_rate": 1.800157297483417e-05, | |
| "loss": 2.2087, | |
| "step": 147 | |
| }, | |
| { | |
| "epoch": 0.04994937563280459, | |
| "grad_norm": 0.6628007292747498, | |
| "learning_rate": 1.7370711923791567e-05, | |
| "loss": 2.2978, | |
| "step": 148 | |
| }, | |
| { | |
| "epoch": 0.05028687141410732, | |
| "grad_norm": 0.6768451929092407, | |
| "learning_rate": 1.6748771394307585e-05, | |
| "loss": 2.2149, | |
| "step": 149 | |
| }, | |
| { | |
| "epoch": 0.05062436719541006, | |
| "grad_norm": 0.8374374508857727, | |
| "learning_rate": 1.6135921418712956e-05, | |
| "loss": 2.1397, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.05062436719541006, | |
| "eval_loss": 1.796425700187683, | |
| "eval_runtime": 114.7901, | |
| "eval_samples_per_second": 43.471, | |
| "eval_steps_per_second": 10.872, | |
| "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": 1.0303319310336e+16, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |