| { | |
| "best_metric": 0.00025325187016278505, | |
| "best_model_checkpoint": "miner_id_24/checkpoint-150", | |
| "epoch": 0.33277870216306155, | |
| "eval_steps": 50, | |
| "global_step": 150, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0022185246810870773, | |
| "grad_norm": 0.4043624699115753, | |
| "learning_rate": 1e-05, | |
| "loss": 0.0218, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.0022185246810870773, | |
| "eval_loss": 0.06729624420404434, | |
| "eval_runtime": 61.4063, | |
| "eval_samples_per_second": 12.36, | |
| "eval_steps_per_second": 3.094, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.004437049362174155, | |
| "grad_norm": 0.5323770642280579, | |
| "learning_rate": 2e-05, | |
| "loss": 0.0272, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.0066555740432612314, | |
| "grad_norm": 0.39081913232803345, | |
| "learning_rate": 3e-05, | |
| "loss": 0.0224, | |
| "step": 3 | |
| }, | |
| { | |
| "epoch": 0.00887409872434831, | |
| "grad_norm": 0.2355053573846817, | |
| "learning_rate": 4e-05, | |
| "loss": 0.014, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.011092623405435386, | |
| "grad_norm": 0.20560254156589508, | |
| "learning_rate": 5e-05, | |
| "loss": 0.007, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.013311148086522463, | |
| "grad_norm": 0.3132166564464569, | |
| "learning_rate": 6e-05, | |
| "loss": 0.003, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.01552967276760954, | |
| "grad_norm": 0.15627335011959076, | |
| "learning_rate": 7e-05, | |
| "loss": 0.0015, | |
| "step": 7 | |
| }, | |
| { | |
| "epoch": 0.01774819744869662, | |
| "grad_norm": 0.5707840919494629, | |
| "learning_rate": 8e-05, | |
| "loss": 0.0024, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.019966722129783693, | |
| "grad_norm": 2.011200189590454, | |
| "learning_rate": 9e-05, | |
| "loss": 0.0111, | |
| "step": 9 | |
| }, | |
| { | |
| "epoch": 0.022185246810870772, | |
| "grad_norm": 0.16989780962467194, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0006, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.024403771491957847, | |
| "grad_norm": 0.2541530728340149, | |
| "learning_rate": 9.999316524962345e-05, | |
| "loss": 0.0019, | |
| "step": 11 | |
| }, | |
| { | |
| "epoch": 0.026622296173044926, | |
| "grad_norm": 0.03248166665434837, | |
| "learning_rate": 9.997266286704631e-05, | |
| "loss": 0.0004, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 0.028840820854132, | |
| "grad_norm": 0.07936106622219086, | |
| "learning_rate": 9.993849845741524e-05, | |
| "loss": 0.0003, | |
| "step": 13 | |
| }, | |
| { | |
| "epoch": 0.03105934553521908, | |
| "grad_norm": 0.5282761454582214, | |
| "learning_rate": 9.989068136093873e-05, | |
| "loss": 0.0007, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 0.033277870216306155, | |
| "grad_norm": 0.09327798336744308, | |
| "learning_rate": 9.98292246503335e-05, | |
| "loss": 0.0004, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.03549639489739324, | |
| "grad_norm": 0.11142992228269577, | |
| "learning_rate": 9.975414512725057e-05, | |
| "loss": 0.0002, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 0.03771491957848031, | |
| "grad_norm": 0.09578566998243332, | |
| "learning_rate": 9.966546331768191e-05, | |
| "loss": 0.0002, | |
| "step": 17 | |
| }, | |
| { | |
| "epoch": 0.03993344425956739, | |
| "grad_norm": 0.12011338770389557, | |
| "learning_rate": 9.956320346634876e-05, | |
| "loss": 0.0011, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 0.04215196894065446, | |
| "grad_norm": 0.1782669723033905, | |
| "learning_rate": 9.944739353007344e-05, | |
| "loss": 0.0004, | |
| "step": 19 | |
| }, | |
| { | |
| "epoch": 0.044370493621741544, | |
| "grad_norm": 0.014049503020942211, | |
| "learning_rate": 9.931806517013612e-05, | |
| "loss": 0.0001, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.04658901830282862, | |
| "grad_norm": 0.0027807666920125484, | |
| "learning_rate": 9.917525374361912e-05, | |
| "loss": 0.0, | |
| "step": 21 | |
| }, | |
| { | |
| "epoch": 0.048807542983915694, | |
| "grad_norm": 0.7871737480163574, | |
| "learning_rate": 9.901899829374047e-05, | |
| "loss": 0.0008, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 0.051026067665002776, | |
| "grad_norm": 0.0016747310291975737, | |
| "learning_rate": 9.884934153917997e-05, | |
| "loss": 0.0, | |
| "step": 23 | |
| }, | |
| { | |
| "epoch": 0.05324459234608985, | |
| "grad_norm": 0.008830898441374302, | |
| "learning_rate": 9.86663298624003e-05, | |
| "loss": 0.0, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 0.05546311702717693, | |
| "grad_norm": 0.2443714439868927, | |
| "learning_rate": 9.847001329696653e-05, | |
| "loss": 0.001, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.057681641708264, | |
| "grad_norm": 0.03958623483777046, | |
| "learning_rate": 9.826044551386744e-05, | |
| "loss": 0.0001, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 0.059900166389351084, | |
| "grad_norm": 1.1243643760681152, | |
| "learning_rate": 9.803768380684242e-05, | |
| "loss": 0.0021, | |
| "step": 27 | |
| }, | |
| { | |
| "epoch": 0.06211869107043816, | |
| "grad_norm": 0.3090110719203949, | |
| "learning_rate": 9.780178907671789e-05, | |
| "loss": 0.0014, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 0.06433721575152523, | |
| "grad_norm": 0.07031821459531784, | |
| "learning_rate": 9.755282581475769e-05, | |
| "loss": 0.0001, | |
| "step": 29 | |
| }, | |
| { | |
| "epoch": 0.06655574043261231, | |
| "grad_norm": 0.12087582051753998, | |
| "learning_rate": 9.729086208503174e-05, | |
| "loss": 0.0003, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.06877426511369938, | |
| "grad_norm": 0.19821757078170776, | |
| "learning_rate": 9.701596950580806e-05, | |
| "loss": 0.0006, | |
| "step": 31 | |
| }, | |
| { | |
| "epoch": 0.07099278979478647, | |
| "grad_norm": 0.3484724760055542, | |
| "learning_rate": 9.672822322997305e-05, | |
| "loss": 0.0008, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 0.07321131447587355, | |
| "grad_norm": 0.2582204341888428, | |
| "learning_rate": 9.642770192448536e-05, | |
| "loss": 0.0005, | |
| "step": 33 | |
| }, | |
| { | |
| "epoch": 0.07542983915696062, | |
| "grad_norm": 0.2562151551246643, | |
| "learning_rate": 9.611448774886924e-05, | |
| "loss": 0.0006, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 0.0776483638380477, | |
| "grad_norm": 0.3555851876735687, | |
| "learning_rate": 9.578866633275288e-05, | |
| "loss": 0.0003, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.07986688851913477, | |
| "grad_norm": 0.0783495083451271, | |
| "learning_rate": 9.545032675245813e-05, | |
| "loss": 0.0015, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 0.08208541320022185, | |
| "grad_norm": 0.005454628728330135, | |
| "learning_rate": 9.509956150664796e-05, | |
| "loss": 0.0, | |
| "step": 37 | |
| }, | |
| { | |
| "epoch": 0.08430393788130892, | |
| "grad_norm": 1.9236397743225098, | |
| "learning_rate": 9.473646649103818e-05, | |
| "loss": 0.0027, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 0.08652246256239601, | |
| "grad_norm": 0.003536924021318555, | |
| "learning_rate": 9.43611409721806e-05, | |
| "loss": 0.0, | |
| "step": 39 | |
| }, | |
| { | |
| "epoch": 0.08874098724348309, | |
| "grad_norm": 10.864886283874512, | |
| "learning_rate": 9.397368756032445e-05, | |
| "loss": 0.0289, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.09095951192457016, | |
| "grad_norm": 0.004927365109324455, | |
| "learning_rate": 9.357421218136386e-05, | |
| "loss": 0.0001, | |
| "step": 41 | |
| }, | |
| { | |
| "epoch": 0.09317803660565724, | |
| "grad_norm": 0.4415525794029236, | |
| "learning_rate": 9.316282404787871e-05, | |
| "loss": 0.0021, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 0.09539656128674431, | |
| "grad_norm": 0.024674855172634125, | |
| "learning_rate": 9.273963562927695e-05, | |
| "loss": 0.0001, | |
| "step": 43 | |
| }, | |
| { | |
| "epoch": 0.09761508596783139, | |
| "grad_norm": 0.36924272775650024, | |
| "learning_rate": 9.230476262104677e-05, | |
| "loss": 0.0011, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 0.09983361064891846, | |
| "grad_norm": 0.09015581756830215, | |
| "learning_rate": 9.185832391312644e-05, | |
| "loss": 0.0005, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 0.10205213533000555, | |
| "grad_norm": 0.01934026926755905, | |
| "learning_rate": 9.140044155740101e-05, | |
| "loss": 0.0002, | |
| "step": 46 | |
| }, | |
| { | |
| "epoch": 0.10427066001109263, | |
| "grad_norm": 0.004875717218965292, | |
| "learning_rate": 9.093124073433463e-05, | |
| "loss": 0.0001, | |
| "step": 47 | |
| }, | |
| { | |
| "epoch": 0.1064891846921797, | |
| "grad_norm": 0.008644173853099346, | |
| "learning_rate": 9.045084971874738e-05, | |
| "loss": 0.0001, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 0.10870770937326678, | |
| "grad_norm": 0.3154628276824951, | |
| "learning_rate": 8.995939984474624e-05, | |
| "loss": 0.0005, | |
| "step": 49 | |
| }, | |
| { | |
| "epoch": 0.11092623405435385, | |
| "grad_norm": 0.23197360336780548, | |
| "learning_rate": 8.945702546981969e-05, | |
| "loss": 0.0007, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.11092623405435385, | |
| "eval_loss": 0.000480437942314893, | |
| "eval_runtime": 62.4124, | |
| "eval_samples_per_second": 12.161, | |
| "eval_steps_per_second": 3.044, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.11314475873544093, | |
| "grad_norm": 0.12439312040805817, | |
| "learning_rate": 8.894386393810563e-05, | |
| "loss": 0.0013, | |
| "step": 51 | |
| }, | |
| { | |
| "epoch": 0.115363283416528, | |
| "grad_norm": 0.01804165169596672, | |
| "learning_rate": 8.842005554284296e-05, | |
| "loss": 0.0001, | |
| "step": 52 | |
| }, | |
| { | |
| "epoch": 0.11758180809761509, | |
| "grad_norm": 0.11951281130313873, | |
| "learning_rate": 8.788574348801675e-05, | |
| "loss": 0.0005, | |
| "step": 53 | |
| }, | |
| { | |
| "epoch": 0.11980033277870217, | |
| "grad_norm": 0.07723747193813324, | |
| "learning_rate": 8.73410738492077e-05, | |
| "loss": 0.0004, | |
| "step": 54 | |
| }, | |
| { | |
| "epoch": 0.12201885745978924, | |
| "grad_norm": 0.009235666133463383, | |
| "learning_rate": 8.678619553365659e-05, | |
| "loss": 0.0001, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 0.12423738214087632, | |
| "grad_norm": 0.002556801540777087, | |
| "learning_rate": 8.622126023955446e-05, | |
| "loss": 0.0, | |
| "step": 56 | |
| }, | |
| { | |
| "epoch": 0.1264559068219634, | |
| "grad_norm": 0.1781330555677414, | |
| "learning_rate": 8.564642241456986e-05, | |
| "loss": 0.0002, | |
| "step": 57 | |
| }, | |
| { | |
| "epoch": 0.12867443150305047, | |
| "grad_norm": 0.005277496296912432, | |
| "learning_rate": 8.506183921362443e-05, | |
| "loss": 0.0, | |
| "step": 58 | |
| }, | |
| { | |
| "epoch": 0.13089295618413754, | |
| "grad_norm": 0.07657326757907867, | |
| "learning_rate": 8.44676704559283e-05, | |
| "loss": 0.0002, | |
| "step": 59 | |
| }, | |
| { | |
| "epoch": 0.13311148086522462, | |
| "grad_norm": 0.012319284491240978, | |
| "learning_rate": 8.386407858128706e-05, | |
| "loss": 0.0001, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.1353300055463117, | |
| "grad_norm": 0.13288556039333344, | |
| "learning_rate": 8.32512286056924e-05, | |
| "loss": 0.0007, | |
| "step": 61 | |
| }, | |
| { | |
| "epoch": 0.13754853022739877, | |
| "grad_norm": 0.053151004016399384, | |
| "learning_rate": 8.262928807620843e-05, | |
| "loss": 0.0004, | |
| "step": 62 | |
| }, | |
| { | |
| "epoch": 0.13976705490848584, | |
| "grad_norm": 0.1312064528465271, | |
| "learning_rate": 8.199842702516583e-05, | |
| "loss": 0.0008, | |
| "step": 63 | |
| }, | |
| { | |
| "epoch": 0.14198557958957295, | |
| "grad_norm": 0.1711154282093048, | |
| "learning_rate": 8.135881792367686e-05, | |
| "loss": 0.0012, | |
| "step": 64 | |
| }, | |
| { | |
| "epoch": 0.14420410427066002, | |
| "grad_norm": 0.028727808967232704, | |
| "learning_rate": 8.07106356344834e-05, | |
| "loss": 0.0002, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 0.1464226289517471, | |
| "grad_norm": 0.007667217403650284, | |
| "learning_rate": 8.005405736415126e-05, | |
| "loss": 0.0001, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 0.14864115363283417, | |
| "grad_norm": 0.0016746758483350277, | |
| "learning_rate": 7.938926261462366e-05, | |
| "loss": 0.0, | |
| "step": 67 | |
| }, | |
| { | |
| "epoch": 0.15085967831392125, | |
| "grad_norm": 0.0075210388749837875, | |
| "learning_rate": 7.871643313414718e-05, | |
| "loss": 0.0, | |
| "step": 68 | |
| }, | |
| { | |
| "epoch": 0.15307820299500832, | |
| "grad_norm": 0.019149813801050186, | |
| "learning_rate": 7.803575286758364e-05, | |
| "loss": 0.0001, | |
| "step": 69 | |
| }, | |
| { | |
| "epoch": 0.1552967276760954, | |
| "grad_norm": 0.008106847293674946, | |
| "learning_rate": 7.734740790612136e-05, | |
| "loss": 0.0, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.15751525235718247, | |
| "grad_norm": 0.04205169901251793, | |
| "learning_rate": 7.66515864363997e-05, | |
| "loss": 0.0002, | |
| "step": 71 | |
| }, | |
| { | |
| "epoch": 0.15973377703826955, | |
| "grad_norm": 0.03955560550093651, | |
| "learning_rate": 7.594847868906076e-05, | |
| "loss": 0.0, | |
| "step": 72 | |
| }, | |
| { | |
| "epoch": 0.16195230171935662, | |
| "grad_norm": 0.002128523774445057, | |
| "learning_rate": 7.52382768867422e-05, | |
| "loss": 0.0, | |
| "step": 73 | |
| }, | |
| { | |
| "epoch": 0.1641708264004437, | |
| "grad_norm": 0.07566439360380173, | |
| "learning_rate": 7.452117519152542e-05, | |
| "loss": 0.0003, | |
| "step": 74 | |
| }, | |
| { | |
| "epoch": 0.16638935108153077, | |
| "grad_norm": 0.011166649870574474, | |
| "learning_rate": 7.379736965185368e-05, | |
| "loss": 0.0, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.16860787576261785, | |
| "grad_norm": 0.0581900030374527, | |
| "learning_rate": 7.30670581489344e-05, | |
| "loss": 0.0002, | |
| "step": 76 | |
| }, | |
| { | |
| "epoch": 0.17082640044370492, | |
| "grad_norm": 0.21121130883693695, | |
| "learning_rate": 7.233044034264034e-05, | |
| "loss": 0.0014, | |
| "step": 77 | |
| }, | |
| { | |
| "epoch": 0.17304492512479203, | |
| "grad_norm": 0.0013943826779723167, | |
| "learning_rate": 7.158771761692464e-05, | |
| "loss": 0.0, | |
| "step": 78 | |
| }, | |
| { | |
| "epoch": 0.1752634498058791, | |
| "grad_norm": 0.000764008320402354, | |
| "learning_rate": 7.083909302476453e-05, | |
| "loss": 0.0, | |
| "step": 79 | |
| }, | |
| { | |
| "epoch": 0.17748197448696618, | |
| "grad_norm": 0.00407798308879137, | |
| "learning_rate": 7.008477123264848e-05, | |
| "loss": 0.0, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.17970049916805325, | |
| "grad_norm": 0.04198778420686722, | |
| "learning_rate": 6.932495846462261e-05, | |
| "loss": 0.0004, | |
| "step": 81 | |
| }, | |
| { | |
| "epoch": 0.18191902384914033, | |
| "grad_norm": 0.13586126267910004, | |
| "learning_rate": 6.855986244591104e-05, | |
| "loss": 0.0009, | |
| "step": 82 | |
| }, | |
| { | |
| "epoch": 0.1841375485302274, | |
| "grad_norm": 0.0022439719177782536, | |
| "learning_rate": 6.778969234612584e-05, | |
| "loss": 0.0, | |
| "step": 83 | |
| }, | |
| { | |
| "epoch": 0.18635607321131448, | |
| "grad_norm": 0.002078837947919965, | |
| "learning_rate": 6.701465872208216e-05, | |
| "loss": 0.0, | |
| "step": 84 | |
| }, | |
| { | |
| "epoch": 0.18857459789240155, | |
| "grad_norm": 0.0007661273120902479, | |
| "learning_rate": 6.623497346023418e-05, | |
| "loss": 0.0, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 0.19079312257348863, | |
| "grad_norm": 0.9077136516571045, | |
| "learning_rate": 6.545084971874738e-05, | |
| "loss": 0.0009, | |
| "step": 86 | |
| }, | |
| { | |
| "epoch": 0.1930116472545757, | |
| "grad_norm": 0.4077296555042267, | |
| "learning_rate": 6.466250186922325e-05, | |
| "loss": 0.0002, | |
| "step": 87 | |
| }, | |
| { | |
| "epoch": 0.19523017193566278, | |
| "grad_norm": 0.5970966815948486, | |
| "learning_rate": 6.387014543809223e-05, | |
| "loss": 0.0012, | |
| "step": 88 | |
| }, | |
| { | |
| "epoch": 0.19744869661674985, | |
| "grad_norm": 0.015372872352600098, | |
| "learning_rate": 6.307399704769099e-05, | |
| "loss": 0.0, | |
| "step": 89 | |
| }, | |
| { | |
| "epoch": 0.19966722129783693, | |
| "grad_norm": 0.003503235289826989, | |
| "learning_rate": 6.227427435703997e-05, | |
| "loss": 0.0, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.201885745978924, | |
| "grad_norm": 1.7559353113174438, | |
| "learning_rate": 6.147119600233758e-05, | |
| "loss": 0.0009, | |
| "step": 91 | |
| }, | |
| { | |
| "epoch": 0.2041042706600111, | |
| "grad_norm": 0.0038733009714633226, | |
| "learning_rate": 6.066498153718735e-05, | |
| "loss": 0.0001, | |
| "step": 92 | |
| }, | |
| { | |
| "epoch": 0.20632279534109818, | |
| "grad_norm": 0.006768415216356516, | |
| "learning_rate": 5.985585137257401e-05, | |
| "loss": 0.0001, | |
| "step": 93 | |
| }, | |
| { | |
| "epoch": 0.20854132002218526, | |
| "grad_norm": 1.5329684019088745, | |
| "learning_rate": 5.90440267166055e-05, | |
| "loss": 0.0012, | |
| "step": 94 | |
| }, | |
| { | |
| "epoch": 0.21075984470327233, | |
| "grad_norm": 0.6506447792053223, | |
| "learning_rate": 5.8229729514036705e-05, | |
| "loss": 0.0029, | |
| "step": 95 | |
| }, | |
| { | |
| "epoch": 0.2129783693843594, | |
| "grad_norm": 0.03461511433124542, | |
| "learning_rate": 5.74131823855921e-05, | |
| "loss": 0.001, | |
| "step": 96 | |
| }, | |
| { | |
| "epoch": 0.21519689406544648, | |
| "grad_norm": 0.12881682813167572, | |
| "learning_rate": 5.6594608567103456e-05, | |
| "loss": 0.0007, | |
| "step": 97 | |
| }, | |
| { | |
| "epoch": 0.21741541874653356, | |
| "grad_norm": 0.4704674482345581, | |
| "learning_rate": 5.577423184847932e-05, | |
| "loss": 0.0027, | |
| "step": 98 | |
| }, | |
| { | |
| "epoch": 0.21963394342762063, | |
| "grad_norm": 0.3575167953968048, | |
| "learning_rate": 5.495227651252315e-05, | |
| "loss": 0.0019, | |
| "step": 99 | |
| }, | |
| { | |
| "epoch": 0.2218524681087077, | |
| "grad_norm": 0.09095858782529831, | |
| "learning_rate": 5.4128967273616625e-05, | |
| "loss": 0.0003, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.2218524681087077, | |
| "eval_loss": 0.0006046874914318323, | |
| "eval_runtime": 62.4244, | |
| "eval_samples_per_second": 12.159, | |
| "eval_steps_per_second": 3.044, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.22407099278979478, | |
| "grad_norm": 0.098163902759552, | |
| "learning_rate": 5.330452921628497e-05, | |
| "loss": 0.0003, | |
| "step": 101 | |
| }, | |
| { | |
| "epoch": 0.22628951747088186, | |
| "grad_norm": 0.15716908872127533, | |
| "learning_rate": 5.247918773366112e-05, | |
| "loss": 0.0014, | |
| "step": 102 | |
| }, | |
| { | |
| "epoch": 0.22850804215196893, | |
| "grad_norm": 0.08430342376232147, | |
| "learning_rate": 5.165316846586541e-05, | |
| "loss": 0.0007, | |
| "step": 103 | |
| }, | |
| { | |
| "epoch": 0.230726566833056, | |
| "grad_norm": 0.12774738669395447, | |
| "learning_rate": 5.0826697238317935e-05, | |
| "loss": 0.0005, | |
| "step": 104 | |
| }, | |
| { | |
| "epoch": 0.23294509151414308, | |
| "grad_norm": 0.06251008808612823, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0002, | |
| "step": 105 | |
| }, | |
| { | |
| "epoch": 0.23516361619523019, | |
| "grad_norm": 0.03128758817911148, | |
| "learning_rate": 4.917330276168208e-05, | |
| "loss": 0.0001, | |
| "step": 106 | |
| }, | |
| { | |
| "epoch": 0.23738214087631726, | |
| "grad_norm": 0.06550537049770355, | |
| "learning_rate": 4.834683153413459e-05, | |
| "loss": 0.0002, | |
| "step": 107 | |
| }, | |
| { | |
| "epoch": 0.23960066555740434, | |
| "grad_norm": 0.01805949956178665, | |
| "learning_rate": 4.7520812266338885e-05, | |
| "loss": 0.0001, | |
| "step": 108 | |
| }, | |
| { | |
| "epoch": 0.2418191902384914, | |
| "grad_norm": 0.07772446423768997, | |
| "learning_rate": 4.669547078371504e-05, | |
| "loss": 0.0002, | |
| "step": 109 | |
| }, | |
| { | |
| "epoch": 0.24403771491957849, | |
| "grad_norm": 0.0164494551718235, | |
| "learning_rate": 4.5871032726383386e-05, | |
| "loss": 0.0001, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.24625623960066556, | |
| "grad_norm": 0.05892806500196457, | |
| "learning_rate": 4.504772348747687e-05, | |
| "loss": 0.0001, | |
| "step": 111 | |
| }, | |
| { | |
| "epoch": 0.24847476428175264, | |
| "grad_norm": 0.06649111211299896, | |
| "learning_rate": 4.4225768151520694e-05, | |
| "loss": 0.0003, | |
| "step": 112 | |
| }, | |
| { | |
| "epoch": 0.2506932889628397, | |
| "grad_norm": 0.11953019350767136, | |
| "learning_rate": 4.3405391432896555e-05, | |
| "loss": 0.0004, | |
| "step": 113 | |
| }, | |
| { | |
| "epoch": 0.2529118136439268, | |
| "grad_norm": 0.2892134487628937, | |
| "learning_rate": 4.2586817614407895e-05, | |
| "loss": 0.0008, | |
| "step": 114 | |
| }, | |
| { | |
| "epoch": 0.25513033832501386, | |
| "grad_norm": 0.04566960781812668, | |
| "learning_rate": 4.17702704859633e-05, | |
| "loss": 0.0001, | |
| "step": 115 | |
| }, | |
| { | |
| "epoch": 0.25734886300610094, | |
| "grad_norm": 0.011184304021298885, | |
| "learning_rate": 4.095597328339452e-05, | |
| "loss": 0.0001, | |
| "step": 116 | |
| }, | |
| { | |
| "epoch": 0.259567387687188, | |
| "grad_norm": 0.005695666652172804, | |
| "learning_rate": 4.0144148627425993e-05, | |
| "loss": 0.0, | |
| "step": 117 | |
| }, | |
| { | |
| "epoch": 0.2617859123682751, | |
| "grad_norm": 0.0013698926195502281, | |
| "learning_rate": 3.933501846281267e-05, | |
| "loss": 0.0, | |
| "step": 118 | |
| }, | |
| { | |
| "epoch": 0.26400443704936216, | |
| "grad_norm": 0.011563248932361603, | |
| "learning_rate": 3.852880399766243e-05, | |
| "loss": 0.0001, | |
| "step": 119 | |
| }, | |
| { | |
| "epoch": 0.26622296173044924, | |
| "grad_norm": 0.018084948882460594, | |
| "learning_rate": 3.772572564296005e-05, | |
| "loss": 0.0001, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.2684414864115363, | |
| "grad_norm": 0.05606401711702347, | |
| "learning_rate": 3.6926002952309016e-05, | |
| "loss": 0.001, | |
| "step": 121 | |
| }, | |
| { | |
| "epoch": 0.2706600110926234, | |
| "grad_norm": 0.024414898827672005, | |
| "learning_rate": 3.612985456190778e-05, | |
| "loss": 0.0001, | |
| "step": 122 | |
| }, | |
| { | |
| "epoch": 0.27287853577371046, | |
| "grad_norm": 0.03431322053074837, | |
| "learning_rate": 3.533749813077677e-05, | |
| "loss": 0.0001, | |
| "step": 123 | |
| }, | |
| { | |
| "epoch": 0.27509706045479754, | |
| "grad_norm": 0.06524939090013504, | |
| "learning_rate": 3.4549150281252636e-05, | |
| "loss": 0.0001, | |
| "step": 124 | |
| }, | |
| { | |
| "epoch": 0.2773155851358846, | |
| "grad_norm": 0.009228132665157318, | |
| "learning_rate": 3.3765026539765834e-05, | |
| "loss": 0.0, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 0.2795341098169717, | |
| "grad_norm": 0.017082631587982178, | |
| "learning_rate": 3.298534127791785e-05, | |
| "loss": 0.0001, | |
| "step": 126 | |
| }, | |
| { | |
| "epoch": 0.2817526344980588, | |
| "grad_norm": 0.19799353182315826, | |
| "learning_rate": 3.221030765387417e-05, | |
| "loss": 0.0009, | |
| "step": 127 | |
| }, | |
| { | |
| "epoch": 0.2839711591791459, | |
| "grad_norm": 0.04512237384915352, | |
| "learning_rate": 3.144013755408895e-05, | |
| "loss": 0.0001, | |
| "step": 128 | |
| }, | |
| { | |
| "epoch": 0.28618968386023297, | |
| "grad_norm": 0.092744380235672, | |
| "learning_rate": 3.0675041535377405e-05, | |
| "loss": 0.0007, | |
| "step": 129 | |
| }, | |
| { | |
| "epoch": 0.28840820854132004, | |
| "grad_norm": 0.16067329049110413, | |
| "learning_rate": 2.991522876735154e-05, | |
| "loss": 0.0004, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.2906267332224071, | |
| "grad_norm": 0.0007857246673665941, | |
| "learning_rate": 2.916090697523549e-05, | |
| "loss": 0.0, | |
| "step": 131 | |
| }, | |
| { | |
| "epoch": 0.2928452579034942, | |
| "grad_norm": 0.00658170273527503, | |
| "learning_rate": 2.8412282383075363e-05, | |
| "loss": 0.0001, | |
| "step": 132 | |
| }, | |
| { | |
| "epoch": 0.29506378258458127, | |
| "grad_norm": 0.0076082381419837475, | |
| "learning_rate": 2.766955965735968e-05, | |
| "loss": 0.0, | |
| "step": 133 | |
| }, | |
| { | |
| "epoch": 0.29728230726566834, | |
| "grad_norm": 0.0009117680601775646, | |
| "learning_rate": 2.693294185106562e-05, | |
| "loss": 0.0, | |
| "step": 134 | |
| }, | |
| { | |
| "epoch": 0.2995008319467554, | |
| "grad_norm": 0.002750788815319538, | |
| "learning_rate": 2.6202630348146324e-05, | |
| "loss": 0.0, | |
| "step": 135 | |
| }, | |
| { | |
| "epoch": 0.3017193566278425, | |
| "grad_norm": 0.005509167443960905, | |
| "learning_rate": 2.547882480847461e-05, | |
| "loss": 0.0001, | |
| "step": 136 | |
| }, | |
| { | |
| "epoch": 0.30393788130892957, | |
| "grad_norm": 0.012034333311021328, | |
| "learning_rate": 2.476172311325783e-05, | |
| "loss": 0.0001, | |
| "step": 137 | |
| }, | |
| { | |
| "epoch": 0.30615640599001664, | |
| "grad_norm": 0.007208544760942459, | |
| "learning_rate": 2.405152131093926e-05, | |
| "loss": 0.0, | |
| "step": 138 | |
| }, | |
| { | |
| "epoch": 0.3083749306711037, | |
| "grad_norm": 0.004525150638073683, | |
| "learning_rate": 2.3348413563600325e-05, | |
| "loss": 0.0001, | |
| "step": 139 | |
| }, | |
| { | |
| "epoch": 0.3105934553521908, | |
| "grad_norm": 0.17208799719810486, | |
| "learning_rate": 2.2652592093878666e-05, | |
| "loss": 0.0001, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.31281198003327787, | |
| "grad_norm": 0.023878416046500206, | |
| "learning_rate": 2.196424713241637e-05, | |
| "loss": 0.0001, | |
| "step": 141 | |
| }, | |
| { | |
| "epoch": 0.31503050471436495, | |
| "grad_norm": 0.019198762252926826, | |
| "learning_rate": 2.128356686585282e-05, | |
| "loss": 0.0001, | |
| "step": 142 | |
| }, | |
| { | |
| "epoch": 0.317249029395452, | |
| "grad_norm": 0.0008248464437201619, | |
| "learning_rate": 2.061073738537635e-05, | |
| "loss": 0.0, | |
| "step": 143 | |
| }, | |
| { | |
| "epoch": 0.3194675540765391, | |
| "grad_norm": 0.3592078387737274, | |
| "learning_rate": 1.9945942635848748e-05, | |
| "loss": 0.0003, | |
| "step": 144 | |
| }, | |
| { | |
| "epoch": 0.32168607875762617, | |
| "grad_norm": 0.0016451748088002205, | |
| "learning_rate": 1.928936436551661e-05, | |
| "loss": 0.0, | |
| "step": 145 | |
| }, | |
| { | |
| "epoch": 0.32390460343871325, | |
| "grad_norm": 0.004575442988425493, | |
| "learning_rate": 1.8641182076323148e-05, | |
| "loss": 0.0001, | |
| "step": 146 | |
| }, | |
| { | |
| "epoch": 0.3261231281198003, | |
| "grad_norm": 0.03082476370036602, | |
| "learning_rate": 1.800157297483417e-05, | |
| "loss": 0.0002, | |
| "step": 147 | |
| }, | |
| { | |
| "epoch": 0.3283416528008874, | |
| "grad_norm": 0.003721684217453003, | |
| "learning_rate": 1.7370711923791567e-05, | |
| "loss": 0.0, | |
| "step": 148 | |
| }, | |
| { | |
| "epoch": 0.33056017748197447, | |
| "grad_norm": 0.0015315449563786387, | |
| "learning_rate": 1.6748771394307585e-05, | |
| "loss": 0.0, | |
| "step": 149 | |
| }, | |
| { | |
| "epoch": 0.33277870216306155, | |
| "grad_norm": 0.02013554982841015, | |
| "learning_rate": 1.6135921418712956e-05, | |
| "loss": 0.0001, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.33277870216306155, | |
| "eval_loss": 0.00025325187016278505, | |
| "eval_runtime": 62.9957, | |
| "eval_samples_per_second": 12.048, | |
| "eval_steps_per_second": 3.016, | |
| "step": 150 | |
| } | |
| ], | |
| "logging_steps": 1, | |
| "max_steps": 200, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 50, | |
| "stateful_callbacks": { | |
| "EarlyStoppingCallback": { | |
| "args": { | |
| "early_stopping_patience": 5, | |
| "early_stopping_threshold": 0.0 | |
| }, | |
| "attributes": { | |
| "early_stopping_patience_counter": 0 | |
| } | |
| }, | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 2.2967530969418957e+17, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |