| { | |
| "best_metric": 3.207045078277588, | |
| "best_model_checkpoint": "miner_id_24/checkpoint-150", | |
| "epoch": 0.25510204081632654, | |
| "eval_steps": 50, | |
| "global_step": 150, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0017006802721088435, | |
| "grad_norm": 0.24977613985538483, | |
| "learning_rate": 1e-05, | |
| "loss": 2.8692, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.0017006802721088435, | |
| "eval_loss": 3.276529550552368, | |
| "eval_runtime": 37.2906, | |
| "eval_samples_per_second": 26.548, | |
| "eval_steps_per_second": 6.65, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.003401360544217687, | |
| "grad_norm": 0.24531881511211395, | |
| "learning_rate": 2e-05, | |
| "loss": 3.0437, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.00510204081632653, | |
| "grad_norm": 0.25416088104248047, | |
| "learning_rate": 3e-05, | |
| "loss": 3.0804, | |
| "step": 3 | |
| }, | |
| { | |
| "epoch": 0.006802721088435374, | |
| "grad_norm": 0.24987466633319855, | |
| "learning_rate": 4e-05, | |
| "loss": 3.0, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.008503401360544218, | |
| "grad_norm": 0.2428084909915924, | |
| "learning_rate": 5e-05, | |
| "loss": 2.9788, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.01020408163265306, | |
| "grad_norm": 0.24101322889328003, | |
| "learning_rate": 6e-05, | |
| "loss": 2.8507, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.011904761904761904, | |
| "grad_norm": 0.2492375373840332, | |
| "learning_rate": 7e-05, | |
| "loss": 3.0046, | |
| "step": 7 | |
| }, | |
| { | |
| "epoch": 0.013605442176870748, | |
| "grad_norm": 0.2558836340904236, | |
| "learning_rate": 8e-05, | |
| "loss": 3.0009, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.015306122448979591, | |
| "grad_norm": 0.2605135142803192, | |
| "learning_rate": 9e-05, | |
| "loss": 3.0991, | |
| "step": 9 | |
| }, | |
| { | |
| "epoch": 0.017006802721088437, | |
| "grad_norm": 0.24015140533447266, | |
| "learning_rate": 0.0001, | |
| "loss": 2.8682, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.01870748299319728, | |
| "grad_norm": 0.23625339567661285, | |
| "learning_rate": 9.999316524962345e-05, | |
| "loss": 2.9808, | |
| "step": 11 | |
| }, | |
| { | |
| "epoch": 0.02040816326530612, | |
| "grad_norm": 0.2354581356048584, | |
| "learning_rate": 9.997266286704631e-05, | |
| "loss": 3.0725, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 0.022108843537414966, | |
| "grad_norm": 0.23762401938438416, | |
| "learning_rate": 9.993849845741524e-05, | |
| "loss": 3.191, | |
| "step": 13 | |
| }, | |
| { | |
| "epoch": 0.023809523809523808, | |
| "grad_norm": 0.22220423817634583, | |
| "learning_rate": 9.989068136093873e-05, | |
| "loss": 2.852, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 0.025510204081632654, | |
| "grad_norm": 0.2376808524131775, | |
| "learning_rate": 9.98292246503335e-05, | |
| "loss": 2.9857, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.027210884353741496, | |
| "grad_norm": 0.2597469091415405, | |
| "learning_rate": 9.975414512725057e-05, | |
| "loss": 3.2452, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 0.02891156462585034, | |
| "grad_norm": 0.25804516673088074, | |
| "learning_rate": 9.966546331768191e-05, | |
| "loss": 3.2068, | |
| "step": 17 | |
| }, | |
| { | |
| "epoch": 0.030612244897959183, | |
| "grad_norm": 0.2507534325122833, | |
| "learning_rate": 9.956320346634876e-05, | |
| "loss": 3.0956, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 0.03231292517006803, | |
| "grad_norm": 0.2603234648704529, | |
| "learning_rate": 9.944739353007344e-05, | |
| "loss": 3.1878, | |
| "step": 19 | |
| }, | |
| { | |
| "epoch": 0.034013605442176874, | |
| "grad_norm": 0.2637031078338623, | |
| "learning_rate": 9.931806517013612e-05, | |
| "loss": 3.2389, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.03571428571428571, | |
| "grad_norm": 0.2582753896713257, | |
| "learning_rate": 9.917525374361912e-05, | |
| "loss": 3.0519, | |
| "step": 21 | |
| }, | |
| { | |
| "epoch": 0.03741496598639456, | |
| "grad_norm": 0.2699371576309204, | |
| "learning_rate": 9.901899829374047e-05, | |
| "loss": 3.1635, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 0.0391156462585034, | |
| "grad_norm": 0.2730785310268402, | |
| "learning_rate": 9.884934153917997e-05, | |
| "loss": 3.1656, | |
| "step": 23 | |
| }, | |
| { | |
| "epoch": 0.04081632653061224, | |
| "grad_norm": 0.2866193652153015, | |
| "learning_rate": 9.86663298624003e-05, | |
| "loss": 3.22, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 0.04251700680272109, | |
| "grad_norm": 0.28086575865745544, | |
| "learning_rate": 9.847001329696653e-05, | |
| "loss": 2.948, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.04421768707482993, | |
| "grad_norm": 0.2861677408218384, | |
| "learning_rate": 9.826044551386744e-05, | |
| "loss": 3.2173, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 0.04591836734693878, | |
| "grad_norm": 0.2805924713611603, | |
| "learning_rate": 9.803768380684242e-05, | |
| "loss": 3.0343, | |
| "step": 27 | |
| }, | |
| { | |
| "epoch": 0.047619047619047616, | |
| "grad_norm": 0.3162534832954407, | |
| "learning_rate": 9.780178907671789e-05, | |
| "loss": 3.313, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 0.04931972789115646, | |
| "grad_norm": 0.3136957287788391, | |
| "learning_rate": 9.755282581475769e-05, | |
| "loss": 3.244, | |
| "step": 29 | |
| }, | |
| { | |
| "epoch": 0.05102040816326531, | |
| "grad_norm": 0.31942903995513916, | |
| "learning_rate": 9.729086208503174e-05, | |
| "loss": 3.2875, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.05272108843537415, | |
| "grad_norm": 0.33962124586105347, | |
| "learning_rate": 9.701596950580806e-05, | |
| "loss": 3.3639, | |
| "step": 31 | |
| }, | |
| { | |
| "epoch": 0.05442176870748299, | |
| "grad_norm": 0.3294471502304077, | |
| "learning_rate": 9.672822322997305e-05, | |
| "loss": 3.3009, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 0.05612244897959184, | |
| "grad_norm": 0.3405294418334961, | |
| "learning_rate": 9.642770192448536e-05, | |
| "loss": 3.3337, | |
| "step": 33 | |
| }, | |
| { | |
| "epoch": 0.05782312925170068, | |
| "grad_norm": 0.33866071701049805, | |
| "learning_rate": 9.611448774886924e-05, | |
| "loss": 3.3841, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 0.05952380952380952, | |
| "grad_norm": 0.40185490250587463, | |
| "learning_rate": 9.578866633275288e-05, | |
| "loss": 3.3815, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.061224489795918366, | |
| "grad_norm": 0.351744145154953, | |
| "learning_rate": 9.545032675245813e-05, | |
| "loss": 3.2629, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 0.06292517006802721, | |
| "grad_norm": 0.36549699306488037, | |
| "learning_rate": 9.509956150664796e-05, | |
| "loss": 3.3077, | |
| "step": 37 | |
| }, | |
| { | |
| "epoch": 0.06462585034013606, | |
| "grad_norm": 0.3804144263267517, | |
| "learning_rate": 9.473646649103818e-05, | |
| "loss": 3.2124, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 0.0663265306122449, | |
| "grad_norm": 0.3786104917526245, | |
| "learning_rate": 9.43611409721806e-05, | |
| "loss": 3.3519, | |
| "step": 39 | |
| }, | |
| { | |
| "epoch": 0.06802721088435375, | |
| "grad_norm": 0.3824675381183624, | |
| "learning_rate": 9.397368756032445e-05, | |
| "loss": 3.344, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.06972789115646258, | |
| "grad_norm": 0.40192821621894836, | |
| "learning_rate": 9.357421218136386e-05, | |
| "loss": 3.197, | |
| "step": 41 | |
| }, | |
| { | |
| "epoch": 0.07142857142857142, | |
| "grad_norm": 0.3984317183494568, | |
| "learning_rate": 9.316282404787871e-05, | |
| "loss": 3.1668, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 0.07312925170068027, | |
| "grad_norm": 0.4219413697719574, | |
| "learning_rate": 9.273963562927695e-05, | |
| "loss": 3.3366, | |
| "step": 43 | |
| }, | |
| { | |
| "epoch": 0.07482993197278912, | |
| "grad_norm": 0.5630561709403992, | |
| "learning_rate": 9.230476262104677e-05, | |
| "loss": 3.2803, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 0.07653061224489796, | |
| "grad_norm": 0.5532132387161255, | |
| "learning_rate": 9.185832391312644e-05, | |
| "loss": 3.2889, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 0.0782312925170068, | |
| "grad_norm": 0.46265870332717896, | |
| "learning_rate": 9.140044155740101e-05, | |
| "loss": 3.2862, | |
| "step": 46 | |
| }, | |
| { | |
| "epoch": 0.07993197278911565, | |
| "grad_norm": 0.5097538828849792, | |
| "learning_rate": 9.093124073433463e-05, | |
| "loss": 3.4773, | |
| "step": 47 | |
| }, | |
| { | |
| "epoch": 0.08163265306122448, | |
| "grad_norm": 0.6026643514633179, | |
| "learning_rate": 9.045084971874738e-05, | |
| "loss": 3.3722, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 0.08333333333333333, | |
| "grad_norm": 0.5687397718429565, | |
| "learning_rate": 8.995939984474624e-05, | |
| "loss": 3.5381, | |
| "step": 49 | |
| }, | |
| { | |
| "epoch": 0.08503401360544217, | |
| "grad_norm": 0.596480667591095, | |
| "learning_rate": 8.945702546981969e-05, | |
| "loss": 3.6079, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.08503401360544217, | |
| "eval_loss": 3.224152088165283, | |
| "eval_runtime": 37.6501, | |
| "eval_samples_per_second": 26.295, | |
| "eval_steps_per_second": 6.587, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.08673469387755102, | |
| "grad_norm": 0.2341763973236084, | |
| "learning_rate": 8.894386393810563e-05, | |
| "loss": 2.9171, | |
| "step": 51 | |
| }, | |
| { | |
| "epoch": 0.08843537414965986, | |
| "grad_norm": 0.24223248660564423, | |
| "learning_rate": 8.842005554284296e-05, | |
| "loss": 3.0614, | |
| "step": 52 | |
| }, | |
| { | |
| "epoch": 0.09013605442176871, | |
| "grad_norm": 0.2515990436077118, | |
| "learning_rate": 8.788574348801675e-05, | |
| "loss": 2.9336, | |
| "step": 53 | |
| }, | |
| { | |
| "epoch": 0.09183673469387756, | |
| "grad_norm": 0.21944257616996765, | |
| "learning_rate": 8.73410738492077e-05, | |
| "loss": 2.8982, | |
| "step": 54 | |
| }, | |
| { | |
| "epoch": 0.0935374149659864, | |
| "grad_norm": 0.21947504580020905, | |
| "learning_rate": 8.678619553365659e-05, | |
| "loss": 2.9304, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 0.09523809523809523, | |
| "grad_norm": 0.21740640699863434, | |
| "learning_rate": 8.622126023955446e-05, | |
| "loss": 2.983, | |
| "step": 56 | |
| }, | |
| { | |
| "epoch": 0.09693877551020408, | |
| "grad_norm": 0.21427614986896515, | |
| "learning_rate": 8.564642241456986e-05, | |
| "loss": 3.0669, | |
| "step": 57 | |
| }, | |
| { | |
| "epoch": 0.09863945578231292, | |
| "grad_norm": 0.21444422006607056, | |
| "learning_rate": 8.506183921362443e-05, | |
| "loss": 3.0359, | |
| "step": 58 | |
| }, | |
| { | |
| "epoch": 0.10034013605442177, | |
| "grad_norm": 0.21520090103149414, | |
| "learning_rate": 8.44676704559283e-05, | |
| "loss": 3.1602, | |
| "step": 59 | |
| }, | |
| { | |
| "epoch": 0.10204081632653061, | |
| "grad_norm": 0.19749152660369873, | |
| "learning_rate": 8.386407858128706e-05, | |
| "loss": 3.0467, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.10374149659863946, | |
| "grad_norm": 0.2105053961277008, | |
| "learning_rate": 8.32512286056924e-05, | |
| "loss": 3.0386, | |
| "step": 61 | |
| }, | |
| { | |
| "epoch": 0.1054421768707483, | |
| "grad_norm": 0.20346972346305847, | |
| "learning_rate": 8.262928807620843e-05, | |
| "loss": 3.0421, | |
| "step": 62 | |
| }, | |
| { | |
| "epoch": 0.10714285714285714, | |
| "grad_norm": 0.2053549736738205, | |
| "learning_rate": 8.199842702516583e-05, | |
| "loss": 3.1259, | |
| "step": 63 | |
| }, | |
| { | |
| "epoch": 0.10884353741496598, | |
| "grad_norm": 0.20625939965248108, | |
| "learning_rate": 8.135881792367686e-05, | |
| "loss": 3.1164, | |
| "step": 64 | |
| }, | |
| { | |
| "epoch": 0.11054421768707483, | |
| "grad_norm": 0.19687773287296295, | |
| "learning_rate": 8.07106356344834e-05, | |
| "loss": 3.2026, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 0.11224489795918367, | |
| "grad_norm": 0.2006532996892929, | |
| "learning_rate": 8.005405736415126e-05, | |
| "loss": 2.971, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 0.11394557823129252, | |
| "grad_norm": 0.20761524140834808, | |
| "learning_rate": 7.938926261462366e-05, | |
| "loss": 3.0802, | |
| "step": 67 | |
| }, | |
| { | |
| "epoch": 0.11564625850340136, | |
| "grad_norm": 0.21374569833278656, | |
| "learning_rate": 7.871643313414718e-05, | |
| "loss": 3.0005, | |
| "step": 68 | |
| }, | |
| { | |
| "epoch": 0.11734693877551021, | |
| "grad_norm": 0.21956601738929749, | |
| "learning_rate": 7.803575286758364e-05, | |
| "loss": 3.1888, | |
| "step": 69 | |
| }, | |
| { | |
| "epoch": 0.11904761904761904, | |
| "grad_norm": 0.22200079262256622, | |
| "learning_rate": 7.734740790612136e-05, | |
| "loss": 2.925, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.12074829931972789, | |
| "grad_norm": 0.2446955293416977, | |
| "learning_rate": 7.66515864363997e-05, | |
| "loss": 3.0428, | |
| "step": 71 | |
| }, | |
| { | |
| "epoch": 0.12244897959183673, | |
| "grad_norm": 0.22348366677761078, | |
| "learning_rate": 7.594847868906076e-05, | |
| "loss": 3.2189, | |
| "step": 72 | |
| }, | |
| { | |
| "epoch": 0.12414965986394558, | |
| "grad_norm": 0.2356300950050354, | |
| "learning_rate": 7.52382768867422e-05, | |
| "loss": 3.2202, | |
| "step": 73 | |
| }, | |
| { | |
| "epoch": 0.12585034013605442, | |
| "grad_norm": 0.24415449798107147, | |
| "learning_rate": 7.452117519152542e-05, | |
| "loss": 3.2609, | |
| "step": 74 | |
| }, | |
| { | |
| "epoch": 0.12755102040816327, | |
| "grad_norm": 0.25401386618614197, | |
| "learning_rate": 7.379736965185368e-05, | |
| "loss": 3.1972, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.1292517006802721, | |
| "grad_norm": 0.25254711508750916, | |
| "learning_rate": 7.30670581489344e-05, | |
| "loss": 3.1182, | |
| "step": 76 | |
| }, | |
| { | |
| "epoch": 0.13095238095238096, | |
| "grad_norm": 0.25126931071281433, | |
| "learning_rate": 7.233044034264034e-05, | |
| "loss": 3.0977, | |
| "step": 77 | |
| }, | |
| { | |
| "epoch": 0.1326530612244898, | |
| "grad_norm": 0.26813384890556335, | |
| "learning_rate": 7.158771761692464e-05, | |
| "loss": 3.2177, | |
| "step": 78 | |
| }, | |
| { | |
| "epoch": 0.13435374149659865, | |
| "grad_norm": 0.29080507159233093, | |
| "learning_rate": 7.083909302476453e-05, | |
| "loss": 3.1212, | |
| "step": 79 | |
| }, | |
| { | |
| "epoch": 0.1360544217687075, | |
| "grad_norm": 0.2834646701812744, | |
| "learning_rate": 7.008477123264848e-05, | |
| "loss": 3.1949, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.1377551020408163, | |
| "grad_norm": 0.2948930859565735, | |
| "learning_rate": 6.932495846462261e-05, | |
| "loss": 3.2997, | |
| "step": 81 | |
| }, | |
| { | |
| "epoch": 0.13945578231292516, | |
| "grad_norm": 0.2953135669231415, | |
| "learning_rate": 6.855986244591104e-05, | |
| "loss": 3.2809, | |
| "step": 82 | |
| }, | |
| { | |
| "epoch": 0.141156462585034, | |
| "grad_norm": 0.30540257692337036, | |
| "learning_rate": 6.778969234612584e-05, | |
| "loss": 3.1185, | |
| "step": 83 | |
| }, | |
| { | |
| "epoch": 0.14285714285714285, | |
| "grad_norm": 0.31249865889549255, | |
| "learning_rate": 6.701465872208216e-05, | |
| "loss": 3.2742, | |
| "step": 84 | |
| }, | |
| { | |
| "epoch": 0.1445578231292517, | |
| "grad_norm": 0.33224502205848694, | |
| "learning_rate": 6.623497346023418e-05, | |
| "loss": 3.0751, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 0.14625850340136054, | |
| "grad_norm": 0.3545996844768524, | |
| "learning_rate": 6.545084971874738e-05, | |
| "loss": 3.422, | |
| "step": 86 | |
| }, | |
| { | |
| "epoch": 0.14795918367346939, | |
| "grad_norm": 0.34536752104759216, | |
| "learning_rate": 6.466250186922325e-05, | |
| "loss": 3.24, | |
| "step": 87 | |
| }, | |
| { | |
| "epoch": 0.14965986394557823, | |
| "grad_norm": 0.34888556599617004, | |
| "learning_rate": 6.387014543809223e-05, | |
| "loss": 3.4163, | |
| "step": 88 | |
| }, | |
| { | |
| "epoch": 0.15136054421768708, | |
| "grad_norm": 0.35664883255958557, | |
| "learning_rate": 6.307399704769099e-05, | |
| "loss": 3.1794, | |
| "step": 89 | |
| }, | |
| { | |
| "epoch": 0.15306122448979592, | |
| "grad_norm": 0.3574293255805969, | |
| "learning_rate": 6.227427435703997e-05, | |
| "loss": 3.1946, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.15476190476190477, | |
| "grad_norm": 0.36754804849624634, | |
| "learning_rate": 6.147119600233758e-05, | |
| "loss": 3.3004, | |
| "step": 91 | |
| }, | |
| { | |
| "epoch": 0.1564625850340136, | |
| "grad_norm": 0.38848114013671875, | |
| "learning_rate": 6.066498153718735e-05, | |
| "loss": 3.1616, | |
| "step": 92 | |
| }, | |
| { | |
| "epoch": 0.15816326530612246, | |
| "grad_norm": 0.39301955699920654, | |
| "learning_rate": 5.985585137257401e-05, | |
| "loss": 3.4017, | |
| "step": 93 | |
| }, | |
| { | |
| "epoch": 0.1598639455782313, | |
| "grad_norm": 0.3985503315925598, | |
| "learning_rate": 5.90440267166055e-05, | |
| "loss": 3.1712, | |
| "step": 94 | |
| }, | |
| { | |
| "epoch": 0.16156462585034015, | |
| "grad_norm": 0.43351513147354126, | |
| "learning_rate": 5.8229729514036705e-05, | |
| "loss": 3.2731, | |
| "step": 95 | |
| }, | |
| { | |
| "epoch": 0.16326530612244897, | |
| "grad_norm": 0.4486982226371765, | |
| "learning_rate": 5.74131823855921e-05, | |
| "loss": 3.2217, | |
| "step": 96 | |
| }, | |
| { | |
| "epoch": 0.1649659863945578, | |
| "grad_norm": 0.4667747914791107, | |
| "learning_rate": 5.6594608567103456e-05, | |
| "loss": 3.061, | |
| "step": 97 | |
| }, | |
| { | |
| "epoch": 0.16666666666666666, | |
| "grad_norm": 0.49286648631095886, | |
| "learning_rate": 5.577423184847932e-05, | |
| "loss": 3.364, | |
| "step": 98 | |
| }, | |
| { | |
| "epoch": 0.1683673469387755, | |
| "grad_norm": 0.5249316692352295, | |
| "learning_rate": 5.495227651252315e-05, | |
| "loss": 3.5878, | |
| "step": 99 | |
| }, | |
| { | |
| "epoch": 0.17006802721088435, | |
| "grad_norm": 0.5330090522766113, | |
| "learning_rate": 5.4128967273616625e-05, | |
| "loss": 3.3553, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.17006802721088435, | |
| "eval_loss": 3.2140052318573, | |
| "eval_runtime": 37.7724, | |
| "eval_samples_per_second": 26.21, | |
| "eval_steps_per_second": 6.566, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.1717687074829932, | |
| "grad_norm": 0.19755986332893372, | |
| "learning_rate": 5.330452921628497e-05, | |
| "loss": 2.8932, | |
| "step": 101 | |
| }, | |
| { | |
| "epoch": 0.17346938775510204, | |
| "grad_norm": 0.20857270061969757, | |
| "learning_rate": 5.247918773366112e-05, | |
| "loss": 2.6802, | |
| "step": 102 | |
| }, | |
| { | |
| "epoch": 0.17517006802721088, | |
| "grad_norm": 0.20948052406311035, | |
| "learning_rate": 5.165316846586541e-05, | |
| "loss": 2.9059, | |
| "step": 103 | |
| }, | |
| { | |
| "epoch": 0.17687074829931973, | |
| "grad_norm": 0.20989733934402466, | |
| "learning_rate": 5.0826697238317935e-05, | |
| "loss": 3.0612, | |
| "step": 104 | |
| }, | |
| { | |
| "epoch": 0.17857142857142858, | |
| "grad_norm": 0.2143418937921524, | |
| "learning_rate": 5e-05, | |
| "loss": 3.0187, | |
| "step": 105 | |
| }, | |
| { | |
| "epoch": 0.18027210884353742, | |
| "grad_norm": 0.20674653351306915, | |
| "learning_rate": 4.917330276168208e-05, | |
| "loss": 2.9624, | |
| "step": 106 | |
| }, | |
| { | |
| "epoch": 0.18197278911564627, | |
| "grad_norm": 0.20312808454036713, | |
| "learning_rate": 4.834683153413459e-05, | |
| "loss": 2.814, | |
| "step": 107 | |
| }, | |
| { | |
| "epoch": 0.1836734693877551, | |
| "grad_norm": 0.1989346295595169, | |
| "learning_rate": 4.7520812266338885e-05, | |
| "loss": 2.9916, | |
| "step": 108 | |
| }, | |
| { | |
| "epoch": 0.18537414965986396, | |
| "grad_norm": 0.19133760035037994, | |
| "learning_rate": 4.669547078371504e-05, | |
| "loss": 2.9558, | |
| "step": 109 | |
| }, | |
| { | |
| "epoch": 0.1870748299319728, | |
| "grad_norm": 0.20038875937461853, | |
| "learning_rate": 4.5871032726383386e-05, | |
| "loss": 3.1106, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.18877551020408162, | |
| "grad_norm": 0.19258815050125122, | |
| "learning_rate": 4.504772348747687e-05, | |
| "loss": 3.1052, | |
| "step": 111 | |
| }, | |
| { | |
| "epoch": 0.19047619047619047, | |
| "grad_norm": 0.19183140993118286, | |
| "learning_rate": 4.4225768151520694e-05, | |
| "loss": 3.2307, | |
| "step": 112 | |
| }, | |
| { | |
| "epoch": 0.1921768707482993, | |
| "grad_norm": 0.1974906325340271, | |
| "learning_rate": 4.3405391432896555e-05, | |
| "loss": 3.1112, | |
| "step": 113 | |
| }, | |
| { | |
| "epoch": 0.19387755102040816, | |
| "grad_norm": 0.19184552133083344, | |
| "learning_rate": 4.2586817614407895e-05, | |
| "loss": 2.9175, | |
| "step": 114 | |
| }, | |
| { | |
| "epoch": 0.195578231292517, | |
| "grad_norm": 0.18929830193519592, | |
| "learning_rate": 4.17702704859633e-05, | |
| "loss": 3.0464, | |
| "step": 115 | |
| }, | |
| { | |
| "epoch": 0.19727891156462585, | |
| "grad_norm": 0.19142581522464752, | |
| "learning_rate": 4.095597328339452e-05, | |
| "loss": 3.1871, | |
| "step": 116 | |
| }, | |
| { | |
| "epoch": 0.1989795918367347, | |
| "grad_norm": 0.19400767982006073, | |
| "learning_rate": 4.0144148627425993e-05, | |
| "loss": 3.163, | |
| "step": 117 | |
| }, | |
| { | |
| "epoch": 0.20068027210884354, | |
| "grad_norm": 0.20371149480342865, | |
| "learning_rate": 3.933501846281267e-05, | |
| "loss": 3.0502, | |
| "step": 118 | |
| }, | |
| { | |
| "epoch": 0.20238095238095238, | |
| "grad_norm": 0.20327185094356537, | |
| "learning_rate": 3.852880399766243e-05, | |
| "loss": 3.2262, | |
| "step": 119 | |
| }, | |
| { | |
| "epoch": 0.20408163265306123, | |
| "grad_norm": 0.20867791771888733, | |
| "learning_rate": 3.772572564296005e-05, | |
| "loss": 3.1472, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.20578231292517007, | |
| "grad_norm": 0.2105218470096588, | |
| "learning_rate": 3.6926002952309016e-05, | |
| "loss": 3.0096, | |
| "step": 121 | |
| }, | |
| { | |
| "epoch": 0.20748299319727892, | |
| "grad_norm": 0.22375613451004028, | |
| "learning_rate": 3.612985456190778e-05, | |
| "loss": 3.3099, | |
| "step": 122 | |
| }, | |
| { | |
| "epoch": 0.20918367346938777, | |
| "grad_norm": 0.22983282804489136, | |
| "learning_rate": 3.533749813077677e-05, | |
| "loss": 3.0288, | |
| "step": 123 | |
| }, | |
| { | |
| "epoch": 0.2108843537414966, | |
| "grad_norm": 0.23791778087615967, | |
| "learning_rate": 3.4549150281252636e-05, | |
| "loss": 3.178, | |
| "step": 124 | |
| }, | |
| { | |
| "epoch": 0.21258503401360543, | |
| "grad_norm": 0.230152890086174, | |
| "learning_rate": 3.3765026539765834e-05, | |
| "loss": 3.079, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 0.21428571428571427, | |
| "grad_norm": 0.253482848405838, | |
| "learning_rate": 3.298534127791785e-05, | |
| "loss": 3.1932, | |
| "step": 126 | |
| }, | |
| { | |
| "epoch": 0.21598639455782312, | |
| "grad_norm": 0.2623364329338074, | |
| "learning_rate": 3.221030765387417e-05, | |
| "loss": 3.3153, | |
| "step": 127 | |
| }, | |
| { | |
| "epoch": 0.21768707482993196, | |
| "grad_norm": 0.26620224118232727, | |
| "learning_rate": 3.144013755408895e-05, | |
| "loss": 3.1083, | |
| "step": 128 | |
| }, | |
| { | |
| "epoch": 0.2193877551020408, | |
| "grad_norm": 0.27223002910614014, | |
| "learning_rate": 3.0675041535377405e-05, | |
| "loss": 2.9927, | |
| "step": 129 | |
| }, | |
| { | |
| "epoch": 0.22108843537414966, | |
| "grad_norm": 0.27560675144195557, | |
| "learning_rate": 2.991522876735154e-05, | |
| "loss": 3.1417, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.2227891156462585, | |
| "grad_norm": 0.2801398038864136, | |
| "learning_rate": 2.916090697523549e-05, | |
| "loss": 3.3287, | |
| "step": 131 | |
| }, | |
| { | |
| "epoch": 0.22448979591836735, | |
| "grad_norm": 0.29159224033355713, | |
| "learning_rate": 2.8412282383075363e-05, | |
| "loss": 3.2613, | |
| "step": 132 | |
| }, | |
| { | |
| "epoch": 0.2261904761904762, | |
| "grad_norm": 0.2984563410282135, | |
| "learning_rate": 2.766955965735968e-05, | |
| "loss": 3.2717, | |
| "step": 133 | |
| }, | |
| { | |
| "epoch": 0.22789115646258504, | |
| "grad_norm": 0.310944527387619, | |
| "learning_rate": 2.693294185106562e-05, | |
| "loss": 3.3415, | |
| "step": 134 | |
| }, | |
| { | |
| "epoch": 0.22959183673469388, | |
| "grad_norm": 0.3141229748725891, | |
| "learning_rate": 2.6202630348146324e-05, | |
| "loss": 3.2254, | |
| "step": 135 | |
| }, | |
| { | |
| "epoch": 0.23129251700680273, | |
| "grad_norm": 0.3253171145915985, | |
| "learning_rate": 2.547882480847461e-05, | |
| "loss": 3.236, | |
| "step": 136 | |
| }, | |
| { | |
| "epoch": 0.23299319727891157, | |
| "grad_norm": 0.3436582386493683, | |
| "learning_rate": 2.476172311325783e-05, | |
| "loss": 3.4277, | |
| "step": 137 | |
| }, | |
| { | |
| "epoch": 0.23469387755102042, | |
| "grad_norm": 0.355596125125885, | |
| "learning_rate": 2.405152131093926e-05, | |
| "loss": 3.3484, | |
| "step": 138 | |
| }, | |
| { | |
| "epoch": 0.23639455782312926, | |
| "grad_norm": 0.38178566098213196, | |
| "learning_rate": 2.3348413563600325e-05, | |
| "loss": 3.2193, | |
| "step": 139 | |
| }, | |
| { | |
| "epoch": 0.23809523809523808, | |
| "grad_norm": 0.36499595642089844, | |
| "learning_rate": 2.2652592093878666e-05, | |
| "loss": 3.2933, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.23979591836734693, | |
| "grad_norm": 0.37546104192733765, | |
| "learning_rate": 2.196424713241637e-05, | |
| "loss": 3.3729, | |
| "step": 141 | |
| }, | |
| { | |
| "epoch": 0.24149659863945577, | |
| "grad_norm": 0.3972339332103729, | |
| "learning_rate": 2.128356686585282e-05, | |
| "loss": 3.3285, | |
| "step": 142 | |
| }, | |
| { | |
| "epoch": 0.24319727891156462, | |
| "grad_norm": 0.38844266533851624, | |
| "learning_rate": 2.061073738537635e-05, | |
| "loss": 3.1859, | |
| "step": 143 | |
| }, | |
| { | |
| "epoch": 0.24489795918367346, | |
| "grad_norm": 0.39924362301826477, | |
| "learning_rate": 1.9945942635848748e-05, | |
| "loss": 3.2693, | |
| "step": 144 | |
| }, | |
| { | |
| "epoch": 0.2465986394557823, | |
| "grad_norm": 0.5246092677116394, | |
| "learning_rate": 1.928936436551661e-05, | |
| "loss": 3.1043, | |
| "step": 145 | |
| }, | |
| { | |
| "epoch": 0.24829931972789115, | |
| "grad_norm": 0.4444296360015869, | |
| "learning_rate": 1.8641182076323148e-05, | |
| "loss": 3.2301, | |
| "step": 146 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "grad_norm": 0.4793895184993744, | |
| "learning_rate": 1.800157297483417e-05, | |
| "loss": 3.1293, | |
| "step": 147 | |
| }, | |
| { | |
| "epoch": 0.25170068027210885, | |
| "grad_norm": 0.5179641842842102, | |
| "learning_rate": 1.7370711923791567e-05, | |
| "loss": 3.4447, | |
| "step": 148 | |
| }, | |
| { | |
| "epoch": 0.2534013605442177, | |
| "grad_norm": 0.5075855255126953, | |
| "learning_rate": 1.6748771394307585e-05, | |
| "loss": 3.4237, | |
| "step": 149 | |
| }, | |
| { | |
| "epoch": 0.25510204081632654, | |
| "grad_norm": 0.5595801472663879, | |
| "learning_rate": 1.6135921418712956e-05, | |
| "loss": 3.5705, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.25510204081632654, | |
| "eval_loss": 3.207045078277588, | |
| "eval_runtime": 37.6664, | |
| "eval_samples_per_second": 26.283, | |
| "eval_steps_per_second": 6.584, | |
| "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": 5.334036721434624e+16, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |