| { | |
| "best_metric": 1.8019505739212036, | |
| "best_model_checkpoint": "miner_id_24/checkpoint-150", | |
| "epoch": 0.09836065573770492, | |
| "eval_steps": 50, | |
| "global_step": 150, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0006557377049180328, | |
| "grad_norm": 0.6599376201629639, | |
| "learning_rate": 1e-05, | |
| "loss": 1.669, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.0006557377049180328, | |
| "eval_loss": 2.312729835510254, | |
| "eval_runtime": 51.6177, | |
| "eval_samples_per_second": 49.77, | |
| "eval_steps_per_second": 12.457, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.0013114754098360656, | |
| "grad_norm": 0.8067207932472229, | |
| "learning_rate": 2e-05, | |
| "loss": 1.8223, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.0019672131147540984, | |
| "grad_norm": 0.7994766235351562, | |
| "learning_rate": 3e-05, | |
| "loss": 1.7245, | |
| "step": 3 | |
| }, | |
| { | |
| "epoch": 0.002622950819672131, | |
| "grad_norm": 0.8394147753715515, | |
| "learning_rate": 4e-05, | |
| "loss": 1.8729, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.003278688524590164, | |
| "grad_norm": 0.8308488726615906, | |
| "learning_rate": 5e-05, | |
| "loss": 1.6856, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.003934426229508197, | |
| "grad_norm": 0.8317843079566956, | |
| "learning_rate": 6e-05, | |
| "loss": 1.7824, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.0045901639344262295, | |
| "grad_norm": 0.8937143087387085, | |
| "learning_rate": 7e-05, | |
| "loss": 1.7508, | |
| "step": 7 | |
| }, | |
| { | |
| "epoch": 0.005245901639344262, | |
| "grad_norm": 0.8892799615859985, | |
| "learning_rate": 8e-05, | |
| "loss": 1.8603, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.005901639344262295, | |
| "grad_norm": 0.9019492864608765, | |
| "learning_rate": 9e-05, | |
| "loss": 1.83, | |
| "step": 9 | |
| }, | |
| { | |
| "epoch": 0.006557377049180328, | |
| "grad_norm": 0.8635932207107544, | |
| "learning_rate": 0.0001, | |
| "loss": 1.7905, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.007213114754098361, | |
| "grad_norm": 0.9219226241111755, | |
| "learning_rate": 9.999316524962345e-05, | |
| "loss": 1.9243, | |
| "step": 11 | |
| }, | |
| { | |
| "epoch": 0.007868852459016393, | |
| "grad_norm": 0.9507675170898438, | |
| "learning_rate": 9.997266286704631e-05, | |
| "loss": 1.8015, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 0.008524590163934427, | |
| "grad_norm": 0.9975212812423706, | |
| "learning_rate": 9.993849845741524e-05, | |
| "loss": 1.7386, | |
| "step": 13 | |
| }, | |
| { | |
| "epoch": 0.009180327868852459, | |
| "grad_norm": 1.0611342191696167, | |
| "learning_rate": 9.989068136093873e-05, | |
| "loss": 1.9632, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 0.009836065573770493, | |
| "grad_norm": 0.955276370048523, | |
| "learning_rate": 9.98292246503335e-05, | |
| "loss": 1.7806, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.010491803278688525, | |
| "grad_norm": 0.9143955707550049, | |
| "learning_rate": 9.975414512725057e-05, | |
| "loss": 1.7652, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 0.011147540983606558, | |
| "grad_norm": 0.9989897012710571, | |
| "learning_rate": 9.966546331768191e-05, | |
| "loss": 1.8439, | |
| "step": 17 | |
| }, | |
| { | |
| "epoch": 0.01180327868852459, | |
| "grad_norm": 1.0064334869384766, | |
| "learning_rate": 9.956320346634876e-05, | |
| "loss": 1.9204, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 0.012459016393442624, | |
| "grad_norm": 1.037940263748169, | |
| "learning_rate": 9.944739353007344e-05, | |
| "loss": 1.8002, | |
| "step": 19 | |
| }, | |
| { | |
| "epoch": 0.013114754098360656, | |
| "grad_norm": 1.0225900411605835, | |
| "learning_rate": 9.931806517013612e-05, | |
| "loss": 1.7015, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.01377049180327869, | |
| "grad_norm": 1.019464135169983, | |
| "learning_rate": 9.917525374361912e-05, | |
| "loss": 1.7533, | |
| "step": 21 | |
| }, | |
| { | |
| "epoch": 0.014426229508196721, | |
| "grad_norm": 1.0816965103149414, | |
| "learning_rate": 9.901899829374047e-05, | |
| "loss": 1.8629, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 0.015081967213114755, | |
| "grad_norm": 1.070054054260254, | |
| "learning_rate": 9.884934153917997e-05, | |
| "loss": 1.7989, | |
| "step": 23 | |
| }, | |
| { | |
| "epoch": 0.015737704918032787, | |
| "grad_norm": 1.165147304534912, | |
| "learning_rate": 9.86663298624003e-05, | |
| "loss": 1.8759, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 0.01639344262295082, | |
| "grad_norm": 1.12631356716156, | |
| "learning_rate": 9.847001329696653e-05, | |
| "loss": 1.8438, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.017049180327868854, | |
| "grad_norm": 1.2178003787994385, | |
| "learning_rate": 9.826044551386744e-05, | |
| "loss": 1.9643, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 0.017704918032786884, | |
| "grad_norm": 1.4308589696884155, | |
| "learning_rate": 9.803768380684242e-05, | |
| "loss": 1.9305, | |
| "step": 27 | |
| }, | |
| { | |
| "epoch": 0.018360655737704918, | |
| "grad_norm": 1.2899497747421265, | |
| "learning_rate": 9.780178907671789e-05, | |
| "loss": 1.8977, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 0.01901639344262295, | |
| "grad_norm": 1.3106162548065186, | |
| "learning_rate": 9.755282581475769e-05, | |
| "loss": 1.916, | |
| "step": 29 | |
| }, | |
| { | |
| "epoch": 0.019672131147540985, | |
| "grad_norm": 1.5448312759399414, | |
| "learning_rate": 9.729086208503174e-05, | |
| "loss": 1.7298, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.020327868852459016, | |
| "grad_norm": 1.5733848810195923, | |
| "learning_rate": 9.701596950580806e-05, | |
| "loss": 1.9298, | |
| "step": 31 | |
| }, | |
| { | |
| "epoch": 0.02098360655737705, | |
| "grad_norm": 1.4610525369644165, | |
| "learning_rate": 9.672822322997305e-05, | |
| "loss": 1.9263, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 0.021639344262295083, | |
| "grad_norm": 1.6689884662628174, | |
| "learning_rate": 9.642770192448536e-05, | |
| "loss": 1.9427, | |
| "step": 33 | |
| }, | |
| { | |
| "epoch": 0.022295081967213116, | |
| "grad_norm": 1.8268836736679077, | |
| "learning_rate": 9.611448774886924e-05, | |
| "loss": 1.9428, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 0.022950819672131147, | |
| "grad_norm": 1.830525517463684, | |
| "learning_rate": 9.578866633275288e-05, | |
| "loss": 1.9356, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.02360655737704918, | |
| "grad_norm": 1.9572194814682007, | |
| "learning_rate": 9.545032675245813e-05, | |
| "loss": 1.898, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 0.024262295081967214, | |
| "grad_norm": 1.9481030702590942, | |
| "learning_rate": 9.509956150664796e-05, | |
| "loss": 1.9033, | |
| "step": 37 | |
| }, | |
| { | |
| "epoch": 0.024918032786885248, | |
| "grad_norm": 2.2485644817352295, | |
| "learning_rate": 9.473646649103818e-05, | |
| "loss": 2.0574, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 0.025573770491803278, | |
| "grad_norm": 2.1793830394744873, | |
| "learning_rate": 9.43611409721806e-05, | |
| "loss": 2.0391, | |
| "step": 39 | |
| }, | |
| { | |
| "epoch": 0.02622950819672131, | |
| "grad_norm": 2.224167585372925, | |
| "learning_rate": 9.397368756032445e-05, | |
| "loss": 1.9815, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.026885245901639345, | |
| "grad_norm": 2.36245059967041, | |
| "learning_rate": 9.357421218136386e-05, | |
| "loss": 2.0383, | |
| "step": 41 | |
| }, | |
| { | |
| "epoch": 0.02754098360655738, | |
| "grad_norm": 2.706737756729126, | |
| "learning_rate": 9.316282404787871e-05, | |
| "loss": 2.249, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 0.02819672131147541, | |
| "grad_norm": 2.7947113513946533, | |
| "learning_rate": 9.273963562927695e-05, | |
| "loss": 2.2375, | |
| "step": 43 | |
| }, | |
| { | |
| "epoch": 0.028852459016393443, | |
| "grad_norm": 2.787977695465088, | |
| "learning_rate": 9.230476262104677e-05, | |
| "loss": 2.0518, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 0.029508196721311476, | |
| "grad_norm": 3.0825870037078857, | |
| "learning_rate": 9.185832391312644e-05, | |
| "loss": 2.1853, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 0.03016393442622951, | |
| "grad_norm": 3.3854265213012695, | |
| "learning_rate": 9.140044155740101e-05, | |
| "loss": 2.0163, | |
| "step": 46 | |
| }, | |
| { | |
| "epoch": 0.03081967213114754, | |
| "grad_norm": 3.194949150085449, | |
| "learning_rate": 9.093124073433463e-05, | |
| "loss": 2.0168, | |
| "step": 47 | |
| }, | |
| { | |
| "epoch": 0.031475409836065574, | |
| "grad_norm": 4.079406261444092, | |
| "learning_rate": 9.045084971874738e-05, | |
| "loss": 2.2669, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 0.032131147540983604, | |
| "grad_norm": 3.9459073543548584, | |
| "learning_rate": 8.995939984474624e-05, | |
| "loss": 2.2286, | |
| "step": 49 | |
| }, | |
| { | |
| "epoch": 0.03278688524590164, | |
| "grad_norm": 8.104619979858398, | |
| "learning_rate": 8.945702546981969e-05, | |
| "loss": 2.3949, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.03278688524590164, | |
| "eval_loss": 2.0653555393218994, | |
| "eval_runtime": 51.7477, | |
| "eval_samples_per_second": 49.645, | |
| "eval_steps_per_second": 12.426, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.03344262295081967, | |
| "grad_norm": 1.7057262659072876, | |
| "learning_rate": 8.894386393810563e-05, | |
| "loss": 1.8153, | |
| "step": 51 | |
| }, | |
| { | |
| "epoch": 0.03409836065573771, | |
| "grad_norm": 1.52045476436615, | |
| "learning_rate": 8.842005554284296e-05, | |
| "loss": 1.84, | |
| "step": 52 | |
| }, | |
| { | |
| "epoch": 0.03475409836065574, | |
| "grad_norm": 1.2717903852462769, | |
| "learning_rate": 8.788574348801675e-05, | |
| "loss": 1.8603, | |
| "step": 53 | |
| }, | |
| { | |
| "epoch": 0.03540983606557377, | |
| "grad_norm": 0.9368915557861328, | |
| "learning_rate": 8.73410738492077e-05, | |
| "loss": 1.8198, | |
| "step": 54 | |
| }, | |
| { | |
| "epoch": 0.036065573770491806, | |
| "grad_norm": 0.7638617753982544, | |
| "learning_rate": 8.678619553365659e-05, | |
| "loss": 1.8271, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 0.036721311475409836, | |
| "grad_norm": 0.7003270387649536, | |
| "learning_rate": 8.622126023955446e-05, | |
| "loss": 1.7082, | |
| "step": 56 | |
| }, | |
| { | |
| "epoch": 0.037377049180327866, | |
| "grad_norm": 0.6691592335700989, | |
| "learning_rate": 8.564642241456986e-05, | |
| "loss": 1.6327, | |
| "step": 57 | |
| }, | |
| { | |
| "epoch": 0.0380327868852459, | |
| "grad_norm": 0.6497421264648438, | |
| "learning_rate": 8.506183921362443e-05, | |
| "loss": 1.6789, | |
| "step": 58 | |
| }, | |
| { | |
| "epoch": 0.038688524590163934, | |
| "grad_norm": 0.6695832014083862, | |
| "learning_rate": 8.44676704559283e-05, | |
| "loss": 1.7499, | |
| "step": 59 | |
| }, | |
| { | |
| "epoch": 0.03934426229508197, | |
| "grad_norm": 0.6742534041404724, | |
| "learning_rate": 8.386407858128706e-05, | |
| "loss": 1.7251, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.04, | |
| "grad_norm": 0.6942402124404907, | |
| "learning_rate": 8.32512286056924e-05, | |
| "loss": 1.7845, | |
| "step": 61 | |
| }, | |
| { | |
| "epoch": 0.04065573770491803, | |
| "grad_norm": 0.6851775050163269, | |
| "learning_rate": 8.262928807620843e-05, | |
| "loss": 1.7387, | |
| "step": 62 | |
| }, | |
| { | |
| "epoch": 0.04131147540983607, | |
| "grad_norm": 0.7770962715148926, | |
| "learning_rate": 8.199842702516583e-05, | |
| "loss": 1.8634, | |
| "step": 63 | |
| }, | |
| { | |
| "epoch": 0.0419672131147541, | |
| "grad_norm": 0.6991023421287537, | |
| "learning_rate": 8.135881792367686e-05, | |
| "loss": 1.7691, | |
| "step": 64 | |
| }, | |
| { | |
| "epoch": 0.04262295081967213, | |
| "grad_norm": 0.7185125350952148, | |
| "learning_rate": 8.07106356344834e-05, | |
| "loss": 1.6845, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 0.043278688524590166, | |
| "grad_norm": 0.7190847992897034, | |
| "learning_rate": 8.005405736415126e-05, | |
| "loss": 1.6519, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 0.043934426229508196, | |
| "grad_norm": 0.7488616704940796, | |
| "learning_rate": 7.938926261462366e-05, | |
| "loss": 1.7423, | |
| "step": 67 | |
| }, | |
| { | |
| "epoch": 0.04459016393442623, | |
| "grad_norm": 0.7755244970321655, | |
| "learning_rate": 7.871643313414718e-05, | |
| "loss": 1.8318, | |
| "step": 68 | |
| }, | |
| { | |
| "epoch": 0.04524590163934426, | |
| "grad_norm": 0.8398361802101135, | |
| "learning_rate": 7.803575286758364e-05, | |
| "loss": 1.7331, | |
| "step": 69 | |
| }, | |
| { | |
| "epoch": 0.04590163934426229, | |
| "grad_norm": 0.8223381042480469, | |
| "learning_rate": 7.734740790612136e-05, | |
| "loss": 1.7701, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.04655737704918033, | |
| "grad_norm": 0.8144294023513794, | |
| "learning_rate": 7.66515864363997e-05, | |
| "loss": 1.8039, | |
| "step": 71 | |
| }, | |
| { | |
| "epoch": 0.04721311475409836, | |
| "grad_norm": 0.884468138217926, | |
| "learning_rate": 7.594847868906076e-05, | |
| "loss": 1.8268, | |
| "step": 72 | |
| }, | |
| { | |
| "epoch": 0.04786885245901639, | |
| "grad_norm": 0.9266506433486938, | |
| "learning_rate": 7.52382768867422e-05, | |
| "loss": 1.8342, | |
| "step": 73 | |
| }, | |
| { | |
| "epoch": 0.04852459016393443, | |
| "grad_norm": 0.9367057085037231, | |
| "learning_rate": 7.452117519152542e-05, | |
| "loss": 1.7206, | |
| "step": 74 | |
| }, | |
| { | |
| "epoch": 0.04918032786885246, | |
| "grad_norm": 1.0088722705841064, | |
| "learning_rate": 7.379736965185368e-05, | |
| "loss": 1.7369, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.049836065573770495, | |
| "grad_norm": 1.0950847864151, | |
| "learning_rate": 7.30670581489344e-05, | |
| "loss": 1.8555, | |
| "step": 76 | |
| }, | |
| { | |
| "epoch": 0.050491803278688525, | |
| "grad_norm": 1.1527807712554932, | |
| "learning_rate": 7.233044034264034e-05, | |
| "loss": 1.8371, | |
| "step": 77 | |
| }, | |
| { | |
| "epoch": 0.051147540983606556, | |
| "grad_norm": 1.2773011922836304, | |
| "learning_rate": 7.158771761692464e-05, | |
| "loss": 1.9268, | |
| "step": 78 | |
| }, | |
| { | |
| "epoch": 0.05180327868852459, | |
| "grad_norm": 1.1889195442199707, | |
| "learning_rate": 7.083909302476453e-05, | |
| "loss": 1.7625, | |
| "step": 79 | |
| }, | |
| { | |
| "epoch": 0.05245901639344262, | |
| "grad_norm": 1.2905858755111694, | |
| "learning_rate": 7.008477123264848e-05, | |
| "loss": 1.68, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.05311475409836065, | |
| "grad_norm": 1.3566590547561646, | |
| "learning_rate": 6.932495846462261e-05, | |
| "loss": 1.7321, | |
| "step": 81 | |
| }, | |
| { | |
| "epoch": 0.05377049180327869, | |
| "grad_norm": 1.4779900312423706, | |
| "learning_rate": 6.855986244591104e-05, | |
| "loss": 1.9474, | |
| "step": 82 | |
| }, | |
| { | |
| "epoch": 0.05442622950819672, | |
| "grad_norm": 1.6690576076507568, | |
| "learning_rate": 6.778969234612584e-05, | |
| "loss": 1.9548, | |
| "step": 83 | |
| }, | |
| { | |
| "epoch": 0.05508196721311476, | |
| "grad_norm": 1.5478459596633911, | |
| "learning_rate": 6.701465872208216e-05, | |
| "loss": 1.7377, | |
| "step": 84 | |
| }, | |
| { | |
| "epoch": 0.05573770491803279, | |
| "grad_norm": 1.7774358987808228, | |
| "learning_rate": 6.623497346023418e-05, | |
| "loss": 1.9184, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 0.05639344262295082, | |
| "grad_norm": 1.7999390363693237, | |
| "learning_rate": 6.545084971874738e-05, | |
| "loss": 2.04, | |
| "step": 86 | |
| }, | |
| { | |
| "epoch": 0.057049180327868855, | |
| "grad_norm": 1.919732689857483, | |
| "learning_rate": 6.466250186922325e-05, | |
| "loss": 2.0443, | |
| "step": 87 | |
| }, | |
| { | |
| "epoch": 0.057704918032786885, | |
| "grad_norm": 1.9718799591064453, | |
| "learning_rate": 6.387014543809223e-05, | |
| "loss": 2.0426, | |
| "step": 88 | |
| }, | |
| { | |
| "epoch": 0.058360655737704915, | |
| "grad_norm": 2.0534744262695312, | |
| "learning_rate": 6.307399704769099e-05, | |
| "loss": 2.0369, | |
| "step": 89 | |
| }, | |
| { | |
| "epoch": 0.05901639344262295, | |
| "grad_norm": 2.051954746246338, | |
| "learning_rate": 6.227427435703997e-05, | |
| "loss": 1.9544, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.05967213114754098, | |
| "grad_norm": 2.204108476638794, | |
| "learning_rate": 6.147119600233758e-05, | |
| "loss": 1.948, | |
| "step": 91 | |
| }, | |
| { | |
| "epoch": 0.06032786885245902, | |
| "grad_norm": 2.153517484664917, | |
| "learning_rate": 6.066498153718735e-05, | |
| "loss": 2.1349, | |
| "step": 92 | |
| }, | |
| { | |
| "epoch": 0.06098360655737705, | |
| "grad_norm": 2.321669578552246, | |
| "learning_rate": 5.985585137257401e-05, | |
| "loss": 1.9175, | |
| "step": 93 | |
| }, | |
| { | |
| "epoch": 0.06163934426229508, | |
| "grad_norm": 2.6689982414245605, | |
| "learning_rate": 5.90440267166055e-05, | |
| "loss": 2.2021, | |
| "step": 94 | |
| }, | |
| { | |
| "epoch": 0.06229508196721312, | |
| "grad_norm": 2.4535908699035645, | |
| "learning_rate": 5.8229729514036705e-05, | |
| "loss": 2.1704, | |
| "step": 95 | |
| }, | |
| { | |
| "epoch": 0.06295081967213115, | |
| "grad_norm": 2.312746524810791, | |
| "learning_rate": 5.74131823855921e-05, | |
| "loss": 1.9127, | |
| "step": 96 | |
| }, | |
| { | |
| "epoch": 0.06360655737704918, | |
| "grad_norm": 3.779184341430664, | |
| "learning_rate": 5.6594608567103456e-05, | |
| "loss": 2.1832, | |
| "step": 97 | |
| }, | |
| { | |
| "epoch": 0.06426229508196721, | |
| "grad_norm": 3.2187552452087402, | |
| "learning_rate": 5.577423184847932e-05, | |
| "loss": 2.1398, | |
| "step": 98 | |
| }, | |
| { | |
| "epoch": 0.06491803278688525, | |
| "grad_norm": 3.261251449584961, | |
| "learning_rate": 5.495227651252315e-05, | |
| "loss": 2.118, | |
| "step": 99 | |
| }, | |
| { | |
| "epoch": 0.06557377049180328, | |
| "grad_norm": 4.284909248352051, | |
| "learning_rate": 5.4128967273616625e-05, | |
| "loss": 2.6812, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.06557377049180328, | |
| "eval_loss": 1.939498782157898, | |
| "eval_runtime": 51.777, | |
| "eval_samples_per_second": 49.617, | |
| "eval_steps_per_second": 12.419, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.06622950819672131, | |
| "grad_norm": 1.073625087738037, | |
| "learning_rate": 5.330452921628497e-05, | |
| "loss": 1.8026, | |
| "step": 101 | |
| }, | |
| { | |
| "epoch": 0.06688524590163934, | |
| "grad_norm": 1.1063910722732544, | |
| "learning_rate": 5.247918773366112e-05, | |
| "loss": 1.9817, | |
| "step": 102 | |
| }, | |
| { | |
| "epoch": 0.06754098360655737, | |
| "grad_norm": 0.8999103307723999, | |
| "learning_rate": 5.165316846586541e-05, | |
| "loss": 1.6759, | |
| "step": 103 | |
| }, | |
| { | |
| "epoch": 0.06819672131147542, | |
| "grad_norm": 0.748047411441803, | |
| "learning_rate": 5.0826697238317935e-05, | |
| "loss": 1.6981, | |
| "step": 104 | |
| }, | |
| { | |
| "epoch": 0.06885245901639345, | |
| "grad_norm": 0.7253121733665466, | |
| "learning_rate": 5e-05, | |
| "loss": 1.6785, | |
| "step": 105 | |
| }, | |
| { | |
| "epoch": 0.06950819672131148, | |
| "grad_norm": 0.6579772233963013, | |
| "learning_rate": 4.917330276168208e-05, | |
| "loss": 1.624, | |
| "step": 106 | |
| }, | |
| { | |
| "epoch": 0.07016393442622951, | |
| "grad_norm": 0.6262582540512085, | |
| "learning_rate": 4.834683153413459e-05, | |
| "loss": 1.625, | |
| "step": 107 | |
| }, | |
| { | |
| "epoch": 0.07081967213114754, | |
| "grad_norm": 0.6289515495300293, | |
| "learning_rate": 4.7520812266338885e-05, | |
| "loss": 1.6347, | |
| "step": 108 | |
| }, | |
| { | |
| "epoch": 0.07147540983606557, | |
| "grad_norm": 0.6527335047721863, | |
| "learning_rate": 4.669547078371504e-05, | |
| "loss": 1.8154, | |
| "step": 109 | |
| }, | |
| { | |
| "epoch": 0.07213114754098361, | |
| "grad_norm": 0.6480495929718018, | |
| "learning_rate": 4.5871032726383386e-05, | |
| "loss": 1.6947, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.07278688524590164, | |
| "grad_norm": 0.6335164904594421, | |
| "learning_rate": 4.504772348747687e-05, | |
| "loss": 1.7071, | |
| "step": 111 | |
| }, | |
| { | |
| "epoch": 0.07344262295081967, | |
| "grad_norm": 0.6182149052619934, | |
| "learning_rate": 4.4225768151520694e-05, | |
| "loss": 1.7596, | |
| "step": 112 | |
| }, | |
| { | |
| "epoch": 0.0740983606557377, | |
| "grad_norm": 0.6145219206809998, | |
| "learning_rate": 4.3405391432896555e-05, | |
| "loss": 1.7287, | |
| "step": 113 | |
| }, | |
| { | |
| "epoch": 0.07475409836065573, | |
| "grad_norm": 0.6250659823417664, | |
| "learning_rate": 4.2586817614407895e-05, | |
| "loss": 1.7327, | |
| "step": 114 | |
| }, | |
| { | |
| "epoch": 0.07540983606557378, | |
| "grad_norm": 0.6553449630737305, | |
| "learning_rate": 4.17702704859633e-05, | |
| "loss": 1.7258, | |
| "step": 115 | |
| }, | |
| { | |
| "epoch": 0.0760655737704918, | |
| "grad_norm": 0.6311564445495605, | |
| "learning_rate": 4.095597328339452e-05, | |
| "loss": 1.7872, | |
| "step": 116 | |
| }, | |
| { | |
| "epoch": 0.07672131147540984, | |
| "grad_norm": 0.6540622115135193, | |
| "learning_rate": 4.0144148627425993e-05, | |
| "loss": 1.5987, | |
| "step": 117 | |
| }, | |
| { | |
| "epoch": 0.07737704918032787, | |
| "grad_norm": 0.7014015913009644, | |
| "learning_rate": 3.933501846281267e-05, | |
| "loss": 1.7665, | |
| "step": 118 | |
| }, | |
| { | |
| "epoch": 0.0780327868852459, | |
| "grad_norm": 0.7268445491790771, | |
| "learning_rate": 3.852880399766243e-05, | |
| "loss": 1.8304, | |
| "step": 119 | |
| }, | |
| { | |
| "epoch": 0.07868852459016394, | |
| "grad_norm": 0.733892023563385, | |
| "learning_rate": 3.772572564296005e-05, | |
| "loss": 1.8077, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.07934426229508197, | |
| "grad_norm": 0.757009744644165, | |
| "learning_rate": 3.6926002952309016e-05, | |
| "loss": 1.8175, | |
| "step": 121 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "grad_norm": 0.7836271524429321, | |
| "learning_rate": 3.612985456190778e-05, | |
| "loss": 1.8589, | |
| "step": 122 | |
| }, | |
| { | |
| "epoch": 0.08065573770491803, | |
| "grad_norm": 0.7717272639274597, | |
| "learning_rate": 3.533749813077677e-05, | |
| "loss": 1.6984, | |
| "step": 123 | |
| }, | |
| { | |
| "epoch": 0.08131147540983606, | |
| "grad_norm": 0.8340203762054443, | |
| "learning_rate": 3.4549150281252636e-05, | |
| "loss": 1.8524, | |
| "step": 124 | |
| }, | |
| { | |
| "epoch": 0.08196721311475409, | |
| "grad_norm": 0.9691823720932007, | |
| "learning_rate": 3.3765026539765834e-05, | |
| "loss": 1.7629, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 0.08262295081967214, | |
| "grad_norm": 0.9481905102729797, | |
| "learning_rate": 3.298534127791785e-05, | |
| "loss": 1.8552, | |
| "step": 126 | |
| }, | |
| { | |
| "epoch": 0.08327868852459017, | |
| "grad_norm": 0.9491969347000122, | |
| "learning_rate": 3.221030765387417e-05, | |
| "loss": 1.7772, | |
| "step": 127 | |
| }, | |
| { | |
| "epoch": 0.0839344262295082, | |
| "grad_norm": 1.0480353832244873, | |
| "learning_rate": 3.144013755408895e-05, | |
| "loss": 1.7154, | |
| "step": 128 | |
| }, | |
| { | |
| "epoch": 0.08459016393442623, | |
| "grad_norm": 1.1652785539627075, | |
| "learning_rate": 3.0675041535377405e-05, | |
| "loss": 1.8561, | |
| "step": 129 | |
| }, | |
| { | |
| "epoch": 0.08524590163934426, | |
| "grad_norm": 1.2230794429779053, | |
| "learning_rate": 2.991522876735154e-05, | |
| "loss": 1.8442, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.0859016393442623, | |
| "grad_norm": 1.2060920000076294, | |
| "learning_rate": 2.916090697523549e-05, | |
| "loss": 1.764, | |
| "step": 131 | |
| }, | |
| { | |
| "epoch": 0.08655737704918033, | |
| "grad_norm": 1.3355447053909302, | |
| "learning_rate": 2.8412282383075363e-05, | |
| "loss": 1.643, | |
| "step": 132 | |
| }, | |
| { | |
| "epoch": 0.08721311475409836, | |
| "grad_norm": 1.3686209917068481, | |
| "learning_rate": 2.766955965735968e-05, | |
| "loss": 1.7939, | |
| "step": 133 | |
| }, | |
| { | |
| "epoch": 0.08786885245901639, | |
| "grad_norm": 1.5673925876617432, | |
| "learning_rate": 2.693294185106562e-05, | |
| "loss": 1.777, | |
| "step": 134 | |
| }, | |
| { | |
| "epoch": 0.08852459016393442, | |
| "grad_norm": 1.5803216695785522, | |
| "learning_rate": 2.6202630348146324e-05, | |
| "loss": 1.9712, | |
| "step": 135 | |
| }, | |
| { | |
| "epoch": 0.08918032786885247, | |
| "grad_norm": 1.4933207035064697, | |
| "learning_rate": 2.547882480847461e-05, | |
| "loss": 1.981, | |
| "step": 136 | |
| }, | |
| { | |
| "epoch": 0.0898360655737705, | |
| "grad_norm": 1.7289679050445557, | |
| "learning_rate": 2.476172311325783e-05, | |
| "loss": 1.9475, | |
| "step": 137 | |
| }, | |
| { | |
| "epoch": 0.09049180327868853, | |
| "grad_norm": 1.685605764389038, | |
| "learning_rate": 2.405152131093926e-05, | |
| "loss": 2.0435, | |
| "step": 138 | |
| }, | |
| { | |
| "epoch": 0.09114754098360656, | |
| "grad_norm": 1.840585470199585, | |
| "learning_rate": 2.3348413563600325e-05, | |
| "loss": 1.9513, | |
| "step": 139 | |
| }, | |
| { | |
| "epoch": 0.09180327868852459, | |
| "grad_norm": 1.848009705543518, | |
| "learning_rate": 2.2652592093878666e-05, | |
| "loss": 1.8782, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.09245901639344262, | |
| "grad_norm": 1.7717748880386353, | |
| "learning_rate": 2.196424713241637e-05, | |
| "loss": 1.6225, | |
| "step": 141 | |
| }, | |
| { | |
| "epoch": 0.09311475409836066, | |
| "grad_norm": 2.065477132797241, | |
| "learning_rate": 2.128356686585282e-05, | |
| "loss": 2.0893, | |
| "step": 142 | |
| }, | |
| { | |
| "epoch": 0.09377049180327869, | |
| "grad_norm": 2.2792515754699707, | |
| "learning_rate": 2.061073738537635e-05, | |
| "loss": 1.4713, | |
| "step": 143 | |
| }, | |
| { | |
| "epoch": 0.09442622950819672, | |
| "grad_norm": 2.1172871589660645, | |
| "learning_rate": 1.9945942635848748e-05, | |
| "loss": 2.0675, | |
| "step": 144 | |
| }, | |
| { | |
| "epoch": 0.09508196721311475, | |
| "grad_norm": 2.2381036281585693, | |
| "learning_rate": 1.928936436551661e-05, | |
| "loss": 1.747, | |
| "step": 145 | |
| }, | |
| { | |
| "epoch": 0.09573770491803278, | |
| "grad_norm": 2.2903246879577637, | |
| "learning_rate": 1.8641182076323148e-05, | |
| "loss": 1.9387, | |
| "step": 146 | |
| }, | |
| { | |
| "epoch": 0.09639344262295083, | |
| "grad_norm": 2.8322818279266357, | |
| "learning_rate": 1.800157297483417e-05, | |
| "loss": 2.3662, | |
| "step": 147 | |
| }, | |
| { | |
| "epoch": 0.09704918032786886, | |
| "grad_norm": 3.2850825786590576, | |
| "learning_rate": 1.7370711923791567e-05, | |
| "loss": 2.2341, | |
| "step": 148 | |
| }, | |
| { | |
| "epoch": 0.09770491803278689, | |
| "grad_norm": 3.280592679977417, | |
| "learning_rate": 1.6748771394307585e-05, | |
| "loss": 2.2783, | |
| "step": 149 | |
| }, | |
| { | |
| "epoch": 0.09836065573770492, | |
| "grad_norm": 4.07296085357666, | |
| "learning_rate": 1.6135921418712956e-05, | |
| "loss": 2.2398, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.09836065573770492, | |
| "eval_loss": 1.8019505739212036, | |
| "eval_runtime": 51.6762, | |
| "eval_samples_per_second": 49.713, | |
| "eval_steps_per_second": 12.443, | |
| "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": 3.2103875179708416e+16, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |