| { | |
| "best_metric": 0.9922726154327393, | |
| "best_model_checkpoint": "miner_id_24/checkpoint-150", | |
| "epoch": 0.49019607843137253, | |
| "eval_steps": 50, | |
| "global_step": 150, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0032679738562091504, | |
| "grad_norm": 0.24469198286533356, | |
| "learning_rate": 1e-05, | |
| "loss": 1.4387, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.0032679738562091504, | |
| "eval_loss": 1.4474108219146729, | |
| "eval_runtime": 12.294, | |
| "eval_samples_per_second": 41.89, | |
| "eval_steps_per_second": 10.493, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.006535947712418301, | |
| "grad_norm": 0.2529772222042084, | |
| "learning_rate": 2e-05, | |
| "loss": 1.4012, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.00980392156862745, | |
| "grad_norm": 0.26472678780555725, | |
| "learning_rate": 3e-05, | |
| "loss": 1.3954, | |
| "step": 3 | |
| }, | |
| { | |
| "epoch": 0.013071895424836602, | |
| "grad_norm": 0.2510676681995392, | |
| "learning_rate": 4e-05, | |
| "loss": 1.3832, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.016339869281045753, | |
| "grad_norm": 0.2522083818912506, | |
| "learning_rate": 5e-05, | |
| "loss": 1.4, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.0196078431372549, | |
| "grad_norm": 0.23163741827011108, | |
| "learning_rate": 6e-05, | |
| "loss": 1.383, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.02287581699346405, | |
| "grad_norm": 0.24272429943084717, | |
| "learning_rate": 7e-05, | |
| "loss": 1.3874, | |
| "step": 7 | |
| }, | |
| { | |
| "epoch": 0.026143790849673203, | |
| "grad_norm": 0.22741547226905823, | |
| "learning_rate": 8e-05, | |
| "loss": 1.3295, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.029411764705882353, | |
| "grad_norm": 0.21041825413703918, | |
| "learning_rate": 9e-05, | |
| "loss": 1.316, | |
| "step": 9 | |
| }, | |
| { | |
| "epoch": 0.032679738562091505, | |
| "grad_norm": 0.19392521679401398, | |
| "learning_rate": 0.0001, | |
| "loss": 1.2768, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.03594771241830065, | |
| "grad_norm": 0.20608149468898773, | |
| "learning_rate": 9.999316524962345e-05, | |
| "loss": 1.2882, | |
| "step": 11 | |
| }, | |
| { | |
| "epoch": 0.0392156862745098, | |
| "grad_norm": 0.22623692452907562, | |
| "learning_rate": 9.997266286704631e-05, | |
| "loss": 1.2364, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 0.042483660130718956, | |
| "grad_norm": 0.26282697916030884, | |
| "learning_rate": 9.993849845741524e-05, | |
| "loss": 1.211, | |
| "step": 13 | |
| }, | |
| { | |
| "epoch": 0.0457516339869281, | |
| "grad_norm": 0.270076185464859, | |
| "learning_rate": 9.989068136093873e-05, | |
| "loss": 1.2835, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 0.049019607843137254, | |
| "grad_norm": 0.21397562325000763, | |
| "learning_rate": 9.98292246503335e-05, | |
| "loss": 1.2031, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.05228758169934641, | |
| "grad_norm": 0.1848316192626953, | |
| "learning_rate": 9.975414512725057e-05, | |
| "loss": 1.1504, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 0.05555555555555555, | |
| "grad_norm": 0.16563019156455994, | |
| "learning_rate": 9.966546331768191e-05, | |
| "loss": 1.1845, | |
| "step": 17 | |
| }, | |
| { | |
| "epoch": 0.058823529411764705, | |
| "grad_norm": 0.17262709140777588, | |
| "learning_rate": 9.956320346634876e-05, | |
| "loss": 1.16, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 0.06209150326797386, | |
| "grad_norm": 0.16582638025283813, | |
| "learning_rate": 9.944739353007344e-05, | |
| "loss": 1.1903, | |
| "step": 19 | |
| }, | |
| { | |
| "epoch": 0.06535947712418301, | |
| "grad_norm": 0.15873746573925018, | |
| "learning_rate": 9.931806517013612e-05, | |
| "loss": 1.1438, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.06862745098039216, | |
| "grad_norm": 0.17037732899188995, | |
| "learning_rate": 9.917525374361912e-05, | |
| "loss": 1.1243, | |
| "step": 21 | |
| }, | |
| { | |
| "epoch": 0.0718954248366013, | |
| "grad_norm": 0.16108785569667816, | |
| "learning_rate": 9.901899829374047e-05, | |
| "loss": 1.1485, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 0.07516339869281045, | |
| "grad_norm": 0.1587645560503006, | |
| "learning_rate": 9.884934153917997e-05, | |
| "loss": 1.1659, | |
| "step": 23 | |
| }, | |
| { | |
| "epoch": 0.0784313725490196, | |
| "grad_norm": 0.1613873392343521, | |
| "learning_rate": 9.86663298624003e-05, | |
| "loss": 1.1371, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 0.08169934640522876, | |
| "grad_norm": 0.14722518622875214, | |
| "learning_rate": 9.847001329696653e-05, | |
| "loss": 1.0705, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.08496732026143791, | |
| "grad_norm": 0.14173099398612976, | |
| "learning_rate": 9.826044551386744e-05, | |
| "loss": 1.0695, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 0.08823529411764706, | |
| "grad_norm": 0.1377929449081421, | |
| "learning_rate": 9.803768380684242e-05, | |
| "loss": 1.1012, | |
| "step": 27 | |
| }, | |
| { | |
| "epoch": 0.0915032679738562, | |
| "grad_norm": 0.1335737556219101, | |
| "learning_rate": 9.780178907671789e-05, | |
| "loss": 1.076, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 0.09477124183006536, | |
| "grad_norm": 0.14067432284355164, | |
| "learning_rate": 9.755282581475769e-05, | |
| "loss": 1.0406, | |
| "step": 29 | |
| }, | |
| { | |
| "epoch": 0.09803921568627451, | |
| "grad_norm": 0.1369902640581131, | |
| "learning_rate": 9.729086208503174e-05, | |
| "loss": 1.0285, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.10130718954248366, | |
| "grad_norm": 0.14028145372867584, | |
| "learning_rate": 9.701596950580806e-05, | |
| "loss": 1.0766, | |
| "step": 31 | |
| }, | |
| { | |
| "epoch": 0.10457516339869281, | |
| "grad_norm": 0.1419658660888672, | |
| "learning_rate": 9.672822322997305e-05, | |
| "loss": 1.0549, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 0.10784313725490197, | |
| "grad_norm": 0.1403641402721405, | |
| "learning_rate": 9.642770192448536e-05, | |
| "loss": 1.1028, | |
| "step": 33 | |
| }, | |
| { | |
| "epoch": 0.1111111111111111, | |
| "grad_norm": 0.1402372419834137, | |
| "learning_rate": 9.611448774886924e-05, | |
| "loss": 1.0423, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 0.11437908496732026, | |
| "grad_norm": 0.1409500241279602, | |
| "learning_rate": 9.578866633275288e-05, | |
| "loss": 1.0762, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.11764705882352941, | |
| "grad_norm": 0.14114561676979065, | |
| "learning_rate": 9.545032675245813e-05, | |
| "loss": 1.0346, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 0.12091503267973856, | |
| "grad_norm": 0.12877798080444336, | |
| "learning_rate": 9.509956150664796e-05, | |
| "loss": 1.0531, | |
| "step": 37 | |
| }, | |
| { | |
| "epoch": 0.12418300653594772, | |
| "grad_norm": 0.13515013456344604, | |
| "learning_rate": 9.473646649103818e-05, | |
| "loss": 0.9911, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 0.12745098039215685, | |
| "grad_norm": 0.1338815987110138, | |
| "learning_rate": 9.43611409721806e-05, | |
| "loss": 1.048, | |
| "step": 39 | |
| }, | |
| { | |
| "epoch": 0.13071895424836602, | |
| "grad_norm": 0.1321924328804016, | |
| "learning_rate": 9.397368756032445e-05, | |
| "loss": 1.0316, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.13398692810457516, | |
| "grad_norm": 0.13545972108840942, | |
| "learning_rate": 9.357421218136386e-05, | |
| "loss": 1.0201, | |
| "step": 41 | |
| }, | |
| { | |
| "epoch": 0.13725490196078433, | |
| "grad_norm": 0.1455843299627304, | |
| "learning_rate": 9.316282404787871e-05, | |
| "loss": 1.0314, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 0.14052287581699346, | |
| "grad_norm": 0.14777617156505585, | |
| "learning_rate": 9.273963562927695e-05, | |
| "loss": 1.0812, | |
| "step": 43 | |
| }, | |
| { | |
| "epoch": 0.1437908496732026, | |
| "grad_norm": 0.1442340612411499, | |
| "learning_rate": 9.230476262104677e-05, | |
| "loss": 1.0044, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 0.14705882352941177, | |
| "grad_norm": 0.142796590924263, | |
| "learning_rate": 9.185832391312644e-05, | |
| "loss": 1.0788, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 0.1503267973856209, | |
| "grad_norm": 0.14356712996959686, | |
| "learning_rate": 9.140044155740101e-05, | |
| "loss": 0.9699, | |
| "step": 46 | |
| }, | |
| { | |
| "epoch": 0.15359477124183007, | |
| "grad_norm": 0.15727904438972473, | |
| "learning_rate": 9.093124073433463e-05, | |
| "loss": 1.103, | |
| "step": 47 | |
| }, | |
| { | |
| "epoch": 0.1568627450980392, | |
| "grad_norm": 0.15104670822620392, | |
| "learning_rate": 9.045084971874738e-05, | |
| "loss": 0.991, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 0.16013071895424835, | |
| "grad_norm": 0.1587590128183365, | |
| "learning_rate": 8.995939984474624e-05, | |
| "loss": 0.9936, | |
| "step": 49 | |
| }, | |
| { | |
| "epoch": 0.16339869281045752, | |
| "grad_norm": 0.3971143960952759, | |
| "learning_rate": 8.945702546981969e-05, | |
| "loss": 1.0333, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.16339869281045752, | |
| "eval_loss": 1.043031096458435, | |
| "eval_runtime": 12.3483, | |
| "eval_samples_per_second": 41.706, | |
| "eval_steps_per_second": 10.447, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.16666666666666666, | |
| "grad_norm": 0.1735205352306366, | |
| "learning_rate": 8.894386393810563e-05, | |
| "loss": 1.1695, | |
| "step": 51 | |
| }, | |
| { | |
| "epoch": 0.16993464052287582, | |
| "grad_norm": 0.17092542350292206, | |
| "learning_rate": 8.842005554284296e-05, | |
| "loss": 1.1097, | |
| "step": 52 | |
| }, | |
| { | |
| "epoch": 0.17320261437908496, | |
| "grad_norm": 0.18938027322292328, | |
| "learning_rate": 8.788574348801675e-05, | |
| "loss": 1.1351, | |
| "step": 53 | |
| }, | |
| { | |
| "epoch": 0.17647058823529413, | |
| "grad_norm": 0.13851259648799896, | |
| "learning_rate": 8.73410738492077e-05, | |
| "loss": 1.1063, | |
| "step": 54 | |
| }, | |
| { | |
| "epoch": 0.17973856209150327, | |
| "grad_norm": 0.13026750087738037, | |
| "learning_rate": 8.678619553365659e-05, | |
| "loss": 1.0843, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 0.1830065359477124, | |
| "grad_norm": 0.11888636648654938, | |
| "learning_rate": 8.622126023955446e-05, | |
| "loss": 1.0522, | |
| "step": 56 | |
| }, | |
| { | |
| "epoch": 0.18627450980392157, | |
| "grad_norm": 0.11656311899423599, | |
| "learning_rate": 8.564642241456986e-05, | |
| "loss": 1.0453, | |
| "step": 57 | |
| }, | |
| { | |
| "epoch": 0.1895424836601307, | |
| "grad_norm": 0.12335702031850815, | |
| "learning_rate": 8.506183921362443e-05, | |
| "loss": 1.0676, | |
| "step": 58 | |
| }, | |
| { | |
| "epoch": 0.19281045751633988, | |
| "grad_norm": 0.1326335370540619, | |
| "learning_rate": 8.44676704559283e-05, | |
| "loss": 1.0474, | |
| "step": 59 | |
| }, | |
| { | |
| "epoch": 0.19607843137254902, | |
| "grad_norm": 0.13129277527332306, | |
| "learning_rate": 8.386407858128706e-05, | |
| "loss": 1.0032, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.19934640522875818, | |
| "grad_norm": 0.1514269858598709, | |
| "learning_rate": 8.32512286056924e-05, | |
| "loss": 1.0552, | |
| "step": 61 | |
| }, | |
| { | |
| "epoch": 0.20261437908496732, | |
| "grad_norm": 0.14056868851184845, | |
| "learning_rate": 8.262928807620843e-05, | |
| "loss": 1.0481, | |
| "step": 62 | |
| }, | |
| { | |
| "epoch": 0.20588235294117646, | |
| "grad_norm": 0.13416630029678345, | |
| "learning_rate": 8.199842702516583e-05, | |
| "loss": 1.075, | |
| "step": 63 | |
| }, | |
| { | |
| "epoch": 0.20915032679738563, | |
| "grad_norm": 0.12393052130937576, | |
| "learning_rate": 8.135881792367686e-05, | |
| "loss": 1.0622, | |
| "step": 64 | |
| }, | |
| { | |
| "epoch": 0.21241830065359477, | |
| "grad_norm": 0.12092893570661545, | |
| "learning_rate": 8.07106356344834e-05, | |
| "loss": 1.0363, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 0.21568627450980393, | |
| "grad_norm": 0.12079605460166931, | |
| "learning_rate": 8.005405736415126e-05, | |
| "loss": 0.983, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 0.21895424836601307, | |
| "grad_norm": 0.12823984026908875, | |
| "learning_rate": 7.938926261462366e-05, | |
| "loss": 1.0169, | |
| "step": 67 | |
| }, | |
| { | |
| "epoch": 0.2222222222222222, | |
| "grad_norm": 0.13442108035087585, | |
| "learning_rate": 7.871643313414718e-05, | |
| "loss": 0.9858, | |
| "step": 68 | |
| }, | |
| { | |
| "epoch": 0.22549019607843138, | |
| "grad_norm": 0.12223077565431595, | |
| "learning_rate": 7.803575286758364e-05, | |
| "loss": 0.9908, | |
| "step": 69 | |
| }, | |
| { | |
| "epoch": 0.22875816993464052, | |
| "grad_norm": 0.12562674283981323, | |
| "learning_rate": 7.734740790612136e-05, | |
| "loss": 1.0032, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.23202614379084968, | |
| "grad_norm": 0.1272905319929123, | |
| "learning_rate": 7.66515864363997e-05, | |
| "loss": 0.9899, | |
| "step": 71 | |
| }, | |
| { | |
| "epoch": 0.23529411764705882, | |
| "grad_norm": 0.12599872052669525, | |
| "learning_rate": 7.594847868906076e-05, | |
| "loss": 1.0017, | |
| "step": 72 | |
| }, | |
| { | |
| "epoch": 0.238562091503268, | |
| "grad_norm": 0.12577585875988007, | |
| "learning_rate": 7.52382768867422e-05, | |
| "loss": 1.0255, | |
| "step": 73 | |
| }, | |
| { | |
| "epoch": 0.24183006535947713, | |
| "grad_norm": 0.1293269693851471, | |
| "learning_rate": 7.452117519152542e-05, | |
| "loss": 1.0599, | |
| "step": 74 | |
| }, | |
| { | |
| "epoch": 0.24509803921568626, | |
| "grad_norm": 0.1317208707332611, | |
| "learning_rate": 7.379736965185368e-05, | |
| "loss": 1.0217, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.24836601307189543, | |
| "grad_norm": 0.13466817140579224, | |
| "learning_rate": 7.30670581489344e-05, | |
| "loss": 1.0332, | |
| "step": 76 | |
| }, | |
| { | |
| "epoch": 0.25163398692810457, | |
| "grad_norm": 0.1261163204908371, | |
| "learning_rate": 7.233044034264034e-05, | |
| "loss": 1.0222, | |
| "step": 77 | |
| }, | |
| { | |
| "epoch": 0.2549019607843137, | |
| "grad_norm": 0.13361579179763794, | |
| "learning_rate": 7.158771761692464e-05, | |
| "loss": 1.0189, | |
| "step": 78 | |
| }, | |
| { | |
| "epoch": 0.2581699346405229, | |
| "grad_norm": 0.1298186331987381, | |
| "learning_rate": 7.083909302476453e-05, | |
| "loss": 1.0339, | |
| "step": 79 | |
| }, | |
| { | |
| "epoch": 0.26143790849673204, | |
| "grad_norm": 0.12694045901298523, | |
| "learning_rate": 7.008477123264848e-05, | |
| "loss": 1.0672, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.2647058823529412, | |
| "grad_norm": 0.12448877096176147, | |
| "learning_rate": 6.932495846462261e-05, | |
| "loss": 1.0152, | |
| "step": 81 | |
| }, | |
| { | |
| "epoch": 0.2679738562091503, | |
| "grad_norm": 0.1287432461977005, | |
| "learning_rate": 6.855986244591104e-05, | |
| "loss": 0.978, | |
| "step": 82 | |
| }, | |
| { | |
| "epoch": 0.27124183006535946, | |
| "grad_norm": 0.13010282814502716, | |
| "learning_rate": 6.778969234612584e-05, | |
| "loss": 1.0344, | |
| "step": 83 | |
| }, | |
| { | |
| "epoch": 0.27450980392156865, | |
| "grad_norm": 0.135304257273674, | |
| "learning_rate": 6.701465872208216e-05, | |
| "loss": 0.9755, | |
| "step": 84 | |
| }, | |
| { | |
| "epoch": 0.2777777777777778, | |
| "grad_norm": 0.1396276205778122, | |
| "learning_rate": 6.623497346023418e-05, | |
| "loss": 0.9595, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 0.28104575163398693, | |
| "grad_norm": 0.13439607620239258, | |
| "learning_rate": 6.545084971874738e-05, | |
| "loss": 1.0071, | |
| "step": 86 | |
| }, | |
| { | |
| "epoch": 0.28431372549019607, | |
| "grad_norm": 0.1342010796070099, | |
| "learning_rate": 6.466250186922325e-05, | |
| "loss": 0.9888, | |
| "step": 87 | |
| }, | |
| { | |
| "epoch": 0.2875816993464052, | |
| "grad_norm": 0.13898822665214539, | |
| "learning_rate": 6.387014543809223e-05, | |
| "loss": 1.0095, | |
| "step": 88 | |
| }, | |
| { | |
| "epoch": 0.2908496732026144, | |
| "grad_norm": 0.13906656205654144, | |
| "learning_rate": 6.307399704769099e-05, | |
| "loss": 1.0027, | |
| "step": 89 | |
| }, | |
| { | |
| "epoch": 0.29411764705882354, | |
| "grad_norm": 0.13596709072589874, | |
| "learning_rate": 6.227427435703997e-05, | |
| "loss": 0.9514, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.2973856209150327, | |
| "grad_norm": 0.14506670832633972, | |
| "learning_rate": 6.147119600233758e-05, | |
| "loss": 0.9667, | |
| "step": 91 | |
| }, | |
| { | |
| "epoch": 0.3006535947712418, | |
| "grad_norm": 0.15467151999473572, | |
| "learning_rate": 6.066498153718735e-05, | |
| "loss": 1.0348, | |
| "step": 92 | |
| }, | |
| { | |
| "epoch": 0.30392156862745096, | |
| "grad_norm": 0.1493193507194519, | |
| "learning_rate": 5.985585137257401e-05, | |
| "loss": 0.9577, | |
| "step": 93 | |
| }, | |
| { | |
| "epoch": 0.30718954248366015, | |
| "grad_norm": 0.14980854094028473, | |
| "learning_rate": 5.90440267166055e-05, | |
| "loss": 0.9604, | |
| "step": 94 | |
| }, | |
| { | |
| "epoch": 0.3104575163398693, | |
| "grad_norm": 0.16146551072597504, | |
| "learning_rate": 5.8229729514036705e-05, | |
| "loss": 1.0109, | |
| "step": 95 | |
| }, | |
| { | |
| "epoch": 0.3137254901960784, | |
| "grad_norm": 0.1532970815896988, | |
| "learning_rate": 5.74131823855921e-05, | |
| "loss": 1.0249, | |
| "step": 96 | |
| }, | |
| { | |
| "epoch": 0.31699346405228757, | |
| "grad_norm": 0.1485988050699234, | |
| "learning_rate": 5.6594608567103456e-05, | |
| "loss": 0.9954, | |
| "step": 97 | |
| }, | |
| { | |
| "epoch": 0.3202614379084967, | |
| "grad_norm": 0.1484474092721939, | |
| "learning_rate": 5.577423184847932e-05, | |
| "loss": 0.9798, | |
| "step": 98 | |
| }, | |
| { | |
| "epoch": 0.3235294117647059, | |
| "grad_norm": 0.16410280764102936, | |
| "learning_rate": 5.495227651252315e-05, | |
| "loss": 0.964, | |
| "step": 99 | |
| }, | |
| { | |
| "epoch": 0.32679738562091504, | |
| "grad_norm": 0.278840571641922, | |
| "learning_rate": 5.4128967273616625e-05, | |
| "loss": 0.8767, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.32679738562091504, | |
| "eval_loss": 1.0070159435272217, | |
| "eval_runtime": 12.3082, | |
| "eval_samples_per_second": 41.842, | |
| "eval_steps_per_second": 10.481, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.3300653594771242, | |
| "grad_norm": 0.184713214635849, | |
| "learning_rate": 5.330452921628497e-05, | |
| "loss": 1.1821, | |
| "step": 101 | |
| }, | |
| { | |
| "epoch": 0.3333333333333333, | |
| "grad_norm": 0.16883322596549988, | |
| "learning_rate": 5.247918773366112e-05, | |
| "loss": 1.1243, | |
| "step": 102 | |
| }, | |
| { | |
| "epoch": 0.3366013071895425, | |
| "grad_norm": 0.1671701818704605, | |
| "learning_rate": 5.165316846586541e-05, | |
| "loss": 1.0867, | |
| "step": 103 | |
| }, | |
| { | |
| "epoch": 0.33986928104575165, | |
| "grad_norm": 0.15528738498687744, | |
| "learning_rate": 5.0826697238317935e-05, | |
| "loss": 1.0852, | |
| "step": 104 | |
| }, | |
| { | |
| "epoch": 0.3431372549019608, | |
| "grad_norm": 0.15602558851242065, | |
| "learning_rate": 5e-05, | |
| "loss": 1.0716, | |
| "step": 105 | |
| }, | |
| { | |
| "epoch": 0.3464052287581699, | |
| "grad_norm": 0.13668398559093475, | |
| "learning_rate": 4.917330276168208e-05, | |
| "loss": 1.0536, | |
| "step": 106 | |
| }, | |
| { | |
| "epoch": 0.34967320261437906, | |
| "grad_norm": 0.1294436901807785, | |
| "learning_rate": 4.834683153413459e-05, | |
| "loss": 0.9405, | |
| "step": 107 | |
| }, | |
| { | |
| "epoch": 0.35294117647058826, | |
| "grad_norm": 0.12812989950180054, | |
| "learning_rate": 4.7520812266338885e-05, | |
| "loss": 1.0272, | |
| "step": 108 | |
| }, | |
| { | |
| "epoch": 0.3562091503267974, | |
| "grad_norm": 0.1318708062171936, | |
| "learning_rate": 4.669547078371504e-05, | |
| "loss": 1.0285, | |
| "step": 109 | |
| }, | |
| { | |
| "epoch": 0.35947712418300654, | |
| "grad_norm": 0.1318873167037964, | |
| "learning_rate": 4.5871032726383386e-05, | |
| "loss": 1.0436, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.3627450980392157, | |
| "grad_norm": 0.13608375191688538, | |
| "learning_rate": 4.504772348747687e-05, | |
| "loss": 1.0168, | |
| "step": 111 | |
| }, | |
| { | |
| "epoch": 0.3660130718954248, | |
| "grad_norm": 0.13542938232421875, | |
| "learning_rate": 4.4225768151520694e-05, | |
| "loss": 1.0793, | |
| "step": 112 | |
| }, | |
| { | |
| "epoch": 0.369281045751634, | |
| "grad_norm": 0.14122723042964935, | |
| "learning_rate": 4.3405391432896555e-05, | |
| "loss": 1.0175, | |
| "step": 113 | |
| }, | |
| { | |
| "epoch": 0.37254901960784315, | |
| "grad_norm": 0.13928431272506714, | |
| "learning_rate": 4.2586817614407895e-05, | |
| "loss": 0.9935, | |
| "step": 114 | |
| }, | |
| { | |
| "epoch": 0.3758169934640523, | |
| "grad_norm": 0.14117971062660217, | |
| "learning_rate": 4.17702704859633e-05, | |
| "loss": 0.9638, | |
| "step": 115 | |
| }, | |
| { | |
| "epoch": 0.3790849673202614, | |
| "grad_norm": 0.14064086973667145, | |
| "learning_rate": 4.095597328339452e-05, | |
| "loss": 1.0097, | |
| "step": 116 | |
| }, | |
| { | |
| "epoch": 0.38235294117647056, | |
| "grad_norm": 0.1440376192331314, | |
| "learning_rate": 4.0144148627425993e-05, | |
| "loss": 0.9901, | |
| "step": 117 | |
| }, | |
| { | |
| "epoch": 0.38562091503267976, | |
| "grad_norm": 0.13447478413581848, | |
| "learning_rate": 3.933501846281267e-05, | |
| "loss": 1.0247, | |
| "step": 118 | |
| }, | |
| { | |
| "epoch": 0.3888888888888889, | |
| "grad_norm": 0.1368880420923233, | |
| "learning_rate": 3.852880399766243e-05, | |
| "loss": 1.0189, | |
| "step": 119 | |
| }, | |
| { | |
| "epoch": 0.39215686274509803, | |
| "grad_norm": 0.1322520673274994, | |
| "learning_rate": 3.772572564296005e-05, | |
| "loss": 1.0052, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.3954248366013072, | |
| "grad_norm": 0.13966742157936096, | |
| "learning_rate": 3.6926002952309016e-05, | |
| "loss": 1.042, | |
| "step": 121 | |
| }, | |
| { | |
| "epoch": 0.39869281045751637, | |
| "grad_norm": 0.12975157797336578, | |
| "learning_rate": 3.612985456190778e-05, | |
| "loss": 0.9768, | |
| "step": 122 | |
| }, | |
| { | |
| "epoch": 0.4019607843137255, | |
| "grad_norm": 0.13822442293167114, | |
| "learning_rate": 3.533749813077677e-05, | |
| "loss": 1.0098, | |
| "step": 123 | |
| }, | |
| { | |
| "epoch": 0.40522875816993464, | |
| "grad_norm": 0.1365102380514145, | |
| "learning_rate": 3.4549150281252636e-05, | |
| "loss": 0.9441, | |
| "step": 124 | |
| }, | |
| { | |
| "epoch": 0.4084967320261438, | |
| "grad_norm": 0.13646253943443298, | |
| "learning_rate": 3.3765026539765834e-05, | |
| "loss": 0.9881, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 0.4117647058823529, | |
| "grad_norm": 0.13747969269752502, | |
| "learning_rate": 3.298534127791785e-05, | |
| "loss": 1.056, | |
| "step": 126 | |
| }, | |
| { | |
| "epoch": 0.4150326797385621, | |
| "grad_norm": 0.13940247893333435, | |
| "learning_rate": 3.221030765387417e-05, | |
| "loss": 0.9747, | |
| "step": 127 | |
| }, | |
| { | |
| "epoch": 0.41830065359477125, | |
| "grad_norm": 0.13401050865650177, | |
| "learning_rate": 3.144013755408895e-05, | |
| "loss": 1.0157, | |
| "step": 128 | |
| }, | |
| { | |
| "epoch": 0.4215686274509804, | |
| "grad_norm": 0.13685061037540436, | |
| "learning_rate": 3.0675041535377405e-05, | |
| "loss": 1.0059, | |
| "step": 129 | |
| }, | |
| { | |
| "epoch": 0.42483660130718953, | |
| "grad_norm": 0.13194219768047333, | |
| "learning_rate": 2.991522876735154e-05, | |
| "loss": 0.9569, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.42810457516339867, | |
| "grad_norm": 0.13972924649715424, | |
| "learning_rate": 2.916090697523549e-05, | |
| "loss": 0.9642, | |
| "step": 131 | |
| }, | |
| { | |
| "epoch": 0.43137254901960786, | |
| "grad_norm": 0.13786868751049042, | |
| "learning_rate": 2.8412282383075363e-05, | |
| "loss": 0.9932, | |
| "step": 132 | |
| }, | |
| { | |
| "epoch": 0.434640522875817, | |
| "grad_norm": 0.14051565527915955, | |
| "learning_rate": 2.766955965735968e-05, | |
| "loss": 0.9972, | |
| "step": 133 | |
| }, | |
| { | |
| "epoch": 0.43790849673202614, | |
| "grad_norm": 0.141611710190773, | |
| "learning_rate": 2.693294185106562e-05, | |
| "loss": 1.0209, | |
| "step": 134 | |
| }, | |
| { | |
| "epoch": 0.4411764705882353, | |
| "grad_norm": 0.1423238068819046, | |
| "learning_rate": 2.6202630348146324e-05, | |
| "loss": 0.9516, | |
| "step": 135 | |
| }, | |
| { | |
| "epoch": 0.4444444444444444, | |
| "grad_norm": 0.14845594763755798, | |
| "learning_rate": 2.547882480847461e-05, | |
| "loss": 0.9477, | |
| "step": 136 | |
| }, | |
| { | |
| "epoch": 0.4477124183006536, | |
| "grad_norm": 0.14403949677944183, | |
| "learning_rate": 2.476172311325783e-05, | |
| "loss": 1.0214, | |
| "step": 137 | |
| }, | |
| { | |
| "epoch": 0.45098039215686275, | |
| "grad_norm": 0.14817410707473755, | |
| "learning_rate": 2.405152131093926e-05, | |
| "loss": 0.9928, | |
| "step": 138 | |
| }, | |
| { | |
| "epoch": 0.4542483660130719, | |
| "grad_norm": 0.15064844489097595, | |
| "learning_rate": 2.3348413563600325e-05, | |
| "loss": 1.0206, | |
| "step": 139 | |
| }, | |
| { | |
| "epoch": 0.45751633986928103, | |
| "grad_norm": 0.14492543041706085, | |
| "learning_rate": 2.2652592093878666e-05, | |
| "loss": 0.9473, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.46078431372549017, | |
| "grad_norm": 0.14751087129116058, | |
| "learning_rate": 2.196424713241637e-05, | |
| "loss": 0.8948, | |
| "step": 141 | |
| }, | |
| { | |
| "epoch": 0.46405228758169936, | |
| "grad_norm": 0.14993497729301453, | |
| "learning_rate": 2.128356686585282e-05, | |
| "loss": 0.9734, | |
| "step": 142 | |
| }, | |
| { | |
| "epoch": 0.4673202614379085, | |
| "grad_norm": 0.1499774307012558, | |
| "learning_rate": 2.061073738537635e-05, | |
| "loss": 0.9313, | |
| "step": 143 | |
| }, | |
| { | |
| "epoch": 0.47058823529411764, | |
| "grad_norm": 0.159506157040596, | |
| "learning_rate": 1.9945942635848748e-05, | |
| "loss": 0.9449, | |
| "step": 144 | |
| }, | |
| { | |
| "epoch": 0.4738562091503268, | |
| "grad_norm": 0.15356262028217316, | |
| "learning_rate": 1.928936436551661e-05, | |
| "loss": 0.9332, | |
| "step": 145 | |
| }, | |
| { | |
| "epoch": 0.477124183006536, | |
| "grad_norm": 0.1673261672258377, | |
| "learning_rate": 1.8641182076323148e-05, | |
| "loss": 1.0145, | |
| "step": 146 | |
| }, | |
| { | |
| "epoch": 0.4803921568627451, | |
| "grad_norm": 0.16216711699962616, | |
| "learning_rate": 1.800157297483417e-05, | |
| "loss": 0.9091, | |
| "step": 147 | |
| }, | |
| { | |
| "epoch": 0.48366013071895425, | |
| "grad_norm": 0.17228259146213531, | |
| "learning_rate": 1.7370711923791567e-05, | |
| "loss": 0.9386, | |
| "step": 148 | |
| }, | |
| { | |
| "epoch": 0.4869281045751634, | |
| "grad_norm": 0.1875411570072174, | |
| "learning_rate": 1.6748771394307585e-05, | |
| "loss": 0.9501, | |
| "step": 149 | |
| }, | |
| { | |
| "epoch": 0.49019607843137253, | |
| "grad_norm": 0.31840279698371887, | |
| "learning_rate": 1.6135921418712956e-05, | |
| "loss": 0.8351, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.49019607843137253, | |
| "eval_loss": 0.9922726154327393, | |
| "eval_runtime": 12.3126, | |
| "eval_samples_per_second": 41.827, | |
| "eval_steps_per_second": 10.477, | |
| "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": 3.717010132933018e+16, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |