| { | |
| "best_metric": 0.017788490280508995, | |
| "best_model_checkpoint": "miner_id_24/checkpoint-150", | |
| "epoch": 0.5808325266214908, | |
| "eval_steps": 50, | |
| "global_step": 150, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.003872216844143272, | |
| "grad_norm": 0.33036765456199646, | |
| "learning_rate": 1e-05, | |
| "loss": 0.297, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.003872216844143272, | |
| "eval_loss": 0.7128530144691467, | |
| "eval_runtime": 30.3619, | |
| "eval_samples_per_second": 14.327, | |
| "eval_steps_per_second": 3.59, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.007744433688286544, | |
| "grad_norm": 0.3850049376487732, | |
| "learning_rate": 2e-05, | |
| "loss": 0.3563, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.011616650532429816, | |
| "grad_norm": 0.37881961464881897, | |
| "learning_rate": 3e-05, | |
| "loss": 0.3861, | |
| "step": 3 | |
| }, | |
| { | |
| "epoch": 0.015488867376573089, | |
| "grad_norm": 0.4100344181060791, | |
| "learning_rate": 4e-05, | |
| "loss": 0.3982, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.01936108422071636, | |
| "grad_norm": 0.40349701046943665, | |
| "learning_rate": 5e-05, | |
| "loss": 0.3863, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.023233301064859633, | |
| "grad_norm": 0.36773568391799927, | |
| "learning_rate": 6e-05, | |
| "loss": 0.3498, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.027105517909002903, | |
| "grad_norm": 0.33091726899147034, | |
| "learning_rate": 7e-05, | |
| "loss": 0.2986, | |
| "step": 7 | |
| }, | |
| { | |
| "epoch": 0.030977734753146177, | |
| "grad_norm": 0.28258058428764343, | |
| "learning_rate": 8e-05, | |
| "loss": 0.2337, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.03484995159728945, | |
| "grad_norm": 0.2912766933441162, | |
| "learning_rate": 9e-05, | |
| "loss": 0.2093, | |
| "step": 9 | |
| }, | |
| { | |
| "epoch": 0.03872216844143272, | |
| "grad_norm": 0.3206344544887543, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1629, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.04259438528557599, | |
| "grad_norm": 0.3448626399040222, | |
| "learning_rate": 9.999316524962345e-05, | |
| "loss": 0.1203, | |
| "step": 11 | |
| }, | |
| { | |
| "epoch": 0.046466602129719266, | |
| "grad_norm": 0.26453205943107605, | |
| "learning_rate": 9.997266286704631e-05, | |
| "loss": 0.0952, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 0.05033881897386254, | |
| "grad_norm": 0.27510973811149597, | |
| "learning_rate": 9.993849845741524e-05, | |
| "loss": 0.0601, | |
| "step": 13 | |
| }, | |
| { | |
| "epoch": 0.05421103581800581, | |
| "grad_norm": 0.19839665293693542, | |
| "learning_rate": 9.989068136093873e-05, | |
| "loss": 0.0684, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 0.05808325266214908, | |
| "grad_norm": 0.31500810384750366, | |
| "learning_rate": 9.98292246503335e-05, | |
| "loss": 0.0835, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.061955469506292354, | |
| "grad_norm": 0.1898542046546936, | |
| "learning_rate": 9.975414512725057e-05, | |
| "loss": 0.0471, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 0.06582768635043562, | |
| "grad_norm": 0.16449055075645447, | |
| "learning_rate": 9.966546331768191e-05, | |
| "loss": 0.0428, | |
| "step": 17 | |
| }, | |
| { | |
| "epoch": 0.0696999031945789, | |
| "grad_norm": 0.17914803326129913, | |
| "learning_rate": 9.956320346634876e-05, | |
| "loss": 0.053, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 0.07357212003872217, | |
| "grad_norm": 0.13265900313854218, | |
| "learning_rate": 9.944739353007344e-05, | |
| "loss": 0.0347, | |
| "step": 19 | |
| }, | |
| { | |
| "epoch": 0.07744433688286544, | |
| "grad_norm": 0.20040208101272583, | |
| "learning_rate": 9.931806517013612e-05, | |
| "loss": 0.038, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.08131655372700872, | |
| "grad_norm": 0.14658556878566742, | |
| "learning_rate": 9.917525374361912e-05, | |
| "loss": 0.0417, | |
| "step": 21 | |
| }, | |
| { | |
| "epoch": 0.08518877057115198, | |
| "grad_norm": 0.10912320762872696, | |
| "learning_rate": 9.901899829374047e-05, | |
| "loss": 0.0336, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 0.08906098741529525, | |
| "grad_norm": 0.10982748866081238, | |
| "learning_rate": 9.884934153917997e-05, | |
| "loss": 0.0278, | |
| "step": 23 | |
| }, | |
| { | |
| "epoch": 0.09293320425943853, | |
| "grad_norm": 0.08335252106189728, | |
| "learning_rate": 9.86663298624003e-05, | |
| "loss": 0.0214, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 0.0968054211035818, | |
| "grad_norm": 0.10161011666059494, | |
| "learning_rate": 9.847001329696653e-05, | |
| "loss": 0.0413, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.10067763794772508, | |
| "grad_norm": 0.09783328324556351, | |
| "learning_rate": 9.826044551386744e-05, | |
| "loss": 0.032, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 0.10454985479186835, | |
| "grad_norm": 0.05879473313689232, | |
| "learning_rate": 9.803768380684242e-05, | |
| "loss": 0.0151, | |
| "step": 27 | |
| }, | |
| { | |
| "epoch": 0.10842207163601161, | |
| "grad_norm": 0.10748795419931412, | |
| "learning_rate": 9.780178907671789e-05, | |
| "loss": 0.0394, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 0.1122942884801549, | |
| "grad_norm": 0.09443118423223495, | |
| "learning_rate": 9.755282581475769e-05, | |
| "loss": 0.0375, | |
| "step": 29 | |
| }, | |
| { | |
| "epoch": 0.11616650532429816, | |
| "grad_norm": 0.08883603662252426, | |
| "learning_rate": 9.729086208503174e-05, | |
| "loss": 0.0325, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.12003872216844143, | |
| "grad_norm": 0.06886650621891022, | |
| "learning_rate": 9.701596950580806e-05, | |
| "loss": 0.0313, | |
| "step": 31 | |
| }, | |
| { | |
| "epoch": 0.12391093901258471, | |
| "grad_norm": 0.07295922935009003, | |
| "learning_rate": 9.672822322997305e-05, | |
| "loss": 0.0276, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 0.12778315585672798, | |
| "grad_norm": 0.06310761719942093, | |
| "learning_rate": 9.642770192448536e-05, | |
| "loss": 0.0234, | |
| "step": 33 | |
| }, | |
| { | |
| "epoch": 0.13165537270087124, | |
| "grad_norm": 0.09747755527496338, | |
| "learning_rate": 9.611448774886924e-05, | |
| "loss": 0.0267, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 0.1355275895450145, | |
| "grad_norm": 0.04697493463754654, | |
| "learning_rate": 9.578866633275288e-05, | |
| "loss": 0.021, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.1393998063891578, | |
| "grad_norm": 0.07634945958852768, | |
| "learning_rate": 9.545032675245813e-05, | |
| "loss": 0.0263, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 0.14327202323330107, | |
| "grad_norm": 0.048437684774398804, | |
| "learning_rate": 9.509956150664796e-05, | |
| "loss": 0.0199, | |
| "step": 37 | |
| }, | |
| { | |
| "epoch": 0.14714424007744434, | |
| "grad_norm": 0.0893416479229927, | |
| "learning_rate": 9.473646649103818e-05, | |
| "loss": 0.0219, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 0.1510164569215876, | |
| "grad_norm": 0.08695591241121292, | |
| "learning_rate": 9.43611409721806e-05, | |
| "loss": 0.0314, | |
| "step": 39 | |
| }, | |
| { | |
| "epoch": 0.15488867376573087, | |
| "grad_norm": 0.0652228444814682, | |
| "learning_rate": 9.397368756032445e-05, | |
| "loss": 0.0223, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.15876089060987417, | |
| "grad_norm": 0.08439888060092926, | |
| "learning_rate": 9.357421218136386e-05, | |
| "loss": 0.0282, | |
| "step": 41 | |
| }, | |
| { | |
| "epoch": 0.16263310745401743, | |
| "grad_norm": 0.05719901621341705, | |
| "learning_rate": 9.316282404787871e-05, | |
| "loss": 0.0226, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 0.1665053242981607, | |
| "grad_norm": 0.05270665884017944, | |
| "learning_rate": 9.273963562927695e-05, | |
| "loss": 0.0232, | |
| "step": 43 | |
| }, | |
| { | |
| "epoch": 0.17037754114230397, | |
| "grad_norm": 0.08705751597881317, | |
| "learning_rate": 9.230476262104677e-05, | |
| "loss": 0.022, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 0.17424975798644723, | |
| "grad_norm": 0.0793544352054596, | |
| "learning_rate": 9.185832391312644e-05, | |
| "loss": 0.0251, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 0.1781219748305905, | |
| "grad_norm": 0.10614676028490067, | |
| "learning_rate": 9.140044155740101e-05, | |
| "loss": 0.0284, | |
| "step": 46 | |
| }, | |
| { | |
| "epoch": 0.1819941916747338, | |
| "grad_norm": 0.05218925327062607, | |
| "learning_rate": 9.093124073433463e-05, | |
| "loss": 0.019, | |
| "step": 47 | |
| }, | |
| { | |
| "epoch": 0.18586640851887706, | |
| "grad_norm": 0.06523724645376205, | |
| "learning_rate": 9.045084971874738e-05, | |
| "loss": 0.025, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 0.18973862536302033, | |
| "grad_norm": 0.05020613223314285, | |
| "learning_rate": 8.995939984474624e-05, | |
| "loss": 0.0168, | |
| "step": 49 | |
| }, | |
| { | |
| "epoch": 0.1936108422071636, | |
| "grad_norm": 0.19285856187343597, | |
| "learning_rate": 8.945702546981969e-05, | |
| "loss": 0.0267, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.1936108422071636, | |
| "eval_loss": 0.022433165460824966, | |
| "eval_runtime": 31.0156, | |
| "eval_samples_per_second": 14.025, | |
| "eval_steps_per_second": 3.514, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.19748305905130686, | |
| "grad_norm": 0.05130339413881302, | |
| "learning_rate": 8.894386393810563e-05, | |
| "loss": 0.0219, | |
| "step": 51 | |
| }, | |
| { | |
| "epoch": 0.20135527589545016, | |
| "grad_norm": 0.054578520357608795, | |
| "learning_rate": 8.842005554284296e-05, | |
| "loss": 0.0223, | |
| "step": 52 | |
| }, | |
| { | |
| "epoch": 0.20522749273959343, | |
| "grad_norm": 0.06454931199550629, | |
| "learning_rate": 8.788574348801675e-05, | |
| "loss": 0.027, | |
| "step": 53 | |
| }, | |
| { | |
| "epoch": 0.2090997095837367, | |
| "grad_norm": 0.06821908801794052, | |
| "learning_rate": 8.73410738492077e-05, | |
| "loss": 0.0162, | |
| "step": 54 | |
| }, | |
| { | |
| "epoch": 0.21297192642787996, | |
| "grad_norm": 0.044611573219299316, | |
| "learning_rate": 8.678619553365659e-05, | |
| "loss": 0.018, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 0.21684414327202323, | |
| "grad_norm": 0.03761984035372734, | |
| "learning_rate": 8.622126023955446e-05, | |
| "loss": 0.0202, | |
| "step": 56 | |
| }, | |
| { | |
| "epoch": 0.2207163601161665, | |
| "grad_norm": 0.0433681458234787, | |
| "learning_rate": 8.564642241456986e-05, | |
| "loss": 0.0154, | |
| "step": 57 | |
| }, | |
| { | |
| "epoch": 0.2245885769603098, | |
| "grad_norm": 0.05336097627878189, | |
| "learning_rate": 8.506183921362443e-05, | |
| "loss": 0.0234, | |
| "step": 58 | |
| }, | |
| { | |
| "epoch": 0.22846079380445306, | |
| "grad_norm": 0.04495273903012276, | |
| "learning_rate": 8.44676704559283e-05, | |
| "loss": 0.0158, | |
| "step": 59 | |
| }, | |
| { | |
| "epoch": 0.23233301064859632, | |
| "grad_norm": 0.05013081058859825, | |
| "learning_rate": 8.386407858128706e-05, | |
| "loss": 0.0212, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.2362052274927396, | |
| "grad_norm": 0.0574067123234272, | |
| "learning_rate": 8.32512286056924e-05, | |
| "loss": 0.0284, | |
| "step": 61 | |
| }, | |
| { | |
| "epoch": 0.24007744433688286, | |
| "grad_norm": 0.03662450611591339, | |
| "learning_rate": 8.262928807620843e-05, | |
| "loss": 0.0222, | |
| "step": 62 | |
| }, | |
| { | |
| "epoch": 0.24394966118102615, | |
| "grad_norm": 0.04296913370490074, | |
| "learning_rate": 8.199842702516583e-05, | |
| "loss": 0.0177, | |
| "step": 63 | |
| }, | |
| { | |
| "epoch": 0.24782187802516942, | |
| "grad_norm": 0.06435459107160568, | |
| "learning_rate": 8.135881792367686e-05, | |
| "loss": 0.0235, | |
| "step": 64 | |
| }, | |
| { | |
| "epoch": 0.25169409486931266, | |
| "grad_norm": 0.04163075610995293, | |
| "learning_rate": 8.07106356344834e-05, | |
| "loss": 0.0226, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 0.25556631171345595, | |
| "grad_norm": 0.03861234709620476, | |
| "learning_rate": 8.005405736415126e-05, | |
| "loss": 0.0233, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 0.25943852855759925, | |
| "grad_norm": 0.039693742990493774, | |
| "learning_rate": 7.938926261462366e-05, | |
| "loss": 0.0186, | |
| "step": 67 | |
| }, | |
| { | |
| "epoch": 0.2633107454017425, | |
| "grad_norm": 0.035804882645606995, | |
| "learning_rate": 7.871643313414718e-05, | |
| "loss": 0.0154, | |
| "step": 68 | |
| }, | |
| { | |
| "epoch": 0.2671829622458858, | |
| "grad_norm": 0.025846365839242935, | |
| "learning_rate": 7.803575286758364e-05, | |
| "loss": 0.0144, | |
| "step": 69 | |
| }, | |
| { | |
| "epoch": 0.271055179090029, | |
| "grad_norm": 0.05968935042619705, | |
| "learning_rate": 7.734740790612136e-05, | |
| "loss": 0.0217, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.2749273959341723, | |
| "grad_norm": 0.06286381930112839, | |
| "learning_rate": 7.66515864363997e-05, | |
| "loss": 0.0209, | |
| "step": 71 | |
| }, | |
| { | |
| "epoch": 0.2787996127783156, | |
| "grad_norm": 0.028835639357566833, | |
| "learning_rate": 7.594847868906076e-05, | |
| "loss": 0.0115, | |
| "step": 72 | |
| }, | |
| { | |
| "epoch": 0.28267182962245885, | |
| "grad_norm": 0.04784088954329491, | |
| "learning_rate": 7.52382768867422e-05, | |
| "loss": 0.0229, | |
| "step": 73 | |
| }, | |
| { | |
| "epoch": 0.28654404646660214, | |
| "grad_norm": 0.05574282631278038, | |
| "learning_rate": 7.452117519152542e-05, | |
| "loss": 0.0256, | |
| "step": 74 | |
| }, | |
| { | |
| "epoch": 0.2904162633107454, | |
| "grad_norm": 0.04102447256445885, | |
| "learning_rate": 7.379736965185368e-05, | |
| "loss": 0.0174, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.2942884801548887, | |
| "grad_norm": 0.0936175063252449, | |
| "learning_rate": 7.30670581489344e-05, | |
| "loss": 0.024, | |
| "step": 76 | |
| }, | |
| { | |
| "epoch": 0.29816069699903197, | |
| "grad_norm": 0.033699460327625275, | |
| "learning_rate": 7.233044034264034e-05, | |
| "loss": 0.0156, | |
| "step": 77 | |
| }, | |
| { | |
| "epoch": 0.3020329138431752, | |
| "grad_norm": 0.06190520524978638, | |
| "learning_rate": 7.158771761692464e-05, | |
| "loss": 0.0245, | |
| "step": 78 | |
| }, | |
| { | |
| "epoch": 0.3059051306873185, | |
| "grad_norm": 0.048285167664289474, | |
| "learning_rate": 7.083909302476453e-05, | |
| "loss": 0.0219, | |
| "step": 79 | |
| }, | |
| { | |
| "epoch": 0.30977734753146174, | |
| "grad_norm": 0.042587533593177795, | |
| "learning_rate": 7.008477123264848e-05, | |
| "loss": 0.0166, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.31364956437560504, | |
| "grad_norm": 0.07332392781972885, | |
| "learning_rate": 6.932495846462261e-05, | |
| "loss": 0.024, | |
| "step": 81 | |
| }, | |
| { | |
| "epoch": 0.31752178121974833, | |
| "grad_norm": 0.12151223421096802, | |
| "learning_rate": 6.855986244591104e-05, | |
| "loss": 0.0241, | |
| "step": 82 | |
| }, | |
| { | |
| "epoch": 0.3213939980638916, | |
| "grad_norm": 0.2832794487476349, | |
| "learning_rate": 6.778969234612584e-05, | |
| "loss": 0.0219, | |
| "step": 83 | |
| }, | |
| { | |
| "epoch": 0.32526621490803487, | |
| "grad_norm": 0.06909225136041641, | |
| "learning_rate": 6.701465872208216e-05, | |
| "loss": 0.0254, | |
| "step": 84 | |
| }, | |
| { | |
| "epoch": 0.3291384317521781, | |
| "grad_norm": 0.3256091773509979, | |
| "learning_rate": 6.623497346023418e-05, | |
| "loss": 0.0196, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 0.3330106485963214, | |
| "grad_norm": 0.05504089593887329, | |
| "learning_rate": 6.545084971874738e-05, | |
| "loss": 0.0228, | |
| "step": 86 | |
| }, | |
| { | |
| "epoch": 0.33688286544046464, | |
| "grad_norm": 0.04585348442196846, | |
| "learning_rate": 6.466250186922325e-05, | |
| "loss": 0.0182, | |
| "step": 87 | |
| }, | |
| { | |
| "epoch": 0.34075508228460794, | |
| "grad_norm": 0.06665895134210587, | |
| "learning_rate": 6.387014543809223e-05, | |
| "loss": 0.0198, | |
| "step": 88 | |
| }, | |
| { | |
| "epoch": 0.34462729912875123, | |
| "grad_norm": 0.047711923718452454, | |
| "learning_rate": 6.307399704769099e-05, | |
| "loss": 0.0206, | |
| "step": 89 | |
| }, | |
| { | |
| "epoch": 0.34849951597289447, | |
| "grad_norm": 0.040118634700775146, | |
| "learning_rate": 6.227427435703997e-05, | |
| "loss": 0.017, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.35237173281703776, | |
| "grad_norm": 0.0628420040011406, | |
| "learning_rate": 6.147119600233758e-05, | |
| "loss": 0.0233, | |
| "step": 91 | |
| }, | |
| { | |
| "epoch": 0.356243949661181, | |
| "grad_norm": 0.05141530930995941, | |
| "learning_rate": 6.066498153718735e-05, | |
| "loss": 0.0246, | |
| "step": 92 | |
| }, | |
| { | |
| "epoch": 0.3601161665053243, | |
| "grad_norm": 0.08067324012517929, | |
| "learning_rate": 5.985585137257401e-05, | |
| "loss": 0.0215, | |
| "step": 93 | |
| }, | |
| { | |
| "epoch": 0.3639883833494676, | |
| "grad_norm": 0.0874338448047638, | |
| "learning_rate": 5.90440267166055e-05, | |
| "loss": 0.021, | |
| "step": 94 | |
| }, | |
| { | |
| "epoch": 0.36786060019361083, | |
| "grad_norm": 0.05889058858156204, | |
| "learning_rate": 5.8229729514036705e-05, | |
| "loss": 0.0233, | |
| "step": 95 | |
| }, | |
| { | |
| "epoch": 0.3717328170377541, | |
| "grad_norm": 0.140172079205513, | |
| "learning_rate": 5.74131823855921e-05, | |
| "loss": 0.0247, | |
| "step": 96 | |
| }, | |
| { | |
| "epoch": 0.37560503388189737, | |
| "grad_norm": 0.05803554505109787, | |
| "learning_rate": 5.6594608567103456e-05, | |
| "loss": 0.014, | |
| "step": 97 | |
| }, | |
| { | |
| "epoch": 0.37947725072604066, | |
| "grad_norm": 0.05956672504544258, | |
| "learning_rate": 5.577423184847932e-05, | |
| "loss": 0.0139, | |
| "step": 98 | |
| }, | |
| { | |
| "epoch": 0.38334946757018395, | |
| "grad_norm": 0.09673935920000076, | |
| "learning_rate": 5.495227651252315e-05, | |
| "loss": 0.0224, | |
| "step": 99 | |
| }, | |
| { | |
| "epoch": 0.3872216844143272, | |
| "grad_norm": 0.12850569188594818, | |
| "learning_rate": 5.4128967273616625e-05, | |
| "loss": 0.0284, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.3872216844143272, | |
| "eval_loss": 0.01937755011022091, | |
| "eval_runtime": 31.0085, | |
| "eval_samples_per_second": 14.028, | |
| "eval_steps_per_second": 3.515, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.3910939012584705, | |
| "grad_norm": 0.05301031470298767, | |
| "learning_rate": 5.330452921628497e-05, | |
| "loss": 0.0234, | |
| "step": 101 | |
| }, | |
| { | |
| "epoch": 0.39496611810261373, | |
| "grad_norm": 0.06627804785966873, | |
| "learning_rate": 5.247918773366112e-05, | |
| "loss": 0.0187, | |
| "step": 102 | |
| }, | |
| { | |
| "epoch": 0.398838334946757, | |
| "grad_norm": 0.05236673355102539, | |
| "learning_rate": 5.165316846586541e-05, | |
| "loss": 0.0171, | |
| "step": 103 | |
| }, | |
| { | |
| "epoch": 0.4027105517909003, | |
| "grad_norm": 0.04712076112627983, | |
| "learning_rate": 5.0826697238317935e-05, | |
| "loss": 0.0195, | |
| "step": 104 | |
| }, | |
| { | |
| "epoch": 0.40658276863504356, | |
| "grad_norm": 0.046172890812158585, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0168, | |
| "step": 105 | |
| }, | |
| { | |
| "epoch": 0.41045498547918685, | |
| "grad_norm": 0.05620027333498001, | |
| "learning_rate": 4.917330276168208e-05, | |
| "loss": 0.0156, | |
| "step": 106 | |
| }, | |
| { | |
| "epoch": 0.4143272023233301, | |
| "grad_norm": 0.06328780949115753, | |
| "learning_rate": 4.834683153413459e-05, | |
| "loss": 0.0203, | |
| "step": 107 | |
| }, | |
| { | |
| "epoch": 0.4181994191674734, | |
| "grad_norm": 0.0666506290435791, | |
| "learning_rate": 4.7520812266338885e-05, | |
| "loss": 0.0267, | |
| "step": 108 | |
| }, | |
| { | |
| "epoch": 0.4220716360116166, | |
| "grad_norm": 0.042762067168951035, | |
| "learning_rate": 4.669547078371504e-05, | |
| "loss": 0.0156, | |
| "step": 109 | |
| }, | |
| { | |
| "epoch": 0.4259438528557599, | |
| "grad_norm": 0.05910148099064827, | |
| "learning_rate": 4.5871032726383386e-05, | |
| "loss": 0.0196, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.4298160696999032, | |
| "grad_norm": 0.046227678656578064, | |
| "learning_rate": 4.504772348747687e-05, | |
| "loss": 0.0193, | |
| "step": 111 | |
| }, | |
| { | |
| "epoch": 0.43368828654404645, | |
| "grad_norm": 0.051421985030174255, | |
| "learning_rate": 4.4225768151520694e-05, | |
| "loss": 0.0183, | |
| "step": 112 | |
| }, | |
| { | |
| "epoch": 0.43756050338818975, | |
| "grad_norm": 0.03285564109683037, | |
| "learning_rate": 4.3405391432896555e-05, | |
| "loss": 0.0148, | |
| "step": 113 | |
| }, | |
| { | |
| "epoch": 0.441432720232333, | |
| "grad_norm": 0.035902880132198334, | |
| "learning_rate": 4.2586817614407895e-05, | |
| "loss": 0.0151, | |
| "step": 114 | |
| }, | |
| { | |
| "epoch": 0.4453049370764763, | |
| "grad_norm": 0.04394988343119621, | |
| "learning_rate": 4.17702704859633e-05, | |
| "loss": 0.0187, | |
| "step": 115 | |
| }, | |
| { | |
| "epoch": 0.4491771539206196, | |
| "grad_norm": 0.04423746466636658, | |
| "learning_rate": 4.095597328339452e-05, | |
| "loss": 0.0172, | |
| "step": 116 | |
| }, | |
| { | |
| "epoch": 0.4530493707647628, | |
| "grad_norm": 0.05456831306219101, | |
| "learning_rate": 4.0144148627425993e-05, | |
| "loss": 0.023, | |
| "step": 117 | |
| }, | |
| { | |
| "epoch": 0.4569215876089061, | |
| "grad_norm": 0.05230424925684929, | |
| "learning_rate": 3.933501846281267e-05, | |
| "loss": 0.019, | |
| "step": 118 | |
| }, | |
| { | |
| "epoch": 0.46079380445304935, | |
| "grad_norm": 0.04211042821407318, | |
| "learning_rate": 3.852880399766243e-05, | |
| "loss": 0.0174, | |
| "step": 119 | |
| }, | |
| { | |
| "epoch": 0.46466602129719264, | |
| "grad_norm": 0.04608823359012604, | |
| "learning_rate": 3.772572564296005e-05, | |
| "loss": 0.0239, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.46853823814133594, | |
| "grad_norm": 0.07439879328012466, | |
| "learning_rate": 3.6926002952309016e-05, | |
| "loss": 0.018, | |
| "step": 121 | |
| }, | |
| { | |
| "epoch": 0.4724104549854792, | |
| "grad_norm": 0.049552951008081436, | |
| "learning_rate": 3.612985456190778e-05, | |
| "loss": 0.0146, | |
| "step": 122 | |
| }, | |
| { | |
| "epoch": 0.4762826718296225, | |
| "grad_norm": 0.051944442093372345, | |
| "learning_rate": 3.533749813077677e-05, | |
| "loss": 0.0223, | |
| "step": 123 | |
| }, | |
| { | |
| "epoch": 0.4801548886737657, | |
| "grad_norm": 0.03937254101037979, | |
| "learning_rate": 3.4549150281252636e-05, | |
| "loss": 0.0135, | |
| "step": 124 | |
| }, | |
| { | |
| "epoch": 0.484027105517909, | |
| "grad_norm": 0.07220064103603363, | |
| "learning_rate": 3.3765026539765834e-05, | |
| "loss": 0.017, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 0.4878993223620523, | |
| "grad_norm": 0.039477597922086716, | |
| "learning_rate": 3.298534127791785e-05, | |
| "loss": 0.0158, | |
| "step": 126 | |
| }, | |
| { | |
| "epoch": 0.49177153920619554, | |
| "grad_norm": 0.04569016396999359, | |
| "learning_rate": 3.221030765387417e-05, | |
| "loss": 0.0124, | |
| "step": 127 | |
| }, | |
| { | |
| "epoch": 0.49564375605033884, | |
| "grad_norm": 0.051465414464473724, | |
| "learning_rate": 3.144013755408895e-05, | |
| "loss": 0.0174, | |
| "step": 128 | |
| }, | |
| { | |
| "epoch": 0.4995159728944821, | |
| "grad_norm": 0.06285690516233444, | |
| "learning_rate": 3.0675041535377405e-05, | |
| "loss": 0.0181, | |
| "step": 129 | |
| }, | |
| { | |
| "epoch": 0.5033881897386253, | |
| "grad_norm": 0.03760867938399315, | |
| "learning_rate": 2.991522876735154e-05, | |
| "loss": 0.0161, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.5072604065827686, | |
| "grad_norm": 0.0777566134929657, | |
| "learning_rate": 2.916090697523549e-05, | |
| "loss": 0.0352, | |
| "step": 131 | |
| }, | |
| { | |
| "epoch": 0.5111326234269119, | |
| "grad_norm": 0.04282362014055252, | |
| "learning_rate": 2.8412282383075363e-05, | |
| "loss": 0.0144, | |
| "step": 132 | |
| }, | |
| { | |
| "epoch": 0.5150048402710552, | |
| "grad_norm": 0.04963918775320053, | |
| "learning_rate": 2.766955965735968e-05, | |
| "loss": 0.0162, | |
| "step": 133 | |
| }, | |
| { | |
| "epoch": 0.5188770571151985, | |
| "grad_norm": 0.050760362297296524, | |
| "learning_rate": 2.693294185106562e-05, | |
| "loss": 0.0173, | |
| "step": 134 | |
| }, | |
| { | |
| "epoch": 0.5227492739593417, | |
| "grad_norm": 0.055553846061229706, | |
| "learning_rate": 2.6202630348146324e-05, | |
| "loss": 0.0197, | |
| "step": 135 | |
| }, | |
| { | |
| "epoch": 0.526621490803485, | |
| "grad_norm": 0.06745583564043045, | |
| "learning_rate": 2.547882480847461e-05, | |
| "loss": 0.021, | |
| "step": 136 | |
| }, | |
| { | |
| "epoch": 0.5304937076476283, | |
| "grad_norm": 0.061450064182281494, | |
| "learning_rate": 2.476172311325783e-05, | |
| "loss": 0.0211, | |
| "step": 137 | |
| }, | |
| { | |
| "epoch": 0.5343659244917716, | |
| "grad_norm": 0.06228160485625267, | |
| "learning_rate": 2.405152131093926e-05, | |
| "loss": 0.0228, | |
| "step": 138 | |
| }, | |
| { | |
| "epoch": 0.5382381413359149, | |
| "grad_norm": 0.055065497756004333, | |
| "learning_rate": 2.3348413563600325e-05, | |
| "loss": 0.0145, | |
| "step": 139 | |
| }, | |
| { | |
| "epoch": 0.542110358180058, | |
| "grad_norm": 0.041081007570028305, | |
| "learning_rate": 2.2652592093878666e-05, | |
| "loss": 0.0155, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.5459825750242013, | |
| "grad_norm": 0.06930065900087357, | |
| "learning_rate": 2.196424713241637e-05, | |
| "loss": 0.0167, | |
| "step": 141 | |
| }, | |
| { | |
| "epoch": 0.5498547918683446, | |
| "grad_norm": 0.056641146540641785, | |
| "learning_rate": 2.128356686585282e-05, | |
| "loss": 0.0173, | |
| "step": 142 | |
| }, | |
| { | |
| "epoch": 0.5537270087124879, | |
| "grad_norm": 0.04107319563627243, | |
| "learning_rate": 2.061073738537635e-05, | |
| "loss": 0.0168, | |
| "step": 143 | |
| }, | |
| { | |
| "epoch": 0.5575992255566312, | |
| "grad_norm": 0.05168822407722473, | |
| "learning_rate": 1.9945942635848748e-05, | |
| "loss": 0.0144, | |
| "step": 144 | |
| }, | |
| { | |
| "epoch": 0.5614714424007744, | |
| "grad_norm": 0.03349588066339493, | |
| "learning_rate": 1.928936436551661e-05, | |
| "loss": 0.0125, | |
| "step": 145 | |
| }, | |
| { | |
| "epoch": 0.5653436592449177, | |
| "grad_norm": 0.0859992504119873, | |
| "learning_rate": 1.8641182076323148e-05, | |
| "loss": 0.0287, | |
| "step": 146 | |
| }, | |
| { | |
| "epoch": 0.569215876089061, | |
| "grad_norm": 0.044449854642152786, | |
| "learning_rate": 1.800157297483417e-05, | |
| "loss": 0.0121, | |
| "step": 147 | |
| }, | |
| { | |
| "epoch": 0.5730880929332043, | |
| "grad_norm": 0.044262442737817764, | |
| "learning_rate": 1.7370711923791567e-05, | |
| "loss": 0.0108, | |
| "step": 148 | |
| }, | |
| { | |
| "epoch": 0.5769603097773476, | |
| "grad_norm": 0.06609819829463959, | |
| "learning_rate": 1.6748771394307585e-05, | |
| "loss": 0.0202, | |
| "step": 149 | |
| }, | |
| { | |
| "epoch": 0.5808325266214908, | |
| "grad_norm": 0.08649075031280518, | |
| "learning_rate": 1.6135921418712956e-05, | |
| "loss": 0.0131, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.5808325266214908, | |
| "eval_loss": 0.017788490280508995, | |
| "eval_runtime": 31.0247, | |
| "eval_samples_per_second": 14.021, | |
| "eval_steps_per_second": 3.513, | |
| "step": 150 | |
| } | |
| ], | |
| "logging_steps": 1, | |
| "max_steps": 200, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 50, | |
| "stateful_callbacks": { | |
| "EarlyStoppingCallback": { | |
| "args": { | |
| "early_stopping_patience": 5, | |
| "early_stopping_threshold": 0.0 | |
| }, | |
| "attributes": { | |
| "early_stopping_patience_counter": 0 | |
| } | |
| }, | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 1.995762988744704e+17, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |