| { | |
| "best_metric": 0.007665493991225958, | |
| "best_model_checkpoint": "limbxy_pose/checkpoint-1780", | |
| "epoch": 20.0, | |
| "eval_steps": 500, | |
| "global_step": 1780, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.2808988764044944, | |
| "grad_norm": 2355601.5, | |
| "learning_rate": 5e-06, | |
| "loss": 0.3253, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.5617977528089888, | |
| "grad_norm": 873035.5, | |
| "learning_rate": 1e-05, | |
| "loss": 0.2139, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.8426966292134831, | |
| "grad_norm": 1139656.125, | |
| "learning_rate": 1.5e-05, | |
| "loss": 0.174, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_loss": 0.14710678160190582, | |
| "eval_rmse": 0.38354501128196716, | |
| "eval_runtime": 9.3728, | |
| "eval_samples_per_second": 106.692, | |
| "eval_steps_per_second": 1.707, | |
| "step": 89 | |
| }, | |
| { | |
| "epoch": 1.1235955056179776, | |
| "grad_norm": 959863.5625, | |
| "learning_rate": 2e-05, | |
| "loss": 0.1598, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.404494382022472, | |
| "grad_norm": 368922.84375, | |
| "learning_rate": 2.5e-05, | |
| "loss": 0.1569, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 1.6853932584269664, | |
| "grad_norm": 1642762.375, | |
| "learning_rate": 3e-05, | |
| "loss": 0.1584, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 1.9662921348314608, | |
| "grad_norm": 2866672.75, | |
| "learning_rate": 3.5e-05, | |
| "loss": 0.1559, | |
| "step": 175 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_loss": 0.14660561084747314, | |
| "eval_rmse": 0.38289114832878113, | |
| "eval_runtime": 9.5918, | |
| "eval_samples_per_second": 104.256, | |
| "eval_steps_per_second": 1.668, | |
| "step": 178 | |
| }, | |
| { | |
| "epoch": 2.247191011235955, | |
| "grad_norm": 140559.03125, | |
| "learning_rate": 4e-05, | |
| "loss": 0.1467, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 2.5280898876404496, | |
| "grad_norm": 2080408.125, | |
| "learning_rate": 4.5e-05, | |
| "loss": 0.1534, | |
| "step": 225 | |
| }, | |
| { | |
| "epoch": 2.808988764044944, | |
| "grad_norm": 2639697.5, | |
| "learning_rate": 5e-05, | |
| "loss": 0.1485, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_loss": 0.16608364880084991, | |
| "eval_rmse": 0.4075336158275604, | |
| "eval_runtime": 9.3878, | |
| "eval_samples_per_second": 106.521, | |
| "eval_steps_per_second": 1.704, | |
| "step": 267 | |
| }, | |
| { | |
| "epoch": 3.0898876404494384, | |
| "grad_norm": 1161722.875, | |
| "learning_rate": 4.918300653594771e-05, | |
| "loss": 0.1613, | |
| "step": 275 | |
| }, | |
| { | |
| "epoch": 3.370786516853933, | |
| "grad_norm": 3452693.5, | |
| "learning_rate": 4.8366013071895424e-05, | |
| "loss": 0.1791, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 3.6516853932584272, | |
| "grad_norm": 1499728.0, | |
| "learning_rate": 4.7549019607843135e-05, | |
| "loss": 0.1599, | |
| "step": 325 | |
| }, | |
| { | |
| "epoch": 3.932584269662921, | |
| "grad_norm": 2438932.5, | |
| "learning_rate": 4.673202614379085e-05, | |
| "loss": 0.1624, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_loss": 0.14177033305168152, | |
| "eval_rmse": 0.3765240013599396, | |
| "eval_runtime": 9.4578, | |
| "eval_samples_per_second": 105.733, | |
| "eval_steps_per_second": 1.692, | |
| "step": 356 | |
| }, | |
| { | |
| "epoch": 4.213483146067416, | |
| "grad_norm": 1148349.625, | |
| "learning_rate": 4.5915032679738564e-05, | |
| "loss": 0.1568, | |
| "step": 375 | |
| }, | |
| { | |
| "epoch": 4.49438202247191, | |
| "grad_norm": 2919198.5, | |
| "learning_rate": 4.5098039215686275e-05, | |
| "loss": 0.1513, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 4.775280898876405, | |
| "grad_norm": 1855904.125, | |
| "learning_rate": 4.4281045751633986e-05, | |
| "loss": 0.1457, | |
| "step": 425 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_loss": 0.14367185533046722, | |
| "eval_rmse": 0.3790406882762909, | |
| "eval_runtime": 9.4902, | |
| "eval_samples_per_second": 105.372, | |
| "eval_steps_per_second": 1.686, | |
| "step": 445 | |
| }, | |
| { | |
| "epoch": 5.056179775280899, | |
| "grad_norm": 3503330.5, | |
| "learning_rate": 4.3464052287581704e-05, | |
| "loss": 0.1512, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 5.337078651685394, | |
| "grad_norm": 334371.0, | |
| "learning_rate": 4.2647058823529415e-05, | |
| "loss": 0.1517, | |
| "step": 475 | |
| }, | |
| { | |
| "epoch": 5.617977528089888, | |
| "grad_norm": 987476.0625, | |
| "learning_rate": 4.1830065359477126e-05, | |
| "loss": 0.1511, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 5.898876404494382, | |
| "grad_norm": 1318160.375, | |
| "learning_rate": 4.101307189542484e-05, | |
| "loss": 0.1635, | |
| "step": 525 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_loss": 0.14237765967845917, | |
| "eval_rmse": 0.37732964754104614, | |
| "eval_runtime": 9.6289, | |
| "eval_samples_per_second": 103.854, | |
| "eval_steps_per_second": 1.662, | |
| "step": 534 | |
| }, | |
| { | |
| "epoch": 6.179775280898877, | |
| "grad_norm": 2755903.5, | |
| "learning_rate": 4.0196078431372555e-05, | |
| "loss": 0.1616, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 6.460674157303371, | |
| "grad_norm": 647614.5625, | |
| "learning_rate": 3.9379084967320266e-05, | |
| "loss": 0.1498, | |
| "step": 575 | |
| }, | |
| { | |
| "epoch": 6.741573033707866, | |
| "grad_norm": 1388723.25, | |
| "learning_rate": 3.8562091503267977e-05, | |
| "loss": 0.1428, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_loss": 0.15840163826942444, | |
| "eval_rmse": 0.39799708127975464, | |
| "eval_runtime": 9.5792, | |
| "eval_samples_per_second": 104.393, | |
| "eval_steps_per_second": 1.67, | |
| "step": 623 | |
| }, | |
| { | |
| "epoch": 7.022471910112359, | |
| "grad_norm": 3193922.25, | |
| "learning_rate": 3.774509803921569e-05, | |
| "loss": 0.15, | |
| "step": 625 | |
| }, | |
| { | |
| "epoch": 7.303370786516854, | |
| "grad_norm": 634225.0, | |
| "learning_rate": 3.6928104575163405e-05, | |
| "loss": 0.1491, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 7.584269662921348, | |
| "grad_norm": 1895702.75, | |
| "learning_rate": 3.611111111111111e-05, | |
| "loss": 0.1521, | |
| "step": 675 | |
| }, | |
| { | |
| "epoch": 7.865168539325842, | |
| "grad_norm": 410481.09375, | |
| "learning_rate": 3.529411764705883e-05, | |
| "loss": 0.1481, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_loss": 0.140840545296669, | |
| "eval_rmse": 0.3752872943878174, | |
| "eval_runtime": 9.3963, | |
| "eval_samples_per_second": 106.424, | |
| "eval_steps_per_second": 1.703, | |
| "step": 712 | |
| }, | |
| { | |
| "epoch": 8.146067415730338, | |
| "grad_norm": 121664.78125, | |
| "learning_rate": 3.447712418300654e-05, | |
| "loss": 0.1497, | |
| "step": 725 | |
| }, | |
| { | |
| "epoch": 8.426966292134832, | |
| "grad_norm": 691517.125, | |
| "learning_rate": 3.366013071895425e-05, | |
| "loss": 0.1423, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 8.707865168539326, | |
| "grad_norm": 2580666.25, | |
| "learning_rate": 3.284313725490196e-05, | |
| "loss": 0.1451, | |
| "step": 775 | |
| }, | |
| { | |
| "epoch": 8.98876404494382, | |
| "grad_norm": 1733368.625, | |
| "learning_rate": 3.202614379084967e-05, | |
| "loss": 0.1494, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_loss": 0.1478467881679535, | |
| "eval_rmse": 0.38450852036476135, | |
| "eval_runtime": 9.5629, | |
| "eval_samples_per_second": 104.571, | |
| "eval_steps_per_second": 1.673, | |
| "step": 801 | |
| }, | |
| { | |
| "epoch": 9.269662921348315, | |
| "grad_norm": 312437.125, | |
| "learning_rate": 3.120915032679739e-05, | |
| "loss": 0.1429, | |
| "step": 825 | |
| }, | |
| { | |
| "epoch": 9.55056179775281, | |
| "grad_norm": 231154.46875, | |
| "learning_rate": 3.0392156862745097e-05, | |
| "loss": 0.1444, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 9.831460674157304, | |
| "grad_norm": 1203652.5, | |
| "learning_rate": 2.957516339869281e-05, | |
| "loss": 0.1417, | |
| "step": 875 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_loss": 0.1544562131166458, | |
| "eval_rmse": 0.393009215593338, | |
| "eval_runtime": 9.6401, | |
| "eval_samples_per_second": 103.733, | |
| "eval_steps_per_second": 1.66, | |
| "step": 890 | |
| }, | |
| { | |
| "epoch": 10.112359550561798, | |
| "grad_norm": 1199470.25, | |
| "learning_rate": 2.8758169934640522e-05, | |
| "loss": 0.158, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 10.393258426966293, | |
| "grad_norm": 192614.96875, | |
| "learning_rate": 2.7941176470588236e-05, | |
| "loss": 0.15, | |
| "step": 925 | |
| }, | |
| { | |
| "epoch": 10.674157303370787, | |
| "grad_norm": 1780062.125, | |
| "learning_rate": 2.7124183006535947e-05, | |
| "loss": 0.1439, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 10.955056179775282, | |
| "grad_norm": 1194267.125, | |
| "learning_rate": 2.630718954248366e-05, | |
| "loss": 0.1421, | |
| "step": 975 | |
| }, | |
| { | |
| "epoch": 11.0, | |
| "eval_loss": 0.1432497501373291, | |
| "eval_rmse": 0.37848347425460815, | |
| "eval_runtime": 9.3096, | |
| "eval_samples_per_second": 107.416, | |
| "eval_steps_per_second": 1.719, | |
| "step": 979 | |
| }, | |
| { | |
| "epoch": 11.235955056179776, | |
| "grad_norm": 527445.1875, | |
| "learning_rate": 2.5490196078431373e-05, | |
| "loss": 0.1392, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 11.51685393258427, | |
| "grad_norm": 186754.59375, | |
| "learning_rate": 2.4673202614379087e-05, | |
| "loss": 0.1455, | |
| "step": 1025 | |
| }, | |
| { | |
| "epoch": 11.797752808988765, | |
| "grad_norm": 184548.09375, | |
| "learning_rate": 2.38562091503268e-05, | |
| "loss": 0.145, | |
| "step": 1050 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "eval_loss": 0.14026711881160736, | |
| "eval_rmse": 0.37452250719070435, | |
| "eval_runtime": 9.4209, | |
| "eval_samples_per_second": 106.147, | |
| "eval_steps_per_second": 1.698, | |
| "step": 1068 | |
| }, | |
| { | |
| "epoch": 12.07865168539326, | |
| "grad_norm": 234157.390625, | |
| "learning_rate": 2.303921568627451e-05, | |
| "loss": 0.1445, | |
| "step": 1075 | |
| }, | |
| { | |
| "epoch": 12.359550561797754, | |
| "grad_norm": 1915960.625, | |
| "learning_rate": 2.2222222222222223e-05, | |
| "loss": 0.144, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 12.640449438202246, | |
| "grad_norm": 664816.875, | |
| "learning_rate": 2.1405228758169934e-05, | |
| "loss": 0.142, | |
| "step": 1125 | |
| }, | |
| { | |
| "epoch": 12.921348314606742, | |
| "grad_norm": 237225.984375, | |
| "learning_rate": 2.058823529411765e-05, | |
| "loss": 0.1466, | |
| "step": 1150 | |
| }, | |
| { | |
| "epoch": 13.0, | |
| "eval_loss": 0.14430756866931915, | |
| "eval_rmse": 0.37987837195396423, | |
| "eval_runtime": 9.2838, | |
| "eval_samples_per_second": 107.715, | |
| "eval_steps_per_second": 1.723, | |
| "step": 1157 | |
| }, | |
| { | |
| "epoch": 13.202247191011235, | |
| "grad_norm": 322684.4375, | |
| "learning_rate": 1.977124183006536e-05, | |
| "loss": 0.1446, | |
| "step": 1175 | |
| }, | |
| { | |
| "epoch": 13.48314606741573, | |
| "grad_norm": 474081.25, | |
| "learning_rate": 1.895424836601307e-05, | |
| "loss": 0.1414, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 13.764044943820224, | |
| "grad_norm": 1984481.125, | |
| "learning_rate": 1.8137254901960785e-05, | |
| "loss": 0.0601, | |
| "step": 1225 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "eval_loss": 0.02082459256052971, | |
| "eval_rmse": 0.14430728554725647, | |
| "eval_runtime": 9.6578, | |
| "eval_samples_per_second": 103.543, | |
| "eval_steps_per_second": 1.657, | |
| "step": 1246 | |
| }, | |
| { | |
| "epoch": 14.044943820224718, | |
| "grad_norm": 3539425.25, | |
| "learning_rate": 1.7320261437908496e-05, | |
| "loss": 0.0386, | |
| "step": 1250 | |
| }, | |
| { | |
| "epoch": 14.325842696629213, | |
| "grad_norm": 975899.9375, | |
| "learning_rate": 1.650326797385621e-05, | |
| "loss": 0.0274, | |
| "step": 1275 | |
| }, | |
| { | |
| "epoch": 14.606741573033707, | |
| "grad_norm": 3280766.0, | |
| "learning_rate": 1.568627450980392e-05, | |
| "loss": 0.0194, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 14.887640449438202, | |
| "grad_norm": 339197.90625, | |
| "learning_rate": 1.4869281045751634e-05, | |
| "loss": 0.0154, | |
| "step": 1325 | |
| }, | |
| { | |
| "epoch": 15.0, | |
| "eval_loss": 0.012424159795045853, | |
| "eval_rmse": 0.11146371811628342, | |
| "eval_runtime": 9.3192, | |
| "eval_samples_per_second": 107.305, | |
| "eval_steps_per_second": 1.717, | |
| "step": 1335 | |
| }, | |
| { | |
| "epoch": 15.168539325842696, | |
| "grad_norm": 325666.3125, | |
| "learning_rate": 1.4052287581699347e-05, | |
| "loss": 0.0106, | |
| "step": 1350 | |
| }, | |
| { | |
| "epoch": 15.44943820224719, | |
| "grad_norm": 407296.9375, | |
| "learning_rate": 1.323529411764706e-05, | |
| "loss": 0.013, | |
| "step": 1375 | |
| }, | |
| { | |
| "epoch": 15.730337078651685, | |
| "grad_norm": 2163835.5, | |
| "learning_rate": 1.2418300653594772e-05, | |
| "loss": 0.0102, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "eval_loss": 0.012837632559239864, | |
| "eval_rmse": 0.1133032739162445, | |
| "eval_runtime": 9.5459, | |
| "eval_samples_per_second": 104.757, | |
| "eval_steps_per_second": 1.676, | |
| "step": 1424 | |
| }, | |
| { | |
| "epoch": 16.01123595505618, | |
| "grad_norm": 1708998.75, | |
| "learning_rate": 1.1601307189542485e-05, | |
| "loss": 0.0127, | |
| "step": 1425 | |
| }, | |
| { | |
| "epoch": 16.292134831460675, | |
| "grad_norm": 324622.65625, | |
| "learning_rate": 1.0784313725490197e-05, | |
| "loss": 0.0063, | |
| "step": 1450 | |
| }, | |
| { | |
| "epoch": 16.573033707865168, | |
| "grad_norm": 533753.9375, | |
| "learning_rate": 9.96732026143791e-06, | |
| "loss": 0.0067, | |
| "step": 1475 | |
| }, | |
| { | |
| "epoch": 16.853932584269664, | |
| "grad_norm": 20416.09765625, | |
| "learning_rate": 9.150326797385621e-06, | |
| "loss": 0.0071, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 17.0, | |
| "eval_loss": 0.01293737068772316, | |
| "eval_rmse": 0.11374256014823914, | |
| "eval_runtime": 9.3021, | |
| "eval_samples_per_second": 107.502, | |
| "eval_steps_per_second": 1.72, | |
| "step": 1513 | |
| }, | |
| { | |
| "epoch": 17.134831460674157, | |
| "grad_norm": 350699.0625, | |
| "learning_rate": 8.333333333333334e-06, | |
| "loss": 0.0086, | |
| "step": 1525 | |
| }, | |
| { | |
| "epoch": 17.415730337078653, | |
| "grad_norm": 1733482.25, | |
| "learning_rate": 7.5163398692810456e-06, | |
| "loss": 0.0079, | |
| "step": 1550 | |
| }, | |
| { | |
| "epoch": 17.696629213483146, | |
| "grad_norm": 2389954.0, | |
| "learning_rate": 6.699346405228758e-06, | |
| "loss": 0.0095, | |
| "step": 1575 | |
| }, | |
| { | |
| "epoch": 17.97752808988764, | |
| "grad_norm": 451414.53125, | |
| "learning_rate": 5.882352941176471e-06, | |
| "loss": 0.0076, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 18.0, | |
| "eval_loss": 0.008463106118142605, | |
| "eval_rmse": 0.09199514985084534, | |
| "eval_runtime": 9.6031, | |
| "eval_samples_per_second": 104.133, | |
| "eval_steps_per_second": 1.666, | |
| "step": 1602 | |
| }, | |
| { | |
| "epoch": 18.258426966292134, | |
| "grad_norm": 1180487.125, | |
| "learning_rate": 5.065359477124184e-06, | |
| "loss": 0.0055, | |
| "step": 1625 | |
| }, | |
| { | |
| "epoch": 18.53932584269663, | |
| "grad_norm": 1367445.375, | |
| "learning_rate": 4.2483660130718954e-06, | |
| "loss": 0.0057, | |
| "step": 1650 | |
| }, | |
| { | |
| "epoch": 18.820224719101123, | |
| "grad_norm": 903850.0, | |
| "learning_rate": 3.431372549019608e-06, | |
| "loss": 0.0057, | |
| "step": 1675 | |
| }, | |
| { | |
| "epoch": 19.0, | |
| "eval_loss": 0.007886779494583607, | |
| "eval_rmse": 0.08880754560232162, | |
| "eval_runtime": 9.2481, | |
| "eval_samples_per_second": 108.13, | |
| "eval_steps_per_second": 1.73, | |
| "step": 1691 | |
| }, | |
| { | |
| "epoch": 19.10112359550562, | |
| "grad_norm": 801771.375, | |
| "learning_rate": 2.6143790849673204e-06, | |
| "loss": 0.0048, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 19.382022471910112, | |
| "grad_norm": 485753.125, | |
| "learning_rate": 1.7973856209150326e-06, | |
| "loss": 0.0049, | |
| "step": 1725 | |
| }, | |
| { | |
| "epoch": 19.662921348314608, | |
| "grad_norm": 370014.5625, | |
| "learning_rate": 9.80392156862745e-07, | |
| "loss": 0.0044, | |
| "step": 1750 | |
| }, | |
| { | |
| "epoch": 19.9438202247191, | |
| "grad_norm": 474841.78125, | |
| "learning_rate": 1.6339869281045752e-07, | |
| "loss": 0.0046, | |
| "step": 1775 | |
| }, | |
| { | |
| "epoch": 20.0, | |
| "eval_loss": 0.007665493991225958, | |
| "eval_rmse": 0.08755281567573547, | |
| "eval_runtime": 9.5227, | |
| "eval_samples_per_second": 105.012, | |
| "eval_steps_per_second": 1.68, | |
| "step": 1780 | |
| }, | |
| { | |
| "epoch": 20.0, | |
| "step": 1780, | |
| "total_flos": 0.0, | |
| "train_loss": 0.10928233507602067, | |
| "train_runtime": 3299.6169, | |
| "train_samples_per_second": 34.319, | |
| "train_steps_per_second": 0.539 | |
| } | |
| ], | |
| "logging_steps": 25, | |
| "max_steps": 1780, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 20, | |
| "save_steps": 500, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 0.0, | |
| "train_batch_size": 64, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |