| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 20.0, | |
| "eval_steps": 500, | |
| "global_step": 10580, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.1890359168241966, | |
| "grad_norm": 0.2088678479194641, | |
| "learning_rate": 0.0003, | |
| "loss": 1.909, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.3780718336483932, | |
| "grad_norm": 0.24642816185951233, | |
| "learning_rate": 0.0003, | |
| "loss": 1.8584, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.5671077504725898, | |
| "grad_norm": 0.27153250575065613, | |
| "learning_rate": 0.0003, | |
| "loss": 1.8175, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.7561436672967864, | |
| "grad_norm": 0.3125714957714081, | |
| "learning_rate": 0.0003, | |
| "loss": 1.8035, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.945179584120983, | |
| "grad_norm": 0.29630693793296814, | |
| "learning_rate": 0.0003, | |
| "loss": 1.7637, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.6288430493273542, | |
| "eval_loss": 1.4994953870773315, | |
| "eval_runtime": 6.5869, | |
| "eval_samples_per_second": 75.908, | |
| "eval_steps_per_second": 9.564, | |
| "step": 529 | |
| }, | |
| { | |
| "epoch": 1.1342155009451795, | |
| "grad_norm": 0.42783793807029724, | |
| "learning_rate": 0.0003, | |
| "loss": 1.5915, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.3232514177693762, | |
| "grad_norm": 0.5046432614326477, | |
| "learning_rate": 0.0003, | |
| "loss": 1.4725, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 1.5122873345935728, | |
| "grad_norm": 0.5139931440353394, | |
| "learning_rate": 0.0003, | |
| "loss": 1.4387, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 1.7013232514177694, | |
| "grad_norm": 0.5190715789794922, | |
| "learning_rate": 0.0003, | |
| "loss": 1.4124, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 1.8903591682419658, | |
| "grad_norm": 0.5980258584022522, | |
| "learning_rate": 0.0003, | |
| "loss": 1.3986, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.6719730941704036, | |
| "eval_loss": 1.171139121055603, | |
| "eval_runtime": 6.5804, | |
| "eval_samples_per_second": 75.983, | |
| "eval_steps_per_second": 9.574, | |
| "step": 1058 | |
| }, | |
| { | |
| "epoch": 2.0793950850661624, | |
| "grad_norm": 0.614499568939209, | |
| "learning_rate": 0.0003, | |
| "loss": 1.207, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 2.268431001890359, | |
| "grad_norm": 0.812962532043457, | |
| "learning_rate": 0.0003, | |
| "loss": 0.9878, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 2.4574669187145557, | |
| "grad_norm": 0.7435634732246399, | |
| "learning_rate": 0.0003, | |
| "loss": 0.9801, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 2.6465028355387523, | |
| "grad_norm": 0.778424859046936, | |
| "learning_rate": 0.0003, | |
| "loss": 0.972, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 2.835538752362949, | |
| "grad_norm": 0.7896220684051514, | |
| "learning_rate": 0.0003, | |
| "loss": 0.9515, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.7147713004484305, | |
| "eval_loss": 0.8765971660614014, | |
| "eval_runtime": 6.6432, | |
| "eval_samples_per_second": 75.265, | |
| "eval_steps_per_second": 9.483, | |
| "step": 1587 | |
| }, | |
| { | |
| "epoch": 3.0245746691871456, | |
| "grad_norm": 0.8365605473518372, | |
| "learning_rate": 0.0003, | |
| "loss": 0.8826, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 3.213610586011342, | |
| "grad_norm": 0.8311352729797363, | |
| "learning_rate": 0.0003, | |
| "loss": 0.5808, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 3.402646502835539, | |
| "grad_norm": 0.7168679237365723, | |
| "learning_rate": 0.0003, | |
| "loss": 0.611, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 3.5916824196597354, | |
| "grad_norm": 0.9022857546806335, | |
| "learning_rate": 0.0003, | |
| "loss": 0.6249, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 3.780718336483932, | |
| "grad_norm": 1.0664108991622925, | |
| "learning_rate": 0.0003, | |
| "loss": 0.6378, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 3.9697542533081287, | |
| "grad_norm": 0.8847127556800842, | |
| "learning_rate": 0.0003, | |
| "loss": 0.642, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.7477668161434977, | |
| "eval_loss": 0.6719540953636169, | |
| "eval_runtime": 7.1688, | |
| "eval_samples_per_second": 69.747, | |
| "eval_steps_per_second": 8.788, | |
| "step": 2116 | |
| }, | |
| { | |
| "epoch": 4.158790170132325, | |
| "grad_norm": 0.765385627746582, | |
| "learning_rate": 0.0003, | |
| "loss": 0.4204, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 4.3478260869565215, | |
| "grad_norm": 0.8121228218078613, | |
| "learning_rate": 0.0003, | |
| "loss": 0.3987, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 4.536862003780718, | |
| "grad_norm": 0.8688237071037292, | |
| "learning_rate": 0.0003, | |
| "loss": 0.4055, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 4.725897920604915, | |
| "grad_norm": 0.8493298292160034, | |
| "learning_rate": 0.0003, | |
| "loss": 0.4276, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 4.914933837429111, | |
| "grad_norm": 0.8626968264579773, | |
| "learning_rate": 0.0003, | |
| "loss": 0.4362, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.7697219730941705, | |
| "eval_loss": 0.5458089709281921, | |
| "eval_runtime": 6.5529, | |
| "eval_samples_per_second": 76.302, | |
| "eval_steps_per_second": 9.614, | |
| "step": 2645 | |
| }, | |
| { | |
| "epoch": 5.103969754253308, | |
| "grad_norm": 0.8268809914588928, | |
| "learning_rate": 0.0003, | |
| "loss": 0.3507, | |
| "step": 2700 | |
| }, | |
| { | |
| "epoch": 5.293005671077505, | |
| "grad_norm": 0.7695613503456116, | |
| "learning_rate": 0.0003, | |
| "loss": 0.2829, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 5.482041587901701, | |
| "grad_norm": 0.8355693221092224, | |
| "learning_rate": 0.0003, | |
| "loss": 0.2987, | |
| "step": 2900 | |
| }, | |
| { | |
| "epoch": 5.671077504725898, | |
| "grad_norm": 0.8978683948516846, | |
| "learning_rate": 0.0003, | |
| "loss": 0.3112, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 5.8601134215500945, | |
| "grad_norm": 0.9739505648612976, | |
| "learning_rate": 0.0003, | |
| "loss": 0.3201, | |
| "step": 3100 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_accuracy": 0.7822780269058296, | |
| "eval_loss": 0.4750530421733856, | |
| "eval_runtime": 6.4801, | |
| "eval_samples_per_second": 77.159, | |
| "eval_steps_per_second": 9.722, | |
| "step": 3174 | |
| }, | |
| { | |
| "epoch": 6.049149338374291, | |
| "grad_norm": 0.8119204044342041, | |
| "learning_rate": 0.0003, | |
| "loss": 0.3033, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 6.238185255198488, | |
| "grad_norm": 0.7731099128723145, | |
| "learning_rate": 0.0003, | |
| "loss": 0.2191, | |
| "step": 3300 | |
| }, | |
| { | |
| "epoch": 6.427221172022684, | |
| "grad_norm": 0.7242079377174377, | |
| "learning_rate": 0.0003, | |
| "loss": 0.2313, | |
| "step": 3400 | |
| }, | |
| { | |
| "epoch": 6.616257088846881, | |
| "grad_norm": 0.8276740908622742, | |
| "learning_rate": 0.0003, | |
| "loss": 0.2433, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 6.805293005671078, | |
| "grad_norm": 0.8737579584121704, | |
| "learning_rate": 0.0003, | |
| "loss": 0.2545, | |
| "step": 3600 | |
| }, | |
| { | |
| "epoch": 6.994328922495274, | |
| "grad_norm": 0.8738126158714294, | |
| "learning_rate": 0.0003, | |
| "loss": 0.2652, | |
| "step": 3700 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_accuracy": 0.7886816143497758, | |
| "eval_loss": 0.4510054588317871, | |
| "eval_runtime": 7.2639, | |
| "eval_samples_per_second": 68.834, | |
| "eval_steps_per_second": 8.673, | |
| "step": 3703 | |
| }, | |
| { | |
| "epoch": 7.183364839319471, | |
| "grad_norm": 0.8802294731140137, | |
| "learning_rate": 0.0003, | |
| "loss": 0.1865, | |
| "step": 3800 | |
| }, | |
| { | |
| "epoch": 7.3724007561436675, | |
| "grad_norm": 0.740033745765686, | |
| "learning_rate": 0.0003, | |
| "loss": 0.1971, | |
| "step": 3900 | |
| }, | |
| { | |
| "epoch": 7.561436672967864, | |
| "grad_norm": 0.8324846029281616, | |
| "learning_rate": 0.0003, | |
| "loss": 0.2043, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 7.750472589792061, | |
| "grad_norm": 0.8003520369529724, | |
| "learning_rate": 0.0003, | |
| "loss": 0.2136, | |
| "step": 4100 | |
| }, | |
| { | |
| "epoch": 7.939508506616257, | |
| "grad_norm": 0.9331985116004944, | |
| "learning_rate": 0.0003, | |
| "loss": 0.2263, | |
| "step": 4200 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_accuracy": 0.791372197309417, | |
| "eval_loss": 0.4372054934501648, | |
| "eval_runtime": 6.6756, | |
| "eval_samples_per_second": 74.9, | |
| "eval_steps_per_second": 9.437, | |
| "step": 4232 | |
| }, | |
| { | |
| "epoch": 8.128544423440454, | |
| "grad_norm": 0.668910801410675, | |
| "learning_rate": 0.0003, | |
| "loss": 0.1888, | |
| "step": 4300 | |
| }, | |
| { | |
| "epoch": 8.31758034026465, | |
| "grad_norm": 0.7265163660049438, | |
| "learning_rate": 0.0003, | |
| "loss": 0.172, | |
| "step": 4400 | |
| }, | |
| { | |
| "epoch": 8.506616257088847, | |
| "grad_norm": 0.6801394820213318, | |
| "learning_rate": 0.0003, | |
| "loss": 0.1861, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 8.695652173913043, | |
| "grad_norm": 0.709992527961731, | |
| "learning_rate": 0.0003, | |
| "loss": 0.1923, | |
| "step": 4600 | |
| }, | |
| { | |
| "epoch": 8.88468809073724, | |
| "grad_norm": 0.7939429879188538, | |
| "learning_rate": 0.0003, | |
| "loss": 0.2035, | |
| "step": 4700 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_accuracy": 0.7940269058295965, | |
| "eval_loss": 0.43346819281578064, | |
| "eval_runtime": 6.6609, | |
| "eval_samples_per_second": 75.064, | |
| "eval_steps_per_second": 9.458, | |
| "step": 4761 | |
| }, | |
| { | |
| "epoch": 9.073724007561436, | |
| "grad_norm": 0.5764950513839722, | |
| "learning_rate": 0.0003, | |
| "loss": 0.1902, | |
| "step": 4800 | |
| }, | |
| { | |
| "epoch": 9.262759924385634, | |
| "grad_norm": 0.7292913198471069, | |
| "learning_rate": 0.0003, | |
| "loss": 0.162, | |
| "step": 4900 | |
| }, | |
| { | |
| "epoch": 9.45179584120983, | |
| "grad_norm": 0.8819996118545532, | |
| "learning_rate": 0.0003, | |
| "loss": 0.168, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 9.640831758034027, | |
| "grad_norm": 0.8124964237213135, | |
| "learning_rate": 0.0003, | |
| "loss": 0.1814, | |
| "step": 5100 | |
| }, | |
| { | |
| "epoch": 9.829867674858223, | |
| "grad_norm": 0.9616438150405884, | |
| "learning_rate": 0.0003, | |
| "loss": 0.1913, | |
| "step": 5200 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_accuracy": 0.7949955156950672, | |
| "eval_loss": 0.43220242857933044, | |
| "eval_runtime": 6.518, | |
| "eval_samples_per_second": 76.71, | |
| "eval_steps_per_second": 9.666, | |
| "step": 5290 | |
| }, | |
| { | |
| "epoch": 10.01890359168242, | |
| "grad_norm": 0.693879246711731, | |
| "learning_rate": 0.0003, | |
| "loss": 0.1943, | |
| "step": 5300 | |
| }, | |
| { | |
| "epoch": 10.207939508506616, | |
| "grad_norm": 0.7181903719902039, | |
| "learning_rate": 0.0003, | |
| "loss": 0.1527, | |
| "step": 5400 | |
| }, | |
| { | |
| "epoch": 10.396975425330814, | |
| "grad_norm": 0.7881558537483215, | |
| "learning_rate": 0.0003, | |
| "loss": 0.1599, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 10.58601134215501, | |
| "grad_norm": 0.8658528327941895, | |
| "learning_rate": 0.0003, | |
| "loss": 0.1683, | |
| "step": 5600 | |
| }, | |
| { | |
| "epoch": 10.775047258979207, | |
| "grad_norm": 0.8280355334281921, | |
| "learning_rate": 0.0003, | |
| "loss": 0.1794, | |
| "step": 5700 | |
| }, | |
| { | |
| "epoch": 10.964083175803403, | |
| "grad_norm": 0.8619106411933899, | |
| "learning_rate": 0.0003, | |
| "loss": 0.188, | |
| "step": 5800 | |
| }, | |
| { | |
| "epoch": 11.0, | |
| "eval_accuracy": 0.7945470852017937, | |
| "eval_loss": 0.43794524669647217, | |
| "eval_runtime": 6.1793, | |
| "eval_samples_per_second": 80.915, | |
| "eval_steps_per_second": 10.195, | |
| "step": 5819 | |
| }, | |
| { | |
| "epoch": 11.1531190926276, | |
| "grad_norm": 0.6532451510429382, | |
| "learning_rate": 0.0003, | |
| "loss": 0.1553, | |
| "step": 5900 | |
| }, | |
| { | |
| "epoch": 11.342155009451796, | |
| "grad_norm": 0.6573035717010498, | |
| "learning_rate": 0.0003, | |
| "loss": 0.152, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 11.531190926275993, | |
| "grad_norm": 0.7298913598060608, | |
| "learning_rate": 0.0003, | |
| "loss": 0.1611, | |
| "step": 6100 | |
| }, | |
| { | |
| "epoch": 11.720226843100189, | |
| "grad_norm": 0.7905253767967224, | |
| "learning_rate": 0.0003, | |
| "loss": 0.1699, | |
| "step": 6200 | |
| }, | |
| { | |
| "epoch": 11.909262759924385, | |
| "grad_norm": 0.5873724222183228, | |
| "learning_rate": 0.0003, | |
| "loss": 0.1777, | |
| "step": 6300 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "eval_accuracy": 0.795695067264574, | |
| "eval_loss": 0.4279196858406067, | |
| "eval_runtime": 7.2474, | |
| "eval_samples_per_second": 68.99, | |
| "eval_steps_per_second": 8.693, | |
| "step": 6348 | |
| }, | |
| { | |
| "epoch": 12.098298676748582, | |
| "grad_norm": 0.6252820491790771, | |
| "learning_rate": 0.0003, | |
| "loss": 0.1608, | |
| "step": 6400 | |
| }, | |
| { | |
| "epoch": 12.287334593572778, | |
| "grad_norm": 0.7915599346160889, | |
| "learning_rate": 0.0003, | |
| "loss": 0.1467, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 12.476370510396976, | |
| "grad_norm": 0.7576895356178284, | |
| "learning_rate": 0.0003, | |
| "loss": 0.1539, | |
| "step": 6600 | |
| }, | |
| { | |
| "epoch": 12.665406427221171, | |
| "grad_norm": 0.7499473690986633, | |
| "learning_rate": 0.0003, | |
| "loss": 0.1648, | |
| "step": 6700 | |
| }, | |
| { | |
| "epoch": 12.854442344045369, | |
| "grad_norm": 0.8016745448112488, | |
| "learning_rate": 0.0003, | |
| "loss": 0.1723, | |
| "step": 6800 | |
| }, | |
| { | |
| "epoch": 13.0, | |
| "eval_accuracy": 0.7956322869955157, | |
| "eval_loss": 0.43264538049697876, | |
| "eval_runtime": 6.6858, | |
| "eval_samples_per_second": 74.785, | |
| "eval_steps_per_second": 9.423, | |
| "step": 6877 | |
| }, | |
| { | |
| "epoch": 13.043478260869565, | |
| "grad_norm": 0.6766420006752014, | |
| "learning_rate": 0.0003, | |
| "loss": 0.1715, | |
| "step": 6900 | |
| }, | |
| { | |
| "epoch": 13.232514177693762, | |
| "grad_norm": 0.6546654105186462, | |
| "learning_rate": 0.0003, | |
| "loss": 0.1461, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 13.421550094517958, | |
| "grad_norm": 0.7697082757949829, | |
| "learning_rate": 0.0003, | |
| "loss": 0.1498, | |
| "step": 7100 | |
| }, | |
| { | |
| "epoch": 13.610586011342155, | |
| "grad_norm": 0.6504223346710205, | |
| "learning_rate": 0.0003, | |
| "loss": 0.1589, | |
| "step": 7200 | |
| }, | |
| { | |
| "epoch": 13.799621928166351, | |
| "grad_norm": 0.834709882736206, | |
| "learning_rate": 0.0003, | |
| "loss": 0.1664, | |
| "step": 7300 | |
| }, | |
| { | |
| "epoch": 13.988657844990549, | |
| "grad_norm": 0.7842487692832947, | |
| "learning_rate": 0.0003, | |
| "loss": 0.1767, | |
| "step": 7400 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "eval_accuracy": 0.7966636771300448, | |
| "eval_loss": 0.43290552496910095, | |
| "eval_runtime": 6.5452, | |
| "eval_samples_per_second": 76.392, | |
| "eval_steps_per_second": 9.625, | |
| "step": 7406 | |
| }, | |
| { | |
| "epoch": 14.177693761814744, | |
| "grad_norm": 0.762667715549469, | |
| "learning_rate": 0.0003, | |
| "loss": 0.1438, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 14.366729678638942, | |
| "grad_norm": 0.6443786025047302, | |
| "learning_rate": 0.0003, | |
| "loss": 0.1467, | |
| "step": 7600 | |
| }, | |
| { | |
| "epoch": 14.555765595463138, | |
| "grad_norm": 0.7382057309150696, | |
| "learning_rate": 0.0003, | |
| "loss": 0.1525, | |
| "step": 7700 | |
| }, | |
| { | |
| "epoch": 14.744801512287335, | |
| "grad_norm": 0.8144531846046448, | |
| "learning_rate": 0.0003, | |
| "loss": 0.1611, | |
| "step": 7800 | |
| }, | |
| { | |
| "epoch": 14.93383742911153, | |
| "grad_norm": 0.7852641940116882, | |
| "learning_rate": 0.0003, | |
| "loss": 0.1666, | |
| "step": 7900 | |
| }, | |
| { | |
| "epoch": 15.0, | |
| "eval_accuracy": 0.7961793721973094, | |
| "eval_loss": 0.43958863615989685, | |
| "eval_runtime": 6.6743, | |
| "eval_samples_per_second": 74.914, | |
| "eval_steps_per_second": 9.439, | |
| "step": 7935 | |
| }, | |
| { | |
| "epoch": 15.122873345935728, | |
| "grad_norm": 0.5839329361915588, | |
| "learning_rate": 0.0003, | |
| "loss": 0.1497, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 15.311909262759924, | |
| "grad_norm": 0.6967843174934387, | |
| "learning_rate": 0.0003, | |
| "loss": 0.1413, | |
| "step": 8100 | |
| }, | |
| { | |
| "epoch": 15.500945179584122, | |
| "grad_norm": 0.7435296177864075, | |
| "learning_rate": 0.0003, | |
| "loss": 0.145, | |
| "step": 8200 | |
| }, | |
| { | |
| "epoch": 15.689981096408317, | |
| "grad_norm": 0.7955066561698914, | |
| "learning_rate": 0.0003, | |
| "loss": 0.1559, | |
| "step": 8300 | |
| }, | |
| { | |
| "epoch": 15.879017013232515, | |
| "grad_norm": 0.6731191873550415, | |
| "learning_rate": 0.0003, | |
| "loss": 0.1642, | |
| "step": 8400 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "eval_accuracy": 0.7965022421524663, | |
| "eval_loss": 0.43914830684661865, | |
| "eval_runtime": 6.6825, | |
| "eval_samples_per_second": 74.822, | |
| "eval_steps_per_second": 9.428, | |
| "step": 8464 | |
| }, | |
| { | |
| "epoch": 16.068052930056712, | |
| "grad_norm": 0.616921603679657, | |
| "learning_rate": 0.0003, | |
| "loss": 0.1555, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 16.257088846880908, | |
| "grad_norm": 0.6410200595855713, | |
| "learning_rate": 0.0003, | |
| "loss": 0.1379, | |
| "step": 8600 | |
| }, | |
| { | |
| "epoch": 16.446124763705104, | |
| "grad_norm": 0.7206813097000122, | |
| "learning_rate": 0.0003, | |
| "loss": 0.1433, | |
| "step": 8700 | |
| }, | |
| { | |
| "epoch": 16.6351606805293, | |
| "grad_norm": 1.1165937185287476, | |
| "learning_rate": 0.0003, | |
| "loss": 0.1526, | |
| "step": 8800 | |
| }, | |
| { | |
| "epoch": 16.8241965973535, | |
| "grad_norm": 0.7585830688476562, | |
| "learning_rate": 0.0003, | |
| "loss": 0.1575, | |
| "step": 8900 | |
| }, | |
| { | |
| "epoch": 17.0, | |
| "eval_accuracy": 0.7966816143497758, | |
| "eval_loss": 0.44049155712127686, | |
| "eval_runtime": 6.6339, | |
| "eval_samples_per_second": 75.37, | |
| "eval_steps_per_second": 9.497, | |
| "step": 8993 | |
| }, | |
| { | |
| "epoch": 17.013232514177695, | |
| "grad_norm": 0.655029833316803, | |
| "learning_rate": 0.0003, | |
| "loss": 0.1624, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 17.20226843100189, | |
| "grad_norm": 0.7475863099098206, | |
| "learning_rate": 0.0003, | |
| "loss": 0.1356, | |
| "step": 9100 | |
| }, | |
| { | |
| "epoch": 17.391304347826086, | |
| "grad_norm": 0.6971881985664368, | |
| "learning_rate": 0.0003, | |
| "loss": 0.1406, | |
| "step": 9200 | |
| }, | |
| { | |
| "epoch": 17.58034026465028, | |
| "grad_norm": 0.6662493944168091, | |
| "learning_rate": 0.0003, | |
| "loss": 0.1472, | |
| "step": 9300 | |
| }, | |
| { | |
| "epoch": 17.76937618147448, | |
| "grad_norm": 0.7289919257164001, | |
| "learning_rate": 0.0003, | |
| "loss": 0.1578, | |
| "step": 9400 | |
| }, | |
| { | |
| "epoch": 17.958412098298677, | |
| "grad_norm": 0.7780668139457703, | |
| "learning_rate": 0.0003, | |
| "loss": 0.1634, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 18.0, | |
| "eval_accuracy": 0.7976233183856503, | |
| "eval_loss": 0.42652928829193115, | |
| "eval_runtime": 6.7415, | |
| "eval_samples_per_second": 74.167, | |
| "eval_steps_per_second": 9.345, | |
| "step": 9522 | |
| }, | |
| { | |
| "epoch": 18.147448015122873, | |
| "grad_norm": 0.5619010925292969, | |
| "learning_rate": 0.0003, | |
| "loss": 0.1401, | |
| "step": 9600 | |
| }, | |
| { | |
| "epoch": 18.33648393194707, | |
| "grad_norm": 0.7481386065483093, | |
| "learning_rate": 0.0003, | |
| "loss": 0.1378, | |
| "step": 9700 | |
| }, | |
| { | |
| "epoch": 18.525519848771268, | |
| "grad_norm": 0.6953665614128113, | |
| "learning_rate": 0.0003, | |
| "loss": 0.144, | |
| "step": 9800 | |
| }, | |
| { | |
| "epoch": 18.714555765595463, | |
| "grad_norm": 0.8559867143630981, | |
| "learning_rate": 0.0003, | |
| "loss": 0.1521, | |
| "step": 9900 | |
| }, | |
| { | |
| "epoch": 18.90359168241966, | |
| "grad_norm": 0.7916247844696045, | |
| "learning_rate": 0.0003, | |
| "loss": 0.1593, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 19.0, | |
| "eval_accuracy": 0.7978116591928252, | |
| "eval_loss": 0.43227851390838623, | |
| "eval_runtime": 6.4486, | |
| "eval_samples_per_second": 77.537, | |
| "eval_steps_per_second": 9.77, | |
| "step": 10051 | |
| }, | |
| { | |
| "epoch": 19.092627599243855, | |
| "grad_norm": 0.638944685459137, | |
| "learning_rate": 0.0003, | |
| "loss": 0.1468, | |
| "step": 10100 | |
| }, | |
| { | |
| "epoch": 19.281663516068054, | |
| "grad_norm": 0.7771846652030945, | |
| "learning_rate": 0.0003, | |
| "loss": 0.1345, | |
| "step": 10200 | |
| }, | |
| { | |
| "epoch": 19.47069943289225, | |
| "grad_norm": 0.5761379599571228, | |
| "learning_rate": 0.0003, | |
| "loss": 0.1404, | |
| "step": 10300 | |
| }, | |
| { | |
| "epoch": 19.659735349716446, | |
| "grad_norm": 0.6479610204696655, | |
| "learning_rate": 0.0003, | |
| "loss": 0.1489, | |
| "step": 10400 | |
| }, | |
| { | |
| "epoch": 19.84877126654064, | |
| "grad_norm": 0.6894132494926453, | |
| "learning_rate": 0.0003, | |
| "loss": 0.153, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 20.0, | |
| "eval_accuracy": 0.7981434977578475, | |
| "eval_loss": 0.43110987544059753, | |
| "eval_runtime": 6.5758, | |
| "eval_samples_per_second": 76.036, | |
| "eval_steps_per_second": 9.581, | |
| "step": 10580 | |
| }, | |
| { | |
| "epoch": 20.0, | |
| "step": 10580, | |
| "total_flos": 9.64242245391745e+17, | |
| "train_loss": 0.3966417311720677, | |
| "train_runtime": 24179.107, | |
| "train_samples_per_second": 14.0, | |
| "train_steps_per_second": 0.438 | |
| } | |
| ], | |
| "logging_steps": 100, | |
| "max_steps": 10580, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 20, | |
| "save_steps": 500, | |
| "total_flos": 9.64242245391745e+17, | |
| "train_batch_size": 1, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |