|
{ |
|
"best_metric": 0.7699832320213318, |
|
"best_model_checkpoint": "./kd_results/microsoft/swin-tiny-patch4-window7-224_alpha0.7_temp3.0_t3/checkpoint-2460", |
|
"epoch": 20.0, |
|
"eval_steps": 500, |
|
"global_step": 3280, |
|
"is_hyper_param_search": false, |
|
"is_local_process_zero": true, |
|
"is_world_process_zero": true, |
|
"log_history": [ |
|
{ |
|
"epoch": 0.1524390243902439, |
|
"grad_norm": 4.97396993637085, |
|
"learning_rate": 3.8109756097560976e-06, |
|
"loss": 0.5849, |
|
"step": 25 |
|
}, |
|
{ |
|
"epoch": 0.3048780487804878, |
|
"grad_norm": 3.1068673133850098, |
|
"learning_rate": 7.621951219512195e-06, |
|
"loss": 0.5864, |
|
"step": 50 |
|
}, |
|
{ |
|
"epoch": 0.4573170731707317, |
|
"grad_norm": 4.72998046875, |
|
"learning_rate": 1.1432926829268294e-05, |
|
"loss": 0.5742, |
|
"step": 75 |
|
}, |
|
{ |
|
"epoch": 0.6097560975609756, |
|
"grad_norm": 2.589412212371826, |
|
"learning_rate": 1.524390243902439e-05, |
|
"loss": 0.5647, |
|
"step": 100 |
|
}, |
|
{ |
|
"epoch": 0.7621951219512195, |
|
"grad_norm": 3.89925479888916, |
|
"learning_rate": 1.9054878048780488e-05, |
|
"loss": 0.5744, |
|
"step": 125 |
|
}, |
|
{ |
|
"epoch": 0.9146341463414634, |
|
"grad_norm": 2.6995656490325928, |
|
"learning_rate": 2.286585365853659e-05, |
|
"loss": 0.5659, |
|
"step": 150 |
|
}, |
|
{ |
|
"epoch": 1.0, |
|
"eval_accuracy": 0.2635574837310195, |
|
"eval_loss": 1.4279733896255493, |
|
"eval_runtime": 41.2596, |
|
"eval_samples_per_second": 22.346, |
|
"eval_steps_per_second": 0.703, |
|
"step": 164 |
|
}, |
|
{ |
|
"epoch": 1.0670731707317074, |
|
"grad_norm": 2.971930980682373, |
|
"learning_rate": 2.6676829268292686e-05, |
|
"loss": 0.5771, |
|
"step": 175 |
|
}, |
|
{ |
|
"epoch": 1.2195121951219512, |
|
"grad_norm": 3.9521827697753906, |
|
"learning_rate": 3.048780487804878e-05, |
|
"loss": 0.5651, |
|
"step": 200 |
|
}, |
|
{ |
|
"epoch": 1.3719512195121952, |
|
"grad_norm": 3.7694523334503174, |
|
"learning_rate": 3.429878048780488e-05, |
|
"loss": 0.5488, |
|
"step": 225 |
|
}, |
|
{ |
|
"epoch": 1.524390243902439, |
|
"grad_norm": 3.5128374099731445, |
|
"learning_rate": 3.8109756097560976e-05, |
|
"loss": 0.5654, |
|
"step": 250 |
|
}, |
|
{ |
|
"epoch": 1.6768292682926829, |
|
"grad_norm": 5.130870342254639, |
|
"learning_rate": 4.1920731707317077e-05, |
|
"loss": 0.5424, |
|
"step": 275 |
|
}, |
|
{ |
|
"epoch": 1.8292682926829267, |
|
"grad_norm": 4.611592769622803, |
|
"learning_rate": 4.573170731707318e-05, |
|
"loss": 0.5299, |
|
"step": 300 |
|
}, |
|
{ |
|
"epoch": 1.9817073170731707, |
|
"grad_norm": 2.701108932495117, |
|
"learning_rate": 4.954268292682927e-05, |
|
"loss": 0.5111, |
|
"step": 325 |
|
}, |
|
{ |
|
"epoch": 2.0, |
|
"eval_accuracy": 0.3720173535791757, |
|
"eval_loss": 1.3039730787277222, |
|
"eval_runtime": 41.7449, |
|
"eval_samples_per_second": 22.087, |
|
"eval_steps_per_second": 0.695, |
|
"step": 328 |
|
}, |
|
{ |
|
"epoch": 2.1341463414634148, |
|
"grad_norm": 3.373603582382202, |
|
"learning_rate": 4.962737127371274e-05, |
|
"loss": 0.4844, |
|
"step": 350 |
|
}, |
|
{ |
|
"epoch": 2.2865853658536586, |
|
"grad_norm": 2.5607736110687256, |
|
"learning_rate": 4.92039295392954e-05, |
|
"loss": 0.4813, |
|
"step": 375 |
|
}, |
|
{ |
|
"epoch": 2.4390243902439024, |
|
"grad_norm": 4.668150901794434, |
|
"learning_rate": 4.878048780487805e-05, |
|
"loss": 0.4456, |
|
"step": 400 |
|
}, |
|
{ |
|
"epoch": 2.591463414634146, |
|
"grad_norm": 6.863236427307129, |
|
"learning_rate": 4.8357046070460704e-05, |
|
"loss": 0.4369, |
|
"step": 425 |
|
}, |
|
{ |
|
"epoch": 2.7439024390243905, |
|
"grad_norm": 4.940046310424805, |
|
"learning_rate": 4.793360433604336e-05, |
|
"loss": 0.4509, |
|
"step": 450 |
|
}, |
|
{ |
|
"epoch": 2.8963414634146343, |
|
"grad_norm": 3.610745906829834, |
|
"learning_rate": 4.7510162601626016e-05, |
|
"loss": 0.4093, |
|
"step": 475 |
|
}, |
|
{ |
|
"epoch": 3.0, |
|
"eval_accuracy": 0.6355748373101953, |
|
"eval_loss": 0.9967854022979736, |
|
"eval_runtime": 41.3413, |
|
"eval_samples_per_second": 22.302, |
|
"eval_steps_per_second": 0.701, |
|
"step": 492 |
|
}, |
|
{ |
|
"epoch": 3.048780487804878, |
|
"grad_norm": 4.550442218780518, |
|
"learning_rate": 4.7086720867208675e-05, |
|
"loss": 0.4146, |
|
"step": 500 |
|
}, |
|
{ |
|
"epoch": 3.201219512195122, |
|
"grad_norm": 4.069523811340332, |
|
"learning_rate": 4.6663279132791334e-05, |
|
"loss": 0.3872, |
|
"step": 525 |
|
}, |
|
{ |
|
"epoch": 3.3536585365853657, |
|
"grad_norm": 3.338857889175415, |
|
"learning_rate": 4.6239837398373986e-05, |
|
"loss": 0.3766, |
|
"step": 550 |
|
}, |
|
{ |
|
"epoch": 3.5060975609756095, |
|
"grad_norm": 2.963149070739746, |
|
"learning_rate": 4.581639566395664e-05, |
|
"loss": 0.3767, |
|
"step": 575 |
|
}, |
|
{ |
|
"epoch": 3.658536585365854, |
|
"grad_norm": 3.808338165283203, |
|
"learning_rate": 4.53929539295393e-05, |
|
"loss": 0.3587, |
|
"step": 600 |
|
}, |
|
{ |
|
"epoch": 3.8109756097560976, |
|
"grad_norm": 3.5818405151367188, |
|
"learning_rate": 4.496951219512196e-05, |
|
"loss": 0.3624, |
|
"step": 625 |
|
}, |
|
{ |
|
"epoch": 3.9634146341463414, |
|
"grad_norm": 3.399240493774414, |
|
"learning_rate": 4.454607046070461e-05, |
|
"loss": 0.3598, |
|
"step": 650 |
|
}, |
|
{ |
|
"epoch": 4.0, |
|
"eval_accuracy": 0.6442516268980477, |
|
"eval_loss": 0.9534724950790405, |
|
"eval_runtime": 40.9142, |
|
"eval_samples_per_second": 22.535, |
|
"eval_steps_per_second": 0.709, |
|
"step": 656 |
|
}, |
|
{ |
|
"epoch": 4.115853658536586, |
|
"grad_norm": 3.605801820755005, |
|
"learning_rate": 4.412262872628726e-05, |
|
"loss": 0.3402, |
|
"step": 675 |
|
}, |
|
{ |
|
"epoch": 4.2682926829268295, |
|
"grad_norm": 2.1427693367004395, |
|
"learning_rate": 4.369918699186992e-05, |
|
"loss": 0.3177, |
|
"step": 700 |
|
}, |
|
{ |
|
"epoch": 4.420731707317073, |
|
"grad_norm": 2.742725133895874, |
|
"learning_rate": 4.3275745257452574e-05, |
|
"loss": 0.3274, |
|
"step": 725 |
|
}, |
|
{ |
|
"epoch": 4.573170731707317, |
|
"grad_norm": 2.6349594593048096, |
|
"learning_rate": 4.285230352303523e-05, |
|
"loss": 0.3243, |
|
"step": 750 |
|
}, |
|
{ |
|
"epoch": 4.725609756097561, |
|
"grad_norm": 2.6220862865448, |
|
"learning_rate": 4.242886178861789e-05, |
|
"loss": 0.3133, |
|
"step": 775 |
|
}, |
|
{ |
|
"epoch": 4.878048780487805, |
|
"grad_norm": 2.377204656600952, |
|
"learning_rate": 4.2005420054200545e-05, |
|
"loss": 0.3284, |
|
"step": 800 |
|
}, |
|
{ |
|
"epoch": 5.0, |
|
"eval_accuracy": 0.7255965292841648, |
|
"eval_loss": 0.84259432554245, |
|
"eval_runtime": 41.4402, |
|
"eval_samples_per_second": 22.249, |
|
"eval_steps_per_second": 0.7, |
|
"step": 820 |
|
}, |
|
{ |
|
"epoch": 5.030487804878049, |
|
"grad_norm": 2.5129756927490234, |
|
"learning_rate": 4.15819783197832e-05, |
|
"loss": 0.3208, |
|
"step": 825 |
|
}, |
|
{ |
|
"epoch": 5.182926829268292, |
|
"grad_norm": 3.733022928237915, |
|
"learning_rate": 4.1158536585365856e-05, |
|
"loss": 0.2929, |
|
"step": 850 |
|
}, |
|
{ |
|
"epoch": 5.335365853658536, |
|
"grad_norm": 1.9925252199172974, |
|
"learning_rate": 4.0735094850948516e-05, |
|
"loss": 0.2878, |
|
"step": 875 |
|
}, |
|
{ |
|
"epoch": 5.487804878048781, |
|
"grad_norm": 1.9667726755142212, |
|
"learning_rate": 4.031165311653117e-05, |
|
"loss": 0.2964, |
|
"step": 900 |
|
}, |
|
{ |
|
"epoch": 5.640243902439025, |
|
"grad_norm": 3.29734206199646, |
|
"learning_rate": 3.988821138211382e-05, |
|
"loss": 0.2913, |
|
"step": 925 |
|
}, |
|
{ |
|
"epoch": 5.7926829268292686, |
|
"grad_norm": 1.7537786960601807, |
|
"learning_rate": 3.946476964769648e-05, |
|
"loss": 0.29, |
|
"step": 950 |
|
}, |
|
{ |
|
"epoch": 5.945121951219512, |
|
"grad_norm": 2.272477149963379, |
|
"learning_rate": 3.904132791327913e-05, |
|
"loss": 0.2933, |
|
"step": 975 |
|
}, |
|
{ |
|
"epoch": 6.0, |
|
"eval_accuracy": 0.7657266811279827, |
|
"eval_loss": 0.8268624544143677, |
|
"eval_runtime": 41.5751, |
|
"eval_samples_per_second": 22.177, |
|
"eval_steps_per_second": 0.698, |
|
"step": 984 |
|
}, |
|
{ |
|
"epoch": 6.097560975609756, |
|
"grad_norm": 2.4584200382232666, |
|
"learning_rate": 3.861788617886179e-05, |
|
"loss": 0.2829, |
|
"step": 1000 |
|
}, |
|
{ |
|
"epoch": 6.25, |
|
"grad_norm": 1.7212045192718506, |
|
"learning_rate": 3.8194444444444444e-05, |
|
"loss": 0.2738, |
|
"step": 1025 |
|
}, |
|
{ |
|
"epoch": 6.402439024390244, |
|
"grad_norm": 1.7124000787734985, |
|
"learning_rate": 3.77710027100271e-05, |
|
"loss": 0.2768, |
|
"step": 1050 |
|
}, |
|
{ |
|
"epoch": 6.554878048780488, |
|
"grad_norm": 1.7901511192321777, |
|
"learning_rate": 3.7347560975609755e-05, |
|
"loss": 0.2715, |
|
"step": 1075 |
|
}, |
|
{ |
|
"epoch": 6.7073170731707314, |
|
"grad_norm": 1.9946818351745605, |
|
"learning_rate": 3.6924119241192415e-05, |
|
"loss": 0.2751, |
|
"step": 1100 |
|
}, |
|
{ |
|
"epoch": 6.859756097560975, |
|
"grad_norm": 2.2430810928344727, |
|
"learning_rate": 3.6500677506775074e-05, |
|
"loss": 0.2734, |
|
"step": 1125 |
|
}, |
|
{ |
|
"epoch": 7.0, |
|
"eval_accuracy": 0.6952277657266811, |
|
"eval_loss": 0.8815469145774841, |
|
"eval_runtime": 41.5411, |
|
"eval_samples_per_second": 22.195, |
|
"eval_steps_per_second": 0.698, |
|
"step": 1148 |
|
}, |
|
{ |
|
"epoch": 7.012195121951219, |
|
"grad_norm": 2.6901626586914062, |
|
"learning_rate": 3.6077235772357726e-05, |
|
"loss": 0.269, |
|
"step": 1150 |
|
}, |
|
{ |
|
"epoch": 7.164634146341464, |
|
"grad_norm": 3.2748658657073975, |
|
"learning_rate": 3.565379403794038e-05, |
|
"loss": 0.2626, |
|
"step": 1175 |
|
}, |
|
{ |
|
"epoch": 7.317073170731708, |
|
"grad_norm": 2.778024435043335, |
|
"learning_rate": 3.523035230352303e-05, |
|
"loss": 0.256, |
|
"step": 1200 |
|
}, |
|
{ |
|
"epoch": 7.469512195121951, |
|
"grad_norm": 1.585498571395874, |
|
"learning_rate": 3.48069105691057e-05, |
|
"loss": 0.2555, |
|
"step": 1225 |
|
}, |
|
{ |
|
"epoch": 7.621951219512195, |
|
"grad_norm": 2.2903597354888916, |
|
"learning_rate": 3.438346883468835e-05, |
|
"loss": 0.2553, |
|
"step": 1250 |
|
}, |
|
{ |
|
"epoch": 7.774390243902439, |
|
"grad_norm": 2.6786723136901855, |
|
"learning_rate": 3.3960027100271e-05, |
|
"loss": 0.2654, |
|
"step": 1275 |
|
}, |
|
{ |
|
"epoch": 7.926829268292683, |
|
"grad_norm": 1.9107999801635742, |
|
"learning_rate": 3.353658536585366e-05, |
|
"loss": 0.2679, |
|
"step": 1300 |
|
}, |
|
{ |
|
"epoch": 8.0, |
|
"eval_accuracy": 0.7678958785249458, |
|
"eval_loss": 0.8078573346138, |
|
"eval_runtime": 41.0894, |
|
"eval_samples_per_second": 22.439, |
|
"eval_steps_per_second": 0.706, |
|
"step": 1312 |
|
}, |
|
{ |
|
"epoch": 8.079268292682928, |
|
"grad_norm": 1.494566798210144, |
|
"learning_rate": 3.3113143631436314e-05, |
|
"loss": 0.2518, |
|
"step": 1325 |
|
}, |
|
{ |
|
"epoch": 8.231707317073171, |
|
"grad_norm": 1.2574859857559204, |
|
"learning_rate": 3.268970189701897e-05, |
|
"loss": 0.2431, |
|
"step": 1350 |
|
}, |
|
{ |
|
"epoch": 8.384146341463415, |
|
"grad_norm": 1.92410409450531, |
|
"learning_rate": 3.226626016260163e-05, |
|
"loss": 0.257, |
|
"step": 1375 |
|
}, |
|
{ |
|
"epoch": 8.536585365853659, |
|
"grad_norm": 1.3554410934448242, |
|
"learning_rate": 3.1842818428184285e-05, |
|
"loss": 0.2503, |
|
"step": 1400 |
|
}, |
|
{ |
|
"epoch": 8.689024390243903, |
|
"grad_norm": 2.443579912185669, |
|
"learning_rate": 3.141937669376694e-05, |
|
"loss": 0.2534, |
|
"step": 1425 |
|
}, |
|
{ |
|
"epoch": 8.841463414634147, |
|
"grad_norm": 1.6061345338821411, |
|
"learning_rate": 3.0995934959349596e-05, |
|
"loss": 0.2483, |
|
"step": 1450 |
|
}, |
|
{ |
|
"epoch": 8.99390243902439, |
|
"grad_norm": 1.1513770818710327, |
|
"learning_rate": 3.0572493224932255e-05, |
|
"loss": 0.2574, |
|
"step": 1475 |
|
}, |
|
{ |
|
"epoch": 9.0, |
|
"eval_accuracy": 0.7863340563991323, |
|
"eval_loss": 0.7823293805122375, |
|
"eval_runtime": 41.8128, |
|
"eval_samples_per_second": 22.051, |
|
"eval_steps_per_second": 0.694, |
|
"step": 1476 |
|
}, |
|
{ |
|
"epoch": 9.146341463414634, |
|
"grad_norm": 1.9186493158340454, |
|
"learning_rate": 3.0149051490514908e-05, |
|
"loss": 0.247, |
|
"step": 1500 |
|
}, |
|
{ |
|
"epoch": 9.298780487804878, |
|
"grad_norm": 1.1330264806747437, |
|
"learning_rate": 2.972560975609756e-05, |
|
"loss": 0.2536, |
|
"step": 1525 |
|
}, |
|
{ |
|
"epoch": 9.451219512195122, |
|
"grad_norm": 1.8567355871200562, |
|
"learning_rate": 2.930216802168022e-05, |
|
"loss": 0.2458, |
|
"step": 1550 |
|
}, |
|
{ |
|
"epoch": 9.603658536585366, |
|
"grad_norm": 1.351461410522461, |
|
"learning_rate": 2.8878726287262875e-05, |
|
"loss": 0.2442, |
|
"step": 1575 |
|
}, |
|
{ |
|
"epoch": 9.75609756097561, |
|
"grad_norm": 2.0739259719848633, |
|
"learning_rate": 2.8455284552845528e-05, |
|
"loss": 0.25, |
|
"step": 1600 |
|
}, |
|
{ |
|
"epoch": 9.908536585365853, |
|
"grad_norm": 0.737319827079773, |
|
"learning_rate": 2.8031842818428184e-05, |
|
"loss": 0.2403, |
|
"step": 1625 |
|
}, |
|
{ |
|
"epoch": 10.0, |
|
"eval_accuracy": 0.7906724511930586, |
|
"eval_loss": 0.7832654118537903, |
|
"eval_runtime": 41.3078, |
|
"eval_samples_per_second": 22.32, |
|
"eval_steps_per_second": 0.702, |
|
"step": 1640 |
|
}, |
|
{ |
|
"epoch": 10.060975609756097, |
|
"grad_norm": 1.5469945669174194, |
|
"learning_rate": 2.7608401084010843e-05, |
|
"loss": 0.2422, |
|
"step": 1650 |
|
}, |
|
{ |
|
"epoch": 10.213414634146341, |
|
"grad_norm": 1.8835623264312744, |
|
"learning_rate": 2.71849593495935e-05, |
|
"loss": 0.2409, |
|
"step": 1675 |
|
}, |
|
{ |
|
"epoch": 10.365853658536585, |
|
"grad_norm": 0.858368456363678, |
|
"learning_rate": 2.676151761517615e-05, |
|
"loss": 0.2451, |
|
"step": 1700 |
|
}, |
|
{ |
|
"epoch": 10.518292682926829, |
|
"grad_norm": 1.7335875034332275, |
|
"learning_rate": 2.633807588075881e-05, |
|
"loss": 0.2423, |
|
"step": 1725 |
|
}, |
|
{ |
|
"epoch": 10.670731707317072, |
|
"grad_norm": 1.0270367860794067, |
|
"learning_rate": 2.5914634146341466e-05, |
|
"loss": 0.2398, |
|
"step": 1750 |
|
}, |
|
{ |
|
"epoch": 10.823170731707316, |
|
"grad_norm": 1.3634330034255981, |
|
"learning_rate": 2.549119241192412e-05, |
|
"loss": 0.2428, |
|
"step": 1775 |
|
}, |
|
{ |
|
"epoch": 10.975609756097562, |
|
"grad_norm": 0.8540765047073364, |
|
"learning_rate": 2.5067750677506774e-05, |
|
"loss": 0.2376, |
|
"step": 1800 |
|
}, |
|
{ |
|
"epoch": 11.0, |
|
"eval_accuracy": 0.7852494577006508, |
|
"eval_loss": 0.7851491570472717, |
|
"eval_runtime": 41.4503, |
|
"eval_samples_per_second": 22.243, |
|
"eval_steps_per_second": 0.7, |
|
"step": 1804 |
|
}, |
|
{ |
|
"epoch": 11.128048780487806, |
|
"grad_norm": 2.378941059112549, |
|
"learning_rate": 2.464430894308943e-05, |
|
"loss": 0.2435, |
|
"step": 1825 |
|
}, |
|
{ |
|
"epoch": 11.28048780487805, |
|
"grad_norm": 1.6375097036361694, |
|
"learning_rate": 2.422086720867209e-05, |
|
"loss": 0.2358, |
|
"step": 1850 |
|
}, |
|
{ |
|
"epoch": 11.432926829268293, |
|
"grad_norm": 1.5316449403762817, |
|
"learning_rate": 2.3797425474254745e-05, |
|
"loss": 0.2367, |
|
"step": 1875 |
|
}, |
|
{ |
|
"epoch": 11.585365853658537, |
|
"grad_norm": 1.458227276802063, |
|
"learning_rate": 2.3373983739837398e-05, |
|
"loss": 0.237, |
|
"step": 1900 |
|
}, |
|
{ |
|
"epoch": 11.737804878048781, |
|
"grad_norm": 1.1965981721878052, |
|
"learning_rate": 2.2950542005420057e-05, |
|
"loss": 0.2451, |
|
"step": 1925 |
|
}, |
|
{ |
|
"epoch": 11.890243902439025, |
|
"grad_norm": 1.0848238468170166, |
|
"learning_rate": 2.252710027100271e-05, |
|
"loss": 0.2399, |
|
"step": 1950 |
|
}, |
|
{ |
|
"epoch": 12.0, |
|
"eval_accuracy": 0.7939262472885033, |
|
"eval_loss": 0.7965609431266785, |
|
"eval_runtime": 41.2064, |
|
"eval_samples_per_second": 22.375, |
|
"eval_steps_per_second": 0.704, |
|
"step": 1968 |
|
}, |
|
{ |
|
"epoch": 12.042682926829269, |
|
"grad_norm": 0.900860071182251, |
|
"learning_rate": 2.210365853658537e-05, |
|
"loss": 0.2385, |
|
"step": 1975 |
|
}, |
|
{ |
|
"epoch": 12.195121951219512, |
|
"grad_norm": 0.6957654356956482, |
|
"learning_rate": 2.1680216802168024e-05, |
|
"loss": 0.235, |
|
"step": 2000 |
|
}, |
|
{ |
|
"epoch": 12.347560975609756, |
|
"grad_norm": 0.9133132100105286, |
|
"learning_rate": 2.1256775067750677e-05, |
|
"loss": 0.2397, |
|
"step": 2025 |
|
}, |
|
{ |
|
"epoch": 12.5, |
|
"grad_norm": 1.3867614269256592, |
|
"learning_rate": 2.0833333333333336e-05, |
|
"loss": 0.2356, |
|
"step": 2050 |
|
}, |
|
{ |
|
"epoch": 12.652439024390244, |
|
"grad_norm": 1.484837532043457, |
|
"learning_rate": 2.040989159891599e-05, |
|
"loss": 0.2359, |
|
"step": 2075 |
|
}, |
|
{ |
|
"epoch": 12.804878048780488, |
|
"grad_norm": 1.6344926357269287, |
|
"learning_rate": 1.9986449864498648e-05, |
|
"loss": 0.2322, |
|
"step": 2100 |
|
}, |
|
{ |
|
"epoch": 12.957317073170731, |
|
"grad_norm": 0.8600764870643616, |
|
"learning_rate": 1.95630081300813e-05, |
|
"loss": 0.231, |
|
"step": 2125 |
|
}, |
|
{ |
|
"epoch": 13.0, |
|
"eval_accuracy": 0.7765726681127982, |
|
"eval_loss": 0.7956083416938782, |
|
"eval_runtime": 41.0434, |
|
"eval_samples_per_second": 22.464, |
|
"eval_steps_per_second": 0.707, |
|
"step": 2132 |
|
}, |
|
{ |
|
"epoch": 13.109756097560975, |
|
"grad_norm": 1.2406657934188843, |
|
"learning_rate": 1.9139566395663956e-05, |
|
"loss": 0.2327, |
|
"step": 2150 |
|
}, |
|
{ |
|
"epoch": 13.262195121951219, |
|
"grad_norm": 0.9290794134140015, |
|
"learning_rate": 1.8716124661246615e-05, |
|
"loss": 0.238, |
|
"step": 2175 |
|
}, |
|
{ |
|
"epoch": 13.414634146341463, |
|
"grad_norm": 0.9094896912574768, |
|
"learning_rate": 1.8292682926829268e-05, |
|
"loss": 0.2332, |
|
"step": 2200 |
|
}, |
|
{ |
|
"epoch": 13.567073170731707, |
|
"grad_norm": 0.9954293370246887, |
|
"learning_rate": 1.7869241192411927e-05, |
|
"loss": 0.2273, |
|
"step": 2225 |
|
}, |
|
{ |
|
"epoch": 13.71951219512195, |
|
"grad_norm": 0.9158800840377808, |
|
"learning_rate": 1.744579945799458e-05, |
|
"loss": 0.2294, |
|
"step": 2250 |
|
}, |
|
{ |
|
"epoch": 13.871951219512194, |
|
"grad_norm": 0.9772486686706543, |
|
"learning_rate": 1.702235772357724e-05, |
|
"loss": 0.2351, |
|
"step": 2275 |
|
}, |
|
{ |
|
"epoch": 14.0, |
|
"eval_accuracy": 0.7917570498915402, |
|
"eval_loss": 0.7792609930038452, |
|
"eval_runtime": 41.3252, |
|
"eval_samples_per_second": 22.311, |
|
"eval_steps_per_second": 0.702, |
|
"step": 2296 |
|
}, |
|
{ |
|
"epoch": 14.024390243902438, |
|
"grad_norm": 0.6987525820732117, |
|
"learning_rate": 1.659891598915989e-05, |
|
"loss": 0.2319, |
|
"step": 2300 |
|
}, |
|
{ |
|
"epoch": 14.176829268292684, |
|
"grad_norm": 1.195956826210022, |
|
"learning_rate": 1.6175474254742547e-05, |
|
"loss": 0.2333, |
|
"step": 2325 |
|
}, |
|
{ |
|
"epoch": 14.329268292682928, |
|
"grad_norm": 1.15598464012146, |
|
"learning_rate": 1.5752032520325206e-05, |
|
"loss": 0.2314, |
|
"step": 2350 |
|
}, |
|
{ |
|
"epoch": 14.481707317073171, |
|
"grad_norm": 0.5450630784034729, |
|
"learning_rate": 1.532859078590786e-05, |
|
"loss": 0.2317, |
|
"step": 2375 |
|
}, |
|
{ |
|
"epoch": 14.634146341463415, |
|
"grad_norm": 1.0192309617996216, |
|
"learning_rate": 1.4905149051490516e-05, |
|
"loss": 0.2286, |
|
"step": 2400 |
|
}, |
|
{ |
|
"epoch": 14.786585365853659, |
|
"grad_norm": 0.8558110594749451, |
|
"learning_rate": 1.448170731707317e-05, |
|
"loss": 0.2285, |
|
"step": 2425 |
|
}, |
|
{ |
|
"epoch": 14.939024390243903, |
|
"grad_norm": 0.7634503245353699, |
|
"learning_rate": 1.4058265582655828e-05, |
|
"loss": 0.2344, |
|
"step": 2450 |
|
}, |
|
{ |
|
"epoch": 15.0, |
|
"eval_accuracy": 0.7960954446854663, |
|
"eval_loss": 0.7699832320213318, |
|
"eval_runtime": 40.7623, |
|
"eval_samples_per_second": 22.619, |
|
"eval_steps_per_second": 0.711, |
|
"step": 2460 |
|
}, |
|
{ |
|
"epoch": 15.091463414634147, |
|
"grad_norm": 1.0911827087402344, |
|
"learning_rate": 1.3634823848238485e-05, |
|
"loss": 0.2216, |
|
"step": 2475 |
|
}, |
|
{ |
|
"epoch": 15.24390243902439, |
|
"grad_norm": 0.665216863155365, |
|
"learning_rate": 1.321138211382114e-05, |
|
"loss": 0.2384, |
|
"step": 2500 |
|
}, |
|
{ |
|
"epoch": 15.396341463414634, |
|
"grad_norm": 0.8301494717597961, |
|
"learning_rate": 1.2787940379403795e-05, |
|
"loss": 0.2296, |
|
"step": 2525 |
|
}, |
|
{ |
|
"epoch": 15.548780487804878, |
|
"grad_norm": 0.9643315076828003, |
|
"learning_rate": 1.2364498644986451e-05, |
|
"loss": 0.2277, |
|
"step": 2550 |
|
}, |
|
{ |
|
"epoch": 15.701219512195122, |
|
"grad_norm": 0.8478797674179077, |
|
"learning_rate": 1.1941056910569107e-05, |
|
"loss": 0.2275, |
|
"step": 2575 |
|
}, |
|
{ |
|
"epoch": 15.853658536585366, |
|
"grad_norm": 0.6218520998954773, |
|
"learning_rate": 1.1517615176151763e-05, |
|
"loss": 0.232, |
|
"step": 2600 |
|
}, |
|
{ |
|
"epoch": 16.0, |
|
"eval_accuracy": 0.7906724511930586, |
|
"eval_loss": 0.7844656705856323, |
|
"eval_runtime": 40.6353, |
|
"eval_samples_per_second": 22.69, |
|
"eval_steps_per_second": 0.714, |
|
"step": 2624 |
|
}, |
|
{ |
|
"epoch": 16.00609756097561, |
|
"grad_norm": 1.0007075071334839, |
|
"learning_rate": 1.1094173441734418e-05, |
|
"loss": 0.2316, |
|
"step": 2625 |
|
}, |
|
{ |
|
"epoch": 16.158536585365855, |
|
"grad_norm": 0.5642134547233582, |
|
"learning_rate": 1.0670731707317074e-05, |
|
"loss": 0.2275, |
|
"step": 2650 |
|
}, |
|
{ |
|
"epoch": 16.3109756097561, |
|
"grad_norm": 0.5428956151008606, |
|
"learning_rate": 1.0247289972899728e-05, |
|
"loss": 0.2265, |
|
"step": 2675 |
|
}, |
|
{ |
|
"epoch": 16.463414634146343, |
|
"grad_norm": 0.49864181876182556, |
|
"learning_rate": 9.823848238482384e-06, |
|
"loss": 0.2246, |
|
"step": 2700 |
|
}, |
|
{ |
|
"epoch": 16.615853658536587, |
|
"grad_norm": 0.8356211185455322, |
|
"learning_rate": 9.400406504065042e-06, |
|
"loss": 0.2287, |
|
"step": 2725 |
|
}, |
|
{ |
|
"epoch": 16.76829268292683, |
|
"grad_norm": 0.8411365151405334, |
|
"learning_rate": 8.976964769647697e-06, |
|
"loss": 0.2235, |
|
"step": 2750 |
|
}, |
|
{ |
|
"epoch": 16.920731707317074, |
|
"grad_norm": 0.6885209083557129, |
|
"learning_rate": 8.553523035230353e-06, |
|
"loss": 0.2318, |
|
"step": 2775 |
|
}, |
|
{ |
|
"epoch": 17.0, |
|
"eval_accuracy": 0.7917570498915402, |
|
"eval_loss": 0.7930197715759277, |
|
"eval_runtime": 41.1598, |
|
"eval_samples_per_second": 22.401, |
|
"eval_steps_per_second": 0.705, |
|
"step": 2788 |
|
}, |
|
{ |
|
"epoch": 17.073170731707318, |
|
"grad_norm": 0.5789854526519775, |
|
"learning_rate": 8.130081300813009e-06, |
|
"loss": 0.2305, |
|
"step": 2800 |
|
}, |
|
{ |
|
"epoch": 17.225609756097562, |
|
"grad_norm": 0.8740878105163574, |
|
"learning_rate": 7.706639566395663e-06, |
|
"loss": 0.2261, |
|
"step": 2825 |
|
}, |
|
{ |
|
"epoch": 17.378048780487806, |
|
"grad_norm": 0.7720724940299988, |
|
"learning_rate": 7.283197831978319e-06, |
|
"loss": 0.2333, |
|
"step": 2850 |
|
}, |
|
{ |
|
"epoch": 17.53048780487805, |
|
"grad_norm": 0.8361527919769287, |
|
"learning_rate": 6.859756097560977e-06, |
|
"loss": 0.2336, |
|
"step": 2875 |
|
}, |
|
{ |
|
"epoch": 17.682926829268293, |
|
"grad_norm": 0.8507777452468872, |
|
"learning_rate": 6.4363143631436324e-06, |
|
"loss": 0.2257, |
|
"step": 2900 |
|
}, |
|
{ |
|
"epoch": 17.835365853658537, |
|
"grad_norm": 1.3327784538269043, |
|
"learning_rate": 6.0128726287262874e-06, |
|
"loss": 0.2223, |
|
"step": 2925 |
|
}, |
|
{ |
|
"epoch": 17.98780487804878, |
|
"grad_norm": 1.0190765857696533, |
|
"learning_rate": 5.589430894308943e-06, |
|
"loss": 0.2241, |
|
"step": 2950 |
|
}, |
|
{ |
|
"epoch": 18.0, |
|
"eval_accuracy": 0.7885032537960954, |
|
"eval_loss": 0.7814083695411682, |
|
"eval_runtime": 40.5372, |
|
"eval_samples_per_second": 22.745, |
|
"eval_steps_per_second": 0.715, |
|
"step": 2952 |
|
}, |
|
{ |
|
"epoch": 18.140243902439025, |
|
"grad_norm": 0.5855722427368164, |
|
"learning_rate": 5.165989159891599e-06, |
|
"loss": 0.2259, |
|
"step": 2975 |
|
}, |
|
{ |
|
"epoch": 18.29268292682927, |
|
"grad_norm": 0.8351572751998901, |
|
"learning_rate": 4.742547425474255e-06, |
|
"loss": 0.2256, |
|
"step": 3000 |
|
}, |
|
{ |
|
"epoch": 18.445121951219512, |
|
"grad_norm": 0.5234180688858032, |
|
"learning_rate": 4.319105691056911e-06, |
|
"loss": 0.2229, |
|
"step": 3025 |
|
}, |
|
{ |
|
"epoch": 18.597560975609756, |
|
"grad_norm": 0.6411917209625244, |
|
"learning_rate": 3.8956639566395666e-06, |
|
"loss": 0.2278, |
|
"step": 3050 |
|
}, |
|
{ |
|
"epoch": 18.75, |
|
"grad_norm": 0.6386851668357849, |
|
"learning_rate": 3.4722222222222224e-06, |
|
"loss": 0.2238, |
|
"step": 3075 |
|
}, |
|
{ |
|
"epoch": 18.902439024390244, |
|
"grad_norm": 0.6463262438774109, |
|
"learning_rate": 3.0487804878048782e-06, |
|
"loss": 0.2291, |
|
"step": 3100 |
|
}, |
|
{ |
|
"epoch": 19.0, |
|
"eval_accuracy": 0.7819956616052061, |
|
"eval_loss": 0.7901421785354614, |
|
"eval_runtime": 41.6462, |
|
"eval_samples_per_second": 22.139, |
|
"eval_steps_per_second": 0.696, |
|
"step": 3116 |
|
}, |
|
{ |
|
"epoch": 19.054878048780488, |
|
"grad_norm": 0.45505502820014954, |
|
"learning_rate": 2.625338753387534e-06, |
|
"loss": 0.2267, |
|
"step": 3125 |
|
}, |
|
{ |
|
"epoch": 19.20731707317073, |
|
"grad_norm": 0.4252510964870453, |
|
"learning_rate": 2.2018970189701894e-06, |
|
"loss": 0.2326, |
|
"step": 3150 |
|
}, |
|
{ |
|
"epoch": 19.359756097560975, |
|
"grad_norm": 0.7417110204696655, |
|
"learning_rate": 1.7784552845528455e-06, |
|
"loss": 0.231, |
|
"step": 3175 |
|
}, |
|
{ |
|
"epoch": 19.51219512195122, |
|
"grad_norm": 0.6761830449104309, |
|
"learning_rate": 1.3550135501355015e-06, |
|
"loss": 0.2263, |
|
"step": 3200 |
|
}, |
|
{ |
|
"epoch": 19.664634146341463, |
|
"grad_norm": 0.6491538286209106, |
|
"learning_rate": 9.315718157181572e-07, |
|
"loss": 0.2268, |
|
"step": 3225 |
|
}, |
|
{ |
|
"epoch": 19.817073170731707, |
|
"grad_norm": 0.6726053357124329, |
|
"learning_rate": 5.081300813008131e-07, |
|
"loss": 0.2231, |
|
"step": 3250 |
|
}, |
|
{ |
|
"epoch": 19.96951219512195, |
|
"grad_norm": 0.7665335536003113, |
|
"learning_rate": 8.468834688346885e-08, |
|
"loss": 0.2227, |
|
"step": 3275 |
|
}, |
|
{ |
|
"epoch": 20.0, |
|
"eval_accuracy": 0.7885032537960954, |
|
"eval_loss": 0.7853749990463257, |
|
"eval_runtime": 41.5331, |
|
"eval_samples_per_second": 22.199, |
|
"eval_steps_per_second": 0.698, |
|
"step": 3280 |
|
}, |
|
{ |
|
"epoch": 20.0, |
|
"step": 3280, |
|
"total_flos": 0.0, |
|
"train_loss": 0.29519598901998706, |
|
"train_runtime": 14634.1552, |
|
"train_samples_per_second": 7.139, |
|
"train_steps_per_second": 0.224 |
|
} |
|
], |
|
"logging_steps": 25, |
|
"max_steps": 3280, |
|
"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": 32, |
|
"trial_name": null, |
|
"trial_params": null |
|
} |
|
|