| { | |
| "best_metric": 0.5063937306404114, | |
| "best_model_checkpoint": "../../experiments_checkpoints/MAdAiLab/Qwen/Qwen1.5_1.8B_ledgar/checkpoint-1800", | |
| "epoch": 1.9189765458422174, | |
| "eval_steps": 100, | |
| "global_step": 1800, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.03, | |
| "grad_norm": 80.4836196899414, | |
| "learning_rate": 4.9555792466240235e-06, | |
| "loss": 7.9887, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.05, | |
| "grad_norm": 60.927364349365234, | |
| "learning_rate": 4.911158493248046e-06, | |
| "loss": 3.1908, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "grad_norm": 49.68091583251953, | |
| "learning_rate": 4.866737739872069e-06, | |
| "loss": 1.7183, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.11, | |
| "grad_norm": 55.176666259765625, | |
| "learning_rate": 4.822316986496091e-06, | |
| "loss": 1.3077, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.11, | |
| "eval_accuracy": 0.7277, | |
| "eval_f1_macro": 0.5770831474844406, | |
| "eval_f1_micro": 0.7277, | |
| "eval_loss": 1.0944937467575073, | |
| "eval_runtime": 25.447, | |
| "eval_samples_per_second": 392.973, | |
| "eval_steps_per_second": 6.17, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "grad_norm": 46.64506530761719, | |
| "learning_rate": 4.777896233120114e-06, | |
| "loss": 1.1393, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "grad_norm": 41.11891174316406, | |
| "learning_rate": 4.733475479744136e-06, | |
| "loss": 1.0243, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.19, | |
| "grad_norm": 34.20009994506836, | |
| "learning_rate": 4.6890547263681595e-06, | |
| "loss": 0.9005, | |
| "step": 175 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "grad_norm": 38.80377197265625, | |
| "learning_rate": 4.644633972992183e-06, | |
| "loss": 0.8627, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "eval_accuracy": 0.7907, | |
| "eval_f1_macro": 0.6657039157603262, | |
| "eval_f1_micro": 0.7907, | |
| "eval_loss": 0.8368468880653381, | |
| "eval_runtime": 25.9484, | |
| "eval_samples_per_second": 385.38, | |
| "eval_steps_per_second": 6.05, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.24, | |
| "grad_norm": 34.973506927490234, | |
| "learning_rate": 4.600213219616206e-06, | |
| "loss": 0.7896, | |
| "step": 225 | |
| }, | |
| { | |
| "epoch": 0.27, | |
| "grad_norm": 29.98388671875, | |
| "learning_rate": 4.555792466240228e-06, | |
| "loss": 0.8307, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "grad_norm": 27.09973907470703, | |
| "learning_rate": 4.51137171286425e-06, | |
| "loss": 0.7846, | |
| "step": 275 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "grad_norm": 36.151161193847656, | |
| "learning_rate": 4.466950959488273e-06, | |
| "loss": 0.7179, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "eval_accuracy": 0.7971, | |
| "eval_f1_macro": 0.6861778340669753, | |
| "eval_f1_micro": 0.7971, | |
| "eval_loss": 0.7824062705039978, | |
| "eval_runtime": 25.9003, | |
| "eval_samples_per_second": 386.095, | |
| "eval_steps_per_second": 6.062, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.35, | |
| "grad_norm": 33.09822463989258, | |
| "learning_rate": 4.422530206112296e-06, | |
| "loss": 0.7133, | |
| "step": 325 | |
| }, | |
| { | |
| "epoch": 0.37, | |
| "grad_norm": 35.52923583984375, | |
| "learning_rate": 4.378109452736319e-06, | |
| "loss": 0.7354, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "grad_norm": 39.79545211791992, | |
| "learning_rate": 4.333688699360342e-06, | |
| "loss": 0.6619, | |
| "step": 375 | |
| }, | |
| { | |
| "epoch": 0.43, | |
| "grad_norm": 30.13161849975586, | |
| "learning_rate": 4.289267945984365e-06, | |
| "loss": 0.6961, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.43, | |
| "eval_accuracy": 0.8138, | |
| "eval_f1_macro": 0.6992465625213966, | |
| "eval_f1_micro": 0.8138, | |
| "eval_loss": 0.6951531171798706, | |
| "eval_runtime": 25.6082, | |
| "eval_samples_per_second": 390.5, | |
| "eval_steps_per_second": 6.131, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.45, | |
| "grad_norm": 27.575519561767578, | |
| "learning_rate": 4.244847192608387e-06, | |
| "loss": 0.7162, | |
| "step": 425 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "grad_norm": 35.084754943847656, | |
| "learning_rate": 4.200426439232409e-06, | |
| "loss": 0.7722, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.51, | |
| "grad_norm": 28.47511863708496, | |
| "learning_rate": 4.156005685856432e-06, | |
| "loss": 0.6866, | |
| "step": 475 | |
| }, | |
| { | |
| "epoch": 0.53, | |
| "grad_norm": 32.34709548950195, | |
| "learning_rate": 4.1115849324804554e-06, | |
| "loss": 0.745, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.53, | |
| "eval_accuracy": 0.8121, | |
| "eval_f1_macro": 0.7033560293953169, | |
| "eval_f1_micro": 0.8121, | |
| "eval_loss": 0.6718780994415283, | |
| "eval_runtime": 25.9161, | |
| "eval_samples_per_second": 385.86, | |
| "eval_steps_per_second": 6.058, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.56, | |
| "grad_norm": 25.5845890045166, | |
| "learning_rate": 4.067164179104478e-06, | |
| "loss": 0.6535, | |
| "step": 525 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "grad_norm": 22.466503143310547, | |
| "learning_rate": 4.022743425728501e-06, | |
| "loss": 0.5969, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 0.61, | |
| "grad_norm": 27.53134536743164, | |
| "learning_rate": 3.978322672352524e-06, | |
| "loss": 0.5926, | |
| "step": 575 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "grad_norm": 31.356454849243164, | |
| "learning_rate": 3.933901918976546e-06, | |
| "loss": 0.6505, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "eval_accuracy": 0.834, | |
| "eval_f1_macro": 0.7469091035082649, | |
| "eval_f1_micro": 0.834, | |
| "eval_loss": 0.6219750046730042, | |
| "eval_runtime": 25.9316, | |
| "eval_samples_per_second": 385.63, | |
| "eval_steps_per_second": 6.054, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "grad_norm": 37.17654800415039, | |
| "learning_rate": 3.889481165600569e-06, | |
| "loss": 0.6171, | |
| "step": 625 | |
| }, | |
| { | |
| "epoch": 0.69, | |
| "grad_norm": 26.71038055419922, | |
| "learning_rate": 3.8450604122245914e-06, | |
| "loss": 0.6218, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 0.72, | |
| "grad_norm": 27.787952423095703, | |
| "learning_rate": 3.8006396588486145e-06, | |
| "loss": 0.6124, | |
| "step": 675 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "grad_norm": 30.405912399291992, | |
| "learning_rate": 3.756218905472637e-06, | |
| "loss": 0.5914, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "eval_accuracy": 0.8362, | |
| "eval_f1_macro": 0.7410957777496914, | |
| "eval_f1_micro": 0.8362, | |
| "eval_loss": 0.6109625101089478, | |
| "eval_runtime": 25.6247, | |
| "eval_samples_per_second": 390.248, | |
| "eval_steps_per_second": 6.127, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.77, | |
| "grad_norm": 30.52012062072754, | |
| "learning_rate": 3.71179815209666e-06, | |
| "loss": 0.5711, | |
| "step": 725 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "grad_norm": 30.88004493713379, | |
| "learning_rate": 3.667377398720683e-06, | |
| "loss": 0.6695, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 0.83, | |
| "grad_norm": 22.504459381103516, | |
| "learning_rate": 3.622956645344705e-06, | |
| "loss": 0.5731, | |
| "step": 775 | |
| }, | |
| { | |
| "epoch": 0.85, | |
| "grad_norm": 21.515512466430664, | |
| "learning_rate": 3.578535891968728e-06, | |
| "loss": 0.5837, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.85, | |
| "eval_accuracy": 0.8385, | |
| "eval_f1_macro": 0.7413235492734335, | |
| "eval_f1_micro": 0.8385, | |
| "eval_loss": 0.5766780972480774, | |
| "eval_runtime": 25.6608, | |
| "eval_samples_per_second": 389.7, | |
| "eval_steps_per_second": 6.118, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "grad_norm": 30.9660587310791, | |
| "learning_rate": 3.534115138592751e-06, | |
| "loss": 0.6085, | |
| "step": 825 | |
| }, | |
| { | |
| "epoch": 0.91, | |
| "grad_norm": 18.883647918701172, | |
| "learning_rate": 3.4896943852167736e-06, | |
| "loss": 0.5121, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 0.93, | |
| "grad_norm": 24.548561096191406, | |
| "learning_rate": 3.4452736318407963e-06, | |
| "loss": 0.5621, | |
| "step": 875 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "grad_norm": 29.833791732788086, | |
| "learning_rate": 3.4008528784648194e-06, | |
| "loss": 0.5218, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "eval_accuracy": 0.849, | |
| "eval_f1_macro": 0.7702797685808792, | |
| "eval_f1_micro": 0.849, | |
| "eval_loss": 0.5365203022956848, | |
| "eval_runtime": 25.9091, | |
| "eval_samples_per_second": 385.964, | |
| "eval_steps_per_second": 6.06, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.99, | |
| "grad_norm": 27.948928833007812, | |
| "learning_rate": 3.3564321250888416e-06, | |
| "loss": 0.5681, | |
| "step": 925 | |
| }, | |
| { | |
| "epoch": 1.01, | |
| "grad_norm": 19.800880432128906, | |
| "learning_rate": 3.3120113717128643e-06, | |
| "loss": 0.4014, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 1.04, | |
| "grad_norm": 19.333465576171875, | |
| "learning_rate": 3.2675906183368874e-06, | |
| "loss": 0.2795, | |
| "step": 975 | |
| }, | |
| { | |
| "epoch": 1.07, | |
| "grad_norm": 22.315195083618164, | |
| "learning_rate": 3.22316986496091e-06, | |
| "loss": 0.2632, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.07, | |
| "eval_accuracy": 0.8562, | |
| "eval_f1_macro": 0.7683569808757446, | |
| "eval_f1_micro": 0.8562, | |
| "eval_loss": 0.5503664016723633, | |
| "eval_runtime": 25.5198, | |
| "eval_samples_per_second": 391.852, | |
| "eval_steps_per_second": 6.152, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.09, | |
| "grad_norm": 24.819501876831055, | |
| "learning_rate": 3.1787491115849327e-06, | |
| "loss": 0.2532, | |
| "step": 1025 | |
| }, | |
| { | |
| "epoch": 1.12, | |
| "grad_norm": 21.534936904907227, | |
| "learning_rate": 3.1343283582089558e-06, | |
| "loss": 0.2311, | |
| "step": 1050 | |
| }, | |
| { | |
| "epoch": 1.15, | |
| "grad_norm": 24.088809967041016, | |
| "learning_rate": 3.0899076048329785e-06, | |
| "loss": 0.3134, | |
| "step": 1075 | |
| }, | |
| { | |
| "epoch": 1.17, | |
| "grad_norm": 27.605493545532227, | |
| "learning_rate": 3.0454868514570007e-06, | |
| "loss": 0.2607, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 1.17, | |
| "eval_accuracy": 0.8525, | |
| "eval_f1_macro": 0.7656891626030512, | |
| "eval_f1_micro": 0.8525, | |
| "eval_loss": 0.5496523380279541, | |
| "eval_runtime": 25.7081, | |
| "eval_samples_per_second": 388.982, | |
| "eval_steps_per_second": 6.107, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 1.2, | |
| "grad_norm": 22.955158233642578, | |
| "learning_rate": 3.0010660980810234e-06, | |
| "loss": 0.2674, | |
| "step": 1125 | |
| }, | |
| { | |
| "epoch": 1.23, | |
| "grad_norm": 19.089893341064453, | |
| "learning_rate": 2.9566453447050464e-06, | |
| "loss": 0.2074, | |
| "step": 1150 | |
| }, | |
| { | |
| "epoch": 1.25, | |
| "grad_norm": 19.285688400268555, | |
| "learning_rate": 2.912224591329069e-06, | |
| "loss": 0.2488, | |
| "step": 1175 | |
| }, | |
| { | |
| "epoch": 1.28, | |
| "grad_norm": 23.45233726501465, | |
| "learning_rate": 2.867803837953092e-06, | |
| "loss": 0.274, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 1.28, | |
| "eval_accuracy": 0.8584, | |
| "eval_f1_macro": 0.7746299057445165, | |
| "eval_f1_micro": 0.8584, | |
| "eval_loss": 0.5439000129699707, | |
| "eval_runtime": 25.9014, | |
| "eval_samples_per_second": 386.079, | |
| "eval_steps_per_second": 6.061, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 1.31, | |
| "grad_norm": 31.231454849243164, | |
| "learning_rate": 2.823383084577115e-06, | |
| "loss": 0.2624, | |
| "step": 1225 | |
| }, | |
| { | |
| "epoch": 1.33, | |
| "grad_norm": 28.1010799407959, | |
| "learning_rate": 2.7789623312011375e-06, | |
| "loss": 0.2992, | |
| "step": 1250 | |
| }, | |
| { | |
| "epoch": 1.36, | |
| "grad_norm": 30.002384185791016, | |
| "learning_rate": 2.7345415778251598e-06, | |
| "loss": 0.2589, | |
| "step": 1275 | |
| }, | |
| { | |
| "epoch": 1.39, | |
| "grad_norm": 23.61323356628418, | |
| "learning_rate": 2.690120824449183e-06, | |
| "loss": 0.2216, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 1.39, | |
| "eval_accuracy": 0.8563, | |
| "eval_f1_macro": 0.7753520513346309, | |
| "eval_f1_micro": 0.8563, | |
| "eval_loss": 0.5687375068664551, | |
| "eval_runtime": 25.9424, | |
| "eval_samples_per_second": 385.47, | |
| "eval_steps_per_second": 6.052, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 1.41, | |
| "grad_norm": 27.56183433532715, | |
| "learning_rate": 2.6457000710732055e-06, | |
| "loss": 0.2845, | |
| "step": 1325 | |
| }, | |
| { | |
| "epoch": 1.44, | |
| "grad_norm": 18.88576316833496, | |
| "learning_rate": 2.601279317697228e-06, | |
| "loss": 0.2685, | |
| "step": 1350 | |
| }, | |
| { | |
| "epoch": 1.47, | |
| "grad_norm": 19.662220001220703, | |
| "learning_rate": 2.5568585643212513e-06, | |
| "loss": 0.2489, | |
| "step": 1375 | |
| }, | |
| { | |
| "epoch": 1.49, | |
| "grad_norm": 22.736656188964844, | |
| "learning_rate": 2.512437810945274e-06, | |
| "loss": 0.2044, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 1.49, | |
| "eval_accuracy": 0.861, | |
| "eval_f1_macro": 0.7820141563614671, | |
| "eval_f1_micro": 0.861, | |
| "eval_loss": 0.5385035276412964, | |
| "eval_runtime": 25.6666, | |
| "eval_samples_per_second": 389.612, | |
| "eval_steps_per_second": 6.117, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 1.52, | |
| "grad_norm": 24.569435119628906, | |
| "learning_rate": 2.4680170575692966e-06, | |
| "loss": 0.2388, | |
| "step": 1425 | |
| }, | |
| { | |
| "epoch": 1.55, | |
| "grad_norm": 17.50179100036621, | |
| "learning_rate": 2.4235963041933193e-06, | |
| "loss": 0.2556, | |
| "step": 1450 | |
| }, | |
| { | |
| "epoch": 1.57, | |
| "grad_norm": 15.387917518615723, | |
| "learning_rate": 2.379175550817342e-06, | |
| "loss": 0.2343, | |
| "step": 1475 | |
| }, | |
| { | |
| "epoch": 1.6, | |
| "grad_norm": 29.757495880126953, | |
| "learning_rate": 2.3347547974413646e-06, | |
| "loss": 0.2508, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 1.6, | |
| "eval_accuracy": 0.8577, | |
| "eval_f1_macro": 0.7710712973870113, | |
| "eval_f1_micro": 0.8577, | |
| "eval_loss": 0.5657808780670166, | |
| "eval_runtime": 25.9754, | |
| "eval_samples_per_second": 384.98, | |
| "eval_steps_per_second": 6.044, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 1.63, | |
| "grad_norm": 24.104217529296875, | |
| "learning_rate": 2.2903340440653877e-06, | |
| "loss": 0.2647, | |
| "step": 1525 | |
| }, | |
| { | |
| "epoch": 1.65, | |
| "grad_norm": 29.48048973083496, | |
| "learning_rate": 2.24591329068941e-06, | |
| "loss": 0.212, | |
| "step": 1550 | |
| }, | |
| { | |
| "epoch": 1.68, | |
| "grad_norm": 11.834880828857422, | |
| "learning_rate": 2.201492537313433e-06, | |
| "loss": 0.1939, | |
| "step": 1575 | |
| }, | |
| { | |
| "epoch": 1.71, | |
| "grad_norm": 24.24506378173828, | |
| "learning_rate": 2.1570717839374557e-06, | |
| "loss": 0.2513, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 1.71, | |
| "eval_accuracy": 0.8589, | |
| "eval_f1_macro": 0.7871987440671023, | |
| "eval_f1_micro": 0.8589, | |
| "eval_loss": 0.5366827845573425, | |
| "eval_runtime": 25.9643, | |
| "eval_samples_per_second": 385.144, | |
| "eval_steps_per_second": 6.047, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 1.73, | |
| "grad_norm": 23.33180046081543, | |
| "learning_rate": 2.112651030561479e-06, | |
| "loss": 0.2409, | |
| "step": 1625 | |
| }, | |
| { | |
| "epoch": 1.76, | |
| "grad_norm": 18.71114730834961, | |
| "learning_rate": 2.068230277185501e-06, | |
| "loss": 0.224, | |
| "step": 1650 | |
| }, | |
| { | |
| "epoch": 1.79, | |
| "grad_norm": 21.95819854736328, | |
| "learning_rate": 2.023809523809524e-06, | |
| "loss": 0.2223, | |
| "step": 1675 | |
| }, | |
| { | |
| "epoch": 1.81, | |
| "grad_norm": 27.065677642822266, | |
| "learning_rate": 1.979388770433547e-06, | |
| "loss": 0.2787, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 1.81, | |
| "eval_accuracy": 0.8653, | |
| "eval_f1_macro": 0.790261134849528, | |
| "eval_f1_micro": 0.8653, | |
| "eval_loss": 0.5133171677589417, | |
| "eval_runtime": 25.5701, | |
| "eval_samples_per_second": 391.081, | |
| "eval_steps_per_second": 6.14, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 1.84, | |
| "grad_norm": 35.288761138916016, | |
| "learning_rate": 1.9349680170575695e-06, | |
| "loss": 0.2709, | |
| "step": 1725 | |
| }, | |
| { | |
| "epoch": 1.87, | |
| "grad_norm": 21.077306747436523, | |
| "learning_rate": 1.8905472636815921e-06, | |
| "loss": 0.2002, | |
| "step": 1750 | |
| }, | |
| { | |
| "epoch": 1.89, | |
| "grad_norm": 25.394838333129883, | |
| "learning_rate": 1.846126510305615e-06, | |
| "loss": 0.2461, | |
| "step": 1775 | |
| }, | |
| { | |
| "epoch": 1.92, | |
| "grad_norm": 26.597759246826172, | |
| "learning_rate": 1.8017057569296375e-06, | |
| "loss": 0.2357, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 1.92, | |
| "eval_accuracy": 0.8669, | |
| "eval_f1_macro": 0.7902403947168268, | |
| "eval_f1_micro": 0.8669, | |
| "eval_loss": 0.5063937306404114, | |
| "eval_runtime": 25.6031, | |
| "eval_samples_per_second": 390.577, | |
| "eval_steps_per_second": 6.132, | |
| "step": 1800 | |
| } | |
| ], | |
| "logging_steps": 25, | |
| "max_steps": 2814, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "save_steps": 100, | |
| "total_flos": 1.074692042564567e+17, | |
| "train_batch_size": 32, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |